The Journal of Family Practice

Earlier this year, the Centers for Disease Control and Prevention (CDC) published a clinical practice guideline aimed at decreasing opioid use in the treatment of chronic pain.¹ It developed this guideline in response to the increasing problem of opioid abuse and opioid-related mortality in the United States.

The CDC notes that an estimated 1.9 million people abused or were dependent on prescription opioid pain medication in 2013.¹ Between 1999 and 2014, more than 165,000 people in the United States died from an overdose of opioid pain medication, with that rate increasing markedly in the past decade.¹ In 2011, an estimated 420,000 emergency department visits were related to the abuse of narcotic pain relievers.²

While the problem of increasing opioid-related abuse and deaths has been apparent for some time, effective interventions have been elusive. Evidence remains sparse on the benefits and harms of long-term opioid therapy for chronic pain, except for those at the end of life. Evidence has been insufficient to determine long-term benefits of opioid therapy vs no opioid therapy, although the potential for harms from high doses of opioids are documented. There is not much evidence comparing nonpharmacologic and non-opioid pharmacologic treatments with long-term opioid therapy.

This lack of an evidence base is reflected in the CDC guideline. Of the guideline’s 12 recommendations, not one has high-level supporting evidence and only one has even moderate-level evidence behind it. Four recommendations are supported by low-level evidence, and 7 by very-low-level evidence. Yet 11 of the 12 are given an A recommendation, meaning that the guideline panel feels that most patients should receive this course of action.

Methodology used to create the guideline

The guideline committee used a modified GRADE approach (Grading of Recommendations Assessment, Development, and Evaluation) to develop the guideline. It is the same system the Advisory Committee on Immunization Practices adopted to assess and make recommendations on vaccines.³ The system’s classification of levels of evidence and recommendation categories are described in FIGURE 1.¹

The committee started by assessing evidence with a report on the long-term effectiveness of opioids for chronic pain, produced by the Agency for Health Care Research and Quality in 2014;⁴ it then augmented that report by performing an updated search for new evidence published since the report came out.⁵ The committee then conducted a “contextual evidence review”⁶ on the following 4 areas:

1. the effectiveness of nonpharmacologic (cognitive behavioral therapy, exercise therapy, interventional treatments, multimodal pain treatment) and non-opioid pharmacologic treatments (acetaminophen, nonsteroidal anti-inflammatory drugs, antidepressants, anticonvulsants)
2. the benefits and harms of opioid therapy
3. clinician and patient values and preferences related to opioids and medication risks, benefits, and use
4. resource allocation, including costs and economic analyses.

The guideline wording indicates that, for this contextual analysis, the committee used a rapid systematic review methodology, in part because of time constraints given the imperative to produce a guideline to address a pressing problem, and because of a recognition that evidence on the questions would be scant and not of high quality.¹ The 12 recommendations are categorized under 3 main headings.

Determining when to initiate or continue opioids for chronic pain

1. Nonpharmacologic therapy and non-opioid pharmacologic therapy are preferred for chronic pain. Consider opioid therapy only if you anticipate that benefits for both pain and function will outweigh risks to the patient. If opioids are used, combine them as appropriate with nonpharmacologic therapy and non-opioid pharmacologic therapy. (Recommendation category: A; evidence type: 3)

Review your state's prescription drug monitoring program to determine whether a patient is receiving opioid dosages, or dangerous combinations, that increase the risk for overdose.2. Before starting opioid therapy for chronic pain, establish treatment goals with the patient, including realistic goals for pain and function, and consider how therapy will be discontinued if the benefits do not outweigh the risks. Continue opioid therapy only if there is clinically meaningful improvement in pain and function that outweighs risks to patient safety. (Recommendation category: A; evidence type: 4)

3. Before starting opioid therapy, and periodically during its course, discuss with patients known risks and realistic benefits of opioid therapy and patient and clinician responsibilities for managing therapy. (Recommendation category: A; evidence type: 3)

Opioid selection, dosage, duration, follow-up, and discontinuation

4. When starting opioid therapy for chronic pain, prescribe immediate-release opioids instead of extended-release/long-acting (ER/LA) agents. (Recommendation category: A; evidence type: 4)

5. When starting opioids, prescribe the lowest effective dosage. Use caution when prescribing opioids at any dosage; carefully reassess the evidence for individual benefits and risks when increasing the dosage to ≥50 morphine milligram equivalents (MME)/d; and avoid increasing the dosage to ≥90 MME/d (or carefully justify such a decision, if made). (Recommendation category: A; evidence type: 3)

When starting opioid therapy for chronic pain, prescribe immediate-release opioids instead of extended-release/long-acting agents.6. Long-term opioid use often begins with treatment of acute pain. When opioids are used for acute pain, prescribe the lowest effective dose of immediate-release opioids at a quantity no greater than is needed for the expected duration of pain severe enough to require opioids. Three days or less will often be sufficient; more than 7 days will rarely be needed. (Recommendation category: A; evidence type: 4)

7. In monitoring opioid therapy for chronic pain, reevaluate benefits and harms with patients within one to 4 weeks of starting opioid therapy or escalating the dose. Also, evaluate the benefits and harms of continued therapy with patients every 3 months or more frequently. If the benefits of continued opioid therapy do not outweigh the harms, optimize other therapies and work with patients to taper opioids to lower dosages or to taper and discontinue them. (Recommendation category: A; evidence type: 4)

Assessing risk and addressing harms of opioid use

8. Before starting opioid therapy, and periodically during its continuation, evaluate risk factors for opioid-related harms. Incorporate strategies into the management plan to mitigate risk; consider offering naloxone when factors are present that increase the risk for opioid overdose—eg, a history of overdose, history of substance use disorder, higher opioid dosages (≥50 MME/d), or concurrent benzodiazepine use. (Recommendation category: A; evidence type: 4)

9. Review the patient’s history of controlled substance prescriptions. Use data from the state prescription drug monitoring program (PDMP) to determine whether the patient is receiving opioid dosages or dangerous combinations that put him or her at high risk for overdose. (State Web sites are available at http://www.pdmpassist.org/content/state-pdmp-websites.) Review PDMP data when starting opioid therapy for chronic pain and periodically during its continuation, at least every 3 months and with each new prescription. (Recommendation category: A; evidence type: 4)

10. Before prescribing opioids for chronic pain, use urine drug testing to assess for prescribed medications, as well as other controlled prescription drugs and illicit drugs, and consider urine drug testing at least annually. (Recommendation category: B; evidence type: 4)

11. Avoid prescribing opioid pain medication and benzodiazepines concurrently whenever possible. (Recommendation category: A; evidence type: 3)

See JFP's "3 in 3" video on urin drug testing at: http://bit.ly/2eL7Ptz.12. For patients with opioid use disorder, offer or arrange for evidence-based treatment (usually medication-assisted treatment with buprenorphine or methadone in combination with behavioral therapies). (Recommendation category: A; evidence type: 2)

Aids for guideline implementation

The CDC has produced materials to assist physicians in implementing this guideline, including checklists for prescribing or continuing opioids. The checklist for initiation of opioids is reproduced in FIGURE 2.⁷

The CDC is addressing a severe public health problem and doing so by using contemporary evidence-based methodology and guideline development processes. The lack of high-quality evidence on the topic and the use of a less-than-optimal evidence review process for some key questions may hamper this effort. However, given the prominence of the CDC, this clinical guideline will likely be considered the standard of care for family physicians.

Earlier this year, the Centers for Disease Control and Prevention (CDC) published a clinical practice guideline aimed at decreasing opioid use in the treatment of chronic pain.¹ It developed this guideline in response to the increasing problem of opioid abuse and opioid-related mortality in the United States.

The CDC notes that an estimated 1.9 million people abused or were dependent on prescription opioid pain medication in 2013.¹ Between 1999 and 2014, more than 165,000 people in the United States died from an overdose of opioid pain medication, with that rate increasing markedly in the past decade.¹ In 2011, an estimated 420,000 emergency department visits were related to the abuse of narcotic pain relievers.²

While the problem of increasing opioid-related abuse and deaths has been apparent for some time, effective interventions have been elusive. Evidence remains sparse on the benefits and harms of long-term opioid therapy for chronic pain, except for those at the end of life. Evidence has been insufficient to determine long-term benefits of opioid therapy vs no opioid therapy, although the potential for harms from high doses of opioids are documented. There is not much evidence comparing nonpharmacologic and non-opioid pharmacologic treatments with long-term opioid therapy.

This lack of an evidence base is reflected in the CDC guideline. Of the guideline’s 12 recommendations, not one has high-level supporting evidence and only one has even moderate-level evidence behind it. Four recommendations are supported by low-level evidence, and 7 by very-low-level evidence. Yet 11 of the 12 are given an A recommendation, meaning that the guideline panel feels that most patients should receive this course of action.

Methodology used to create the guideline

The guideline committee used a modified GRADE approach (Grading of Recommendations Assessment, Development, and Evaluation) to develop the guideline. It is the same system the Advisory Committee on Immunization Practices adopted to assess and make recommendations on vaccines.³ The system’s classification of levels of evidence and recommendation categories are described in FIGURE 1.¹

The committee started by assessing evidence with a report on the long-term effectiveness of opioids for chronic pain, produced by the Agency for Health Care Research and Quality in 2014;⁴ it then augmented that report by performing an updated search for new evidence published since the report came out.⁵ The committee then conducted a “contextual evidence review”⁶ on the following 4 areas:

1. the effectiveness of nonpharmacologic (cognitive behavioral therapy, exercise therapy, interventional treatments, multimodal pain treatment) and non-opioid pharmacologic treatments (acetaminophen, nonsteroidal anti-inflammatory drugs, antidepressants, anticonvulsants)
2. the benefits and harms of opioid therapy
3. clinician and patient values and preferences related to opioids and medication risks, benefits, and use
4. resource allocation, including costs and economic analyses.

The guideline wording indicates that, for this contextual analysis, the committee used a rapid systematic review methodology, in part because of time constraints given the imperative to produce a guideline to address a pressing problem, and because of a recognition that evidence on the questions would be scant and not of high quality.¹ The 12 recommendations are categorized under 3 main headings.

Determining when to initiate or continue opioids for chronic pain

1. Nonpharmacologic therapy and non-opioid pharmacologic therapy are preferred for chronic pain. Consider opioid therapy only if you anticipate that benefits for both pain and function will outweigh risks to the patient. If opioids are used, combine them as appropriate with nonpharmacologic therapy and non-opioid pharmacologic therapy. (Recommendation category: A; evidence type: 3)

Review your state's prescription drug monitoring program to determine whether a patient is receiving opioid dosages, or dangerous combinations, that increase the risk for overdose.2. Before starting opioid therapy for chronic pain, establish treatment goals with the patient, including realistic goals for pain and function, and consider how therapy will be discontinued if the benefits do not outweigh the risks. Continue opioid therapy only if there is clinically meaningful improvement in pain and function that outweighs risks to patient safety. (Recommendation category: A; evidence type: 4)

3. Before starting opioid therapy, and periodically during its course, discuss with patients known risks and realistic benefits of opioid therapy and patient and clinician responsibilities for managing therapy. (Recommendation category: A; evidence type: 3)

Opioid selection, dosage, duration, follow-up, and discontinuation

4. When starting opioid therapy for chronic pain, prescribe immediate-release opioids instead of extended-release/long-acting (ER/LA) agents. (Recommendation category: A; evidence type: 4)

5. When starting opioids, prescribe the lowest effective dosage. Use caution when prescribing opioids at any dosage; carefully reassess the evidence for individual benefits and risks when increasing the dosage to ≥50 morphine milligram equivalents (MME)/d; and avoid increasing the dosage to ≥90 MME/d (or carefully justify such a decision, if made). (Recommendation category: A; evidence type: 3)

When starting opioid therapy for chronic pain, prescribe immediate-release opioids instead of extended-release/long-acting agents.6. Long-term opioid use often begins with treatment of acute pain. When opioids are used for acute pain, prescribe the lowest effective dose of immediate-release opioids at a quantity no greater than is needed for the expected duration of pain severe enough to require opioids. Three days or less will often be sufficient; more than 7 days will rarely be needed. (Recommendation category: A; evidence type: 4)

7. In monitoring opioid therapy for chronic pain, reevaluate benefits and harms with patients within one to 4 weeks of starting opioid therapy or escalating the dose. Also, evaluate the benefits and harms of continued therapy with patients every 3 months or more frequently. If the benefits of continued opioid therapy do not outweigh the harms, optimize other therapies and work with patients to taper opioids to lower dosages or to taper and discontinue them. (Recommendation category: A; evidence type: 4)

Assessing risk and addressing harms of opioid use

8. Before starting opioid therapy, and periodically during its continuation, evaluate risk factors for opioid-related harms. Incorporate strategies into the management plan to mitigate risk; consider offering naloxone when factors are present that increase the risk for opioid overdose—eg, a history of overdose, history of substance use disorder, higher opioid dosages (≥50 MME/d), or concurrent benzodiazepine use. (Recommendation category: A; evidence type: 4)

9. Review the patient’s history of controlled substance prescriptions. Use data from the state prescription drug monitoring program (PDMP) to determine whether the patient is receiving opioid dosages or dangerous combinations that put him or her at high risk for overdose. (State Web sites are available at http://www.pdmpassist.org/content/state-pdmp-websites.) Review PDMP data when starting opioid therapy for chronic pain and periodically during its continuation, at least every 3 months and with each new prescription. (Recommendation category: A; evidence type: 4)

10. Before prescribing opioids for chronic pain, use urine drug testing to assess for prescribed medications, as well as other controlled prescription drugs and illicit drugs, and consider urine drug testing at least annually. (Recommendation category: B; evidence type: 4)

11. Avoid prescribing opioid pain medication and benzodiazepines concurrently whenever possible. (Recommendation category: A; evidence type: 3)

See JFP's "3 in 3" video on urin drug testing at: http://bit.ly/2eL7Ptz.12. For patients with opioid use disorder, offer or arrange for evidence-based treatment (usually medication-assisted treatment with buprenorphine or methadone in combination with behavioral therapies). (Recommendation category: A; evidence type: 2)

Aids for guideline implementation

The CDC has produced materials to assist physicians in implementing this guideline, including checklists for prescribing or continuing opioids. The checklist for initiation of opioids is reproduced in FIGURE 2.⁷

The CDC is addressing a severe public health problem and doing so by using contemporary evidence-based methodology and guideline development processes. The lack of high-quality evidence on the topic and the use of a less-than-optimal evidence review process for some key questions may hamper this effort. However, given the prominence of the CDC, this clinical guideline will likely be considered the standard of care for family physicians.

Earlier this year, the Centers for Disease Control and Prevention (CDC) published a clinical practice guideline aimed at decreasing opioid use in the treatment of chronic pain.¹ It developed this guideline in response to the increasing problem of opioid abuse and opioid-related mortality in the United States.

The CDC notes that an estimated 1.9 million people abused or were dependent on prescription opioid pain medication in 2013.¹ Between 1999 and 2014, more than 165,000 people in the United States died from an overdose of opioid pain medication, with that rate increasing markedly in the past decade.¹ In 2011, an estimated 420,000 emergency department visits were related to the abuse of narcotic pain relievers.²

While the problem of increasing opioid-related abuse and deaths has been apparent for some time, effective interventions have been elusive. Evidence remains sparse on the benefits and harms of long-term opioid therapy for chronic pain, except for those at the end of life. Evidence has been insufficient to determine long-term benefits of opioid therapy vs no opioid therapy, although the potential for harms from high doses of opioids are documented. There is not much evidence comparing nonpharmacologic and non-opioid pharmacologic treatments with long-term opioid therapy.

This lack of an evidence base is reflected in the CDC guideline. Of the guideline’s 12 recommendations, not one has high-level supporting evidence and only one has even moderate-level evidence behind it. Four recommendations are supported by low-level evidence, and 7 by very-low-level evidence. Yet 11 of the 12 are given an A recommendation, meaning that the guideline panel feels that most patients should receive this course of action.

Methodology used to create the guideline

The guideline committee used a modified GRADE approach (Grading of Recommendations Assessment, Development, and Evaluation) to develop the guideline. It is the same system the Advisory Committee on Immunization Practices adopted to assess and make recommendations on vaccines.³ The system’s classification of levels of evidence and recommendation categories are described in FIGURE 1.¹

The committee started by assessing evidence with a report on the long-term effectiveness of opioids for chronic pain, produced by the Agency for Health Care Research and Quality in 2014;⁴ it then augmented that report by performing an updated search for new evidence published since the report came out.⁵ The committee then conducted a “contextual evidence review”⁶ on the following 4 areas:

1. the effectiveness of nonpharmacologic (cognitive behavioral therapy, exercise therapy, interventional treatments, multimodal pain treatment) and non-opioid pharmacologic treatments (acetaminophen, nonsteroidal anti-inflammatory drugs, antidepressants, anticonvulsants)
2. the benefits and harms of opioid therapy
3. clinician and patient values and preferences related to opioids and medication risks, benefits, and use
4. resource allocation, including costs and economic analyses.

