No proof of a causal relationship between antiepileptic drug treatment and incidence of dementia. Comment on: Use of antiepileptic drugs and dementia risk—An analysis of Finnish health register and German health insurance data

We have serious concerns about the validity of the conclusions drawn in the study by Heidi Taipale and colleagues1 on the use of antiepileptic drugs (AEDs) and the risk of dementia or Alzheimer's disease, which has been widely publicized by media and the lay press as conclusive evidence that “antiepileptic drugs increase risk of Alzheimer's disease, dementia,” as written in DG News Neurology on April 9, 2018 (http://dgnews.docguide.com/antiepileptic-drugs-increase-risk-alzheimer-s-disease-dementia). Specifically, the alarming news is that “Continuous use of antiepileptic drugs for a period exceeding 1 year was associated with a 15% increased risk of Alzheimer's disease and with a 30% increased risk of dementia,” respectively 20% and 60% for drugs that impair cognitive function. The study in question is an epidemiological study that analyzed large health care databases from Germany and Finland and found that the intake of AEDs with known negative psychotropic effects for >2 years was associated with a small increase in the risk of dementia. Although the authors point out that they simply described an association, the media and the lay public interpret the finding as indicative of a cause-effect relationship. This is reminiscent of how several years ago AEDs were blamed as a cause for an increased risk of suicide (U.S. Food and Drug Administration warning in January 2008), a claim which has been subsequently questioned or minimized by many studies.2 Misinterpretation of the suggested association between AEDs and dementia as a causal relationship would have even worse implications, because it could undermine physicians’ confidence in the risk-to-benefit ratio of prescribing necessary medications and lead to reduced patient adherence, with potentially life-threatening consequences. Overall, up to 1% of the population needs chronic treatment with AEDs because of epilepsy.3 Not treating epilepsy is generally not an option, and if patients stop treatment abruptly they may experience severe seizures, status epilepticus, or sudden unexpected death in epilepsy. Importantly, the study by Taipale et al is affected by potential flaws that deserve attention. The time lag between AED exposure and the diagnosis of dementia was too short to exclude the possibility that AEDs were prescribed to treat early symptoms of dementia. In this respect, it should be considered that several AEDs, in particular carbamazepine and valproate, are sometimes used as mood stabilizers in people with dementia and behavioral disturbances. Moreover, epileptic seizures can present as early symptoms of a dementing brain disease.4 Taipale et al published comparable studies on associations with the risk of Alzheimer's disease (eg, opioid use, antipsychotics, postmenopausal hormone therapy, proton pump inhibitors, month and season of birth). A scientific question and its answer, however, must be put into context, and lack of context is a major problem in their study on AEDs. First, there is no scientifically plausible reason to investigate the relation of AED use and dementia or Alzheimer's disease without an implicit hypothesis of a causal relationship. Purely associative research without any hypothesis is likely to reveal a number of spurious correlations. We are not aware of any evidence that AEDs may cause dementia or even Alzheimer's disease. The authors cite an early Canadian study5 that discusses epilepsy and its pharmacological treatment in elderly people as risk factors for dementia. This 5-year observational study in persons older than 65 years also reports an association between AEDs and dementia. However, the diagnosis was based on a weak criterion for dementia (falling below a cutoff in a cognitive screening test), and, not surprisingly, the significance of the association was lost when the underlying pathology was considered, that is, when patients with head trauma or stroke were excluded. This leads to the second major problem with the study by Taipale et al. Given the risk of high rates of a misdiagnosis of dementia,6, 7 the reliability of the diagnosis in the German and Finnish databases can be questioned. It remains unclear whether the diagnoses were based on evidence of progressive mental decline (at least 1 follow-up evaluation) and/or biomarkers indicative of dementia. Often, as in the Canadian study,5 dementia in elderly people may simply be defined as failing a psychometric dementia screening test. Patients taking AEDs have a high risk of being cognitively impaired by their medications.8 This risk increases with higher doses, and when AEDs with more prominent negative psychotropic cognitive effects are taken. The major affected domains are attention and executive functions, but memory and language performance can also be affected, and these effects are likely to be more prominent in elderly people with age-related central nervous system (CNS) impairment and, potentially, concurrent intake of other CNS-active medications.9 These symptomatic (and reversible) AED effects will influence performance on dementia screening tests, thereby increasing the number of persons who score low on such a test and may be suspected of, or misdiagnosed with, dementia. Not surprisingly, the AEDs found by Taipale et al to be associated with an increased risk of dementia were those more likely to exert adverse effects on cognition. This leads to the third issue that needs to be addressed to place the findings by Taipale et al into context. It remains unclear which diseases beyond epilepsy required AED treatment in their population, and to