Clinical Review

From Temple University School of Medicine, Philadelphia, PA.

Abstract

• Objective: To outline the use and utility of gastric electric stimulation (GES) as a therapeutic intervention for gastroparesis.
• Methods: Review of the literature.
• Results: Gastroparesis is characterized by delayed gastric emptying, with symptoms of nausea, vomiting, early satiety, postprandial fullness, and abdominal pain. Some patients with gastroparesis do not respond to medical intervention, and for these patients surgical intervention may be warranted. GES utilizes high-frequency gastric neurostimulation to facilitate gastric emptying and reduce symptoms of gastroparesis. It is indicated for patients with idiopathic and diabetic gastroparesis who have nausea and vomiting as their primary symptoms and who have not responded to medical therapy. GES has also been used in postsurgical and pediatric gastroparesis patients. Optimizing the outcome of this surgical treatment through proper patient selection and meticulous surgical technique is essential as there are inherent risks to the procedure. Nonblinded studies of GES for medically refractory gastroparesis have demonstrated therapeutic symptomatic benefit, whereas randomized controlled trials have not. New interventions such as pyloromyotomy and pyloroplasty are reasonable alternatives or addendums to GES.
• Conclusion: GES may be considered among the therapies available for treating patients with refractory symptoms of gastroparesis. More studies, specifically those comparing GES, pyloromyotomy, GES combined with pyloromyotomy, and placebo, are needed to help guide therapy selection for refractory gastroparesis.

Keywords: diabetes; gastroparesis; dysmotility; gastric emptying; electric stimulation.

Gastroparesis is a chronic dysmotility disorder characterized by delayed gastric emptying with associated symptoms of nausea, vomiting, early satiety, postprandial fullness, and abdominal pain. Medical treatments for gastroparesis include dietary modifications, glucose control in those with diabetes, prokinetic medications, antiemetic medications, and symptom modulators, but unfortunately patients frequently do not respond to these treatments. In patients refractory to medical therapy, surgical treatments can be considered.

Gastric electric stimulation (GES; Enterra [Medtronic, Minneapolis, MN]) was approved via a Food and Drug Administration (FDA) Humanitarian Use Device (HUD) exemption for the treatment of medically refractory gastroparesis in 2000. Understanding the indications, risks, outcomes, and alternatives to GES is essential to providing appropriate care for patients with medically refractory gastroparesis. This article outlines the use and utility of GES as a therapeutic intervention for gastroparesis.

Types of Gastroparesis

Gastroparesis is a chronic symptomatic disorder of the stomach manifested by delayed gastric emptying without evidence of gastric outlet obstruction or ulceration.¹ The pathophysiology of gastroparesis appears to involve abnormalities in functioning of several elements including the autonomic nervous system, especially the vagus nerve, smooth muscle cells, enteric neurons, and interstitial cells of Cajal.

Idiopathic gastroparesis and diabetic gastroparesis are the 2 most common types of gastroparesis.² Symptomatic delayed gastric emptying with no primary underlying abnormality predisposing to gastroparesis is categorized as idiopathic gastroparesis.³ A small subset of patients with idiopathic gastroparesis report an initial infectious prodrome such as gastroenteritis or respiratory infection. It has been suggested that this postinfectious gastroparesis results from viral injury to the neural innervation of the stomach or the interstitial cells of Cajal in the stomach.⁴ Viruses that have been implicated in the development of gastroparesis include cytomegalovirus, Epstein-Barr virus, Norwalk virus, rotavirus, herpes zoster, and varicella zoster.^5-9

Diabetic gastroparesis is characterized as onset of symptoms of gastroparesis in patients with diabetes, with concomitant delayed gastric emptying. It is often attributed to chronic hyperglycemia-induced damage to the vagus nerve, and is frequently observed in association with other diabetic complications such as neuropathy, retinopathy, and nephropathy.¹⁰

Gastroparesis that develops following surgery is classified as postsurgical gastroparesis. In the past, this form of gastroparesis most commonly occurred after ulcer surgery, often performed with vagotomy. These types of surgeries are performed less frequently in the era of proton pump inhibitor therapy and treatments for Helicobacter pylori. Presently, Nissen fundoplication and bariatric surgery are the more common surgical procedures associated with gastroparesis.³ Long-term use of medications that delay gastric emptying, such as opiate narcotic medications, can lead to gastroparesis and represent another form of iatrogenic gastroparesis. Other forms of gastroparesis (atypical gastroparesis) arise due to various underlying etiologies, including neurological disorders (eg, Parkinson disease, multiple sclerosis), metabolic or endocrine conditions (eg, hypothyroidism), autoimmune disorders, connective tissue and collagen vascular disorders (eg, systemic lupus erythematosus, scleroderma, Sjögren syndrome, Ehlers-Danlos syndrome), or eating disorders (eg, anorexia, bulimia).³

Epidemiology

There is a female preponderance in patients with gastroparesis. Data from the Rochester Epidemiology Project, a database of linked medical records for residents of Olmsted County, MN, showed that the age-adjusted prevalence of definite gastroparesis per 100,000 inhabitants was 37.8 for women and 9.6 for men.¹¹ More recent estimates have suggested a much higher prevalence of probable gastroparesis (approximately 1.8%) in the general population using symptoms suggestive of gastroparesis.¹² Hospitalization rates for gastroparesis have increased since 2000, which could reflect rising prevalence and/or the effects of heightened awareness about and better identification of gastroparesis.¹³ This increase may also be due in part to the rising rate of diabetes leading to more cases of diabetic gastroparesis; withdrawal of some gastroparesis treatments from the market (cisapride, tegaserod) leading to hospitalizations for symptoms not adequately being treated; and hospitalizations needed for insertion of the gastric electric stimulator.

Gastroparesis Symptoms

The main symptoms of gastroparesis are early satiety, postprandial fullness, bloating, nausea, and vomiting.¹⁴ Nausea (> 90% of patients) and early satiety (60% of patients) are the most common symptoms.¹⁵ Abdominal pain is often present in patients with gastroparesis but is usually not the predominant symptom. The pain can be multifactorial, with somatic, visceral, and neuropathic components.^16-18 Moderate to severe abdominal pain has been found more often in patients with idiopathic gastroparesis and in association with opiate use.¹⁶ Symptoms of gastroparesis may be persistent or present as episodic flares. Due to the symptoms, some patients will experience weight loss and malnutrition and, in severe cases, dehydration.¹⁹

Although the definition of gastroparesis is a delay in gastric emptying along with symptoms, symptoms correlate poorly with the degree of delayed gastric emptying. The symptoms that appear to have the strongest correlation with gastric emptying are nausea, vomiting, early satiety, and postprandial fullness, whereas symptoms such as abdominal pain and bloating have little correlation. Furthermore, improving gastric emptying does not necessarily lead to improved symptoms, and symptom improvement does not always lead to improved gastric emptying times.²⁰ Between 5% and 12% of patients with diabetes report symptoms consistent with gastroparesis, though many of these patients have normal gastric emptying. The symptoms of gastroparesis overlap with those of functional dyspepsia, as both may have motor and sensory alterations.²¹

The Gastroparesis Cardinal Symptom Index (GCSI), a subset of the Patient Assessment of Gastrointestinal Disorders Symptom Severity Index (PAGI-SYM), is a questionnaire that is commonly used to establish symptom severity in patients with gastroparesis. It is comprised of 3 subscales—nausea and vomiting, postprandial fullness and early satiety, and bloating—which are averaged to provide a total GCSI score. Symptoms over the 2 weeks prior to administration of the questionnaire are assessed and rated from 0 (none) to 5 (very severe).²² Grading the severity of gastroparesis may take into account symptoms, quality of life, and gastric emptying. One commonly used grading system assigns a grade from 1 to 3, with grade 1 being mild gastroparesis, grade 2 being compensated gastroparesis, and grade 3 being gastric failure with refractory symptoms that are uncontrolled.^18,23 Quality-of-life surveys also suggest that gastroparesis independent of other factors leads to a worse quality of life.²⁴

Indications for GES

Gastric electric stimulator implantation is a surgical procedure with inherent risks and complications and is reserved for patients with intractable symptoms of gastroparesis who remain symptomatic despite treatment attempts with dietary management, antiemetic agents (eg, compazine, phenergan, and ondansetron), and prokinetic agents (eg, metoclopramide, erythromycin, and domperidone). Symptom modulators such as nortriptyline and mirtazapine are occasionally tried.

Surgical intervention can be considered upon failure of medical treatment measures. At least a year of documented care provided by a physician specializing in gastroparesis is suggested for surgical consideration. The gastric electric neurostimulator is approved by the FDA as a HUD for the care of patients with idiopathic and diabetic gastroparesis, performed on a compassionate basis. GES implantation requires Institutional Review Board approval at the institution, and patients are required to have documented delayed gastric emptying.

It is important to remember that the GES device is incompatible with magnetic resonance imaging (MRI) and explantation of the device is necessary prior to MRI. As such, in patients with anticipated need of frequent MRI, such as those with multiple sclerosis, serious consideration should be given to alternative strategies prior to focusing on this modality.

Device Placement

GES was devised to improve gastric emptying. The Enterra GES system uses high-frequency, low-energy electric stimulation. An alternative method is true gastric pacing that uses high-energy, low-frequency stimulation to entrain the gastric slow waves and subsequent contractions at 3 cycles per minute (cpm). Gastric pacing has greater energy requirements than GES, which makes the size of the stimulator too large to be practical. In pilot animal studies, GES produced an accelerating effect on gastric emptying, but in human studies GES had an inconsistent effect on gastric emptying. Studies have suggested that GES influences the proximal stomach, with a reduction of gastric tone,²⁵ and also that GES has an afferent modulatory mechanism.²⁶

The Enterra GES is placed surgically under general anesthesia, commonly via laparotomy or minimal access surgical techniques (laparoscopically or robotically assisted). Preoperative intravenous antibiotics are given. The system consists of a pair of electrodes connected to a pulse generator. The 2 stimulation leads are inserted into the gastric muscularis propria 1 cm apart along the greater curvature 10 cm proximal to the pylorus. Upper endoscopy is performed to ensure that the leads do not penetrate through the mucosa into the stomach lumen; if this occurs, repositioning of the lead is necessary. A horizontal incision through the skin is made, and the distal ends of the stimulating wires are tunneled through the abdominal wall and connected to the pulse generator. The impedance (resistance) between the wires is measured to ensure the appropriate range (200-800 Ohms). The neurostimulator with the distal ends of the stimulating wires is then placed into the subcutaneous pocket and sutured to the underlying fascia. The pulse generator delivers a high-frequency, low-energy, 0.1-second train of pulses at a frequency of 12 cpm. Within each pulse train, individual pulses oscillate at a frequency of 14 cycles per second. The voltage of the stimulations is set to provide a current of 5 milliamps (mA; remembering that voltage = current × resistance).

Patients are often hospitalized with a recovery time of 1 to 3 days. Immediate postoperative care usually includes intravenous fluids, controlling any postoperative ileus, advancing diet, and providing analgesic pain medications. Hospital length of stay can be impacted by surgical technique.²⁵ Patients are seen several weeks after discharge for assessment of the incision and toleration of diet. Medications for gastroparesis that patients were taking prior to the GES implantation are usually continued postoperatively, with a goal of reducing these medications over time. Patients are then followed every 3 to 12 months, depending on their clinical condition.

At follow-up visits, medications are reviewed and new treatments can be added if appropriate. The gastric stimulator is interrogated to determine if changes in resistance occurred; if necessary, minor readjustments can be made to keep the current at desired levels (5 mA). For persistent symptoms with GES treatment, the stimulator parameters can be adjusted after 3 months of follow up, typically first increasing the current from 5 to 7.5 mA and then to 10 mA. After this, the frequency can be increased from 14 Hz to 28 Hz, and then to 55 Hz. Rarely, the ON duration is increased from 0.1 to 1 second. Increasing the ON time can worsen symptoms in some patients, cause abdominal pain, and decrease the battery life from the usual 7 years.

Complications of GES

In an analysis of the Manufacturer and User Facility Device Experience (MAUDE) databank, Bielefeldt identified 1587 reports of adverse effects related to the gastric electric stimulator from January 2001 to October 2015.²⁷ The most common adverse effects are reviewed here.

Skin erosion/wound dehiscence is one of the most common reported complications; it may be related to superficial placement or inadequate securing of the device to the fascia. Abscess can develop postoperatively due to hematogenous seeding or may be a sign of lead erosion into the lumen, tracking along the leads into subcutaneous tissue.²⁸ It is important to warn patients to protect the area over the device from needle injections as this also can lead to hematoma formation and direct contamination of the device. If the device gets infected, it cannot be salvaged and requires explantation. Implantation of a new device can be attempted once all wound issues resolve.

Device migration/flipping most often occurs because the device is inadequately fixed to the underlying fascia, but occasionally it can occur from patients flipping the device around. Flipping can occur due to superficial pocket location within subcutaneous tissue, especially in obese patients. Migration/flipping can lead to prominence of the contour of the device and discomfort, ultimately requiring surgical correction.

Perforation and erosion of the leads. With time, leads can erode into the stomach, although this is rare. Usually erosion is associated with loss of device function. Endoscopy confirms this finding. In rare cases, infection can track proximally along the lead and present as a surgical site infection at the pulse generator. This complication often requires explantation of the neurostimulator leads and pulse generator.

Intestinal obstruction. Although rare, the intestines can get wrapped around the leads of the device, causing different degrees of obstruction (Figure 1). Positioning the device in the left upper quadrant minimizes the intraabdominal length of the leads and pulls them maximally out, coiling under the device (Figure 2). In cases where other locations are used either due to a hostile upper abdominal region (skin infection, presence of gastrostomy or other devices) or surgeon’s preference, the GES device can be implanted in the lower abdomen (Figure 3). In these circumstances, carefully draping the omentum over the bowels might help to prevent this complication. Tacking of the leads to the parietal peritoneum with sutures can also be preventative. In cases of obstruction requiring intervention by laparotomy or minimal access techniques (laparoscopy or robotic assisted surgery), all efforts are made to preserve the neurostimulator leads. In cases that require bowel resection, lead contamination is a serious concern, but lead explantation is not mandatory. Close postoperative monitoring for the development of lead infection is required.

Hematoma and seroma. Postoperative hematomas can occur from inadequate hemostasis, and seromas can occur in the stimulator pocket. Small hematomas may be observed if not complicated (Figure 4). In cases of large hematomas with skin compromise or dehiscence, prompt washout and drainage is required. In ideal cases, the device can be preserved. Relocation to another site might be required if skin necrosis develops. The possibility of device contamination also must be considered; after resolution of wound issues, implantation of a new device may be tried. Seromas at the generator pocket site are a frequent occurrence but are often benign, self-limiting, and generally resolve over 4 to 6 weeks.

Incisional hernia. Hernias can develop after any abdominal surgery and are not unique to GES implantation. Use of minimally invasive technique for the GES implantation minimizes this complication.

Electric shock sensations may occur from breakage of the plastic lining covering the stimulator wires or from fluid buildup around the insertion of the wires into the stimulator. Shocks can also occur due to shortening of the leads on the muscles of the abdominal wall. Patients describe periodic muscle cramps with the frequency of the device (every 5 seconds). To prevent this complication, freshly implanted leads should be covered by an omental flap to isolate them from the abdominal wall. In patients who continue to feel shocks despite all efforts, the possibility of visceral hypersensitivity should be considered. A trial of symptom modulators such as nortriptyline and lowering of the output amperage below the minimal recommended setting of 5 mA can be undertaken. If these interventions do not work, the device must be turned off for a period of time. Occasionally, replacement of the leads or explantation of the device must be considered.

Lack of effect/persistent symptoms. If a patient presents with lack of improvement after device implantation, a thorough workup should be undertaken to ensure that the device is functioning properly. In the case of abnormal impedance values, an abdominal x-ray study can be performed to rule out lead migration (Figure 5). If no abnormalities are detected, the output of the device can be increased. After adjusting device settings, the patient should be assessed for improvement over at least a 1- to 3-month period. One report suggests that in patients not responding to GES, repositioning the location of the stimulator leads on the stomach can be helpful.²⁹

Outcomes of GES

Study results of investigative GES models in animals and select patients were published in 1997.^30,31 Following these reports, 2 large multicenter studies were conducted to demonstrate the efficacy of GES for the treatment of refractory gastroparesis. The Gastric Electrical Mechanical Stimulation Study (GEMS) was an open-label, multicenter study of 38 patients who received percutaneous and later permanent GES devices.³² Marked reduction in weekly vomiting and nausea was observed at 4 weeks, with a 90% reduction in nausea and vomiting frequency at 11 months. Following this, a second multicenter study (Worldwide Anti-Vomiting Electrical Stimulation Study [WAVES]) involving a double-blind sham stimulation controlled trial with 33 idiopathic and diabetic gastroparesis patients was performed.³³ During the blinded portion of this study, there was a noticeable decrease in vomiting frequency, particularly in the patients with diabetic gastroparesis. Patient preference was for the stimulator ON as compared to OFF. The FDA’s HUD exemption for the Enterra GES device in 2000 was based on these studies.

Four independent double-blind studies of GES have been conducted (Table 1).^33-37 It has been difficult to demonstrate improvement during the double-blind period with gastric stimulation compared to no stimulation. Despite total symptom severity improvement and individual symptom improvements in these studies, a recent meta-analysis demonstrated a summative insignificant difference between the GES ON versus OFF states.³⁸

In contrast to the double-blind studies, numerous open-label studies have demonstrated clinical improvements in patients with diabetic and idiopathic gastroparesis (Table 2),^{32,33,35,36,39-55} leading some to question whether the demonstrable efficacy reflects a placebo effect or regression to the mean. Patients may perceive an operative, aggressive intervention as likely to be effectual in comparison to incremental medication efforts, thus creating a placebo effect. It should also be noted that not all open-label studies have demonstrated improvement with GES. Indeed, Jones et al reported no significant difference in nausea and vomiting at 6-month follow-up, and recommended that physicians exercise caution with GES as a therapeutic strategy given the cost and lack of confirmed demonstrable effect.⁵⁶ Thus, the clinical successes demonstrated in open-label studies must be weighed not only against the lack of unequivocal improvement, but also against the potential deleterious effects of the surgery.

In an open-label study that employed the GCSI to follow symptoms of gastroparesis, 29 patients underwent GES implantation over an 18-month period, with follow-up in 28 patients.⁴⁴ GES resulted in clinical improvement in 50% of patients with refractory gastroparesis. The overall GCSI significantly decreased, with improvement in the nausea/vomiting subscore and the post-prandial fullness subscore, but no improvement in the bloating subscore or abdominal pain. The decrease in GCSI was greater for patients with diabetic versus idiopathic gastroparesis. Patients with the main symptom of nausea/vomiting had a greater improvement than patients with the main symptom of abdominal pain. Patients taking narcotic analgesics at the time of implant had a poorer response compared to patients who were not. In this study, 3 clinical parameters were associated with a favorable clinical response: (1) diabetic rather than idiopathic gastroparesis, (2) nausea/vomiting rather than abdominal pain as the primary symptom, and (3) independence from narcotic analgesics prior to stimulator implantation. Knowledge of these 3 factors may allow improved patient selection for GES.

A large prospective study by Heckert et al detailed marked improvements with GES and the patterns of those improvements.⁵⁵ Nausea, vomiting, loss of appetite, and early satiety improved significantly with stimulator use, with a greater improvement in vomiting in patients with diabetic gastroparesis than in those with the idiopathic form. Although GES improved symptoms in 75% of all patients, patients with diabetes had a post-GES Clinical Patient Grading Assessment score that was statistically higher than the score among patients with idiopathic gastroparesis. This difference is thought to be due to the neuromolecular mechanism of diabetic gastroparesis, where blunting of the enteric nervous system may contribute to symptomatology.

Several studies have demonstrated a clinical response to GES in patients with postsurgical gastroparesis. A study by Oubre et al showed that GES led to weekly vomiting improvements as well as a reduction in total symptom severity score.⁵⁷ A study by McCallum et al further demonstrated improved symptoms, quality of life, nutritional status, and hospitalization requirements.⁵⁸ GES has also been shown to improve gastroparesis symptoms in pediatric populations.^47,59 Thus, although not a direct indication, GES has been shown to be beneficial in various subtypes of gastroparesis.

Additionally, irrespective of gastroparesis type, the improved symptomatology with GES appears to be durable, with one study showing persistent clinical improvements up to 8 years after device placement.⁶⁰ The improvements were persistent and incremental. Likewise, McCallum et al showed that continued reductions in total symptom severity scores were evident in all gastroparesis types up to 10 years after stimulator implantation.⁶¹ The success of the procedures in part comes from careful selection of patients. Clinical parameters that are associated with favorable clinical response include diabetic gastroparesis subtype, nausea/vomiting predominance, and independence from narcotic analgesics prior to stimulator placement.⁶²

GES has also been noted to improve other patient care metrics besides symptomatology, including nutritional status, reduced need for nutritional supplementation, and improved HbA1c.^63-65 Additionally, a study by Cutts et al established that health care resource utilization significantly improved at 12, 24, and 36 months following GES placement, as compared to patients receiving standard medical therapy.⁶⁶ This decreased resource utilization was also reflected in decreased costs in the GES group compared with the standard care group.

Surgical Alternatives to GES

Pyloric interventions such as pyloroplasty and pyloromyotomy are other surgical treatment modalities offered for gastroparesis. Whereas GES uses neurostimulation to facilitate gastric emptying and potentially improve fundic accommodation, pyloric interventions are intended to increase gastric emptying by reducing outflow resistance from the pyloric sphincter.

Pyloric Interventions

Various studies have shown significant improvements with pyloric interventions, similar to the improvements seen with GES. One such study involving 177 patients demonstrated an 86% improvement in gastric emptying, with symptom severity scores for nausea, vomiting, bloating, abdominal pain, and early satiety decreasing significantly at 3 months following pyloroplasty.⁶⁷ A significant advantage of pyloric interventions is that pyloromyotomy can be performed endoscopically (gastric peroral endoscopic pyloromyotomy [G-POEM] or peroral pyloromyotomy [POP]), thus minimizing the risks of open surgery. A recent review that included a pooled analysis of 7 studies of G-POEM for gastroparesis demonstrated 100% technical success, with clinical efficacy in 81.5% of the procedures as assessed by the GCSI.⁶⁸ Additionally, the intraoperative and perioperative complication rates were 6.6% and 7.6%, respectively, suggesting that G-POEM is a safe and clinically beneficial therapeutic option. Few studies comparing the outcomes of pyloric interventions to GES have been performed.

Recently, GES has been combined with pyloric interventions to maximize therapeutic potential. This allows simultaneous neurologic and functional interventions to expedite gastric emptying and improve patient symptomatology. Davis et al demonstrated significant improvement in 21 patients who underwent GES placement and pyloroplasty, with 71% improvement in total symptom severity.⁶⁹ Notably, dual surgery did not increase the incidence of infection or adverse surgical outcomes. Although this study did not directly compare dual surgery to GES alone, the results are nonetheless favorable. GES provides a strong antiemetic and anti-nausea effect, whereas the pyloromyotomy provides improvement in gastric emptying.

Feeding/Venting Tubes

Feeding jejunostomy tubes and venting gastrostomy tubes can be used alone or in combination with GES. Feeding jejunostomy is performed for malnutrition and weight loss that accompanies the refractory symptoms of early satiety, nausea, and vomiting. Venting gastrostomy tubes allow for removal of retained gastric contents that may cause distension, nausea, and vomiting. Gastrojejunostomy tubes can also be placed endoscopically or by interventional radiology.

Gastrectomy

Gastrectomy can provide therapeutic benefit through elimination of the gastric reservoir function and consequent removal of afferent neural impulses. In select patient populations, outcomes of gastrectomy have compared favorably with those of GES. For example, one study demonstrated favorable outcomes of Roux-en-Y gastrectomy in morbidly obese patients with gastroparesis.⁷⁰ In another study, favorable outcomes were reported in a cohort of 103 patients, with gastrectomy demonstrating 87% symptom improvement (nausea, vomiting, epigastric pain) compared to just 63% improvement with GES.⁷¹ However, the dramatic impact on anatomy and physiology and the invasiveness of the procedure need to be weighed against the therapeutic benefit. For example, in the same study, the 30-day morbidity was 23% for gastrectomy versus just 8% for the GES implant.⁷¹

The gastric electrical neurostimulator (Enterra; Medtronic, Inc.) is approved for treatment of idiopathic and diabetic gastroparesis that is refractory to medical treatment, performed on a compassionate basis. Patients with diabetic gastroparesis respond to GES better than do patients with the idiopathic form. Of the symptoms of gastroparesis, primarily nausea and vomiting improve. Thus, GES favors patients with diabetic gastroparesis who have primarily nausea and vomiting, rather than, for instance, patients with idiopathic gastroparesis who have primarily abdominal pain and may be taking narcotics. Some centers provide GES for postsurgical patients and children with gastroparesis.

The 3 main surgical interventions for medically refractory gastroparesis are GES, pyloric intervention (pyloroplasty or pyloromyotomy), and gastrectomy. Of the 3 interventions, gastrectomy is the most radical given its dramatic effect on anatomy and is thus not preferred. The clinical decision then becomes: GES, pyloric intervention, or both? There are limited data to support a definitive answer to this question.

In a single-center retrospective analysis of prospective data (electronic medical record), Arthur et al compared outcomes of GES patients with medically refractory gastroparesis who received various surgical interventions.⁷² In total, 33 stimulator, 7 pyloroplasty, 2 gastrectomy, and 16 combined stimulator and pyloroplasty patients were analyzed for postoperative symptom improvement. Pyloroplasty alone demonstrated the least symptom improvement, combination GES and pyloroplasty demonstrated increased improvement, and GES alone demonstrated the most improvement. The results of this study suggest that barring contraindication, placement of a gastric stimulator as the initial treatment is best, with pyloroplasty reserved for patients who do not achieve adequate symptom control. Limitations of the study include its single-center design and low patient numbers for pyloroplasty in isolation.

In contrast, a recent retrospective systematic review synthesized the outcomes of various studies of GES and pyloric interventions for medically refractory gastroparesis.⁷³ A therapeutic effect was found for each surgical intervention, with pyloric surgery patients demonstrating a greater response to intervention than GES patients. Unfortunately, attempts to analyze combination interventions were hindered by a lack of power.

Conclusion

Initial management of gastroparesis is medical (lifestyle and diet changes), with antiemetic and prokinetic agents used in refractory cases. Following failure of this therapy, placement of a GES device is a surgical intervention that has been approved under FDA humanitarian device exemption to help ameliorate symptomatology. Improvement with GES has been demonstrated in nonblinded studies, but the lack of randomized controlled trials demonstrating benefit suggests the possibility of an underlying placebo effect. Additionally, new medical procedures such as G-POEM complicate the decision of which intervention should be attempted first. More studies, specifically comparing GES, pyloric interventions, and combined GES with pyloric intervention to placebo, are needed to fully understand what therapy is best for refractory gastroparesis.

Corresponding author: Henry P. Parkman, MD, Gastroenterology Section, Temple University School of Medicine, 3401 North Broad Street, Philadelphia, PA 19140; [email protected].

Financial disclosures: None.

From Temple University School of Medicine, Philadelphia, PA.

Abstract

• Objective: To outline the use and utility of gastric electric stimulation (GES) as a therapeutic intervention for gastroparesis.
• Methods: Review of the literature.
• Results: Gastroparesis is characterized by delayed gastric emptying, with symptoms of nausea, vomiting, early satiety, postprandial fullness, and abdominal pain. Some patients with gastroparesis do not respond to medical intervention, and for these patients surgical intervention may be warranted. GES utilizes high-frequency gastric neurostimulation to facilitate gastric emptying and reduce symptoms of gastroparesis. It is indicated for patients with idiopathic and diabetic gastroparesis who have nausea and vomiting as their primary symptoms and who have not responded to medical therapy. GES has also been used in postsurgical and pediatric gastroparesis patients. Optimizing the outcome of this surgical treatment through proper patient selection and meticulous surgical technique is essential as there are inherent risks to the procedure. Nonblinded studies of GES for medically refractory gastroparesis have demonstrated therapeutic symptomatic benefit, whereas randomized controlled trials have not. New interventions such as pyloromyotomy and pyloroplasty are reasonable alternatives or addendums to GES.
• Conclusion: GES may be considered among the therapies available for treating patients with refractory symptoms of gastroparesis. More studies, specifically those comparing GES, pyloromyotomy, GES combined with pyloromyotomy, and placebo, are needed to help guide therapy selection for refractory gastroparesis.

Keywords: diabetes; gastroparesis; dysmotility; gastric emptying; electric stimulation.

Gastroparesis is a chronic dysmotility disorder characterized by delayed gastric emptying with associated symptoms of nausea, vomiting, early satiety, postprandial fullness, and abdominal pain. Medical treatments for gastroparesis include dietary modifications, glucose control in those with diabetes, prokinetic medications, antiemetic medications, and symptom modulators, but unfortunately patients frequently do not respond to these treatments. In patients refractory to medical therapy, surgical treatments can be considered.

Gastric electric stimulation (GES; Enterra [Medtronic, Minneapolis, MN]) was approved via a Food and Drug Administration (FDA) Humanitarian Use Device (HUD) exemption for the treatment of medically refractory gastroparesis in 2000. Understanding the indications, risks, outcomes, and alternatives to GES is essential to providing appropriate care for patients with medically refractory gastroparesis. This article outlines the use and utility of GES as a therapeutic intervention for gastroparesis.

Types of Gastroparesis

Gastroparesis is a chronic symptomatic disorder of the stomach manifested by delayed gastric emptying without evidence of gastric outlet obstruction or ulceration.¹ The pathophysiology of gastroparesis appears to involve abnormalities in functioning of several elements including the autonomic nervous system, especially the vagus nerve, smooth muscle cells, enteric neurons, and interstitial cells of Cajal.

Idiopathic gastroparesis and diabetic gastroparesis are the 2 most common types of gastroparesis.² Symptomatic delayed gastric emptying with no primary underlying abnormality predisposing to gastroparesis is categorized as idiopathic gastroparesis.³ A small subset of patients with idiopathic gastroparesis report an initial infectious prodrome such as gastroenteritis or respiratory infection. It has been suggested that this postinfectious gastroparesis results from viral injury to the neural innervation of the stomach or the interstitial cells of Cajal in the stomach.⁴ Viruses that have been implicated in the development of gastroparesis include cytomegalovirus, Epstein-Barr virus, Norwalk virus, rotavirus, herpes zoster, and varicella zoster.^5-9

Diabetic gastroparesis is characterized as onset of symptoms of gastroparesis in patients with diabetes, with concomitant delayed gastric emptying. It is often attributed to chronic hyperglycemia-induced damage to the vagus nerve, and is frequently observed in association with other diabetic complications such as neuropathy, retinopathy, and nephropathy.¹⁰

Gastroparesis that develops following surgery is classified as postsurgical gastroparesis. In the past, this form of gastroparesis most commonly occurred after ulcer surgery, often performed with vagotomy. These types of surgeries are performed less frequently in the era of proton pump inhibitor therapy and treatments for Helicobacter pylori. Presently, Nissen fundoplication and bariatric surgery are the more common surgical procedures associated with gastroparesis.³ Long-term use of medications that delay gastric emptying, such as opiate narcotic medications, can lead to gastroparesis and represent another form of iatrogenic gastroparesis. Other forms of gastroparesis (atypical gastroparesis) arise due to various underlying etiologies, including neurological disorders (eg, Parkinson disease, multiple sclerosis), metabolic or endocrine conditions (eg, hypothyroidism), autoimmune disorders, connective tissue and collagen vascular disorders (eg, systemic lupus erythematosus, scleroderma, Sjögren syndrome, Ehlers-Danlos syndrome), or eating disorders (eg, anorexia, bulimia).³

Epidemiology

There is a female preponderance in patients with gastroparesis. Data from the Rochester Epidemiology Project, a database of linked medical records for residents of Olmsted County, MN, showed that the age-adjusted prevalence of definite gastroparesis per 100,000 inhabitants was 37.8 for women and 9.6 for men.¹¹ More recent estimates have suggested a much higher prevalence of probable gastroparesis (approximately 1.8%) in the general population using symptoms suggestive of gastroparesis.¹² Hospitalization rates for gastroparesis have increased since 2000, which could reflect rising prevalence and/or the effects of heightened awareness about and better identification of gastroparesis.¹³ This increase may also be due in part to the rising rate of diabetes leading to more cases of diabetic gastroparesis; withdrawal of some gastroparesis treatments from the market (cisapride, tegaserod) leading to hospitalizations for symptoms not adequately being treated; and hospitalizations needed for insertion of the gastric electric stimulator.

Gastroparesis Symptoms

The main symptoms of gastroparesis are early satiety, postprandial fullness, bloating, nausea, and vomiting.¹⁴ Nausea (> 90% of patients) and early satiety (60% of patients) are the most common symptoms.¹⁵ Abdominal pain is often present in patients with gastroparesis but is usually not the predominant symptom. The pain can be multifactorial, with somatic, visceral, and neuropathic components.^16-18 Moderate to severe abdominal pain has been found more often in patients with idiopathic gastroparesis and in association with opiate use.¹⁶ Symptoms of gastroparesis may be persistent or present as episodic flares. Due to the symptoms, some patients will experience weight loss and malnutrition and, in severe cases, dehydration.¹⁹

Although the definition of gastroparesis is a delay in gastric emptying along with symptoms, symptoms correlate poorly with the degree of delayed gastric emptying. The symptoms that appear to have the strongest correlation with gastric emptying are nausea, vomiting, early satiety, and postprandial fullness, whereas symptoms such as abdominal pain and bloating have little correlation. Furthermore, improving gastric emptying does not necessarily lead to improved symptoms, and symptom improvement does not always lead to improved gastric emptying times.²⁰ Between 5% and 12% of patients with diabetes report symptoms consistent with gastroparesis, though many of these patients have normal gastric emptying. The symptoms of gastroparesis overlap with those of functional dyspepsia, as both may have motor and sensory alterations.²¹

The Gastroparesis Cardinal Symptom Index (GCSI), a subset of the Patient Assessment of Gastrointestinal Disorders Symptom Severity Index (PAGI-SYM), is a questionnaire that is commonly used to establish symptom severity in patients with gastroparesis. It is comprised of 3 subscales—nausea and vomiting, postprandial fullness and early satiety, and bloating—which are averaged to provide a total GCSI score. Symptoms over the 2 weeks prior to administration of the questionnaire are assessed and rated from 0 (none) to 5 (very severe).²² Grading the severity of gastroparesis may take into account symptoms, quality of life, and gastric emptying. One commonly used grading system assigns a grade from 1 to 3, with grade 1 being mild gastroparesis, grade 2 being compensated gastroparesis, and grade 3 being gastric failure with refractory symptoms that are uncontrolled.^18,23 Quality-of-life surveys also suggest that gastroparesis independent of other factors leads to a worse quality of life.²⁴

Indications for GES

Gastric electric stimulator implantation is a surgical procedure with inherent risks and complications and is reserved for patients with intractable symptoms of gastroparesis who remain symptomatic despite treatment attempts with dietary management, antiemetic agents (eg, compazine, phenergan, and ondansetron), and prokinetic agents (eg, metoclopramide, erythromycin, and domperidone). Symptom modulators such as nortriptyline and mirtazapine are occasionally tried.

Surgical intervention can be considered upon failure of medical treatment measures. At least a year of documented care provided by a physician specializing in gastroparesis is suggested for surgical consideration. The gastric electric neurostimulator is approved by the FDA as a HUD for the care of patients with idiopathic and diabetic gastroparesis, performed on a compassionate basis. GES implantation requires Institutional Review Board approval at the institution, and patients are required to have documented delayed gastric emptying.

It is important to remember that the GES device is incompatible with magnetic resonance imaging (MRI) and explantation of the device is necessary prior to MRI. As such, in patients with anticipated need of frequent MRI, such as those with multiple sclerosis, serious consideration should be given to alternative strategies prior to focusing on this modality.

Device Placement

GES was devised to improve gastric emptying. The Enterra GES system uses high-frequency, low-energy electric stimulation. An alternative method is true gastric pacing that uses high-energy, low-frequency stimulation to entrain the gastric slow waves and subsequent contractions at 3 cycles per minute (cpm). Gastric pacing has greater energy requirements than GES, which makes the size of the stimulator too large to be practical. In pilot animal studies, GES produced an accelerating effect on gastric emptying, but in human studies GES had an inconsistent effect on gastric emptying. Studies have suggested that GES influences the proximal stomach, with a reduction of gastric tone,²⁵ and also that GES has an afferent modulatory mechanism.²⁶

The Enterra GES is placed surgically under general anesthesia, commonly via laparotomy or minimal access surgical techniques (laparoscopically or robotically assisted). Preoperative intravenous antibiotics are given. The system consists of a pair of electrodes connected to a pulse generator. The 2 stimulation leads are inserted into the gastric muscularis propria 1 cm apart along the greater curvature 10 cm proximal to the pylorus. Upper endoscopy is performed to ensure that the leads do not penetrate through the mucosa into the stomach lumen; if this occurs, repositioning of the lead is necessary. A horizontal incision through the skin is made, and the distal ends of the stimulating wires are tunneled through the abdominal wall and connected to the pulse generator. The impedance (resistance) between the wires is measured to ensure the appropriate range (200-800 Ohms). The neurostimulator with the distal ends of the stimulating wires is then placed into the subcutaneous pocket and sutured to the underlying fascia. The pulse generator delivers a high-frequency, low-energy, 0.1-second train of pulses at a frequency of 12 cpm. Within each pulse train, individual pulses oscillate at a frequency of 14 cycles per second. The voltage of the stimulations is set to provide a current of 5 milliamps (mA; remembering that voltage = current × resistance).

Patients are often hospitalized with a recovery time of 1 to 3 days. Immediate postoperative care usually includes intravenous fluids, controlling any postoperative ileus, advancing diet, and providing analgesic pain medications. Hospital length of stay can be impacted by surgical technique.²⁵ Patients are seen several weeks after discharge for assessment of the incision and toleration of diet. Medications for gastroparesis that patients were taking prior to the GES implantation are usually continued postoperatively, with a goal of reducing these medications over time. Patients are then followed every 3 to 12 months, depending on their clinical condition.

At follow-up visits, medications are reviewed and new treatments can be added if appropriate. The gastric stimulator is interrogated to determine if changes in resistance occurred; if necessary, minor readjustments can be made to keep the current at desired levels (5 mA). For persistent symptoms with GES treatment, the stimulator parameters can be adjusted after 3 months of follow up, typically first increasing the current from 5 to 7.5 mA and then to 10 mA. After this, the frequency can be increased from 14 Hz to 28 Hz, and then to 55 Hz. Rarely, the ON duration is increased from 0.1 to 1 second. Increasing the ON time can worsen symptoms in some patients, cause abdominal pain, and decrease the battery life from the usual 7 years.

Complications of GES

In an analysis of the Manufacturer and User Facility Device Experience (MAUDE) databank, Bielefeldt identified 1587 reports of adverse effects related to the gastric electric stimulator from January 2001 to October 2015.²⁷ The most common adverse effects are reviewed here.

Skin erosion/wound dehiscence is one of the most common reported complications; it may be related to superficial placement or inadequate securing of the device to the fascia. Abscess can develop postoperatively due to hematogenous seeding or may be a sign of lead erosion into the lumen, tracking along the leads into subcutaneous tissue.²⁸ It is important to warn patients to protect the area over the device from needle injections as this also can lead to hematoma formation and direct contamination of the device. If the device gets infected, it cannot be salvaged and requires explantation. Implantation of a new device can be attempted once all wound issues resolve.

Device migration/flipping most often occurs because the device is inadequately fixed to the underlying fascia, but occasionally it can occur from patients flipping the device around. Flipping can occur due to superficial pocket location within subcutaneous tissue, especially in obese patients. Migration/flipping can lead to prominence of the contour of the device and discomfort, ultimately requiring surgical correction.

Perforation and erosion of the leads. With time, leads can erode into the stomach, although this is rare. Usually erosion is associated with loss of device function. Endoscopy confirms this finding. In rare cases, infection can track proximally along the lead and present as a surgical site infection at the pulse generator. This complication often requires explantation of the neurostimulator leads and pulse generator.

Intestinal obstruction. Although rare, the intestines can get wrapped around the leads of the device, causing different degrees of obstruction (Figure 1). Positioning the device in the left upper quadrant minimizes the intraabdominal length of the leads and pulls them maximally out, coiling under the device (Figure 2). In cases where other locations are used either due to a hostile upper abdominal region (skin infection, presence of gastrostomy or other devices) or surgeon’s preference, the GES device can be implanted in the lower abdomen (Figure 3). In these circumstances, carefully draping the omentum over the bowels might help to prevent this complication. Tacking of the leads to the parietal peritoneum with sutures can also be preventative. In cases of obstruction requiring intervention by laparotomy or minimal access techniques (laparoscopy or robotic assisted surgery), all efforts are made to preserve the neurostimulator leads. In cases that require bowel resection, lead contamination is a serious concern, but lead explantation is not mandatory. Close postoperative monitoring for the development of lead infection is required.

Hematoma and seroma. Postoperative hematomas can occur from inadequate hemostasis, and seromas can occur in the stimulator pocket. Small hematomas may be observed if not complicated (Figure 4). In cases of large hematomas with skin compromise or dehiscence, prompt washout and drainage is required. In ideal cases, the device can be preserved. Relocation to another site might be required if skin necrosis develops. The possibility of device contamination also must be considered; after resolution of wound issues, implantation of a new device may be tried. Seromas at the generator pocket site are a frequent occurrence but are often benign, self-limiting, and generally resolve over 4 to 6 weeks.

Incisional hernia. Hernias can develop after any abdominal surgery and are not unique to GES implantation. Use of minimally invasive technique for the GES implantation minimizes this complication.