The guideline wording indicates that, for this contextual analysis, the committee used a rapid systematic review methodology, in part because of time constraints given the imperative to produce a guideline to address a pressing problem, and because of a recognition that evidence on the questions would be scant and not of high quality.¹ The 12 recommendations are categorized under 3 main headings.

Determining when to initiate or continue opioids for chronic pain

1. Nonpharmacologic therapy and non-opioid pharmacologic therapy are preferred for chronic pain. Consider opioid therapy only if you anticipate that benefits for both pain and function will outweigh risks to the patient. If opioids are used, combine them as appropriate with nonpharmacologic therapy and non-opioid pharmacologic therapy. (Recommendation category: A; evidence type: 3)

Review your state's prescription drug monitoring program to determine whether a patient is receiving opioid dosages, or dangerous combinations, that increase the risk for overdose.2. Before starting opioid therapy for chronic pain, establish treatment goals with the patient, including realistic goals for pain and function, and consider how therapy will be discontinued if the benefits do not outweigh the risks. Continue opioid therapy only if there is clinically meaningful improvement in pain and function that outweighs risks to patient safety. (Recommendation category: A; evidence type: 4)

3. Before starting opioid therapy, and periodically during its course, discuss with patients known risks and realistic benefits of opioid therapy and patient and clinician responsibilities for managing therapy. (Recommendation category: A; evidence type: 3)

Opioid selection, dosage, duration, follow-up, and discontinuation

4. When starting opioid therapy for chronic pain, prescribe immediate-release opioids instead of extended-release/long-acting (ER/LA) agents. (Recommendation category: A; evidence type: 4)

5. When starting opioids, prescribe the lowest effective dosage. Use caution when prescribing opioids at any dosage; carefully reassess the evidence for individual benefits and risks when increasing the dosage to ≥50 morphine milligram equivalents (MME)/d; and avoid increasing the dosage to ≥90 MME/d (or carefully justify such a decision, if made). (Recommendation category: A; evidence type: 3)

When starting opioid therapy for chronic pain, prescribe immediate-release opioids instead of extended-release/long-acting agents.6. Long-term opioid use often begins with treatment of acute pain. When opioids are used for acute pain, prescribe the lowest effective dose of immediate-release opioids at a quantity no greater than is needed for the expected duration of pain severe enough to require opioids. Three days or less will often be sufficient; more than 7 days will rarely be needed. (Recommendation category: A; evidence type: 4)

7. In monitoring opioid therapy for chronic pain, reevaluate benefits and harms with patients within one to 4 weeks of starting opioid therapy or escalating the dose. Also, evaluate the benefits and harms of continued therapy with patients every 3 months or more frequently. If the benefits of continued opioid therapy do not outweigh the harms, optimize other therapies and work with patients to taper opioids to lower dosages or to taper and discontinue them. (Recommendation category: A; evidence type: 4)

Assessing risk and addressing harms of opioid use

8. Before starting opioid therapy, and periodically during its continuation, evaluate risk factors for opioid-related harms. Incorporate strategies into the management plan to mitigate risk; consider offering naloxone when factors are present that increase the risk for opioid overdose—eg, a history of overdose, history of substance use disorder, higher opioid dosages (≥50 MME/d), or concurrent benzodiazepine use. (Recommendation category: A; evidence type: 4)

9. Review the patient’s history of controlled substance prescriptions. Use data from the state prescription drug monitoring program (PDMP) to determine whether the patient is receiving opioid dosages or dangerous combinations that put him or her at high risk for overdose. (State Web sites are available at http://www.pdmpassist.org/content/state-pdmp-websites.) Review PDMP data when starting opioid therapy for chronic pain and periodically during its continuation, at least every 3 months and with each new prescription. (Recommendation category: A; evidence type: 4)

10. Before prescribing opioids for chronic pain, use urine drug testing to assess for prescribed medications, as well as other controlled prescription drugs and illicit drugs, and consider urine drug testing at least annually. (Recommendation category: B; evidence type: 4)

11. Avoid prescribing opioid pain medication and benzodiazepines concurrently whenever possible. (Recommendation category: A; evidence type: 3)

See JFP's "3 in 3" video on urin drug testing at: http://bit.ly/2eL7Ptz.12. For patients with opioid use disorder, offer or arrange for evidence-based treatment (usually medication-assisted treatment with buprenorphine or methadone in combination with behavioral therapies). (Recommendation category: A; evidence type: 2)

Aids for guideline implementation

The CDC has produced materials to assist physicians in implementing this guideline, including checklists for prescribing or continuing opioids. The checklist for initiation of opioids is reproduced in FIGURE 2.⁷

The CDC is addressing a severe public health problem and doing so by using contemporary evidence-based methodology and guideline development processes. The lack of high-quality evidence on the topic and the use of a less-than-optimal evidence review process for some key questions may hamper this effort. However, given the prominence of the CDC, this clinical guideline will likely be considered the standard of care for family physicians.

Nasal obstruction is one of the most common reasons that patients visit their primary care providers.^1,2 Often described by patients as nasal congestion or the inability to adequately breathe out of one or both nostrils during the day and/or night, nasal obstruction commonly interferes with a patient’s ability to eat, sleep, and function, thereby significantly impacting quality of life. Overlapping presentations can make discerning the exact cause of nasal obstruction difficult.

To improve diagnosis and treatment, we review here the evidence-based recommendations for the most common causes of nasal obstruction: rhinitis, rhinosinusitis (RS), drug-induced nasal obstruction, and mechanical/structural abnormalities (TABLE 1^3-14).

Rhinitis/rhinosinusitis: It all begins with inflammation

Sneezing, rhinorrhea, nasal congestion, and nasal itching are complaints that signal rhinitis, which affects 30 to 60 million people in the United States annually.³ Rhinitis can be allergic, non-allergic, infectious, hormonal, or occupational in nature. All forms of rhinitis share inflammation as the cause of the nasal obstruction. The most common form is allergic rhinitis (AR), which includes seasonal AR and perennial AR. Seasonal AR is typically caused by outdoor allergens and waxes and wanes with pollen seasons. Perennial AR is caused mostly by indoor allergens, such as dust mites, molds, cockroaches, and pet dander; it persists all or most of the year.⁶ Causes of non-allergic rhinitis (NAR) include environmental irritants such as cigarette smoke, perfume, and car exhaust; medications; and hormonal changes,⁶ but most causes of NAR are unknown.^3,6

While AR can begin at any age, most people develop symptoms in childhood or as young adults, whereas NAR tends to begin later in life. Nasal itching can help to distinguish AR from NAR. NAR symptoms tend to be perennial and include postnasal drainage. If symptoms persist longer than 12 weeks despite treatment, the condition becomes known as chronic rhinosinusitis (CRS).

Treatment of rhinitis: Tiered and often continuous

Treatment of AR and NAR is similar and multitiered beginning with the avoidance of irritants and/or allergens whenever possible, moving on to pharmacotherapy, and, at least for AR, ending with allergen immunotherapy. Treatment is often an ongoing process and typically requires continuous therapy as opposed to treatment on an as-needed basis.³ It is unnecessary to perform allergy testing before making a presumed diagnosis of NAR and starting treatment.⁶

Intranasal corticosteroids. Currently, intranasal glucocorticosteroids (INGCs) are the most effective monotherapy for AR and NAR and have few adverse effects when used at prescribed doses.^3,4 For mild to intermittent symptoms, begin with the maximum dosage of an INGC for the patient’s age and proceed with incremental reductions to identify the lowest effective dose.³ If INGCs alone are ineffective, studies have shown that the addition of an intranasal second-generation antihistamine can be of some benefit.^3,4 In fact, an INGC and an intranasal antihistamine—along with saline nasal irrigation—is recommended for both AR and NAR resistant to single therapy.^3,6,15 If intranasal antihistamines are not an option, oral therapy can be initiated.

Start with second-generation antihistamines and consider LRAs. For oral therapy, start with second-generation antihistamines (loratadine, cetirizine, fexofenadine). First-generation antihistamines (diphenhydramine, hydroxyzine, chlorpheniramine), although widely available at relatively low cost, can cause several significant adverse effects including sedation, impaired cognitive function, and agitation in children.^3,4 Because second-generation antihistamines have fewer adverse effects, they are recommended as first-line therapy when oral antihistamine therapy is desired, such as for nasal congestion, sneezing, and itchy, watery eyes.

When prescribing intranasal corticosteroids for allergic and nonallergic rhinitis, begin with the maximum dosage and then incrementally reduce the amount to identify the lowest effective dose.Of note: A 2014 meta-analysis found that a leukotriene receptor antagonist (LRA) (montelukast) had efficacy similar to oral antihistamines for symptom relief in AR, and that LRAs may be better suited to nighttime symptoms (difficulty falling asleep, nighttime awakenings, congestion on awakening), while antihistamines may provide better relief of daytime symptoms (pruritus, rhinorrhea, sneezing).¹⁶ Although further head-to-head, double-blind randomized controlled trials (RCTs) are needed to confirm the results and investigate possible gender differences in symptom response, consider an LRA for first-line therapy in patients with AR who have predominantly nighttime symptoms.

What about pregnant women and the elderly?

It is important to consider teratogenicity when selecting medications for pregnant patients, especially during the first trimester.³ Nasal cromolyn has the most reassuring safety profile in pregnancy. Cetirizine, chlorpheniramine, loratadine, diphenhydramine, and tripelennamine may be used in pregnancy. The US Food and Drug Administration considers them to have a low risk of fetal harm, based on human data, whereas it views many other antihistamines as probably safe, based on limited or no human data. Most INGCs are not expected to cause fetal harm, but limited human data are available. Avoid prescribing oral decongestants to women who are in the first trimester of pregnancy due to the risk of gastroschisis in newborns.¹⁷

Elderly patients represent another population for which adverse effects must be carefully considered. Allergies in individuals >65 years of age are uncommon. Rhinitis in this age group is often secondary to cholinergic hyperactivity, alpha-adrenergic hyperactivity, or rhinosinusitis. Given elderly patients’ increased susceptibility to the potential adverse central nervous system (CNS) and anticholinergic effects of antihistamines, non-sedating medications are recommended. Oral decongestants also should be used with caution in this population, not only because of CNS effects, but also because of heart and bladder effects³ (TABLE 2¹⁸).

For drug-induced rhinitis, stop the offending drug and consider an INGC

Several types of medications, both oral and inhaled, are known to cause rhinitis. The use of alpha-adrenergic decongestant sprays for more than 5 to 7 days can induce rebound congestion on withdrawal, known as rhinitis medicamentosa.³ Repeated use of intranasal cocaine and methamphetamines can also result in rebound congestion. Oral medications that can result in rhinitis or congestion include angiotensin-converting enzyme (ACE) inhibitors, beta-blockers, nonsteroidal anti-inflammatory drugs (NSAIDS), oral contraceptives, and even antidepressants.³

The treatment for drug-induced rhinitis is termination of the offending agent. INGCs can be used to help decrease inflammation and control symptoms once the offending agent is discontinued.

Mechanical/structural causes of obstruction are wide-ranging

Mechanical/structural causes of nasal obstruction range from foreign bodies to anatomical variations including nasal polyps, a deviated septum, adenoidal hypertrophy, foreign bodies, and tumors. Because more than one etiology may be at work, it is best to first treat any non-mechanical causes of obstruction, such as ARS or NARS.

Nasal polyposis often requires both a medical and surgical approach

Nasal polyps are benign growths arising from the mucosa of the nasal sinuses and nasal cavities and affecting up to 4% of the population.⁷ Their etiology is unclear, but we do know that nasal polyps result from underlying inflammation.⁷ Uncommon in children outside of those affected by cystic fibrosis,⁷ nasal polyposis can be associated with disease processes such as AR and sinusitis. Polyps are also associated with clinical syndromes such as aspirin-exacerbated respiratory disease (AERD) syndrome, which involves upper and lower respiratory tract symptoms in patients with asthma who have taken aspirin or other NSAIDs.⁹

Leukotriene receptor agonists may be better suited to nighttime symptoms of nasal obstruction, while oral antihistamines may be better suited to daytime symptoms.Symptoms vary with the location and size of the polyps, but generally include nasal congestion, alteration in smell, and rhinorrhea. The goals of treatment are to restore or improve nasal breathing and olfaction and prevent recurrence.⁸ This often requires both a medical and surgical approach.

Topical corticosteroids are effective at reducing both the size of polyps and associated symptoms (rhinorrhea, rhinitis).⁸ And research has shown that steroids reduce the need for both primary and repeat surgical polypectomies.⁴ Other treatments to consider prior to surgery (if no symptom reduction occurs with INGCs) include systemic (oral) corticosteroids, intra-polyp steroid injections, macrolide antibiotics, and nasal washes.^7,14

When symptoms of polyposis are refractory to medical management, functional endoscopic sinus surgery (FESS) is the surgical procedure of choice.³ In addition to refractory symptoms, indications for FESS include the need to correct anatomic deformities believed to be contributing to the persistence of disease and the need to debulk advanced nasal polyposis.³ The principal goal is to restore patency to the ostiomeatal unit.³

Several studies have reported a high success rate for FESS in improving the symptoms of CRS.^3,19-23 In a 1992 study, for example, 98% of patients reported improvement following surgery,¹⁹ and in a follow-up report approximately 6 years later, 98% of patients continued to report subjective improvement.²²

For septal etiologies, consider septoplasty

Deviation of the nasal septum is a common structural etiology for nasal obstruction arising primarily from congenital, genetic, or traumatic causes.²⁴ Turbulent airflow from the septal deviation often causes turbinate hypertrophy, which creates (or exacerbates) the obstructive symptoms from the septal deviation.²⁵

Septoplasty is the most common ear, nose, and throat operation in adults.²⁶ Reduction of nasal symptoms has been reported in up to 89% of patients who receive this surgery, according to one single-center, non-randomized trial.²⁷ Currently, at least one multicenter, randomized trial is underway that aims to develop evidence-based guidelines for septoplasty.²⁶

Septal perforation is another etiology that can present with nasal obstruction symptoms. Causes include traumatic perforation, inflammatory or collagen vascular diseases, infections, overuse of vasoconstrictive medications, and malignancy.^28,29 A careful inspection of the nasal septum is necessary to identify a perforation; this may require nasal endoscopy.

Although adenoidectomy is commonly performed to correct adenoid hypertrophy in children, current evidence regarding the efficacy of the procedure is inconclusive.Anterior, rather than posterior, perforations are more likely to cause symptoms of nasal obstruction. Posterior perforations rarely require treatment unless malignancy is suspected, in which case referral for biopsy is recommended. Anterior perforations are treated initially with avoidance of any causative agent if, for example, the problem is drug- or medication-induced, and then with humidification and emollients.^28,29

For anterior perforations, septal silicone buttons can be used for recalcitrant symptoms. However, observational studies indicate that for long-term symptom resolution, silicone buttons are effective in only about one-third of patients.²⁹

For patients with persistent symptoms despite the above measures, surgical repair with various flap techniques is an option. A meta-analysis of case studies involving various techniques concluded that there is a wide variety of options, and that surgeons must weigh factors such as the characteristics and etiology of the perforation and their own experience and expertise when choosing from among available methods.³⁰ Additional good quality research is necessary before clear recommendations regarding technique can be made.

Adenoid hypertrophy: Consider corticosteroid nasal drops

Adenoid hypertrophy is a common cause of chronic nasal obstruction in children. Although adenoidectomy is commonly performed to correct the problem, current evidence regarding the efficacy of the procedure is inconclusive.¹⁰ Evidence demonstrates corticosteroid nasal drops significantly reduce symptoms of nasal obstruction in children and may provide an effective alternative to surgical resection.¹⁸ Studies have also demonstrated that treatment with oral LRAs significantly reduces adenoid size and nasal obstruction symptoms.^12,13

Foreign bodies: Don’t forget “a mother’s kiss”

Foreign bodies are the most common cause of nasal obstruction in the pediatric population. There is a paucity of high-quality evidence on removal of these objects; however, a number of retrospective reviews and case series support that most objects can be removed in the office or emergency department without otolaryngologic referral.^31,32

Techniques for removal include positive pressure, which is best used for smooth or soft objects. Positive pressure techniques include having the patient blow their own nose or having a parent use a mouth-to-mouth–type blowing technique (ie, the “mother’s kiss” method).³² Refer patients to Otolaryngology if the obstruction involves:³¹

• objects not easily visualized by anterior rhinoscopy
• chronic or impacted objects
• button batteries or magnets
• penetrating or hooked objects
• any object that cannot be removed during an initial attempt.

