Bipolar Disorder

Despite numerous drug treatment innovations, just about all patients with bipolar disorder that I have treated reported improvement after starting a lithium regimen.

In the 1970s, the clinical literature began highlighting numerous drug therapies for different bipolar symptoms. Before then, I had been taught to treat all “manic-depressive” patients with lithium—regardless of whether mood swings, bouts of anger, depression, or mania were the dominant symptoms.

So I experimented. I prescribed lithium to any potential bipolar patient who did not meet DSM criteria for another mental illness. I discovered the following:

1. A family history of any mental illness, especially alcohol abuse and depression, is a strong indicator of bipolar disorder and of potential positive response to lithium.
2. The existence of mood swings, especially without cause, confirms the diagnosis of bipolar disorder when paired with family history.
3. Lithium, 900 mg/d, works fine as acute or maintenance therapy. I would decrease the dosage for smaller people (eg, 600 mg/d for a patient weighing approximately 125 lbs). I would only increase the dosage—to 1,200 mg/d—for patients with severe mania.
4. Gauging lithium blood levels is a waste of time, assuming you have checked for kidney disease. Across 3 decades in practice, the only patient I have ever seen with an abnormally high lithium blood count also suffered renal failure.
5. Side effects I have seen most commonly with lithium are:
   • weight gain in women, in which case another medication should be prescribed
   • tremor, which should warrant a check of the patient’s caffeine intake.

Other side effects (such as diarrhea and GI upset) are usually mild and easy to control by adding other medications.

Despite numerous drug treatment innovations, just about all patients with bipolar disorder that I have treated reported improvement after starting a lithium regimen.

In the 1970s, the clinical literature began highlighting numerous drug therapies for different bipolar symptoms. Before then, I had been taught to treat all “manic-depressive” patients with lithium—regardless of whether mood swings, bouts of anger, depression, or mania were the dominant symptoms.

So I experimented. I prescribed lithium to any potential bipolar patient who did not meet DSM criteria for another mental illness. I discovered the following:

1. A family history of any mental illness, especially alcohol abuse and depression, is a strong indicator of bipolar disorder and of potential positive response to lithium.
2. The existence of mood swings, especially without cause, confirms the diagnosis of bipolar disorder when paired with family history.
3. Lithium, 900 mg/d, works fine as acute or maintenance therapy. I would decrease the dosage for smaller people (eg, 600 mg/d for a patient weighing approximately 125 lbs). I would only increase the dosage—to 1,200 mg/d—for patients with severe mania.
4. Gauging lithium blood levels is a waste of time, assuming you have checked for kidney disease. Across 3 decades in practice, the only patient I have ever seen with an abnormally high lithium blood count also suffered renal failure.
5. Side effects I have seen most commonly with lithium are:
   • weight gain in women, in which case another medication should be prescribed
   • tremor, which should warrant a check of the patient’s caffeine intake.

Other side effects (such as diarrhea and GI upset) are usually mild and easy to control by adding other medications.

Despite numerous drug treatment innovations, just about all patients with bipolar disorder that I have treated reported improvement after starting a lithium regimen.

In the 1970s, the clinical literature began highlighting numerous drug therapies for different bipolar symptoms. Before then, I had been taught to treat all “manic-depressive” patients with lithium—regardless of whether mood swings, bouts of anger, depression, or mania were the dominant symptoms.

So I experimented. I prescribed lithium to any potential bipolar patient who did not meet DSM criteria for another mental illness. I discovered the following:

1. A family history of any mental illness, especially alcohol abuse and depression, is a strong indicator of bipolar disorder and of potential positive response to lithium.
2. The existence of mood swings, especially without cause, confirms the diagnosis of bipolar disorder when paired with family history.
3. Lithium, 900 mg/d, works fine as acute or maintenance therapy. I would decrease the dosage for smaller people (eg, 600 mg/d for a patient weighing approximately 125 lbs). I would only increase the dosage—to 1,200 mg/d—for patients with severe mania.
4. Gauging lithium blood levels is a waste of time, assuming you have checked for kidney disease. Across 3 decades in practice, the only patient I have ever seen with an abnormally high lithium blood count also suffered renal failure.
5. Side effects I have seen most commonly with lithium are:
   • weight gain in women, in which case another medication should be prescribed
   • tremor, which should warrant a check of the patient’s caffeine intake.

Other side effects (such as diarrhea and GI upset) are usually mild and easy to control by adding other medications.

Prescribing drug therapy for pregnant bipolar women requires psychiatrists to balance the potential for neonatal malformations against the high risk of relapse when patients discontinue their medications.¹ To help you achieve this balance, we offer an evidence-based approach that includes:

• analysis of the FDA’s teratogenicity categories for psychotropics
• review of the safety profiles of drugs used in mood stabilization
• an algorithm for managing patients who are considering conception or are pregnant.

PSYCHOTROPIC RISKS TO OFFSPRING

All psychotropic medications diffuse across the placenta, which exposes the fetus to some degree. Risks include teratogenicity, obstetrical complications, perinatal syndromes, and long-term postnatal behavioral sequelae.

Teratogenicity. A medication is considered teratogenic when prenatal exposure significantly increases the risk of congenital deformities over the baseline risk, which is 2% in the United States.² The cause of most congenital malformations is unknown. Risk for teratogenicity occurs in the first 12 weeks of gestation, as organs are formed.

Table 1

FDA Use-in-Pregnancy ratings for medications The FDA system weighs the degree to which research findings have ruled out risk to the fetus

Obstetrical complications include preterm delivery, low birth weight, and delivery complications such as low Apgar scores or behavioral effects requiring intensive care.

Perinatal syndromes include physical and behavioral symptoms noticed shortly after birth (such as jitteriness). These consequences are putatively related to drug use at or near birth and have limited duration.

Postnatal behavioral sequelae include long-term neurobehavioral abnormalities in children who were exposed to psychotropics in utero.

BALANCING RISKS

Risks with medication. The FDA’s “use in pregnancy” rating system (Table 1) uses available data to assess the degree of teratogenic risk. These guidelines can be confusing and are one of many tools to use when considering a possible drug treatment.

Most psychotropics are category “C” or “D,” which imply a chance of harm to the exposed fetus. Category “B” drugs would appear safer, but this rating could simply indicate a lack of adequate human data or that no data have shown harm in animals.

Moreover, a category “D” drug may be chosen more often during pregnancy than a category “C” drug. This may occur when more human data exist on using the category “D” drug in patients with a particular disorder (such as using lithium versus valproate or olanzapine in pregnant bipolar women).

No psychotropics are classified as “A,” meaning either some risks are associated with every psychotropic or the risk of some agents has not been adequately explored. Furthermore, no psychotropics are FDA-approved for use during pregnancy.

Risks without medication. Teratogenicity notwithstanding, psychotropic intervention is the most effective treatment for women with bipolar disorder. Patients who discontinue mood-stabilizing medication after conception increase their risk of relapse into depression or mania,³ either of which could lead to complications and untoward effects on the fetus.

Depression during pregnancy has been linked to low birth weight and preterm delivery.^4,5 These effects may be mediated by the illness itself or by other factors that indirectly affect birth outcomes. For example, depression during pregnancy is associated with decreased appetite, substance use and abuse, and lower use of prenatal care.⁶

Untreated mania may also be associated with perinatal risks, as a pregnant patient in a manic state may engage in impulsive, high-risk behaviors that endanger her and the fetus.⁷

MOOD STABILIZERS

The FDA categorizes as “D” the three most commonly used mood stabilizers: lithium, valproate, and carbamazepine (Table 2). This rating implies that studies have demonstrated fetal risk but the drug’s potential benefit may still outweigh the risk.

Lithium. The International Registry of Lithium reported increased rates of cardiovascular malformations— such as Ebstein’s anomaly—in children whose mothers took lithium during pregnancy.

Relative risk for Ebstein’s anomaly in children with fetal exposure to lithium may be 20 times higher than the risk in unexposed children, although the absolute risk with lithium exposure remains low (1 in 1,000 births).^1,8

No significant neurobehavioral teratogenicity has been reported in infants exposed in utero to lithium, although few cases have been studied. One study reported that 22 lithium-exposed infants attained developmental milestones at a pace comparable to that of unexposed controls.⁹

“Floppy baby” syndrome, in which infants experience hypotonicity and cyanosis, is the most recognized adverse effect in infants exposed to lithium in utero.¹⁰ Its frequency is unknown, but rare. Neonatal hypothyroidism and nephrogenic diabetes insipidus have also been documented.

Anticonvulsants. To date, no studies have examined the outcomes of children whose mothers took anticonvulsants for bipolar disorder during pregnancy, though the research concerning epileptic mothers is extensive.

Neural tube defects. Data associate anticonvulsant exposure with a significantly greater risk for malformations than in the general population. Specifically, anticonvulsants may cause neural tube defects such as spina bifida, ancephaly, and encephaly in 2 to 5% of those exposed, as well as craniofacial anomalies, microcephaly, growth retardation, and heart defects.^11-14

Table 2

FDA’s teratogenicity ratings of mood stabilizers and other antimanic agents

More minor malformations—such as rotated ears, depressed nasal bridge, short nose, elongated upper lip, and fingernail hypoplasia—have been reported in infants exposed to anticonvulsants in utero.¹⁴ These malformations disappear with age.¹³ Teratogenicity increases with the use of multiple anticonvulsants and possibly with higher maternal plasma levels and toxic metabolites.¹⁵

Conclusion. The three most commonly used mood stabilizers are all teratogenic. The least risk may occur with lithium (0.1%) versus valproate (2 to 5%) or carbamazepine (1 to 3%). These risks must be weighed against the up to 50% chance of relapse with medication discontinuation.³

ANTIPSYCHOTICS

Antipsychotics are often used to treat mania because of their rapid effects and sedative properties. Most antipsychotics—specifically, haloperidol, olanzapine, and risperidone—are labeled “C,” specifying that fetal risk cannot be ruled out.

Table 3

FDA’s teratogenicity ratings of common antidepressants

Chlorpromazine and haloperidol have been most studied during pregnancy but in relation to treating hyperemesis gravidarum and psychosis, not bipolar disorder. Results regarding antipsychotics’ teratogenic and behavioral risks are mixed,^16-21 probably because the various compounds have different effects on the fetus.

The underlying illness—rather than the medications—may increase the rate of anomalies seen with exposure to antipsychotics:

• Rieder et al²² reported an increased rate of perinatal death in infants of schizophrenic mothers but no significant association between the mothers’ use of antipsychotics and perinatal death.
• Sobel²³ compared psychotic women with and without histories of chlorpromazine exposure during pregnancy. Rates of fetal damage were similar and approximately twice that of the general population.

A meta-analysis of 74,337 live births revealed that first-trimester exposure to low-potency antipsychotics increases the relative risk of fetal anomalies in nonpsychotic women. Phenothiazines may increase the 2% baseline incidence of malformations to 2.4%.¹ No specific organ malformation following fetal exposure to phenothiazines has been consistently identified.

Olanzapine was recently approved for treating mania. Very little data exist regarding its impact on fetal development when used during pregnancy, although studies on small numbers of women have not revealed teratogenicity.^24,25

Conclusion. Psychotic illness itself may increase the risk of poor fetal outcome to a greater extent than does antipsychotic use. Prenatal exposure to low-potency phenothiazines may further increase this risk, although only slightly. The effect of prenatal exposure to atypical antipsychotics requires further study.

BENZODIAZEPINES

Benzodiazepines are rarely a primary treatment for mania or depression. Thus, a comprehensive review of their effect on fetal outcome is beyond the scope of this review. A meta-analysis of exposure in the first trimester suggests a very small but significant increase in risk for cleft palate.¹ The absolute risk is <1 in 1,000 cases.

ANTIDEPRESSANTS

Whereas treatment of acute mania is considered a medical emergency, women with bipolar disorder may also relapse into depression during pregnancy. An antidepressant should not be used without a mood stabilizer when treating bipolar I disorder, although a mood stabilizer alone may be inadequate to treat depression. Using tricyclics and selective serotonin reuptake inhibitors (SSRIs) during pregnancy has not been associated with teratogenicity (Table 3),²⁶ although perinatal effects have been reported.¹

Tricyclics. In case-control studies involving more than 300,000 live births, 414 incidences of first-trimester exposure to tricyclics were followed. Information from these patients found no significant association between fetal exposure to tricyclics and increased rates of congenital malformations.¹ The few studies that have been performed suggest no long-term effects from in utero exposure.²⁶ Although these results suggest that prenatal exposure to tricyclics is relatively safe, more research is needed.

SSRIs. To date, no significant teratogenic effects of SSRIs have been identified in offspring of treated women.

The manufacturer’s register for fluoxetine contains approximately 2,000 cases of treated patients, with no excess cases of congenital anomalies or malformations following prenatal exposure. Citalopram has the next largest database of in utero exposure (n=365), again with no increased risk for teratogenicity. Several smaller systematic reports are available on in utero exposure to sertraline, paroxetine, or escitalopram.²⁶

Most studies of pregnant women taking fluoxetine in the first trimester have found no increased risk of obstetrical complications—including spontaneous pregnancy loss, preterm labor, or low birth weight—compared with women not taking fluoxetine. Taking fluoxetine during the third trimester may increase the risk for perinatal complications,²⁷ although this has been inconsistently reported and requires further study. Effects of other SSRIs in the third trimester have not been systematically explored.

Case reports and one controlled study have addressed possible neonatal perinatal symptoms from in utero exposure to SSRIs.^28,29 Preliminary data show no adverse neurobehavioral function in exposed neonates.²⁶

Electroconvulsive therapy (ECT) has been proven effective for acute mania and depression, demonstrating few deleterious effects on neonates. ECT has few side effects and may be safer than drug therapy in this population. Two reviews support the efficacy and relative safety of ECT treatment during pregnancy, although more evidence is needed.^30,31

RECOMMENDATIONS

Discuss pregnancy and medication risks with all bipolar women, regardless of proximal plans for pregnancy. If psychotropic medication is used, prescribe carefully during the first trimester, using the minimum number of drugs and the lowest dosages needed to restore or maintain well-being.³²

Pros and cons of switching. Some clinicians may encourage a patient to taper a medication during the first trimester because of its unknown or high teratogenicity. Depending on the patient’s illness severity, this might not be the optimal decision. A more conservative option would be to switch to a lower-risk drug during pregnancy.

Lithium has both antidepressant and antimanic properties and is less teratogenic compared with first-trimester exposure to an anticonvulsant. However, if lithium has not been successful for the woman’s mania prophylaxis in the past and she has demonstrated antimanic response to an anticonvulsant, switching to lithium or another anticonvulsant is not recommended.

Algorithm Suggested approach to the bipolar patient who wishes to conceive or is pregnant

Low maternal folate levels are often associated with neural tube defects from any cause.³³ Valproate lowers folate levels by inhibiting one of the enzymes necessary for its formation, which may be a mechanism for the increased risk of spina bifida.³⁴

Folate supplementation. To date, no study has demonstrated that giving folate supplements to women taking anticonvulsants during pregnancy reduces the risk of neural tube defects.³⁵ Nonetheless, we recommend that women who continue to take valproate or carbamazepine during pregnancy receive folate, 3 to 4 mg/d, as a precaution.

Treating manic relapse. Data show high rates of relapse in patients who stop taking lithium, particularly if done abruptly.³ Counsel women taking lithium to plan their pregnancies to allow enough time to taper off the medication prior to conception, if they want to try this. Lithium should be decreased slowly—approximately 50% every 2 weeks—to avoid relapse.

Treat aggressively if relapse occurs during pregnancy. Consider:

• psychiatric hospitalization in case of suicidality or psychosis
• reinstituting drug therapy with a less-teratogenic agent
• ECT for a manic or depressive episode.

As the pregnancy advances and the mother’s volume of distribution increases, dosage increases may be needed to maintain therapeutic drug levels.

Treating depressive relapse. Should depression occur in pregnancy, SSRIs or tricyclics added to mood stabilizer therapy have been shown to be effective, with few teratogenic effects.

Cognitive-behavioral and interpersonal psychotherapies also have shown efficacy in pregnant women with major depressive disorder³⁶ and may be effective for women with bipolar disorder in pregnancy. Cognitive psychotherapies, when used with medication, have been reported effective in preventing relapse in nongravid bipolar patients.^36-37

Related resources

Psychiatric disorders during pregnancy. Massachusetts General Hospital Center for Women’s Health. Perinatal Resource Center. www.womensmentalhealth.org/topics/pregnancy_lib.html

Drug brand names

• Amitriptyline • Elavil
• Bupropion • Wellbutrin
• Carbamazepine • Tegretol
• Chlorpromazine • Thorazine
• Citalopram • Celexa
• Clomipramine • Anafranil
• Desipramine • Norpramin
• Fluoxetine • Prozac
• Fluvoxamine • Luvox
• Gabapentin • Neurontin
• Haloperidol • Haldol
• Imipramine • Tofranil
• Lamotrigine • Lamictal
• Lithium • Lithobid et al
• Methylphenidate • Ritalin et al
• Nortriptyline • Pamelor
• Olanzapine • Zyprexa
• Paroxetine • Paxil
• Phenelzine • Nardil
• Risperidone • Risperdal
• Sertraline • Zoloft
• Topiramate • Topamax
• Tranylcypromine • Parnate
• Trifluoperazine • Stelazine
• Valproate • Depakote et al

Disclosure

Dr. Altshuler receives research support from Abbott Laboratories, is a consultant to Abbott Laboratories, Forest Laboratories, and Eli Lilly & Co., and is a speaker for GlaxoSmithKline and Janssen Pharmaceutica.

Ms. Richards reports no financial relationship with any company whose products are mentioned in this article, or with manufacturers of competing products.

Dr. Yonkers receives research support from GlaxoSmithKline and Berlex Laboratories, is a consultant to GlaxoSmithKline, and is a speaker for Eli Lilly and Co., Pfizer Inc., GlaxoSmithKline, and Wyeth Pharmaceuticals.

Prescribing drug therapy for pregnant bipolar women requires psychiatrists to balance the potential for neonatal malformations against the high risk of relapse when patients discontinue their medications.¹ To help you achieve this balance, we offer an evidence-based approach that includes:

• analysis of the FDA’s teratogenicity categories for psychotropics
• review of the safety profiles of drugs used in mood stabilization
• an algorithm for managing patients who are considering conception or are pregnant.

PSYCHOTROPIC RISKS TO OFFSPRING

All psychotropic medications diffuse across the placenta, which exposes the fetus to some degree. Risks include teratogenicity, obstetrical complications, perinatal syndromes, and long-term postnatal behavioral sequelae.

Teratogenicity. A medication is considered teratogenic when prenatal exposure significantly increases the risk of congenital deformities over the baseline risk, which is 2% in the United States.² The cause of most congenital malformations is unknown. Risk for teratogenicity occurs in the first 12 weeks of gestation, as organs are formed.

Table 1

FDA Use-in-Pregnancy ratings for medications The FDA system weighs the degree to which research findings have ruled out risk to the fetus

Obstetrical complications include preterm delivery, low birth weight, and delivery complications such as low Apgar scores or behavioral effects requiring intensive care.

Perinatal syndromes include physical and behavioral symptoms noticed shortly after birth (such as jitteriness). These consequences are putatively related to drug use at or near birth and have limited duration.

Postnatal behavioral sequelae include long-term neurobehavioral abnormalities in children who were exposed to psychotropics in utero.

BALANCING RISKS

Risks with medication. The FDA’s “use in pregnancy” rating system (Table 1) uses available data to assess the degree of teratogenic risk. These guidelines can be confusing and are one of many tools to use when considering a possible drug treatment.

Most psychotropics are category “C” or “D,” which imply a chance of harm to the exposed fetus. Category “B” drugs would appear safer, but this rating could simply indicate a lack of adequate human data or that no data have shown harm in animals.

Moreover, a category “D” drug may be chosen more often during pregnancy than a category “C” drug. This may occur when more human data exist on using the category “D” drug in patients with a particular disorder (such as using lithium versus valproate or olanzapine in pregnant bipolar women).

No psychotropics are classified as “A,” meaning either some risks are associated with every psychotropic or the risk of some agents has not been adequately explored. Furthermore, no psychotropics are FDA-approved for use during pregnancy.

Risks without medication. Teratogenicity notwithstanding, psychotropic intervention is the most effective treatment for women with bipolar disorder. Patients who discontinue mood-stabilizing medication after conception increase their risk of relapse into depression or mania,³ either of which could lead to complications and untoward effects on the fetus.

Depression during pregnancy has been linked to low birth weight and preterm delivery.^4,5 These effects may be mediated by the illness itself or by other factors that indirectly affect birth outcomes. For example, depression during pregnancy is associated with decreased appetite, substance use and abuse, and lower use of prenatal care.⁶

Untreated mania may also be associated with perinatal risks, as a pregnant patient in a manic state may engage in impulsive, high-risk behaviors that endanger her and the fetus.⁷

MOOD STABILIZERS

The FDA categorizes as “D” the three most commonly used mood stabilizers: lithium, valproate, and carbamazepine (Table 2). This rating implies that studies have demonstrated fetal risk but the drug’s potential benefit may still outweigh the risk.

Lithium. The International Registry of Lithium reported increased rates of cardiovascular malformations— such as Ebstein’s anomaly—in children whose mothers took lithium during pregnancy.

Relative risk for Ebstein’s anomaly in children with fetal exposure to lithium may be 20 times higher than the risk in unexposed children, although the absolute risk with lithium exposure remains low (1 in 1,000 births).^1,8

No significant neurobehavioral teratogenicity has been reported in infants exposed in utero to lithium, although few cases have been studied. One study reported that 22 lithium-exposed infants attained developmental milestones at a pace comparable to that of unexposed controls.⁹

“Floppy baby” syndrome, in which infants experience hypotonicity and cyanosis, is the most recognized adverse effect in infants exposed to lithium in utero.¹⁰ Its frequency is unknown, but rare. Neonatal hypothyroidism and nephrogenic diabetes insipidus have also been documented.

Anticonvulsants. To date, no studies have examined the outcomes of children whose mothers took anticonvulsants for bipolar disorder during pregnancy, though the research concerning epileptic mothers is extensive.

Neural tube defects. Data associate anticonvulsant exposure with a significantly greater risk for malformations than in the general population. Specifically, anticonvulsants may cause neural tube defects such as spina bifida, ancephaly, and encephaly in 2 to 5% of those exposed, as well as craniofacial anomalies, microcephaly, growth retardation, and heart defects.^11-14

Table 2

FDA’s teratogenicity ratings of mood stabilizers and other antimanic agents

More minor malformations—such as rotated ears, depressed nasal bridge, short nose, elongated upper lip, and fingernail hypoplasia—have been reported in infants exposed to anticonvulsants in utero.¹⁴ These malformations disappear with age.¹³ Teratogenicity increases with the use of multiple anticonvulsants and possibly with higher maternal plasma levels and toxic metabolites.¹⁵

Conclusion. The three most commonly used mood stabilizers are all teratogenic. The least risk may occur with lithium (0.1%) versus valproate (2 to 5%) or carbamazepine (1 to 3%). These risks must be weighed against the up to 50% chance of relapse with medication discontinuation.³

ANTIPSYCHOTICS

Antipsychotics are often used to treat mania because of their rapid effects and sedative properties. Most antipsychotics—specifically, haloperidol, olanzapine, and risperidone—are labeled “C,” specifying that fetal risk cannot be ruled out.

Table 3

FDA’s teratogenicity ratings of common antidepressants

Chlorpromazine and haloperidol have been most studied during pregnancy but in relation to treating hyperemesis gravidarum and psychosis, not bipolar disorder. Results regarding antipsychotics’ teratogenic and behavioral risks are mixed,^16-21 probably because the various compounds have different effects on the fetus.

The underlying illness—rather than the medications—may increase the rate of anomalies seen with exposure to antipsychotics:

• Rieder et al²² reported an increased rate of perinatal death in infants of schizophrenic mothers but no significant association between the mothers’ use of antipsychotics and perinatal death.
• Sobel²³ compared psychotic women with and without histories of chlorpromazine exposure during pregnancy. Rates of fetal damage were similar and approximately twice that of the general population.

A meta-analysis of 74,337 live births revealed that first-trimester exposure to low-potency antipsychotics increases the relative risk of fetal anomalies in nonpsychotic women. Phenothiazines may increase the 2% baseline incidence of malformations to 2.4%.¹ No specific organ malformation following fetal exposure to phenothiazines has been consistently identified.

Olanzapine was recently approved for treating mania. Very little data exist regarding its impact on fetal development when used during pregnancy, although studies on small numbers of women have not revealed teratogenicity.^24,25

Conclusion. Psychotic illness itself may increase the risk of poor fetal outcome to a greater extent than does antipsychotic use. Prenatal exposure to low-potency phenothiazines may further increase this risk, although only slightly. The effect of prenatal exposure to atypical antipsychotics requires further study.

BENZODIAZEPINES

Benzodiazepines are rarely a primary treatment for mania or depression. Thus, a comprehensive review of their effect on fetal outcome is beyond the scope of this review. A meta-analysis of exposure in the first trimester suggests a very small but significant increase in risk for cleft palate.¹ The absolute risk is <1 in 1,000 cases.

ANTIDEPRESSANTS

Whereas treatment of acute mania is considered a medical emergency, women with bipolar disorder may also relapse into depression during pregnancy. An antidepressant should not be used without a mood stabilizer when treating bipolar I disorder, although a mood stabilizer alone may be inadequate to treat depression. Using tricyclics and selective serotonin reuptake inhibitors (SSRIs) during pregnancy has not been associated with teratogenicity (Table 3),²⁶ although perinatal effects have been reported.¹

Tricyclics. In case-control studies involving more than 300,000 live births, 414 incidences of first-trimester exposure to tricyclics were followed. Information from these patients found no significant association between fetal exposure to tricyclics and increased rates of congenital malformations.¹ The few studies that have been performed suggest no long-term effects from in utero exposure.²⁶ Although these results suggest that prenatal exposure to tricyclics is relatively safe, more research is needed.

SSRIs. To date, no significant teratogenic effects of SSRIs have been identified in offspring of treated women.

The manufacturer’s register for fluoxetine contains approximately 2,000 cases of treated patients, with no excess cases of congenital anomalies or malformations following prenatal exposure. Citalopram has the next largest database of in utero exposure (n=365), again with no increased risk for teratogenicity. Several smaller systematic reports are available on in utero exposure to sertraline, paroxetine, or escitalopram.²⁶

Most studies of pregnant women taking fluoxetine in the first trimester have found no increased risk of obstetrical complications—including spontaneous pregnancy loss, preterm labor, or low birth weight—compared with women not taking fluoxetine. Taking fluoxetine during the third trimester may increase the risk for perinatal complications,²⁷ although this has been inconsistently reported and requires further study. Effects of other SSRIs in the third trimester have not been systematically explored.

Case reports and one controlled study have addressed possible neonatal perinatal symptoms from in utero exposure to SSRIs.^28,29 Preliminary data show no adverse neurobehavioral function in exposed neonates.²⁶

Electroconvulsive therapy (ECT) has been proven effective for acute mania and depression, demonstrating few deleterious effects on neonates. ECT has few side effects and may be safer than drug therapy in this population. Two reviews support the efficacy and relative safety of ECT treatment during pregnancy, although more evidence is needed.^30,31

RECOMMENDATIONS

Discuss pregnancy and medication risks with all bipolar women, regardless of proximal plans for pregnancy. If psychotropic medication is used, prescribe carefully during the first trimester, using the minimum number of drugs and the lowest dosages needed to restore or maintain well-being.³²

Pros and cons of switching. Some clinicians may encourage a patient to taper a medication during the first trimester because of its unknown or high teratogenicity. Depending on the patient’s illness severity, this might not be the optimal decision. A more conservative option would be to switch to a lower-risk drug during pregnancy.

Lithium has both antidepressant and antimanic properties and is less teratogenic compared with first-trimester exposure to an anticonvulsant. However, if lithium has not been successful for the woman’s mania prophylaxis in the past and she has demonstrated antimanic response to an anticonvulsant, switching to lithium or another anticonvulsant is not recommended.

Algorithm Suggested approach to the bipolar patient who wishes to conceive or is pregnant

Low maternal folate levels are often associated with neural tube defects from any cause.³³ Valproate lowers folate levels by inhibiting one of the enzymes necessary for its formation, which may be a mechanism for the increased risk of spina bifida.³⁴

Folate supplementation. To date, no study has demonstrated that giving folate supplements to women taking anticonvulsants during pregnancy reduces the risk of neural tube defects.³⁵ Nonetheless, we recommend that women who continue to take valproate or carbamazepine during pregnancy receive folate, 3 to 4 mg/d, as a precaution.