what extent these diseases could by themselves affect cognitive functioning. Many patients with epilepsy, because of their seizures or underlying pathologies, may fulfill psychometrically the criteria of a mild cognitive impairment. However, there is no evidence that a dementing mental decline occurs as a result of having seizures or seizures being treated.10, 11 If there is progressive cognitive decline, there should be an underlying pathological condition causing it. It is clear that in these cases differential diagnosis is challenging and misdiagnoses of dementia can easily occur. Although Taipale et al claim to have adjusted for the confounding effects of comorbidities, the accuracy and ascertainment rate of relevant comorbidities in the study databases are uncertain. The study also did not account for the types, severity, and duration of the comorbidities (eg, epilepsy, depression). Moreover, this was a case-control study reporting neither the absolute risks of dementia/Alzheimer's disease in patients with AED intake and in those patients without AED use nor the relative risks, but only odds ratios. Odds ratios may considerably “overestimate” relative risks.12 Unfortunately, odds ratios are often misinterpreted as relative risks, especially but not exclusively by the lay press (see the abovementioned DG News Neurology article stating that “Continuous use of antiepileptic drugs […] was associated with a 15% increased risk of Alzheimer's disease”). The strength of the reported association was low and achieved statistical significance only because of the large sample sizes. In addition, there are some obvious discrepancies between the German and Finnish datasets, which were not discussed by the authors. For example, in the Finnish dataset 4.4% of the patients with Alzheimer's disease were on regular AED treatment and 1.6% had epilepsy; that is, the vast majority of patients treated with AEDs had no epilepsy. What were the indications for regular AED treatment in these patients? In contrast, in the German dataset only 2.7% of the patients with dementia were on regular AED treatment, and 2.6% had epilepsy. In conclusion, the study by Taipale et al has important methodological weaknesses and its findings are likely to be invalidated by the influence of major confounders. For the reasons stated above, noncritical reporting of these findings could have serious negative consequences for the care and well-being of people with epilepsy. It should also be emphasized that AEDs can exert not only negative but also positive effects on cognition and behavior,13 and that some of these drugs can have neuroprotective effects in experimental models14-16 and are even being considered for investigation as a potential therapeutic option in neurodegenerative diseases.17 A causal relationship between AED treatment and the risk of developing dementia has not been demonstrated, and may be considered unlikely based on longstanding clinical experience with these drugs as an often lifetime treatment in people with epilepsy. Please note that a short version of this open letter with similar content has already been submitted to and accepted by the Journal of the American Geriatrics Society,18 which published the work of Taipale and colleagues that is in question here. C.H. reports grants from EU Grant E-pilepsy and from the EpiCare European Network, and personal fees and nonfinancial support from UCB, Eisai, Desitin Pharmaceuticals, GW Pharmaceuticals, and the journal Seizure. E.B. has received research funds from the European Union, the Italian Medicines Agency, the Italian Ministry of Health, and the American ALS Association; he has also received speaker or consultancy fees from UCB, Eisai, and Shire. C.E.E. reports honoraria and consultation fees from UCB, Desitin Pharmaceuticals, Bial, and Eisai, and grants from Deutsche Forschungsgemeinschaft, Bundesministerium für Bildung und Forschung, and Marga and Walter Boll Stiftung. R.K. has received research funds from the Academy of Finland and Saastamoinen Foundationen and has received speaker's honoraria from Eisai, UCB, and Orion and honoraria for membership on advisory boards from Eisai, Fennomedical, GW Pharmaceuticals, Pfizer, Sage Therapeutics, Takeda, and UCB. K.M. has received research funding from the Swedish Foundation of Strategic Research and through Swedish State Funds for Research at University Hospitals. T.W.M. has received honoraria for speaking engagements from Eisai and Desitin and has received financial support from Desitin for visiting scientific meetings. E.P. has received research funds from the European Union, the Italian Medicines Agency, the Italian Ministry of Health, and the Italian Ministry for Education, University, and Research; he has also received speaker's or consultancy fees from Eisai, GW Pharmaceuticals, LivaNova, Medichem, Mylan, Sandoz, Sanofi, Sun Pharmaceutical Industries, Takeda, and UCB. E.T. has received research funding from the European Union and the FWF Austrian Science Fund; his institution holds research grants from Biogen Idec, Bayer, Red Bull, and Merck; he has also acted as a paid consultant to Eisai, Ever Pharma, Biogen Idec, Medtronics, LivaNova, and Bial, and has received speaker's honoraria from Bial, Eisai, GL Pharma, Novartis, GlaxoSmithKline, Boehringer, Viropharma, LivaNova, Actavis, and UCB in the past 3 years. J.-A.W. has received honoraria from Eisai and UCB. We confirm that we have read the Journal's position on issues involved in ethical publication and affirm that this report is consistent with those guidelines.

Récupéré en direct depuis OpenAlex et désinversé. Les résumés ne sont pas conservés dans cette base de données : les index inversés représentent 8,6 Go des 9,3 Go de texte de la base, et le serveur dispose de 13 Go libres.