Nasal tumors: More common in older men

Nasal tumors occur most often in the nasal cavity itself and are more common in men ≥60 years.³³ There is no notable racial predominance.³³ Other risk factors include human papillomavirus (HPV) infection, tobacco smoke, and occupational exposure to inhaled wood dust, glues, and adhesives.^34-37

Most foreign bodies can be removed in the office or emergency department without referral to Otolaryngology.Benign tumors occurring in the nasal cavity are a diverse group of disorders, including inverted papillomas, squamous papillomas, pyogenic granulomas, and other less common lesions, all of which typically present with nasal obstruction as a symptom. Many of these lesions cause local tissue destruction or have a high incidence of recurrence. These tumors are treated universally with nasoendoscopic resection.³⁸

Malignant nasal tumors are rare but serious causes of nasal obstruction, making up 3% of all head and neck cancers.³⁹ Most nasal cancers present when they are locally advanced and cause unilateral nasal obstruction, lacrimation, and epistaxis. These symptoms are typically refractory to initial medical management and present as CRS. This diagnosis should be suspected in certain patient groups, such as those who have been exposed to wood dust (eg, construction workers or those who work in wood mills).³⁶

Computed tomography is the gold standard imaging method for CRS; however, if nasal cancer is suspected, referral for biopsy and histopathologic examination is necessary for a final diagnosis.³⁹ Because of the nonspecific nature of their initial presentation, many nasal tumors are at an advanced stage and carry a poor prognosis by the time they are diagnosed.³⁹

CORRESPONDENCE
Margaret A. Bayard, MD, MPH, FAAFP, Naval Hospital Camp Pendleton, 200 Mercy Circle, Camp Pendleton, CA 92005; [email protected].

Nasal obstruction is one of the most common reasons that patients visit their primary care providers.^1,2 Often described by patients as nasal congestion or the inability to adequately breathe out of one or both nostrils during the day and/or night, nasal obstruction commonly interferes with a patient’s ability to eat, sleep, and function, thereby significantly impacting quality of life. Overlapping presentations can make discerning the exact cause of nasal obstruction difficult.

To improve diagnosis and treatment, we review here the evidence-based recommendations for the most common causes of nasal obstruction: rhinitis, rhinosinusitis (RS), drug-induced nasal obstruction, and mechanical/structural abnormalities (TABLE 1^3-14).

Rhinitis/rhinosinusitis: It all begins with inflammation

Sneezing, rhinorrhea, nasal congestion, and nasal itching are complaints that signal rhinitis, which affects 30 to 60 million people in the United States annually.³ Rhinitis can be allergic, non-allergic, infectious, hormonal, or occupational in nature. All forms of rhinitis share inflammation as the cause of the nasal obstruction. The most common form is allergic rhinitis (AR), which includes seasonal AR and perennial AR. Seasonal AR is typically caused by outdoor allergens and waxes and wanes with pollen seasons. Perennial AR is caused mostly by indoor allergens, such as dust mites, molds, cockroaches, and pet dander; it persists all or most of the year.⁶ Causes of non-allergic rhinitis (NAR) include environmental irritants such as cigarette smoke, perfume, and car exhaust; medications; and hormonal changes,⁶ but most causes of NAR are unknown.^3,6

While AR can begin at any age, most people develop symptoms in childhood or as young adults, whereas NAR tends to begin later in life. Nasal itching can help to distinguish AR from NAR. NAR symptoms tend to be perennial and include postnasal drainage. If symptoms persist longer than 12 weeks despite treatment, the condition becomes known as chronic rhinosinusitis (CRS).

Treatment of rhinitis: Tiered and often continuous

Treatment of AR and NAR is similar and multitiered beginning with the avoidance of irritants and/or allergens whenever possible, moving on to pharmacotherapy, and, at least for AR, ending with allergen immunotherapy. Treatment is often an ongoing process and typically requires continuous therapy as opposed to treatment on an as-needed basis.³ It is unnecessary to perform allergy testing before making a presumed diagnosis of NAR and starting treatment.⁶

Intranasal corticosteroids. Currently, intranasal glucocorticosteroids (INGCs) are the most effective monotherapy for AR and NAR and have few adverse effects when used at prescribed doses.^3,4 For mild to intermittent symptoms, begin with the maximum dosage of an INGC for the patient’s age and proceed with incremental reductions to identify the lowest effective dose.³ If INGCs alone are ineffective, studies have shown that the addition of an intranasal second-generation antihistamine can be of some benefit.^3,4 In fact, an INGC and an intranasal antihistamine—along with saline nasal irrigation—is recommended for both AR and NAR resistant to single therapy.^3,6,15 If intranasal antihistamines are not an option, oral therapy can be initiated.

Start with second-generation antihistamines and consider LRAs. For oral therapy, start with second-generation antihistamines (loratadine, cetirizine, fexofenadine). First-generation antihistamines (diphenhydramine, hydroxyzine, chlorpheniramine), although widely available at relatively low cost, can cause several significant adverse effects including sedation, impaired cognitive function, and agitation in children.^3,4 Because second-generation antihistamines have fewer adverse effects, they are recommended as first-line therapy when oral antihistamine therapy is desired, such as for nasal congestion, sneezing, and itchy, watery eyes.

When prescribing intranasal corticosteroids for allergic and nonallergic rhinitis, begin with the maximum dosage and then incrementally reduce the amount to identify the lowest effective dose.Of note: A 2014 meta-analysis found that a leukotriene receptor antagonist (LRA) (montelukast) had efficacy similar to oral antihistamines for symptom relief in AR, and that LRAs may be better suited to nighttime symptoms (difficulty falling asleep, nighttime awakenings, congestion on awakening), while antihistamines may provide better relief of daytime symptoms (pruritus, rhinorrhea, sneezing).¹⁶ Although further head-to-head, double-blind randomized controlled trials (RCTs) are needed to confirm the results and investigate possible gender differences in symptom response, consider an LRA for first-line therapy in patients with AR who have predominantly nighttime symptoms.

What about pregnant women and the elderly?

It is important to consider teratogenicity when selecting medications for pregnant patients, especially during the first trimester.³ Nasal cromolyn has the most reassuring safety profile in pregnancy. Cetirizine, chlorpheniramine, loratadine, diphenhydramine, and tripelennamine may be used in pregnancy. The US Food and Drug Administration considers them to have a low risk of fetal harm, based on human data, whereas it views many other antihistamines as probably safe, based on limited or no human data. Most INGCs are not expected to cause fetal harm, but limited human data are available. Avoid prescribing oral decongestants to women who are in the first trimester of pregnancy due to the risk of gastroschisis in newborns.¹⁷

Elderly patients represent another population for which adverse effects must be carefully considered. Allergies in individuals >65 years of age are uncommon. Rhinitis in this age group is often secondary to cholinergic hyperactivity, alpha-adrenergic hyperactivity, or rhinosinusitis. Given elderly patients’ increased susceptibility to the potential adverse central nervous system (CNS) and anticholinergic effects of antihistamines, non-sedating medications are recommended. Oral decongestants also should be used with caution in this population, not only because of CNS effects, but also because of heart and bladder effects³ (TABLE 2¹⁸).

For drug-induced rhinitis, stop the offending drug and consider an INGC

Several types of medications, both oral and inhaled, are known to cause rhinitis. The use of alpha-adrenergic decongestant sprays for more than 5 to 7 days can induce rebound congestion on withdrawal, known as rhinitis medicamentosa.³ Repeated use of intranasal cocaine and methamphetamines can also result in rebound congestion. Oral medications that can result in rhinitis or congestion include angiotensin-converting enzyme (ACE) inhibitors, beta-blockers, nonsteroidal anti-inflammatory drugs (NSAIDS), oral contraceptives, and even antidepressants.³

The treatment for drug-induced rhinitis is termination of the offending agent. INGCs can be used to help decrease inflammation and control symptoms once the offending agent is discontinued.

Mechanical/structural causes of obstruction are wide-ranging

Mechanical/structural causes of nasal obstruction range from foreign bodies to anatomical variations including nasal polyps, a deviated septum, adenoidal hypertrophy, foreign bodies, and tumors. Because more than one etiology may be at work, it is best to first treat any non-mechanical causes of obstruction, such as ARS or NARS.

Nasal polyposis often requires both a medical and surgical approach

Nasal polyps are benign growths arising from the mucosa of the nasal sinuses and nasal cavities and affecting up to 4% of the population.⁷ Their etiology is unclear, but we do know that nasal polyps result from underlying inflammation.⁷ Uncommon in children outside of those affected by cystic fibrosis,⁷ nasal polyposis can be associated with disease processes such as AR and sinusitis. Polyps are also associated with clinical syndromes such as aspirin-exacerbated respiratory disease (AERD) syndrome, which involves upper and lower respiratory tract symptoms in patients with asthma who have taken aspirin or other NSAIDs.⁹

Leukotriene receptor agonists may be better suited to nighttime symptoms of nasal obstruction, while oral antihistamines may be better suited to daytime symptoms.Symptoms vary with the location and size of the polyps, but generally include nasal congestion, alteration in smell, and rhinorrhea. The goals of treatment are to restore or improve nasal breathing and olfaction and prevent recurrence.⁸ This often requires both a medical and surgical approach.

Topical corticosteroids are effective at reducing both the size of polyps and associated symptoms (rhinorrhea, rhinitis).⁸ And research has shown that steroids reduce the need for both primary and repeat surgical polypectomies.⁴ Other treatments to consider prior to surgery (if no symptom reduction occurs with INGCs) include systemic (oral) corticosteroids, intra-polyp steroid injections, macrolide antibiotics, and nasal washes.^7,14

When symptoms of polyposis are refractory to medical management, functional endoscopic sinus surgery (FESS) is the surgical procedure of choice.³ In addition to refractory symptoms, indications for FESS include the need to correct anatomic deformities believed to be contributing to the persistence of disease and the need to debulk advanced nasal polyposis.³ The principal goal is to restore patency to the ostiomeatal unit.³

Several studies have reported a high success rate for FESS in improving the symptoms of CRS.^3,19-23 In a 1992 study, for example, 98% of patients reported improvement following surgery,¹⁹ and in a follow-up report approximately 6 years later, 98% of patients continued to report subjective improvement.²²

For septal etiologies, consider septoplasty

Deviation of the nasal septum is a common structural etiology for nasal obstruction arising primarily from congenital, genetic, or traumatic causes.²⁴ Turbulent airflow from the septal deviation often causes turbinate hypertrophy, which creates (or exacerbates) the obstructive symptoms from the septal deviation.²⁵

Septoplasty is the most common ear, nose, and throat operation in adults.²⁶ Reduction of nasal symptoms has been reported in up to 89% of patients who receive this surgery, according to one single-center, non-randomized trial.²⁷ Currently, at least one multicenter, randomized trial is underway that aims to develop evidence-based guidelines for septoplasty.²⁶

Septal perforation is another etiology that can present with nasal obstruction symptoms. Causes include traumatic perforation, inflammatory or collagen vascular diseases, infections, overuse of vasoconstrictive medications, and malignancy.^28,29 A careful inspection of the nasal septum is necessary to identify a perforation; this may require nasal endoscopy.

Although adenoidectomy is commonly performed to correct adenoid hypertrophy in children, current evidence regarding the efficacy of the procedure is inconclusive.Anterior, rather than posterior, perforations are more likely to cause symptoms of nasal obstruction. Posterior perforations rarely require treatment unless malignancy is suspected, in which case referral for biopsy is recommended. Anterior perforations are treated initially with avoidance of any causative agent if, for example, the problem is drug- or medication-induced, and then with humidification and emollients.^28,29

For anterior perforations, septal silicone buttons can be used for recalcitrant symptoms. However, observational studies indicate that for long-term symptom resolution, silicone buttons are effective in only about one-third of patients.²⁹

For patients with persistent symptoms despite the above measures, surgical repair with various flap techniques is an option. A meta-analysis of case studies involving various techniques concluded that there is a wide variety of options, and that surgeons must weigh factors such as the characteristics and etiology of the perforation and their own experience and expertise when choosing from among available methods.³⁰ Additional good quality research is necessary before clear recommendations regarding technique can be made.

Adenoid hypertrophy: Consider corticosteroid nasal drops

Adenoid hypertrophy is a common cause of chronic nasal obstruction in children. Although adenoidectomy is commonly performed to correct the problem, current evidence regarding the efficacy of the procedure is inconclusive.¹⁰ Evidence demonstrates corticosteroid nasal drops significantly reduce symptoms of nasal obstruction in children and may provide an effective alternative to surgical resection.¹⁸ Studies have also demonstrated that treatment with oral LRAs significantly reduces adenoid size and nasal obstruction symptoms.^12,13

Foreign bodies: Don’t forget “a mother’s kiss”

Foreign bodies are the most common cause of nasal obstruction in the pediatric population. There is a paucity of high-quality evidence on removal of these objects; however, a number of retrospective reviews and case series support that most objects can be removed in the office or emergency department without otolaryngologic referral.^31,32

Techniques for removal include positive pressure, which is best used for smooth or soft objects. Positive pressure techniques include having the patient blow their own nose or having a parent use a mouth-to-mouth–type blowing technique (ie, the “mother’s kiss” method).³² Refer patients to Otolaryngology if the obstruction involves:³¹

• objects not easily visualized by anterior rhinoscopy
• chronic or impacted objects
• button batteries or magnets
• penetrating or hooked objects
• any object that cannot be removed during an initial attempt.

Nasal tumors: More common in older men

Nasal tumors occur most often in the nasal cavity itself and are more common in men ≥60 years.³³ There is no notable racial predominance.³³ Other risk factors include human papillomavirus (HPV) infection, tobacco smoke, and occupational exposure to inhaled wood dust, glues, and adhesives.^34-37

Most foreign bodies can be removed in the office or emergency department without referral to Otolaryngology.Benign tumors occurring in the nasal cavity are a diverse group of disorders, including inverted papillomas, squamous papillomas, pyogenic granulomas, and other less common lesions, all of which typically present with nasal obstruction as a symptom. Many of these lesions cause local tissue destruction or have a high incidence of recurrence. These tumors are treated universally with nasoendoscopic resection.³⁸

Malignant nasal tumors are rare but serious causes of nasal obstruction, making up 3% of all head and neck cancers.³⁹ Most nasal cancers present when they are locally advanced and cause unilateral nasal obstruction, lacrimation, and epistaxis. These symptoms are typically refractory to initial medical management and present as CRS. This diagnosis should be suspected in certain patient groups, such as those who have been exposed to wood dust (eg, construction workers or those who work in wood mills).³⁶

Computed tomography is the gold standard imaging method for CRS; however, if nasal cancer is suspected, referral for biopsy and histopathologic examination is necessary for a final diagnosis.³⁹ Because of the nonspecific nature of their initial presentation, many nasal tumors are at an advanced stage and carry a poor prognosis by the time they are diagnosed.³⁹

CORRESPONDENCE
Margaret A. Bayard, MD, MPH, FAAFP, Naval Hospital Camp Pendleton, 200 Mercy Circle, Camp Pendleton, CA 92005; [email protected].

Nasal obstruction is one of the most common reasons that patients visit their primary care providers.^1,2 Often described by patients as nasal congestion or the inability to adequately breathe out of one or both nostrils during the day and/or night, nasal obstruction commonly interferes with a patient’s ability to eat, sleep, and function, thereby significantly impacting quality of life. Overlapping presentations can make discerning the exact cause of nasal obstruction difficult.

To improve diagnosis and treatment, we review here the evidence-based recommendations for the most common causes of nasal obstruction: rhinitis, rhinosinusitis (RS), drug-induced nasal obstruction, and mechanical/structural abnormalities (TABLE 1^3-14).

Rhinitis/rhinosinusitis: It all begins with inflammation

Sneezing, rhinorrhea, nasal congestion, and nasal itching are complaints that signal rhinitis, which affects 30 to 60 million people in the United States annually.³ Rhinitis can be allergic, non-allergic, infectious, hormonal, or occupational in nature. All forms of rhinitis share inflammation as the cause of the nasal obstruction. The most common form is allergic rhinitis (AR), which includes seasonal AR and perennial AR. Seasonal AR is typically caused by outdoor allergens and waxes and wanes with pollen seasons. Perennial AR is caused mostly by indoor allergens, such as dust mites, molds, cockroaches, and pet dander; it persists all or most of the year.⁶ Causes of non-allergic rhinitis (NAR) include environmental irritants such as cigarette smoke, perfume, and car exhaust; medications; and hormonal changes,⁶ but most causes of NAR are unknown.^3,6

While AR can begin at any age, most people develop symptoms in childhood or as young adults, whereas NAR tends to begin later in life. Nasal itching can help to distinguish AR from NAR. NAR symptoms tend to be perennial and include postnasal drainage. If symptoms persist longer than 12 weeks despite treatment, the condition becomes known as chronic rhinosinusitis (CRS).