Electric shock sensations may occur from breakage of the plastic lining covering the stimulator wires or from fluid buildup around the insertion of the wires into the stimulator. Shocks can also occur due to shortening of the leads on the muscles of the abdominal wall. Patients describe periodic muscle cramps with the frequency of the device (every 5 seconds). To prevent this complication, freshly implanted leads should be covered by an omental flap to isolate them from the abdominal wall. In patients who continue to feel shocks despite all efforts, the possibility of visceral hypersensitivity should be considered. A trial of symptom modulators such as nortriptyline and lowering of the output amperage below the minimal recommended setting of 5 mA can be undertaken. If these interventions do not work, the device must be turned off for a period of time. Occasionally, replacement of the leads or explantation of the device must be considered.

Lack of effect/persistent symptoms. If a patient presents with lack of improvement after device implantation, a thorough workup should be undertaken to ensure that the device is functioning properly. In the case of abnormal impedance values, an abdominal x-ray study can be performed to rule out lead migration (Figure 5). If no abnormalities are detected, the output of the device can be increased. After adjusting device settings, the patient should be assessed for improvement over at least a 1- to 3-month period. One report suggests that in patients not responding to GES, repositioning the location of the stimulator leads on the stomach can be helpful.²⁹

Outcomes of GES

Study results of investigative GES models in animals and select patients were published in 1997.^30,31 Following these reports, 2 large multicenter studies were conducted to demonstrate the efficacy of GES for the treatment of refractory gastroparesis. The Gastric Electrical Mechanical Stimulation Study (GEMS) was an open-label, multicenter study of 38 patients who received percutaneous and later permanent GES devices.³² Marked reduction in weekly vomiting and nausea was observed at 4 weeks, with a 90% reduction in nausea and vomiting frequency at 11 months. Following this, a second multicenter study (Worldwide Anti-Vomiting Electrical Stimulation Study [WAVES]) involving a double-blind sham stimulation controlled trial with 33 idiopathic and diabetic gastroparesis patients was performed.³³ During the blinded portion of this study, there was a noticeable decrease in vomiting frequency, particularly in the patients with diabetic gastroparesis. Patient preference was for the stimulator ON as compared to OFF. The FDA’s HUD exemption for the Enterra GES device in 2000 was based on these studies.

Four independent double-blind studies of GES have been conducted (Table 1).^33-37 It has been difficult to demonstrate improvement during the double-blind period with gastric stimulation compared to no stimulation. Despite total symptom severity improvement and individual symptom improvements in these studies, a recent meta-analysis demonstrated a summative insignificant difference between the GES ON versus OFF states.³⁸

In contrast to the double-blind studies, numerous open-label studies have demonstrated clinical improvements in patients with diabetic and idiopathic gastroparesis (Table 2),^{32,33,35,36,39-55} leading some to question whether the demonstrable efficacy reflects a placebo effect or regression to the mean. Patients may perceive an operative, aggressive intervention as likely to be effectual in comparison to incremental medication efforts, thus creating a placebo effect. It should also be noted that not all open-label studies have demonstrated improvement with GES. Indeed, Jones et al reported no significant difference in nausea and vomiting at 6-month follow-up, and recommended that physicians exercise caution with GES as a therapeutic strategy given the cost and lack of confirmed demonstrable effect.⁵⁶ Thus, the clinical successes demonstrated in open-label studies must be weighed not only against the lack of unequivocal improvement, but also against the potential deleterious effects of the surgery.

In an open-label study that employed the GCSI to follow symptoms of gastroparesis, 29 patients underwent GES implantation over an 18-month period, with follow-up in 28 patients.⁴⁴ GES resulted in clinical improvement in 50% of patients with refractory gastroparesis. The overall GCSI significantly decreased, with improvement in the nausea/vomiting subscore and the post-prandial fullness subscore, but no improvement in the bloating subscore or abdominal pain. The decrease in GCSI was greater for patients with diabetic versus idiopathic gastroparesis. Patients with the main symptom of nausea/vomiting had a greater improvement than patients with the main symptom of abdominal pain. Patients taking narcotic analgesics at the time of implant had a poorer response compared to patients who were not. In this study, 3 clinical parameters were associated with a favorable clinical response: (1) diabetic rather than idiopathic gastroparesis, (2) nausea/vomiting rather than abdominal pain as the primary symptom, and (3) independence from narcotic analgesics prior to stimulator implantation. Knowledge of these 3 factors may allow improved patient selection for GES.

A large prospective study by Heckert et al detailed marked improvements with GES and the patterns of those improvements.⁵⁵ Nausea, vomiting, loss of appetite, and early satiety improved significantly with stimulator use, with a greater improvement in vomiting in patients with diabetic gastroparesis than in those with the idiopathic form. Although GES improved symptoms in 75% of all patients, patients with diabetes had a post-GES Clinical Patient Grading Assessment score that was statistically higher than the score among patients with idiopathic gastroparesis. This difference is thought to be due to the neuromolecular mechanism of diabetic gastroparesis, where blunting of the enteric nervous system may contribute to symptomatology.

Several studies have demonstrated a clinical response to GES in patients with postsurgical gastroparesis. A study by Oubre et al showed that GES led to weekly vomiting improvements as well as a reduction in total symptom severity score.⁵⁷ A study by McCallum et al further demonstrated improved symptoms, quality of life, nutritional status, and hospitalization requirements.⁵⁸ GES has also been shown to improve gastroparesis symptoms in pediatric populations.^47,59 Thus, although not a direct indication, GES has been shown to be beneficial in various subtypes of gastroparesis.

Additionally, irrespective of gastroparesis type, the improved symptomatology with GES appears to be durable, with one study showing persistent clinical improvements up to 8 years after device placement.⁶⁰ The improvements were persistent and incremental. Likewise, McCallum et al showed that continued reductions in total symptom severity scores were evident in all gastroparesis types up to 10 years after stimulator implantation.⁶¹ The success of the procedures in part comes from careful selection of patients. Clinical parameters that are associated with favorable clinical response include diabetic gastroparesis subtype, nausea/vomiting predominance, and independence from narcotic analgesics prior to stimulator placement.⁶²

GES has also been noted to improve other patient care metrics besides symptomatology, including nutritional status, reduced need for nutritional supplementation, and improved HbA1c.^63-65 Additionally, a study by Cutts et al established that health care resource utilization significantly improved at 12, 24, and 36 months following GES placement, as compared to patients receiving standard medical therapy.⁶⁶ This decreased resource utilization was also reflected in decreased costs in the GES group compared with the standard care group.

Surgical Alternatives to GES

Pyloric interventions such as pyloroplasty and pyloromyotomy are other surgical treatment modalities offered for gastroparesis. Whereas GES uses neurostimulation to facilitate gastric emptying and potentially improve fundic accommodation, pyloric interventions are intended to increase gastric emptying by reducing outflow resistance from the pyloric sphincter.

Pyloric Interventions

Various studies have shown significant improvements with pyloric interventions, similar to the improvements seen with GES. One such study involving 177 patients demonstrated an 86% improvement in gastric emptying, with symptom severity scores for nausea, vomiting, bloating, abdominal pain, and early satiety decreasing significantly at 3 months following pyloroplasty.⁶⁷ A significant advantage of pyloric interventions is that pyloromyotomy can be performed endoscopically (gastric peroral endoscopic pyloromyotomy [G-POEM] or peroral pyloromyotomy [POP]), thus minimizing the risks of open surgery. A recent review that included a pooled analysis of 7 studies of G-POEM for gastroparesis demonstrated 100% technical success, with clinical efficacy in 81.5% of the procedures as assessed by the GCSI.⁶⁸ Additionally, the intraoperative and perioperative complication rates were 6.6% and 7.6%, respectively, suggesting that G-POEM is a safe and clinically beneficial therapeutic option. Few studies comparing the outcomes of pyloric interventions to GES have been performed.

Recently, GES has been combined with pyloric interventions to maximize therapeutic potential. This allows simultaneous neurologic and functional interventions to expedite gastric emptying and improve patient symptomatology. Davis et al demonstrated significant improvement in 21 patients who underwent GES placement and pyloroplasty, with 71% improvement in total symptom severity.⁶⁹ Notably, dual surgery did not increase the incidence of infection or adverse surgical outcomes. Although this study did not directly compare dual surgery to GES alone, the results are nonetheless favorable. GES provides a strong antiemetic and anti-nausea effect, whereas the pyloromyotomy provides improvement in gastric emptying.

Feeding/Venting Tubes

Feeding jejunostomy tubes and venting gastrostomy tubes can be used alone or in combination with GES. Feeding jejunostomy is performed for malnutrition and weight loss that accompanies the refractory symptoms of early satiety, nausea, and vomiting. Venting gastrostomy tubes allow for removal of retained gastric contents that may cause distension, nausea, and vomiting. Gastrojejunostomy tubes can also be placed endoscopically or by interventional radiology.

Gastrectomy

Gastrectomy can provide therapeutic benefit through elimination of the gastric reservoir function and consequent removal of afferent neural impulses. In select patient populations, outcomes of gastrectomy have compared favorably with those of GES. For example, one study demonstrated favorable outcomes of Roux-en-Y gastrectomy in morbidly obese patients with gastroparesis.⁷⁰ In another study, favorable outcomes were reported in a cohort of 103 patients, with gastrectomy demonstrating 87% symptom improvement (nausea, vomiting, epigastric pain) compared to just 63% improvement with GES.⁷¹ However, the dramatic impact on anatomy and physiology and the invasiveness of the procedure need to be weighed against the therapeutic benefit. For example, in the same study, the 30-day morbidity was 23% for gastrectomy versus just 8% for the GES implant.⁷¹

The gastric electrical neurostimulator (Enterra; Medtronic, Inc.) is approved for treatment of idiopathic and diabetic gastroparesis that is refractory to medical treatment, performed on a compassionate basis. Patients with diabetic gastroparesis respond to GES better than do patients with the idiopathic form. Of the symptoms of gastroparesis, primarily nausea and vomiting improve. Thus, GES favors patients with diabetic gastroparesis who have primarily nausea and vomiting, rather than, for instance, patients with idiopathic gastroparesis who have primarily abdominal pain and may be taking narcotics. Some centers provide GES for postsurgical patients and children with gastroparesis.

The 3 main surgical interventions for medically refractory gastroparesis are GES, pyloric intervention (pyloroplasty or pyloromyotomy), and gastrectomy. Of the 3 interventions, gastrectomy is the most radical given its dramatic effect on anatomy and is thus not preferred. The clinical decision then becomes: GES, pyloric intervention, or both? There are limited data to support a definitive answer to this question.

In a single-center retrospective analysis of prospective data (electronic medical record), Arthur et al compared outcomes of GES patients with medically refractory gastroparesis who received various surgical interventions.⁷² In total, 33 stimulator, 7 pyloroplasty, 2 gastrectomy, and 16 combined stimulator and pyloroplasty patients were analyzed for postoperative symptom improvement. Pyloroplasty alone demonstrated the least symptom improvement, combination GES and pyloroplasty demonstrated increased improvement, and GES alone demonstrated the most improvement. The results of this study suggest that barring contraindication, placement of a gastric stimulator as the initial treatment is best, with pyloroplasty reserved for patients who do not achieve adequate symptom control. Limitations of the study include its single-center design and low patient numbers for pyloroplasty in isolation.

In contrast, a recent retrospective systematic review synthesized the outcomes of various studies of GES and pyloric interventions for medically refractory gastroparesis.⁷³ A therapeutic effect was found for each surgical intervention, with pyloric surgery patients demonstrating a greater response to intervention than GES patients. Unfortunately, attempts to analyze combination interventions were hindered by a lack of power.

Conclusion

Initial management of gastroparesis is medical (lifestyle and diet changes), with antiemetic and prokinetic agents used in refractory cases. Following failure of this therapy, placement of a GES device is a surgical intervention that has been approved under FDA humanitarian device exemption to help ameliorate symptomatology. Improvement with GES has been demonstrated in nonblinded studies, but the lack of randomized controlled trials demonstrating benefit suggests the possibility of an underlying placebo effect. Additionally, new medical procedures such as G-POEM complicate the decision of which intervention should be attempted first. More studies, specifically comparing GES, pyloric interventions, and combined GES with pyloric intervention to placebo, are needed to fully understand what therapy is best for refractory gastroparesis.

Corresponding author: Henry P. Parkman, MD, Gastroenterology Section, Temple University School of Medicine, 3401 North Broad Street, Philadelphia, PA 19140; [email protected].

Financial disclosures: None.

From Temple University School of Medicine, Philadelphia, PA.

Abstract

• Objective: To outline the use and utility of gastric electric stimulation (GES) as a therapeutic intervention for gastroparesis.
• Methods: Review of the literature.
• Results: Gastroparesis is characterized by delayed gastric emptying, with symptoms of nausea, vomiting, early satiety, postprandial fullness, and abdominal pain. Some patients with gastroparesis do not respond to medical intervention, and for these patients surgical intervention may be warranted. GES utilizes high-frequency gastric neurostimulation to facilitate gastric emptying and reduce symptoms of gastroparesis. It is indicated for patients with idiopathic and diabetic gastroparesis who have nausea and vomiting as their primary symptoms and who have not responded to medical therapy. GES has also been used in postsurgical and pediatric gastroparesis patients. Optimizing the outcome of this surgical treatment through proper patient selection and meticulous surgical technique is essential as there are inherent risks to the procedure. Nonblinded studies of GES for medically refractory gastroparesis have demonstrated therapeutic symptomatic benefit, whereas randomized controlled trials have not. New interventions such as pyloromyotomy and pyloroplasty are reasonable alternatives or addendums to GES.
• Conclusion: GES may be considered among the therapies available for treating patients with refractory symptoms of gastroparesis. More studies, specifically those comparing GES, pyloromyotomy, GES combined with pyloromyotomy, and placebo, are needed to help guide therapy selection for refractory gastroparesis.

Keywords: diabetes; gastroparesis; dysmotility; gastric emptying; electric stimulation.

Gastroparesis is a chronic dysmotility disorder characterized by delayed gastric emptying with associated symptoms of nausea, vomiting, early satiety, postprandial fullness, and abdominal pain. Medical treatments for gastroparesis include dietary modifications, glucose control in those with diabetes, prokinetic medications, antiemetic medications, and symptom modulators, but unfortunately patients frequently do not respond to these treatments. In patients refractory to medical therapy, surgical treatments can be considered.

Gastric electric stimulation (GES; Enterra [Medtronic, Minneapolis, MN]) was approved via a Food and Drug Administration (FDA) Humanitarian Use Device (HUD) exemption for the treatment of medically refractory gastroparesis in 2000. Understanding the indications, risks, outcomes, and alternatives to GES is essential to providing appropriate care for patients with medically refractory gastroparesis. This article outlines the use and utility of GES as a therapeutic intervention for gastroparesis.

Types of Gastroparesis

Gastroparesis is a chronic symptomatic disorder of the stomach manifested by delayed gastric emptying without evidence of gastric outlet obstruction or ulceration.¹ The pathophysiology of gastroparesis appears to involve abnormalities in functioning of several elements including the autonomic nervous system, especially the vagus nerve, smooth muscle cells, enteric neurons, and interstitial cells of Cajal.

Idiopathic gastroparesis and diabetic gastroparesis are the 2 most common types of gastroparesis.² Symptomatic delayed gastric emptying with no primary underlying abnormality predisposing to gastroparesis is categorized as idiopathic gastroparesis.³ A small subset of patients with idiopathic gastroparesis report an initial infectious prodrome such as gastroenteritis or respiratory infection. It has been suggested that this postinfectious gastroparesis results from viral injury to the neural innervation of the stomach or the interstitial cells of Cajal in the stomach.⁴ Viruses that have been implicated in the development of gastroparesis include cytomegalovirus, Epstein-Barr virus, Norwalk virus, rotavirus, herpes zoster, and varicella zoster.^5-9

Diabetic gastroparesis is characterized as onset of symptoms of gastroparesis in patients with diabetes, with concomitant delayed gastric emptying. It is often attributed to chronic hyperglycemia-induced damage to the vagus nerve, and is frequently observed in association with other diabetic complications such as neuropathy, retinopathy, and nephropathy.¹⁰

Gastroparesis that develops following surgery is classified as postsurgical gastroparesis. In the past, this form of gastroparesis most commonly occurred after ulcer surgery, often performed with vagotomy. These types of surgeries are performed less frequently in the era of proton pump inhibitor therapy and treatments for Helicobacter pylori. Presently, Nissen fundoplication and bariatric surgery are the more common surgical procedures associated with gastroparesis.³ Long-term use of medications that delay gastric emptying, such as opiate narcotic medications, can lead to gastroparesis and represent another form of iatrogenic gastroparesis. Other forms of gastroparesis (atypical gastroparesis) arise due to various underlying etiologies, including neurological disorders (eg, Parkinson disease, multiple sclerosis), metabolic or endocrine conditions (eg, hypothyroidism), autoimmune disorders, connective tissue and collagen vascular disorders (eg, systemic lupus erythematosus, scleroderma, Sjögren syndrome, Ehlers-Danlos syndrome), or eating disorders (eg, anorexia, bulimia).³

Epidemiology

There is a female preponderance in patients with gastroparesis. Data from the Rochester Epidemiology Project, a database of linked medical records for residents of Olmsted County, MN, showed that the age-adjusted prevalence of definite gastroparesis per 100,000 inhabitants was 37.8 for women and 9.6 for men.¹¹ More recent estimates have suggested a much higher prevalence of probable gastroparesis (approximately 1.8%) in the general population using symptoms suggestive of gastroparesis.¹² Hospitalization rates for gastroparesis have increased since 2000, which could reflect rising prevalence and/or the effects of heightened awareness about and better identification of gastroparesis.¹³ This increase may also be due in part to the rising rate of diabetes leading to more cases of diabetic gastroparesis; withdrawal of some gastroparesis treatments from the market (cisapride, tegaserod) leading to hospitalizations for symptoms not adequately being treated; and hospitalizations needed for insertion of the gastric electric stimulator.

Gastroparesis Symptoms

The main symptoms of gastroparesis are early satiety, postprandial fullness, bloating, nausea, and vomiting.¹⁴ Nausea (> 90% of patients) and early satiety (60% of patients) are the most common symptoms.¹⁵ Abdominal pain is often present in patients with gastroparesis but is usually not the predominant symptom. The pain can be multifactorial, with somatic, visceral, and neuropathic components.^16-18 Moderate to severe abdominal pain has been found more often in patients with idiopathic gastroparesis and in association with opiate use.¹⁶ Symptoms of gastroparesis may be persistent or present as episodic flares. Due to the symptoms, some patients will experience weight loss and malnutrition and, in severe cases, dehydration.¹⁹

Although the definition of gastroparesis is a delay in gastric emptying along with symptoms, symptoms correlate poorly with the degree of delayed gastric emptying. The symptoms that appear to have the strongest correlation with gastric emptying are nausea, vomiting, early satiety, and postprandial fullness, whereas symptoms such as abdominal pain and bloating have little correlation. Furthermore, improving gastric emptying does not necessarily lead to improved symptoms, and symptom improvement does not always lead to improved gastric emptying times.²⁰ Between 5% and 12% of patients with diabetes report symptoms consistent with gastroparesis, though many of these patients have normal gastric emptying. The symptoms of gastroparesis overlap with those of functional dyspepsia, as both may have motor and sensory alterations.²¹

The Gastroparesis Cardinal Symptom Index (GCSI), a subset of the Patient Assessment of Gastrointestinal Disorders Symptom Severity Index (PAGI-SYM), is a questionnaire that is commonly used to establish symptom severity in patients with gastroparesis. It is comprised of 3 subscales—nausea and vomiting, postprandial fullness and early satiety, and bloating—which are averaged to provide a total GCSI score. Symptoms over the 2 weeks prior to administration of the questionnaire are assessed and rated from 0 (none) to 5 (very severe).²² Grading the severity of gastroparesis may take into account symptoms, quality of life, and gastric emptying. One commonly used grading system assigns a grade from 1 to 3, with grade 1 being mild gastroparesis, grade 2 being compensated gastroparesis, and grade 3 being gastric failure with refractory symptoms that are uncontrolled.^18,23 Quality-of-life surveys also suggest that gastroparesis independent of other factors leads to a worse quality of life.²⁴

Indications for GES

Gastric electric stimulator implantation is a surgical procedure with inherent risks and complications and is reserved for patients with intractable symptoms of gastroparesis who remain symptomatic despite treatment attempts with dietary management, antiemetic agents (eg, compazine, phenergan, and ondansetron), and prokinetic agents (eg, metoclopramide, erythromycin, and domperidone). Symptom modulators such as nortriptyline and mirtazapine are occasionally tried.

Surgical intervention can be considered upon failure of medical treatment measures. At least a year of documented care provided by a physician specializing in gastroparesis is suggested for surgical consideration. The gastric electric neurostimulator is approved by the FDA as a HUD for the care of patients with idiopathic and diabetic gastroparesis, performed on a compassionate basis. GES implantation requires Institutional Review Board approval at the institution, and patients are required to have documented delayed gastric emptying.

It is important to remember that the GES device is incompatible with magnetic resonance imaging (MRI) and explantation of the device is necessary prior to MRI. As such, in patients with anticipated need of frequent MRI, such as those with multiple sclerosis, serious consideration should be given to alternative strategies prior to focusing on this modality.

Device Placement

GES was devised to improve gastric emptying. The Enterra GES system uses high-frequency, low-energy electric stimulation. An alternative method is true gastric pacing that uses high-energy, low-frequency stimulation to entrain the gastric slow waves and subsequent contractions at 3 cycles per minute (cpm). Gastric pacing has greater energy requirements than GES, which makes the size of the stimulator too large to be practical. In pilot animal studies, GES produced an accelerating effect on gastric emptying, but in human studies GES had an inconsistent effect on gastric emptying. Studies have suggested that GES influences the proximal stomach, with a reduction of gastric tone,²⁵ and also that GES has an afferent modulatory mechanism.²⁶

The Enterra GES is placed surgically under general anesthesia, commonly via laparotomy or minimal access surgical techniques (laparoscopically or robotically assisted). Preoperative intravenous antibiotics are given. The system consists of a pair of electrodes connected to a pulse generator. The 2 stimulation leads are inserted into the gastric muscularis propria 1 cm apart along the greater curvature 10 cm proximal to the pylorus. Upper endoscopy is performed to ensure that the leads do not penetrate through the mucosa into the stomach lumen; if this occurs, repositioning of the lead is necessary. A horizontal incision through the skin is made, and the distal ends of the stimulating wires are tunneled through the abdominal wall and connected to the pulse generator. The impedance (resistance) between the wires is measured to ensure the appropriate range (200-800 Ohms). The neurostimulator with the distal ends of the stimulating wires is then placed into the subcutaneous pocket and sutured to the underlying fascia. The pulse generator delivers a high-frequency, low-energy, 0.1-second train of pulses at a frequency of 12 cpm. Within each pulse train, individual pulses oscillate at a frequency of 14 cycles per second. The voltage of the stimulations is set to provide a current of 5 milliamps (mA; remembering that voltage = current × resistance).

Patients are often hospitalized with a recovery time of 1 to 3 days. Immediate postoperative care usually includes intravenous fluids, controlling any postoperative ileus, advancing diet, and providing analgesic pain medications. Hospital length of stay can be impacted by surgical technique.²⁵ Patients are seen several weeks after discharge for assessment of the incision and toleration of diet. Medications for gastroparesis that patients were taking prior to the GES implantation are usually continued postoperatively, with a goal of reducing these medications over time. Patients are then followed every 3 to 12 months, depending on their clinical condition.

At follow-up visits, medications are reviewed and new treatments can be added if appropriate. The gastric stimulator is interrogated to determine if changes in resistance occurred; if necessary, minor readjustments can be made to keep the current at desired levels (5 mA). For persistent symptoms with GES treatment, the stimulator parameters can be adjusted after 3 months of follow up, typically first increasing the current from 5 to 7.5 mA and then to 10 mA. After this, the frequency can be increased from 14 Hz to 28 Hz, and then to 55 Hz. Rarely, the ON duration is increased from 0.1 to 1 second. Increasing the ON time can worsen symptoms in some patients, cause abdominal pain, and decrease the battery life from the usual 7 years.

Complications of GES

In an analysis of the Manufacturer and User Facility Device Experience (MAUDE) databank, Bielefeldt identified 1587 reports of adverse effects related to the gastric electric stimulator from January 2001 to October 2015.²⁷ The most common adverse effects are reviewed here.

Skin erosion/wound dehiscence is one of the most common reported complications; it may be related to superficial placement or inadequate securing of the device to the fascia. Abscess can develop postoperatively due to hematogenous seeding or may be a sign of lead erosion into the lumen, tracking along the leads into subcutaneous tissue.²⁸ It is important to warn patients to protect the area over the device from needle injections as this also can lead to hematoma formation and direct contamination of the device. If the device gets infected, it cannot be salvaged and requires explantation. Implantation of a new device can be attempted once all wound issues resolve.

Device migration/flipping most often occurs because the device is inadequately fixed to the underlying fascia, but occasionally it can occur from patients flipping the device around. Flipping can occur due to superficial pocket location within subcutaneous tissue, especially in obese patients. Migration/flipping can lead to prominence of the contour of the device and discomfort, ultimately requiring surgical correction.

Perforation and erosion of the leads. With time, leads can erode into the stomach, although this is rare. Usually erosion is associated with loss of device function. Endoscopy confirms this finding. In rare cases, infection can track proximally along the lead and present as a surgical site infection at the pulse generator. This complication often requires explantation of the neurostimulator leads and pulse generator.

Intestinal obstruction. Although rare, the intestines can get wrapped around the leads of the device, causing different degrees of obstruction (Figure 1). Positioning the device in the left upper quadrant minimizes the intraabdominal length of the leads and pulls them maximally out, coiling under the device (Figure 2). In cases where other locations are used either due to a hostile upper abdominal region (skin infection, presence of gastrostomy or other devices) or surgeon’s preference, the GES device can be implanted in the lower abdomen (Figure 3). In these circumstances, carefully draping the omentum over the bowels might help to prevent this complication. Tacking of the leads to the parietal peritoneum with sutures can also be preventative. In cases of obstruction requiring intervention by laparotomy or minimal access techniques (laparoscopy or robotic assisted surgery), all efforts are made to preserve the neurostimulator leads. In cases that require bowel resection, lead contamination is a serious concern, but lead explantation is not mandatory. Close postoperative monitoring for the development of lead infection is required.

Hematoma and seroma. Postoperative hematomas can occur from inadequate hemostasis, and seromas can occur in the stimulator pocket. Small hematomas may be observed if not complicated (Figure 4). In cases of large hematomas with skin compromise or dehiscence, prompt washout and drainage is required. In ideal cases, the device can be preserved. Relocation to another site might be required if skin necrosis develops. The possibility of device contamination also must be considered; after resolution of wound issues, implantation of a new device may be tried. Seromas at the generator pocket site are a frequent occurrence but are often benign, self-limiting, and generally resolve over 4 to 6 weeks.

Incisional hernia. Hernias can develop after any abdominal surgery and are not unique to GES implantation. Use of minimally invasive technique for the GES implantation minimizes this complication.

Electric shock sensations may occur from breakage of the plastic lining covering the stimulator wires or from fluid buildup around the insertion of the wires into the stimulator. Shocks can also occur due to shortening of the leads on the muscles of the abdominal wall. Patients describe periodic muscle cramps with the frequency of the device (every 5 seconds). To prevent this complication, freshly implanted leads should be covered by an omental flap to isolate them from the abdominal wall. In patients who continue to feel shocks despite all efforts, the possibility of visceral hypersensitivity should be considered. A trial of symptom modulators such as nortriptyline and lowering of the output amperage below the minimal recommended setting of 5 mA can be undertaken. If these interventions do not work, the device must be turned off for a period of time. Occasionally, replacement of the leads or explantation of the device must be considered.

Lack of effect/persistent symptoms. If a patient presents with lack of improvement after device implantation, a thorough workup should be undertaken to ensure that the device is functioning properly. In the case of abnormal impedance values, an abdominal x-ray study can be performed to rule out lead migration (Figure 5). If no abnormalities are detected, the output of the device can be increased. After adjusting device settings, the patient should be assessed for improvement over at least a 1- to 3-month period. One report suggests that in patients not responding to GES, repositioning the location of the stimulator leads on the stomach can be helpful.²⁹

Outcomes of GES

Study results of investigative GES models in animals and select patients were published in 1997.^30,31 Following these reports, 2 large multicenter studies were conducted to demonstrate the efficacy of GES for the treatment of refractory gastroparesis. The Gastric Electrical Mechanical Stimulation Study (GEMS) was an open-label, multicenter study of 38 patients who received percutaneous and later permanent GES devices.³² Marked reduction in weekly vomiting and nausea was observed at 4 weeks, with a 90% reduction in nausea and vomiting frequency at 11 months. Following this, a second multicenter study (Worldwide Anti-Vomiting Electrical Stimulation Study [WAVES]) involving a double-blind sham stimulation controlled trial with 33 idiopathic and diabetic gastroparesis patients was performed.³³ During the blinded portion of this study, there was a noticeable decrease in vomiting frequency, particularly in the patients with diabetic gastroparesis. Patient preference was for the stimulator ON as compared to OFF. The FDA’s HUD exemption for the Enterra GES device in 2000 was based on these studies.

Four independent double-blind studies of GES have been conducted (Table 1).^33-37 It has been difficult to demonstrate improvement during the double-blind period with gastric stimulation compared to no stimulation. Despite total symptom severity improvement and individual symptom improvements in these studies, a recent meta-analysis demonstrated a summative insignificant difference between the GES ON versus OFF states.³⁸

In contrast to the double-blind studies, numerous open-label studies have demonstrated clinical improvements in patients with diabetic and idiopathic gastroparesis (Table 2),^{32,33,35,36,39-55} leading some to question whether the demonstrable efficacy reflects a placebo effect or regression to the mean. Patients may perceive an operative, aggressive intervention as likely to be effectual in comparison to incremental medication efforts, thus creating a placebo effect. It should also be noted that not all open-label studies have demonstrated improvement with GES. Indeed, Jones et al reported no significant difference in nausea and vomiting at 6-month follow-up, and recommended that physicians exercise caution with GES as a therapeutic strategy given the cost and lack of confirmed demonstrable effect.⁵⁶ Thus, the clinical successes demonstrated in open-label studies must be weighed not only against the lack of unequivocal improvement, but also against the potential deleterious effects of the surgery.

In an open-label study that employed the GCSI to follow symptoms of gastroparesis, 29 patients underwent GES implantation over an 18-month period, with follow-up in 28 patients.⁴⁴ GES resulted in clinical improvement in 50% of patients with refractory gastroparesis. The overall GCSI significantly decreased, with improvement in the nausea/vomiting subscore and the post-prandial fullness subscore, but no improvement in the bloating subscore or abdominal pain. The decrease in GCSI was greater for patients with diabetic versus idiopathic gastroparesis. Patients with the main symptom of nausea/vomiting had a greater improvement than patients with the main symptom of abdominal pain. Patients taking narcotic analgesics at the time of implant had a poorer response compared to patients who were not. In this study, 3 clinical parameters were associated with a favorable clinical response: (1) diabetic rather than idiopathic gastroparesis, (2) nausea/vomiting rather than abdominal pain as the primary symptom, and (3) independence from narcotic analgesics prior to stimulator implantation. Knowledge of these 3 factors may allow improved patient selection for GES.

A large prospective study by Heckert et al detailed marked improvements with GES and the patterns of those improvements.⁵⁵ Nausea, vomiting, loss of appetite, and early satiety improved significantly with stimulator use, with a greater improvement in vomiting in patients with diabetic gastroparesis than in those with the idiopathic form. Although GES improved symptoms in 75% of all patients, patients with diabetes had a post-GES Clinical Patient Grading Assessment score that was statistically higher than the score among patients with idiopathic gastroparesis. This difference is thought to be due to the neuromolecular mechanism of diabetic gastroparesis, where blunting of the enteric nervous system may contribute to symptomatology.

Several studies have demonstrated a clinical response to GES in patients with postsurgical gastroparesis. A study by Oubre et al showed that GES led to weekly vomiting improvements as well as a reduction in total symptom severity score.⁵⁷ A study by McCallum et al further demonstrated improved symptoms, quality of life, nutritional status, and hospitalization requirements.⁵⁸ GES has also been shown to improve gastroparesis symptoms in pediatric populations.^47,59 Thus, although not a direct indication, GES has been shown to be beneficial in various subtypes of gastroparesis.

Additionally, irrespective of gastroparesis type, the improved symptomatology with GES appears to be durable, with one study showing persistent clinical improvements up to 8 years after device placement.⁶⁰ The improvements were persistent and incremental. Likewise, McCallum et al showed that continued reductions in total symptom severity scores were evident in all gastroparesis types up to 10 years after stimulator implantation.⁶¹ The success of the procedures in part comes from careful selection of patients. Clinical parameters that are associated with favorable clinical response include diabetic gastroparesis subtype, nausea/vomiting predominance, and independence from narcotic analgesics prior to stimulator placement.⁶²

GES has also been noted to improve other patient care metrics besides symptomatology, including nutritional status, reduced need for nutritional supplementation, and improved HbA1c.^63-65 Additionally, a study by Cutts et al established that health care resource utilization significantly improved at 12, 24, and 36 months following GES placement, as compared to patients receiving standard medical therapy.⁶⁶ This decreased resource utilization was also reflected in decreased costs in the GES group compared with the standard care group.

Surgical Alternatives to GES

Pyloric interventions such as pyloroplasty and pyloromyotomy are other surgical treatment modalities offered for gastroparesis. Whereas GES uses neurostimulation to facilitate gastric emptying and potentially improve fundic accommodation, pyloric interventions are intended to increase gastric emptying by reducing outflow resistance from the pyloric sphincter.

Pyloric Interventions

Various studies have shown significant improvements with pyloric interventions, similar to the improvements seen with GES. One such study involving 177 patients demonstrated an 86% improvement in gastric emptying, with symptom severity scores for nausea, vomiting, bloating, abdominal pain, and early satiety decreasing significantly at 3 months following pyloroplasty.⁶⁷ A significant advantage of pyloric interventions is that pyloromyotomy can be performed endoscopically (gastric peroral endoscopic pyloromyotomy [G-POEM] or peroral pyloromyotomy [POP]), thus minimizing the risks of open surgery. A recent review that included a pooled analysis of 7 studies of G-POEM for gastroparesis demonstrated 100% technical success, with clinical efficacy in 81.5% of the procedures as assessed by the GCSI.⁶⁸ Additionally, the intraoperative and perioperative complication rates were 6.6% and 7.6%, respectively, suggesting that G-POEM is a safe and clinically beneficial therapeutic option. Few studies comparing the outcomes of pyloric interventions to GES have been performed.

Recently, GES has been combined with pyloric interventions to maximize therapeutic potential. This allows simultaneous neurologic and functional interventions to expedite gastric emptying and improve patient symptomatology. Davis et al demonstrated significant improvement in 21 patients who underwent GES placement and pyloroplasty, with 71% improvement in total symptom severity.⁶⁹ Notably, dual surgery did not increase the incidence of infection or adverse surgical outcomes. Although this study did not directly compare dual surgery to GES alone, the results are nonetheless favorable. GES provides a strong antiemetic and anti-nausea effect, whereas the pyloromyotomy provides improvement in gastric emptying.

Feeding/Venting Tubes

Feeding jejunostomy tubes and venting gastrostomy tubes can be used alone or in combination with GES. Feeding jejunostomy is performed for malnutrition and weight loss that accompanies the refractory symptoms of early satiety, nausea, and vomiting. Venting gastrostomy tubes allow for removal of retained gastric contents that may cause distension, nausea, and vomiting. Gastrojejunostomy tubes can also be placed endoscopically or by interventional radiology.

Gastrectomy

Gastrectomy can provide therapeutic benefit through elimination of the gastric reservoir function and consequent removal of afferent neural impulses. In select patient populations, outcomes of gastrectomy have compared favorably with those of GES. For example, one study demonstrated favorable outcomes of Roux-en-Y gastrectomy in morbidly obese patients with gastroparesis.⁷⁰ In another study, favorable outcomes were reported in a cohort of 103 patients, with gastrectomy demonstrating 87% symptom improvement (nausea, vomiting, epigastric pain) compared to just 63% improvement with GES.⁷¹ However, the dramatic impact on anatomy and physiology and the invasiveness of the procedure need to be weighed against the therapeutic benefit. For example, in the same study, the 30-day morbidity was 23% for gastrectomy versus just 8% for the GES implant.⁷¹

The gastric electrical neurostimulator (Enterra; Medtronic, Inc.) is approved for treatment of idiopathic and diabetic gastroparesis that is refractory to medical treatment, performed on a compassionate basis. Patients with diabetic gastroparesis respond to GES better than do patients with the idiopathic form. Of the symptoms of gastroparesis, primarily nausea and vomiting improve. Thus, GES favors patients with diabetic gastroparesis who have primarily nausea and vomiting, rather than, for instance, patients with idiopathic gastroparesis who have primarily abdominal pain and may be taking narcotics. Some centers provide GES for postsurgical patients and children with gastroparesis.

The 3 main surgical interventions for medically refractory gastroparesis are GES, pyloric intervention (pyloroplasty or pyloromyotomy), and gastrectomy. Of the 3 interventions, gastrectomy is the most radical given its dramatic effect on anatomy and is thus not preferred. The clinical decision then becomes: GES, pyloric intervention, or both? There are limited data to support a definitive answer to this question.

In a single-center retrospective analysis of prospective data (electronic medical record), Arthur et al compared outcomes of GES patients with medically refractory gastroparesis who received various surgical interventions.⁷² In total, 33 stimulator, 7 pyloroplasty, 2 gastrectomy, and 16 combined stimulator and pyloroplasty patients were analyzed for postoperative symptom improvement. Pyloroplasty alone demonstrated the least symptom improvement, combination GES and pyloroplasty demonstrated increased improvement, and GES alone demonstrated the most improvement. The results of this study suggest that barring contraindication, placement of a gastric stimulator as the initial treatment is best, with pyloroplasty reserved for patients who do not achieve adequate symptom control. Limitations of the study include its single-center design and low patient numbers for pyloroplasty in isolation.

In contrast, a recent retrospective systematic review synthesized the outcomes of various studies of GES and pyloric interventions for medically refractory gastroparesis.⁷³ A therapeutic effect was found for each surgical intervention, with pyloric surgery patients demonstrating a greater response to intervention than GES patients. Unfortunately, attempts to analyze combination interventions were hindered by a lack of power.

Conclusion

Initial management of gastroparesis is medical (lifestyle and diet changes), with antiemetic and prokinetic agents used in refractory cases. Following failure of this therapy, placement of a GES device is a surgical intervention that has been approved under FDA humanitarian device exemption to help ameliorate symptomatology. Improvement with GES has been demonstrated in nonblinded studies, but the lack of randomized controlled trials demonstrating benefit suggests the possibility of an underlying placebo effect. Additionally, new medical procedures such as G-POEM complicate the decision of which intervention should be attempted first. More studies, specifically comparing GES, pyloric interventions, and combined GES with pyloric intervention to placebo, are needed to fully understand what therapy is best for refractory gastroparesis.

Corresponding author: Henry P. Parkman, MD, Gastroenterology Section, Temple University School of Medicine, 3401 North Broad Street, Philadelphia, PA 19140; [email protected].

Financial disclosures: None.

Approximately 55 years ago, tricyclic antidepressants (TCAs) began to be used to treat neuropathic pain.¹ Eventually, clinical trials emerged suggesting the utility of TCAs for other chronic pain conditions, such as fibromyalgia (FM) and migraine prophylaxis. However, despite TCAs’ effectiveness in mitigating painful conditions, their adverse effects limited their use.

Pharmacologic advancements have led to the development of other antidepressant classes, including selective serotonin reuptake inhibitors (SSRIs) and serotonin-norepinephrine reuptake inhibitors (SNRIs), and the use of these agents has come to eclipse that of TCAs. In the realm of pain management, such developments have raised the hope of possible alternative co-analgesic agents that could avoid the adverse effects associated with TCAs. Some of these agents have demonstrated efficacy for managing chronic pain states, while others have demonstrated only limited utility.

This article provides a synopsis of systematic reviews and meta-analyses examining the role of antidepressant therapy for managing several chronic pain conditions, including pain associated with neuropathy, FM, headache, and irritable bowel syndrome (IBS). Because the literature base is rapidly evolving, it is necessary to revisit the information gleaned from clinical data with respect to treatment effectiveness, and to clarify how antidepressants might be positioned in the management of chronic pain.

The effectiveness of antidepressants for pain

The pathophysiologic processes that precipitate and maintain chronic pain conditions are complex (Box 1^2-10). The pain-mitigating effects of antidepressants can be thought of in terms of direct analgesic effects and indirect effects (Box 2^{2,3,8,10,11-33}).

Box 1

During the last several decades, anti­depressants have been used to address—and have demonstrated clinical utility for—a variety of chronic pain states. However, antidepressants are not a panacea; some chronic pain conditions are more responsive to antidepressants than are others. The chronic painful states most amenable to antidepressants are those that result primarily from a process of neural sensitization, as opposed to acute somatic or visceral nociception. Hence, several meta-analyses and evidence-based reviews have long suggested the usefulness of antidepressants for mitigating pain associated with neuropathy,^34,35 FM,^36,37 headache,³⁸ and IBS.^39,40

Box 2

Continue to: Neuropathic pain

Several treatment guidelines advocate for the use of antidepressants for neuropathic pain.^41-44 For decades, TCAs have been employed off-label to successfully treat many patients with neuropathic pain states. Early investigations suggested that TCAs were robustly efficacious in managing patients with neuropathy.^45-48 Calculated number-needed-to-treat (NNT) values for TCAs were quite low (ie, reflecting that few patients would need to be treated to yield a positive response in one patient compared with placebo), and were comparable to, if not slightly better than, the NNTs generated for anticonvulsants and α2-δ ligands, such as gabapentin or pregabalin.^45-48

Unfortunately, early studies involving TCAs conducted many years ago do not meet contemporary standards of methodological rigor; they featured relatively small samples of patients assessed for brief post-treatment intervals with variable outcome measures. Thus, the NNT values obtained in meta-analyses based on these studies may overestimate treatment benefits.⁴⁹ Further, NNT values derived from meta-analyses tended to combine all drugs within a particular antidepressant class (eg, amitriptyline, nortriptyline, desipramine, and imipramine among the TCAs) employed at diverse doses. Taken together, these limitations raise questions about the results of those meta-analyses.

