Epilepsy: risk of suicidal behavior with antiepileptic drugs

Risk of suicidal behavior
with antiepileptic drugs
Maurizio Pompili and Ross J. Baldessarini appropriate for such disease staging.3 last, Antiepileptic drugs (AeDs) are receiving increasing attention for a primary end points in clinical trials, change possible association with suicidal thoughts and behaviors. Several recent studies examining this association, however, have yielded inconsistent findings, particularly in terms of the risk of suicidal behavior conferred concentrations, since the latter would not by specific AeDs. In patients with epilepsy, heightened suicide risk is also follow-up unless the therapy in question attributable to comorbid psychiatric conditions.
reverses the disease course and, hence, CsF controls or patients with epilepsy who did in conclusion, the studies by De meyer public-health challenge, accounting for at not receive aeDs. indeed, the prevalence et al. and vemuri et al. have furthered our least 1 million deaths per year worldwide.1 disorders (including mood, psychotic and orders, including epilepsy, are associated personality disorders, and substance abuse) with markedly increased rates of clinical depression and suicidal acts.2,3 the poten- 12.5-fold higher, respectively, in patients tial contribution of psychotropic drugs to exhibiting suicidal behavior than in con- suicide—as potential causes of or treat- trols. andersohn et al. made efforts to adjust ments for suicidal thinking or behavior— for differences in risk factors among the various patient subgroups.6 nevertheless, in intervention studies involving indivi duals interest followed the regulatory approval in the effectiveness of the statistical proce- with proven aD pathology who have not yet 2003 of clozapine as an effective agent for dures in compensating for comorbid factors lowering the incidence of suicidal thoughts associ ated with suicide remains uncertain. or acts among patients with schizophrenia.1 the findings of andersohn et al. encour- Department of Neurology and Alzheimer a series of FDa investigations—based on age close consideration of risk factors for Center, Vrije Universiteit Medical Center, post hoc meta-analyses of adverse-effect suicide that are unrelated to treatment in Boelelaan 1117, 1081 HV Amsterdam, The Netherlands (N. D. Prins). Department from controlled trials involving patients of Neurology, Erasmus Medical Center, s-Gravendijkwal 230, 3015 CE Rotterdam, The Netherlands (J. C. van Swieten). epilepsy—considered the risk of suicidal ideation and suicide attempts associ ated with use of various types of central y active drugs, particularly anti depressants in juve- nile patients and anticonvulsants in adults. the findings from these post hoc analyses in a second recent study, arana et al. indicated that antiepileptic drugs (aeDs) examined clinical findings related to suicide might increase such risk.1,4,5 two recently risk among 5.13 million general practice Competing interestsThe authors declare no competing interests.
patients in the uK.3 these researchers found that among patients who did not show evi- 1. De Meyer, G. et al. Diagnosis-independent andersohn et al. conducted a case– control dence of mood disorders or epilepsy, the Alzheimer disease biomarker signature in study involving 9,420 patients with epi lepsy cognitively normal elderly people. Arch. Neurol. who were treated with aeDs of various classes suicide per 100,000 person-years was 15.0 2. vemuri, P. et al. serial MRi and CsF biomarkers (table 1), including several ‘newer’ aeDs that (95% Ci 14.6–15.5). this rate was 2.5-fold in normal aging, MCi, and AD. Neurology 75, have been associated with increased rates of clinical depression.6 these re searchers were not treated with anticonvulsants (38.2, 3. Jack, C. R. Jr et al. Hypothetical model of found that the current or recent use of these dynamic biomarkers of the Alzheimer’s pathological cascade. Lancet Neurol. 9, increasing this rate by only 26.2% (48.2, mate, vigabatrin and, particularly, levitira- 4. Perrin, R. J., Fagan, A. M. & Holtzman, D. M. cetam) was associated with an approxi mately the observations of arana et al.3 are in Multimodal techniques for diagnosis and prognosis of Alzheimer’s disease. Nature 461, threefold increased risk of suicid al behavior line with the results of a small veterans 5. Kester, M. i. et al. Diagnostic impact of CsF Patients exhibiting suicidal behavior had and colleagues (table 1).7 the va investi- biomarkers in a local hospital memory clinic. Dement. Geriatr. Cogn. Disord. 29, 491–497 far more psychiatric risk factors commonly associated with suicide than did nonsuicidal geriatric patients with epilepsy seemed to volume 6 | DeCemBer 2010 | 651
0 Macmil an Publishers Limited. Al rights reserved Table 1 | Relative risk of suicidal thoughts or acts associated with AeD treatment
risk factors for suicide, all adding to the dif- ficulty of drawing conclusions. moreover, to further complicate matters, Gibbons et al.8 found that aeD treatment led to a decrease in the risk of suicide attempts in patients on top of the issues outlined above, the reported investigations examining suicide risk and aeDs have been strikingly inconsis- tent in their rankings of the relative risks associated with particular drugs (table 1). never theless, three agents—levetiracetam, lamo trigine and topiramate—were among the top three anticonvulsants associated with the highest observed risk of suicidal be havior in at least two of the six reported analyses (in four of the five studies that found an increase in the risk of suicidal behaviors, levetiracetam appeared in the top three, and lamotrigine and topiramate appeared in the top three twice in these five studies).5–7,9,10 of note, however, topiramate, but not lamo- trigine or levetiracetam, has been associated with clinical depression. moreover, these three drugs differ in their pharmacodynamic mechanisms, precluding conclusions about potential biological mechanisms that might Given the varied and largely inconclusive *AeD showing the greatest change in relative risk of suicidal behaviors (derived from outcomes of AeD treatment versus findings regarding the putative effects of placebo treatment) listed first. ‡significant increase in relative risk. §significant decrease in relative risk. Abbreviation: AeD, antiepileptic drug.
