Author Affiliation: Department of Cardiovascular Diseases, University of Leuven, Leuven, Belgium (firstname.lastname@example.org).
In Reply: Many letters questioned our use of 24-hour urine collections. Volume and levels of potassium and creatinine were approximately 20% lower in the low-sodium than in the high-sodium tertile, whereas sodium was 60% lower. All measurements should have shown a similar gradient from the high to the low tertile if urine collections were incomplete. Administration of para-aminobenzoic acid to assess completeness, as suggested by Dr Aleksandrova and colleagues, could introduce error1 and is impractical in population studies. As Drs Rebholz and He request, we classified participants by sodium-to-creatinine ratio, including those with incomplete urine collections. The adjusted hazard ratios (HRs) for CVD mortality in the low, medium, and high tertiles were 1.54, 1.04, and 0.63 (P = .01).
Staessen JA, EPOGH Investigators FT. Urinary Sodium Excretion and Cardiovascular Disease Mortality—Reply. JAMA. 2011;306(10):1083-1087. doi:10.1001/jama.2011.1297