Treating manic relapse. Data show high rates of relapse in patients who stop taking lithium, particularly if done abruptly.³ Counsel women taking lithium to plan their pregnancies to allow enough time to taper off the medication prior to conception, if they want to try this. Lithium should be decreased slowly—approximately 50% every 2 weeks—to avoid relapse.

Treat aggressively if relapse occurs during pregnancy. Consider:

• psychiatric hospitalization in case of suicidality or psychosis
• reinstituting drug therapy with a less-teratogenic agent
• ECT for a manic or depressive episode.

As the pregnancy advances and the mother’s volume of distribution increases, dosage increases may be needed to maintain therapeutic drug levels.

Treating depressive relapse. Should depression occur in pregnancy, SSRIs or tricyclics added to mood stabilizer therapy have been shown to be effective, with few teratogenic effects.

Cognitive-behavioral and interpersonal psychotherapies also have shown efficacy in pregnant women with major depressive disorder³⁶ and may be effective for women with bipolar disorder in pregnancy. Cognitive psychotherapies, when used with medication, have been reported effective in preventing relapse in nongravid bipolar patients.^36-37

Related resources

Psychiatric disorders during pregnancy. Massachusetts General Hospital Center for Women’s Health. Perinatal Resource Center. www.womensmentalhealth.org/topics/pregnancy_lib.html

Drug brand names

• Amitriptyline • Elavil
• Bupropion • Wellbutrin
• Carbamazepine • Tegretol
• Chlorpromazine • Thorazine
• Citalopram • Celexa
• Clomipramine • Anafranil
• Desipramine • Norpramin
• Fluoxetine • Prozac
• Fluvoxamine • Luvox
• Gabapentin • Neurontin
• Haloperidol • Haldol
• Imipramine • Tofranil
• Lamotrigine • Lamictal
• Lithium • Lithobid et al
• Methylphenidate • Ritalin et al
• Nortriptyline • Pamelor
• Olanzapine • Zyprexa
• Paroxetine • Paxil
• Phenelzine • Nardil
• Risperidone • Risperdal
• Sertraline • Zoloft
• Topiramate • Topamax
• Tranylcypromine • Parnate
• Trifluoperazine • Stelazine
• Valproate • Depakote et al

Disclosure

Dr. Altshuler receives research support from Abbott Laboratories, is a consultant to Abbott Laboratories, Forest Laboratories, and Eli Lilly & Co., and is a speaker for GlaxoSmithKline and Janssen Pharmaceutica.

Ms. Richards reports no financial relationship with any company whose products are mentioned in this article, or with manufacturers of competing products.

Dr. Yonkers receives research support from GlaxoSmithKline and Berlex Laboratories, is a consultant to GlaxoSmithKline, and is a speaker for Eli Lilly and Co., Pfizer Inc., GlaxoSmithKline, and Wyeth Pharmaceuticals.

Prescribing drug therapy for pregnant bipolar women requires psychiatrists to balance the potential for neonatal malformations against the high risk of relapse when patients discontinue their medications.¹ To help you achieve this balance, we offer an evidence-based approach that includes:

• analysis of the FDA’s teratogenicity categories for psychotropics
• review of the safety profiles of drugs used in mood stabilization
• an algorithm for managing patients who are considering conception or are pregnant.

PSYCHOTROPIC RISKS TO OFFSPRING

All psychotropic medications diffuse across the placenta, which exposes the fetus to some degree. Risks include teratogenicity, obstetrical complications, perinatal syndromes, and long-term postnatal behavioral sequelae.

Teratogenicity. A medication is considered teratogenic when prenatal exposure significantly increases the risk of congenital deformities over the baseline risk, which is 2% in the United States.² The cause of most congenital malformations is unknown. Risk for teratogenicity occurs in the first 12 weeks of gestation, as organs are formed.

Table 1

FDA Use-in-Pregnancy ratings for medications The FDA system weighs the degree to which research findings have ruled out risk to the fetus

Obstetrical complications include preterm delivery, low birth weight, and delivery complications such as low Apgar scores or behavioral effects requiring intensive care.

Perinatal syndromes include physical and behavioral symptoms noticed shortly after birth (such as jitteriness). These consequences are putatively related to drug use at or near birth and have limited duration.

Postnatal behavioral sequelae include long-term neurobehavioral abnormalities in children who were exposed to psychotropics in utero.

BALANCING RISKS

Risks with medication. The FDA’s “use in pregnancy” rating system (Table 1) uses available data to assess the degree of teratogenic risk. These guidelines can be confusing and are one of many tools to use when considering a possible drug treatment.

Most psychotropics are category “C” or “D,” which imply a chance of harm to the exposed fetus. Category “B” drugs would appear safer, but this rating could simply indicate a lack of adequate human data or that no data have shown harm in animals.

Moreover, a category “D” drug may be chosen more often during pregnancy than a category “C” drug. This may occur when more human data exist on using the category “D” drug in patients with a particular disorder (such as using lithium versus valproate or olanzapine in pregnant bipolar women).

No psychotropics are classified as “A,” meaning either some risks are associated with every psychotropic or the risk of some agents has not been adequately explored. Furthermore, no psychotropics are FDA-approved for use during pregnancy.

Risks without medication. Teratogenicity notwithstanding, psychotropic intervention is the most effective treatment for women with bipolar disorder. Patients who discontinue mood-stabilizing medication after conception increase their risk of relapse into depression or mania,³ either of which could lead to complications and untoward effects on the fetus.

Depression during pregnancy has been linked to low birth weight and preterm delivery.^4,5 These effects may be mediated by the illness itself or by other factors that indirectly affect birth outcomes. For example, depression during pregnancy is associated with decreased appetite, substance use and abuse, and lower use of prenatal care.⁶

Untreated mania may also be associated with perinatal risks, as a pregnant patient in a manic state may engage in impulsive, high-risk behaviors that endanger her and the fetus.⁷

MOOD STABILIZERS

The FDA categorizes as “D” the three most commonly used mood stabilizers: lithium, valproate, and carbamazepine (Table 2). This rating implies that studies have demonstrated fetal risk but the drug’s potential benefit may still outweigh the risk.

Lithium. The International Registry of Lithium reported increased rates of cardiovascular malformations— such as Ebstein’s anomaly—in children whose mothers took lithium during pregnancy.

Relative risk for Ebstein’s anomaly in children with fetal exposure to lithium may be 20 times higher than the risk in unexposed children, although the absolute risk with lithium exposure remains low (1 in 1,000 births).^1,8

No significant neurobehavioral teratogenicity has been reported in infants exposed in utero to lithium, although few cases have been studied. One study reported that 22 lithium-exposed infants attained developmental milestones at a pace comparable to that of unexposed controls.⁹

“Floppy baby” syndrome, in which infants experience hypotonicity and cyanosis, is the most recognized adverse effect in infants exposed to lithium in utero.¹⁰ Its frequency is unknown, but rare. Neonatal hypothyroidism and nephrogenic diabetes insipidus have also been documented.

Anticonvulsants. To date, no studies have examined the outcomes of children whose mothers took anticonvulsants for bipolar disorder during pregnancy, though the research concerning epileptic mothers is extensive.

Neural tube defects. Data associate anticonvulsant exposure with a significantly greater risk for malformations than in the general population. Specifically, anticonvulsants may cause neural tube defects such as spina bifida, ancephaly, and encephaly in 2 to 5% of those exposed, as well as craniofacial anomalies, microcephaly, growth retardation, and heart defects.^11-14

Table 2

FDA’s teratogenicity ratings of mood stabilizers and other antimanic agents

More minor malformations—such as rotated ears, depressed nasal bridge, short nose, elongated upper lip, and fingernail hypoplasia—have been reported in infants exposed to anticonvulsants in utero.¹⁴ These malformations disappear with age.¹³ Teratogenicity increases with the use of multiple anticonvulsants and possibly with higher maternal plasma levels and toxic metabolites.¹⁵

Conclusion. The three most commonly used mood stabilizers are all teratogenic. The least risk may occur with lithium (0.1%) versus valproate (2 to 5%) or carbamazepine (1 to 3%). These risks must be weighed against the up to 50% chance of relapse with medication discontinuation.³

ANTIPSYCHOTICS

Antipsychotics are often used to treat mania because of their rapid effects and sedative properties. Most antipsychotics—specifically, haloperidol, olanzapine, and risperidone—are labeled “C,” specifying that fetal risk cannot be ruled out.

Table 3

FDA’s teratogenicity ratings of common antidepressants

Chlorpromazine and haloperidol have been most studied during pregnancy but in relation to treating hyperemesis gravidarum and psychosis, not bipolar disorder. Results regarding antipsychotics’ teratogenic and behavioral risks are mixed,^16-21 probably because the various compounds have different effects on the fetus.

The underlying illness—rather than the medications—may increase the rate of anomalies seen with exposure to antipsychotics:

• Rieder et al²² reported an increased rate of perinatal death in infants of schizophrenic mothers but no significant association between the mothers’ use of antipsychotics and perinatal death.
• Sobel²³ compared psychotic women with and without histories of chlorpromazine exposure during pregnancy. Rates of fetal damage were similar and approximately twice that of the general population.

A meta-analysis of 74,337 live births revealed that first-trimester exposure to low-potency antipsychotics increases the relative risk of fetal anomalies in nonpsychotic women. Phenothiazines may increase the 2% baseline incidence of malformations to 2.4%.¹ No specific organ malformation following fetal exposure to phenothiazines has been consistently identified.

Olanzapine was recently approved for treating mania. Very little data exist regarding its impact on fetal development when used during pregnancy, although studies on small numbers of women have not revealed teratogenicity.^24,25

Conclusion. Psychotic illness itself may increase the risk of poor fetal outcome to a greater extent than does antipsychotic use. Prenatal exposure to low-potency phenothiazines may further increase this risk, although only slightly. The effect of prenatal exposure to atypical antipsychotics requires further study.

BENZODIAZEPINES

Benzodiazepines are rarely a primary treatment for mania or depression. Thus, a comprehensive review of their effect on fetal outcome is beyond the scope of this review. A meta-analysis of exposure in the first trimester suggests a very small but significant increase in risk for cleft palate.¹ The absolute risk is <1 in 1,000 cases.

ANTIDEPRESSANTS

Whereas treatment of acute mania is considered a medical emergency, women with bipolar disorder may also relapse into depression during pregnancy. An antidepressant should not be used without a mood stabilizer when treating bipolar I disorder, although a mood stabilizer alone may be inadequate to treat depression. Using tricyclics and selective serotonin reuptake inhibitors (SSRIs) during pregnancy has not been associated with teratogenicity (Table 3),²⁶ although perinatal effects have been reported.¹

Tricyclics. In case-control studies involving more than 300,000 live births, 414 incidences of first-trimester exposure to tricyclics were followed. Information from these patients found no significant association between fetal exposure to tricyclics and increased rates of congenital malformations.¹ The few studies that have been performed suggest no long-term effects from in utero exposure.²⁶ Although these results suggest that prenatal exposure to tricyclics is relatively safe, more research is needed.

SSRIs. To date, no significant teratogenic effects of SSRIs have been identified in offspring of treated women.

The manufacturer’s register for fluoxetine contains approximately 2,000 cases of treated patients, with no excess cases of congenital anomalies or malformations following prenatal exposure. Citalopram has the next largest database of in utero exposure (n=365), again with no increased risk for teratogenicity. Several smaller systematic reports are available on in utero exposure to sertraline, paroxetine, or escitalopram.²⁶

Most studies of pregnant women taking fluoxetine in the first trimester have found no increased risk of obstetrical complications—including spontaneous pregnancy loss, preterm labor, or low birth weight—compared with women not taking fluoxetine. Taking fluoxetine during the third trimester may increase the risk for perinatal complications,²⁷ although this has been inconsistently reported and requires further study. Effects of other SSRIs in the third trimester have not been systematically explored.

Case reports and one controlled study have addressed possible neonatal perinatal symptoms from in utero exposure to SSRIs.^28,29 Preliminary data show no adverse neurobehavioral function in exposed neonates.²⁶

Electroconvulsive therapy (ECT) has been proven effective for acute mania and depression, demonstrating few deleterious effects on neonates. ECT has few side effects and may be safer than drug therapy in this population. Two reviews support the efficacy and relative safety of ECT treatment during pregnancy, although more evidence is needed.^30,31

RECOMMENDATIONS

Discuss pregnancy and medication risks with all bipolar women, regardless of proximal plans for pregnancy. If psychotropic medication is used, prescribe carefully during the first trimester, using the minimum number of drugs and the lowest dosages needed to restore or maintain well-being.³²

Pros and cons of switching. Some clinicians may encourage a patient to taper a medication during the first trimester because of its unknown or high teratogenicity. Depending on the patient’s illness severity, this might not be the optimal decision. A more conservative option would be to switch to a lower-risk drug during pregnancy.

Lithium has both antidepressant and antimanic properties and is less teratogenic compared with first-trimester exposure to an anticonvulsant. However, if lithium has not been successful for the woman’s mania prophylaxis in the past and she has demonstrated antimanic response to an anticonvulsant, switching to lithium or another anticonvulsant is not recommended.

Algorithm Suggested approach to the bipolar patient who wishes to conceive or is pregnant

Low maternal folate levels are often associated with neural tube defects from any cause.³³ Valproate lowers folate levels by inhibiting one of the enzymes necessary for its formation, which may be a mechanism for the increased risk of spina bifida.³⁴

Folate supplementation. To date, no study has demonstrated that giving folate supplements to women taking anticonvulsants during pregnancy reduces the risk of neural tube defects.³⁵ Nonetheless, we recommend that women who continue to take valproate or carbamazepine during pregnancy receive folate, 3 to 4 mg/d, as a precaution.

Treating manic relapse. Data show high rates of relapse in patients who stop taking lithium, particularly if done abruptly.³ Counsel women taking lithium to plan their pregnancies to allow enough time to taper off the medication prior to conception, if they want to try this. Lithium should be decreased slowly—approximately 50% every 2 weeks—to avoid relapse.

Treat aggressively if relapse occurs during pregnancy. Consider:

• psychiatric hospitalization in case of suicidality or psychosis
• reinstituting drug therapy with a less-teratogenic agent
• ECT for a manic or depressive episode.

As the pregnancy advances and the mother’s volume of distribution increases, dosage increases may be needed to maintain therapeutic drug levels.

Treating depressive relapse. Should depression occur in pregnancy, SSRIs or tricyclics added to mood stabilizer therapy have been shown to be effective, with few teratogenic effects.

Cognitive-behavioral and interpersonal psychotherapies also have shown efficacy in pregnant women with major depressive disorder³⁶ and may be effective for women with bipolar disorder in pregnancy. Cognitive psychotherapies, when used with medication, have been reported effective in preventing relapse in nongravid bipolar patients.^36-37

Related resources

Psychiatric disorders during pregnancy. Massachusetts General Hospital Center for Women’s Health. Perinatal Resource Center. www.womensmentalhealth.org/topics/pregnancy_lib.html

Drug brand names

• Amitriptyline • Elavil
• Bupropion • Wellbutrin
• Carbamazepine • Tegretol
• Chlorpromazine • Thorazine
• Citalopram • Celexa
• Clomipramine • Anafranil
• Desipramine • Norpramin
• Fluoxetine • Prozac
• Fluvoxamine • Luvox
• Gabapentin • Neurontin
• Haloperidol • Haldol
• Imipramine • Tofranil
• Lamotrigine • Lamictal
• Lithium • Lithobid et al
• Methylphenidate • Ritalin et al
• Nortriptyline • Pamelor
• Olanzapine • Zyprexa
• Paroxetine • Paxil
• Phenelzine • Nardil
• Risperidone • Risperdal
• Sertraline • Zoloft
• Topiramate • Topamax
• Tranylcypromine • Parnate
• Trifluoperazine • Stelazine
• Valproate • Depakote et al

Disclosure

Dr. Altshuler receives research support from Abbott Laboratories, is a consultant to Abbott Laboratories, Forest Laboratories, and Eli Lilly & Co., and is a speaker for GlaxoSmithKline and Janssen Pharmaceutica.

Ms. Richards reports no financial relationship with any company whose products are mentioned in this article, or with manufacturers of competing products.

Dr. Yonkers receives research support from GlaxoSmithKline and Berlex Laboratories, is a consultant to GlaxoSmithKline, and is a speaker for Eli Lilly and Co., Pfizer Inc., GlaxoSmithKline, and Wyeth Pharmaceuticals.

To me, the main difference between MDs and PhDs* is that MDs—at some point—must stop gathering data and make decisions.

I once heard Dr. Albert (Mickey) Stunkard say that when he was a physician fellow at Stanford University’s Center for Advanced Studies in the Behavioral Sciences he was at first energized—and a little intimidated—by the scintillating conversations taking place around him. Eventually, though, all the discourse reminded him of those long, philosophical discussions he and his classmates had had in their college dorms (“Well, on one hand you have communism, and on the other hand you have fascism… ”).

Physicians do not have the luxury of endless debate. At some point, we need to do something or else let our patients die of old age while waiting. One issue about which I have had to make decisions over the years—and which has troubled me the most—is whether to treat pregnant patients with psychotropics. Generally, I try to avoid using drugs in these cases, but sometimes I decide that the mother’s need for drug therapy outweighs the potential risks to her offspring.

Dr. Lori Altshuler and colleagues’ article in this issue is the best summary I have seen of what is known about the risks of using psychotropics in pregnant bipolar women. Each time I treat a woman with bipolar disorder, I will remember this discussion and the algorithm these authors suggest for making therapeutic decisions.

This excellent article may not be the final word on the subject. It can, however, help us with an important clinical decision we often have to make—and that is what Current Psychiatry is all about.

To me, the main difference between MDs and PhDs* is that MDs—at some point—must stop gathering data and make decisions.

I once heard Dr. Albert (Mickey) Stunkard say that when he was a physician fellow at Stanford University’s Center for Advanced Studies in the Behavioral Sciences he was at first energized—and a little intimidated—by the scintillating conversations taking place around him. Eventually, though, all the discourse reminded him of those long, philosophical discussions he and his classmates had had in their college dorms (“Well, on one hand you have communism, and on the other hand you have fascism… ”).

Physicians do not have the luxury of endless debate. At some point, we need to do something or else let our patients die of old age while waiting. One issue about which I have had to make decisions over the years—and which has troubled me the most—is whether to treat pregnant patients with psychotropics. Generally, I try to avoid using drugs in these cases, but sometimes I decide that the mother’s need for drug therapy outweighs the potential risks to her offspring.

Dr. Lori Altshuler and colleagues’ article in this issue is the best summary I have seen of what is known about the risks of using psychotropics in pregnant bipolar women. Each time I treat a woman with bipolar disorder, I will remember this discussion and the algorithm these authors suggest for making therapeutic decisions.

This excellent article may not be the final word on the subject. It can, however, help us with an important clinical decision we often have to make—and that is what Current Psychiatry is all about.

To me, the main difference between MDs and PhDs* is that MDs—at some point—must stop gathering data and make decisions.

I once heard Dr. Albert (Mickey) Stunkard say that when he was a physician fellow at Stanford University’s Center for Advanced Studies in the Behavioral Sciences he was at first energized—and a little intimidated—by the scintillating conversations taking place around him. Eventually, though, all the discourse reminded him of those long, philosophical discussions he and his classmates had had in their college dorms (“Well, on one hand you have communism, and on the other hand you have fascism… ”).

Physicians do not have the luxury of endless debate. At some point, we need to do something or else let our patients die of old age while waiting. One issue about which I have had to make decisions over the years—and which has troubled me the most—is whether to treat pregnant patients with psychotropics. Generally, I try to avoid using drugs in these cases, but sometimes I decide that the mother’s need for drug therapy outweighs the potential risks to her offspring.

Dr. Lori Altshuler and colleagues’ article in this issue is the best summary I have seen of what is known about the risks of using psychotropics in pregnant bipolar women. Each time I treat a woman with bipolar disorder, I will remember this discussion and the algorithm these authors suggest for making therapeutic decisions.

This excellent article may not be the final word on the subject. It can, however, help us with an important clinical decision we often have to make—and that is what Current Psychiatry is all about.

Children and adolescents with bipolar disorder are often referred to psychiatrists because of disruptive behaviors at home and in school. They exhibit poor academic performance, disturbed interpersonal relationships, increased rates of substance abuse, legal difficulties, multiple hospitalizations, and high rates of suicide attempts and completions.^1,2 Many have comorbid psychiatric problems—particularly attention-deficit/hyperactivity disorder (ADHD).

Although few studies have examined this complex diagnosis, we do know that bipolar disorder presents differently in children and adolescents than in adults. Prodromal symptoms can appear early—before kindergarten in some children. Early recognition therefore is key to effectively treating these sick and often complicated patients.

How often a clinician encounters a child or adolescent with bipolar disorder depends largely on the practice setting (Box).^1,3,4 Wherever you practice, however, you can recognize and treat pediatric bipolar disorder if you keep in mind that its presentation and disease progression differ from the adult type.

Pediatric versus adult symptoms

Prodromal symptoms—such as episodes of depressed mood or hopelessness and excessive mood lability—have been detected in youths who later were diagnosed with bipolar disorder. More than one-half of 494 adult members of the Depression and Bipolar Support Alliance have reported that they first exhibited signs of bipolar illness before age 19, with distribution by age as follows:

• 5% before age 5
• 12% at ages 5 to 9
• 14% at ages 10 to 14
• 28% at ages 15 to 19.⁵

Box

Table 1

COMMON PRESENTING SYMPTOMS OF PEDIATRIC BIPOLAR DISORDER

In a similar study,⁶ 58 adult patients with bipolar I disorder reported an average interval of 9 to 12 years between the emergence of bipolar symptoms and the onset of a major affective disorder.

Common initial symptoms of pediatric bipolar disorder are listed in Table 1. Most of these symptoms occur in discrete episodes and represent a change from the child’s normal functioning.

Many children and adolescents are labeled “bipolar” without careful consideration of this disorder’s diagnostic complexities and subtypes. Bipolarity in young patients can be difficult to establish because of:

• variability of symptom expression, depending on the illness’ context and phase
• effects of development on symptom expression
• mood and behavioral effects of psychotropic medications the patient is taking.

Pediatric bipolar patients often present with a mixed or “dysphoric” picture characterized by frequent short periods of intense mood lability and irritability rather than classic euphoric mania.^3,7 Clinicians who evaluate children with pediatric bipolar disorders often try to fit them into the DSM-IV “rapid cycling” subtype. This subtype does not fit bipolar children very well, however, because they often lack clear episodes of mania. Rather, researchers are reporting that bipolar children cycle far more frequently than the four episodes/year in DSM-IV’s diagnostic criteria:

• Continuous, daily cycling from mania or hypomania to euthymia or depression was seen in 81% of a well-defined group of pediatric bipolar patients.^7,8
• A high rate of rapid cycling and onset of a first manic episode at mean age 7 was reported in 90 children and adolescents (mean age 11) with bipolar I disorder.⁹

The picture that emerges from independent research groups is that multiple daily mood swings and irritability are much more common than euphoria in prepubertal children with bipolar disorder.^8,10

Making the diagnosis

DSM-IV’s diagnostic classification system for bipolar disorders is complex, involving:

• five types of episodes (manic, hypomanic, mixed, depressed, unspecified)
• four severity levels (mild, moderate, severe without psychosis, severe with psychosis)
• and three course specifiers (with or without inter-episode recovery, seasonal pattern, rapid cycling).

Table 2

PEDIATRIC BIPOLAR DISORDER SUBTYPES: DIAGNOSTIC CHARACTERISTICS AND ASSOCIATED FEATURES

Table 3

COMPLICATING FACTORS IN PEDIATRIC BIPOLAR DISORDER

DSM-IV criteria for mania—which were developed from data on adults with bipolar disorders—do not take into account developmental differences between bipolar adults and bipolar children and adolescents.

Diagnostic characteristics of the pediatric bipolar disorder subtypes are compared in Table 2. Generally:

• Pediatric patients with bipolar I disorder have multiple daily mood swings, a “mixed” type of episode with short periods of euphoria and longer periods of irritability, and comorbidities such as ADHD, oppositional defiant disorder, or conduct disorder.^3,11,12
• Bipolar II disorder presents more typically in adolescence and is usually noticed clinically as a major depressive episode. Past episodes of hypomania may have been missed unless a careful history was taken.
• Cyclothymia is difficult to diagnose because the hypomania and depressive symptoms are not as severe as in bipolar types I or II. Prospective mood charting can help the clinician diagnose cyclothymia (see “Related Resources”).
• Bipolar disorder NOS represents the largest group of patients with bipolar symptoms. This diagnosis is made when bipolar symptoms are present but not of sufficient severity or duration to warrant a diagnosis of bipolar I, II or cyclothymia. Bipolar NOS also can be diagnosed when a bipolar disorder is secondary to a general medical condition, such as fetal alcohol syndrome or alcohol-related neurodevelopmental disorder.

Differential diagnosis. Medications and medical disorders may exacerbate or mimic pediatric bipolar symptoms (Table 3), so it is important to assess these potential confounds before initiating treatment. Psychiatric comorbidities also frequently complicate the presentation of pediatric bipolar disorder and its response to treatment (Table 4). ADHD is the most common comorbidity, with rates as high as 98% in bipolar children.^3,13

Outcomes

Long-term outcomes of children and adolescents with bipolar disorders have not been well studied. In the only prospective follow-up investigation of adolescent inpatients with mania, Strober et al found that most of 54 patients (96%) recovered from an index affective episode, but nearly one-half (44%) experienced one or more relapses within 5 years.¹⁴ The rate of recovery varied according to the index episode’s polarity. Recovery was faster in patients with pure mania or mixed states, and multiple relapses occurred more frequently in those with mixed or cycling episodes. Twenty percent of the patients attempted suicide.

Recently, Geller et al reported the results of the first large, prospective, follow-up study of children with bipolar disorder.¹⁵ In 89 outpatients (mean age 11) with bipolar I disorder, comprehensive assessments at baseline and at 6, 12, 18, and 24 months showed that 65% recovered from mania but 55% relapsed after recovery. Mean time to recovery was 36 weeks, and relapse occurred after a mean of 28.6 weeks. Children living with their intact biological families were twice as likely to recover as those in other living arrangements.

The poor outcomes of these bipolar children highlight the need for earlier recognition and more effective treatments.

Treating acute mania

Many psychotropic medications used to treat adults with bipolar disorders are also used for children and adolescents. To date, only two double-blind, placebo-controlled studies^13,16 and one uncontrolled maintenance treatment study¹⁷ have examined treatment of acute mania in pediatric bipolar disorder.

Lithium is the most studied medication for pediatric bipolar disorder and the only FDA-approved medication for treating acute mania and bipolar disorder in patients ages 12 to 18. Approximately 40 to 50% of children and adolescents with bipolar disorder respond to lithium monotherapy.^18,19

In general, lithium should be titrated to 30 mg/kg/d in two or three divided doses; this typically produces a serum level of 0.8 to 1.2 mEq/L. Common side effects in children and adolescents include nausea, polyuria, polydipsia, tremor, acne, and weight gain. Lithium levels and thyroid function should be monitored, as in adult patients.

Only one prospective, placebo-controlled study has examined lithium use in children and adolescents with bipolar disorders. Twenty-five adolescents with comorbid bipolar and substance use disorders were treated with lithium or placebo for 6 weeks. Positive urine toxicology screens decreased significantly, and global assessment of functioning improved in 46% of those receiving lithium vs. 8% of those receiving placebo.¹³ This study demonstrated lithium’s efficacy in treating bipolar adolescents with comorbid substance abuse but did not measure its effect on mood.

Risk factors for poor lithium response in children and adolescents with bipolar disorder include prepubertal onset and comorbid ADHD.²⁰

Divalproex. No placebo-controlled studies of antiepileptics in pediatric bipolar disorder have been published. Open-label studies of divalproex have reported response rates of 53 to 82% in manic adolescents.^{18, 21-23} Several case reports and series have described successful use of carbamazepine as monotherapy and adjunctive treatment in children and adolescents with bipolar disorder.^24,25

Table 4

RATES OF COMMON COMORBIDITIES IN PEDIATRIC BIPOLAR DISORDERS

One 6-week, random-assignment, prospective study compared lithium, divalproex, and carbamazepine in treating 42 acutely manic or hypomanic patients ages 8 to 18.¹⁸ In this open study, all three mood stabilizers demonstrated efficacy in treating a mixed or manic episode in youths with bipolar I or II disorder. Response rates—based on a 50% improvement in Young Mania Rating Scale baseline scores—were divalproex 53%, lithium 38%, and carbamazepine 38%.