Subsequent meta-analyses, which employed strict criteria to eliminate data from studies with potential sources of bias and used a primary outcome of frequencies of patients reporting at least 30% pain reduction compared with a placebo-controlled sample, suggest that the effectiveness of TCAs as a class for treating neuropathic pain is not as compelling as once was thought. Meta-analyses of studies employing specific TCAs revealed that there was little evidence to support the use of desipramine,⁵⁰ imipramine,⁵¹ or nortriptyline⁵² in managing diabetic neuropathy or postherpetic neuralgia. Studies evaluating amitriptyline (dose range 12.5 to 150 mg/d), found low-level evidence of effectiveness; the benefit was expected to be present for a small subset (approximately 25%) of patients with neuropathic pain.⁵³

There is moderate-quality evidence that duloxetine (60 to 120 mg/d) can produce a ≥50% improvement in pain severity ratings among patients with diabetic peripheral neuropathy.⁵⁴ Although head-to-head studies with other antidepressants are limited, it appears that duloxetine and amitriptyline have comparable efficacy, even though the NNTs for amitriptyline were derived from lower-quality studies than those for duloxetine. Duloxetine is the only antidepressant to receive FDA approval for managing diabetic neuropathy. By contrast, studies assessing the utility of venlafaxine in neuropathic pain comprised small samples for brief durations, which limits the ability to draw clear (unbiased) support for its usefulness.⁵⁵

Given the diversity of pathophysiologic processes underlying the disturbances that cause neuropathic pain disorders, it is unsurprising that the effectiveness of amitriptyline and duloxetine were not generalizable to all neuropathic pain states. Although amitriptyline produced pain-mitigating effects in patients with diabetic neuropathy and post-herpetic neuralgia, and duloxetine mitigated pain among patients with diabetic neuropathy, there was no evidence to suggest their effectiveness in phantom limb pain or human immunodeficiency virus-related and spinal cord-related neuropathies.^{35,53,54,56-58}

Continue to: Fibromyalgia

As with the issues encountered in interpreting the effectiveness of antidepressants in neuropathic pain, interpreting results gleaned from clinical trials of antidepressants for treating FM are fraught with similar difficulties. Although amitriptyline has been a first-line treatment for FM for many years, the evidence upon which such recommendations were based consisted of low-level studies that had a significant potential for bias.⁵⁹ Large randomized trials would offer more compelling data regarding the efficacy of amitriptyline, but the prohibitive costs of such studies makes it unlikely they will be conducted. Amitriptyline (25 to 50 mg/d) was effective in mitigating FM-related pain in a small percentage of patients studied, with an estimated NNT of 4.⁵⁹ Adverse effects, often contributing to treatment discontinuation, were encountered more frequently among patients who received amitriptyline compared with placebo.

Selective serotonin reuptake inhibitors failed to demonstrate significant pain relief (estimated NNT of 10), or improvement in fatigue or sleep problems, even though the studies upon which such conclusions were based were low-level studies with a high potential for bias.⁶⁰ Although SSRIs have limited utility for mitigating pain, they are still quite useful for reducing depression among patients with FM.⁶⁰

By contrast, the SNRIs duloxetine and milnacipran provided clinically relevant benefit over placebo in the frequency of patients reporting pain relief of ≥30%, as well as patients’ global impression of change.⁶¹ These agents, however, failed to provide clinically relevant benefit over placebo in improving health-related quality of life, reducing sleep problems, or improving fatigue. Nonetheless, duloxetine and milnacipran are FDA-approved for managing pain in FM. Studies assessing the efficacy of venlafaxine in the treatment of FM to date have been limited by small sample sizes, inconsistent dosing, lack of a placebo control, and lack of blinding, which limits the ability to clearly delineate the role of venlafaxine in managing FM.⁶²

Mirtazapine (15 to 45 mg/d) showed a clinically relevant benefit compared with placebo for participant-reported pain relief of ≥30% and sleep disturbances. There was no benefit in terms of participant-reported improvement of quality of life, fatigue, or negative mood.⁶³ The evidence was considered to be of low quality overall.

Headache

Amitriptyline has been employed off-label to address headache prophylaxis since 1964.⁶⁴ Compared with placebo, it is efficacious in ameliorating migraine frequency and intensity as well as the frequency of tension headache.^65,66 However, SSRIs and SNRIs (venlafaxine) failed to produce significant reductions in migraine frequency or severity or the frequencies of tension headache when compared with placebo.^67,68

Continue to: Irritable bowel syndrome

Early studies addressing antidepressant efficacy in IBS reveal inconsistencies. For example, whereas some suggest that TCAs are effective in mitigating chronic, severe abdominal pain,^39,40 others concluded that TCAs failed to demonstrate a significant analgesic benefit.⁶⁹ A recent meta-analysis that restricted analysis of efficacy to randomized controlled trials (RCTs) with more rigorous methodological adherence found that pain relief in IBS is possible with both TCAs as well as SSRIs. However, adverse effects were more commonly encountered with TCAs than with SSRIs. Some of the inconsistencies in treatment efficacy reported in early studies may be due to variations in responsiveness of subsets of IBS patients. Specifically, the utility of TCAs appears to be best among patients with diarrheal-type (as opposed to constipation-type) IBS, presumably due to TCAs’ anticholinergic effects, whereas SSRIs may provide more of a benefit for patients with predominantly constipation-type IBS.^40,70

Other chronic pain conditions

Antidepressants have been used to assist in the management of several other pain conditions, including oral-facial pain, interstitial cystitis, non-cardiac chest pain, and others. The role of antidepressants for such conditions remains unclear due to limitations in the prevailing empirical work, such as few trials, small sample sizes, variations in outcome measures, and insufficient randomization and blinding.^71-76 The interpretation of results from systematic reviews and meta-analyses is limited because of these shortcomings.⁷⁷ Hence, it has not always been possible to determine whether, and to what extent, patients with such conditions may benefit from antidepressants.

Neuromodulatory effects and efficacy for pain

The interplay of norepinephrine (NE) and serotonin (5-HT) neurotransmitter systems and cellular mechanisms involved in the descending modulation of pain pathways is complex. Experimental animal models of pain modulation suggest that 5-HT can both inhibit as well as promote pain perception by different physiological mechanisms, in contrast to NE, which is predominately inhibitory. While 5-HT in the descending modulating system can inhibit pain transmission ascending to the brain from the periphery, it appears that an intact noradrenergic system is necessary for the inhibitory influences of the serotonergic system to be appreciated.^16,78,79 Deficiencies in one or both of these neurotransmitter systems may contribute to hyperactive pain processing, and thereby precipitate or maintain chronic pain.

Pain mitigation may be achieved best by enhancing both neurotransmitters simultaneously, less so by enhancing NE alone, and least by enhancing 5-HT alone.⁶ The ability to impact pain modulation would, therefore, depend on the degree to which an antidepressant capitalizes on both noradrenergic and serotonergic neurotransmission. Antidepressants commonly employed to manage pain are presented in Table 1^{47,60,68,80-88} according to their primary neurotransmitter effects. Thus, the literature summarized above suggests that antidepressants that influence both NE and 5-HT transmission have greater analgesic effects than antidepressants with more specific effects, such as influencing 5-HT reuptake alone.^80-85 It is unsurprising, therefore, that the SSRIs have not been demonstrated to be as consistently analgesic.^{47,60,68,80,86-88}

Similarly, pharmacodynamic and pharmacokinetic differences within antidepressant classes may influence analgesic effectiveness. Simultaneous effects on NE and 5-HT are achieved at low doses with duloxetine and milnacipran. By contrast, 5-HT effects predominate at low doses for venlafaxine. To achieve pain-mitigating effects, higher doses of venlafaxine generally are required.⁸⁹ Therefore, inconsistencies across studies regarding the analgesic benefits of venlafaxine may be attributable to variability in dosing; patients treated with lower doses may not have experienced sufficient NE effects to garner positive results.

Continue to: The differences in analgesic efficacy...

The differences in analgesic efficacy among specific TCAs may be understood in a similar fashion. Specifically, tertiary TCAs (imipramine and amitriptyline) inhibit both 5-HT and NE reuptake.^6,90 Secondary amines (desipramine and nortriptyline) predominantly impact NE reuptake, possibly accounting for the lesser pain-mitigating benefit achieved with these agents, such as for treating neuro­pathic pain. Further, in vivo imipramine and amitriptyline are rapidly metabolized to secondary amines that are potent and selective NE reuptake inhibitors. In this way, the secondary amines may substantially lose the ability to modulate pain transmission because of the loss of concurrent 5-HT influences.⁹⁰

Clinical pearls

The following practical points can help guide clinicians regarding the usefulness of antidepressants for pain management:

• Antidepressants can alleviate symptoms of depression and pain. The pain-mitigating effects of antidepressants are possible even among chronic pain patients who are not depressed. Antidepressants may confer benefits for chronic pain patients with depression and other comorbid conditions, such as somatic symptom and related disorders.
• Antidepressants are useful for select chronic pain states. Although the noradrenergic and serotonergic antidepressants (SNRIs and, to some extent, amitriptyline) appear to have efficacy for neuropathic pain and FM, the benefits of SSRIs appear to be less robust. On the other hand, SSRIs and TCAs may have potential benefit for patients with IBS. However, the results of meta-analyses are limited in the ability to provide information about which patients will best respond to which specific antidepressant or how well. Future research directed at identifying characteristics that can predict which patients are likely to benefit from one antidepressant vs another would help inform how best to tailor treatment to individual needs.
• The pain-mitigating effects of antidepressants often emerge early in the course of treatment (often before mood-elevating effects are observed). For example, in the case of amitriptyline, pain relief may be possible for some patients at doses generally lower than those required for mood-elevating effects. To date, there is limited information in the literature to determine what constitutes a sufficient duration of treatment, or when treatment should be modified.
• Failure to reduce pain should raise questions about whether the dose should be increased, an alternative agent should be tried, or combinations with other analgesic agents should be considered. Failure to achieve pain-mitigating effects with one antidepressant does not mean failure with others. Hence, failure to achieve desired effects with one agent might warrant an empirical trial with another agent. Presently, too few double-blind RCTs have been conducted to assess the pain-mitigating effects of other antidepressants (eg, bupropion and newer SNRIs such as desvenlafaxine and levomilnacipran). Meta-analysis of the analgesic effectiveness of these agents or comparisons to the efficacy of other antidepressant classes is, therefore, impossible at this time.

Because many chronic pain states are complex, patients will seldom experience clinically relevant benefit from any one intervention.⁵³ The bigger implication for clinical research is to determine whether there is a sequence or combination of medication use that will provide overall better clinical effectiveness.⁵³ Only limited data are available exploring the utility of combining pharmacologic approaches to address pain.⁹¹ For example, preliminary evidence suggests that combinations of complementary strategies, such as duloxetine combined with pregabalin, may result in significantly greater numbers of FM patients achieving ≥30% pain reduction compared with monotherapy with either agent alone or placebo.⁹²

• Antidepressant selection may need to be based on medication-related adverse effect profiles and the potential for drug interactions. These factors are useful to consider in delineating multimodal treatment regimens for chronic pain in light of patients’ comorbidities and co-medication regimen. For example, the adverse effects of TCAs (anticholinergic and alpha-adrenergic influences) limit their utility for treating pain. Some of these effects can be more problematic in select populations, such as older adults or those with orthostatic difficulties, among others. TCAs are contraindicated in patients with closed-angle glaucoma, recent myocardial infarction, cardiac arrhythmias, poorly controlled seizures, or severe benign prostatic hypertrophy. Although the pain-mitigating effects of SNRIs have not been demonstrated to significantly exceed those of TCAs,^68,93,94 SNRIs would offer an advantage of greater tolerability of adverse effects and relative safety in patients with comorbid medical conditions that would otherwise preclude TCA use. The adverse effects and common drug interactions associated with antidepressants are summarized in Table 2⁹⁵.

Conclusion

Chronic, nonmalignant pain conditions afflict many patients and significantly impair their ability to function. Because of heightened concerns related to the appropriateness of, and restricting inordinate access to, long-term opioid analgesics, clinicians need to explore the usefulness of co-analgesic agents, such as antidepressants. Significant comorbidities exist between psychiatric disorders and chronic pain, and psychiatrists are uniquely positioned to diagnose and treat psychiatric comorbidities, as well as pain, among their patients, especially since they understand the kinetics and dynamics of antidepressants.

Bottom Line

Antidepressants can alleviate symptoms of depression and pain. Noradrenergic and serotonergic antidepressants appear to have efficacy for pain associated with neuropathy and fibromyalgia, while selective serotonin reuptake inhibitors and tricyclic antidepressants may have benefit for patients with irritable bowel syndrome. However, evidence regarding which patients will best respond to which specific antidepressant is limited.

Continue to: Related Resources

• Williams AM, Knox ED. When to prescribe antidepressants to treat comorbid depression and pain disorders. Current Psychiatry. 2017;16(1):55-58.
• Maletic V, Demuri B. Chronic pain and depression: treatment of 2 culprits in common. Current Psychiatry. 2016;15(3):41,47-50,52.

Drug Brand Names

Amitriptyline • Elavil, Endep
Bupropion • Wellbutrin, Zyban
Carisoprodol • Rela, Soma
Cyclobenzaprine • Amrix, Flexeril
Desipramine • Norpramin
Desvenlafaxine • Pristiq
Duloxetine • Cymbalta
Fluoxetine • Prozac
Gabapentin • Horizant, Neurontin
Imipramine • Tofranil
Levomilnacipran • Fetzima
Methadone • Dolophine, Methadose
Milnacipran • Savella
Mirtazapine • Remeron
Nortriptyline • Pamelor
Paroxetine • Paxil
Pregabalin • Lyrica, Lyrica CR
Tapentadol • Nucynta
Tramadol • Ultram
Trazodone • Desyrel, Oleptro
Venlafaxine • Effexor
Warfarin • Coumadin, Jantoven

Approximately 55 years ago, tricyclic antidepressants (TCAs) began to be used to treat neuropathic pain.¹ Eventually, clinical trials emerged suggesting the utility of TCAs for other chronic pain conditions, such as fibromyalgia (FM) and migraine prophylaxis. However, despite TCAs’ effectiveness in mitigating painful conditions, their adverse effects limited their use.

Pharmacologic advancements have led to the development of other antidepressant classes, including selective serotonin reuptake inhibitors (SSRIs) and serotonin-norepinephrine reuptake inhibitors (SNRIs), and the use of these agents has come to eclipse that of TCAs. In the realm of pain management, such developments have raised the hope of possible alternative co-analgesic agents that could avoid the adverse effects associated with TCAs. Some of these agents have demonstrated efficacy for managing chronic pain states, while others have demonstrated only limited utility.

This article provides a synopsis of systematic reviews and meta-analyses examining the role of antidepressant therapy for managing several chronic pain conditions, including pain associated with neuropathy, FM, headache, and irritable bowel syndrome (IBS). Because the literature base is rapidly evolving, it is necessary to revisit the information gleaned from clinical data with respect to treatment effectiveness, and to clarify how antidepressants might be positioned in the management of chronic pain.

The effectiveness of antidepressants for pain

The pathophysiologic processes that precipitate and maintain chronic pain conditions are complex (Box 1^2-10). The pain-mitigating effects of antidepressants can be thought of in terms of direct analgesic effects and indirect effects (Box 2^{2,3,8,10,11-33}).

Box 1

During the last several decades, anti­depressants have been used to address—and have demonstrated clinical utility for—a variety of chronic pain states. However, antidepressants are not a panacea; some chronic pain conditions are more responsive to antidepressants than are others. The chronic painful states most amenable to antidepressants are those that result primarily from a process of neural sensitization, as opposed to acute somatic or visceral nociception. Hence, several meta-analyses and evidence-based reviews have long suggested the usefulness of antidepressants for mitigating pain associated with neuropathy,^34,35 FM,^36,37 headache,³⁸ and IBS.^39,40

Box 2

Continue to: Neuropathic pain

Several treatment guidelines advocate for the use of antidepressants for neuropathic pain.^41-44 For decades, TCAs have been employed off-label to successfully treat many patients with neuropathic pain states. Early investigations suggested that TCAs were robustly efficacious in managing patients with neuropathy.^45-48 Calculated number-needed-to-treat (NNT) values for TCAs were quite low (ie, reflecting that few patients would need to be treated to yield a positive response in one patient compared with placebo), and were comparable to, if not slightly better than, the NNTs generated for anticonvulsants and α2-δ ligands, such as gabapentin or pregabalin.^45-48

Unfortunately, early studies involving TCAs conducted many years ago do not meet contemporary standards of methodological rigor; they featured relatively small samples of patients assessed for brief post-treatment intervals with variable outcome measures. Thus, the NNT values obtained in meta-analyses based on these studies may overestimate treatment benefits.⁴⁹ Further, NNT values derived from meta-analyses tended to combine all drugs within a particular antidepressant class (eg, amitriptyline, nortriptyline, desipramine, and imipramine among the TCAs) employed at diverse doses. Taken together, these limitations raise questions about the results of those meta-analyses.

Subsequent meta-analyses, which employed strict criteria to eliminate data from studies with potential sources of bias and used a primary outcome of frequencies of patients reporting at least 30% pain reduction compared with a placebo-controlled sample, suggest that the effectiveness of TCAs as a class for treating neuropathic pain is not as compelling as once was thought. Meta-analyses of studies employing specific TCAs revealed that there was little evidence to support the use of desipramine,⁵⁰ imipramine,⁵¹ or nortriptyline⁵² in managing diabetic neuropathy or postherpetic neuralgia. Studies evaluating amitriptyline (dose range 12.5 to 150 mg/d), found low-level evidence of effectiveness; the benefit was expected to be present for a small subset (approximately 25%) of patients with neuropathic pain.⁵³

There is moderate-quality evidence that duloxetine (60 to 120 mg/d) can produce a ≥50% improvement in pain severity ratings among patients with diabetic peripheral neuropathy.⁵⁴ Although head-to-head studies with other antidepressants are limited, it appears that duloxetine and amitriptyline have comparable efficacy, even though the NNTs for amitriptyline were derived from lower-quality studies than those for duloxetine. Duloxetine is the only antidepressant to receive FDA approval for managing diabetic neuropathy. By contrast, studies assessing the utility of venlafaxine in neuropathic pain comprised small samples for brief durations, which limits the ability to draw clear (unbiased) support for its usefulness.⁵⁵

Given the diversity of pathophysiologic processes underlying the disturbances that cause neuropathic pain disorders, it is unsurprising that the effectiveness of amitriptyline and duloxetine were not generalizable to all neuropathic pain states. Although amitriptyline produced pain-mitigating effects in patients with diabetic neuropathy and post-herpetic neuralgia, and duloxetine mitigated pain among patients with diabetic neuropathy, there was no evidence to suggest their effectiveness in phantom limb pain or human immunodeficiency virus-related and spinal cord-related neuropathies.^{35,53,54,56-58}

Continue to: Fibromyalgia

As with the issues encountered in interpreting the effectiveness of antidepressants in neuropathic pain, interpreting results gleaned from clinical trials of antidepressants for treating FM are fraught with similar difficulties. Although amitriptyline has been a first-line treatment for FM for many years, the evidence upon which such recommendations were based consisted of low-level studies that had a significant potential for bias.⁵⁹ Large randomized trials would offer more compelling data regarding the efficacy of amitriptyline, but the prohibitive costs of such studies makes it unlikely they will be conducted. Amitriptyline (25 to 50 mg/d) was effective in mitigating FM-related pain in a small percentage of patients studied, with an estimated NNT of 4.⁵⁹ Adverse effects, often contributing to treatment discontinuation, were encountered more frequently among patients who received amitriptyline compared with placebo.

Selective serotonin reuptake inhibitors failed to demonstrate significant pain relief (estimated NNT of 10), or improvement in fatigue or sleep problems, even though the studies upon which such conclusions were based were low-level studies with a high potential for bias.⁶⁰ Although SSRIs have limited utility for mitigating pain, they are still quite useful for reducing depression among patients with FM.⁶⁰

By contrast, the SNRIs duloxetine and milnacipran provided clinically relevant benefit over placebo in the frequency of patients reporting pain relief of ≥30%, as well as patients’ global impression of change.⁶¹ These agents, however, failed to provide clinically relevant benefit over placebo in improving health-related quality of life, reducing sleep problems, or improving fatigue. Nonetheless, duloxetine and milnacipran are FDA-approved for managing pain in FM. Studies assessing the efficacy of venlafaxine in the treatment of FM to date have been limited by small sample sizes, inconsistent dosing, lack of a placebo control, and lack of blinding, which limits the ability to clearly delineate the role of venlafaxine in managing FM.⁶²

Mirtazapine (15 to 45 mg/d) showed a clinically relevant benefit compared with placebo for participant-reported pain relief of ≥30% and sleep disturbances. There was no benefit in terms of participant-reported improvement of quality of life, fatigue, or negative mood.⁶³ The evidence was considered to be of low quality overall.

Headache

Amitriptyline has been employed off-label to address headache prophylaxis since 1964.⁶⁴ Compared with placebo, it is efficacious in ameliorating migraine frequency and intensity as well as the frequency of tension headache.^65,66 However, SSRIs and SNRIs (venlafaxine) failed to produce significant reductions in migraine frequency or severity or the frequencies of tension headache when compared with placebo.^67,68

Continue to: Irritable bowel syndrome

Early studies addressing antidepressant efficacy in IBS reveal inconsistencies. For example, whereas some suggest that TCAs are effective in mitigating chronic, severe abdominal pain,^39,40 others concluded that TCAs failed to demonstrate a significant analgesic benefit.⁶⁹ A recent meta-analysis that restricted analysis of efficacy to randomized controlled trials (RCTs) with more rigorous methodological adherence found that pain relief in IBS is possible with both TCAs as well as SSRIs. However, adverse effects were more commonly encountered with TCAs than with SSRIs. Some of the inconsistencies in treatment efficacy reported in early studies may be due to variations in responsiveness of subsets of IBS patients. Specifically, the utility of TCAs appears to be best among patients with diarrheal-type (as opposed to constipation-type) IBS, presumably due to TCAs’ anticholinergic effects, whereas SSRIs may provide more of a benefit for patients with predominantly constipation-type IBS.^40,70

Other chronic pain conditions

Antidepressants have been used to assist in the management of several other pain conditions, including oral-facial pain, interstitial cystitis, non-cardiac chest pain, and others. The role of antidepressants for such conditions remains unclear due to limitations in the prevailing empirical work, such as few trials, small sample sizes, variations in outcome measures, and insufficient randomization and blinding.^71-76 The interpretation of results from systematic reviews and meta-analyses is limited because of these shortcomings.⁷⁷ Hence, it has not always been possible to determine whether, and to what extent, patients with such conditions may benefit from antidepressants.

Neuromodulatory effects and efficacy for pain

The interplay of norepinephrine (NE) and serotonin (5-HT) neurotransmitter systems and cellular mechanisms involved in the descending modulation of pain pathways is complex. Experimental animal models of pain modulation suggest that 5-HT can both inhibit as well as promote pain perception by different physiological mechanisms, in contrast to NE, which is predominately inhibitory. While 5-HT in the descending modulating system can inhibit pain transmission ascending to the brain from the periphery, it appears that an intact noradrenergic system is necessary for the inhibitory influences of the serotonergic system to be appreciated.^16,78,79 Deficiencies in one or both of these neurotransmitter systems may contribute to hyperactive pain processing, and thereby precipitate or maintain chronic pain.

Pain mitigation may be achieved best by enhancing both neurotransmitters simultaneously, less so by enhancing NE alone, and least by enhancing 5-HT alone.⁶ The ability to impact pain modulation would, therefore, depend on the degree to which an antidepressant capitalizes on both noradrenergic and serotonergic neurotransmission. Antidepressants commonly employed to manage pain are presented in Table 1^{47,60,68,80-88} according to their primary neurotransmitter effects. Thus, the literature summarized above suggests that antidepressants that influence both NE and 5-HT transmission have greater analgesic effects than antidepressants with more specific effects, such as influencing 5-HT reuptake alone.^80-85 It is unsurprising, therefore, that the SSRIs have not been demonstrated to be as consistently analgesic.^{47,60,68,80,86-88}

Similarly, pharmacodynamic and pharmacokinetic differences within antidepressant classes may influence analgesic effectiveness. Simultaneous effects on NE and 5-HT are achieved at low doses with duloxetine and milnacipran. By contrast, 5-HT effects predominate at low doses for venlafaxine. To achieve pain-mitigating effects, higher doses of venlafaxine generally are required.⁸⁹ Therefore, inconsistencies across studies regarding the analgesic benefits of venlafaxine may be attributable to variability in dosing; patients treated with lower doses may not have experienced sufficient NE effects to garner positive results.

Continue to: The differences in analgesic efficacy...

The differences in analgesic efficacy among specific TCAs may be understood in a similar fashion. Specifically, tertiary TCAs (imipramine and amitriptyline) inhibit both 5-HT and NE reuptake.^6,90 Secondary amines (desipramine and nortriptyline) predominantly impact NE reuptake, possibly accounting for the lesser pain-mitigating benefit achieved with these agents, such as for treating neuro­pathic pain. Further, in vivo imipramine and amitriptyline are rapidly metabolized to secondary amines that are potent and selective NE reuptake inhibitors. In this way, the secondary amines may substantially lose the ability to modulate pain transmission because of the loss of concurrent 5-HT influences.⁹⁰

Clinical pearls

The following practical points can help guide clinicians regarding the usefulness of antidepressants for pain management:

• Antidepressants can alleviate symptoms of depression and pain. The pain-mitigating effects of antidepressants are possible even among chronic pain patients who are not depressed. Antidepressants may confer benefits for chronic pain patients with depression and other comorbid conditions, such as somatic symptom and related disorders.
• Antidepressants are useful for select chronic pain states. Although the noradrenergic and serotonergic antidepressants (SNRIs and, to some extent, amitriptyline) appear to have efficacy for neuropathic pain and FM, the benefits of SSRIs appear to be less robust. On the other hand, SSRIs and TCAs may have potential benefit for patients with IBS. However, the results of meta-analyses are limited in the ability to provide information about which patients will best respond to which specific antidepressant or how well. Future research directed at identifying characteristics that can predict which patients are likely to benefit from one antidepressant vs another would help inform how best to tailor treatment to individual needs.
• The pain-mitigating effects of antidepressants often emerge early in the course of treatment (often before mood-elevating effects are observed). For example, in the case of amitriptyline, pain relief may be possible for some patients at doses generally lower than those required for mood-elevating effects. To date, there is limited information in the literature to determine what constitutes a sufficient duration of treatment, or when treatment should be modified.
• Failure to reduce pain should raise questions about whether the dose should be increased, an alternative agent should be tried, or combinations with other analgesic agents should be considered. Failure to achieve pain-mitigating effects with one antidepressant does not mean failure with others. Hence, failure to achieve desired effects with one agent might warrant an empirical trial with another agent. Presently, too few double-blind RCTs have been conducted to assess the pain-mitigating effects of other antidepressants (eg, bupropion and newer SNRIs such as desvenlafaxine and levomilnacipran). Meta-analysis of the analgesic effectiveness of these agents or comparisons to the efficacy of other antidepressant classes is, therefore, impossible at this time.

Because many chronic pain states are complex, patients will seldom experience clinically relevant benefit from any one intervention.⁵³ The bigger implication for clinical research is to determine whether there is a sequence or combination of medication use that will provide overall better clinical effectiveness.⁵³ Only limited data are available exploring the utility of combining pharmacologic approaches to address pain.⁹¹ For example, preliminary evidence suggests that combinations of complementary strategies, such as duloxetine combined with pregabalin, may result in significantly greater numbers of FM patients achieving ≥30% pain reduction compared with monotherapy with either agent alone or placebo.⁹²

• Antidepressant selection may need to be based on medication-related adverse effect profiles and the potential for drug interactions. These factors are useful to consider in delineating multimodal treatment regimens for chronic pain in light of patients’ comorbidities and co-medication regimen. For example, the adverse effects of TCAs (anticholinergic and alpha-adrenergic influences) limit their utility for treating pain. Some of these effects can be more problematic in select populations, such as older adults or those with orthostatic difficulties, among others. TCAs are contraindicated in patients with closed-angle glaucoma, recent myocardial infarction, cardiac arrhythmias, poorly controlled seizures, or severe benign prostatic hypertrophy. Although the pain-mitigating effects of SNRIs have not been demonstrated to significantly exceed those of TCAs,^68,93,94 SNRIs would offer an advantage of greater tolerability of adverse effects and relative safety in patients with comorbid medical conditions that would otherwise preclude TCA use. The adverse effects and common drug interactions associated with antidepressants are summarized in Table 2⁹⁵.

Conclusion

Chronic, nonmalignant pain conditions afflict many patients and significantly impair their ability to function. Because of heightened concerns related to the appropriateness of, and restricting inordinate access to, long-term opioid analgesics, clinicians need to explore the usefulness of co-analgesic agents, such as antidepressants. Significant comorbidities exist between psychiatric disorders and chronic pain, and psychiatrists are uniquely positioned to diagnose and treat psychiatric comorbidities, as well as pain, among their patients, especially since they understand the kinetics and dynamics of antidepressants.

Bottom Line

Antidepressants can alleviate symptoms of depression and pain. Noradrenergic and serotonergic antidepressants appear to have efficacy for pain associated with neuropathy and fibromyalgia, while selective serotonin reuptake inhibitors and tricyclic antidepressants may have benefit for patients with irritable bowel syndrome. However, evidence regarding which patients will best respond to which specific antidepressant is limited.

Continue to: Related Resources

• Williams AM, Knox ED. When to prescribe antidepressants to treat comorbid depression and pain disorders. Current Psychiatry. 2017;16(1):55-58.
• Maletic V, Demuri B. Chronic pain and depression: treatment of 2 culprits in common. Current Psychiatry. 2016;15(3):41,47-50,52.

Drug Brand Names

Amitriptyline • Elavil, Endep
Bupropion • Wellbutrin, Zyban
Carisoprodol • Rela, Soma
Cyclobenzaprine • Amrix, Flexeril
Desipramine • Norpramin
Desvenlafaxine • Pristiq
Duloxetine • Cymbalta
Fluoxetine • Prozac
Gabapentin • Horizant, Neurontin
Imipramine • Tofranil
Levomilnacipran • Fetzima
Methadone • Dolophine, Methadose
Milnacipran • Savella
Mirtazapine • Remeron
Nortriptyline • Pamelor
Paroxetine • Paxil
Pregabalin • Lyrica, Lyrica CR
Tapentadol • Nucynta
Tramadol • Ultram
Trazodone • Desyrel, Oleptro
Venlafaxine • Effexor
Warfarin • Coumadin, Jantoven

Approximately 55 years ago, tricyclic antidepressants (TCAs) began to be used to treat neuropathic pain.¹ Eventually, clinical trials emerged suggesting the utility of TCAs for other chronic pain conditions, such as fibromyalgia (FM) and migraine prophylaxis. However, despite TCAs’ effectiveness in mitigating painful conditions, their adverse effects limited their use.

Pharmacologic advancements have led to the development of other antidepressant classes, including selective serotonin reuptake inhibitors (SSRIs) and serotonin-norepinephrine reuptake inhibitors (SNRIs), and the use of these agents has come to eclipse that of TCAs. In the realm of pain management, such developments have raised the hope of possible alternative co-analgesic agents that could avoid the adverse effects associated with TCAs. Some of these agents have demonstrated efficacy for managing chronic pain states, while others have demonstrated only limited utility.

This article provides a synopsis of systematic reviews and meta-analyses examining the role of antidepressant therapy for managing several chronic pain conditions, including pain associated with neuropathy, FM, headache, and irritable bowel syndrome (IBS). Because the literature base is rapidly evolving, it is necessary to revisit the information gleaned from clinical data with respect to treatment effectiveness, and to clarify how antidepressants might be positioned in the management of chronic pain.

The effectiveness of antidepressants for pain

The pathophysiologic processes that precipitate and maintain chronic pain conditions are complex (Box 1^2-10). The pain-mitigating effects of antidepressants can be thought of in terms of direct analgesic effects and indirect effects (Box 2^{2,3,8,10,11-33}).

Box 1

During the last several decades, anti­depressants have been used to address—and have demonstrated clinical utility for—a variety of chronic pain states. However, antidepressants are not a panacea; some chronic pain conditions are more responsive to antidepressants than are others. The chronic painful states most amenable to antidepressants are those that result primarily from a process of neural sensitization, as opposed to acute somatic or visceral nociception. Hence, several meta-analyses and evidence-based reviews have long suggested the usefulness of antidepressants for mitigating pain associated with neuropathy,^34,35 FM,^36,37 headache,³⁸ and IBS.^39,40

Box 2

Continue to: Neuropathic pain

Several treatment guidelines advocate for the use of antidepressants for neuropathic pain.^41-44 For decades, TCAs have been employed off-label to successfully treat many patients with neuropathic pain states. Early investigations suggested that TCAs were robustly efficacious in managing patients with neuropathy.^45-48 Calculated number-needed-to-treat (NNT) values for TCAs were quite low (ie, reflecting that few patients would need to be treated to yield a positive response in one patient compared with placebo), and were comparable to, if not slightly better than, the NNTs generated for anticonvulsants and α2-δ ligands, such as gabapentin or pregabalin.^45-48

Unfortunately, early studies involving TCAs conducted many years ago do not meet contemporary standards of methodological rigor; they featured relatively small samples of patients assessed for brief post-treatment intervals with variable outcome measures. Thus, the NNT values obtained in meta-analyses based on these studies may overestimate treatment benefits.⁴⁹ Further, NNT values derived from meta-analyses tended to combine all drugs within a particular antidepressant class (eg, amitriptyline, nortriptyline, desipramine, and imipramine among the TCAs) employed at diverse doses. Taken together, these limitations raise questions about the results of those meta-analyses.

Subsequent meta-analyses, which employed strict criteria to eliminate data from studies with potential sources of bias and used a primary outcome of frequencies of patients reporting at least 30% pain reduction compared with a placebo-controlled sample, suggest that the effectiveness of TCAs as a class for treating neuropathic pain is not as compelling as once was thought. Meta-analyses of studies employing specific TCAs revealed that there was little evidence to support the use of desipramine,⁵⁰ imipramine,⁵¹ or nortriptyline⁵² in managing diabetic neuropathy or postherpetic neuralgia. Studies evaluating amitriptyline (dose range 12.5 to 150 mg/d), found low-level evidence of effectiveness; the benefit was expected to be present for a small subset (approximately 25%) of patients with neuropathic pain.⁵³

There is moderate-quality evidence that duloxetine (60 to 120 mg/d) can produce a ≥50% improvement in pain severity ratings among patients with diabetic peripheral neuropathy.⁵⁴ Although head-to-head studies with other antidepressants are limited, it appears that duloxetine and amitriptyline have comparable efficacy, even though the NNTs for amitriptyline were derived from lower-quality studies than those for duloxetine. Duloxetine is the only antidepressant to receive FDA approval for managing diabetic neuropathy. By contrast, studies assessing the utility of venlafaxine in neuropathic pain comprised small samples for brief durations, which limits the ability to draw clear (unbiased) support for its usefulness.⁵⁵

Given the diversity of pathophysiologic processes underlying the disturbances that cause neuropathic pain disorders, it is unsurprising that the effectiveness of amitriptyline and duloxetine were not generalizable to all neuropathic pain states. Although amitriptyline produced pain-mitigating effects in patients with diabetic neuropathy and post-herpetic neuralgia, and duloxetine mitigated pain among patients with diabetic neuropathy, there was no evidence to suggest their effectiveness in phantom limb pain or human immunodeficiency virus-related and spinal cord-related neuropathies.^{35,53,54,56-58}

Continue to: Fibromyalgia

As with the issues encountered in interpreting the effectiveness of antidepressants in neuropathic pain, interpreting results gleaned from clinical trials of antidepressants for treating FM are fraught with similar difficulties. Although amitriptyline has been a first-line treatment for FM for many years, the evidence upon which such recommendations were based consisted of low-level studies that had a significant potential for bias.⁵⁹ Large randomized trials would offer more compelling data regarding the efficacy of amitriptyline, but the prohibitive costs of such studies makes it unlikely they will be conducted. Amitriptyline (25 to 50 mg/d) was effective in mitigating FM-related pain in a small percentage of patients studied, with an estimated NNT of 4.⁵⁹ Adverse effects, often contributing to treatment discontinuation, were encountered more frequently among patients who received amitriptyline compared with placebo.

Selective serotonin reuptake inhibitors failed to demonstrate significant pain relief (estimated NNT of 10), or improvement in fatigue or sleep problems, even though the studies upon which such conclusions were based were low-level studies with a high potential for bias.⁶⁰ Although SSRIs have limited utility for mitigating pain, they are still quite useful for reducing depression among patients with FM.⁶⁰

By contrast, the SNRIs duloxetine and milnacipran provided clinically relevant benefit over placebo in the frequency of patients reporting pain relief of ≥30%, as well as patients’ global impression of change.⁶¹ These agents, however, failed to provide clinically relevant benefit over placebo in improving health-related quality of life, reducing sleep problems, or improving fatigue. Nonetheless, duloxetine and milnacipran are FDA-approved for managing pain in FM. Studies assessing the efficacy of venlafaxine in the treatment of FM to date have been limited by small sample sizes, inconsistent dosing, lack of a placebo control, and lack of blinding, which limits the ability to clearly delineate the role of venlafaxine in managing FM.⁶²

Mirtazapine (15 to 45 mg/d) showed a clinically relevant benefit compared with placebo for participant-reported pain relief of ≥30% and sleep disturbances. There was no benefit in terms of participant-reported improvement of quality of life, fatigue, or negative mood.⁶³ The evidence was considered to be of low quality overall.

Headache

Amitriptyline has been employed off-label to address headache prophylaxis since 1964.⁶⁴ Compared with placebo, it is efficacious in ameliorating migraine frequency and intensity as well as the frequency of tension headache.^65,66 However, SSRIs and SNRIs (venlafaxine) failed to produce significant reductions in migraine frequency or severity or the frequencies of tension headache when compared with placebo.^67,68

Continue to: Irritable bowel syndrome

Early studies addressing antidepressant efficacy in IBS reveal inconsistencies. For example, whereas some suggest that TCAs are effective in mitigating chronic, severe abdominal pain,^39,40 others concluded that TCAs failed to demonstrate a significant analgesic benefit.⁶⁹ A recent meta-analysis that restricted analysis of efficacy to randomized controlled trials (RCTs) with more rigorous methodological adherence found that pain relief in IBS is possible with both TCAs as well as SSRIs. However, adverse effects were more commonly encountered with TCAs than with SSRIs. Some of the inconsistencies in treatment efficacy reported in early studies may be due to variations in responsiveness of subsets of IBS patients. Specifically, the utility of TCAs appears to be best among patients with diarrheal-type (as opposed to constipation-type) IBS, presumably due to TCAs’ anticholinergic effects, whereas SSRIs may provide more of a benefit for patients with predominantly constipation-type IBS.^40,70

Other chronic pain conditions

Antidepressants have been used to assist in the management of several other pain conditions, including oral-facial pain, interstitial cystitis, non-cardiac chest pain, and others. The role of antidepressants for such conditions remains unclear due to limitations in the prevailing empirical work, such as few trials, small sample sizes, variations in outcome measures, and insufficient randomization and blinding.^71-76 The interpretation of results from systematic reviews and meta-analyses is limited because of these shortcomings.⁷⁷ Hence, it has not always been possible to determine whether, and to what extent, patients with such conditions may benefit from antidepressants.

Neuromodulatory effects and efficacy for pain

The interplay of norepinephrine (NE) and serotonin (5-HT) neurotransmitter systems and cellular mechanisms involved in the descending modulation of pain pathways is complex. Experimental animal models of pain modulation suggest that 5-HT can both inhibit as well as promote pain perception by different physiological mechanisms, in contrast to NE, which is predominately inhibitory. While 5-HT in the descending modulating system can inhibit pain transmission ascending to the brain from the periphery, it appears that an intact noradrenergic system is necessary for the inhibitory influences of the serotonergic system to be appreciated.^16,78,79 Deficiencies in one or both of these neurotransmitter systems may contribute to hyperactive pain processing, and thereby precipitate or maintain chronic pain.

Pain mitigation may be achieved best by enhancing both neurotransmitters simultaneously, less so by enhancing NE alone, and least by enhancing 5-HT alone.⁶ The ability to impact pain modulation would, therefore, depend on the degree to which an antidepressant capitalizes on both noradrenergic and serotonergic neurotransmission. Antidepressants commonly employed to manage pain are presented in Table 1^{47,60,68,80-88} according to their primary neurotransmitter effects. Thus, the literature summarized above suggests that antidepressants that influence both NE and 5-HT transmission have greater analgesic effects than antidepressants with more specific effects, such as influencing 5-HT reuptake alone.^80-85 It is unsurprising, therefore, that the SSRIs have not been demonstrated to be as consistently analgesic.^{47,60,68,80,86-88}

Similarly, pharmacodynamic and pharmacokinetic differences within antidepressant classes may influence analgesic effectiveness. Simultaneous effects on NE and 5-HT are achieved at low doses with duloxetine and milnacipran. By contrast, 5-HT effects predominate at low doses for venlafaxine. To achieve pain-mitigating effects, higher doses of venlafaxine generally are required.⁸⁹ Therefore, inconsistencies across studies regarding the analgesic benefits of venlafaxine may be attributable to variability in dosing; patients treated with lower doses may not have experienced sufficient NE effects to garner positive results.

