Child & Adolescent Clinical Psychopharmacology
4th Edition

CHAPTER 8
Mood Stabilizers: Lithium Carbonate and Antiepileptics
Lithium Carbonate (Lithotabs, Eskalith, Lithane, Lithobid) and Lithium Citrate (Cibalith-S)
Currently, lithium carbonate is approved by the FDA only for the treatment of manic episodes of bipolar disorders and for maintenance therapy of bipolar patients with a history of mania; the drug is approved only for persons 12 years of age and older. Over the past three decades, however, lithium carbonate has been investigated in the treatment of many child and adolescent disorders, but especially in the treatment of children with severe aggression directed toward self or others, children with bipolar or similar disorders, and behaviorally disturbed children whose parents are known lithium responders. One major impetus for this research was that standard antipsychotic agents, which were frequently used to control severe behavioral disorders and sometimes mania, not only could cause cognitive dulling when used in sufficient dosage to control symptoms, but also carried significant risk of causing tardive dyskinesia when used on a long-term basis (Platt et al. 1984).
Pharmacokinetics of Lithium Carbonate
The lithium ion is readily absorbed from the gastrointestinal tract and is most commonly administered in the form of lithium carbonate (Li2CO3), a highly soluble salt. Peak plasma concentrations occur within 2 to 4 hours, and complete absorption takes place within approximately 8 hours (Baldessarini, 1990). Approximately 95% of a single dose of lithium is excreted by the kidneys, with up to two-thirds of an acute dose being excreted within 6 to 12 hours. The serum half-life is approximately 20 to 24 hours. Depletion of the sodium ion causes a clinically significant degree of lithium retention by the kidneys. Steady-state serum lithium levels typically occur within 5 to 8 days of repeated identical daily doses of lithium carbonate. Although lithium pharmacokinetics differ considerably among individuals, they are fairly stable over time for a given person (Baldessarini and Stephens, 1970).
Vitiello et al. (1988) studied the pharmacokinetics of lithium carbonate in nine children aged 9 to 12 years. The children had a trend toward a shorter elimination half-life of lithium and a significantly higher total renal clearance of lithium. The clinical significance of this rate is that a steady state of lithium serum levels is reached more rapidly in children than in adults, and therapeutic levels can be achieved more quickly.
P.260

Contraindications for Lithium Carbonate Administration
Administration of lithium carbonate is relatively contraindicated in individuals with significant renal or cardiovascular disease, severe debilitation, severe dehydration, or sodium depletion, because these conditions are associated with a very high risk of lithium toxicity. Patients with such disorders should be thoroughly assessed, usually in consultation with the person providing medical care, before beginning lithium therapy.
Except under urgent circumstances, adolescents who are likely to become pregnant should not be administered lithium; this is particularly true of those in early pregnancy. Lithium carbonate is associated with a significant increase in cardiac teratogenicity, especially with Ebstein’s anomaly. A significantly increased incidence of other cardiac anomalies has also been reported. Kallen and Tandberg (1983) reported that 7% of the infants of women who used lithium in early pregnancy had serious heart defects other than Ebstein’s anomaly.
Significant thyroid disease is a relative contraindication to lithium carbonate therapy; however, with careful monitoring of thyroid function and the use of supplemental thyroid preparations when necessary, it may be used when other drugs are not effective and the potential benefits outweigh the risks.
Interactions of Lithium Carbonate with Other Drugs
There are several reports that increased neuroleptic toxicity with an encephalopathic syndrome or neuroleptic malignant syndrome may occur when lithium and neuroleptics are used concomitantly, but this has usually been seen with high doses. The simultaneous use of lithium and neuroleptic agents, however, may be indicated in some cases of mania or schizoaffective psychoses, and many patients have received both a neuroleptic and lithium with no untoward effects.
Elevations in lithium serum concentration and increased risk of neurotoxic lithium effects may occur when carbamazepine and lithium are used simultaneously, because carbamazepine decreases lithium renal clearance.
Many other drugs may increase or decrease serum lithium levels by influencing its absorption or excretion by the kidneys; for example, tetracyclines increase lithium levels.
Lithium Toxicity
One major difficulty associated with the administration of lithium carbonate is its low therapeutic index; lithium toxicity is closely related to serum lithium levels and may occur at doses of lithium carbonate close to those necessary to achieve therapeutic serum lithium levels. Adverse or side effects are those unwanted symptoms that occur at therapeutic serum lithium levels, whereas toxic effects occur when serum lithium levels exceed therapeutic levels. However, this is not
P.261

an absolute, as patients who are unusually sensitive to lithium may develop toxic signs at serum levels below 1.0 mEq/L (PDR, 2000).
Lithium toxicity may be heralded by diarrhea, vomiting, mild ataxia, coarse tremor, muscular weakness and fasciculations (twitches), drowsiness, sedation, slurred speech, and impaired coordination. Patients and/or their caretakers must be made familiar with the symptoms of early lithium toxicity and instructed to discontinue lithium immediately and contact their physician if such signs occur. Increasingly severe and life-threatening toxic effects, including cardiac arrhythmias and severe central nervous system difficulties such as impaired consciousness, confusion, stupor, seizures, coma, and death, may occur with further elevations in serum lithium levels.
No specific treatment for lithium toxicity is available. If signs of early lithium toxicity appear, the drug should be withheld, lithium levels determined, and the medication resumed at a lower dosage only after 24 to 48 hours. Severe lithium toxicity is life-threatening and requires hospital admission, treatments to reduce the concentration of the lithium ion, and supportive measures.
Lithium’s low therapeutic index and its pharmacokinetics make it necessary to administer lithium carbonate tablets or immediate-release capsules in divided doses, usually three or four times daily, to maintain therapeutic serum levels without toxicity. Even controlled-release tablets must be administered every 12 hours. It is essential that a laboratory capable of determining serum lithium levels rapidly and accurately be readily available to the clinician. For accuracy and serial comparisons, determinations of serum lithium levels should be made when lithium concentrations are relatively stable, and at the same time each day. Typically, blood is drawn 12 hours after the last dose of lithium and immediately before the morning dose (trough level).
Saliva lithium levels have also been used to monitor lithium levels in children, which avoids the necessity of repeated venipunctures, an upsetting experience for some children and adolescents. Perry et al. (1984) reported that saliva lithium levels in 15 children diagnosed with undersocialized aggressive conduct disorder averaged approximately 2.5 times higher than serum lithium levels, with saliva/serum ratios for individual children ranging from 1.56 to 3.99. Weller et al. (1987) found that saliva lithium levels were approximately 1.82 times those of serum levels in 14 prepubertal children receiving lithium for treatment of bipolar disorder. Saliva/serum lithium ratios in these children ranged from 1.50 to 2.32. Vitiello et al. (1988) reported saliva lithium levels to be 2.84 times those of serum lithium levels in nine children, six of whom were diagnosed with conduct disorder and three of whom were diagnosed with adjustment disorders. Saliva/serum lithium ratios in these children ranged from 2.08 to 3.88. Bernstein (1988) found that in adults the ratio of saliva lithium levels to serum lithium levels varied from 1:1 to as high as 3:1. Despite the rather marked interindividual variability in the saliva/serum ratio, it appears to be relatively constant for a given individual. Therefore, to be clinically useful, a stable saliva/serum lithium level ratio must be calculated for each patient.
Although some patients who are unusually sensitive to lithium may exhibit toxic effects at serum levels below 1 mEq/L, for most patients mild-to-moderate toxic effects occur at serum levels between 1.5 and 2 mEq/L, and moderate-to-severe reactions occur at levels of 2 mEq/L and above. Younger subjects may be at greater risk than adults for developing adverse effects at lower serum lithium levels. Many adverse effects have been reported to occur in children at serum levels well below 1 mEq/L (Campbell et al. 1984a). The most common adverse effects of lithium
P.262

carbonate in 36 children, aged 3 to 13 years and diagnosed with conduct disorder (N = 24), infantile autism (N = 8), or other (N = 4) were weight gain in 44.4%, excessive sedation in 27.8%, decreased motor activity in 25%, and stomachache, vomiting, tremor, and/or irritability in 19.4% of patients (Campbell et al. 1984a).
Lithium decreases sodium reuptake by the renal tubules; hence adequate sodium intake must be maintained. This is especially important if there is significant sodium loss during illness (e.g., sweating, vomiting, or diarrhea) or because of changes in diet or elimination of electrolytes. The importance of adequate ingestion of ordinary table salt and fluids should be emphasized. Caution during hot weather or vigorous exertion has been advised, because additional salt loss and concomitant dehydration secondary to pronounced diaphoresis may cause the serum lithium levels of patients on maintenance lithium to increase and move into the toxic range. This may also be true of sweating caused by elevated body temperature secondary to infection or heat without exercise (e.g., sauna), but some evidence suggests that heavy sweating caused by exercise may result in lowered rather than elevated serum lithium levels. Jefferson et al. (1982) studied four healthy athletes who were stabilized on lithium for 1 week before running a 20-km race. At the end of the race, the subjects were dehydrated but their serum lithium levels had decreased by 20%. The authors found that the sweat-to-serum ratio for the lithium ion was approximately four times greater than that for the sodium ion. These authors concluded that strenuous exercise with extensive perspiration was more likely to decrease rather than increase serum lithium levels, and patients were more likely to require either no change or an increase, rather than a decrease, in dosage of lithium to maintain therapeutic levels. The authors do caution, however, that any conditions that significantly alter fluid and electrolyte balance, including strenuous exercise with heavy sweating, should be carefully monitored with serum lithium levels.
Untoward Effects of Lithium Carbonate
Lithium carbonate is frequently reported to have adverse effects early in the course of treatment. Most of these diminish or disappear during the first weeks of treatment.
These early adverse effects include fine tremor (unresponsive to antiparkinsonism drugs), polydipsia, and polyuria that may occur during initial treatment and persist or be variably present throughout treatment. Nausea and malaise or general discomfort may initially occur but usually subside with ongoing treatment. Weight gain, headache, and other gastrointestinal complaints such as diarrhea may also occur. Taking lithium with meals or after meals or increasing the dosage more gradually may be helpful in controlling gastrointestinal symptoms.
Later adverse effects are often related to serum level, including levels in the therapeutic range; these include continued hand tremor that may worsen, polydipsia, polyuria, weight gain and edema, thyroid and renal abnormalities, dermatologic abnormalities (including acne), fatigue, leukocytosis, and other symptoms. As serum levels increase, toxicity increases and other, more severe untoward effects, discussed earlier under toxicity, appear.
Abnormalities in renal functioning (diminution of renal concentrating ability) and morphologic structure (glomerular and interstitial fibrosis and nephron atrophy) have been reported in adults on long-term lithium maintenance. Occasional proteinuria was reported in a 14-year-old girl (Lena et al. 1978). Vetro et al. (1985) reported that after 1 year of lithium treatment, one child developed polyuria with
P.263