In general, divalproex is started at 20 mg/kg/d, which typically produces a serum level of 80 to 120 μg/ml. Common side effects in children include weight gain, nausea, sedation, and tremor.

A possible association between divalproex and polycystic ovary syndrome (PCOS) has been reported in women with epilepsy.²⁶ The mechanism for PCOS has been hypothesized to be obesity secondary to divalproex, resulting in elevated insulin and androgen levels. Recently, Rasgon et al²⁷ reported that epilepsy—and not the anticonvulsants used to treat it—may increase the risk of PCOS. In contrast, O’Donovan et al reported higher rates of menstrual irregularities and PCOS in women with bipolar disorder who were taking divalproex than in those who were not taking divalproex and in healthy controls.²⁸

Until we learn more about this association, clinicians should monitor bipolar female adolescents treated with divalproex for any signs of PCOS, which include menstrual abnormalities, hirsutism, and acne.

Carbamazepine is used widely for seizure management but less commonly than divalproex in pediatric bipolar disorder. This anticonvulsant must be titrated slowly and requires frequent monitoring of blood levels, which can be a problem in children with needle phobia.

Carbamazepine is usually titrated to 15 mg/kg/d to produce a serum level of 7 to 10 μg/ml. Its most common side effects are sedation, rash, nausea, and hyponatremia. Aplastic anemia and severe dermatologic reactions, such as Stevens-Johnson syndrome, occur uncommonly.²⁸

Atypical antipsychotics. Recent case series and open-label reports suggest that atypical antipsychotics such as clozapine,²⁹ risperidone,³⁰ olanzapine,^31-33 and quetiapine¹⁶ are effective in treating pediatric bipolar disorder. However, clinically significant weight gain may be associated with the use of olanzapine and risperidone.³⁴ Ziprasidone may increase QTc prolongation, and safety data are limited in children and adolescents. Therefore, ziprasidone should be used with caution in pediatric bipolar disorder, and ECGs should be monitored.

In the only double-blind, placebo-controlled study of an atypical antipsychotic in pediatric bipolar disorder, manic symptoms were more greatly reduced in 15 adolescents given quetiapine plus divalproex than in 15 patients who received divalproex alone. Quetiapine was titrated to 450 mg/d across 7 days and was well-tolerated. The findings suggest that a mood stabilizer plus an atypical antipsychotic may be more effective than a mood stabilizer alone for treating adolescent mania.¹⁶

Long-term treatment

In addition to treating acute affective episodes, lithium may also help prevent recurrent affective episodes in younger patients. In the only maintenance treatment study for pediatric bipolar disorder, Strober et al prospectively evaluated 37 adolescents whose bipolar disorder had been stabilized with lithium during hospitalization.¹⁷ After 18 months of follow-up, 35% of patients had discontinued lithium, and their relapse rate was 92% (compared with 38% in patients who were lithium-compliant.

It is reasonable to maintain a child or adolescent who has had a single manic episode on mood-stabilizing treatment for several years and then—if the patient is euthymic and asymptomatic—to slowly taper the mood stabilizer(s) over several months. If mood symptoms recur, the agent(s) should be reintroduced.

If a child with bipolar disorder does not respond or only partially responds to a mood stabilizer, it may be necessary to add a second mood stabilizer or an atypical antipsychotic. A bipolar child or adolescent with psychotic symptoms should be maintained on an antipsychotic (typical or atypical) for at least 1 month, even if the psychosis has resolved.³⁵

Treatment of comorbid ADHD. Most children with bipolar disorder have comorbid ADHD, and mood stabilization is necessary prior to starting stimulant medications.³⁶ In bipolar patients, sustained-release psychostimulants may reduce rebound symptoms more effectively than immediate-release formulations. Typical dosages for a child with bipolar disorder and ADHD would be Concerta, 36 mg/d, or Adderall XR, 10 to 20 mg/d.

Psychosocial interventions

Most psychotherapeutic interventions have not been systematically studied in pediatric bipolar disorder but may be beneficial. In a recent study, Fristad et al reported the efficacy of multifamily psychoeducational group therapy for treating bipolar children and adolescents and their families.³⁷

Other useful psychosocial tactics include:

• Minimize periods of overstimulation (for example, these patients do not do well at shopping malls).
• Maintain good sleep hygiene.
• Address medication nonadherence immediately.
• Discuss the risk of substance abuse with the patient.
• Encourage mood charting by the patient and or parent.

Related resources

• Mood charts appropriate for pediatric patients.
• Child and Adolescent Bipolar Foundation. www.cabf.org
• Findling RF, Kowatch RA, Post RM. Pediatric bipolar disorders: a handbook for clinicians. London, Martin Dunitz Press, 2002.
• Geller B, DelBello MP (eds). Child and early adolescent bipolar disorder: theory, assessment, and treatment. New York: Guilford Publications, 2002.

Drug brand names

• Clozapine • Clozaril
• Divalproex sodium • Depakote
• Olanzapine • Zyprexa
• Risperidone • Risperdal
• Quetiapine • Seroquel
• Ziprasidone • Geodon

Disclosure

Dr. Kowatch receives grant support from, serves as a consultant to, or is on the speakers bureau of Novartis Pharmaceuticals Corp., Abbott Laboratories, Solvay Pharmaceuticals, Eli Lilly and Co., Janssen Pharmaceutica, GlaxoSmithKline, and AstraZeneca Pharmaceuticals.

Dr. DelBello receives grant support from, serves as a consultant to, or is on the speakers bureau of Abbott Laboratories, AstraZeneca Pharmaceuticals, Eli Lilly and Co., Ortho-McNeil Pharmaceutical, Shire Pharmaceutical Group, Janssen Pharmaceutica, Pfizer Inc., and GlaxoSmithKline.

Children and adolescents with bipolar disorder are often referred to psychiatrists because of disruptive behaviors at home and in school. They exhibit poor academic performance, disturbed interpersonal relationships, increased rates of substance abuse, legal difficulties, multiple hospitalizations, and high rates of suicide attempts and completions.^1,2 Many have comorbid psychiatric problems—particularly attention-deficit/hyperactivity disorder (ADHD).

Although few studies have examined this complex diagnosis, we do know that bipolar disorder presents differently in children and adolescents than in adults. Prodromal symptoms can appear early—before kindergarten in some children. Early recognition therefore is key to effectively treating these sick and often complicated patients.

How often a clinician encounters a child or adolescent with bipolar disorder depends largely on the practice setting (Box).^1,3,4 Wherever you practice, however, you can recognize and treat pediatric bipolar disorder if you keep in mind that its presentation and disease progression differ from the adult type.

Pediatric versus adult symptoms

Prodromal symptoms—such as episodes of depressed mood or hopelessness and excessive mood lability—have been detected in youths who later were diagnosed with bipolar disorder. More than one-half of 494 adult members of the Depression and Bipolar Support Alliance have reported that they first exhibited signs of bipolar illness before age 19, with distribution by age as follows:

• 5% before age 5
• 12% at ages 5 to 9
• 14% at ages 10 to 14
• 28% at ages 15 to 19.⁵

Box

Table 1

COMMON PRESENTING SYMPTOMS OF PEDIATRIC BIPOLAR DISORDER

In a similar study,⁶ 58 adult patients with bipolar I disorder reported an average interval of 9 to 12 years between the emergence of bipolar symptoms and the onset of a major affective disorder.

Common initial symptoms of pediatric bipolar disorder are listed in Table 1. Most of these symptoms occur in discrete episodes and represent a change from the child’s normal functioning.

Many children and adolescents are labeled “bipolar” without careful consideration of this disorder’s diagnostic complexities and subtypes. Bipolarity in young patients can be difficult to establish because of:

• variability of symptom expression, depending on the illness’ context and phase
• effects of development on symptom expression
• mood and behavioral effects of psychotropic medications the patient is taking.

Pediatric bipolar patients often present with a mixed or “dysphoric” picture characterized by frequent short periods of intense mood lability and irritability rather than classic euphoric mania.^3,7 Clinicians who evaluate children with pediatric bipolar disorders often try to fit them into the DSM-IV “rapid cycling” subtype. This subtype does not fit bipolar children very well, however, because they often lack clear episodes of mania. Rather, researchers are reporting that bipolar children cycle far more frequently than the four episodes/year in DSM-IV’s diagnostic criteria:

• Continuous, daily cycling from mania or hypomania to euthymia or depression was seen in 81% of a well-defined group of pediatric bipolar patients.^7,8
• A high rate of rapid cycling and onset of a first manic episode at mean age 7 was reported in 90 children and adolescents (mean age 11) with bipolar I disorder.⁹

The picture that emerges from independent research groups is that multiple daily mood swings and irritability are much more common than euphoria in prepubertal children with bipolar disorder.^8,10

Making the diagnosis

DSM-IV’s diagnostic classification system for bipolar disorders is complex, involving:

• five types of episodes (manic, hypomanic, mixed, depressed, unspecified)
• four severity levels (mild, moderate, severe without psychosis, severe with psychosis)
• and three course specifiers (with or without inter-episode recovery, seasonal pattern, rapid cycling).

Table 2

PEDIATRIC BIPOLAR DISORDER SUBTYPES: DIAGNOSTIC CHARACTERISTICS AND ASSOCIATED FEATURES

Table 3

COMPLICATING FACTORS IN PEDIATRIC BIPOLAR DISORDER

DSM-IV criteria for mania—which were developed from data on adults with bipolar disorders—do not take into account developmental differences between bipolar adults and bipolar children and adolescents.

Diagnostic characteristics of the pediatric bipolar disorder subtypes are compared in Table 2. Generally:

• Pediatric patients with bipolar I disorder have multiple daily mood swings, a “mixed” type of episode with short periods of euphoria and longer periods of irritability, and comorbidities such as ADHD, oppositional defiant disorder, or conduct disorder.^3,11,12
• Bipolar II disorder presents more typically in adolescence and is usually noticed clinically as a major depressive episode. Past episodes of hypomania may have been missed unless a careful history was taken.
• Cyclothymia is difficult to diagnose because the hypomania and depressive symptoms are not as severe as in bipolar types I or II. Prospective mood charting can help the clinician diagnose cyclothymia (see “Related Resources”).
• Bipolar disorder NOS represents the largest group of patients with bipolar symptoms. This diagnosis is made when bipolar symptoms are present but not of sufficient severity or duration to warrant a diagnosis of bipolar I, II or cyclothymia. Bipolar NOS also can be diagnosed when a bipolar disorder is secondary to a general medical condition, such as fetal alcohol syndrome or alcohol-related neurodevelopmental disorder.

Differential diagnosis. Medications and medical disorders may exacerbate or mimic pediatric bipolar symptoms (Table 3), so it is important to assess these potential confounds before initiating treatment. Psychiatric comorbidities also frequently complicate the presentation of pediatric bipolar disorder and its response to treatment (Table 4). ADHD is the most common comorbidity, with rates as high as 98% in bipolar children.^3,13

Outcomes

Long-term outcomes of children and adolescents with bipolar disorders have not been well studied. In the only prospective follow-up investigation of adolescent inpatients with mania, Strober et al found that most of 54 patients (96%) recovered from an index affective episode, but nearly one-half (44%) experienced one or more relapses within 5 years.¹⁴ The rate of recovery varied according to the index episode’s polarity. Recovery was faster in patients with pure mania or mixed states, and multiple relapses occurred more frequently in those with mixed or cycling episodes. Twenty percent of the patients attempted suicide.

Recently, Geller et al reported the results of the first large, prospective, follow-up study of children with bipolar disorder.¹⁵ In 89 outpatients (mean age 11) with bipolar I disorder, comprehensive assessments at baseline and at 6, 12, 18, and 24 months showed that 65% recovered from mania but 55% relapsed after recovery. Mean time to recovery was 36 weeks, and relapse occurred after a mean of 28.6 weeks. Children living with their intact biological families were twice as likely to recover as those in other living arrangements.

The poor outcomes of these bipolar children highlight the need for earlier recognition and more effective treatments.

Treating acute mania

Many psychotropic medications used to treat adults with bipolar disorders are also used for children and adolescents. To date, only two double-blind, placebo-controlled studies^13,16 and one uncontrolled maintenance treatment study¹⁷ have examined treatment of acute mania in pediatric bipolar disorder.

Lithium is the most studied medication for pediatric bipolar disorder and the only FDA-approved medication for treating acute mania and bipolar disorder in patients ages 12 to 18. Approximately 40 to 50% of children and adolescents with bipolar disorder respond to lithium monotherapy.^18,19

In general, lithium should be titrated to 30 mg/kg/d in two or three divided doses; this typically produces a serum level of 0.8 to 1.2 mEq/L. Common side effects in children and adolescents include nausea, polyuria, polydipsia, tremor, acne, and weight gain. Lithium levels and thyroid function should be monitored, as in adult patients.

Only one prospective, placebo-controlled study has examined lithium use in children and adolescents with bipolar disorders. Twenty-five adolescents with comorbid bipolar and substance use disorders were treated with lithium or placebo for 6 weeks. Positive urine toxicology screens decreased significantly, and global assessment of functioning improved in 46% of those receiving lithium vs. 8% of those receiving placebo.¹³ This study demonstrated lithium’s efficacy in treating bipolar adolescents with comorbid substance abuse but did not measure its effect on mood.

Risk factors for poor lithium response in children and adolescents with bipolar disorder include prepubertal onset and comorbid ADHD.²⁰

Divalproex. No placebo-controlled studies of antiepileptics in pediatric bipolar disorder have been published. Open-label studies of divalproex have reported response rates of 53 to 82% in manic adolescents.^{18, 21-23} Several case reports and series have described successful use of carbamazepine as monotherapy and adjunctive treatment in children and adolescents with bipolar disorder.^24,25

Table 4

RATES OF COMMON COMORBIDITIES IN PEDIATRIC BIPOLAR DISORDERS

One 6-week, random-assignment, prospective study compared lithium, divalproex, and carbamazepine in treating 42 acutely manic or hypomanic patients ages 8 to 18.¹⁸ In this open study, all three mood stabilizers demonstrated efficacy in treating a mixed or manic episode in youths with bipolar I or II disorder. Response rates—based on a 50% improvement in Young Mania Rating Scale baseline scores—were divalproex 53%, lithium 38%, and carbamazepine 38%.

In general, divalproex is started at 20 mg/kg/d, which typically produces a serum level of 80 to 120 μg/ml. Common side effects in children include weight gain, nausea, sedation, and tremor.

A possible association between divalproex and polycystic ovary syndrome (PCOS) has been reported in women with epilepsy.²⁶ The mechanism for PCOS has been hypothesized to be obesity secondary to divalproex, resulting in elevated insulin and androgen levels. Recently, Rasgon et al²⁷ reported that epilepsy—and not the anticonvulsants used to treat it—may increase the risk of PCOS. In contrast, O’Donovan et al reported higher rates of menstrual irregularities and PCOS in women with bipolar disorder who were taking divalproex than in those who were not taking divalproex and in healthy controls.²⁸

Until we learn more about this association, clinicians should monitor bipolar female adolescents treated with divalproex for any signs of PCOS, which include menstrual abnormalities, hirsutism, and acne.

Carbamazepine is used widely for seizure management but less commonly than divalproex in pediatric bipolar disorder. This anticonvulsant must be titrated slowly and requires frequent monitoring of blood levels, which can be a problem in children with needle phobia.

Carbamazepine is usually titrated to 15 mg/kg/d to produce a serum level of 7 to 10 μg/ml. Its most common side effects are sedation, rash, nausea, and hyponatremia. Aplastic anemia and severe dermatologic reactions, such as Stevens-Johnson syndrome, occur uncommonly.²⁸

Atypical antipsychotics. Recent case series and open-label reports suggest that atypical antipsychotics such as clozapine,²⁹ risperidone,³⁰ olanzapine,^31-33 and quetiapine¹⁶ are effective in treating pediatric bipolar disorder. However, clinically significant weight gain may be associated with the use of olanzapine and risperidone.³⁴ Ziprasidone may increase QTc prolongation, and safety data are limited in children and adolescents. Therefore, ziprasidone should be used with caution in pediatric bipolar disorder, and ECGs should be monitored.

In the only double-blind, placebo-controlled study of an atypical antipsychotic in pediatric bipolar disorder, manic symptoms were more greatly reduced in 15 adolescents given quetiapine plus divalproex than in 15 patients who received divalproex alone. Quetiapine was titrated to 450 mg/d across 7 days and was well-tolerated. The findings suggest that a mood stabilizer plus an atypical antipsychotic may be more effective than a mood stabilizer alone for treating adolescent mania.¹⁶

Long-term treatment

In addition to treating acute affective episodes, lithium may also help prevent recurrent affective episodes in younger patients. In the only maintenance treatment study for pediatric bipolar disorder, Strober et al prospectively evaluated 37 adolescents whose bipolar disorder had been stabilized with lithium during hospitalization.¹⁷ After 18 months of follow-up, 35% of patients had discontinued lithium, and their relapse rate was 92% (compared with 38% in patients who were lithium-compliant.

It is reasonable to maintain a child or adolescent who has had a single manic episode on mood-stabilizing treatment for several years and then—if the patient is euthymic and asymptomatic—to slowly taper the mood stabilizer(s) over several months. If mood symptoms recur, the agent(s) should be reintroduced.

If a child with bipolar disorder does not respond or only partially responds to a mood stabilizer, it may be necessary to add a second mood stabilizer or an atypical antipsychotic. A bipolar child or adolescent with psychotic symptoms should be maintained on an antipsychotic (typical or atypical) for at least 1 month, even if the psychosis has resolved.³⁵

Treatment of comorbid ADHD. Most children with bipolar disorder have comorbid ADHD, and mood stabilization is necessary prior to starting stimulant medications.³⁶ In bipolar patients, sustained-release psychostimulants may reduce rebound symptoms more effectively than immediate-release formulations. Typical dosages for a child with bipolar disorder and ADHD would be Concerta, 36 mg/d, or Adderall XR, 10 to 20 mg/d.

Psychosocial interventions

Most psychotherapeutic interventions have not been systematically studied in pediatric bipolar disorder but may be beneficial. In a recent study, Fristad et al reported the efficacy of multifamily psychoeducational group therapy for treating bipolar children and adolescents and their families.³⁷

Other useful psychosocial tactics include:

• Minimize periods of overstimulation (for example, these patients do not do well at shopping malls).
• Maintain good sleep hygiene.
• Address medication nonadherence immediately.
• Discuss the risk of substance abuse with the patient.
• Encourage mood charting by the patient and or parent.

Related resources

• Mood charts appropriate for pediatric patients.
• Child and Adolescent Bipolar Foundation. www.cabf.org
• Findling RF, Kowatch RA, Post RM. Pediatric bipolar disorders: a handbook for clinicians. London, Martin Dunitz Press, 2002.
• Geller B, DelBello MP (eds). Child and early adolescent bipolar disorder: theory, assessment, and treatment. New York: Guilford Publications, 2002.

Drug brand names

• Clozapine • Clozaril
• Divalproex sodium • Depakote
• Olanzapine • Zyprexa
• Risperidone • Risperdal
• Quetiapine • Seroquel
• Ziprasidone • Geodon

Disclosure

Dr. Kowatch receives grant support from, serves as a consultant to, or is on the speakers bureau of Novartis Pharmaceuticals Corp., Abbott Laboratories, Solvay Pharmaceuticals, Eli Lilly and Co., Janssen Pharmaceutica, GlaxoSmithKline, and AstraZeneca Pharmaceuticals.

Dr. DelBello receives grant support from, serves as a consultant to, or is on the speakers bureau of Abbott Laboratories, AstraZeneca Pharmaceuticals, Eli Lilly and Co., Ortho-McNeil Pharmaceutical, Shire Pharmaceutical Group, Janssen Pharmaceutica, Pfizer Inc., and GlaxoSmithKline.

Children and adolescents with bipolar disorder are often referred to psychiatrists because of disruptive behaviors at home and in school. They exhibit poor academic performance, disturbed interpersonal relationships, increased rates of substance abuse, legal difficulties, multiple hospitalizations, and high rates of suicide attempts and completions.^1,2 Many have comorbid psychiatric problems—particularly attention-deficit/hyperactivity disorder (ADHD).

Although few studies have examined this complex diagnosis, we do know that bipolar disorder presents differently in children and adolescents than in adults. Prodromal symptoms can appear early—before kindergarten in some children. Early recognition therefore is key to effectively treating these sick and often complicated patients.

How often a clinician encounters a child or adolescent with bipolar disorder depends largely on the practice setting (Box).^1,3,4 Wherever you practice, however, you can recognize and treat pediatric bipolar disorder if you keep in mind that its presentation and disease progression differ from the adult type.

Pediatric versus adult symptoms

Prodromal symptoms—such as episodes of depressed mood or hopelessness and excessive mood lability—have been detected in youths who later were diagnosed with bipolar disorder. More than one-half of 494 adult members of the Depression and Bipolar Support Alliance have reported that they first exhibited signs of bipolar illness before age 19, with distribution by age as follows:

• 5% before age 5
• 12% at ages 5 to 9
• 14% at ages 10 to 14
• 28% at ages 15 to 19.⁵

Box

Table 1

COMMON PRESENTING SYMPTOMS OF PEDIATRIC BIPOLAR DISORDER

In a similar study,⁶ 58 adult patients with bipolar I disorder reported an average interval of 9 to 12 years between the emergence of bipolar symptoms and the onset of a major affective disorder.

Common initial symptoms of pediatric bipolar disorder are listed in Table 1. Most of these symptoms occur in discrete episodes and represent a change from the child’s normal functioning.

Many children and adolescents are labeled “bipolar” without careful consideration of this disorder’s diagnostic complexities and subtypes. Bipolarity in young patients can be difficult to establish because of:

• variability of symptom expression, depending on the illness’ context and phase
• effects of development on symptom expression
• mood and behavioral effects of psychotropic medications the patient is taking.

Pediatric bipolar patients often present with a mixed or “dysphoric” picture characterized by frequent short periods of intense mood lability and irritability rather than classic euphoric mania.^3,7 Clinicians who evaluate children with pediatric bipolar disorders often try to fit them into the DSM-IV “rapid cycling” subtype. This subtype does not fit bipolar children very well, however, because they often lack clear episodes of mania. Rather, researchers are reporting that bipolar children cycle far more frequently than the four episodes/year in DSM-IV’s diagnostic criteria:

• Continuous, daily cycling from mania or hypomania to euthymia or depression was seen in 81% of a well-defined group of pediatric bipolar patients.^7,8
• A high rate of rapid cycling and onset of a first manic episode at mean age 7 was reported in 90 children and adolescents (mean age 11) with bipolar I disorder.⁹

The picture that emerges from independent research groups is that multiple daily mood swings and irritability are much more common than euphoria in prepubertal children with bipolar disorder.^8,10

Making the diagnosis

DSM-IV’s diagnostic classification system for bipolar disorders is complex, involving:

• five types of episodes (manic, hypomanic, mixed, depressed, unspecified)
• four severity levels (mild, moderate, severe without psychosis, severe with psychosis)
• and three course specifiers (with or without inter-episode recovery, seasonal pattern, rapid cycling).

Table 2

PEDIATRIC BIPOLAR DISORDER SUBTYPES: DIAGNOSTIC CHARACTERISTICS AND ASSOCIATED FEATURES

Table 3

COMPLICATING FACTORS IN PEDIATRIC BIPOLAR DISORDER

DSM-IV criteria for mania—which were developed from data on adults with bipolar disorders—do not take into account developmental differences between bipolar adults and bipolar children and adolescents.

Diagnostic characteristics of the pediatric bipolar disorder subtypes are compared in Table 2. Generally:

• Pediatric patients with bipolar I disorder have multiple daily mood swings, a “mixed” type of episode with short periods of euphoria and longer periods of irritability, and comorbidities such as ADHD, oppositional defiant disorder, or conduct disorder.^3,11,12
• Bipolar II disorder presents more typically in adolescence and is usually noticed clinically as a major depressive episode. Past episodes of hypomania may have been missed unless a careful history was taken.
• Cyclothymia is difficult to diagnose because the hypomania and depressive symptoms are not as severe as in bipolar types I or II. Prospective mood charting can help the clinician diagnose cyclothymia (see “Related Resources”).
• Bipolar disorder NOS represents the largest group of patients with bipolar symptoms. This diagnosis is made when bipolar symptoms are present but not of sufficient severity or duration to warrant a diagnosis of bipolar I, II or cyclothymia. Bipolar NOS also can be diagnosed when a bipolar disorder is secondary to a general medical condition, such as fetal alcohol syndrome or alcohol-related neurodevelopmental disorder.

Differential diagnosis. Medications and medical disorders may exacerbate or mimic pediatric bipolar symptoms (Table 3), so it is important to assess these potential confounds before initiating treatment. Psychiatric comorbidities also frequently complicate the presentation of pediatric bipolar disorder and its response to treatment (Table 4). ADHD is the most common comorbidity, with rates as high as 98% in bipolar children.^3,13

Outcomes

Long-term outcomes of children and adolescents with bipolar disorders have not been well studied. In the only prospective follow-up investigation of adolescent inpatients with mania, Strober et al found that most of 54 patients (96%) recovered from an index affective episode, but nearly one-half (44%) experienced one or more relapses within 5 years.¹⁴ The rate of recovery varied according to the index episode’s polarity. Recovery was faster in patients with pure mania or mixed states, and multiple relapses occurred more frequently in those with mixed or cycling episodes. Twenty percent of the patients attempted suicide.

Recently, Geller et al reported the results of the first large, prospective, follow-up study of children with bipolar disorder.¹⁵ In 89 outpatients (mean age 11) with bipolar I disorder, comprehensive assessments at baseline and at 6, 12, 18, and 24 months showed that 65% recovered from mania but 55% relapsed after recovery. Mean time to recovery was 36 weeks, and relapse occurred after a mean of 28.6 weeks. Children living with their intact biological families were twice as likely to recover as those in other living arrangements.

The poor outcomes of these bipolar children highlight the need for earlier recognition and more effective treatments.

Treating acute mania

Many psychotropic medications used to treat adults with bipolar disorders are also used for children and adolescents. To date, only two double-blind, placebo-controlled studies^13,16 and one uncontrolled maintenance treatment study¹⁷ have examined treatment of acute mania in pediatric bipolar disorder.

Lithium is the most studied medication for pediatric bipolar disorder and the only FDA-approved medication for treating acute mania and bipolar disorder in patients ages 12 to 18. Approximately 40 to 50% of children and adolescents with bipolar disorder respond to lithium monotherapy.^18,19

In general, lithium should be titrated to 30 mg/kg/d in two or three divided doses; this typically produces a serum level of 0.8 to 1.2 mEq/L. Common side effects in children and adolescents include nausea, polyuria, polydipsia, tremor, acne, and weight gain. Lithium levels and thyroid function should be monitored, as in adult patients.

Only one prospective, placebo-controlled study has examined lithium use in children and adolescents with bipolar disorders. Twenty-five adolescents with comorbid bipolar and substance use disorders were treated with lithium or placebo for 6 weeks. Positive urine toxicology screens decreased significantly, and global assessment of functioning improved in 46% of those receiving lithium vs. 8% of those receiving placebo.¹³ This study demonstrated lithium’s efficacy in treating bipolar adolescents with comorbid substance abuse but did not measure its effect on mood.

Risk factors for poor lithium response in children and adolescents with bipolar disorder include prepubertal onset and comorbid ADHD.²⁰

Divalproex. No placebo-controlled studies of antiepileptics in pediatric bipolar disorder have been published. Open-label studies of divalproex have reported response rates of 53 to 82% in manic adolescents.^{18, 21-23} Several case reports and series have described successful use of carbamazepine as monotherapy and adjunctive treatment in children and adolescents with bipolar disorder.^24,25

Table 4

RATES OF COMMON COMORBIDITIES IN PEDIATRIC BIPOLAR DISORDERS

One 6-week, random-assignment, prospective study compared lithium, divalproex, and carbamazepine in treating 42 acutely manic or hypomanic patients ages 8 to 18.¹⁸ In this open study, all three mood stabilizers demonstrated efficacy in treating a mixed or manic episode in youths with bipolar I or II disorder. Response rates—based on a 50% improvement in Young Mania Rating Scale baseline scores—were divalproex 53%, lithium 38%, and carbamazepine 38%.

In general, divalproex is started at 20 mg/kg/d, which typically produces a serum level of 80 to 120 μg/ml. Common side effects in children include weight gain, nausea, sedation, and tremor.