Continue to: The differences in analgesic efficacy...

The differences in analgesic efficacy among specific TCAs may be understood in a similar fashion. Specifically, tertiary TCAs (imipramine and amitriptyline) inhibit both 5-HT and NE reuptake.^6,90 Secondary amines (desipramine and nortriptyline) predominantly impact NE reuptake, possibly accounting for the lesser pain-mitigating benefit achieved with these agents, such as for treating neuro­pathic pain. Further, in vivo imipramine and amitriptyline are rapidly metabolized to secondary amines that are potent and selective NE reuptake inhibitors. In this way, the secondary amines may substantially lose the ability to modulate pain transmission because of the loss of concurrent 5-HT influences.⁹⁰

Clinical pearls

The following practical points can help guide clinicians regarding the usefulness of antidepressants for pain management:

• Antidepressants can alleviate symptoms of depression and pain. The pain-mitigating effects of antidepressants are possible even among chronic pain patients who are not depressed. Antidepressants may confer benefits for chronic pain patients with depression and other comorbid conditions, such as somatic symptom and related disorders.
• Antidepressants are useful for select chronic pain states. Although the noradrenergic and serotonergic antidepressants (SNRIs and, to some extent, amitriptyline) appear to have efficacy for neuropathic pain and FM, the benefits of SSRIs appear to be less robust. On the other hand, SSRIs and TCAs may have potential benefit for patients with IBS. However, the results of meta-analyses are limited in the ability to provide information about which patients will best respond to which specific antidepressant or how well. Future research directed at identifying characteristics that can predict which patients are likely to benefit from one antidepressant vs another would help inform how best to tailor treatment to individual needs.
• The pain-mitigating effects of antidepressants often emerge early in the course of treatment (often before mood-elevating effects are observed). For example, in the case of amitriptyline, pain relief may be possible for some patients at doses generally lower than those required for mood-elevating effects. To date, there is limited information in the literature to determine what constitutes a sufficient duration of treatment, or when treatment should be modified.
• Failure to reduce pain should raise questions about whether the dose should be increased, an alternative agent should be tried, or combinations with other analgesic agents should be considered. Failure to achieve pain-mitigating effects with one antidepressant does not mean failure with others. Hence, failure to achieve desired effects with one agent might warrant an empirical trial with another agent. Presently, too few double-blind RCTs have been conducted to assess the pain-mitigating effects of other antidepressants (eg, bupropion and newer SNRIs such as desvenlafaxine and levomilnacipran). Meta-analysis of the analgesic effectiveness of these agents or comparisons to the efficacy of other antidepressant classes is, therefore, impossible at this time.

Because many chronic pain states are complex, patients will seldom experience clinically relevant benefit from any one intervention.⁵³ The bigger implication for clinical research is to determine whether there is a sequence or combination of medication use that will provide overall better clinical effectiveness.⁵³ Only limited data are available exploring the utility of combining pharmacologic approaches to address pain.⁹¹ For example, preliminary evidence suggests that combinations of complementary strategies, such as duloxetine combined with pregabalin, may result in significantly greater numbers of FM patients achieving ≥30% pain reduction compared with monotherapy with either agent alone or placebo.⁹²

• Antidepressant selection may need to be based on medication-related adverse effect profiles and the potential for drug interactions. These factors are useful to consider in delineating multimodal treatment regimens for chronic pain in light of patients’ comorbidities and co-medication regimen. For example, the adverse effects of TCAs (anticholinergic and alpha-adrenergic influences) limit their utility for treating pain. Some of these effects can be more problematic in select populations, such as older adults or those with orthostatic difficulties, among others. TCAs are contraindicated in patients with closed-angle glaucoma, recent myocardial infarction, cardiac arrhythmias, poorly controlled seizures, or severe benign prostatic hypertrophy. Although the pain-mitigating effects of SNRIs have not been demonstrated to significantly exceed those of TCAs,^68,93,94 SNRIs would offer an advantage of greater tolerability of adverse effects and relative safety in patients with comorbid medical conditions that would otherwise preclude TCA use. The adverse effects and common drug interactions associated with antidepressants are summarized in Table 2⁹⁵.

Conclusion

Chronic, nonmalignant pain conditions afflict many patients and significantly impair their ability to function. Because of heightened concerns related to the appropriateness of, and restricting inordinate access to, long-term opioid analgesics, clinicians need to explore the usefulness of co-analgesic agents, such as antidepressants. Significant comorbidities exist between psychiatric disorders and chronic pain, and psychiatrists are uniquely positioned to diagnose and treat psychiatric comorbidities, as well as pain, among their patients, especially since they understand the kinetics and dynamics of antidepressants.

Bottom Line

Antidepressants can alleviate symptoms of depression and pain. Noradrenergic and serotonergic antidepressants appear to have efficacy for pain associated with neuropathy and fibromyalgia, while selective serotonin reuptake inhibitors and tricyclic antidepressants may have benefit for patients with irritable bowel syndrome. However, evidence regarding which patients will best respond to which specific antidepressant is limited.

Continue to: Related Resources

• Williams AM, Knox ED. When to prescribe antidepressants to treat comorbid depression and pain disorders. Current Psychiatry. 2017;16(1):55-58.
• Maletic V, Demuri B. Chronic pain and depression: treatment of 2 culprits in common. Current Psychiatry. 2016;15(3):41,47-50,52.

Drug Brand Names

Amitriptyline • Elavil, Endep
Bupropion • Wellbutrin, Zyban
Carisoprodol • Rela, Soma
Cyclobenzaprine • Amrix, Flexeril
Desipramine • Norpramin
Desvenlafaxine • Pristiq
Duloxetine • Cymbalta
Fluoxetine • Prozac
Gabapentin • Horizant, Neurontin
Imipramine • Tofranil
Levomilnacipran • Fetzima
Methadone • Dolophine, Methadose
Milnacipran • Savella
Mirtazapine • Remeron
Nortriptyline • Pamelor
Paroxetine • Paxil
Pregabalin • Lyrica, Lyrica CR
Tapentadol • Nucynta
Tramadol • Ultram
Trazodone • Desyrel, Oleptro
Venlafaxine • Effexor
Warfarin • Coumadin, Jantoven

THE CASE

Ms. E, a 42-year-old woman, visited her new physician for a physical exam. When asked about alcohol intake, she reported that she drank 3 to 4 beers after work and sometimes 5 to 8 beers a day on the weekends. Occasionally, she exceeded those amounts, but she didn’t feel guilty about her drinking. She was often late to work and said her relationship with her boyfriend was strained. A review of systems was positive for fatigue, poor concentration, abdominal pain, and weight gain. Her body mass index was 41, pulse 100 beats/min, blood pressure 125/75 mm Hg, and she was afebrile. Her physical exam was otherwise within normal limits.

How would you proceed with this patient?

Alcohol use disorder (AUD) is a common and often untreated condition that is increasingly prevalent in the United States.¹ The Diagnostic and Statistical Manual of Mental Disorders, 5th Edition (DSM-5) characterizes AUD as a combination of signs and symptoms typifying alcohol abuse and dependence (discussed in a bit).²

Data from the 2015 National Survey on Drug Use and Health (NSDUH) showed 15.7 million Americans with AUD, affecting 6.2% of the population ages 18 years or older and 2.5% of adolescents ages 12 to 17 years.³

Alcohol use and AUD account for an estimated 3.8% of all global deaths and 4.6% of global disability-adjusted life years.⁴ AUD adversely affects several systems (TABLE 1⁵), and patients with AUD are sicker and more likely to die younger than those without AUD.⁴ In the United States, prevalence of AUD has increased in recent years among women, older adults, racial minorities, and individuals with a low education level.⁶

Screening for AUD is reasonable and straightforward, although diagnosis and treatment of AUD in primary care settings may be challenging due to competing clinical priorities; lack of training, resources, and support; and skepticism about the efficacy of behavioral and pharmacologic treatments.^7,8 However, family physicians are in an excellent position to diagnose and help address the complex biopsychosocial needs of patients with AUD, often in collaboration with colleagues and community organizations.

Signs and symptoms of AUD

In clinical practice, at least 2 of the following 11 behaviors or symptoms are required to diagnose AUD²:

• consuming larger amounts of alcohol over a longer period than intended
• persistent desire or unsuccessful efforts to cut down or control alcohol use
• making a significant effort to obtain, use, or recover from alcohol

In moderate-to-severe cases:

• cravings or urges to use alcohol
• recurrent failure to fulfill major work, school, or social obligations
• continued alcohol use despite recurrent social and interpersonal problems
• giving up social, occupational, and recreational activities due to alcohol
• using alcohol in physically dangerous situations
• continued alcohol use despite having physical or psychological problems
• tolerance to alcohol’s effects
• withdrawal symptoms.

Continue to: Patients meet criteria for mild AUD severity if...

Patients meet criteria for mild AUD severity if they exhibit 2 or 3 symptoms, moderate AUD with 4 or 5 symptoms, and severe AUD if there are 6 or more symptoms.²

USPSTF recommends screening for alcohol use disorder with the 10-question AUDIT, 3-question AUDIT-C, or single-question NIAAA tool.

Those who meet criteria for AUD and are able to stop using alcohol are deemed to be in early remission if the criteria have gone unfulfilled for at least 3 months and less than 12 months. Patients are considered to be in sustained remission if they have not met criteria for AUD at any time during a period of 12 months or longer.

How to detect AUD

Several clues in a patient’s history can suggest AUD (TABLE 2^9,10). Most imbibers are unaware of the dangers and may consider themselves merely “social drinkers.” Binge drinking may be an early indicator of vulnerability to AUD and should be assessed as part of a thorough clinical evaluation.¹¹ The US Preventive Services Task Force (USPSTF) recommends (Grade B) that clinicians screen adults ages 18 years or older for alcohol misuse.¹²

Studies demonstrate that both genetic and environmental factors play important roles in the development of AUD.¹³ A family history of excessive alcohol use increases the risk of AUD. Comorbidity of AUD and other mental health conditions is extremely common. For example, high rates of association between major depressive disorder and AUD have been observed.¹⁴

Tools to use in screening and diagnosing AUD

Screening for AUD during an office visit can be done fairly quickly. While 96% of primary care physicians screen for alcohol misuse in some way, only 38% use 1 of the 3 tools recommended by the USPSTF¹⁵—the Alcohol Use Disorders Identification Test (AUDIT), the abbreviated AUDIT-C, or the National Institute on Alcohol Abuse and Alcoholism (NIAAA) single question screen—which detect the full spectrum of alcohol misuse in adults.¹² Although the commonly used CAGE questionnaire is one of the most studied self-report tools, it has lower sensitivity at a lower level of alcohol intake.¹⁶

Continue to: The NIAAA single-question screen asks...

The NIAAA single-question screen asks how many times in the past year the patient had ≥4 drinks (women) or ≥5 drinks (men) in a day.¹⁵ The sensitivity and specificity of single-question screening are 82% to 87% and 61% to 79%, respectively, and the test has been validated in several different settings.¹² The AUDIT screening tool, freely available from the World Health Organization, is a 10-item questionnaire that probes an individual’s alcohol intake, alcohol dependence, and adverse consequences of alcohol use. Administration of the AUDIT typically requires only 2 minutes. AUDIT-C¹⁷ is an abbreviated version of the AUDIT questionnaire that asks 3 consumption questions to screen for AUD.

It was found that AUDIT scores in the range of 8 to 15 indicate a medium-level alcohol problem, whereas a score of ≥16 indicates a high-level alcohol problem. The AUDIT-C is scored from 0 to 12, with ≥4 indicating a problem in men and ≥3 a problem in women.

THE CASE

The physician had used the NIAAA single- question screen to determine that Ms. E drank more than 4 beers per day during social events and weekends, which occurred 2 to 3 times per month over the past year. She lives alone and said that she’d been seeing less and less of her boyfriend lately. Her score on the Patient Health Questionnaire (PHQ), which screens for depression, was 11, indicating moderate impairment. Her response on the CAGE questionnaire was negative for a problem with alcohol. However, her AUDIT score was 17, indicating a high-level alcohol problem. Based on these findings, her physician expressed concern that her alcohol use might be contributing to her symptoms and difficulties.

The CAGE questionnaire has lower sensitivity when alcohol intake is lower.

Although she did not have a history of increasing usage per day, a persistent desire to cut down, significant effort to obtain alcohol, or cravings, she was having work troubles and continued to drink even though it was straining relationships, promoting weight gain, and causing abdominal pain.

The physician asked her to schedule a return visit and ordered several blood studies. He also offered to connect her with a colleague with whom he collaborated who could speak with her about possible alcohol use disorders and depression.

Continue to: Selecting blood work in screening for AUD

Lab tests used to measure hepatic injury due to alcohol include gamma-glutamyl-transferase, aspartate aminotransferase (AST), alanine aminotransferase (ALT), and macrocytic volume, although the indices of hepatic damage have low specificity. Elevated serum ethanol levels can reveal recent alcohol use, and vitamin deficiencies and other abnormalities can be used to differentiate other causes of hepatic inflammation and co-existing health issues (TABLE 3^10,18). A number of as-yet-unvalidated biomarkers are being studied to assist in screening, diagnosing, and treating AUD.¹⁸

What treatment approaches work for AUD?

Family physicians can efficiently and productively address AUD by using alcohol screening and brief intervention, which have been shown to reduce risky drinking. Reimbursement for this service is covered by such CPT codes as 99408, 99409, or H0049, or with other evaluation and management (E/M) codes by using modifier 25.

Treatment of AUD varies and should be customized to each patient’s needs, readiness, preferences, and resources. Individual and group counseling approaches can be effective, and medications are available for inpatient and outpatient settings. Psychotherapy options include brief interventions, 12-step programs (eg, Alcoholics Anonymous—https://www.aa.org/pages/en_US/find-aa-resources),motivational enhancement therapy, and cognitive behavioral therapy. Although it is beyond the scope of this article to describe these options in detail, resources are available for those who wish to learn more.^19-21

Psychopharmacologic management includes US Food and Drug Administration (FDA)-approved medications such as disulfiram, naltrexone, and acamprosate, and off-label uses of other medications (TABLE 4⁹). Not enough empiric evidence is available to judge the effectiveness of these medications in adolescents, and the FDA has not approved them for such use. Evidence from meta-analyses comparing naltrexone and acamprosate have shown naltrexone to be more efficacious in reducing heavy drinking and cravings, while acamprosate is effective in promoting abstinence.^22,23 Naltrexone combined with behavioral intervention reduces the heavy drinking days and percentage of abstinence days.²⁴

Current guideline recommendations from the American Psychiatric Association²⁵ include:

• Naltrexone and acamprosate are recommended to treat patients with moderate-to-severe AUD in specific circumstances (eg, when nonpharmacologic approaches have failed to produce an effect or when patients prefer to use one of these medications).
• Topiramate and gabapentin are also suggested as medications for patients with moderate-to-severe AUD, but typically after first trying naltrexone and acamprosate.
• Disulfiram generally should not be used as first-line treatment. It produces physical reactions (eg, flushing) if alcohol is consumed within 12 to 24 hours of medication use.

Continue to: THE CASE

Ms. E was open to the idea of decreasing her alcohol use and agreed that she was depressed. Her lab tests at follow-up were normal other than an elevated AST/ALT of 90/80 U/L. She received brief counseling from her family physician combined with cognitive behavioral therapy by a psychologist colleague. A subsequent ultrasound of her liver showed mild hepatomegaly and moderate-to-severe steatosis with nodularity of the liver, raising the possibility of cirrhosis.

The NIAAA single-question screen asks how many times in the past year the patient had ≥4 drinks (women) or ≥5 drinks (men) in a day.

She continued to get counseling for her AUD and for her comorbid depression in addition to taking a selective serotonin reuptake inhibitor. She is now in early remission for her alcohol use.

CORRESPONDENCE
Jaividhya Dasarathy, MD, Department of Family Medicine, Metro Health Medical Center, 2500 MetroHealth Drive, Cleveland, OH 44109; [email protected].

THE CASE

Ms. E, a 42-year-old woman, visited her new physician for a physical exam. When asked about alcohol intake, she reported that she drank 3 to 4 beers after work and sometimes 5 to 8 beers a day on the weekends. Occasionally, she exceeded those amounts, but she didn’t feel guilty about her drinking. She was often late to work and said her relationship with her boyfriend was strained. A review of systems was positive for fatigue, poor concentration, abdominal pain, and weight gain. Her body mass index was 41, pulse 100 beats/min, blood pressure 125/75 mm Hg, and she was afebrile. Her physical exam was otherwise within normal limits.

How would you proceed with this patient?

Alcohol use disorder (AUD) is a common and often untreated condition that is increasingly prevalent in the United States.¹ The Diagnostic and Statistical Manual of Mental Disorders, 5th Edition (DSM-5) characterizes AUD as a combination of signs and symptoms typifying alcohol abuse and dependence (discussed in a bit).²

Data from the 2015 National Survey on Drug Use and Health (NSDUH) showed 15.7 million Americans with AUD, affecting 6.2% of the population ages 18 years or older and 2.5% of adolescents ages 12 to 17 years.³

Alcohol use and AUD account for an estimated 3.8% of all global deaths and 4.6% of global disability-adjusted life years.⁴ AUD adversely affects several systems (TABLE 1⁵), and patients with AUD are sicker and more likely to die younger than those without AUD.⁴ In the United States, prevalence of AUD has increased in recent years among women, older adults, racial minorities, and individuals with a low education level.⁶

Screening for AUD is reasonable and straightforward, although diagnosis and treatment of AUD in primary care settings may be challenging due to competing clinical priorities; lack of training, resources, and support; and skepticism about the efficacy of behavioral and pharmacologic treatments.^7,8 However, family physicians are in an excellent position to diagnose and help address the complex biopsychosocial needs of patients with AUD, often in collaboration with colleagues and community organizations.

Signs and symptoms of AUD

In clinical practice, at least 2 of the following 11 behaviors or symptoms are required to diagnose AUD²:

• consuming larger amounts of alcohol over a longer period than intended
• persistent desire or unsuccessful efforts to cut down or control alcohol use
• making a significant effort to obtain, use, or recover from alcohol

In moderate-to-severe cases:

• cravings or urges to use alcohol
• recurrent failure to fulfill major work, school, or social obligations
• continued alcohol use despite recurrent social and interpersonal problems
• giving up social, occupational, and recreational activities due to alcohol
• using alcohol in physically dangerous situations
• continued alcohol use despite having physical or psychological problems
• tolerance to alcohol’s effects
• withdrawal symptoms.

Continue to: Patients meet criteria for mild AUD severity if...

Patients meet criteria for mild AUD severity if they exhibit 2 or 3 symptoms, moderate AUD with 4 or 5 symptoms, and severe AUD if there are 6 or more symptoms.²

USPSTF recommends screening for alcohol use disorder with the 10-question AUDIT, 3-question AUDIT-C, or single-question NIAAA tool.

Those who meet criteria for AUD and are able to stop using alcohol are deemed to be in early remission if the criteria have gone unfulfilled for at least 3 months and less than 12 months. Patients are considered to be in sustained remission if they have not met criteria for AUD at any time during a period of 12 months or longer.

How to detect AUD

Several clues in a patient’s history can suggest AUD (TABLE 2^9,10). Most imbibers are unaware of the dangers and may consider themselves merely “social drinkers.” Binge drinking may be an early indicator of vulnerability to AUD and should be assessed as part of a thorough clinical evaluation.¹¹ The US Preventive Services Task Force (USPSTF) recommends (Grade B) that clinicians screen adults ages 18 years or older for alcohol misuse.¹²

Studies demonstrate that both genetic and environmental factors play important roles in the development of AUD.¹³ A family history of excessive alcohol use increases the risk of AUD. Comorbidity of AUD and other mental health conditions is extremely common. For example, high rates of association between major depressive disorder and AUD have been observed.¹⁴

Tools to use in screening and diagnosing AUD

Screening for AUD during an office visit can be done fairly quickly. While 96% of primary care physicians screen for alcohol misuse in some way, only 38% use 1 of the 3 tools recommended by the USPSTF¹⁵—the Alcohol Use Disorders Identification Test (AUDIT), the abbreviated AUDIT-C, or the National Institute on Alcohol Abuse and Alcoholism (NIAAA) single question screen—which detect the full spectrum of alcohol misuse in adults.¹² Although the commonly used CAGE questionnaire is one of the most studied self-report tools, it has lower sensitivity at a lower level of alcohol intake.¹⁶

Continue to: The NIAAA single-question screen asks...

The NIAAA single-question screen asks how many times in the past year the patient had ≥4 drinks (women) or ≥5 drinks (men) in a day.¹⁵ The sensitivity and specificity of single-question screening are 82% to 87% and 61% to 79%, respectively, and the test has been validated in several different settings.¹² The AUDIT screening tool, freely available from the World Health Organization, is a 10-item questionnaire that probes an individual’s alcohol intake, alcohol dependence, and adverse consequences of alcohol use. Administration of the AUDIT typically requires only 2 minutes. AUDIT-C¹⁷ is an abbreviated version of the AUDIT questionnaire that asks 3 consumption questions to screen for AUD.

It was found that AUDIT scores in the range of 8 to 15 indicate a medium-level alcohol problem, whereas a score of ≥16 indicates a high-level alcohol problem. The AUDIT-C is scored from 0 to 12, with ≥4 indicating a problem in men and ≥3 a problem in women.

THE CASE

The physician had used the NIAAA single- question screen to determine that Ms. E drank more than 4 beers per day during social events and weekends, which occurred 2 to 3 times per month over the past year. She lives alone and said that she’d been seeing less and less of her boyfriend lately. Her score on the Patient Health Questionnaire (PHQ), which screens for depression, was 11, indicating moderate impairment. Her response on the CAGE questionnaire was negative for a problem with alcohol. However, her AUDIT score was 17, indicating a high-level alcohol problem. Based on these findings, her physician expressed concern that her alcohol use might be contributing to her symptoms and difficulties.

The CAGE questionnaire has lower sensitivity when alcohol intake is lower.

Although she did not have a history of increasing usage per day, a persistent desire to cut down, significant effort to obtain alcohol, or cravings, she was having work troubles and continued to drink even though it was straining relationships, promoting weight gain, and causing abdominal pain.

The physician asked her to schedule a return visit and ordered several blood studies. He also offered to connect her with a colleague with whom he collaborated who could speak with her about possible alcohol use disorders and depression.

Continue to: Selecting blood work in screening for AUD

Lab tests used to measure hepatic injury due to alcohol include gamma-glutamyl-transferase, aspartate aminotransferase (AST), alanine aminotransferase (ALT), and macrocytic volume, although the indices of hepatic damage have low specificity. Elevated serum ethanol levels can reveal recent alcohol use, and vitamin deficiencies and other abnormalities can be used to differentiate other causes of hepatic inflammation and co-existing health issues (TABLE 3^10,18). A number of as-yet-unvalidated biomarkers are being studied to assist in screening, diagnosing, and treating AUD.¹⁸

What treatment approaches work for AUD?

Family physicians can efficiently and productively address AUD by using alcohol screening and brief intervention, which have been shown to reduce risky drinking. Reimbursement for this service is covered by such CPT codes as 99408, 99409, or H0049, or with other evaluation and management (E/M) codes by using modifier 25.

Treatment of AUD varies and should be customized to each patient’s needs, readiness, preferences, and resources. Individual and group counseling approaches can be effective, and medications are available for inpatient and outpatient settings. Psychotherapy options include brief interventions, 12-step programs (eg, Alcoholics Anonymous—https://www.aa.org/pages/en_US/find-aa-resources),motivational enhancement therapy, and cognitive behavioral therapy. Although it is beyond the scope of this article to describe these options in detail, resources are available for those who wish to learn more.^19-21

Psychopharmacologic management includes US Food and Drug Administration (FDA)-approved medications such as disulfiram, naltrexone, and acamprosate, and off-label uses of other medications (TABLE 4⁹). Not enough empiric evidence is available to judge the effectiveness of these medications in adolescents, and the FDA has not approved them for such use. Evidence from meta-analyses comparing naltrexone and acamprosate have shown naltrexone to be more efficacious in reducing heavy drinking and cravings, while acamprosate is effective in promoting abstinence.^22,23 Naltrexone combined with behavioral intervention reduces the heavy drinking days and percentage of abstinence days.²⁴

Current guideline recommendations from the American Psychiatric Association²⁵ include:

• Naltrexone and acamprosate are recommended to treat patients with moderate-to-severe AUD in specific circumstances (eg, when nonpharmacologic approaches have failed to produce an effect or when patients prefer to use one of these medications).
• Topiramate and gabapentin are also suggested as medications for patients with moderate-to-severe AUD, but typically after first trying naltrexone and acamprosate.
• Disulfiram generally should not be used as first-line treatment. It produces physical reactions (eg, flushing) if alcohol is consumed within 12 to 24 hours of medication use.

Continue to: THE CASE

Ms. E was open to the idea of decreasing her alcohol use and agreed that she was depressed. Her lab tests at follow-up were normal other than an elevated AST/ALT of 90/80 U/L. She received brief counseling from her family physician combined with cognitive behavioral therapy by a psychologist colleague. A subsequent ultrasound of her liver showed mild hepatomegaly and moderate-to-severe steatosis with nodularity of the liver, raising the possibility of cirrhosis.

The NIAAA single-question screen asks how many times in the past year the patient had ≥4 drinks (women) or ≥5 drinks (men) in a day.

She continued to get counseling for her AUD and for her comorbid depression in addition to taking a selective serotonin reuptake inhibitor. She is now in early remission for her alcohol use.

CORRESPONDENCE
Jaividhya Dasarathy, MD, Department of Family Medicine, Metro Health Medical Center, 2500 MetroHealth Drive, Cleveland, OH 44109; [email protected].

THE CASE

Ms. E, a 42-year-old woman, visited her new physician for a physical exam. When asked about alcohol intake, she reported that she drank 3 to 4 beers after work and sometimes 5 to 8 beers a day on the weekends. Occasionally, she exceeded those amounts, but she didn’t feel guilty about her drinking. She was often late to work and said her relationship with her boyfriend was strained. A review of systems was positive for fatigue, poor concentration, abdominal pain, and weight gain. Her body mass index was 41, pulse 100 beats/min, blood pressure 125/75 mm Hg, and she was afebrile. Her physical exam was otherwise within normal limits.

How would you proceed with this patient?

Alcohol use disorder (AUD) is a common and often untreated condition that is increasingly prevalent in the United States.¹ The Diagnostic and Statistical Manual of Mental Disorders, 5th Edition (DSM-5) characterizes AUD as a combination of signs and symptoms typifying alcohol abuse and dependence (discussed in a bit).²

Data from the 2015 National Survey on Drug Use and Health (NSDUH) showed 15.7 million Americans with AUD, affecting 6.2% of the population ages 18 years or older and 2.5% of adolescents ages 12 to 17 years.³

Alcohol use and AUD account for an estimated 3.8% of all global deaths and 4.6% of global disability-adjusted life years.⁴ AUD adversely affects several systems (TABLE 1⁵), and patients with AUD are sicker and more likely to die younger than those without AUD.⁴ In the United States, prevalence of AUD has increased in recent years among women, older adults, racial minorities, and individuals with a low education level.⁶

Screening for AUD is reasonable and straightforward, although diagnosis and treatment of AUD in primary care settings may be challenging due to competing clinical priorities; lack of training, resources, and support; and skepticism about the efficacy of behavioral and pharmacologic treatments.^7,8 However, family physicians are in an excellent position to diagnose and help address the complex biopsychosocial needs of patients with AUD, often in collaboration with colleagues and community organizations.

Signs and symptoms of AUD

In clinical practice, at least 2 of the following 11 behaviors or symptoms are required to diagnose AUD²:

• consuming larger amounts of alcohol over a longer period than intended
• persistent desire or unsuccessful efforts to cut down or control alcohol use
• making a significant effort to obtain, use, or recover from alcohol

In moderate-to-severe cases:

• cravings or urges to use alcohol
• recurrent failure to fulfill major work, school, or social obligations
• continued alcohol use despite recurrent social and interpersonal problems
• giving up social, occupational, and recreational activities due to alcohol
• using alcohol in physically dangerous situations
• continued alcohol use despite having physical or psychological problems
• tolerance to alcohol’s effects
• withdrawal symptoms.

Continue to: Patients meet criteria for mild AUD severity if...

Patients meet criteria for mild AUD severity if they exhibit 2 or 3 symptoms, moderate AUD with 4 or 5 symptoms, and severe AUD if there are 6 or more symptoms.²

USPSTF recommends screening for alcohol use disorder with the 10-question AUDIT, 3-question AUDIT-C, or single-question NIAAA tool.

Those who meet criteria for AUD and are able to stop using alcohol are deemed to be in early remission if the criteria have gone unfulfilled for at least 3 months and less than 12 months. Patients are considered to be in sustained remission if they have not met criteria for AUD at any time during a period of 12 months or longer.

How to detect AUD

Several clues in a patient’s history can suggest AUD (TABLE 2^9,10). Most imbibers are unaware of the dangers and may consider themselves merely “social drinkers.” Binge drinking may be an early indicator of vulnerability to AUD and should be assessed as part of a thorough clinical evaluation.¹¹ The US Preventive Services Task Force (USPSTF) recommends (Grade B) that clinicians screen adults ages 18 years or older for alcohol misuse.¹²

Studies demonstrate that both genetic and environmental factors play important roles in the development of AUD.¹³ A family history of excessive alcohol use increases the risk of AUD. Comorbidity of AUD and other mental health conditions is extremely common. For example, high rates of association between major depressive disorder and AUD have been observed.¹⁴

Tools to use in screening and diagnosing AUD

Screening for AUD during an office visit can be done fairly quickly. While 96% of primary care physicians screen for alcohol misuse in some way, only 38% use 1 of the 3 tools recommended by the USPSTF¹⁵—the Alcohol Use Disorders Identification Test (AUDIT), the abbreviated AUDIT-C, or the National Institute on Alcohol Abuse and Alcoholism (NIAAA) single question screen—which detect the full spectrum of alcohol misuse in adults.¹² Although the commonly used CAGE questionnaire is one of the most studied self-report tools, it has lower sensitivity at a lower level of alcohol intake.¹⁶

Continue to: The NIAAA single-question screen asks...

The NIAAA single-question screen asks how many times in the past year the patient had ≥4 drinks (women) or ≥5 drinks (men) in a day.¹⁵ The sensitivity and specificity of single-question screening are 82% to 87% and 61% to 79%, respectively, and the test has been validated in several different settings.¹² The AUDIT screening tool, freely available from the World Health Organization, is a 10-item questionnaire that probes an individual’s alcohol intake, alcohol dependence, and adverse consequences of alcohol use. Administration of the AUDIT typically requires only 2 minutes. AUDIT-C¹⁷ is an abbreviated version of the AUDIT questionnaire that asks 3 consumption questions to screen for AUD.

It was found that AUDIT scores in the range of 8 to 15 indicate a medium-level alcohol problem, whereas a score of ≥16 indicates a high-level alcohol problem. The AUDIT-C is scored from 0 to 12, with ≥4 indicating a problem in men and ≥3 a problem in women.

THE CASE

The physician had used the NIAAA single- question screen to determine that Ms. E drank more than 4 beers per day during social events and weekends, which occurred 2 to 3 times per month over the past year. She lives alone and said that she’d been seeing less and less of her boyfriend lately. Her score on the Patient Health Questionnaire (PHQ), which screens for depression, was 11, indicating moderate impairment. Her response on the CAGE questionnaire was negative for a problem with alcohol. However, her AUDIT score was 17, indicating a high-level alcohol problem. Based on these findings, her physician expressed concern that her alcohol use might be contributing to her symptoms and difficulties.

The CAGE questionnaire has lower sensitivity when alcohol intake is lower.

Although she did not have a history of increasing usage per day, a persistent desire to cut down, significant effort to obtain alcohol, or cravings, she was having work troubles and continued to drink even though it was straining relationships, promoting weight gain, and causing abdominal pain.

The physician asked her to schedule a return visit and ordered several blood studies. He also offered to connect her with a colleague with whom he collaborated who could speak with her about possible alcohol use disorders and depression.

Continue to: Selecting blood work in screening for AUD

Lab tests used to measure hepatic injury due to alcohol include gamma-glutamyl-transferase, aspartate aminotransferase (AST), alanine aminotransferase (ALT), and macrocytic volume, although the indices of hepatic damage have low specificity. Elevated serum ethanol levels can reveal recent alcohol use, and vitamin deficiencies and other abnormalities can be used to differentiate other causes of hepatic inflammation and co-existing health issues (TABLE 3^10,18). A number of as-yet-unvalidated biomarkers are being studied to assist in screening, diagnosing, and treating AUD.¹⁸

What treatment approaches work for AUD?

Family physicians can efficiently and productively address AUD by using alcohol screening and brief intervention, which have been shown to reduce risky drinking. Reimbursement for this service is covered by such CPT codes as 99408, 99409, or H0049, or with other evaluation and management (E/M) codes by using modifier 25.

Treatment of AUD varies and should be customized to each patient’s needs, readiness, preferences, and resources. Individual and group counseling approaches can be effective, and medications are available for inpatient and outpatient settings. Psychotherapy options include brief interventions, 12-step programs (eg, Alcoholics Anonymous—https://www.aa.org/pages/en_US/find-aa-resources),motivational enhancement therapy, and cognitive behavioral therapy. Although it is beyond the scope of this article to describe these options in detail, resources are available for those who wish to learn more.^19-21

Psychopharmacologic management includes US Food and Drug Administration (FDA)-approved medications such as disulfiram, naltrexone, and acamprosate, and off-label uses of other medications (TABLE 4⁹). Not enough empiric evidence is available to judge the effectiveness of these medications in adolescents, and the FDA has not approved them for such use. Evidence from meta-analyses comparing naltrexone and acamprosate have shown naltrexone to be more efficacious in reducing heavy drinking and cravings, while acamprosate is effective in promoting abstinence.^22,23 Naltrexone combined with behavioral intervention reduces the heavy drinking days and percentage of abstinence days.²⁴

Current guideline recommendations from the American Psychiatric Association²⁵ include:

• Naltrexone and acamprosate are recommended to treat patients with moderate-to-severe AUD in specific circumstances (eg, when nonpharmacologic approaches have failed to produce an effect or when patients prefer to use one of these medications).
• Topiramate and gabapentin are also suggested as medications for patients with moderate-to-severe AUD, but typically after first trying naltrexone and acamprosate.
• Disulfiram generally should not be used as first-line treatment. It produces physical reactions (eg, flushing) if alcohol is consumed within 12 to 24 hours of medication use.

Continue to: THE CASE

Ms. E was open to the idea of decreasing her alcohol use and agreed that she was depressed. Her lab tests at follow-up were normal other than an elevated AST/ALT of 90/80 U/L. She received brief counseling from her family physician combined with cognitive behavioral therapy by a psychologist colleague. A subsequent ultrasound of her liver showed mild hepatomegaly and moderate-to-severe steatosis with nodularity of the liver, raising the possibility of cirrhosis.

The NIAAA single-question screen asks how many times in the past year the patient had ≥4 drinks (women) or ≥5 drinks (men) in a day.

She continued to get counseling for her AUD and for her comorbid depression in addition to taking a selective serotonin reuptake inhibitor. She is now in early remission for her alcohol use.

CORRESPONDENCE
Jaividhya Dasarathy, MD, Department of Family Medicine, Metro Health Medical Center, 2500 MetroHealth Drive, Cleveland, OH 44109; [email protected].

Migraine is a highly disabling primary headache disorder that affects more than 44 million Americans annually.¹ The disorder causes pain, photophobia, phonophobia, and nausea that can last for hours, even days. Migraine headaches are 2 times more common in women than in men; although migraine is most common in people 30 to 39 years of age, all ages are affected.^2,3 Frequency of migraine headache is variable; chronic migraineurs experience more than 15 headache days a month.

©Cath Riley/Science Source

Recent estimates indicate that the cost of acute and chronic migraine headaches reaches approximately $78 million a year in the United States. ⁴ This high burden of disease has made effective migraine treatment options absolutely essential. Recent advances in our understanding of migraine pathophysiology have led to new therapeutic targets; there are now many novel treatment approaches on the horizon.

In this article, we review the diagnosis and management of migraine in detail. Our emphasis is on evidence-based approaches to acute and prophylactic treatment, including tried-and-true options and newly emerging therapies.

Neuronal dysfunction and a genetic predisposition

Although migraine was once thought to be caused by abnormalities of vasodilation, current research suggests that the disorder has its origins in primary neuronal dysfunction. There appears to be a genetic predisposition toward widespread neuronal hyperexcitability in migraineurs.⁵ In addition, hypothalamic neurons are thought to initiate migraine by responding to changes in brain homeostasis. Increased parasympathetic tone might activate meningeal pain receptors or lower the threshold for transmitting pain signals from the thalamus to the cortex.⁶

Prodromal symptoms and aura appear to originate from multiple areas across the brain, including the hypothalamus, cortex, limbic system, and brainstem. This widespread brain involvement might explain why some headache sufferers concurrently experience a variety of symptoms, including fatigue, depression, muscle pain, and an abnormal sensitivity to light, sound, and smell.^6,7

After taking the initial history (headache onset, location, duration, associated symptoms), focus attention on assessing the risk of intracranial pathology.

Although the exact mechanisms behind each of these symptoms have yet to be defined precisely, waves of neuronal depolarization—known as cortical spreading depression—are suspected to cause migraine aura.^8-10 Cortical spreading depression activates the trigeminal pain pathway and leads to the release of pro-inflammatory markers such as calcitonin gene-related protein (CGRP).⁶ A better understanding of these complex signaling pathways has helped provide potential therapeutic targets for new migraine drugs.

Diagnosis: Close patient inquiry is most helpful

The International Headache Society (IHS) criteria for primary headache disorders serve as the basis for the diagnosis of migraine and its subtypes, which include migraine without aura and migraine with aura. Due to variability of presentation, migraine with aura is further subdivided into migraine with typical aura (with and without headache), migraine with brainstem aura, hemiplegic migraine, and retinal migraine.¹¹

Continue to: How is migraine defined?

How is migraine defined? Simply, migraine is classically defined as a unilateral, pulsating headache of moderate to severe intensity lasting 4 to 72 hours, associated with photophobia and phonophobia or nausea and vomiting, or both.¹¹ Often visual in nature, aura is a set of neurologic symptoms that lasts for minutes and precedes the onset of the headache. The visual aura is often described as a scintillating scotoma that begins near the point of visual fixation and then spreads left or right. Other aura symptoms include tingling or numbness (second most common), speech disturbance (aphasia), motor changes and, in rare cases, a combination of these in succession. By definition, all of these symptoms fully resolve between attacks.¹¹

Validated valuable questionnaires. To help with accurate and timely diagnosis, researchers have developed and validated simplified questionnaires that can be completed independently by patients presenting to primary care (TABLE 1^12,13):

• ID Migraine is a set of 3 questions that scores positive when a patient endorses at least 2 of the 3 symptoms. ¹²
• MS-Q is similar to the ID Migraine but includes 5 items. A score of ≥4 is a positive screen. ¹³

The sensitivity and specificity of MS-Q (0.93 and 0.81, respectively) are slightly higher than those of ID Migraine (0.81 and 0.75).¹³

Remember POUND. This mnemonic device can also be used during history-taking to aid in diagnostic accuracy. Migraine is highly likely (92%) in patients who endorse 4 of the following 5 symptoms and unlikely (17%) in those who endorse ≤2 symptoms¹⁴: Pulsatile quality of headache 4 to 72 hOurs in duration, Unilateral location, Nausea or vomiting, and Disabling intensity.

Differential Dx. Although the differential diagnosis of headache is broad (TABLE 2^14,15), the history alone can often guide clinicians towards the correct assessment. After taking the initial history (headache onset, location, duration, and associated symptoms), focus your attention on assessing the risk of intracranial pathology. This is best accomplished by assessing specific details of the history (TABLE 3¹⁴) and findings on physical examination¹⁵:

• blood pressure measurement (seated, legs uncrossed, feet flat on the floor; having rested for 5 minutes; arm well supported)
• cranial nerve exam
• extremity strength testing
• eye exam (vision, extra-ocular muscles, visual fields, pupillary reactivity, and funduscopic exam)
• gait (tandem walk)
• reflexes.

Continue to: Further testing needed?

Further testing needed? Neuroimaging should be considered only in patients with an abnormal neurologic exam, atypical headache features, or certain risk factors, such as an immune deficiency. There is no role for electroencephalography or other diagnostic testing in migraine.¹⁶

Take a multipronged approach to treatment

As with other complex, chronic conditions, the treatment of migraine should take a multifaceted approach, including management of acute symptoms as well as prevention of future headaches. In 2015, the American Headache Society published a systematic review that specified particular treatment goals for migraine sufferers. ¹⁷ These goals include:

• headache reduction
• headache relief
• decreased disability from headache
• elimination of nausea and vomiting
• elimination of photophobia and phonophobia.

Our review, which follows, of therapeutic options focuses on the management of migraine in adults. Approaches in special populations (older adults, pregnant women, and children) are discussed afterward.

Pharmacotherapy for acute migraine

Acute migraine should be treated with an abortive medication at the onset of headache. The immediate goal is to relieve pain within 2 hours and prevent its recurrence within the subsequent 48 hours (TABLE 4^12,18-20).

Electroencephalography and other diagnostic testing have no role in the workup of migraine.