daytime enuresis and impaired renal concentration. Other parameters of renal function did not change, and polyuria ceased within a few days of lithium’s being discontinued. Five other children on long-term lithium therapy showed transient albuminuria that remitted spontaneously, and discontinuation of treatment was not necessary (Vetro et al. 1985).
Lithium may also interfere with thyroid function, with decreased circulating thyroid hormones and increased thyroid-stimulating hormone (TSH). Vetro et al. (1985) reported that two children developed goiter with normal function after 1.5 to 2 years of lithium therapy.
Neuroleptic malignant syndrome has been reported in a few patients who were administered neuroleptic drugs and lithium simultaneously.
Dostal (1972) reported specific adverse effects of lithium in 14 retarded adolescent males, that interfered with patient management despite significant therapeutic gains. Polydipsia, polyuria, and nocturnal enuresis were so severe as to alienate staff who cared for the youngsters. These symptoms remitted within 2 weeks of discontinuing lithium (Dostal, 1972).
Premedication Workup and Periodic Monitoring for Lithium Treatment
Routine Laboratory Tests
Complete Blood Cell Count with Differential
Lithium frequently causes a clinically insignificant and reversible elevation of white blood cells, with counts commonly between 10,000 and 15,000 cells/mm3. The lithium-induced leukocytosis characteristically shows neutrophilia (increased polymorphonuclear leukocytes) and lymphocytopenia (Reisberg and Gershon, 1979). Thus leukocytosis can usually be differentiated from one caused by infection, because the increase in neutrophils is in more mature forms, whereas in infection younger forms predominate. Lithium may also increase platelet counts.
Serum Electrolytes
Serum electrolyte levels should be determined in particular to verify that sodium ion levels are normal, because hyponatremia decreases lithium excretion by the renal tubules.
Pregnancy Test
Lithium crosses the placenta, and data from birth registries suggest teratogenicity with increased abnormalities, including cardiac malformations, especially Ebstein’s anomaly. Lithium is relatively contraindicated during pregnancy but especially during the first trimester. Infants born to mothers taking lithium appear to be at increased risk for hypotonia, lethargy, cyanosis, and ECG changes (USPDI, 1990). All females who could be pregnant should be tested before initiation of lithium therapy and warned that, because of lithium’s teratogenic potential for the fetus, they should take care not to become pregnant while taking the medication.
Renal Function Tests
Baseline assessment of renal functioning is essential, because the kidney is the primary route of elimination of lithium. For healthy children and adolescents, a baseline serum creatinine, blood urea nitrogen (BUN) level, and urinalysis are usually adequate (Jefferson et al. 1987). If kidney disease is suspected or abnormalities are found, a more thorough evaluation, including tests such as urinalysis (including specific gravity), 24-hour urine volume, and 24-hour urine for creatinine clearance and protein, should be performed and the patient referred to a renal consultant if necessary.
P.264

Thyroid Function Tests
Lithium causes thyroid abnormalities primarily by decreasing the release of thyroid hormones. This causes such findings as euthyroid goiter; hypothyroidism; decreased triiodothyronine (T3), thyroxine (T4), and protein-bound iodine (PBI) levels; and elevated 131I and thyroid-stimulating hormone (TSH) levels between 5% and 15% of patients receiving long-term lithium therapy (Jefferson et al. 1987). Hypothyroidism resulting from lithium treatment is thought to be related to preexisting Hashimoto’s thyroiditis, suggesting that determining antithyroid antibodies as part of the workup may be useful (Rosse et al. 1989). Recommended baseline studies include thyroxine (T4), triiodothyronine resin uptake (T3RU), and TSH levels.
Cardiovascular Function Tests
Various cardiac conduction and repolarization abnormalities (e.g., bradycardia) and reversible ECG abnormalities have been reported in a large percentage of adults receiving lithium. ECG changes commonly include benign, reversible T-wave changes (flattening, isoelectricity, and inversion of T waves), which are dose-dependent, and an increase in the PQ interval (Jefferson et al. 1987). It has been hypothesized that lithium’s cardiotoxic effects result from its displacing and substituting for intracellular potassium. A baseline ECG should be obtained routinely in patients > 40 years of age or those who have any history or clinical suggestions of cardiovascular disease. Although not considered mandatory in young, healthy patients, a baseline ECG is justifiable and useful to have for comparison, should cardiovascular abnormalities develop at some later time. If patients have or develop cardiac abnormalities, frequent ECG monitoring should be done and the advice of a cardiac consultant sought. In other patients it is prudent to repeat the ECG at the time of scheduled routine physical examinations.
Calcium Metabolism Tests
Lithium may increase renal calcium reabsorption, resulting in hypocalcuria (Jefferson et al. 1987). Lithium may also cause hyperparathyroidism with hypercalcemia and hypophosphatemia, with resulting decreased bone formation or density in children. If abnormal results occur, parathyroid hormone (parathormone) levels may be determined. Lithium may also replace calcium in bone formation, especially in immature bones (USPDI, 1990). A baseline calcium level should be determined in children and adolescents, but a baseline parathormone level is not usually recommended.
Electroencephalogram
Bennett et al. (1983) reported that optimal doses of lithium caused worsening of conduct-disordered children’s EEGs in statistically significant numbers. Paroxysmal and focal EEG abnormalities, in particular, were increased over pretreatment EEGs. EEG worsening, however, did not correlate with clinical symptoms of toxicity, and children receiving lithium showed significantly more behavioral improvement than those receiving placebo.
Although an EEG is not required as a baseline workup for normal healthy youngsters, if EEG abnormalities or a seizure disorder is known to exist, a baseline EEG should be obtained and the EEG periodically monitored. Lithium levels should be determined the morning that the EEG is performed, to facilitate correlation of EEG changes and serum lithium levels.
Periodic Monitoring
Because there is little information on the long-term effects of lithium on the development and maturation of children and adolescents, periodic monitoring of thyroid, kidney, and cardiac functioning is particularly important. It is recommended that
P.265

TSH, BUN, and serum creatinine levels be determined at approximately 6-month intervals. When there is a concern about renal function, 24-hour urine volume, creatinine clearance, and protein excretion should also be determined. If a suggestion of thyroid abnormality arises, T3 and T4 levels should also be determined (Rosse et al. 1989).
Titration of Lithium Dosage
Schou (1969) noted that early untoward effects, such as nausea, diarrhea, muscle weakness, thirst, urinary frequency, hand tremor, and a dazed feeling, may be caused by a too rapid rise in serum lithium levels. Lithium is also a gastric irritant. A low initial dose of lithium taken after meals, which slows absorption, and gradual
P.266

increases in dose will often avert the development of these symptoms. When they develop, they usually subside spontaneously within a few days.
Serum lithium levels should be monitored twice weekly during the acute manic phase and until both serum level and clinical condition have stabilized. In the maintenance phase of therapy during remission, serum lithium levels and thyroid, kidney, and cardiac functions should be periodically monitored. The National Institute of Mental Health/National Institutes of Health (NIMH/NIH) Consensus Development Panel (1985) recommends that serum lithium levels be determined at intervals of 1 to 3 months and that TSH and serum creatinine values be determined every 6 to 12 months. Because there is less experience in long-term administration of lithium carbonate to children and adolescents than that to adults, the author recommends monitoring at the shorter recommended intervals; that is, determining the lithium level at least bimonthly and TSH and serum creatinine levels every 6 months.
Typically, doses of approximately 1,800 mg/day will achieve the serum lithium levels between 1 and 1.5 mEq/L necessary to control symptoms during acute mania. During long-term maintenance, serum lithium levels usually range between 0.6 and 1.2 mEq/L; this usually requires a divided daily dose between 900 mg and 1,200 mg (PDR, 2000). Berg et al. (1974), however, reported that a 14-year-old girl and her father, who were both diagnosed with bipolar manic-depressive disorder, required daily doses of lithium as high as 2,400 mg to achieve therapeutic levels.
Kutcher et al. (1990) reported differences in lithium responsiveness in adolescents with bipolar disorder only and with comorbid personality disorder. When assessed during a period of relative euthymia following discharge from an inpatient unit, 35% (N = 7) of 20 adolescents (mean age, 17.5 years) diagnosed with bipolar disorder and having had at least one manic and one depressive episode were diagnosed with at least one comorbid personality disorder. None of the subjects diagnosed with comorbid personality disorder improved on lithium, whereas 6 of the 13 subjects with bipolar disorder only improved on lithium (P = .05).
The NIMH/NIH Consensus Development Panel (1985) notes that criteria for prophylactic use of lithium in children and adolescents do not yet exist, and thus the preventive use of lithium must be based on clinical judgment. The risks versus the benefits for this age range are not yet firmly established, although available data suggest that the potential problems are similar to those encountered in adults (NIMH/NIH, 1985).
Use of Lithium Carbonate in Children Under 12 Years of Age
The therapeutic dosages of lithium carbonate used in treating children over 5 years of age with various disorders do not differ significantly from those used in treating older adolescents and adults, and the principles of administration are essentially the same (Campbell et al. 1984a). This higher dose per body weight ratio may reflect the fact that higher renal lithium clearance may occur in children and adolescents than in adults.
Weller et al. (1986) published a guide for determining the initial total daily lithium dose for prepubertal children 6 to 12 years of age. The guide and summary of how it is used are presented in Table 8.1. Lower initial doses should be used for children diagnosed with mental retardation or organicity (central nervous system damage) (E.B. Weller, personal communication, 1990).
The purpose of this guide is to reach therapeutic serum lithium levels (0.6 to 1.2 mEq/L) as rapidly as possible using currently available tablet strengths without
P.267

undue risk of reaching toxic serum levels. The authors administered lithium to ten subjects diagnosed with manic-depressive illness and five subjects diagnosed with conduct disorder, following these guidelines. Thirteen of the 15 subjects had serum lithium levels in the therapeutic range after only 5 days of treatment. Side effects were reported to be minimal, primarily mild nausea, abdominal pain, polydipsia and polyuria, and increase in preexisting enuresis. Most were transient and none required discontinuation of lithium. As discussed earlier, some adverse effects of lithium appear to be related to excessively rapid increases in serum lithium level. It remains to be determined whether the use of the proposed lithium dosage guide will cause significantly more adverse effects or will increase their severity more than would a more gradual titration of lithium. In cases where very rapid control of symptoms is critical, however, it may be prove to be especially useful.
Table 8.1 • Lithium Carbonate Dosage Guide for Prepubertal School-Aged Childrena,b
Weight 8 AM
Dose
12 Noon
Dose
6 PM
Dose
Total
Daily Dose
< 25 kg 150 mg 150 mg 300 mg 600 mg
25–40 kg 300 mg 300 mg 300 mg 900 mg
40–50 kg 300 mg 300 mg 600 mg 1,200 mg
50–60 kg 600 mg 300 mg 600 mg 1,500 mg
aDose specified in schedule should be maintained at least 5 days with serum lithium levels drawn every other day 12 hours after ingestion of the last lithium dose until two consecutive levels appear in the therapeutic range (0.6–1.2 mEq/L). Dose may then be adjusted based on serum level, side effects, or clinical response. Do not exceed 1.4 mEq/L serum level.
bLower initial dose should be used for children diagnosed with mental retardation or organicity.
From Weller EB, Weller RA, Fristed MA. Lithium dosage guide for prepubertal children: A Preliminary report. J Am Acad Child Psychiatry 1986:25;92–95.
Reports of Interest
Lithium Carbonate in the Treatment of Mood Disorders (Mania, Bipolar Disorder), Behavioral Disorders with Mood Swings, and/or Patients Whose Parent(s) are Lithium Responders
DeLong and Aldershof (1987) reported successful treatment with lithium carbonate in 66% of 59 children diagnosed with bipolar affective disorder; 82% of 11 children with emotionally unstable character disorder; and 71% of seven offspring of a lithium-responsive parent.
Varanka et al. (1988) treated ten prepubertal children (nine males, one female; mean age, 9 years, 6 months ± 2 years; range, 6 years, 9 months to 12 years, 7 months) with lithium carbonate. All ten children were diagnosed with manic episode with psychotic features. Doses ranged from 1,150 to 1,800 mg/day (32 to 63 mg/kg/day). Therapeutic lithium levels of 0.6 to 1.4 mEq/L were reached in all cases within 3 to 5 days. Substantial improvement was observed in all the children within an average of 11 days (range, 3 to 24 days) after therapy was begun.
P.268

