Little is known regarding serum albumin's epidemiologic relation to chronic disease. The relation of serum albumin to blood pressure was assessed in a longitudinal study of men who have been seen at 3- to 5-year intervals since the early 1960s. The authors analyzed data from over 20 years of observation using cross-sectional multiple regression models of blood pressure that allow for the correlation between repeated measures on the same individual (GLMIC models), longitudinal GLMIC models that incorporate terms for the interaction of time with serum albumin at baseline, and models of the slope of individuals' blood pressure over time. Serum albumin levels were found to have a consistently strong relation with both systolic blood pressure and diastolic blood pressure in the cross-sectional GLMIC models only. This relation did not change appreciably when covariates for age, body mass index, alcohol ingestion, smoking, serum calcium, hematocrit, heart rate, and antihypertensive medications were added. A rise in serum albumin of 1 g/dl was associated with 1.79-mmHg and 0.91-mmHg increases in systolic blood pressure and diastolic blood pressure, respectively. This phenomenon may be related to experimental studies linking tryptophan, the only amino acid to bind noncovalently to serum albumin, to a blood pressure-lowering effect mediated by promotion of 5-hydroxytryptamine synthesis in the brain.