Treatment of rhinitis: Tiered and often continuous

Treatment of AR and NAR is similar and multitiered beginning with the avoidance of irritants and/or allergens whenever possible, moving on to pharmacotherapy, and, at least for AR, ending with allergen immunotherapy. Treatment is often an ongoing process and typically requires continuous therapy as opposed to treatment on an as-needed basis.³ It is unnecessary to perform allergy testing before making a presumed diagnosis of NAR and starting treatment.⁶

Intranasal corticosteroids. Currently, intranasal glucocorticosteroids (INGCs) are the most effective monotherapy for AR and NAR and have few adverse effects when used at prescribed doses.^3,4 For mild to intermittent symptoms, begin with the maximum dosage of an INGC for the patient’s age and proceed with incremental reductions to identify the lowest effective dose.³ If INGCs alone are ineffective, studies have shown that the addition of an intranasal second-generation antihistamine can be of some benefit.^3,4 In fact, an INGC and an intranasal antihistamine—along with saline nasal irrigation—is recommended for both AR and NAR resistant to single therapy.^3,6,15 If intranasal antihistamines are not an option, oral therapy can be initiated.

Start with second-generation antihistamines and consider LRAs. For oral therapy, start with second-generation antihistamines (loratadine, cetirizine, fexofenadine). First-generation antihistamines (diphenhydramine, hydroxyzine, chlorpheniramine), although widely available at relatively low cost, can cause several significant adverse effects including sedation, impaired cognitive function, and agitation in children.^3,4 Because second-generation antihistamines have fewer adverse effects, they are recommended as first-line therapy when oral antihistamine therapy is desired, such as for nasal congestion, sneezing, and itchy, watery eyes.

When prescribing intranasal corticosteroids for allergic and nonallergic rhinitis, begin with the maximum dosage and then incrementally reduce the amount to identify the lowest effective dose.Of note: A 2014 meta-analysis found that a leukotriene receptor antagonist (LRA) (montelukast) had efficacy similar to oral antihistamines for symptom relief in AR, and that LRAs may be better suited to nighttime symptoms (difficulty falling asleep, nighttime awakenings, congestion on awakening), while antihistamines may provide better relief of daytime symptoms (pruritus, rhinorrhea, sneezing).¹⁶ Although further head-to-head, double-blind randomized controlled trials (RCTs) are needed to confirm the results and investigate possible gender differences in symptom response, consider an LRA for first-line therapy in patients with AR who have predominantly nighttime symptoms.

What about pregnant women and the elderly?

It is important to consider teratogenicity when selecting medications for pregnant patients, especially during the first trimester.³ Nasal cromolyn has the most reassuring safety profile in pregnancy. Cetirizine, chlorpheniramine, loratadine, diphenhydramine, and tripelennamine may be used in pregnancy. The US Food and Drug Administration considers them to have a low risk of fetal harm, based on human data, whereas it views many other antihistamines as probably safe, based on limited or no human data. Most INGCs are not expected to cause fetal harm, but limited human data are available. Avoid prescribing oral decongestants to women who are in the first trimester of pregnancy due to the risk of gastroschisis in newborns.¹⁷

Elderly patients represent another population for which adverse effects must be carefully considered. Allergies in individuals >65 years of age are uncommon. Rhinitis in this age group is often secondary to cholinergic hyperactivity, alpha-adrenergic hyperactivity, or rhinosinusitis. Given elderly patients’ increased susceptibility to the potential adverse central nervous system (CNS) and anticholinergic effects of antihistamines, non-sedating medications are recommended. Oral decongestants also should be used with caution in this population, not only because of CNS effects, but also because of heart and bladder effects³ (TABLE 2¹⁸).

For drug-induced rhinitis, stop the offending drug and consider an INGC

Several types of medications, both oral and inhaled, are known to cause rhinitis. The use of alpha-adrenergic decongestant sprays for more than 5 to 7 days can induce rebound congestion on withdrawal, known as rhinitis medicamentosa.³ Repeated use of intranasal cocaine and methamphetamines can also result in rebound congestion. Oral medications that can result in rhinitis or congestion include angiotensin-converting enzyme (ACE) inhibitors, beta-blockers, nonsteroidal anti-inflammatory drugs (NSAIDS), oral contraceptives, and even antidepressants.³

The treatment for drug-induced rhinitis is termination of the offending agent. INGCs can be used to help decrease inflammation and control symptoms once the offending agent is discontinued.

Mechanical/structural causes of obstruction are wide-ranging

Mechanical/structural causes of nasal obstruction range from foreign bodies to anatomical variations including nasal polyps, a deviated septum, adenoidal hypertrophy, foreign bodies, and tumors. Because more than one etiology may be at work, it is best to first treat any non-mechanical causes of obstruction, such as ARS or NARS.

Nasal polyposis often requires both a medical and surgical approach

Nasal polyps are benign growths arising from the mucosa of the nasal sinuses and nasal cavities and affecting up to 4% of the population.⁷ Their etiology is unclear, but we do know that nasal polyps result from underlying inflammation.⁷ Uncommon in children outside of those affected by cystic fibrosis,⁷ nasal polyposis can be associated with disease processes such as AR and sinusitis. Polyps are also associated with clinical syndromes such as aspirin-exacerbated respiratory disease (AERD) syndrome, which involves upper and lower respiratory tract symptoms in patients with asthma who have taken aspirin or other NSAIDs.⁹

Leukotriene receptor agonists may be better suited to nighttime symptoms of nasal obstruction, while oral antihistamines may be better suited to daytime symptoms.Symptoms vary with the location and size of the polyps, but generally include nasal congestion, alteration in smell, and rhinorrhea. The goals of treatment are to restore or improve nasal breathing and olfaction and prevent recurrence.⁸ This often requires both a medical and surgical approach.

Topical corticosteroids are effective at reducing both the size of polyps and associated symptoms (rhinorrhea, rhinitis).⁸ And research has shown that steroids reduce the need for both primary and repeat surgical polypectomies.⁴ Other treatments to consider prior to surgery (if no symptom reduction occurs with INGCs) include systemic (oral) corticosteroids, intra-polyp steroid injections, macrolide antibiotics, and nasal washes.^7,14

When symptoms of polyposis are refractory to medical management, functional endoscopic sinus surgery (FESS) is the surgical procedure of choice.³ In addition to refractory symptoms, indications for FESS include the need to correct anatomic deformities believed to be contributing to the persistence of disease and the need to debulk advanced nasal polyposis.³ The principal goal is to restore patency to the ostiomeatal unit.³

Several studies have reported a high success rate for FESS in improving the symptoms of CRS.^3,19-23 In a 1992 study, for example, 98% of patients reported improvement following surgery,¹⁹ and in a follow-up report approximately 6 years later, 98% of patients continued to report subjective improvement.²²

For septal etiologies, consider septoplasty

Deviation of the nasal septum is a common structural etiology for nasal obstruction arising primarily from congenital, genetic, or traumatic causes.²⁴ Turbulent airflow from the septal deviation often causes turbinate hypertrophy, which creates (or exacerbates) the obstructive symptoms from the septal deviation.²⁵

Septoplasty is the most common ear, nose, and throat operation in adults.²⁶ Reduction of nasal symptoms has been reported in up to 89% of patients who receive this surgery, according to one single-center, non-randomized trial.²⁷ Currently, at least one multicenter, randomized trial is underway that aims to develop evidence-based guidelines for septoplasty.²⁶

Septal perforation is another etiology that can present with nasal obstruction symptoms. Causes include traumatic perforation, inflammatory or collagen vascular diseases, infections, overuse of vasoconstrictive medications, and malignancy.^28,29 A careful inspection of the nasal septum is necessary to identify a perforation; this may require nasal endoscopy.

Although adenoidectomy is commonly performed to correct adenoid hypertrophy in children, current evidence regarding the efficacy of the procedure is inconclusive.Anterior, rather than posterior, perforations are more likely to cause symptoms of nasal obstruction. Posterior perforations rarely require treatment unless malignancy is suspected, in which case referral for biopsy is recommended. Anterior perforations are treated initially with avoidance of any causative agent if, for example, the problem is drug- or medication-induced, and then with humidification and emollients.^28,29

For anterior perforations, septal silicone buttons can be used for recalcitrant symptoms. However, observational studies indicate that for long-term symptom resolution, silicone buttons are effective in only about one-third of patients.²⁹

For patients with persistent symptoms despite the above measures, surgical repair with various flap techniques is an option. A meta-analysis of case studies involving various techniques concluded that there is a wide variety of options, and that surgeons must weigh factors such as the characteristics and etiology of the perforation and their own experience and expertise when choosing from among available methods.³⁰ Additional good quality research is necessary before clear recommendations regarding technique can be made.

Adenoid hypertrophy: Consider corticosteroid nasal drops

Adenoid hypertrophy is a common cause of chronic nasal obstruction in children. Although adenoidectomy is commonly performed to correct the problem, current evidence regarding the efficacy of the procedure is inconclusive.¹⁰ Evidence demonstrates corticosteroid nasal drops significantly reduce symptoms of nasal obstruction in children and may provide an effective alternative to surgical resection.¹⁸ Studies have also demonstrated that treatment with oral LRAs significantly reduces adenoid size and nasal obstruction symptoms.^12,13

Foreign bodies: Don’t forget “a mother’s kiss”

Foreign bodies are the most common cause of nasal obstruction in the pediatric population. There is a paucity of high-quality evidence on removal of these objects; however, a number of retrospective reviews and case series support that most objects can be removed in the office or emergency department without otolaryngologic referral.^31,32

Techniques for removal include positive pressure, which is best used for smooth or soft objects. Positive pressure techniques include having the patient blow their own nose or having a parent use a mouth-to-mouth–type blowing technique (ie, the “mother’s kiss” method).³² Refer patients to Otolaryngology if the obstruction involves:³¹

• objects not easily visualized by anterior rhinoscopy
• chronic or impacted objects
• button batteries or magnets
• penetrating or hooked objects
• any object that cannot be removed during an initial attempt.

Nasal tumors: More common in older men

Nasal tumors occur most often in the nasal cavity itself and are more common in men ≥60 years.³³ There is no notable racial predominance.³³ Other risk factors include human papillomavirus (HPV) infection, tobacco smoke, and occupational exposure to inhaled wood dust, glues, and adhesives.^34-37

Most foreign bodies can be removed in the office or emergency department without referral to Otolaryngology.Benign tumors occurring in the nasal cavity are a diverse group of disorders, including inverted papillomas, squamous papillomas, pyogenic granulomas, and other less common lesions, all of which typically present with nasal obstruction as a symptom. Many of these lesions cause local tissue destruction or have a high incidence of recurrence. These tumors are treated universally with nasoendoscopic resection.³⁸

Malignant nasal tumors are rare but serious causes of nasal obstruction, making up 3% of all head and neck cancers.³⁹ Most nasal cancers present when they are locally advanced and cause unilateral nasal obstruction, lacrimation, and epistaxis. These symptoms are typically refractory to initial medical management and present as CRS. This diagnosis should be suspected in certain patient groups, such as those who have been exposed to wood dust (eg, construction workers or those who work in wood mills).³⁶

Computed tomography is the gold standard imaging method for CRS; however, if nasal cancer is suspected, referral for biopsy and histopathologic examination is necessary for a final diagnosis.³⁹ Because of the nonspecific nature of their initial presentation, many nasal tumors are at an advanced stage and carry a poor prognosis by the time they are diagnosed.³⁹

CORRESPONDENCE
Margaret A. Bayard, MD, MPH, FAAFP, Naval Hospital Camp Pendleton, 200 Mercy Circle, Camp Pendleton, CA 92005; [email protected].

ABSTRACT

Purpose Pain management with opioids in primary care is challenging. The objective of this study was to identify the number of opioid-related tasks in our clinics and determine whether opioid-related tasks occur more often in a residency setting.

Methods This was a retrospective observational review of an electronic health record (EHR) system to evaluate tasks related to the use of opioids and other controlled substances. Tasks are created in the EHR when patients call the clinic; the task-box system is a means of communication within the EHR. The study setting was 2 university-based family medicine clinics. Clinic 1 has faculty and resident providers in an urban area. Clinic 2 has only faculty providers in a suburban area. We reviewed all tasks recorded in November 2010.

Results A total of 3193 patients were seen at the clinics. In addition, 1028 call-related tasks were created, 220 of which (21.4%) were opioid-related. More than half of the tasks were about chronic (ongoing) patient issues. More than one‑third of the tasks required follow-up phone calls. Multiple logistic regression analysis showed more opioid-related tasks in the residency setting (Clinic 1) compared with the nonresidency setting (Clinic 2), (23.1% vs 16.7%; P<.001). However, multiple logistic regression analysis did not show any correlations between opioid-related tasks and who addressed the tasks or the day tasks were created.

Conclusions Primary care physicians prescribe significant amounts of opioids. Due to the nature of opioid use and abuse, a well-planned protocol customized to the practice or institution is required to streamline this process and decrease the number of unnecessary phone calls and follow-ups.

Pain management with opioids in primary care is challenging,^1,2 and many physicians find it unsatisfying and burdensome.³ More than 60 million patient visits for chronic pain occur annually in the United States, consuming large amounts of time and resources.⁴ Contributing to the challenge is the need to ensure patient safety and satisfaction, as well as staff satisfaction with pain management.^5-8 Opioid-related death is a major cause of iatrogenic mortality in the United States:^9,10 From 1999 to 2006, fatal opioid-involved intoxications more than tripled from 4000 to 13,800.⁷

At issue for many providers, as well as patients and staff, is dissatisfaction with current systems in place for managing chronic non-cancer pain with opioids.^2,3,8,11 In developing this study, we decided to focus on the systems aspect of care with 2 primary outcome measures in mind. Specifically, we sought to identify the tasks related to managing opioids and other controlled substances in 2 primary care clinics in a university-based family medicine program and to determine what proportion of all routine tasks in these 2 clinics could be attributed to opioid-related issues. With our secondary outcome measures, we sought to compare the number of opioid-related tasks in the residency setting with those in a nonresidency setting, and to identify factors that might be associated with an increase in the number of opioid-related tasks.

METHODS

Setting and design

We conducted a retrospective observational pilot study reviewing our electronic health record (EHR) system (Allscripts TouchWorks) at 2 of our outpatient family medicine clinics at the University of Colorado. When patients call the clinics, or when patient-care-related concerns need to be addressed, an electronic task message is created and sent to the appropriate task box for staff or provider response. The task box system is how staff and providers communicate within the EHR. Each provider has a personal task box, and there are other task boxes in the system (eg, triage, medication refill) for urgent and non-urgent patient care issues.

Nearly a quarter of clinic tasks were opioid related.For example, when a patient calls to request a refill, a medical assistant (MA), care team assistant (CTA), or nurse will create a task for the medication refill box. If the task is urgent, it is marked with a red asterisk and a triage provider will address the task that same day. Non-urgent triage tasks will be addressed by the patient’s primary care provider within 2 to 3 days. Depending on the issue at hand, the task may or may not require phone calls to the patient, pharmacy, or insurance company.

Clinic 1, in urban Denver, has 13 physicians (many of them part-time clinical faculty), one nurse practitioner (NP), one physician assistant (PA), and 18 family medicine residents. Clinic 2, in a suburb of Denver, has 5 physicians (only one is part-time) and one nurse practitioner. Clinic 1 is divided into 3 pods, and each has the same number of attending physicians, residents, and MAs, and either a PA or NP.

We reviewed, one by one, all tasks created from November 1 to 30, 2010. One of the study’s investigators categorized each task according to the following descriptors: who created the task, who addressed the task, what day of the week the task was created, urgency of the task, whether the task required a follow-up phone call, and whether the task was related to opioid/controlled-substance issues. The task was categorized as acute if the issue was related to a condition that had been present for fewer than 3 weeks. Chronic tasks were created for conditions present for ≥3 weeks. At the time the study was completed, our EHR had no portal through which we could communicate with patients.

ANALYSIS

We conducted statistical analyses with the IBM SPSS, version 22.0 (SPSS, Inc, Chicago, Illinois). We used descriptive statistics to examine the frequency and percentage for all variables. We used a chi-squared (χ2) test to assess the differences between the 2 clinics, and used a binary multiple logistic regression model to determine possible factors related to opioid-related tasks. P values <.05 were considered statistically significant. The Colorado Multiple Institutional Review Board approved this study.

RESULTS

Clinics 1 and 2, respectively, saw 2007 and 1186 patients during the study period (TABLE 1). The additional 1028 tasks generated by phone calls were almost equally distributed among the 3 pods of Clinic 1 (290, 202, and 260) and Clinic 2 (276). For data analysis, we compared Clinic 1 with Clinic 2 and also compared the 3 pods of Clinic 1 individually with Clinic 2. Both approaches produced similar results.

