Abstract P092: New Predictive Equation for 24-Hour Urine Sodium Excretion from Spot Urine Samples Among Older and Hypertensive Adults: The MESA and CARDIA Urinary Sodium Study
Background: Previous analyses indicated the mean bias in existing predictive equations for 24-h urinary sodium (UNa) excretion using spot urine specimens varied significantly among different gender-race groups of adults aged 45-79 years. Our goal is to develop a new equation for 24-hour UNa which is unbiased across gender-race groups of older adults.
Methods: 24-h urine samples were obtained from 554 MESA and CARDIA participants aged 45-79 years. One third (n=189) provided a second 24-h urine sample. Four timed voids (morning, afternoon, evening, and overnight) and the 24-h collection were analyzed for sodium (Na) concentration. With each spot specimen, we used the first day samples to build gender-race specific equations for 24-h UNa. Candidate predictors included age, age-squared, gender, race, height, weight, body mass index (BMI), hypertensive status, diuretic use, estimated Glomerular Filtration Rate (eGFR), and timed-spot Na/creatinine (Cr)/potassium (K)/chloride (Cl) concentrations and excretions. Stepwise regression and the lasso procedures were used to develop parsimonious candidate equations for predicting 24-h UNa. We chose the best fitted model and then used the second day sample to compare the mean biases in 24-h UNa of the resulted equation with existing equations.
Results: Across the timing of specimens, existing equations have significant mean bias for one or more gender-race groups. The proposed equation generally produced less or equivalent bias than previous equations; predictors included age, weight, BMI, hypertensive status, diuretic use, eGFR, and timed-spot urinary Na/Cr/K excretions. The proposed equation performed best using the overnight urine specimen, with unbiased prediction of 24-h UNa overall (mean bias=21.3) and across gender-race subgroups (mean bias <300 mg/d) (Figure).
Conclusion: The proposed equation may result in less biased estimates for the overall and subgroup mean UNa excretion among older US biracial adults, especially using an overnight urine specimen.
Author Disclosures: L. Zhao: None. N.B. Allen: None. C.M. Loria: None. L. Van Horn: None. C. Wang: None. M.E. Cogswell: None. J. Wright: None. A. Carriquiry: None. K. Liu: None.
- © 2015 by American Heart Association, Inc.