To Name :
To Email :
From Name :
From Email :
Comments :

Evidence-Based Reviews

Are anticonvulsants safe for pediatric bipolar disorder?

Using antiepileptic agents as mood stabilizers requires caution.

Vol. 4, No. 8 / August 2005

Are anticonvulsants safe and effective mood stabilizers for children and adolescents with bipolar disorder? The answer is unclear because most bipolar disorder treatment trials have included adults only, and clinicians are desperate for data.1

To help you care for young patients, we report what is known about the potential benefits and risks of using mood stabilizers and anticonvulsants in bipolar youth. We base our dosing, target serum level, and monitoring recommendations on clinical experience and the limited published evidence.


Bipolar disorder’s “atypical” presentation in children—often more irritability and explosiveness than euphoria—can complicate diagnosis. Bipolar children and adolescents often have comorbid attention-deficit/hyperactivity disorder (ADHD), other disruptive behavior disorders, or anxiety disorders. Thus, comorbidities and presenting symptoms often dictate medication choice.

An expert consensus guideline acknowledges that more evidence on pediatric bipolar disorder is needed. In the meantime, the guideline suggests trying valproate or lithium first to treat nonpsychotic mania in pediatric bipolar patients.1 It also recommends three atypical antipsychotics— olanzapine, quetiapine, and risperidone—as potential first-line treatments. Valproate and lithium may be preferred because of atypicals’ risk of weight gain and metabolic syndrome.

Trying other anticonvulsants may be justified for bipolar youths who are not functioning well with first-line agents. Lamotrigine, for example, has antidepressant and antimanic effects.2 When you try anticonvulsants that lack double-blind, placebo-controlled trials, we recommend that you:

  • obtain consent from the parents and child
  • monitor carefully for side effects.


Lithium is one of the most well-studied medications for pediatric bipolar disorder and the only mood stabilizer FDA-approved for children and adolescents (Table 1).3 Although approved for ages 12 and older, lithium has been used in younger children in practice and in clinical trials.

Table 1

FDA-approval status of medications used to treat bipolar disorder


Indications for adults

Indications for children


Acute manic episode and acute mixed episode

Not approved


Maintenance therapy

Not approved


Acute manic episode and maintenance therapy

Age ≥ 12 years


Not approved

Not approved


Not approved

Not approved


Acute manic episode

Not approved

Source: Reference 3

Efficacy. In an open-label study of 100 adolescents with type I bipolar disorder,4 63% met response criteria after 4 weeks of lithium and 26% showed manic symptom remission. Symptoms worsened in both groups, however, when 40 responders were randomly assigned to continue or discontinue lithium for 2 weeks.5 The authors speculated that these conflicting results might indicate that mood stabilization requires longer treatment. Contrary to earlier reports,6 manic adolescents with comorbid ADHD did not show poor response to lithium.

In the only double-blind, placebo-controlled trial of lithium in adolescents with bipolar disorder, some subjects had secondary substance dependency disorders.7 For 6 weeks, 25 outpatient adolescents received lithium (13 patients) or placebo (12 patients). Lithium was effective in treating bipolar and substance dependency symptoms, with significantly improved clinical global assessment scores and decreased positive urine assays for drugs. Little difference was seen in mood item scores on the Schedule for Affective Disorders and Schizophrenia, child version (KSADS-1986), whether patients were taking lithium or placebo.

Pediatric dosing. For bipolar patients ages 6 to 12, use the child’s weight to determine lithium dosage (Table 2).8 Maintain serum levels between 0.8 and 1.2 mEq/L,9 and check them frequently when starting therapy.10 After mood stabilization, check levels every 1 to 3 months or when you suspect noncompliance. Obtain renal and thyroid function values at baseline and every 4 to 6 months.

Table 2

Guide to dosing lithium for prepubertal school-aged children*


Doses (mg)

Child’s weight (kg)

8 AM

12 PM

6 PM

Total daily






25 to 40





40 to 50





50 to 60





* Maintain specified dose at least 5 days, drawing serum levels 12 hrs after the last lithium dose until two consecutive levels appear in the therapeutic range (0.6 to 1.2 mEq/L). Dose may then be adjusted based on serum level, side effects, or clinical response. Do not exceed 1.4 mEq/L.

Source: Reference 8

Safety. Common side effects reported in adolescents include weight gain (55%), polydipsia (33%), polyuria (25%), headache (23%), tremor (20%), and GI complaints (up to 18%).4 Neurologic side effects are associated with higher serum lithium levels (0.91 to 1.36 mEq/L)10 and occur more often in younger than in older children.11 The cardiac defect Ebstein’s anomaly occurs in approximately 0.05 to 0.1% of children exposed to lithium in utero (Box).12-16


Birth-defect risks to consider when prescribing mood stabilizers

Consider teratogenicity when choosing mood stabilizers for bipolar adolescent girls who may be sexually active. Lithium, valproate, and carbamazepine are labeled pregnancy category D because of their potential to cause birth defects.

