Article Text


Antiepileptic drugs in psychiatry
  1. F Ovsiew
  1. Correspondence to:
 F Ovsiew
 Department of Psychiatry MC 3077, University of Chicago Hospitals, 5841 S. Maryland, Chicago, IL 60637, USA;

Statistics from

The potential risks of AEDs widely used in psychiatric patients are not fully understood

Antiepileptic drugs (AEDs) have been widely used by psychiatrists to treat disorders other than epilepsy for two decades. Here I survey those uses and selectively review the side effects of AEDs in the psychiatric context.

AEDs are now so widely used for non-epileptic indications that the appellation “anti-epileptic drug” may verge on mislabelling (to paraphrase David Healy’s comment about serotonin-reuptake inhibitors and “antidepressants”). Nonetheless, no better designation is available, and I refer to these drugs as AEDs. Oddly, the newer AEDs may prove be more useful in psychiatric illness than in epilepsy, where their greater efficacy than the older agents has not yet been conclusively demonstrated.1

An assumption that the mechanisms of action of AEDs are the same in epilepsy and in psychiatric disorders would go beyond the evidence. Activity against kindling has been proposed to explain the psychotropic effect,2 but this seems to account for neither the spectrum of psychiatric benefits nor the range of beneficial drugs. A division of psychotropic effects into “sedating” (mediated by GABAergic action) and “activating” (mediated by antiglutamatergic action)3 likewise may fail to capture the diversity of psychotropic “principles” by which the diverse category of AEDs exerts psychotropic effects.4

Although I review scientific data regarding the use of AEDs in psychiatry, a non-scientific factor, of uncertain power, may partly account for the rise in prescription of AEDs. The interest of pharmaceutical firms in expanding the use of their products and their willingness, at least in some instances, to use improper means have recently been highlighted by the guilty plea of a major company in a case involving marketing an AED for off-label indications.5 Since the scientific evidence in the realm reviewed is far from complete, the clinician needs to be alert for non-scientific influences on practice.


Lithium remains the agent for bipolar mood disorder with the best documentation of efficacy across the phases of the disorder, depressed and manic, acute and maintenance.6 However, AEDs are now commonly used in the treatment of bipolar mood disorder. Limited data suggest that over the course of the 1990s the use of AEDs approximately doubled, while the use of lithium fell below that of AEDs.7 At least three findings account for this development. First, certain syndromes of bipolar mood disorder respond poorly to lithium. Patients responding poorly to lithium include those with mixed bipolar states (states that include a mixture of manic and depressive features, often presenting in atypical fashion, for example as characterological irritability) and patients with rapid cycling (defined as having more than four episodes of mood disorder in a year). Second, bipolar mood disorder may be much more common than assumed: though conventional epidemiology has it that unipolar major depression is several times more common than bipolar mood disorder, recent studies suggest that half of patients presenting with major depression may have bipolarity.8 Third, patients with bipolar disorder often suffer from sub-syndromal depressive symptoms during much of their lives, but the use of antidepressants in these patients risks evoking mania or exacerbating cycling.9

Thus agents with efficacy for bipolar depressive or mixed states or for lithium refractory patients are particularly welcome, and considerable high quality evidence—and much more anecdotal or uncontrolled evidence—supports the use of AEDs for these indications. Reports of psychotropic effects of valpromide (an amide derivative of valproic acid available in Europe but not in the United States) and carbamazepine emerged soon after their introduction.4 Divalproex (an enteric-coated formulation of valproic acid and sodium valproate, the most widely used preparation of valproic acid in the United States because it has fewer gastrointestinal side effects than valproic acid) and carbamazepine have now been shown to be effective in the treatment of acute mania, with a spectrum of action that includes the “non-classical” manic states of dysphoric mania and rapid cycling.3,10 Efficacy for prophylaxis is less securely supported by clinical trial evidence, although many clinicians employ these agents as well as the newer AEDs in this fashion. Evidence for antidepressant efficacy is modest. Oxcarbazepine appears, on limited evidence, to have similar efficacy to carbamazepine but better tolerability.11,12

Of the newer AEDs, lamotrigine has occasioned particular interest. Its efficacy for bipolar depression was foreshadowed by reports of efficacy against depressive symptoms in epileptic patients.13 Evidence of its acute and prophylactic effects against bipolar depression is now available, with efficacy against manic relapse in maintenance treatment also finding support.14 Whether it is effective as acute treatment for mania is a less pressing clinical matter, because the required slow titration precludes reliance on its use in many such circumstances. Topiramate is less well studied, but open trials indicate good tolerability as well as efficacy for non-classical mania.15 Levetiracetam is of considerable interest for bipolar disorder, with reports of benefit for mania, depression, and rapid cycling, but adequate data are not available.16,17 Limited uncontrolled data suggest utility of zonisamide and tiagabine for bipolar disorder.18,19 On present evidence, gabapentin does not appear effective for bipolar disorder.20,21

However, gabapentin—an attractive compound because of a favourable side effect profile and a lack of drug interactions—appears on limited evidence to be of considerable use in anxiety states, including social phobia, post-traumatic stress disorder, panic disorder, and opiate withdrawal.22–24 The role of other AEDs in the treatment of anxiety is less well studied.