In the general population, mild, infrequent migraines can be managed with acetaminophen and nonsteroidal anti-inflammatory drugs (NSAIDs).²¹

Continue to: For moderate-to-severe migraine...

For moderate-to-severe migraine, triptans, which target serotonin receptors, are the drug of choice for most patients.²¹ Triptans are superior to placebo in achieving a pain-free state at 2 and 24 hours after administration; eletriptan has the most desirable outcome, with 68% of patients pain free at 2 hours and 54% pain free at 24 hours.²² Triptans are available as sublingual tablets and nasal sprays, as well as subcutaneous injections for patients with significant associated nausea and vomiting. Avoid prescribing triptans for patients with known vascular disease (eg, history of stroke, myocardial infarction, peripheral vascular disease, uncontrolled hypertension, or signs and symptoms of these conditions), as well as for patients with severe hepatic impairment.

Treat migraine with a multifaceted approach, including management of acute symptoms and prevention of future headaches.

Importantly, although triptans all have a similar mechanism of action, patients might respond differently to different drugs within the class. If a patient does not get adequate headache relief from an appropriate dosage of a given triptan during a particular migraine episode, a different triptan can be tried during the next migraine.²² Additionally, if a patient experiences an adverse effect from one triptan, this does not necessarily mean that a trial of another triptan at a later time is contraindicated.

For patients who have an incomplete response to migraine treatment or for those with frequent recurrence, the combination formulation of sumatriptan, 85 mg, and naproxen, 500 mg, showed the highest rate of resolution of headache within 2 hours compared with either drug alone.²³ A similar result might be found by combining a triptan known to be effective for a patient and an NSAID other than naproxen. If migraine persists despite initial treatment of an attack, a different class of medication should be tried during the course of that attack to attain relief of symptoms of that migraine.²¹

When a patient is seen in an acute care setting (eg, emergency department, urgent care center) while suffering a migraine, additional treatment options are available. Intravenous (IV) anti-emetics are useful for relieving the pain of migraine and nausea, and can be used in combination with an IV NSAID (eg, ketorolac).²¹ The most effective anti-emetics are dopamine receptor type-2 blockers, including chlorpromazine, droperidol, metoclopramide, and prochlorperazine, which has the highest level of efficacy.²⁴ Note that these medications do present the risk of a dystonic reaction; diphenhydramine is therefore often used in tandem to mitigate such a response.

Looking ahead. Although triptans are the current first-line therapy for acute migraine, their effectiveness is limited. Only 20% of patients report sustained relief of pain in the 2 to 24 hours after treatment, and the response can vary from episode to episode.²⁵

Continue to: With better understading of the pathophysiology of migraine...

With better understanding of the pathophysiology of migraine, a host of novel anti-migraine drugs are on the horizon.

CGRP receptor antagonists. The neuropeptide CGRP, which mediates central and peripheral nervous system pain signaling, has been noted to be elevated during acute migraine attacks²⁶; clinical trials are therefore underway to evaluate the safety and efficacy of CGRP receptor antagonists.¹⁸ These agents appear to be better tolerated than triptans, have fewer vascular and central nervous system adverse effects, and present less of a risk of medication overuse headache.¹⁸ Liver toxicity has been seen with some medications in this class and remains an important concern in their development.¹⁹

Phase 3 clinical trials for 1 drug in this class, ubrogepant, were completed in late 2017; full analysis of the data is not yet available. Primary outcomes being evaluated include relief of pain at 2 hours and relief from the most bothersome symptoms again at 2 hours.²⁷

Selective serotonin-HT_1f receptor agonists, such as lasmiditan, offer another potential approach. Although the exact mechanism of action of these agents is not entirely clear, clinical trials have supported their efficacy and safety.²⁰ Importantly, ongoing trials are specifically targeting patients with known cardiovascular risk factors because they are most likely to benefit from the nonvasoconstrictive mechanism of action.^28,29 Adverse effects reported primarily include dizziness, fatigue, and vertigo.

Strategies for managing recurrent episodic migraine

Because of the risk of medication overuse headache with acute treatment, daily preventive therapy for migraine is indicated for any patient with ³⁰ :

• ≥6 headache days a month
• ≥4 headache days a month with some impairment
• ≥3 headache days a month with severe impairment.

Continue to: Treatment begins by having patients identify...

Treatment begins by having patients identify, and then avoid, migraine triggers (TABLE 5). This can be accomplished by having patients keep a headache diary, in which they can enter notations about personal and environmental situations that precede a headache.

For the individual patient, some triggers are modifiable; others are not. Helping a patient develop strategies for coping with triggers, rather than aiming for complete avoidance, might help her (him) manage those that are inescapable (eg stress, menstruation, etc).³¹ For many patients, however, this is not an adequate intervention and other approaches must be explored. When considering which therapy might be best for a given patient, evaluate her (his) comorbidities and assess that particular treatment for potential secondary benefits and the possibility of adverse effects. Pay attention to the choice of preventive therapy in women who are considering pregnancy because many available treatments are potentially teratogenic.

Oral medications. Oral agents from several classes of drugs can be used for migraine prophylaxis, including anti-epileptics,antidepressants, and antihypertensives (TABLE 6^{20,29,30,32-41}). Selected anti-epileptics (divalproex sodium, sodium valproate, topiramate) and beta-blockers (metoprolol, propranolol, and timolol) have the strongest evidence to support their use.³² Overall, regular use of prophylactic medications can reduce headache frequency by 50% for approximately 40% to 45% of patients who take them.²⁹ However, adherence may be limited by adverse effects or perceived lack of efficacy, thus reducing their potential for benefit.⁴²

OnabotulinumtoxinA. In patients with chronic migraine (≥15 headache days a month for at least 3 months) who have failed oral medications, the American Academy of Neurology (AAN) recommends the use of onabotulinumtoxinA.³⁰ The treatment regimen comprises 31 injections at various sites on the head, neck, and shoulders every 3 months.³³

A 2010 large randomized controlled trial showed a decrease in the frequency of headache days for patients receiving onabotulinumtoxinA compared to placebo after a 24-week treatment period (7.8 fewer headache days a month, compared to 6.4 fewer in the placebo group).³³ A recent systematic review also noted a reduction of 2 headache days a month compared with placebo; the authors cautioned, however, that data with which to evaluate onabotulinumtoxinA in comparison to other prophylactic agents are limited.⁴³

Continue to: In both studies...

In both studies, the risk of adverse drug events due to onabotulinumtoxinA was high and led to a significant rate of discontinuation.^33,43 Despite this, onabotulinumtoxinA remains the only Food and Drug Administration (FDA)–approved treatment for chronic migraine, making it reasonable to consider for appropriate patients.

Acupuncture. A 2016 Cochrane review found benefit for patients using acupuncture compared with sham acupuncture.³⁴ When acupuncture was compared with prophylactic agents such as beta-blockers, calcium-channel blockers, and anti-epileptics, however, there was no significant difference between the procedure and pharmacotherapy. Patients willing and able to try acupuncture might see a reduction in the overall number of headaches. Acupuncture has few adverse effects; however, long-term data are lacking.³⁴

Exercise is not supported by robust data for its role as a prophylactic treatment. It is generally considered safe in most populations, however, and can be pursued with little out-of-pocket cost.³⁵

Cognitive behavioral therapy (CBT). The AAN recommends CBT, relaxation therapy, and biofeedback therapy. Accessibility of these services remains limited for many patients, and cost can be prohibitive.¹⁶

Supplements used to help prevent migraine include the root of Petasites hybridus (butterbur), magnesium, vitamin B2 (riboflavin), Tanacetum parthenium (feverfew), and coenzyme Q10.¹⁶ Although the strength of evidence for these therapies is limited by small trials, their overall risk of adverse effects is low, and they might be easier for patients to obtain than acupuncture or CBT.

Continue to: Butterbur, in particular...

Butterbur, in particular, has been found to be beneficial for migraine prevention in 2 small placebo-controlled trials. In a randomized controlled study of 245 patients P hybridus, (specifically, the German formulation, Petadolex), 75 mg BID, reduced the frequency of migraine attack by 48% at 4 months, compared to placebo (number needed to treat, 5.3).⁴⁴ No difference was found at lower dosages. The most common reported adverse effect was burping.

Regrettably, unpurified butterbur extract contains pyrrolizidine alkaloids, potentially hepatotoxic and carcinogenic compounds. Because of variations in purification in production facilities in the United States, butterbur supplements might not have all of these compounds removed—and so should be used with caution.⁴¹

Magnesium. Studies evaluating the use of magnesium have demonstrated varied results; differences in methods and dosing have limited broad application of findings. As with most supplements considered for prophylactic treatment, magnesium dosing is poorly understood, and bioavailability varies in its different forms. Oral supplementation can be given as magnesium dicitrate, 600 mg/d.⁴⁵

Recently, products containing various combinations of feverfew, coenzyme Q10, riboflavin, magnesium, and other supplements have shown benefit in early clinical trials.^36,37

Neural stimulation. Over the past few years, a variety of transcutaneous nerve stimulator devices have gained FDA approval for use in migraine prophylaxis. The long-term safety and efficacy of these devices is not yet well understood, but they appear to provide headache relief in the short term and decrease the frequency of headache.³⁸ Use of the noninvasive stimulators is limited today by high cost and poor coverage by US health care insurers.

Continue to: Newly available medical therapy

Newly available medical therapy. The FDA recently approved erenumab, a fully human monoclonal antibody for prevention of migraine in adults. This is the first drug in the CGRP antagonist class to be approved for this indication. Trials of this once-monthly, self-injectable drug show promising results for patients whose migraines have been refractory to other therapies.

A recent large trial evaluated 955 adults with migraine, randomizing them to receive erenumab, 70 mg; erenumab, 140 mg; or placebo over 28 weeks.³⁹ The groups receiving erenumab had a nearly 2-fold higher odds of having their migraine reduced by 50%, compared with placebo (number needed to treat with the 140-mg dose, 4.27). Similar numbers of participants from all groups discontinued the study.³⁹ Phase 3 trials that are not yet formally published have produced similarly beneficial results.^40,46 The FDA has listed injection site reaction and constipation as the most reported adverse effects.⁴⁰

Three other anti-CGRP antibodies are likely to be approved in the near future: fremanezumab, galcanezumab, and eptinezumab.

The approach to migraine in special populations

Management of acute and chronic migraine in children, pregnant women, and older adults requires special attention: Treatment approaches are different than they are for adults 19 to 65 years of age.

Pediatric patients. Migraine is the most common acute and recurrent headache syndrome in children. Headaches differ from those of adult migraine as a result of variations in brain maturation, plasticity, and cognitive development.⁴⁷ Migraine attacks are often of shorter duration in children, lasting 1 to 2 hours, but can still be preceded by visual aura.⁴⁸ Just as with adults, imaging, electroencephalography, lumbar puncture, and routine labs should be considered only if a child has an abnormal neurological exam or other concerning features (TABLE 2^14,15).

Continue to: The general approach to migraine treatment...

The general approach to migraine treatment in the pediatric population includes education of the child and family about symptom management. Acetaminophen, NSAIDs, and triptans are approved for abortive therapy in children and should be used for acute headache relief in the same way that they are used in adults. Oral rizatriptan, the most well studied triptan in the pediatric population, is approved for use in children as young as 6 years⁴⁹; the pediatric dosage is 5 mg/d for patients weighing 20 to 39 kg and 10 mg/d for patients weighing more than 40 kg (same as the adult dosage).

Don’t prescribe triptans for patients with known vascular disease or severe hepatic impairment.

Oral almotriptan and zolmitriptan are also approved for use in children 12 to 17 years of age. Usual dosages are: almotriptan, 12.5 mg at onset, can repeat in 2 hours as needed (maximum dosage, 25 mg/d); and zolmitriptan, 2.5 mg at onset, can repeat in 2 hours as needed (maximum dosage, 10 mg/d).⁵⁰

For children who are unable to swallow pills or who are vomiting, a non-oral route of administration is preferable. Rizatriptan is available as an orally disintegrating tablet. Zolmitriptan is available in a nasal spray at a dose of 5 mg for children 12 years and older. Sumatriptan is not approved for use in patients younger than 18 years; however, recent studies have shown that it might have good efficacy and tolerability.⁵⁰

Daily prophylactic treatment for recurrent migraine in the pediatric population is an evolving subject; published guidelines do not exist. It is reasonable to consider treatment using the same guidelines as those in place for adults.⁵¹ Topiramate, 1 to 2 mg/kg/d, is the only therapy approved by the FDA for episodic migraine preventive therapy in adolescents.⁵⁰

If a patient doesn’t get adequate headache relief from an appropriate dosage of a given triptan, try a different triptan during the next migraine.

Notably, a nonpharmacotherapeutic approach may be more effective for pediatric prevention. In 2017, a large double-blind, placebo-controlled trial investigated the use of amitriptyline, topiramate, and placebo for the treatment of recurrent migraine in children 8 to 17 years of age. An interim analysis of the 328 children enrolled found no significant differences in reduction of headache frequency with treatment compared with placebo over a 24-week period; the trial was stopped early due to futility.⁵²

Continue to: The study did show...

The study did show, however, that reducing migraine triggers provided a high level of benefit to study participants. Stress is one of the most common migraine triggers in children; lack of sleep, exposure to a warm climate, and exposure to video games are also notable triggers.⁵³ CBT may augment the efficacy of standard migraine medications in the pediatric population and may help prevent recurrence of episodes.⁵⁴

Pregnancy. The treatment of migraine is different in pregnant women than it is in nonpregnant adults because of a concern over adverse effects on fetal development. For acute headache treatment, first-line therapies include trigger avoidance and acetaminophen, 1000 mg (maximum dosage, 4000 mg/d).⁵⁵ If this is ineffective, a 10-mg dose of metoclopramide, as often as every 6 hours (not an FDA-approved indication), can be considered. During the second trimester, NSAIDs can be considered second-line therapy.

Triptans—specifically, sumatriptan and rizatriptan—can also be considered if first-line therapies fail.⁵⁶ Triptan-exposed pregnant women with migraine have a rate of congenital malformations, spontaneous abortions, and prematurity that is similar to what is seen in pregnant women with migraine who have not been exposed to triptans. However, when triptan-exposed women are compared with healthy, non-migraine-suffering women, the rate of spontaneous abortion appears to be increased in the triptan-exposed population.⁵⁷

Ergotamine is contraindicated during pregnancy because of its potential to induce uterine contractions and vasospasm, which can be detrimental to the fetus.⁵⁶Nonpharmacotherapeutic interventions such as heat, ice, massage, rest, and avoidance of triggers are as successful in the pregnant population as in the nonpregnant population. For migraine prevention, coenzyme Q10, vitamins B2 and B6 (pyridoxine), and oral magnesium can be considered. Feverfew and butterbur should be avoided because of concerns about fetal malformation and preterm labor.⁵⁸

Older adults. Choosing appropriate migraine therapy for older adults requires special consideration because of changes in drug metabolism and risks associated with drug adverse effects. Additionally, few studies of migraine drugs have included large populations of adults older than 65 years; medications should therefore be prescribed cautiously in this population, with particular attention to drug–drug interactions.

Continue to: Just as for younger adults...

Just as for younger adults, mild symptoms can be managed effectively with acetaminophen. NSAIDs may be used as well, but carry increased risks of gastric bleeding and elevation in blood pressure.⁵⁹ The use of triptans is acceptable for the appropriate patient, but should be avoided in patients with known vascular disease.⁶⁰ Antiemetics present an increased risk of extrapyramidal adverse effects in the elderly and should be used with caution at the lowest effective dosage.⁵⁹ Novel mechanisms of action make some of the newer agents potentially safer for use in older adults when treating acute migraine.

Stress is one of the most common migraine triggers in children.

For migraine prevention in older adults, particular attention should be paid to reducing triggers and minimizing polypharmacy.

More and more, successful treatment is within reach

With many clinical trials evaluating novel drugs underway, and additional studies contributing to our understanding of nonpharmacotherapeutic approaches to migraine treatment, improved headache control may become increasingly common over the next few years.

CORRESPONDENCE
Kathryn McGrath, MD, Department of Family and Community Medicine, Thomas Jefferson University, 1015 Walnut St, Philadelphia PA 19107; [email protected].

Migraine is a highly disabling primary headache disorder that affects more than 44 million Americans annually.¹ The disorder causes pain, photophobia, phonophobia, and nausea that can last for hours, even days. Migraine headaches are 2 times more common in women than in men; although migraine is most common in people 30 to 39 years of age, all ages are affected.^2,3 Frequency of migraine headache is variable; chronic migraineurs experience more than 15 headache days a month.

©Cath Riley/Science Source

Recent estimates indicate that the cost of acute and chronic migraine headaches reaches approximately $78 million a year in the United States. ⁴ This high burden of disease has made effective migraine treatment options absolutely essential. Recent advances in our understanding of migraine pathophysiology have led to new therapeutic targets; there are now many novel treatment approaches on the horizon.

In this article, we review the diagnosis and management of migraine in detail. Our emphasis is on evidence-based approaches to acute and prophylactic treatment, including tried-and-true options and newly emerging therapies.

Neuronal dysfunction and a genetic predisposition

Although migraine was once thought to be caused by abnormalities of vasodilation, current research suggests that the disorder has its origins in primary neuronal dysfunction. There appears to be a genetic predisposition toward widespread neuronal hyperexcitability in migraineurs.⁵ In addition, hypothalamic neurons are thought to initiate migraine by responding to changes in brain homeostasis. Increased parasympathetic tone might activate meningeal pain receptors or lower the threshold for transmitting pain signals from the thalamus to the cortex.⁶

Prodromal symptoms and aura appear to originate from multiple areas across the brain, including the hypothalamus, cortex, limbic system, and brainstem. This widespread brain involvement might explain why some headache sufferers concurrently experience a variety of symptoms, including fatigue, depression, muscle pain, and an abnormal sensitivity to light, sound, and smell.^6,7

After taking the initial history (headache onset, location, duration, associated symptoms), focus attention on assessing the risk of intracranial pathology.

Although the exact mechanisms behind each of these symptoms have yet to be defined precisely, waves of neuronal depolarization—known as cortical spreading depression—are suspected to cause migraine aura.^8-10 Cortical spreading depression activates the trigeminal pain pathway and leads to the release of pro-inflammatory markers such as calcitonin gene-related protein (CGRP).⁶ A better understanding of these complex signaling pathways has helped provide potential therapeutic targets for new migraine drugs.

Diagnosis: Close patient inquiry is most helpful

The International Headache Society (IHS) criteria for primary headache disorders serve as the basis for the diagnosis of migraine and its subtypes, which include migraine without aura and migraine with aura. Due to variability of presentation, migraine with aura is further subdivided into migraine with typical aura (with and without headache), migraine with brainstem aura, hemiplegic migraine, and retinal migraine.¹¹

Continue to: How is migraine defined?

How is migraine defined? Simply, migraine is classically defined as a unilateral, pulsating headache of moderate to severe intensity lasting 4 to 72 hours, associated with photophobia and phonophobia or nausea and vomiting, or both.¹¹ Often visual in nature, aura is a set of neurologic symptoms that lasts for minutes and precedes the onset of the headache. The visual aura is often described as a scintillating scotoma that begins near the point of visual fixation and then spreads left or right. Other aura symptoms include tingling or numbness (second most common), speech disturbance (aphasia), motor changes and, in rare cases, a combination of these in succession. By definition, all of these symptoms fully resolve between attacks.¹¹

Validated valuable questionnaires. To help with accurate and timely diagnosis, researchers have developed and validated simplified questionnaires that can be completed independently by patients presenting to primary care (TABLE 1^12,13):

• ID Migraine is a set of 3 questions that scores positive when a patient endorses at least 2 of the 3 symptoms. ¹²
• MS-Q is similar to the ID Migraine but includes 5 items. A score of ≥4 is a positive screen. ¹³

The sensitivity and specificity of MS-Q (0.93 and 0.81, respectively) are slightly higher than those of ID Migraine (0.81 and 0.75).¹³

Remember POUND. This mnemonic device can also be used during history-taking to aid in diagnostic accuracy. Migraine is highly likely (92%) in patients who endorse 4 of the following 5 symptoms and unlikely (17%) in those who endorse ≤2 symptoms¹⁴: Pulsatile quality of headache 4 to 72 hOurs in duration, Unilateral location, Nausea or vomiting, and Disabling intensity.

Differential Dx. Although the differential diagnosis of headache is broad (TABLE 2^14,15), the history alone can often guide clinicians towards the correct assessment. After taking the initial history (headache onset, location, duration, and associated symptoms), focus your attention on assessing the risk of intracranial pathology. This is best accomplished by assessing specific details of the history (TABLE 3¹⁴) and findings on physical examination¹⁵:

• blood pressure measurement (seated, legs uncrossed, feet flat on the floor; having rested for 5 minutes; arm well supported)
• cranial nerve exam
• extremity strength testing
• eye exam (vision, extra-ocular muscles, visual fields, pupillary reactivity, and funduscopic exam)
• gait (tandem walk)
• reflexes.

Continue to: Further testing needed?

Further testing needed? Neuroimaging should be considered only in patients with an abnormal neurologic exam, atypical headache features, or certain risk factors, such as an immune deficiency. There is no role for electroencephalography or other diagnostic testing in migraine.¹⁶

Take a multipronged approach to treatment

As with other complex, chronic conditions, the treatment of migraine should take a multifaceted approach, including management of acute symptoms as well as prevention of future headaches. In 2015, the American Headache Society published a systematic review that specified particular treatment goals for migraine sufferers. ¹⁷ These goals include:

• headache reduction
• headache relief
• decreased disability from headache
• elimination of nausea and vomiting
• elimination of photophobia and phonophobia.

Our review, which follows, of therapeutic options focuses on the management of migraine in adults. Approaches in special populations (older adults, pregnant women, and children) are discussed afterward.

Pharmacotherapy for acute migraine

Acute migraine should be treated with an abortive medication at the onset of headache. The immediate goal is to relieve pain within 2 hours and prevent its recurrence within the subsequent 48 hours (TABLE 4^12,18-20).

Electroencephalography and other diagnostic testing have no role in the workup of migraine.

In the general population, mild, infrequent migraines can be managed with acetaminophen and nonsteroidal anti-inflammatory drugs (NSAIDs).²¹

Continue to: For moderate-to-severe migraine...

For moderate-to-severe migraine, triptans, which target serotonin receptors, are the drug of choice for most patients.²¹ Triptans are superior to placebo in achieving a pain-free state at 2 and 24 hours after administration; eletriptan has the most desirable outcome, with 68% of patients pain free at 2 hours and 54% pain free at 24 hours.²² Triptans are available as sublingual tablets and nasal sprays, as well as subcutaneous injections for patients with significant associated nausea and vomiting. Avoid prescribing triptans for patients with known vascular disease (eg, history of stroke, myocardial infarction, peripheral vascular disease, uncontrolled hypertension, or signs and symptoms of these conditions), as well as for patients with severe hepatic impairment.

Treat migraine with a multifaceted approach, including management of acute symptoms and prevention of future headaches.

Importantly, although triptans all have a similar mechanism of action, patients might respond differently to different drugs within the class. If a patient does not get adequate headache relief from an appropriate dosage of a given triptan during a particular migraine episode, a different triptan can be tried during the next migraine.²² Additionally, if a patient experiences an adverse effect from one triptan, this does not necessarily mean that a trial of another triptan at a later time is contraindicated.

For patients who have an incomplete response to migraine treatment or for those with frequent recurrence, the combination formulation of sumatriptan, 85 mg, and naproxen, 500 mg, showed the highest rate of resolution of headache within 2 hours compared with either drug alone.²³ A similar result might be found by combining a triptan known to be effective for a patient and an NSAID other than naproxen. If migraine persists despite initial treatment of an attack, a different class of medication should be tried during the course of that attack to attain relief of symptoms of that migraine.²¹

When a patient is seen in an acute care setting (eg, emergency department, urgent care center) while suffering a migraine, additional treatment options are available. Intravenous (IV) anti-emetics are useful for relieving the pain of migraine and nausea, and can be used in combination with an IV NSAID (eg, ketorolac).²¹ The most effective anti-emetics are dopamine receptor type-2 blockers, including chlorpromazine, droperidol, metoclopramide, and prochlorperazine, which has the highest level of efficacy.²⁴ Note that these medications do present the risk of a dystonic reaction; diphenhydramine is therefore often used in tandem to mitigate such a response.

Looking ahead. Although triptans are the current first-line therapy for acute migraine, their effectiveness is limited. Only 20% of patients report sustained relief of pain in the 2 to 24 hours after treatment, and the response can vary from episode to episode.²⁵

Continue to: With better understading of the pathophysiology of migraine...

With better understanding of the pathophysiology of migraine, a host of novel anti-migraine drugs are on the horizon.

CGRP receptor antagonists. The neuropeptide CGRP, which mediates central and peripheral nervous system pain signaling, has been noted to be elevated during acute migraine attacks²⁶; clinical trials are therefore underway to evaluate the safety and efficacy of CGRP receptor antagonists.¹⁸ These agents appear to be better tolerated than triptans, have fewer vascular and central nervous system adverse effects, and present less of a risk of medication overuse headache.¹⁸ Liver toxicity has been seen with some medications in this class and remains an important concern in their development.¹⁹

Phase 3 clinical trials for 1 drug in this class, ubrogepant, were completed in late 2017; full analysis of the data is not yet available. Primary outcomes being evaluated include relief of pain at 2 hours and relief from the most bothersome symptoms again at 2 hours.²⁷

Selective serotonin-HT_1f receptor agonists, such as lasmiditan, offer another potential approach. Although the exact mechanism of action of these agents is not entirely clear, clinical trials have supported their efficacy and safety.²⁰ Importantly, ongoing trials are specifically targeting patients with known cardiovascular risk factors because they are most likely to benefit from the nonvasoconstrictive mechanism of action.^28,29 Adverse effects reported primarily include dizziness, fatigue, and vertigo.

Strategies for managing recurrent episodic migraine

Because of the risk of medication overuse headache with acute treatment, daily preventive therapy for migraine is indicated for any patient with ³⁰ :

• ≥6 headache days a month
• ≥4 headache days a month with some impairment
• ≥3 headache days a month with severe impairment.

Continue to: Treatment begins by having patients identify...

Treatment begins by having patients identify, and then avoid, migraine triggers (TABLE 5). This can be accomplished by having patients keep a headache diary, in which they can enter notations about personal and environmental situations that precede a headache.

For the individual patient, some triggers are modifiable; others are not. Helping a patient develop strategies for coping with triggers, rather than aiming for complete avoidance, might help her (him) manage those that are inescapable (eg stress, menstruation, etc).³¹ For many patients, however, this is not an adequate intervention and other approaches must be explored. When considering which therapy might be best for a given patient, evaluate her (his) comorbidities and assess that particular treatment for potential secondary benefits and the possibility of adverse effects. Pay attention to the choice of preventive therapy in women who are considering pregnancy because many available treatments are potentially teratogenic.

Oral medications. Oral agents from several classes of drugs can be used for migraine prophylaxis, including anti-epileptics,antidepressants, and antihypertensives (TABLE 6^{20,29,30,32-41}). Selected anti-epileptics (divalproex sodium, sodium valproate, topiramate) and beta-blockers (metoprolol, propranolol, and timolol) have the strongest evidence to support their use.³² Overall, regular use of prophylactic medications can reduce headache frequency by 50% for approximately 40% to 45% of patients who take them.²⁹ However, adherence may be limited by adverse effects or perceived lack of efficacy, thus reducing their potential for benefit.⁴²

OnabotulinumtoxinA. In patients with chronic migraine (≥15 headache days a month for at least 3 months) who have failed oral medications, the American Academy of Neurology (AAN) recommends the use of onabotulinumtoxinA.³⁰ The treatment regimen comprises 31 injections at various sites on the head, neck, and shoulders every 3 months.³³

A 2010 large randomized controlled trial showed a decrease in the frequency of headache days for patients receiving onabotulinumtoxinA compared to placebo after a 24-week treatment period (7.8 fewer headache days a month, compared to 6.4 fewer in the placebo group).³³ A recent systematic review also noted a reduction of 2 headache days a month compared with placebo; the authors cautioned, however, that data with which to evaluate onabotulinumtoxinA in comparison to other prophylactic agents are limited.⁴³

Continue to: In both studies...

In both studies, the risk of adverse drug events due to onabotulinumtoxinA was high and led to a significant rate of discontinuation.^33,43 Despite this, onabotulinumtoxinA remains the only Food and Drug Administration (FDA)–approved treatment for chronic migraine, making it reasonable to consider for appropriate patients.

Acupuncture. A 2016 Cochrane review found benefit for patients using acupuncture compared with sham acupuncture.³⁴ When acupuncture was compared with prophylactic agents such as beta-blockers, calcium-channel blockers, and anti-epileptics, however, there was no significant difference between the procedure and pharmacotherapy. Patients willing and able to try acupuncture might see a reduction in the overall number of headaches. Acupuncture has few adverse effects; however, long-term data are lacking.³⁴

Exercise is not supported by robust data for its role as a prophylactic treatment. It is generally considered safe in most populations, however, and can be pursued with little out-of-pocket cost.³⁵

Cognitive behavioral therapy (CBT). The AAN recommends CBT, relaxation therapy, and biofeedback therapy. Accessibility of these services remains limited for many patients, and cost can be prohibitive.¹⁶

Supplements used to help prevent migraine include the root of Petasites hybridus (butterbur), magnesium, vitamin B2 (riboflavin), Tanacetum parthenium (feverfew), and coenzyme Q10.¹⁶ Although the strength of evidence for these therapies is limited by small trials, their overall risk of adverse effects is low, and they might be easier for patients to obtain than acupuncture or CBT.

Continue to: Butterbur, in particular...

Butterbur, in particular, has been found to be beneficial for migraine prevention in 2 small placebo-controlled trials. In a randomized controlled study of 245 patients P hybridus, (specifically, the German formulation, Petadolex), 75 mg BID, reduced the frequency of migraine attack by 48% at 4 months, compared to placebo (number needed to treat, 5.3).⁴⁴ No difference was found at lower dosages. The most common reported adverse effect was burping.

Regrettably, unpurified butterbur extract contains pyrrolizidine alkaloids, potentially hepatotoxic and carcinogenic compounds. Because of variations in purification in production facilities in the United States, butterbur supplements might not have all of these compounds removed—and so should be used with caution.⁴¹

Magnesium. Studies evaluating the use of magnesium have demonstrated varied results; differences in methods and dosing have limited broad application of findings. As with most supplements considered for prophylactic treatment, magnesium dosing is poorly understood, and bioavailability varies in its different forms. Oral supplementation can be given as magnesium dicitrate, 600 mg/d.⁴⁵

Recently, products containing various combinations of feverfew, coenzyme Q10, riboflavin, magnesium, and other supplements have shown benefit in early clinical trials.^36,37

Neural stimulation. Over the past few years, a variety of transcutaneous nerve stimulator devices have gained FDA approval for use in migraine prophylaxis. The long-term safety and efficacy of these devices is not yet well understood, but they appear to provide headache relief in the short term and decrease the frequency of headache.³⁸ Use of the noninvasive stimulators is limited today by high cost and poor coverage by US health care insurers.

Continue to: Newly available medical therapy

Newly available medical therapy. The FDA recently approved erenumab, a fully human monoclonal antibody for prevention of migraine in adults. This is the first drug in the CGRP antagonist class to be approved for this indication. Trials of this once-monthly, self-injectable drug show promising results for patients whose migraines have been refractory to other therapies.

A recent large trial evaluated 955 adults with migraine, randomizing them to receive erenumab, 70 mg; erenumab, 140 mg; or placebo over 28 weeks.³⁹ The groups receiving erenumab had a nearly 2-fold higher odds of having their migraine reduced by 50%, compared with placebo (number needed to treat with the 140-mg dose, 4.27). Similar numbers of participants from all groups discontinued the study.³⁹ Phase 3 trials that are not yet formally published have produced similarly beneficial results.^40,46 The FDA has listed injection site reaction and constipation as the most reported adverse effects.⁴⁰

Three other anti-CGRP antibodies are likely to be approved in the near future: fremanezumab, galcanezumab, and eptinezumab.

The approach to migraine in special populations

Management of acute and chronic migraine in children, pregnant women, and older adults requires special attention: Treatment approaches are different than they are for adults 19 to 65 years of age.

Pediatric patients. Migraine is the most common acute and recurrent headache syndrome in children. Headaches differ from those of adult migraine as a result of variations in brain maturation, plasticity, and cognitive development.⁴⁷ Migraine attacks are often of shorter duration in children, lasting 1 to 2 hours, but can still be preceded by visual aura.⁴⁸ Just as with adults, imaging, electroencephalography, lumbar puncture, and routine labs should be considered only if a child has an abnormal neurological exam or other concerning features (TABLE 2^14,15).

Continue to: The general approach to migraine treatment...

The general approach to migraine treatment in the pediatric population includes education of the child and family about symptom management. Acetaminophen, NSAIDs, and triptans are approved for abortive therapy in children and should be used for acute headache relief in the same way that they are used in adults. Oral rizatriptan, the most well studied triptan in the pediatric population, is approved for use in children as young as 6 years⁴⁹; the pediatric dosage is 5 mg/d for patients weighing 20 to 39 kg and 10 mg/d for patients weighing more than 40 kg (same as the adult dosage).

Don’t prescribe triptans for patients with known vascular disease or severe hepatic impairment.

Oral almotriptan and zolmitriptan are also approved for use in children 12 to 17 years of age. Usual dosages are: almotriptan, 12.5 mg at onset, can repeat in 2 hours as needed (maximum dosage, 25 mg/d); and zolmitriptan, 2.5 mg at onset, can repeat in 2 hours as needed (maximum dosage, 10 mg/d).⁵⁰

For children who are unable to swallow pills or who are vomiting, a non-oral route of administration is preferable. Rizatriptan is available as an orally disintegrating tablet. Zolmitriptan is available in a nasal spray at a dose of 5 mg for children 12 years and older. Sumatriptan is not approved for use in patients younger than 18 years; however, recent studies have shown that it might have good efficacy and tolerability.⁵⁰

Daily prophylactic treatment for recurrent migraine in the pediatric population is an evolving subject; published guidelines do not exist. It is reasonable to consider treatment using the same guidelines as those in place for adults.⁵¹ Topiramate, 1 to 2 mg/kg/d, is the only therapy approved by the FDA for episodic migraine preventive therapy in adolescents.⁵⁰

If a patient doesn’t get adequate headache relief from an appropriate dosage of a given triptan, try a different triptan during the next migraine.

Notably, a nonpharmacotherapeutic approach may be more effective for pediatric prevention. In 2017, a large double-blind, placebo-controlled trial investigated the use of amitriptyline, topiramate, and placebo for the treatment of recurrent migraine in children 8 to 17 years of age. An interim analysis of the 328 children enrolled found no significant differences in reduction of headache frequency with treatment compared with placebo over a 24-week period; the trial was stopped early due to futility.⁵²

Continue to: The study did show...

The study did show, however, that reducing migraine triggers provided a high level of benefit to study participants. Stress is one of the most common migraine triggers in children; lack of sleep, exposure to a warm climate, and exposure to video games are also notable triggers.⁵³ CBT may augment the efficacy of standard migraine medications in the pediatric population and may help prevent recurrence of episodes.⁵⁴

Pregnancy. The treatment of migraine is different in pregnant women than it is in nonpregnant adults because of a concern over adverse effects on fetal development. For acute headache treatment, first-line therapies include trigger avoidance and acetaminophen, 1000 mg (maximum dosage, 4000 mg/d).⁵⁵ If this is ineffective, a 10-mg dose of metoclopramide, as often as every 6 hours (not an FDA-approved indication), can be considered. During the second trimester, NSAIDs can be considered second-line therapy.

Triptans—specifically, sumatriptan and rizatriptan—can also be considered if first-line therapies fail.⁵⁶ Triptan-exposed pregnant women with migraine have a rate of congenital malformations, spontaneous abortions, and prematurity that is similar to what is seen in pregnant women with migraine who have not been exposed to triptans. However, when triptan-exposed women are compared with healthy, non-migraine-suffering women, the rate of spontaneous abortion appears to be increased in the triptan-exposed population.⁵⁷

Ergotamine is contraindicated during pregnancy because of its potential to induce uterine contractions and vasospasm, which can be detrimental to the fetus.⁵⁶Nonpharmacotherapeutic interventions such as heat, ice, massage, rest, and avoidance of triggers are as successful in the pregnant population as in the nonpregnant population. For migraine prevention, coenzyme Q10, vitamins B2 and B6 (pyridoxine), and oral magnesium can be considered. Feverfew and butterbur should be avoided because of concerns about fetal malformation and preterm labor.⁵⁸

Older adults. Choosing appropriate migraine therapy for older adults requires special consideration because of changes in drug metabolism and risks associated with drug adverse effects. Additionally, few studies of migraine drugs have included large populations of adults older than 65 years; medications should therefore be prescribed cautiously in this population, with particular attention to drug–drug interactions.

Continue to: Just as for younger adults...

Just as for younger adults, mild symptoms can be managed effectively with acetaminophen. NSAIDs may be used as well, but carry increased risks of gastric bleeding and elevation in blood pressure.⁵⁹ The use of triptans is acceptable for the appropriate patient, but should be avoided in patients with known vascular disease.⁶⁰ Antiemetics present an increased risk of extrapyramidal adverse effects in the elderly and should be used with caution at the lowest effective dosage.⁵⁹ Novel mechanisms of action make some of the newer agents potentially safer for use in older adults when treating acute migraine.

Stress is one of the most common migraine triggers in children.

For migraine prevention in older adults, particular attention should be paid to reducing triggers and minimizing polypharmacy.

More and more, successful treatment is within reach

With many clinical trials evaluating novel drugs underway, and additional studies contributing to our understanding of nonpharmacotherapeutic approaches to migraine treatment, improved headache control may become increasingly common over the next few years.

CORRESPONDENCE
Kathryn McGrath, MD, Department of Family and Community Medicine, Thomas Jefferson University, 1015 Walnut St, Philadelphia PA 19107; [email protected].

Migraine is a highly disabling primary headache disorder that affects more than 44 million Americans annually.¹ The disorder causes pain, photophobia, phonophobia, and nausea that can last for hours, even days. Migraine headaches are 2 times more common in women than in men; although migraine is most common in people 30 to 39 years of age, all ages are affected.^2,3 Frequency of migraine headache is variable; chronic migraineurs experience more than 15 headache days a month.

©Cath Riley/Science Source

Recent estimates indicate that the cost of acute and chronic migraine headaches reaches approximately $78 million a year in the United States. ⁴ This high burden of disease has made effective migraine treatment options absolutely essential. Recent advances in our understanding of migraine pathophysiology have led to new therapeutic targets; there are now many novel treatment approaches on the horizon.

In this article, we review the diagnosis and management of migraine in detail. Our emphasis is on evidence-based approaches to acute and prophylactic treatment, including tried-and-true options and newly emerging therapies.

Neuronal dysfunction and a genetic predisposition

Although migraine was once thought to be caused by abnormalities of vasodilation, current research suggests that the disorder has its origins in primary neuronal dysfunction. There appears to be a genetic predisposition toward widespread neuronal hyperexcitability in migraineurs.⁵ In addition, hypothalamic neurons are thought to initiate migraine by responding to changes in brain homeostasis. Increased parasympathetic tone might activate meningeal pain receptors or lower the threshold for transmitting pain signals from the thalamus to the cortex.⁶

Prodromal symptoms and aura appear to originate from multiple areas across the brain, including the hypothalamus, cortex, limbic system, and brainstem. This widespread brain involvement might explain why some headache sufferers concurrently experience a variety of symptoms, including fatigue, depression, muscle pain, and an abnormal sensitivity to light, sound, and smell.^6,7

After taking the initial history (headache onset, location, duration, associated symptoms), focus attention on assessing the risk of intracranial pathology.

Although the exact mechanisms behind each of these symptoms have yet to be defined precisely, waves of neuronal depolarization—known as cortical spreading depression—are suspected to cause migraine aura.^8-10 Cortical spreading depression activates the trigeminal pain pathway and leads to the release of pro-inflammatory markers such as calcitonin gene-related protein (CGRP).⁶ A better understanding of these complex signaling pathways has helped provide potential therapeutic targets for new migraine drugs.

Diagnosis: Close patient inquiry is most helpful

The International Headache Society (IHS) criteria for primary headache disorders serve as the basis for the diagnosis of migraine and its subtypes, which include migraine without aura and migraine with aura. Due to variability of presentation, migraine with aura is further subdivided into migraine with typical aura (with and without headache), migraine with brainstem aura, hemiplegic migraine, and retinal migraine.¹¹

Continue to: How is migraine defined?

How is migraine defined? Simply, migraine is classically defined as a unilateral, pulsating headache of moderate to severe intensity lasting 4 to 72 hours, associated with photophobia and phonophobia or nausea and vomiting, or both.¹¹ Often visual in nature, aura is a set of neurologic symptoms that lasts for minutes and precedes the onset of the headache. The visual aura is often described as a scintillating scotoma that begins near the point of visual fixation and then spreads left or right. Other aura symptoms include tingling or numbness (second most common), speech disturbance (aphasia), motor changes and, in rare cases, a combination of these in succession. By definition, all of these symptoms fully resolve between attacks.¹¹

Validated valuable questionnaires. To help with accurate and timely diagnosis, researchers have developed and validated simplified questionnaires that can be completed independently by patients presenting to primary care (TABLE 1^12,13):

• ID Migraine is a set of 3 questions that scores positive when a patient endorses at least 2 of the 3 symptoms. ¹²
• MS-Q is similar to the ID Migraine but includes 5 items. A score of ≥4 is a positive screen. ¹³

The sensitivity and specificity of MS-Q (0.93 and 0.81, respectively) are slightly higher than those of ID Migraine (0.81 and 0.75).¹³

Remember POUND. This mnemonic device can also be used during history-taking to aid in diagnostic accuracy. Migraine is highly likely (92%) in patients who endorse 4 of the following 5 symptoms and unlikely (17%) in those who endorse ≤2 symptoms¹⁴: Pulsatile quality of headache 4 to 72 hOurs in duration, Unilateral location, Nausea or vomiting, and Disabling intensity.

