Drug Therapy

Islam is the second most common religion in the world, and there are 1.6 billion Muslims, many in areas where diabetes is prevalent. Each year observant Muslims fast during the daylight hours for the holy month of Ramadan. It is estimated that 50 million diabetic people fast between dawn and sundown during Ramadan, and Muslims are not the only group of patients who fast for religious or other reasons. It is important for healthcare providers to guide patients with diabetes in avoiding problems related to prolonged fasting.

See related article

In this issue of the Cleveland Clinic Journal of Medicine, Drs. A.V. and Zagar address management of diabetes specifically relating to Ramadan fasting, with considerations that also apply to other diabetic patients who fast for religious or for medical reasons.

Fortunately, we now have antihyperglycemic agents that are unlikely to cause hypoglycemia if used alone or in combination, as long as the regimen does not include insulin or a sulfonylurea. These include:

• Metformin and thiazolidinediones (pioglitazone and rosig­litazone), which improve insulin sensitivity
• Glucagon-like peptide 1 (GLP-1) agonists (exenatide, liraglutide, dulaglutide, and albaglutide), which facilitate insulin release in a glucose-dependent fashion
• Dipeptidyl peptidase 4 inhibitors (sitagliptin, saxagliptin, alogliptin, and linagliptin), which augment endogenous incretin hormones, primarily GLP-1, and also facilitate insulin production in a glucose-dependent fashion
• Alpha glucosidase inhibitors (acarbose and miglitol), which slow carbohydrate absorption.

Introduced in recent years, the sodium-glucose cotransporter 2 (SGLT-2) inhibitors canagliflozin, dapagliflozin, and empagliflozin lower blood glucose by reducing the renal threshold for reabsorption of glucose, coupled with reabsorption of sodium leading to daily urinary excretion of about 200 calories. These agents alone or taken with any of the agents above should not cause hypoglycemia. However, they can lead to dehydration if fasting precludes the intake of water as well as food.

The primary concern during fasting is hypoglycemia when diabetes regimens involve insulin or insulin secretagogues, most commonly sulfonylureas. Long-acting basal insulin should not require adjustment during fasting if the dose is not excessive. The amount and timing of short-acting analogues administered before meals should be adjusted to the timing of meals, and doses should be adjusted proportionally to the anticipated carbohydrate intake. Premixed insulins such as intermediate-acting (protamine suspension) insulin and a short-acting insulin in 70/30, 75/25, or 50/50 ratios should be avoided. They do not lend themselves to changes in timing, and the short-acting component is fixed and cannot be changed for varied intake without changing the intermediate-acting portion, which functions as the basal insulin.

Sulfonylurea doses can be reduced or the larger dose moved to before the evening meal, but these agents still pose a risk of hypoglycemia during fasting hours. And as Drs. A.V. and Zagar state, glimepiride, glipizide, and gliclazide are the only agents in the class that should be considered; glyburide (ie, glibenclamide) poses too great a risk of hypoglycemia. On the other hand, the short-acting secretagogue nateglinide can be used safely before meals without much risk of hypoglycemia.

We have focused primarily on hypoglycemia risk. But if antihyperglycemic agents are halted completely or if the reduction is too severe, patients are at risk for hyperglycemia and even diabetic ketoacidosis. Careful monitoring of blood glucose levels during the fasting period is most important for patients taking agents that can cause hypoglycemia, and patients should be advised to break the fast if dangerously low glycemic levels occur. Similarly, if severe hyperglycemia or ketoacidosis develops, patients should be advised to seek medical advice promptly.

Islam is the second most common religion in the world, and there are 1.6 billion Muslims, many in areas where diabetes is prevalent. Each year observant Muslims fast during the daylight hours for the holy month of Ramadan. It is estimated that 50 million diabetic people fast between dawn and sundown during Ramadan, and Muslims are not the only group of patients who fast for religious or other reasons. It is important for healthcare providers to guide patients with diabetes in avoiding problems related to prolonged fasting.

See related article

In this issue of the Cleveland Clinic Journal of Medicine, Drs. A.V. and Zagar address management of diabetes specifically relating to Ramadan fasting, with considerations that also apply to other diabetic patients who fast for religious or for medical reasons.

Fortunately, we now have antihyperglycemic agents that are unlikely to cause hypoglycemia if used alone or in combination, as long as the regimen does not include insulin or a sulfonylurea. These include:

• Metformin and thiazolidinediones (pioglitazone and rosig­litazone), which improve insulin sensitivity
• Glucagon-like peptide 1 (GLP-1) agonists (exenatide, liraglutide, dulaglutide, and albaglutide), which facilitate insulin release in a glucose-dependent fashion
• Dipeptidyl peptidase 4 inhibitors (sitagliptin, saxagliptin, alogliptin, and linagliptin), which augment endogenous incretin hormones, primarily GLP-1, and also facilitate insulin production in a glucose-dependent fashion
• Alpha glucosidase inhibitors (acarbose and miglitol), which slow carbohydrate absorption.

Introduced in recent years, the sodium-glucose cotransporter 2 (SGLT-2) inhibitors canagliflozin, dapagliflozin, and empagliflozin lower blood glucose by reducing the renal threshold for reabsorption of glucose, coupled with reabsorption of sodium leading to daily urinary excretion of about 200 calories. These agents alone or taken with any of the agents above should not cause hypoglycemia. However, they can lead to dehydration if fasting precludes the intake of water as well as food.

The primary concern during fasting is hypoglycemia when diabetes regimens involve insulin or insulin secretagogues, most commonly sulfonylureas. Long-acting basal insulin should not require adjustment during fasting if the dose is not excessive. The amount and timing of short-acting analogues administered before meals should be adjusted to the timing of meals, and doses should be adjusted proportionally to the anticipated carbohydrate intake. Premixed insulins such as intermediate-acting (protamine suspension) insulin and a short-acting insulin in 70/30, 75/25, or 50/50 ratios should be avoided. They do not lend themselves to changes in timing, and the short-acting component is fixed and cannot be changed for varied intake without changing the intermediate-acting portion, which functions as the basal insulin.

Sulfonylurea doses can be reduced or the larger dose moved to before the evening meal, but these agents still pose a risk of hypoglycemia during fasting hours. And as Drs. A.V. and Zagar state, glimepiride, glipizide, and gliclazide are the only agents in the class that should be considered; glyburide (ie, glibenclamide) poses too great a risk of hypoglycemia. On the other hand, the short-acting secretagogue nateglinide can be used safely before meals without much risk of hypoglycemia.

We have focused primarily on hypoglycemia risk. But if antihyperglycemic agents are halted completely or if the reduction is too severe, patients are at risk for hyperglycemia and even diabetic ketoacidosis. Careful monitoring of blood glucose levels during the fasting period is most important for patients taking agents that can cause hypoglycemia, and patients should be advised to break the fast if dangerously low glycemic levels occur. Similarly, if severe hyperglycemia or ketoacidosis develops, patients should be advised to seek medical advice promptly.

Islam is the second most common religion in the world, and there are 1.6 billion Muslims, many in areas where diabetes is prevalent. Each year observant Muslims fast during the daylight hours for the holy month of Ramadan. It is estimated that 50 million diabetic people fast between dawn and sundown during Ramadan, and Muslims are not the only group of patients who fast for religious or other reasons. It is important for healthcare providers to guide patients with diabetes in avoiding problems related to prolonged fasting.

See related article

In this issue of the Cleveland Clinic Journal of Medicine, Drs. A.V. and Zagar address management of diabetes specifically relating to Ramadan fasting, with considerations that also apply to other diabetic patients who fast for religious or for medical reasons.

Fortunately, we now have antihyperglycemic agents that are unlikely to cause hypoglycemia if used alone or in combination, as long as the regimen does not include insulin or a sulfonylurea. These include:

• Metformin and thiazolidinediones (pioglitazone and rosig­litazone), which improve insulin sensitivity
• Glucagon-like peptide 1 (GLP-1) agonists (exenatide, liraglutide, dulaglutide, and albaglutide), which facilitate insulin release in a glucose-dependent fashion
• Dipeptidyl peptidase 4 inhibitors (sitagliptin, saxagliptin, alogliptin, and linagliptin), which augment endogenous incretin hormones, primarily GLP-1, and also facilitate insulin production in a glucose-dependent fashion
• Alpha glucosidase inhibitors (acarbose and miglitol), which slow carbohydrate absorption.

Introduced in recent years, the sodium-glucose cotransporter 2 (SGLT-2) inhibitors canagliflozin, dapagliflozin, and empagliflozin lower blood glucose by reducing the renal threshold for reabsorption of glucose, coupled with reabsorption of sodium leading to daily urinary excretion of about 200 calories. These agents alone or taken with any of the agents above should not cause hypoglycemia. However, they can lead to dehydration if fasting precludes the intake of water as well as food.

The primary concern during fasting is hypoglycemia when diabetes regimens involve insulin or insulin secretagogues, most commonly sulfonylureas. Long-acting basal insulin should not require adjustment during fasting if the dose is not excessive. The amount and timing of short-acting analogues administered before meals should be adjusted to the timing of meals, and doses should be adjusted proportionally to the anticipated carbohydrate intake. Premixed insulins such as intermediate-acting (protamine suspension) insulin and a short-acting insulin in 70/30, 75/25, or 50/50 ratios should be avoided. They do not lend themselves to changes in timing, and the short-acting component is fixed and cannot be changed for varied intake without changing the intermediate-acting portion, which functions as the basal insulin.

Sulfonylurea doses can be reduced or the larger dose moved to before the evening meal, but these agents still pose a risk of hypoglycemia during fasting hours. And as Drs. A.V. and Zagar state, glimepiride, glipizide, and gliclazide are the only agents in the class that should be considered; glyburide (ie, glibenclamide) poses too great a risk of hypoglycemia. On the other hand, the short-acting secretagogue nateglinide can be used safely before meals without much risk of hypoglycemia.

We have focused primarily on hypoglycemia risk. But if antihyperglycemic agents are halted completely or if the reduction is too severe, patients are at risk for hyperglycemia and even diabetic ketoacidosis. Careful monitoring of blood glucose levels during the fasting period is most important for patients taking agents that can cause hypoglycemia, and patients should be advised to break the fast if dangerously low glycemic levels occur. Similarly, if severe hyperglycemia or ketoacidosis develops, patients should be advised to seek medical advice promptly.

Contrary to common belief, pregnancy does not confer protection against depression.^1,2 In fact, pregnant women are just as likely as nonpregnant women to become or remain depressed, and up to 12.7% of pregnant women meet criteria for depression.¹

In the postpartum period, women are particularly vulnerable to a major depressive episode, whether a first episode or a recurrence. The estimated prevalence of a depressive episode in the first 3 postpartum  months is 19.2%,² making postpartum depression the most common complication of childbearing.² At the same time, peripartum depression remains largely underrecognized and undertreated.³

As evidence mounts regarding the deleterious impact of untreated mental illness on the mother, the developing fetus, and the infant, early detection and intervention for peripartum depression are paramount.³

DEPRESSION DURING PREGNANCY: SIGNIFICANT CONSEQUENCES

Although the rates of depression in pregnant and nonpregnant women are similar, depression in pregnancy carries additional significant consequences. Further, many depressed pregnant women believe their depression will lift once their baby is born, though it is well documented that depression during pregnancy is the strongest predictor of postpartum depression and that if left untreated it can be devastating for mother, infant, and family.⁴

Compared with nondepressed pregnant women, depressed pregnant women have poorer overall health status,⁵ are more likely to engage in behaviors that pose risk to the developing fetus such as smoking,⁵ alcohol consumption, and substance use,⁶ and have poor nutrition and inadequate weight gain.^7,8

Pregnant women who are depressed and are also experiencing domestic violence are especially at risk for poor prenatal care as they tend to miss more prenatal appointments.⁹ Evidence also suggests that depressed pregnant women are less attached to the fetus and more likely to have elective terminations.^10,11

Depression in pregnancy is associated with higher rates of adverse pregnancy outcomes such as preterm birth, low birth weight, operative delivery, and longer predelivery hospital stay.^3,12 Depression and anxiety during pregnancy have been associated with prenatal hypertension,¹³ gestational diabetes,¹⁴ preeclampsia,¹⁵ and HELLP syndrome (ie, hemolysis, elevated liver enzymes, and low platelet count).¹⁵ Depression and anxiety during pregnancy are associated with subsequent poorer infant attachment^16,17 and an overall unfavorable impact on infant and child development.¹⁸

Risk factors for depression during pregnancy include past episodes of depression, current anxiety, poor social support, unintended pregnancy, life stress, being single, domestic violence, and being on Medicaid.¹⁹

Undoubtedly the most devastating consequence of severe depression during pregnancy is suicide. Rates of suicide are lower in peripartum women,²⁰ but when suicide does occur, pregnant women tend to use more violent means than nonpregnant women. Pregnant adolescents represent a particularly high-risk group.²¹

POSTPARTUM DEPRESSION

Postpartum depression is the most common complication of childbearing. Although the precise pathogenesis is undetermined, there is converging evidence of a subset of women particularly sensitive to dramatic fluctuations in levels of estradiol and progesterone that occur during childbirth.^22,23 There is also evidence that dysregulation of the hypothalamic-pituitary-adrenal axis contributes to the development of postpartum depression in certain women.²⁴ Further, women who have depression or anxiety during pregnancy are much more likely to experience postpartum depression than those who are not symptomatic during pregnancy.⁴ A history of peripartum depression or other lifetime depressive episodes, poverty, conflict with a primary partner, poor social support, stressful life events, and low self-esteem are strongly associated with postpartum depression.²⁵

When unrecognized and untreated, postpartum depression can have profound and persistent effects on the mother and the developing infant.^18,26 Mothers with postpartum depression are much more likely than mothers without depression to have impaired bonding,²⁷ to be less responsive to their infant’s needs,¹⁷ and to be more likely to miss well-baby checkups.²⁸

Postpartum depression’s effects on maternal-infant interactions can include maternal withdrawal, disengagement, intrusion, and hostility and can lead to long-term effects on child development, including poor cognitive functioning, emotional maladjustment, and behavioral inhibition.^29,30 Infants and children of mothers with untreated postpartum depression have been shown to exhibit a higher incidence of colic, excessive crying, sleep problems, and irritability.^31,32 Women with postpartum depression may be less likely to initiate or maintain breastfeeding, and depressive symptoms have been noted to precede the discontinuation of breastfeeding.^33–35

Risk factors for postpartum depression

Characteristics to look for in the prenatal care of pregnant women include the following:

• Depression during pregnancy
• History of postpartum or other depressive episode
• Poverty
• Conflict with primary partner
• Poor social support
• Low self-esteem
• Single status.

Primary care providers are often the first point of contact for depressed women. The diagnosis of major depression in pregnant and postpartum women is challenging because of changes in sleep, appetite, and energy brought on by pregnancy, complications of delivery, and demands of caring for a newborn.³⁶ Many pregnant and postpartum women are reluctant to disclose their symptoms due to a sense of shame and guilt for being depressed during a time in their life that society commonly regards as joyful, and this contributes to under-detection.

In the first few days postpartum, fatigue, emotionality, irritability, and worry over the infant’s well-being affect up to 75% of women. This period, typically referred to as the “baby blues” or “postpartum blues,” is not considered a disorder and responds well to support, reassurance, and adequate sleep, and it typically resolves within 2 weeks.^37,38 Table 1 lists features that help distinguish postpartum blues from major depression.

Signs of major depressive disorder

Major depressive disorder is a serious and disabling condition. To meet criteria for major depressive disorder, women must report depressed mood and loss of interest or pleasure in normally pleasurable activities for at least 2 weeks. Completing the symptom profile, at least 5 of the following must be present: sleep disturbance (insomnia or hypersomnia), lack of energy, feelings of worthlessness or low self-esteem, guilt, difficulty concentrating, indecisiveness, psychomotor retardation or agitation, and thoughts of suicide or death.

The Diagnostic and Statistical Manual of Mental Disorders (5th edition) recognizes that postpartum depression commonly begins during pregnancy, and now uses “peripartum onset” as the specifier for major depressive disorder that occurs during pregnancy, postpartum, or both.³⁹ Other hallmark symptoms with peripartum onset include a lack of interest in or attachment to the pregnancy or infant, and anxiety and worry often accompanied by intrusive, unwanted thoughts of harm befalling the infant.⁴⁰

Postpartum psychosis

Postpartum psychosis is a far less common presentation, occurring in 1 to 2 per 1,000 births, but it constitutes a psychiatric emergency requiring immediate referral to a psychiatric care setting. Women at highest risk are those with a personal or family history of bipolar disorder.

The clinical presentation is most commonly characterized by confusion, agitation, hallucinations, delusional beliefs, and disorientation. Suicide and infanticide, while rare, are more likely to occur in the context of a psychotic episode.⁴¹

SCREENING RECOMMENDATIONS

Screening for depression is routine in primary care settings and is no less important for peripartum women.

In 2016, the US Preventive Services Task Force issued a recommendation that all pregnant and postpartum women be screened for depression,⁴² highlighting the need for all medical providers to be alert to the potentially serious consequences of unrecognized and untreated maternal psychiatric illness.

The American College of Obstetricians and Gynecologists (ACOG) recommends screening for depression and anxiety at least once during the peripartum period,⁴³ and the American Academy of Pediatrics recommends screening mothers for depression at the 1-, 2-, and 4-month well-baby visits.⁴⁴

The peripartum period is associated with changes in sleep, appetite, and energy levels, but these are also typical of depression. Taking this into account, the Edinburgh Postnatal Depression Scale (EPDS) was developed to screen for depression specifically in this population.⁴⁵ The EPDS is a validated and widely used 10-item self-reporting questionnaire with a high degree of sensitivity and specificity; it is easily administered and quickly scored. A cutoff score of 13 (of a maximum of 30) is considered indicative of depressed mood and signals the need for further assessment.

Source: Kroenke K, Spitzer RL, Williams JB. The PHQ-9: Validity of a brief depression severity measure. J Gen In-tern Med 2001; 16:606–616. No permission required to reproduce, translate, display, or distribute.

ACOG, the American Academy of Pediatrics, and the US Preventive Services Task Force recommend a standardized validated tool and cite both the EPDS (https://psychology-tools.com/epds/) and the Patient Health Questionnaire-9 (PHQ-9) (Figure 1) as appropriate to screen for peripartum depression.^42–44 Primary care providers tend to be most familiar with the PHQ-9, a highly sensitive and specific 9-item depression screen that has been validated in primary care and obstetric clinic patients.⁴⁶ A score on the PHQ-9 ranging from 5 to 10 indicates mild depression, 10 to 14 moderate depression, 15 to 19 moderate to severe depression, and greater than 19 severe depression.

Many women prefer nondrug therapy

The gold standard treatment for moderate to severe major depressive disorder is psychotherapy plus pharmacotherapy. Yet many peripartum women voice concerns about exposure to pharmacologic treatment, and studies have shown that many women prefer nonpharmacologic intervention.⁴⁷

Evidence-based psychotherapies that have demonstrated efficacy in peripartum women include cognitive behavioral therapy⁴⁸ and interpersonal psychotherapy when administered by a psychotherapist trained in these treatments. Pregnant and breastfeeding women often express preference for psychotherapy and complementary and alternative treatments as a means of avoiding fetal and infant exposure to antidepressants.⁴⁷

For mild to moderate depression, complementary therapies such as exercise, yoga, bright light therapy, and acupuncture have shown efficacy and can be used alone or adjunctively.⁴⁹ Because a poor marital relationship is consistently associated with peripartum depression,²⁵ primary care physicians who routinely address social support and screen for family conflict are well positioned to detect this significant correlate and to recommend marital or family therapy as a primary or adjunctive treatment.

When to consider drug therapy

The decision to recommend drug therapy must be individualized and based on the severity of symptoms, functional impairment, number and frequency of depressive episodes, history of response to medications, and the preferences of the patient, with the recognition that no decision is risk-free and that antidepressants enter the amniotic fluid, so fetal exposure is unavoidable.

Table 2 lists common antidepressants. The antidepressants most commonly prescribed, especially in the primary care setting, are selective serotonin reuptake inhibitors (SSRIs), which are favored because of their effectiveness, low side-effect profile, and lack of overdose toxicity.

Serotonin syndrome is no more likely to occur in pregnant than in nonpregnant women. Close monitoring for this condition is warranted only when patients are taking very  high doses of SSRIs or SSRIs in combination with other serotonergic agonists.

Prescribing antidepressants for pregnant or breastfeeding women requires thoughtful consideration of the patient’s preferences, as well as weighing the risks and benefits of fetal and infant exposure to maternal depression vs exposure to medications. Additional considerations include monotherapy, avoiding medication changes, choosing drugs that have been effective in the past, and avoiding drugs with known drug-drug interactions or teratogenic effects.⁵⁰

There is increasing consensus that the short- and long-term consequences of undertreatment or nontreatment of maternal depression outweigh the risk of fetal exposure to SSRIs.^3,51,52 Cohen et al⁵³ have recommended that if a woman is on an antidepressant and learns she is pregnant, she should not discontinue it because of the likelihood of relapse; they found a 68% relapse rate in women who discontinued their antidepressant in the first trimester of pregnancy.⁵³

In a comprehensive review of studies published between 1996 and 2012 that examined antidepressant use during pregnancy, Byatt et al⁵⁴ found little or no evidence of increased teratogenic risk with antidepressants with the exception of paroxetine, which is associated with a small but significant increased risk of cardiac malformation during first-trimester exposure.⁵⁴

These conclusions were underscored in a large cohort study in the United Kingdom.⁵⁵ In addition, a joint task force of the American Psychiatric Association and ACOG reviewed studies looking at the association between depression, antidepressants, and birth outcomes including miscarriage, preterm birth, cardiac abnormalities (resulting from first trimester exposure), persistent pulmonary hypertension (related to second- and third-trimester exposure), and neonatal adaptation syndrome (associated with third-trimester exposure).⁸ They concluded that the available data neither support nor refute a link between the use of antidepressants and several of the above outcomes. No increase in risk of congenital malformations (including cardiac abnormalities) was found. An increased risk of persistent pulmonary hypertension was noted, although the absolute risk of this disorder remained low, at 3 to 6 per 1,000 infants exposed to SSRIs in utero.^8,56

Neonatal adaptation syndrome

Neonatal adaptation syndrome is characterized by jitteriness, irritability, decreased muscle tone, and feeding difficulty in the neonate. It can occur in 15% to 30% of infants exposed to SSRIs antenatally.^57,58 These symptoms, however, are transient and typically resolve within 7 to 10 days after birth. A more recent study suggested that neurobehavioral symptoms for some infants extend beyond 2 weeks and that concomitant exposure to benzodiazepines results in even higher rates of this syndrome.⁵⁹ There is no evidence that tapering or discontinuing antidepressants near term is necessary, safe, or effective in preventing transient neonatal complications. However, this approach would increase the risk of relapse for the mother.

Autism spectrum disorders

The possible association between antidepressants and autism spectrum disorders in pregnancy has captured much attention in recent years. One study based on healthcare claims⁶⁰ and one registry-based study⁶¹ associated in utero exposure to antidepressants with autism liability in children. However, a large-scale Danish registry-based study did not replicate this association.⁶² In addition, 2 recent cohort studies, identifying children with autism spectrum disorder or attention-deficit hyperactivity disorder from electronic health records, found that neither disorder was significantly associated with prenatal antidepressant exposure in crude or adjusted models. However, both studies found a significant association with the use of antidepressants before pregnancy, indicating that the risk of autism observed with prenatal antidepressant exposure is likely confounded by the severity of maternal illness.^63,64

Concerns about drug therapy during breastfeeding

For infants of breastfeeding women, exposure to antidepressants through breast milk is minimal. Amounts in breast milk depend on the timing of the antidepressant dose, timing of feeding, and genetically influenced metabolic activity in mother and infant. The current literature supports antidepressant use for breastfeeding mothers of healthy full-term infants.⁶⁵

The 2 most widely studied antidepressants in breastfed infants are paroxetine and sertraline. It has been shown that very little can be detected in the infant’s serum, with relative infant doses ranging from 0.4% to 2.8%.⁶⁵ While clinicians are cautioned against prescribing paroxetine for pregnant women, the drug remains a suitable alternative for breastfeeding women.

If an antidepressant is started postpartum, the recommendation is to start with a low dose and then slowly titrate upward while monitoring the infant for adverse effects.^65,66 Possible adverse effects in breastfeeding infants include irritability, sedation, poor weight gain, and a change in feeding patterns.⁶⁷ Adverse events are most likely to occur in newborns up to 8 weeks of age, and infants born prematurely or with medical problems may be particularly at risk.^65,68

Helping patients weigh risks and benefits of drug therapy

Women may hear about the risks of medications to the fetus and during breastfeeding and so may be reluctant to seek or accept intervention. Often, the information is not from a reliable, scientifically based source. Primary care physicians are well positioned to guide peripartum women in risk-benefit analysis of proper treatment of their depression vs no treatment or undertreatment. In addition, establishing referral sources—ideally with a peripartum mental health specialist—is advisable. Online resources that clinicians can refer patients to for help in managing peripartum depression include the following:

• www.postpartum.net
• www.womensmentalhealth.org
• www.mothertobaby.org (for pharmacologic guidance).

INCREASED AWARENESS IS KEY

Primary care physicians must remain alert to the high prevalence of depression in women of childbearing age and embrace routine screening for depression. (See the sidebar, “The primary care management of peripartum depression.”) Since half of pregnancies are unintended, awareness of the risks of undetected and untreated peripartum depression to the mother, developing fetus, and infant is essential. Untreated antepartum depression has been linked to poor pregnancy outcomes, nutritional deficits, and substance abuse. Untreated postpartum depression negatively affects mother-infant attachment, infant, and child development and maternal self care.

Not treating depression is hazardous

Drug treatment during pregnancy and breastfeeding poses challenges for the patient and physician due to the inevitability of fetal and infant exposure, but lack of treatment can be hazardous.

To date, the evidence on the use of antidepressants in pregnant and lactating women is reassuring. Specialized peripartum psychiatric partial hospital programs⁶⁹ and inpatient programs⁷⁰ exist for women who need a higher level of care. There is also substantial evidence that psychotherapy, especially cognitive behavioral therapy and interpersonal therapy, is highly effective, and emerging data on complementary and alternative treatments are promising. Coordinated care between primary care and behavioral healthcare providers with expertise in treating peripartum depression is most likely to yield optimal outcomes.

Contrary to common belief, pregnancy does not confer protection against depression.^1,2 In fact, pregnant women are just as likely as nonpregnant women to become or remain depressed, and up to 12.7% of pregnant women meet criteria for depression.¹

In the postpartum period, women are particularly vulnerable to a major depressive episode, whether a first episode or a recurrence. The estimated prevalence of a depressive episode in the first 3 postpartum  months is 19.2%,² making postpartum depression the most common complication of childbearing.² At the same time, peripartum depression remains largely underrecognized and undertreated.³

As evidence mounts regarding the deleterious impact of untreated mental illness on the mother, the developing fetus, and the infant, early detection and intervention for peripartum depression are paramount.³

DEPRESSION DURING PREGNANCY: SIGNIFICANT CONSEQUENCES

Although the rates of depression in pregnant and nonpregnant women are similar, depression in pregnancy carries additional significant consequences. Further, many depressed pregnant women believe their depression will lift once their baby is born, though it is well documented that depression during pregnancy is the strongest predictor of postpartum depression and that if left untreated it can be devastating for mother, infant, and family.⁴

Compared with nondepressed pregnant women, depressed pregnant women have poorer overall health status,⁵ are more likely to engage in behaviors that pose risk to the developing fetus such as smoking,⁵ alcohol consumption, and substance use,⁶ and have poor nutrition and inadequate weight gain.^7,8

Pregnant women who are depressed and are also experiencing domestic violence are especially at risk for poor prenatal care as they tend to miss more prenatal appointments.⁹ Evidence also suggests that depressed pregnant women are less attached to the fetus and more likely to have elective terminations.^10,11

Depression in pregnancy is associated with higher rates of adverse pregnancy outcomes such as preterm birth, low birth weight, operative delivery, and longer predelivery hospital stay.^3,12 Depression and anxiety during pregnancy have been associated with prenatal hypertension,¹³ gestational diabetes,¹⁴ preeclampsia,¹⁵ and HELLP syndrome (ie, hemolysis, elevated liver enzymes, and low platelet count).¹⁵ Depression and anxiety during pregnancy are associated with subsequent poorer infant attachment^16,17 and an overall unfavorable impact on infant and child development.¹⁸

Risk factors for depression during pregnancy include past episodes of depression, current anxiety, poor social support, unintended pregnancy, life stress, being single, domestic violence, and being on Medicaid.¹⁹

Undoubtedly the most devastating consequence of severe depression during pregnancy is suicide. Rates of suicide are lower in peripartum women,²⁰ but when suicide does occur, pregnant women tend to use more violent means than nonpregnant women. Pregnant adolescents represent a particularly high-risk group.²¹

POSTPARTUM DEPRESSION

Postpartum depression is the most common complication of childbearing. Although the precise pathogenesis is undetermined, there is converging evidence of a subset of women particularly sensitive to dramatic fluctuations in levels of estradiol and progesterone that occur during childbirth.^22,23 There is also evidence that dysregulation of the hypothalamic-pituitary-adrenal axis contributes to the development of postpartum depression in certain women.²⁴ Further, women who have depression or anxiety during pregnancy are much more likely to experience postpartum depression than those who are not symptomatic during pregnancy.⁴ A history of peripartum depression or other lifetime depressive episodes, poverty, conflict with a primary partner, poor social support, stressful life events, and low self-esteem are strongly associated with postpartum depression.²⁵

When unrecognized and untreated, postpartum depression can have profound and persistent effects on the mother and the developing infant.^18,26 Mothers with postpartum depression are much more likely than mothers without depression to have impaired bonding,²⁷ to be less responsive to their infant’s needs,¹⁷ and to be more likely to miss well-baby checkups.²⁸

Postpartum depression’s effects on maternal-infant interactions can include maternal withdrawal, disengagement, intrusion, and hostility and can lead to long-term effects on child development, including poor cognitive functioning, emotional maladjustment, and behavioral inhibition.^29,30 Infants and children of mothers with untreated postpartum depression have been shown to exhibit a higher incidence of colic, excessive crying, sleep problems, and irritability.^31,32 Women with postpartum depression may be less likely to initiate or maintain breastfeeding, and depressive symptoms have been noted to precede the discontinuation of breastfeeding.^33–35

Risk factors for postpartum depression

Characteristics to look for in the prenatal care of pregnant women include the following:

• Depression during pregnancy
• History of postpartum or other depressive episode
• Poverty
• Conflict with primary partner
• Poor social support
• Low self-esteem
• Single status.

Primary care providers are often the first point of contact for depressed women. The diagnosis of major depression in pregnant and postpartum women is challenging because of changes in sleep, appetite, and energy brought on by pregnancy, complications of delivery, and demands of caring for a newborn.³⁶ Many pregnant and postpartum women are reluctant to disclose their symptoms due to a sense of shame and guilt for being depressed during a time in their life that society commonly regards as joyful, and this contributes to under-detection.

In the first few days postpartum, fatigue, emotionality, irritability, and worry over the infant’s well-being affect up to 75% of women. This period, typically referred to as the “baby blues” or “postpartum blues,” is not considered a disorder and responds well to support, reassurance, and adequate sleep, and it typically resolves within 2 weeks.^37,38 Table 1 lists features that help distinguish postpartum blues from major depression.

Signs of major depressive disorder

Major depressive disorder is a serious and disabling condition. To meet criteria for major depressive disorder, women must report depressed mood and loss of interest or pleasure in normally pleasurable activities for at least 2 weeks. Completing the symptom profile, at least 5 of the following must be present: sleep disturbance (insomnia or hypersomnia), lack of energy, feelings of worthlessness or low self-esteem, guilt, difficulty concentrating, indecisiveness, psychomotor retardation or agitation, and thoughts of suicide or death.

The Diagnostic and Statistical Manual of Mental Disorders (5th edition) recognizes that postpartum depression commonly begins during pregnancy, and now uses “peripartum onset” as the specifier for major depressive disorder that occurs during pregnancy, postpartum, or both.³⁹ Other hallmark symptoms with peripartum onset include a lack of interest in or attachment to the pregnancy or infant, and anxiety and worry often accompanied by intrusive, unwanted thoughts of harm befalling the infant.⁴⁰

Postpartum psychosis

Postpartum psychosis is a far less common presentation, occurring in 1 to 2 per 1,000 births, but it constitutes a psychiatric emergency requiring immediate referral to a psychiatric care setting. Women at highest risk are those with a personal or family history of bipolar disorder.

