Assessment of serum albumin concentration as a risk factor for stroke and coronary disease in African Americans and whites.

Relatively high serum albumin levels compared to low levels have been associated with reduced cardiovascular mortality and coronary heart disease and stroke incidence. The nature of these associations remains undefined. Therefore, widely accepted criteria of Bradford Hill were used to assess the likelihood that the association is a causal one. These criteria were applied to the findings of published studies of serum albumin and risk of cardiovascular disease. Favoring a causal association are lack of temporal ambiguity in cohort studies, consistency of findings among studies and biologically plausible mechanisms. Not favoring a causal association are lack of specificity of the association, and inconsistency among age-, sex-groups. Further studies are needed to elucidate mechanisms for the effect of serum albumin on incidence of cardiovascular disease and death.