Drinking alcohol is linked to an increased risk of new-onset epilepsy, with greater consumption tied to greater risk, but more research is needed before any definitive conclusions can be drawn.

Results of an updated meta-analysis are consistent with those of a previous meta-analysis but contrast with some prior cohort studies.

“Further large cohort studies of the general population are required to assert a definite causal relationship between alcohol consumption and epilepsy and to identify a potential threshold,” Prof Yun Hak Kim, Departments of Biomedical Informatics and Anatomy, School of Medicine, Pusan National University, Republic of Korea, said in a press release.

Conflicting Findings

Much of the research into the impact of alcohol on epilepsy risk has focused on provoked seizures related to alcohol intoxication or withdrawal, but few studies have investigated the effect of alcohol on unprovoked seizures. In addition, the research in this area has been conflicting.

A 2010 meta-analysis that included six case-control studies showed that alcohol users had an increased risk of unprovoked seizure or epilepsy with a pooled relative risk (RR) of 2.19 (95% CI, 1.82 – 2.63). This analysis also showed a dose-dependent relationship with relative risks increasing with more grams of alcohol consumed daily.

However, some recent large cohort studies showed that moderate alcohol consumption was associated with a lower risk of epilepsy.

The updated meta-analysis included eight studies ― three cohort studies not included in the previous meta-analysis and five case-control studies.

The study excluded two case-control studies included in the previous meta-analysis. One of these studies used duplicated data, and the other included epilepsy patients and did not present results of subgroup analysis for patients experiencing their first seizure.

Results of the new analysis showed the pooled odds ratio for newly diagnosed epilepsy was 1.70 (95% CI, 1.16 – 2.49) in alcohol users vs nondrinkers.

A dose-response analysis of case-control studies carried out using the cubic spline analysis showed a significant positive dose-response relationship. A dose-response graph showed a steep increase in risk above about 150 g/day and 250 g/day of alcohol consumption.

However, a subgroup analysis showed that epilepsy risk was only found in the case-control studies. In fact, two of the three cohort studies showed that alcohol consumption was associated with a lower risk of epilepsy, although this was not significant.

Cohort studies often include more control subjects and longer follow-up periods and are less prone to bias, such as selection and recall biases, the investigators note.

“Therefore, cohort studies usually provide a stronger association between exposure and disease than case-control studies, despite having limitations for diseases with low incidence levels,” they write.

More Research Needed

They add that most case-control studies included in the new meta-analysis assessed alcohol consumption only in the 6 months prior to the onset of seizures. Research shows it usually takes heavy drinkers 5 or more years to develop repetitive unprovoked seizures.

“Considering these temporal relationships and differences in study design, alcohol may not actually increase the risk of epilepsy, as seen in our subgroup analysis for cohort studies,” the investigators write.

They note that the cohort studies in the meta-analysis were variously limited to young women, elderly patients, and post–subdural hematoma patients.

“This limitation makes it difficult to confirm or generalize the results of the subgroup analysis.”

To resolve this “discrepancy,” further large cohort studies of the general population over a longer period are needed, the investigators write.

Examining the risk of bias within studies, the authors evaluated three cohort studies as “good” quality. Of the case-control studies, they rated two as “good,” one as “fair,” and two as “poor.”

For primary prevention, an assessment of the risk of alcohol consumption in various clinical situations, such as the time relation of alcohol consumption with seizures, will be important, lead author Prof Kyoung Nam Woo, Department of Neurology, Pusan National University, said in the release.

To increase the applicability to the general population, future studies should be conducted in which the potential confounders such as age, sex, and smoking have been adjusted.”

Commenting on the study for Medscape Medical News, Jacqueline French, MD, professor, NYU Comprehensive Epilepsy Center, New York City, echoed the authors in noting a number of weaknesses in the study.

The analysis was unable to exclude alcohol withdrawal seizures. Also, while some studies suggested a positive relationship, others suggested a negative relationship, she said.

“The authors suggest further work is needed before a definitive determination is made, and I agree.”

SOURCE: medscape.com, Pauline Anderson