Drug Treatment of Bipolar Disorder
Choice of drug can be difficult because all drugs have significant adverse effects, drug interactions are common, and no drug is universally effective. Selection should be based on what has previously been effective and well-tolerated in a given patient. If there is no prior experience (or it is unknown), choice is based on the patient’s medical history (vis-à-vis the adverse effects of the specific mood stabilizer) and the severity of symptoms.
For severe manic psychosis, in which immediate patient safety and management is compromised, urgent behavioral control usually requires a sedating 2nd-generation antipsychotic, sometimes supplemented initially with a benzodiazepine such as lorazepam or clonazepam 2 to 4 mg IM or po tid.
For less severe acute episodes in patients without contraindications (eg, renal disorders), lithium is a good first choice for both mania and depressive episodes. Because its onset is slow (4 to 10 days), patients with significant symptoms may also be given an anticonvulsant or a 2nd-generation antipsychotic.
For patients with depression, lamotrigine may be a good choice of anticonvulsant.
For bipolar depression, the best evidence suggests using quetiapine or lurasidone alone or the combination of fluoxetine and olanzapine.
Once remission is achieved, preventive treatment with mood stabilizers is indicated for all bipolar I patients. If episodes recur during maintenance treatment, clinicians should determine whether adherence is poor and, if so, whether nonadherence preceded or followed recurrence. Reasons for nonadherence should be explored to determine whether a change in mood stabilizer type or dosing would render treatment more acceptable.
As many as two thirds of patients with uncomplicated bipolar disorder respond to lithium, which attenuates bipolar mood swings but has no effect on normal mood.
Whether lithium or another mood stabilizer is being used, breakthroughs are more likely in patients who have mixed states, rapid-cycling forms of bipolar disorder, comorbid anxiety, substance abuse, or a neurologic disorder.
Lithium carbonate is started at 300 mg po bid or tid and titrated, based on steady-state blood levels and tolerance, to a range of 0.8 to 1.2 mEq/L. Levels should be drawn after 5 days at a stable dose and 12 h after the last dose. Target drug levels for maintenance are lower, about 0.6 to 0.7 mEq/L. Higher maintenance levels are more protective against manic (but not depressive) episodes but have more adverse effects. Adolescents, whose glomerular function is excellent, need higher doses; elderly patients need lower doses.
Lithium can cause sedation and cognitive impairment directly or indirectly (by causing hypothyroidism) and often exacerbates acne and psoriasis. The most common acute, mild adverse effects are fine tremor, fasciculation, nausea, diarrhea, polyuria, polydipsia, and weight gain (partly attributed to drinking high-calorie beverages). These effects are usually transient and often respond to decreasing the dose slightly, dividing the dose (eg, tid), or using slow-release forms. Once dosage is established, the entire dose should be given after the evening meal. This dosing may improve adherence. A beta-blocker (eg, atenolol 25 to 50 mg po once/day) can control severe tremor; however, some beta-blockers (eg, propranolol) may worsen depression.
Acute lithium toxicity is manifested initially by gross tremor, increased deep tendon reflexes, persistent headache, vomiting, and confusion and may progress to stupor, seizures, and arrhythmias. Toxicity is more likely to occur in the following:
Thiazide diuretics, ACE inhibitors, and NSAIDs other than aspirin may contribute to hyperlithemia. Lithium blood levels should be measured every 6 mo and whenever the dose is changed.
Long-term adverse effects of lithium include
Hypothyroidism, particularly when there is a family history of hypothyroidism
Renal damage involving the distal tubule (mainly in patients with a history of renal parenchymal disease)
Therefore, TSH levels should be monitored when lithium is started and annually thereafter if there is a family history of thyroid dysfunction or every other year for all other patients. Levels should also be measured whenever symptoms suggest thyroid dysfunction (including when mania recurs) because hypothyroidism may blunt the effect of mood stabilizers. BUN and creatinine should be measured at baseline, 2 or 3 times during the first 6 mo, and then once or twice a year.
Anticonvulsants that act as mood stabilizers, especially valproate and carbamazepine, are often used for acute mania and for mixed states (mania and depression). Lamotrigine is effective for mood-cycling and for depression. The precise mechanism of action for anticonvulsants in bipolar disorder is unknown but may involve gamma-aminobutyric acid mechanisms and ultimately G-protein signaling systems. Their main advantages over lithium include a wider therapeutic margin and lack of renal toxicity.