A possible association between divalproex and polycystic ovary syndrome (PCOS) has been reported in women with epilepsy.²⁶ The mechanism for PCOS has been hypothesized to be obesity secondary to divalproex, resulting in elevated insulin and androgen levels. Recently, Rasgon et al²⁷ reported that epilepsy—and not the anticonvulsants used to treat it—may increase the risk of PCOS. In contrast, O’Donovan et al reported higher rates of menstrual irregularities and PCOS in women with bipolar disorder who were taking divalproex than in those who were not taking divalproex and in healthy controls.²⁸

Until we learn more about this association, clinicians should monitor bipolar female adolescents treated with divalproex for any signs of PCOS, which include menstrual abnormalities, hirsutism, and acne.

Carbamazepine is used widely for seizure management but less commonly than divalproex in pediatric bipolar disorder. This anticonvulsant must be titrated slowly and requires frequent monitoring of blood levels, which can be a problem in children with needle phobia.

Carbamazepine is usually titrated to 15 mg/kg/d to produce a serum level of 7 to 10 μg/ml. Its most common side effects are sedation, rash, nausea, and hyponatremia. Aplastic anemia and severe dermatologic reactions, such as Stevens-Johnson syndrome, occur uncommonly.²⁸

Atypical antipsychotics. Recent case series and open-label reports suggest that atypical antipsychotics such as clozapine,²⁹ risperidone,³⁰ olanzapine,^31-33 and quetiapine¹⁶ are effective in treating pediatric bipolar disorder. However, clinically significant weight gain may be associated with the use of olanzapine and risperidone.³⁴ Ziprasidone may increase QTc prolongation, and safety data are limited in children and adolescents. Therefore, ziprasidone should be used with caution in pediatric bipolar disorder, and ECGs should be monitored.

In the only double-blind, placebo-controlled study of an atypical antipsychotic in pediatric bipolar disorder, manic symptoms were more greatly reduced in 15 adolescents given quetiapine plus divalproex than in 15 patients who received divalproex alone. Quetiapine was titrated to 450 mg/d across 7 days and was well-tolerated. The findings suggest that a mood stabilizer plus an atypical antipsychotic may be more effective than a mood stabilizer alone for treating adolescent mania.¹⁶

Long-term treatment

In addition to treating acute affective episodes, lithium may also help prevent recurrent affective episodes in younger patients. In the only maintenance treatment study for pediatric bipolar disorder, Strober et al prospectively evaluated 37 adolescents whose bipolar disorder had been stabilized with lithium during hospitalization.¹⁷ After 18 months of follow-up, 35% of patients had discontinued lithium, and their relapse rate was 92% (compared with 38% in patients who were lithium-compliant.

It is reasonable to maintain a child or adolescent who has had a single manic episode on mood-stabilizing treatment for several years and then—if the patient is euthymic and asymptomatic—to slowly taper the mood stabilizer(s) over several months. If mood symptoms recur, the agent(s) should be reintroduced.

If a child with bipolar disorder does not respond or only partially responds to a mood stabilizer, it may be necessary to add a second mood stabilizer or an atypical antipsychotic. A bipolar child or adolescent with psychotic symptoms should be maintained on an antipsychotic (typical or atypical) for at least 1 month, even if the psychosis has resolved.³⁵

Treatment of comorbid ADHD. Most children with bipolar disorder have comorbid ADHD, and mood stabilization is necessary prior to starting stimulant medications.³⁶ In bipolar patients, sustained-release psychostimulants may reduce rebound symptoms more effectively than immediate-release formulations. Typical dosages for a child with bipolar disorder and ADHD would be Concerta, 36 mg/d, or Adderall XR, 10 to 20 mg/d.

Psychosocial interventions

Most psychotherapeutic interventions have not been systematically studied in pediatric bipolar disorder but may be beneficial. In a recent study, Fristad et al reported the efficacy of multifamily psychoeducational group therapy for treating bipolar children and adolescents and their families.³⁷

Other useful psychosocial tactics include:

• Minimize periods of overstimulation (for example, these patients do not do well at shopping malls).
• Maintain good sleep hygiene.
• Address medication nonadherence immediately.
• Discuss the risk of substance abuse with the patient.
• Encourage mood charting by the patient and or parent.

Related resources

• Mood charts appropriate for pediatric patients.
• Child and Adolescent Bipolar Foundation. www.cabf.org
• Findling RF, Kowatch RA, Post RM. Pediatric bipolar disorders: a handbook for clinicians. London, Martin Dunitz Press, 2002.
• Geller B, DelBello MP (eds). Child and early adolescent bipolar disorder: theory, assessment, and treatment. New York: Guilford Publications, 2002.

Drug brand names

• Clozapine • Clozaril
• Divalproex sodium • Depakote
• Olanzapine • Zyprexa
• Risperidone • Risperdal
• Quetiapine • Seroquel
• Ziprasidone • Geodon

Disclosure

Dr. Kowatch receives grant support from, serves as a consultant to, or is on the speakers bureau of Novartis Pharmaceuticals Corp., Abbott Laboratories, Solvay Pharmaceuticals, Eli Lilly and Co., Janssen Pharmaceutica, GlaxoSmithKline, and AstraZeneca Pharmaceuticals.

Dr. DelBello receives grant support from, serves as a consultant to, or is on the speakers bureau of Abbott Laboratories, AstraZeneca Pharmaceuticals, Eli Lilly and Co., Ortho-McNeil Pharmaceutical, Shire Pharmaceutical Group, Janssen Pharmaceutica, Pfizer Inc., and GlaxoSmithKline.

Children and adolescents with bipolar disorder are often referred to psychiatrists because of disruptive behaviors at home and in school. They exhibit poor academic performance, disturbed interpersonal relationships, increased rates of substance abuse, legal difficulties, multiple hospitalizations, and high rates of suicide attempts and completions.^1,2 Many have comorbid psychiatric problems—particularly attention-deficit/hyperactivity disorder (ADHD).

Although few studies have examined this complex diagnosis, we do know that bipolar disorder presents differently in children and adolescents than in adults. Prodromal symptoms can appear early—before kindergarten in some children. Early recognition therefore is key to effectively treating these sick and often complicated patients.

How often a clinician encounters a child or adolescent with bipolar disorder depends largely on the practice setting (Box).^1,3,4 Wherever you practice, however, you can recognize and treat pediatric bipolar disorder if you keep in mind that its presentation and disease progression differ from the adult type.

Pediatric versus adult symptoms

Prodromal symptoms—such as episodes of depressed mood or hopelessness and excessive mood lability—have been detected in youths who later were diagnosed with bipolar disorder. More than one-half of 494 adult members of the Depression and Bipolar Support Alliance have reported that they first exhibited signs of bipolar illness before age 19, with distribution by age as follows:

• 5% before age 5
• 12% at ages 5 to 9
• 14% at ages 10 to 14
• 28% at ages 15 to 19.⁵

Box

Table 1

COMMON PRESENTING SYMPTOMS OF PEDIATRIC BIPOLAR DISORDER

In a similar study,⁶ 58 adult patients with bipolar I disorder reported an average interval of 9 to 12 years between the emergence of bipolar symptoms and the onset of a major affective disorder.

Common initial symptoms of pediatric bipolar disorder are listed in Table 1. Most of these symptoms occur in discrete episodes and represent a change from the child’s normal functioning.

Many children and adolescents are labeled “bipolar” without careful consideration of this disorder’s diagnostic complexities and subtypes. Bipolarity in young patients can be difficult to establish because of:

• variability of symptom expression, depending on the illness’ context and phase
• effects of development on symptom expression
• mood and behavioral effects of psychotropic medications the patient is taking.

Pediatric bipolar patients often present with a mixed or “dysphoric” picture characterized by frequent short periods of intense mood lability and irritability rather than classic euphoric mania.^3,7 Clinicians who evaluate children with pediatric bipolar disorders often try to fit them into the DSM-IV “rapid cycling” subtype. This subtype does not fit bipolar children very well, however, because they often lack clear episodes of mania. Rather, researchers are reporting that bipolar children cycle far more frequently than the four episodes/year in DSM-IV’s diagnostic criteria:

• Continuous, daily cycling from mania or hypomania to euthymia or depression was seen in 81% of a well-defined group of pediatric bipolar patients.^7,8
• A high rate of rapid cycling and onset of a first manic episode at mean age 7 was reported in 90 children and adolescents (mean age 11) with bipolar I disorder.⁹

The picture that emerges from independent research groups is that multiple daily mood swings and irritability are much more common than euphoria in prepubertal children with bipolar disorder.^8,10

Making the diagnosis

DSM-IV’s diagnostic classification system for bipolar disorders is complex, involving:

• five types of episodes (manic, hypomanic, mixed, depressed, unspecified)
• four severity levels (mild, moderate, severe without psychosis, severe with psychosis)
• and three course specifiers (with or without inter-episode recovery, seasonal pattern, rapid cycling).

Table 2

PEDIATRIC BIPOLAR DISORDER SUBTYPES: DIAGNOSTIC CHARACTERISTICS AND ASSOCIATED FEATURES

Table 3

COMPLICATING FACTORS IN PEDIATRIC BIPOLAR DISORDER

DSM-IV criteria for mania—which were developed from data on adults with bipolar disorders—do not take into account developmental differences between bipolar adults and bipolar children and adolescents.

Diagnostic characteristics of the pediatric bipolar disorder subtypes are compared in Table 2. Generally:

• Pediatric patients with bipolar I disorder have multiple daily mood swings, a “mixed” type of episode with short periods of euphoria and longer periods of irritability, and comorbidities such as ADHD, oppositional defiant disorder, or conduct disorder.^3,11,12
• Bipolar II disorder presents more typically in adolescence and is usually noticed clinically as a major depressive episode. Past episodes of hypomania may have been missed unless a careful history was taken.
• Cyclothymia is difficult to diagnose because the hypomania and depressive symptoms are not as severe as in bipolar types I or II. Prospective mood charting can help the clinician diagnose cyclothymia (see “Related Resources”).
• Bipolar disorder NOS represents the largest group of patients with bipolar symptoms. This diagnosis is made when bipolar symptoms are present but not of sufficient severity or duration to warrant a diagnosis of bipolar I, II or cyclothymia. Bipolar NOS also can be diagnosed when a bipolar disorder is secondary to a general medical condition, such as fetal alcohol syndrome or alcohol-related neurodevelopmental disorder.

Differential diagnosis. Medications and medical disorders may exacerbate or mimic pediatric bipolar symptoms (Table 3), so it is important to assess these potential confounds before initiating treatment. Psychiatric comorbidities also frequently complicate the presentation of pediatric bipolar disorder and its response to treatment (Table 4). ADHD is the most common comorbidity, with rates as high as 98% in bipolar children.^3,13

Outcomes

Long-term outcomes of children and adolescents with bipolar disorders have not been well studied. In the only prospective follow-up investigation of adolescent inpatients with mania, Strober et al found that most of 54 patients (96%) recovered from an index affective episode, but nearly one-half (44%) experienced one or more relapses within 5 years.¹⁴ The rate of recovery varied according to the index episode’s polarity. Recovery was faster in patients with pure mania or mixed states, and multiple relapses occurred more frequently in those with mixed or cycling episodes. Twenty percent of the patients attempted suicide.

Recently, Geller et al reported the results of the first large, prospective, follow-up study of children with bipolar disorder.¹⁵ In 89 outpatients (mean age 11) with bipolar I disorder, comprehensive assessments at baseline and at 6, 12, 18, and 24 months showed that 65% recovered from mania but 55% relapsed after recovery. Mean time to recovery was 36 weeks, and relapse occurred after a mean of 28.6 weeks. Children living with their intact biological families were twice as likely to recover as those in other living arrangements.

The poor outcomes of these bipolar children highlight the need for earlier recognition and more effective treatments.

Treating acute mania

Many psychotropic medications used to treat adults with bipolar disorders are also used for children and adolescents. To date, only two double-blind, placebo-controlled studies^13,16 and one uncontrolled maintenance treatment study¹⁷ have examined treatment of acute mania in pediatric bipolar disorder.

Lithium is the most studied medication for pediatric bipolar disorder and the only FDA-approved medication for treating acute mania and bipolar disorder in patients ages 12 to 18. Approximately 40 to 50% of children and adolescents with bipolar disorder respond to lithium monotherapy.^18,19

In general, lithium should be titrated to 30 mg/kg/d in two or three divided doses; this typically produces a serum level of 0.8 to 1.2 mEq/L. Common side effects in children and adolescents include nausea, polyuria, polydipsia, tremor, acne, and weight gain. Lithium levels and thyroid function should be monitored, as in adult patients.

Only one prospective, placebo-controlled study has examined lithium use in children and adolescents with bipolar disorders. Twenty-five adolescents with comorbid bipolar and substance use disorders were treated with lithium or placebo for 6 weeks. Positive urine toxicology screens decreased significantly, and global assessment of functioning improved in 46% of those receiving lithium vs. 8% of those receiving placebo.¹³ This study demonstrated lithium’s efficacy in treating bipolar adolescents with comorbid substance abuse but did not measure its effect on mood.

Risk factors for poor lithium response in children and adolescents with bipolar disorder include prepubertal onset and comorbid ADHD.²⁰

Divalproex. No placebo-controlled studies of antiepileptics in pediatric bipolar disorder have been published. Open-label studies of divalproex have reported response rates of 53 to 82% in manic adolescents.^{18, 21-23} Several case reports and series have described successful use of carbamazepine as monotherapy and adjunctive treatment in children and adolescents with bipolar disorder.^24,25

Table 4

RATES OF COMMON COMORBIDITIES IN PEDIATRIC BIPOLAR DISORDERS

One 6-week, random-assignment, prospective study compared lithium, divalproex, and carbamazepine in treating 42 acutely manic or hypomanic patients ages 8 to 18.¹⁸ In this open study, all three mood stabilizers demonstrated efficacy in treating a mixed or manic episode in youths with bipolar I or II disorder. Response rates—based on a 50% improvement in Young Mania Rating Scale baseline scores—were divalproex 53%, lithium 38%, and carbamazepine 38%.

In general, divalproex is started at 20 mg/kg/d, which typically produces a serum level of 80 to 120 μg/ml. Common side effects in children include weight gain, nausea, sedation, and tremor.

A possible association between divalproex and polycystic ovary syndrome (PCOS) has been reported in women with epilepsy.²⁶ The mechanism for PCOS has been hypothesized to be obesity secondary to divalproex, resulting in elevated insulin and androgen levels. Recently, Rasgon et al²⁷ reported that epilepsy—and not the anticonvulsants used to treat it—may increase the risk of PCOS. In contrast, O’Donovan et al reported higher rates of menstrual irregularities and PCOS in women with bipolar disorder who were taking divalproex than in those who were not taking divalproex and in healthy controls.²⁸

Until we learn more about this association, clinicians should monitor bipolar female adolescents treated with divalproex for any signs of PCOS, which include menstrual abnormalities, hirsutism, and acne.

Carbamazepine is used widely for seizure management but less commonly than divalproex in pediatric bipolar disorder. This anticonvulsant must be titrated slowly and requires frequent monitoring of blood levels, which can be a problem in children with needle phobia.

Carbamazepine is usually titrated to 15 mg/kg/d to produce a serum level of 7 to 10 μg/ml. Its most common side effects are sedation, rash, nausea, and hyponatremia. Aplastic anemia and severe dermatologic reactions, such as Stevens-Johnson syndrome, occur uncommonly.²⁸

Atypical antipsychotics. Recent case series and open-label reports suggest that atypical antipsychotics such as clozapine,²⁹ risperidone,³⁰ olanzapine,^31-33 and quetiapine¹⁶ are effective in treating pediatric bipolar disorder. However, clinically significant weight gain may be associated with the use of olanzapine and risperidone.³⁴ Ziprasidone may increase QTc prolongation, and safety data are limited in children and adolescents. Therefore, ziprasidone should be used with caution in pediatric bipolar disorder, and ECGs should be monitored.

In the only double-blind, placebo-controlled study of an atypical antipsychotic in pediatric bipolar disorder, manic symptoms were more greatly reduced in 15 adolescents given quetiapine plus divalproex than in 15 patients who received divalproex alone. Quetiapine was titrated to 450 mg/d across 7 days and was well-tolerated. The findings suggest that a mood stabilizer plus an atypical antipsychotic may be more effective than a mood stabilizer alone for treating adolescent mania.¹⁶

Long-term treatment

In addition to treating acute affective episodes, lithium may also help prevent recurrent affective episodes in younger patients. In the only maintenance treatment study for pediatric bipolar disorder, Strober et al prospectively evaluated 37 adolescents whose bipolar disorder had been stabilized with lithium during hospitalization.¹⁷ After 18 months of follow-up, 35% of patients had discontinued lithium, and their relapse rate was 92% (compared with 38% in patients who were lithium-compliant.

It is reasonable to maintain a child or adolescent who has had a single manic episode on mood-stabilizing treatment for several years and then—if the patient is euthymic and asymptomatic—to slowly taper the mood stabilizer(s) over several months. If mood symptoms recur, the agent(s) should be reintroduced.

If a child with bipolar disorder does not respond or only partially responds to a mood stabilizer, it may be necessary to add a second mood stabilizer or an atypical antipsychotic. A bipolar child or adolescent with psychotic symptoms should be maintained on an antipsychotic (typical or atypical) for at least 1 month, even if the psychosis has resolved.³⁵

Treatment of comorbid ADHD. Most children with bipolar disorder have comorbid ADHD, and mood stabilization is necessary prior to starting stimulant medications.³⁶ In bipolar patients, sustained-release psychostimulants may reduce rebound symptoms more effectively than immediate-release formulations. Typical dosages for a child with bipolar disorder and ADHD would be Concerta, 36 mg/d, or Adderall XR, 10 to 20 mg/d.

Psychosocial interventions

Most psychotherapeutic interventions have not been systematically studied in pediatric bipolar disorder but may be beneficial. In a recent study, Fristad et al reported the efficacy of multifamily psychoeducational group therapy for treating bipolar children and adolescents and their families.³⁷

Other useful psychosocial tactics include:

• Minimize periods of overstimulation (for example, these patients do not do well at shopping malls).
• Maintain good sleep hygiene.
• Address medication nonadherence immediately.
• Discuss the risk of substance abuse with the patient.
• Encourage mood charting by the patient and or parent.

Related resources

• Mood charts appropriate for pediatric patients.
• Child and Adolescent Bipolar Foundation. www.cabf.org
• Findling RF, Kowatch RA, Post RM. Pediatric bipolar disorders: a handbook for clinicians. London, Martin Dunitz Press, 2002.
• Geller B, DelBello MP (eds). Child and early adolescent bipolar disorder: theory, assessment, and treatment. New York: Guilford Publications, 2002.

Drug brand names

• Clozapine • Clozaril
• Divalproex sodium • Depakote
• Olanzapine • Zyprexa
• Risperidone • Risperdal
• Quetiapine • Seroquel
• Ziprasidone • Geodon

Disclosure

Dr. Kowatch receives grant support from, serves as a consultant to, or is on the speakers bureau of Novartis Pharmaceuticals Corp., Abbott Laboratories, Solvay Pharmaceuticals, Eli Lilly and Co., Janssen Pharmaceutica, GlaxoSmithKline, and AstraZeneca Pharmaceuticals.

Dr. DelBello receives grant support from, serves as a consultant to, or is on the speakers bureau of Abbott Laboratories, AstraZeneca Pharmaceuticals, Eli Lilly and Co., Ortho-McNeil Pharmaceutical, Shire Pharmaceutical Group, Janssen Pharmaceutica, Pfizer Inc., and GlaxoSmithKline.

Children and adolescents with bipolar disorder are often referred to psychiatrists because of disruptive behaviors at home and in school. They exhibit poor academic performance, disturbed interpersonal relationships, increased rates of substance abuse, legal difficulties, multiple hospitalizations, and high rates of suicide attempts and completions.^1,2 Many have comorbid psychiatric problems—particularly attention-deficit/hyperactivity disorder (ADHD).

Although few studies have examined this complex diagnosis, we do know that bipolar disorder presents differently in children and adolescents than in adults. Prodromal symptoms can appear early—before kindergarten in some children. Early recognition therefore is key to effectively treating these sick and often complicated patients.

How often a clinician encounters a child or adolescent with bipolar disorder depends largely on the practice setting (Box).^1,3,4 Wherever you practice, however, you can recognize and treat pediatric bipolar disorder if you keep in mind that its presentation and disease progression differ from the adult type.

Pediatric versus adult symptoms

Prodromal symptoms—such as episodes of depressed mood or hopelessness and excessive mood lability—have been detected in youths who later were diagnosed with bipolar disorder. More than one-half of 494 adult members of the Depression and Bipolar Support Alliance have reported that they first exhibited signs of bipolar illness before age 19, with distribution by age as follows:

• 5% before age 5
• 12% at ages 5 to 9
• 14% at ages 10 to 14
• 28% at ages 15 to 19.⁵

Box

Table 1

COMMON PRESENTING SYMPTOMS OF PEDIATRIC BIPOLAR DISORDER

In a similar study,⁶ 58 adult patients with bipolar I disorder reported an average interval of 9 to 12 years between the emergence of bipolar symptoms and the onset of a major affective disorder.

Common initial symptoms of pediatric bipolar disorder are listed in Table 1. Most of these symptoms occur in discrete episodes and represent a change from the child’s normal functioning.

Many children and adolescents are labeled “bipolar” without careful consideration of this disorder’s diagnostic complexities and subtypes. Bipolarity in young patients can be difficult to establish because of:

• variability of symptom expression, depending on the illness’ context and phase
• effects of development on symptom expression
• mood and behavioral effects of psychotropic medications the patient is taking.

Pediatric bipolar patients often present with a mixed or “dysphoric” picture characterized by frequent short periods of intense mood lability and irritability rather than classic euphoric mania.^3,7 Clinicians who evaluate children with pediatric bipolar disorders often try to fit them into the DSM-IV “rapid cycling” subtype. This subtype does not fit bipolar children very well, however, because they often lack clear episodes of mania. Rather, researchers are reporting that bipolar children cycle far more frequently than the four episodes/year in DSM-IV’s diagnostic criteria:

• Continuous, daily cycling from mania or hypomania to euthymia or depression was seen in 81% of a well-defined group of pediatric bipolar patients.^7,8
• A high rate of rapid cycling and onset of a first manic episode at mean age 7 was reported in 90 children and adolescents (mean age 11) with bipolar I disorder.⁹

The picture that emerges from independent research groups is that multiple daily mood swings and irritability are much more common than euphoria in prepubertal children with bipolar disorder.^8,10

Making the diagnosis

DSM-IV’s diagnostic classification system for bipolar disorders is complex, involving:

• five types of episodes (manic, hypomanic, mixed, depressed, unspecified)
• four severity levels (mild, moderate, severe without psychosis, severe with psychosis)
• and three course specifiers (with or without inter-episode recovery, seasonal pattern, rapid cycling).

Table 2

PEDIATRIC BIPOLAR DISORDER SUBTYPES: DIAGNOSTIC CHARACTERISTICS AND ASSOCIATED FEATURES

Table 3

COMPLICATING FACTORS IN PEDIATRIC BIPOLAR DISORDER

DSM-IV criteria for mania—which were developed from data on adults with bipolar disorders—do not take into account developmental differences between bipolar adults and bipolar children and adolescents.

Diagnostic characteristics of the pediatric bipolar disorder subtypes are compared in Table 2. Generally:

• Pediatric patients with bipolar I disorder have multiple daily mood swings, a “mixed” type of episode with short periods of euphoria and longer periods of irritability, and comorbidities such as ADHD, oppositional defiant disorder, or conduct disorder.^3,11,12
• Bipolar II disorder presents more typically in adolescence and is usually noticed clinically as a major depressive episode. Past episodes of hypomania may have been missed unless a careful history was taken.
• Cyclothymia is difficult to diagnose because the hypomania and depressive symptoms are not as severe as in bipolar types I or II. Prospective mood charting can help the clinician diagnose cyclothymia (see “Related Resources”).
• Bipolar disorder NOS represents the largest group of patients with bipolar symptoms. This diagnosis is made when bipolar symptoms are present but not of sufficient severity or duration to warrant a diagnosis of bipolar I, II or cyclothymia. Bipolar NOS also can be diagnosed when a bipolar disorder is secondary to a general medical condition, such as fetal alcohol syndrome or alcohol-related neurodevelopmental disorder.

Differential diagnosis. Medications and medical disorders may exacerbate or mimic pediatric bipolar symptoms (Table 3), so it is important to assess these potential confounds before initiating treatment. Psychiatric comorbidities also frequently complicate the presentation of pediatric bipolar disorder and its response to treatment (Table 4). ADHD is the most common comorbidity, with rates as high as 98% in bipolar children.^3,13

Outcomes

Long-term outcomes of children and adolescents with bipolar disorders have not been well studied. In the only prospective follow-up investigation of adolescent inpatients with mania, Strober et al found that most of 54 patients (96%) recovered from an index affective episode, but nearly one-half (44%) experienced one or more relapses within 5 years.¹⁴ The rate of recovery varied according to the index episode’s polarity. Recovery was faster in patients with pure mania or mixed states, and multiple relapses occurred more frequently in those with mixed or cycling episodes. Twenty percent of the patients attempted suicide.

Recently, Geller et al reported the results of the first large, prospective, follow-up study of children with bipolar disorder.¹⁵ In 89 outpatients (mean age 11) with bipolar I disorder, comprehensive assessments at baseline and at 6, 12, 18, and 24 months showed that 65% recovered from mania but 55% relapsed after recovery. Mean time to recovery was 36 weeks, and relapse occurred after a mean of 28.6 weeks. Children living with their intact biological families were twice as likely to recover as those in other living arrangements.

The poor outcomes of these bipolar children highlight the need for earlier recognition and more effective treatments.

Treating acute mania

Many psychotropic medications used to treat adults with bipolar disorders are also used for children and adolescents. To date, only two double-blind, placebo-controlled studies^13,16 and one uncontrolled maintenance treatment study¹⁷ have examined treatment of acute mania in pediatric bipolar disorder.

Lithium is the most studied medication for pediatric bipolar disorder and the only FDA-approved medication for treating acute mania and bipolar disorder in patients ages 12 to 18. Approximately 40 to 50% of children and adolescents with bipolar disorder respond to lithium monotherapy.^18,19

In general, lithium should be titrated to 30 mg/kg/d in two or three divided doses; this typically produces a serum level of 0.8 to 1.2 mEq/L. Common side effects in children and adolescents include nausea, polyuria, polydipsia, tremor, acne, and weight gain. Lithium levels and thyroid function should be monitored, as in adult patients.

Only one prospective, placebo-controlled study has examined lithium use in children and adolescents with bipolar disorders. Twenty-five adolescents with comorbid bipolar and substance use disorders were treated with lithium or placebo for 6 weeks. Positive urine toxicology screens decreased significantly, and global assessment of functioning improved in 46% of those receiving lithium vs. 8% of those receiving placebo.¹³ This study demonstrated lithium’s efficacy in treating bipolar adolescents with comorbid substance abuse but did not measure its effect on mood.

Risk factors for poor lithium response in children and adolescents with bipolar disorder include prepubertal onset and comorbid ADHD.²⁰

Divalproex. No placebo-controlled studies of antiepileptics in pediatric bipolar disorder have been published. Open-label studies of divalproex have reported response rates of 53 to 82% in manic adolescents.^{18, 21-23} Several case reports and series have described successful use of carbamazepine as monotherapy and adjunctive treatment in children and adolescents with bipolar disorder.^24,25

Table 4

RATES OF COMMON COMORBIDITIES IN PEDIATRIC BIPOLAR DISORDERS

One 6-week, random-assignment, prospective study compared lithium, divalproex, and carbamazepine in treating 42 acutely manic or hypomanic patients ages 8 to 18.¹⁸ In this open study, all three mood stabilizers demonstrated efficacy in treating a mixed or manic episode in youths with bipolar I or II disorder. Response rates—based on a 50% improvement in Young Mania Rating Scale baseline scores—were divalproex 53%, lithium 38%, and carbamazepine 38%.

In general, divalproex is started at 20 mg/kg/d, which typically produces a serum level of 80 to 120 μg/ml. Common side effects in children include weight gain, nausea, sedation, and tremor.

