Intended for healthcare professionals


Risk for schizophrenia and schizophrenia-like psychosis among patients with epilepsy: population based cohort study

BMJ 2005; 331 doi: (Published 30 June 2005) Cite this as: BMJ 2005;331:23

Schizophrenia and Epilepsy: Untenable Link from Flawed Research Design

The recently published BMJ article by Qin et al, “Risk for
schizophrenia and schizophrenia-like psychosis among patients with
epilepsy: Population based cohort study” offers an untenable
conclusion.(1) The first sentence of the authors’ discussion asserts,
“…this large population study shows that people with a history of epilepsy
have nearly 2.5 times the risk of developing schizophrenia and nearly
three times the risk of developing a schizophrenia-like psychosis compared
with the general population.” While it is possible that the authors’
declaration of risk might have future validity, it is unsupportable based
upon their current research. An analysis of their work, partially funded
by the USA-based Stanley Medical Research Institute, ignores the effects
of antiepileptic medication-induced neuropsychiatric sequelae in an
epileptic cohort.

The oversight is significant. It is a problem addressed briefly by
Dinah KC Murray, London, in her previous response to the Qin article.(2)
The authors failed to account for the putative role of anticonvulsant
medications in the production of psychotic symptoms, including psychoses
that resemble schizophrenic-like disorders. This omission erroneously
assumes that anticonvulsant-treated epilepsy in Denmark, and elsewhere,
has the same natural course as unmedicated epilepsy. The evidence
regarding anticonvulsant-mediated adverse side-effects suggests otherwise.

In 1994, Srinivasan and Richens described a ‘schizophrenia-like
syndrome' in vigabartin.(3) That paper outlined at least three subtypes
of psychosis associated with vigabartin, an irreversible inhibitor of GABA
transaminase whose net effect is to enhance regional GABAergic
activity.(4) In similar fashion, but not to the same degree, all of the
major anticonvulsants list psychosis or psychotic-like symptoms in their
presentation of adverse effects.

The number one and number two best-selling anti-seizure drugs in the
United States (according to the 21 April 2005 story in the Wall Street
Journal, “FDA Request Reviews of Epilepsy Drugs-Paper”) dramatize the
problem.(5) Best-seller gabapentin (Neurontin-Pfizer), a GABAergic
anticonvulsant similar to vigabartin but whose different mechanism of
action also enhances GABA activity, induces comparable problems. The
adverse event data for gabapentin, as observed and recorded during all
clinical trials in adults and adolescents with epilepsy, according to 2004
Physicians’ Desk Reference(PDR), pages 2562-3, reveals the following
neuropsychiatric side-effects: Apathy, hallucinations, agitation,
depression, paranoia, depersonalization, euphoria, emotional lability,
nervousness, abnormal thinking, encephalopathy, suicidal gestures and

The second most widely prescribed anticonvulsant in the United
States, topiramate (Topamax-Ortho-McNeil; division of Johnson and
Johnson), has a mechanism of action similar to vigabatrin and gabapentin,
suggesting that it also potentiates the activity of the inhibitory
neurotransmitter, GABA. The 2004 PDR(pages 2486-88) lists its adverse
neuropsychiatric events as follows: Psychomotor slowing, impaired
concentration, encephalopathy, personality disorder, speech or language
problems, confusion, exacerbation of mood disturbances(including
irritability and depression), hallucinations, aggressive reaction,
paranoid reaction, delusions, paranoia, manic reactions, suicidal
gestures, suicide attempts and psychosis.(6)

A similar picture of adverse neuropsychiatric side-effects has been
reported with all of the other major anticonvulsants, including
carbamazepine, valproic acid, the oxazolidinediones, the succinimides, and
the hydantoins. In many cases, the co-administration of these
anticonvulsants increases the risk of neurotoxicity and the production of
psychotic symptoms.(4)

In order to substantiate their putative link between epilepsy and
schizophrenia-like psychosis, Qin et al. would have to affirm three
assumptions: (a) that medicated and unmedicated epileptic cohorts are
separable and have natural history courses that are comparable, (b) that
antiepileptic drugs, otherwise known to produce psychotic sequelae, do not
produce long-term clinical conditions similar to schizophrenia-like
psychosis, and, (c) that antiepileptic drugs, despite producing serious
mood and thought disturbances as side-effects, do not induce those
conditions in sufficient numbers in a given cohort to be addressed in
linkage studies. Since these assumptions have not been addressed, the
author’s conclusions are not valid.

Stefan P. Kruszewski, M.D.

Harrisburg, Pennsylvania 17112 USA

1. Qin P, Huilan X, Laursen TM, et al. 2005(17 June) Risk for
schizophrenia and schizophrenia-like psychosis among patients with
epilepsy: population based cohort study. BMJ Publishing Group Ltd. BMJ,

2. Murray, D, July 3, 2005 response to: Qin P, Huilan X, Laursen TM,
et al. 2005(17 June). Risk for schizophrenia and schizophrenia-like
psychosis among patients with epilepsy: population based cohort study. BMJ
Publishing Group Ltd. BMJ, doi:10.1136/bmj.38488.462037.8F.

3. Srinivasan, J & Richens A. (1994) A Risk-Benefit Assessment
of Vigabartin in the treatment of neurological disorders. Drug Saf. 10:

4. Brown TM, Stoudemire A. (1998) Psychiatric Side Effects of
Prescription and Over-the-Counter Medications: Recognition and Management.
Washington, DC. American Psychiatric Press, Inc.

5. Dow Jones Newswire .2005(21 April). FDA Requests Reviews of
Epilepsy Drugs-Paper. The Wall Street Journal Online.

6. Murray, Lori, Senior Ed. (2004) PDR 58 Edition 2004: Physicians’
Desk Reference. Montvale, NJ, Thompson PDR.

Competing interests:
None declared

Competing interests: No competing interests

05 July 2005
Stefan P Kruszewski
Harrisburg, Pennsylvania USA