All psychotic symptoms remitted, and mood normalized. The children became less irritable and destructive; their thought processes, motor activity, and attention spans improved remarkably. Untoward effects, such as fatigue, diminished appetite, abdominal discomfort, nausea, urinary frequency, and tremor, were infrequent and so mild that they did not necessitate discontinuation of lithium.
Carlson et al. (1992) reported data on 11 hospitalized children (age range, 5 years, 11 months to 12 years) whom the authors thought were likely to be lithium responders. Diagnoses and symptomatology varied but included bipolar disorder (two), manic episode (three), bipolar not otherwise specified (four); disruptive behavior disorders (nine), exhibited psychotic symptoms (seven), and multiple placements in the seclusion room for explosive behaviors (six). Four subjects had first-degree relatives with histories of bipolar illness. Seven of the subjects participated in a double-blind, placebo-controlled protocol, whereas the other four received lithium on an open basis. Lithium dosage ranged from 600 to 1,500 mg/day, resulting in serum lithium levels that ranged between 0.7 and 1.1 mEq/L. In general, positive clinical responses increased with time, and improvements in self-control, aggression, and anxiety/agitation were greater at 8 weeks than at 4 weeks in all the subjects. These improvements, however, could not be distinguished from the effects of a longer time in the hospital, because the seven children in the double-blind crossover study maintained their gains on placebo. Only three subjects improved sufficiently on lithium to be discharged on that drug. Three patients diagnosed with bipolar disorder or bipolar disorder not otherwise specified (NOS) and one diagnosed with major depressive disorder did not have adequate therapeutic responses to lithium; however, they improved and were discharged when given an open trial of desipramine. There was no evidence that lithium caused worsening or improvement in attention, cognitive functioning, or learning performance on several rating scales (Carlson et al. 1992).
Lithium Carbonate in the Treatment of Acute Mania in Adolescents
Kafantaris et al. (2003) conducted a 4-week, open trial of lithium carbonate in treating acute mania in 100 adolescents (mean age 15.23 years, age range, 12 to 18 years; 50 males, 50 females) who had been diagnosed with bipolar I disorder and met DSM-IV criteria for a current manic or mixed episode and had a score of ≥ 16 on the Young Mania Rating Scale (YMRS). ADHD was a codiagnosis in 31% of patients. Immediate-release lithium was rapidly titrated to therapeutic serum levels between 0.6 and 1.2 mEq/L using Cooper’s technique (Cooper et al. 1973). Subjects with severe aggression and/or psychosis were treated concomitantly with antipsychotics. Mean lithium serum level at the end of week 1 was 0.90 ± 0.25 mEq/L; at endpoint (week 4), the mean serum level was 0.93 ± 0.21 mEq/L and the mean dose was 1,355 ± 389 mg/day.
Subjects were rated weekly on the YMRS, Hamilton Depression Rating Scale (HDRS, 17 item), Brief PsychiatricRating Scale (BPRS), Clinical Global Impressions-Improvement (CGI-I) Scale, and the Children’s Global Assessment Scale (CGAS). Responders were defined as having both a decrease of > 33% from baseline YMRS score and a ≤ 2 rating (much or very much improved) on the CGI-I. At the end of week 4, all the ratings showed significant improvement (P < .001). Sixty-three patients met responder criteria by the end of week 2. Remission of manic symptoms (YMRS score < 6) occurred in 26 patients by week 4 and only 4 of the 23 patients with suicidal ideation at baseline had such symptoms by week 4. The authors reported that the presence of baseline psychotic features
P.269

(with antipsychotic treatment), prominent depressive symptoms, comorbid diagnoses including ADHD, early onset of mood disorders, severity of mania at initial presentation and hospitalization did not impact significantly on response to lithium at week 4.
Adverse events present at week 4 ratings in > 10% of patients included weight gain (1 to 12 pounds), 55.3%; polydipsia, 33.3%; polyuria 25.5%; headache 23.5%; tremor 19.6%; gastrointestinal pain, 17.6%; nausea, 15.7%; vomiting 13.7%; anorexia 13.7%; and diarrhea 13.7%.
This study suggests that lithium is an effective treatment for moderate-to-severe acute mania in adolescents (Kafantaris et al. 2003).
Lithium Carbonate versus Valproic Acid in the Maintenance Treatment of Children and Adolescents Diagnosed with Bipolar Disorder
The paper by Findling et al. (2005), which found no clinically significant differences between the two drugs for this indication is summarized in the section on valproic acid.
Lithium Carbonate in the Treatment of Adolescents Diagnosed with Bipolar Disorder and Substance Dependency
Geller et al. (1998) conducted a 6-week, double-blind, placebo-controlled, parallel-groups study comparing lithium and placebo in the treatment of 25 outpatients (16 males, 9 females; mean age, 16.3 ± 1.2 years) diagnosed by DSM-III-R (APA, 1987) criteria with a bipolar disorder or major depressive disorder with one or more predictors of future bipolar disorder and substance dependency disorder. The mean age of onset of substance abuse disorders was approximately 6 years after the mean age of onset of subjects’ mood disorders. Subjects did not have to agree to stop their substance abuse to participate in the study. Thirteen subjects were assigned to the lithium group; of these, ten completed the study. Twelve were assigned to the placebo group and 11 completed the study.
Efficacy was determined by ratings on the Children’s Global Assessment Scale (CGAS) and random weekly urine drug assays. “Responders” were required to have a score of ≥ 65. Lithium was initiated with a 600-mg dose; serum lithium levels were determined 24 hours later, and an initial target dose was calculated for each subject using a nomogram to yield a serum lithium level between 0.9 and 1.3 mEq/L. The dose was further individually adjusted as necessary; total dose was divided and given at 7:00 AM and 7:00 PM daily. The subjects on lithium improved significantly more than those on placebo. Six (60%) of the 10 completers on lithium were “responders,” compared with 1 (9.1%) of the 11 completers on placebo (P = .024). The mean daily lithium dose for the ten completers was 1,733 ± 428 mg; the responders’ daily dose was significantly higher (1,975 ± 240 mg) than that of the nonresponders (1,368 ± 399 mg; P = .02), but there was no significant difference in their serum lithium levels (responders, 0.88 ± 0.27 mEq/L vs. nonresponders, 0.85 ± 0.3 mEq/L). After 3 weeks, the percentage of the weekly random urine tests that was positive for drug assays was significantly lower in the lithium group than that in the placebo group (P = .042) for the 21 completers. The ratings of untoward effects on the acute lithium side effects scale showed that lithium was well tolerated. Only polyuria and polydipsia occurred significantly more frequently in the lithium group than in the placebo group. The authors concluded that the results after 6 weeks in their sample were encouraging but further research for these chronic disorders was needed with larger samples and long-term maintenance on the lithium.
P.270

Lithium Carbonate in the Treatment of Disorders with Severe Aggression, Especially when Accompanied by Explosive Affect, Including Self-Injurious Behavior
In a double-blind, placebo-controlled study of 61 treatment-resistant hospitalized children (age range, 5.2 to 12.9 years) diagnosed with undersocialized aggressive conduct disorder, both haloperidol and lithium were found to be superior to placebo in ameliorating behavioral symptoms (Campbell et al. 1984b). Optimal doses of lithium carbonate ranged from 500 to 2,000 mg/day (mean, 1,166 mg/day); corresponding serum levels ranged from 0.32 to 1.51 mEq/L (mean, 0.99 mEq/L), and saliva levels ranged from 0.81 to 5.05 mEq/L (mean, 2.52 mEq/L). The authors noted that lithium caused fewer and milder untoward effects than did haloperidol and that these effects did not appear to interfere significantly with the children’s daily routines. There was also a suggestion that lithium was particularly effective in diminishing the explosive affect and that other improvements followed (Campbell et al. 1984b).
In 1995, Campbell et al. (1995) reported a double-blind, placebo-controlled study that was designed to replicate their 1984 study. Fifty treatment-resistant inpatients (46 males, 4 females; mean age, 9.4 ± 1.8 years; age range, 5.1 to 12.0 years) diagnosed with conduct disorder, undersocialized aggressive type by DSM-III (APA, 1980a) criteria and having chronic severe explosive aggressiveness were treated with lithium carbonate only or placebo. Following a 2-week, placebo baseline period during which baseline assessments were conducted and placebo responders were eliminated, the 50 remaining subjects were randomly assigned to placebo (N = 25) or lithium (N = 25) for a 6-week period; this was followed by 2 weeks of posttreatment placebo. Efficacy was assessed by ratings on the Global Clinical Judgments (Consensus) Scale, Children’s Psychiatric Rating Scale (CPRS), Clinical Global Impressions (CGI), Clinical Global Impressions-Severity (CGI-S), and Improvement (CGI-I) Scales, Conners Teacher Questionnaire (CTQ), and the Parent-Teacher Questionnaire (PTQ). Lithium carbonate was begun at 600 mg/day and titrated individually over a 2-week period with a maximum permitted dose of 2,100 mg/day or serum lithium of 1.8 mEq/L or equivalent saliva lithium level. The mean optimal dose of lithium was 1,248 mg/day (range, 600 to 1,800 mg/day); the mean serum lithium level was 1.12 mEq/L (range, 0.53 to 1.79 mEq/L; and the mean saliva lithium level was 2.5 mEq/L (range, 1.45 to 4.44 mEq/L).
On the Global Clinical Judgments (consensus) Scale, 68% (17/25) of subjects on lithium were rated as moderately or markedly improved while only 40% (10/25) of subjects on placebo were so rated (P = .003). Further refining this measure, 40% (10/25) of the subjects of lithium were “markedly” improved versus only 4% (1/25) of the subjects on placebo. The CGI-I scores after 6 weeks were also significantly better for the lithium group (P = .044); although it was not significant whether the lithium group improved more on the CGI-S. The authors concluded that these data supported the conclusions of their earlier study and that lithium carbonate can be efficacious in treating children with conduct disorder and explosive aggressiveness who have not responded to psychosocial treatments or medication with methylphenidate or standard neuroleptics.
Vetro et al. (1985) treated 17 children, aged 3 years to 12 years, with lithium, who were hospitalized for hyperaggressivity, active destruction of property, severely disturbed social adjustment, and unresponsiveness to discipline. Ten of the children
P.271

had not responded to prior pharmacotherapy, including haloperidol and concomitant individual and family therapy. Lithium carbonate was titrated slowly over 2 to 3 weeks to achieve serum levels in the therapeutic range (0.6 to 1.2 mEq/L). Mean serum lithium level was 0.68 mEq/L ± 0.30 mEq/L. The authors reported that 13 of the children improved enough that their abilities to adapt to their environment could be described as good, and their aggressivity had been reduced to tolerable levels. Three of the four cases that did not improve had poor compliance in taking the medication at home. The authors also noted that these children usually required continuous treatment with lithium for longer than 6 months.
DeLong and Aldershof (1987) reported that rage, aggressive outbursts, and, interestingly, encopresis responded favorably to lithium pharmacotherapy in children with behavioral disorders associated with a variety of neurologic and medical diseases, including mental retardation.
Lithium Carbonate in the Treatment of Children and Adolescents Diagnosed with Conduct Disorder
Malone et al. (2000) conducted a 6-week, double-blind, placebo-controlled, parallel-groups study comparing lithium carbonate and placebo in the treatment of 40 inpatients (33 males, seven females; mean age, 12.5 years; age range, 9.5 to 17.1 years) who were diagnosed with conduct disorder by DSM-III-R (APA, 1987) criteria and hospitalized for chronic, severe aggressive behavior. Eighty-six inpatients entered the study, however, 46 were eliminated during the initial 2-week single-blind placebo baseline; 40 of this group did not meet the protocol’s aggression criteria. All 40 remaining subjects entered the 4-week treatment phase and completed the protocol; 20 subjects were assigned randomly to each group.
Efficacy was determined by ratings on the Global Clinical Judgments (Consensus) Scale (GCJCS), the Clinical Global Impressions (CGI), and the Overt Aggression Scale (OAS). Lithium was initiated with a 600-mg dose; serum lithium levels were determined 24 hours later and an initial target dose was calculated for each subject using a nomogram. Subsequent lithium doses were increased by 300 mg daily and given in three equal doses to reach the target dose. At the end of the study, optimal mean lithium dose was 1,425 ± 321 mg/day (range, 900 to 2,100 mg/day) with a mean steady-state therapeutic lithium level of 1.07 ± 0.19 mmol/L (range, 0.78 to 1.55 mmol/L).
On the GCJCS, 16 (80%) of the lithium group versus six (30%) of the placebo group were rated as “marked” or “moderately” improved on the criterion for responders (P = .004). Significantly more of the lithium group were also rated as responders on the CGI (17 [30%] vs. four [20%] of the placebo group; P = .004). On the OAS, the lithium group continued to show improvement over the 4-week period, whereas the placebo group showed an initial decline at week 1 but then remained rather stable. The lithium group’s mean decrease from baseline was significantly greater than that of the placebo group, with a significant interaction between treatment group and time (P = .04). Although untoward effects were frequent, they were usually mild and similar for both placebo and lithium groups. Only three adverse effects occurred significantly more on lithium: nausea in 12 of 20, vomiting in 11 of 20, and urinary frequency 11 of 20 (P ≤ .05 in all cases). The authors noted that the aggressive behavior of 40 (47.1%) of their initial 85 subjects improved significantly during the first 2 weeks secondary to hospitalization and treatment with placebo alone. For the 40 subjects who remained aggressive and
P.272

