Editorial: Postoperative antiepileptic drugs

Abstract

In this issue, Komotar and colleagues1 report the results of their “meta-analysis” of publications describing the incidence of postoperative seizures with and without the use of prophylactic antiepileptic drugs (AEDs) in patients with supratentorial meningiomas. They identified 16 articles, 13 of which provided seizure outcome for patients treated with AED prophylaxis, 3 of which did the same for patients not treated with AEDs, and 3 of which compared patient groups. The patients in these series varied with respect to their duration of anticonvulsant treatment and the agent and dose used, as well as their length of follow-up. The authors’ findings agree with recent case series that have demonstrated no clear evidence that AEDs prevent the occurrence of postoperative seizures in these patients. Furthermore, the paper provides a good review of the literature, and can serve as the reference basis for future trials. The authors acknowledge the limitations of their approach to summarizing and analyzing the existing literature and suggest that more prospective studies would be needed to more completely answer the various practical questions that arise in clinical decision-making in this setting. A number of further questions do arise: 1. In other conditions, such as brain injury, early AED use decreases the risk of early, but not late, seizures. Do the authors think that there are potentially different risks or benefits resulting from strategies that would potentially reduce the risk of early versus late seizures in patients following surgery for meningioma? 2. Is there evidence that the occurrence of an early seizure changes the patient’s outcome with respect to the development of late seizures or neurological deficits? 3. Although the authors mention the potential importance of tumor location in the genesis of postoperative seizures, neither the text nor the tables compare tumor location between patients who received AEDs and those who did not. Are there data on this question, and if so, how would this inform the design of future studies? The authors do acknowledge that “meta-analyses” are methodologically imperfect and concede that their findings, though convincing, cannot be interpreted with the same certainty as a randomized controlled trial. The term “meta-analysis” seems to have myriad definitions, some rigorous, and others descriptive. It is not equivalent to a structured review, which is how I would characterize this paper. From the statistical perspective, a meta-analysis is a technique that seeks to combine results from different existing studies to obtain a more accurate estimate of effect size. The analysis should weight the sample sizes involved, such that larger samples would have more statistical impact. There are a number of statistical methods to produce this aggregate estimation of effect. Simply combining subjects into a pooled sample does not achieve this goal. “Meta-analysis,” however defined, also has a certain mystique attached to it. It is important to recognize that the results of such an analysis can be highly questionable when the included studies employ varying research methods or definitions or when the study selection process is somehow biased. Ultimately, a randomized controlled and blinded clinical trial will be necessary to answer the question of whether or not prophylactic AED therapy should be used to prevent early and late seizures after meningioma surgery. This paper provides the impetus and foundation for such a definitive trial.