Using Spot Urine Samples to Estimate Sodium and Potassium intake in Healthy vs CKD Patients: Results from a Controlled Feeding Study

Abstract

Abstract Objectives Evaluating sodium (Na) and potassium (K) intake is important as both have implications for blood pressure and cardiovascular health, especially in individuals with chronic kidney disease (CKD). Spot urine samples may be used to estimate Na and K intake with several published equations, but the accuracy has not been thoroughly explored. Our objective was to compare estimated 24-hour urinary Na and K excretion (e24hUNa and e24hUK), calculated from a spot urine sample using published equations, to measured 24hUNa and 24hUK and Na and K intake, and to determine if there are differences between healthy and CKD participants. Methods This is a secondary analysis of a controlled feeding study in participants with moderate CKD matched to healthy adults (n = 16). Participants consumed a controlled diet for 9 days, providing ∼2400 mg Na/d and ∼3000 mg K/d. On days 7 and 8, participants collected all urine in an inpatient setting. Urine Na and K were analyzed by ICP and urine creatinine by the Jaffe reaction. The day 7 fasting urine sample was used to calculate e24hUNa using 2 equations and e24huK using 1 equation. Log-transformed Na intake, measured 24hUNa, and e24hUNa were compared using a general linear mixed model and partial correlations, and agreement between Na intake and e24hUNa was assessed using Bland-Altman plots. Similar analyses were run for K. Results Participants were aged 54.6 ± 13.0 y, n = 8 were female, n = 6 were black, and n = 10 were white. In CKD participants, eGFR was 40.7 ± 7.9 mL/min. Average Na intake on day 7 was 2138 ± 302 mg and K intake was 2528 ± 254 mg. Na intake, measured 24hUNa, and e24hUNa were not significantly different (P = 0.17) or correlated with each other (all P > 0.20) and there was no significant interaction with CKD status (P = 0.51). However, both e24hUNa were ∼500 mg higher than Na intake. For K, e24hUK was significantly different from K intake (P < 0.001) and measured 24hUK (P = 0.02), with no interaction with CKD (P = 0.53). K intake, measured 24hUK, and e24hUK were not correlated (all P > 0.05). Bland-Altman plots show poor agreement between both e24hUNa and Na intake and e24hUK and K intake. Conclusions Estimated 24hUNa and estimated 24hUK are poor predictors of Na and K intake, respectively, in both healthy and CKD participants. Results should be confirmed in a larger sample. Funding Sources Indiana CTSI (funded by NCATS CTSA program), NIDDK.