The clinical presentation is most commonly characterized by confusion, agitation, hallucinations, delusional beliefs, and disorientation. Suicide and infanticide, while rare, are more likely to occur in the context of a psychotic episode.⁴¹

SCREENING RECOMMENDATIONS

Screening for depression is routine in primary care settings and is no less important for peripartum women.

In 2016, the US Preventive Services Task Force issued a recommendation that all pregnant and postpartum women be screened for depression,⁴² highlighting the need for all medical providers to be alert to the potentially serious consequences of unrecognized and untreated maternal psychiatric illness.

The American College of Obstetricians and Gynecologists (ACOG) recommends screening for depression and anxiety at least once during the peripartum period,⁴³ and the American Academy of Pediatrics recommends screening mothers for depression at the 1-, 2-, and 4-month well-baby visits.⁴⁴

The peripartum period is associated with changes in sleep, appetite, and energy levels, but these are also typical of depression. Taking this into account, the Edinburgh Postnatal Depression Scale (EPDS) was developed to screen for depression specifically in this population.⁴⁵ The EPDS is a validated and widely used 10-item self-reporting questionnaire with a high degree of sensitivity and specificity; it is easily administered and quickly scored. A cutoff score of 13 (of a maximum of 30) is considered indicative of depressed mood and signals the need for further assessment.

Source: Kroenke K, Spitzer RL, Williams JB. The PHQ-9: Validity of a brief depression severity measure. J Gen In-tern Med 2001; 16:606–616. No permission required to reproduce, translate, display, or distribute.

ACOG, the American Academy of Pediatrics, and the US Preventive Services Task Force recommend a standardized validated tool and cite both the EPDS (https://psychology-tools.com/epds/) and the Patient Health Questionnaire-9 (PHQ-9) (Figure 1) as appropriate to screen for peripartum depression.^42–44 Primary care providers tend to be most familiar with the PHQ-9, a highly sensitive and specific 9-item depression screen that has been validated in primary care and obstetric clinic patients.⁴⁶ A score on the PHQ-9 ranging from 5 to 10 indicates mild depression, 10 to 14 moderate depression, 15 to 19 moderate to severe depression, and greater than 19 severe depression.

Many women prefer nondrug therapy

The gold standard treatment for moderate to severe major depressive disorder is psychotherapy plus pharmacotherapy. Yet many peripartum women voice concerns about exposure to pharmacologic treatment, and studies have shown that many women prefer nonpharmacologic intervention.⁴⁷

Evidence-based psychotherapies that have demonstrated efficacy in peripartum women include cognitive behavioral therapy⁴⁸ and interpersonal psychotherapy when administered by a psychotherapist trained in these treatments. Pregnant and breastfeeding women often express preference for psychotherapy and complementary and alternative treatments as a means of avoiding fetal and infant exposure to antidepressants.⁴⁷

For mild to moderate depression, complementary therapies such as exercise, yoga, bright light therapy, and acupuncture have shown efficacy and can be used alone or adjunctively.⁴⁹ Because a poor marital relationship is consistently associated with peripartum depression,²⁵ primary care physicians who routinely address social support and screen for family conflict are well positioned to detect this significant correlate and to recommend marital or family therapy as a primary or adjunctive treatment.

When to consider drug therapy

The decision to recommend drug therapy must be individualized and based on the severity of symptoms, functional impairment, number and frequency of depressive episodes, history of response to medications, and the preferences of the patient, with the recognition that no decision is risk-free and that antidepressants enter the amniotic fluid, so fetal exposure is unavoidable.

Table 2 lists common antidepressants. The antidepressants most commonly prescribed, especially in the primary care setting, are selective serotonin reuptake inhibitors (SSRIs), which are favored because of their effectiveness, low side-effect profile, and lack of overdose toxicity.

Serotonin syndrome is no more likely to occur in pregnant than in nonpregnant women. Close monitoring for this condition is warranted only when patients are taking very  high doses of SSRIs or SSRIs in combination with other serotonergic agonists.

Prescribing antidepressants for pregnant or breastfeeding women requires thoughtful consideration of the patient’s preferences, as well as weighing the risks and benefits of fetal and infant exposure to maternal depression vs exposure to medications. Additional considerations include monotherapy, avoiding medication changes, choosing drugs that have been effective in the past, and avoiding drugs with known drug-drug interactions or teratogenic effects.⁵⁰

There is increasing consensus that the short- and long-term consequences of undertreatment or nontreatment of maternal depression outweigh the risk of fetal exposure to SSRIs.^3,51,52 Cohen et al⁵³ have recommended that if a woman is on an antidepressant and learns she is pregnant, she should not discontinue it because of the likelihood of relapse; they found a 68% relapse rate in women who discontinued their antidepressant in the first trimester of pregnancy.⁵³

In a comprehensive review of studies published between 1996 and 2012 that examined antidepressant use during pregnancy, Byatt et al⁵⁴ found little or no evidence of increased teratogenic risk with antidepressants with the exception of paroxetine, which is associated with a small but significant increased risk of cardiac malformation during first-trimester exposure.⁵⁴

These conclusions were underscored in a large cohort study in the United Kingdom.⁵⁵ In addition, a joint task force of the American Psychiatric Association and ACOG reviewed studies looking at the association between depression, antidepressants, and birth outcomes including miscarriage, preterm birth, cardiac abnormalities (resulting from first trimester exposure), persistent pulmonary hypertension (related to second- and third-trimester exposure), and neonatal adaptation syndrome (associated with third-trimester exposure).⁸ They concluded that the available data neither support nor refute a link between the use of antidepressants and several of the above outcomes. No increase in risk of congenital malformations (including cardiac abnormalities) was found. An increased risk of persistent pulmonary hypertension was noted, although the absolute risk of this disorder remained low, at 3 to 6 per 1,000 infants exposed to SSRIs in utero.^8,56

Neonatal adaptation syndrome

Neonatal adaptation syndrome is characterized by jitteriness, irritability, decreased muscle tone, and feeding difficulty in the neonate. It can occur in 15% to 30% of infants exposed to SSRIs antenatally.^57,58 These symptoms, however, are transient and typically resolve within 7 to 10 days after birth. A more recent study suggested that neurobehavioral symptoms for some infants extend beyond 2 weeks and that concomitant exposure to benzodiazepines results in even higher rates of this syndrome.⁵⁹ There is no evidence that tapering or discontinuing antidepressants near term is necessary, safe, or effective in preventing transient neonatal complications. However, this approach would increase the risk of relapse for the mother.

Autism spectrum disorders

The possible association between antidepressants and autism spectrum disorders in pregnancy has captured much attention in recent years. One study based on healthcare claims⁶⁰ and one registry-based study⁶¹ associated in utero exposure to antidepressants with autism liability in children. However, a large-scale Danish registry-based study did not replicate this association.⁶² In addition, 2 recent cohort studies, identifying children with autism spectrum disorder or attention-deficit hyperactivity disorder from electronic health records, found that neither disorder was significantly associated with prenatal antidepressant exposure in crude or adjusted models. However, both studies found a significant association with the use of antidepressants before pregnancy, indicating that the risk of autism observed with prenatal antidepressant exposure is likely confounded by the severity of maternal illness.^63,64

Concerns about drug therapy during breastfeeding

For infants of breastfeeding women, exposure to antidepressants through breast milk is minimal. Amounts in breast milk depend on the timing of the antidepressant dose, timing of feeding, and genetically influenced metabolic activity in mother and infant. The current literature supports antidepressant use for breastfeeding mothers of healthy full-term infants.⁶⁵

The 2 most widely studied antidepressants in breastfed infants are paroxetine and sertraline. It has been shown that very little can be detected in the infant’s serum, with relative infant doses ranging from 0.4% to 2.8%.⁶⁵ While clinicians are cautioned against prescribing paroxetine for pregnant women, the drug remains a suitable alternative for breastfeeding women.

If an antidepressant is started postpartum, the recommendation is to start with a low dose and then slowly titrate upward while monitoring the infant for adverse effects.^65,66 Possible adverse effects in breastfeeding infants include irritability, sedation, poor weight gain, and a change in feeding patterns.⁶⁷ Adverse events are most likely to occur in newborns up to 8 weeks of age, and infants born prematurely or with medical problems may be particularly at risk.^65,68

Helping patients weigh risks and benefits of drug therapy

Women may hear about the risks of medications to the fetus and during breastfeeding and so may be reluctant to seek or accept intervention. Often, the information is not from a reliable, scientifically based source. Primary care physicians are well positioned to guide peripartum women in risk-benefit analysis of proper treatment of their depression vs no treatment or undertreatment. In addition, establishing referral sources—ideally with a peripartum mental health specialist—is advisable. Online resources that clinicians can refer patients to for help in managing peripartum depression include the following:

• www.postpartum.net
• www.womensmentalhealth.org
• www.mothertobaby.org (for pharmacologic guidance).

INCREASED AWARENESS IS KEY

Primary care physicians must remain alert to the high prevalence of depression in women of childbearing age and embrace routine screening for depression. (See the sidebar, “The primary care management of peripartum depression.”) Since half of pregnancies are unintended, awareness of the risks of undetected and untreated peripartum depression to the mother, developing fetus, and infant is essential. Untreated antepartum depression has been linked to poor pregnancy outcomes, nutritional deficits, and substance abuse. Untreated postpartum depression negatively affects mother-infant attachment, infant, and child development and maternal self care.

Not treating depression is hazardous

Drug treatment during pregnancy and breastfeeding poses challenges for the patient and physician due to the inevitability of fetal and infant exposure, but lack of treatment can be hazardous.

To date, the evidence on the use of antidepressants in pregnant and lactating women is reassuring. Specialized peripartum psychiatric partial hospital programs⁶⁹ and inpatient programs⁷⁰ exist for women who need a higher level of care. There is also substantial evidence that psychotherapy, especially cognitive behavioral therapy and interpersonal therapy, is highly effective, and emerging data on complementary and alternative treatments are promising. Coordinated care between primary care and behavioral healthcare providers with expertise in treating peripartum depression is most likely to yield optimal outcomes.

Contrary to common belief, pregnancy does not confer protection against depression.^1,2 In fact, pregnant women are just as likely as nonpregnant women to become or remain depressed, and up to 12.7% of pregnant women meet criteria for depression.¹

In the postpartum period, women are particularly vulnerable to a major depressive episode, whether a first episode or a recurrence. The estimated prevalence of a depressive episode in the first 3 postpartum  months is 19.2%,² making postpartum depression the most common complication of childbearing.² At the same time, peripartum depression remains largely underrecognized and undertreated.³

As evidence mounts regarding the deleterious impact of untreated mental illness on the mother, the developing fetus, and the infant, early detection and intervention for peripartum depression are paramount.³

DEPRESSION DURING PREGNANCY: SIGNIFICANT CONSEQUENCES

Although the rates of depression in pregnant and nonpregnant women are similar, depression in pregnancy carries additional significant consequences. Further, many depressed pregnant women believe their depression will lift once their baby is born, though it is well documented that depression during pregnancy is the strongest predictor of postpartum depression and that if left untreated it can be devastating for mother, infant, and family.⁴

Compared with nondepressed pregnant women, depressed pregnant women have poorer overall health status,⁵ are more likely to engage in behaviors that pose risk to the developing fetus such as smoking,⁵ alcohol consumption, and substance use,⁶ and have poor nutrition and inadequate weight gain.^7,8

Pregnant women who are depressed and are also experiencing domestic violence are especially at risk for poor prenatal care as they tend to miss more prenatal appointments.⁹ Evidence also suggests that depressed pregnant women are less attached to the fetus and more likely to have elective terminations.^10,11

Depression in pregnancy is associated with higher rates of adverse pregnancy outcomes such as preterm birth, low birth weight, operative delivery, and longer predelivery hospital stay.^3,12 Depression and anxiety during pregnancy have been associated with prenatal hypertension,¹³ gestational diabetes,¹⁴ preeclampsia,¹⁵ and HELLP syndrome (ie, hemolysis, elevated liver enzymes, and low platelet count).¹⁵ Depression and anxiety during pregnancy are associated with subsequent poorer infant attachment^16,17 and an overall unfavorable impact on infant and child development.¹⁸

Risk factors for depression during pregnancy include past episodes of depression, current anxiety, poor social support, unintended pregnancy, life stress, being single, domestic violence, and being on Medicaid.¹⁹

Undoubtedly the most devastating consequence of severe depression during pregnancy is suicide. Rates of suicide are lower in peripartum women,²⁰ but when suicide does occur, pregnant women tend to use more violent means than nonpregnant women. Pregnant adolescents represent a particularly high-risk group.²¹

POSTPARTUM DEPRESSION

Postpartum depression is the most common complication of childbearing. Although the precise pathogenesis is undetermined, there is converging evidence of a subset of women particularly sensitive to dramatic fluctuations in levels of estradiol and progesterone that occur during childbirth.^22,23 There is also evidence that dysregulation of the hypothalamic-pituitary-adrenal axis contributes to the development of postpartum depression in certain women.²⁴ Further, women who have depression or anxiety during pregnancy are much more likely to experience postpartum depression than those who are not symptomatic during pregnancy.⁴ A history of peripartum depression or other lifetime depressive episodes, poverty, conflict with a primary partner, poor social support, stressful life events, and low self-esteem are strongly associated with postpartum depression.²⁵

When unrecognized and untreated, postpartum depression can have profound and persistent effects on the mother and the developing infant.^18,26 Mothers with postpartum depression are much more likely than mothers without depression to have impaired bonding,²⁷ to be less responsive to their infant’s needs,¹⁷ and to be more likely to miss well-baby checkups.²⁸

Postpartum depression’s effects on maternal-infant interactions can include maternal withdrawal, disengagement, intrusion, and hostility and can lead to long-term effects on child development, including poor cognitive functioning, emotional maladjustment, and behavioral inhibition.^29,30 Infants and children of mothers with untreated postpartum depression have been shown to exhibit a higher incidence of colic, excessive crying, sleep problems, and irritability.^31,32 Women with postpartum depression may be less likely to initiate or maintain breastfeeding, and depressive symptoms have been noted to precede the discontinuation of breastfeeding.^33–35

Risk factors for postpartum depression

Characteristics to look for in the prenatal care of pregnant women include the following:

• Depression during pregnancy
• History of postpartum or other depressive episode
• Poverty
• Conflict with primary partner
• Poor social support
• Low self-esteem
• Single status.

Primary care providers are often the first point of contact for depressed women. The diagnosis of major depression in pregnant and postpartum women is challenging because of changes in sleep, appetite, and energy brought on by pregnancy, complications of delivery, and demands of caring for a newborn.³⁶ Many pregnant and postpartum women are reluctant to disclose their symptoms due to a sense of shame and guilt for being depressed during a time in their life that society commonly regards as joyful, and this contributes to under-detection.

In the first few days postpartum, fatigue, emotionality, irritability, and worry over the infant’s well-being affect up to 75% of women. This period, typically referred to as the “baby blues” or “postpartum blues,” is not considered a disorder and responds well to support, reassurance, and adequate sleep, and it typically resolves within 2 weeks.^37,38 Table 1 lists features that help distinguish postpartum blues from major depression.

Signs of major depressive disorder

Major depressive disorder is a serious and disabling condition. To meet criteria for major depressive disorder, women must report depressed mood and loss of interest or pleasure in normally pleasurable activities for at least 2 weeks. Completing the symptom profile, at least 5 of the following must be present: sleep disturbance (insomnia or hypersomnia), lack of energy, feelings of worthlessness or low self-esteem, guilt, difficulty concentrating, indecisiveness, psychomotor retardation or agitation, and thoughts of suicide or death.

The Diagnostic and Statistical Manual of Mental Disorders (5th edition) recognizes that postpartum depression commonly begins during pregnancy, and now uses “peripartum onset” as the specifier for major depressive disorder that occurs during pregnancy, postpartum, or both.³⁹ Other hallmark symptoms with peripartum onset include a lack of interest in or attachment to the pregnancy or infant, and anxiety and worry often accompanied by intrusive, unwanted thoughts of harm befalling the infant.⁴⁰

Postpartum psychosis

Postpartum psychosis is a far less common presentation, occurring in 1 to 2 per 1,000 births, but it constitutes a psychiatric emergency requiring immediate referral to a psychiatric care setting. Women at highest risk are those with a personal or family history of bipolar disorder.

The clinical presentation is most commonly characterized by confusion, agitation, hallucinations, delusional beliefs, and disorientation. Suicide and infanticide, while rare, are more likely to occur in the context of a psychotic episode.⁴¹

SCREENING RECOMMENDATIONS

Screening for depression is routine in primary care settings and is no less important for peripartum women.

In 2016, the US Preventive Services Task Force issued a recommendation that all pregnant and postpartum women be screened for depression,⁴² highlighting the need for all medical providers to be alert to the potentially serious consequences of unrecognized and untreated maternal psychiatric illness.

The American College of Obstetricians and Gynecologists (ACOG) recommends screening for depression and anxiety at least once during the peripartum period,⁴³ and the American Academy of Pediatrics recommends screening mothers for depression at the 1-, 2-, and 4-month well-baby visits.⁴⁴

The peripartum period is associated with changes in sleep, appetite, and energy levels, but these are also typical of depression. Taking this into account, the Edinburgh Postnatal Depression Scale (EPDS) was developed to screen for depression specifically in this population.⁴⁵ The EPDS is a validated and widely used 10-item self-reporting questionnaire with a high degree of sensitivity and specificity; it is easily administered and quickly scored. A cutoff score of 13 (of a maximum of 30) is considered indicative of depressed mood and signals the need for further assessment.

Source: Kroenke K, Spitzer RL, Williams JB. The PHQ-9: Validity of a brief depression severity measure. J Gen In-tern Med 2001; 16:606–616. No permission required to reproduce, translate, display, or distribute.

ACOG, the American Academy of Pediatrics, and the US Preventive Services Task Force recommend a standardized validated tool and cite both the EPDS (https://psychology-tools.com/epds/) and the Patient Health Questionnaire-9 (PHQ-9) (Figure 1) as appropriate to screen for peripartum depression.^42–44 Primary care providers tend to be most familiar with the PHQ-9, a highly sensitive and specific 9-item depression screen that has been validated in primary care and obstetric clinic patients.⁴⁶ A score on the PHQ-9 ranging from 5 to 10 indicates mild depression, 10 to 14 moderate depression, 15 to 19 moderate to severe depression, and greater than 19 severe depression.

Many women prefer nondrug therapy

The gold standard treatment for moderate to severe major depressive disorder is psychotherapy plus pharmacotherapy. Yet many peripartum women voice concerns about exposure to pharmacologic treatment, and studies have shown that many women prefer nonpharmacologic intervention.⁴⁷

Evidence-based psychotherapies that have demonstrated efficacy in peripartum women include cognitive behavioral therapy⁴⁸ and interpersonal psychotherapy when administered by a psychotherapist trained in these treatments. Pregnant and breastfeeding women often express preference for psychotherapy and complementary and alternative treatments as a means of avoiding fetal and infant exposure to antidepressants.⁴⁷

For mild to moderate depression, complementary therapies such as exercise, yoga, bright light therapy, and acupuncture have shown efficacy and can be used alone or adjunctively.⁴⁹ Because a poor marital relationship is consistently associated with peripartum depression,²⁵ primary care physicians who routinely address social support and screen for family conflict are well positioned to detect this significant correlate and to recommend marital or family therapy as a primary or adjunctive treatment.

When to consider drug therapy

The decision to recommend drug therapy must be individualized and based on the severity of symptoms, functional impairment, number and frequency of depressive episodes, history of response to medications, and the preferences of the patient, with the recognition that no decision is risk-free and that antidepressants enter the amniotic fluid, so fetal exposure is unavoidable.

Table 2 lists common antidepressants. The antidepressants most commonly prescribed, especially in the primary care setting, are selective serotonin reuptake inhibitors (SSRIs), which are favored because of their effectiveness, low side-effect profile, and lack of overdose toxicity.

Serotonin syndrome is no more likely to occur in pregnant than in nonpregnant women. Close monitoring for this condition is warranted only when patients are taking very  high doses of SSRIs or SSRIs in combination with other serotonergic agonists.

Prescribing antidepressants for pregnant or breastfeeding women requires thoughtful consideration of the patient’s preferences, as well as weighing the risks and benefits of fetal and infant exposure to maternal depression vs exposure to medications. Additional considerations include monotherapy, avoiding medication changes, choosing drugs that have been effective in the past, and avoiding drugs with known drug-drug interactions or teratogenic effects.⁵⁰

There is increasing consensus that the short- and long-term consequences of undertreatment or nontreatment of maternal depression outweigh the risk of fetal exposure to SSRIs.^3,51,52 Cohen et al⁵³ have recommended that if a woman is on an antidepressant and learns she is pregnant, she should not discontinue it because of the likelihood of relapse; they found a 68% relapse rate in women who discontinued their antidepressant in the first trimester of pregnancy.⁵³

In a comprehensive review of studies published between 1996 and 2012 that examined antidepressant use during pregnancy, Byatt et al⁵⁴ found little or no evidence of increased teratogenic risk with antidepressants with the exception of paroxetine, which is associated with a small but significant increased risk of cardiac malformation during first-trimester exposure.⁵⁴

These conclusions were underscored in a large cohort study in the United Kingdom.⁵⁵ In addition, a joint task force of the American Psychiatric Association and ACOG reviewed studies looking at the association between depression, antidepressants, and birth outcomes including miscarriage, preterm birth, cardiac abnormalities (resulting from first trimester exposure), persistent pulmonary hypertension (related to second- and third-trimester exposure), and neonatal adaptation syndrome (associated with third-trimester exposure).⁸ They concluded that the available data neither support nor refute a link between the use of antidepressants and several of the above outcomes. No increase in risk of congenital malformations (including cardiac abnormalities) was found. An increased risk of persistent pulmonary hypertension was noted, although the absolute risk of this disorder remained low, at 3 to 6 per 1,000 infants exposed to SSRIs in utero.^8,56

Neonatal adaptation syndrome

Neonatal adaptation syndrome is characterized by jitteriness, irritability, decreased muscle tone, and feeding difficulty in the neonate. It can occur in 15% to 30% of infants exposed to SSRIs antenatally.^57,58 These symptoms, however, are transient and typically resolve within 7 to 10 days after birth. A more recent study suggested that neurobehavioral symptoms for some infants extend beyond 2 weeks and that concomitant exposure to benzodiazepines results in even higher rates of this syndrome.⁵⁹ There is no evidence that tapering or discontinuing antidepressants near term is necessary, safe, or effective in preventing transient neonatal complications. However, this approach would increase the risk of relapse for the mother.

Autism spectrum disorders

The possible association between antidepressants and autism spectrum disorders in pregnancy has captured much attention in recent years. One study based on healthcare claims⁶⁰ and one registry-based study⁶¹ associated in utero exposure to antidepressants with autism liability in children. However, a large-scale Danish registry-based study did not replicate this association.⁶² In addition, 2 recent cohort studies, identifying children with autism spectrum disorder or attention-deficit hyperactivity disorder from electronic health records, found that neither disorder was significantly associated with prenatal antidepressant exposure in crude or adjusted models. However, both studies found a significant association with the use of antidepressants before pregnancy, indicating that the risk of autism observed with prenatal antidepressant exposure is likely confounded by the severity of maternal illness.^63,64

Concerns about drug therapy during breastfeeding

For infants of breastfeeding women, exposure to antidepressants through breast milk is minimal. Amounts in breast milk depend on the timing of the antidepressant dose, timing of feeding, and genetically influenced metabolic activity in mother and infant. The current literature supports antidepressant use for breastfeeding mothers of healthy full-term infants.⁶⁵

The 2 most widely studied antidepressants in breastfed infants are paroxetine and sertraline. It has been shown that very little can be detected in the infant’s serum, with relative infant doses ranging from 0.4% to 2.8%.⁶⁵ While clinicians are cautioned against prescribing paroxetine for pregnant women, the drug remains a suitable alternative for breastfeeding women.

If an antidepressant is started postpartum, the recommendation is to start with a low dose and then slowly titrate upward while monitoring the infant for adverse effects.^65,66 Possible adverse effects in breastfeeding infants include irritability, sedation, poor weight gain, and a change in feeding patterns.⁶⁷ Adverse events are most likely to occur in newborns up to 8 weeks of age, and infants born prematurely or with medical problems may be particularly at risk.^65,68

Helping patients weigh risks and benefits of drug therapy

Women may hear about the risks of medications to the fetus and during breastfeeding and so may be reluctant to seek or accept intervention. Often, the information is not from a reliable, scientifically based source. Primary care physicians are well positioned to guide peripartum women in risk-benefit analysis of proper treatment of their depression vs no treatment or undertreatment. In addition, establishing referral sources—ideally with a peripartum mental health specialist—is advisable. Online resources that clinicians can refer patients to for help in managing peripartum depression include the following:

• www.postpartum.net
• www.womensmentalhealth.org
• www.mothertobaby.org (for pharmacologic guidance).

INCREASED AWARENESS IS KEY

Primary care physicians must remain alert to the high prevalence of depression in women of childbearing age and embrace routine screening for depression. (See the sidebar, “The primary care management of peripartum depression.”) Since half of pregnancies are unintended, awareness of the risks of undetected and untreated peripartum depression to the mother, developing fetus, and infant is essential. Untreated antepartum depression has been linked to poor pregnancy outcomes, nutritional deficits, and substance abuse. Untreated postpartum depression negatively affects mother-infant attachment, infant, and child development and maternal self care.

Not treating depression is hazardous

Drug treatment during pregnancy and breastfeeding poses challenges for the patient and physician due to the inevitability of fetal and infant exposure, but lack of treatment can be hazardous.

To date, the evidence on the use of antidepressants in pregnant and lactating women is reassuring. Specialized peripartum psychiatric partial hospital programs⁶⁹ and inpatient programs⁷⁰ exist for women who need a higher level of care. There is also substantial evidence that psychotherapy, especially cognitive behavioral therapy and interpersonal therapy, is highly effective, and emerging data on complementary and alternative treatments are promising. Coordinated care between primary care and behavioral healthcare providers with expertise in treating peripartum depression is most likely to yield optimal outcomes.

Chronic constipation has a variety of possible causes and mechanisms. Although traditional conservative treatments are still valid and first-line, if these fail, clinicians can choose from a growing list of new treatments, tailored to the cause in the individual patient.

This article discusses how defecation works (or doesn’t), the types of chronic constipation, the available diagnostic tools, and traditional and newer treatments, including some still in development.

THE EPIDEMIOLOGY OF CONSTIPATION

Chronic constipation is one of the most common gastrointestinal disorders, affecting about 15% of all adults and 30% of those over the age of 60.¹ It can be a primary disorder or secondary to other factors.

Constipation is more prevalent in women and in institutionalized elderly people.² It is associated with lower socioeconomic status, depression, less self-reported physical activity, certain medications, and stressful life events.³ Given its high prevalence and its impact on quality of life, it is also associated with significant utilization of healthcare resources.⁴

Constipation defined by Rome IV criteria

Physicians and patients may disagree about what constitutes constipation. Physicians primarily regard it as infrequent bowel movements, while patients tend to have a broader definition. According to the Rome IV criteria,⁵ chronic constipation is defined by the presence of the following for at least 3 months (with symptom onset at least 6 months prior to diagnosis):

 (1) Two or more of the following for more than 25% of defecations:

• Straining
• Lumpy or hard stools
• Sensation of incomplete evacuation
• Sensation of anorectal obstruction or blockage
• Manual maneuvers to facilitate evacuation
• Fewer than 3 spontaneous bowel movements per week.

 (2) Loose stools are rarely present without the use of laxatives.

 (3) The patient does not meet the criteria for diagnosis of irritable bowel syndrome.

DEFECATION IS COMPLEX

Defecation begins when the rectum fills with stool, causing relaxation of the internal anal sphincter and the urge to defecate. The external anal sphincter, which is under voluntary control, can then either contract to delay defecation or relax to allow the stool to be expelled.⁶

Colonic muscles propel stool toward the rectum in repetitive localized contractions that help mix and promote absorption of the content, and larger coordinated (high-amplitude propagating) contractions that, in healthy individuals, move the stool forward from the proximal to the distal colon multiple times daily. These contractions usually occur in the morning and are accentuated by gastric distention from food and the resulting gastrocolic reflex.

Serotonin (5-HT) is released by enterochromaffin cells in response to distention of the gut wall. It mediates peristaltic movements of the gastrointestinal tract by binding to receptors (especially 5-HT4), stimulating release of neurotransmitters such as acetylcholine, causing smooth-muscle contraction behind the luminal contents and propelling them forward.

PRIMARY CONSTIPATION DISORDERS

The American Gastroenterological Association⁷ classifies constipation into 3 groups on the basis of colonic transit time and anorectal function:

Normal-transit constipation

Stool normally takes 20 to 72 hours to pass through the colon, with transit time affected by diet, drugs, level of physical activity, and emotional status.⁸

Normal-transit constipation is the most common type of constipation. The term is sometimes used interchangeably with constipation-predominant irritable bowel syndrome, but the latter is a distinct entity characterized by abdominal pain relieved by defecation as the primary symptom, as well as having occasional loose stools. These 2 conditions can be hard to tell apart, especially if the patient cannot describe the symptoms precisely.

Slow-transit constipation

Slow-transit constipation—also called delayed-transit constipation, colonoparesis, colonic inertia, and pseudo-obstruction—is defined as prolonged stool transit in the colon, ie, for more than 5 days.⁹ It can be the result of colonic smooth muscle dysfunction, compromised colonic neural pathways, or both, leading to slow colon peristalsis.

Factors that can affect colonic motility such as opioid use and hypothyroidism should be carefully considered in these patients. Opioids are notorious for causing constipation by decreasing bowel tone and contractility and thereby increasing colonic transit time. They also tighten up the anal sphincters, resulting in decreased rectal evacuation.¹⁰

Outlet dysfunction

Outlet dysfunction, also called pelvic floor dysfunction or defecatory disorder, is associated with incomplete rectal evacuation. It can be a consequence of weak rectal expulsion forces (slow colonic transit, rectal hyposensitivity), functional resistance to rectal evacuation (high anal resting pressure, anismus, incomplete relaxation of the anal sphincter, dyssynergic defecation), or structural outlet obstruction (excessive perineal descent, rectoceles, rectal intussusception). About 50% of patients with outlet dysfunction have concurrent slow-transit constipation.

Dyssynergic defecation is the most common outlet dysfunction disorder, accounting for about half of the cases referred to tertiary centers. It is defined as a paradoxical elevation in anal sphincter tone or less than 20% relaxation of the resting anal sphincter pressure with weak abdominal and pelvic propulsive forces.¹¹ Anorectal biofeedback is a therapeutic option for dyssynergic defecation, as we discuss later in this article.

SECONDARY CONSTIPATION

Constipation can be secondary to several conditions and factors (Table 1), including:

• Neurologic disorders that affect gastrointestinal motility (eg, Hirschsprung disease, Parkinson disease, multiple sclerosis, spinal cord injury, stroke, spinal or ganglionic  tumor, hypothyroidism, amyloidosis, diabetes mellitus, hypercalcemia)
• Drugs used to treat neurologic disorders
• Mechanical obstruction
• Diet (eg, low fiber, decreased fluid intake).

EVALUATION OF CONSTIPATION

It is crucial for physicians to efficiently use the available diagnostic tools for constipation to tailor the treatment to the patient.

Evaluation of chronic constipation begins with a thorough history and physical examination to rule out secondary constipation (Figure 1). Red flags such as unintentional weight loss, blood in the stool, rectal pain, fever, and iron-deficiency anemia should prompt referral for colonoscopy to evaluate for malignancy, colitis, or other potential colonic abnormalities.¹²

A detailed perineal and rectal examination can help diagnose defecatory disorders and should include evaluation of the resting anal tone and the sphincter during simulated evacuation.

Laboratory tests of thyroid function, electrolytes, and a complete blood cell count should be ordered if clinically indicated.¹³

Further tests

Further diagnostic tests can be considered if symptoms persist despite conservative treatment or if a defecatory disorder is suspected. These include anorectal manometry, colonic transit studies, defecography, and colonic manometry.

