Showing papers by "Timothy S. Naimi published in 2017"

Selection biases in observational studies affect associations between 'moderate' alcohol consumption and mortality.

Abstract: Selection biases may lead to systematic overestimate of protective effects from 'moderate' alcohol consumption. Overall, most sources of selection bias favor low-volume drinkers in relation to non-drinkers. Studies that attempt to address these types of bias generally find attenuated or non-significant relationships between low-volume alcohol consumption and cardiovascular disease, which is the major source of possible protective effects on mortality from low-volume consumption. Furthermore, observed mortality effects among established low-volume consumers are of limited relevance to health-related decisions about whether to initiate consumption or to continue drinking purposefully into old age. Short of randomized trials with mortality end-points, there are a number of approaches that can minimize selection bias involving low-volume alcohol consumption.

Alcohol Consumption and Mortality From Coronary Heart Disease: An Updated Meta-Analysis of Cohort Studies

Abstract: Objective:Previous meta-analyses estimate that low-volume alcohol consumption protects against coronary heart disease (CHD). Potential errors in studies include systematic misclassification of drinkers as abstainers, inadequate measurement, and selection bias across the life course.Method:Prospective studies of alcohol consumption and CHD mortality were identified in scholarly databases and reference lists. Studies were coded for potential abstainer biases and other study characteristics. The alcohol–CHD risk relationship was estimated in mixed models with controls for potential biases. Stratified analyses were performed based on variables identified as potential effect modifiers.Results:Fully adjusted meta-analysis of all 45 studies found significantly reduced CHD mortality for current low-volume drinkers (relative risk [RR] = 0.80, 95% CI [0.69, 0.93]) and all current drinkers (RR = 0.88, 95% CI [0.78, 0.99]). There was evidence of effect modification by cohort age, gender, ethnicity, and heart health a...

Alcohol Policies and Alcohol-Related Motor Vehicle Crash Fatalities Among Young People in the US

Abstract: BACKGROUND: Motor vehicle crashes (MVCs) are a leading cause of death among young people in the United States. We examined the relationship between states’ alcohol policy environments and alcohol-related MVC fatalities among children, adolescents, and young adults under the minimum legal drinking age of 21 years. METHODS: We used the Alcohol Policy Scale (APS), an assessment of 29 alcohol policies across 50 states and Washington, DC, developed with the assistance of an interdisciplinary Delphi panel. Using the Fatality Analysis Reporting System, we examined APS scores in relation to fatalities of people ≤20 years old from 2000 to 2013 occurring in crashes in which ≥1 involved driver had a blood alcohol content ≥0.08%. Logistic regression was used with a 1-year lag between policies and MVC fatalities and adjusted for potential confounders. RESULTS: Of 84 756 MVC fatalities of those ≤20 years old during the study period, 23 757 (28.0%) were alcohol related, including deaths of 11 006 (46.3%) drivers, 10 212 (43.0%) passengers, and 2539 (10.7%) pedestrians, cyclists, and others. People killed in alcohol-related MVCs were predominantly male (72.7%) and older (65.5% were 18–20 years old), and 51.2% were non-Hispanic white. Restrictive policy environments were associated with fewer fatalities (adjusted odds ratio, 0.91 per 10-percentage-point increase in APS score; 95% confidence interval, 0.89–0.94). The association was observed for drivers and passengers, male and female decendents, and children, adolescents, and young adults. CONCLUSIONS: More restrictive alcohol policies are associated with reduced alcohol-related MVC mortality among young people. Studies should scrutinize the relationship between policies and fatalities to highlight mechanisms.

Alcohol Policies and Alcohol-Involved Homicide Victimization in the United States

Abstract: Objective:The purpose of this study was to examine the associations between the alcohol policy environment and alcohol involvement in homicide victims in the United States, overall and by sociodemographic groups.Method:To characterize the alcohol policy environment, the presence, efficacy, and degree of implementation of 29 alcohol policies were used to determine Alcohol Policy Scale (APS) scores by state and year. Data about homicide victims from 17 states from 2003 to 2012 were obtained from the National Violent Death Reporting System. APS scores were used as lagged exposure variables in generalized estimating equation logistic regression models to predict the individual-level odds of alcohol involvement (i.e., blood alcohol concentration [BAC] > 0.00% vs. = 0.00% and BAC > 0.08% vs. < 0.079%) among homicide victims.Results:A 10 percentage point increase in APS score (representing a more restrictive policy environment) was associated with reduced odds of alcohol-involved homicide with BAC greater than 0...

