Relationships between urinary cadmium levels and blood pressure were examined in a sample of 951 adult men and women who participated in the Second National Health and Nutritional Examination Survey (NHANES II). Among all participants, positive relationships were seen between urinary cadmium levels and both systolic and diastolic blood pressure (p less than 0.05 and p less than 0.01, respectively), after adjusting for age, sex, race, relative body weight, smoking status, and hypertensive medication use. However, analyses for subgroups determined by sex and smoking status were inconsistent. Among current smokers, urinary cadmium levels were significantly positively associated with both systolic and diastolic blood pressure for women, and with diastolic blood pressure for men. Yet among former smokers and lifelong nonsmokers of both sexes, urinary cadmium was not significantly associated with either systolic or diastolic blood pressure. Evidence that some hypertensive medications increase urinary cadmium excretion suggests that the positive associations seen among current smokers may reflect high urinary cadmium levels among hypertensives induced by hypertensive treatment. After treated hypertensives were removed from the analysis, regression coefficients relating blood pressure to cadmium dropped by a factor of two and lost statistical significance. We conclude that the present data provide little support for a causal association between systemic cadmium and hypertension at nonoccupational exposure levels. Further, conflicting results of previous studies may reflect failure to control adequately for age, smoking status, and hypertensive treatment.