Cardiovascular

Alcohol, myocardial infarction and survival: published estimates require correction and reanalysis.

TL;DR

The published abstract attributes an odds ratio of 0.57 to myocardial infarction, whereas the reported counts yield an odds ratio of 1.41, and the findings support only exploratory associations and do not establish cardioprotection or a survival benefit from alcohol consumption.

Key Findings

The odds ratio for myocardial infarction reported in the original study's abstract (0.57) is inconsistent with the value calculable from the published data (1.41).

  • The published abstract attributes an odds ratio of 0.57 to myocardial infarction.
  • Independent recalculation from the reported counts yields an odds ratio of 1.41 for myocardial infarction.
  • The 0.57 estimate actually relates to all-cause mortality, not myocardial infarction.
  • The recalculation was performed using published aggregate data from a 30-year cohort study of 973 male automotive workers.

The original study contains additional concerns beyond the misattributed odds ratio, including survival-time terminology errors and numerical inconsistencies.

  • Concerns identified include survival-time terminology issues.
  • Numerical inconsistencies were identified across tables and figures in the original publication.
  • Methodological concerns include competing-risk methods, multiplicity, and residual confounding.
  • Sick-quitter bias was identified as a specific form of residual confounding not adequately addressed.

The findings of the original 30-year cohort study support only exploratory associations and do not establish cardioprotection or a survival benefit from alcohol consumption.

  • The study population consisted of 973 male automotive workers followed for 30 years in a Swedish occupational cohort.
  • The commentary used critical analysis with independent recalculation of odds ratios from published aggregate data.
  • The authors concluded that the published record requires correction and reconciliation with the underlying data and statistical analyses.
  • The outcome measures assessed in the original study included myocardial infarction, stroke, all-cause mortality, and survival-time estimates.

What This Means

This research involves a critical review of a previously published 30-year study that examined whether alcohol consumption was associated with heart attacks and other cardiovascular outcomes in nearly 1,000 Swedish male automotive workers. The authors of this commentary independently recalculated the statistics using the same data that had been published, and found a major error: the original study's abstract claimed that moderate drinkers had a much lower odds of heart attack (an odds ratio of 0.57, suggesting protection), but when the actual patient counts from the paper's own tables were used, the real odds ratio for heart attack was 1.41 — meaning drinkers actually had higher odds of heart attack, not lower. The 0.57 figure correctly applied to a different outcome, all-cause mortality, not heart attacks. Beyond this numerical mix-up, the commentary identified several other methodological problems in the original study, including issues with how survival time was described, inconsistencies between the numbers presented in tables versus figures, inadequate handling of statistical methods for competing risks (e.g., the fact that dying from one cause prevents dying from another), and a well-known bias called 'sick-quitter bias,' where people who stopped drinking due to illness are incorrectly grouped with lifetime non-drinkers, making moderate drinkers appear healthier by comparison. This research suggests that the original study's conclusion — that alcohol consumption provides heart protection or a survival benefit — is not supported by its own data when correctly analyzed. The authors call for a formal correction to the published record and caution that the findings should be interpreted only as exploratory. This matters because studies claiming alcohol is cardioprotective are frequently cited in public health discussions, and errors in such research can have broad consequences for how people and policymakers think about alcohol and heart health.

Have a question about this study?

Citation

Koulaouzidis G, Koulaouzidis A. (2026). Alcohol, myocardial infarction and survival: published estimates require correction and reanalysis.. Scandinavian journal of primary health care. https://doi.org/10.1080/02813432.2026.2728001