Abstract P413: Do Methodological Issues Account for the Inconsistent Results Observed in Cohort Studies of Sodium Intake and Clinical Cardiovascular Disease Outcomes?

Introduction: As sodium intake (Na) rises, so does blood pressure. However, the results of cohort studies that relate Na to cardiovascular disease (CVD) are inconsistent. We systematically reviewed the quality of cohort studies for methodological issues that could bias assessments of the relationship between Na and CVD. Hypothesis: Methodological limitations are common in cohort studies assessing the link between Na and CVD. Methods: We included all known articles that examined the association between Na and CVD or all-cause mortality, coronary heart disease, stroke or heart failure. We divided articles based on 1) type of study population, i.e. general population versus targeted samples at elevated risk of CVD and 2) method of Na assessment. We looked for the presence of random and systematic error in Na assessment, analytic issues such as under-adjustment, and potential for residual confounding and reverse causality. Results: We identified 26 studies with 31 independent samples. Of the samples, 13 found at least one direct association between Na and CVD, 10 at least one inverse or J-shaped association, and eight only null associations. 20 were from the general population and 11 from targeted groups. Urine collection was used to assess Na in 13 (24-hour collections in 11) and diet-based measures in 18. Overall, 81% of the 31 samples had a high potential for random errors in sodium assessment, eight of the 17 with sufficient information had systematic errors in Na assessment, 29% were conducted in targeted groups with high potential for reverse causality, and 52% were inadequately adjusted. Conclusions: Methodological issues are common in observational studies linking Na and CVD.