Are there non-linear relationships between alcohol consumption and long-term health?: a systematic review of observational studies employing approaches to improve causal inference

The association between alcohol consumption and all-cause mortality: An umbrella review of systematic reviews using lifetime abstainers or low-volume drinkers as a reference group

BACKGROUND AND AIMS

Systematic reviews of the relationship between alcohol consumption and all-cause mortality have reported different relative risk (RR) curves, possibly due to the choice of reference group. Results have varied from 'J-shaped' curves, where low-volume consumption is associated with reduced risk, to monotonically increased risk with increasing consumption. We summarised the evidence on alcohol consumption and all-cause mortality exclusively from systematic reviews using lifetime abstainers or low-volume/occasional drinkers as the reference group.

METHODS

We conducted a systematic umbrella review of systematic reviews of the relationship between alcohol consumption and all-cause mortality in prospective cohort studies using a reference group of lifetime abstainers or low-volume/occasional drinkers. Several databases (PubMed/Medline/Embase/PsycINFO/Cochrane Library) were searched to March 2022. Reviews were assessed for risk of bias, and those with reference groups containing former drinkers were excluded.

RESULTS

From 2149 articles retrieved, 25 systematic reviews were identified, and five did not include former drinkers in the reference group. Four of the five included reviews had high risk of bias. Three reviews reported a J-shaped relationship between alcohol consumption and all-cause mortality with significant decreased risk for low-volume drinking (RR range 0.84 to 0.95), while two reviews did not. The one review at low risk of bias reported monotonically increased risk with greater consumption (RRs = 1.02, 1.13, 1.33 and 1.52 for low-, medium-, high- and higher-volume drinking, respectively, compared with occasional drinking). All five reviews reported significantly increased risk with higher levels of alcohol consumption (RR range 1.28 to 3.70). Sub-group analyses were reported by sex and age; however, there were evidence gaps for many important factors. Conversely, 17 of 20 excluded systematic reviews reported decreased mortality risk for low-volume drinking.

CONCLUSIONS

Over 70% of systematic reviews and meta-analyses published to March 2022 of all-cause mortality risk associated with alcohol consumption did not exclude former drinkers from the reference group and may therefore be biased by the 'sick-quitter effect'.

Is low-level alcohol consumption really health-protective? A critical review of approaches to promote causal inference and recent applications

Heavy and disordered alcohol consumption is a known risk factor for several health conditions and is associated with considerable disease burden. However, at low-to-moderate levels, evidence suggests that drinking is associated with reduced risk for certain health outcomes. Whether these findings represent genuine protective effects or mere methodological artifacts remains unclear, but has substantial consequences for policy and practice. This critical review introduces methodological advances capable of enhancing causal inference from observational research, focusing on the 'G-methods' and Mendelian Randomization. We also present and evaluate recent research applying these methods and compare findings to the existing evidence base. Future directions are proposed for improving our causal understanding of the relationships between alcohol and long-term health outcomes.

Behavior-related risk factors and time to death among persons with alcohol consumption versus persons without: A general population study with mortality follow-up after 20 years

BACKGROUND

Evidence shows that low to moderate alcohol consumers seem to live longer than abstainers. Insufficient consideration of subgroups among abstainers and of further behavior-related risk factors for death might be reasons. The aim of this study was to compare alcohol lifetime abstainers, former drinkers, and current consumers with regard to mortality considering tobacco smoking, body overweight, and physical inactivity.

METHODS

A general adult population sample of residents aged 18 to 64 had been drawn at random in northern Germany. Among eligible persons, 4093 (70.2%) participated. Assessments included alcohol consumption by the Alcohol Use Disorders Identification Test Consumption in addition to lifetime alcohol abstinence and former drinking. A score of behavior-related risk factors was built from tobacco smoking, body overweight, and physical inactivity. Twenty years later, a mortality follow-up was conducted. Data of 4028 study participants were analyzed.

