Hydes · BMC public health 2019 · epidemiological risk modeling study · n=?

A comparison of gender-linked population cancer risks between alcohol and tobacco: how many cigarettes are there in a bottle of wine?

Cited 327 times in the scientific literature.

Level 5 - mechanism / opinion, no new human data

Epidemiological risk modeling and secondary estimation using published relative risks and population attributable fractions without primary clinical or cohort data.

PubMed 30917803 · doi:10.1186/s12889-019-6576-9 · record verified 2026-08-29

What was done

Researchers calculated the absolute lifetime cancer risk associated with moderate alcohol consumption versus low-level smoking. Using alcohol- and tobacco-attributable fractions subtracted from general population lifetime cancer risks, they estimated baseline cancer risk for non-smoking, alcohol-abstaining adults. They then applied relative risk estimates for drinking 10 units of alcohol per week (equivalent to one bottle of wine) versus smoking 10 cigarettes per week to establish cigarette-equivalent harm estimates across increasing levels of consumption.

What was found

One bottle of wine per week was associated with an absolute lifetime cancer risk increase of 1.0% in non-smoking men and 1.4% in non-smoking women. In terms of absolute cancer risk, one bottle of wine per week equaled smoking 5 cigarettes per week for men and 10 cigarettes per week for women. This gender difference was driven by moderate drinking leading to an absolute breast cancer risk increase of 0.8% in female non-smokers.

Why it matters

Translating alcohol-related absolute cancer risks into cigarette equivalents provides an intuitive metric to help communicate the carcinogenic risks of moderate drinking, particularly for breast cancer in women.

Limits

The study is a mathematical modeling exercise using secondary population-level statistics and relative risk estimates, rather than an observational cohort with primary patient data. It examines cancer endpoints exclusively, omitting non-cancer morbidity and mortality where alcohol and tobacco effects differ substantially. The abstract does not account for drinking patterns (such as binge drinking) or interactive effects between simultaneous smoking and drinking.

Cited by