Alcohol, Cardiovascular Disease, and Industry Influence: A Meta-review
Golder, S.; Lau, O.; Hartwell, G.; Blanchard, L.; Gibson, A.; Crookes, C.; Foster Davies, L.; Glover, R.
Show abstract
Objectives: This meta-review compares methodological and reporting approaches between systematic reviews examining alcohol dose and cardiovascular disease (CVD) and assesses whether alcohol industry involvement is associated with divergent conclusions. Methods: KSR Evidence was searched 6 May 2025 to update a cohort of 60 systematic reviews from previous review. Reviews were included if they examined any dose-response relationship between alcohol consumption and CVD. Two reviewers independently screened records and extracted data on review characteristics, and citations. Methodological quality was appraised using AMSTAR 2. For a matched sample of reviews with and without known alcohol industry funding, the overlap of included primary studies was compared using Corrected Covered Area (CCA) analysis. Results: Thirty additional systematic reviews met the inclusion criteria, yielding 90 systematic reviews (1996-2025). Most (60.0%, 54/90) concluded that alcohol had a cardioprotective effect, whereas 31.1% (28/90) concluded no evidence of protection, and 8.9% (8/90) were inconclusive. Twenty reviews (22.0%) had declared or inferred alcohol industry funding or author connection; all but one reported a protective effect at lower doses, the other was inconclusive. Industry-connected reviews were cited more often (mean 575.9 vs 193.0, p=0.0002) and more commonly examined overall CVD rather than specific conditions (such as hypertension or stroke). Study overlap was low (CCA 2.59%) and 99% of reviews were rated as critically low quality. Conclusions: The fragmented evidence base is of poor methodological quality with selective inclusion of studies. Alcohol industry connections are strongly associated, with conclusions favouring alcohol consumption, highlighting the need for independent high-quality systematic reviews.
Matching journals
The top 5 journals account for 50% of the predicted probability mass.