RESULTS

At baseline, former alcohol consumers but not current low to moderate alcohol drinkers had more behavior-related risk factors than lifetime abstainers. At follow-up, former alcohol drinkers with two or more behavior-related risk factors had a shorter time to death than lifetime abstainers with 0 or one behavior-related risk factor (hazard ratio 3.43, 95% confidence interval: 1.63-7.20). Low to moderate alcohol drinkers did not survive longer than lifetime alcohol abstainers with 0 or one behavior-related risk factor.

CONCLUSION

The results provide evidence against the assumption that alcohol consumption has a beneficial effect on health and longevity.

EQ-5D-5L Population Norms and Quality-Adjusted Life Expectancy by Sociodemographic Characteristics and Modifiable Risk Factors for Adults in Queensland, Australia

OBJECTIVES

Over half of Australia's disease burden is due to morbidity, predominantly chronic conditions. Health-related quality of life instruments provide measures of morbidity and health status across different dimensions with EQ-5D being one of the most widely used. This study reports EQ-5D-5L general population norms for Queensland, Australia using the recently published Australian value set.

METHODS

Population survey results from cross-sectional computer-assisted telephone interviews for Queensland adults in 2022 and 2023 were analyzed. EQ-5D-5L, as well as modifiable risk factors and sociodemographic data were collected. Using the recently published final Australian EQ-5D-5L value set, mean utility scores were calculated for Queensland, as well as by sociodemographic characteristics, including remoteness and socioeconomic area-based measures, and modifiable risk factors, such as smoking and body mass index. Results were combined with life tables to estimate quality-adjusted life expectancy (QALE) for subgroups with different lifestyles.

RESULTS

The EQ-5D utility score for the Queensland adult population was 0.916. Smoking daily, being obese or older in age, or living in the most disadvantaged socioeconomic area were associated with lower mean scores. QALE was 6.1 and 7.9 years shorter than the life expectancy for Queensland males and females, respectively, but generally, those who reported having healthier lifestyles had higher mean utility scores and thus longer QALE.

CONCLUSIONS

In addition to reporting Queensland EQ-5D-5L general population norms, these results demonstrate potential QALE gains in people following healthier lifestyles. The results support investment in prevention and may motivate further studies in this important area.

The dose-response relationship of pre-menopausal alcohol consumption with age at menopause: a population study of 280 497 women in Norway

BACKGROUND

Previous research suggests that alcohol consumption is associated with high age at menopause. Yet, knowledge about the dose-response relationship is inconsistent. Thus, we studied the pattern of the association of pre-menopausal alcohol consumption with age at natural menopause.

METHODS

We performed a retrospective population-based study using self-reported data from 280 497 women aged 50-69 years attending the Norwegian breast cancer screening programme (BreastScreen Norway) during 2006-15. Associations of weekly alcohol consumption between the age of 20 and 49 years with age at menopause were estimated as hazard ratios (HRs) using Cox proportional hazard models with restricted cubic splines to allow for non-linear associations. We adjusted for year and place of birth, number of childbirths, educational level, body mass index and smoking habits.

RESULTS

Mean age at natural menopause was 51.20 years (interquartile range: 49-54 years). The adjusted HR of reaching menopause was highest for women with no alcohol consumption (reference) and the HR decreased by alcohol consumption up to 50 grams per week (adjusted HR 0.87; 95% CI: 0.86-0.88). Above 50 grams, there was no further decrease in the HR of reaching menopause (P for non-linearity of <0.001).

CONCLUSIONS

Women who did not consume alcohol were youngest at menopause. The lack of a dose-response association among alcohol consumers implies virtually no relation of alcohol consumption with age at menopause. Our findings may suggest that characteristics of the women who did not consume alcohol, not accounted for in the data analyses, explain their younger age at menopause.