entered the medication phase of the protocol, lithium was a safe and effective treatment. The authors noted that determining the long-term efficacy and safety of lithium in such subjects will require further research.
Lithium Carbonate in the Treatment of ADHD
Greenhill et al. (1973) and DeLong and Aldershof (1987) reported that lithium was not effective or worsened symptoms in the treatment of children with earlier equivalent diagnoses of ADHD.
Antiepileptics/Mood Stabilizers
The use of antiepileptic drugs for the treatment of psychiatric disorders in children and adolescents was reviewed by Stores, 1978. He concluded that “while some of the antiepileptic drugs show possibilities as psychotropic agents, their use in children with nonepileptic conditions such as behavior or learning disorders of childhood cannot be justified except as a carefully controlled research exercise” (p. 314).
Currently, most clinical interest in the off-label use of antiepileptic drugs to treat psychiatric disorders in children and adolescents is focused on valproic acid (valproate) and carbamazepine; their safety and efficacy in treating these disorders remains to be elucidated. Carbamazepine and valproic acid are being used with increasing frequency to treat many psychiatric and neuropsychiatric disorders that have failed to respond satisfactorily to more standard therapies. Both drugs have been used independently and as adjunctive agents to treat adults and children with bipolar disorder and mania as mood stabilizers and to treat patients with aggressiveness directed either toward self or others, and behavioral dyscontrol. Some mentally retarded persons with concomitant affective symptomatology and disordered behavior have also shown significant clinical benefit from these drugs, allowing decrease or cessation of antipsychotic drugs. Further research is necessary to determine which specific disorders, symptoms, and patients or subgroups of patients are most likely to respond well (e.g., patients with various abnormal EEG findings and patients who are mentally retarded or have other evidence of abnormal central nervous system functioning compared with affectually or behaviorally disordered patients without signs of central nervous system dysfunction). It will also be important to ascertain further the adverse effects in children and adolescents when these drugs are used for off-label indications rather than to treat seizure disorders. Both have rare but potentially fatal adverse effects and must be used cautiously and monitored carefully.
Valproic Acid (Depakene); Divalproex Sodium (Valproic Acid and Valproate Sodium [Depakote; Depacon])
Valproic acid and divalproex sodium (a stable coordination compound of valproic acid and valproate sodium) both dissociate to the valproate ion in the gastrointestinal tract and have antiepileptic properties. These drugs are indicated for the treatment of simple and complex absence seizures and adjunctively in patients with multiple seizure types, which include absence seizures. Divalproex sodium has also been approved by the FDA for advertising as safe and effective in the treatment of manic episodes associated with bipolar disorder for up to 3 weeks and the prophylaxis of migraine headaches in adults (PDR, 2006).
Pharmacokinetics of Valproic Acid
Following oral administration, valproic acid and divalproex sodium dissociate to the valproate ion, which is the active agent, in the gastrointestinal tract. Administration of valproic acid with food may slow the absorption rate but does not interfere with clinical efficacy. Food does not significantly affect the total amount of valproate absorbed and may be helpful in reducing gastrointestinal irritation in some patients.
Peak plasma concentration after a single dose usually occurs between 1 and 4 hours after ingestion of valproic acid, 4 hours after ingestion of sodium valproate tablets, and 3.3 hours after taking sodium valproate “sprinkles.” Valproic acid is metabolized almost entirely by the liver; the metabolites are excreted primarily in the urine. Plasma valproate half-life is between 6 and 16 hours; the more rapid metabolism rates occur most frequently in patients receiving valproic acid and other antiepileptics that induce enzymes that increase the metabolism rate of valproate.
In a retroactive chart review of 16 males (age range, 5 to 14 years; mean age, 9.3 years) hospitalized for mood stabilization, Good et al. (2001) found that a relatively conservative total loading dose of 15 mg/kg/day of divalproex sodium given in two equal doses resulted in therapeutic trough plasma valproate levels on day 5 of therapy in 13 (81.3%) cases. The initial dose was calculated for one subgroup using actual weight and for a second subgroup using adjusted ideal body weight (IBW). For the latter group, Adjusted IBW = IBW + 40% (Current Weight–IBW). All subjects were also taking atypical antipsychotics, and some were taking stimulants as well during this period. The authors noted several findings of clinical interest: of the eight patients experiencing untoward effects (mostly sedation and nausea), six (75%) had valproate plasma levels of > 90 μg/mL. Patients who were ≥ 15%
P.274

over IBW and who were dosed according to actual body weight were significantly more likely to have supratherapeutic (> 120 μg/mL) valproate plasma levels than normal-weight subjects or overweight subjects whose doses were determined by adjusted IBW. Based on this study, it would seem prudent to calculate and use adjusted IBW for significantly overweight children and adolescents if it is decided to administer a loading dose of valproate to rapidly achieve therapeutic plasma levels.
Contraindications for Valproic Acid Administration
Valproic acid can cause severe hepatotoxicity, including fatal hepatic failure. Children under 2 years of age are at increased risk. It should not be administered to anyone with hepatic disease, significant liver dysfunction, or known hypersensitivity to the drug.
Because valproic acid has been reported to cause teratogenic effects in the fetus, it should be administered with caution to women who are likely to become pregnant, and they should be warned to notify their physician immediately if they become pregnant.
Interactions with Other Drugs
Valproate may potentiate the action of central nervous system depressants such as alcohol and benzodiazepines.
Coadministration with clonazepam may induce absence seizures in patients with a history of absence-type seizures.
Coadministration with risperidone (4 mg/day) did not affect the predose or average plasma concentrations and exposure area under the curve (AUC) of valproate (a total of 1,000 mg administered in three divided doses) but there was a 20% increase in valproate peak plasma concentration after concomitant administration of risperidone.
Ambrosini and Sheikh (1998) have reported two cases in which coadministration of valproic acid and guanfacine resulted in significantly increased levels of valproic acid. It was suggested that this was secondary to drug-drug competition at the level of hepatic glucuronidation.
Other drug interactions have been reported.
Untoward Effects of Valproic Acid
The most serious side effect of valproic acid is hepatic failure, which can be fatal; it occurs most frequently within the first 6 months of treatment. Children under 2 years of age are at increased risk. The risk of hepatotoxicity decreases considerably as patients become progressively older. Hence liver function must be monitored carefully and frequently, especially during the first 6 months of treatment.
Nausea, vomiting, and indigestion may occur early in treatment with valproic acid and usually are transient. Sedation may occur, and untoward psychiatric effects such as emotional upset, depression, psychosis, aggression, hyperactivity, and behavioral deterioration have been reported. Thrombocytopenia, other hematologic abnormalities, and many other untoward effects have been reported.
Valproate and Polycystic Ovaries
Isojarvi et al. (1993) published an article noting that there was an association between valproate use in treating epileptic women and polycystic ovaries and hyperandrogenism (elevated serum testosterone concentrations). The finding was
P.275

more pronounced in women who had begun treatment with valproate before 20 years of age than in women who began valproate treatment at 20 years of age or older. Sussman and Ginsberg (1998) published a critical review of valproate and polycystic ovary syndrome (PCOS), concluding that the available evidence suggests that early and long-term treatment with valproic acid is a causal or precipitating factor in the development of polycystic ovary syndrome in epileptic women, particularly if they are overweight; relative risk factors for nonepileptic adolescents are at present unknown. Johnston (1999) basically concurs. Piontek and Wisner (2000) have suggested clinical guidelines for the appropriate clinical management of women with reproductive capacity who are treated with valproate. Although risk of PCOS does not preclude the use of valproic acid/divalproex sodium in adolescent females, risks and benefits must be discussed, informed consent obtained, and careful monitoring maintained. Further research is urgently needed to clarify this issue.
Reports of Interest
Valproic Acid in the Treatment of Adolescents Diagnosed with Mania
West et al. (1994) treated 11 adolescents (nine males and two females; age range, 12 to 17 years; mean, 14.5 years) with valproate, on an open basis, who were diagnosed with bipolar disorder (five were manic type and six were mixed type) and hospitalized for acute mania. Seven patients had comorbid diagnoses of ADHD. All the patients had unsatisfactory responses to antipsychotic drugs alone (N = 5) or in combination with lithium (N = 6). Valproate was added to the medications already being prescribed and was begun at a dose of 250 mg twice daily and titrated upward based on clinical response and adverse effects. Optimal doses ranged from 500 to 2,000 mg/day (mean, 1,068 mg/day), and serum levels ranged from 38 to 94 μg/mL (mean, 74 μg/mL). The length of inpatient valproate administration ranged from 6 to 26 days, with a mean of 17 days. Three patients showed marked improvement, with virtual or complete remission of manic symptoms; six patients improved moderately, with significant reduction of symptoms; and two showed
P.277