Anorectal manometry and the rectal balloon expulsion test are usually done first because of their high sensitivity (88%) and specificity (89%) for defecatory disorders.¹⁴ These tests measure the function of the internal and external anal sphincters at rest and with straining and assess rectal sensitivity and compliance. Anorectal manometry is also used in biofeedback therapy in patients with dyssynergic defecation.¹⁵

Colonic transit time can be measured if anorectal manometry and the balloon expulsion test are normal. The study uses radiopaque markers, radioisotopes, or wireless motility capsules to confirm slow-transit constipation and to identify areas of delayed transit in the colon.¹⁶

Defecography is usually the next step in diagnosis if anorectal manometry and balloon expulsion tests are inconclusive or if an anatomic abnormality of the pelvic floor is suspected. It can be done with a variety of techniques. Barium defecography can identify anatomic defects, scintigraphy can quantify evacuation of artificial stools, and magnetic resonance defecography visualizes anatomic landmarks to assess pelvic floor motion without exposing the patient to radiation.^17,18

Colonic manometry is most useful in patients with refractory slow-transit constipation and can identify patients with isolated colonic motor dysfunction with no pelvic floor dysfunction who may benefit from subtotal colectomy and end-ileostomy.⁷

TRADITIONAL TREATMENTS STILL THE MAINSTAY

Nonpharmacologic treatments are the first-line options for patients with normal-transit and slow-transit constipation and should precede diagnostic testing. Lifestyle modifications and dietary changes (Table 2) aim to augment the known factors that stimulate the gastrocolic reflex and increase intestinal motility by high-amplitude propagated contractions.

Increasing physical activity increases intestinal gas clearance, decreases bloating, and lessens constipation.^19,20

Toilet training is an integral part of lifestyle modifications.²¹

Diet. Drinking hot caffeinated beverages, eating breakfast within an hour of waking up, and consuming fiber in the morning (25–30 g of fiber daily) have traditionally been recommended as the first-line measures for chronic constipation. Dehydrated patients with constipation also benefit from increasing their fluid intake.²²

LAXATIVES

Fiber (bulk-forming laxatives) for normal-transit constipation

Fiber remains a key part of the initial management of chronic constipation, as it is cheap, available, and safe. Increasing fiber intake is effective for normal-transit constipation, but patients with slow-transit constipation or refractory outlet dysfunction are less likely to benefit.²³ Other laxatives are incorporated into the regimen if first-line nonpharmacologic interventions fail (Table 3).

Bulk-forming laxatives include insoluble fiber (wheat bran) and soluble fiber (psyllium, methylcellulose, inulin, calcium polycarbo­phil). Insoluble fiber, though often used, has little impact on symptoms of chronic constipation after 1 month of use, and up to 60% of patients report adverse effects from it.²⁴ On the other hand, clinical trials have shown that soluble fiber such as psyllium facilitates defecation and improves functional bowel symptoms in patients with normal-transit constipation.²⁵

Patients should be instructed to increase their dietary fiber intake gradually to avoid adverse effects and should be told to expect significant symptomatic improvement only after a few weeks. They should also be informed that increasing dietary fiber intake can cause bloating but that the bloating is temporary. If it continues, a different fiber can be tried.

Osmotic laxatives

Osmotic laxatives are often employed as a first- line laxative treatment option for patients with constipation. They draw water into the lumen by osmosis, helping to soften stool and speed intestinal transit. They include macrogols (inert polymers of ethylene glycol), nonabsorbable carbohydrates (lactulose, sorbitol), magnesium products, and sodium phosphate products.

Polyethylene glycol, the most studied osmotic laxative, has been shown to maintain therapeutic efficacy for up to 2 years, though it is not generally used this long.²⁶ A meta-analysis of 10 randomized clinical trials found it to be superior to lactulose in improving stool consistency and frequency, and rates of adverse effects were similar to those with placebo.²⁷

Lactulose and sorbitol are semisynthetic disaccharides that are not absorbed from the gastrointestinal tract. Apart from the osmotic effect of the disaccharide, these sugars are metabolized by colonic bacteria to acetic acid and other short-chain fatty acids, resulting in acidification of the stool, which exerts an osmotic effect in the colonic lumen.

Lactulose and sorbitol were shown to have similar efficacy in increasing the frequency of bowel movements in a small study, though patients taking lactulose had a higher rate of nausea.²⁸

The usual recommended dose is 15 to 30 mL once or twice daily.

Adverse effects include gas, bloating, and abdominal distention (due to fermentation by colonic bacteria) and can limit long-term use.

Magnesium citrate and magnesium hydroxide are strong osmotic laxatives, but so far no clinical trial has been done to assess their efficacy in constipation. Although the risk of hypermagnesemia is low with magnesium-based products, this group of laxatives is generally avoided in patients with renal or cardiac disease.²⁹

Sodium phosphate enemas (Fleet enemas) are used for bowel cleansing before certain procedures but have only limited use in constipation because of potential adverse effects such as hyperphosphatemia, hypocalcemia, and the rarer but more serious complication of acute phosphate nephropathy.³⁰

Stimulant laxatives for short-term use only

Stimulant laxatives include glycerin, bisacodyl, senna, and sodium picosulfate. Sodium piosulfate and bisacodyl have been validated for treatment of chronic constipation for up to 4 weeks.^31–33

Stimulant laxative suppositories should be used 30 minutes after meals to augment the physiologic gastrocolic reflex.

As more evidence is available for osmotic laxatives such as polyethylene glycol, they tend to be preferred over stimulant agents, especially for long-term use. Clinicians have traditionally hesitated to prescribe stimulant laxatives for long-term use, as they were thought to damage the enteric nervous system.³⁴ Although more recent studies have not shown this potential effect,³⁵ more research is warranted on the use of stimulant laxatives for longer than 4 weeks.

Stool softeners enhance the interaction of stool and water, leading to softer stool and easier evacuation. Docusate sodium and docusate calcium are thought to facilitate the mixing of aqueous and fatty substances, thereby softening the stool.

However, there is little evidence to support the use of docusate for constipation in hospitalized adults or in ambulatory care. A recent review reported that docusate was no better than placebo in diminishing symptoms of constipation.³⁶

INTESTINAL SECRETAGOGUES

The secretagogues include lubiprostone, linaclotide, and plecanatide. These medications are preferred therapy for patients with normal- or slow-transit constipation once conservative therapies have failed. Even though there is no current consensus, lifestyle measures and conservative treatment options should be tried for about 8 weeks.

Lubiprostone and linaclotide are approved by the US Food and Drug Administration (FDA) for both constipation and constipation-predominant irritable bowel syndrome. They activate chloride channels on the apical surface of enterocytes, increasing intestinal secretion of chloride, which in turn increases luminal sodium efflux to maintain electroneutrality, leading to secretion of water into the intestinal lumen. This eventually facilitates intestinal transit and increases the passage of stool.

Lubiprostone

Lubiprostone, a prostaglandin E1 derivative, is approved for treating chronic constipation, constipation-predominant irritable bowel syndrome in women, and opioid-induced constipation in patients with chronic noncancer pain.

Adverse effects in clinical trials were nausea (up to 30%) and headache.^37,38

Linaclotide

Linaclotide, a minimally absorbed 14-amino acid peptide, increases intestinal secretion of chloride and bicarbonate, increasing intestinal fluid and promoting intestinal transit.³⁹ It also decreases the firing rate of the visceral afferent pain fibers and helps reduce visceral pain, especially in patients with constipation-predominant irritable bowel syndrome.⁴⁰ It is approved for chronic constipation and constipation-predominant irritable bowel syndrome.^41–43

Dosage starts at 145 μg/day for chronic constipation, and can be titrated up to 290 μg if there is no response or if a diagnosis of constipation-predominant irritable bowel syndrome is under consideration. Linaclotide should be taken 30 to 60 minutes before breakfast to reduce the likelihood of diarrhea.⁴⁴

Adverse effects. Diarrhea led to treatment discontinuation in 4.5% of patients in one study.⁴²

Plecanatide

Plecanatide is a guanylate cyclase-c agonist with a mode of action similar to that of linaclotide. It was recently approved by the FDA for chronic idiopathic constipation in adults. The recommended dose is 3 mg once daily.

Data from phase 2 trials in chronic constipation showed improvement in straining, abdominal discomfort, and stool frequency after 14 days of treatment.⁴⁵

A phase 3 trial showed that plecanatide was more effective than placebo when used for 12 weeks in 951 patients with chronic constipation (P = .009).⁴⁶ The most common adverse effect reported was diarrhea.

SEROTONIN RECEPTOR AGONISTS

Activation of serotonin 5-HT4 receptors in the gut leads to release of acetylcholine, which in turn induces mucosal secretion by activating submucosal neurons and increasing gut motility.⁴⁷

Two 5-HT4 receptor agonists were withdrawn from the market (cisapride in 2000 and tegaserod in 2007) due to serious cardiovascular adverse events (fatal arrhythmias, heart attacks, and strokes) resulting from their affinity for hERG-K+ cardiac channels.  

The newer agents prucalopride,⁴⁸ velusetrag, and naronapride are highly selective 5-HT4 agonists with low affinity for hERG-K+ receptors and do not have proarrhythmic properties, based on extensive assessment in clinical trials.

Prucalopride

Prucalopride has been shown to accelerate gastrointestinal and colonic transit in patients with chronic constipation, with improvement in bowel movements, symptoms of chronic constipation, and quality of life.^49–52

Adverse effects reported with its use have been headache, nausea, abdominal pain, and cramps.

Prucalopride is approved in Europe and Canada for chronic constipation in women but is not yet approved in the United States.

Dosage is 2 mg orally once daily. Caution is advised in elderly patients, in whom the preferred maximum dose is 1 mg daily, as there are only limited data available on the safety of this medication in the elderly.

Velusetrag

Velusetrag has been shown to increase colonic motility and improve symptoms of chronic constipation. In a phase 2 trial,⁵³ the most effective dose was 15 mg once daily. Higher doses were associated with a higher incidence of adverse effects such as diarrhea, headache, nausea, and vomiting.

Naronapride

Naronapride (ATI-7505) is in phase 2 trials for chronic constipation. Reported adverse effects were headache, diarrhea, nausea, and vomiting.⁵⁴

BILE SALT ABSORPTION INHIBITORS

Bile acids exert prosecretory and prokinetic effects by increasing colonic secretion of water and electrolytes through the activation of adenylate cyclase. This happens as a result of their deconjugation after passage into the colon.

Elobixibat is an ileal bile acid transporter inhibitor that prevents absorption of nonconjugated bile salts in the distal ileum. It has few side effects because its systemic absorption is minimal. Phase 3 trials are under way. Dosage is 5 to 20 mg daily. Adverse effects are few because systemic absorption is minimal, but include abdominal pain and diarrhea.^55,56

Opioids cause constipation by binding to mu receptors in the enteric nervous system. Activation of these receptors decreases bowel tone and contractility, which increases transit time. Stimulation of these receptors also increases anal sphincter tone, resulting in decreased rectal evacuation.⁵⁷

Though underrecognized, opioid-induced constipation affects 40% of patients who take these drugs for nonmalignant pain and 90% of those taking them for cancer pain. Patients with this condition were found to take more time off work and feel more impaired in their domestic and work-related obligations than patients who did not develop constipation with use of opioids.⁵⁸

Initial management of opioid-induced constipation includes increasing intake of fluids and dietary fiber (fiber alone can worsen abdominal pain in this condition by increasing stool bulk without a concomitant improvement in peristalsis) and increasing physical activity. It is common clinical practice to use a stool softener along with a stimulant laxative if lifestyle modifications are inadequate.⁵⁹ If these measures are ineffective, osmotic agents can be added.

If these conventional measures fail, a peripherally acting mu-opioid receptor antagonist such as methylnaltrexone or naloxegol should be considered.

Methylnaltrexone

Methylnaltrexone^60,61 is a peripherally acting mu receptor antagonist with a rapid onset of action. It does not cross the blood-brain barrier, as it contains a methyl group. It was approved by the FDA in 2008 to treat opioid-induced constipation in adults with advanced illnesses when other approaches are ineffective.

Adverse effects. Although the mu receptor antagonist alvimopan had been shown to be associated with cardiovascular events hypothesized to be a consequence of opioid withdrawal, methylnaltrexone has been deemed to have a safe cardiovascular profile without any potential effects on platelets, corrected QT interval, metabolism, heart rate, or blood pressure.⁶¹ Side effects include abdominal pain, nausea, diarrhea, hot flashes, tremor, and chills.

Contraindications. Methylnaltrexone is contraindicated in patients with structural diseases of the gastrointestinal tract, ie, peptic ulcer disease, inflammatory bowel disease, diverticulitis, stomach or intestinal cancer) since it can increase the risk of perforation.

Dosing is 1 dose subcutaneously every other day, as needed, and no more than 1 dose in a 24-hour period. Dosage is based on weight: 0.15 mg/kg/dose for patients weighing less than 38 kg or more than 114 kg; 8 mg for those weighing 38 to 62 kg; and 12 mg for those weighing 62 to 114 kg.⁶²

Naloxegol

Naloxegol, FDA-approved for treating opioid-induced constipation in 2014, consists of naloxone conjugated with polyethylene glycol, which prevents it from crossing the blood-brain barrier and diminishing the central effects of opioid-induced analgesia. Unlike methylnaltrexone, which is given by subcutaneous injection, naloxegol is taken orally.

Adverse effects reported in clinical trials^63,64 were abdominal pain, diarrhea, nausea, headache, and flatulence. No clinically relevant association with QT and corrected QT interval prolongation or cardiac repolarization was noted.⁶⁴

Dosing is 25 mg by mouth once daily, which can be decreased to 12.5 mg if the initial dose is difficult to tolerate. It should be taken on an empty stomach at least 1 hour before the first meal of the day or 2 hours after the meal. In patients with renal impairment (creatinine clearance < 60 mL/min), the dose is 12.5 mg once daily.⁶⁵

CONSTIPATION-PREDOMINANT IRRITABLE BOWEL SYNDROME

Irritable bowel syndrome is the reason for 3.1 million office visits and 59 million prescriptions in the United States every year, with patients equally distributed between diarrhea-predominant, constipation-predominant, and mixed subtypes.⁶⁶

To be diagnosed with constipation-predominant irritable bowel syndrome, patients must meet the Rome IV criteria, more than 25% of bowel movements should have Bristol stool form types 1 or 2, and less than 25% of bowel movements should have Bristol stool form types 6 or 7. In practice, patients reporting that their bowel movements are usually constipated often suffices to make the diagnosis.⁵

Osmotic laxatives are often tried first, but despite improving stool frequency and consistency, they have little efficacy in satisfying complaints of bloating or abdominal pain in patients with constipation-predominant irritable syndrome.⁶⁷ Stimulant laxatives have not yet been tested in clinical trials. Lubiprostone and linaclotide are FDA-approved for this condition; in women, lubiprostone is approved only for those over age 18.

Antidepressant therapy

Patients often derive additional benefit from treatment with antidepressants. A meta-analysis demonstrated a number needed to treat of 4 for selective serotonin reuptake inhibitors and tricyclic antidepressants in managing abdominal pain associated with irritable bowel syndrome.⁶⁸ The major limiting factor is usually adverse effects of these drugs.

For constipation-predominant irritable bowel syndrome, selective serotonin reuptake inhibitors are preferred over tricyclics because of their additional prokinetic properties. Starting at a low dose and titrating upward slowly avoids potential adverse effects.

Cognitive behavioral therapy has also been beneficial in treating irritable bowel syndrome.⁶⁹

Adjunctive therapies

Adjunctive therapies including peppermint oil, probiotics (eg, Lactobacillus, Bifidobacterium), and acupuncture have also shown promise in managing irritable bowel syndrome, but more data are needed on the use of these therapies for constipation-predominant irritable bowel syndrome before any definite conclusions can be drawn.⁷⁰ Other emerging pharmacologic therapies are plecanatide (discussed earlier) and tenapanor.

Peppermint oil is an antispasmodic that inhibits calcium channels, leading to relaxation of smooth muscles in the gastrointestinal tract. Different dosages and treatment durations have been studied—450 to 900 mg daily in 2 to 3 divided doses over 1 to 3 months.^71,72 The most common adverse effect reported was gastroesophageal reflux, related in part to the oil’s relaxing effect on the lower esophageal sphincter. Observation of this led to the development of enteric-coated preparations that have the potential to bypass the upper gastrointestinal tract.⁷³

Tenapanor inhibits the sodium-hydrogen exchanger 3 channel (a regulator of sodium and water uptake in intestinal lumen), which in turn leads to a higher sodium level in the entire gastrointestinal tract (whereas linaclotide’s action is limited to the duodenum and jejunem), resulting in more fluid volume and increased luminal transit.⁷⁴ It was found effective in a phase 2 clinical trial,⁷⁵ and the most effective dose was 50 mg twice daily.

Since tenapanor is minimally absorbed, it has few side effects, the major ones being diarrhea (11.2% vs 0% with placebo) and urinary tract infection (5.6% vs 4.4% with placebo).⁷⁵ Further study is needed to confirm these findings.

Tenapanor also has the advantage of inhibiting luminal phosphorus absorption. This has led to exploration of its use as a phosphate binder in patients with end-stage renal disease.

DYSSYNERGIC DEFECATION AND ANORECTAL BIOFEEDBACK

According to the Rome IV criteria,⁵ dyssynergic defecation is present if the criteria for chronic constipation are met, if a dyssynergic pattern of defecation is confirmed by manometry, imaging, or electromyography, and if 1 or more of the following are present: inability to expel an artificial stool (a 50-mL water-filled balloon) within 1 minute, prolonged colonic transit time, inability to evacuate, or 50% or more retention of barium during defecography.⁵

Even though biofeedback has been controversial as a treatment for dyssynergic defecation because of conflicting results in older studies,⁷⁶ 3 trials have shown it to be better than placebo, laxatives, and muscle relaxants, with symptomatic improvement in 70% of patients.^77–79

Biofeedback therapy involves an instrument-based auditory or visual tool (using electromyographic sensors or anorectal manometry) to help patients coordinate abdominal, rectal, puborectalis, and anal sphincter muscles and produce a propulsive force using their abdominal muscles to achieve complete evacuation. Important components of this therapy include:

Proper evacuation positioning (brace-pump technique, which involves sitting on the toilet leaning forward with forearms resting on thighs, shoulders relaxed, and feet placed on a small footstool

Breathing relaxation and training exercises during defecation (no straining, keeping a normal pattern of breathing, and avoiding holding the breath while defecating)

Use of the abdominal muscles by pushing the abdomen forward, along with relaxation of the anal sphincter.⁸⁰

The anorectal feedback program usually consists of 6 weekly sessions of 45 to 60 minutes each. Limitations of this therapy include unavailability, lack of trained therapists, lack of insurance coverage, and inapplicability to certain patient groups, such as those with dementia or learning disabilities.

SURGERY FOR CHRONIC CONSTIPATION

Surgery for constipation is reserved for patients who continue to have symptoms despite optimal medical therapy.

Total abdominal colectomy and ileorectal anastomosis

Total abdominal colectomy with ileorectal anastomosis is a surgical option for medically intractable slow-transit constipation. Before considering surgery, complete diagnostic testing should be done, including colonic manometry and documentation of whether the patient also has outlet dysfunction. 

Even though it has shown excellent outcomes and satisfaction rates as high as 100% in patients with pure slow-transit constipation,^81–83 results in older studies in patients with mixed disorders (eg, slow-transit constipation with features of outlet dysfunction) were less predictable.⁸⁴ More recent studies have reported comparable long-term morbidity and postoperative satisfaction rates in those with pure slow-transit constipation and those with a mixed disorder, indicating that careful patient selection is likely the key to a favorable outcome.⁸⁵

Partial colectomies based on segmental colon transit time measurements can also be considered in some patients.⁸⁶

Stapled transanal resection

Stapled transanal resection involves circumferential transanal stapling of the redundant rectal mucosa. It is an option for patients with defecatory disorders, specifically large rectoceles and rectal intussusception not amenable to therapy with pelvic floor retraining exercises.⁸⁷

The efficacy of this procedure in controlling symptoms and improving quality of life is around 77% to 81% at 12 months, though complication rates as high as 46% and disappointing long-term outcomes have been a deterrent to its widespread acceptance in the United States.^88–91

Chronic constipation has a variety of possible causes and mechanisms. Although traditional conservative treatments are still valid and first-line, if these fail, clinicians can choose from a growing list of new treatments, tailored to the cause in the individual patient.

This article discusses how defecation works (or doesn’t), the types of chronic constipation, the available diagnostic tools, and traditional and newer treatments, including some still in development.

THE EPIDEMIOLOGY OF CONSTIPATION

Chronic constipation is one of the most common gastrointestinal disorders, affecting about 15% of all adults and 30% of those over the age of 60.¹ It can be a primary disorder or secondary to other factors.

Constipation is more prevalent in women and in institutionalized elderly people.² It is associated with lower socioeconomic status, depression, less self-reported physical activity, certain medications, and stressful life events.³ Given its high prevalence and its impact on quality of life, it is also associated with significant utilization of healthcare resources.⁴

Constipation defined by Rome IV criteria

Physicians and patients may disagree about what constitutes constipation. Physicians primarily regard it as infrequent bowel movements, while patients tend to have a broader definition. According to the Rome IV criteria,⁵ chronic constipation is defined by the presence of the following for at least 3 months (with symptom onset at least 6 months prior to diagnosis):

 (1) Two or more of the following for more than 25% of defecations:

• Straining
• Lumpy or hard stools
• Sensation of incomplete evacuation
• Sensation of anorectal obstruction or blockage
• Manual maneuvers to facilitate evacuation
• Fewer than 3 spontaneous bowel movements per week.

 (2) Loose stools are rarely present without the use of laxatives.

 (3) The patient does not meet the criteria for diagnosis of irritable bowel syndrome.

DEFECATION IS COMPLEX

Defecation begins when the rectum fills with stool, causing relaxation of the internal anal sphincter and the urge to defecate. The external anal sphincter, which is under voluntary control, can then either contract to delay defecation or relax to allow the stool to be expelled.⁶

Colonic muscles propel stool toward the rectum in repetitive localized contractions that help mix and promote absorption of the content, and larger coordinated (high-amplitude propagating) contractions that, in healthy individuals, move the stool forward from the proximal to the distal colon multiple times daily. These contractions usually occur in the morning and are accentuated by gastric distention from food and the resulting gastrocolic reflex.

Serotonin (5-HT) is released by enterochromaffin cells in response to distention of the gut wall. It mediates peristaltic movements of the gastrointestinal tract by binding to receptors (especially 5-HT4), stimulating release of neurotransmitters such as acetylcholine, causing smooth-muscle contraction behind the luminal contents and propelling them forward.

PRIMARY CONSTIPATION DISORDERS

The American Gastroenterological Association⁷ classifies constipation into 3 groups on the basis of colonic transit time and anorectal function:

Normal-transit constipation

Stool normally takes 20 to 72 hours to pass through the colon, with transit time affected by diet, drugs, level of physical activity, and emotional status.⁸

Normal-transit constipation is the most common type of constipation. The term is sometimes used interchangeably with constipation-predominant irritable bowel syndrome, but the latter is a distinct entity characterized by abdominal pain relieved by defecation as the primary symptom, as well as having occasional loose stools. These 2 conditions can be hard to tell apart, especially if the patient cannot describe the symptoms precisely.

Slow-transit constipation

Slow-transit constipation—also called delayed-transit constipation, colonoparesis, colonic inertia, and pseudo-obstruction—is defined as prolonged stool transit in the colon, ie, for more than 5 days.⁹ It can be the result of colonic smooth muscle dysfunction, compromised colonic neural pathways, or both, leading to slow colon peristalsis.

Factors that can affect colonic motility such as opioid use and hypothyroidism should be carefully considered in these patients. Opioids are notorious for causing constipation by decreasing bowel tone and contractility and thereby increasing colonic transit time. They also tighten up the anal sphincters, resulting in decreased rectal evacuation.¹⁰

Outlet dysfunction

Outlet dysfunction, also called pelvic floor dysfunction or defecatory disorder, is associated with incomplete rectal evacuation. It can be a consequence of weak rectal expulsion forces (slow colonic transit, rectal hyposensitivity), functional resistance to rectal evacuation (high anal resting pressure, anismus, incomplete relaxation of the anal sphincter, dyssynergic defecation), or structural outlet obstruction (excessive perineal descent, rectoceles, rectal intussusception). About 50% of patients with outlet dysfunction have concurrent slow-transit constipation.

Dyssynergic defecation is the most common outlet dysfunction disorder, accounting for about half of the cases referred to tertiary centers. It is defined as a paradoxical elevation in anal sphincter tone or less than 20% relaxation of the resting anal sphincter pressure with weak abdominal and pelvic propulsive forces.¹¹ Anorectal biofeedback is a therapeutic option for dyssynergic defecation, as we discuss later in this article.

SECONDARY CONSTIPATION

Constipation can be secondary to several conditions and factors (Table 1), including:

• Neurologic disorders that affect gastrointestinal motility (eg, Hirschsprung disease, Parkinson disease, multiple sclerosis, spinal cord injury, stroke, spinal or ganglionic  tumor, hypothyroidism, amyloidosis, diabetes mellitus, hypercalcemia)
• Drugs used to treat neurologic disorders
• Mechanical obstruction
• Diet (eg, low fiber, decreased fluid intake).

EVALUATION OF CONSTIPATION

It is crucial for physicians to efficiently use the available diagnostic tools for constipation to tailor the treatment to the patient.

Evaluation of chronic constipation begins with a thorough history and physical examination to rule out secondary constipation (Figure 1). Red flags such as unintentional weight loss, blood in the stool, rectal pain, fever, and iron-deficiency anemia should prompt referral for colonoscopy to evaluate for malignancy, colitis, or other potential colonic abnormalities.¹²

A detailed perineal and rectal examination can help diagnose defecatory disorders and should include evaluation of the resting anal tone and the sphincter during simulated evacuation.

Laboratory tests of thyroid function, electrolytes, and a complete blood cell count should be ordered if clinically indicated.¹³

Further tests

Further diagnostic tests can be considered if symptoms persist despite conservative treatment or if a defecatory disorder is suspected. These include anorectal manometry, colonic transit studies, defecography, and colonic manometry.

Anorectal manometry and the rectal balloon expulsion test are usually done first because of their high sensitivity (88%) and specificity (89%) for defecatory disorders.¹⁴ These tests measure the function of the internal and external anal sphincters at rest and with straining and assess rectal sensitivity and compliance. Anorectal manometry is also used in biofeedback therapy in patients with dyssynergic defecation.¹⁵

Colonic transit time can be measured if anorectal manometry and the balloon expulsion test are normal. The study uses radiopaque markers, radioisotopes, or wireless motility capsules to confirm slow-transit constipation and to identify areas of delayed transit in the colon.¹⁶

Defecography is usually the next step in diagnosis if anorectal manometry and balloon expulsion tests are inconclusive or if an anatomic abnormality of the pelvic floor is suspected. It can be done with a variety of techniques. Barium defecography can identify anatomic defects, scintigraphy can quantify evacuation of artificial stools, and magnetic resonance defecography visualizes anatomic landmarks to assess pelvic floor motion without exposing the patient to radiation.^17,18

Colonic manometry is most useful in patients with refractory slow-transit constipation and can identify patients with isolated colonic motor dysfunction with no pelvic floor dysfunction who may benefit from subtotal colectomy and end-ileostomy.⁷

TRADITIONAL TREATMENTS STILL THE MAINSTAY

Nonpharmacologic treatments are the first-line options for patients with normal-transit and slow-transit constipation and should precede diagnostic testing. Lifestyle modifications and dietary changes (Table 2) aim to augment the known factors that stimulate the gastrocolic reflex and increase intestinal motility by high-amplitude propagated contractions.

Increasing physical activity increases intestinal gas clearance, decreases bloating, and lessens constipation.^19,20

Toilet training is an integral part of lifestyle modifications.²¹

Diet. Drinking hot caffeinated beverages, eating breakfast within an hour of waking up, and consuming fiber in the morning (25–30 g of fiber daily) have traditionally been recommended as the first-line measures for chronic constipation. Dehydrated patients with constipation also benefit from increasing their fluid intake.²²

LAXATIVES

Fiber (bulk-forming laxatives) for normal-transit constipation

Fiber remains a key part of the initial management of chronic constipation, as it is cheap, available, and safe. Increasing fiber intake is effective for normal-transit constipation, but patients with slow-transit constipation or refractory outlet dysfunction are less likely to benefit.²³ Other laxatives are incorporated into the regimen if first-line nonpharmacologic interventions fail (Table 3).

Bulk-forming laxatives include insoluble fiber (wheat bran) and soluble fiber (psyllium, methylcellulose, inulin, calcium polycarbo­phil). Insoluble fiber, though often used, has little impact on symptoms of chronic constipation after 1 month of use, and up to 60% of patients report adverse effects from it.²⁴ On the other hand, clinical trials have shown that soluble fiber such as psyllium facilitates defecation and improves functional bowel symptoms in patients with normal-transit constipation.²⁵

Patients should be instructed to increase their dietary fiber intake gradually to avoid adverse effects and should be told to expect significant symptomatic improvement only after a few weeks. They should also be informed that increasing dietary fiber intake can cause bloating but that the bloating is temporary. If it continues, a different fiber can be tried.

Osmotic laxatives

Osmotic laxatives are often employed as a first- line laxative treatment option for patients with constipation. They draw water into the lumen by osmosis, helping to soften stool and speed intestinal transit. They include macrogols (inert polymers of ethylene glycol), nonabsorbable carbohydrates (lactulose, sorbitol), magnesium products, and sodium phosphate products.

Polyethylene glycol, the most studied osmotic laxative, has been shown to maintain therapeutic efficacy for up to 2 years, though it is not generally used this long.²⁶ A meta-analysis of 10 randomized clinical trials found it to be superior to lactulose in improving stool consistency and frequency, and rates of adverse effects were similar to those with placebo.²⁷

Lactulose and sorbitol are semisynthetic disaccharides that are not absorbed from the gastrointestinal tract. Apart from the osmotic effect of the disaccharide, these sugars are metabolized by colonic bacteria to acetic acid and other short-chain fatty acids, resulting in acidification of the stool, which exerts an osmotic effect in the colonic lumen.

Lactulose and sorbitol were shown to have similar efficacy in increasing the frequency of bowel movements in a small study, though patients taking lactulose had a higher rate of nausea.²⁸

The usual recommended dose is 15 to 30 mL once or twice daily.

Adverse effects include gas, bloating, and abdominal distention (due to fermentation by colonic bacteria) and can limit long-term use.

Magnesium citrate and magnesium hydroxide are strong osmotic laxatives, but so far no clinical trial has been done to assess their efficacy in constipation. Although the risk of hypermagnesemia is low with magnesium-based products, this group of laxatives is generally avoided in patients with renal or cardiac disease.²⁹

Sodium phosphate enemas (Fleet enemas) are used for bowel cleansing before certain procedures but have only limited use in constipation because of potential adverse effects such as hyperphosphatemia, hypocalcemia, and the rarer but more serious complication of acute phosphate nephropathy.³⁰

Stimulant laxatives for short-term use only

Stimulant laxatives include glycerin, bisacodyl, senna, and sodium picosulfate. Sodium piosulfate and bisacodyl have been validated for treatment of chronic constipation for up to 4 weeks.^31–33

Stimulant laxative suppositories should be used 30 minutes after meals to augment the physiologic gastrocolic reflex.

As more evidence is available for osmotic laxatives such as polyethylene glycol, they tend to be preferred over stimulant agents, especially for long-term use. Clinicians have traditionally hesitated to prescribe stimulant laxatives for long-term use, as they were thought to damage the enteric nervous system.³⁴ Although more recent studies have not shown this potential effect,³⁵ more research is warranted on the use of stimulant laxatives for longer than 4 weeks.

Stool softeners enhance the interaction of stool and water, leading to softer stool and easier evacuation. Docusate sodium and docusate calcium are thought to facilitate the mixing of aqueous and fatty substances, thereby softening the stool.

However, there is little evidence to support the use of docusate for constipation in hospitalized adults or in ambulatory care. A recent review reported that docusate was no better than placebo in diminishing symptoms of constipation.³⁶

INTESTINAL SECRETAGOGUES

The secretagogues include lubiprostone, linaclotide, and plecanatide. These medications are preferred therapy for patients with normal- or slow-transit constipation once conservative therapies have failed. Even though there is no current consensus, lifestyle measures and conservative treatment options should be tried for about 8 weeks.

Lubiprostone and linaclotide are approved by the US Food and Drug Administration (FDA) for both constipation and constipation-predominant irritable bowel syndrome. They activate chloride channels on the apical surface of enterocytes, increasing intestinal secretion of chloride, which in turn increases luminal sodium efflux to maintain electroneutrality, leading to secretion of water into the intestinal lumen. This eventually facilitates intestinal transit and increases the passage of stool.

Lubiprostone

Lubiprostone, a prostaglandin E1 derivative, is approved for treating chronic constipation, constipation-predominant irritable bowel syndrome in women, and opioid-induced constipation in patients with chronic noncancer pain.