Most tasks (54% for Clinic 1 and 99% for Clinic 2) were created by MAs and CTAs. At Clinic 1, tasks were also created by residents (17%), PA/NPs (8%), attending physicians (7%), and others/clinical nurses (14%). Tasks at Clinic 1 were addressed by attending physicians (49%), residents (25%), PA/NPs (25%), and others (1%). At Clinic 2, tasks were addressed by attending physicians (75%) and PA/NPs (25%). Approximately half of the tasks (51%) in both clinics were created during weekdays, compared with the day after weekends/holidays (28%), the day before weekends/holidays (17%), and during weekends/holidays (4%). Chronic patient issues, acute patient issues, and other issues accounted for 54%, 29%, and 17% of tasks, respectively. Follow-up phone calls to patients, pharmacies, or others occurred in 37% of tasks. Two hundred twenty tasks (21%) in the clinics combined were related to opioids and controlled substances.

Multiple logistic regression analysis of data from both clinics (TABLE 2) showed more opioid-related tasks in Clinic 1 compared with Clinic 2 (P<.001), and that these tasks were more often related to chronic issues than to acute issues (P<.001). Tasks created by MAs, CTAs, clinical nurses, and others were more likely to be opioid-related compared with the tasks created by attending physicians, residents, NPs, or a PA (25% vs 15%; P<.05). Compared with non-opioid-related tasks, opioid-related tasks required more follow-up phone calls (P<.001). Follow-up phone calls to pharmacies occurred more often with opioid-related tasks than with non-opioid tasks (11% vs 5%), while follow-up phone calls to patients occurred more often for non-opioid related tasks than opioid-related tasks (28% vs 18%). No correlations with task creation were found for who addressed the opioid-related task or the day the task was created.

DISCUSSION

This study demonstrated that our process of handling patient issues related to opioids accounts for a large proportion of all tasks. Dealing with tasks is time consuming, not only for attending physicians and residents but also for clinic nurses and staff. Almost a quarter of clinic tasks were opioid related. As has been shown in previous studies,^5-8 chronic pain management with opioids is an unsatisfying task for staff and care providers at our clinics. We also found that tasks created by non-providers were more likely to be opioid-related than were tasks created by providers. This is most likely due to the fact that non-providers cannot write prescriptions and they have to ask providers for further reviews.

In this study, the larger urban practice with residents had proportionately more opioid-related tasks than the smaller suburban practice. Despite their different locations, these 2 clinics have relatively similar patient populations with relatively similar insurance coverage (TABLE 3). One reason for the difference noted in opioid-related tasks could be the composition of the provider pools (ie, part-time vs full-time) at each clinic. About half of the providers at Clinic 1 were residents; no residents served at Clinic 2. The variable and part-time nature of a resident’s clinic schedule could have led to discrepancies in opioid management, possibly leading in turn to an increase in phone calls and tasks. However, this finding could also be due to patients’ preferences for seeing less ex­perienced providers for opioid management issues.^12,13

Khalid et al found that, compared with attending physicians, residents had more patients on chronic opioids who displayed concerning behaviors, including early refills and refills from multiple providers.¹³ The higher number of part-time providers at Clinic 1 in our study may have also caused insufficient continuity of care at that site. Nevertheless, this model of practice is used in many academic primary care institutions.⁴ Another possible reason for the difference could be a lack of resident training on current guidelines for managing opiates for chronic pain.^3,13,14 Again, this was a pilot study and we drew no solid conclusion about the reasons for differences between these 2 clinics.

It is obvious, however, that we spend a significant amount of time and resources dealing with chronic pain management. Our institution created an opioid/controlled-substance patient registry about 3 years ago. The data for 2014 showed that 22.8% and 18% of patients seen at least once at Clinic 1 and Clinic 2, respectively, were prescribed opioids/controlled substances (TABLE 3).

Possible solutions to reduce tasks related to opioid management. For both small and large practices, one way to reduce the number of tasks related to opioid management and, therefore, the time allocated to completing those tasks, would be to have a clear protocol to follow.^{3,4,8,11,14,15} The protocol may include the creation of an opioid/controlled-substance registry and the development and implementation of clinical decision support programs.

We also recommend the dissemination of tools for clinical management at the point of care. These can include a controlled-substance risk assessment tool for aberrant behaviors, a controlled-substance informed consent form, a functional and quality-of-life assessment, electronic clinical-note templates in the EHR, urine drug screening, and routine use of existing state pharmacy prescription drug monitoring programs. Also essential would be the provision of routine educational programs for clinicians regarding chronic pain management based on existing evidence and guidelines. (See “Opioids for chronic pain: The CDC’s 12 recommendations.”) It has been demonstrated that an EHR opioid dashboard or an EHR-based protocol improved adherence to guidelines for prescribing opiates.¹⁶

This study has several limitations. First, this was a small pilot study completed over a short period of time, although we believe the findings are likely representative of the prescribing practices in the 2 clinics we evaluated. Second, it was a retrospective study, which was appropriate for evaluating our questions. Third, we were unable to account for other factors that could potentially confound the results, including, but not limited to, the amount of time allocated to each task, and the total number of patients at each clinic who were on opioids for management of chronic pain during the study period. However, due to our recent addition of an opioid/controlled-substance patient registry, we were able to add information for the year 2014 (TABLE 3). Multi-center large scale studies are required to evaluate this further.

ACKNOWLEDGEMENTS
We thank Dr. Corey Lyon for his editorial assistance.

CORRESPONDENCE
Morteza Khodaee, MD, AFW Family Medicine Clinic, 3055 Roslyn Street, Denver, CO 80238; [email protected].

ABSTRACT

Purpose Pain management with opioids in primary care is challenging. The objective of this study was to identify the number of opioid-related tasks in our clinics and determine whether opioid-related tasks occur more often in a residency setting.

Methods This was a retrospective observational review of an electronic health record (EHR) system to evaluate tasks related to the use of opioids and other controlled substances. Tasks are created in the EHR when patients call the clinic; the task-box system is a means of communication within the EHR. The study setting was 2 university-based family medicine clinics. Clinic 1 has faculty and resident providers in an urban area. Clinic 2 has only faculty providers in a suburban area. We reviewed all tasks recorded in November 2010.

Results A total of 3193 patients were seen at the clinics. In addition, 1028 call-related tasks were created, 220 of which (21.4%) were opioid-related. More than half of the tasks were about chronic (ongoing) patient issues. More than one‑third of the tasks required follow-up phone calls. Multiple logistic regression analysis showed more opioid-related tasks in the residency setting (Clinic 1) compared with the nonresidency setting (Clinic 2), (23.1% vs 16.7%; P<.001). However, multiple logistic regression analysis did not show any correlations between opioid-related tasks and who addressed the tasks or the day tasks were created.

Conclusions Primary care physicians prescribe significant amounts of opioids. Due to the nature of opioid use and abuse, a well-planned protocol customized to the practice or institution is required to streamline this process and decrease the number of unnecessary phone calls and follow-ups.

Pain management with opioids in primary care is challenging,^1,2 and many physicians find it unsatisfying and burdensome.³ More than 60 million patient visits for chronic pain occur annually in the United States, consuming large amounts of time and resources.⁴ Contributing to the challenge is the need to ensure patient safety and satisfaction, as well as staff satisfaction with pain management.^5-8 Opioid-related death is a major cause of iatrogenic mortality in the United States:^9,10 From 1999 to 2006, fatal opioid-involved intoxications more than tripled from 4000 to 13,800.⁷

At issue for many providers, as well as patients and staff, is dissatisfaction with current systems in place for managing chronic non-cancer pain with opioids.^2,3,8,11 In developing this study, we decided to focus on the systems aspect of care with 2 primary outcome measures in mind. Specifically, we sought to identify the tasks related to managing opioids and other controlled substances in 2 primary care clinics in a university-based family medicine program and to determine what proportion of all routine tasks in these 2 clinics could be attributed to opioid-related issues. With our secondary outcome measures, we sought to compare the number of opioid-related tasks in the residency setting with those in a nonresidency setting, and to identify factors that might be associated with an increase in the number of opioid-related tasks.

METHODS

Setting and design

We conducted a retrospective observational pilot study reviewing our electronic health record (EHR) system (Allscripts TouchWorks) at 2 of our outpatient family medicine clinics at the University of Colorado. When patients call the clinics, or when patient-care-related concerns need to be addressed, an electronic task message is created and sent to the appropriate task box for staff or provider response. The task box system is how staff and providers communicate within the EHR. Each provider has a personal task box, and there are other task boxes in the system (eg, triage, medication refill) for urgent and non-urgent patient care issues.

Nearly a quarter of clinic tasks were opioid related.For example, when a patient calls to request a refill, a medical assistant (MA), care team assistant (CTA), or nurse will create a task for the medication refill box. If the task is urgent, it is marked with a red asterisk and a triage provider will address the task that same day. Non-urgent triage tasks will be addressed by the patient’s primary care provider within 2 to 3 days. Depending on the issue at hand, the task may or may not require phone calls to the patient, pharmacy, or insurance company.

Clinic 1, in urban Denver, has 13 physicians (many of them part-time clinical faculty), one nurse practitioner (NP), one physician assistant (PA), and 18 family medicine residents. Clinic 2, in a suburb of Denver, has 5 physicians (only one is part-time) and one nurse practitioner. Clinic 1 is divided into 3 pods, and each has the same number of attending physicians, residents, and MAs, and either a PA or NP.

We reviewed, one by one, all tasks created from November 1 to 30, 2010. One of the study’s investigators categorized each task according to the following descriptors: who created the task, who addressed the task, what day of the week the task was created, urgency of the task, whether the task required a follow-up phone call, and whether the task was related to opioid/controlled-substance issues. The task was categorized as acute if the issue was related to a condition that had been present for fewer than 3 weeks. Chronic tasks were created for conditions present for ≥3 weeks. At the time the study was completed, our EHR had no portal through which we could communicate with patients.

ANALYSIS

We conducted statistical analyses with the IBM SPSS, version 22.0 (SPSS, Inc, Chicago, Illinois). We used descriptive statistics to examine the frequency and percentage for all variables. We used a chi-squared (χ2) test to assess the differences between the 2 clinics, and used a binary multiple logistic regression model to determine possible factors related to opioid-related tasks. P values <.05 were considered statistically significant. The Colorado Multiple Institutional Review Board approved this study.

RESULTS

Clinics 1 and 2, respectively, saw 2007 and 1186 patients during the study period (TABLE 1). The additional 1028 tasks generated by phone calls were almost equally distributed among the 3 pods of Clinic 1 (290, 202, and 260) and Clinic 2 (276). For data analysis, we compared Clinic 1 with Clinic 2 and also compared the 3 pods of Clinic 1 individually with Clinic 2. Both approaches produced similar results.

Most tasks (54% for Clinic 1 and 99% for Clinic 2) were created by MAs and CTAs. At Clinic 1, tasks were also created by residents (17%), PA/NPs (8%), attending physicians (7%), and others/clinical nurses (14%). Tasks at Clinic 1 were addressed by attending physicians (49%), residents (25%), PA/NPs (25%), and others (1%). At Clinic 2, tasks were addressed by attending physicians (75%) and PA/NPs (25%). Approximately half of the tasks (51%) in both clinics were created during weekdays, compared with the day after weekends/holidays (28%), the day before weekends/holidays (17%), and during weekends/holidays (4%). Chronic patient issues, acute patient issues, and other issues accounted for 54%, 29%, and 17% of tasks, respectively. Follow-up phone calls to patients, pharmacies, or others occurred in 37% of tasks. Two hundred twenty tasks (21%) in the clinics combined were related to opioids and controlled substances.

Multiple logistic regression analysis of data from both clinics (TABLE 2) showed more opioid-related tasks in Clinic 1 compared with Clinic 2 (P<.001), and that these tasks were more often related to chronic issues than to acute issues (P<.001). Tasks created by MAs, CTAs, clinical nurses, and others were more likely to be opioid-related compared with the tasks created by attending physicians, residents, NPs, or a PA (25% vs 15%; P<.05). Compared with non-opioid-related tasks, opioid-related tasks required more follow-up phone calls (P<.001). Follow-up phone calls to pharmacies occurred more often with opioid-related tasks than with non-opioid tasks (11% vs 5%), while follow-up phone calls to patients occurred more often for non-opioid related tasks than opioid-related tasks (28% vs 18%). No correlations with task creation were found for who addressed the opioid-related task or the day the task was created.

DISCUSSION

This study demonstrated that our process of handling patient issues related to opioids accounts for a large proportion of all tasks. Dealing with tasks is time consuming, not only for attending physicians and residents but also for clinic nurses and staff. Almost a quarter of clinic tasks were opioid related. As has been shown in previous studies,^5-8 chronic pain management with opioids is an unsatisfying task for staff and care providers at our clinics. We also found that tasks created by non-providers were more likely to be opioid-related than were tasks created by providers. This is most likely due to the fact that non-providers cannot write prescriptions and they have to ask providers for further reviews.

In this study, the larger urban practice with residents had proportionately more opioid-related tasks than the smaller suburban practice. Despite their different locations, these 2 clinics have relatively similar patient populations with relatively similar insurance coverage (TABLE 3). One reason for the difference noted in opioid-related tasks could be the composition of the provider pools (ie, part-time vs full-time) at each clinic. About half of the providers at Clinic 1 were residents; no residents served at Clinic 2. The variable and part-time nature of a resident’s clinic schedule could have led to discrepancies in opioid management, possibly leading in turn to an increase in phone calls and tasks. However, this finding could also be due to patients’ preferences for seeing less ex­perienced providers for opioid management issues.^12,13

Khalid et al found that, compared with attending physicians, residents had more patients on chronic opioids who displayed concerning behaviors, including early refills and refills from multiple providers.¹³ The higher number of part-time providers at Clinic 1 in our study may have also caused insufficient continuity of care at that site. Nevertheless, this model of practice is used in many academic primary care institutions.⁴ Another possible reason for the difference could be a lack of resident training on current guidelines for managing opiates for chronic pain.^3,13,14 Again, this was a pilot study and we drew no solid conclusion about the reasons for differences between these 2 clinics.

It is obvious, however, that we spend a significant amount of time and resources dealing with chronic pain management. Our institution created an opioid/controlled-substance patient registry about 3 years ago. The data for 2014 showed that 22.8% and 18% of patients seen at least once at Clinic 1 and Clinic 2, respectively, were prescribed opioids/controlled substances (TABLE 3).

Possible solutions to reduce tasks related to opioid management. For both small and large practices, one way to reduce the number of tasks related to opioid management and, therefore, the time allocated to completing those tasks, would be to have a clear protocol to follow.^{3,4,8,11,14,15} The protocol may include the creation of an opioid/controlled-substance registry and the development and implementation of clinical decision support programs.

We also recommend the dissemination of tools for clinical management at the point of care. These can include a controlled-substance risk assessment tool for aberrant behaviors, a controlled-substance informed consent form, a functional and quality-of-life assessment, electronic clinical-note templates in the EHR, urine drug screening, and routine use of existing state pharmacy prescription drug monitoring programs. Also essential would be the provision of routine educational programs for clinicians regarding chronic pain management based on existing evidence and guidelines. (See “Opioids for chronic pain: The CDC’s 12 recommendations.”) It has been demonstrated that an EHR opioid dashboard or an EHR-based protocol improved adherence to guidelines for prescribing opiates.¹⁶

This study has several limitations. First, this was a small pilot study completed over a short period of time, although we believe the findings are likely representative of the prescribing practices in the 2 clinics we evaluated. Second, it was a retrospective study, which was appropriate for evaluating our questions. Third, we were unable to account for other factors that could potentially confound the results, including, but not limited to, the amount of time allocated to each task, and the total number of patients at each clinic who were on opioids for management of chronic pain during the study period. However, due to our recent addition of an opioid/controlled-substance patient registry, we were able to add information for the year 2014 (TABLE 3). Multi-center large scale studies are required to evaluate this further.

ACKNOWLEDGEMENTS
We thank Dr. Corey Lyon for his editorial assistance.

CORRESPONDENCE
Morteza Khodaee, MD, AFW Family Medicine Clinic, 3055 Roslyn Street, Denver, CO 80238; [email protected].

ABSTRACT

Purpose Pain management with opioids in primary care is challenging. The objective of this study was to identify the number of opioid-related tasks in our clinics and determine whether opioid-related tasks occur more often in a residency setting.

Methods This was a retrospective observational review of an electronic health record (EHR) system to evaluate tasks related to the use of opioids and other controlled substances. Tasks are created in the EHR when patients call the clinic; the task-box system is a means of communication within the EHR. The study setting was 2 university-based family medicine clinics. Clinic 1 has faculty and resident providers in an urban area. Clinic 2 has only faculty providers in a suburban area. We reviewed all tasks recorded in November 2010.