only slight improvement. The only adverse effects reported were sedation in two patients.
Papatheodorou et al. (1995) reported an open-label, 7-week study in which the efficacy and safety of divalproex sodium was assessed in the treatment of 15 subjects (two males, 13 females; mean age, 17.3 years, with 10 subjects being 15 to 18 years old and 5 being 19 or 20 years old) who were diagnosed by DSM-III-R (APA, 1987) criteria with bipolar disorder, in an acute manic phase. Efficacy was evaluated using ratings on the Modified Mania Rating Scale (MMRS), the Brief Psychiatric Rating Scale (BPRS), the Global Assessment Scale (GAS), the Clinical Global Impressions Scale (CGI), and the Valproic Acid Side Effects Scale (VA-SES). Following a 2-day entry phase during which baseline evaluations were performed, subjects began 7 weeks of treatment with divalproex sodium. Medication was administered in three divided doses and individually titrated. Thirteen patients completed the 7-week study; one patient was discontinued for lack of clinical response and one patient withdrew because of “subjectively intolerable sedation and dizziness.” All 13 completers required some additional medication for symptom control (e.g., agitation) during the study. Mean dose at the end of 7 weeks was 1,423.08 mg/day (range, 750 to 2,000 mg/day) and the mean serum valproic acid level (12 to 14 hours after the evening dose and before the morning dose) was 642.85 ± 183.08 μmol/L (range, 360 to 923 μmol/L).
The 13 completers’ ratings on the MMRS, BPRS, GAS, and CGI were all very significantly lower (P < .0001 for all four scales) than at baseline. An analysis of variance (ANOVA) found a significant reduction in the MMRS within 1 week on valproex (P < .016), which continued to increase throughout the treatment period. Overall untoward effects were benign and their frequency was reported to decrease over the duration of the study, with a very low number being reported at the end of the study. The most significant adverse effect was the development of hypocortisolemia and hypothyroidism in one patient. Liver function tests remained normal except for one patient with transiently elevated enzyme levels that reverted to normal without change in dosage. The data suggest that divalproex sodium is safe and efficacious in the acute (short-term) treatment of mania in adolescents; further research is indicated.
Valproic Acid versus Lithium in the Maintenance Treatment of Children and Adolescents Diagnosed with Bipolar Disorder
Findling et al. (2005) conducted a double-blind study to determine whether Divalproex sodium (DVPX) or lithium was superior as the only drug in maintenance treatment of 139 subjects (age range 5 to 17 years; mean age 10.8 ± 3.5 years, 93 (66.9%) males 46 (33.1%) females) who were diagnosed with Bipolar I (131, 94.2%) or Bipolar II (8, 5.8%) disorder and stabilized on a combination of lithium carbonate and valproex sodium during acute treatment. Sixty subjects who met remission criteria (Children’s Depression Rating Scale score < 40, Young Mania Rating Scale score < 12.5 and a Children’s Global Assessment Scale score > 51) for a minimum of 4 weeks were than randomized to monotherapy with either lithium (N = 30) or divalproex (N = 30) for up to 76 weeks; subjects were dropped from the study if they violated protocol or required additional clinical intervention. Subjects were tapered off the nonmaintenance/discontinued drug over a period of 8 weeks to minimize discontinuation rebound relapse. Subjects maintained on lithium were maintained at lithium serum concentrations between 0.6 and 1.2 mmol/L and those on valproex were maintained with valproate plasma concentrations between 50 and 100 μg/mL. Primary measures of effectiveness were time to premature discontinuation because
P.278

of emerging mood symptoms of relapse and premature discontinuation for any reason.
Median survival time to mood relapse for subjects on lithium was 114 ± 57.4 days for lithium and 112 ± 56 days for subjects on valproex and was not statistically different (P = .55); overall, 38 (63.3%) subjects relapsed. There was also no significant difference between the lithium and valproex groups in the 12 (20%) who dropped out for any reason (P = .72). At the study’s conclusion, the mean lithium serum level was 0.84 ± 0.3 mmol/L and the mean valproate plasma level was 75.3 ± 29.4 μg/mL (Findling et al. 2005). Only six subjects (10%), three in each treatment group completed the 76-week protocol, a vivid indication of the chronic and debilitating course of pediatric bipolar disorder.
Regarding adverse events, comparing lithium to valproex, emesis (30% vs. 3%), enuresis (30% vs. 6.7%), and increased thirst (16.5% vs. 0%) were significantly more frequent in the lithium group; other frequent adverse events, which were not significantly different between lithium and valproex were headache (13.3% vs. 23.3%), tremor (20.0% vs. 16.7%), stomach pain (10.0% vs. 23.3%), nausea (16.7% vs. 6.7%), diarrhea (13.3% vs. 6.7%) and decreased appetite (10% vs. 10%).
The authors concluded that there was no clinically significant difference between lithium and valproex monotherapy in maintaining the youth who were stabilized on combination lithium/valproex therapy for bipolar disorder (Findling et al. 2005).
Valproic Acid in the Treatment of Children and Adolescents Diagnosed with Mental Retardation and Mood Disorders
Kastner et al. (1990) reported treating three patients with valproic acid, a 16-year-old moderately retarded male and 13- and 8-year-old profoundly retarded girls, all of whom also had symptoms of mood disorder. These symptoms included self-injurious behaviors such as face gouging and head banging, irritability, aggressiveness, hyperactivity, sleep disturbance, and paroxysms of crying. All had unsatisfactory responses to several trials of other medications. All three patients showed excellent response to valproic acid and at follow-up had maintained their gains for 7 to 10 months. Maintenance doses were 2,700 mg/day (plasma level, 109 μg/mL) for the 16-year-old, 3,000 mg/day (plasma level, 75 μg/mL) for the 13-year-old, and 1,500 mg/day (plasma level, 111 μg/mL) for the 8-year-old. The authors noted that the plasma levels were high or just above the typical therapeutic upper range and that no hepatic abnormalities developed in their patients.
In a 2-year prospective study, Kastner et al. (1993) administered valproic acid to 21 patients diagnosed with mental retardation who also had behavioral symptoms of irritability, sleep disturbance, aggressive or self-injurious behavior, and behavioral cycling that were interpreted as symptomatic of an affective disorder. Eighteen patients completed the study. (Two could not be followed up and one developed acute hyperammonemia and was dropped from the study.) Twelve of the patients completing the study were 18 years old or younger; the degree of their retardation ranged from moderate to profound. Valproic acid was titrated upward until symptoms remitted or untoward effects prevented further increase to plasma levels between 50 and 125 μg/mL. Patients’ ratings on the Clinical Global Impressions of Severity (CGI-S) Scale after 2 years on medication were significantly improved (P < .001) from ratings at baseline. Patients with a diagnosis of epilepsy or a suspicion of seizures correlated with a positive response (P < .005). Of note, nine of the ten patients who were receiving neuroleptic drugs
P.279

at the beginning of the study were no longer being prescribed these drugs at the study’s completion.
Divalproex Sodium in the Treatment of Adolescents with Explosive Mood Disorder
In an open-label, 5-week study, Donovan et al. (1997) treated with divalproex sodium ten outpatient adolescents (eight males, two females; age range, 15 to 17 years) who were diagnosed by DSM-III-R (APA, 1987) criteria with disruptive behavioral disorders (seven, conduct disorder; two, oppositional defiant disorder; and one, ADHD). Most had comorbid drug abuse or dependency (five, marijuana abuse; three, marijuana dependency; and one, alcohol abuse). All ten subjects had severe unpredictable mood swings and a low threshold/high amplitude for dyscontrol once irritable with frequent and severe temper tantrums (“explosive mood disorder”), which preceded drug abuse by at least 1 year. Efficacy was determined based on multiple informants’ (subjects, parents, and teachers) reports of temper outbursts and mood lability and the Global Assessment of Functioning (GAF; Axis V of the DSM-III-R diagnoses) Scale.
Divalproex sodium was initiated at a dose of 250 mg/day and titrated to 1,000 mg/day over a period of 2 to 4 weeks. The mean plasma valproate level after receiving 1,000 mg/day of divalproex for 1 week was 75 μg/mL, range 45 to 113 μg/mL. At the end of the fifth week, all ten subjects showed very significant improvement on all three measures; nine subjects had no temper outbursts during the fifth week and six subjects had no significant mood lability. Their mean number of temper outbursts decreased from 6.5 ± 4.5 at baseline to 0.1 ± 0.3 after 5 weeks (P < .001). The mean mood lability score (0 = least to 4 = greater frequency, duration, and autonomy of mood swings) decreased from 3.8 ± 0.4 at baseline to 0.5 ± 0.7 after 5 weeks (P < .000). The mean GAF score improved from 37.8 ± 7.0 at baseline to 65.7 ± 10.2 after 5 weeks (P < .000). Divalproex was well tolerated with only two patients reporting mild sedation and transient nausea. There were no serious untoward effects, and liver function tests showed no significant changes. Improvements were maintained while on medication during follow-up; however, five subjects independently discontinued medication for at least 5 days and rapidly relapsed; improvement recurred within a few days of resuming medication. A sixth patient took medication sporadically during follow-up and maintained gains for approximately 6 weeks, when partial relapse occurred. The data suggest that divalproex sodium may be safe and efficacious in such adolescents; further studies should be undertaken (Donovan et al. 1997).
Donovan et al. (2000) conducted a 12-week, randomly assigned, double-blind, placebo-controlled, crossover study of divalproex sodium in the treatment of 20 outpatients (16 males, 4 females; mean age, 13.8 ± 2.4 years; age range, 10 to 18 years), all of whom were diagnosed with conduct disorder or oppositional defiant disorder by DSM-IV (APA, 1994) criteria and chronic explosive temper (more than four episodes monthly of rage, property destruction, or fighting with minimal provocation) and mood lability (multiple daily unpredictable shifts in mood from normal to irritable and withdrawn to boisterous behavior). Four subjects were diagnosed with comorbid ADHD and six with marijuana abuse. Efficacy was assessed by ratings on the Modified Overt Aggression Scale and on six items from the anger-hostility subscale of the Symptom Checklist-90 (SCL-90); it was decided a priori that “responders” had to have a ≥ 70% reduction from baseline scores on both rating scales. The first 6 weeks of the study consisted of a parallel-groups design, with ten subjects randomly assigned to valproate or placebo. Divalproex was gradually titrated to 10 mg/lb/day over the first 2 weeks; if the plasma level
P.280

of valproate was < 90 μg/mL at that time, a single increase of 250 mg/day was added. (To preserve the blind, a similar number of increases were made in the placebo group.) Doses ranged from 750 to 1,500 mg/day, and the mean plasma valproate level was 82.2 ± 19.1 μg/mL. At the end of this 6-week phase, eight (80%) of the ten patients receiving divalproex were rated as responders versus no responders in the ten subjects on placebo (P < .001). Seventeen subjects completed phase I (during the first 2 weeks, one subject on divalproex dropped out as he was incarcerated for parole violation and two subjects on placebo dropped out for lack of clinical improvement). Fifteen subjects (eight responders to divalproex and seven nonresponders to placebo) entered the crossover phase of the study (weeks 7 to 12) and all completed it. Six (86%) of the seven placebo nonresponders during phase I responded to divalproex during phase II. Six of the eight responders to divalproex during phase I began relapsing between 1 and 2 weeks into phase II, and at the end of week 12, their average Modified Overt Aggression Scale score had worsened to only 33% over baseline and their average anger-hostility scores on the subscale of the SCL-90 declined to 27% over baseline. Of the 15 subjects completing the entire study, 12 met “responder” criteria only during the medication phase, indicating that divalproex is significantly better than placebo (P = .003) in this population.
Divalproex Sodium in the Treatment of Aggression in Children and Adolescents Diagnosed with Pervasive Developmental Disorders
Hellings et al. (2005) conducted a double-blind, placebo-controlled study to evaluate the efficacy of valproate (VPA) in treating aggressive symptoms in 30 subjects (20 male and 10 female; age range 6 to 20 years) who were diagnosed with a pervasive developmental disorder (PDD) by DSM-IV criteria (27 were diagnosed with autistic disorder, 1 with PDDNOS and 2 with Asperger’s disorder). Comorbid diagnoses, with the exception of Tourette’s disorder were permitted. No other psychotropic medications or antiseizure medications were permitted. Subjects exhibited significant aggression toward themselves or others, or to property, a minimum of three times weekly. Twenty-six subjects had IQs in the mentally retarded range. Subjects were randomly assigned to liquid placebo (N = 14) or liquid VPA (N = 16) for a period of 8 weeks, following a 1-week lead-in on placebo. In the VPA group, the liquid placebo was gradually replaced by liquid VPA beginning with a 250 mg/5 mL dose. VPA liquid (250 mg/5 mL) was added every 3 days to reach a target dose of 20 mg/kg/day. A psychiatrist not involved in ratings, adjusted the VPA to achieve trough plasma levels of 70 to 100 μg/mL after measurement at the end of 2 and 4 weeks. Mean VPA trough plasma levels were 75.5 μg/mL at week 4 and 77.8 μ/mL at week 8.
There were no statistically significant differences between the two groups on the primary outcome measure, the Aberrant Behavior Checklist—Community Scale (ABC-C; P = .65), or the secondary outcome measures, the Clinical Global Impressions-Improvement Subscale (CGI-I; P = .16) and the Overt Aggression Scale (OAS; P = .96). The CGI-Severity Subscale (CGI-S) also showed no statistical difference between the groups (P = .96).
Adverse effects were usually mild. One subject on VPA developed a rash and dropped out of the study. Increased appetite was the only adverse effect that was significantly greater in the VPA group (P = .03). Gastrointestinal complaints, sedation, headache, chills, and fever did not differ. Two subjects on VPA had elevations of ammonia above the normal range of 21 to 50 μmol/L and the parent
P.281