Adverse effects in clinical trials were nausea (up to 30%) and headache.^37,38

Linaclotide

Linaclotide, a minimally absorbed 14-amino acid peptide, increases intestinal secretion of chloride and bicarbonate, increasing intestinal fluid and promoting intestinal transit.³⁹ It also decreases the firing rate of the visceral afferent pain fibers and helps reduce visceral pain, especially in patients with constipation-predominant irritable bowel syndrome.⁴⁰ It is approved for chronic constipation and constipation-predominant irritable bowel syndrome.^41–43

Dosage starts at 145 μg/day for chronic constipation, and can be titrated up to 290 μg if there is no response or if a diagnosis of constipation-predominant irritable bowel syndrome is under consideration. Linaclotide should be taken 30 to 60 minutes before breakfast to reduce the likelihood of diarrhea.⁴⁴

Adverse effects. Diarrhea led to treatment discontinuation in 4.5% of patients in one study.⁴²

Plecanatide

Plecanatide is a guanylate cyclase-c agonist with a mode of action similar to that of linaclotide. It was recently approved by the FDA for chronic idiopathic constipation in adults. The recommended dose is 3 mg once daily.

Data from phase 2 trials in chronic constipation showed improvement in straining, abdominal discomfort, and stool frequency after 14 days of treatment.⁴⁵

A phase 3 trial showed that plecanatide was more effective than placebo when used for 12 weeks in 951 patients with chronic constipation (P = .009).⁴⁶ The most common adverse effect reported was diarrhea.

SEROTONIN RECEPTOR AGONISTS

Activation of serotonin 5-HT4 receptors in the gut leads to release of acetylcholine, which in turn induces mucosal secretion by activating submucosal neurons and increasing gut motility.⁴⁷

Two 5-HT4 receptor agonists were withdrawn from the market (cisapride in 2000 and tegaserod in 2007) due to serious cardiovascular adverse events (fatal arrhythmias, heart attacks, and strokes) resulting from their affinity for hERG-K+ cardiac channels.  

The newer agents prucalopride,⁴⁸ velusetrag, and naronapride are highly selective 5-HT4 agonists with low affinity for hERG-K+ receptors and do not have proarrhythmic properties, based on extensive assessment in clinical trials.

Prucalopride

Prucalopride has been shown to accelerate gastrointestinal and colonic transit in patients with chronic constipation, with improvement in bowel movements, symptoms of chronic constipation, and quality of life.^49–52

Adverse effects reported with its use have been headache, nausea, abdominal pain, and cramps.

Prucalopride is approved in Europe and Canada for chronic constipation in women but is not yet approved in the United States.

Dosage is 2 mg orally once daily. Caution is advised in elderly patients, in whom the preferred maximum dose is 1 mg daily, as there are only limited data available on the safety of this medication in the elderly.

Velusetrag

Velusetrag has been shown to increase colonic motility and improve symptoms of chronic constipation. In a phase 2 trial,⁵³ the most effective dose was 15 mg once daily. Higher doses were associated with a higher incidence of adverse effects such as diarrhea, headache, nausea, and vomiting.

Naronapride

Naronapride (ATI-7505) is in phase 2 trials for chronic constipation. Reported adverse effects were headache, diarrhea, nausea, and vomiting.⁵⁴

BILE SALT ABSORPTION INHIBITORS

Bile acids exert prosecretory and prokinetic effects by increasing colonic secretion of water and electrolytes through the activation of adenylate cyclase. This happens as a result of their deconjugation after passage into the colon.

Elobixibat is an ileal bile acid transporter inhibitor that prevents absorption of nonconjugated bile salts in the distal ileum. It has few side effects because its systemic absorption is minimal. Phase 3 trials are under way. Dosage is 5 to 20 mg daily. Adverse effects are few because systemic absorption is minimal, but include abdominal pain and diarrhea.^55,56

Opioids cause constipation by binding to mu receptors in the enteric nervous system. Activation of these receptors decreases bowel tone and contractility, which increases transit time. Stimulation of these receptors also increases anal sphincter tone, resulting in decreased rectal evacuation.⁵⁷

Though underrecognized, opioid-induced constipation affects 40% of patients who take these drugs for nonmalignant pain and 90% of those taking them for cancer pain. Patients with this condition were found to take more time off work and feel more impaired in their domestic and work-related obligations than patients who did not develop constipation with use of opioids.⁵⁸

Initial management of opioid-induced constipation includes increasing intake of fluids and dietary fiber (fiber alone can worsen abdominal pain in this condition by increasing stool bulk without a concomitant improvement in peristalsis) and increasing physical activity. It is common clinical practice to use a stool softener along with a stimulant laxative if lifestyle modifications are inadequate.⁵⁹ If these measures are ineffective, osmotic agents can be added.

If these conventional measures fail, a peripherally acting mu-opioid receptor antagonist such as methylnaltrexone or naloxegol should be considered.

Methylnaltrexone

Methylnaltrexone^60,61 is a peripherally acting mu receptor antagonist with a rapid onset of action. It does not cross the blood-brain barrier, as it contains a methyl group. It was approved by the FDA in 2008 to treat opioid-induced constipation in adults with advanced illnesses when other approaches are ineffective.

Adverse effects. Although the mu receptor antagonist alvimopan had been shown to be associated with cardiovascular events hypothesized to be a consequence of opioid withdrawal, methylnaltrexone has been deemed to have a safe cardiovascular profile without any potential effects on platelets, corrected QT interval, metabolism, heart rate, or blood pressure.⁶¹ Side effects include abdominal pain, nausea, diarrhea, hot flashes, tremor, and chills.

Contraindications. Methylnaltrexone is contraindicated in patients with structural diseases of the gastrointestinal tract, ie, peptic ulcer disease, inflammatory bowel disease, diverticulitis, stomach or intestinal cancer) since it can increase the risk of perforation.

Dosing is 1 dose subcutaneously every other day, as needed, and no more than 1 dose in a 24-hour period. Dosage is based on weight: 0.15 mg/kg/dose for patients weighing less than 38 kg or more than 114 kg; 8 mg for those weighing 38 to 62 kg; and 12 mg for those weighing 62 to 114 kg.⁶²

Naloxegol

Naloxegol, FDA-approved for treating opioid-induced constipation in 2014, consists of naloxone conjugated with polyethylene glycol, which prevents it from crossing the blood-brain barrier and diminishing the central effects of opioid-induced analgesia. Unlike methylnaltrexone, which is given by subcutaneous injection, naloxegol is taken orally.

Adverse effects reported in clinical trials^63,64 were abdominal pain, diarrhea, nausea, headache, and flatulence. No clinically relevant association with QT and corrected QT interval prolongation or cardiac repolarization was noted.⁶⁴

Dosing is 25 mg by mouth once daily, which can be decreased to 12.5 mg if the initial dose is difficult to tolerate. It should be taken on an empty stomach at least 1 hour before the first meal of the day or 2 hours after the meal. In patients with renal impairment (creatinine clearance < 60 mL/min), the dose is 12.5 mg once daily.⁶⁵

CONSTIPATION-PREDOMINANT IRRITABLE BOWEL SYNDROME

Irritable bowel syndrome is the reason for 3.1 million office visits and 59 million prescriptions in the United States every year, with patients equally distributed between diarrhea-predominant, constipation-predominant, and mixed subtypes.⁶⁶

To be diagnosed with constipation-predominant irritable bowel syndrome, patients must meet the Rome IV criteria, more than 25% of bowel movements should have Bristol stool form types 1 or 2, and less than 25% of bowel movements should have Bristol stool form types 6 or 7. In practice, patients reporting that their bowel movements are usually constipated often suffices to make the diagnosis.⁵

Osmotic laxatives are often tried first, but despite improving stool frequency and consistency, they have little efficacy in satisfying complaints of bloating or abdominal pain in patients with constipation-predominant irritable syndrome.⁶⁷ Stimulant laxatives have not yet been tested in clinical trials. Lubiprostone and linaclotide are FDA-approved for this condition; in women, lubiprostone is approved only for those over age 18.

Antidepressant therapy

Patients often derive additional benefit from treatment with antidepressants. A meta-analysis demonstrated a number needed to treat of 4 for selective serotonin reuptake inhibitors and tricyclic antidepressants in managing abdominal pain associated with irritable bowel syndrome.⁶⁸ The major limiting factor is usually adverse effects of these drugs.

For constipation-predominant irritable bowel syndrome, selective serotonin reuptake inhibitors are preferred over tricyclics because of their additional prokinetic properties. Starting at a low dose and titrating upward slowly avoids potential adverse effects.

Cognitive behavioral therapy has also been beneficial in treating irritable bowel syndrome.⁶⁹

Adjunctive therapies

Adjunctive therapies including peppermint oil, probiotics (eg, Lactobacillus, Bifidobacterium), and acupuncture have also shown promise in managing irritable bowel syndrome, but more data are needed on the use of these therapies for constipation-predominant irritable bowel syndrome before any definite conclusions can be drawn.⁷⁰ Other emerging pharmacologic therapies are plecanatide (discussed earlier) and tenapanor.

Peppermint oil is an antispasmodic that inhibits calcium channels, leading to relaxation of smooth muscles in the gastrointestinal tract. Different dosages and treatment durations have been studied—450 to 900 mg daily in 2 to 3 divided doses over 1 to 3 months.^71,72 The most common adverse effect reported was gastroesophageal reflux, related in part to the oil’s relaxing effect on the lower esophageal sphincter. Observation of this led to the development of enteric-coated preparations that have the potential to bypass the upper gastrointestinal tract.⁷³

Tenapanor inhibits the sodium-hydrogen exchanger 3 channel (a regulator of sodium and water uptake in intestinal lumen), which in turn leads to a higher sodium level in the entire gastrointestinal tract (whereas linaclotide’s action is limited to the duodenum and jejunem), resulting in more fluid volume and increased luminal transit.⁷⁴ It was found effective in a phase 2 clinical trial,⁷⁵ and the most effective dose was 50 mg twice daily.

Since tenapanor is minimally absorbed, it has few side effects, the major ones being diarrhea (11.2% vs 0% with placebo) and urinary tract infection (5.6% vs 4.4% with placebo).⁷⁵ Further study is needed to confirm these findings.

Tenapanor also has the advantage of inhibiting luminal phosphorus absorption. This has led to exploration of its use as a phosphate binder in patients with end-stage renal disease.

DYSSYNERGIC DEFECATION AND ANORECTAL BIOFEEDBACK

According to the Rome IV criteria,⁵ dyssynergic defecation is present if the criteria for chronic constipation are met, if a dyssynergic pattern of defecation is confirmed by manometry, imaging, or electromyography, and if 1 or more of the following are present: inability to expel an artificial stool (a 50-mL water-filled balloon) within 1 minute, prolonged colonic transit time, inability to evacuate, or 50% or more retention of barium during defecography.⁵

Even though biofeedback has been controversial as a treatment for dyssynergic defecation because of conflicting results in older studies,⁷⁶ 3 trials have shown it to be better than placebo, laxatives, and muscle relaxants, with symptomatic improvement in 70% of patients.^77–79

Biofeedback therapy involves an instrument-based auditory or visual tool (using electromyographic sensors or anorectal manometry) to help patients coordinate abdominal, rectal, puborectalis, and anal sphincter muscles and produce a propulsive force using their abdominal muscles to achieve complete evacuation. Important components of this therapy include:

Proper evacuation positioning (brace-pump technique, which involves sitting on the toilet leaning forward with forearms resting on thighs, shoulders relaxed, and feet placed on a small footstool

Breathing relaxation and training exercises during defecation (no straining, keeping a normal pattern of breathing, and avoiding holding the breath while defecating)

Use of the abdominal muscles by pushing the abdomen forward, along with relaxation of the anal sphincter.⁸⁰

The anorectal feedback program usually consists of 6 weekly sessions of 45 to 60 minutes each. Limitations of this therapy include unavailability, lack of trained therapists, lack of insurance coverage, and inapplicability to certain patient groups, such as those with dementia or learning disabilities.

SURGERY FOR CHRONIC CONSTIPATION

Surgery for constipation is reserved for patients who continue to have symptoms despite optimal medical therapy.

Total abdominal colectomy and ileorectal anastomosis

Total abdominal colectomy with ileorectal anastomosis is a surgical option for medically intractable slow-transit constipation. Before considering surgery, complete diagnostic testing should be done, including colonic manometry and documentation of whether the patient also has outlet dysfunction. 

Even though it has shown excellent outcomes and satisfaction rates as high as 100% in patients with pure slow-transit constipation,^81–83 results in older studies in patients with mixed disorders (eg, slow-transit constipation with features of outlet dysfunction) were less predictable.⁸⁴ More recent studies have reported comparable long-term morbidity and postoperative satisfaction rates in those with pure slow-transit constipation and those with a mixed disorder, indicating that careful patient selection is likely the key to a favorable outcome.⁸⁵

Partial colectomies based on segmental colon transit time measurements can also be considered in some patients.⁸⁶

Stapled transanal resection

Stapled transanal resection involves circumferential transanal stapling of the redundant rectal mucosa. It is an option for patients with defecatory disorders, specifically large rectoceles and rectal intussusception not amenable to therapy with pelvic floor retraining exercises.⁸⁷

The efficacy of this procedure in controlling symptoms and improving quality of life is around 77% to 81% at 12 months, though complication rates as high as 46% and disappointing long-term outcomes have been a deterrent to its widespread acceptance in the United States.^88–91

Chronic constipation has a variety of possible causes and mechanisms. Although traditional conservative treatments are still valid and first-line, if these fail, clinicians can choose from a growing list of new treatments, tailored to the cause in the individual patient.

This article discusses how defecation works (or doesn’t), the types of chronic constipation, the available diagnostic tools, and traditional and newer treatments, including some still in development.

THE EPIDEMIOLOGY OF CONSTIPATION

Chronic constipation is one of the most common gastrointestinal disorders, affecting about 15% of all adults and 30% of those over the age of 60.¹ It can be a primary disorder or secondary to other factors.

Constipation is more prevalent in women and in institutionalized elderly people.² It is associated with lower socioeconomic status, depression, less self-reported physical activity, certain medications, and stressful life events.³ Given its high prevalence and its impact on quality of life, it is also associated with significant utilization of healthcare resources.⁴

Constipation defined by Rome IV criteria

Physicians and patients may disagree about what constitutes constipation. Physicians primarily regard it as infrequent bowel movements, while patients tend to have a broader definition. According to the Rome IV criteria,⁵ chronic constipation is defined by the presence of the following for at least 3 months (with symptom onset at least 6 months prior to diagnosis):

 (1) Two or more of the following for more than 25% of defecations:

• Straining
• Lumpy or hard stools
• Sensation of incomplete evacuation
• Sensation of anorectal obstruction or blockage
• Manual maneuvers to facilitate evacuation
• Fewer than 3 spontaneous bowel movements per week.

 (2) Loose stools are rarely present without the use of laxatives.

 (3) The patient does not meet the criteria for diagnosis of irritable bowel syndrome.

DEFECATION IS COMPLEX

Defecation begins when the rectum fills with stool, causing relaxation of the internal anal sphincter and the urge to defecate. The external anal sphincter, which is under voluntary control, can then either contract to delay defecation or relax to allow the stool to be expelled.⁶

Colonic muscles propel stool toward the rectum in repetitive localized contractions that help mix and promote absorption of the content, and larger coordinated (high-amplitude propagating) contractions that, in healthy individuals, move the stool forward from the proximal to the distal colon multiple times daily. These contractions usually occur in the morning and are accentuated by gastric distention from food and the resulting gastrocolic reflex.

Serotonin (5-HT) is released by enterochromaffin cells in response to distention of the gut wall. It mediates peristaltic movements of the gastrointestinal tract by binding to receptors (especially 5-HT4), stimulating release of neurotransmitters such as acetylcholine, causing smooth-muscle contraction behind the luminal contents and propelling them forward.

PRIMARY CONSTIPATION DISORDERS

The American Gastroenterological Association⁷ classifies constipation into 3 groups on the basis of colonic transit time and anorectal function:

Normal-transit constipation

Stool normally takes 20 to 72 hours to pass through the colon, with transit time affected by diet, drugs, level of physical activity, and emotional status.⁸

Normal-transit constipation is the most common type of constipation. The term is sometimes used interchangeably with constipation-predominant irritable bowel syndrome, but the latter is a distinct entity characterized by abdominal pain relieved by defecation as the primary symptom, as well as having occasional loose stools. These 2 conditions can be hard to tell apart, especially if the patient cannot describe the symptoms precisely.

Slow-transit constipation

Slow-transit constipation—also called delayed-transit constipation, colonoparesis, colonic inertia, and pseudo-obstruction—is defined as prolonged stool transit in the colon, ie, for more than 5 days.⁹ It can be the result of colonic smooth muscle dysfunction, compromised colonic neural pathways, or both, leading to slow colon peristalsis.

Factors that can affect colonic motility such as opioid use and hypothyroidism should be carefully considered in these patients. Opioids are notorious for causing constipation by decreasing bowel tone and contractility and thereby increasing colonic transit time. They also tighten up the anal sphincters, resulting in decreased rectal evacuation.¹⁰

Outlet dysfunction

Outlet dysfunction, also called pelvic floor dysfunction or defecatory disorder, is associated with incomplete rectal evacuation. It can be a consequence of weak rectal expulsion forces (slow colonic transit, rectal hyposensitivity), functional resistance to rectal evacuation (high anal resting pressure, anismus, incomplete relaxation of the anal sphincter, dyssynergic defecation), or structural outlet obstruction (excessive perineal descent, rectoceles, rectal intussusception). About 50% of patients with outlet dysfunction have concurrent slow-transit constipation.

Dyssynergic defecation is the most common outlet dysfunction disorder, accounting for about half of the cases referred to tertiary centers. It is defined as a paradoxical elevation in anal sphincter tone or less than 20% relaxation of the resting anal sphincter pressure with weak abdominal and pelvic propulsive forces.¹¹ Anorectal biofeedback is a therapeutic option for dyssynergic defecation, as we discuss later in this article.

SECONDARY CONSTIPATION

Constipation can be secondary to several conditions and factors (Table 1), including:

• Neurologic disorders that affect gastrointestinal motility (eg, Hirschsprung disease, Parkinson disease, multiple sclerosis, spinal cord injury, stroke, spinal or ganglionic  tumor, hypothyroidism, amyloidosis, diabetes mellitus, hypercalcemia)
• Drugs used to treat neurologic disorders
• Mechanical obstruction
• Diet (eg, low fiber, decreased fluid intake).

EVALUATION OF CONSTIPATION

It is crucial for physicians to efficiently use the available diagnostic tools for constipation to tailor the treatment to the patient.

Evaluation of chronic constipation begins with a thorough history and physical examination to rule out secondary constipation (Figure 1). Red flags such as unintentional weight loss, blood in the stool, rectal pain, fever, and iron-deficiency anemia should prompt referral for colonoscopy to evaluate for malignancy, colitis, or other potential colonic abnormalities.¹²

A detailed perineal and rectal examination can help diagnose defecatory disorders and should include evaluation of the resting anal tone and the sphincter during simulated evacuation.

Laboratory tests of thyroid function, electrolytes, and a complete blood cell count should be ordered if clinically indicated.¹³

Further tests

Further diagnostic tests can be considered if symptoms persist despite conservative treatment or if a defecatory disorder is suspected. These include anorectal manometry, colonic transit studies, defecography, and colonic manometry.

Anorectal manometry and the rectal balloon expulsion test are usually done first because of their high sensitivity (88%) and specificity (89%) for defecatory disorders.¹⁴ These tests measure the function of the internal and external anal sphincters at rest and with straining and assess rectal sensitivity and compliance. Anorectal manometry is also used in biofeedback therapy in patients with dyssynergic defecation.¹⁵

Colonic transit time can be measured if anorectal manometry and the balloon expulsion test are normal. The study uses radiopaque markers, radioisotopes, or wireless motility capsules to confirm slow-transit constipation and to identify areas of delayed transit in the colon.¹⁶

Defecography is usually the next step in diagnosis if anorectal manometry and balloon expulsion tests are inconclusive or if an anatomic abnormality of the pelvic floor is suspected. It can be done with a variety of techniques. Barium defecography can identify anatomic defects, scintigraphy can quantify evacuation of artificial stools, and magnetic resonance defecography visualizes anatomic landmarks to assess pelvic floor motion without exposing the patient to radiation.^17,18

Colonic manometry is most useful in patients with refractory slow-transit constipation and can identify patients with isolated colonic motor dysfunction with no pelvic floor dysfunction who may benefit from subtotal colectomy and end-ileostomy.⁷

TRADITIONAL TREATMENTS STILL THE MAINSTAY

Nonpharmacologic treatments are the first-line options for patients with normal-transit and slow-transit constipation and should precede diagnostic testing. Lifestyle modifications and dietary changes (Table 2) aim to augment the known factors that stimulate the gastrocolic reflex and increase intestinal motility by high-amplitude propagated contractions.

Increasing physical activity increases intestinal gas clearance, decreases bloating, and lessens constipation.^19,20

Toilet training is an integral part of lifestyle modifications.²¹

Diet. Drinking hot caffeinated beverages, eating breakfast within an hour of waking up, and consuming fiber in the morning (25–30 g of fiber daily) have traditionally been recommended as the first-line measures for chronic constipation. Dehydrated patients with constipation also benefit from increasing their fluid intake.²²

LAXATIVES

Fiber (bulk-forming laxatives) for normal-transit constipation

Fiber remains a key part of the initial management of chronic constipation, as it is cheap, available, and safe. Increasing fiber intake is effective for normal-transit constipation, but patients with slow-transit constipation or refractory outlet dysfunction are less likely to benefit.²³ Other laxatives are incorporated into the regimen if first-line nonpharmacologic interventions fail (Table 3).

Bulk-forming laxatives include insoluble fiber (wheat bran) and soluble fiber (psyllium, methylcellulose, inulin, calcium polycarbo­phil). Insoluble fiber, though often used, has little impact on symptoms of chronic constipation after 1 month of use, and up to 60% of patients report adverse effects from it.²⁴ On the other hand, clinical trials have shown that soluble fiber such as psyllium facilitates defecation and improves functional bowel symptoms in patients with normal-transit constipation.²⁵

Patients should be instructed to increase their dietary fiber intake gradually to avoid adverse effects and should be told to expect significant symptomatic improvement only after a few weeks. They should also be informed that increasing dietary fiber intake can cause bloating but that the bloating is temporary. If it continues, a different fiber can be tried.

Osmotic laxatives

Osmotic laxatives are often employed as a first- line laxative treatment option for patients with constipation. They draw water into the lumen by osmosis, helping to soften stool and speed intestinal transit. They include macrogols (inert polymers of ethylene glycol), nonabsorbable carbohydrates (lactulose, sorbitol), magnesium products, and sodium phosphate products.

Polyethylene glycol, the most studied osmotic laxative, has been shown to maintain therapeutic efficacy for up to 2 years, though it is not generally used this long.²⁶ A meta-analysis of 10 randomized clinical trials found it to be superior to lactulose in improving stool consistency and frequency, and rates of adverse effects were similar to those with placebo.²⁷

Lactulose and sorbitol are semisynthetic disaccharides that are not absorbed from the gastrointestinal tract. Apart from the osmotic effect of the disaccharide, these sugars are metabolized by colonic bacteria to acetic acid and other short-chain fatty acids, resulting in acidification of the stool, which exerts an osmotic effect in the colonic lumen.

Lactulose and sorbitol were shown to have similar efficacy in increasing the frequency of bowel movements in a small study, though patients taking lactulose had a higher rate of nausea.²⁸

The usual recommended dose is 15 to 30 mL once or twice daily.

Adverse effects include gas, bloating, and abdominal distention (due to fermentation by colonic bacteria) and can limit long-term use.

Magnesium citrate and magnesium hydroxide are strong osmotic laxatives, but so far no clinical trial has been done to assess their efficacy in constipation. Although the risk of hypermagnesemia is low with magnesium-based products, this group of laxatives is generally avoided in patients with renal or cardiac disease.²⁹

Sodium phosphate enemas (Fleet enemas) are used for bowel cleansing before certain procedures but have only limited use in constipation because of potential adverse effects such as hyperphosphatemia, hypocalcemia, and the rarer but more serious complication of acute phosphate nephropathy.³⁰

Stimulant laxatives for short-term use only

Stimulant laxatives include glycerin, bisacodyl, senna, and sodium picosulfate. Sodium piosulfate and bisacodyl have been validated for treatment of chronic constipation for up to 4 weeks.^31–33

Stimulant laxative suppositories should be used 30 minutes after meals to augment the physiologic gastrocolic reflex.

As more evidence is available for osmotic laxatives such as polyethylene glycol, they tend to be preferred over stimulant agents, especially for long-term use. Clinicians have traditionally hesitated to prescribe stimulant laxatives for long-term use, as they were thought to damage the enteric nervous system.³⁴ Although more recent studies have not shown this potential effect,³⁵ more research is warranted on the use of stimulant laxatives for longer than 4 weeks.

Stool softeners enhance the interaction of stool and water, leading to softer stool and easier evacuation. Docusate sodium and docusate calcium are thought to facilitate the mixing of aqueous and fatty substances, thereby softening the stool.

However, there is little evidence to support the use of docusate for constipation in hospitalized adults or in ambulatory care. A recent review reported that docusate was no better than placebo in diminishing symptoms of constipation.³⁶

INTESTINAL SECRETAGOGUES

The secretagogues include lubiprostone, linaclotide, and plecanatide. These medications are preferred therapy for patients with normal- or slow-transit constipation once conservative therapies have failed. Even though there is no current consensus, lifestyle measures and conservative treatment options should be tried for about 8 weeks.

Lubiprostone and linaclotide are approved by the US Food and Drug Administration (FDA) for both constipation and constipation-predominant irritable bowel syndrome. They activate chloride channels on the apical surface of enterocytes, increasing intestinal secretion of chloride, which in turn increases luminal sodium efflux to maintain electroneutrality, leading to secretion of water into the intestinal lumen. This eventually facilitates intestinal transit and increases the passage of stool.

Lubiprostone

Lubiprostone, a prostaglandin E1 derivative, is approved for treating chronic constipation, constipation-predominant irritable bowel syndrome in women, and opioid-induced constipation in patients with chronic noncancer pain.

Adverse effects in clinical trials were nausea (up to 30%) and headache.^37,38

Linaclotide

Linaclotide, a minimally absorbed 14-amino acid peptide, increases intestinal secretion of chloride and bicarbonate, increasing intestinal fluid and promoting intestinal transit.³⁹ It also decreases the firing rate of the visceral afferent pain fibers and helps reduce visceral pain, especially in patients with constipation-predominant irritable bowel syndrome.⁴⁰ It is approved for chronic constipation and constipation-predominant irritable bowel syndrome.^41–43

Dosage starts at 145 μg/day for chronic constipation, and can be titrated up to 290 μg if there is no response or if a diagnosis of constipation-predominant irritable bowel syndrome is under consideration. Linaclotide should be taken 30 to 60 minutes before breakfast to reduce the likelihood of diarrhea.⁴⁴

Adverse effects. Diarrhea led to treatment discontinuation in 4.5% of patients in one study.⁴²

Plecanatide

Plecanatide is a guanylate cyclase-c agonist with a mode of action similar to that of linaclotide. It was recently approved by the FDA for chronic idiopathic constipation in adults. The recommended dose is 3 mg once daily.

Data from phase 2 trials in chronic constipation showed improvement in straining, abdominal discomfort, and stool frequency after 14 days of treatment.⁴⁵

A phase 3 trial showed that plecanatide was more effective than placebo when used for 12 weeks in 951 patients with chronic constipation (P = .009).⁴⁶ The most common adverse effect reported was diarrhea.

SEROTONIN RECEPTOR AGONISTS

Activation of serotonin 5-HT4 receptors in the gut leads to release of acetylcholine, which in turn induces mucosal secretion by activating submucosal neurons and increasing gut motility.⁴⁷

Two 5-HT4 receptor agonists were withdrawn from the market (cisapride in 2000 and tegaserod in 2007) due to serious cardiovascular adverse events (fatal arrhythmias, heart attacks, and strokes) resulting from their affinity for hERG-K+ cardiac channels.  

The newer agents prucalopride,⁴⁸ velusetrag, and naronapride are highly selective 5-HT4 agonists with low affinity for hERG-K+ receptors and do not have proarrhythmic properties, based on extensive assessment in clinical trials.

Prucalopride

Prucalopride has been shown to accelerate gastrointestinal and colonic transit in patients with chronic constipation, with improvement in bowel movements, symptoms of chronic constipation, and quality of life.^49–52

Adverse effects reported with its use have been headache, nausea, abdominal pain, and cramps.

Prucalopride is approved in Europe and Canada for chronic constipation in women but is not yet approved in the United States.

Dosage is 2 mg orally once daily. Caution is advised in elderly patients, in whom the preferred maximum dose is 1 mg daily, as there are only limited data available on the safety of this medication in the elderly.

Velusetrag

Velusetrag has been shown to increase colonic motility and improve symptoms of chronic constipation. In a phase 2 trial,⁵³ the most effective dose was 15 mg once daily. Higher doses were associated with a higher incidence of adverse effects such as diarrhea, headache, nausea, and vomiting.

Naronapride

Naronapride (ATI-7505) is in phase 2 trials for chronic constipation. Reported adverse effects were headache, diarrhea, nausea, and vomiting.⁵⁴

BILE SALT ABSORPTION INHIBITORS

Bile acids exert prosecretory and prokinetic effects by increasing colonic secretion of water and electrolytes through the activation of adenylate cyclase. This happens as a result of their deconjugation after passage into the colon.

Elobixibat is an ileal bile acid transporter inhibitor that prevents absorption of nonconjugated bile salts in the distal ileum. It has few side effects because its systemic absorption is minimal. Phase 3 trials are under way. Dosage is 5 to 20 mg daily. Adverse effects are few because systemic absorption is minimal, but include abdominal pain and diarrhea.^55,56

Opioids cause constipation by binding to mu receptors in the enteric nervous system. Activation of these receptors decreases bowel tone and contractility, which increases transit time. Stimulation of these receptors also increases anal sphincter tone, resulting in decreased rectal evacuation.⁵⁷

Though underrecognized, opioid-induced constipation affects 40% of patients who take these drugs for nonmalignant pain and 90% of those taking them for cancer pain. Patients with this condition were found to take more time off work and feel more impaired in their domestic and work-related obligations than patients who did not develop constipation with use of opioids.⁵⁸

Initial management of opioid-induced constipation includes increasing intake of fluids and dietary fiber (fiber alone can worsen abdominal pain in this condition by increasing stool bulk without a concomitant improvement in peristalsis) and increasing physical activity. It is common clinical practice to use a stool softener along with a stimulant laxative if lifestyle modifications are inadequate.⁵⁹ If these measures are ineffective, osmotic agents can be added.

If these conventional measures fail, a peripherally acting mu-opioid receptor antagonist such as methylnaltrexone or naloxegol should be considered.

Methylnaltrexone

Methylnaltrexone^60,61 is a peripherally acting mu receptor antagonist with a rapid onset of action. It does not cross the blood-brain barrier, as it contains a methyl group. It was approved by the FDA in 2008 to treat opioid-induced constipation in adults with advanced illnesses when other approaches are ineffective.

Adverse effects. Although the mu receptor antagonist alvimopan had been shown to be associated with cardiovascular events hypothesized to be a consequence of opioid withdrawal, methylnaltrexone has been deemed to have a safe cardiovascular profile without any potential effects on platelets, corrected QT interval, metabolism, heart rate, or blood pressure.⁶¹ Side effects include abdominal pain, nausea, diarrhea, hot flashes, tremor, and chills.

Contraindications. Methylnaltrexone is contraindicated in patients with structural diseases of the gastrointestinal tract, ie, peptic ulcer disease, inflammatory bowel disease, diverticulitis, stomach or intestinal cancer) since it can increase the risk of perforation.

Dosing is 1 dose subcutaneously every other day, as needed, and no more than 1 dose in a 24-hour period. Dosage is based on weight: 0.15 mg/kg/dose for patients weighing less than 38 kg or more than 114 kg; 8 mg for those weighing 38 to 62 kg; and 12 mg for those weighing 62 to 114 kg.⁶²

Naloxegol

Naloxegol, FDA-approved for treating opioid-induced constipation in 2014, consists of naloxone conjugated with polyethylene glycol, which prevents it from crossing the blood-brain barrier and diminishing the central effects of opioid-induced analgesia. Unlike methylnaltrexone, which is given by subcutaneous injection, naloxegol is taken orally.