Lithium treatment has been associated with increased risk of cardiac defects, specifically Ebstein’s anomaly (malformation of the tricuspid valve). Its incidence in children of women who used lithium during pregnancy is estimated to be 1:1,000 (0.10%) to 2:1,000 (0.05%)— 20 to 40 times the rate in the general population.12

Valproate.Results from the North American Antiepileptic Drug (AED) Pregnancy Registry showed a 10.7% rate of major congenital malformations (MCM)— including neural tube defects (spina bifida) and cardiac defects (pulmonary atresia)—in children of women who used valproate during pregnancy. The rate of births with MCMs in the general population is 2.9%.13

Carbamazepine.Data from the Australian Pregnancy Registry showed no significant increase in malformation rates in infants of carbamazepine users compared with those of women receiving no antiepileptics.14 Other studies, however, have linked carbamazepine with an increased risk of craniofacial defects (11%), neural tube defects (0.5 to 1%), and cardiac malformations.12

Lamotrigine.The teratogenic effects of the newer anticonvulsants are unclear. An 11-year study of lamotrigine15 found MCM risk after first-trimester exposure to lamotrigine to be similar to the general population’s MCM risk.

Combination therapy.Teratogenic risk appears to increase when multiple antiepileptic drugs are used (9.9% risk in polytherapy vs 6.2% in monotherapy).16


Efficacy. No double-blind, placebo-controlled study has shown valproate to be effective in treating bipolar disorder in children and adolescents. When used as monotherapy in open-label studies, valproate has produced response rates of:

  • 53% in a 6-week, randomized, open-label trial in which 42 outpatients (mean age 11.4 years) with bipolar disorder type I or II received lithium, divalproex sodium, or carbamazepine9
  • 61% in an open-label study of 40 patients ages 7 to 19 with a manic, hypomanic, or mixed episode who received divalproex for 2 to 8 weeks17
  • 80% in an 8-week open-label trial of 40 patients ages 6 to 17 with bipolar disorder type I (77.5%) or type II (22.5%) and a Young Mania Rating Scale (YMRS)score ≥ 14.18

In a prospective trial, 90 patients ages 5 to 17 with bipolar disorder type I or II were treated with lithium plus divalproex sodium. After up to 20 weeks, 47% met criteria for depressive and manic symptom remission.19 A chart review has showed valproate’s efficacy in treating aggression and irritability in adolescent mania.20

Safety: Black-box warnings. Valproate therapy carries risks of hepatic failure, pancreatitis, and birth defects. Monitor blood counts and hepatic enzymes throughout therapy (Table 3).3 Rare yet potentially fatal hepatic toxicity appears to occur most often in children age <2 who are treated with anticonvulsant combinations.21 Other studies suggest:

  • an association with congenital malformations, including spina bifida and pulmonary atresia, in children exposed to valproate in utero6
  • a link between valproate and hyperammonemic encephalopathy, especially in patients with urea cycle disorders22
  • potential for benign thrombocytopenia23
  • increased incidence of polycystic ovary syndrome—ovarian cysts, hyperandrogenism, chronic anovulation—in peripubertal mentally retarded women treated with valproate for seizure disorders.24

Because of these risks, use caution when prescribing valproate to bipolar adolescent girls. Monitor menstrual cycle regularity, and collaborate with a gynecologist to watch for potentially dangerous effects.

Table 3

Mood stabilizers’ side effects and recommended monitoring


Major side effects



Allergic skin rash, drowsiness, blood dyscrasias, diplopia

CBC with reticulocytes, iron, LFTs, urinalysis, BUN, TFTs, sodium, serum carbamazepine levels


Stevens-Johnson syndrome, headache, dizziness, ataxia, somnolence, nausea, diplopia, blurred vision, rhinitis

No serum monitoring recommended


Polyuria, polydipsia, nausea, diarrhea, tininecleatremor, enuresis, fatigue, ataxia, leukocytosis, malaise, cardiac arrhythmias, weight gain

BUN/creatinine, crearance, TFTs, calcium/phosphorus, ECG, serum lithium levels every 1 to 3 months once stabilized


Dizziness, somnolence/fatigue, ataxia/gait disturbance, vertigo, headache, tremor, rash, hyponatremia, hypersensitivity reaction, GI symptoms, diplopia

