Pharmacoepidemiology of the effect of caffeine on blood pressure.

In experimental studies, caffeine increases blood pressure in caffeine-naive or nontolerant individuals, but not in regular caffeine consumers. Using an epidemiologic approach, we examined the hypothesis that serum-caffeine concentration would be positively associated with blood pressure in infrequent (but not habitual) caffeine users in a group of bus drivers. Infrequent and habitual users of caffeine showed no differences in systolic or diastolic blood pressures when there is no measurable caffeine in the serum. However, at serum concentrations of caffeine typical of those achieved after one to two cups of coffee, infrequent users demonstrated greater systolic and diastolic pressures, averaging +5.3 mm Hg and +3.6 mm Hg, respectively, compared with habitual users. The magnitude of difference remained after adjustment for age, body mass index, race, sex, and tobacco and alcohol use. These elevations are large enough to exaggerate the prevalence of hypertension, if such assessments are based on cross-sectional surveys that fail to assess both proximate caffeine consumption and usual caffeine consumption habits.