of one of these subjects reported cognitive slowing and slurred speech at times (ammonia was 98 μmol/L at the end of the study).
The authors noted that there was high intrasubject variability with large differences in the frequency and severity of aggression in different weeks, and high intersubject variability with large standard deviations for each of the outcome measures, which weakened study power. Following completion of the study, ten subjects on VPA elected to continue on the drug and six on placebo elected to an open trial of VPA. Ten of these 16 subjects continued to demonstrate a positive and sustained response. The authors concluded that although this study did not demonstrate efficacy of VPA, there might be a subgroup of aggressive children and adolescents with PDD who respond favorably to VPA and that a larger, multisite study is indicated.
Carbamazepine (Tegretol; Carbatrol, Equetro)
Carbamazepine is an anticonvulsant indicated for the treatment of psychomotor and grand mal seizures. It is also a specific analgesic for trigeminal neuralgia.
Pharmacokinetics of Carbamazepine
Peak serum levels occur 4 to 5 hours after ingestion of standard carbamazepine tablets. Initial serum half-life values range from 25 to 65 hours; however, carbamazepine is an autoinducer of its own metabolism. Autoinduction stabilizes over 3 to 5 weeks at a fixed dose, with half-life decreasing to 12 to 17 hours. In children there is a poor correlation between dose and serum level of carbamazepine.
Contraindications for Carbamazepine Administration
Known hypersensitivity to carbamazepine or tricyclic antidepressants, a history of previous bone marrow depression, and the ingestion of an MAOI within the previous 14 days are contraindications.
Interactions of Carbamazepine with Other Drugs
Carbamazepine has been reported to decrease the serum half-lives of haloperidol, phenytoin, theophylline, and other drugs.
When coadministered with risperidone (6 mg/day) over a 3-week period, plasma concentrations of risperidone and 9-hydroxyrisperidone were decreased by approximately 50%. Plasma levels of carbamazepine did not appear to be effected.
When coadministered with olanzepine, carbamazepine in doses of 200 mg twice daily caused approximately 50% increase in the clearance of olanzapine. This was thought to be secondary to carbamazepine’s being a potent inducer of CYP1A2 activity. Higher daily doses of carbamazepine may cause an even greater increase in olanzapine clearance.
P.282

Carbamazepine serum levels are markedly reduced by the simultaneous use of phenobarbital, phenytoin, or primidone.
Increased lithium serum concentrations and increased risk of neurotoxic lithium effects may occur when carbamazepine and lithium are used simultaneously, because carbamazepine decreases lithium renal clearance.
The FDA has advised that carbamazepine may lose up to one-third of its potency if stored under humid conditions such as in a bathroom. Supplies should be kept tightly closed and in a dry location.
Untoward Effects of Carbamazepine
Evans et al. (1987) attributed the increased interest in the use of carbamazepine in child and adolescent psychiatry in part to both the increased awareness of the serious untoward long-term complications of standard neuroleptic drugs and the finding that the untoward effects of carbamazepine are less formidable than initially thought. In particular, the serious blood dyscrasias, agranulocytosis, and aplastic anemia are very rare. The risk of developing these disorders when treated with carbamazepine is five to eight times that of the general population. Agranulocytosis occurs in approximately six per million and aplastic anemia in approximately two per million of the untreated general population (PDR, 2000).
The most frequently reported untoward effects are dizziness, drowsiness, unsteadiness, nausea, and vomiting. These are more likely to occur if treatment is not begun with the low doses recommended. As noted earlier, aplastic anemia and agranulocytosis, although rare, have been reported. Hence a complete baseline hematologic evaluation must be done and complete blood cell count with differential and platelets must be repeated and monitored closely throughout the treatment.
Pleak et al. (1988) reported that adverse behavioral and neurologic reactions developed in 6 of their 20 male subjects, aged 10 to 16, who were diagnosed with various disorders, but primarily with ADHD and conduct disorder, and who were participating in an ongoing protocol evaluating the efficacy of carbamazepine in treating severe aggressive outbursts in child and adolescent inpatients. The untoward effects included a severe manic episode in a 16-year-old, hypomania in a 10-year-old, and increased irritability, impulsivity, and aggressiveness and/or worsening of behavior in two subjects aged 14 and 15. Two 11-year-old boys developed EEG abnormalities, with sharp waves and spikes. One of these boys improved behaviorally but had his first two absence seizures in several years. The authors caution that patients must be monitored carefully for the development of adverse neuropsychiatric effects.
Carbamazepine and the Induction of Mania
Three additional cases of carbamazepine-induced mania have been reported in children (Reiss and O’Donnell, 1984; Myers and Carrera, 1989). Myers and Carrera speculated that when adverse behavioral effects such as irritability, insomnia, agitation, talkativeness, and prepubescent hypersexuality occur with carbamazepine administration, they may be symptoms of an unrecognized hypomania or mania.
P.283

Reports of Interest
Evans et al. (1987) reviewed the use of carbamazepine in treating children and adolescents with psychiatric disorders, including hyperkinesis, aggression, impulsivity, and emotional lability. They noted the lack of systematic, well-controlled studies. These authors also addressed the important issue of the behavioral toxicity of carbamazepine and noted that, in their clinical experience in treating hyperactive and conduct-disordered children with carbamazepine, untoward effects on mood and behavior such as irritability, aggressiveness, increased hyperactivity, emotional lability, angry outbursts, and insomnia commonly occurred and sometimes
P.284

resembled the target symptoms for which the drug was being prescribed. Although a few studies have appeared since then, many of their concerns remain valid today.
Carbamazepine in the Treatment of Children and Adolescents with Various Diagnoses
Although approved only for treating certain kinds of seizure disorders and neuralgias, carbamazepine has been used to treat many psychiatric disorders. In adults, perhaps the best known of these is the treatment of lithium-resistant bipolar disorder. There is evidence that carbamazepine has acute antimanic and antidepressive effects as well as longer-term prophylactic action in treating bipolar disorder (Post, 1987). Post notes that more severe and dysphoric mania and a rapidly cycling course, variables associated with a poor response to lithium, appear to correspond to better responses to carbamazepine. It has been suggested that the efficacy of carbamazepine in psychiatric disorders may be secondary to its hypothesized ability to inhibit limbic system kindling. Kessler et al. (1989) reported that three psychotic adults improved markedly when carbamazepine was substituted for their neuroleptic medication and suggested criteria to identify effectively ill patients who may have a primary or superimposed organic mood disorder and who might benefit from carbamazepine.
There are many reports, particularly in the European literature, of the use of carbamazepine on an open basis to treat children and adolescents with psychiatric disorders; however, few placebo-controlled studies have been published. In 1988, Pleak et al. noted that carbamazepine was so frequently chosen to treat aggressive children and adolescents who did not respond satisfactorily to standard treatments that it had become “somewhat of a ‘vogue’ medication” (p. 502).
Remschmidt (1976) reviewed data from 28 clinical trials (seven double-blind and 21 open studies) with a total of > 800 nonepileptic child and adolescent subjects who were treated with carbamazepine. Positive clinical results were found for target symptoms of hyperactivity or hypoactivity, impaired concentration, aggressive behavioral disturbances, and dysphoric mood disorders. In addition to these behavioral effects, Remschmidt suggested that these patients experienced positive mood changes, increased initiative, and decreased anxiety.
Groh (1976) reported on 62 nonepileptic children treated with carbamazepine for various abnormal behavioral patterns. Of the 27 who showed improvement, most had a “dysphoric or dysthymic syndrome,” the most important features of which were emotional lability and moodiness, which were thought to cause most of the other behavioral abnormalities.
Kuhn-Gebhart (1976) reported symptom improvement in a large number of nonepileptic children who were treated with carbamazepine for a wide variety of behavioral disorders. The author reported that 30 of the last 50 patients treated showed good or very good responses, 10 had discernible improvement, 9 had no change in behavior, and 1 deteriorated. The author noted that the more abnormal the EEGs of these nonepileptic patients are in general, the better the response; that many of the good responders came from stable homes; and that poorer results were more frequent in subjects from unfavorable homes.
Puente (1976) reported an open study in which carbamazepine was administered to 72 children with various behavioral disorders who did not have evidence of neurologic disease. Fifty-six children completed the study. The usual optimal dose was 300 mg/day (range, 100 to 600 mg/day). Carbamazepine was given for an average of 12 weeks (range, 9 to 23 weeks). Twenty symptoms were rated on a severity scale at the beginning and end of the treatment. Individual symptoms were present in as many as 55 and in as few as 2 of the 56 children. Over the
P.285

course of treatment, a decrease in symptom expression of 70% or more occurred in 17 of 20 symptoms in at least 60% of the subjects. Interestingly, all 6 children (100%) with night terrors responded positively, as did 16 (94%) of the 17 children with other sleep disturbances. Anxiety, present in 47 children, improved in 34 (72%). Enuresis improved in 8 of 9 children (89%), and aggressiveness, present in 46 children, improved in 32 (70%). The most frequent untoward effects were transient drowsiness (20%), nausea and vomiting (4%), and urticaria (4%).
Carbamazepine in the Treatment of Juvenile-Onset Bipolar I Disorder
Woolston (1999) reported three cases diagnosed with bipolar I disorder whom he treated successfully with carbamazepine. One case, a 16-year-old female, had experienced several cycles of mania followed by depression that had been managed with various neuroleptics and lithium for approximately 4 years. The patient was noncompliant with lithium at least three times, resulting in manic episodes within a month that were followed by severe depression. Following the last of these episodes, she was begun on carbamazepine, 150 mg/day, which was increased to 300 mg/day 3 weeks later and continued at that dosage. Her serum carbamazepine level was 7 μg/mL. The patient became euthymic within 3 weeks and remained so on 300 mg/day of carbamazepine over the next 4 years, with the exception of three brief hypomanic episodes that responded to the addition of a short course of haloperidol 1 mg/day. No untoward effects were reported and blood counts and liver function remained within normal limits.
A 14-year-old male who had been treated for his mania with lithium for approximately 2 years was seen after he discontinued his lithium because it made him tired and dysphoric and he subsequently developed another manic episode. Carbamazepine was begun at a dose of 100 mg/day and increased to 200 mg/day, 5 days later. His mania improved significantly within 15 days and he did not experience the unpleasant symptoms he associated with lithium. His carbamazepine serum level was 8 μg/mL. He remained euthymic on carbamazepine 200 mg/day for the next 3 years with serum levels ranging from 6 μg/mL to 9 μg/mL. No untoward effects were reported and blood counts and liver chemistries remained normal throughout his treatment.
The third case was a 12.5-year-old girl also diagnosed with spastic cerebral palsy and mild mental retardation. She was treated briefly with risperidone for persistent euphoric mood, decreased need for sleep, and intermittent hallucinations. After 3 weeks, she had increased manic symptoms with flight of ideas, pressured speech, motor restlessness, and nearly continuous visual hallucinations with poor reality testing. Risperidone was discontinued and carbamazepine 100 mg/day was begun. After 1 week she showed improvement in sleep and reality testing and no untoward effects. Carbamazepine was then increased to 200 mg/day. Six days later, hallucinations had totally remitted, she was euthymic, had no evidence of a thought disorder, and had her normal sleep pattern. Her serum carbamazepine level was 8 μg/mL. The patient was continued on maintenance carbamazepine, 200 mg/day. Over the next 2 years she developed two brief periods of hypomania, both of which responded rapidly to an additional 50 mg of carbamazepine. She remained euthymic on her final maintenance daily dose of 300 mg of carbamazepine.
Carbamazepine in the Treatment of Children Diagnosed with Conduct Disorder
Kafantaris et al. (1992) reported an open pilot study in which ten children (nine male, one female; age range, 5.25 to 10.92 years; mean, 8.27 years) diagnosed with conduct disorder and hospitalized for symptoms of explosive aggressiveness
P.286

