(Morris et al., 2001) and the study conducted by Wilson and colleagues (1980) included a greater proportion of African American and Hispanic women than are in the general population.

In view of the interactive effects of sodium and potassium highlighted in this report, it is useful to examine intakes of sodium and potassium expressed as the ratio of sodium intake (in mmol/day) to potassium intake (mmol/day) for the various lifestage groups. Appendix Table D-11 includes these data from NHANES III. Under 1 year of age, the median sodium:potassium ratio is less than one. The ratio then rises rapidly to just above two for children 4 through 8 years of age, and remains above two into adulthood, but then drops somewhat in middle- and older-aged adults. The progressive rise in this ratio at an early age reflects a greater increase in dietary sodium intake compared with the increase in dietary potassium intake. A similar pattern is present in both men and women.

Sodium chloride consumption is one of several dietary factors that contribute to increased blood pressure. Other dietary factors that raise blood pressure are excess weight, inadequate potassium intake, high alcohol consumption, and a suboptimal dietary pattern (see the following sections). Physical inactivity also increases blood pressure. Increased blood pressure is associated with several chronic diseases, including stroke, coronary heart disease, renal disease, and left ventricular hypertrophy.

Cardiovascular Disease and High Blood Pressure. Data from numerous observational studies provide persuasive evidence of the direct relationship between blood pressure and cardiovascular disease. A review of each epidemiologic study is beyond the scope of this report. However, several meta-analyses have aggregated data across these studies (Lewington et al., 2002; MacMahon et al., 1990). The most recent and largest meta-analysis to date pooled data from 61 prospective observational studies that together enrolled almost 1 million adults, including persons with hypertension (Lewington et al., 2002). Individual-level records were available for each participant in each study. Those individuals with pre-existing vascular disease were excluded.