Fortunately, electroconvulsive therapy can safely be undertaken in patients treated with AEDs (and indeed in patients with epilepsy).25,26 Often it will be necessary to reduce the dose or hold a dose prior to the treatment so as not to elevate the seizure threshold, especially later in a course of ECT when the seizure threshold has risen.


The potential use of AEDs in schizophrenia, the subject of many open trials, was highlighted by a blinded, placebo controlled study in which lamotrigine was shown to be effective (in combination with clozapine) in treatment refractory schizophrenic patients.27,28 The anti-glutamatergic action of lamotrigine is of particular interest because of a current focus on glutamatergic mechanisms in schizophrenia; lamotrigine via inhibition of glutamate release may be synergistic with clozapine, which is a glutamate antagonist.

The overlap of mood disorder, especially sub-syndromal bipolar mood disorder with rapid cycling and dysphoria or irritability, and borderline personality disorder (BPD) is a matter of active investigation.29 The problem of BPD is a major one, because of the morbidity associated with the disorder—a 10% lifetime suicide risk is estimated—and because of the patients’ intense utilisation of mental health resources. An empirically validated pharmacological approach to these patients is desirable; study data and clinical experience show roles for antipsychotics and antidepressants. Carbamazepine reduced impulsivity in early trials.30 In recent studies, divalproex and lamotrigine showed benefit for aggression, mood stability, and impulsivity even in the absence of a diagnosable mood disorder.31,32


AEDs have come to be widely used for agitation and aggression in the demented elderly and in children and adults with learning disability. A neuroprotective role for divalproex has even been suggested.33 In general, the data supporting the practice of treating agitation with AEDs are not of high quality; a recent Cochrane review found major methodological flaws in existing studies and could not support the use of divalproex in dementia.34 Further, clinicians need to be aware of the special risks of adverse behavioural effects of AEDs in the developmental disability population.35 Nonetheless, open trial data and clinical anecdotes continue to encourage clinicians to use AEDs as less toxic alternatives to antipsychotics in these populations.33,36


Weight gain produced by psychiatric drugs and the related “metabolic syndrome” of insulin resistance, dyslipidaemia, abdominal obesity, and elevated blood pressure emerged as a focus of psychiatric concern with the advent of the newer antipsychotic agents. However, weight gain is an important consideration with antidepressants, lithium, carbamazepine, and divalproex.37 The potential of divalproex to cause fatty liver (by virtue of its association with the metabolic syndrome, not by idiosyncratic hepatotoxicity) is an additional concern.38 Perhaps this latter risk extends to other agents that promote weight gain. The advent of AEDs that are weight neutral, such as lamotrigine and levetiracetam, or produce weight loss, such as topiramate and zonisamide, is thus an important development.39 Topiramate may find a specific use in the treatment of bulimia.40,41 Zonisamide has been the subject of an impressively favourable trial in which it was prescribed solely for weight loss.42 Especially in psychiatric patients who are on other psychotropic drugs that promote weight gain, these considerations are of substantial importance to patients’ general health and to their compliance with psychiatric regimens.


Reproductive endocrine disorders are common in epileptic women, but the role of AEDs in the dysfunction remains controversial.43,44 Evidence from several centres demonstrates that valproate produces elevated androgen levels,45–48 although the elevations may not correlate with symptomatic measures of androgen excess (the polycystic ovary syndrome).45,47 This effect appears not to depend solely on valproate induced weight gain. Of particular interest is whether women with epilepsy and women with bipolar mood disorder are equally affected when valproate is administered. In one of the few direct tests of this question, valproate produced elevated androgens (compared with lithium treated bipolar patients) in both epileptic and bipolar women, but produced more clinical symptoms of hyperandrogenism in the epileptics than in the bipolars.49 A differential effect is biologically plausible because of the known effects of epilepsy itself on reproductive endocrine function, so research on epileptic patients cannot easily be extrapolated to psychiatric populations. In view of the wide use of valproate in women with psychiatric disorders—not to mention female adolescents and children—such focused research is urgently required.