Selection bias and relationships between alcohol consumption and mortality

Abstract: We appreciate the insightful comments from the eminent Drs Britton & Bell, Gmel and Roerecke. While we agree with Britton & Bell’s [1] assertion that modeling potential benefits from low-volume consumption has little impact on reducing world-wide estimates of harms, scientific conclusions about health effects of ‘moderate’ alcohol consumption shape general attitudes about alcohol consumption. In particular, concepts of health benefit are used to oppose alcohol control policies, factor into development of public health guidelines and complicate public health and clinical messaging. Our concerns with selection bias [2] were made with regard to limitations about observational evidence, and life-course analysis (such as that by Britton & Bell) can greatly reduce selection bias. We agree that selection bias applies to both drinkers and non-drinkers, but overall favors positive findings for drinkers. This is because many former drinkers (who, typically, are removed from drinker–non-drinker comparisons) may quit for reasons of ill health, including that related to alcohol consumption itself. Conversely, life-time abstainers appear to have poorer health and social profiles, even at young ages [3]. Finally, in all-cause mortality studies, death preceding cohort inception—particularly among younger people—is due more probably to alcohol consumption among drinkers than to a lack of consumption among non-drinkers. Randomized trials can address some aspects of selection bias that are difficult to overcome with life-course analysis (e.g. performing true intention-to-treat analyses), and are the best way to mitigate confounding. Gmel’s well-argued case for applying Occam’s razor to simplify the argument and accept the J-curve as evidence of ‘a true beneficial effect’ [4] is growing more difficult to defend. For starters, a J-curve suggesting initial mortality benefit that inflects positively above low levels of consumption does not beg a simple explanation, and can be an epidemiological red flag for spurious associations. Beyond the recent meta-analysis finding of no protection for all-cause mortality among low-volume drinkers after adjusting for study quality and abstainer biases [5], and the Mendelian randomization study finding no protection against coronary heart disease (CHD) events [6], arguments for biological plausibility have weakened. High-density lipoprotein cholesterol (HDL-c), which is raised by alcohol consumption in experimental studies, has been discredited as a causal factor for CHD on the basis of a Mendelian randomization study [7] a metaanalysis of statin trials that adjusted for their low-density lipoprotein (LDL) effects [8] and multiple failed drug trials of HDL-raising pharmacological agents [9]. In observational and Mendelian randomization studies increased alcohol consumption is associated with higher blood pressure and hypertension, and in observational studies increased alcohol consumption is associated with increased carotid intima media thickness and coronary calcification [10]. In a meta-analysis of experimental studies, low-volume alcohol consumption does not lower blood glucose among diabetics [11], and two randomized studies that found benefit from the Mediterranean diet for CHD outcomes were not randomized with respect to alcohol [12,13], suggesting that diet is the driver behind the ‘French paradox’. It is precisely because of the conflicted science and methodological limitations, including selection bias, that we should apply the Precautionary Principle, rather than Occam’s razor, to the matter at hand. This requires that the burden of proof rests on demonstrating benefit with high-level evidence before recommending alcohol as a preventative agent. Indeed, we echo Roerecke’s [14] note of caution, particularly with respect to policy development and drinking guidelines. Compared with benefits, the addictive and adverse health effects are not only larger but are more scientifically compelling, to the extent that they are led by conditions that are wholly alcohol-attributable or that have high relative risk estimates, conditions for which alcohol is one of a relatively small number of risk factors, and/or conditions where the latency between exposure and outcome is short. Observed relationships between low-volume alcohol consumption and CHD lack these attributes. While we agree with Roerecke that it is important to assess cause-specific outcomes, when it comes to a substance associated with multiple outcomes across the life-course, all-cause mortality seems the most important outcome to gauge accurately.