Results A total of 3193 patients were seen at the clinics. In addition, 1028 call-related tasks were created, 220 of which (21.4%) were opioid-related. More than half of the tasks were about chronic (ongoing) patient issues. More than one‑third of the tasks required follow-up phone calls. Multiple logistic regression analysis showed more opioid-related tasks in the residency setting (Clinic 1) compared with the nonresidency setting (Clinic 2), (23.1% vs 16.7%; P<.001). However, multiple logistic regression analysis did not show any correlations between opioid-related tasks and who addressed the tasks or the day tasks were created.

Conclusions Primary care physicians prescribe significant amounts of opioids. Due to the nature of opioid use and abuse, a well-planned protocol customized to the practice or institution is required to streamline this process and decrease the number of unnecessary phone calls and follow-ups.

Pain management with opioids in primary care is challenging,^1,2 and many physicians find it unsatisfying and burdensome.³ More than 60 million patient visits for chronic pain occur annually in the United States, consuming large amounts of time and resources.⁴ Contributing to the challenge is the need to ensure patient safety and satisfaction, as well as staff satisfaction with pain management.^5-8 Opioid-related death is a major cause of iatrogenic mortality in the United States:^9,10 From 1999 to 2006, fatal opioid-involved intoxications more than tripled from 4000 to 13,800.⁷

At issue for many providers, as well as patients and staff, is dissatisfaction with current systems in place for managing chronic non-cancer pain with opioids.^2,3,8,11 In developing this study, we decided to focus on the systems aspect of care with 2 primary outcome measures in mind. Specifically, we sought to identify the tasks related to managing opioids and other controlled substances in 2 primary care clinics in a university-based family medicine program and to determine what proportion of all routine tasks in these 2 clinics could be attributed to opioid-related issues. With our secondary outcome measures, we sought to compare the number of opioid-related tasks in the residency setting with those in a nonresidency setting, and to identify factors that might be associated with an increase in the number of opioid-related tasks.

METHODS

Setting and design

We conducted a retrospective observational pilot study reviewing our electronic health record (EHR) system (Allscripts TouchWorks) at 2 of our outpatient family medicine clinics at the University of Colorado. When patients call the clinics, or when patient-care-related concerns need to be addressed, an electronic task message is created and sent to the appropriate task box for staff or provider response. The task box system is how staff and providers communicate within the EHR. Each provider has a personal task box, and there are other task boxes in the system (eg, triage, medication refill) for urgent and non-urgent patient care issues.

Nearly a quarter of clinic tasks were opioid related.For example, when a patient calls to request a refill, a medical assistant (MA), care team assistant (CTA), or nurse will create a task for the medication refill box. If the task is urgent, it is marked with a red asterisk and a triage provider will address the task that same day. Non-urgent triage tasks will be addressed by the patient’s primary care provider within 2 to 3 days. Depending on the issue at hand, the task may or may not require phone calls to the patient, pharmacy, or insurance company.

Clinic 1, in urban Denver, has 13 physicians (many of them part-time clinical faculty), one nurse practitioner (NP), one physician assistant (PA), and 18 family medicine residents. Clinic 2, in a suburb of Denver, has 5 physicians (only one is part-time) and one nurse practitioner. Clinic 1 is divided into 3 pods, and each has the same number of attending physicians, residents, and MAs, and either a PA or NP.

We reviewed, one by one, all tasks created from November 1 to 30, 2010. One of the study’s investigators categorized each task according to the following descriptors: who created the task, who addressed the task, what day of the week the task was created, urgency of the task, whether the task required a follow-up phone call, and whether the task was related to opioid/controlled-substance issues. The task was categorized as acute if the issue was related to a condition that had been present for fewer than 3 weeks. Chronic tasks were created for conditions present for ≥3 weeks. At the time the study was completed, our EHR had no portal through which we could communicate with patients.

ANALYSIS

We conducted statistical analyses with the IBM SPSS, version 22.0 (SPSS, Inc, Chicago, Illinois). We used descriptive statistics to examine the frequency and percentage for all variables. We used a chi-squared (χ2) test to assess the differences between the 2 clinics, and used a binary multiple logistic regression model to determine possible factors related to opioid-related tasks. P values <.05 were considered statistically significant. The Colorado Multiple Institutional Review Board approved this study.

RESULTS

Clinics 1 and 2, respectively, saw 2007 and 1186 patients during the study period (TABLE 1). The additional 1028 tasks generated by phone calls were almost equally distributed among the 3 pods of Clinic 1 (290, 202, and 260) and Clinic 2 (276). For data analysis, we compared Clinic 1 with Clinic 2 and also compared the 3 pods of Clinic 1 individually with Clinic 2. Both approaches produced similar results.

Most tasks (54% for Clinic 1 and 99% for Clinic 2) were created by MAs and CTAs. At Clinic 1, tasks were also created by residents (17%), PA/NPs (8%), attending physicians (7%), and others/clinical nurses (14%). Tasks at Clinic 1 were addressed by attending physicians (49%), residents (25%), PA/NPs (25%), and others (1%). At Clinic 2, tasks were addressed by attending physicians (75%) and PA/NPs (25%). Approximately half of the tasks (51%) in both clinics were created during weekdays, compared with the day after weekends/holidays (28%), the day before weekends/holidays (17%), and during weekends/holidays (4%). Chronic patient issues, acute patient issues, and other issues accounted for 54%, 29%, and 17% of tasks, respectively. Follow-up phone calls to patients, pharmacies, or others occurred in 37% of tasks. Two hundred twenty tasks (21%) in the clinics combined were related to opioids and controlled substances.

Multiple logistic regression analysis of data from both clinics (TABLE 2) showed more opioid-related tasks in Clinic 1 compared with Clinic 2 (P<.001), and that these tasks were more often related to chronic issues than to acute issues (P<.001). Tasks created by MAs, CTAs, clinical nurses, and others were more likely to be opioid-related compared with the tasks created by attending physicians, residents, NPs, or a PA (25% vs 15%; P<.05). Compared with non-opioid-related tasks, opioid-related tasks required more follow-up phone calls (P<.001). Follow-up phone calls to pharmacies occurred more often with opioid-related tasks than with non-opioid tasks (11% vs 5%), while follow-up phone calls to patients occurred more often for non-opioid related tasks than opioid-related tasks (28% vs 18%). No correlations with task creation were found for who addressed the opioid-related task or the day the task was created.

DISCUSSION

This study demonstrated that our process of handling patient issues related to opioids accounts for a large proportion of all tasks. Dealing with tasks is time consuming, not only for attending physicians and residents but also for clinic nurses and staff. Almost a quarter of clinic tasks were opioid related. As has been shown in previous studies,^5-8 chronic pain management with opioids is an unsatisfying task for staff and care providers at our clinics. We also found that tasks created by non-providers were more likely to be opioid-related than were tasks created by providers. This is most likely due to the fact that non-providers cannot write prescriptions and they have to ask providers for further reviews.

In this study, the larger urban practice with residents had proportionately more opioid-related tasks than the smaller suburban practice. Despite their different locations, these 2 clinics have relatively similar patient populations with relatively similar insurance coverage (TABLE 3). One reason for the difference noted in opioid-related tasks could be the composition of the provider pools (ie, part-time vs full-time) at each clinic. About half of the providers at Clinic 1 were residents; no residents served at Clinic 2. The variable and part-time nature of a resident’s clinic schedule could have led to discrepancies in opioid management, possibly leading in turn to an increase in phone calls and tasks. However, this finding could also be due to patients’ preferences for seeing less ex­perienced providers for opioid management issues.^12,13

Khalid et al found that, compared with attending physicians, residents had more patients on chronic opioids who displayed concerning behaviors, including early refills and refills from multiple providers.¹³ The higher number of part-time providers at Clinic 1 in our study may have also caused insufficient continuity of care at that site. Nevertheless, this model of practice is used in many academic primary care institutions.⁴ Another possible reason for the difference could be a lack of resident training on current guidelines for managing opiates for chronic pain.^3,13,14 Again, this was a pilot study and we drew no solid conclusion about the reasons for differences between these 2 clinics.

It is obvious, however, that we spend a significant amount of time and resources dealing with chronic pain management. Our institution created an opioid/controlled-substance patient registry about 3 years ago. The data for 2014 showed that 22.8% and 18% of patients seen at least once at Clinic 1 and Clinic 2, respectively, were prescribed opioids/controlled substances (TABLE 3).

Possible solutions to reduce tasks related to opioid management. For both small and large practices, one way to reduce the number of tasks related to opioid management and, therefore, the time allocated to completing those tasks, would be to have a clear protocol to follow.^{3,4,8,11,14,15} The protocol may include the creation of an opioid/controlled-substance registry and the development and implementation of clinical decision support programs.

We also recommend the dissemination of tools for clinical management at the point of care. These can include a controlled-substance risk assessment tool for aberrant behaviors, a controlled-substance informed consent form, a functional and quality-of-life assessment, electronic clinical-note templates in the EHR, urine drug screening, and routine use of existing state pharmacy prescription drug monitoring programs. Also essential would be the provision of routine educational programs for clinicians regarding chronic pain management based on existing evidence and guidelines. (See “Opioids for chronic pain: The CDC’s 12 recommendations.”) It has been demonstrated that an EHR opioid dashboard or an EHR-based protocol improved adherence to guidelines for prescribing opiates.¹⁶

This study has several limitations. First, this was a small pilot study completed over a short period of time, although we believe the findings are likely representative of the prescribing practices in the 2 clinics we evaluated. Second, it was a retrospective study, which was appropriate for evaluating our questions. Third, we were unable to account for other factors that could potentially confound the results, including, but not limited to, the amount of time allocated to each task, and the total number of patients at each clinic who were on opioids for management of chronic pain during the study period. However, due to our recent addition of an opioid/controlled-substance patient registry, we were able to add information for the year 2014 (TABLE 3). Multi-center large scale studies are required to evaluate this further.

ACKNOWLEDGEMENTS
We thank Dr. Corey Lyon for his editorial assistance.

CORRESPONDENCE
Morteza Khodaee, MD, AFW Family Medicine Clinic, 3055 Roslyn Street, Denver, CO 80238; [email protected].

Over the past decade, androgen replacement prescriptions for men ≥40 years of age have increased 3-fold, according to one study.¹ While one could argue this trend represents greater attention to an underdiagnosed problem, the study of prescription claims for almost 11 million men found that a quarter of them did not have a testosterone level documented in the 12 months prior to receiving treatment.¹

At the same time, sales of testosterone products totaled about $2.4 billion dollars in 2013, a number projected to top $4 billion by 2017.² The increase in prescribing is thought to be due, at least in part, to direct-to-consumer marketing techniques encouraging patients to seek medical attention if they are experiencing non-specific symptoms, such as fatigue and lack of energy, because their “problem” could be due to low testosterone.

Testosterone begins to decrease after age 40

The Endocrine Society defines “androgen deficiency” as low serum testosterone (generally <280 ng/dL for healthy young men) along with signs and symptoms of hypogonadism, including decreased sexual function; loss of axillary and/or pubic hair; low bone mineral density; loss of motivation and/or concentration; poor mood or depression; decline in cognitive function; and loss of muscle strength and work capacity (TABLE 1).³

Primary vs secondary hypogonadism. Primary (or hypogonadotropic) hypogonadism results when the testes fail to produce adequate testosterone in the presence of normal serum luteinizing hormone (LH) and follicle stimulating hormone (FSH) levels. Secondary hypogonadism is pituitary or hypothalamic in origin. Patients with primary hypogonadism will have high LH and FSH levels, whereas patients with secondary hypogonadism will have low or normal LH and FSH levels.⁴ The Endocrine Society recommends checking LH and FSH levels in all patients with hypogonadism to differentiate the primary from the secondary type.³ Patients with late onset primary hypogonadism do not require any further evaluation. In young men, it is important to consider Klinefelter syndrome. This diagnosis can be determined with a karyotype. In patients with secondary hypogonadism, checking serum iron, prolactin, and other pituitary hormones, and getting a magnetic resonance imaging scan of the sella turcica may be indicated. This will rule out infiltrative diseases, such as hemochromatosis, prolactinoma, and hypothalamic or pituitary neoplasm.

Testosterone is present in the body in 3 forms: free testosterone, albumin-bound testosterone, and testosterone bound to sex hormone-binding globulin (SHBG). In young healthy men, only 1% to 2% of testosterone is free, about 40% is albumin-bound and readily dissociates to free testosterone, and the remainder is tightly bound to SHBG, which does not readily dissociate and is therefore biologically unavailable.⁵ The amount of SHBG increases with age, decreasing the amount of bioavailable testosterone.

Serum levels of testosterone remain approximately stable until about age 40. After age 40, total levels of testosterone decrease by 1% to 2% annually, and serum free testosterone levels decrease by 2% to 3% annually.⁶ Testing of free testosterone levels is recommended when a patient falls in the low normal range of total testosterone (see below).

Testosterone screening: How and for whom?

Do not measure testosterone levels while patients are taking glucocorticoids or opioids, or have an acute or subacute illness.The Endocrine Society, consistent with the American Urological Association and the European Association of Urology, recommends against screening the general population for testosterone deficiency, fearing overdiagnosis and treatment of asymptomatic men.^3,7,8

The Endocrine Society’s recommendation for targeted screening states that for men with chronic diseases (eg, diabetes mellitus, end-stage renal disease, and chronic obstructive lung disease), measurement of testosterone may be indicated by symptoms such as sexual dysfunction, unexplained weight loss, weakness, or mobility limitation. The recommendation also states that in men with other conditions (eg, pituitary mass, human immunodeficiency virus (HIV)-associated weight loss, low-trauma fracture, or treatment with medications that affect testosterone production), measurement of testosterone may be indicated, regardless of symptoms.³ The United States Preventive Services Task Force does not have any specific recommendations regarding screening for hypogonadism in men.

Start with total serum testosterone

Measuring total serum testosterone should be the initial test for suspected testosterone deficiency. Testosterone levels vary throughout the day, peaking in the morning. As a result, levels should generally be measured before 10 am.

Lab values to watch for. Again, the lower limit of the normal testosterone range in healthy young men is 280 to 300 ng/dL, but may vary depending on the laboratory or assay used.³ If the level is abnormal (<280 ng/dL), repeat the test at least a month later prior to initiating testosterone replacement.³ For men with values in the low normal range and clinical symptoms, obtain levels of free testosterone to confirm the diagnosis.

Patients with chronic diseases, such as obesity, diabetes mellitus, liver disease, nephrotic syndrome, or thyroid disease, are more likely to have an increase in SHBG. For these patients, check free testosterone levels in the setting of symptoms and a low-to-normal total testosterone level.⁹ If a patient has symptoms of hypogonadism and a total testosterone level in the low normal range, as well as a free testosterone level that is less than the lower limit of normal for a laboratory (typically around 50 ng/dL), it is reasonable to offer testosterone replacement.

Do not prescribe testosterone for men with symptoms associated with hypogonadism (eg, fatigue or decreased libido) who do not have a low serum testosterone level.Medications such as glucocorticoids and opioids can affect testosterone levels, as can acute or subacute illness.¹⁰ Therefore, do not measure testosterone levels while a patient is receiving these medications, and wait until a patient has recovered from being ill before doing any testing.

Temper your response with older men. Many men >65 years old may have testosterone levels below the normal range for healthy, young counterparts. This decline is of uncertain clinical significance; it remains unclear if lower levels in older men result in health problems. Some have suggested establishing age-adjusted normal values, and recommend not initiating testosterone replacement therapy in older men until serum levels are below 200 ng/dL, rather than 280 ng/dL, which is the generally accepted lower limit for younger populations.^3,11,12

Testosterone replacement works when indicated

When clinically indicated (ie, when a patient’s testosterone level is below 280 ng/dL and the patient is experiencing a variety of symptoms associated with hypogonadism), research has shown testosterone replacement therapy can improve sexual function, mood, and, in some cases, lean body mass and physical function.^11,13

Keep in mind that the Endocrine Society and most professional organizations strongly discourage testosterone replacement in eugonadal men.³ Because of suppression of the HPG axis, men who discontinue testosterone replacement will typically experience symptoms of hypogonadism. Consequently, testosterone replacement should NOT be given to men with symptoms associated with hypogonadism (eg, fatigue or decreased libido) who do not have a low serum testosterone level.³

Testosterone is available in various forms, including oral, parenteral, pellets, transdermal gels and solutions, and as a buccal system. (Testosterone formulations and dosing information are described in TABLE 2.²) Oral formulations are generally not recommended due to potential hepatotoxicity and adverse effects on lipids.² In addition, oral formulations have short half-lives, making it difficult to achieve and maintain normal testosterone levels.