Reproductive endocrine disorders associated with AEDs are not confined to women. In epileptic men, AEDs including carbamazepine, oxcarbazepine, and valproate were found to produce alterations of sex hormones and of sperm morphology and function.50,51 This issue is entirely unstudied in male patients with psychiatric disorders.


The literature on the cognitive, emotional, and behavioural impact of AEDs on epileptic patients has been held up as a warning to psychiatrists about the potential for adverse effects in psychiatric patients.52 This potential was recognised early in the psychiatric literature when carbamazepine used in borderline personality disorder was reported to cause depression.53 Such effects might be particularly likely because of vulnerability indicated by psychiatric history,54–56 and might be hard to recognise because of concurrent primary psychiatric symptoms.

However, adverse behavioural effects in epileptic patients might be a consequence of interaction between the epileptic brain and the AED, so such effects might be less likely in psychiatric populations. As an example, topiramate has salient adverse cognitive effects in epileptics, including executive cognitive dysfunction and a distinctive anomia, but these effects do not seem to be as prominently mentioned in the limited literature on the psychiatric use of topiramate.15,57,58 This might be because of their subtle nature, or it might be because of psychiatrists’ inadequate sensitivity. On the other hand, since the occurrence of the anomia appears to be related to a left temporal seizure focus and that of depression and adverse cognitive effects to the presence of hippocampal sclerosis,56,59 and the occurrence of adverse behavioural effects may be critically related to abolition of seizures,60 such phenomena may simply be less common in patients without organic disease.61 While certainly the warning provided by the epilepsy literature must be taken seriously in psychiatry, only systematic observations of AED use in psychiatric patients will resolve the issue.

Similarly, reports of parkinsonism and cognitive impairment due to valproic acid, which have repeatedly appeared in the neurological literature, are almost completely absent from the psychiatric literature.62–64 Either psychiatrists have been oblivious to this important phenomenon, or the drug has different effects in epileptics and patients with mood disorder.


AEDs cause bone loss and an increased risk of fracture by multiple mechanisms, notably by reducing serum vitamin D and increasing bone turnover. Enzyme inducers and enzyme inhibitors share this side effect, but evidence on its occurrence with the newer AEDs is limited. For example, lamotrigine may not produce bone loss.65 The compelling evidence of risk has led experts to propose that epileptic patients on AEDs should be screened at least with a baseline bone density scan and annual measurements of serum vitamin D, calcium, and alkaline phosphatase and should be treated with supplemental calcium and vitamin D.65,66 The elevated risk of fractures is not related solely to trauma during convulsions; the risk of fracture not related to a seizure is also elevated several-fold in patients on AEDs. Thus the risk could reasonably be expected to be present in patients treated with the same agents for psychiatric indications, especially in the elderly or developmentally disabled patient with limited mobility and exposure to sunlight. However, to my knowledge evidence on this point is not available. The issue is virtually unmentioned in the psychiatric literature, and recommendations for screening and prophylaxis for bone loss in psychiatric patients have not been provided.67


A crucial aspect of the use of AEDs in psychiatric patients is their use when the psychiatric patient also has epilepsy. The prevalence of psychiatric illness in epilepsy is high, particularly depression, which is under-recognised but has a substantial impact on quality of life.68,69 As a first step in assessing such situations, consideration of adverse effects of the AEDs being used to treat epilepsy is appropriate.52 This caution applies particularly to the learning disability population, in which such side effects may be both more common and more difficult to recognise.35 The potential psychotropic effects of AEDs can, in part, guide the choice of agent for epilepsy; for example, the benefit of lamotrigine for depression has already been mentioned.13

When other psychotropic drugs are to be used in patients with epilepsy, consideration of drug interactions is necessary and complex. Certain antidepressants produce marked changes in AED levels, while others (such as citalopram, escitalopram, and sertraline) do not. De Leon70 recently provided a useful summary of the literature on dosage adjustment of antipsychotic drugs necessitated by concurrent administration of AEDs.


This review has surveyed the utility and selected aspects of the adverse effects of AEDs for psychiatric indications. Without doubt these agents are a welcome advance. That their use has, in some respects, gone beyond the systematic evidence for efficacy testifies to the difficulty and urgency of providing effective treatment to many psychiatric patients. (The use of AEDs in epilepsy routinely goes beyond the evidence, when drugs approved on the basis of add on trials are used as first line treatment or monotherapy.) That the potential risks of AEDs in psychiatric patients are not fully understood signals a need for research and for caution on the part of clinicians, for caution does not have to mean undertreatment.

The potential risks of AEDs widely used in psychiatric patients are not fully understood


View Abstract

Request permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.