were treated with carbamazepine. Five of the subjects had previously failed to respond to a trial of lithium. After a 1-week baseline period, carbamazepine was administered in three divided doses, beginning at a total of 200 mg/day and titrated to a maximum of 800 mg/day, or a serum level of 12 μg/mL over a period of 3 to 5 weeks. Optimal dose range was 600 to 800 mg/day (mean, 630 mg/day) with serum levels from 4.8 to 10.4 μg/mL (mean, 6.2 μg/mL). Target symptoms of aggressiveness and explosiveness declined significantly on all measures compared with baseline ratings. On the Global Clinical Consensus Ratings, four subjects were rated as markedly improved, four as moderately improved, one as slightly improved, and one as not improved. Three of the lithium nonresponders showed marked improvement, and one showed moderate improvement; the fifth did not respond to either drug. Untoward effects during regulation and at optimal dose included fatigue (2 of 10 cases), blurred vision (2 of 10), and dizziness (1 of 10). Untoward effects above optimal dose included diplopia (2 of 10), mild ataxia (2 of 10), mild dysarthria (1 of 10), headache (2 of 10), and lethargy (1 of 10). One child experienced worsening of preexisting behavioral symptoms and loosening of associations, which were thought to be manifestations of behavioral toxicity. Overall, the untoward effects were transient and were decreased to tolerable levels or eliminated by reduction in dose. White blood cell counts remained within normal limits, although four children had reductions from baseline determinations.
Cueva et al. (1996) reported a 9-week, double-blind, placebo-controlled study comparing carbamazepine and placebo in 22 children (20 males, 2 females; mean age, 8.97 years; age range, 5.33 to 11.7 years) who were diagnosed with conduct disorder, solitary aggressive type, by DSM-III-R (APA, 1987) criteria and hospitalized for treatment-resistant aggressiveness and explosiveness. Thirty-eight children who met protocol criteria entered the initial 2-week placebo washout period. At the end of this period, 14 were eliminated because they no longer met study or aggression criteria. Of the 24 subjects who entered the treatment phase of the study, 13 were assigned to carbamazepine and 11 to placebo; 22 subjects completed the study. Efficacy was measured using the Children’s Psychiatric Rating Scale (CPRS), the NIMH Clinical Global Impressions-Severity (CGI-S) and, CGI-Improvement (CGI-I) scales, the Overt Aggression Scale (OAS), and the Global Clinical Judgments (Consensus) Scale, with a blind rating by all clinical staff occurring just before the code is broken. Medication was dispensed in two capsules given three times daily throughout the study. Carbamazepine was initiated at a dose of 200 mg/day and increased over a 2-week period in predetermined steps of 200 mg/dose to a maximum of 1,000 mg/day or until therapeutic effects were observed or untoward effects prevented further increase. Mean optimal dose of carbamazepine was 683 mg/day (range, 400 to 800 mg/day), with a mean serum carbamazepine level of 6.81 μg/mL (range, 4.98 to 9.1 μg/mL), for the 11 subjects for whom values were available.
The results showed no significant differences in the clinical improvement of aggression between carbamazepine and placebo on any of the rating scales. Both groups improved on the aggression factor of the CPRS over time and both improved similarly on the Global Clinical Judgments (Consensus) Scale as rated by clinical staff. Carbamazepine treatment resulted in significantly more untoward effects than placebo. Twelve of the 13 subjects on carbamazepine reported a total of 57 untoward effects, whereas only 6 of 11 subjects on placebo reported six untoward effects. Two subjects on carbamazepine developed marked leukopenia with between 2,000 and 3,000 WBC/mm3, and four developed moderate leukopenia
P.287

with between 3,000 and 3,500 WBC/mm3; one subject on placebo also developed moderate leukopenia. The leukopenia was transient in all seven cases. Other untoward effects experienced by subjects on carbamazepine included dizziness (N = 7, 54%), rash (N = 6, 46%), headache (46%), diplopia (N = 5, 38%), drowsiness (N = 4, 31%), nausea (31%), ataxia (N = 3, 23%), and vomiting (23%) (Cueva et al. 1996).
Carbamazepine in the Treatment of Children and Adolescents with Symptoms of ADHD
Silva et al. (1996) searched the world literature for reports in which carbamazepine was used to treat children and adolescents with behavioral problems and hyperactivity. Twenty-nine such reports were located; ten of them provided information suitable for a meta-analysis. Seven studies, with a total of 189 patients, were open; 70% of subjects experienced at least a marked improvement in target symptoms (significance ranged from P < .001 to P = .05). There was a significant correlation between longer treatment and positive outcome. In the three double-blind studies, 53 subjects were assigned to carbamazepine and 52 to placebo. Thirty-eight (72%) subjects on carbamazepine and 14 (27%) subjects on placebo were rated as moderately to markedly improved. Meta-analysis of these three studies found carbamazepine significantly (P = .018) more efficacious than placebo in diminishing target symptoms. The most common untoward effects in the studies reviewed were sedation and rash. The authors concluded that carbamazepine merited further study as a possible second-line treatment for children and adolescents with ADHD that is not responsive to stimulant medication or when stimulant medication cannot be tolerated.
Phenytoin, Diphenylhydantoin (Dilantin)
Contraindications for Phenytoin Administration
Known hypersensitivity to the phenytoin or a related drug is a contraindication.
Interactions of Phenytoin with Other Drugs
Acute alcohol intake may increase serum phenytoin levels, whereas chronic alcohol use may decrease levels.
Tricyclic antidepressants may precipitate seizures in susceptible patients, necessitating increased phenytoin doses.
Specific drugs have been reported to increase, decrease, or either increase or decrease phenytoin levels. Obtaining serum phenytoin levels may help clarify the situation when necessary. Some drugs that increase phenytoin levels are alcohol (when acutely ingested), benzodiazepines, phenothiazines, salicylates, and methylphenidate. Some drugs that decrease phenytoin levels are carbamazepine, alcohol (with chronic abuse), and molindone.
Interactions of phenytoin and phenobarbital, valproic acid, and sodium valproate are unpredictable, and serum levels of the drugs involved may either increase or decrease.
Use of Phenytoin in Child and Adolescent Psychiatry
Reports of Interest
Three double-blind, placebo-controlled studies that treated children and adolescents with phenytoin (diphenylhydantoin) reported that it was not significantly better than placebo.
Lefkowitz (1969) reviewed some of the earlier literature in which phenytoin was administered, primarily on an open basis, to nonepileptic children with psychiatric
P.288

disorders with discrepant results. Lefkowitz compared the efficacy of placebo and phenytoin in treating disruptive behavior in male juvenile delinquents (mean age, 14 years, 11 months; range, 13 years to 16 years, 3 months) in a residential treatment center. Each group contained 25 subjects. Phenytoin or placebo was administered in doses of 100 mg twice daily for 76 days. Both groups showed marked reductions in disruptive behavior. Phenytoin, however, was not significantly better than placebo on any of the 11 behavioral measures. In fact, placebo was significantly more efficacious than phenytoin in diminishing distress, unhappiness, negativism, and aggressiveness. The author suggested that mild toxic effects of phenytoin, such as insomnia, irritability, quarrelsomeness, ataxia, and gastric distress, may have accounted for the superiority of placebo.
Looker and Conners (1970) administered phenytoin to 17 children and adolescents (mean age, 9.1 years; range, 5.5 years to 14.5 years) who had severe temper tantrums and suspected minimal brain dysfunction. Eleven subjects had normal EEGs, three had mildly abnormal EEGs, and three had abnormal EEGs, but no subject had clinical seizures. Subjects were placed on a 9-week, double-blind, placebo-controlled, crossover protocol, and phenytoin was titrated to achieve blood levels of at least 10 μg/mL. Twelve of the 13 subjects who had final blood levels determined had adequate levels to suppress epileptic discharge. Scores on the Continuous Performance Test, the Porteus Maze Test, parent questionnaires for all subjects, and school questionnaires for 11 subjects showed no statistically significant differences between phenytoin and placebo. The authors noted, however, that some individual subjects appeared to respond positively and rather dramatically to phenytoin.
Conners et al. (1971) treated 43 particularly aggressive or disturbed delinquent males (mean age, 12 years; range, 9 to 14 years) living in a residential training school with phenytoin (200 mg/day), methylphenidate (20 mg/day), or placebo administered for 2 weeks in a double-blind protocol. Although the authors noted some limitations in their study, they found no significant difference between drugs and placebo on ratings by cottage parents, teachers, clinicians, and scores on the Rosenzweig Picture Frustration Test and Porteus Maze Test.
Overall, although there are individual patients without seizure disorder who appear to benefit from phenytoin, as yet there is no convincing evidence from double-blind placebo-controlled studies attesting to the effectiveness of phenytoin prescribed for psychiatric symptoms.
Gabapentin (Neurontin)
Gabapentin is an anticonvulsant whose mechanism of action is unknown. Gab-apentin is indicated as adjunctive therapy in the treatment of partial seizures with and without secondary generalization in adults with epilepsy.
Pharmacokinetics of Gabapentin
Gabapentin is not significantly metabolized in humans. It is eliminated unchanged by renal secretion, which is directly proportional to creatinine clearance. Half-life is 5 to 7 hours, and food has no effect on its absorption or excretion. Bioavailability of gabapentin decreases with dose, with a greater percentage of lower doses being available; at doses of approximately 600 mg/day and higher, it stabilizes at approximately 60% of the dose being available.
P.289

Contraindications for the Administration of Gabapentin
Gabapentin is contraindicated for patients with known hypersensitivity to the drug.
Interactions of Gabapentin with Other Drugs
There are no significant interactions of gabapentin with other commonly used antiepileptic drugs.
Untoward Effects of Gabapentin
The most common untoward effects reported somnolence, dizziness, ataxia, fatigue, and nystagmus. Many other effects have been reported.
Lamotrigine (Lamictal)
P.290