Long-acting parenteral testosterone is effective but must be given as an intramuscular injection, usually at 2- to 4-week intervals. These preparations produce fluctuations in serum testosterone levels, with supranormal levels occurring soon after injection and subnormal levels occurring immediately prior to subsequent injections.¹⁴

Pellets that contain 75 mg of testosterone are implanted subcutaneously. The usual dose is 2 pellets (150 mg), but may be as high as 6 (450 mg). The dose can be titrated based on follow-up serum testosterone levels. The therapeutic effects of the pellets continue, on average, for 3 to 4 months, and up to as long as 6 months.

Transdermal testosterone preparations are the most commonly prescribed. These include gels, patches, and solutions. They are easy to use and achieve more stable serum levels that remain in a normal range with daily use.¹⁵

• Gels. Considerations when prescribing testosterone gel forms include the possibility of spread to female partners or children, leading to virilization and precocious puberty. The gel should be applied to the skin but not the genitals, and should be covered with clothing after drying for at least 5 to 10 minutes.
• Patches can be applied to the back, abdomen, or extremities. A skin rash occurs in about one-third of men who use testosterone patches and may lead to discontinuation.¹⁶
• Solutions are applied to each under­arm daily. The starting dose is 60 mg under each arm; the dose can be adjusted based on follow-up serum testosterone levels.
• Buccal testosterone is applied to the buccal mucosa every 12 hours. It achieves therapeutic levels without large fluctuations. The tablet softens and forms to the gum, but does not dissolve and needs to be removed after 12 hours. The most common adverse effects are mucosal irritation and taste alteration.

Contraindications

Contraindications to testosterone replacement include heart failure, hepatic dysfunction (cirrhosis), prostate cancer, and breast cancer. Current guidelines also recommend not giving testosterone to men with severe lower urinary tract symptoms (due to benign prostate hyperplasia) with an International Prostate Symptom Score (IPSS) score >19.³ And, as mentioned earlier, the Endocrine Society strongly discourages testosterone replacement in eugonadal men.

After prescribing, monitoring is required

Men receiving testosterone replacement should have their testosterone levels checked at 3, 6, and 12 months after initiation of therapy, and annually thereafter.³ Therapy should be adjusted to achieve testosterone levels in the mid-normal range. Additional laboratory monitoring should include a serum hematocrit at baseline, at 6 months, and then annually if hematocrit remains in the normal range. Such testing is required because testosterone stimulates production of red blood cells from the bone marrow, which can lead to polycythemia. Discontinue therapy or reduce the dosage if a patient’s hematocrit rises above 54%, as there is a risk of thrombosis, although, in general, these events appear to be rare.^3,8

Obtain a lipid panel, liver function tests. Lipid abnormalities—primarily a decrease in high-density lipoprotein (HDL) cholesterol—may occur with testosterone replacement. Obtain a lipid panel and liver function tests at baseline and then yearly during replacement therapy.

Keep an eye on PSA. Although testosterone replacement does not increase the risk of prostate cancer, the Endocrine Society still recommends obtaining a prostate specific antigen (PSA) level and performing a digital rectal exam in men 40 years of age and older prior to initiating testosterone therapy.

Do not prescribe testosterone replacement if the patient’s PSA level is >4 ng/mL (or >3 ng/mL in high-risk groups) or if there is a palpable nodule or significant prostatic hypertrophy. Repeat the PSA in 6 months and then annually as long as testosterone therapy is continued. Further evaluation for prostate cancer is warranted if the PSA increases more than 0.4 ng/dL/year.^3,17

Testosterone replacement raises issues of abuse and CV risk

On October 25, 2016, the US Food and Drug Administration (FDA) approved class-wide labeling changes for all prescription testosterone products, alerting prescribers to the agent’s abuse potential and the serious cardiac and mental health adverse outcomes that have been reported as a result of such abuse. In addition, the FDA is revising the Abuse and Dependence section to include new safety information regarding the risks associated with abuse of testosterone and other anabolic androgenic steroids.¹⁸

Prior to this announcement, the FDA had mandated in 2015 that product labels include information about a possible increased risk of myocardial infarction (MI) and stroke in people using testosterone. This warning was based on 2 published studies that showed increased cardiovascular risk.^19,20 However, a third larger study showed no increase in risk.²¹ All 3 of these studies were retrospective and had methodologic limitations, including differing baseline testosterone levels, insufficient documentation of baseline levels, and inadequate monitoring of response to therapy.

A recent statement by the American Association of Clinical Endocrinologists and the American College of Endocrinology in response to the older FDA warning cites the need for randomized controlled trials (RCTs) to elucidate whether an association exists between testosterone replacement and cardiovascular risk.²²

Of note, researchers have shown that androgen deprivation therapy (ADT) in patients with prostate cancer impacts cardiovascular risk factors (ie, it increases body fat and decreases lean body mass, increases total cholesterol, and increases insulin resistance and risk of diabetes). ADT may also be associated with increased cardiovascular mortality, although data are conflicting.²³

Investigators have shown that testosterone replacement positively affects certain risk factors for cardiovascular disease (CVD) including increasing lean muscle mass and improving laboratory values associated with the metabolic syndrome.²⁴ A large retrospective cohort study of male veterans with documented low total testosterone levels who received their medical care at the Veterans Health Administration (VHA) found that those who received testosterone replacement and achieved normal testosterone levels had lower all-cause, cardiovascular, and stroke mortality than controls.²¹ The men who did not achieve normal testosterone levels also had lower all-cause mortality (but significantly less than those with normalization of serum testosterone levels), but no change in stroke or cardiovascular mortality.

Since this study was retrospective, there were significant limitations, including unknown baseline characteristics of patients in each group. The CVD risks associated with testosterone therapy in middle-aged and older men should be discussed by physicians and their patients on an individual basis. Some experts believe that patients who have had an MI, revascularization, or a stroke within the past 6 months are not good candidates for replacement therapy.²⁵

About 20% to 40% of men with erectile dysfunction have low testosterone, although testosterone replacement does not always improve the condition.Until there are better data from prospective RCTs, it may be prudent to make sure that traditional CVD risk factors including smoking, hypertension, hyperlipidemia, and diabetes have been assessed and are appropriately managed in men prescribed testosterone replacement.

Testosterone helps with ED in certain cases

Testosterone deficiency is associated with sexual dysfunction in men, including decreased libido and erectile dysfunction (ED). About 20% to 40% of men with ED will have low testosterone, although replacement does not always improve the condition.²

Current guidelines do not recommend testosterone replacement to treat ED or sexual dysfunction in the absence of a low serum testosterone level and recommend evaluating for other causes of sexual problems in men.³ In one study, men who did not have documented hypogonadism received testosterone replacement therapy for sexual dysfunction including ED or ejaculator dysfunction. These patients saw no improvement in symptoms.²⁶

CORRESPONDENCE
J. Andrew Hoover, MD, Department of Family and Community Medicine, Lancaster General Hospital, 540 North Duke Street, Lancaster, PA 17604; [email protected].

Over the past decade, androgen replacement prescriptions for men ≥40 years of age have increased 3-fold, according to one study.¹ While one could argue this trend represents greater attention to an underdiagnosed problem, the study of prescription claims for almost 11 million men found that a quarter of them did not have a testosterone level documented in the 12 months prior to receiving treatment.¹

At the same time, sales of testosterone products totaled about $2.4 billion dollars in 2013, a number projected to top $4 billion by 2017.² The increase in prescribing is thought to be due, at least in part, to direct-to-consumer marketing techniques encouraging patients to seek medical attention if they are experiencing non-specific symptoms, such as fatigue and lack of energy, because their “problem” could be due to low testosterone.

Testosterone begins to decrease after age 40

The Endocrine Society defines “androgen deficiency” as low serum testosterone (generally <280 ng/dL for healthy young men) along with signs and symptoms of hypogonadism, including decreased sexual function; loss of axillary and/or pubic hair; low bone mineral density; loss of motivation and/or concentration; poor mood or depression; decline in cognitive function; and loss of muscle strength and work capacity (TABLE 1).³

Primary vs secondary hypogonadism. Primary (or hypogonadotropic) hypogonadism results when the testes fail to produce adequate testosterone in the presence of normal serum luteinizing hormone (LH) and follicle stimulating hormone (FSH) levels. Secondary hypogonadism is pituitary or hypothalamic in origin. Patients with primary hypogonadism will have high LH and FSH levels, whereas patients with secondary hypogonadism will have low or normal LH and FSH levels.⁴ The Endocrine Society recommends checking LH and FSH levels in all patients with hypogonadism to differentiate the primary from the secondary type.³ Patients with late onset primary hypogonadism do not require any further evaluation. In young men, it is important to consider Klinefelter syndrome. This diagnosis can be determined with a karyotype. In patients with secondary hypogonadism, checking serum iron, prolactin, and other pituitary hormones, and getting a magnetic resonance imaging scan of the sella turcica may be indicated. This will rule out infiltrative diseases, such as hemochromatosis, prolactinoma, and hypothalamic or pituitary neoplasm.

Testosterone is present in the body in 3 forms: free testosterone, albumin-bound testosterone, and testosterone bound to sex hormone-binding globulin (SHBG). In young healthy men, only 1% to 2% of testosterone is free, about 40% is albumin-bound and readily dissociates to free testosterone, and the remainder is tightly bound to SHBG, which does not readily dissociate and is therefore biologically unavailable.⁵ The amount of SHBG increases with age, decreasing the amount of bioavailable testosterone.

Serum levels of testosterone remain approximately stable until about age 40. After age 40, total levels of testosterone decrease by 1% to 2% annually, and serum free testosterone levels decrease by 2% to 3% annually.⁶ Testing of free testosterone levels is recommended when a patient falls in the low normal range of total testosterone (see below).

Testosterone screening: How and for whom?

Do not measure testosterone levels while patients are taking glucocorticoids or opioids, or have an acute or subacute illness.The Endocrine Society, consistent with the American Urological Association and the European Association of Urology, recommends against screening the general population for testosterone deficiency, fearing overdiagnosis and treatment of asymptomatic men.^3,7,8

The Endocrine Society’s recommendation for targeted screening states that for men with chronic diseases (eg, diabetes mellitus, end-stage renal disease, and chronic obstructive lung disease), measurement of testosterone may be indicated by symptoms such as sexual dysfunction, unexplained weight loss, weakness, or mobility limitation. The recommendation also states that in men with other conditions (eg, pituitary mass, human immunodeficiency virus (HIV)-associated weight loss, low-trauma fracture, or treatment with medications that affect testosterone production), measurement of testosterone may be indicated, regardless of symptoms.³ The United States Preventive Services Task Force does not have any specific recommendations regarding screening for hypogonadism in men.

Start with total serum testosterone

Measuring total serum testosterone should be the initial test for suspected testosterone deficiency. Testosterone levels vary throughout the day, peaking in the morning. As a result, levels should generally be measured before 10 am.

Lab values to watch for. Again, the lower limit of the normal testosterone range in healthy young men is 280 to 300 ng/dL, but may vary depending on the laboratory or assay used.³ If the level is abnormal (<280 ng/dL), repeat the test at least a month later prior to initiating testosterone replacement.³ For men with values in the low normal range and clinical symptoms, obtain levels of free testosterone to confirm the diagnosis.

Patients with chronic diseases, such as obesity, diabetes mellitus, liver disease, nephrotic syndrome, or thyroid disease, are more likely to have an increase in SHBG. For these patients, check free testosterone levels in the setting of symptoms and a low-to-normal total testosterone level.⁹ If a patient has symptoms of hypogonadism and a total testosterone level in the low normal range, as well as a free testosterone level that is less than the lower limit of normal for a laboratory (typically around 50 ng/dL), it is reasonable to offer testosterone replacement.

Do not prescribe testosterone for men with symptoms associated with hypogonadism (eg, fatigue or decreased libido) who do not have a low serum testosterone level.Medications such as glucocorticoids and opioids can affect testosterone levels, as can acute or subacute illness.¹⁰ Therefore, do not measure testosterone levels while a patient is receiving these medications, and wait until a patient has recovered from being ill before doing any testing.

Temper your response with older men. Many men >65 years old may have testosterone levels below the normal range for healthy, young counterparts. This decline is of uncertain clinical significance; it remains unclear if lower levels in older men result in health problems. Some have suggested establishing age-adjusted normal values, and recommend not initiating testosterone replacement therapy in older men until serum levels are below 200 ng/dL, rather than 280 ng/dL, which is the generally accepted lower limit for younger populations.^3,11,12

Testosterone replacement works when indicated

When clinically indicated (ie, when a patient’s testosterone level is below 280 ng/dL and the patient is experiencing a variety of symptoms associated with hypogonadism), research has shown testosterone replacement therapy can improve sexual function, mood, and, in some cases, lean body mass and physical function.^11,13

Keep in mind that the Endocrine Society and most professional organizations strongly discourage testosterone replacement in eugonadal men.³ Because of suppression of the HPG axis, men who discontinue testosterone replacement will typically experience symptoms of hypogonadism. Consequently, testosterone replacement should NOT be given to men with symptoms associated with hypogonadism (eg, fatigue or decreased libido) who do not have a low serum testosterone level.³

Testosterone is available in various forms, including oral, parenteral, pellets, transdermal gels and solutions, and as a buccal system. (Testosterone formulations and dosing information are described in TABLE 2.²) Oral formulations are generally not recommended due to potential hepatotoxicity and adverse effects on lipids.² In addition, oral formulations have short half-lives, making it difficult to achieve and maintain normal testosterone levels.

Long-acting parenteral testosterone is effective but must be given as an intramuscular injection, usually at 2- to 4-week intervals. These preparations produce fluctuations in serum testosterone levels, with supranormal levels occurring soon after injection and subnormal levels occurring immediately prior to subsequent injections.¹⁴

Pellets that contain 75 mg of testosterone are implanted subcutaneously. The usual dose is 2 pellets (150 mg), but may be as high as 6 (450 mg). The dose can be titrated based on follow-up serum testosterone levels. The therapeutic effects of the pellets continue, on average, for 3 to 4 months, and up to as long as 6 months.

Transdermal testosterone preparations are the most commonly prescribed. These include gels, patches, and solutions. They are easy to use and achieve more stable serum levels that remain in a normal range with daily use.¹⁵

• Gels. Considerations when prescribing testosterone gel forms include the possibility of spread to female partners or children, leading to virilization and precocious puberty. The gel should be applied to the skin but not the genitals, and should be covered with clothing after drying for at least 5 to 10 minutes.
• Patches can be applied to the back, abdomen, or extremities. A skin rash occurs in about one-third of men who use testosterone patches and may lead to discontinuation.¹⁶
• Solutions are applied to each under­arm daily. The starting dose is 60 mg under each arm; the dose can be adjusted based on follow-up serum testosterone levels.
• Buccal testosterone is applied to the buccal mucosa every 12 hours. It achieves therapeutic levels without large fluctuations. The tablet softens and forms to the gum, but does not dissolve and needs to be removed after 12 hours. The most common adverse effects are mucosal irritation and taste alteration.

Contraindications

Contraindications to testosterone replacement include heart failure, hepatic dysfunction (cirrhosis), prostate cancer, and breast cancer. Current guidelines also recommend not giving testosterone to men with severe lower urinary tract symptoms (due to benign prostate hyperplasia) with an International Prostate Symptom Score (IPSS) score >19.³ And, as mentioned earlier, the Endocrine Society strongly discourages testosterone replacement in eugonadal men.

After prescribing, monitoring is required

Men receiving testosterone replacement should have their testosterone levels checked at 3, 6, and 12 months after initiation of therapy, and annually thereafter.³ Therapy should be adjusted to achieve testosterone levels in the mid-normal range. Additional laboratory monitoring should include a serum hematocrit at baseline, at 6 months, and then annually if hematocrit remains in the normal range. Such testing is required because testosterone stimulates production of red blood cells from the bone marrow, which can lead to polycythemia. Discontinue therapy or reduce the dosage if a patient’s hematocrit rises above 54%, as there is a risk of thrombosis, although, in general, these events appear to be rare.^3,8

Obtain a lipid panel, liver function tests. Lipid abnormalities—primarily a decrease in high-density lipoprotein (HDL) cholesterol—may occur with testosterone replacement. Obtain a lipid panel and liver function tests at baseline and then yearly during replacement therapy.

Keep an eye on PSA. Although testosterone replacement does not increase the risk of prostate cancer, the Endocrine Society still recommends obtaining a prostate specific antigen (PSA) level and performing a digital rectal exam in men 40 years of age and older prior to initiating testosterone therapy.

