Abstract
OBJECTIVE: To determine whether the estimates of the size of the association between blood pressure and sodium intake derived from studies of individuals within populations can be quantitatively reconciled with our estimates derived from comparisons of the average blood pressure and sodium intake between different populations. DESIGN: Examination of data from 14 published studies that correlated blood pressure recordings in individuals against measurements of their 24 hour sodium intake (within population studies). MAIN OUTCOME MEASURE: Comparison of observed differences in blood pressure per 100 mmol/24 h difference in sodium intake in each within population study with predicted differences calculated from the between population data, after allowing for the underestimation of the true association of blood pressure with sodium intake caused by the large day to day variation in 24 hour sodium intake within individuals. RESULTS: The underestimation bias inherent in the within populations studies reduced the regression slope of blood pressure on single measures of 24 hour sodium intake to between a half and a quarter of the true value (for example, in one study from 6.0 to 2.4 mm Hg/100 mmol/24 h). Estimates from between population comparisons of the regression slope of blood pressure on sodium intake, after adjustment to take this underestimation bias into account, were similar to the values actually observed in the within population studies. CONCLUSION: The within population studies confirm our estimates from between population comparisons of the magnitude of the association between blood pressure and sodium intake.