Adverse effects reported in clinical trials^63,64 were abdominal pain, diarrhea, nausea, headache, and flatulence. No clinically relevant association with QT and corrected QT interval prolongation or cardiac repolarization was noted.⁶⁴

Dosing is 25 mg by mouth once daily, which can be decreased to 12.5 mg if the initial dose is difficult to tolerate. It should be taken on an empty stomach at least 1 hour before the first meal of the day or 2 hours after the meal. In patients with renal impairment (creatinine clearance < 60 mL/min), the dose is 12.5 mg once daily.⁶⁵

CONSTIPATION-PREDOMINANT IRRITABLE BOWEL SYNDROME

Irritable bowel syndrome is the reason for 3.1 million office visits and 59 million prescriptions in the United States every year, with patients equally distributed between diarrhea-predominant, constipation-predominant, and mixed subtypes.⁶⁶

To be diagnosed with constipation-predominant irritable bowel syndrome, patients must meet the Rome IV criteria, more than 25% of bowel movements should have Bristol stool form types 1 or 2, and less than 25% of bowel movements should have Bristol stool form types 6 or 7. In practice, patients reporting that their bowel movements are usually constipated often suffices to make the diagnosis.⁵

Osmotic laxatives are often tried first, but despite improving stool frequency and consistency, they have little efficacy in satisfying complaints of bloating or abdominal pain in patients with constipation-predominant irritable syndrome.⁶⁷ Stimulant laxatives have not yet been tested in clinical trials. Lubiprostone and linaclotide are FDA-approved for this condition; in women, lubiprostone is approved only for those over age 18.

Antidepressant therapy

Patients often derive additional benefit from treatment with antidepressants. A meta-analysis demonstrated a number needed to treat of 4 for selective serotonin reuptake inhibitors and tricyclic antidepressants in managing abdominal pain associated with irritable bowel syndrome.⁶⁸ The major limiting factor is usually adverse effects of these drugs.

For constipation-predominant irritable bowel syndrome, selective serotonin reuptake inhibitors are preferred over tricyclics because of their additional prokinetic properties. Starting at a low dose and titrating upward slowly avoids potential adverse effects.

Cognitive behavioral therapy has also been beneficial in treating irritable bowel syndrome.⁶⁹

Adjunctive therapies

Adjunctive therapies including peppermint oil, probiotics (eg, Lactobacillus, Bifidobacterium), and acupuncture have also shown promise in managing irritable bowel syndrome, but more data are needed on the use of these therapies for constipation-predominant irritable bowel syndrome before any definite conclusions can be drawn.⁷⁰ Other emerging pharmacologic therapies are plecanatide (discussed earlier) and tenapanor.

Peppermint oil is an antispasmodic that inhibits calcium channels, leading to relaxation of smooth muscles in the gastrointestinal tract. Different dosages and treatment durations have been studied—450 to 900 mg daily in 2 to 3 divided doses over 1 to 3 months.^71,72 The most common adverse effect reported was gastroesophageal reflux, related in part to the oil’s relaxing effect on the lower esophageal sphincter. Observation of this led to the development of enteric-coated preparations that have the potential to bypass the upper gastrointestinal tract.⁷³

Tenapanor inhibits the sodium-hydrogen exchanger 3 channel (a regulator of sodium and water uptake in intestinal lumen), which in turn leads to a higher sodium level in the entire gastrointestinal tract (whereas linaclotide’s action is limited to the duodenum and jejunem), resulting in more fluid volume and increased luminal transit.⁷⁴ It was found effective in a phase 2 clinical trial,⁷⁵ and the most effective dose was 50 mg twice daily.

Since tenapanor is minimally absorbed, it has few side effects, the major ones being diarrhea (11.2% vs 0% with placebo) and urinary tract infection (5.6% vs 4.4% with placebo).⁷⁵ Further study is needed to confirm these findings.

Tenapanor also has the advantage of inhibiting luminal phosphorus absorption. This has led to exploration of its use as a phosphate binder in patients with end-stage renal disease.

DYSSYNERGIC DEFECATION AND ANORECTAL BIOFEEDBACK

According to the Rome IV criteria,⁵ dyssynergic defecation is present if the criteria for chronic constipation are met, if a dyssynergic pattern of defecation is confirmed by manometry, imaging, or electromyography, and if 1 or more of the following are present: inability to expel an artificial stool (a 50-mL water-filled balloon) within 1 minute, prolonged colonic transit time, inability to evacuate, or 50% or more retention of barium during defecography.⁵

Even though biofeedback has been controversial as a treatment for dyssynergic defecation because of conflicting results in older studies,⁷⁶ 3 trials have shown it to be better than placebo, laxatives, and muscle relaxants, with symptomatic improvement in 70% of patients.^77–79

Biofeedback therapy involves an instrument-based auditory or visual tool (using electromyographic sensors or anorectal manometry) to help patients coordinate abdominal, rectal, puborectalis, and anal sphincter muscles and produce a propulsive force using their abdominal muscles to achieve complete evacuation. Important components of this therapy include:

Proper evacuation positioning (brace-pump technique, which involves sitting on the toilet leaning forward with forearms resting on thighs, shoulders relaxed, and feet placed on a small footstool

Breathing relaxation and training exercises during defecation (no straining, keeping a normal pattern of breathing, and avoiding holding the breath while defecating)

Use of the abdominal muscles by pushing the abdomen forward, along with relaxation of the anal sphincter.⁸⁰

The anorectal feedback program usually consists of 6 weekly sessions of 45 to 60 minutes each. Limitations of this therapy include unavailability, lack of trained therapists, lack of insurance coverage, and inapplicability to certain patient groups, such as those with dementia or learning disabilities.

SURGERY FOR CHRONIC CONSTIPATION

Surgery for constipation is reserved for patients who continue to have symptoms despite optimal medical therapy.

Total abdominal colectomy and ileorectal anastomosis

Total abdominal colectomy with ileorectal anastomosis is a surgical option for medically intractable slow-transit constipation. Before considering surgery, complete diagnostic testing should be done, including colonic manometry and documentation of whether the patient also has outlet dysfunction. 

Even though it has shown excellent outcomes and satisfaction rates as high as 100% in patients with pure slow-transit constipation,^81–83 results in older studies in patients with mixed disorders (eg, slow-transit constipation with features of outlet dysfunction) were less predictable.⁸⁴ More recent studies have reported comparable long-term morbidity and postoperative satisfaction rates in those with pure slow-transit constipation and those with a mixed disorder, indicating that careful patient selection is likely the key to a favorable outcome.⁸⁵

Partial colectomies based on segmental colon transit time measurements can also be considered in some patients.⁸⁶

Stapled transanal resection

Stapled transanal resection involves circumferential transanal stapling of the redundant rectal mucosa. It is an option for patients with defecatory disorders, specifically large rectoceles and rectal intussusception not amenable to therapy with pelvic floor retraining exercises.⁸⁷

The efficacy of this procedure in controlling symptoms and improving quality of life is around 77% to 81% at 12 months, though complication rates as high as 46% and disappointing long-term outcomes have been a deterrent to its widespread acceptance in the United States.^88–91

To the Editor: I enjoyed the article “Serotonin syndrome: Preventing, recognizing, and treating it.”¹ I am a relatively new internal medicine physician, out of residency only 1 year, and sadly I felt that the psychiatric training I received was minimal at best. Therefore, I was very excited to read more about serotonin syndrome since such a large percentage of my patients are on selective serotonin reuptake inhibitors (SSRIs) or serotonin-norepinephrine reuptake inhibitors.

Could you speak to the time frame it takes for serotonin syndrome to develop? For instance, if someone is taking an SSRI and develops a terrible yeast infection, would 3 doses of fluconazole be enough to tip the scales? Or as-needed sumatriptan, with some ondansetron for migraine? The problem I have is that patients often require short doses of many medications that can interact, and I routinely sigh, briefly explain the possibility of serotonin syndrome, and then click through the flashing red warning signs on the electronic medical record and send patients out with their meds—though in honesty I do not know the likelihood of developing even mild symptoms of serotonin syndrome with short courses of interacting medications.

To the Editor: I enjoyed the article “Serotonin syndrome: Preventing, recognizing, and treating it.”¹ I am a relatively new internal medicine physician, out of residency only 1 year, and sadly I felt that the psychiatric training I received was minimal at best. Therefore, I was very excited to read more about serotonin syndrome since such a large percentage of my patients are on selective serotonin reuptake inhibitors (SSRIs) or serotonin-norepinephrine reuptake inhibitors.

Could you speak to the time frame it takes for serotonin syndrome to develop? For instance, if someone is taking an SSRI and develops a terrible yeast infection, would 3 doses of fluconazole be enough to tip the scales? Or as-needed sumatriptan, with some ondansetron for migraine? The problem I have is that patients often require short doses of many medications that can interact, and I routinely sigh, briefly explain the possibility of serotonin syndrome, and then click through the flashing red warning signs on the electronic medical record and send patients out with their meds—though in honesty I do not know the likelihood of developing even mild symptoms of serotonin syndrome with short courses of interacting medications.

To the Editor: I enjoyed the article “Serotonin syndrome: Preventing, recognizing, and treating it.”¹ I am a relatively new internal medicine physician, out of residency only 1 year, and sadly I felt that the psychiatric training I received was minimal at best. Therefore, I was very excited to read more about serotonin syndrome since such a large percentage of my patients are on selective serotonin reuptake inhibitors (SSRIs) or serotonin-norepinephrine reuptake inhibitors.

Could you speak to the time frame it takes for serotonin syndrome to develop? For instance, if someone is taking an SSRI and develops a terrible yeast infection, would 3 doses of fluconazole be enough to tip the scales? Or as-needed sumatriptan, with some ondansetron for migraine? The problem I have is that patients often require short doses of many medications that can interact, and I routinely sigh, briefly explain the possibility of serotonin syndrome, and then click through the flashing red warning signs on the electronic medical record and send patients out with their meds—though in honesty I do not know the likelihood of developing even mild symptoms of serotonin syndrome with short courses of interacting medications.

In Reply: The questions posed by Dr. Rose reflect critical issues primary care physicians encounter when prescribing medications for patients who are taking serotonergic agents. “Switching strategies” have been described for starting or discontinuing serotonergic antidepressants.¹ Options range from conservative exchanges requiring 5 half-lives between discontinuation of 1 antidepressant and initiation of another vs a direct cross-taper exchange. Decisions regarding specific patients should take into account previous adverse effects from serotonergic medications and half-lives of discontinued antidepressants. To our knowledge, switching strategies have not been validated and are based on expert opinion. Scenarios are complicated further if patients have already been prescribed 2 or more antidepressants and 1 medication is exchanged or dose-adjusted while another is added. With this degree of complexity, we recommend referral to a psychiatrist.

Dr. Rose’s questions on prescribing nonpsychiatric serotonergic drugs concurrently with antidepressants broaches a topic with even less evidence. Some data exist about nonpsychiatric serotonergic drugs given in combination with triptans. Soldin et al² reviewed the US Food and Drug Administration’s Adverse Event Reporting System and discovered 38 cases of serotonin syndrome in patients using triptans. Eleven of these patients were using triptans without concomitant antidepressants. Though definitive evidence is lacking for safe prescribing practice with triptans, the authors noted that most cases of triptan-induced serotonin toxicity occur within hours of triptan ingestion.²

The evidence on the risk of serotonin syndrome with other medications is limited to case reports. In regard to linezolid, a review suggested that when linezolid was administered to a patient on long-term citalopram, a prolonged serotonin syndrome was precipitated, which is not an issue with other antidepressants.³ The World Health Organization has issued warnings for serotonin toxicity with ondansetron and other 5-HT3 receptor antagonists based on case reports.^4,5 No data are available for the appropriate prescribing of 5-HT3 antagonists with antidepressants. A review of cases suggests a link between fluconazole and severe serotonin toxicity in patients taking citalopram; however, no prescribing guidelines have been established for fluconazole either.⁶

Dr. Rose asks important clinical questions, but evidence-based answers are not available.  We can only recommend that patients be advised to report symptoms immediately after starting any medication associated with serotonin syndrome. For patients on multiple antidepressants, psychiatric assistance is advised. An observational cohort study of patients using antidepressants while exposed to other suspect drugs may better delineate effects of several pharmaceuticals on the serotonergic axis. Only then may safe prescribing practices be validated with evidence.

In Reply: The questions posed by Dr. Rose reflect critical issues primary care physicians encounter when prescribing medications for patients who are taking serotonergic agents. “Switching strategies” have been described for starting or discontinuing serotonergic antidepressants.¹ Options range from conservative exchanges requiring 5 half-lives between discontinuation of 1 antidepressant and initiation of another vs a direct cross-taper exchange. Decisions regarding specific patients should take into account previous adverse effects from serotonergic medications and half-lives of discontinued antidepressants. To our knowledge, switching strategies have not been validated and are based on expert opinion. Scenarios are complicated further if patients have already been prescribed 2 or more antidepressants and 1 medication is exchanged or dose-adjusted while another is added. With this degree of complexity, we recommend referral to a psychiatrist.

Dr. Rose’s questions on prescribing nonpsychiatric serotonergic drugs concurrently with antidepressants broaches a topic with even less evidence. Some data exist about nonpsychiatric serotonergic drugs given in combination with triptans. Soldin et al² reviewed the US Food and Drug Administration’s Adverse Event Reporting System and discovered 38 cases of serotonin syndrome in patients using triptans. Eleven of these patients were using triptans without concomitant antidepressants. Though definitive evidence is lacking for safe prescribing practice with triptans, the authors noted that most cases of triptan-induced serotonin toxicity occur within hours of triptan ingestion.²

The evidence on the risk of serotonin syndrome with other medications is limited to case reports. In regard to linezolid, a review suggested that when linezolid was administered to a patient on long-term citalopram, a prolonged serotonin syndrome was precipitated, which is not an issue with other antidepressants.³ The World Health Organization has issued warnings for serotonin toxicity with ondansetron and other 5-HT3 receptor antagonists based on case reports.^4,5 No data are available for the appropriate prescribing of 5-HT3 antagonists with antidepressants. A review of cases suggests a link between fluconazole and severe serotonin toxicity in patients taking citalopram; however, no prescribing guidelines have been established for fluconazole either.⁶

Dr. Rose asks important clinical questions, but evidence-based answers are not available.  We can only recommend that patients be advised to report symptoms immediately after starting any medication associated with serotonin syndrome. For patients on multiple antidepressants, psychiatric assistance is advised. An observational cohort study of patients using antidepressants while exposed to other suspect drugs may better delineate effects of several pharmaceuticals on the serotonergic axis. Only then may safe prescribing practices be validated with evidence.

In Reply: The questions posed by Dr. Rose reflect critical issues primary care physicians encounter when prescribing medications for patients who are taking serotonergic agents. “Switching strategies” have been described for starting or discontinuing serotonergic antidepressants.¹ Options range from conservative exchanges requiring 5 half-lives between discontinuation of 1 antidepressant and initiation of another vs a direct cross-taper exchange. Decisions regarding specific patients should take into account previous adverse effects from serotonergic medications and half-lives of discontinued antidepressants. To our knowledge, switching strategies have not been validated and are based on expert opinion. Scenarios are complicated further if patients have already been prescribed 2 or more antidepressants and 1 medication is exchanged or dose-adjusted while another is added. With this degree of complexity, we recommend referral to a psychiatrist.

Dr. Rose’s questions on prescribing nonpsychiatric serotonergic drugs concurrently with antidepressants broaches a topic with even less evidence. Some data exist about nonpsychiatric serotonergic drugs given in combination with triptans. Soldin et al² reviewed the US Food and Drug Administration’s Adverse Event Reporting System and discovered 38 cases of serotonin syndrome in patients using triptans. Eleven of these patients were using triptans without concomitant antidepressants. Though definitive evidence is lacking for safe prescribing practice with triptans, the authors noted that most cases of triptan-induced serotonin toxicity occur within hours of triptan ingestion.²

The evidence on the risk of serotonin syndrome with other medications is limited to case reports. In regard to linezolid, a review suggested that when linezolid was administered to a patient on long-term citalopram, a prolonged serotonin syndrome was precipitated, which is not an issue with other antidepressants.³ The World Health Organization has issued warnings for serotonin toxicity with ondansetron and other 5-HT3 receptor antagonists based on case reports.^4,5 No data are available for the appropriate prescribing of 5-HT3 antagonists with antidepressants. A review of cases suggests a link between fluconazole and severe serotonin toxicity in patients taking citalopram; however, no prescribing guidelines have been established for fluconazole either.⁶

Dr. Rose asks important clinical questions, but evidence-based answers are not available.  We can only recommend that patients be advised to report symptoms immediately after starting any medication associated with serotonin syndrome. For patients on multiple antidepressants, psychiatric assistance is advised. An observational cohort study of patients using antidepressants while exposed to other suspect drugs may better delineate effects of several pharmaceuticals on the serotonergic axis. Only then may safe prescribing practices be validated with evidence.

Frail elderly patients are at high risk of adverse clinical outcomes, including those due to polypharmacy. Several groups tackle “deprescribing” by developing lists of medications that are potentially inappropriate for the elderly, such as the Beers or STOPP/START criteria.^1–4

In contrast, our group (the Palliative and Therapeutic Harmonization [PATH] program and the Dalhousie Academic Detailing Service) has developed evidence-based, frailty-specific guidelines for treating hypertension⁵ and diabetes,⁶ in which we advocate less-stringent treatment targets and tapering or discontinuing medications, as needed.

The PATH program⁷ is a clinical approach that prioritizes the consideration of frailty when making treatment decisions. The Dalhousie Academic Detailing Service collaborates with the Nova Scotia Health Authority to research and develop evidence-informed educational messages about the treatment of common medical conditions.

Here, we address lipid-lowering therapy in this population.

CONSIDERING FRAILTY

Frailty is defined in several ways. The Fried model^8,9 identifies frailty when 3 of the following characteristics are present: unintentional weight loss, exhaustion, muscle weakness, slow walking speed, or low levels of activity. The Clinical Frailty Scale^10,11 and the Frailty Assessment for Care-planning Tool (FACT)⁵ use deficits in cognition, function, and mobility to define frailty. According to these scales, people are considered severely frail when they require assistance with basic activities of daily living (such as bathing or dressing), owing to cognitive or physical deficits from any cause.

In reviewing the evidence, we consider five questions:

• What is the quality of the evidence? (Up to 48% of clinical practice guideline recommendations may be based on low-level evidence or expert opinion.¹²)
• How did the study population compare with the frail?
• Are study outcomes and potential benefits clinically relevant to those who are frail?
• How long did it take for the clinical benefit of a treatment to become apparent, and are the frail elderly likely to live that long?
• Have the harms of treatment been sufficiently considered?

WHAT IS THE QUALITY OF THE EVIDENCE?

We found no studies that specifically evaluated the benefit of lipid-lowering for severely frail older adults. Therefore, we examined randomized controlled trials that enrolled non-frail older adults,^13–28 subgroup analyses of randomized controlled trials,^29,30 meta-analyses that analyzed subgroups of elderly populations,^31,32 and publications describing the study designs of randomized controlled trials.^33–37

Most of the evidence comes from post hoc subgroup analyses of elderly populations. Although meta-analysis is commonly used to compare subgroups, the Cochrane handbook and others consider subgroup comparisons observational by nature.^38,39 (See Table 1 for lipid-lowering studies discussed in this article.)

Studies of statins for primary prevention of cardiovascular disease

For evidence of benefit from lipid-lowering for primary prevention (ie, to reduce the risk of cardiovascular events in patients with no known cardiovascular disease at baseline but at increased risk), we reviewed the meta-analysis conducted by the Cholesterol Treatment Trialists’ (CTT) Collaborators.³² Since this meta-analysis included the major trials that enrolled elderly patients, individual publications of post hoc, elderly subgroups were, for the most part, not examined individually. The exception to this approach was a decision to report on the PROSPER¹³ and JUPITER²⁸ trials separately, because PROSPER is the most representative of the elderly population and JUPITER reached the lowest LDL-C of primary prevention trials published to date and included a large elderly subgroup (n = 5,695).

Savarese et al⁴⁰ evaluated the benefits of statins for older adults who did not have established cardiovascular disease. We did not report on this meta-analysis, as not all of the subjects that populated the meta-analysis were representative of a typical prevention population. For instance, in the Anglo-Scandinavian Cardiac Outcomes Trial lipid-lowering arm,⁴¹ 14% of the subjects had had a previous stroke or transient ischemic attack. In the Antihypertensive and Lipid-Lowering Treatment Trial,⁴² 16% of the population had a family history of premature coronary heart disease.

In addition, all the trials in the Savarese meta-analysis were also included in the CTT meta-analysis.³² The CTT reports on baseline risk using patient-level data stratified by age and risk, which may be more relevant to the question of primary prevention for older adults, as highlighted in our review.

PROSPER (Prospective Study of Pravastatin in the Elderly at Risk),¹³ a well-conducted, double-blind, randomized controlled trial with low probability of bias, compared pravastatin 40 mg and placebo. It was the only study that specifically enrolled older adults, with prespecified analysis of primary and secondary prevention subgroups. The primary prevention subgroup accounted for 56% of the 5,084 participants.

JUPITER (Justification for the Use of Statins in Prevention)²⁸ compared rosuvastatin 20 mg and placebo in 17,802 participants. All had low-density lipoprotein cholesterol (LDL-C) levels below 3.4 mmol/L (130 mg/dL) and elevated levels of the inflammatory biomarker high-sensitivity C-reactive protein (hsCRP), ie, 2 mg/L or higher. Subsequently, Glynn et al performed a post hoc, exploratory subgroup analysis of elderly participants (N = 5,695).²⁹

The JUPITER trial had several limitations.^43,44 The planned follow-up period was 5 years, but the trial was stopped early at 1.9 years, after a statistically significant difference was detected in the primary composite outcome of reduction in all vascular events. Studies that are stopped early may exaggerate positive findings.⁴⁵

Further, JUPITER’s patients were a select group, with normal LDL-C levels, elevated hsCRP values, and without diabetes. Of 90,000 patients screened, 72,000 (80%) did not meet the inclusion criteria and were not enrolled. This high rate of exclusion limits the generalizability of study findings beyond the shortcomings of post hoc subgroup analysis.

The meta-analysis performed by the CTT Collaborators³² used individual participant data from large-scale randomized trials of lipid-modifying treatment. This analysis was specific to people at low risk of vascular disease. In a supplementary appendix, the authors described the reduction in major vascular events for each 1.0 mmol/L decrease in LDL-C in three age categories: under age 60, ages 61 to 70, and over age 70.

The authors also stratified the results by risk category and provided information about those with a risk of major vascular events of less than 20%, which would be more representative of a purer primary prevention population.

For the elderly subgroup at low risk, the CTT Collaborators³² only reported a composite of major vascular events (coronary death, nonfatal myocardial infarction [MI], ischemic stroke, or revascularization) and did not describe individual outcomes, such as prevention of coronary heart disease.

Study results are based on postrandomization findings and therefore may be observational, not experimental.⁴⁶

Studies of statins for secondary prevention of cardiovascular disease

The aim of secondary prevention is to reduce the risk of recurrent cardiovascular events in patients who already have cardiovascular disease.

To address the question of whether statins reduce cardiovascular risk, we reviewed:

PROSPER,¹³ which included a preplanned analysis of the secondary prevention population.

Afilalo et al,^31,47 who performed a meta-analysis of the elderly subgroups of nine major secondary prevention studies (19,569 patients) using published and unpublished data.

To address the question of whether statins benefit individuals with heart failure, we found two relevant studies:

GISSI-HF (Gruppo Italiano per lo Studio della Sopravvivenza nell’Insufficienza Cardiaca Heart Failure)²⁵ and CORONA (Controlled Rosuvastatin Multinational Trial in Heart Failure),²⁶ which were large, international, well-conducted randomized controlled trials that examined statin use in heart failure.

To answer the question of whether statins benefit individuals after a stroke or transient ischemic attack, we found one relevant study:

SPARCL (Stroke Prevention by Aggressive Reduction in Cholesterol Levels),²⁷ which evaluated the benefit of statins in older adults with a history of stroke or transient ischemic attack. It was a prospective, double-blind, placebo-controlled, international trial conducted at 205 centers. One to 6 months after their cerebrovascular event, patients were randomized to receive either atorvastatin 80 mg or placebo. Given the young age of patients in this trial (mean age 63), we also reviewed a post hoc subgroup analysis of the elderly patients in SPARCL (age > 65).³⁰

HOW DID THE STUDY POPULATION COMPARE WITH THOSE WHO ARE FRAIL?

Frail older adults are almost always excluded from large-scale clinical trials,⁴⁸ leading to uncertainty about whether the conclusions can be applied to those with advanced frailty.

Although age is an imperfect proxy measure of frailty,⁴⁹ we consider the age of the study population as well as their comorbidities.

Participants in the studies we reviewed were generally younger and healthier than those who are frail, with mean ages of about 75 or less (Table 1).

PROSPER was the most representative study, as it specifically enrolled older adults, albeit without frailty,¹³ and excluded people with poor cognitive function as defined by a Mini Mental State Examination score less than 24.

JUPITER enrolled a select population, as described above. The median age in the elderly subgroup was 74 (interquartile range 72–78).²⁹

The Afilalo et al³¹ meta-analysis primarily included studies of young-elderly patients, with a mean age of less than 70. PROSPER¹³ was an exception.

The GISSI-HF study,²⁵ which examined the benefit of statins in heart failure, described their study population as frail, although the mean age was only 68. Compared with those in GISSI-HF, the CORONA patients²⁶ with heart failure were older (mean age 73) and had more severe heart failure. Accordingly, it is possible that many of the CORONA participants were frail.

ARE STUDY OUTCOMES CLINICALLY RELEVANT TO THOSE WHO ARE FRAIL?

Because baseline cardiovascular risk increases with age, the elderly should, in theory, experience greater absolute benefit from lipid-lowering. However, there is uncertainty about whether this is true in practice.

Some, but not all, epidemiologic studies show a weaker relationship between cholesterol levels and cardiovascular morbidity and mortality rates in older compared to younger adults.^50,51 This may be because those with high cholesterol levels die before they get old (time-related bias), or because those with life-threatening illness may have lower cholesterol levels.⁵⁰ In addition, classic risk factors such as age, sex, systolic blood pressure, cholesterol values, diabetes, smoking, and left ventricular hypertrophy on electrocardiography may have less power to predict cardiovascular risk among older patients.⁵²

The goal of treatment in frailty is to prevent further disability or improve quality of life. Therefore, meaningful outcomes for lipid-lowering therapy should include symptomatic nonfatal MI and its associated morbidity (eg, heart failure and persistent angina) or symptomatic nonfatal stroke leading to disability. Outcomes without sustained clinical impact, such as transient ischemic attack, nondisabling stroke, or silent MI, while potentially important in other populations, are less relevant in severe frailty. Notably, in many statin studies, outcomes include asymptomatic heart disease (eg, silent MI and “suspected events”) and nondisabling stroke (eg, mild stroke, transient ischemic attack). When symptomatic outcomes are not reported separately, the impact of the reported benefit on quality of life and function is uncertain.

The outcome of all-cause mortality is generally recognized as a gold standard for determining treatment benefit. However, since advanced frailty is characterized by multiple competing causes for mortality, a reduction in all-cause mortality that is achieved by addressing a single issue in nonfrail populations may not extend to the frail.

To more fully understand the impact of lipid-lowering therapy on quality of life and function, we examined the following questions:

Do statins as primary prevention reduce symptomatic heart disease?

Outcomes for coronary heart disease from PROSPER and JUPITER are summarized in Table 2.

PROSPER. In the PROSPER primary prevention group,¹³ statin therapy did not reduce the combined outcome of coronary heart disease death and nonfatal MI.

The JUPITER trial demonstrated a statistically significant benefit for preventing MI in the elderly subpopulation (ages 70–97),²⁹ but the number needed to treat was high (211 for 2 years), with a wide confidence interval (CI) (95% CI 106–32,924). The trial did not adequately differentiate between symptomatic and asymptomatic events, making it difficult to determine outcome relevance. Also, due to the methodologic limitations of JUPITER as described above, its results should be interpreted with caution.^43,44

The CTT Collaborators³² did not report individual outcomes (eg, coronary heart disease) for the elderly low-risk subgroup and, therefore, this meta-analysis does not answer the question of whether statins reduce symptomatic heart disease in primary prevention populations.

Taken together, these findings do not provide convincing evidence that statin therapy as primary prevention reduces the incidence of symptomatic heart disease for severely frail older adults.

Do statins as secondary prevention reduce symptomatic heart disease?

Most studies defined secondary prevention narrowly as treatment for patients with established coronary artery disease. For instance, in the Afilalo et al meta-analysis,³¹ the small number of studies that included individuals with other forms of vascular disease (such as peripheral vascular disease) enrolled few participants with noncardiac conditions (eg, 29% in PROSPER¹³ and 13% in the Heart Protection Study²⁰).

Therefore, any evidence of benefit for secondary prevention demonstrated in these studies is most applicable to patients with coronary heart disease, with less certainty for those with other forms of cardiovascular disease.

In PROSPER,¹³ the secondary prevention group experienced benefit in the combined outcome of coronary heart disease death or nonfatal MI. In the treatment group, 12.7% experienced this outcome compared with 16.8% with placebo, an absolute risk reduction of 4.1% in 3 years (P = .004, number needed to treat 25, 95% CI 15–77). This measure includes coronary heart disease death, an outcome that may not be generalizable to those who are frail. In addition, the outcome of nonfatal MI includes both symptomatic and suspected events. As such, there is uncertainty whether the realized benefit is clinically relevant to frail older adults.

The Afilalo et al meta-analysis³¹ showed that the number needed to treat to prevent one nonfatal MI was 38 (95% CI 16–118) over 5 years (Table 2). However, this outcome included both symptomatic and asymptomatic (silent) events.

Based on the available data, we conclude that it is not possible to determine whether statins reduce symptomatic heart disease as secondary prevention for older adults who are frail.

Do statins reduce heart disease in combined populations?

In the combined primary and secondary population from PROSPER,¹³ pravastatin decreased the risk of nonfatal symptomatic MI from 4.3% in the placebo group to 3.4%, a relatively small reduction in absolute risk (0.9%) and not statistically significant by our chi-square calculation (P = .099).

Do statins prevent a first symptomatic stroke in people with or without preexisting cardiovascular disease?

Preventing strokes that cause functional decline is an important outcome for the frail elderly. Stroke outcomes from PROSPER,¹³ JUPITER,²⁹ and the Afilalo et al meta-analysis³¹ are summarized in Table 3.

For primary prevention:

In PROSPER (primary prevention),¹³ there was no statistically significant benefit in the combined outcome of fatal and nonfatal stroke or the single outcome of transient ischemic attack after 3.2 years.

JUPITER,²⁹ in contrast, found that rosuvastatin 20 mg reduced strokes in primary prevention, but the absolute benefit was small. In 2 years, 0.8% of the treatment group had strokes, compared with 1.4% with placebo, an absolute risk reduction of 0.6% (P = .023, number needed to treat 161, 95% CI 86–1,192).

Neither PROSPER nor JUPITER differentiated between disabling and nondisabling strokes.

For secondary prevention:

In PROSPER (secondary prevention),¹³ there was no statistically significant benefit in the combined outcome of fatal and nonfatal stroke or the single outcome of transient ischemic attack after 3.2 years.

The Afilalo et al secondary prevention meta-analysis demonstrated a 25% relative reduction in stroke (relative risk 0.75, 95% CI 0.56–0.94, number needed to treat 58, 95% CI 27–177).³¹

Notably, the stroke outcome in Afilalo included both disabling and nondisabling strokes. For example, in the Heart Protection Study,²⁰ the largest study in the Afilalo et al meta-analysis, approximately 50% of nonfatal, classifiable strokes in the overall study population (ie, both younger and older patients) were not disabling. Including disabling and nondisabling strokes in a composite outcome confounds the clinical meaningfulness of these findings in frailty, as the number needed to treat to prevent one disabling stroke cannot be calculated from the data provided.

Do statins prevent a second (symptomatic) stroke in people with a previous stroke?

SPARCL²⁷ (Table 3) examined the question of whether statins decrease the risk of recurrent ischemic stroke for patients with a prior history of stroke or transient ischemic attack. There was a statistically significant reduction in the primary composite outcome of fatal and nonfatal stroke, with 11.2% of the treatment group and 13.1% of the placebo group experiencing this outcome, an absolute risk reduction of 1.9% at 5 years (P = .03; number needed to treat 52, 95% CI 26–1,303). However, the difference in nonfatal stroke, which is the outcome of interest for frailty (since mortality has uncertain relevance), was not statistically significant (10.4% with treatment vs 11.8% with placebo, P =.11).

An exploratory subgroup analysis of SPARCL patients based on age³⁰ showed a smaller, nonsignificant reduction in the primary end point of fatal and nonfatal stroke in the group over age 65 (relative risk 0.90, 95% confidence interval 0.73–1.11, P = .33) compared with the younger group (age < 65) (relative risk 0.74, 95% CI 0.57–0.96, P = .02).