Differential Dx. Although the differential diagnosis of headache is broad (TABLE 2^14,15), the history alone can often guide clinicians towards the correct assessment. After taking the initial history (headache onset, location, duration, and associated symptoms), focus your attention on assessing the risk of intracranial pathology. This is best accomplished by assessing specific details of the history (TABLE 3¹⁴) and findings on physical examination¹⁵:

• blood pressure measurement (seated, legs uncrossed, feet flat on the floor; having rested for 5 minutes; arm well supported)
• cranial nerve exam
• extremity strength testing
• eye exam (vision, extra-ocular muscles, visual fields, pupillary reactivity, and funduscopic exam)
• gait (tandem walk)
• reflexes.

Continue to: Further testing needed?

Further testing needed? Neuroimaging should be considered only in patients with an abnormal neurologic exam, atypical headache features, or certain risk factors, such as an immune deficiency. There is no role for electroencephalography or other diagnostic testing in migraine.¹⁶

Take a multipronged approach to treatment

As with other complex, chronic conditions, the treatment of migraine should take a multifaceted approach, including management of acute symptoms as well as prevention of future headaches. In 2015, the American Headache Society published a systematic review that specified particular treatment goals for migraine sufferers. ¹⁷ These goals include:

• headache reduction
• headache relief
• decreased disability from headache
• elimination of nausea and vomiting
• elimination of photophobia and phonophobia.

Our review, which follows, of therapeutic options focuses on the management of migraine in adults. Approaches in special populations (older adults, pregnant women, and children) are discussed afterward.

Pharmacotherapy for acute migraine

Acute migraine should be treated with an abortive medication at the onset of headache. The immediate goal is to relieve pain within 2 hours and prevent its recurrence within the subsequent 48 hours (TABLE 4^12,18-20).

Electroencephalography and other diagnostic testing have no role in the workup of migraine.

In the general population, mild, infrequent migraines can be managed with acetaminophen and nonsteroidal anti-inflammatory drugs (NSAIDs).²¹

Continue to: For moderate-to-severe migraine...

For moderate-to-severe migraine, triptans, which target serotonin receptors, are the drug of choice for most patients.²¹ Triptans are superior to placebo in achieving a pain-free state at 2 and 24 hours after administration; eletriptan has the most desirable outcome, with 68% of patients pain free at 2 hours and 54% pain free at 24 hours.²² Triptans are available as sublingual tablets and nasal sprays, as well as subcutaneous injections for patients with significant associated nausea and vomiting. Avoid prescribing triptans for patients with known vascular disease (eg, history of stroke, myocardial infarction, peripheral vascular disease, uncontrolled hypertension, or signs and symptoms of these conditions), as well as for patients with severe hepatic impairment.

Treat migraine with a multifaceted approach, including management of acute symptoms and prevention of future headaches.

Importantly, although triptans all have a similar mechanism of action, patients might respond differently to different drugs within the class. If a patient does not get adequate headache relief from an appropriate dosage of a given triptan during a particular migraine episode, a different triptan can be tried during the next migraine.²² Additionally, if a patient experiences an adverse effect from one triptan, this does not necessarily mean that a trial of another triptan at a later time is contraindicated.

For patients who have an incomplete response to migraine treatment or for those with frequent recurrence, the combination formulation of sumatriptan, 85 mg, and naproxen, 500 mg, showed the highest rate of resolution of headache within 2 hours compared with either drug alone.²³ A similar result might be found by combining a triptan known to be effective for a patient and an NSAID other than naproxen. If migraine persists despite initial treatment of an attack, a different class of medication should be tried during the course of that attack to attain relief of symptoms of that migraine.²¹

When a patient is seen in an acute care setting (eg, emergency department, urgent care center) while suffering a migraine, additional treatment options are available. Intravenous (IV) anti-emetics are useful for relieving the pain of migraine and nausea, and can be used in combination with an IV NSAID (eg, ketorolac).²¹ The most effective anti-emetics are dopamine receptor type-2 blockers, including chlorpromazine, droperidol, metoclopramide, and prochlorperazine, which has the highest level of efficacy.²⁴ Note that these medications do present the risk of a dystonic reaction; diphenhydramine is therefore often used in tandem to mitigate such a response.

Looking ahead. Although triptans are the current first-line therapy for acute migraine, their effectiveness is limited. Only 20% of patients report sustained relief of pain in the 2 to 24 hours after treatment, and the response can vary from episode to episode.²⁵

Continue to: With better understading of the pathophysiology of migraine...

With better understanding of the pathophysiology of migraine, a host of novel anti-migraine drugs are on the horizon.

CGRP receptor antagonists. The neuropeptide CGRP, which mediates central and peripheral nervous system pain signaling, has been noted to be elevated during acute migraine attacks²⁶; clinical trials are therefore underway to evaluate the safety and efficacy of CGRP receptor antagonists.¹⁸ These agents appear to be better tolerated than triptans, have fewer vascular and central nervous system adverse effects, and present less of a risk of medication overuse headache.¹⁸ Liver toxicity has been seen with some medications in this class and remains an important concern in their development.¹⁹

Phase 3 clinical trials for 1 drug in this class, ubrogepant, were completed in late 2017; full analysis of the data is not yet available. Primary outcomes being evaluated include relief of pain at 2 hours and relief from the most bothersome symptoms again at 2 hours.²⁷

Selective serotonin-HT_1f receptor agonists, such as lasmiditan, offer another potential approach. Although the exact mechanism of action of these agents is not entirely clear, clinical trials have supported their efficacy and safety.²⁰ Importantly, ongoing trials are specifically targeting patients with known cardiovascular risk factors because they are most likely to benefit from the nonvasoconstrictive mechanism of action.^28,29 Adverse effects reported primarily include dizziness, fatigue, and vertigo.

Strategies for managing recurrent episodic migraine

Because of the risk of medication overuse headache with acute treatment, daily preventive therapy for migraine is indicated for any patient with ³⁰ :

• ≥6 headache days a month
• ≥4 headache days a month with some impairment
• ≥3 headache days a month with severe impairment.

Continue to: Treatment begins by having patients identify...

Treatment begins by having patients identify, and then avoid, migraine triggers (TABLE 5). This can be accomplished by having patients keep a headache diary, in which they can enter notations about personal and environmental situations that precede a headache.

For the individual patient, some triggers are modifiable; others are not. Helping a patient develop strategies for coping with triggers, rather than aiming for complete avoidance, might help her (him) manage those that are inescapable (eg stress, menstruation, etc).³¹ For many patients, however, this is not an adequate intervention and other approaches must be explored. When considering which therapy might be best for a given patient, evaluate her (his) comorbidities and assess that particular treatment for potential secondary benefits and the possibility of adverse effects. Pay attention to the choice of preventive therapy in women who are considering pregnancy because many available treatments are potentially teratogenic.

Oral medications. Oral agents from several classes of drugs can be used for migraine prophylaxis, including anti-epileptics,antidepressants, and antihypertensives (TABLE 6^{20,29,30,32-41}). Selected anti-epileptics (divalproex sodium, sodium valproate, topiramate) and beta-blockers (metoprolol, propranolol, and timolol) have the strongest evidence to support their use.³² Overall, regular use of prophylactic medications can reduce headache frequency by 50% for approximately 40% to 45% of patients who take them.²⁹ However, adherence may be limited by adverse effects or perceived lack of efficacy, thus reducing their potential for benefit.⁴²

OnabotulinumtoxinA. In patients with chronic migraine (≥15 headache days a month for at least 3 months) who have failed oral medications, the American Academy of Neurology (AAN) recommends the use of onabotulinumtoxinA.³⁰ The treatment regimen comprises 31 injections at various sites on the head, neck, and shoulders every 3 months.³³

A 2010 large randomized controlled trial showed a decrease in the frequency of headache days for patients receiving onabotulinumtoxinA compared to placebo after a 24-week treatment period (7.8 fewer headache days a month, compared to 6.4 fewer in the placebo group).³³ A recent systematic review also noted a reduction of 2 headache days a month compared with placebo; the authors cautioned, however, that data with which to evaluate onabotulinumtoxinA in comparison to other prophylactic agents are limited.⁴³

Continue to: In both studies...

In both studies, the risk of adverse drug events due to onabotulinumtoxinA was high and led to a significant rate of discontinuation.^33,43 Despite this, onabotulinumtoxinA remains the only Food and Drug Administration (FDA)–approved treatment for chronic migraine, making it reasonable to consider for appropriate patients.

Acupuncture. A 2016 Cochrane review found benefit for patients using acupuncture compared with sham acupuncture.³⁴ When acupuncture was compared with prophylactic agents such as beta-blockers, calcium-channel blockers, and anti-epileptics, however, there was no significant difference between the procedure and pharmacotherapy. Patients willing and able to try acupuncture might see a reduction in the overall number of headaches. Acupuncture has few adverse effects; however, long-term data are lacking.³⁴

Exercise is not supported by robust data for its role as a prophylactic treatment. It is generally considered safe in most populations, however, and can be pursued with little out-of-pocket cost.³⁵

Cognitive behavioral therapy (CBT). The AAN recommends CBT, relaxation therapy, and biofeedback therapy. Accessibility of these services remains limited for many patients, and cost can be prohibitive.¹⁶

Supplements used to help prevent migraine include the root of Petasites hybridus (butterbur), magnesium, vitamin B2 (riboflavin), Tanacetum parthenium (feverfew), and coenzyme Q10.¹⁶ Although the strength of evidence for these therapies is limited by small trials, their overall risk of adverse effects is low, and they might be easier for patients to obtain than acupuncture or CBT.

Continue to: Butterbur, in particular...

Butterbur, in particular, has been found to be beneficial for migraine prevention in 2 small placebo-controlled trials. In a randomized controlled study of 245 patients P hybridus, (specifically, the German formulation, Petadolex), 75 mg BID, reduced the frequency of migraine attack by 48% at 4 months, compared to placebo (number needed to treat, 5.3).⁴⁴ No difference was found at lower dosages. The most common reported adverse effect was burping.

Regrettably, unpurified butterbur extract contains pyrrolizidine alkaloids, potentially hepatotoxic and carcinogenic compounds. Because of variations in purification in production facilities in the United States, butterbur supplements might not have all of these compounds removed—and so should be used with caution.⁴¹

Magnesium. Studies evaluating the use of magnesium have demonstrated varied results; differences in methods and dosing have limited broad application of findings. As with most supplements considered for prophylactic treatment, magnesium dosing is poorly understood, and bioavailability varies in its different forms. Oral supplementation can be given as magnesium dicitrate, 600 mg/d.⁴⁵

Recently, products containing various combinations of feverfew, coenzyme Q10, riboflavin, magnesium, and other supplements have shown benefit in early clinical trials.^36,37

Neural stimulation. Over the past few years, a variety of transcutaneous nerve stimulator devices have gained FDA approval for use in migraine prophylaxis. The long-term safety and efficacy of these devices is not yet well understood, but they appear to provide headache relief in the short term and decrease the frequency of headache.³⁸ Use of the noninvasive stimulators is limited today by high cost and poor coverage by US health care insurers.

Continue to: Newly available medical therapy

Newly available medical therapy. The FDA recently approved erenumab, a fully human monoclonal antibody for prevention of migraine in adults. This is the first drug in the CGRP antagonist class to be approved for this indication. Trials of this once-monthly, self-injectable drug show promising results for patients whose migraines have been refractory to other therapies.

A recent large trial evaluated 955 adults with migraine, randomizing them to receive erenumab, 70 mg; erenumab, 140 mg; or placebo over 28 weeks.³⁹ The groups receiving erenumab had a nearly 2-fold higher odds of having their migraine reduced by 50%, compared with placebo (number needed to treat with the 140-mg dose, 4.27). Similar numbers of participants from all groups discontinued the study.³⁹ Phase 3 trials that are not yet formally published have produced similarly beneficial results.^40,46 The FDA has listed injection site reaction and constipation as the most reported adverse effects.⁴⁰

Three other anti-CGRP antibodies are likely to be approved in the near future: fremanezumab, galcanezumab, and eptinezumab.

The approach to migraine in special populations

Management of acute and chronic migraine in children, pregnant women, and older adults requires special attention: Treatment approaches are different than they are for adults 19 to 65 years of age.

Pediatric patients. Migraine is the most common acute and recurrent headache syndrome in children. Headaches differ from those of adult migraine as a result of variations in brain maturation, plasticity, and cognitive development.⁴⁷ Migraine attacks are often of shorter duration in children, lasting 1 to 2 hours, but can still be preceded by visual aura.⁴⁸ Just as with adults, imaging, electroencephalography, lumbar puncture, and routine labs should be considered only if a child has an abnormal neurological exam or other concerning features (TABLE 2^14,15).

Continue to: The general approach to migraine treatment...

The general approach to migraine treatment in the pediatric population includes education of the child and family about symptom management. Acetaminophen, NSAIDs, and triptans are approved for abortive therapy in children and should be used for acute headache relief in the same way that they are used in adults. Oral rizatriptan, the most well studied triptan in the pediatric population, is approved for use in children as young as 6 years⁴⁹; the pediatric dosage is 5 mg/d for patients weighing 20 to 39 kg and 10 mg/d for patients weighing more than 40 kg (same as the adult dosage).

Don’t prescribe triptans for patients with known vascular disease or severe hepatic impairment.

Oral almotriptan and zolmitriptan are also approved for use in children 12 to 17 years of age. Usual dosages are: almotriptan, 12.5 mg at onset, can repeat in 2 hours as needed (maximum dosage, 25 mg/d); and zolmitriptan, 2.5 mg at onset, can repeat in 2 hours as needed (maximum dosage, 10 mg/d).⁵⁰

For children who are unable to swallow pills or who are vomiting, a non-oral route of administration is preferable. Rizatriptan is available as an orally disintegrating tablet. Zolmitriptan is available in a nasal spray at a dose of 5 mg for children 12 years and older. Sumatriptan is not approved for use in patients younger than 18 years; however, recent studies have shown that it might have good efficacy and tolerability.⁵⁰

Daily prophylactic treatment for recurrent migraine in the pediatric population is an evolving subject; published guidelines do not exist. It is reasonable to consider treatment using the same guidelines as those in place for adults.⁵¹ Topiramate, 1 to 2 mg/kg/d, is the only therapy approved by the FDA for episodic migraine preventive therapy in adolescents.⁵⁰

If a patient doesn’t get adequate headache relief from an appropriate dosage of a given triptan, try a different triptan during the next migraine.

Notably, a nonpharmacotherapeutic approach may be more effective for pediatric prevention. In 2017, a large double-blind, placebo-controlled trial investigated the use of amitriptyline, topiramate, and placebo for the treatment of recurrent migraine in children 8 to 17 years of age. An interim analysis of the 328 children enrolled found no significant differences in reduction of headache frequency with treatment compared with placebo over a 24-week period; the trial was stopped early due to futility.⁵²

Continue to: The study did show...

The study did show, however, that reducing migraine triggers provided a high level of benefit to study participants. Stress is one of the most common migraine triggers in children; lack of sleep, exposure to a warm climate, and exposure to video games are also notable triggers.⁵³ CBT may augment the efficacy of standard migraine medications in the pediatric population and may help prevent recurrence of episodes.⁵⁴

Pregnancy. The treatment of migraine is different in pregnant women than it is in nonpregnant adults because of a concern over adverse effects on fetal development. For acute headache treatment, first-line therapies include trigger avoidance and acetaminophen, 1000 mg (maximum dosage, 4000 mg/d).⁵⁵ If this is ineffective, a 10-mg dose of metoclopramide, as often as every 6 hours (not an FDA-approved indication), can be considered. During the second trimester, NSAIDs can be considered second-line therapy.

Triptans—specifically, sumatriptan and rizatriptan—can also be considered if first-line therapies fail.⁵⁶ Triptan-exposed pregnant women with migraine have a rate of congenital malformations, spontaneous abortions, and prematurity that is similar to what is seen in pregnant women with migraine who have not been exposed to triptans. However, when triptan-exposed women are compared with healthy, non-migraine-suffering women, the rate of spontaneous abortion appears to be increased in the triptan-exposed population.⁵⁷

Ergotamine is contraindicated during pregnancy because of its potential to induce uterine contractions and vasospasm, which can be detrimental to the fetus.⁵⁶Nonpharmacotherapeutic interventions such as heat, ice, massage, rest, and avoidance of triggers are as successful in the pregnant population as in the nonpregnant population. For migraine prevention, coenzyme Q10, vitamins B2 and B6 (pyridoxine), and oral magnesium can be considered. Feverfew and butterbur should be avoided because of concerns about fetal malformation and preterm labor.⁵⁸

Older adults. Choosing appropriate migraine therapy for older adults requires special consideration because of changes in drug metabolism and risks associated with drug adverse effects. Additionally, few studies of migraine drugs have included large populations of adults older than 65 years; medications should therefore be prescribed cautiously in this population, with particular attention to drug–drug interactions.

Continue to: Just as for younger adults...

Just as for younger adults, mild symptoms can be managed effectively with acetaminophen. NSAIDs may be used as well, but carry increased risks of gastric bleeding and elevation in blood pressure.⁵⁹ The use of triptans is acceptable for the appropriate patient, but should be avoided in patients with known vascular disease.⁶⁰ Antiemetics present an increased risk of extrapyramidal adverse effects in the elderly and should be used with caution at the lowest effective dosage.⁵⁹ Novel mechanisms of action make some of the newer agents potentially safer for use in older adults when treating acute migraine.

Stress is one of the most common migraine triggers in children.

For migraine prevention in older adults, particular attention should be paid to reducing triggers and minimizing polypharmacy.

More and more, successful treatment is within reach

With many clinical trials evaluating novel drugs underway, and additional studies contributing to our understanding of nonpharmacotherapeutic approaches to migraine treatment, improved headache control may become increasingly common over the next few years.

CORRESPONDENCE
Kathryn McGrath, MD, Department of Family and Community Medicine, Thomas Jefferson University, 1015 Walnut St, Philadelphia PA 19107; [email protected].

Aplastic anemia is a rare hematologic disorder marked by pancytopenia and a hypocellular marrow. Aplastic anemia results from either inherited or acquired causes, and the treatment approach varies significantly between the 2 causes. This article reviews the treatment of inherited and acquired forms of aplastic anemia. The approach to evaluation and diagnosis of aplastic anemia is reviewed in a separate article.

Inherited Aplastic Anemia

First-line treatment options for patients with inherited marrow failure syndromes (IMFS) are androgen therapy and hematopoietic stem cell transplant (HSCT). When evaluating patients for HSCT, it is critical to identify the presence of an IMFS, as the risk and mortality associated with the conditioning regimen, stem cell source, graft-versus-host disease (GVHD), and secondary malignancies differ between patients with IMFS and those with acquired marrow failure syndromes or hematologic malignancies.

Potential sibling donors need to be screened for donor candidacy as well as for the inherited defect.¹ Among patients with Fanconi anemia or a telomere biology disorder, the stem cell source must be considered, with bone marrow demonstrating lower rates of acute GVHD than a peripheral blood stem cell source.^2-4 In IMFS patients, the donor cell type may affect the choice of conditioning regimen.^5,6 Reduced-intensity conditioning in lieu of myeloablative conditioning without total body irradiation has proved feasible in patients with Fanconi anemia, and is associated with a reduced risk of secondary malignancies.^5,6 Incorporation of fludarabine in the conditioning regimen of patients without a matched sibling donor is associated with superior engraftment and survival^2,5,7 compared to cyclophosphamide conditioning, which was historically used in matched related donors.^6,8 The addition of fludarabine appears to be especially beneficial in older patients, in whom its use is associated with lower rates of graft failure, likely due to increased immunosuppression at the time of engraftment.^7,9 Fludarabine has also been incorporated into conditioning regimens for patients with a telomere biology disorder, but outcomes data is limited.⁵

For patients presenting with acute myeloid leukemia (AML) or a high-risk myelodysplastic syndrome (MDS) who are subsequently diagnosed with an IMFS, treatment can be more complex, as these patients are at high risk for toxicity from standard chemotherapy. Limited data suggests that induction therapy and transplantation are feasible in this group of patients, and this approach is associated with increased overall survival (OS) despite lower OS rates than those of IMFS patients who present prior to the development of MDS or AML.^10,11 Further work is needed to determine the optimal induction regimen that balances the risks of treatment-related mortality and complications associated with conditioning regimens, risk of relapse, and risk of secondary malignancies, especially in the cohort of patients diagnosed at an older age.

Acquired Aplastic Anemia

Supportive Care

While the workup and treatment plan is being established, attention should be directed at supportive care for prevention of complications. The most common complications leading to death in patients with significant pancytopenia and neutropenia are opportunistic infections and hemorrhagic complications.¹²

Transfusion support is critical to avoid symptomatic anemia and hemorrhagic complications related to thrombocytopenia, which typically occur with platelet counts lower than 10,000 cells/µL. However, transfusion carries the risk of alloimmunization (which may persist for years following transfusion) and transfusion-related graft versus host disease (trGVHD), and thus use of transfusion should be minimized when possible.^13,14 Transfusion support is often required to prevent complications associated with thrombocytopenia and anemia; all blood products given to patients with aplastic anemia should be irradiated and leukoreduced to reduce the risk of both alloimmunization and trGVHD. Guidelines from the British Society for Haematology recommend routine screening for Rh and Kell antibodies to reduce the risk of alloimmunization.¹⁵ Infectious complications remain a common cause of morbidity and mortality in patients with aplastic anemia who have prolonged neutropenia (defined as an absolute neutrophil count [ANC] < 500 cells/µL).^16-19 Therefore, patients should receive broad-spectrum antibiotics with antipseudomonal coverage. In a study by Tichelli and colleagues evaluating the role of granulocyte-colony stimulating factor (G-CSF) in patients with SAA receiving immunosuppressive therapy, 55% of all patient deaths were secondary to infection.²⁰ There was no OS benefit seen in patients who received G-CSF, though a significantly lower rate of infection was observed in the G-CSF arm compared to those not receiving G-CSF (56% versus 81%, P = 0.006).This difference was largely driven by a decrease in infectious episodes in patients with very severe aplastic anemia (VSAA) treated with G-CSF as compared to those who did not receive this therapy (22% versus 48%, P = 0.014).²⁰

Angio-invasive pulmonary aspergillosis and Zygomycetes (eg, Rhizopus, Mucor species) remain major causes of mortality related to opportunistic mycotic infections in patients with aplastic anemia.¹⁸ The infectious risk is directly related to the duration and severity of neutropenia, with one study demonstrating a significant increase in risk in AML patients with neutropenia lasting longer than 3 weeks.²¹ Invasive fungal infections carry a high mortality in patients with severe neutropenia, though due to earlier recognition and empiric antifungal therapy with extended-spectrum azoles, overall mortality secondary to invasive fungal infections is declining.^19,22

While neutropenia related to cytotoxic chemotherapy is commonly associated with gram-negative bacteria due to disruption of mucosal barriers, patients with aplastic anemia have an increased incidence of gram-positive bacteremia with staphylococcal species compared to other neutropenic populations.^18,19 This appears to be changing with time. Valdez and colleagues demonstrated a decrease in prevalence of coagulase-negative staphylococcal infections, increased prevalence of gram-positive bacilli bacteremia, and no change in prevalence of gram-negative bacteremia in patients with aplastic anemia treated between 1989 and 2008.²² Gram-negative bacteremia caused by Stenotrophomonas maltophila, Escherichia coli, Klebsiella pneumoniae, Citrobacter, and Proteus has also been reported.¹⁹ Despite a lack of clinical trials investigating the role of antifungal and antibacterial prophylaxis for patients with aplastic anemia, most centers initiate antifungal prophylaxis in patients with severe aplastic anema (SAA) or VSAA with an anti-mold agent such as voriconazole or posaconazole (which has the additional benefit compared to voriconazole of covering Mucor species).^17,23 This is especially true for patients who have received ATG or undergone HSCT. For antimicrobial prophylaxis, a fluoroquinolone antibiotic with a spectrum of activity against Pseudomonas should be considered for patients with an ANC < 500 cells/µL.¹⁷ Acyclovir or valacyclovir prophylaxis is recommended for varicella-zoster virus and herpes simplex virus. Cytomegalovirus reactivation is minimal in patients with aplastic anemia, unless multiple courses of ATG are used.

Iron overload is another complication the provider must be aware of in the setting of increased transfusions in aplastic anemia patients. Lee and colleagues demonstrated that iron chelation therapy using deferasirox is effective at reducing serum ferritin levels in patients with aplastic anemia (median ferritin level: 3254 ng/mL prior to therapy, 1854 ng/mL following), and is associated with no serious adverse events (most common adverse events included nausea, diarrhea, vomiting, and rash).²⁴ Approximately 25% of patients in this trial demonstrated an increase in creatinine, with patients taking concomitant cyclosporine more affected than those on chelation therapy alone.²⁴ For patients following HSCT or with improved hematopoiesis following immunosuppressive therapy, phlebotomy can be used to treat iron overload in lieu of chelation therapy.¹⁵

Approach to Therapy

The main treatment options for SAA and VSAA include allogeneic bone marrow transplant and immunosuppression. The deciding factors as to which treatment is best initially depends on the availability of HLA-matched related donors and age (Figure 1 and Figure 2). Survival is decreased in patients with SAA or VSAA who delay initiation of therapy, and therefore prompt referral for HLA typing and evaluation for bone marrow transplant is a very important first step in managing aplastic anemia.

Matched Sibling Donor Transplant

Current standards of care recommend HLA-matched sibling donor transplant for patients with SAA or VSAA who are younger than 50 years of age, with the caveat that integration of fludarabine and reduced cyclophosphamide dosing along with ATG shows the best overall outcomes. Locasciulli and colleagues examined outcomes in patients given either immunosuppressive therapy or sibling HSCT between 1991-1996 and 1997-2002, respectively, and found that sibling HSCT was associated with a superior 10-year OS compared to immunosuppressive therapy (73% versus 68%).²⁵ Interestingly in this study, there was no OS improvement seen with immunosuppressive therapy alone (69% versus 73%) between the 2 time periods, despite increased OS in both sibling HSCT (74% and 80%) and matched unrelated donor HSCT (38% and 65%).²⁵ Though total body irradiation has been used in the past, it is typically not included in current conditioning regimens for matched related donor transplants.²⁶

Current conditioning regimens typically use a combination of cyclophosphamide and ATG^27,28 with or without fludarabine. Fludarabine-based conditioning regimens have shown promise in patients undergoing sibling HSCT. Maury and colleagues evaluated the role of fludarabine in addition to low-dose cyclophosphamide and ATG compared to cyclophosphamide alone or in combination with ATG in patients over age 30 undergoing sibling HSCT.⁹ There was a nonsignificant improvement in 5-year OS in the fludarabine arm compared to controls (77% ± 8% versus 60% ± 3%, P = 0.14) in the pooled analysis, but when adjusted for age the fludarabine arm had a significantly lower relative risk (RR) of death (RR, 0.44; P = 0.04) compared to the control arm. Shin et al reported outcomes with fludarabine/cyclophosphamide/ATG, with excellent overall outcomes and no difference in patients older or younger than 40 years.²⁹ In addition, Kim et al evaluated their experience with patients older than 40 years of age receiving matched related donors, finding comparable outcomes in those aged 41 to 50 years compared to younger patients. Outcomes did decline in those over the age of 50 years.³⁰ Long-term data for matched related donor transplant for aplastic anemia shows excellent long-term outcomes, with minimal chronic GVHD and good performance status.³¹ Hence, these factors support the role of matched related donor transplant as the initial treatment in SAA and VSAA.

Regarding the role of transplant for patients who lack a matched related donor, a growing body of literature demonstrating identical outcomes between matched related and matched unrelated donor (MUD) transplants for pediatric patients^32,33 supports recent recommendations for upfront unrelated donor transplantation for aplastic anemia.^34,35

Immunosuppressive Therapy

For patients without an HLA-matched sibling donor or those who are older than 50 years of age, immunosuppressive therapy is the first-line therapy. ATG and cyclosporine A are the treatments of choice.³⁶ The potential effectiveness of immunosuppressive therapy in treating aplastic anemia was initially observed in patients in whom autologous transplant failed but who still experienced hematopoietic reconstitution despite the failed graft; this observation led to the hypothesis that the conditioning regimen may have an effect on hematopoiesis.^16,36,37

Anti-thymocyte globulin. Immunosuppressive therapy with ATG has been used for the treatment of aplastic anemia since the 1980s.³⁸ Historically, rabbit ATG had been used, but a 2011 study of horse ATG demonstrated superior hematological response at 6 months compared to rabbit ATG (68% versus 37%).¹⁶ Superior survival was also seen with horse ATG compared to rabbit ATG (3-year OS: 96% versus 76%). Due to these results, horse ATG is preferred over rabbit ATG. ATG should be used in combination with cyclosporine A to optimize outcomes.

Cyclosporine A. Early studies also demonstrated the efficacy of cyclosporine A in the treatment of aplastic anemia, with response rates equivalent to that of ATG monotherapy.³⁹ Recent publications still note the efficacy of cyclosporine A in the treatment of aplastic anemia. Its role as an affordable option for single-agent therapy in developing countries is intriguing.³⁹

The combination of the ATG and cyclosporine A was proven superior to either agent alone in a study by Frickhofen et al.³⁷ In this study patients were randomly assigned to a control arm that received ATG plus methylprednisolone or to an arm that received ATG plus cyclosporine A and methylprednisolone. At 6 months, 70% of patients in the cyclosporine A arm had a complete remission (CR) or partial remission compared to 46% in the control arm.⁴⁰ Further work confirmed the long-term efficacy of this regimen, reporting a 7-year OS of 55%.⁴¹ Among a pediatric population, immunosuppressive therapy was associated with an 83% 10-year OS.⁴²

It is recommended that patients remain on cyclosporine therapy for a minimum of 6 months, after which a gradual taper may be considered, although there is variation among practitioners, with some continuing immunosuppressive therapy for a minimum of 12 months due to a proportion of patients being cyclosporine dependent.^42,43 A study found that within a population of patients who responded to immunosuppressive therapy, 18% became cyclosporine dependent.⁴² The median duration of cyclosporine A treatment at full dose was 12 months, with tapering completed over a median of 19 months after patients had been in a stable CR for a minimum of 3 months. Relapse occurred more often when patients were tapered quickly (decrease ≥ 0.8 mg/kg/month) compared to slowly (0.4-0.7 mg/kg/month) or very slowly (< 0.3 mg/kg/month).

Immunosuppressive therapy plus eltrombopag. Townsley and colleagues recently investigated incorporating the use of the thrombopoietin receptor agonist eltrombopag with immunosuppressive therapy as first-line therapy in aplastic anemia.⁴⁴ When given at a dose of 150 mg daily in patients aged 12 years and older or 75 mg daily in patients younger than 12 years, in conjunction with cyclosporine A and ATG, patients demonstrated markedly improved hematological response compared to historical treatment with standard immunosuppressive therapy alone.⁴⁴ In the patient cohort administered eltrombopag starting on day 1 and continuing for 6 months, the complete response rate was 58%. Eltrombopag led to improvement in all cell lines among all treatment subgroups, and OS (censored for patients who proceeded to transplant) was 99% at 2 years.⁴⁵ Overall, toxicities associated with this therapy were low, with liver enzyme elevations most commonly observed.⁴⁴ Recently, a phase 2 trial of immunosuppressive therapy with or without eltrombopag was reported. Of the 38 patients enrolled, overall response, complete response, and time to response were not statistically different.⁴⁶ With this recent finding, the role of eltrombopag in addition to immunosuppressive therapy is not clearly defined, and further studies are warranted.

OS for patients who do not respond to immunosuppressive therapy is approximately 57% at 5 years, largely due to improved supportive measures among this patient population.^4,22 Therefore, it is important to recognize those patients who have a low chance of response so that second-line therapy can be pursued to improve outcomes.

Matched Unrelated Donor Transplant

For patients with refractory disease following immunosuppressive therapy who lack a matched sibling donor, MUD HSCT is considered standard therapy given the marked improvement in overall outcomes with modulating conditioning regimens and high-resolution HLA typing. A European Society for Blood and Marrow Transplantation analysis comparing matched sibling HSCT to MUD HSCT noted significantly higher rates of acute grade II-IV and grade III-V GVHD (grade II-IV 13% versus 25%, grade III-IV 5% versus 10%) among patients undergoing MUD transplant.⁴⁷ Chronic GVHD rates were 14% in the sibling group, as compared to 26% in the MUD group. Additional benefits seen in this analysis included improved survival when transplanted under age 20 years (84% versus 72%), when transplanted within 6 months of diagnosis (85% versus 72%), the use of ATG in the conditioning regimen (81% versus 73%), and when the donor and recipient were cytomegalovirus-negative compared to other combinations (82% versus 76%).⁴⁷ Interestingly, this study demonstrated that OS was not significantly increased when using a sibling HSCT compared to a MUD HSCT, likely as a result of improved understanding of conditioning regimens, GVHD prophylaxis, and supportive care.

Additional studies of MUD HSCT have shown outcomes similar to those seen in sibling HSCT.^4,43 A French study found a significant increase in survival in patients undergoing MUD HSCT compared to historical cohorts (2000-2005: OS 52%; 2006-2012: OS 74%).³³ The majority of patients underwent conditioning with cyclophosphamide or a combination of busulfan and cyclophosphamide, with or without fludarabine; 81% of patients included underwent in vivo T-cell depletion, and a bone marrow donor source was utilized. OS was significantly lower in patients over age 30 years undergoing MUD HSCT (57%) compared to those under age 30 years (70%). Improved OS was also seen when patients underwent transplant within 1 year of diagnosis and when a 10/10 matched donor (compared to a 9/10 mismatched donor) was utilized.⁴ A 2015 study investigated the role of MUD HSCT as frontline therapy instead of immunosuppressive therapy in patients without a matched sibling donor.³³ The 2-year OS was 96% in the MUD HSCT cohort compared to 91%, 94%, and 74% in historical cohorts of sibling HSCT, frontline immunosuppressive therapy, and second-line MUD HSCT following failed immunosuppressive therapy, respectively. Additionally, event-free survival in the MUD HSCT cohort (defined by the authors as death, lack of response, relapse, occurrence of clonal evolution/clinical paroxysmal nocturnal hemoglobinuria, malignancies developing over follow‐up, and transplant for patients receiving immunosuppressive therapy frontline) was similar compared to sibling HSCT and superior to frontline immunosuppressive therapy and second-line MUD HSCT. Furthermore, Samarasinghe et al highlighted the importance of in vivo T-cell depletion with either ATG or alemtuzumab (anti-CD52 monoclonal antibody) in the prevention of acute and chronic GVHD in both sibling HSCT and MUD HSCT.⁴⁸

With continued improvement of less toxic and more immunomodulating conditioning regimens, utilization of bone marrow as a donor cell source, in vivo T-cell depletion, and use of GVHD and antimicrobial prophylaxis, more clinical evidence supports elevating MUD HSCT in the treatment plan for patients without a matched sibling donor.⁴⁹ However, there is still a large population of patients without matched sibling or unrelated donor options. In an effort to expand the transplant pool and thus avoid clonal hematopoiesis, clinically significant paroxysmal nocturnal hemoglobinuria, and relapsed aplastic anemia, more work continues to recognize the expanding role of alternative donor transplants (cord blood and haploidentical) as another viable treatment strategy for aplastic anemia after immunosuppressive therapy failure.⁵⁰

Summary

Aplastic anemia is a rare but potentially life-threatening disorder characterized by pancytopenia and a marked reduction in the hematopoietic stem cell compartment. Treatment should be instituted as soon as the dignosis of aplastic anemia is established. Treatment outcomes are excellent with modern supportive care and the current approach to allogeneic transplantation, and therefore referral to a bone marrow transplant program to evaluate for early transplantation is the new standard of care.

Aplastic anemia is a rare hematologic disorder marked by pancytopenia and a hypocellular marrow. Aplastic anemia results from either inherited or acquired causes, and the treatment approach varies significantly between the 2 causes. This article reviews the treatment of inherited and acquired forms of aplastic anemia. The approach to evaluation and diagnosis of aplastic anemia is reviewed in a separate article.

Inherited Aplastic Anemia

First-line treatment options for patients with inherited marrow failure syndromes (IMFS) are androgen therapy and hematopoietic stem cell transplant (HSCT). When evaluating patients for HSCT, it is critical to identify the presence of an IMFS, as the risk and mortality associated with the conditioning regimen, stem cell source, graft-versus-host disease (GVHD), and secondary malignancies differ between patients with IMFS and those with acquired marrow failure syndromes or hematologic malignancies.

Potential sibling donors need to be screened for donor candidacy as well as for the inherited defect.¹ Among patients with Fanconi anemia or a telomere biology disorder, the stem cell source must be considered, with bone marrow demonstrating lower rates of acute GVHD than a peripheral blood stem cell source.^2-4 In IMFS patients, the donor cell type may affect the choice of conditioning regimen.^5,6 Reduced-intensity conditioning in lieu of myeloablative conditioning without total body irradiation has proved feasible in patients with Fanconi anemia, and is associated with a reduced risk of secondary malignancies.^5,6 Incorporation of fludarabine in the conditioning regimen of patients without a matched sibling donor is associated with superior engraftment and survival^2,5,7 compared to cyclophosphamide conditioning, which was historically used in matched related donors.^6,8 The addition of fludarabine appears to be especially beneficial in older patients, in whom its use is associated with lower rates of graft failure, likely due to increased immunosuppression at the time of engraftment.^7,9 Fludarabine has also been incorporated into conditioning regimens for patients with a telomere biology disorder, but outcomes data is limited.⁵

For patients presenting with acute myeloid leukemia (AML) or a high-risk myelodysplastic syndrome (MDS) who are subsequently diagnosed with an IMFS, treatment can be more complex, as these patients are at high risk for toxicity from standard chemotherapy. Limited data suggests that induction therapy and transplantation are feasible in this group of patients, and this approach is associated with increased overall survival (OS) despite lower OS rates than those of IMFS patients who present prior to the development of MDS or AML.^10,11 Further work is needed to determine the optimal induction regimen that balances the risks of treatment-related mortality and complications associated with conditioning regimens, risk of relapse, and risk of secondary malignancies, especially in the cohort of patients diagnosed at an older age.

Acquired Aplastic Anemia

Supportive Care

While the workup and treatment plan is being established, attention should be directed at supportive care for prevention of complications. The most common complications leading to death in patients with significant pancytopenia and neutropenia are opportunistic infections and hemorrhagic complications.¹²

Transfusion support is critical to avoid symptomatic anemia and hemorrhagic complications related to thrombocytopenia, which typically occur with platelet counts lower than 10,000 cells/µL. However, transfusion carries the risk of alloimmunization (which may persist for years following transfusion) and transfusion-related graft versus host disease (trGVHD), and thus use of transfusion should be minimized when possible.^13,14 Transfusion support is often required to prevent complications associated with thrombocytopenia and anemia; all blood products given to patients with aplastic anemia should be irradiated and leukoreduced to reduce the risk of both alloimmunization and trGVHD. Guidelines from the British Society for Haematology recommend routine screening for Rh and Kell antibodies to reduce the risk of alloimmunization.¹⁵ Infectious complications remain a common cause of morbidity and mortality in patients with aplastic anemia who have prolonged neutropenia (defined as an absolute neutrophil count [ANC] < 500 cells/µL).^16-19 Therefore, patients should receive broad-spectrum antibiotics with antipseudomonal coverage. In a study by Tichelli and colleagues evaluating the role of granulocyte-colony stimulating factor (G-CSF) in patients with SAA receiving immunosuppressive therapy, 55% of all patient deaths were secondary to infection.²⁰ There was no OS benefit seen in patients who received G-CSF, though a significantly lower rate of infection was observed in the G-CSF arm compared to those not receiving G-CSF (56% versus 81%, P = 0.006).This difference was largely driven by a decrease in infectious episodes in patients with very severe aplastic anemia (VSAA) treated with G-CSF as compared to those who did not receive this therapy (22% versus 48%, P = 0.014).²⁰

Angio-invasive pulmonary aspergillosis and Zygomycetes (eg, Rhizopus, Mucor species) remain major causes of mortality related to opportunistic mycotic infections in patients with aplastic anemia.¹⁸ The infectious risk is directly related to the duration and severity of neutropenia, with one study demonstrating a significant increase in risk in AML patients with neutropenia lasting longer than 3 weeks.²¹ Invasive fungal infections carry a high mortality in patients with severe neutropenia, though due to earlier recognition and empiric antifungal therapy with extended-spectrum azoles, overall mortality secondary to invasive fungal infections is declining.^19,22

While neutropenia related to cytotoxic chemotherapy is commonly associated with gram-negative bacteria due to disruption of mucosal barriers, patients with aplastic anemia have an increased incidence of gram-positive bacteremia with staphylococcal species compared to other neutropenic populations.^18,19 This appears to be changing with time. Valdez and colleagues demonstrated a decrease in prevalence of coagulase-negative staphylococcal infections, increased prevalence of gram-positive bacilli bacteremia, and no change in prevalence of gram-negative bacteremia in patients with aplastic anemia treated between 1989 and 2008.²² Gram-negative bacteremia caused by Stenotrophomonas maltophila, Escherichia coli, Klebsiella pneumoniae, Citrobacter, and Proteus has also been reported.¹⁹ Despite a lack of clinical trials investigating the role of antifungal and antibacterial prophylaxis for patients with aplastic anemia, most centers initiate antifungal prophylaxis in patients with severe aplastic anema (SAA) or VSAA with an anti-mold agent such as voriconazole or posaconazole (which has the additional benefit compared to voriconazole of covering Mucor species).^17,23 This is especially true for patients who have received ATG or undergone HSCT. For antimicrobial prophylaxis, a fluoroquinolone antibiotic with a spectrum of activity against Pseudomonas should be considered for patients with an ANC < 500 cells/µL.¹⁷ Acyclovir or valacyclovir prophylaxis is recommended for varicella-zoster virus and herpes simplex virus. Cytomegalovirus reactivation is minimal in patients with aplastic anemia, unless multiple courses of ATG are used.