aeDs on the risk of suicidal behavior and fatalities, clinical prudence indicates that patients with epilepsy should be assessed be more strongly associated with suicide suici dal behavior conferred by aeDs includes routinely and treated for psychiatric comor- actual mortality, is a class effect of anti convul- bidities that are important risk factors for of this study was its limited statistical power sants or is specific to certain agents, and can suicide. in addition, on the basis of the to assess the suicide risk associated with dynamic mechanisms have not been resolved levetiracetam or, potential y, lamotrigine or topiramate warrants heightened patient care with epilepsy or a psychiatric disorder, sui- and further research in relation to possible cidal ideation occurred more frequently and (table 1) have used the prevalent but highly the rate of occasional suicide attempts was ambiguous concept of ‘suicidality’, which higher during treatment with seven of the includes suicidal thoughts and acts (sui- Department of Neuroscience, Mental Health cides and suicide attempts). However, the and Sensory Functions, Suicide Prevention Center, Sant’Andrea Hospital, Sapienza trigine and topiramate) than during treat- epidemiology and clinical predictive value University of Rome, Via di Grottarossa 1035, ment with placebo.5 nevertheless, aeDs were of suicidal thinking and behaviors, as well as 00189 Rome, Italy (M. Pompili). Harvard not associated with a significant increase in their responses to treatment, are not neces- Medical School and Mailman Research Center, suicide risk among patients with psychi atric sarily comparable.1 several recent studies, McLean Hospital, 115 Mill Street, Belmont, MA disorders.4,5 of note, these findings were including the two considered here,3,6 have derived incidental y from reports of adverse examined the rates of suicidal behaviors events in trials not designed specifical y to and other violent self-injurious acts, and risk of suicide conferred by specific aeDs. convulsants or placebo in the trials analyzed ascribing suicide risk to specific drugs is dif- by the FDa, the detailed clinical compar- ficult, even with large samples, as such acts AcknowledgmentsR. J. Baldessarini was supported, in part, by a grant ability of patients with epilepsy between the are rare.5–10 in addition, the available studies from the Bruce J. Anderson Foundation and by the treatment arms is uncertain, as is the con- (table 1) have varied greatly in metho dology, McLean Private Donors Psychopharmacology and tribution of psychiatric factors to the risk of size, statistical power and the adequacy of suici dal behavior in patients with epilepsy. measures taken to control potential y con- the issues of whether the putative risk of founding variables, par ticularly psychiatric The authors declare no competing interests.
652 | DECEMBER 2010 | voluME 6
0 Macmil an Publishers Limited. Al rights reserved 1. Pompili, M. & Tatarelli, R. (eds) Evidence-Based 6. Andersohn, F., schade, R., willich, s. n. & Practice in Suicidology: A Sourcebook (Hogrefe, Garge, e. Use of antiepileptic drugs in epilepsy and the risk of self-harm or suicidal behavior. 2. Blumer, D. et al. suicide in epilepsy: Neurology 75, 335–340 (2010).