The applicability of these results to the frail elderly is uncertain, since the subgroup analysis was not powered to determine outcomes based on age stratification and there were differences between groups in characteristics such as blood pressure and smoking status. In addition, the outcome of interest, nonfatal stroke, is not provided for the elderly subgroup.

In conclusion, in both primary and secondary prevention populations, the evidence that statins reduce nonfatal, symptomatic stroke rates for older adults is uncertain.

Do statins decrease all-cause mortality for primary or secondary prevention?

Due to competing risks for death, the outcome of mortality may not be relevant to those who are frail; however, studies showed the following:

For primary prevention, there was no decrease in mortality in PROSPER¹³ or in the elderly subgroup of JUPITER.²⁹

For secondary prevention, an analysis of PROSPER trial data by Afilalo et al³¹ showed a significant 18% decrease in all-cause mortality (relative risk 0.82, 95% CI 0.69–0.98) using pravastatin 40 mg.

A decrease in all-cause mortality with statins was also reported in the pooled result of the Afilalo et al meta-analysis.³¹

What are the reported composite outcomes for primary and secondary prevention?

While we were most interested in the symptomatic outcomes described above, we recognize that the small numbers of events make it difficult to draw firm conclusions. Therefore, we also considered composite primary outcomes, even though most included multiple measures that have varying associations with disability and relevancy to frail older adults.

For primary prevention, in the PROSPER preplanned subgroup analysis,¹³ there was no statistical benefit for any outcome, including the primary composite measure. In contrast, the elderly subpopulation in the JUPITER trial²⁸ showed a treatment benefit with rosuvastatin 20 mg compared with placebo for the primary composite outcome of MI, stroke, cardiovascular death, hospitalization for unstable angina, or revascularization. The number needed to treat for 2 years was 62 (95% CI 39–148).

In the CTT meta-analysis,³² patients at all levels of baseline risk showed benefit up to age 70. However, there was no statistically significant benefit in the composite primary outcome of coronary deaths, nonfatal myocardial infarction, ischemic stroke, or revascularization in the population most representative of elderly primary prevention—those who were more than 70 years old with a 5-year baseline risk of less than 20%.

For secondary prevention, in PROSPER,¹³ the subpopulation of patients treated for secondary prevention experienced benefit in the primary composite outcome of coronary heart disease death, nonfatal MI, or fatal or nonfatal stroke, achieving a 4% absolute risk reduction with a number needed to treat of 23 (95% CI 14–81) over 3 years.

Do statins decrease disability?

PROSPER was the only study that reported on disability. Compared with placebo, pravastatin did not decrease disability in the total population as measured by basic and instrumental activities of daily living scales.

Do statins help patients with heart failure?

Neither GISSI-HF²⁵ nor CORONA²⁶ found significant benefit from rosuvastatin 10 mg, despite LDL-C lowering of 27% in GISSI-HF and 45% in CORONA.

Do ezetimibe or other nonstatin lipid-lowering agents improve outcomes?

There is no definitive evidence that ezetimibe provides clinically meaningful benefit as a single agent.

For combination therapy, the IMPROVE-IT (Improved Reduction of Outcomes: Vytorin Efficacy International Trial)⁵³ showed that adding ezetimibe 10 mg to simvastatin 40 mg after an acute coronary syndrome reduced the risk of nonfatal myocardial infarction compared with simvastatin monotherapy (event rate 12.8% vs 14.4%; hazard ratio 0.87, 95% CI 0.80–0.95; P = .002) for a population with a mean age of 64. The risk of any stroke was also reduced; strokes occurred in 4.2% of those receiving combination therapy vs 4.8% with monotherapy (hazard ratio 0.86, 95% CI 0.73–1.00, P = .05). After a median of 6 years, 42% of patients in each group had discontinued treatment. Given the very specific clinical scenario of acute coronary syndrome and the young age of the patients in this trial, we do not think that this study justifies the use of ezetimibe for severely frail older adults.

There is no evidence that other combinations (ie, a statin plus another lipid-lowering drug) improve clinical outcomes for either primary or secondary prevention in any population.⁵⁴

WILL FRAIL PATIENTS LIVE LONG ENOUGH TO BENEFIT?

It is often difficult to determine the number of years that are needed to achieve benefit, as most trials do not provide a statistical analysis of varying time frames.

The PROSPER trial¹³ lasted 3.2 years. From the Kaplan-Meier curves in PROSPER, we estimate that it took about 1.5 years to achieve a 1% absolute risk reduction and 2.5 years for a 2% absolute risk reduction in coronary heart disease death and nonfatal MI in the combined primary and secondary groups.

JUPITER²⁸ was stopped early at 1.9 years. The Afilalo et al meta-analysis³¹ was based on follow-up over 4.9 years.

IMPROVE-IT⁵³ reported event rates at 7 years. The authors note that benefit in the primary composite outcome appeared to emerge at 1 year, although no statistical support is given for this statement and divergence in the Kaplan-Meier curves is not visually apparent.

The duration of other studies ranged between 2.7 and 4.9 years (Table 1).^26–28

It has been suggested that statins should be considered for elderly patients who have a life expectancy of at least 5 years.³ However, many older adults have already been taking statins for many years, which makes it difficult to interpret the available timeframe evidence.

In a multicenter, unblinded, randomized trial,⁵⁵ statins were either stopped or continued in older adults who had a short life expectancy and a median survival of approximately 7 months. Causes of death were evenly divided between cancer and noncancer diagnoses, and 22% of the patients were cognitively impaired. Discontinuing statin therapy did not increase mortality or cardiovascular events within 60 days. Nevertheless, stopping statin therapy did not achieve noninferiority for the primary end point, the proportion of participants who died within 60 days. Statin discontinuation was associated with improved quality of life, although the study was not blinded, which could have influenced results.

HAVE THE HARMS BEEN SUFFICIENTLY CONSIDERED?

Frail older adults commonly take multiple medications and are more vulnerable to adverse events.⁵⁶

Many statins require dose reduction with severe renal impairment (creatinine clearance < 30 mL/min/1.73 m²), which would be a common consideration in severely frail older adults.

Myopathy

Myopathy, which includes myalgias and muscle weakness, is a statin-related adverse event that can impair quality of life. Myopathy typically develops within the first 6 months but can occur at any time during statin treatment.⁵⁷ When muscle-related adverse effects occur, they may affect the elderly more significantly, particularly their ability to perform activities of daily living, rise from a chair, or mobilize independently. Another concern is that older adults with dementia may not be able to accurately report muscle-related symptoms.

It is difficult to ascertain the true prevalence of myopathy, especially in advanced age and frailty. Randomized controlled trials report incidence rates of 1.5% to 5%, which is comparable to placebo.^57,58 However, inconsistent definitions of myopathy and exclusion of subjects with previous statin intolerance or adverse effects during run-in periods limit interpretability.⁵⁷ Clinical experience suggests that muscle complaints may be relatively common.^59–61

Advanced age, female sex, low body mass index, and multisystem disease are all associated with frailty and have also been described as risk factors for statin-associated muscle syndromes.⁶¹ Physiologic changes associated with frailty, such as reduced muscle strength, decreased lean body mass, impaired functional mobility, decreased reserve capacity, and altered drug metabolism may increase the risk and severity of myopathy.⁶²

Adverse cognitive events

Meta-analyses of randomized clinical trials and narrative reviews find no definitive relationship between statin therapy and adverse cognitive events.^63–67 Nevertheless, there have been case reports of memory loss associated with the use of statins, and the US Food and Drug Administration has issued a warning that statins have been associated with memory loss and confusion.⁶⁸

It may be difficult to determine whether a statin is causing or aggravating cognitive symptoms among individuals with dementia without a trial withdrawal of the drug.

OUR RECOMMENDATIONS

The recommendations below are intended for adults with severe or very severe frailty (ie, a score of 7 or 8 on the Clinical Frailty Scale¹¹ or FACT⁵ and therefore apply to most older adults living in long-term care facilities.

Primary prevention

There is no reason to prescribe or continue statins for primary prevention, as it is unlikely that they would provide benefit for outcomes that are relevant in this population.

Secondary prevention

Statin treatment is probably not necessary for secondary prevention in those with severe frailty, although there may be extenuating circumstances that justify statin use.

Heart failure

There is no reason to start or continue statins for heart failure, as there is insufficient evidence that they are effective for this indication in any population.

Ezetimibe

There is no evidence that ezetimibe reduces cardiovascular events in any population when used as monotherapy. For a select population with acute coronary syndromes, ezetimibe has a modest effect. Given the very specific clinical scenario of acute coronary syndrome, we do not think that the available evidence justifies the use of ezetimibe for severely frail older adults.

Agents other than ezetimibe combined with statins

There is no reason to start or continue other lipid-lowering drugs in conjunction with statins.

Statin dosing

As statin adverse effects have the potential to increase with advancing age and frailty, lower doses may be appropriate.⁶⁸

Adverse events

Consider stopping statins on a trial basis if there is concern regarding myopathy, drug interactions, or other adverse effects.

BOTTOM LINE: DO STATINS IMPROVE QUALITY OF LIFE OR FUNCTION?

In primary prevention for older adults, there is doubt that statins prevent cardiovascular disease and stroke-related events because the main study involving the elderly did not show a benefit in the primary prevention subgroup.¹³ Additionally, there is no conclusive evidence that statin treatment decreases mortality in primary prevention.^13,29

There is insufficient information to determine whether the frail elderly should receive statins for secondary prevention. Although there is evidence that treatment decreases measures of coronary heart disease and stroke, it is unclear whether it improves quality of life or function for those who are frail. To answer this question, we need more information about whether reported outcomes (such as stroke and MI) are associated with disability, which is not provided in many of the studies we reviewed. When disability was specifically considered in the PROSPER trial for the combined population of primary and secondary prevention, treatment with statins had no impact on basic and instrumental activities of daily living.

Some experts may not agree with our interpretation of the complex evidence presented in this article. Others may ask, “What is the harm in using statins, even if there is no definitive benefit?” However, the harms associated with statin therapy for the frail are poorly defined. In the face of these uncertainties and in the absence of definitive improvement in quality of life, we believe that “less is more” in the context of severe frailty.⁶⁹

The cost of medications should also be considered, especially in long-term care facilities, where there is an added expense of drug administration that diverts human resources away from interactions that are more congruent with respecting the lifestage of frailty.

Careful review of evidence before applying clinical practice guidelines to those who are frail should become the norm. When considering treatment of frail patients, the five questions described in this review shed light on the applicability of clinical trial evidence. Therapies that are highly effective in healthier populations may be less effective when individuals are severely frail. Accordingly, we propose that medications should only be used if they improve quality of life or function.

Frail elderly patients are at high risk of adverse clinical outcomes, including those due to polypharmacy. Several groups tackle “deprescribing” by developing lists of medications that are potentially inappropriate for the elderly, such as the Beers or STOPP/START criteria.^1–4

In contrast, our group (the Palliative and Therapeutic Harmonization [PATH] program and the Dalhousie Academic Detailing Service) has developed evidence-based, frailty-specific guidelines for treating hypertension⁵ and diabetes,⁶ in which we advocate less-stringent treatment targets and tapering or discontinuing medications, as needed.

The PATH program⁷ is a clinical approach that prioritizes the consideration of frailty when making treatment decisions. The Dalhousie Academic Detailing Service collaborates with the Nova Scotia Health Authority to research and develop evidence-informed educational messages about the treatment of common medical conditions.

Here, we address lipid-lowering therapy in this population.

CONSIDERING FRAILTY

Frailty is defined in several ways. The Fried model^8,9 identifies frailty when 3 of the following characteristics are present: unintentional weight loss, exhaustion, muscle weakness, slow walking speed, or low levels of activity. The Clinical Frailty Scale^10,11 and the Frailty Assessment for Care-planning Tool (FACT)⁵ use deficits in cognition, function, and mobility to define frailty. According to these scales, people are considered severely frail when they require assistance with basic activities of daily living (such as bathing or dressing), owing to cognitive or physical deficits from any cause.

In reviewing the evidence, we consider five questions:

• What is the quality of the evidence? (Up to 48% of clinical practice guideline recommendations may be based on low-level evidence or expert opinion.¹²)
• How did the study population compare with the frail?
• Are study outcomes and potential benefits clinically relevant to those who are frail?
• How long did it take for the clinical benefit of a treatment to become apparent, and are the frail elderly likely to live that long?
• Have the harms of treatment been sufficiently considered?

WHAT IS THE QUALITY OF THE EVIDENCE?

We found no studies that specifically evaluated the benefit of lipid-lowering for severely frail older adults. Therefore, we examined randomized controlled trials that enrolled non-frail older adults,^13–28 subgroup analyses of randomized controlled trials,^29,30 meta-analyses that analyzed subgroups of elderly populations,^31,32 and publications describing the study designs of randomized controlled trials.^33–37

Most of the evidence comes from post hoc subgroup analyses of elderly populations. Although meta-analysis is commonly used to compare subgroups, the Cochrane handbook and others consider subgroup comparisons observational by nature.^38,39 (See Table 1 for lipid-lowering studies discussed in this article.)

Studies of statins for primary prevention of cardiovascular disease

For evidence of benefit from lipid-lowering for primary prevention (ie, to reduce the risk of cardiovascular events in patients with no known cardiovascular disease at baseline but at increased risk), we reviewed the meta-analysis conducted by the Cholesterol Treatment Trialists’ (CTT) Collaborators.³² Since this meta-analysis included the major trials that enrolled elderly patients, individual publications of post hoc, elderly subgroups were, for the most part, not examined individually. The exception to this approach was a decision to report on the PROSPER¹³ and JUPITER²⁸ trials separately, because PROSPER is the most representative of the elderly population and JUPITER reached the lowest LDL-C of primary prevention trials published to date and included a large elderly subgroup (n = 5,695).

Savarese et al⁴⁰ evaluated the benefits of statins for older adults who did not have established cardiovascular disease. We did not report on this meta-analysis, as not all of the subjects that populated the meta-analysis were representative of a typical prevention population. For instance, in the Anglo-Scandinavian Cardiac Outcomes Trial lipid-lowering arm,⁴¹ 14% of the subjects had had a previous stroke or transient ischemic attack. In the Antihypertensive and Lipid-Lowering Treatment Trial,⁴² 16% of the population had a family history of premature coronary heart disease.

In addition, all the trials in the Savarese meta-analysis were also included in the CTT meta-analysis.³² The CTT reports on baseline risk using patient-level data stratified by age and risk, which may be more relevant to the question of primary prevention for older adults, as highlighted in our review.

PROSPER (Prospective Study of Pravastatin in the Elderly at Risk),¹³ a well-conducted, double-blind, randomized controlled trial with low probability of bias, compared pravastatin 40 mg and placebo. It was the only study that specifically enrolled older adults, with prespecified analysis of primary and secondary prevention subgroups. The primary prevention subgroup accounted for 56% of the 5,084 participants.

JUPITER (Justification for the Use of Statins in Prevention)²⁸ compared rosuvastatin 20 mg and placebo in 17,802 participants. All had low-density lipoprotein cholesterol (LDL-C) levels below 3.4 mmol/L (130 mg/dL) and elevated levels of the inflammatory biomarker high-sensitivity C-reactive protein (hsCRP), ie, 2 mg/L or higher. Subsequently, Glynn et al performed a post hoc, exploratory subgroup analysis of elderly participants (N = 5,695).²⁹

The JUPITER trial had several limitations.^43,44 The planned follow-up period was 5 years, but the trial was stopped early at 1.9 years, after a statistically significant difference was detected in the primary composite outcome of reduction in all vascular events. Studies that are stopped early may exaggerate positive findings.⁴⁵

Further, JUPITER’s patients were a select group, with normal LDL-C levels, elevated hsCRP values, and without diabetes. Of 90,000 patients screened, 72,000 (80%) did not meet the inclusion criteria and were not enrolled. This high rate of exclusion limits the generalizability of study findings beyond the shortcomings of post hoc subgroup analysis.

The meta-analysis performed by the CTT Collaborators³² used individual participant data from large-scale randomized trials of lipid-modifying treatment. This analysis was specific to people at low risk of vascular disease. In a supplementary appendix, the authors described the reduction in major vascular events for each 1.0 mmol/L decrease in LDL-C in three age categories: under age 60, ages 61 to 70, and over age 70.

The authors also stratified the results by risk category and provided information about those with a risk of major vascular events of less than 20%, which would be more representative of a purer primary prevention population.

For the elderly subgroup at low risk, the CTT Collaborators³² only reported a composite of major vascular events (coronary death, nonfatal myocardial infarction [MI], ischemic stroke, or revascularization) and did not describe individual outcomes, such as prevention of coronary heart disease.

Study results are based on postrandomization findings and therefore may be observational, not experimental.⁴⁶

Studies of statins for secondary prevention of cardiovascular disease

The aim of secondary prevention is to reduce the risk of recurrent cardiovascular events in patients who already have cardiovascular disease.

To address the question of whether statins reduce cardiovascular risk, we reviewed:

PROSPER,¹³ which included a preplanned analysis of the secondary prevention population.

Afilalo et al,^31,47 who performed a meta-analysis of the elderly subgroups of nine major secondary prevention studies (19,569 patients) using published and unpublished data.

To address the question of whether statins benefit individuals with heart failure, we found two relevant studies:

GISSI-HF (Gruppo Italiano per lo Studio della Sopravvivenza nell’Insufficienza Cardiaca Heart Failure)²⁵ and CORONA (Controlled Rosuvastatin Multinational Trial in Heart Failure),²⁶ which were large, international, well-conducted randomized controlled trials that examined statin use in heart failure.

To answer the question of whether statins benefit individuals after a stroke or transient ischemic attack, we found one relevant study:

SPARCL (Stroke Prevention by Aggressive Reduction in Cholesterol Levels),²⁷ which evaluated the benefit of statins in older adults with a history of stroke or transient ischemic attack. It was a prospective, double-blind, placebo-controlled, international trial conducted at 205 centers. One to 6 months after their cerebrovascular event, patients were randomized to receive either atorvastatin 80 mg or placebo. Given the young age of patients in this trial (mean age 63), we also reviewed a post hoc subgroup analysis of the elderly patients in SPARCL (age > 65).³⁰

HOW DID THE STUDY POPULATION COMPARE WITH THOSE WHO ARE FRAIL?

Frail older adults are almost always excluded from large-scale clinical trials,⁴⁸ leading to uncertainty about whether the conclusions can be applied to those with advanced frailty.

Although age is an imperfect proxy measure of frailty,⁴⁹ we consider the age of the study population as well as their comorbidities.

Participants in the studies we reviewed were generally younger and healthier than those who are frail, with mean ages of about 75 or less (Table 1).

PROSPER was the most representative study, as it specifically enrolled older adults, albeit without frailty,¹³ and excluded people with poor cognitive function as defined by a Mini Mental State Examination score less than 24.

JUPITER enrolled a select population, as described above. The median age in the elderly subgroup was 74 (interquartile range 72–78).²⁹

The Afilalo et al³¹ meta-analysis primarily included studies of young-elderly patients, with a mean age of less than 70. PROSPER¹³ was an exception.

The GISSI-HF study,²⁵ which examined the benefit of statins in heart failure, described their study population as frail, although the mean age was only 68. Compared with those in GISSI-HF, the CORONA patients²⁶ with heart failure were older (mean age 73) and had more severe heart failure. Accordingly, it is possible that many of the CORONA participants were frail.

ARE STUDY OUTCOMES CLINICALLY RELEVANT TO THOSE WHO ARE FRAIL?

Because baseline cardiovascular risk increases with age, the elderly should, in theory, experience greater absolute benefit from lipid-lowering. However, there is uncertainty about whether this is true in practice.

Some, but not all, epidemiologic studies show a weaker relationship between cholesterol levels and cardiovascular morbidity and mortality rates in older compared to younger adults.^50,51 This may be because those with high cholesterol levels die before they get old (time-related bias), or because those with life-threatening illness may have lower cholesterol levels.⁵⁰ In addition, classic risk factors such as age, sex, systolic blood pressure, cholesterol values, diabetes, smoking, and left ventricular hypertrophy on electrocardiography may have less power to predict cardiovascular risk among older patients.⁵²

The goal of treatment in frailty is to prevent further disability or improve quality of life. Therefore, meaningful outcomes for lipid-lowering therapy should include symptomatic nonfatal MI and its associated morbidity (eg, heart failure and persistent angina) or symptomatic nonfatal stroke leading to disability. Outcomes without sustained clinical impact, such as transient ischemic attack, nondisabling stroke, or silent MI, while potentially important in other populations, are less relevant in severe frailty. Notably, in many statin studies, outcomes include asymptomatic heart disease (eg, silent MI and “suspected events”) and nondisabling stroke (eg, mild stroke, transient ischemic attack). When symptomatic outcomes are not reported separately, the impact of the reported benefit on quality of life and function is uncertain.

The outcome of all-cause mortality is generally recognized as a gold standard for determining treatment benefit. However, since advanced frailty is characterized by multiple competing causes for mortality, a reduction in all-cause mortality that is achieved by addressing a single issue in nonfrail populations may not extend to the frail.

To more fully understand the impact of lipid-lowering therapy on quality of life and function, we examined the following questions:

Do statins as primary prevention reduce symptomatic heart disease?

Outcomes for coronary heart disease from PROSPER and JUPITER are summarized in Table 2.

PROSPER. In the PROSPER primary prevention group,¹³ statin therapy did not reduce the combined outcome of coronary heart disease death and nonfatal MI.

The JUPITER trial demonstrated a statistically significant benefit for preventing MI in the elderly subpopulation (ages 70–97),²⁹ but the number needed to treat was high (211 for 2 years), with a wide confidence interval (CI) (95% CI 106–32,924). The trial did not adequately differentiate between symptomatic and asymptomatic events, making it difficult to determine outcome relevance. Also, due to the methodologic limitations of JUPITER as described above, its results should be interpreted with caution.^43,44

The CTT Collaborators³² did not report individual outcomes (eg, coronary heart disease) for the elderly low-risk subgroup and, therefore, this meta-analysis does not answer the question of whether statins reduce symptomatic heart disease in primary prevention populations.

Taken together, these findings do not provide convincing evidence that statin therapy as primary prevention reduces the incidence of symptomatic heart disease for severely frail older adults.

Do statins as secondary prevention reduce symptomatic heart disease?

Most studies defined secondary prevention narrowly as treatment for patients with established coronary artery disease. For instance, in the Afilalo et al meta-analysis,³¹ the small number of studies that included individuals with other forms of vascular disease (such as peripheral vascular disease) enrolled few participants with noncardiac conditions (eg, 29% in PROSPER¹³ and 13% in the Heart Protection Study²⁰).

Therefore, any evidence of benefit for secondary prevention demonstrated in these studies is most applicable to patients with coronary heart disease, with less certainty for those with other forms of cardiovascular disease.

In PROSPER,¹³ the secondary prevention group experienced benefit in the combined outcome of coronary heart disease death or nonfatal MI. In the treatment group, 12.7% experienced this outcome compared with 16.8% with placebo, an absolute risk reduction of 4.1% in 3 years (P = .004, number needed to treat 25, 95% CI 15–77). This measure includes coronary heart disease death, an outcome that may not be generalizable to those who are frail. In addition, the outcome of nonfatal MI includes both symptomatic and suspected events. As such, there is uncertainty whether the realized benefit is clinically relevant to frail older adults.

The Afilalo et al meta-analysis³¹ showed that the number needed to treat to prevent one nonfatal MI was 38 (95% CI 16–118) over 5 years (Table 2). However, this outcome included both symptomatic and asymptomatic (silent) events.

Based on the available data, we conclude that it is not possible to determine whether statins reduce symptomatic heart disease as secondary prevention for older adults who are frail.

Do statins reduce heart disease in combined populations?

In the combined primary and secondary population from PROSPER,¹³ pravastatin decreased the risk of nonfatal symptomatic MI from 4.3% in the placebo group to 3.4%, a relatively small reduction in absolute risk (0.9%) and not statistically significant by our chi-square calculation (P = .099).

Do statins prevent a first symptomatic stroke in people with or without preexisting cardiovascular disease?

Preventing strokes that cause functional decline is an important outcome for the frail elderly. Stroke outcomes from PROSPER,¹³ JUPITER,²⁹ and the Afilalo et al meta-analysis³¹ are summarized in Table 3.

For primary prevention:

In PROSPER (primary prevention),¹³ there was no statistically significant benefit in the combined outcome of fatal and nonfatal stroke or the single outcome of transient ischemic attack after 3.2 years.

JUPITER,²⁹ in contrast, found that rosuvastatin 20 mg reduced strokes in primary prevention, but the absolute benefit was small. In 2 years, 0.8% of the treatment group had strokes, compared with 1.4% with placebo, an absolute risk reduction of 0.6% (P = .023, number needed to treat 161, 95% CI 86–1,192).

Neither PROSPER nor JUPITER differentiated between disabling and nondisabling strokes.

For secondary prevention:

In PROSPER (secondary prevention),¹³ there was no statistically significant benefit in the combined outcome of fatal and nonfatal stroke or the single outcome of transient ischemic attack after 3.2 years.

The Afilalo et al secondary prevention meta-analysis demonstrated a 25% relative reduction in stroke (relative risk 0.75, 95% CI 0.56–0.94, number needed to treat 58, 95% CI 27–177).³¹

Notably, the stroke outcome in Afilalo included both disabling and nondisabling strokes. For example, in the Heart Protection Study,²⁰ the largest study in the Afilalo et al meta-analysis, approximately 50% of nonfatal, classifiable strokes in the overall study population (ie, both younger and older patients) were not disabling. Including disabling and nondisabling strokes in a composite outcome confounds the clinical meaningfulness of these findings in frailty, as the number needed to treat to prevent one disabling stroke cannot be calculated from the data provided.

Do statins prevent a second (symptomatic) stroke in people with a previous stroke?

SPARCL²⁷ (Table 3) examined the question of whether statins decrease the risk of recurrent ischemic stroke for patients with a prior history of stroke or transient ischemic attack. There was a statistically significant reduction in the primary composite outcome of fatal and nonfatal stroke, with 11.2% of the treatment group and 13.1% of the placebo group experiencing this outcome, an absolute risk reduction of 1.9% at 5 years (P = .03; number needed to treat 52, 95% CI 26–1,303). However, the difference in nonfatal stroke, which is the outcome of interest for frailty (since mortality has uncertain relevance), was not statistically significant (10.4% with treatment vs 11.8% with placebo, P =.11).

An exploratory subgroup analysis of SPARCL patients based on age³⁰ showed a smaller, nonsignificant reduction in the primary end point of fatal and nonfatal stroke in the group over age 65 (relative risk 0.90, 95% confidence interval 0.73–1.11, P = .33) compared with the younger group (age < 65) (relative risk 0.74, 95% CI 0.57–0.96, P = .02).

The applicability of these results to the frail elderly is uncertain, since the subgroup analysis was not powered to determine outcomes based on age stratification and there were differences between groups in characteristics such as blood pressure and smoking status. In addition, the outcome of interest, nonfatal stroke, is not provided for the elderly subgroup.

In conclusion, in both primary and secondary prevention populations, the evidence that statins reduce nonfatal, symptomatic stroke rates for older adults is uncertain.

Do statins decrease all-cause mortality for primary or secondary prevention?

Due to competing risks for death, the outcome of mortality may not be relevant to those who are frail; however, studies showed the following:

For primary prevention, there was no decrease in mortality in PROSPER¹³ or in the elderly subgroup of JUPITER.²⁹

For secondary prevention, an analysis of PROSPER trial data by Afilalo et al³¹ showed a significant 18% decrease in all-cause mortality (relative risk 0.82, 95% CI 0.69–0.98) using pravastatin 40 mg.

A decrease in all-cause mortality with statins was also reported in the pooled result of the Afilalo et al meta-analysis.³¹

What are the reported composite outcomes for primary and secondary prevention?

While we were most interested in the symptomatic outcomes described above, we recognize that the small numbers of events make it difficult to draw firm conclusions. Therefore, we also considered composite primary outcomes, even though most included multiple measures that have varying associations with disability and relevancy to frail older adults.

For primary prevention, in the PROSPER preplanned subgroup analysis,¹³ there was no statistical benefit for any outcome, including the primary composite measure. In contrast, the elderly subpopulation in the JUPITER trial²⁸ showed a treatment benefit with rosuvastatin 20 mg compared with placebo for the primary composite outcome of MI, stroke, cardiovascular death, hospitalization for unstable angina, or revascularization. The number needed to treat for 2 years was 62 (95% CI 39–148).

In the CTT meta-analysis,³² patients at all levels of baseline risk showed benefit up to age 70. However, there was no statistically significant benefit in the composite primary outcome of coronary deaths, nonfatal myocardial infarction, ischemic stroke, or revascularization in the population most representative of elderly primary prevention—those who were more than 70 years old with a 5-year baseline risk of less than 20%.

For secondary prevention, in PROSPER,¹³ the subpopulation of patients treated for secondary prevention experienced benefit in the primary composite outcome of coronary heart disease death, nonfatal MI, or fatal or nonfatal stroke, achieving a 4% absolute risk reduction with a number needed to treat of 23 (95% CI 14–81) over 3 years.

Do statins decrease disability?

PROSPER was the only study that reported on disability. Compared with placebo, pravastatin did not decrease disability in the total population as measured by basic and instrumental activities of daily living scales.

Do statins help patients with heart failure?

Neither GISSI-HF²⁵ nor CORONA²⁶ found significant benefit from rosuvastatin 10 mg, despite LDL-C lowering of 27% in GISSI-HF and 45% in CORONA.

Do ezetimibe or other nonstatin lipid-lowering agents improve outcomes?

There is no definitive evidence that ezetimibe provides clinically meaningful benefit as a single agent.

For combination therapy, the IMPROVE-IT (Improved Reduction of Outcomes: Vytorin Efficacy International Trial)⁵³ showed that adding ezetimibe 10 mg to simvastatin 40 mg after an acute coronary syndrome reduced the risk of nonfatal myocardial infarction compared with simvastatin monotherapy (event rate 12.8% vs 14.4%; hazard ratio 0.87, 95% CI 0.80–0.95; P = .002) for a population with a mean age of 64. The risk of any stroke was also reduced; strokes occurred in 4.2% of those receiving combination therapy vs 4.8% with monotherapy (hazard ratio 0.86, 95% CI 0.73–1.00, P = .05). After a median of 6 years, 42% of patients in each group had discontinued treatment. Given the very specific clinical scenario of acute coronary syndrome and the young age of the patients in this trial, we do not think that this study justifies the use of ezetimibe for severely frail older adults.

There is no evidence that other combinations (ie, a statin plus another lipid-lowering drug) improve clinical outcomes for either primary or secondary prevention in any population.⁵⁴

WILL FRAIL PATIENTS LIVE LONG ENOUGH TO BENEFIT?

It is often difficult to determine the number of years that are needed to achieve benefit, as most trials do not provide a statistical analysis of varying time frames.

The PROSPER trial¹³ lasted 3.2 years. From the Kaplan-Meier curves in PROSPER, we estimate that it took about 1.5 years to achieve a 1% absolute risk reduction and 2.5 years for a 2% absolute risk reduction in coronary heart disease death and nonfatal MI in the combined primary and secondary groups.

JUPITER²⁸ was stopped early at 1.9 years. The Afilalo et al meta-analysis³¹ was based on follow-up over 4.9 years.

IMPROVE-IT⁵³ reported event rates at 7 years. The authors note that benefit in the primary composite outcome appeared to emerge at 1 year, although no statistical support is given for this statement and divergence in the Kaplan-Meier curves is not visually apparent.

The duration of other studies ranged between 2.7 and 4.9 years (Table 1).^26–28

It has been suggested that statins should be considered for elderly patients who have a life expectancy of at least 5 years.³ However, many older adults have already been taking statins for many years, which makes it difficult to interpret the available timeframe evidence.

In a multicenter, unblinded, randomized trial,⁵⁵ statins were either stopped or continued in older adults who had a short life expectancy and a median survival of approximately 7 months. Causes of death were evenly divided between cancer and noncancer diagnoses, and 22% of the patients were cognitively impaired. Discontinuing statin therapy did not increase mortality or cardiovascular events within 60 days. Nevertheless, stopping statin therapy did not achieve noninferiority for the primary end point, the proportion of participants who died within 60 days. Statin discontinuation was associated with improved quality of life, although the study was not blinded, which could have influenced results.

HAVE THE HARMS BEEN SUFFICIENTLY CONSIDERED?