A possible association between divalproex and polycystic ovary syndrome (PCOS) has been reported in women with epilepsy.²⁶ The mechanism for PCOS has been hypothesized to be obesity secondary to divalproex, resulting in elevated insulin and androgen levels. Recently, Rasgon et al²⁷ reported that epilepsy—and not the anticonvulsants used to treat it—may increase the risk of PCOS. In contrast, O’Donovan et al reported higher rates of menstrual irregularities and PCOS in women with bipolar disorder who were taking divalproex than in those who were not taking divalproex and in healthy controls.²⁸

Until we learn more about this association, clinicians should monitor bipolar female adolescents treated with divalproex for any signs of PCOS, which include menstrual abnormalities, hirsutism, and acne.

Carbamazepine is used widely for seizure management but less commonly than divalproex in pediatric bipolar disorder. This anticonvulsant must be titrated slowly and requires frequent monitoring of blood levels, which can be a problem in children with needle phobia.

Carbamazepine is usually titrated to 15 mg/kg/d to produce a serum level of 7 to 10 μg/ml. Its most common side effects are sedation, rash, nausea, and hyponatremia. Aplastic anemia and severe dermatologic reactions, such as Stevens-Johnson syndrome, occur uncommonly.²⁸

Atypical antipsychotics. Recent case series and open-label reports suggest that atypical antipsychotics such as clozapine,²⁹ risperidone,³⁰ olanzapine,^31-33 and quetiapine¹⁶ are effective in treating pediatric bipolar disorder. However, clinically significant weight gain may be associated with the use of olanzapine and risperidone.³⁴ Ziprasidone may increase QTc prolongation, and safety data are limited in children and adolescents. Therefore, ziprasidone should be used with caution in pediatric bipolar disorder, and ECGs should be monitored.

In the only double-blind, placebo-controlled study of an atypical antipsychotic in pediatric bipolar disorder, manic symptoms were more greatly reduced in 15 adolescents given quetiapine plus divalproex than in 15 patients who received divalproex alone. Quetiapine was titrated to 450 mg/d across 7 days and was well-tolerated. The findings suggest that a mood stabilizer plus an atypical antipsychotic may be more effective than a mood stabilizer alone for treating adolescent mania.¹⁶

Long-term treatment

In addition to treating acute affective episodes, lithium may also help prevent recurrent affective episodes in younger patients. In the only maintenance treatment study for pediatric bipolar disorder, Strober et al prospectively evaluated 37 adolescents whose bipolar disorder had been stabilized with lithium during hospitalization.¹⁷ After 18 months of follow-up, 35% of patients had discontinued lithium, and their relapse rate was 92% (compared with 38% in patients who were lithium-compliant.

It is reasonable to maintain a child or adolescent who has had a single manic episode on mood-stabilizing treatment for several years and then—if the patient is euthymic and asymptomatic—to slowly taper the mood stabilizer(s) over several months. If mood symptoms recur, the agent(s) should be reintroduced.

If a child with bipolar disorder does not respond or only partially responds to a mood stabilizer, it may be necessary to add a second mood stabilizer or an atypical antipsychotic. A bipolar child or adolescent with psychotic symptoms should be maintained on an antipsychotic (typical or atypical) for at least 1 month, even if the psychosis has resolved.³⁵

Treatment of comorbid ADHD. Most children with bipolar disorder have comorbid ADHD, and mood stabilization is necessary prior to starting stimulant medications.³⁶ In bipolar patients, sustained-release psychostimulants may reduce rebound symptoms more effectively than immediate-release formulations. Typical dosages for a child with bipolar disorder and ADHD would be Concerta, 36 mg/d, or Adderall XR, 10 to 20 mg/d.

Psychosocial interventions

Most psychotherapeutic interventions have not been systematically studied in pediatric bipolar disorder but may be beneficial. In a recent study, Fristad et al reported the efficacy of multifamily psychoeducational group therapy for treating bipolar children and adolescents and their families.³⁷

Other useful psychosocial tactics include:

• Minimize periods of overstimulation (for example, these patients do not do well at shopping malls).
• Maintain good sleep hygiene.
• Address medication nonadherence immediately.
• Discuss the risk of substance abuse with the patient.
• Encourage mood charting by the patient and or parent.

Related resources

• Mood charts appropriate for pediatric patients.
• Child and Adolescent Bipolar Foundation. www.cabf.org
• Findling RF, Kowatch RA, Post RM. Pediatric bipolar disorders: a handbook for clinicians. London, Martin Dunitz Press, 2002.
• Geller B, DelBello MP (eds). Child and early adolescent bipolar disorder: theory, assessment, and treatment. New York: Guilford Publications, 2002.

Drug brand names

• Clozapine • Clozaril
• Divalproex sodium • Depakote
• Olanzapine • Zyprexa
• Risperidone • Risperdal
• Quetiapine • Seroquel
• Ziprasidone • Geodon

Disclosure

Dr. Kowatch receives grant support from, serves as a consultant to, or is on the speakers bureau of Novartis Pharmaceuticals Corp., Abbott Laboratories, Solvay Pharmaceuticals, Eli Lilly and Co., Janssen Pharmaceutica, GlaxoSmithKline, and AstraZeneca Pharmaceuticals.

Dr. DelBello receives grant support from, serves as a consultant to, or is on the speakers bureau of Abbott Laboratories, AstraZeneca Pharmaceuticals, Eli Lilly and Co., Ortho-McNeil Pharmaceutical, Shire Pharmaceutical Group, Janssen Pharmaceutica, Pfizer Inc., and GlaxoSmithKline.

Children and adolescents with bipolar disorder are often referred to psychiatrists because of disruptive behaviors at home and in school. They exhibit poor academic performance, disturbed interpersonal relationships, increased rates of substance abuse, legal difficulties, multiple hospitalizations, and high rates of suicide attempts and completions.^1,2 Many have comorbid psychiatric problems—particularly attention-deficit/hyperactivity disorder (ADHD).

Although few studies have examined this complex diagnosis, we do know that bipolar disorder presents differently in children and adolescents than in adults. Prodromal symptoms can appear early—before kindergarten in some children. Early recognition therefore is key to effectively treating these sick and often complicated patients.

How often a clinician encounters a child or adolescent with bipolar disorder depends largely on the practice setting (Box).^1,3,4 Wherever you practice, however, you can recognize and treat pediatric bipolar disorder if you keep in mind that its presentation and disease progression differ from the adult type.

Pediatric versus adult symptoms

Prodromal symptoms—such as episodes of depressed mood or hopelessness and excessive mood lability—have been detected in youths who later were diagnosed with bipolar disorder. More than one-half of 494 adult members of the Depression and Bipolar Support Alliance have reported that they first exhibited signs of bipolar illness before age 19, with distribution by age as follows:

• 5% before age 5
• 12% at ages 5 to 9
• 14% at ages 10 to 14
• 28% at ages 15 to 19.⁵

Box

Table 1

COMMON PRESENTING SYMPTOMS OF PEDIATRIC BIPOLAR DISORDER

In a similar study,⁶ 58 adult patients with bipolar I disorder reported an average interval of 9 to 12 years between the emergence of bipolar symptoms and the onset of a major affective disorder.

Common initial symptoms of pediatric bipolar disorder are listed in Table 1. Most of these symptoms occur in discrete episodes and represent a change from the child’s normal functioning.

Many children and adolescents are labeled “bipolar” without careful consideration of this disorder’s diagnostic complexities and subtypes. Bipolarity in young patients can be difficult to establish because of:

• variability of symptom expression, depending on the illness’ context and phase
• effects of development on symptom expression
• mood and behavioral effects of psychotropic medications the patient is taking.

Pediatric bipolar patients often present with a mixed or “dysphoric” picture characterized by frequent short periods of intense mood lability and irritability rather than classic euphoric mania.^3,7 Clinicians who evaluate children with pediatric bipolar disorders often try to fit them into the DSM-IV “rapid cycling” subtype. This subtype does not fit bipolar children very well, however, because they often lack clear episodes of mania. Rather, researchers are reporting that bipolar children cycle far more frequently than the four episodes/year in DSM-IV’s diagnostic criteria:

• Continuous, daily cycling from mania or hypomania to euthymia or depression was seen in 81% of a well-defined group of pediatric bipolar patients.^7,8
• A high rate of rapid cycling and onset of a first manic episode at mean age 7 was reported in 90 children and adolescents (mean age 11) with bipolar I disorder.⁹

The picture that emerges from independent research groups is that multiple daily mood swings and irritability are much more common than euphoria in prepubertal children with bipolar disorder.^8,10

Making the diagnosis

DSM-IV’s diagnostic classification system for bipolar disorders is complex, involving:

• five types of episodes (manic, hypomanic, mixed, depressed, unspecified)
• four severity levels (mild, moderate, severe without psychosis, severe with psychosis)
• and three course specifiers (with or without inter-episode recovery, seasonal pattern, rapid cycling).

Table 2

PEDIATRIC BIPOLAR DISORDER SUBTYPES: DIAGNOSTIC CHARACTERISTICS AND ASSOCIATED FEATURES

Table 3

COMPLICATING FACTORS IN PEDIATRIC BIPOLAR DISORDER

DSM-IV criteria for mania—which were developed from data on adults with bipolar disorders—do not take into account developmental differences between bipolar adults and bipolar children and adolescents.

Diagnostic characteristics of the pediatric bipolar disorder subtypes are compared in Table 2. Generally:

• Pediatric patients with bipolar I disorder have multiple daily mood swings, a “mixed” type of episode with short periods of euphoria and longer periods of irritability, and comorbidities such as ADHD, oppositional defiant disorder, or conduct disorder.^3,11,12
• Bipolar II disorder presents more typically in adolescence and is usually noticed clinically as a major depressive episode. Past episodes of hypomania may have been missed unless a careful history was taken.
• Cyclothymia is difficult to diagnose because the hypomania and depressive symptoms are not as severe as in bipolar types I or II. Prospective mood charting can help the clinician diagnose cyclothymia (see “Related Resources”).
• Bipolar disorder NOS represents the largest group of patients with bipolar symptoms. This diagnosis is made when bipolar symptoms are present but not of sufficient severity or duration to warrant a diagnosis of bipolar I, II or cyclothymia. Bipolar NOS also can be diagnosed when a bipolar disorder is secondary to a general medical condition, such as fetal alcohol syndrome or alcohol-related neurodevelopmental disorder.

Differential diagnosis. Medications and medical disorders may exacerbate or mimic pediatric bipolar symptoms (Table 3), so it is important to assess these potential confounds before initiating treatment. Psychiatric comorbidities also frequently complicate the presentation of pediatric bipolar disorder and its response to treatment (Table 4). ADHD is the most common comorbidity, with rates as high as 98% in bipolar children.^3,13

Outcomes

Long-term outcomes of children and adolescents with bipolar disorders have not been well studied. In the only prospective follow-up investigation of adolescent inpatients with mania, Strober et al found that most of 54 patients (96%) recovered from an index affective episode, but nearly one-half (44%) experienced one or more relapses within 5 years.¹⁴ The rate of recovery varied according to the index episode’s polarity. Recovery was faster in patients with pure mania or mixed states, and multiple relapses occurred more frequently in those with mixed or cycling episodes. Twenty percent of the patients attempted suicide.

Recently, Geller et al reported the results of the first large, prospective, follow-up study of children with bipolar disorder.¹⁵ In 89 outpatients (mean age 11) with bipolar I disorder, comprehensive assessments at baseline and at 6, 12, 18, and 24 months showed that 65% recovered from mania but 55% relapsed after recovery. Mean time to recovery was 36 weeks, and relapse occurred after a mean of 28.6 weeks. Children living with their intact biological families were twice as likely to recover as those in other living arrangements.

The poor outcomes of these bipolar children highlight the need for earlier recognition and more effective treatments.

Treating acute mania

Many psychotropic medications used to treat adults with bipolar disorders are also used for children and adolescents. To date, only two double-blind, placebo-controlled studies^13,16 and one uncontrolled maintenance treatment study¹⁷ have examined treatment of acute mania in pediatric bipolar disorder.

Lithium is the most studied medication for pediatric bipolar disorder and the only FDA-approved medication for treating acute mania and bipolar disorder in patients ages 12 to 18. Approximately 40 to 50% of children and adolescents with bipolar disorder respond to lithium monotherapy.^18,19

In general, lithium should be titrated to 30 mg/kg/d in two or three divided doses; this typically produces a serum level of 0.8 to 1.2 mEq/L. Common side effects in children and adolescents include nausea, polyuria, polydipsia, tremor, acne, and weight gain. Lithium levels and thyroid function should be monitored, as in adult patients.

Only one prospective, placebo-controlled study has examined lithium use in children and adolescents with bipolar disorders. Twenty-five adolescents with comorbid bipolar and substance use disorders were treated with lithium or placebo for 6 weeks. Positive urine toxicology screens decreased significantly, and global assessment of functioning improved in 46% of those receiving lithium vs. 8% of those receiving placebo.¹³ This study demonstrated lithium’s efficacy in treating bipolar adolescents with comorbid substance abuse but did not measure its effect on mood.

Risk factors for poor lithium response in children and adolescents with bipolar disorder include prepubertal onset and comorbid ADHD.²⁰

Divalproex. No placebo-controlled studies of antiepileptics in pediatric bipolar disorder have been published. Open-label studies of divalproex have reported response rates of 53 to 82% in manic adolescents.^{18, 21-23} Several case reports and series have described successful use of carbamazepine as monotherapy and adjunctive treatment in children and adolescents with bipolar disorder.^24,25

Table 4

RATES OF COMMON COMORBIDITIES IN PEDIATRIC BIPOLAR DISORDERS

One 6-week, random-assignment, prospective study compared lithium, divalproex, and carbamazepine in treating 42 acutely manic or hypomanic patients ages 8 to 18.¹⁸ In this open study, all three mood stabilizers demonstrated efficacy in treating a mixed or manic episode in youths with bipolar I or II disorder. Response rates—based on a 50% improvement in Young Mania Rating Scale baseline scores—were divalproex 53%, lithium 38%, and carbamazepine 38%.

In general, divalproex is started at 20 mg/kg/d, which typically produces a serum level of 80 to 120 μg/ml. Common side effects in children include weight gain, nausea, sedation, and tremor.

A possible association between divalproex and polycystic ovary syndrome (PCOS) has been reported in women with epilepsy.²⁶ The mechanism for PCOS has been hypothesized to be obesity secondary to divalproex, resulting in elevated insulin and androgen levels. Recently, Rasgon et al²⁷ reported that epilepsy—and not the anticonvulsants used to treat it—may increase the risk of PCOS. In contrast, O’Donovan et al reported higher rates of menstrual irregularities and PCOS in women with bipolar disorder who were taking divalproex than in those who were not taking divalproex and in healthy controls.²⁸

Until we learn more about this association, clinicians should monitor bipolar female adolescents treated with divalproex for any signs of PCOS, which include menstrual abnormalities, hirsutism, and acne.

Carbamazepine is used widely for seizure management but less commonly than divalproex in pediatric bipolar disorder. This anticonvulsant must be titrated slowly and requires frequent monitoring of blood levels, which can be a problem in children with needle phobia.

Carbamazepine is usually titrated to 15 mg/kg/d to produce a serum level of 7 to 10 μg/ml. Its most common side effects are sedation, rash, nausea, and hyponatremia. Aplastic anemia and severe dermatologic reactions, such as Stevens-Johnson syndrome, occur uncommonly.²⁸

Atypical antipsychotics. Recent case series and open-label reports suggest that atypical antipsychotics such as clozapine,²⁹ risperidone,³⁰ olanzapine,^31-33 and quetiapine¹⁶ are effective in treating pediatric bipolar disorder. However, clinically significant weight gain may be associated with the use of olanzapine and risperidone.³⁴ Ziprasidone may increase QTc prolongation, and safety data are limited in children and adolescents. Therefore, ziprasidone should be used with caution in pediatric bipolar disorder, and ECGs should be monitored.

In the only double-blind, placebo-controlled study of an atypical antipsychotic in pediatric bipolar disorder, manic symptoms were more greatly reduced in 15 adolescents given quetiapine plus divalproex than in 15 patients who received divalproex alone. Quetiapine was titrated to 450 mg/d across 7 days and was well-tolerated. The findings suggest that a mood stabilizer plus an atypical antipsychotic may be more effective than a mood stabilizer alone for treating adolescent mania.¹⁶

Long-term treatment

In addition to treating acute affective episodes, lithium may also help prevent recurrent affective episodes in younger patients. In the only maintenance treatment study for pediatric bipolar disorder, Strober et al prospectively evaluated 37 adolescents whose bipolar disorder had been stabilized with lithium during hospitalization.¹⁷ After 18 months of follow-up, 35% of patients had discontinued lithium, and their relapse rate was 92% (compared with 38% in patients who were lithium-compliant.

It is reasonable to maintain a child or adolescent who has had a single manic episode on mood-stabilizing treatment for several years and then—if the patient is euthymic and asymptomatic—to slowly taper the mood stabilizer(s) over several months. If mood symptoms recur, the agent(s) should be reintroduced.

If a child with bipolar disorder does not respond or only partially responds to a mood stabilizer, it may be necessary to add a second mood stabilizer or an atypical antipsychotic. A bipolar child or adolescent with psychotic symptoms should be maintained on an antipsychotic (typical or atypical) for at least 1 month, even if the psychosis has resolved.³⁵

Treatment of comorbid ADHD. Most children with bipolar disorder have comorbid ADHD, and mood stabilization is necessary prior to starting stimulant medications.³⁶ In bipolar patients, sustained-release psychostimulants may reduce rebound symptoms more effectively than immediate-release formulations. Typical dosages for a child with bipolar disorder and ADHD would be Concerta, 36 mg/d, or Adderall XR, 10 to 20 mg/d.

Psychosocial interventions

Most psychotherapeutic interventions have not been systematically studied in pediatric bipolar disorder but may be beneficial. In a recent study, Fristad et al reported the efficacy of multifamily psychoeducational group therapy for treating bipolar children and adolescents and their families.³⁷

Other useful psychosocial tactics include:

• Minimize periods of overstimulation (for example, these patients do not do well at shopping malls).
• Maintain good sleep hygiene.
• Address medication nonadherence immediately.
• Discuss the risk of substance abuse with the patient.
• Encourage mood charting by the patient and or parent.

Related resources

• Mood charts appropriate for pediatric patients.
• Child and Adolescent Bipolar Foundation. www.cabf.org
• Findling RF, Kowatch RA, Post RM. Pediatric bipolar disorders: a handbook for clinicians. London, Martin Dunitz Press, 2002.
• Geller B, DelBello MP (eds). Child and early adolescent bipolar disorder: theory, assessment, and treatment. New York: Guilford Publications, 2002.

Drug brand names

• Clozapine • Clozaril
• Divalproex sodium • Depakote
• Olanzapine • Zyprexa
• Risperidone • Risperdal
• Quetiapine • Seroquel
• Ziprasidone • Geodon

Disclosure

Dr. Kowatch receives grant support from, serves as a consultant to, or is on the speakers bureau of Novartis Pharmaceuticals Corp., Abbott Laboratories, Solvay Pharmaceuticals, Eli Lilly and Co., Janssen Pharmaceutica, GlaxoSmithKline, and AstraZeneca Pharmaceuticals.

Dr. DelBello receives grant support from, serves as a consultant to, or is on the speakers bureau of Abbott Laboratories, AstraZeneca Pharmaceuticals, Eli Lilly and Co., Ortho-McNeil Pharmaceutical, Shire Pharmaceutical Group, Janssen Pharmaceutica, Pfizer Inc., and GlaxoSmithKline.

As a mood stabilizer for patients with bipolar disorder, lithium was the darling of U.S. psychiatry from the 1970s to well into the 1990s. It then began an ill-deserved, gradual fall from grace and today could be considered a pharmaceutical endangered species. But why?

Did lithium lose effectiveness? Is it too toxic? Is its side effect burden too heavy? Does it interact adversely with too many medicines? Is it too cumbersome to use? Was it just a fad whose time came and went—a psychiatric pet rock? Did it fall prey to the marketing might behind patent-protected drugs? Was it replaced by more effective and safer drugs?

You are partially correct if you checked “all of the above,” because all contain a kernel of truth. At the same time, each is an exaggeration that does grave injustice to a remarkable medication. In addition, psychiatry appears to pay only lip service to convincing evidence that lithium is the only mood stabilizer that reduces the risk of suicide during long-term treatment.¹

Some psychiatrists rationalize that “lithium is too difficult to use, so I never prescribe it.”^2,3 My response is simply, “try it, and I think you’ll like it.” Measuring serum lithium concentrations is simple, accurate, and inexpensive. And we know quite a bit about how lithium dosage and blood level relate to response and tolerability.

Where does lithium stand?

Lithium is the first solid element in the periodic table (atomic number 3, atomic weight 6.94) (Box 1). As a treatment for bipolar disorder, lithium’s rise to prominence in the United States was far from rapid. Its antimanic properties were described by John Cade in Australia in 1949 in an open-label case series, but it was not FDA-approved for 20 years—for acute manic episodes in 1970 and for maintenance therapy “in those manic depressive patients with a history of mania” in 1974. Today, lithium shares FDA-approved manic episode billing with chlorpromazine (1973), divalproex (1995), and olanzapine (2000), but it remains the only FDA-approved drug for maintenance (although the FDA is considering a bipolar depression maintenance indication for lamotrigine).

Box 1

If you examine lithium’s status relative to other bipolar medications, you’ll find some inconsistencies. For example:

• Clinical practice guidelines from the Department of Veterans Affairs (January 1999) recommended lithium as the first-line agent for acute and prophylactic treatment of manic and mixed states, bipolar depression, and rapid cycling.⁴
• The Expert Consensus Guidelines (April 2000) gave at least equal billing—if not preferred status—to divalproex for those indications.⁵
• The American Psychiatric Association’s (APA) revised guidelines (April 2002) gave the nod to lithium for classic elated mania and bipolar depression but to divalproex for mixed mania and rapid cycling.⁶ Divalproex was rated comparable to lithium for maintenance therapy, despite the lack of convincing data.
• The European perspective (January 2002) is most similar to that of the Department of Veterans Affairs, favoring lithium for acute mania, bipolar depression, and long-term treatment.⁷

There is no clear winner (or loser) in the battle for bipolar marketplace supremacy. The belief that one drug does everything is a fantasy for all but a small minority of patients with bipolar disorder. Polypharmacy is the rule, and rational polypharmacy the goal. To exclude lithium from the arsenal of bipolar drugs would be folly, yet lithium prescribing seems to have become a vanishing art. One of my psychopharmacologist colleagues recently expressed bewilderment at the number of “treatment-resistant” bipolar patients referred to him who had never been treated with lithium.

Diagnosis matters

Lithium is most effective in patients with euphoric mania, full remission between episodes, and normal interepisode functioning. Its potential benefits, however, clearly extend to all other aspects of bipolar disorder, to augmentation for treatment-resistant major depressive disorder, to schizoaffective disorder, and—at times—to aggressive states. As the bipolar spectrum expands, it is hardly surprising that the effectiveness of lithium (or any other drug) lessens as we approach the periphery of the spectrum.

Blood levels and dosing

Recommended lithium serum concentrations are given as ranges, realizing that individual variability makes exact numbers impractical. Package inserts for lithium products list serum concentrations between 1.0 and 1.5 mEq/L for acute mania and 0.6 and 1.2 mEq/L for maintenance therapy. The APA’s revised guidelines are a bit more conservative, recommending 0.5 to 1.2 mEq/L for acute mania and waffling somewhat on maintenance.⁴ Many patients on maintenance therapy do well at levels between 0.6 and 0.8 mEq/L, and some prosper at even lower levels.

To avoid obtaining a misleading blood level:

• Samples should be drawn in the morning as close as possible to 12 hours after the last dose.
• Steady state conditions should exist, usually meaning 4 or 5 days on the same dosage without any missed or extra doses (Box 2).

Start treatment using divided dosages, but—following stabilization—once-daily dosing is possible for many patients. If lithium is taken as a single daily dose, 12-hour blood levels will be somewhat higher than with multiple daily dosing. Single and multiple daily dosing are similarly effective, but once-daily dosing may have a compliance and tolerability edge in some patients.

Considering individual patient variability, a lithium carbonate dosage of 1,200 to 1,800 mg/d is likely to be therapeutic for mania and 900 to 1,200 mg/d for maintenance in otherwise healthy, nongeriatric adults.

Starting and maintaining lithium

Medical history. Assuming that lithium theapy is indicated, obtain a detailed medical history. Focus on findings that increase the risk of lithium toxicity, such as renal impairment, drug interactions, and unstable fluid-electrolyte balance.

Although lithium is not contraindicated in patients with renal disease, using an alternate drug is probably preferable. On the other hand, because lithium does not adversely affect the liver or pancreas, it may be preferred to some other mood stabilizers if these organs are diseased.

A thorough diet and drug history is also important. Because low-sodium diets reduce renal lithium clearance, lower doses may be required to reach a given serum concentration. Some drugs alter lithium excretion and can increase or decrease blood levels (see “Drug combinations,”).

Advise women of childbearing age about lithium’s teratogenic potential (which is considerably less than that of carbamazepine or valproate). The risk of cardiovascular malformation of the fetus has been estimated at 1/1,000 to 1/2,000 births among women who took lithium during the first trimester of pregnancy.⁸

Box 2

Baseline lab tests. Assessing renal function is essential. A serum creatinine level will usually suffice, unless a history of renal disease suggests the need for a more extensive evaluation, such as creatinine clearance, renal ultrasound, or nephrology consultation.

A urinalysis is often part of the package. Because thyroid dysfunction can alter mood and lithium can disrupt thyroid function, baseline TSH and T4 tests are recommended. CBC is optional (lithium can cause leukocytosis). The medical history should determine whether additional blood work is necessary. An ECG is sometimes advised in older patients, especially if the history suggests cardiovascular disease. Finally, don’t forget a pregnancy test in women of childbearing potential (Box 3).

Monitoring. Early in the course of therapy, lithium blood levels are usually obtained at 5- to 7-day intervals until the patient is stabilized. After that, assuming all is well, routine monitoring can occur every 3, 4, or even 6 months, depending on the individual’s reliability and stability. Because ongoing assessment of renal and thyroid function is also important, it makes sense to obtain:

• a serum creatinine measurement linked to each lithium level
• and a serum TSH yearly or at the slightest indication of thyroid dysfunction, such as fatigue, weight gain, cognitive impairment, cold intolerance, or depression.

Stopping lithium. Lithium can be discontinued abruptly without side effects if it is ineffective or not tolerated. Stopping lithium after successful long-term use is another story. There is a high likelihood of illness recurrence and a small but real possibility that lithium will be ineffective when restarted. Also, abrupt or rapid discontinuation (within 1 to 14 days) is believed to increase the likelihood of earlier recurrence, compared with more gradual discontinuation.⁹

Side effects and toxicity

One reason for lithium’s slide in popularity is its perceived side-effect profile. Toxic amounts can be lethal, and therapeutic amounts can be bothersome. Yet concerns are often exaggerated because of lack of familiarity with the drug.^10,11

Intoxication. Lithium does have a narrow therapeutic index, with toxicity related to serum concentration and duration of exposure. Acute overdoses, while not benign, are often better tolerated than gradual, more tissue-saturating exposures. Idiosyncratic factors are also involved, as evidenced by documented toxicity at “therapeutic” levels and tolerability despite very high levels.

Early warnings of impending toxicity include:

• neurologic findings such as dysarthria, new or worsening tremor, and ataxia
• gastrointestinal symptoms such as anorexia, nausea, vomiting, and diarrhea.

Severe toxicity can be fatal or cause permanent neurologic (often cerebellar) damage. Causes of intoxication range from deliberate overdose to renal impairment, low-sodium diets, drug interactions, and dehydration. At particular risk are patients with lithium-induced polyuria whose access to fluid replacement is compromised.

Treatment involves reducing absorption, increasing excretion, and restoring fluid-electrolyte balance. Severe intoxication, especially if renal function is impaired, is best treated with hemodialysis.

Box 3

Neurologic. Mild neurologic complaints such as memory impairment, slow reaction time, lack of spontaneity, and lost creativity have been ascribed to lithium and may lead to noncompliance. Under such circumstances, other diagnostic considerations include breakthrough depression, hypothyroidism, other illness, hypercalcemia, other medications, and absence of hypomania.

Like valproate, lithium can cause a benign postural tremor that is usually tolerable and often transient. Should the tremor be problematic, treatment considerations include dosage reduction, switching to a slow-release preparation, reducing caffeine intake, avoiding other tremor-causing drugs such as theophylline or stimulants, and treating associated anxiety. If an anti-tremor drug is needed, a beta-blocker such as propranolol is used most commonly; other options to consider are primidone and gabapentin. Don’t forget that a worsening tremor may indicate impending toxicity.