Improving interventional causal predictions in regulatory risk assessment

In 2022, the US EPA published an important risk assessment concluding that "Compared to the current annual standard, meeting a revised annual standard with a lower level is estimated to reduce PM2.5-associated health risks in the 30 annually-controlled study areas by about 7-9% for a level of 11.0 µg/m3… and 30-37% for a level of 8.0 µg/m3." These are interventional causal predictions: they predict percentage reductions in mortality risks caused by different counterfactual reductions in fine particulate (PM2.5) levels. Valid causal predictions are possible if: (1) Study designs are used that can support valid causal inferences about the effects of interventions (e.g., quasi-experiments with appropriate control groups); (2) Appropriate causal models and methods are used to analyze the data; (3) Model assumptions are satisfied (at least approximately); and (4) Non-causal sources of exposure-response associations such as confounding, measurement error, and model misspecification are appropriately modeled and adjusted for. This paper examines two long-term mortality studies selected by the EPA to predict reductions in PM2.5-associated risk. Both papers use Cox proportional hazards (PH) models. For these models, none of these four conditions is satisfied, making it difficult to interpret or validate their causal predictions. Scientists, reviewers, regulators, and members of the public can benefit from more trustworthy and credible risk assessments and causal predictions by insisting that risk assessments supporting interventional causal conclusions be based on study designs, methods, and models that are appropriate for predicting effects caused by interventions.

Are health behaviors associated with mental health among tertiary education students? A systematic review of cohort studies

OBJECTIVE

To evaluate the association between health behaviors with mental health among tertiary education students.

METHODS

Six databases were searched until September 2021 for prospective cohort studies evaluating the association between health behavior(s) (dietary intake, physical activity, sedentary behavior, alcohol intake, sleep, smoking or illicit drug use) and mental health. Two independent reviewers screened records for inclusion, extracted data and completed risk of bias assessments.

RESULTS

33 studies were included (14 assessed sleep, 14 alcohol intake, 13 physical activity, 8 smoking, 6 sedentary behavior, 4 diet, 1 illicit drug use). A consistent association between poor sleep, and physical inactivity with increased risk of poor psychological wellbeing, and between poor sleep and increased mental ill-health related outcomes was demonstrated.

CONCLUSION

Findings suggest interventions to address poor sleep and physical inactivity among students may positively impact mental health. Further research of other health behaviors, and their association with mental health, is required.

A comprehensive evaluation of the longitudinal association between alcohol consumption and a measure of inflammation: Multiverse and vibration of effects analyses

BACKGROUND

Moderate alcohol consumption appears to be associated with reduced inflammation. Determining whether this association is robust to common variations in research parameters has wide-reaching implications for our understanding of disease aetiology and public health policy. We aimed to conduct comprehensive multiverse and vibration of effects analyses evaluating the associations between alcohol consumption and a measure of inflammation.

METHODS

A secondary analysis of the 1970 British Birth Cohort Study was performed, using data from 1970 through 2016. Measurements of alcohol consumption were taken in early/mid-adulthood (ages 34 and 42), and level of inflammation marker high-sensitivity C-reactive protein (hsCRP) at age 46. Multiverse analyses were applied to comparisons of low-to-moderate consumption and consumption above various international drinking guidelines with an 'abstinent' reference. Research parameters of interest related to: definitions of drinking and reference groups; alcohol consumption measurement year; outcome variable transformation; and breadth of covariate adjustment. After identifying various analytic options within these parameters and running the analysis over each unique option combination, specification curve plots, volcano plots, effect ranges, and variance decomposition metrics were used to assess consistency of results.

RESULTS

A total of 3101 individuals were included in the final analyses, with primary analyses limited to those where occasional consumers served as reference. All combinations of research specifications resulted in lower levels of inflammation amongst low-to-moderate consumers compared to occasional consumers (1st percentile effect: -0.21; 99th percentile effect: -0.04). Estimates comparing above-guidelines drinking with occasional consumers were less definitive (1st percentile effect: -0.26; 99th percentile effect: 0.43).

CONCLUSIONS

The association between low-to-moderate drinking and lower hsCRP levels is largely robust to common variations in researcher-defined parameters, warranting further research to establish whether this relationship is causal. The association between above-guidelines drinking and hsCRP levels is less definitive.

Moderate Alcohol Consumption and Depression: A Marginal Structural Model Approach Promoting Causal Inference

OBJECTIVE

Moderate alcohol consumption is associated with decreased risk for depression, but it remains unclear whether this is a causal relationship or a methodological artifact. To compare the effects of consistent abstinence and occasional, moderate, and above-guideline alcohol consumption throughout early to middle adulthood on depression at age 50, the authors conducted a secondary analysis of the National Longitudinal Survey of Youth 1979 cohort and employed a marginal structural model (MSM) approach.