Frail older adults commonly take multiple medications and are more vulnerable to adverse events.⁵⁶

Many statins require dose reduction with severe renal impairment (creatinine clearance < 30 mL/min/1.73 m²), which would be a common consideration in severely frail older adults.

Myopathy

Myopathy, which includes myalgias and muscle weakness, is a statin-related adverse event that can impair quality of life. Myopathy typically develops within the first 6 months but can occur at any time during statin treatment.⁵⁷ When muscle-related adverse effects occur, they may affect the elderly more significantly, particularly their ability to perform activities of daily living, rise from a chair, or mobilize independently. Another concern is that older adults with dementia may not be able to accurately report muscle-related symptoms.

It is difficult to ascertain the true prevalence of myopathy, especially in advanced age and frailty. Randomized controlled trials report incidence rates of 1.5% to 5%, which is comparable to placebo.^57,58 However, inconsistent definitions of myopathy and exclusion of subjects with previous statin intolerance or adverse effects during run-in periods limit interpretability.⁵⁷ Clinical experience suggests that muscle complaints may be relatively common.^59–61

Advanced age, female sex, low body mass index, and multisystem disease are all associated with frailty and have also been described as risk factors for statin-associated muscle syndromes.⁶¹ Physiologic changes associated with frailty, such as reduced muscle strength, decreased lean body mass, impaired functional mobility, decreased reserve capacity, and altered drug metabolism may increase the risk and severity of myopathy.⁶²

Adverse cognitive events

Meta-analyses of randomized clinical trials and narrative reviews find no definitive relationship between statin therapy and adverse cognitive events.^63–67 Nevertheless, there have been case reports of memory loss associated with the use of statins, and the US Food and Drug Administration has issued a warning that statins have been associated with memory loss and confusion.⁶⁸

It may be difficult to determine whether a statin is causing or aggravating cognitive symptoms among individuals with dementia without a trial withdrawal of the drug.

OUR RECOMMENDATIONS

The recommendations below are intended for adults with severe or very severe frailty (ie, a score of 7 or 8 on the Clinical Frailty Scale¹¹ or FACT⁵ and therefore apply to most older adults living in long-term care facilities.

Primary prevention

There is no reason to prescribe or continue statins for primary prevention, as it is unlikely that they would provide benefit for outcomes that are relevant in this population.

Secondary prevention

Statin treatment is probably not necessary for secondary prevention in those with severe frailty, although there may be extenuating circumstances that justify statin use.

Heart failure

There is no reason to start or continue statins for heart failure, as there is insufficient evidence that they are effective for this indication in any population.

Ezetimibe

There is no evidence that ezetimibe reduces cardiovascular events in any population when used as monotherapy. For a select population with acute coronary syndromes, ezetimibe has a modest effect. Given the very specific clinical scenario of acute coronary syndrome, we do not think that the available evidence justifies the use of ezetimibe for severely frail older adults.

Agents other than ezetimibe combined with statins

There is no reason to start or continue other lipid-lowering drugs in conjunction with statins.

Statin dosing

As statin adverse effects have the potential to increase with advancing age and frailty, lower doses may be appropriate.⁶⁸

Adverse events

Consider stopping statins on a trial basis if there is concern regarding myopathy, drug interactions, or other adverse effects.

BOTTOM LINE: DO STATINS IMPROVE QUALITY OF LIFE OR FUNCTION?

In primary prevention for older adults, there is doubt that statins prevent cardiovascular disease and stroke-related events because the main study involving the elderly did not show a benefit in the primary prevention subgroup.¹³ Additionally, there is no conclusive evidence that statin treatment decreases mortality in primary prevention.^13,29

There is insufficient information to determine whether the frail elderly should receive statins for secondary prevention. Although there is evidence that treatment decreases measures of coronary heart disease and stroke, it is unclear whether it improves quality of life or function for those who are frail. To answer this question, we need more information about whether reported outcomes (such as stroke and MI) are associated with disability, which is not provided in many of the studies we reviewed. When disability was specifically considered in the PROSPER trial for the combined population of primary and secondary prevention, treatment with statins had no impact on basic and instrumental activities of daily living.

Some experts may not agree with our interpretation of the complex evidence presented in this article. Others may ask, “What is the harm in using statins, even if there is no definitive benefit?” However, the harms associated with statin therapy for the frail are poorly defined. In the face of these uncertainties and in the absence of definitive improvement in quality of life, we believe that “less is more” in the context of severe frailty.⁶⁹

The cost of medications should also be considered, especially in long-term care facilities, where there is an added expense of drug administration that diverts human resources away from interactions that are more congruent with respecting the lifestage of frailty.

Careful review of evidence before applying clinical practice guidelines to those who are frail should become the norm. When considering treatment of frail patients, the five questions described in this review shed light on the applicability of clinical trial evidence. Therapies that are highly effective in healthier populations may be less effective when individuals are severely frail. Accordingly, we propose that medications should only be used if they improve quality of life or function.

Frail elderly patients are at high risk of adverse clinical outcomes, including those due to polypharmacy. Several groups tackle “deprescribing” by developing lists of medications that are potentially inappropriate for the elderly, such as the Beers or STOPP/START criteria.^1–4

In contrast, our group (the Palliative and Therapeutic Harmonization [PATH] program and the Dalhousie Academic Detailing Service) has developed evidence-based, frailty-specific guidelines for treating hypertension⁵ and diabetes,⁶ in which we advocate less-stringent treatment targets and tapering or discontinuing medications, as needed.

The PATH program⁷ is a clinical approach that prioritizes the consideration of frailty when making treatment decisions. The Dalhousie Academic Detailing Service collaborates with the Nova Scotia Health Authority to research and develop evidence-informed educational messages about the treatment of common medical conditions.

Here, we address lipid-lowering therapy in this population.

CONSIDERING FRAILTY

Frailty is defined in several ways. The Fried model^8,9 identifies frailty when 3 of the following characteristics are present: unintentional weight loss, exhaustion, muscle weakness, slow walking speed, or low levels of activity. The Clinical Frailty Scale^10,11 and the Frailty Assessment for Care-planning Tool (FACT)⁵ use deficits in cognition, function, and mobility to define frailty. According to these scales, people are considered severely frail when they require assistance with basic activities of daily living (such as bathing or dressing), owing to cognitive or physical deficits from any cause.

In reviewing the evidence, we consider five questions:

• What is the quality of the evidence? (Up to 48% of clinical practice guideline recommendations may be based on low-level evidence or expert opinion.¹²)
• How did the study population compare with the frail?
• Are study outcomes and potential benefits clinically relevant to those who are frail?
• How long did it take for the clinical benefit of a treatment to become apparent, and are the frail elderly likely to live that long?
• Have the harms of treatment been sufficiently considered?

WHAT IS THE QUALITY OF THE EVIDENCE?

We found no studies that specifically evaluated the benefit of lipid-lowering for severely frail older adults. Therefore, we examined randomized controlled trials that enrolled non-frail older adults,^13–28 subgroup analyses of randomized controlled trials,^29,30 meta-analyses that analyzed subgroups of elderly populations,^31,32 and publications describing the study designs of randomized controlled trials.^33–37

Most of the evidence comes from post hoc subgroup analyses of elderly populations. Although meta-analysis is commonly used to compare subgroups, the Cochrane handbook and others consider subgroup comparisons observational by nature.^38,39 (See Table 1 for lipid-lowering studies discussed in this article.)

Studies of statins for primary prevention of cardiovascular disease

For evidence of benefit from lipid-lowering for primary prevention (ie, to reduce the risk of cardiovascular events in patients with no known cardiovascular disease at baseline but at increased risk), we reviewed the meta-analysis conducted by the Cholesterol Treatment Trialists’ (CTT) Collaborators.³² Since this meta-analysis included the major trials that enrolled elderly patients, individual publications of post hoc, elderly subgroups were, for the most part, not examined individually. The exception to this approach was a decision to report on the PROSPER¹³ and JUPITER²⁸ trials separately, because PROSPER is the most representative of the elderly population and JUPITER reached the lowest LDL-C of primary prevention trials published to date and included a large elderly subgroup (n = 5,695).

Savarese et al⁴⁰ evaluated the benefits of statins for older adults who did not have established cardiovascular disease. We did not report on this meta-analysis, as not all of the subjects that populated the meta-analysis were representative of a typical prevention population. For instance, in the Anglo-Scandinavian Cardiac Outcomes Trial lipid-lowering arm,⁴¹ 14% of the subjects had had a previous stroke or transient ischemic attack. In the Antihypertensive and Lipid-Lowering Treatment Trial,⁴² 16% of the population had a family history of premature coronary heart disease.

In addition, all the trials in the Savarese meta-analysis were also included in the CTT meta-analysis.³² The CTT reports on baseline risk using patient-level data stratified by age and risk, which may be more relevant to the question of primary prevention for older adults, as highlighted in our review.

PROSPER (Prospective Study of Pravastatin in the Elderly at Risk),¹³ a well-conducted, double-blind, randomized controlled trial with low probability of bias, compared pravastatin 40 mg and placebo. It was the only study that specifically enrolled older adults, with prespecified analysis of primary and secondary prevention subgroups. The primary prevention subgroup accounted for 56% of the 5,084 participants.

JUPITER (Justification for the Use of Statins in Prevention)²⁸ compared rosuvastatin 20 mg and placebo in 17,802 participants. All had low-density lipoprotein cholesterol (LDL-C) levels below 3.4 mmol/L (130 mg/dL) and elevated levels of the inflammatory biomarker high-sensitivity C-reactive protein (hsCRP), ie, 2 mg/L or higher. Subsequently, Glynn et al performed a post hoc, exploratory subgroup analysis of elderly participants (N = 5,695).²⁹

The JUPITER trial had several limitations.^43,44 The planned follow-up period was 5 years, but the trial was stopped early at 1.9 years, after a statistically significant difference was detected in the primary composite outcome of reduction in all vascular events. Studies that are stopped early may exaggerate positive findings.⁴⁵

Further, JUPITER’s patients were a select group, with normal LDL-C levels, elevated hsCRP values, and without diabetes. Of 90,000 patients screened, 72,000 (80%) did not meet the inclusion criteria and were not enrolled. This high rate of exclusion limits the generalizability of study findings beyond the shortcomings of post hoc subgroup analysis.

The meta-analysis performed by the CTT Collaborators³² used individual participant data from large-scale randomized trials of lipid-modifying treatment. This analysis was specific to people at low risk of vascular disease. In a supplementary appendix, the authors described the reduction in major vascular events for each 1.0 mmol/L decrease in LDL-C in three age categories: under age 60, ages 61 to 70, and over age 70.

The authors also stratified the results by risk category and provided information about those with a risk of major vascular events of less than 20%, which would be more representative of a purer primary prevention population.

For the elderly subgroup at low risk, the CTT Collaborators³² only reported a composite of major vascular events (coronary death, nonfatal myocardial infarction [MI], ischemic stroke, or revascularization) and did not describe individual outcomes, such as prevention of coronary heart disease.

Study results are based on postrandomization findings and therefore may be observational, not experimental.⁴⁶

Studies of statins for secondary prevention of cardiovascular disease

The aim of secondary prevention is to reduce the risk of recurrent cardiovascular events in patients who already have cardiovascular disease.

To address the question of whether statins reduce cardiovascular risk, we reviewed:

PROSPER,¹³ which included a preplanned analysis of the secondary prevention population.

Afilalo et al,^31,47 who performed a meta-analysis of the elderly subgroups of nine major secondary prevention studies (19,569 patients) using published and unpublished data.

To address the question of whether statins benefit individuals with heart failure, we found two relevant studies:

GISSI-HF (Gruppo Italiano per lo Studio della Sopravvivenza nell’Insufficienza Cardiaca Heart Failure)²⁵ and CORONA (Controlled Rosuvastatin Multinational Trial in Heart Failure),²⁶ which were large, international, well-conducted randomized controlled trials that examined statin use in heart failure.

To answer the question of whether statins benefit individuals after a stroke or transient ischemic attack, we found one relevant study:

SPARCL (Stroke Prevention by Aggressive Reduction in Cholesterol Levels),²⁷ which evaluated the benefit of statins in older adults with a history of stroke or transient ischemic attack. It was a prospective, double-blind, placebo-controlled, international trial conducted at 205 centers. One to 6 months after their cerebrovascular event, patients were randomized to receive either atorvastatin 80 mg or placebo. Given the young age of patients in this trial (mean age 63), we also reviewed a post hoc subgroup analysis of the elderly patients in SPARCL (age > 65).³⁰

HOW DID THE STUDY POPULATION COMPARE WITH THOSE WHO ARE FRAIL?

Frail older adults are almost always excluded from large-scale clinical trials,⁴⁸ leading to uncertainty about whether the conclusions can be applied to those with advanced frailty.

Although age is an imperfect proxy measure of frailty,⁴⁹ we consider the age of the study population as well as their comorbidities.

Participants in the studies we reviewed were generally younger and healthier than those who are frail, with mean ages of about 75 or less (Table 1).

PROSPER was the most representative study, as it specifically enrolled older adults, albeit without frailty,¹³ and excluded people with poor cognitive function as defined by a Mini Mental State Examination score less than 24.

JUPITER enrolled a select population, as described above. The median age in the elderly subgroup was 74 (interquartile range 72–78).²⁹

The Afilalo et al³¹ meta-analysis primarily included studies of young-elderly patients, with a mean age of less than 70. PROSPER¹³ was an exception.

The GISSI-HF study,²⁵ which examined the benefit of statins in heart failure, described their study population as frail, although the mean age was only 68. Compared with those in GISSI-HF, the CORONA patients²⁶ with heart failure were older (mean age 73) and had more severe heart failure. Accordingly, it is possible that many of the CORONA participants were frail.

ARE STUDY OUTCOMES CLINICALLY RELEVANT TO THOSE WHO ARE FRAIL?

Because baseline cardiovascular risk increases with age, the elderly should, in theory, experience greater absolute benefit from lipid-lowering. However, there is uncertainty about whether this is true in practice.

Some, but not all, epidemiologic studies show a weaker relationship between cholesterol levels and cardiovascular morbidity and mortality rates in older compared to younger adults.^50,51 This may be because those with high cholesterol levels die before they get old (time-related bias), or because those with life-threatening illness may have lower cholesterol levels.⁵⁰ In addition, classic risk factors such as age, sex, systolic blood pressure, cholesterol values, diabetes, smoking, and left ventricular hypertrophy on electrocardiography may have less power to predict cardiovascular risk among older patients.⁵²

The goal of treatment in frailty is to prevent further disability or improve quality of life. Therefore, meaningful outcomes for lipid-lowering therapy should include symptomatic nonfatal MI and its associated morbidity (eg, heart failure and persistent angina) or symptomatic nonfatal stroke leading to disability. Outcomes without sustained clinical impact, such as transient ischemic attack, nondisabling stroke, or silent MI, while potentially important in other populations, are less relevant in severe frailty. Notably, in many statin studies, outcomes include asymptomatic heart disease (eg, silent MI and “suspected events”) and nondisabling stroke (eg, mild stroke, transient ischemic attack). When symptomatic outcomes are not reported separately, the impact of the reported benefit on quality of life and function is uncertain.

The outcome of all-cause mortality is generally recognized as a gold standard for determining treatment benefit. However, since advanced frailty is characterized by multiple competing causes for mortality, a reduction in all-cause mortality that is achieved by addressing a single issue in nonfrail populations may not extend to the frail.

To more fully understand the impact of lipid-lowering therapy on quality of life and function, we examined the following questions:

Do statins as primary prevention reduce symptomatic heart disease?

Outcomes for coronary heart disease from PROSPER and JUPITER are summarized in Table 2.

PROSPER. In the PROSPER primary prevention group,¹³ statin therapy did not reduce the combined outcome of coronary heart disease death and nonfatal MI.

The JUPITER trial demonstrated a statistically significant benefit for preventing MI in the elderly subpopulation (ages 70–97),²⁹ but the number needed to treat was high (211 for 2 years), with a wide confidence interval (CI) (95% CI 106–32,924). The trial did not adequately differentiate between symptomatic and asymptomatic events, making it difficult to determine outcome relevance. Also, due to the methodologic limitations of JUPITER as described above, its results should be interpreted with caution.^43,44

The CTT Collaborators³² did not report individual outcomes (eg, coronary heart disease) for the elderly low-risk subgroup and, therefore, this meta-analysis does not answer the question of whether statins reduce symptomatic heart disease in primary prevention populations.

Taken together, these findings do not provide convincing evidence that statin therapy as primary prevention reduces the incidence of symptomatic heart disease for severely frail older adults.

Do statins as secondary prevention reduce symptomatic heart disease?

Most studies defined secondary prevention narrowly as treatment for patients with established coronary artery disease. For instance, in the Afilalo et al meta-analysis,³¹ the small number of studies that included individuals with other forms of vascular disease (such as peripheral vascular disease) enrolled few participants with noncardiac conditions (eg, 29% in PROSPER¹³ and 13% in the Heart Protection Study²⁰).

Therefore, any evidence of benefit for secondary prevention demonstrated in these studies is most applicable to patients with coronary heart disease, with less certainty for those with other forms of cardiovascular disease.

In PROSPER,¹³ the secondary prevention group experienced benefit in the combined outcome of coronary heart disease death or nonfatal MI. In the treatment group, 12.7% experienced this outcome compared with 16.8% with placebo, an absolute risk reduction of 4.1% in 3 years (P = .004, number needed to treat 25, 95% CI 15–77). This measure includes coronary heart disease death, an outcome that may not be generalizable to those who are frail. In addition, the outcome of nonfatal MI includes both symptomatic and suspected events. As such, there is uncertainty whether the realized benefit is clinically relevant to frail older adults.

The Afilalo et al meta-analysis³¹ showed that the number needed to treat to prevent one nonfatal MI was 38 (95% CI 16–118) over 5 years (Table 2). However, this outcome included both symptomatic and asymptomatic (silent) events.

Based on the available data, we conclude that it is not possible to determine whether statins reduce symptomatic heart disease as secondary prevention for older adults who are frail.

Do statins reduce heart disease in combined populations?

In the combined primary and secondary population from PROSPER,¹³ pravastatin decreased the risk of nonfatal symptomatic MI from 4.3% in the placebo group to 3.4%, a relatively small reduction in absolute risk (0.9%) and not statistically significant by our chi-square calculation (P = .099).

Do statins prevent a first symptomatic stroke in people with or without preexisting cardiovascular disease?

Preventing strokes that cause functional decline is an important outcome for the frail elderly. Stroke outcomes from PROSPER,¹³ JUPITER,²⁹ and the Afilalo et al meta-analysis³¹ are summarized in Table 3.

For primary prevention:

In PROSPER (primary prevention),¹³ there was no statistically significant benefit in the combined outcome of fatal and nonfatal stroke or the single outcome of transient ischemic attack after 3.2 years.

JUPITER,²⁹ in contrast, found that rosuvastatin 20 mg reduced strokes in primary prevention, but the absolute benefit was small. In 2 years, 0.8% of the treatment group had strokes, compared with 1.4% with placebo, an absolute risk reduction of 0.6% (P = .023, number needed to treat 161, 95% CI 86–1,192).

Neither PROSPER nor JUPITER differentiated between disabling and nondisabling strokes.

For secondary prevention:

In PROSPER (secondary prevention),¹³ there was no statistically significant benefit in the combined outcome of fatal and nonfatal stroke or the single outcome of transient ischemic attack after 3.2 years.

The Afilalo et al secondary prevention meta-analysis demonstrated a 25% relative reduction in stroke (relative risk 0.75, 95% CI 0.56–0.94, number needed to treat 58, 95% CI 27–177).³¹

Notably, the stroke outcome in Afilalo included both disabling and nondisabling strokes. For example, in the Heart Protection Study,²⁰ the largest study in the Afilalo et al meta-analysis, approximately 50% of nonfatal, classifiable strokes in the overall study population (ie, both younger and older patients) were not disabling. Including disabling and nondisabling strokes in a composite outcome confounds the clinical meaningfulness of these findings in frailty, as the number needed to treat to prevent one disabling stroke cannot be calculated from the data provided.

Do statins prevent a second (symptomatic) stroke in people with a previous stroke?

SPARCL²⁷ (Table 3) examined the question of whether statins decrease the risk of recurrent ischemic stroke for patients with a prior history of stroke or transient ischemic attack. There was a statistically significant reduction in the primary composite outcome of fatal and nonfatal stroke, with 11.2% of the treatment group and 13.1% of the placebo group experiencing this outcome, an absolute risk reduction of 1.9% at 5 years (P = .03; number needed to treat 52, 95% CI 26–1,303). However, the difference in nonfatal stroke, which is the outcome of interest for frailty (since mortality has uncertain relevance), was not statistically significant (10.4% with treatment vs 11.8% with placebo, P =.11).

An exploratory subgroup analysis of SPARCL patients based on age³⁰ showed a smaller, nonsignificant reduction in the primary end point of fatal and nonfatal stroke in the group over age 65 (relative risk 0.90, 95% confidence interval 0.73–1.11, P = .33) compared with the younger group (age < 65) (relative risk 0.74, 95% CI 0.57–0.96, P = .02).

The applicability of these results to the frail elderly is uncertain, since the subgroup analysis was not powered to determine outcomes based on age stratification and there were differences between groups in characteristics such as blood pressure and smoking status. In addition, the outcome of interest, nonfatal stroke, is not provided for the elderly subgroup.

In conclusion, in both primary and secondary prevention populations, the evidence that statins reduce nonfatal, symptomatic stroke rates for older adults is uncertain.

Do statins decrease all-cause mortality for primary or secondary prevention?

Due to competing risks for death, the outcome of mortality may not be relevant to those who are frail; however, studies showed the following:

For primary prevention, there was no decrease in mortality in PROSPER¹³ or in the elderly subgroup of JUPITER.²⁹

For secondary prevention, an analysis of PROSPER trial data by Afilalo et al³¹ showed a significant 18% decrease in all-cause mortality (relative risk 0.82, 95% CI 0.69–0.98) using pravastatin 40 mg.

A decrease in all-cause mortality with statins was also reported in the pooled result of the Afilalo et al meta-analysis.³¹

What are the reported composite outcomes for primary and secondary prevention?

While we were most interested in the symptomatic outcomes described above, we recognize that the small numbers of events make it difficult to draw firm conclusions. Therefore, we also considered composite primary outcomes, even though most included multiple measures that have varying associations with disability and relevancy to frail older adults.

For primary prevention, in the PROSPER preplanned subgroup analysis,¹³ there was no statistical benefit for any outcome, including the primary composite measure. In contrast, the elderly subpopulation in the JUPITER trial²⁸ showed a treatment benefit with rosuvastatin 20 mg compared with placebo for the primary composite outcome of MI, stroke, cardiovascular death, hospitalization for unstable angina, or revascularization. The number needed to treat for 2 years was 62 (95% CI 39–148).

In the CTT meta-analysis,³² patients at all levels of baseline risk showed benefit up to age 70. However, there was no statistically significant benefit in the composite primary outcome of coronary deaths, nonfatal myocardial infarction, ischemic stroke, or revascularization in the population most representative of elderly primary prevention—those who were more than 70 years old with a 5-year baseline risk of less than 20%.

For secondary prevention, in PROSPER,¹³ the subpopulation of patients treated for secondary prevention experienced benefit in the primary composite outcome of coronary heart disease death, nonfatal MI, or fatal or nonfatal stroke, achieving a 4% absolute risk reduction with a number needed to treat of 23 (95% CI 14–81) over 3 years.

Do statins decrease disability?

PROSPER was the only study that reported on disability. Compared with placebo, pravastatin did not decrease disability in the total population as measured by basic and instrumental activities of daily living scales.

Do statins help patients with heart failure?

Neither GISSI-HF²⁵ nor CORONA²⁶ found significant benefit from rosuvastatin 10 mg, despite LDL-C lowering of 27% in GISSI-HF and 45% in CORONA.

Do ezetimibe or other nonstatin lipid-lowering agents improve outcomes?

There is no definitive evidence that ezetimibe provides clinically meaningful benefit as a single agent.

For combination therapy, the IMPROVE-IT (Improved Reduction of Outcomes: Vytorin Efficacy International Trial)⁵³ showed that adding ezetimibe 10 mg to simvastatin 40 mg after an acute coronary syndrome reduced the risk of nonfatal myocardial infarction compared with simvastatin monotherapy (event rate 12.8% vs 14.4%; hazard ratio 0.87, 95% CI 0.80–0.95; P = .002) for a population with a mean age of 64. The risk of any stroke was also reduced; strokes occurred in 4.2% of those receiving combination therapy vs 4.8% with monotherapy (hazard ratio 0.86, 95% CI 0.73–1.00, P = .05). After a median of 6 years, 42% of patients in each group had discontinued treatment. Given the very specific clinical scenario of acute coronary syndrome and the young age of the patients in this trial, we do not think that this study justifies the use of ezetimibe for severely frail older adults.

There is no evidence that other combinations (ie, a statin plus another lipid-lowering drug) improve clinical outcomes for either primary or secondary prevention in any population.⁵⁴

WILL FRAIL PATIENTS LIVE LONG ENOUGH TO BENEFIT?

It is often difficult to determine the number of years that are needed to achieve benefit, as most trials do not provide a statistical analysis of varying time frames.

The PROSPER trial¹³ lasted 3.2 years. From the Kaplan-Meier curves in PROSPER, we estimate that it took about 1.5 years to achieve a 1% absolute risk reduction and 2.5 years for a 2% absolute risk reduction in coronary heart disease death and nonfatal MI in the combined primary and secondary groups.

JUPITER²⁸ was stopped early at 1.9 years. The Afilalo et al meta-analysis³¹ was based on follow-up over 4.9 years.

IMPROVE-IT⁵³ reported event rates at 7 years. The authors note that benefit in the primary composite outcome appeared to emerge at 1 year, although no statistical support is given for this statement and divergence in the Kaplan-Meier curves is not visually apparent.

The duration of other studies ranged between 2.7 and 4.9 years (Table 1).^26–28

It has been suggested that statins should be considered for elderly patients who have a life expectancy of at least 5 years.³ However, many older adults have already been taking statins for many years, which makes it difficult to interpret the available timeframe evidence.

In a multicenter, unblinded, randomized trial,⁵⁵ statins were either stopped or continued in older adults who had a short life expectancy and a median survival of approximately 7 months. Causes of death were evenly divided between cancer and noncancer diagnoses, and 22% of the patients were cognitively impaired. Discontinuing statin therapy did not increase mortality or cardiovascular events within 60 days. Nevertheless, stopping statin therapy did not achieve noninferiority for the primary end point, the proportion of participants who died within 60 days. Statin discontinuation was associated with improved quality of life, although the study was not blinded, which could have influenced results.

HAVE THE HARMS BEEN SUFFICIENTLY CONSIDERED?

Frail older adults commonly take multiple medications and are more vulnerable to adverse events.⁵⁶

Many statins require dose reduction with severe renal impairment (creatinine clearance < 30 mL/min/1.73 m²), which would be a common consideration in severely frail older adults.

Myopathy

Myopathy, which includes myalgias and muscle weakness, is a statin-related adverse event that can impair quality of life. Myopathy typically develops within the first 6 months but can occur at any time during statin treatment.⁵⁷ When muscle-related adverse effects occur, they may affect the elderly more significantly, particularly their ability to perform activities of daily living, rise from a chair, or mobilize independently. Another concern is that older adults with dementia may not be able to accurately report muscle-related symptoms.

It is difficult to ascertain the true prevalence of myopathy, especially in advanced age and frailty. Randomized controlled trials report incidence rates of 1.5% to 5%, which is comparable to placebo.^57,58 However, inconsistent definitions of myopathy and exclusion of subjects with previous statin intolerance or adverse effects during run-in periods limit interpretability.⁵⁷ Clinical experience suggests that muscle complaints may be relatively common.^59–61

Advanced age, female sex, low body mass index, and multisystem disease are all associated with frailty and have also been described as risk factors for statin-associated muscle syndromes.⁶¹ Physiologic changes associated with frailty, such as reduced muscle strength, decreased lean body mass, impaired functional mobility, decreased reserve capacity, and altered drug metabolism may increase the risk and severity of myopathy.⁶²

Adverse cognitive events

Meta-analyses of randomized clinical trials and narrative reviews find no definitive relationship between statin therapy and adverse cognitive events.^63–67 Nevertheless, there have been case reports of memory loss associated with the use of statins, and the US Food and Drug Administration has issued a warning that statins have been associated with memory loss and confusion.⁶⁸

It may be difficult to determine whether a statin is causing or aggravating cognitive symptoms among individuals with dementia without a trial withdrawal of the drug.

OUR RECOMMENDATIONS

The recommendations below are intended for adults with severe or very severe frailty (ie, a score of 7 or 8 on the Clinical Frailty Scale¹¹ or FACT⁵ and therefore apply to most older adults living in long-term care facilities.

Primary prevention

There is no reason to prescribe or continue statins for primary prevention, as it is unlikely that they would provide benefit for outcomes that are relevant in this population.

Secondary prevention

Statin treatment is probably not necessary for secondary prevention in those with severe frailty, although there may be extenuating circumstances that justify statin use.

Heart failure

There is no reason to start or continue statins for heart failure, as there is insufficient evidence that they are effective for this indication in any population.

Ezetimibe

There is no evidence that ezetimibe reduces cardiovascular events in any population when used as monotherapy. For a select population with acute coronary syndromes, ezetimibe has a modest effect. Given the very specific clinical scenario of acute coronary syndrome, we do not think that the available evidence justifies the use of ezetimibe for severely frail older adults.

Agents other than ezetimibe combined with statins

There is no reason to start or continue other lipid-lowering drugs in conjunction with statins.

Statin dosing

As statin adverse effects have the potential to increase with advancing age and frailty, lower doses may be appropriate.⁶⁸

Adverse events

Consider stopping statins on a trial basis if there is concern regarding myopathy, drug interactions, or other adverse effects.

BOTTOM LINE: DO STATINS IMPROVE QUALITY OF LIFE OR FUNCTION?

In primary prevention for older adults, there is doubt that statins prevent cardiovascular disease and stroke-related events because the main study involving the elderly did not show a benefit in the primary prevention subgroup.¹³ Additionally, there is no conclusive evidence that statin treatment decreases mortality in primary prevention.^13,29

There is insufficient information to determine whether the frail elderly should receive statins for secondary prevention. Although there is evidence that treatment decreases measures of coronary heart disease and stroke, it is unclear whether it improves quality of life or function for those who are frail. To answer this question, we need more information about whether reported outcomes (such as stroke and MI) are associated with disability, which is not provided in many of the studies we reviewed. When disability was specifically considered in the PROSPER trial for the combined population of primary and secondary prevention, treatment with statins had no impact on basic and instrumental activities of daily living.

Some experts may not agree with our interpretation of the complex evidence presented in this article. Others may ask, “What is the harm in using statins, even if there is no definitive benefit?” However, the harms associated with statin therapy for the frail are poorly defined. In the face of these uncertainties and in the absence of definitive improvement in quality of life, we believe that “less is more” in the context of severe frailty.⁶⁹

The cost of medications should also be considered, especially in long-term care facilities, where there is an added expense of drug administration that diverts human resources away from interactions that are more congruent with respecting the lifestage of frailty.

Careful review of evidence before applying clinical practice guidelines to those who are frail should become the norm. When considering treatment of frail patients, the five questions described in this review shed light on the applicability of clinical trial evidence. Therapies that are highly effective in healthier populations may be less effective when individuals are severely frail. Accordingly, we propose that medications should only be used if they improve quality of life or function.

To the Editor: I enjoyed Dr. Galicia-Castillo’s article about long-term opioid therapy in older adults,¹ which reaffirmed the imperative to “start low and go slow” to minimize the risk of addiction. However, the article missed an opportunity to raise awareness regarding another extremely important side effect of chronic prescription opioid consumption, that of ingestion prior to sleep, with consequent cessation of breathing leading to death.

According to the Drug Enforcement Administration,² most narcotic deaths are a result of respiratory depression. And the American Pain Society has stated, “No patient has succumbed to [opioid] respiratory depression while awake.”³

Dr. Galicia-Castillo noted that the prevalence of central sleep apnea in chronic opioid users is 24%, based on a review by Correa et al.⁴ As alarming as this number is, other investigators have estimated it to be even higher—as high as 50% to 90%.⁵

Walker et al,⁶ in a study of 60 patients, found that the higher the opioid dose the patients were on, the more episodes of obstructive sleep apnea and central sleep apnea per hour they had. Yet prescribing a low dose does not adequately protect the chronic opioid user. Farney et al⁷ reported that oxygen saturation dropped precipitously—from 98% to 70%—15 minutes after a patient took just 7.5 mg of hydrocodone in the middle of the night. Mogri et al⁸ reported that a patient had 91 apnea events within 1 hour of taking 15 mg of oxycodone at 2 am.