Sodium levels (particularly ifirst 3 months)


Hyperchloremic metabolic acidosis, oligohydrosis and hyperthermia, acute myopia, somnolence/fatigue, nausea, anorexia/weight loss, paresthesia, tremor, difficulty concentrating

BUN/creatinine, sodium bicarbonate


Irritability/restlessness, ataxia, headache, weight gain, hyperammonemic encephalopathy, alopecia, GI upset, pancreatitis,sedation, thrombocytopenia, liver failure, polycystic ovaries/hyperandrogenism, teratogenic effects,rash

Ammonia, LFTs, bilirubin, CBC with platelets, serum valproate levels

BUN: blood urea nitrogen; CBC: complete blood count; ECG: electrocardiography; LFT: liver function tests; TFTs: thyroid function tests

Note: Bolded items included in black-box warnings

Source: Reference 3

Body weight. Valproate has been associated with weight gain. In a study of 372 bipolar adults, 21% reported a 5% weight-gain during 52 weeks of maintenance therapy, compared with 13% of patients on lithium and 7% on placebo.25 Shortterm studies of adjunctive valproate in pediatric bipolar patients raise similar concerns.26 Thus, monitor for weight gain and serum lipid changes in youths starting valproate therapy.


Carbamazepine is used less often than lithium or divalproex for bipolar disorder. It tends to be used adjunctively when lithium alone is ineffective.

Efficacy. In an open-label study,9 42 patients ages 8 to 18 with bipolar disorder type I or II were randomly assigned to lithium, divalproex sodium, or carbamazepine monotherapy for 6 weeks. Response rates—measured as a ≥ 50% change from baseline in YMRS scores—were 53% with divalproex, 38% with lithium, and 38% with carbamazepine.

A retrospective review of 44 hospitalized bipolar patients ages 5 to 12 treated for at least 7 days with lithium, valproate, or carbamazepine reported higher (ie, worse) Clinical Global Impression of Improvement scores with carbamazepine.27 Small sample sizes, particularly in the carbamazepine group, limited this naturalistic study.

Safety: Black-box warnings. Carbamazepine’s hematologic “black box” warns of increased risk of aplastic anemia, agranulocytosis, leukopenia, and thrombocytopenia. Risks associated with carbamazepine have been estimated at:

  • aplastic anemia: 5.1/million patient years
  • agranulocytosis: 1.4/million patient years.28

Leukopenia is relatively more common and occurs in approximately 20% of children receiving carbamazepine.29 Consider stopping carbamazepine when the white cell count falls below 3,000/mm3 (or the neutrophil count drops to <1,000/mm3).29 Advise children and parents to watch for leukopenia’s signs and symptoms, including fever, infections, sore throat, and mouth ulcers.3

Body weight. Carbamazepine is not associated with significant weight gain, which could be clinically important for some patients.

Drug interactions. Carbamazepine activates the cytochrome P-450 liver enzyme system, increasing the metabolism of many medications and decreasing their blood levels. Consider monitoring serum levels when using carbamazepine with valproate, imipramine, corticosteroids, warfarin, oral contraceptives, and some antibiotics. Because carbamazepine induces its own metabolism, you might need to increase its dosage if its effects appear to be waning.3

Carbamazepine and tricyclic antidepressants may show cross-sensitivity because of structural similarity. Do not use monoamine oxidase inhibitors with carbamazepine; discontinue them at least 14 days before starting carbamazepine.3


Oxcarbazepine has similar efficacy to carbamazepine but less side effect risk and does not require plasma level monitoring. A weaker inducer of CYP-450, it causes fewer clinically important drug-drug interactions and may be useful for patients who respond to carbamazepine but cannot tolerate its side effects.30

Efficacy. Case studies31,32 have been encouraging, but no published, double-blind, placebo-controlled studies support using oxcarbazepine in bipolar children and adolescents.

Safety. Oxcarbazepine appears to be generally well-tolerated but can cause potentially serious reactions—including hyponatremia.33 Somnolence, emesis, and ataxia are the most common side effects in pediatric patients.3

Hyponatremia —plasma sodium 125 mEq/L—occurs in 2.5% of adults taking oxcarbazepine3 and has been reported in a similar percentage of children.34 This potentially severe reaction—characterized by nausea, lethargy, malaise, headache, confusion, decreased seizure threshold, or simply decreased serum sodium35—is usually noted within the first 12 weeks of therapy. The risk increases with concomitant use of other sodium-altering drugs, such as antidepressants or antipsychotics.36

Did you miss this content?
An under-recognized epidemic of elder abuse needs your awareness and action