Very rarely, lithium has been associated with pseudotumor cerebri (benign intracranial hypertension), peripheral neuropathy, and a myasthenia gravis-like syndrome.

Cardiovascular. Like many drugs, lithium can cause benign ST-T wave changes on ECG.

More serious, but fortunately quite uncommon, is lithium-induced sinus node dysfunction manifesting as a variety of bradyarrhythmias and, at times, syncopal episodes. Since normal aging is associated with a gradual loss of sinus node pacemaker cells, the elderly may be especially sensitive to this problem. Unless a pacemaker is implanted, sinus node dysfunction usually requires lithium discontinuation.

Endocrine. The association between lithium and goiter and hypothyroidism is well-recognized, with elevated risk in women and in patients with pre-existing thyroid disease. Both clinical and symptomatic subclinical hypothyroidism will improve with supplemental thyroid hormone. Less well appreciated are reports of hyperthyroidism occurring during lithium therapy or shortly after its discontinuation. Because subclinical hyperthyroidism may not be benign, careful attention must be paid to maintaining thyroid function well within the normal range.

Reports continue to accrue of lithium-related hypercalcemia and increased parathyroid hormone levels, with an occasional patient developing parathyroid hyperplasia or adenoma requiring surgical intervention.¹² No specific guidelines have been established for monitoring serum calcium, but some authors have recommended periodic testing.

Weight. At least one-third of patients on lithium gain weight for a variety of reasons, such as altered lipid and carbohydrate metabolism, use of high-calorie fluids to combat polydipsia and polyuria, hypothyroidism, and the use of other drugs associated with weight gain. If weight gain occurs, recognize it early (weigh your patients) and institute appropriate dietary and exercise measures.

Hematologic. A mild, benign leukocytosis is seen sometimes during treatment with lithium. This effect has been harnessed to treat some neutropenic conditions. Lithium does not increase the risk of blood dyscrasias.

Dermatologic. Acne, psoriasis, and follicular keratosis may first appear or worsen during lithium therapy. Occasionally, otherwise successful lithium therapy has been rendered impossible by a dramatic dermatologic flare-up. Hair loss has also been associated with lithium use for unclear reasons, although hypothyroidism is occasionally a factor.

Renal. Impaired urinary concentrating ability and polyuria are common adverse effects. Both may be reversed with timely treatment discontinuation, but they may persist even after discontinuation in patients on long-term lithium treatment.¹²

Polyuria is largely nephrogenic in origin and, at times, can be voluminous, cause great inconvenience, and pose a risk of dehydration and lithium intoxication. Patients sometimes believe that thirst drives the polyuria and attempt to deal with it by restricting fluid intake, which can be quite dangerous. More appropriate interventions include dosage reduction (if possible) and the use of a thiazide and/or potassium-sparing diuretic. If diuretics are used, serum lithium concentrations may rise. Debate remains as to whether slow-or controlled-release preparations or single daily dosing are “kinder to the kidney.”

Box 4

In recent years, there has been a disturbing increase in reports of elevated serum creatinine and reduced creatinine clearance associated with long-term lithium use.¹³ Because renal impairment has many causes, evaluation by a nephrologist is strongly advised. Then if the finger of causation points strongly at lithium, a careful risk/benefit analysis is in order. Even if lithium is discontinued—and especially if it is continued—regular renal function assessment is essential.

Rarely, lithium can cause a nephrotic syndrome (proteinuria, edema, decreased serum albumin, and increased serum lipids) that tends to be reversible with drug discontinuation.

Drug combinations

First the good news. Lithium tends to combine well with all the anticonvulsant mood stabilizers, making it the favored drug for combination therapies. Lithium/antidepressant combinations can be useful for treatment-resistant depression, although serotonin syndrome occasionally has been reported when lithium is combined with selective serotonin reuptake inhibitors.^10,11 Using lithium with atypical antipsychotics is common, often effective, and usually well-tolerated.

Drug-drug interactions. Some nonpsychiatric drugs are associated with reduced renal lithium clearance and potential lithium toxicity. Because nonpsychiatrists usually prescribe these drugs, encourage patients taking lithium to ask their doctors about the possibility of interactions whenever a new drug is prescribed. Pharmacists can be particularly helpful in avoiding drug-drug interactions.

In patients taking diuretics, serum lithium concentrations are definitely increased by thiazides, possibly by potassium-sparing types, and occasionally by loop types. Osmotic and xanthine diuretics do just the opposite. Because diuretics are often used in medically unstable patients, assume that all can disrupt lithium balance.

Most nonsteroidal anti-inflammatory drugs can increase serum lithium levels, although dose and treatment duration are important variables. Aspirin and acetaminophen should not cause problems. The effect of COX-2 inhibitors on lithium levels has not been studied adequately, so these drugs should remain under suspicion.¹⁴

Lithium toxicity has been reported with angiotensin-converting enzyme (ACE) inhibitors, and their package inserts caution about this possibility. More recently, a few cases of lithium toxicity have been reported in patients taking angiotensin II receptor type-1 (AT-1) antagonists (e.g., candesartan, losartan, valsartan).¹⁵

Other, less well-substantiated pharmacokinetic and pharmacodynamic interactions that have been reported with lithium and other drugs can be researched, by using a computer-based drug interaction program or consulting with a drug information center.

Patient and clinician education

Both patients and clinicians have an obligation to ensure that lithium (or any other drug) is used safely and effectively (Box 4). Excellent sources of continuing education are listed below in “Related resources.” Rather than fall prey to the illusion that lithium therapy is a “vanishing art,” it would be better for clinicians to heed these words from the APA’s 2002 practice guidelines for bipolar disorder:

“No other treatment has performed as well as lithium in as many aspects of long-term care of bipolar disorder patients, and despite some risks and limitations lithium remains the standard against which all proposed alternatives are compared.”⁶

Related resources

• Depression and Bipolar Support Alliance. www.dbsalliance.org
• Lithium Information Center. www.miminc.org

Drug brand names

• Chlorpromazine • Thorazine
• Divalproex • Depakote
• Gabapentin • Neurontin
• Lamotrigine • Lamictal
• Olanzapine • Zyprexa
• Primidone • Mysoline
• Propranolol • Inderal

Disclosure

Dr. Jefferson receives grant/research support from Abbott Laboratories, Bristol-Myers Squibb Co., Forest Laboratories Inc., GlaxoSmithKline, Eli Lilly and Co., Novartis Pharmaceuticals Corp., Organon, Janssen Pharmaceutica, Pfizer Inc., Solvay, and Wyeth Pharmaceuticals. He also serves as a consultant to GlaxoSmithKline, Novartis Pharmaceuticals Corp., Solvay, and UCB Pharma.

As a mood stabilizer for patients with bipolar disorder, lithium was the darling of U.S. psychiatry from the 1970s to well into the 1990s. It then began an ill-deserved, gradual fall from grace and today could be considered a pharmaceutical endangered species. But why?

Did lithium lose effectiveness? Is it too toxic? Is its side effect burden too heavy? Does it interact adversely with too many medicines? Is it too cumbersome to use? Was it just a fad whose time came and went—a psychiatric pet rock? Did it fall prey to the marketing might behind patent-protected drugs? Was it replaced by more effective and safer drugs?

You are partially correct if you checked “all of the above,” because all contain a kernel of truth. At the same time, each is an exaggeration that does grave injustice to a remarkable medication. In addition, psychiatry appears to pay only lip service to convincing evidence that lithium is the only mood stabilizer that reduces the risk of suicide during long-term treatment.¹

Some psychiatrists rationalize that “lithium is too difficult to use, so I never prescribe it.”^2,3 My response is simply, “try it, and I think you’ll like it.” Measuring serum lithium concentrations is simple, accurate, and inexpensive. And we know quite a bit about how lithium dosage and blood level relate to response and tolerability.

Where does lithium stand?

Lithium is the first solid element in the periodic table (atomic number 3, atomic weight 6.94) (Box 1). As a treatment for bipolar disorder, lithium’s rise to prominence in the United States was far from rapid. Its antimanic properties were described by John Cade in Australia in 1949 in an open-label case series, but it was not FDA-approved for 20 years—for acute manic episodes in 1970 and for maintenance therapy “in those manic depressive patients with a history of mania” in 1974. Today, lithium shares FDA-approved manic episode billing with chlorpromazine (1973), divalproex (1995), and olanzapine (2000), but it remains the only FDA-approved drug for maintenance (although the FDA is considering a bipolar depression maintenance indication for lamotrigine).

Box 1

If you examine lithium’s status relative to other bipolar medications, you’ll find some inconsistencies. For example:

• Clinical practice guidelines from the Department of Veterans Affairs (January 1999) recommended lithium as the first-line agent for acute and prophylactic treatment of manic and mixed states, bipolar depression, and rapid cycling.⁴
• The Expert Consensus Guidelines (April 2000) gave at least equal billing—if not preferred status—to divalproex for those indications.⁵
• The American Psychiatric Association’s (APA) revised guidelines (April 2002) gave the nod to lithium for classic elated mania and bipolar depression but to divalproex for mixed mania and rapid cycling.⁶ Divalproex was rated comparable to lithium for maintenance therapy, despite the lack of convincing data.
• The European perspective (January 2002) is most similar to that of the Department of Veterans Affairs, favoring lithium for acute mania, bipolar depression, and long-term treatment.⁷

There is no clear winner (or loser) in the battle for bipolar marketplace supremacy. The belief that one drug does everything is a fantasy for all but a small minority of patients with bipolar disorder. Polypharmacy is the rule, and rational polypharmacy the goal. To exclude lithium from the arsenal of bipolar drugs would be folly, yet lithium prescribing seems to have become a vanishing art. One of my psychopharmacologist colleagues recently expressed bewilderment at the number of “treatment-resistant” bipolar patients referred to him who had never been treated with lithium.

Diagnosis matters

Lithium is most effective in patients with euphoric mania, full remission between episodes, and normal interepisode functioning. Its potential benefits, however, clearly extend to all other aspects of bipolar disorder, to augmentation for treatment-resistant major depressive disorder, to schizoaffective disorder, and—at times—to aggressive states. As the bipolar spectrum expands, it is hardly surprising that the effectiveness of lithium (or any other drug) lessens as we approach the periphery of the spectrum.

Blood levels and dosing

Recommended lithium serum concentrations are given as ranges, realizing that individual variability makes exact numbers impractical. Package inserts for lithium products list serum concentrations between 1.0 and 1.5 mEq/L for acute mania and 0.6 and 1.2 mEq/L for maintenance therapy. The APA’s revised guidelines are a bit more conservative, recommending 0.5 to 1.2 mEq/L for acute mania and waffling somewhat on maintenance.⁴ Many patients on maintenance therapy do well at levels between 0.6 and 0.8 mEq/L, and some prosper at even lower levels.

To avoid obtaining a misleading blood level:

• Samples should be drawn in the morning as close as possible to 12 hours after the last dose.
• Steady state conditions should exist, usually meaning 4 or 5 days on the same dosage without any missed or extra doses (Box 2).

Start treatment using divided dosages, but—following stabilization—once-daily dosing is possible for many patients. If lithium is taken as a single daily dose, 12-hour blood levels will be somewhat higher than with multiple daily dosing. Single and multiple daily dosing are similarly effective, but once-daily dosing may have a compliance and tolerability edge in some patients.

Considering individual patient variability, a lithium carbonate dosage of 1,200 to 1,800 mg/d is likely to be therapeutic for mania and 900 to 1,200 mg/d for maintenance in otherwise healthy, nongeriatric adults.

Starting and maintaining lithium

Medical history. Assuming that lithium theapy is indicated, obtain a detailed medical history. Focus on findings that increase the risk of lithium toxicity, such as renal impairment, drug interactions, and unstable fluid-electrolyte balance.

Although lithium is not contraindicated in patients with renal disease, using an alternate drug is probably preferable. On the other hand, because lithium does not adversely affect the liver or pancreas, it may be preferred to some other mood stabilizers if these organs are diseased.

A thorough diet and drug history is also important. Because low-sodium diets reduce renal lithium clearance, lower doses may be required to reach a given serum concentration. Some drugs alter lithium excretion and can increase or decrease blood levels (see “Drug combinations,”).

Advise women of childbearing age about lithium’s teratogenic potential (which is considerably less than that of carbamazepine or valproate). The risk of cardiovascular malformation of the fetus has been estimated at 1/1,000 to 1/2,000 births among women who took lithium during the first trimester of pregnancy.⁸

Box 2

Baseline lab tests. Assessing renal function is essential. A serum creatinine level will usually suffice, unless a history of renal disease suggests the need for a more extensive evaluation, such as creatinine clearance, renal ultrasound, or nephrology consultation.

A urinalysis is often part of the package. Because thyroid dysfunction can alter mood and lithium can disrupt thyroid function, baseline TSH and T4 tests are recommended. CBC is optional (lithium can cause leukocytosis). The medical history should determine whether additional blood work is necessary. An ECG is sometimes advised in older patients, especially if the history suggests cardiovascular disease. Finally, don’t forget a pregnancy test in women of childbearing potential (Box 3).

Monitoring. Early in the course of therapy, lithium blood levels are usually obtained at 5- to 7-day intervals until the patient is stabilized. After that, assuming all is well, routine monitoring can occur every 3, 4, or even 6 months, depending on the individual’s reliability and stability. Because ongoing assessment of renal and thyroid function is also important, it makes sense to obtain:

• a serum creatinine measurement linked to each lithium level
• and a serum TSH yearly or at the slightest indication of thyroid dysfunction, such as fatigue, weight gain, cognitive impairment, cold intolerance, or depression.

Stopping lithium. Lithium can be discontinued abruptly without side effects if it is ineffective or not tolerated. Stopping lithium after successful long-term use is another story. There is a high likelihood of illness recurrence and a small but real possibility that lithium will be ineffective when restarted. Also, abrupt or rapid discontinuation (within 1 to 14 days) is believed to increase the likelihood of earlier recurrence, compared with more gradual discontinuation.⁹

Side effects and toxicity

One reason for lithium’s slide in popularity is its perceived side-effect profile. Toxic amounts can be lethal, and therapeutic amounts can be bothersome. Yet concerns are often exaggerated because of lack of familiarity with the drug.^10,11

Intoxication. Lithium does have a narrow therapeutic index, with toxicity related to serum concentration and duration of exposure. Acute overdoses, while not benign, are often better tolerated than gradual, more tissue-saturating exposures. Idiosyncratic factors are also involved, as evidenced by documented toxicity at “therapeutic” levels and tolerability despite very high levels.

Early warnings of impending toxicity include:

• neurologic findings such as dysarthria, new or worsening tremor, and ataxia
• gastrointestinal symptoms such as anorexia, nausea, vomiting, and diarrhea.

Severe toxicity can be fatal or cause permanent neurologic (often cerebellar) damage. Causes of intoxication range from deliberate overdose to renal impairment, low-sodium diets, drug interactions, and dehydration. At particular risk are patients with lithium-induced polyuria whose access to fluid replacement is compromised.

Treatment involves reducing absorption, increasing excretion, and restoring fluid-electrolyte balance. Severe intoxication, especially if renal function is impaired, is best treated with hemodialysis.

Box 3

Neurologic. Mild neurologic complaints such as memory impairment, slow reaction time, lack of spontaneity, and lost creativity have been ascribed to lithium and may lead to noncompliance. Under such circumstances, other diagnostic considerations include breakthrough depression, hypothyroidism, other illness, hypercalcemia, other medications, and absence of hypomania.

Like valproate, lithium can cause a benign postural tremor that is usually tolerable and often transient. Should the tremor be problematic, treatment considerations include dosage reduction, switching to a slow-release preparation, reducing caffeine intake, avoiding other tremor-causing drugs such as theophylline or stimulants, and treating associated anxiety. If an anti-tremor drug is needed, a beta-blocker such as propranolol is used most commonly; other options to consider are primidone and gabapentin. Don’t forget that a worsening tremor may indicate impending toxicity.

Very rarely, lithium has been associated with pseudotumor cerebri (benign intracranial hypertension), peripheral neuropathy, and a myasthenia gravis-like syndrome.

Cardiovascular. Like many drugs, lithium can cause benign ST-T wave changes on ECG.

More serious, but fortunately quite uncommon, is lithium-induced sinus node dysfunction manifesting as a variety of bradyarrhythmias and, at times, syncopal episodes. Since normal aging is associated with a gradual loss of sinus node pacemaker cells, the elderly may be especially sensitive to this problem. Unless a pacemaker is implanted, sinus node dysfunction usually requires lithium discontinuation.

Endocrine. The association between lithium and goiter and hypothyroidism is well-recognized, with elevated risk in women and in patients with pre-existing thyroid disease. Both clinical and symptomatic subclinical hypothyroidism will improve with supplemental thyroid hormone. Less well appreciated are reports of hyperthyroidism occurring during lithium therapy or shortly after its discontinuation. Because subclinical hyperthyroidism may not be benign, careful attention must be paid to maintaining thyroid function well within the normal range.

Reports continue to accrue of lithium-related hypercalcemia and increased parathyroid hormone levels, with an occasional patient developing parathyroid hyperplasia or adenoma requiring surgical intervention.¹² No specific guidelines have been established for monitoring serum calcium, but some authors have recommended periodic testing.

Weight. At least one-third of patients on lithium gain weight for a variety of reasons, such as altered lipid and carbohydrate metabolism, use of high-calorie fluids to combat polydipsia and polyuria, hypothyroidism, and the use of other drugs associated with weight gain. If weight gain occurs, recognize it early (weigh your patients) and institute appropriate dietary and exercise measures.

Hematologic. A mild, benign leukocytosis is seen sometimes during treatment with lithium. This effect has been harnessed to treat some neutropenic conditions. Lithium does not increase the risk of blood dyscrasias.

Dermatologic. Acne, psoriasis, and follicular keratosis may first appear or worsen during lithium therapy. Occasionally, otherwise successful lithium therapy has been rendered impossible by a dramatic dermatologic flare-up. Hair loss has also been associated with lithium use for unclear reasons, although hypothyroidism is occasionally a factor.

Renal. Impaired urinary concentrating ability and polyuria are common adverse effects. Both may be reversed with timely treatment discontinuation, but they may persist even after discontinuation in patients on long-term lithium treatment.¹²

Polyuria is largely nephrogenic in origin and, at times, can be voluminous, cause great inconvenience, and pose a risk of dehydration and lithium intoxication. Patients sometimes believe that thirst drives the polyuria and attempt to deal with it by restricting fluid intake, which can be quite dangerous. More appropriate interventions include dosage reduction (if possible) and the use of a thiazide and/or potassium-sparing diuretic. If diuretics are used, serum lithium concentrations may rise. Debate remains as to whether slow-or controlled-release preparations or single daily dosing are “kinder to the kidney.”

Box 4

In recent years, there has been a disturbing increase in reports of elevated serum creatinine and reduced creatinine clearance associated with long-term lithium use.¹³ Because renal impairment has many causes, evaluation by a nephrologist is strongly advised. Then if the finger of causation points strongly at lithium, a careful risk/benefit analysis is in order. Even if lithium is discontinued—and especially if it is continued—regular renal function assessment is essential.

Rarely, lithium can cause a nephrotic syndrome (proteinuria, edema, decreased serum albumin, and increased serum lipids) that tends to be reversible with drug discontinuation.

Drug combinations

First the good news. Lithium tends to combine well with all the anticonvulsant mood stabilizers, making it the favored drug for combination therapies. Lithium/antidepressant combinations can be useful for treatment-resistant depression, although serotonin syndrome occasionally has been reported when lithium is combined with selective serotonin reuptake inhibitors.^10,11 Using lithium with atypical antipsychotics is common, often effective, and usually well-tolerated.

Drug-drug interactions. Some nonpsychiatric drugs are associated with reduced renal lithium clearance and potential lithium toxicity. Because nonpsychiatrists usually prescribe these drugs, encourage patients taking lithium to ask their doctors about the possibility of interactions whenever a new drug is prescribed. Pharmacists can be particularly helpful in avoiding drug-drug interactions.

In patients taking diuretics, serum lithium concentrations are definitely increased by thiazides, possibly by potassium-sparing types, and occasionally by loop types. Osmotic and xanthine diuretics do just the opposite. Because diuretics are often used in medically unstable patients, assume that all can disrupt lithium balance.

Most nonsteroidal anti-inflammatory drugs can increase serum lithium levels, although dose and treatment duration are important variables. Aspirin and acetaminophen should not cause problems. The effect of COX-2 inhibitors on lithium levels has not been studied adequately, so these drugs should remain under suspicion.¹⁴

Lithium toxicity has been reported with angiotensin-converting enzyme (ACE) inhibitors, and their package inserts caution about this possibility. More recently, a few cases of lithium toxicity have been reported in patients taking angiotensin II receptor type-1 (AT-1) antagonists (e.g., candesartan, losartan, valsartan).¹⁵

Other, less well-substantiated pharmacokinetic and pharmacodynamic interactions that have been reported with lithium and other drugs can be researched, by using a computer-based drug interaction program or consulting with a drug information center.

Patient and clinician education

Both patients and clinicians have an obligation to ensure that lithium (or any other drug) is used safely and effectively (Box 4). Excellent sources of continuing education are listed below in “Related resources.” Rather than fall prey to the illusion that lithium therapy is a “vanishing art,” it would be better for clinicians to heed these words from the APA’s 2002 practice guidelines for bipolar disorder:

“No other treatment has performed as well as lithium in as many aspects of long-term care of bipolar disorder patients, and despite some risks and limitations lithium remains the standard against which all proposed alternatives are compared.”⁶

Related resources

• Depression and Bipolar Support Alliance. www.dbsalliance.org
• Lithium Information Center. www.miminc.org

Drug brand names

• Chlorpromazine • Thorazine
• Divalproex • Depakote
• Gabapentin • Neurontin
• Lamotrigine • Lamictal
• Olanzapine • Zyprexa
• Primidone • Mysoline
• Propranolol • Inderal

Disclosure

Dr. Jefferson receives grant/research support from Abbott Laboratories, Bristol-Myers Squibb Co., Forest Laboratories Inc., GlaxoSmithKline, Eli Lilly and Co., Novartis Pharmaceuticals Corp., Organon, Janssen Pharmaceutica, Pfizer Inc., Solvay, and Wyeth Pharmaceuticals. He also serves as a consultant to GlaxoSmithKline, Novartis Pharmaceuticals Corp., Solvay, and UCB Pharma.

As a mood stabilizer for patients with bipolar disorder, lithium was the darling of U.S. psychiatry from the 1970s to well into the 1990s. It then began an ill-deserved, gradual fall from grace and today could be considered a pharmaceutical endangered species. But why?

Did lithium lose effectiveness? Is it too toxic? Is its side effect burden too heavy? Does it interact adversely with too many medicines? Is it too cumbersome to use? Was it just a fad whose time came and went—a psychiatric pet rock? Did it fall prey to the marketing might behind patent-protected drugs? Was it replaced by more effective and safer drugs?

You are partially correct if you checked “all of the above,” because all contain a kernel of truth. At the same time, each is an exaggeration that does grave injustice to a remarkable medication. In addition, psychiatry appears to pay only lip service to convincing evidence that lithium is the only mood stabilizer that reduces the risk of suicide during long-term treatment.¹

Some psychiatrists rationalize that “lithium is too difficult to use, so I never prescribe it.”^2,3 My response is simply, “try it, and I think you’ll like it.” Measuring serum lithium concentrations is simple, accurate, and inexpensive. And we know quite a bit about how lithium dosage and blood level relate to response and tolerability.

Where does lithium stand?

Lithium is the first solid element in the periodic table (atomic number 3, atomic weight 6.94) (Box 1). As a treatment for bipolar disorder, lithium’s rise to prominence in the United States was far from rapid. Its antimanic properties were described by John Cade in Australia in 1949 in an open-label case series, but it was not FDA-approved for 20 years—for acute manic episodes in 1970 and for maintenance therapy “in those manic depressive patients with a history of mania” in 1974. Today, lithium shares FDA-approved manic episode billing with chlorpromazine (1973), divalproex (1995), and olanzapine (2000), but it remains the only FDA-approved drug for maintenance (although the FDA is considering a bipolar depression maintenance indication for lamotrigine).

Box 1

If you examine lithium’s status relative to other bipolar medications, you’ll find some inconsistencies. For example:

• Clinical practice guidelines from the Department of Veterans Affairs (January 1999) recommended lithium as the first-line agent for acute and prophylactic treatment of manic and mixed states, bipolar depression, and rapid cycling.⁴
• The Expert Consensus Guidelines (April 2000) gave at least equal billing—if not preferred status—to divalproex for those indications.⁵
• The American Psychiatric Association’s (APA) revised guidelines (April 2002) gave the nod to lithium for classic elated mania and bipolar depression but to divalproex for mixed mania and rapid cycling.⁶ Divalproex was rated comparable to lithium for maintenance therapy, despite the lack of convincing data.
• The European perspective (January 2002) is most similar to that of the Department of Veterans Affairs, favoring lithium for acute mania, bipolar depression, and long-term treatment.⁷

There is no clear winner (or loser) in the battle for bipolar marketplace supremacy. The belief that one drug does everything is a fantasy for all but a small minority of patients with bipolar disorder. Polypharmacy is the rule, and rational polypharmacy the goal. To exclude lithium from the arsenal of bipolar drugs would be folly, yet lithium prescribing seems to have become a vanishing art. One of my psychopharmacologist colleagues recently expressed bewilderment at the number of “treatment-resistant” bipolar patients referred to him who had never been treated with lithium.

Diagnosis matters

Lithium is most effective in patients with euphoric mania, full remission between episodes, and normal interepisode functioning. Its potential benefits, however, clearly extend to all other aspects of bipolar disorder, to augmentation for treatment-resistant major depressive disorder, to schizoaffective disorder, and—at times—to aggressive states. As the bipolar spectrum expands, it is hardly surprising that the effectiveness of lithium (or any other drug) lessens as we approach the periphery of the spectrum.

Blood levels and dosing

Recommended lithium serum concentrations are given as ranges, realizing that individual variability makes exact numbers impractical. Package inserts for lithium products list serum concentrations between 1.0 and 1.5 mEq/L for acute mania and 0.6 and 1.2 mEq/L for maintenance therapy. The APA’s revised guidelines are a bit more conservative, recommending 0.5 to 1.2 mEq/L for acute mania and waffling somewhat on maintenance.⁴ Many patients on maintenance therapy do well at levels between 0.6 and 0.8 mEq/L, and some prosper at even lower levels.

To avoid obtaining a misleading blood level:

• Samples should be drawn in the morning as close as possible to 12 hours after the last dose.
• Steady state conditions should exist, usually meaning 4 or 5 days on the same dosage without any missed or extra doses (Box 2).

Start treatment using divided dosages, but—following stabilization—once-daily dosing is possible for many patients. If lithium is taken as a single daily dose, 12-hour blood levels will be somewhat higher than with multiple daily dosing. Single and multiple daily dosing are similarly effective, but once-daily dosing may have a compliance and tolerability edge in some patients.

Considering individual patient variability, a lithium carbonate dosage of 1,200 to 1,800 mg/d is likely to be therapeutic for mania and 900 to 1,200 mg/d for maintenance in otherwise healthy, nongeriatric adults.

Starting and maintaining lithium

Medical history. Assuming that lithium theapy is indicated, obtain a detailed medical history. Focus on findings that increase the risk of lithium toxicity, such as renal impairment, drug interactions, and unstable fluid-electrolyte balance.

Although lithium is not contraindicated in patients with renal disease, using an alternate drug is probably preferable. On the other hand, because lithium does not adversely affect the liver or pancreas, it may be preferred to some other mood stabilizers if these organs are diseased.

A thorough diet and drug history is also important. Because low-sodium diets reduce renal lithium clearance, lower doses may be required to reach a given serum concentration. Some drugs alter lithium excretion and can increase or decrease blood levels (see “Drug combinations,”).

Advise women of childbearing age about lithium’s teratogenic potential (which is considerably less than that of carbamazepine or valproate). The risk of cardiovascular malformation of the fetus has been estimated at 1/1,000 to 1/2,000 births among women who took lithium during the first trimester of pregnancy.⁸

Box 2

Baseline lab tests. Assessing renal function is essential. A serum creatinine level will usually suffice, unless a history of renal disease suggests the need for a more extensive evaluation, such as creatinine clearance, renal ultrasound, or nephrology consultation.

