September 1992

Coffee and Coronary Heart Disease

Author Affiliations

From the Departments of Medicine (Dr Myers) and Clinical Epidemiology (Dr Basinski), Sunnybrook Health Science Centre and Departments of Medicine (Dr Myers) and Family and Community Medicine (Dr Basinski), University of Toronto, Ontario.

Arch Intern Med. 1992;152(9):1767-1772. doi:10.1001/archinte.1992.00400210009002

Objective.—  We determined if coffee consumption is associated with an increased risk of developing coronary heart disease.

Data Identification.—  Articles published between 1966 and August 1991 examining a possible link between coffee and coronary heart disease were identified by a computeraided literature search (Medline) and by standard bibliographic searches.

Study Selection.—  All prospective cohort studies providing data on daily coffee consumption and coronary events (acute myocardial infarction and/or coronary death) were included.

Data Extraction.—  Data from each published article were extracted. Additional unpublished data augmenting those published for one study were also included. Each cohort was categorized by reported daily coffee consumption. Incidence of coronary events at each level of coffee consumption was the primary outcome.

Results.—  Eleven prospective studies were included. The coronary events for subjects consuming little or no coffee (≤1 cup per day) were compared with event rates for those consuming greater amounts of coffee. The studies exhibited heterogeneity of results. The typical odds ratios and 95% confidence intervals across studies were estimated by logistic regression analysis. Coffee intake from 1 to 4 cups per day was not associated with any increase in coronary heart disease occurrence compared with 1 cup or less per day (odds ratio, 1.01; confidence interval [0.93,1.11]). The odds ratios for 4 to 6 and 6 cups or more per day compared with up to 1 cup per day were 1.01 (0.90,1.12) and 1.09 (0.97,1.22), respectively.

Conclusions.—  There is no association between coffee consumption and the occurrence of coronary heart disease. This conclusion holds in the absence of adjustment for other coronary risk factors.(Arch Intern Med. 1992;152:1767-1772)