Do not prescribe testosterone replacement if the patient’s PSA level is >4 ng/mL (or >3 ng/mL in high-risk groups) or if there is a palpable nodule or significant prostatic hypertrophy. Repeat the PSA in 6 months and then annually as long as testosterone therapy is continued. Further evaluation for prostate cancer is warranted if the PSA increases more than 0.4 ng/dL/year.^3,17

Testosterone replacement raises issues of abuse and CV risk

On October 25, 2016, the US Food and Drug Administration (FDA) approved class-wide labeling changes for all prescription testosterone products, alerting prescribers to the agent’s abuse potential and the serious cardiac and mental health adverse outcomes that have been reported as a result of such abuse. In addition, the FDA is revising the Abuse and Dependence section to include new safety information regarding the risks associated with abuse of testosterone and other anabolic androgenic steroids.¹⁸

Prior to this announcement, the FDA had mandated in 2015 that product labels include information about a possible increased risk of myocardial infarction (MI) and stroke in people using testosterone. This warning was based on 2 published studies that showed increased cardiovascular risk.^19,20 However, a third larger study showed no increase in risk.²¹ All 3 of these studies were retrospective and had methodologic limitations, including differing baseline testosterone levels, insufficient documentation of baseline levels, and inadequate monitoring of response to therapy.

A recent statement by the American Association of Clinical Endocrinologists and the American College of Endocrinology in response to the older FDA warning cites the need for randomized controlled trials (RCTs) to elucidate whether an association exists between testosterone replacement and cardiovascular risk.²²

Of note, researchers have shown that androgen deprivation therapy (ADT) in patients with prostate cancer impacts cardiovascular risk factors (ie, it increases body fat and decreases lean body mass, increases total cholesterol, and increases insulin resistance and risk of diabetes). ADT may also be associated with increased cardiovascular mortality, although data are conflicting.²³

Investigators have shown that testosterone replacement positively affects certain risk factors for cardiovascular disease (CVD) including increasing lean muscle mass and improving laboratory values associated with the metabolic syndrome.²⁴ A large retrospective cohort study of male veterans with documented low total testosterone levels who received their medical care at the Veterans Health Administration (VHA) found that those who received testosterone replacement and achieved normal testosterone levels had lower all-cause, cardiovascular, and stroke mortality than controls.²¹ The men who did not achieve normal testosterone levels also had lower all-cause mortality (but significantly less than those with normalization of serum testosterone levels), but no change in stroke or cardiovascular mortality.

Since this study was retrospective, there were significant limitations, including unknown baseline characteristics of patients in each group. The CVD risks associated with testosterone therapy in middle-aged and older men should be discussed by physicians and their patients on an individual basis. Some experts believe that patients who have had an MI, revascularization, or a stroke within the past 6 months are not good candidates for replacement therapy.²⁵

About 20% to 40% of men with erectile dysfunction have low testosterone, although testosterone replacement does not always improve the condition.Until there are better data from prospective RCTs, it may be prudent to make sure that traditional CVD risk factors including smoking, hypertension, hyperlipidemia, and diabetes have been assessed and are appropriately managed in men prescribed testosterone replacement.

Testosterone helps with ED in certain cases

Testosterone deficiency is associated with sexual dysfunction in men, including decreased libido and erectile dysfunction (ED). About 20% to 40% of men with ED will have low testosterone, although replacement does not always improve the condition.²

Current guidelines do not recommend testosterone replacement to treat ED or sexual dysfunction in the absence of a low serum testosterone level and recommend evaluating for other causes of sexual problems in men.³ In one study, men who did not have documented hypogonadism received testosterone replacement therapy for sexual dysfunction including ED or ejaculator dysfunction. These patients saw no improvement in symptoms.²⁶

CORRESPONDENCE
J. Andrew Hoover, MD, Department of Family and Community Medicine, Lancaster General Hospital, 540 North Duke Street, Lancaster, PA 17604; [email protected].

Over the past decade, androgen replacement prescriptions for men ≥40 years of age have increased 3-fold, according to one study.¹ While one could argue this trend represents greater attention to an underdiagnosed problem, the study of prescription claims for almost 11 million men found that a quarter of them did not have a testosterone level documented in the 12 months prior to receiving treatment.¹

At the same time, sales of testosterone products totaled about $2.4 billion dollars in 2013, a number projected to top $4 billion by 2017.² The increase in prescribing is thought to be due, at least in part, to direct-to-consumer marketing techniques encouraging patients to seek medical attention if they are experiencing non-specific symptoms, such as fatigue and lack of energy, because their “problem” could be due to low testosterone.

Testosterone begins to decrease after age 40

The Endocrine Society defines “androgen deficiency” as low serum testosterone (generally <280 ng/dL for healthy young men) along with signs and symptoms of hypogonadism, including decreased sexual function; loss of axillary and/or pubic hair; low bone mineral density; loss of motivation and/or concentration; poor mood or depression; decline in cognitive function; and loss of muscle strength and work capacity (TABLE 1).³

Primary vs secondary hypogonadism. Primary (or hypogonadotropic) hypogonadism results when the testes fail to produce adequate testosterone in the presence of normal serum luteinizing hormone (LH) and follicle stimulating hormone (FSH) levels. Secondary hypogonadism is pituitary or hypothalamic in origin. Patients with primary hypogonadism will have high LH and FSH levels, whereas patients with secondary hypogonadism will have low or normal LH and FSH levels.⁴ The Endocrine Society recommends checking LH and FSH levels in all patients with hypogonadism to differentiate the primary from the secondary type.³ Patients with late onset primary hypogonadism do not require any further evaluation. In young men, it is important to consider Klinefelter syndrome. This diagnosis can be determined with a karyotype. In patients with secondary hypogonadism, checking serum iron, prolactin, and other pituitary hormones, and getting a magnetic resonance imaging scan of the sella turcica may be indicated. This will rule out infiltrative diseases, such as hemochromatosis, prolactinoma, and hypothalamic or pituitary neoplasm.

Testosterone is present in the body in 3 forms: free testosterone, albumin-bound testosterone, and testosterone bound to sex hormone-binding globulin (SHBG). In young healthy men, only 1% to 2% of testosterone is free, about 40% is albumin-bound and readily dissociates to free testosterone, and the remainder is tightly bound to SHBG, which does not readily dissociate and is therefore biologically unavailable.⁵ The amount of SHBG increases with age, decreasing the amount of bioavailable testosterone.

Serum levels of testosterone remain approximately stable until about age 40. After age 40, total levels of testosterone decrease by 1% to 2% annually, and serum free testosterone levels decrease by 2% to 3% annually.⁶ Testing of free testosterone levels is recommended when a patient falls in the low normal range of total testosterone (see below).

Testosterone screening: How and for whom?

Do not measure testosterone levels while patients are taking glucocorticoids or opioids, or have an acute or subacute illness.The Endocrine Society, consistent with the American Urological Association and the European Association of Urology, recommends against screening the general population for testosterone deficiency, fearing overdiagnosis and treatment of asymptomatic men.^3,7,8

The Endocrine Society’s recommendation for targeted screening states that for men with chronic diseases (eg, diabetes mellitus, end-stage renal disease, and chronic obstructive lung disease), measurement of testosterone may be indicated by symptoms such as sexual dysfunction, unexplained weight loss, weakness, or mobility limitation. The recommendation also states that in men with other conditions (eg, pituitary mass, human immunodeficiency virus (HIV)-associated weight loss, low-trauma fracture, or treatment with medications that affect testosterone production), measurement of testosterone may be indicated, regardless of symptoms.³ The United States Preventive Services Task Force does not have any specific recommendations regarding screening for hypogonadism in men.

Start with total serum testosterone

Measuring total serum testosterone should be the initial test for suspected testosterone deficiency. Testosterone levels vary throughout the day, peaking in the morning. As a result, levels should generally be measured before 10 am.

Lab values to watch for. Again, the lower limit of the normal testosterone range in healthy young men is 280 to 300 ng/dL, but may vary depending on the laboratory or assay used.³ If the level is abnormal (<280 ng/dL), repeat the test at least a month later prior to initiating testosterone replacement.³ For men with values in the low normal range and clinical symptoms, obtain levels of free testosterone to confirm the diagnosis.

Patients with chronic diseases, such as obesity, diabetes mellitus, liver disease, nephrotic syndrome, or thyroid disease, are more likely to have an increase in SHBG. For these patients, check free testosterone levels in the setting of symptoms and a low-to-normal total testosterone level.⁹ If a patient has symptoms of hypogonadism and a total testosterone level in the low normal range, as well as a free testosterone level that is less than the lower limit of normal for a laboratory (typically around 50 ng/dL), it is reasonable to offer testosterone replacement.

Do not prescribe testosterone for men with symptoms associated with hypogonadism (eg, fatigue or decreased libido) who do not have a low serum testosterone level.Medications such as glucocorticoids and opioids can affect testosterone levels, as can acute or subacute illness.¹⁰ Therefore, do not measure testosterone levels while a patient is receiving these medications, and wait until a patient has recovered from being ill before doing any testing.

Temper your response with older men. Many men >65 years old may have testosterone levels below the normal range for healthy, young counterparts. This decline is of uncertain clinical significance; it remains unclear if lower levels in older men result in health problems. Some have suggested establishing age-adjusted normal values, and recommend not initiating testosterone replacement therapy in older men until serum levels are below 200 ng/dL, rather than 280 ng/dL, which is the generally accepted lower limit for younger populations.^3,11,12

Testosterone replacement works when indicated

When clinically indicated (ie, when a patient’s testosterone level is below 280 ng/dL and the patient is experiencing a variety of symptoms associated with hypogonadism), research has shown testosterone replacement therapy can improve sexual function, mood, and, in some cases, lean body mass and physical function.^11,13

Keep in mind that the Endocrine Society and most professional organizations strongly discourage testosterone replacement in eugonadal men.³ Because of suppression of the HPG axis, men who discontinue testosterone replacement will typically experience symptoms of hypogonadism. Consequently, testosterone replacement should NOT be given to men with symptoms associated with hypogonadism (eg, fatigue or decreased libido) who do not have a low serum testosterone level.³

Testosterone is available in various forms, including oral, parenteral, pellets, transdermal gels and solutions, and as a buccal system. (Testosterone formulations and dosing information are described in TABLE 2.²) Oral formulations are generally not recommended due to potential hepatotoxicity and adverse effects on lipids.² In addition, oral formulations have short half-lives, making it difficult to achieve and maintain normal testosterone levels.

Long-acting parenteral testosterone is effective but must be given as an intramuscular injection, usually at 2- to 4-week intervals. These preparations produce fluctuations in serum testosterone levels, with supranormal levels occurring soon after injection and subnormal levels occurring immediately prior to subsequent injections.¹⁴

Pellets that contain 75 mg of testosterone are implanted subcutaneously. The usual dose is 2 pellets (150 mg), but may be as high as 6 (450 mg). The dose can be titrated based on follow-up serum testosterone levels. The therapeutic effects of the pellets continue, on average, for 3 to 4 months, and up to as long as 6 months.

Transdermal testosterone preparations are the most commonly prescribed. These include gels, patches, and solutions. They are easy to use and achieve more stable serum levels that remain in a normal range with daily use.¹⁵

• Gels. Considerations when prescribing testosterone gel forms include the possibility of spread to female partners or children, leading to virilization and precocious puberty. The gel should be applied to the skin but not the genitals, and should be covered with clothing after drying for at least 5 to 10 minutes.
• Patches can be applied to the back, abdomen, or extremities. A skin rash occurs in about one-third of men who use testosterone patches and may lead to discontinuation.¹⁶
• Solutions are applied to each under­arm daily. The starting dose is 60 mg under each arm; the dose can be adjusted based on follow-up serum testosterone levels.
• Buccal testosterone is applied to the buccal mucosa every 12 hours. It achieves therapeutic levels without large fluctuations. The tablet softens and forms to the gum, but does not dissolve and needs to be removed after 12 hours. The most common adverse effects are mucosal irritation and taste alteration.

Contraindications

Contraindications to testosterone replacement include heart failure, hepatic dysfunction (cirrhosis), prostate cancer, and breast cancer. Current guidelines also recommend not giving testosterone to men with severe lower urinary tract symptoms (due to benign prostate hyperplasia) with an International Prostate Symptom Score (IPSS) score >19.³ And, as mentioned earlier, the Endocrine Society strongly discourages testosterone replacement in eugonadal men.

After prescribing, monitoring is required

Men receiving testosterone replacement should have their testosterone levels checked at 3, 6, and 12 months after initiation of therapy, and annually thereafter.³ Therapy should be adjusted to achieve testosterone levels in the mid-normal range. Additional laboratory monitoring should include a serum hematocrit at baseline, at 6 months, and then annually if hematocrit remains in the normal range. Such testing is required because testosterone stimulates production of red blood cells from the bone marrow, which can lead to polycythemia. Discontinue therapy or reduce the dosage if a patient’s hematocrit rises above 54%, as there is a risk of thrombosis, although, in general, these events appear to be rare.^3,8

Obtain a lipid panel, liver function tests. Lipid abnormalities—primarily a decrease in high-density lipoprotein (HDL) cholesterol—may occur with testosterone replacement. Obtain a lipid panel and liver function tests at baseline and then yearly during replacement therapy.

Keep an eye on PSA. Although testosterone replacement does not increase the risk of prostate cancer, the Endocrine Society still recommends obtaining a prostate specific antigen (PSA) level and performing a digital rectal exam in men 40 years of age and older prior to initiating testosterone therapy.

Do not prescribe testosterone replacement if the patient’s PSA level is >4 ng/mL (or >3 ng/mL in high-risk groups) or if there is a palpable nodule or significant prostatic hypertrophy. Repeat the PSA in 6 months and then annually as long as testosterone therapy is continued. Further evaluation for prostate cancer is warranted if the PSA increases more than 0.4 ng/dL/year.^3,17

Testosterone replacement raises issues of abuse and CV risk

On October 25, 2016, the US Food and Drug Administration (FDA) approved class-wide labeling changes for all prescription testosterone products, alerting prescribers to the agent’s abuse potential and the serious cardiac and mental health adverse outcomes that have been reported as a result of such abuse. In addition, the FDA is revising the Abuse and Dependence section to include new safety information regarding the risks associated with abuse of testosterone and other anabolic androgenic steroids.¹⁸

Prior to this announcement, the FDA had mandated in 2015 that product labels include information about a possible increased risk of myocardial infarction (MI) and stroke in people using testosterone. This warning was based on 2 published studies that showed increased cardiovascular risk.^19,20 However, a third larger study showed no increase in risk.²¹ All 3 of these studies were retrospective and had methodologic limitations, including differing baseline testosterone levels, insufficient documentation of baseline levels, and inadequate monitoring of response to therapy.

A recent statement by the American Association of Clinical Endocrinologists and the American College of Endocrinology in response to the older FDA warning cites the need for randomized controlled trials (RCTs) to elucidate whether an association exists between testosterone replacement and cardiovascular risk.²²

Of note, researchers have shown that androgen deprivation therapy (ADT) in patients with prostate cancer impacts cardiovascular risk factors (ie, it increases body fat and decreases lean body mass, increases total cholesterol, and increases insulin resistance and risk of diabetes). ADT may also be associated with increased cardiovascular mortality, although data are conflicting.²³

Investigators have shown that testosterone replacement positively affects certain risk factors for cardiovascular disease (CVD) including increasing lean muscle mass and improving laboratory values associated with the metabolic syndrome.²⁴ A large retrospective cohort study of male veterans with documented low total testosterone levels who received their medical care at the Veterans Health Administration (VHA) found that those who received testosterone replacement and achieved normal testosterone levels had lower all-cause, cardiovascular, and stroke mortality than controls.²¹ The men who did not achieve normal testosterone levels also had lower all-cause mortality (but significantly less than those with normalization of serum testosterone levels), but no change in stroke or cardiovascular mortality.

Since this study was retrospective, there were significant limitations, including unknown baseline characteristics of patients in each group. The CVD risks associated with testosterone therapy in middle-aged and older men should be discussed by physicians and their patients on an individual basis. Some experts believe that patients who have had an MI, revascularization, or a stroke within the past 6 months are not good candidates for replacement therapy.²⁵

About 20% to 40% of men with erectile dysfunction have low testosterone, although testosterone replacement does not always improve the condition.Until there are better data from prospective RCTs, it may be prudent to make sure that traditional CVD risk factors including smoking, hypertension, hyperlipidemia, and diabetes have been assessed and are appropriately managed in men prescribed testosterone replacement.

Testosterone helps with ED in certain cases

Testosterone deficiency is associated with sexual dysfunction in men, including decreased libido and erectile dysfunction (ED). About 20% to 40% of men with ED will have low testosterone, although replacement does not always improve the condition.²

Current guidelines do not recommend testosterone replacement to treat ED or sexual dysfunction in the absence of a low serum testosterone level and recommend evaluating for other causes of sexual problems in men.³ In one study, men who did not have documented hypogonadism received testosterone replacement therapy for sexual dysfunction including ED or ejaculator dysfunction. These patients saw no improvement in symptoms.²⁶

CORRESPONDENCE
J. Andrew Hoover, MD, Department of Family and Community Medicine, Lancaster General Hospital, 540 North Duke Street, Lancaster, PA 17604; [email protected].