METHODS

Baseline was set at 1994, when individuals were ages 29-37. The MSM incorporated measurements of alcohol consumption in 1994, 2002, and 2006, baseline and time-varying covariates, and repeated measurements with the Center for Epidemiologic Studies Depression Scale-Short Form (CES-D-SF). A total of 5,667 eligible participants provided valid data at baseline, 3,593 of whom provided valid outcome data. The authors used all observed data to predict CES-D-SF means and rates of probable depression for hypothetical trajectories of consistent alcohol consumption.

RESULTS

The results approximated J-curve relationships. Specifically, both consistent occasional and consistent moderate drinkers were predicted to have reduced CES-D-SF scores and rates of probable depression at age 50 compared with consistent abstainers (CES-D-SF scores: b=-0.84, 95% CI=-1.47, -0.11; probable depression: odds ratio=0.58, 95% CI=0.36, 0.88 for consistent occasional drinkers vs. abstainers; CES-D-SF scores: b=-1.08, 95% CI=-1.88, -0.20; probable depression: odds ratio=0.59, 95% CI=0.26, 1.13 for consistent moderate drinkers vs. consistent abstainers). Consistent above-guideline drinkers were predicted to have slightly increased risk compared with consistent abstainers, but this was not significant. In sex-stratified analyses, results were similar for females and males.

CONCLUSIONS

This study contributes preliminary evidence that associations between moderate alcohol consumption and reduced risk for depression may reflect genuine causal effects. Further research using diverse methodologies that promote causal inference is required.

The Alcohol Use Disorders Identification Test and Mortality 20 Years later

The objective was to test whether Alcohol Use Disorders Identification Test (AUDIT) results may predict mortality after 20 years. An observational study was conducted. It included an interview of a general population sample aged 18 to 64 in northern Germany in the years 1996 − 1997 (baseline) and a mortality-follow-up in the years 2017 − 2018. Study participants were 3581 persons who had consumed alcohol during the last 12 months prior to the baseline assessment. It included the AUDIT which was filled in by study participants. At follow-up, death cases were ascertained including the date of death. Official records and death certificates from local health authorities were used. Cox proportional hazards regression revealed that the AUDIT predicted time to death. The hazard ratio was 1.70 (95% confidence interval: 1.43 − 2.02) with the lowest AUDIT zone of values as the reference group. Competing risks regression analysis for diagnosis-specific mortality data revealed that the AUDIT predicted cardiovascular mortality (subhazard ratio, 1.84; 95% confidence interval, 1.49 − 2.27). It is concluded that the alcohol screening predicted total and cardiovascular mortality in this adult general population sample.

Do MAFLD Patients with Harmful Alcohol Consumption Have a Different Dietary Intake?

The term metabolic-associated fatty liver disease (MAFLD) has been proposed to define positively fatty liver disease in the form associated with metabolic risk factors. The aim of this study was to assess the dietary intake of MAFLD and explore a possible relationship between its inflammatory characteristics (assessed by Dietary Inflammatory Index—DII^®), the degree of liver fibrosis (assessed by transient elastography), and the amount of alcohol intake. MAFLD patients were included (n = 161) and were classified, according to the amount of alcoholic intake, as MAFLD without alcohol intake (n = 77) and MAFLD with alcohol intake (n = 84), with 19 presenting harmful alcoholic consumption. Dietary intake was 1868 ± 415 kcal/day and did not present differences in energy or nutrient intake based on the presence of metabolic comorbidities. Patients with MAFLD and alcohol intake consumed significantly more energy and presented a tendency for higher intake of carbohydrates and sugar. Patients with harmful alcohol intake presented a higher intake of total fat and cholesterol compared with moderate alcohol intake. There were no differences in DII^® based on fibrosis severity or the amount of alcohol consumption. This work contributes to the characterization of baseline dietary intake in MAFLD patients, paving the way to design more suited dietary interventional trials.