investigate any potential genetic dif ferences 7. vanCott, A. C. et al. suicide-related behaviors in prevention. Epilepsy Behav. 3, 232–241 older patients with new anti-epileptic drug use: data from the vA hospital system. BMC Med. 8, 3. Arana, A., wentworth, C. e., Ayuso-Mateos, J. L. & Arellano, F. M. suicide-related events in 8. Gibbons, R. D., Hur, K., Brown, C. H. & hybridiza tion to evaluate Dna copy number patients treated with antiepileptic drugs. Mann, J. J. Relationship between antiepileptic in 34 adult patients with medulloblastomas, N. Engl. J. Med. 363, 542–551 (2010).
drugs and suicide attempts in patients with and utilized interphase fluorescent in situ 4. Pompili, M., Tatarelli, R., Girardi, P., Tondo, L. & bipolar disorder. Arch. Gen. Psychiatry 66, hybridization (FisH) to validate their find- anticonvulsant treatment. Pharmacoepidemiol. 9. Patorno, e. et al. Anticonvulsant medications ings in a further 112 adults and 303 children Drug Saf. 19, 525–528 (2010).
and the risk of suicide, attempted suicide, or who had this type of tumor.1 By use of these 5. statistical review and evaluation: antiepileptic violent death. JAMA 303, 1401–1409 (2010).
drugs and suicidality. FDA [online], 10. Olesen, J. B. et al. Antiepileptic drugs and risk of suicide: a nationwide study. Pharmacoepidemiol. that age- dependent differences in genomic Drug Saf. 19, 518–524 (2010).
profiles exist between adult and pediatric medullo blastomas. For example, analysis of the interphase FisH data revealed that MYC or MYCN oncogene amplification and Genetic variation in pediatric
gain of chromosomes 1q, 2, 6q, 7 and 17q and adult brain tumors
cyclin-dependent kinase 6 (CDK6) gene amplification and 17q gain was frequently Alba A. Brandes and Enrico Franceschi Two new studies suggest that pediatric medulloblastomas and high- markers are better than clinical variables at grade gliomas are genetically different from the same tumors in adults. predicting survival, the researchers devised an algorithm that stratified adult patients Age-dependent gene expression might affect tumor biology; therefore, into three risk groups on the basis of chromo- therapies for adult medulloblastomas or gliomas might not produce the same clinical outcomes in pediatric patients, and vice versa.
according to the re searchers, tumors with both 10q deletion and 17q gain are associ- trials of potential medulloblastoma thera- ated with a poor prognosis. tumors that have pies have rarely been conducted in adults.3 either 10q loss or 17q gain are associated with the same morphology can have diverse with a 44% overall survival rate at 5 years genetic profiles. Genetic profiles can aid the blastomas have been assumed to be the same. after diagnosis, and tumors without 17q gain identification of specific patient subgroups; Consequently, in adults these tumors are or 10q loss are associated with a good prog- frequently treated with pediatric protocols, nosis at 5 years. thus, this algorithm could smokers with lung adeno carcinoma is known with simple adjustments being made to the aid the identification of adult patients with to harbor exon 19 epidermal growth factor drug regimens to account for dif ferences in medulloblastoma who have a poor prognosis receptor mutations. medulloblastomas and drug tolerance. likewise, because our under- and could benefit from more-tailored thera- high-grade gliomas are two types of brain standing of pediatric high-grade gliomas is peutic approaches. Pediatric patients had tumor that can affect both adults and chil- limited compared with adult gliomas, these poor prognoses if their tumors showed gains dren, and tumors in these different patient tumors in children are often treated with of chromosome 6q and/or 17q or amplifica- anticancer therapies that are routinely used tion of MYC or MYCN. By contrast, loss of tumor location, histo pathology and/or gene chromosome 6q was associated with a good expression. two studies published in the Journal of Clinical Oncology now show that in the other study, Paugh and colleagues pediatric high-grade gliomas and medullo- blastomas are genetical y distinct from the adult forms of these tumors.1,2 these obser- vations indicate that these two types of identification of significant targets in ferently to anti cancer treatments in adults in a study by Korshunov et al., analysis of clinical and histological data revealed that population, but are more frequent in chil- desmoplastic tumors were found frequently dren. in fact, these tumors are so rare in in adults, while large-cell– anaplastic tumors adults that the best treatment option for number alterations in three chromosomes. this type of cancer has not been determined in this patient population, and prospective most often in pediatric patients, and adult more frequent in pediatric tumors than in volume 6 | DeCemBer 2010 | 653
0 Macmil an Publishers Limited. Al rights reserved

Source: https://iasp.info/pdf/papers/pompili_baldessarini_Nat_Rev_Neurol_2010.pdf