Lamotrigine is an antiepileptic drug of the phenyltriazine class and is chemically unrelated to antiepileptic drugs currently in use. Its mechanism of action is unknown. Lamotrigine is approved for use in the adjunctive treatment of partial seizures and in the adjunctive treatment of the generalized seizures of Lennox-Gastaut syndrome in patients ≥ 2 years old. Lamotrigine is also approved for use in adult patients diagnosed with partial seizures who are being treated with a single hepatic enzyme-inducing antiepileptic drug (EIAED; e.g., carbamazepine, phenytoin, phenobarbital, primidone, or valproate) to convert them to monotherapy with lamotrigine. Lamotrigine has also been approved for the maintenance treatment of bipolar I disorder to decrease the frequency of mood episodes (depression, amnia, hypomania, mixed episodes) in patients at least 18 years old who are being treated for acute mood episode with standard therapy. Lamotrigine has not been proved effective in treating acute mood episodes.
Safety and efficacy in patients under 16 years of age (except for Lennox-Gastaut syndrome) have not been established. A boxed warning notes that serious rashes, including Stevens-Johnson syndrome, requiring hospitalization and discontinuation of lamotrigine occur in approximately 8/1,000 (0.8%) of patients under 16 years of age. Because the rate of serious rash is greater in pediatric patients than in adults (approximately threefold greater), it bears emphasis that due to the preceding considerations, currently lamotrigine cannot be recommended for use in child and adolescent psychiatry, although its use as an adjunctive mood stabilizer is being explored.
Oxcarbazepine (Trileptal)
Oxcarbazepine is an antiepileptic drug; pharmacologic activity is exerted primarily through its 10-monohydroxy metabolite (MHD). Its mechanism of action is unknown; however, in vitro studies have indicated that oxcarbazepine and MHD produce blockade of voltage-sensitive sodium channels resulting in stabilization of hyperexcited neural membranes.
Pharmacokinetics of Oxcarbazepine
Oxcarbazepine taken orally is completely absorbed and extensively metabolized to MHD. The half-life of oxcarbazepine is approximately 2 hours and the half-life of MHD is approximately 9 hours. Food does not appear to affect the bioavailability of either oxcarbazepine or MHD. The median peak plasma level is 4.5 hours (range, 3 to 13 hours) for film-coated tablets and 6 hours for the oral suspension preparation; both preparations have similar bioavailability. Steady-state plasma concentrations are achieved in 2 to 3 days when a given dose is administered twice daily. Clearance of oxcarbazepine and its metabolites is primarily (over 95%) through the kidneys. Clearance in children under 8 years of age is 30% to 40% greater than in older children and adults, and in a controlled clinical trial such patients had the highest maintenance doses.
Contraindications for the Administration of Oxcarbazepine
Hypersensitivitity to oxcarbazepine or its components is a contraindication to the administration of oxcarbazepine. Approximately 25% to 30% of patients who had hypersensitivity reactions to carbamazepine are likely to do so with oxcarbazepine, hence they should be asked about any such prior exposure.
Interactions of Oxcarbazepine with Other Drugs
Oxcarbazepine can inhibit CYP2C19 and induce CYP3A4 and CYP3A5, which can potentially significantly effect plasma concentrations of other drugs. Drugs that induce cytochrome P450, including some other antiepileptic drugs, can result in decreases in plasma levels of oxcarbazepine and MHD. Plasma levels of phenytoin increased by up to 40% when oxcarbazepine was give in doses > 1,200 mg/day, however, phenobarbital levels increased by only 15%. Carbamazepine, phenytoin, and phenobarbital, which are all strong inducers of cytochrome P450 enzymes decreased the plasma level of MHD by 29% to 40%.
Coadministration of oxcarbazepine with an oral hormonal contraceptive de-creased plasma concentrations of ethinylestradiol and levonorgestrel, which may decrease the effectiveness of the contraceptive.
P.292

Untoward Effects of Oxcarbazepine
The most common untoward effects reported in pediatric patients being treated for a partial seizure disorder include: fatigue, vomiting, nausea, headache, somnolence, dizziness, ataxia, nystagmus, diplopia, vision abnormalities, and emotional liability. One manufacturer noted that approximately 9.2% (14) of 152 pediatric patients who were treated with oxcarbazepine but had not been treated previously with antiepileptic drugs discontinued the drug because of untoward effects. Although relatively infrequent as total untoward effects (< 1%), “rash” was responsible for 5.3% (8) and “maculopapular rash” for 1.3% (2) of those discontinuing. In a second group of 456 pediatric patients who were being treated with oxcarbazepine as a monotherapy or adjunctive therapy and who had been previously treated with antiepileptic drugs, 11% (50) discontinued the drug because of untoward effects. Patients discontinued for the following reasons: somnolence 2.4% (11), vomiting 2.0% (9), ataxia 1.8% (8), diplopia 1.3% (6), dizziness 1.3% (6), fatigue 1.1% (5), and nystagmus 1.1% (5).
Oxcarbazepine has a reduced risk for leukopenia, rashes, drug interactions, and enzyme autoinduction compared with carbamazepine (see references sited by Teitelbaum, 2001).
Hyponatremia
Clinically significant hyponatremia (sodium < 125 mmol/L) may occur during treatment with oxcarbazepine. This usually occurs within 3 months of initiation of therapy but may also occur after over a year of treatment. In 14 studies with a total of 1,524 patients, 38 (2.5%) developed a sodium of < 125 mmol/L sometime during treatment compared with no patients on placebo or active control (other antiepileptic drugs). Symptoms that may reflect hyponatremia such as nausea, malaise, headache, lethargy, confusion, obtundation, or increase in seizure frequency or severity should be the cause for determining sodium plasma levels. Monitoring of sodium levels during treatment should be considered.
Reports of Interest
Teitelbaum (2001) reported that a 6-year-old female diagnosed with bipolar I disorder who had been hospitalized four times in the preceding year for out-of-control behavior with extreme aggression toward others and destruction of property and who did not have an adequate clinical response to trials with lithium, lamotrigine (some benefit but discontinued because of rash), valproate, gabapentin, clonidine, guanfacine, risperidone, olanzepine, quetiapine, fluoxetine, and methylphenidate (worsened symptoms), improved significantly when oxcarbazepine was added to her medication regime of 3 months, which consisted of lithium carbonate 150 mg three times daily (0.7 mEq/L serum level) and guanfacine 0.5 mg three times daily. Oxcarbazepine was added to these medications at an initial dose at 150 mg twice daily with full mood stabilization after 6 weeks. Three months after the initiation of oxcarbazepine, the lithium dose was decreased to 150 mg twice daily; good control continued to be maintained for 7 months. The author noted that the girl experienced some transient mild exacerbation of symptoms during a school vacation and again during a viral illness, but was being maintained on lithium 150 mg twice daily; guanfacine 0.5 mg twice daily and oxcarbazepine 150 mg twice daily (Teitelbaum, 2001).
Staller et al. (2005) conducted a retrospective chart review of 14 outpatients (age range, 6 to 17 years; 6 males and 8 females) to whom oxcarbazepine was prescribed to address their moderate-to-severe problems with anger, irritability, and aggression but many subjects had additional symptoms including depression, mania, anxiety, disruptiveness, oppositionality, and psychosis. Subjects were rated as moderately ill (6), markedly ill (7); and severely ill (1) on the Clinical Global Impressions-Severity scale. Subjects’ Axis I diagnoses, including multiple diagnoses in 11, included bipolar disorder (5); other mood disorders (3); ADHD (4); disruptive behavior disorder (3); and PDD spectrum disorders (2). Ten (71.4%) of the subjects
P.294

had failed to respond adequately to prior drug trials. During treatment with oxcarbazepine, three subjects received oxcarbazepine only; three subjects were already receiving one additional medication; and seven subjects were receiving two additional medications and one subject was receiving three additional medications; the medications included atypical antipsychotics (7); SSRIs (4); stimulants (2), alpha agonists (2) antihistamines (2) beta blocker (1) and valproate (1). The average daily dose of oxcarbazepine was 878 mg with a range from 600 to 1,800 mg/day. Duration on oxcarbazepine averaged 9.8 months with a range of 0.5 to 30 months. Clinical improvement was rated on the CGI-I scale. The ratings were “much improved” in seven subjects; “minimally improved” in two and unchanged in five. AEs including dizziness, muscle aches, and tremors resulted in discontinuation in two (14%) of the subjects, but they were not considered to be serious AEs.
Topiramate (Topamax)
Topiramate, an antiepileptic drug, is a sulfamate-substituted monosacharide. The mechanisms responsible for its antiepileptic and migraine prophylaxis effects have not been elucidated; however, preclinicial studies suggest that at clinically effective concentrations, topiramate blocks voltage-dependent sodium channels, augments the activity of the neurotransmitter gamma-aminobutyrate at some subtypes of the GABA-A receptor, antagonizes the AMPA/kainate subtype of the glutamate receptor, and inhibits the carbonic anhydrase enzyme, particularly isoenzymes II and IV (package insert).
Pharmacokinetics of Topiramate
The bioavailability of topiramate is not affected by food. Peak plasma concentrations occur in approximately 2 hours. Mean plasma elimination half-life is 21 hours after single or multiple doses and steady state occurs in approximately 4 days at a given dose. Topiramate is not extensively metabolized and approximately 70% is eliminated unchanged through the kidneys. Pediatric patients, aged 4 to 17 years, have approximately a 50% higher clearance than adults. Consequently such patients have a shorter elimination half-life than adults and their plasma concentration of topiramate may be lower than that of adults receiving the same dose.
Contraindications for the Administration of Topiramate
Topiramate is contraindicated in patients with a history of sensitivity to topiramate or any of the components included in the pill. Clearance may be significantly reduced in patients with renal or hepatic impairment.
Interactions of Topiramate with Other Drugs
Hyperammonemia with or without encephalopathy has been associated with the combined use of topiramate and valproic acid in patients who have not developed these symptoms when treated with either drug alone. Patients who develop symptoms such as acute alterations in the level of consciousness or cognitive functioning in combination with lethargy or vomiting, which may be associated with hyperammonemic encephalopathy should have their serum ammonia levels determined.
P.295

Topiramate can decrease the AUC and maximum serum concentration of Lithium by up to 20%.
Many other drug interactions are possible; see package insert.
Untoward Effects of Topiramate
Topiramate is associated with metabolic acidosis, which if chronic and untreated may cause osteomalacia/rickets and may reduce growth rate and maximal stature in pediatric patients. Treatment-emergent adverse events in children in the age-group 10 through 16 years who were being treated with monotherapy (400 mg/day) for epilepsy that occurred with an incidence of at least 5% and which were more frequent than those at lower doses (50 mg/day) included: fever (9%), paresthesias (16%), diarrhea (11%), weight loss (21%), anorexia (14%), mood problems (11%), difficulty with concentration/attention (9%), and alopecia (5%).
Report of Interest
DelBello et al. (2005) reported a 4-week, multicenter, randomized, double-blind, placebo-controlled, parallel group trial of topiramate in treating mania in 56 children and adolescent means age 13.8 ± 2.56 years, age range 6 to 17 years, who were diagnosed with bipolar disorder type I; 33 subjects (58.9%) had codiagnoses of ADHD. The primary outcome measure of efficacy was the Young Mania Rating Scale (YMRS) scores at baseline and 4, 7, 14, 21, and 28 days of treatment. Secondary outcome measures included baseline and weekly scores on the Clinical
P.296

Global Impressions-Improvement Scale (CGI-I), the Brief Psychiatric Rating Scale for Children (BPRS-C), the Children’s Global Assessment Scale (C-GAS) and the Childrens’ Depression Rating Scale (CDRS).
Although the study was designed to enroll approximately 230 subjects, only 56 were enrolled; 29 subjects were assigned to topiramate and 27 to placebo. This was because a study of topiramate in adult subjects hospitalized with acute mania showed no benefit of topiramate over placebo and hence it was thought topiramate would be unlikely to benefit younger subjects and that it was not ethical to continue and the study was stopped at that time. Because of this, the study was inconclusive as it was not adequately powered to detect treatment differences. The baseline YMRS score for the topiramate group was 31.7 ± 5.53 and it was decreased by -9.7 ± 9.65 at endpoint. The baseline YMRS score for the placebo group was 29.9 ± 6.01 and it was decrease by -4.7 ± 9.79 at endpoint, less than one-half the improvement seen in the topiramate group. However, there was no significant group difference at any visit for the change in YMRS score, the total BPRS-C score, total CDRS score, or CGAS score.
Treatment-emergent adverse events occurring in > 10% of subjects and greater for topiramate than for placebo included decreased appetite (27.6% vs. 0%), nausea (24.1% vs. 0%), diarrhea (13.8% vs. 7.4%), paresthesia (13.8% vs. 3.7%), somnolence (13.8% vs. 3.7%), insomnia (10.3% vs. 3.7%) and rash (10.3% vs. 3.7%). Mean change in body weight from baseline to endpoint was significantly different, with the topiramate group losing a mean of -1.76 ± 2.03 kg and the placebo group gaining a mean of 0.95 ± 1.45 kg (P < .001). No subject experienced a serious adverse event. The authors noted that this study suggests that topiramate may be effective in treating bipolar disorder type I in children and adolescents, although it was not so in adults, and that adequately powered, controlled studies in this age-group are needed to determine whether this is the case.