A urinalysis is often part of the package. Because thyroid dysfunction can alter mood and lithium can disrupt thyroid function, baseline TSH and T4 tests are recommended. CBC is optional (lithium can cause leukocytosis). The medical history should determine whether additional blood work is necessary. An ECG is sometimes advised in older patients, especially if the history suggests cardiovascular disease. Finally, don’t forget a pregnancy test in women of childbearing potential (Box 3).

Monitoring. Early in the course of therapy, lithium blood levels are usually obtained at 5- to 7-day intervals until the patient is stabilized. After that, assuming all is well, routine monitoring can occur every 3, 4, or even 6 months, depending on the individual’s reliability and stability. Because ongoing assessment of renal and thyroid function is also important, it makes sense to obtain:

• a serum creatinine measurement linked to each lithium level
• and a serum TSH yearly or at the slightest indication of thyroid dysfunction, such as fatigue, weight gain, cognitive impairment, cold intolerance, or depression.

Stopping lithium. Lithium can be discontinued abruptly without side effects if it is ineffective or not tolerated. Stopping lithium after successful long-term use is another story. There is a high likelihood of illness recurrence and a small but real possibility that lithium will be ineffective when restarted. Also, abrupt or rapid discontinuation (within 1 to 14 days) is believed to increase the likelihood of earlier recurrence, compared with more gradual discontinuation.⁹

Side effects and toxicity

One reason for lithium’s slide in popularity is its perceived side-effect profile. Toxic amounts can be lethal, and therapeutic amounts can be bothersome. Yet concerns are often exaggerated because of lack of familiarity with the drug.^10,11

Intoxication. Lithium does have a narrow therapeutic index, with toxicity related to serum concentration and duration of exposure. Acute overdoses, while not benign, are often better tolerated than gradual, more tissue-saturating exposures. Idiosyncratic factors are also involved, as evidenced by documented toxicity at “therapeutic” levels and tolerability despite very high levels.

Early warnings of impending toxicity include:

• neurologic findings such as dysarthria, new or worsening tremor, and ataxia
• gastrointestinal symptoms such as anorexia, nausea, vomiting, and diarrhea.

Severe toxicity can be fatal or cause permanent neurologic (often cerebellar) damage. Causes of intoxication range from deliberate overdose to renal impairment, low-sodium diets, drug interactions, and dehydration. At particular risk are patients with lithium-induced polyuria whose access to fluid replacement is compromised.

Treatment involves reducing absorption, increasing excretion, and restoring fluid-electrolyte balance. Severe intoxication, especially if renal function is impaired, is best treated with hemodialysis.

Box 3

Neurologic. Mild neurologic complaints such as memory impairment, slow reaction time, lack of spontaneity, and lost creativity have been ascribed to lithium and may lead to noncompliance. Under such circumstances, other diagnostic considerations include breakthrough depression, hypothyroidism, other illness, hypercalcemia, other medications, and absence of hypomania.

Like valproate, lithium can cause a benign postural tremor that is usually tolerable and often transient. Should the tremor be problematic, treatment considerations include dosage reduction, switching to a slow-release preparation, reducing caffeine intake, avoiding other tremor-causing drugs such as theophylline or stimulants, and treating associated anxiety. If an anti-tremor drug is needed, a beta-blocker such as propranolol is used most commonly; other options to consider are primidone and gabapentin. Don’t forget that a worsening tremor may indicate impending toxicity.

Very rarely, lithium has been associated with pseudotumor cerebri (benign intracranial hypertension), peripheral neuropathy, and a myasthenia gravis-like syndrome.

Cardiovascular. Like many drugs, lithium can cause benign ST-T wave changes on ECG.

More serious, but fortunately quite uncommon, is lithium-induced sinus node dysfunction manifesting as a variety of bradyarrhythmias and, at times, syncopal episodes. Since normal aging is associated with a gradual loss of sinus node pacemaker cells, the elderly may be especially sensitive to this problem. Unless a pacemaker is implanted, sinus node dysfunction usually requires lithium discontinuation.

Endocrine. The association between lithium and goiter and hypothyroidism is well-recognized, with elevated risk in women and in patients with pre-existing thyroid disease. Both clinical and symptomatic subclinical hypothyroidism will improve with supplemental thyroid hormone. Less well appreciated are reports of hyperthyroidism occurring during lithium therapy or shortly after its discontinuation. Because subclinical hyperthyroidism may not be benign, careful attention must be paid to maintaining thyroid function well within the normal range.

Reports continue to accrue of lithium-related hypercalcemia and increased parathyroid hormone levels, with an occasional patient developing parathyroid hyperplasia or adenoma requiring surgical intervention.¹² No specific guidelines have been established for monitoring serum calcium, but some authors have recommended periodic testing.

Weight. At least one-third of patients on lithium gain weight for a variety of reasons, such as altered lipid and carbohydrate metabolism, use of high-calorie fluids to combat polydipsia and polyuria, hypothyroidism, and the use of other drugs associated with weight gain. If weight gain occurs, recognize it early (weigh your patients) and institute appropriate dietary and exercise measures.

Hematologic. A mild, benign leukocytosis is seen sometimes during treatment with lithium. This effect has been harnessed to treat some neutropenic conditions. Lithium does not increase the risk of blood dyscrasias.

Dermatologic. Acne, psoriasis, and follicular keratosis may first appear or worsen during lithium therapy. Occasionally, otherwise successful lithium therapy has been rendered impossible by a dramatic dermatologic flare-up. Hair loss has also been associated with lithium use for unclear reasons, although hypothyroidism is occasionally a factor.

Renal. Impaired urinary concentrating ability and polyuria are common adverse effects. Both may be reversed with timely treatment discontinuation, but they may persist even after discontinuation in patients on long-term lithium treatment.¹²

Polyuria is largely nephrogenic in origin and, at times, can be voluminous, cause great inconvenience, and pose a risk of dehydration and lithium intoxication. Patients sometimes believe that thirst drives the polyuria and attempt to deal with it by restricting fluid intake, which can be quite dangerous. More appropriate interventions include dosage reduction (if possible) and the use of a thiazide and/or potassium-sparing diuretic. If diuretics are used, serum lithium concentrations may rise. Debate remains as to whether slow-or controlled-release preparations or single daily dosing are “kinder to the kidney.”

Box 4

In recent years, there has been a disturbing increase in reports of elevated serum creatinine and reduced creatinine clearance associated with long-term lithium use.¹³ Because renal impairment has many causes, evaluation by a nephrologist is strongly advised. Then if the finger of causation points strongly at lithium, a careful risk/benefit analysis is in order. Even if lithium is discontinued—and especially if it is continued—regular renal function assessment is essential.

Rarely, lithium can cause a nephrotic syndrome (proteinuria, edema, decreased serum albumin, and increased serum lipids) that tends to be reversible with drug discontinuation.

Drug combinations

First the good news. Lithium tends to combine well with all the anticonvulsant mood stabilizers, making it the favored drug for combination therapies. Lithium/antidepressant combinations can be useful for treatment-resistant depression, although serotonin syndrome occasionally has been reported when lithium is combined with selective serotonin reuptake inhibitors.^10,11 Using lithium with atypical antipsychotics is common, often effective, and usually well-tolerated.

Drug-drug interactions. Some nonpsychiatric drugs are associated with reduced renal lithium clearance and potential lithium toxicity. Because nonpsychiatrists usually prescribe these drugs, encourage patients taking lithium to ask their doctors about the possibility of interactions whenever a new drug is prescribed. Pharmacists can be particularly helpful in avoiding drug-drug interactions.

In patients taking diuretics, serum lithium concentrations are definitely increased by thiazides, possibly by potassium-sparing types, and occasionally by loop types. Osmotic and xanthine diuretics do just the opposite. Because diuretics are often used in medically unstable patients, assume that all can disrupt lithium balance.

Most nonsteroidal anti-inflammatory drugs can increase serum lithium levels, although dose and treatment duration are important variables. Aspirin and acetaminophen should not cause problems. The effect of COX-2 inhibitors on lithium levels has not been studied adequately, so these drugs should remain under suspicion.¹⁴

Lithium toxicity has been reported with angiotensin-converting enzyme (ACE) inhibitors, and their package inserts caution about this possibility. More recently, a few cases of lithium toxicity have been reported in patients taking angiotensin II receptor type-1 (AT-1) antagonists (e.g., candesartan, losartan, valsartan).¹⁵

Other, less well-substantiated pharmacokinetic and pharmacodynamic interactions that have been reported with lithium and other drugs can be researched, by using a computer-based drug interaction program or consulting with a drug information center.

Patient and clinician education

Both patients and clinicians have an obligation to ensure that lithium (or any other drug) is used safely and effectively (Box 4). Excellent sources of continuing education are listed below in “Related resources.” Rather than fall prey to the illusion that lithium therapy is a “vanishing art,” it would be better for clinicians to heed these words from the APA’s 2002 practice guidelines for bipolar disorder:

“No other treatment has performed as well as lithium in as many aspects of long-term care of bipolar disorder patients, and despite some risks and limitations lithium remains the standard against which all proposed alternatives are compared.”⁶

Related resources

• Depression and Bipolar Support Alliance. www.dbsalliance.org
• Lithium Information Center. www.miminc.org

Drug brand names

• Chlorpromazine • Thorazine
• Divalproex • Depakote
• Gabapentin • Neurontin
• Lamotrigine • Lamictal
• Olanzapine • Zyprexa
• Primidone • Mysoline
• Propranolol • Inderal

Disclosure

Dr. Jefferson receives grant/research support from Abbott Laboratories, Bristol-Myers Squibb Co., Forest Laboratories Inc., GlaxoSmithKline, Eli Lilly and Co., Novartis Pharmaceuticals Corp., Organon, Janssen Pharmaceutica, Pfizer Inc., Solvay, and Wyeth Pharmaceuticals. He also serves as a consultant to GlaxoSmithKline, Novartis Pharmaceuticals Corp., Solvay, and UCB Pharma.

As a mood stabilizer for patients with bipolar disorder, lithium was the darling of U.S. psychiatry from the 1970s to well into the 1990s. It then began an ill-deserved, gradual fall from grace and today could be considered a pharmaceutical endangered species. But why?

Did lithium lose effectiveness? Is it too toxic? Is its side effect burden too heavy? Does it interact adversely with too many medicines? Is it too cumbersome to use? Was it just a fad whose time came and went—a psychiatric pet rock? Did it fall prey to the marketing might behind patent-protected drugs? Was it replaced by more effective and safer drugs?

You are partially correct if you checked “all of the above,” because all contain a kernel of truth. At the same time, each is an exaggeration that does grave injustice to a remarkable medication. In addition, psychiatry appears to pay only lip service to convincing evidence that lithium is the only mood stabilizer that reduces the risk of suicide during long-term treatment.¹

Some psychiatrists rationalize that “lithium is too difficult to use, so I never prescribe it.”^2,3 My response is simply, “try it, and I think you’ll like it.” Measuring serum lithium concentrations is simple, accurate, and inexpensive. And we know quite a bit about how lithium dosage and blood level relate to response and tolerability.

Where does lithium stand?

Lithium is the first solid element in the periodic table (atomic number 3, atomic weight 6.94) (Box 1). As a treatment for bipolar disorder, lithium’s rise to prominence in the United States was far from rapid. Its antimanic properties were described by John Cade in Australia in 1949 in an open-label case series, but it was not FDA-approved for 20 years—for acute manic episodes in 1970 and for maintenance therapy “in those manic depressive patients with a history of mania” in 1974. Today, lithium shares FDA-approved manic episode billing with chlorpromazine (1973), divalproex (1995), and olanzapine (2000), but it remains the only FDA-approved drug for maintenance (although the FDA is considering a bipolar depression maintenance indication for lamotrigine).

Box 1

If you examine lithium’s status relative to other bipolar medications, you’ll find some inconsistencies. For example:

• Clinical practice guidelines from the Department of Veterans Affairs (January 1999) recommended lithium as the first-line agent for acute and prophylactic treatment of manic and mixed states, bipolar depression, and rapid cycling.⁴
• The Expert Consensus Guidelines (April 2000) gave at least equal billing—if not preferred status—to divalproex for those indications.⁵
• The American Psychiatric Association’s (APA) revised guidelines (April 2002) gave the nod to lithium for classic elated mania and bipolar depression but to divalproex for mixed mania and rapid cycling.⁶ Divalproex was rated comparable to lithium for maintenance therapy, despite the lack of convincing data.
• The European perspective (January 2002) is most similar to that of the Department of Veterans Affairs, favoring lithium for acute mania, bipolar depression, and long-term treatment.⁷

There is no clear winner (or loser) in the battle for bipolar marketplace supremacy. The belief that one drug does everything is a fantasy for all but a small minority of patients with bipolar disorder. Polypharmacy is the rule, and rational polypharmacy the goal. To exclude lithium from the arsenal of bipolar drugs would be folly, yet lithium prescribing seems to have become a vanishing art. One of my psychopharmacologist colleagues recently expressed bewilderment at the number of “treatment-resistant” bipolar patients referred to him who had never been treated with lithium.

Diagnosis matters

Lithium is most effective in patients with euphoric mania, full remission between episodes, and normal interepisode functioning. Its potential benefits, however, clearly extend to all other aspects of bipolar disorder, to augmentation for treatment-resistant major depressive disorder, to schizoaffective disorder, and—at times—to aggressive states. As the bipolar spectrum expands, it is hardly surprising that the effectiveness of lithium (or any other drug) lessens as we approach the periphery of the spectrum.

Blood levels and dosing

Recommended lithium serum concentrations are given as ranges, realizing that individual variability makes exact numbers impractical. Package inserts for lithium products list serum concentrations between 1.0 and 1.5 mEq/L for acute mania and 0.6 and 1.2 mEq/L for maintenance therapy. The APA’s revised guidelines are a bit more conservative, recommending 0.5 to 1.2 mEq/L for acute mania and waffling somewhat on maintenance.⁴ Many patients on maintenance therapy do well at levels between 0.6 and 0.8 mEq/L, and some prosper at even lower levels.

To avoid obtaining a misleading blood level:

• Samples should be drawn in the morning as close as possible to 12 hours after the last dose.
• Steady state conditions should exist, usually meaning 4 or 5 days on the same dosage without any missed or extra doses (Box 2).

Start treatment using divided dosages, but—following stabilization—once-daily dosing is possible for many patients. If lithium is taken as a single daily dose, 12-hour blood levels will be somewhat higher than with multiple daily dosing. Single and multiple daily dosing are similarly effective, but once-daily dosing may have a compliance and tolerability edge in some patients.

Considering individual patient variability, a lithium carbonate dosage of 1,200 to 1,800 mg/d is likely to be therapeutic for mania and 900 to 1,200 mg/d for maintenance in otherwise healthy, nongeriatric adults.

Starting and maintaining lithium

Medical history. Assuming that lithium theapy is indicated, obtain a detailed medical history. Focus on findings that increase the risk of lithium toxicity, such as renal impairment, drug interactions, and unstable fluid-electrolyte balance.

Although lithium is not contraindicated in patients with renal disease, using an alternate drug is probably preferable. On the other hand, because lithium does not adversely affect the liver or pancreas, it may be preferred to some other mood stabilizers if these organs are diseased.

A thorough diet and drug history is also important. Because low-sodium diets reduce renal lithium clearance, lower doses may be required to reach a given serum concentration. Some drugs alter lithium excretion and can increase or decrease blood levels (see “Drug combinations,”).

Advise women of childbearing age about lithium’s teratogenic potential (which is considerably less than that of carbamazepine or valproate). The risk of cardiovascular malformation of the fetus has been estimated at 1/1,000 to 1/2,000 births among women who took lithium during the first trimester of pregnancy.⁸

Box 2

Baseline lab tests. Assessing renal function is essential. A serum creatinine level will usually suffice, unless a history of renal disease suggests the need for a more extensive evaluation, such as creatinine clearance, renal ultrasound, or nephrology consultation.

A urinalysis is often part of the package. Because thyroid dysfunction can alter mood and lithium can disrupt thyroid function, baseline TSH and T4 tests are recommended. CBC is optional (lithium can cause leukocytosis). The medical history should determine whether additional blood work is necessary. An ECG is sometimes advised in older patients, especially if the history suggests cardiovascular disease. Finally, don’t forget a pregnancy test in women of childbearing potential (Box 3).

Monitoring. Early in the course of therapy, lithium blood levels are usually obtained at 5- to 7-day intervals until the patient is stabilized. After that, assuming all is well, routine monitoring can occur every 3, 4, or even 6 months, depending on the individual’s reliability and stability. Because ongoing assessment of renal and thyroid function is also important, it makes sense to obtain:

• a serum creatinine measurement linked to each lithium level
• and a serum TSH yearly or at the slightest indication of thyroid dysfunction, such as fatigue, weight gain, cognitive impairment, cold intolerance, or depression.

Stopping lithium. Lithium can be discontinued abruptly without side effects if it is ineffective or not tolerated. Stopping lithium after successful long-term use is another story. There is a high likelihood of illness recurrence and a small but real possibility that lithium will be ineffective when restarted. Also, abrupt or rapid discontinuation (within 1 to 14 days) is believed to increase the likelihood of earlier recurrence, compared with more gradual discontinuation.⁹

Side effects and toxicity

One reason for lithium’s slide in popularity is its perceived side-effect profile. Toxic amounts can be lethal, and therapeutic amounts can be bothersome. Yet concerns are often exaggerated because of lack of familiarity with the drug.^10,11

Intoxication. Lithium does have a narrow therapeutic index, with toxicity related to serum concentration and duration of exposure. Acute overdoses, while not benign, are often better tolerated than gradual, more tissue-saturating exposures. Idiosyncratic factors are also involved, as evidenced by documented toxicity at “therapeutic” levels and tolerability despite very high levels.

Early warnings of impending toxicity include:

• neurologic findings such as dysarthria, new or worsening tremor, and ataxia
• gastrointestinal symptoms such as anorexia, nausea, vomiting, and diarrhea.

Severe toxicity can be fatal or cause permanent neurologic (often cerebellar) damage. Causes of intoxication range from deliberate overdose to renal impairment, low-sodium diets, drug interactions, and dehydration. At particular risk are patients with lithium-induced polyuria whose access to fluid replacement is compromised.

Treatment involves reducing absorption, increasing excretion, and restoring fluid-electrolyte balance. Severe intoxication, especially if renal function is impaired, is best treated with hemodialysis.

Box 3

Neurologic. Mild neurologic complaints such as memory impairment, slow reaction time, lack of spontaneity, and lost creativity have been ascribed to lithium and may lead to noncompliance. Under such circumstances, other diagnostic considerations include breakthrough depression, hypothyroidism, other illness, hypercalcemia, other medications, and absence of hypomania.

Like valproate, lithium can cause a benign postural tremor that is usually tolerable and often transient. Should the tremor be problematic, treatment considerations include dosage reduction, switching to a slow-release preparation, reducing caffeine intake, avoiding other tremor-causing drugs such as theophylline or stimulants, and treating associated anxiety. If an anti-tremor drug is needed, a beta-blocker such as propranolol is used most commonly; other options to consider are primidone and gabapentin. Don’t forget that a worsening tremor may indicate impending toxicity.

Very rarely, lithium has been associated with pseudotumor cerebri (benign intracranial hypertension), peripheral neuropathy, and a myasthenia gravis-like syndrome.

Cardiovascular. Like many drugs, lithium can cause benign ST-T wave changes on ECG.

More serious, but fortunately quite uncommon, is lithium-induced sinus node dysfunction manifesting as a variety of bradyarrhythmias and, at times, syncopal episodes. Since normal aging is associated with a gradual loss of sinus node pacemaker cells, the elderly may be especially sensitive to this problem. Unless a pacemaker is implanted, sinus node dysfunction usually requires lithium discontinuation.

Endocrine. The association between lithium and goiter and hypothyroidism is well-recognized, with elevated risk in women and in patients with pre-existing thyroid disease. Both clinical and symptomatic subclinical hypothyroidism will improve with supplemental thyroid hormone. Less well appreciated are reports of hyperthyroidism occurring during lithium therapy or shortly after its discontinuation. Because subclinical hyperthyroidism may not be benign, careful attention must be paid to maintaining thyroid function well within the normal range.

Reports continue to accrue of lithium-related hypercalcemia and increased parathyroid hormone levels, with an occasional patient developing parathyroid hyperplasia or adenoma requiring surgical intervention.¹² No specific guidelines have been established for monitoring serum calcium, but some authors have recommended periodic testing.

Weight. At least one-third of patients on lithium gain weight for a variety of reasons, such as altered lipid and carbohydrate metabolism, use of high-calorie fluids to combat polydipsia and polyuria, hypothyroidism, and the use of other drugs associated with weight gain. If weight gain occurs, recognize it early (weigh your patients) and institute appropriate dietary and exercise measures.

Hematologic. A mild, benign leukocytosis is seen sometimes during treatment with lithium. This effect has been harnessed to treat some neutropenic conditions. Lithium does not increase the risk of blood dyscrasias.

Dermatologic. Acne, psoriasis, and follicular keratosis may first appear or worsen during lithium therapy. Occasionally, otherwise successful lithium therapy has been rendered impossible by a dramatic dermatologic flare-up. Hair loss has also been associated with lithium use for unclear reasons, although hypothyroidism is occasionally a factor.

Renal. Impaired urinary concentrating ability and polyuria are common adverse effects. Both may be reversed with timely treatment discontinuation, but they may persist even after discontinuation in patients on long-term lithium treatment.¹²

Polyuria is largely nephrogenic in origin and, at times, can be voluminous, cause great inconvenience, and pose a risk of dehydration and lithium intoxication. Patients sometimes believe that thirst drives the polyuria and attempt to deal with it by restricting fluid intake, which can be quite dangerous. More appropriate interventions include dosage reduction (if possible) and the use of a thiazide and/or potassium-sparing diuretic. If diuretics are used, serum lithium concentrations may rise. Debate remains as to whether slow-or controlled-release preparations or single daily dosing are “kinder to the kidney.”

Box 4

In recent years, there has been a disturbing increase in reports of elevated serum creatinine and reduced creatinine clearance associated with long-term lithium use.¹³ Because renal impairment has many causes, evaluation by a nephrologist is strongly advised. Then if the finger of causation points strongly at lithium, a careful risk/benefit analysis is in order. Even if lithium is discontinued—and especially if it is continued—regular renal function assessment is essential.

Rarely, lithium can cause a nephrotic syndrome (proteinuria, edema, decreased serum albumin, and increased serum lipids) that tends to be reversible with drug discontinuation.

Drug combinations

First the good news. Lithium tends to combine well with all the anticonvulsant mood stabilizers, making it the favored drug for combination therapies. Lithium/antidepressant combinations can be useful for treatment-resistant depression, although serotonin syndrome occasionally has been reported when lithium is combined with selective serotonin reuptake inhibitors.^10,11 Using lithium with atypical antipsychotics is common, often effective, and usually well-tolerated.

Drug-drug interactions. Some nonpsychiatric drugs are associated with reduced renal lithium clearance and potential lithium toxicity. Because nonpsychiatrists usually prescribe these drugs, encourage patients taking lithium to ask their doctors about the possibility of interactions whenever a new drug is prescribed. Pharmacists can be particularly helpful in avoiding drug-drug interactions.

In patients taking diuretics, serum lithium concentrations are definitely increased by thiazides, possibly by potassium-sparing types, and occasionally by loop types. Osmotic and xanthine diuretics do just the opposite. Because diuretics are often used in medically unstable patients, assume that all can disrupt lithium balance.

Most nonsteroidal anti-inflammatory drugs can increase serum lithium levels, although dose and treatment duration are important variables. Aspirin and acetaminophen should not cause problems. The effect of COX-2 inhibitors on lithium levels has not been studied adequately, so these drugs should remain under suspicion.¹⁴

Lithium toxicity has been reported with angiotensin-converting enzyme (ACE) inhibitors, and their package inserts caution about this possibility. More recently, a few cases of lithium toxicity have been reported in patients taking angiotensin II receptor type-1 (AT-1) antagonists (e.g., candesartan, losartan, valsartan).¹⁵

Other, less well-substantiated pharmacokinetic and pharmacodynamic interactions that have been reported with lithium and other drugs can be researched, by using a computer-based drug interaction program or consulting with a drug information center.

Patient and clinician education

Both patients and clinicians have an obligation to ensure that lithium (or any other drug) is used safely and effectively (Box 4). Excellent sources of continuing education are listed below in “Related resources.” Rather than fall prey to the illusion that lithium therapy is a “vanishing art,” it would be better for clinicians to heed these words from the APA’s 2002 practice guidelines for bipolar disorder:

“No other treatment has performed as well as lithium in as many aspects of long-term care of bipolar disorder patients, and despite some risks and limitations lithium remains the standard against which all proposed alternatives are compared.”⁶

Related resources

• Depression and Bipolar Support Alliance. www.dbsalliance.org
• Lithium Information Center. www.miminc.org

Drug brand names

• Chlorpromazine • Thorazine
• Divalproex • Depakote
• Gabapentin • Neurontin
• Lamotrigine • Lamictal
• Olanzapine • Zyprexa
• Primidone • Mysoline
• Propranolol • Inderal

Disclosure

Dr. Jefferson receives grant/research support from Abbott Laboratories, Bristol-Myers Squibb Co., Forest Laboratories Inc., GlaxoSmithKline, Eli Lilly and Co., Novartis Pharmaceuticals Corp., Organon, Janssen Pharmaceutica, Pfizer Inc., Solvay, and Wyeth Pharmaceuticals. He also serves as a consultant to GlaxoSmithKline, Novartis Pharmaceuticals Corp., Solvay, and UCB Pharma.

As a mood stabilizer for patients with bipolar disorder, lithium was the darling of U.S. psychiatry from the 1970s to well into the 1990s. It then began an ill-deserved, gradual fall from grace and today could be considered a pharmaceutical endangered species. But why?

Did lithium lose effectiveness? Is it too toxic? Is its side effect burden too heavy? Does it interact adversely with too many medicines? Is it too cumbersome to use? Was it just a fad whose time came and went—a psychiatric pet rock? Did it fall prey to the marketing might behind patent-protected drugs? Was it replaced by more effective and safer drugs?

You are partially correct if you checked “all of the above,” because all contain a kernel of truth. At the same time, each is an exaggeration that does grave injustice to a remarkable medication. In addition, psychiatry appears to pay only lip service to convincing evidence that lithium is the only mood stabilizer that reduces the risk of suicide during long-term treatment.¹

Some psychiatrists rationalize that “lithium is too difficult to use, so I never prescribe it.”^2,3 My response is simply, “try it, and I think you’ll like it.” Measuring serum lithium concentrations is simple, accurate, and inexpensive. And we know quite a bit about how lithium dosage and blood level relate to response and tolerability.

Where does lithium stand?

Lithium is the first solid element in the periodic table (atomic number 3, atomic weight 6.94) (Box 1). As a treatment for bipolar disorder, lithium’s rise to prominence in the United States was far from rapid. Its antimanic properties were described by John Cade in Australia in 1949 in an open-label case series, but it was not FDA-approved for 20 years—for acute manic episodes in 1970 and for maintenance therapy “in those manic depressive patients with a history of mania” in 1974. Today, lithium shares FDA-approved manic episode billing with chlorpromazine (1973), divalproex (1995), and olanzapine (2000), but it remains the only FDA-approved drug for maintenance (although the FDA is considering a bipolar depression maintenance indication for lamotrigine).

Box 1

If you examine lithium’s status relative to other bipolar medications, you’ll find some inconsistencies. For example:

• Clinical practice guidelines from the Department of Veterans Affairs (January 1999) recommended lithium as the first-line agent for acute and prophylactic treatment of manic and mixed states, bipolar depression, and rapid cycling.⁴
• The Expert Consensus Guidelines (April 2000) gave at least equal billing—if not preferred status—to divalproex for those indications.⁵
• The American Psychiatric Association’s (APA) revised guidelines (April 2002) gave the nod to lithium for classic elated mania and bipolar depression but to divalproex for mixed mania and rapid cycling.⁶ Divalproex was rated comparable to lithium for maintenance therapy, despite the lack of convincing data.
• The European perspective (January 2002) is most similar to that of the Department of Veterans Affairs, favoring lithium for acute mania, bipolar depression, and long-term treatment.⁷

There is no clear winner (or loser) in the battle for bipolar marketplace supremacy. The belief that one drug does everything is a fantasy for all but a small minority of patients with bipolar disorder. Polypharmacy is the rule, and rational polypharmacy the goal. To exclude lithium from the arsenal of bipolar drugs would be folly, yet lithium prescribing seems to have become a vanishing art. One of my psychopharmacologist colleagues recently expressed bewilderment at the number of “treatment-resistant” bipolar patients referred to him who had never been treated with lithium.

