Arthur L. Klatsky, M.D., Kaiser Permanente Medical Center, Oakland, CA, USA
Objective and Background: Sparse reports suggest that coffee drinking may reduce risk of cerebrovascular disease (CVD). We explored this hypothesis in a large population.
Methods: We studied risk of death and hospitalization attributed to CVD in a free-living multiethnic cohort of 130,054 persons that supplied baseline data in 1978-85. Cox proportional hazards models adjusted for 8 covariates were used to analyze 2,062 deaths and 5,108 persons hospitalized through 2008. These models yielded hazard ratios (HR) and 95% confidence intervals (CI).
Results: For all CVD deaths the adjusted HR (CI) versus non-coffee drinkers was 0.9 (0.8-1.0) for 1-3 cups/day and 0.8 (0.7-0.9, p < 0.001) for 4+ cups/day. This inverse relation to CVD deaths included inverse relations intracerebral hemorrhage, ischemic stroke, and non-specific CVD and these relationships were independent of evidence of baseline cardiorespiratory disease. Other major causes of cardiovascular death were unrelated to coffee. Tea drinking was unrelated to death from CVD. The risk of any CVD hospitalization showed no coffee association with HRs of 1.0 for persons reporting < 1, 1-3, and 4+ cups/day. There were disparities in hospitalization risk for some CVD subgroups, with lower risk of intracerebral hemorrhage and increased risk of subarachnoid hemorrhage.
Conclusions: (1) Independent of baseline cardiorespiratory history, coffee drinking is related to lower risk of death from CVD and its major subgroups.
(2) While causality is only one possibility, a hypothetical protective effect may exist for caffeine or another coffee ingredient.