For valproate, a loading dose of 20 to 30 mg/kg is given, then 250 to 500 mg po tid (extended-release formulation can be used); target blood levels are between 50 and 125 μg/mL. This approach does not result in more adverse effects than does gradual titration. Adverse effects include nausea, headache, sedation, dizziness, and weight gain; rare serious effects include hepatotoxicity and pancreatitis.
Carbamazepine should not be loaded; it should be started at 200 mg po bid and be increased gradually in 200-mg/day increments to target levels between 4 and 12 μg/mL (maximum, 800 mg bid). Adverse effects include nausea, dizziness, sedation, and unsteadiness. Very severe effects include aplastic anemia and agranulocytosis.
Lamotrigine is started at 25 mg po once/day for 2 wk, then 50 mg once/day for 2 wk, then 100 mg/day for 1 wk, and then can be increased by 50 mg each week as needed up to 200 mg once/day. Dosage is lower for patients taking valproate and higher for patients taking carbamazepine. Lamotrigine can cause rash and, rarely, the life-threatening Stevens-Johnson syndrome, particularly if the dosage is increased more rapidly than recommended. While taking lamotrigine, patients should be encouraged to report any new rash, hives, fever, swollen glands, sores in the mouth and on the eyes, and swelling of the lips or tongue.
Acute manic psychosis is being increasingly managed with 2nd-generation antipsychotics, such as
In addition, evidence suggests that these drugs may enhance the effects of mood stabilizers after the acute phase.
Although any of these drugs may have extrapyramidal adverse effects and cause akathisia, risk is lower with more sedating drugs such as quetiapine and olanzapine. Less immediate adverse effects include substantial weight gain and development of the metabolic syndrome (including weight gain, excess abdominal fat, insulin resistance, and dyslipidemia); risk may be lower with the least sedating 2nd-generation antipsychotics, ziprasidone and aripiprazole.
For extremely hyperactive psychotic patients with poor food and fluid intake, an antipsychotic given IM plus supportive care in addition to lithium or an anticonvulsant may be appropriate.
Lithium use during pregnancy has been associated with an increased risk of cardiovascular malformations (particularly Ebstein anomaly). However, the absolute risk of this particular malformation is quite low. Taking lithium during pregnancy appears to increase the relative risk of any congenital anomaly by about 2-fold, a risk similar to the 2- to 3-fold increased risk of congenital anomalies associated with use of carbamazepine or lamotrigine and is substantially lower than the risk associated with use of valproate.
With valproate, risk of neural tube defects and other congenital malformations appears to be 2 to 7 times higher than that with other commonly used anticonvulsants. Valproate increases the risk of neural tube defects, congenital heart defects, genitourinary anomalies, musculoskeletal abnormalities, and cleft lip or palate. Also, cognitive outcomes (eg, IQ scores) in children of women who took valproate during pregnancy are worse than those with other anticonvulsants; risk appears to be dose-related. Valproate also appears to increase risk of attention-deficit/hyperactivity disorder and autism spectrum disorders (1).
Extensive study of the use of 1st-generation antipsychotics and tricyclic antidepressants during early pregnancy has not revealed causes for concern. The same appears to be true of SSRIs, except for paroxetine. Data about the risks of 2nd-generation antipsychotics to the fetus are sparse as yet, even though these drugs are being more widely used for all phases of bipolar disorder.
Use of drugs (particularly lithium and SSRIs) before parturition may have carry-over effects on neonates.
Treatment decisions are complicated by the fact that with unplanned pregnancy, teratogenic effects may already have taken place by the time practitioners become aware of the issue. Consultation with a perinatal psychiatrist should be considered. In all cases, discussing the risks and benefits of treatment with patients is important.
1. Tomson T, Battino D, Perucca E: Valproic acid after five decades of use in epilepsy: Time to reconsider the indications of a time-honoured drug. Lancet Neurol 15(2): 210–218, 2016. Epub ahead of print (2015 Dec 3). doi: 10.1016/S1474-4422(15)00314-2.