REVIEW OF: Topiwala A et al, BMJ 2017 Jun 6;357:j2353. doi:10.1136/bmj.j2353

It’s well established that long-term, heavy alcohol use can damage the brain and can cause problems such as Wernicke-Korsakoff syndrome and alcohol-related dementia. But what about the vast majority of our patients, who drink moderately? A recent study reports an association between light-to-moderate drinking and hippocampal atrophy.

In the Whitehall II imaging study, 550 civil servants in the UK without alcohol use disorder were enrolled. All participants were given the CAGE questionnaire, and to be included, they had to score < 2 (considered not to have an alcohol use disorder). Participants were then followed over 30 years, from 1985 to 2015. Patients were asked to track their weekly consumption and then report on an annual basis through a questionnaire.

Participants’ drinking habits were ascertained by asking them to estimate how many drinks they had per week. Based on their answers, they were divided into the following groups: “abstinent” (< 1 unit per week), “light” (1 to < 7 units per week), “moderate” (7 to < 14 units per week for women or 7 to < 21 units per week for men), and “unsafe” (14+ units per week for women or 21+ units per week for men). A standard 6 oz glass of wine would equal 2.4 units of alcohol, and a standard 12 oz beer roughly 2 units. The primary outcomes were the results of cognitive function tests (MoCA and MMSE) and brain imaging results from annual MRI scans administered over the 30 years.

The results of the study showed an association between higher alcohol use and reduced gray matter density, hippocampal atrophy, and reduced white matter microstructural integrity. Even at only moderate alcohol consumption, the risk of hippocampal atrophy was significant (OR 3.4, p = 0.007). For the cognitive tests, higher alcohol consumption predicted faster decline of lexical fluency (naming words that begin with the same letter), but no difference in semantic fluency (naming words belonging to the same category) or word recall. Furthermore, there was no difference in the results of the full MoCA or MMSE.

CATR’s Take
This study shows a correlation between long-term drinking, even light-to-moderate drinking, and brain abnormalities. The clinical relevance of this, though, is not clear. Drinkers did not do any more poorly than abstainers on 4 out of 5 cognitive tests, so it’s possible that the MRI findings seen do not lead to actual cognitive difficulties. It’s also important to point out that the findings were correlations and do not imply causality. The results are consistent with the hypothesis that drinking can cause brain changes without clear cognitive consequences. Sharing these findings with patients may be helpful, particularly if they are curious about the topic. But the study does not convincingly argue that everybody should cease drinking alcohol.