Significance of frequency patterns in ‘moderate’ drinkers for low-risk drinking guidelines
Until recently, average levels of drinking have been the main criterion in assessing the health effects of alcohol in epidemiological studies, and guidelines on safe drinking have focused on the average volume (quantity/amount) of drinking based on number of drinks per week . A substantial body of literature and meta-analyses show a U- or J-shaped relationship between total weekly alcohol intake and cardiovascular disease or all-cause mortality with the lowest risk among light–moderate drinkers (≤ three drinks a day) and risk of harm increasing thereafter [2-4]. While the protective effect of light–moderate drinking has been the major topic of debate over the last three decades , attention has turned more recently to the patterns of drinking among low-risk drinkers (defined as those who consume on average ≤ two drinks/day in men and ≤ one drink/day in women) . There has been growing recognition that patterns and frequency of drinking may have an influence on health outcomes in particular coronary heart disease (CHD) risk [7-9]. Many studies have shown that those who drink more regularly show a lower risk of CHD than those who drink infrequently, even among low average alcohol drinkers or when alcohol volume has been taken into account [8-10]. This is often interpreted as regular light–moderate drinking being most beneficial for CHD [9, 10].
The importance of confounding in studying the health effects of alcohol was brought to attention more than 2 decades ago, when interpreting the lower risk associated with light–moderate drinking compared to life-long abstainers . The recent study by Naimi et al. raises again the important question of confounding, but in the context of whether the lower risk seen in regular light drinkers compared to those who drink less frequently, despite similar volumes of alcohol/week, is due to confounding . In the study of more than 168 000 men and women classified as ‘low-risk drinkers’ (≤ seven units/week for women/≤ 14 drinks/week for men), it was shown that regular light drinkers have more advantageous characteristics than infrequent light drinkers and were more likely to drink within safe limits. Binge drinkers (defined as > five drinks/occasion for men and > four drinks/occasion for women) or those who exceeded US dietary guidelines limits (≥ two drinks during drinking days for women/≥ three drinks during drinking days for men), have more adverse characteristics than non-binge drinkers and those who did not exceed guideline limits even though average weekly consumption may be similar. While this cross-sectional study cannot illustrate whether these advantageous characteristic do indeed account for the difference in risk between frequent and infrequent light drinkers seen in other prospective studies, they raise the possibility that confounding may explain this difference. The findings by Naimi et al. have potentially important implications in light of the inferences which have been drawn from studies which have not taken binge drinking into account when examining frequency of drinking [9, 10].
While this study examines the characteristics of frequent and infrequent drinkers, the study has not differentiated binge drinkers and non-binge drinkers clearly within the infrequent light drinkers. The authors have only dichotomized the population into frequent and infrequent drinkers or binge drinking and non-binge drinking, and it is not evident from the study whether infrequent drinkers who consume within recommended limits are significantly different from regular light drinkers or from binge drinkers. One would need to see three groups presented within this category of low-risk drinkers: regular light drinking (15+ days/month), infrequent light drinkers but who do not binge and binge drinkers. It is likely that even infrequent light drinkers who are not binge drinkers have fewer advantageous life-style characteristics compared to regular light drinkers, and multiple advantageous characteristics in frequent drinkers may well contribute to the differences in risk between infrequent light drinkers and regular light drinkers seen in many studies [8-10]. This is similar to the concept of wine drinkers having lower risk than non-wine drinkers which is, to a large extent, associated with their advantageous life-style characteristics . However, binge drinkers, who are usually included among infrequent drinkers, are likely to form a separate group. In many large prospective studies which have examined patterns and frequency of drinking in relation to health, the number of binge drinkers in the study, particularly in women, is too small to make any inferences, or binge drinkers have not been separated out [9, 10]. Studies on the effects of binge drinking have shown adverse effects of binge drinking on CHD [14-17], even if average consumption on a weekly basis is moderate [15-17]. It is this group in whom alcohol may well have different effects on health. Larger prospective studies, where binge drinking is prevalent, detailed assessment of drinking patterns and extensive control of confounders would help to address the analysis of health-related effects of patterns of drinking in low-risk drinkers. Future studies on the health effect of low-risk drinking should focus on the potential harmful health effects of binge drinking among those classified as low-risk drinkers based on average weekly consumption, rather than on whether increased frequency of drinking among these low-risk drinkers is protective against CHD.
Declaration of interests