Opioids, benzodiazepines, barbiturates, and ethanol individually and additively suppress medullary reflex ventilatory drive during sleep, especially during non–rapid-eye-movement (non-REM) sleep.⁶ During waking hours, in contrast, there is redundant backup of cerebral cortical drive, ensuring that we keep breathing. Therefore, people are most vulnerable to dying of opioid ingestion during sleep.

Moreover, oxygen desaturation during episodes of sleep apnea may precipitate seizures (which may be lethal) or coronary vasospasm with consequent malignant arrhythmias and myocardial ischemia.

Continuous positive airway pressure protects against obstructive sleep apnea, but not against central sleep apnea.⁹

Patients need to be aware of the danger, and physicians need to consider the pharmacokinetic profiles of the opioid preparations they prescribe. If patients are taking an opioid that has a short half-life, such as immediate-release oxycodone, they should not take it within 5 hours of sleep. Longer-lasting preparations need a longer interval, and some, such as extended-release tramadol, may need to be taken only on awakening.

Safe sleep can be facilitated by medications that are sedating but do not compromise ventilation. Optimal agents also enhance restorative REM and stage III and IV deep-sleep duration, and some may have the additional benefit of reducing the risk of cancer.^10,11 Such agents may include baclofen, cyproheptadine, gabapentin, mirtazepine, and melatonin. Nonpharmacologic measures include sleep hygiene, aerobic exercise, and cognitive behavioral therapy.

A retrospective study¹² found that 301 (60.4%) of 498 patients who died while on opioid therapy and whose death was judged to be related to the opioid were also taking benzodiazepines. Patients who take opioids should avoid taking benzodiazepines, barbiturates, or alcohol before going to sleep, and physicians should be extremely cautious about prescribing benzodiazepines and barbiturates to patients who are on opioids. 

To the Editor: I enjoyed Dr. Galicia-Castillo’s article about long-term opioid therapy in older adults,¹ which reaffirmed the imperative to “start low and go slow” to minimize the risk of addiction. However, the article missed an opportunity to raise awareness regarding another extremely important side effect of chronic prescription opioid consumption, that of ingestion prior to sleep, with consequent cessation of breathing leading to death.

According to the Drug Enforcement Administration,² most narcotic deaths are a result of respiratory depression. And the American Pain Society has stated, “No patient has succumbed to [opioid] respiratory depression while awake.”³

Dr. Galicia-Castillo noted that the prevalence of central sleep apnea in chronic opioid users is 24%, based on a review by Correa et al.⁴ As alarming as this number is, other investigators have estimated it to be even higher—as high as 50% to 90%.⁵

Walker et al,⁶ in a study of 60 patients, found that the higher the opioid dose the patients were on, the more episodes of obstructive sleep apnea and central sleep apnea per hour they had. Yet prescribing a low dose does not adequately protect the chronic opioid user. Farney et al⁷ reported that oxygen saturation dropped precipitously—from 98% to 70%—15 minutes after a patient took just 7.5 mg of hydrocodone in the middle of the night. Mogri et al⁸ reported that a patient had 91 apnea events within 1 hour of taking 15 mg of oxycodone at 2 am.

Opioids, benzodiazepines, barbiturates, and ethanol individually and additively suppress medullary reflex ventilatory drive during sleep, especially during non–rapid-eye-movement (non-REM) sleep.⁶ During waking hours, in contrast, there is redundant backup of cerebral cortical drive, ensuring that we keep breathing. Therefore, people are most vulnerable to dying of opioid ingestion during sleep.

Moreover, oxygen desaturation during episodes of sleep apnea may precipitate seizures (which may be lethal) or coronary vasospasm with consequent malignant arrhythmias and myocardial ischemia.

Continuous positive airway pressure protects against obstructive sleep apnea, but not against central sleep apnea.⁹

Patients need to be aware of the danger, and physicians need to consider the pharmacokinetic profiles of the opioid preparations they prescribe. If patients are taking an opioid that has a short half-life, such as immediate-release oxycodone, they should not take it within 5 hours of sleep. Longer-lasting preparations need a longer interval, and some, such as extended-release tramadol, may need to be taken only on awakening.

Safe sleep can be facilitated by medications that are sedating but do not compromise ventilation. Optimal agents also enhance restorative REM and stage III and IV deep-sleep duration, and some may have the additional benefit of reducing the risk of cancer.^10,11 Such agents may include baclofen, cyproheptadine, gabapentin, mirtazepine, and melatonin. Nonpharmacologic measures include sleep hygiene, aerobic exercise, and cognitive behavioral therapy.

A retrospective study¹² found that 301 (60.4%) of 498 patients who died while on opioid therapy and whose death was judged to be related to the opioid were also taking benzodiazepines. Patients who take opioids should avoid taking benzodiazepines, barbiturates, or alcohol before going to sleep, and physicians should be extremely cautious about prescribing benzodiazepines and barbiturates to patients who are on opioids. 

To the Editor: I enjoyed Dr. Galicia-Castillo’s article about long-term opioid therapy in older adults,¹ which reaffirmed the imperative to “start low and go slow” to minimize the risk of addiction. However, the article missed an opportunity to raise awareness regarding another extremely important side effect of chronic prescription opioid consumption, that of ingestion prior to sleep, with consequent cessation of breathing leading to death.

According to the Drug Enforcement Administration,² most narcotic deaths are a result of respiratory depression. And the American Pain Society has stated, “No patient has succumbed to [opioid] respiratory depression while awake.”³

Dr. Galicia-Castillo noted that the prevalence of central sleep apnea in chronic opioid users is 24%, based on a review by Correa et al.⁴ As alarming as this number is, other investigators have estimated it to be even higher—as high as 50% to 90%.⁵

Walker et al,⁶ in a study of 60 patients, found that the higher the opioid dose the patients were on, the more episodes of obstructive sleep apnea and central sleep apnea per hour they had. Yet prescribing a low dose does not adequately protect the chronic opioid user. Farney et al⁷ reported that oxygen saturation dropped precipitously—from 98% to 70%—15 minutes after a patient took just 7.5 mg of hydrocodone in the middle of the night. Mogri et al⁸ reported that a patient had 91 apnea events within 1 hour of taking 15 mg of oxycodone at 2 am.

Opioids, benzodiazepines, barbiturates, and ethanol individually and additively suppress medullary reflex ventilatory drive during sleep, especially during non–rapid-eye-movement (non-REM) sleep.⁶ During waking hours, in contrast, there is redundant backup of cerebral cortical drive, ensuring that we keep breathing. Therefore, people are most vulnerable to dying of opioid ingestion during sleep.

Moreover, oxygen desaturation during episodes of sleep apnea may precipitate seizures (which may be lethal) or coronary vasospasm with consequent malignant arrhythmias and myocardial ischemia.

Continuous positive airway pressure protects against obstructive sleep apnea, but not against central sleep apnea.⁹

Patients need to be aware of the danger, and physicians need to consider the pharmacokinetic profiles of the opioid preparations they prescribe. If patients are taking an opioid that has a short half-life, such as immediate-release oxycodone, they should not take it within 5 hours of sleep. Longer-lasting preparations need a longer interval, and some, such as extended-release tramadol, may need to be taken only on awakening.

Safe sleep can be facilitated by medications that are sedating but do not compromise ventilation. Optimal agents also enhance restorative REM and stage III and IV deep-sleep duration, and some may have the additional benefit of reducing the risk of cancer.^10,11 Such agents may include baclofen, cyproheptadine, gabapentin, mirtazepine, and melatonin. Nonpharmacologic measures include sleep hygiene, aerobic exercise, and cognitive behavioral therapy.

A retrospective study¹² found that 301 (60.4%) of 498 patients who died while on opioid therapy and whose death was judged to be related to the opioid were also taking benzodiazepines. Patients who take opioids should avoid taking benzodiazepines, barbiturates, or alcohol before going to sleep, and physicians should be extremely cautious about prescribing benzodiazepines and barbiturates to patients who are on opioids. 

In Reply: Dr. Geller makes some excellent points about sleep and opioid use.

Opioids pose risks,¹ just like any other type of medication. In particular, opioids have been linked to sleep-disordered breathing, which affects 70% to 85% of patients taking opioids.^2–4

Other options can be used in some older adults, but they are not always successful. Ideally, nonpharmacologic strategies and nonopioid medications such as acetaminophen, nonsteroidal anti-inflammatory agents, antidepressants, and anticonvulsants should be used, although these medications have their own side effects. Optimum pain control may offer the potential for significant improvement in function, and opioids are but one tool in the clinician’s kit.

Ongoing discussions of the risks and benefits are necessary, along with continuous re-evaluation of the need for and effect of opioids.

In Reply: Dr. Geller makes some excellent points about sleep and opioid use.

Opioids pose risks,¹ just like any other type of medication. In particular, opioids have been linked to sleep-disordered breathing, which affects 70% to 85% of patients taking opioids.^2–4

Other options can be used in some older adults, but they are not always successful. Ideally, nonpharmacologic strategies and nonopioid medications such as acetaminophen, nonsteroidal anti-inflammatory agents, antidepressants, and anticonvulsants should be used, although these medications have their own side effects. Optimum pain control may offer the potential for significant improvement in function, and opioids are but one tool in the clinician’s kit.

Ongoing discussions of the risks and benefits are necessary, along with continuous re-evaluation of the need for and effect of opioids.

In Reply: Dr. Geller makes some excellent points about sleep and opioid use.

Opioids pose risks,¹ just like any other type of medication. In particular, opioids have been linked to sleep-disordered breathing, which affects 70% to 85% of patients taking opioids.^2–4

Other options can be used in some older adults, but they are not always successful. Ideally, nonpharmacologic strategies and nonopioid medications such as acetaminophen, nonsteroidal anti-inflammatory agents, antidepressants, and anticonvulsants should be used, although these medications have their own side effects. Optimum pain control may offer the potential for significant improvement in function, and opioids are but one tool in the clinician’s kit.

Ongoing discussions of the risks and benefits are necessary, along with continuous re-evaluation of the need for and effect of opioids.

The article by Hagerty and Rich in this issue of the Cleveland Clinic Journal of Medicine¹ covers an important topic—whether elderly patients with atrial fibrillation should receive anticoagulant therapy for it, or whether the risk of bleeding with this therapy outweighs the benefit of preventing stroke.

See related article

BETTER RISK PREDICTORS ARE NEEDED

Prediction tools are available for estimating the risk of stroke in patients with atrial fibrillation without anticoagulation^2,3 and to estimate bleeding risk from anticoagulation^4–7 (Table 1). Both tools have limitations, but as Hagerty and Rich point out, the stroke risk scales are likely better than the bleeding risk scales.

For example, Fang et al⁸ note that the risk of intracranial hemorrhage increases significantly after age 85. The bleeding risk scales use lower age cutoffs than this, perhaps increasing their sensitivity but decreasing their specificity.

Although HAS-BLED^5,6 includes antiplatelet drugs such as nonsteroidal anti-inflammatory drugs and aspirin as risk factors for bleeding, ATRIA⁴ and HEMORR₂HAGES⁷ do not.

Other drugs such as macrolides, quinolones, and high-dose corticosteroids raise the international normalized ratio (INR). These are typically used short-term, but can cause major fluctuations in the INR that may not be detected by monthly INR checks. Incorporating the short-term use of such drugs into bleeding risk scales would be difficult if not impossible a priori. Yet clinicians should be aware that these drugs can affect bleeding risk.

As Hagerty and Rich note,¹ the bleeding risk scores were developed for warfarin, and their applicability to patients treated with novel oral anticoagulants is uncertain.

All three of the available bleeding risk scales consider prior bleeding as a risk factor, but the severity of the prior bleeding varies. Although it is understandable to include major bleeding as a risk factor since it carries an increased risk of death, minor bleeding can affect morbidity and quality of life. Only the ATRIA score⁴ considers both major and minor bleeding, while HEMORR₂HAGES⁷ does not specify bleeding severity, and HAS-BLED^5,6 considers only major bleeding. Clearly, there is a need to update these existing bleeding risk scores so that they can apply to novel oral anticoagulants and consider both major and minor bleeding.

As the authors note, for predicting the risk of stroke, the CHA₂DS₂-VASc score³ provides more precision than the CHADS₂ score² at the lower end of the benefit spectrum. Unfortunately, there is no similar screening tool to predict bleeding risk from anticoagulation with greater precision in the middle to lower part of the risk spectrum.

THE PATIENT’S PREFERENCES MATTER

The patient’s life expectancy and personal preferences are important independent factors that affect the decision of whether to anticoagulate or not. It is the responsibility of clinicians who care for older adults to make sure that these two important considerations are included in any anticoagulation decision-making for this group of patients.

The article by Hagerty and Rich in this issue of the Cleveland Clinic Journal of Medicine¹ covers an important topic—whether elderly patients with atrial fibrillation should receive anticoagulant therapy for it, or whether the risk of bleeding with this therapy outweighs the benefit of preventing stroke.

See related article

BETTER RISK PREDICTORS ARE NEEDED

Prediction tools are available for estimating the risk of stroke in patients with atrial fibrillation without anticoagulation^2,3 and to estimate bleeding risk from anticoagulation^4–7 (Table 1). Both tools have limitations, but as Hagerty and Rich point out, the stroke risk scales are likely better than the bleeding risk scales.

For example, Fang et al⁸ note that the risk of intracranial hemorrhage increases significantly after age 85. The bleeding risk scales use lower age cutoffs than this, perhaps increasing their sensitivity but decreasing their specificity.

Although HAS-BLED^5,6 includes antiplatelet drugs such as nonsteroidal anti-inflammatory drugs and aspirin as risk factors for bleeding, ATRIA⁴ and HEMORR₂HAGES⁷ do not.

Other drugs such as macrolides, quinolones, and high-dose corticosteroids raise the international normalized ratio (INR). These are typically used short-term, but can cause major fluctuations in the INR that may not be detected by monthly INR checks. Incorporating the short-term use of such drugs into bleeding risk scales would be difficult if not impossible a priori. Yet clinicians should be aware that these drugs can affect bleeding risk.

As Hagerty and Rich note,¹ the bleeding risk scores were developed for warfarin, and their applicability to patients treated with novel oral anticoagulants is uncertain.

All three of the available bleeding risk scales consider prior bleeding as a risk factor, but the severity of the prior bleeding varies. Although it is understandable to include major bleeding as a risk factor since it carries an increased risk of death, minor bleeding can affect morbidity and quality of life. Only the ATRIA score⁴ considers both major and minor bleeding, while HEMORR₂HAGES⁷ does not specify bleeding severity, and HAS-BLED^5,6 considers only major bleeding. Clearly, there is a need to update these existing bleeding risk scores so that they can apply to novel oral anticoagulants and consider both major and minor bleeding.

As the authors note, for predicting the risk of stroke, the CHA₂DS₂-VASc score³ provides more precision than the CHADS₂ score² at the lower end of the benefit spectrum. Unfortunately, there is no similar screening tool to predict bleeding risk from anticoagulation with greater precision in the middle to lower part of the risk spectrum.

THE PATIENT’S PREFERENCES MATTER

The patient’s life expectancy and personal preferences are important independent factors that affect the decision of whether to anticoagulate or not. It is the responsibility of clinicians who care for older adults to make sure that these two important considerations are included in any anticoagulation decision-making for this group of patients.

The article by Hagerty and Rich in this issue of the Cleveland Clinic Journal of Medicine¹ covers an important topic—whether elderly patients with atrial fibrillation should receive anticoagulant therapy for it, or whether the risk of bleeding with this therapy outweighs the benefit of preventing stroke.

See related article

BETTER RISK PREDICTORS ARE NEEDED

Prediction tools are available for estimating the risk of stroke in patients with atrial fibrillation without anticoagulation^2,3 and to estimate bleeding risk from anticoagulation^4–7 (Table 1). Both tools have limitations, but as Hagerty and Rich point out, the stroke risk scales are likely better than the bleeding risk scales.

For example, Fang et al⁸ note that the risk of intracranial hemorrhage increases significantly after age 85. The bleeding risk scales use lower age cutoffs than this, perhaps increasing their sensitivity but decreasing their specificity.

Although HAS-BLED^5,6 includes antiplatelet drugs such as nonsteroidal anti-inflammatory drugs and aspirin as risk factors for bleeding, ATRIA⁴ and HEMORR₂HAGES⁷ do not.

Other drugs such as macrolides, quinolones, and high-dose corticosteroids raise the international normalized ratio (INR). These are typically used short-term, but can cause major fluctuations in the INR that may not be detected by monthly INR checks. Incorporating the short-term use of such drugs into bleeding risk scales would be difficult if not impossible a priori. Yet clinicians should be aware that these drugs can affect bleeding risk.

As Hagerty and Rich note,¹ the bleeding risk scores were developed for warfarin, and their applicability to patients treated with novel oral anticoagulants is uncertain.

All three of the available bleeding risk scales consider prior bleeding as a risk factor, but the severity of the prior bleeding varies. Although it is understandable to include major bleeding as a risk factor since it carries an increased risk of death, minor bleeding can affect morbidity and quality of life. Only the ATRIA score⁴ considers both major and minor bleeding, while HEMORR₂HAGES⁷ does not specify bleeding severity, and HAS-BLED^5,6 considers only major bleeding. Clearly, there is a need to update these existing bleeding risk scores so that they can apply to novel oral anticoagulants and consider both major and minor bleeding.

As the authors note, for predicting the risk of stroke, the CHA₂DS₂-VASc score³ provides more precision than the CHADS₂ score² at the lower end of the benefit spectrum. Unfortunately, there is no similar screening tool to predict bleeding risk from anticoagulation with greater precision in the middle to lower part of the risk spectrum.

THE PATIENT’S PREFERENCES MATTER

The patient’s life expectancy and personal preferences are important independent factors that affect the decision of whether to anticoagulate or not. It is the responsibility of clinicians who care for older adults to make sure that these two important considerations are included in any anticoagulation decision-making for this group of patients.

Roflumilast has been shown to reduce rates of acute exacerbation in patients with severe chronic obstructive pulmonary disease (COPD), ie, forced expiratory volume in 1 second (FEV₁) less than 50% with symptoms of chronic bronchitis and a history of exacerbations.

Roflumilast is a selective phosphodiesterase 4 (PDE4) inhibitor that acts on airway smooth muscle cells and various inflammatory cells. By blocking PDE4, roflumilast raises cyclic adenosine monophosphate levels within these cells, curtailing the inflammatory response.^1,2

Roflumilast is not a bronchodilator, although modest improvements in FEV₁ have been documented in clinical trials when it was used as maintenance therapy.

TRIALS OF ROFLUMILAST

Several trials have investigated the efficacy of roflumilast in COPD (Table 1).

The RECORD trial

The RECORD trial¹ in 2005 was the first large randomized controlled trial of roflumilast in moderate to severe COPD. At a dose of 500 µg orally daily, there was a modest but statistically significant improvement in the postbronchodilator FEV₁. There was also improvement in the St. George Respiratory Questionnaire score in the treatment arm, but this was not statistically significant. The study also found a reduction in acute exacerbations of COPD with roflumilast, which was a secondary end point.¹

The results of this study spurred interest in roflumilast as well as criticism of the design of the study. First, COPD patients on inhaled maintenance therapy such as an inhaled corticosteroid and long-acting beta-agonist combination or a long-acting muscarinic antagonist had their medications held during the study. Second, the average FEV₁ was 54% of predicted, indicative of a study population with less severe disease.¹

The RATIO trial

Taking into account the results of the RECORD trial, the RATIO trial³ in 2007 recruited patients with more severe COPD—ie, Global Initiative for Chronic Obstructive Lung Disease (GOLD) class III and IV—and included the rate of acute exacerbations as a primary end point. Maintenance therapy with inhaled corticosteroids was continued in patients already taking them. However, long-acting beta-agonists and long-acting muscarinic antagonist therapies were held.³

Again, roflumilast improved postbronchodilator FEV₁ compared with placebo. A reduction in acute exacerbations was seen but was not statistically significant except in subgroup analysis, where a statistically significant reduction in acute exacerbations was noted for patients with very severe (GOLD class IV) COPD.³

Post hoc analysis from the RATIO trial suggested that patients with chronic bronchitis and patients with a history of frequent exacerbations were more likely to respond to roflumilast.²

The EOS and HELIOS trials

In 2009, the results of the EOS and HELIOS trials of roflumilast in patients with severe COPD were published.⁴ These trials allowed continuation of long-acting beta-agonists and muscarinic antagonists. The prebronchodilator FEV₁ improved modestly when roflumilast was added to a long-acting bronchodilator. These studies ran for only 24 weeks, and the rate of acute exacerbations was not a primary end point, although the results did show a trend toward reduction of exacerbations.⁴

The AURA and HERMES trials

Also in 2009 was the publication of the results of two 52-week placebo-controlled trials (AURA and HERMES) of roflumilast in patients with severe COPD with chronic bronchitis and a history of frequent exacerbations.⁵ Maintenance therapy with long-acting beta-agonists was continued, whereas inhaled corticosteroids and long-acting muscarinic antagonists were held. Statistically significant improvements in prebronchodilator FEV₁ and reduction in the rate of exacerbations were observed in the roflumilast group (17% reduction, 95% confidence interval 8–25, P < .0003).⁵

The REACT trial

The REACT trial⁶ randomized 1,945 patients with severe COPD already on maximal recommended combination inhaled corticosteroid and long-acting beta-agonist therapy to receive either roflumilast or placebo. The patients’ ratio of FEV₁ to forced vital capacity was less than 70%, their postbronchodilator FEV₁ was less than 50%, and they had chronic bronchitis and a history of at least two acute exacerbations during the past year. They had also been on combination therapy for the previous year. Patients who were on long-acting muscarinic-antagonist therapy (70% of the cohort) were included, and continued with their medication.

Patients were followed for 52 weeks. There was a significant reduction in the rate of exacerbations in the roflumilast group vs placebo (0.823 vs 0.959; risk ratio 0.858; 95% confidence interval 0.740–0.995; P = .0424).⁶ As in previous trials, the roflumilast group showed an improvement in postbronchodilator FEV₁. The study also showed a reduction in hospital admissions in the treatment group.⁶

ADVERSE EFFECTS OF ROFLUMILAST

Roflumilast is known to have adverse effects significant enough to reduce compliance, the most common being diarrhea, weight loss, and nausea.^2,6,7 In the REACT trial,⁶ 11% of patients in the roflumilast group vs 5% in the placebo group dropped out of the study because of adverse drug effects. Diarrhea was reported in 10% and weight loss in 9% of patients taking roflumilast. Weight loss has been shown to be reversible upon stopping roflumilast.² There has been no evidence of increased risk of death or serious adverse events in studies of roflumilast in patients with COPD.² However, the benefit-to-harm ratio suggests that roflumilast provides a net benefit only in patients at high risk of severe exacerbations.⁷

Roflumilast has been shown to reduce rates of acute exacerbation in patients with severe chronic obstructive pulmonary disease (COPD), ie, forced expiratory volume in 1 second (FEV₁) less than 50% with symptoms of chronic bronchitis and a history of exacerbations.

Roflumilast is a selective phosphodiesterase 4 (PDE4) inhibitor that acts on airway smooth muscle cells and various inflammatory cells. By blocking PDE4, roflumilast raises cyclic adenosine monophosphate levels within these cells, curtailing the inflammatory response.^1,2

Roflumilast is not a bronchodilator, although modest improvements in FEV₁ have been documented in clinical trials when it was used as maintenance therapy.

TRIALS OF ROFLUMILAST

Several trials have investigated the efficacy of roflumilast in COPD (Table 1).

The RECORD trial

The RECORD trial¹ in 2005 was the first large randomized controlled trial of roflumilast in moderate to severe COPD. At a dose of 500 µg orally daily, there was a modest but statistically significant improvement in the postbronchodilator FEV₁. There was also improvement in the St. George Respiratory Questionnaire score in the treatment arm, but this was not statistically significant. The study also found a reduction in acute exacerbations of COPD with roflumilast, which was a secondary end point.¹

The results of this study spurred interest in roflumilast as well as criticism of the design of the study. First, COPD patients on inhaled maintenance therapy such as an inhaled corticosteroid and long-acting beta-agonist combination or a long-acting muscarinic antagonist had their medications held during the study. Second, the average FEV₁ was 54% of predicted, indicative of a study population with less severe disease.¹

The RATIO trial

Taking into account the results of the RECORD trial, the RATIO trial³ in 2007 recruited patients with more severe COPD—ie, Global Initiative for Chronic Obstructive Lung Disease (GOLD) class III and IV—and included the rate of acute exacerbations as a primary end point. Maintenance therapy with inhaled corticosteroids was continued in patients already taking them. However, long-acting beta-agonists and long-acting muscarinic antagonist therapies were held.³

Again, roflumilast improved postbronchodilator FEV₁ compared with placebo. A reduction in acute exacerbations was seen but was not statistically significant except in subgroup analysis, where a statistically significant reduction in acute exacerbations was noted for patients with very severe (GOLD class IV) COPD.³

Post hoc analysis from the RATIO trial suggested that patients with chronic bronchitis and patients with a history of frequent exacerbations were more likely to respond to roflumilast.²

The EOS and HELIOS trials

In 2009, the results of the EOS and HELIOS trials of roflumilast in patients with severe COPD were published.⁴ These trials allowed continuation of long-acting beta-agonists and muscarinic antagonists. The prebronchodilator FEV₁ improved modestly when roflumilast was added to a long-acting bronchodilator. These studies ran for only 24 weeks, and the rate of acute exacerbations was not a primary end point, although the results did show a trend toward reduction of exacerbations.⁴

The AURA and HERMES trials

Also in 2009 was the publication of the results of two 52-week placebo-controlled trials (AURA and HERMES) of roflumilast in patients with severe COPD with chronic bronchitis and a history of frequent exacerbations.⁵ Maintenance therapy with long-acting beta-agonists was continued, whereas inhaled corticosteroids and long-acting muscarinic antagonists were held. Statistically significant improvements in prebronchodilator FEV₁ and reduction in the rate of exacerbations were observed in the roflumilast group (17% reduction, 95% confidence interval 8–25, P < .0003).⁵

The REACT trial

The REACT trial⁶ randomized 1,945 patients with severe COPD already on maximal recommended combination inhaled corticosteroid and long-acting beta-agonist therapy to receive either roflumilast or placebo. The patients’ ratio of FEV₁ to forced vital capacity was less than 70%, their postbronchodilator FEV₁ was less than 50%, and they had chronic bronchitis and a history of at least two acute exacerbations during the past year. They had also been on combination therapy for the previous year. Patients who were on long-acting muscarinic-antagonist therapy (70% of the cohort) were included, and continued with their medication.

Patients were followed for 52 weeks. There was a significant reduction in the rate of exacerbations in the roflumilast group vs placebo (0.823 vs 0.959; risk ratio 0.858; 95% confidence interval 0.740–0.995; P = .0424).⁶ As in previous trials, the roflumilast group showed an improvement in postbronchodilator FEV₁. The study also showed a reduction in hospital admissions in the treatment group.⁶

ADVERSE EFFECTS OF ROFLUMILAST

Roflumilast is known to have adverse effects significant enough to reduce compliance, the most common being diarrhea, weight loss, and nausea.^2,6,7 In the REACT trial,⁶ 11% of patients in the roflumilast group vs 5% in the placebo group dropped out of the study because of adverse drug effects. Diarrhea was reported in 10% and weight loss in 9% of patients taking roflumilast. Weight loss has been shown to be reversible upon stopping roflumilast.² There has been no evidence of increased risk of death or serious adverse events in studies of roflumilast in patients with COPD.² However, the benefit-to-harm ratio suggests that roflumilast provides a net benefit only in patients at high risk of severe exacerbations.⁷

Roflumilast has been shown to reduce rates of acute exacerbation in patients with severe chronic obstructive pulmonary disease (COPD), ie, forced expiratory volume in 1 second (FEV₁) less than 50% with symptoms of chronic bronchitis and a history of exacerbations.

Roflumilast is a selective phosphodiesterase 4 (PDE4) inhibitor that acts on airway smooth muscle cells and various inflammatory cells. By blocking PDE4, roflumilast raises cyclic adenosine monophosphate levels within these cells, curtailing the inflammatory response.^1,2

Roflumilast is not a bronchodilator, although modest improvements in FEV₁ have been documented in clinical trials when it was used as maintenance therapy.

TRIALS OF ROFLUMILAST

Several trials have investigated the efficacy of roflumilast in COPD (Table 1).

The RECORD trial

The RECORD trial¹ in 2005 was the first large randomized controlled trial of roflumilast in moderate to severe COPD. At a dose of 500 µg orally daily, there was a modest but statistically significant improvement in the postbronchodilator FEV₁. There was also improvement in the St. George Respiratory Questionnaire score in the treatment arm, but this was not statistically significant. The study also found a reduction in acute exacerbations of COPD with roflumilast, which was a secondary end point.¹

The results of this study spurred interest in roflumilast as well as criticism of the design of the study. First, COPD patients on inhaled maintenance therapy such as an inhaled corticosteroid and long-acting beta-agonist combination or a long-acting muscarinic antagonist had their medications held during the study. Second, the average FEV₁ was 54% of predicted, indicative of a study population with less severe disease.¹

The RATIO trial

Taking into account the results of the RECORD trial, the RATIO trial³ in 2007 recruited patients with more severe COPD—ie, Global Initiative for Chronic Obstructive Lung Disease (GOLD) class III and IV—and included the rate of acute exacerbations as a primary end point. Maintenance therapy with inhaled corticosteroids was continued in patients already taking them. However, long-acting beta-agonists and long-acting muscarinic antagonist therapies were held.³

Again, roflumilast improved postbronchodilator FEV₁ compared with placebo. A reduction in acute exacerbations was seen but was not statistically significant except in subgroup analysis, where a statistically significant reduction in acute exacerbations was noted for patients with very severe (GOLD class IV) COPD.³

Post hoc analysis from the RATIO trial suggested that patients with chronic bronchitis and patients with a history of frequent exacerbations were more likely to respond to roflumilast.²

The EOS and HELIOS trials

In 2009, the results of the EOS and HELIOS trials of roflumilast in patients with severe COPD were published.⁴ These trials allowed continuation of long-acting beta-agonists and muscarinic antagonists. The prebronchodilator FEV₁ improved modestly when roflumilast was added to a long-acting bronchodilator. These studies ran for only 24 weeks, and the rate of acute exacerbations was not a primary end point, although the results did show a trend toward reduction of exacerbations.⁴

The AURA and HERMES trials

Also in 2009 was the publication of the results of two 52-week placebo-controlled trials (AURA and HERMES) of roflumilast in patients with severe COPD with chronic bronchitis and a history of frequent exacerbations.⁵ Maintenance therapy with long-acting beta-agonists was continued, whereas inhaled corticosteroids and long-acting muscarinic antagonists were held. Statistically significant improvements in prebronchodilator FEV₁ and reduction in the rate of exacerbations were observed in the roflumilast group (17% reduction, 95% confidence interval 8–25, P < .0003).⁵

The REACT trial

The REACT trial⁶ randomized 1,945 patients with severe COPD already on maximal recommended combination inhaled corticosteroid and long-acting beta-agonist therapy to receive either roflumilast or placebo. The patients’ ratio of FEV₁ to forced vital capacity was less than 70%, their postbronchodilator FEV₁ was less than 50%, and they had chronic bronchitis and a history of at least two acute exacerbations during the past year. They had also been on combination therapy for the previous year. Patients who were on long-acting muscarinic-antagonist therapy (70% of the cohort) were included, and continued with their medication.

Patients were followed for 52 weeks. There was a significant reduction in the rate of exacerbations in the roflumilast group vs placebo (0.823 vs 0.959; risk ratio 0.858; 95% confidence interval 0.740–0.995; P = .0424).⁶ As in previous trials, the roflumilast group showed an improvement in postbronchodilator FEV₁. The study also showed a reduction in hospital admissions in the treatment group.⁶

ADVERSE EFFECTS OF ROFLUMILAST

Roflumilast is known to have adverse effects significant enough to reduce compliance, the most common being diarrhea, weight loss, and nausea.^2,6,7 In the REACT trial,⁶ 11% of patients in the roflumilast group vs 5% in the placebo group dropped out of the study because of adverse drug effects. Diarrhea was reported in 10% and weight loss in 9% of patients taking roflumilast. Weight loss has been shown to be reversible upon stopping roflumilast.² There has been no evidence of increased risk of death or serious adverse events in studies of roflumilast in patients with COPD.² However, the benefit-to-harm ratio suggests that roflumilast provides a net benefit only in patients at high risk of severe exacerbations.⁷