Diagnosis matters

Lithium is most effective in patients with euphoric mania, full remission between episodes, and normal interepisode functioning. Its potential benefits, however, clearly extend to all other aspects of bipolar disorder, to augmentation for treatment-resistant major depressive disorder, to schizoaffective disorder, and—at times—to aggressive states. As the bipolar spectrum expands, it is hardly surprising that the effectiveness of lithium (or any other drug) lessens as we approach the periphery of the spectrum.

Blood levels and dosing

Recommended lithium serum concentrations are given as ranges, realizing that individual variability makes exact numbers impractical. Package inserts for lithium products list serum concentrations between 1.0 and 1.5 mEq/L for acute mania and 0.6 and 1.2 mEq/L for maintenance therapy. The APA’s revised guidelines are a bit more conservative, recommending 0.5 to 1.2 mEq/L for acute mania and waffling somewhat on maintenance.⁴ Many patients on maintenance therapy do well at levels between 0.6 and 0.8 mEq/L, and some prosper at even lower levels.

To avoid obtaining a misleading blood level:

• Samples should be drawn in the morning as close as possible to 12 hours after the last dose.
• Steady state conditions should exist, usually meaning 4 or 5 days on the same dosage without any missed or extra doses (Box 2).

Start treatment using divided dosages, but—following stabilization—once-daily dosing is possible for many patients. If lithium is taken as a single daily dose, 12-hour blood levels will be somewhat higher than with multiple daily dosing. Single and multiple daily dosing are similarly effective, but once-daily dosing may have a compliance and tolerability edge in some patients.

Considering individual patient variability, a lithium carbonate dosage of 1,200 to 1,800 mg/d is likely to be therapeutic for mania and 900 to 1,200 mg/d for maintenance in otherwise healthy, nongeriatric adults.

Starting and maintaining lithium

Medical history. Assuming that lithium theapy is indicated, obtain a detailed medical history. Focus on findings that increase the risk of lithium toxicity, such as renal impairment, drug interactions, and unstable fluid-electrolyte balance.

Although lithium is not contraindicated in patients with renal disease, using an alternate drug is probably preferable. On the other hand, because lithium does not adversely affect the liver or pancreas, it may be preferred to some other mood stabilizers if these organs are diseased.

A thorough diet and drug history is also important. Because low-sodium diets reduce renal lithium clearance, lower doses may be required to reach a given serum concentration. Some drugs alter lithium excretion and can increase or decrease blood levels (see “Drug combinations,”).

Advise women of childbearing age about lithium’s teratogenic potential (which is considerably less than that of carbamazepine or valproate). The risk of cardiovascular malformation of the fetus has been estimated at 1/1,000 to 1/2,000 births among women who took lithium during the first trimester of pregnancy.⁸

Box 2

Baseline lab tests. Assessing renal function is essential. A serum creatinine level will usually suffice, unless a history of renal disease suggests the need for a more extensive evaluation, such as creatinine clearance, renal ultrasound, or nephrology consultation.

A urinalysis is often part of the package. Because thyroid dysfunction can alter mood and lithium can disrupt thyroid function, baseline TSH and T4 tests are recommended. CBC is optional (lithium can cause leukocytosis). The medical history should determine whether additional blood work is necessary. An ECG is sometimes advised in older patients, especially if the history suggests cardiovascular disease. Finally, don’t forget a pregnancy test in women of childbearing potential (Box 3).

Monitoring. Early in the course of therapy, lithium blood levels are usually obtained at 5- to 7-day intervals until the patient is stabilized. After that, assuming all is well, routine monitoring can occur every 3, 4, or even 6 months, depending on the individual’s reliability and stability. Because ongoing assessment of renal and thyroid function is also important, it makes sense to obtain:

• a serum creatinine measurement linked to each lithium level
• and a serum TSH yearly or at the slightest indication of thyroid dysfunction, such as fatigue, weight gain, cognitive impairment, cold intolerance, or depression.

Stopping lithium. Lithium can be discontinued abruptly without side effects if it is ineffective or not tolerated. Stopping lithium after successful long-term use is another story. There is a high likelihood of illness recurrence and a small but real possibility that lithium will be ineffective when restarted. Also, abrupt or rapid discontinuation (within 1 to 14 days) is believed to increase the likelihood of earlier recurrence, compared with more gradual discontinuation.⁹

Side effects and toxicity

One reason for lithium’s slide in popularity is its perceived side-effect profile. Toxic amounts can be lethal, and therapeutic amounts can be bothersome. Yet concerns are often exaggerated because of lack of familiarity with the drug.^10,11

Intoxication. Lithium does have a narrow therapeutic index, with toxicity related to serum concentration and duration of exposure. Acute overdoses, while not benign, are often better tolerated than gradual, more tissue-saturating exposures. Idiosyncratic factors are also involved, as evidenced by documented toxicity at “therapeutic” levels and tolerability despite very high levels.

Early warnings of impending toxicity include:

• neurologic findings such as dysarthria, new or worsening tremor, and ataxia
• gastrointestinal symptoms such as anorexia, nausea, vomiting, and diarrhea.

Severe toxicity can be fatal or cause permanent neurologic (often cerebellar) damage. Causes of intoxication range from deliberate overdose to renal impairment, low-sodium diets, drug interactions, and dehydration. At particular risk are patients with lithium-induced polyuria whose access to fluid replacement is compromised.

Treatment involves reducing absorption, increasing excretion, and restoring fluid-electrolyte balance. Severe intoxication, especially if renal function is impaired, is best treated with hemodialysis.

Box 3

Neurologic. Mild neurologic complaints such as memory impairment, slow reaction time, lack of spontaneity, and lost creativity have been ascribed to lithium and may lead to noncompliance. Under such circumstances, other diagnostic considerations include breakthrough depression, hypothyroidism, other illness, hypercalcemia, other medications, and absence of hypomania.

Like valproate, lithium can cause a benign postural tremor that is usually tolerable and often transient. Should the tremor be problematic, treatment considerations include dosage reduction, switching to a slow-release preparation, reducing caffeine intake, avoiding other tremor-causing drugs such as theophylline or stimulants, and treating associated anxiety. If an anti-tremor drug is needed, a beta-blocker such as propranolol is used most commonly; other options to consider are primidone and gabapentin. Don’t forget that a worsening tremor may indicate impending toxicity.

Very rarely, lithium has been associated with pseudotumor cerebri (benign intracranial hypertension), peripheral neuropathy, and a myasthenia gravis-like syndrome.

Cardiovascular. Like many drugs, lithium can cause benign ST-T wave changes on ECG.

More serious, but fortunately quite uncommon, is lithium-induced sinus node dysfunction manifesting as a variety of bradyarrhythmias and, at times, syncopal episodes. Since normal aging is associated with a gradual loss of sinus node pacemaker cells, the elderly may be especially sensitive to this problem. Unless a pacemaker is implanted, sinus node dysfunction usually requires lithium discontinuation.

Endocrine. The association between lithium and goiter and hypothyroidism is well-recognized, with elevated risk in women and in patients with pre-existing thyroid disease. Both clinical and symptomatic subclinical hypothyroidism will improve with supplemental thyroid hormone. Less well appreciated are reports of hyperthyroidism occurring during lithium therapy or shortly after its discontinuation. Because subclinical hyperthyroidism may not be benign, careful attention must be paid to maintaining thyroid function well within the normal range.

Reports continue to accrue of lithium-related hypercalcemia and increased parathyroid hormone levels, with an occasional patient developing parathyroid hyperplasia or adenoma requiring surgical intervention.¹² No specific guidelines have been established for monitoring serum calcium, but some authors have recommended periodic testing.

Weight. At least one-third of patients on lithium gain weight for a variety of reasons, such as altered lipid and carbohydrate metabolism, use of high-calorie fluids to combat polydipsia and polyuria, hypothyroidism, and the use of other drugs associated with weight gain. If weight gain occurs, recognize it early (weigh your patients) and institute appropriate dietary and exercise measures.

Hematologic. A mild, benign leukocytosis is seen sometimes during treatment with lithium. This effect has been harnessed to treat some neutropenic conditions. Lithium does not increase the risk of blood dyscrasias.

Dermatologic. Acne, psoriasis, and follicular keratosis may first appear or worsen during lithium therapy. Occasionally, otherwise successful lithium therapy has been rendered impossible by a dramatic dermatologic flare-up. Hair loss has also been associated with lithium use for unclear reasons, although hypothyroidism is occasionally a factor.

Renal. Impaired urinary concentrating ability and polyuria are common adverse effects. Both may be reversed with timely treatment discontinuation, but they may persist even after discontinuation in patients on long-term lithium treatment.¹²

Polyuria is largely nephrogenic in origin and, at times, can be voluminous, cause great inconvenience, and pose a risk of dehydration and lithium intoxication. Patients sometimes believe that thirst drives the polyuria and attempt to deal with it by restricting fluid intake, which can be quite dangerous. More appropriate interventions include dosage reduction (if possible) and the use of a thiazide and/or potassium-sparing diuretic. If diuretics are used, serum lithium concentrations may rise. Debate remains as to whether slow-or controlled-release preparations or single daily dosing are “kinder to the kidney.”

Box 4

In recent years, there has been a disturbing increase in reports of elevated serum creatinine and reduced creatinine clearance associated with long-term lithium use.¹³ Because renal impairment has many causes, evaluation by a nephrologist is strongly advised. Then if the finger of causation points strongly at lithium, a careful risk/benefit analysis is in order. Even if lithium is discontinued—and especially if it is continued—regular renal function assessment is essential.

Rarely, lithium can cause a nephrotic syndrome (proteinuria, edema, decreased serum albumin, and increased serum lipids) that tends to be reversible with drug discontinuation.

Drug combinations

First the good news. Lithium tends to combine well with all the anticonvulsant mood stabilizers, making it the favored drug for combination therapies. Lithium/antidepressant combinations can be useful for treatment-resistant depression, although serotonin syndrome occasionally has been reported when lithium is combined with selective serotonin reuptake inhibitors.^10,11 Using lithium with atypical antipsychotics is common, often effective, and usually well-tolerated.

Drug-drug interactions. Some nonpsychiatric drugs are associated with reduced renal lithium clearance and potential lithium toxicity. Because nonpsychiatrists usually prescribe these drugs, encourage patients taking lithium to ask their doctors about the possibility of interactions whenever a new drug is prescribed. Pharmacists can be particularly helpful in avoiding drug-drug interactions.

In patients taking diuretics, serum lithium concentrations are definitely increased by thiazides, possibly by potassium-sparing types, and occasionally by loop types. Osmotic and xanthine diuretics do just the opposite. Because diuretics are often used in medically unstable patients, assume that all can disrupt lithium balance.

Most nonsteroidal anti-inflammatory drugs can increase serum lithium levels, although dose and treatment duration are important variables. Aspirin and acetaminophen should not cause problems. The effect of COX-2 inhibitors on lithium levels has not been studied adequately, so these drugs should remain under suspicion.¹⁴

Lithium toxicity has been reported with angiotensin-converting enzyme (ACE) inhibitors, and their package inserts caution about this possibility. More recently, a few cases of lithium toxicity have been reported in patients taking angiotensin II receptor type-1 (AT-1) antagonists (e.g., candesartan, losartan, valsartan).¹⁵

Other, less well-substantiated pharmacokinetic and pharmacodynamic interactions that have been reported with lithium and other drugs can be researched, by using a computer-based drug interaction program or consulting with a drug information center.

Patient and clinician education

Both patients and clinicians have an obligation to ensure that lithium (or any other drug) is used safely and effectively (Box 4). Excellent sources of continuing education are listed below in “Related resources.” Rather than fall prey to the illusion that lithium therapy is a “vanishing art,” it would be better for clinicians to heed these words from the APA’s 2002 practice guidelines for bipolar disorder:

“No other treatment has performed as well as lithium in as many aspects of long-term care of bipolar disorder patients, and despite some risks and limitations lithium remains the standard against which all proposed alternatives are compared.”⁶

Related resources

• Depression and Bipolar Support Alliance. www.dbsalliance.org
• Lithium Information Center. www.miminc.org

Drug brand names

• Chlorpromazine • Thorazine
• Divalproex • Depakote
• Gabapentin • Neurontin
• Lamotrigine • Lamictal
• Olanzapine • Zyprexa
• Primidone • Mysoline
• Propranolol • Inderal

Disclosure

Dr. Jefferson receives grant/research support from Abbott Laboratories, Bristol-Myers Squibb Co., Forest Laboratories Inc., GlaxoSmithKline, Eli Lilly and Co., Novartis Pharmaceuticals Corp., Organon, Janssen Pharmaceutica, Pfizer Inc., Solvay, and Wyeth Pharmaceuticals. He also serves as a consultant to GlaxoSmithKline, Novartis Pharmaceuticals Corp., Solvay, and UCB Pharma.

As a mood stabilizer for patients with bipolar disorder, lithium was the darling of U.S. psychiatry from the 1970s to well into the 1990s. It then began an ill-deserved, gradual fall from grace and today could be considered a pharmaceutical endangered species. But why?

Did lithium lose effectiveness? Is it too toxic? Is its side effect burden too heavy? Does it interact adversely with too many medicines? Is it too cumbersome to use? Was it just a fad whose time came and went—a psychiatric pet rock? Did it fall prey to the marketing might behind patent-protected drugs? Was it replaced by more effective and safer drugs?

You are partially correct if you checked “all of the above,” because all contain a kernel of truth. At the same time, each is an exaggeration that does grave injustice to a remarkable medication. In addition, psychiatry appears to pay only lip service to convincing evidence that lithium is the only mood stabilizer that reduces the risk of suicide during long-term treatment.¹

Some psychiatrists rationalize that “lithium is too difficult to use, so I never prescribe it.”^2,3 My response is simply, “try it, and I think you’ll like it.” Measuring serum lithium concentrations is simple, accurate, and inexpensive. And we know quite a bit about how lithium dosage and blood level relate to response and tolerability.

Where does lithium stand?

Lithium is the first solid element in the periodic table (atomic number 3, atomic weight 6.94) (Box 1). As a treatment for bipolar disorder, lithium’s rise to prominence in the United States was far from rapid. Its antimanic properties were described by John Cade in Australia in 1949 in an open-label case series, but it was not FDA-approved for 20 years—for acute manic episodes in 1970 and for maintenance therapy “in those manic depressive patients with a history of mania” in 1974. Today, lithium shares FDA-approved manic episode billing with chlorpromazine (1973), divalproex (1995), and olanzapine (2000), but it remains the only FDA-approved drug for maintenance (although the FDA is considering a bipolar depression maintenance indication for lamotrigine).

Box 1

If you examine lithium’s status relative to other bipolar medications, you’ll find some inconsistencies. For example:

• Clinical practice guidelines from the Department of Veterans Affairs (January 1999) recommended lithium as the first-line agent for acute and prophylactic treatment of manic and mixed states, bipolar depression, and rapid cycling.⁴
• The Expert Consensus Guidelines (April 2000) gave at least equal billing—if not preferred status—to divalproex for those indications.⁵
• The American Psychiatric Association’s (APA) revised guidelines (April 2002) gave the nod to lithium for classic elated mania and bipolar depression but to divalproex for mixed mania and rapid cycling.⁶ Divalproex was rated comparable to lithium for maintenance therapy, despite the lack of convincing data.
• The European perspective (January 2002) is most similar to that of the Department of Veterans Affairs, favoring lithium for acute mania, bipolar depression, and long-term treatment.⁷

There is no clear winner (or loser) in the battle for bipolar marketplace supremacy. The belief that one drug does everything is a fantasy for all but a small minority of patients with bipolar disorder. Polypharmacy is the rule, and rational polypharmacy the goal. To exclude lithium from the arsenal of bipolar drugs would be folly, yet lithium prescribing seems to have become a vanishing art. One of my psychopharmacologist colleagues recently expressed bewilderment at the number of “treatment-resistant” bipolar patients referred to him who had never been treated with lithium.

Diagnosis matters

Lithium is most effective in patients with euphoric mania, full remission between episodes, and normal interepisode functioning. Its potential benefits, however, clearly extend to all other aspects of bipolar disorder, to augmentation for treatment-resistant major depressive disorder, to schizoaffective disorder, and—at times—to aggressive states. As the bipolar spectrum expands, it is hardly surprising that the effectiveness of lithium (or any other drug) lessens as we approach the periphery of the spectrum.

Blood levels and dosing

Recommended lithium serum concentrations are given as ranges, realizing that individual variability makes exact numbers impractical. Package inserts for lithium products list serum concentrations between 1.0 and 1.5 mEq/L for acute mania and 0.6 and 1.2 mEq/L for maintenance therapy. The APA’s revised guidelines are a bit more conservative, recommending 0.5 to 1.2 mEq/L for acute mania and waffling somewhat on maintenance.⁴ Many patients on maintenance therapy do well at levels between 0.6 and 0.8 mEq/L, and some prosper at even lower levels.

To avoid obtaining a misleading blood level:

• Samples should be drawn in the morning as close as possible to 12 hours after the last dose.
• Steady state conditions should exist, usually meaning 4 or 5 days on the same dosage without any missed or extra doses (Box 2).

Start treatment using divided dosages, but—following stabilization—once-daily dosing is possible for many patients. If lithium is taken as a single daily dose, 12-hour blood levels will be somewhat higher than with multiple daily dosing. Single and multiple daily dosing are similarly effective, but once-daily dosing may have a compliance and tolerability edge in some patients.

Considering individual patient variability, a lithium carbonate dosage of 1,200 to 1,800 mg/d is likely to be therapeutic for mania and 900 to 1,200 mg/d for maintenance in otherwise healthy, nongeriatric adults.

Starting and maintaining lithium

Medical history. Assuming that lithium theapy is indicated, obtain a detailed medical history. Focus on findings that increase the risk of lithium toxicity, such as renal impairment, drug interactions, and unstable fluid-electrolyte balance.

Although lithium is not contraindicated in patients with renal disease, using an alternate drug is probably preferable. On the other hand, because lithium does not adversely affect the liver or pancreas, it may be preferred to some other mood stabilizers if these organs are diseased.

A thorough diet and drug history is also important. Because low-sodium diets reduce renal lithium clearance, lower doses may be required to reach a given serum concentration. Some drugs alter lithium excretion and can increase or decrease blood levels (see “Drug combinations,”).

Advise women of childbearing age about lithium’s teratogenic potential (which is considerably less than that of carbamazepine or valproate). The risk of cardiovascular malformation of the fetus has been estimated at 1/1,000 to 1/2,000 births among women who took lithium during the first trimester of pregnancy.⁸

Box 2

Baseline lab tests. Assessing renal function is essential. A serum creatinine level will usually suffice, unless a history of renal disease suggests the need for a more extensive evaluation, such as creatinine clearance, renal ultrasound, or nephrology consultation.

A urinalysis is often part of the package. Because thyroid dysfunction can alter mood and lithium can disrupt thyroid function, baseline TSH and T4 tests are recommended. CBC is optional (lithium can cause leukocytosis). The medical history should determine whether additional blood work is necessary. An ECG is sometimes advised in older patients, especially if the history suggests cardiovascular disease. Finally, don’t forget a pregnancy test in women of childbearing potential (Box 3).

Monitoring. Early in the course of therapy, lithium blood levels are usually obtained at 5- to 7-day intervals until the patient is stabilized. After that, assuming all is well, routine monitoring can occur every 3, 4, or even 6 months, depending on the individual’s reliability and stability. Because ongoing assessment of renal and thyroid function is also important, it makes sense to obtain:

• a serum creatinine measurement linked to each lithium level
• and a serum TSH yearly or at the slightest indication of thyroid dysfunction, such as fatigue, weight gain, cognitive impairment, cold intolerance, or depression.

Stopping lithium. Lithium can be discontinued abruptly without side effects if it is ineffective or not tolerated. Stopping lithium after successful long-term use is another story. There is a high likelihood of illness recurrence and a small but real possibility that lithium will be ineffective when restarted. Also, abrupt or rapid discontinuation (within 1 to 14 days) is believed to increase the likelihood of earlier recurrence, compared with more gradual discontinuation.⁹

Side effects and toxicity

One reason for lithium’s slide in popularity is its perceived side-effect profile. Toxic amounts can be lethal, and therapeutic amounts can be bothersome. Yet concerns are often exaggerated because of lack of familiarity with the drug.^10,11

Intoxication. Lithium does have a narrow therapeutic index, with toxicity related to serum concentration and duration of exposure. Acute overdoses, while not benign, are often better tolerated than gradual, more tissue-saturating exposures. Idiosyncratic factors are also involved, as evidenced by documented toxicity at “therapeutic” levels and tolerability despite very high levels.

Early warnings of impending toxicity include:

• neurologic findings such as dysarthria, new or worsening tremor, and ataxia
• gastrointestinal symptoms such as anorexia, nausea, vomiting, and diarrhea.

Severe toxicity can be fatal or cause permanent neurologic (often cerebellar) damage. Causes of intoxication range from deliberate overdose to renal impairment, low-sodium diets, drug interactions, and dehydration. At particular risk are patients with lithium-induced polyuria whose access to fluid replacement is compromised.

Treatment involves reducing absorption, increasing excretion, and restoring fluid-electrolyte balance. Severe intoxication, especially if renal function is impaired, is best treated with hemodialysis.

Box 3

Neurologic. Mild neurologic complaints such as memory impairment, slow reaction time, lack of spontaneity, and lost creativity have been ascribed to lithium and may lead to noncompliance. Under such circumstances, other diagnostic considerations include breakthrough depression, hypothyroidism, other illness, hypercalcemia, other medications, and absence of hypomania.

Like valproate, lithium can cause a benign postural tremor that is usually tolerable and often transient. Should the tremor be problematic, treatment considerations include dosage reduction, switching to a slow-release preparation, reducing caffeine intake, avoiding other tremor-causing drugs such as theophylline or stimulants, and treating associated anxiety. If an anti-tremor drug is needed, a beta-blocker such as propranolol is used most commonly; other options to consider are primidone and gabapentin. Don’t forget that a worsening tremor may indicate impending toxicity.

Very rarely, lithium has been associated with pseudotumor cerebri (benign intracranial hypertension), peripheral neuropathy, and a myasthenia gravis-like syndrome.

Cardiovascular. Like many drugs, lithium can cause benign ST-T wave changes on ECG.

More serious, but fortunately quite uncommon, is lithium-induced sinus node dysfunction manifesting as a variety of bradyarrhythmias and, at times, syncopal episodes. Since normal aging is associated with a gradual loss of sinus node pacemaker cells, the elderly may be especially sensitive to this problem. Unless a pacemaker is implanted, sinus node dysfunction usually requires lithium discontinuation.

Endocrine. The association between lithium and goiter and hypothyroidism is well-recognized, with elevated risk in women and in patients with pre-existing thyroid disease. Both clinical and symptomatic subclinical hypothyroidism will improve with supplemental thyroid hormone. Less well appreciated are reports of hyperthyroidism occurring during lithium therapy or shortly after its discontinuation. Because subclinical hyperthyroidism may not be benign, careful attention must be paid to maintaining thyroid function well within the normal range.

Reports continue to accrue of lithium-related hypercalcemia and increased parathyroid hormone levels, with an occasional patient developing parathyroid hyperplasia or adenoma requiring surgical intervention.¹² No specific guidelines have been established for monitoring serum calcium, but some authors have recommended periodic testing.

Weight. At least one-third of patients on lithium gain weight for a variety of reasons, such as altered lipid and carbohydrate metabolism, use of high-calorie fluids to combat polydipsia and polyuria, hypothyroidism, and the use of other drugs associated with weight gain. If weight gain occurs, recognize it early (weigh your patients) and institute appropriate dietary and exercise measures.

Hematologic. A mild, benign leukocytosis is seen sometimes during treatment with lithium. This effect has been harnessed to treat some neutropenic conditions. Lithium does not increase the risk of blood dyscrasias.

Dermatologic. Acne, psoriasis, and follicular keratosis may first appear or worsen during lithium therapy. Occasionally, otherwise successful lithium therapy has been rendered impossible by a dramatic dermatologic flare-up. Hair loss has also been associated with lithium use for unclear reasons, although hypothyroidism is occasionally a factor.

Renal. Impaired urinary concentrating ability and polyuria are common adverse effects. Both may be reversed with timely treatment discontinuation, but they may persist even after discontinuation in patients on long-term lithium treatment.¹²

Polyuria is largely nephrogenic in origin and, at times, can be voluminous, cause great inconvenience, and pose a risk of dehydration and lithium intoxication. Patients sometimes believe that thirst drives the polyuria and attempt to deal with it by restricting fluid intake, which can be quite dangerous. More appropriate interventions include dosage reduction (if possible) and the use of a thiazide and/or potassium-sparing diuretic. If diuretics are used, serum lithium concentrations may rise. Debate remains as to whether slow-or controlled-release preparations or single daily dosing are “kinder to the kidney.”

Box 4

In recent years, there has been a disturbing increase in reports of elevated serum creatinine and reduced creatinine clearance associated with long-term lithium use.¹³ Because renal impairment has many causes, evaluation by a nephrologist is strongly advised. Then if the finger of causation points strongly at lithium, a careful risk/benefit analysis is in order. Even if lithium is discontinued—and especially if it is continued—regular renal function assessment is essential.

Rarely, lithium can cause a nephrotic syndrome (proteinuria, edema, decreased serum albumin, and increased serum lipids) that tends to be reversible with drug discontinuation.

Drug combinations

First the good news. Lithium tends to combine well with all the anticonvulsant mood stabilizers, making it the favored drug for combination therapies. Lithium/antidepressant combinations can be useful for treatment-resistant depression, although serotonin syndrome occasionally has been reported when lithium is combined with selective serotonin reuptake inhibitors.^10,11 Using lithium with atypical antipsychotics is common, often effective, and usually well-tolerated.

Drug-drug interactions. Some nonpsychiatric drugs are associated with reduced renal lithium clearance and potential lithium toxicity. Because nonpsychiatrists usually prescribe these drugs, encourage patients taking lithium to ask their doctors about the possibility of interactions whenever a new drug is prescribed. Pharmacists can be particularly helpful in avoiding drug-drug interactions.

In patients taking diuretics, serum lithium concentrations are definitely increased by thiazides, possibly by potassium-sparing types, and occasionally by loop types. Osmotic and xanthine diuretics do just the opposite. Because diuretics are often used in medically unstable patients, assume that all can disrupt lithium balance.

Most nonsteroidal anti-inflammatory drugs can increase serum lithium levels, although dose and treatment duration are important variables. Aspirin and acetaminophen should not cause problems. The effect of COX-2 inhibitors on lithium levels has not been studied adequately, so these drugs should remain under suspicion.¹⁴

Lithium toxicity has been reported with angiotensin-converting enzyme (ACE) inhibitors, and their package inserts caution about this possibility. More recently, a few cases of lithium toxicity have been reported in patients taking angiotensin II receptor type-1 (AT-1) antagonists (e.g., candesartan, losartan, valsartan).¹⁵

Other, less well-substantiated pharmacokinetic and pharmacodynamic interactions that have been reported with lithium and other drugs can be researched, by using a computer-based drug interaction program or consulting with a drug information center.

Patient and clinician education

Both patients and clinicians have an obligation to ensure that lithium (or any other drug) is used safely and effectively (Box 4). Excellent sources of continuing education are listed below in “Related resources.” Rather than fall prey to the illusion that lithium therapy is a “vanishing art,” it would be better for clinicians to heed these words from the APA’s 2002 practice guidelines for bipolar disorder:

“No other treatment has performed as well as lithium in as many aspects of long-term care of bipolar disorder patients, and despite some risks and limitations lithium remains the standard against which all proposed alternatives are compared.”⁶

Related resources

• Depression and Bipolar Support Alliance. www.dbsalliance.org
• Lithium Information Center. www.miminc.org

Drug brand names

• Chlorpromazine • Thorazine
• Divalproex • Depakote
• Gabapentin • Neurontin
• Lamotrigine • Lamictal
• Olanzapine • Zyprexa
• Primidone • Mysoline
• Propranolol • Inderal

Disclosure

Dr. Jefferson receives grant/research support from Abbott Laboratories, Bristol-Myers Squibb Co., Forest Laboratories Inc., GlaxoSmithKline, Eli Lilly and Co., Novartis Pharmaceuticals Corp., Organon, Janssen Pharmaceutica, Pfizer Inc., Solvay, and Wyeth Pharmaceuticals. He also serves as a consultant to GlaxoSmithKline, Novartis Pharmaceuticals Corp., Solvay, and UCB Pharma.