Iron overload is another complication the provider must be aware of in the setting of increased transfusions in aplastic anemia patients. Lee and colleagues demonstrated that iron chelation therapy using deferasirox is effective at reducing serum ferritin levels in patients with aplastic anemia (median ferritin level: 3254 ng/mL prior to therapy, 1854 ng/mL following), and is associated with no serious adverse events (most common adverse events included nausea, diarrhea, vomiting, and rash).²⁴ Approximately 25% of patients in this trial demonstrated an increase in creatinine, with patients taking concomitant cyclosporine more affected than those on chelation therapy alone.²⁴ For patients following HSCT or with improved hematopoiesis following immunosuppressive therapy, phlebotomy can be used to treat iron overload in lieu of chelation therapy.¹⁵

Approach to Therapy

The main treatment options for SAA and VSAA include allogeneic bone marrow transplant and immunosuppression. The deciding factors as to which treatment is best initially depends on the availability of HLA-matched related donors and age (Figure 1 and Figure 2). Survival is decreased in patients with SAA or VSAA who delay initiation of therapy, and therefore prompt referral for HLA typing and evaluation for bone marrow transplant is a very important first step in managing aplastic anemia.

Matched Sibling Donor Transplant

Current standards of care recommend HLA-matched sibling donor transplant for patients with SAA or VSAA who are younger than 50 years of age, with the caveat that integration of fludarabine and reduced cyclophosphamide dosing along with ATG shows the best overall outcomes. Locasciulli and colleagues examined outcomes in patients given either immunosuppressive therapy or sibling HSCT between 1991-1996 and 1997-2002, respectively, and found that sibling HSCT was associated with a superior 10-year OS compared to immunosuppressive therapy (73% versus 68%).²⁵ Interestingly in this study, there was no OS improvement seen with immunosuppressive therapy alone (69% versus 73%) between the 2 time periods, despite increased OS in both sibling HSCT (74% and 80%) and matched unrelated donor HSCT (38% and 65%).²⁵ Though total body irradiation has been used in the past, it is typically not included in current conditioning regimens for matched related donor transplants.²⁶

Current conditioning regimens typically use a combination of cyclophosphamide and ATG^27,28 with or without fludarabine. Fludarabine-based conditioning regimens have shown promise in patients undergoing sibling HSCT. Maury and colleagues evaluated the role of fludarabine in addition to low-dose cyclophosphamide and ATG compared to cyclophosphamide alone or in combination with ATG in patients over age 30 undergoing sibling HSCT.⁹ There was a nonsignificant improvement in 5-year OS in the fludarabine arm compared to controls (77% ± 8% versus 60% ± 3%, P = 0.14) in the pooled analysis, but when adjusted for age the fludarabine arm had a significantly lower relative risk (RR) of death (RR, 0.44; P = 0.04) compared to the control arm. Shin et al reported outcomes with fludarabine/cyclophosphamide/ATG, with excellent overall outcomes and no difference in patients older or younger than 40 years.²⁹ In addition, Kim et al evaluated their experience with patients older than 40 years of age receiving matched related donors, finding comparable outcomes in those aged 41 to 50 years compared to younger patients. Outcomes did decline in those over the age of 50 years.³⁰ Long-term data for matched related donor transplant for aplastic anemia shows excellent long-term outcomes, with minimal chronic GVHD and good performance status.³¹ Hence, these factors support the role of matched related donor transplant as the initial treatment in SAA and VSAA.

Regarding the role of transplant for patients who lack a matched related donor, a growing body of literature demonstrating identical outcomes between matched related and matched unrelated donor (MUD) transplants for pediatric patients^32,33 supports recent recommendations for upfront unrelated donor transplantation for aplastic anemia.^34,35

Immunosuppressive Therapy

For patients without an HLA-matched sibling donor or those who are older than 50 years of age, immunosuppressive therapy is the first-line therapy. ATG and cyclosporine A are the treatments of choice.³⁶ The potential effectiveness of immunosuppressive therapy in treating aplastic anemia was initially observed in patients in whom autologous transplant failed but who still experienced hematopoietic reconstitution despite the failed graft; this observation led to the hypothesis that the conditioning regimen may have an effect on hematopoiesis.^16,36,37

Anti-thymocyte globulin. Immunosuppressive therapy with ATG has been used for the treatment of aplastic anemia since the 1980s.³⁸ Historically, rabbit ATG had been used, but a 2011 study of horse ATG demonstrated superior hematological response at 6 months compared to rabbit ATG (68% versus 37%).¹⁶ Superior survival was also seen with horse ATG compared to rabbit ATG (3-year OS: 96% versus 76%). Due to these results, horse ATG is preferred over rabbit ATG. ATG should be used in combination with cyclosporine A to optimize outcomes.

Cyclosporine A. Early studies also demonstrated the efficacy of cyclosporine A in the treatment of aplastic anemia, with response rates equivalent to that of ATG monotherapy.³⁹ Recent publications still note the efficacy of cyclosporine A in the treatment of aplastic anemia. Its role as an affordable option for single-agent therapy in developing countries is intriguing.³⁹

The combination of the ATG and cyclosporine A was proven superior to either agent alone in a study by Frickhofen et al.³⁷ In this study patients were randomly assigned to a control arm that received ATG plus methylprednisolone or to an arm that received ATG plus cyclosporine A and methylprednisolone. At 6 months, 70% of patients in the cyclosporine A arm had a complete remission (CR) or partial remission compared to 46% in the control arm.⁴⁰ Further work confirmed the long-term efficacy of this regimen, reporting a 7-year OS of 55%.⁴¹ Among a pediatric population, immunosuppressive therapy was associated with an 83% 10-year OS.⁴²

It is recommended that patients remain on cyclosporine therapy for a minimum of 6 months, after which a gradual taper may be considered, although there is variation among practitioners, with some continuing immunosuppressive therapy for a minimum of 12 months due to a proportion of patients being cyclosporine dependent.^42,43 A study found that within a population of patients who responded to immunosuppressive therapy, 18% became cyclosporine dependent.⁴² The median duration of cyclosporine A treatment at full dose was 12 months, with tapering completed over a median of 19 months after patients had been in a stable CR for a minimum of 3 months. Relapse occurred more often when patients were tapered quickly (decrease ≥ 0.8 mg/kg/month) compared to slowly (0.4-0.7 mg/kg/month) or very slowly (< 0.3 mg/kg/month).

Immunosuppressive therapy plus eltrombopag. Townsley and colleagues recently investigated incorporating the use of the thrombopoietin receptor agonist eltrombopag with immunosuppressive therapy as first-line therapy in aplastic anemia.⁴⁴ When given at a dose of 150 mg daily in patients aged 12 years and older or 75 mg daily in patients younger than 12 years, in conjunction with cyclosporine A and ATG, patients demonstrated markedly improved hematological response compared to historical treatment with standard immunosuppressive therapy alone.⁴⁴ In the patient cohort administered eltrombopag starting on day 1 and continuing for 6 months, the complete response rate was 58%. Eltrombopag led to improvement in all cell lines among all treatment subgroups, and OS (censored for patients who proceeded to transplant) was 99% at 2 years.⁴⁵ Overall, toxicities associated with this therapy were low, with liver enzyme elevations most commonly observed.⁴⁴ Recently, a phase 2 trial of immunosuppressive therapy with or without eltrombopag was reported. Of the 38 patients enrolled, overall response, complete response, and time to response were not statistically different.⁴⁶ With this recent finding, the role of eltrombopag in addition to immunosuppressive therapy is not clearly defined, and further studies are warranted.

OS for patients who do not respond to immunosuppressive therapy is approximately 57% at 5 years, largely due to improved supportive measures among this patient population.^4,22 Therefore, it is important to recognize those patients who have a low chance of response so that second-line therapy can be pursued to improve outcomes.

Matched Unrelated Donor Transplant

For patients with refractory disease following immunosuppressive therapy who lack a matched sibling donor, MUD HSCT is considered standard therapy given the marked improvement in overall outcomes with modulating conditioning regimens and high-resolution HLA typing. A European Society for Blood and Marrow Transplantation analysis comparing matched sibling HSCT to MUD HSCT noted significantly higher rates of acute grade II-IV and grade III-V GVHD (grade II-IV 13% versus 25%, grade III-IV 5% versus 10%) among patients undergoing MUD transplant.⁴⁷ Chronic GVHD rates were 14% in the sibling group, as compared to 26% in the MUD group. Additional benefits seen in this analysis included improved survival when transplanted under age 20 years (84% versus 72%), when transplanted within 6 months of diagnosis (85% versus 72%), the use of ATG in the conditioning regimen (81% versus 73%), and when the donor and recipient were cytomegalovirus-negative compared to other combinations (82% versus 76%).⁴⁷ Interestingly, this study demonstrated that OS was not significantly increased when using a sibling HSCT compared to a MUD HSCT, likely as a result of improved understanding of conditioning regimens, GVHD prophylaxis, and supportive care.

Additional studies of MUD HSCT have shown outcomes similar to those seen in sibling HSCT.^4,43 A French study found a significant increase in survival in patients undergoing MUD HSCT compared to historical cohorts (2000-2005: OS 52%; 2006-2012: OS 74%).³³ The majority of patients underwent conditioning with cyclophosphamide or a combination of busulfan and cyclophosphamide, with or without fludarabine; 81% of patients included underwent in vivo T-cell depletion, and a bone marrow donor source was utilized. OS was significantly lower in patients over age 30 years undergoing MUD HSCT (57%) compared to those under age 30 years (70%). Improved OS was also seen when patients underwent transplant within 1 year of diagnosis and when a 10/10 matched donor (compared to a 9/10 mismatched donor) was utilized.⁴ A 2015 study investigated the role of MUD HSCT as frontline therapy instead of immunosuppressive therapy in patients without a matched sibling donor.³³ The 2-year OS was 96% in the MUD HSCT cohort compared to 91%, 94%, and 74% in historical cohorts of sibling HSCT, frontline immunosuppressive therapy, and second-line MUD HSCT following failed immunosuppressive therapy, respectively. Additionally, event-free survival in the MUD HSCT cohort (defined by the authors as death, lack of response, relapse, occurrence of clonal evolution/clinical paroxysmal nocturnal hemoglobinuria, malignancies developing over follow‐up, and transplant for patients receiving immunosuppressive therapy frontline) was similar compared to sibling HSCT and superior to frontline immunosuppressive therapy and second-line MUD HSCT. Furthermore, Samarasinghe et al highlighted the importance of in vivo T-cell depletion with either ATG or alemtuzumab (anti-CD52 monoclonal antibody) in the prevention of acute and chronic GVHD in both sibling HSCT and MUD HSCT.⁴⁸

With continued improvement of less toxic and more immunomodulating conditioning regimens, utilization of bone marrow as a donor cell source, in vivo T-cell depletion, and use of GVHD and antimicrobial prophylaxis, more clinical evidence supports elevating MUD HSCT in the treatment plan for patients without a matched sibling donor.⁴⁹ However, there is still a large population of patients without matched sibling or unrelated donor options. In an effort to expand the transplant pool and thus avoid clonal hematopoiesis, clinically significant paroxysmal nocturnal hemoglobinuria, and relapsed aplastic anemia, more work continues to recognize the expanding role of alternative donor transplants (cord blood and haploidentical) as another viable treatment strategy for aplastic anemia after immunosuppressive therapy failure.⁵⁰

Summary

Aplastic anemia is a rare but potentially life-threatening disorder characterized by pancytopenia and a marked reduction in the hematopoietic stem cell compartment. Treatment should be instituted as soon as the dignosis of aplastic anemia is established. Treatment outcomes are excellent with modern supportive care and the current approach to allogeneic transplantation, and therefore referral to a bone marrow transplant program to evaluate for early transplantation is the new standard of care.

Aplastic anemia is a rare hematologic disorder marked by pancytopenia and a hypocellular marrow. Aplastic anemia results from either inherited or acquired causes, and the treatment approach varies significantly between the 2 causes. This article reviews the treatment of inherited and acquired forms of aplastic anemia. The approach to evaluation and diagnosis of aplastic anemia is reviewed in a separate article.

Inherited Aplastic Anemia

First-line treatment options for patients with inherited marrow failure syndromes (IMFS) are androgen therapy and hematopoietic stem cell transplant (HSCT). When evaluating patients for HSCT, it is critical to identify the presence of an IMFS, as the risk and mortality associated with the conditioning regimen, stem cell source, graft-versus-host disease (GVHD), and secondary malignancies differ between patients with IMFS and those with acquired marrow failure syndromes or hematologic malignancies.

Potential sibling donors need to be screened for donor candidacy as well as for the inherited defect.¹ Among patients with Fanconi anemia or a telomere biology disorder, the stem cell source must be considered, with bone marrow demonstrating lower rates of acute GVHD than a peripheral blood stem cell source.^2-4 In IMFS patients, the donor cell type may affect the choice of conditioning regimen.^5,6 Reduced-intensity conditioning in lieu of myeloablative conditioning without total body irradiation has proved feasible in patients with Fanconi anemia, and is associated with a reduced risk of secondary malignancies.^5,6 Incorporation of fludarabine in the conditioning regimen of patients without a matched sibling donor is associated with superior engraftment and survival^2,5,7 compared to cyclophosphamide conditioning, which was historically used in matched related donors.^6,8 The addition of fludarabine appears to be especially beneficial in older patients, in whom its use is associated with lower rates of graft failure, likely due to increased immunosuppression at the time of engraftment.^7,9 Fludarabine has also been incorporated into conditioning regimens for patients with a telomere biology disorder, but outcomes data is limited.⁵

For patients presenting with acute myeloid leukemia (AML) or a high-risk myelodysplastic syndrome (MDS) who are subsequently diagnosed with an IMFS, treatment can be more complex, as these patients are at high risk for toxicity from standard chemotherapy. Limited data suggests that induction therapy and transplantation are feasible in this group of patients, and this approach is associated with increased overall survival (OS) despite lower OS rates than those of IMFS patients who present prior to the development of MDS or AML.^10,11 Further work is needed to determine the optimal induction regimen that balances the risks of treatment-related mortality and complications associated with conditioning regimens, risk of relapse, and risk of secondary malignancies, especially in the cohort of patients diagnosed at an older age.

Acquired Aplastic Anemia

Supportive Care

While the workup and treatment plan is being established, attention should be directed at supportive care for prevention of complications. The most common complications leading to death in patients with significant pancytopenia and neutropenia are opportunistic infections and hemorrhagic complications.¹²

Transfusion support is critical to avoid symptomatic anemia and hemorrhagic complications related to thrombocytopenia, which typically occur with platelet counts lower than 10,000 cells/µL. However, transfusion carries the risk of alloimmunization (which may persist for years following transfusion) and transfusion-related graft versus host disease (trGVHD), and thus use of transfusion should be minimized when possible.^13,14 Transfusion support is often required to prevent complications associated with thrombocytopenia and anemia; all blood products given to patients with aplastic anemia should be irradiated and leukoreduced to reduce the risk of both alloimmunization and trGVHD. Guidelines from the British Society for Haematology recommend routine screening for Rh and Kell antibodies to reduce the risk of alloimmunization.¹⁵ Infectious complications remain a common cause of morbidity and mortality in patients with aplastic anemia who have prolonged neutropenia (defined as an absolute neutrophil count [ANC] < 500 cells/µL).^16-19 Therefore, patients should receive broad-spectrum antibiotics with antipseudomonal coverage. In a study by Tichelli and colleagues evaluating the role of granulocyte-colony stimulating factor (G-CSF) in patients with SAA receiving immunosuppressive therapy, 55% of all patient deaths were secondary to infection.²⁰ There was no OS benefit seen in patients who received G-CSF, though a significantly lower rate of infection was observed in the G-CSF arm compared to those not receiving G-CSF (56% versus 81%, P = 0.006).This difference was largely driven by a decrease in infectious episodes in patients with very severe aplastic anemia (VSAA) treated with G-CSF as compared to those who did not receive this therapy (22% versus 48%, P = 0.014).²⁰

Angio-invasive pulmonary aspergillosis and Zygomycetes (eg, Rhizopus, Mucor species) remain major causes of mortality related to opportunistic mycotic infections in patients with aplastic anemia.¹⁸ The infectious risk is directly related to the duration and severity of neutropenia, with one study demonstrating a significant increase in risk in AML patients with neutropenia lasting longer than 3 weeks.²¹ Invasive fungal infections carry a high mortality in patients with severe neutropenia, though due to earlier recognition and empiric antifungal therapy with extended-spectrum azoles, overall mortality secondary to invasive fungal infections is declining.^19,22

While neutropenia related to cytotoxic chemotherapy is commonly associated with gram-negative bacteria due to disruption of mucosal barriers, patients with aplastic anemia have an increased incidence of gram-positive bacteremia with staphylococcal species compared to other neutropenic populations.^18,19 This appears to be changing with time. Valdez and colleagues demonstrated a decrease in prevalence of coagulase-negative staphylococcal infections, increased prevalence of gram-positive bacilli bacteremia, and no change in prevalence of gram-negative bacteremia in patients with aplastic anemia treated between 1989 and 2008.²² Gram-negative bacteremia caused by Stenotrophomonas maltophila, Escherichia coli, Klebsiella pneumoniae, Citrobacter, and Proteus has also been reported.¹⁹ Despite a lack of clinical trials investigating the role of antifungal and antibacterial prophylaxis for patients with aplastic anemia, most centers initiate antifungal prophylaxis in patients with severe aplastic anema (SAA) or VSAA with an anti-mold agent such as voriconazole or posaconazole (which has the additional benefit compared to voriconazole of covering Mucor species).^17,23 This is especially true for patients who have received ATG or undergone HSCT. For antimicrobial prophylaxis, a fluoroquinolone antibiotic with a spectrum of activity against Pseudomonas should be considered for patients with an ANC < 500 cells/µL.¹⁷ Acyclovir or valacyclovir prophylaxis is recommended for varicella-zoster virus and herpes simplex virus. Cytomegalovirus reactivation is minimal in patients with aplastic anemia, unless multiple courses of ATG are used.

Iron overload is another complication the provider must be aware of in the setting of increased transfusions in aplastic anemia patients. Lee and colleagues demonstrated that iron chelation therapy using deferasirox is effective at reducing serum ferritin levels in patients with aplastic anemia (median ferritin level: 3254 ng/mL prior to therapy, 1854 ng/mL following), and is associated with no serious adverse events (most common adverse events included nausea, diarrhea, vomiting, and rash).²⁴ Approximately 25% of patients in this trial demonstrated an increase in creatinine, with patients taking concomitant cyclosporine more affected than those on chelation therapy alone.²⁴ For patients following HSCT or with improved hematopoiesis following immunosuppressive therapy, phlebotomy can be used to treat iron overload in lieu of chelation therapy.¹⁵

Approach to Therapy

The main treatment options for SAA and VSAA include allogeneic bone marrow transplant and immunosuppression. The deciding factors as to which treatment is best initially depends on the availability of HLA-matched related donors and age (Figure 1 and Figure 2). Survival is decreased in patients with SAA or VSAA who delay initiation of therapy, and therefore prompt referral for HLA typing and evaluation for bone marrow transplant is a very important first step in managing aplastic anemia.

Matched Sibling Donor Transplant

Current standards of care recommend HLA-matched sibling donor transplant for patients with SAA or VSAA who are younger than 50 years of age, with the caveat that integration of fludarabine and reduced cyclophosphamide dosing along with ATG shows the best overall outcomes. Locasciulli and colleagues examined outcomes in patients given either immunosuppressive therapy or sibling HSCT between 1991-1996 and 1997-2002, respectively, and found that sibling HSCT was associated with a superior 10-year OS compared to immunosuppressive therapy (73% versus 68%).²⁵ Interestingly in this study, there was no OS improvement seen with immunosuppressive therapy alone (69% versus 73%) between the 2 time periods, despite increased OS in both sibling HSCT (74% and 80%) and matched unrelated donor HSCT (38% and 65%).²⁵ Though total body irradiation has been used in the past, it is typically not included in current conditioning regimens for matched related donor transplants.²⁶

Current conditioning regimens typically use a combination of cyclophosphamide and ATG^27,28 with or without fludarabine. Fludarabine-based conditioning regimens have shown promise in patients undergoing sibling HSCT. Maury and colleagues evaluated the role of fludarabine in addition to low-dose cyclophosphamide and ATG compared to cyclophosphamide alone or in combination with ATG in patients over age 30 undergoing sibling HSCT.⁹ There was a nonsignificant improvement in 5-year OS in the fludarabine arm compared to controls (77% ± 8% versus 60% ± 3%, P = 0.14) in the pooled analysis, but when adjusted for age the fludarabine arm had a significantly lower relative risk (RR) of death (RR, 0.44; P = 0.04) compared to the control arm. Shin et al reported outcomes with fludarabine/cyclophosphamide/ATG, with excellent overall outcomes and no difference in patients older or younger than 40 years.²⁹ In addition, Kim et al evaluated their experience with patients older than 40 years of age receiving matched related donors, finding comparable outcomes in those aged 41 to 50 years compared to younger patients. Outcomes did decline in those over the age of 50 years.³⁰ Long-term data for matched related donor transplant for aplastic anemia shows excellent long-term outcomes, with minimal chronic GVHD and good performance status.³¹ Hence, these factors support the role of matched related donor transplant as the initial treatment in SAA and VSAA.

Regarding the role of transplant for patients who lack a matched related donor, a growing body of literature demonstrating identical outcomes between matched related and matched unrelated donor (MUD) transplants for pediatric patients^32,33 supports recent recommendations for upfront unrelated donor transplantation for aplastic anemia.^34,35

Immunosuppressive Therapy

For patients without an HLA-matched sibling donor or those who are older than 50 years of age, immunosuppressive therapy is the first-line therapy. ATG and cyclosporine A are the treatments of choice.³⁶ The potential effectiveness of immunosuppressive therapy in treating aplastic anemia was initially observed in patients in whom autologous transplant failed but who still experienced hematopoietic reconstitution despite the failed graft; this observation led to the hypothesis that the conditioning regimen may have an effect on hematopoiesis.^16,36,37

Anti-thymocyte globulin. Immunosuppressive therapy with ATG has been used for the treatment of aplastic anemia since the 1980s.³⁸ Historically, rabbit ATG had been used, but a 2011 study of horse ATG demonstrated superior hematological response at 6 months compared to rabbit ATG (68% versus 37%).¹⁶ Superior survival was also seen with horse ATG compared to rabbit ATG (3-year OS: 96% versus 76%). Due to these results, horse ATG is preferred over rabbit ATG. ATG should be used in combination with cyclosporine A to optimize outcomes.

Cyclosporine A. Early studies also demonstrated the efficacy of cyclosporine A in the treatment of aplastic anemia, with response rates equivalent to that of ATG monotherapy.³⁹ Recent publications still note the efficacy of cyclosporine A in the treatment of aplastic anemia. Its role as an affordable option for single-agent therapy in developing countries is intriguing.³⁹

The combination of the ATG and cyclosporine A was proven superior to either agent alone in a study by Frickhofen et al.³⁷ In this study patients were randomly assigned to a control arm that received ATG plus methylprednisolone or to an arm that received ATG plus cyclosporine A and methylprednisolone. At 6 months, 70% of patients in the cyclosporine A arm had a complete remission (CR) or partial remission compared to 46% in the control arm.⁴⁰ Further work confirmed the long-term efficacy of this regimen, reporting a 7-year OS of 55%.⁴¹ Among a pediatric population, immunosuppressive therapy was associated with an 83% 10-year OS.⁴²

It is recommended that patients remain on cyclosporine therapy for a minimum of 6 months, after which a gradual taper may be considered, although there is variation among practitioners, with some continuing immunosuppressive therapy for a minimum of 12 months due to a proportion of patients being cyclosporine dependent.^42,43 A study found that within a population of patients who responded to immunosuppressive therapy, 18% became cyclosporine dependent.⁴² The median duration of cyclosporine A treatment at full dose was 12 months, with tapering completed over a median of 19 months after patients had been in a stable CR for a minimum of 3 months. Relapse occurred more often when patients were tapered quickly (decrease ≥ 0.8 mg/kg/month) compared to slowly (0.4-0.7 mg/kg/month) or very slowly (< 0.3 mg/kg/month).

Immunosuppressive therapy plus eltrombopag. Townsley and colleagues recently investigated incorporating the use of the thrombopoietin receptor agonist eltrombopag with immunosuppressive therapy as first-line therapy in aplastic anemia.⁴⁴ When given at a dose of 150 mg daily in patients aged 12 years and older or 75 mg daily in patients younger than 12 years, in conjunction with cyclosporine A and ATG, patients demonstrated markedly improved hematological response compared to historical treatment with standard immunosuppressive therapy alone.⁴⁴ In the patient cohort administered eltrombopag starting on day 1 and continuing for 6 months, the complete response rate was 58%. Eltrombopag led to improvement in all cell lines among all treatment subgroups, and OS (censored for patients who proceeded to transplant) was 99% at 2 years.⁴⁵ Overall, toxicities associated with this therapy were low, with liver enzyme elevations most commonly observed.⁴⁴ Recently, a phase 2 trial of immunosuppressive therapy with or without eltrombopag was reported. Of the 38 patients enrolled, overall response, complete response, and time to response were not statistically different.⁴⁶ With this recent finding, the role of eltrombopag in addition to immunosuppressive therapy is not clearly defined, and further studies are warranted.

OS for patients who do not respond to immunosuppressive therapy is approximately 57% at 5 years, largely due to improved supportive measures among this patient population.^4,22 Therefore, it is important to recognize those patients who have a low chance of response so that second-line therapy can be pursued to improve outcomes.

Matched Unrelated Donor Transplant

For patients with refractory disease following immunosuppressive therapy who lack a matched sibling donor, MUD HSCT is considered standard therapy given the marked improvement in overall outcomes with modulating conditioning regimens and high-resolution HLA typing. A European Society for Blood and Marrow Transplantation analysis comparing matched sibling HSCT to MUD HSCT noted significantly higher rates of acute grade II-IV and grade III-V GVHD (grade II-IV 13% versus 25%, grade III-IV 5% versus 10%) among patients undergoing MUD transplant.⁴⁷ Chronic GVHD rates were 14% in the sibling group, as compared to 26% in the MUD group. Additional benefits seen in this analysis included improved survival when transplanted under age 20 years (84% versus 72%), when transplanted within 6 months of diagnosis (85% versus 72%), the use of ATG in the conditioning regimen (81% versus 73%), and when the donor and recipient were cytomegalovirus-negative compared to other combinations (82% versus 76%).⁴⁷ Interestingly, this study demonstrated that OS was not significantly increased when using a sibling HSCT compared to a MUD HSCT, likely as a result of improved understanding of conditioning regimens, GVHD prophylaxis, and supportive care.

Additional studies of MUD HSCT have shown outcomes similar to those seen in sibling HSCT.^4,43 A French study found a significant increase in survival in patients undergoing MUD HSCT compared to historical cohorts (2000-2005: OS 52%; 2006-2012: OS 74%).³³ The majority of patients underwent conditioning with cyclophosphamide or a combination of busulfan and cyclophosphamide, with or without fludarabine; 81% of patients included underwent in vivo T-cell depletion, and a bone marrow donor source was utilized. OS was significantly lower in patients over age 30 years undergoing MUD HSCT (57%) compared to those under age 30 years (70%). Improved OS was also seen when patients underwent transplant within 1 year of diagnosis and when a 10/10 matched donor (compared to a 9/10 mismatched donor) was utilized.⁴ A 2015 study investigated the role of MUD HSCT as frontline therapy instead of immunosuppressive therapy in patients without a matched sibling donor.³³ The 2-year OS was 96% in the MUD HSCT cohort compared to 91%, 94%, and 74% in historical cohorts of sibling HSCT, frontline immunosuppressive therapy, and second-line MUD HSCT following failed immunosuppressive therapy, respectively. Additionally, event-free survival in the MUD HSCT cohort (defined by the authors as death, lack of response, relapse, occurrence of clonal evolution/clinical paroxysmal nocturnal hemoglobinuria, malignancies developing over follow‐up, and transplant for patients receiving immunosuppressive therapy frontline) was similar compared to sibling HSCT and superior to frontline immunosuppressive therapy and second-line MUD HSCT. Furthermore, Samarasinghe et al highlighted the importance of in vivo T-cell depletion with either ATG or alemtuzumab (anti-CD52 monoclonal antibody) in the prevention of acute and chronic GVHD in both sibling HSCT and MUD HSCT.⁴⁸

With continued improvement of less toxic and more immunomodulating conditioning regimens, utilization of bone marrow as a donor cell source, in vivo T-cell depletion, and use of GVHD and antimicrobial prophylaxis, more clinical evidence supports elevating MUD HSCT in the treatment plan for patients without a matched sibling donor.⁴⁹ However, there is still a large population of patients without matched sibling or unrelated donor options. In an effort to expand the transplant pool and thus avoid clonal hematopoiesis, clinically significant paroxysmal nocturnal hemoglobinuria, and relapsed aplastic anemia, more work continues to recognize the expanding role of alternative donor transplants (cord blood and haploidentical) as another viable treatment strategy for aplastic anemia after immunosuppressive therapy failure.⁵⁰

Summary

Aplastic anemia is a rare but potentially life-threatening disorder characterized by pancytopenia and a marked reduction in the hematopoietic stem cell compartment. Treatment should be instituted as soon as the dignosis of aplastic anemia is established. Treatment outcomes are excellent with modern supportive care and the current approach to allogeneic transplantation, and therefore referral to a bone marrow transplant program to evaluate for early transplantation is the new standard of care.

Aplastic anemia is a clinical and pathological entity of bone marrow failure that causes progressive loss of hematopoietic progenitor stem cells (HPSC), resulting in pancytopenia.¹ Patients may present along a spectrum, ranging from being asymptomatic with incidental findings on peripheral blood testing to having life-threatening neutropenic infections or bleeding. Aplastic anemia results from either inherited or acquired causes, and the pathophysiology and treatment approach vary significantly between these 2 causes. Therefore, recognition of inherited marrow failure diseases, such as Fanconi anemia and telomere biology disorders, is critical to establishing the management plan. This article reviews the epidemiology, pathophysiology, clinical presentation, and diagnosis of aplastic anemia. Treatment of aplastic anemia is reviewed in a separate article.

Epidemiology

Aplastic anemia is a rare disorder, with an incidence of approximately 1.5 to 7 cases per million individuals per year.^2,3 A recent Scandinavian study reported that the incidence of aplastic anemia among the Swedish population is 2.3 cases per million individuals per year, with a median age at diagnosis of 60 years and a slight female predominance (52% versus 48%, respectively).² This data is congruent with prior observations made in Barcelona, where the incidence was 2.34 cases per million individuals per year, albeit with a slightly higher incidence in males compared to females (2.54 versus 2.16, respectively).⁴ The incidence of aplastic anemia varies globally, with a disproportionate increase in incidence seen among Asian populations, with rates as high as 8.8 per million individuals per year.^3-5 This variation in incidence in Asia versus other countries has not been well explained. There appears to be a bimodal distribution, with incidence peaks seen in young adults and in older adults.^2,3,6

Pathophysiology

Acquired Aplastic Anemia

The leading hypothesis as to the cause of most cases of acquired aplastic anemia is that a dysregulated immune system destroys hematopoietic progenitor cells. Inciting etiologies implicated in the development of acquired aplastic anemia include pregnancy, infection, medications, and exposure to certain chemicals, such as benzene.^1,7 The historical understanding of acquired aplastic anemia implicates cytotoxic T-lymphocyte–mediated destruction of CD34+ hematopoietic stem cells.^1,8,9 This hypothesis served as the basis for treatment of acquired aplastic anemia with immunosuppressive therapy, predominantly anti-thymocyte globulin (ATG) combined with cyclosporine A.^1,8 More recent work has focused on cytokine interactions, particularly the suppressive role of interferon (IFN)-γ on hematopoietic stem cells independent of T-lymphocyte–mediated hematopoietic destruction, which has been demonstrated in a murine model.⁸ The interaction of IFN-γ with the hematopoietic stem cells pool is dynamic. IFN-γ levels are elevated during an acute inflammatory response such as a viral infection, providing further basis for the immune-mediated nature of the acquired disease.¹⁰ Specifically, in vitro studies suggest the effects of IFN-γ on HPSC may be secondary to interruption of thrombopoietin and its respective signaling pathways, which play a key role in hematopoietic stem cell renewal.¹¹ Eltrombopag, a thrombopoietin receptor antagonist, has shown promise in the treatment of refractory aplastic anemia, with studies indicating that its effectiveness is independent of IFN-γ levels.^11,12

Inherited Aplastic Anemia

The inherited marrow failure syndromes (IMFSs) are a group of disorders characterized by cellular maintenance and repair defects, leading to cytopenias, increased cancer risk, structural defects, and risk of end organ damage, such as liver cirrhosis and pulmonary fibrosis.^13-15 The most common diseases include Fanconi anemia, dyskeratosis congenita/telomere biology disorders, Diamond-Blackfan anemia, and Shwachman-Diamond syndrome, but with the advent of whole exome sequencing new syndromes continue to be discovered. While classically these disorders present in children, adult presentations of these syndromes are now commonplace. Broadly, the pathophysiology of inherited aplastic anemia relates to the defective hematopoietic progenitor cells and an accelerated decline of the hematopoietic stem cell compartment.

The most common IMFS, Fanconi anemia and telomere biology disorders, are associated with numerous mutations in DNA damage repair pathways and telomere maintenance pathways. TERT, DKC, and TERC mutations are most commonly associated with dyskeratosis congenita, but may also be found infrequently in patients with aplastic anemia presenting at an older age in the absence of the classic phenotypical features.^1,16,17 The recognition of an underlying genetic disorder or telomere biology disorder leading to constitutional aplastic anemia is significant, as these conditions are associated not only with marrow failure, but also endocrinopathies, organ fibrosis, and solid organ malignancies.^13-15 In particular, mutations in the TERT and TERC genes have been associated with dyskeratosis congenita as well as pulmonary fibrosis and cirrhosis.^18,19 The implications of early diagnosis of an IMFS lie in the approach to treatment and prognosis.

Clonal Disorders and Secondary Malignancies

Myelodysplastic syndrome (MDS) and secondary acute myeloid leukemia (AML) are 2 clonal disorders that may arise from a background of aplastic anemia.^9,20,21 Hypoplastic MDS can be difficult to differentiate from aplastic anemia at diagnosis based on morphology alone, although recent work has demonstrated that molecular testing for somatic mutations in ASXL1, DNMT3A, and BCOR can aid in differentiating a subset of aplastic anemia patients who are more likely to progress to MDS.²¹ Clonal populations of cells harboring 6p uniparental disomy are seen in more than 10% of patients with aplastic anemia on cytogenetic analysis, which can help differentiate the diseases.⁹ Yoshizato and colleagues found lower rates of ASXL1 and DNMT3A mutations in patients with aplastic anemia as compared with patients with MDS or AML. In this study, patients with aplastic anemia had higher rates of mutations in PIGA (reflecting the increased paroxysmal nocturnal hemoglobinuria [PNH] clonality seen in aplastic anemia) and BCOR.⁹ Mutations were also found in genes commonly mutated in MDS and AML, including TET2, RUNX1, TP53, and JAK2, albeit at lower frequencies.⁹ These mutations as a whole have not predicted response to therapy or prognosis. However, when performing survival analysis in patients with specific mutations, those commonly encountered in MDS/AML (ASXL1, DNMT3A, TP53, RUNX1, CSMD1) are associated with faster progression to overt MDS/AML and decreased overall survival (OS),^20,21 suggesting these mutations may represent early clonality that can lead to clonal evolution and the development of secondary malignancies. Conversely, mutations in BCOR and BCORL appear to identify patients who may have a favorable outcome in response to immunosuppressive therapy and, similar to patients with PIGA mutations, improved OS.⁹

Paroxysmal Nocturnal Hemoglobinuria

In addition to having an increased risk of myelodysplasia and malignancy due to the development of a dominant pre-malignant clone, patients with aplastic anemia often harbor progenitor cell clones associated with PNH.^1,17 PNH clones have been identified in more than 50% of patients with aplastic anemia.^22,23 PNH represents a clonal disorder of hematopoiesis in which cells harbor X-linked somatic mutations in the PIGA gene; this gene encodes a protein responsible for the synthesis of glycosylphosphatidylinositol (GPI) anchors on the cell surface.^22,24 The lack of these cell surface proteins, specifically CD55 (also known as decay accelerating factor) and CD59 (also known as membrane inhibitor of reactive lysis), predisposes red cells to increased complement-mediated lysis.²⁵ The exact mechanism for the development of these clones in patients with aplastic anemia is not fully understood. Current theories hypothesize that these clones are protected from the immune-mediated destruction of normal hematopoietic stem cells due to the absence of the cell surface proteins.^1,20 The role of these clones over time in patients with aplastic anemia is less clear, though recent work demonstrated that despite differences in clonality over the disease course, aplastic anemia patients with small PNH clones are less likely to develop overt hemolysis and larger PNH clones compared to patients harboring larger (≥ 50%) PNH clones at diagnosis.^23,26,27 Additionally, PNH clones in patients with aplastic anemia infrequently become clinically significant.²⁷ It should be noted that these conditions exist along a continuum; that is, patients with aplastic anemia may develop PNH clones, while conversely patients with PNH may develop aplastic anemia.²⁰ Patients with PNH clones should be followed via peripheral blood flow cytometry in addition to complete blood count to track clonal stability and identify clinically significant PNH among aplastic anemia patients.²⁸

Patients with aplastic anemia typically are diagnosed either due to asymptomatic cytopenias found on peripheral blood sampling, symptomatic anemia, bleeding secondary to thrombocytopenia, or wound healing and infectious complications related to neutropenia.²⁹ A thorough history to understand the timing of symptoms, recent infectious symptoms/exposure, habits, and chemical or toxin exposures (including medications, travel, and supplements) helps guide diagnostic testing. Family history is also critical, with attention given to premature graying, pulmonary, renal, and liver disease, and blood disorders.

Patients with an IMFS, (eg, Fanconi anemia or dyskeratosis congenita) may have associated phenotypical findings such as urogenital abnormalities or short stature; in addition, those with dyskeratosis congenita may present with the classic triad of oral leukoplakia, lacy skin pigmentation, and dystrophic nails.⁷ However, in patients with IMFS, classic phenotypical findings may be lacking in up to 30% to 40% of patients.⁷ As described previously, while congenital malformations are common in Fanconi anemia and dyskeratosis congenita, a third of patients may have no or only subtle phenotypical abnormalities, including alterations in skin or hair pigmentation, skeletal and growth abnormalities, and endocrine disorders.³⁰ The International Fanconi Anemia Registry identified central nervous system, genitourinary, skin and musculoskeletal, ophthalmic, and gastrointestinal system malformations among children with Fanconi anemia.^31,32 Patients with dyskeratosis congenita may present with pulmonary fibrosis, hepatic cirrhosis, or premature graying, as highlighted in a recent study by DiNardo and colleagues.³³ Therefore, physicians must have a heightened index of suspicion in patients with subtle phenotypical findings and associated cytopenias.

Diagnosis

Differential Diagnosis

The diagnosis of aplastic anemia should be suspected in any patient presenting with pancytopenia. Aplastic anemia is a diagnosis of exclusion.³⁴ Other conditions associated with peripheral blood pancytopenia should be considered including infections (HIV, hepatitis, parvovirus B19, cytomegalovirus, Epstein-Barr virus, varicella-zoster virus), nutritional deficiencies (vitamin B12, folate, copper, zinc), autoimmune disease (systemic lupus erythematosus, rheumatoid arthritis, hemophagocytic lymphohistiocytosis), hypersplenism, marrow-occupying diseases (eg, leukemia, lymphoma, MDS), solid malignancies, and fibrosis (Table).⁷

Diagnostic Evaluation

The workup for aplastic anemia should include a thorough history and physical exam to search simultaneously for alternative diagnoses and clues pointing to potential etiologic agents.⁷ Diagnostic tests to be performed include a complete blood count with differential, reticulocyte count, immature platelet fraction, flow cytometry (to rule out lymphoproliferative disorders and atypical myeloid cells and to evaluate for PNH), and bone marrow biopsy with subsequent cytogenetic, immunohistochemical, and molecular testing.³⁵ The typical findings in aplastic anemia include peripheral blood pancytopenia without dysplastic features and bone marrow biopsy demonstrating a hypocellular marrow.⁷ A relative lymphocytosis in the peripheral blood is common.⁷ In patients with a significant PNH clone, a macrocytosis along with elevated lactate dehydrogenase and elevated reticulocyte and granulocyte counts may be present.³⁶

The diagnosis (based on the Camitta criteria³⁷ and modified Camitta criteria³⁸ for severe aplastic anemia) requires 2 of the following findings on peripheral blood samples:

• Absolute neutrophil count (ANC) < 500 cells/µL
• Platelet count < 20,000 cells/µL
• Reticulocyte count < 1% corrected or < 20,000 cells/µL.³⁵

In addition to peripheral blood findings, bone marrow biopsy is essential for the diagnosis, and should demonstrate a markedly hypocellular marrow (cellularity < 25%), occasionally with an increase in T lymphocytes.^7,39 Because marrow cellularity varies with age and can be challenging to assess, additional biopsies may be needed to confirm the diagnosis.²⁹ A 1- to 2-cm core biopsy is necessary to confirm hypocellularity, as small areas of residual hematopoiesis may be present and obscure the diagnosis.³⁵

Excluding Hypocellular MDS and IMFS

A diagnostic challenge is the exclusion of hypocellular MDS, especially in the older adult presenting with aplastic anemia, as patients with aplastic anemia may have some degree of erythroid dysplasia on bone marrow morphology.³⁶ The presence of a PNH clone on flow cytometry can aid in diagnosing aplastic anemia and excluding MDS,³⁴ although PNH clones can be present in refractory anemia MDS. Patients with aplastic anemia have a lower ratio of CD34+ cells compared to those with hypoplastic MDS, with one study demonstrating a mean CD34+ percentage of < 0.5% in aplastic anemia versus 3.7% in hypoplastic MDS.⁴⁰ Cytogenetic and molecular testing can also aid in making this distinction by identifying mutations commonly implicated in MDS.⁷ The presence of monosomy 7 (-7) in aplastic anemia patients is associated with a poor overall prognosis.^34,41

Peripheral blood screening using chromosome breakage analysis (done using either mitomycin C or diepoxybutane as in vitro DNA-crosslinking agents) and telomere length testing (of peripheral blood leukocytes) is necessary to exclude the main IMFS, Fanconi anemia and telomere biology disorders, respectively. Ruling out these conditions is imperative, as the approach to treatment varies significantly between IMFS and aplastic anemia. Patients with shortened telomeres should undergo genetic screening for mutations in the telomere maintenance genes to evaluate the underlying defect leading to shortened telomeres. Patients with increased peripheral blood breakage should have genetic testing to detect mutations associated with Fanconi anemia.

