Alcohol-related mortality is more pronounced in lower than in higher socioeconomic groups in Western countries. Part of the explanation is differences in drinking patterns. However, differences in vulnerability to health consequences of alcohol consumption across socioeconomic groups may also play a role. We investigated the joint effect of alcohol consumption and educational level on the rate of alcohol-related medical events.
We pooled seven prospective cohorts from Denmark that enrolled 74,278 men and women age 30–70 years (study period, 1981 to 2009). We measured alcohol consumption at baseline using self-administrated questionnaires. Information on highest attained education 1 year before study entry and hospital and mortality data on alcohol-related medical events were obtained through linkage to nationwide registries. We performed analyses using the Aalen additive hazards model.
During follow-up (1,085,049 person-years), a total of 1718 alcohol-related events occurred. The joint effect of very high alcohol consumption (>21 [>28] drinks per week in women [men]) and low education on alcohol-related events exceeded the sum of their separate effects. Among men, we observed 289 (95% confidence interval = 123, 457) extra events per 100,000 person-years owing to education–alcohol interaction (P < 0.001). Similarly, among women, we observed 239 (95% confidence interval = 90, 388) extra events per 100,000 person-years owing to this interaction (P < 0.001).
High alcohol consumption is associated with a higher risk of alcohol-related medical events among those with low compared with high education. This interaction may be explained by differences in vulnerability and drinking patterns across educational groups.
See video abstract at, http://links.lww.com/EDE/B267.
Supplemental Digital Content is available in the text.
From the aSection of Social Medicine, Department of Public Health, University of Copenhagen, Denmark; bAstraZeneca, Nordic Baltic MC, Medical & Regulatory Nordic Baltic, Södertälje, Sweden; cFundacao Oswaldo Cruz, Brazil; dSection of Biostatistics, Department of Public Health, University of Copenhagen, Denmark; eCenter for Statistical Science, Peking University; fResearch Center for Prevention and Health, Rigshospitalet-Glostrup University Hospital, Denmark; gDepartment of Cardiology and the Copenhagen City Heart Study, Bispebjerg University Hospital, Denmark; and hThe Danish Cancer Society Research Centre, Copenhagen Ø, Denmark.
Submitted 8 April 2016; accepted 17 July 2017.
The other authors have no conflicts to report.
The study was funded by the Danish Cancer Society, the Commission of Social Inequality in Cancer (Grant SU08004).
H.N.C., I.A., U.A.H. analyzed the data and contributed to interpretation of the data. H.N.C. drafted the article. F.D., I.A., T.L., N.H.R. developed the study concept and analytical strategy, supervised the interpretation of data, and reviewed the article. M.O., E.P., A.T. were responsible for the acquisition of data and helped revising the article critically. U.A.H., P.K.A. contributed to the development of the analytical strategy and gave important input to the content of the method and discussion section of the text. All authors have approved the final version to be submitted.
While drafting the article, H.N.C. changed her affiliation and is now employed as an epidemiologist by AstraZeneca.
The SIC investigators welcome collaboration on specific projects. Further details of the SIC cohort can be obtained by contacting the steering committee at the Department of Public Health, University of Copenhagen, at email@example.com.
Supplemental digital content is available through direct URL citations in the HTML and PDF version of this article (www.epidem.com)
Correspondence: Ingelise Andersen, CSS, Oester Farimagsgade 5a, DK-1014 Copenhagen K, Denmark. E-mail: firstname.lastname@example.org