A Reply to <scp>R</scp>oerecke &amp; <scp>R</scp>ehm: Continuing Questions about Alcohol and Health Benefits

I am grateful for the opportunity for further debate and to reiterate my view that (i) Roerecke & Rehm 1 conducted a technically excellent meta-analysis on a largely poor literature and (ii) that plausible arguments remain against there being health benefits from moderate drinking. Given the fundamental significance of the issue for alcohol policy, I suggest that it is important to keep the shortcomings of this literature in clear view—as well as some other awkward findings for the health benefit hypothesis. I checked through their list of included studies and accept they were adequate for gaining an estimate of average daily alcohol consumption. I further agree that if measures of heavy episodic drinking had been employed in more than a couple of studies then the appearance of health benefits might have been even more pronounced. However, significant problems remain. They note that ‘most’ of their studies controlled for age and smoking status and ‘some’ also controlled for physical activity. Naimi et al. 2 identified some 27 risk factors (e.g. nutrition, physical exercise, body weight, diabetes) that differentiated abstainers significantly from moderate drinkers, and so there is still some room for improvement here. Roerecke & Rehm 3 acknowledge that many of their studies created bias by including former drinkers in the abstainer reference group 4, and they attempted to adjust for this in the analysis. However, there is another potential form of abstainer bias to deal with: the inclusion of occasional drinkers in the abstainer reference group. The idea here is that as people become more frail with age they are inclined not only to give up drinking, but also to cut down 5. If they are included with abstainers this could make moderate drinkers look good by comparison 4. Occasional drinkers may also be more healthy than abstainers for reasons unrelated to their drinking which could cause bias from another direction 6: many studies combine occasional with moderate drinkers, thus also making the latter look good when compared with abstainers. Finally, how adequate is it to estimate average alcohol consumption on the basis of only 2 weeks drinking, which was the minimum criterion used for study inclusion? Drinking patterns are highly variable, and assessments over only a few days can fail to catch occasional heavy drinkers and exaggerate the number of abstainers (e.g. Rehm et al. 7). In summary, I still think this literature is characterized accurately as having many shortcomings. Roerecke & Rehm 3 did not address the additional concern regarding other, less plausible health benefits associated with moderate drinking, such as a reduced risk of liver cirrhosis morbidity for men 8. A recent review paper identified studies claiming health benefits from moderate drinking for: Alzheimer's, asthma, the common cold, depression, gallstones, deafness, urinary tract symptoms, osteoporosis, obesity, arthritis, various cancers and hip fracture—to mention a few 9. Does this mean that alcohol is a universal panacea or are, perhaps, systematic biases at work? Some relevant and interesting new data suggest that even life-time abstainers may be biased towards being less healthy than light or moderate drinkers. Ng Fat & Shelton 10 examined a group of young adults who elected to be abstinent and compared them with their drinking peers, and found them to be significantly more likely to have poor health as well as low income. Somewhat older data from Scandinavia suggest that individuals who abstain completely for ideological reasons are much healthier than other abstainers—and just as healthy as moderate drinkers 11. If these findings make one inclined to throw up hands in despair of finding a watertight reference group, this may be a useful reality check. This is an area of research in need of both continued improvement and a little caution when it comes to interpretation. None.

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.