Classification

Once the diagnosis of aplastic anemia has been made, the patient should be classified according to the severity of their disease. Disease severity is determined based on peripheral blood ANC:³⁴ non-severe aplastic anemia (NSAA), ANC > 500 polymorphonuclear neutrophils (PMNs)/µL; severe aplastic anemia (SAA), 200–500 PMNs/µL; and very severe (VSAA), 0–200 PMNs/µL.^4,34 Disease classification is important, as VSAA is associated with a decreased OS compared to SAA.² Disease classification may affect treatment decisions, as patients with NSAA may be observed for a short period of time, while conversely patients with SAA have a worse prognosis with delays in therapy.^42-44

Summary

Aplastic anemia is a rare but potentially life-threatening disorder characterized by pancytopenia and a marked reduction in the hematopoietic stem cell compartment. It can be acquired or associated with an IMFS, and the treatment and prognosis vary dramatically between these 2 etiologies. Work-up and diagnosis involves investigating IMFSs and ruling out malignant or infectious etiologies for pancytopenia. After aplastic anemia has been diagnosed, the patient should be classified according to the severity of their disease based on peripheral blood ANC.

Aplastic anemia is a clinical and pathological entity of bone marrow failure that causes progressive loss of hematopoietic progenitor stem cells (HPSC), resulting in pancytopenia.¹ Patients may present along a spectrum, ranging from being asymptomatic with incidental findings on peripheral blood testing to having life-threatening neutropenic infections or bleeding. Aplastic anemia results from either inherited or acquired causes, and the pathophysiology and treatment approach vary significantly between these 2 causes. Therefore, recognition of inherited marrow failure diseases, such as Fanconi anemia and telomere biology disorders, is critical to establishing the management plan. This article reviews the epidemiology, pathophysiology, clinical presentation, and diagnosis of aplastic anemia. Treatment of aplastic anemia is reviewed in a separate article.

Epidemiology

Aplastic anemia is a rare disorder, with an incidence of approximately 1.5 to 7 cases per million individuals per year.^2,3 A recent Scandinavian study reported that the incidence of aplastic anemia among the Swedish population is 2.3 cases per million individuals per year, with a median age at diagnosis of 60 years and a slight female predominance (52% versus 48%, respectively).² This data is congruent with prior observations made in Barcelona, where the incidence was 2.34 cases per million individuals per year, albeit with a slightly higher incidence in males compared to females (2.54 versus 2.16, respectively).⁴ The incidence of aplastic anemia varies globally, with a disproportionate increase in incidence seen among Asian populations, with rates as high as 8.8 per million individuals per year.^3-5 This variation in incidence in Asia versus other countries has not been well explained. There appears to be a bimodal distribution, with incidence peaks seen in young adults and in older adults.^2,3,6

Pathophysiology

Acquired Aplastic Anemia

The leading hypothesis as to the cause of most cases of acquired aplastic anemia is that a dysregulated immune system destroys hematopoietic progenitor cells. Inciting etiologies implicated in the development of acquired aplastic anemia include pregnancy, infection, medications, and exposure to certain chemicals, such as benzene.^1,7 The historical understanding of acquired aplastic anemia implicates cytotoxic T-lymphocyte–mediated destruction of CD34+ hematopoietic stem cells.^1,8,9 This hypothesis served as the basis for treatment of acquired aplastic anemia with immunosuppressive therapy, predominantly anti-thymocyte globulin (ATG) combined with cyclosporine A.^1,8 More recent work has focused on cytokine interactions, particularly the suppressive role of interferon (IFN)-γ on hematopoietic stem cells independent of T-lymphocyte–mediated hematopoietic destruction, which has been demonstrated in a murine model.⁸ The interaction of IFN-γ with the hematopoietic stem cells pool is dynamic. IFN-γ levels are elevated during an acute inflammatory response such as a viral infection, providing further basis for the immune-mediated nature of the acquired disease.¹⁰ Specifically, in vitro studies suggest the effects of IFN-γ on HPSC may be secondary to interruption of thrombopoietin and its respective signaling pathways, which play a key role in hematopoietic stem cell renewal.¹¹ Eltrombopag, a thrombopoietin receptor antagonist, has shown promise in the treatment of refractory aplastic anemia, with studies indicating that its effectiveness is independent of IFN-γ levels.^11,12

Inherited Aplastic Anemia

The inherited marrow failure syndromes (IMFSs) are a group of disorders characterized by cellular maintenance and repair defects, leading to cytopenias, increased cancer risk, structural defects, and risk of end organ damage, such as liver cirrhosis and pulmonary fibrosis.^13-15 The most common diseases include Fanconi anemia, dyskeratosis congenita/telomere biology disorders, Diamond-Blackfan anemia, and Shwachman-Diamond syndrome, but with the advent of whole exome sequencing new syndromes continue to be discovered. While classically these disorders present in children, adult presentations of these syndromes are now commonplace. Broadly, the pathophysiology of inherited aplastic anemia relates to the defective hematopoietic progenitor cells and an accelerated decline of the hematopoietic stem cell compartment.

The most common IMFS, Fanconi anemia and telomere biology disorders, are associated with numerous mutations in DNA damage repair pathways and telomere maintenance pathways. TERT, DKC, and TERC mutations are most commonly associated with dyskeratosis congenita, but may also be found infrequently in patients with aplastic anemia presenting at an older age in the absence of the classic phenotypical features.^1,16,17 The recognition of an underlying genetic disorder or telomere biology disorder leading to constitutional aplastic anemia is significant, as these conditions are associated not only with marrow failure, but also endocrinopathies, organ fibrosis, and solid organ malignancies.^13-15 In particular, mutations in the TERT and TERC genes have been associated with dyskeratosis congenita as well as pulmonary fibrosis and cirrhosis.^18,19 The implications of early diagnosis of an IMFS lie in the approach to treatment and prognosis.

Clonal Disorders and Secondary Malignancies

Myelodysplastic syndrome (MDS) and secondary acute myeloid leukemia (AML) are 2 clonal disorders that may arise from a background of aplastic anemia.^9,20,21 Hypoplastic MDS can be difficult to differentiate from aplastic anemia at diagnosis based on morphology alone, although recent work has demonstrated that molecular testing for somatic mutations in ASXL1, DNMT3A, and BCOR can aid in differentiating a subset of aplastic anemia patients who are more likely to progress to MDS.²¹ Clonal populations of cells harboring 6p uniparental disomy are seen in more than 10% of patients with aplastic anemia on cytogenetic analysis, which can help differentiate the diseases.⁹ Yoshizato and colleagues found lower rates of ASXL1 and DNMT3A mutations in patients with aplastic anemia as compared with patients with MDS or AML. In this study, patients with aplastic anemia had higher rates of mutations in PIGA (reflecting the increased paroxysmal nocturnal hemoglobinuria [PNH] clonality seen in aplastic anemia) and BCOR.⁹ Mutations were also found in genes commonly mutated in MDS and AML, including TET2, RUNX1, TP53, and JAK2, albeit at lower frequencies.⁹ These mutations as a whole have not predicted response to therapy or prognosis. However, when performing survival analysis in patients with specific mutations, those commonly encountered in MDS/AML (ASXL1, DNMT3A, TP53, RUNX1, CSMD1) are associated with faster progression to overt MDS/AML and decreased overall survival (OS),^20,21 suggesting these mutations may represent early clonality that can lead to clonal evolution and the development of secondary malignancies. Conversely, mutations in BCOR and BCORL appear to identify patients who may have a favorable outcome in response to immunosuppressive therapy and, similar to patients with PIGA mutations, improved OS.⁹

Paroxysmal Nocturnal Hemoglobinuria

In addition to having an increased risk of myelodysplasia and malignancy due to the development of a dominant pre-malignant clone, patients with aplastic anemia often harbor progenitor cell clones associated with PNH.^1,17 PNH clones have been identified in more than 50% of patients with aplastic anemia.^22,23 PNH represents a clonal disorder of hematopoiesis in which cells harbor X-linked somatic mutations in the PIGA gene; this gene encodes a protein responsible for the synthesis of glycosylphosphatidylinositol (GPI) anchors on the cell surface.^22,24 The lack of these cell surface proteins, specifically CD55 (also known as decay accelerating factor) and CD59 (also known as membrane inhibitor of reactive lysis), predisposes red cells to increased complement-mediated lysis.²⁵ The exact mechanism for the development of these clones in patients with aplastic anemia is not fully understood. Current theories hypothesize that these clones are protected from the immune-mediated destruction of normal hematopoietic stem cells due to the absence of the cell surface proteins.^1,20 The role of these clones over time in patients with aplastic anemia is less clear, though recent work demonstrated that despite differences in clonality over the disease course, aplastic anemia patients with small PNH clones are less likely to develop overt hemolysis and larger PNH clones compared to patients harboring larger (≥ 50%) PNH clones at diagnosis.^23,26,27 Additionally, PNH clones in patients with aplastic anemia infrequently become clinically significant.²⁷ It should be noted that these conditions exist along a continuum; that is, patients with aplastic anemia may develop PNH clones, while conversely patients with PNH may develop aplastic anemia.²⁰ Patients with PNH clones should be followed via peripheral blood flow cytometry in addition to complete blood count to track clonal stability and identify clinically significant PNH among aplastic anemia patients.²⁸

Patients with aplastic anemia typically are diagnosed either due to asymptomatic cytopenias found on peripheral blood sampling, symptomatic anemia, bleeding secondary to thrombocytopenia, or wound healing and infectious complications related to neutropenia.²⁹ A thorough history to understand the timing of symptoms, recent infectious symptoms/exposure, habits, and chemical or toxin exposures (including medications, travel, and supplements) helps guide diagnostic testing. Family history is also critical, with attention given to premature graying, pulmonary, renal, and liver disease, and blood disorders.

Patients with an IMFS, (eg, Fanconi anemia or dyskeratosis congenita) may have associated phenotypical findings such as urogenital abnormalities or short stature; in addition, those with dyskeratosis congenita may present with the classic triad of oral leukoplakia, lacy skin pigmentation, and dystrophic nails.⁷ However, in patients with IMFS, classic phenotypical findings may be lacking in up to 30% to 40% of patients.⁷ As described previously, while congenital malformations are common in Fanconi anemia and dyskeratosis congenita, a third of patients may have no or only subtle phenotypical abnormalities, including alterations in skin or hair pigmentation, skeletal and growth abnormalities, and endocrine disorders.³⁰ The International Fanconi Anemia Registry identified central nervous system, genitourinary, skin and musculoskeletal, ophthalmic, and gastrointestinal system malformations among children with Fanconi anemia.^31,32 Patients with dyskeratosis congenita may present with pulmonary fibrosis, hepatic cirrhosis, or premature graying, as highlighted in a recent study by DiNardo and colleagues.³³ Therefore, physicians must have a heightened index of suspicion in patients with subtle phenotypical findings and associated cytopenias.

Diagnosis

Differential Diagnosis

The diagnosis of aplastic anemia should be suspected in any patient presenting with pancytopenia. Aplastic anemia is a diagnosis of exclusion.³⁴ Other conditions associated with peripheral blood pancytopenia should be considered including infections (HIV, hepatitis, parvovirus B19, cytomegalovirus, Epstein-Barr virus, varicella-zoster virus), nutritional deficiencies (vitamin B12, folate, copper, zinc), autoimmune disease (systemic lupus erythematosus, rheumatoid arthritis, hemophagocytic lymphohistiocytosis), hypersplenism, marrow-occupying diseases (eg, leukemia, lymphoma, MDS), solid malignancies, and fibrosis (Table).⁷

Diagnostic Evaluation

The workup for aplastic anemia should include a thorough history and physical exam to search simultaneously for alternative diagnoses and clues pointing to potential etiologic agents.⁷ Diagnostic tests to be performed include a complete blood count with differential, reticulocyte count, immature platelet fraction, flow cytometry (to rule out lymphoproliferative disorders and atypical myeloid cells and to evaluate for PNH), and bone marrow biopsy with subsequent cytogenetic, immunohistochemical, and molecular testing.³⁵ The typical findings in aplastic anemia include peripheral blood pancytopenia without dysplastic features and bone marrow biopsy demonstrating a hypocellular marrow.⁷ A relative lymphocytosis in the peripheral blood is common.⁷ In patients with a significant PNH clone, a macrocytosis along with elevated lactate dehydrogenase and elevated reticulocyte and granulocyte counts may be present.³⁶

The diagnosis (based on the Camitta criteria³⁷ and modified Camitta criteria³⁸ for severe aplastic anemia) requires 2 of the following findings on peripheral blood samples:

• Absolute neutrophil count (ANC) < 500 cells/µL
• Platelet count < 20,000 cells/µL
• Reticulocyte count < 1% corrected or < 20,000 cells/µL.³⁵

In addition to peripheral blood findings, bone marrow biopsy is essential for the diagnosis, and should demonstrate a markedly hypocellular marrow (cellularity < 25%), occasionally with an increase in T lymphocytes.^7,39 Because marrow cellularity varies with age and can be challenging to assess, additional biopsies may be needed to confirm the diagnosis.²⁹ A 1- to 2-cm core biopsy is necessary to confirm hypocellularity, as small areas of residual hematopoiesis may be present and obscure the diagnosis.³⁵

Excluding Hypocellular MDS and IMFS

A diagnostic challenge is the exclusion of hypocellular MDS, especially in the older adult presenting with aplastic anemia, as patients with aplastic anemia may have some degree of erythroid dysplasia on bone marrow morphology.³⁶ The presence of a PNH clone on flow cytometry can aid in diagnosing aplastic anemia and excluding MDS,³⁴ although PNH clones can be present in refractory anemia MDS. Patients with aplastic anemia have a lower ratio of CD34+ cells compared to those with hypoplastic MDS, with one study demonstrating a mean CD34+ percentage of < 0.5% in aplastic anemia versus 3.7% in hypoplastic MDS.⁴⁰ Cytogenetic and molecular testing can also aid in making this distinction by identifying mutations commonly implicated in MDS.⁷ The presence of monosomy 7 (-7) in aplastic anemia patients is associated with a poor overall prognosis.^34,41

Peripheral blood screening using chromosome breakage analysis (done using either mitomycin C or diepoxybutane as in vitro DNA-crosslinking agents) and telomere length testing (of peripheral blood leukocytes) is necessary to exclude the main IMFS, Fanconi anemia and telomere biology disorders, respectively. Ruling out these conditions is imperative, as the approach to treatment varies significantly between IMFS and aplastic anemia. Patients with shortened telomeres should undergo genetic screening for mutations in the telomere maintenance genes to evaluate the underlying defect leading to shortened telomeres. Patients with increased peripheral blood breakage should have genetic testing to detect mutations associated with Fanconi anemia.

Classification

Once the diagnosis of aplastic anemia has been made, the patient should be classified according to the severity of their disease. Disease severity is determined based on peripheral blood ANC:³⁴ non-severe aplastic anemia (NSAA), ANC > 500 polymorphonuclear neutrophils (PMNs)/µL; severe aplastic anemia (SAA), 200–500 PMNs/µL; and very severe (VSAA), 0–200 PMNs/µL.^4,34 Disease classification is important, as VSAA is associated with a decreased OS compared to SAA.² Disease classification may affect treatment decisions, as patients with NSAA may be observed for a short period of time, while conversely patients with SAA have a worse prognosis with delays in therapy.^42-44

Summary

Aplastic anemia is a rare but potentially life-threatening disorder characterized by pancytopenia and a marked reduction in the hematopoietic stem cell compartment. It can be acquired or associated with an IMFS, and the treatment and prognosis vary dramatically between these 2 etiologies. Work-up and diagnosis involves investigating IMFSs and ruling out malignant or infectious etiologies for pancytopenia. After aplastic anemia has been diagnosed, the patient should be classified according to the severity of their disease based on peripheral blood ANC.

Aplastic anemia is a clinical and pathological entity of bone marrow failure that causes progressive loss of hematopoietic progenitor stem cells (HPSC), resulting in pancytopenia.¹ Patients may present along a spectrum, ranging from being asymptomatic with incidental findings on peripheral blood testing to having life-threatening neutropenic infections or bleeding. Aplastic anemia results from either inherited or acquired causes, and the pathophysiology and treatment approach vary significantly between these 2 causes. Therefore, recognition of inherited marrow failure diseases, such as Fanconi anemia and telomere biology disorders, is critical to establishing the management plan. This article reviews the epidemiology, pathophysiology, clinical presentation, and diagnosis of aplastic anemia. Treatment of aplastic anemia is reviewed in a separate article.

Epidemiology

Aplastic anemia is a rare disorder, with an incidence of approximately 1.5 to 7 cases per million individuals per year.^2,3 A recent Scandinavian study reported that the incidence of aplastic anemia among the Swedish population is 2.3 cases per million individuals per year, with a median age at diagnosis of 60 years and a slight female predominance (52% versus 48%, respectively).² This data is congruent with prior observations made in Barcelona, where the incidence was 2.34 cases per million individuals per year, albeit with a slightly higher incidence in males compared to females (2.54 versus 2.16, respectively).⁴ The incidence of aplastic anemia varies globally, with a disproportionate increase in incidence seen among Asian populations, with rates as high as 8.8 per million individuals per year.^3-5 This variation in incidence in Asia versus other countries has not been well explained. There appears to be a bimodal distribution, with incidence peaks seen in young adults and in older adults.^2,3,6

Pathophysiology

Acquired Aplastic Anemia

The leading hypothesis as to the cause of most cases of acquired aplastic anemia is that a dysregulated immune system destroys hematopoietic progenitor cells. Inciting etiologies implicated in the development of acquired aplastic anemia include pregnancy, infection, medications, and exposure to certain chemicals, such as benzene.^1,7 The historical understanding of acquired aplastic anemia implicates cytotoxic T-lymphocyte–mediated destruction of CD34+ hematopoietic stem cells.^1,8,9 This hypothesis served as the basis for treatment of acquired aplastic anemia with immunosuppressive therapy, predominantly anti-thymocyte globulin (ATG) combined with cyclosporine A.^1,8 More recent work has focused on cytokine interactions, particularly the suppressive role of interferon (IFN)-γ on hematopoietic stem cells independent of T-lymphocyte–mediated hematopoietic destruction, which has been demonstrated in a murine model.⁸ The interaction of IFN-γ with the hematopoietic stem cells pool is dynamic. IFN-γ levels are elevated during an acute inflammatory response such as a viral infection, providing further basis for the immune-mediated nature of the acquired disease.¹⁰ Specifically, in vitro studies suggest the effects of IFN-γ on HPSC may be secondary to interruption of thrombopoietin and its respective signaling pathways, which play a key role in hematopoietic stem cell renewal.¹¹ Eltrombopag, a thrombopoietin receptor antagonist, has shown promise in the treatment of refractory aplastic anemia, with studies indicating that its effectiveness is independent of IFN-γ levels.^11,12

Inherited Aplastic Anemia

The inherited marrow failure syndromes (IMFSs) are a group of disorders characterized by cellular maintenance and repair defects, leading to cytopenias, increased cancer risk, structural defects, and risk of end organ damage, such as liver cirrhosis and pulmonary fibrosis.^13-15 The most common diseases include Fanconi anemia, dyskeratosis congenita/telomere biology disorders, Diamond-Blackfan anemia, and Shwachman-Diamond syndrome, but with the advent of whole exome sequencing new syndromes continue to be discovered. While classically these disorders present in children, adult presentations of these syndromes are now commonplace. Broadly, the pathophysiology of inherited aplastic anemia relates to the defective hematopoietic progenitor cells and an accelerated decline of the hematopoietic stem cell compartment.

The most common IMFS, Fanconi anemia and telomere biology disorders, are associated with numerous mutations in DNA damage repair pathways and telomere maintenance pathways. TERT, DKC, and TERC mutations are most commonly associated with dyskeratosis congenita, but may also be found infrequently in patients with aplastic anemia presenting at an older age in the absence of the classic phenotypical features.^1,16,17 The recognition of an underlying genetic disorder or telomere biology disorder leading to constitutional aplastic anemia is significant, as these conditions are associated not only with marrow failure, but also endocrinopathies, organ fibrosis, and solid organ malignancies.^13-15 In particular, mutations in the TERT and TERC genes have been associated with dyskeratosis congenita as well as pulmonary fibrosis and cirrhosis.^18,19 The implications of early diagnosis of an IMFS lie in the approach to treatment and prognosis.

Clonal Disorders and Secondary Malignancies

Myelodysplastic syndrome (MDS) and secondary acute myeloid leukemia (AML) are 2 clonal disorders that may arise from a background of aplastic anemia.^9,20,21 Hypoplastic MDS can be difficult to differentiate from aplastic anemia at diagnosis based on morphology alone, although recent work has demonstrated that molecular testing for somatic mutations in ASXL1, DNMT3A, and BCOR can aid in differentiating a subset of aplastic anemia patients who are more likely to progress to MDS.²¹ Clonal populations of cells harboring 6p uniparental disomy are seen in more than 10% of patients with aplastic anemia on cytogenetic analysis, which can help differentiate the diseases.⁹ Yoshizato and colleagues found lower rates of ASXL1 and DNMT3A mutations in patients with aplastic anemia as compared with patients with MDS or AML. In this study, patients with aplastic anemia had higher rates of mutations in PIGA (reflecting the increased paroxysmal nocturnal hemoglobinuria [PNH] clonality seen in aplastic anemia) and BCOR.⁹ Mutations were also found in genes commonly mutated in MDS and AML, including TET2, RUNX1, TP53, and JAK2, albeit at lower frequencies.⁹ These mutations as a whole have not predicted response to therapy or prognosis. However, when performing survival analysis in patients with specific mutations, those commonly encountered in MDS/AML (ASXL1, DNMT3A, TP53, RUNX1, CSMD1) are associated with faster progression to overt MDS/AML and decreased overall survival (OS),^20,21 suggesting these mutations may represent early clonality that can lead to clonal evolution and the development of secondary malignancies. Conversely, mutations in BCOR and BCORL appear to identify patients who may have a favorable outcome in response to immunosuppressive therapy and, similar to patients with PIGA mutations, improved OS.⁹

Paroxysmal Nocturnal Hemoglobinuria

In addition to having an increased risk of myelodysplasia and malignancy due to the development of a dominant pre-malignant clone, patients with aplastic anemia often harbor progenitor cell clones associated with PNH.^1,17 PNH clones have been identified in more than 50% of patients with aplastic anemia.^22,23 PNH represents a clonal disorder of hematopoiesis in which cells harbor X-linked somatic mutations in the PIGA gene; this gene encodes a protein responsible for the synthesis of glycosylphosphatidylinositol (GPI) anchors on the cell surface.^22,24 The lack of these cell surface proteins, specifically CD55 (also known as decay accelerating factor) and CD59 (also known as membrane inhibitor of reactive lysis), predisposes red cells to increased complement-mediated lysis.²⁵ The exact mechanism for the development of these clones in patients with aplastic anemia is not fully understood. Current theories hypothesize that these clones are protected from the immune-mediated destruction of normal hematopoietic stem cells due to the absence of the cell surface proteins.^1,20 The role of these clones over time in patients with aplastic anemia is less clear, though recent work demonstrated that despite differences in clonality over the disease course, aplastic anemia patients with small PNH clones are less likely to develop overt hemolysis and larger PNH clones compared to patients harboring larger (≥ 50%) PNH clones at diagnosis.^23,26,27 Additionally, PNH clones in patients with aplastic anemia infrequently become clinically significant.²⁷ It should be noted that these conditions exist along a continuum; that is, patients with aplastic anemia may develop PNH clones, while conversely patients with PNH may develop aplastic anemia.²⁰ Patients with PNH clones should be followed via peripheral blood flow cytometry in addition to complete blood count to track clonal stability and identify clinically significant PNH among aplastic anemia patients.²⁸

Patients with aplastic anemia typically are diagnosed either due to asymptomatic cytopenias found on peripheral blood sampling, symptomatic anemia, bleeding secondary to thrombocytopenia, or wound healing and infectious complications related to neutropenia.²⁹ A thorough history to understand the timing of symptoms, recent infectious symptoms/exposure, habits, and chemical or toxin exposures (including medications, travel, and supplements) helps guide diagnostic testing. Family history is also critical, with attention given to premature graying, pulmonary, renal, and liver disease, and blood disorders.

Patients with an IMFS, (eg, Fanconi anemia or dyskeratosis congenita) may have associated phenotypical findings such as urogenital abnormalities or short stature; in addition, those with dyskeratosis congenita may present with the classic triad of oral leukoplakia, lacy skin pigmentation, and dystrophic nails.⁷ However, in patients with IMFS, classic phenotypical findings may be lacking in up to 30% to 40% of patients.⁷ As described previously, while congenital malformations are common in Fanconi anemia and dyskeratosis congenita, a third of patients may have no or only subtle phenotypical abnormalities, including alterations in skin or hair pigmentation, skeletal and growth abnormalities, and endocrine disorders.³⁰ The International Fanconi Anemia Registry identified central nervous system, genitourinary, skin and musculoskeletal, ophthalmic, and gastrointestinal system malformations among children with Fanconi anemia.^31,32 Patients with dyskeratosis congenita may present with pulmonary fibrosis, hepatic cirrhosis, or premature graying, as highlighted in a recent study by DiNardo and colleagues.³³ Therefore, physicians must have a heightened index of suspicion in patients with subtle phenotypical findings and associated cytopenias.

Diagnosis

Differential Diagnosis

The diagnosis of aplastic anemia should be suspected in any patient presenting with pancytopenia. Aplastic anemia is a diagnosis of exclusion.³⁴ Other conditions associated with peripheral blood pancytopenia should be considered including infections (HIV, hepatitis, parvovirus B19, cytomegalovirus, Epstein-Barr virus, varicella-zoster virus), nutritional deficiencies (vitamin B12, folate, copper, zinc), autoimmune disease (systemic lupus erythematosus, rheumatoid arthritis, hemophagocytic lymphohistiocytosis), hypersplenism, marrow-occupying diseases (eg, leukemia, lymphoma, MDS), solid malignancies, and fibrosis (Table).⁷

Diagnostic Evaluation

The workup for aplastic anemia should include a thorough history and physical exam to search simultaneously for alternative diagnoses and clues pointing to potential etiologic agents.⁷ Diagnostic tests to be performed include a complete blood count with differential, reticulocyte count, immature platelet fraction, flow cytometry (to rule out lymphoproliferative disorders and atypical myeloid cells and to evaluate for PNH), and bone marrow biopsy with subsequent cytogenetic, immunohistochemical, and molecular testing.³⁵ The typical findings in aplastic anemia include peripheral blood pancytopenia without dysplastic features and bone marrow biopsy demonstrating a hypocellular marrow.⁷ A relative lymphocytosis in the peripheral blood is common.⁷ In patients with a significant PNH clone, a macrocytosis along with elevated lactate dehydrogenase and elevated reticulocyte and granulocyte counts may be present.³⁶

The diagnosis (based on the Camitta criteria³⁷ and modified Camitta criteria³⁸ for severe aplastic anemia) requires 2 of the following findings on peripheral blood samples:

• Absolute neutrophil count (ANC) < 500 cells/µL
• Platelet count < 20,000 cells/µL
• Reticulocyte count < 1% corrected or < 20,000 cells/µL.³⁵

In addition to peripheral blood findings, bone marrow biopsy is essential for the diagnosis, and should demonstrate a markedly hypocellular marrow (cellularity < 25%), occasionally with an increase in T lymphocytes.^7,39 Because marrow cellularity varies with age and can be challenging to assess, additional biopsies may be needed to confirm the diagnosis.²⁹ A 1- to 2-cm core biopsy is necessary to confirm hypocellularity, as small areas of residual hematopoiesis may be present and obscure the diagnosis.³⁵

Excluding Hypocellular MDS and IMFS

A diagnostic challenge is the exclusion of hypocellular MDS, especially in the older adult presenting with aplastic anemia, as patients with aplastic anemia may have some degree of erythroid dysplasia on bone marrow morphology.³⁶ The presence of a PNH clone on flow cytometry can aid in diagnosing aplastic anemia and excluding MDS,³⁴ although PNH clones can be present in refractory anemia MDS. Patients with aplastic anemia have a lower ratio of CD34+ cells compared to those with hypoplastic MDS, with one study demonstrating a mean CD34+ percentage of < 0.5% in aplastic anemia versus 3.7% in hypoplastic MDS.⁴⁰ Cytogenetic and molecular testing can also aid in making this distinction by identifying mutations commonly implicated in MDS.⁷ The presence of monosomy 7 (-7) in aplastic anemia patients is associated with a poor overall prognosis.^34,41

Peripheral blood screening using chromosome breakage analysis (done using either mitomycin C or diepoxybutane as in vitro DNA-crosslinking agents) and telomere length testing (of peripheral blood leukocytes) is necessary to exclude the main IMFS, Fanconi anemia and telomere biology disorders, respectively. Ruling out these conditions is imperative, as the approach to treatment varies significantly between IMFS and aplastic anemia. Patients with shortened telomeres should undergo genetic screening for mutations in the telomere maintenance genes to evaluate the underlying defect leading to shortened telomeres. Patients with increased peripheral blood breakage should have genetic testing to detect mutations associated with Fanconi anemia.

Classification

Once the diagnosis of aplastic anemia has been made, the patient should be classified according to the severity of their disease. Disease severity is determined based on peripheral blood ANC:³⁴ non-severe aplastic anemia (NSAA), ANC > 500 polymorphonuclear neutrophils (PMNs)/µL; severe aplastic anemia (SAA), 200–500 PMNs/µL; and very severe (VSAA), 0–200 PMNs/µL.^4,34 Disease classification is important, as VSAA is associated with a decreased OS compared to SAA.² Disease classification may affect treatment decisions, as patients with NSAA may be observed for a short period of time, while conversely patients with SAA have a worse prognosis with delays in therapy.^42-44

Summary

Aplastic anemia is a rare but potentially life-threatening disorder characterized by pancytopenia and a marked reduction in the hematopoietic stem cell compartment. It can be acquired or associated with an IMFS, and the treatment and prognosis vary dramatically between these 2 etiologies. Work-up and diagnosis involves investigating IMFSs and ruling out malignant or infectious etiologies for pancytopenia. After aplastic anemia has been diagnosed, the patient should be classified according to the severity of their disease based on peripheral blood ANC.

The past year was an exciting one in obstetrics. The landmark ARRIVE trial presented at the Society for Maternal-Fetal Medicine’s (SMFM) annual meeting and subsequently published in the New England Journal of Medicine contradicted a long-held belief about the safety of elective labor induction. In a large randomized trial, Cahill and colleagues took a controversial but practical clinical question about second-stage labor management and answered it for the practicing obstetrician in the trenches. Finally, the American College of Obstetricians and Gynecologists (ACOG) placed new emphasis on the oft overlooked but increasingly more complicated postpartum period, offering guidance to support improving care for women in this transitional period.

Ultimately, this was the year of the patient, as research, clinical guidelines, and education focused on how to achieve the best in safety and quality of care for delivery planning, the delivery itself, and the so-called fourth trimester.

ARRIVE: Labor induction at 39 weeks reduces CD rate with no difference in perinatal death or serious outcomes 

Grobman WA, Rice MM, Reddy UM, et al; for the Eunice Kennedy Shriver National Institute of Child Health and Human Development Maternal-Fetal Medicine Units Network. Labor induction versus expectant management in low-risk nulliparous women. N Engl J Med. 2018;379:513-523.

The term "elective induction of labor" has long had a negative connotation because of its association with increased CD rates and adverse perinatal outcomes. This view was based on results from older observational studies that compared outcomes for labor induction with those of spontaneous labor. In more recent observational studies that more appropriately compared labor induction with expectant management, however, elective induction of labor appears to be associated with similar CD rates and perinatal outcomes. 
To test the hypothesis that elective induction would have a lower risk for perinatal death or severe neonatal complications than expectant management in low-risk nulliparous women, Grobman and colleagues conducted A Randomized Trial of Induction Versus Expectant Management (ARRIVE).1 

Study population, timing of delivery, and trial outcomes 

This randomized controlled trial included 6,106 women at 41 US centers in the Maternal-Fetal Medicine Units Network of the Eunice Kennedy Shriver National Institute of Child Health and Human Development. Study participants were low-risk nulliparous women with a singleton vertex fetus who were randomly assigned to induction of labor at 39 to 39 4/7 weeks (n = 3,062) or expectant management (n = 3,044) until 40 5/7 to 42 2/7 weeks. 

"Low risk" was defined as having no maternal or fetal indication for delivery prior to 40 5/7 weeks. Reliable gestational dating was required.  

While no specific protocol for induction of labor management was required, there were 2 requests: 1) Cervical ripening was requested for an unfavorable cervix (63% of participants had a modified Bishop score <5), and 2) a duration of at least 12 hours after cervical ripening, rupture of membranes, and use of uterine stimulant was requested before performing a CD for "failed induction" (if medically appropriate). 

The primary outcome was a composite of perinatal death or serious neonatal complications. The main secondary outcome was CD. 

Potentially game-changing findings 

The investigators found that there was no statistically significant difference between the elective induction and expectant management groups for the primary composite perinatal outcome (4.3% vs 5.4%; P = .049, with P<.046 prespecified for significance). In addition, the rate of CD was significantly lower in the labor induction group than in the expectant management group (18.6% vs 22.2%; P<.001). 
Other significant findings in secondary outcomes included the following: 

• Hypertensive disorders of pregnancy were significantly lower in the labor induction group compared with the expectant management group (9.1% vs 14.1%; P<.001). 
• The labor induction group had a longer length of stay in the labor and delivery unit but a shorter postpartum hospital stay. 
• The labor induction group reported less pain and more control during labor. 

Results refute negative notion of elective labor induction 

The authors concluded that in a low-risk nulliparous patient population, elective induction of labor at 39 weeks does not increase the risk for adverse perinatal outcomes and decreases the rate of CD and hypertensive disorders of pregnancy. Additionally, they noted that induction at 39 weeks should not be avoided with the goal of preventing CD, as even women with an unfavorable cervix had a lower rate of CD in the induction group compared with the expectant management group.

Continue to: Immediate pushing in second stage...

Immediate pushing in second stage offers benefits and is preferable to delayed pushing 

Cahill AG, Srinivas SK, Tita AT, et al. Effect of immediate vs delayed pushing on rates of spontaneous vaginal delivery among nulliparous women receiving neuraxial analgesia: a randomized clinical trial. JAMA. 2018;320:1444-1454. 

In a randomized trial of 2,414 women, Cahill and colleagues sought to answer a seemingly simple question: What is the best timing for pushing during the second stage of labor--immediate or delayed? 

Practical management of the second stage of labor (defined as complete cervical dilation to the delivery of the infant) varies by provider and setting, and previous data on pushing efforts are conflicting. Delayed pushing, or "laboring down," has been suggested to allow passive fetal rotation and to conserve maternal energy for pushing. Older studies have shown that delayed pushing decreases the rate of operative delivery. More recent study data have not demonstrated a difference between immediate and delayed pushing techniques on vaginal delivery rates and have noted that increased maternal and neonatal morbidities are associated with a longer second stage of labor. 

The recent trial by Cahill and colleagues was designed to determine the effect of these 2 techniques on spontaneous vaginal delivery rates and on maternal and neonatal morbidities.⁴ 

Large study population 

This randomized pragmatic trial was conducted at 6 centers in the United States. Study participants (2,404 women completed the study) were nulliparous women at 37 or more weeks' gestation with neuraxial anesthesia who were randomly assigned at complete cervical dilation either to immediate pushing (n = 1,200) or to delayed pushing, that is, instructed to wait 60 minutes before starting to push (n = 1,204). The obstetric provider determined the rest of the labor management. 

The primary outcome was the rate of spontaneous vaginal delivery. Secondary outcomes included duration of the second stage of labor, duration of active pushing, operative vaginal delivery, CD, and several maternal assessments (postpartum hemorrhage, chorioamnionitis, endometritis, and perineal lacerations). 

Both groups had similar vaginal delivery rates, differences in some measures 

There was no difference in the primary outcome between the 2 groups: The spontaneous vaginal delivery rate was 85.9% (n = 1,031) in the immediate pushing group and 86.5% (n = 1,041) in the delayed pushing group (P = .67).  
Analysis of secondary outcomes revealed several significant differences: 

• decreased total time for the second stage of labor in the immediate pushing group compared with the delayed pushing group (102.4 vs 134.2 minutes) but longer active pushing time (83.7 vs 74.5 minutes) 
• a lower rate of postpartum hemorrhage, chorioamnionitis in the second stage, neonatal acidemia, and suspected neonatal sepsis in the immediate pushing group 
• a higher rate of third-degree perineal lacerations in the immediate pushing group. 

 No difference was found between groups in rates of operative vaginal deliveries, CDs, endometritis, overall perineal lacerations, or spontaneous vaginal delivery by fetal station or occiput position. 

Authors' takeaway 

The authors concluded that since delayed pushing does not increase spontaneous vaginal delivery rates and increases the duration of the second stage of labor and both maternal and neonatal morbidity, immediate pushing may be preferred in this patient population. 

ACOG aims to optimize postpartum care 

American College of Obstetricians and Gynecologists. ACOG Committee Opinion No. 736. Optimizing postpartum care. Obstet Gynecol. 2018;131:e140-e150. 

In May 2018, ACOG released "Optimizing postpartum care," a committee opinion that proposes a new model of comprehensive postpartum care focused on improving both short- and long-term health outcomes for women and infants. (This replaces the June 2016 committee opinion No. 666.) Described as "the fourth trimester," the postpartum period is a critical transitional period in which both pregnancy-related and pre-existing conditions may affect maternal, neonatal, and family status; half of pregnancy-related maternal deaths occur during the postpartum period.⁶ 

The postpartum visit: Often a lost opportunity 

ACOG cites that up to 40% of women in the United States do not attend their postpartum visit.6 Many aspects of the postpartum visit, including follow-up for chronic diseases, mental health screening, and contraceptive counseling, provide opportunities for acute intervention as well as establishment of healthy behaviors. Some studies have shown that postpartum depression, breastfeeding, and patient satisfaction outcomes improve as a result of postpartum engagement. 

Continue to: ACOG's recommendations...

ACOG's recommendations 

Ongoing process. ACOG's first proposed change concerns the structure of the postpartum visit itself, which traditionally has been a single visit with a provider at approximately 6 weeks postpartum. Postpartum care plans actually should be started before birth, during regular prenatal care, and adjusted in the hospital as needed so that the provider can educate patients about the issues they may face and resources they may need during this time. This prenatal preparation hopefully will encourage more patients to attend their postpartum visits. 

Increased provider contact. Another proposed change is that after delivery, the patient should have contact with a provider within the first 3 weeks postpartum. For high-risk patients, this may involve an in-person clinic visit as soon as 3 to 10 days postpartum (for hypertensive disorders of pregnancy) or at 1 to 2 weeks (for postpartum depression screening, incision checks, and lactation issues). For lower-risk patients, a phone call may be appropriate and/or preferred. Ongoing follow-up for all patients before the final postpartum visit should be individualized. 

Postpartum visit and care transition. ACOG recommends a comprehensive postpartum visit at 4 to 12 weeks to fully evaluate the woman's physical, social, and psychologic well-being and to serve as a transition from pregnancy care to well-woman care. This is a large order and includes evaluation of the following: 

• mood and emotional well-being 
• infant care and feeding 
• sexuality, contraception, and birth spacing 
• sleep and fatigue 
• physical recovery from birth 
• chronic disease management and transition to primary care provider 
• health maintenance 
• review of labor and delivery course if needed 
• review of risks and recommendations for future pregnancies. 

After these components are addressed, it is expected that the patient will be transitioned to a primary care provider (who may continue to be the ObGyn, as appropriate) to coordinate her future care in the primary medical home.  

Useful resource for adopting new paradigm 

ACOG's recommendations are somewhat daunting, and these changes will require education and resources, a significant increase in obstetric provider time and effort, and consideration of policy change regarding such issues as parental leave and postpartum care reimbursement. As a start, ACOG has developed an online aid for health care providers called "Postpartum toolkit" (https://www.acog.org/About-ACOG/ACOG-Departments/Toolkits-for-Health-Care-Providers/Postpartum-Toolkit), which provides education and resources for all steps in the process and can be individualized for each practice and patient.⁷