Twenty-Four-Hour Urine Phosphorus as a Biomarker of Dietary Phosphorus Intake and Absorption in CKD: A Secondary Analysis from a Controlled Diet Balance Study

Abstract

Background and objectives: Twenty-four-hour urine phosphorus is commonly used as a surrogate measure for phosphorus intake and absorption in research studies, but its reliability and accuracy are unproven in health or CKD. This secondary analysis sought to determine the reliability and accuracy of 24-hour urine phosphorus as a biomarker of phosphorus intake and absorption in moderate CKD.

Design, setting, participants, & measurements: Eight patients with stage 3-4 CKD participated in 2-week balance studies with tightly controlled phosphorus and calcium intakes. Thirteen 24-hour urine collections per patient were analyzed for variability and reliability of 24-hour urine phosphorus and phosphorus-to-creatinine ratio. The accuracy of 24-hour urine phosphorus to predict phosphorus intake was determined using a published equation. The relationships of 24-hour urine phosphorus with phosphorus intake, net absorption, and retention were determined.

Results: There was wide day-to-day variation in 24-hour urine phosphorus within and among subjects (coefficient of variation of 30% and 37%, respectively). Two 24-hour urine measures were needed to achieve ≥75% reliability. Estimating dietary phosphorus intake from a single 24-hour urine resulted in underestimation up to 98% in some patients and overestimation up to 79% in others. Twenty-four-hour urine phosphorus negatively correlated with whole-body retention but was not related to net absorption.

Conclusions: From a sample of eight patients with moderate CKD on a tightly controlled dietary intake, 24-hour urine phosphorus was highly variable and did not relate to dietary phosphorus intake or absorption, rather it inversely related to phosphorus retention.

Trial registration: ClinicalTrials.gov NCT01161407.

Keywords: Biological Phenomena; Biomarkers; Humans; Phosphorus; Phosphorus Absorption; Phosphorus, Dietary; Physiological Phenomena; Renal Insufficiency, Chronic; Reproducibility of Results; Urine Specimen Collection; calcium; chronic kidney disease; creatinine; nutrition.

Copyright © 2018 by the American Society of Nephrology.

Figures

Graphical abstract

Figure 1.

Twenty four-hour urine phosphorus was highly variable within and among patients with CKD. Daily variation in subjects in (A) 24-hour urine phosphorus (absolute values) and (B) 24-hour urine phosphorus (% variation above and below the 13-day mean [set at zero] for each subject); (C) 24-hour urine creatinine (absolute values) and (D) 24-hour urine creatinine (% variation above and below the 13-day mean [set at zero] for each subject); (E) 24-hour urine phosphorus-to-creatinine ratio (absolute values) and (F) 24-hour urine phosphorus-to-creatinine (% variation above and below the 13-day-mean [set at zero] for each subject); and (G) predicted dietary phosphorus intake calculated on the basis of 24-hour urine phosphorus (see Materials and Methods). In (G), the measured, controlled level of phosphorus intake is shown by the horizontal black line. In (B, D, and F), the mean for each subject is set at zero and the % fluctuation each day above or below the mean is shown; zero is indicated by a horizontal black dashed line. In all panels, different color lines represent individual subjects. Cr, creatinine; P, phosphorus; P/Cr, phosphorus-to-creatinine.

Figure 1.

Twenty four-hour urine phosphorus was highly variable within and among patients with CKD. Daily variation in subjects in (A) 24-hour urine phosphorus (absolute values) and (B) 24-hour urine phosphorus (% variation above and below the 13-day mean [set at zero] for each subject); (C) 24-hour urine creatinine (absolute values) and (D) 24-hour urine creatinine (% variation above and below the 13-day mean [set at zero] for each subject); (E) 24-hour urine phosphorus-to-creatinine ratio (absolute values) and (F) 24-hour urine phosphorus-to-creatinine (% variation above and below the 13-day-mean [set at zero] for each subject); and (G) predicted dietary phosphorus intake calculated on the basis of 24-hour urine phosphorus (see Materials and Methods). In (G), the measured, controlled level of phosphorus intake is shown by the horizontal black line. In (B, D, and F), the mean for each subject is set at zero and the % fluctuation each day above or below the mean is shown; zero is indicated by a horizontal black dashed line. In all panels, different color lines represent individual subjects. Cr, creatinine; P, phosphorus; P/Cr, phosphorus-to-creatinine.

Figure 2.

At least two 24-hour urine collections were needed for a reliable measure of 24-hour urine phosphorus or phosphorus-to-creatinine ratio. Reliability of (A) 24-hour urine phosphorus and (B) 24-hour urine phosphorus-to-creatinine. n is the number of 24-hour urine collections needed to achieve various levels of reliability. The solid line represents the estimated reliability with n samples, and 95% CIs around this estimate is indicated by dashed lines. Reliability is affected by the variability of the measure and is the extent to which a measurement is predicted to give the same result when repeated. P, phosphorus; P/Cr, phosphorus-to-creatinine.

Figure 3.

Twenty four-hour urine phosphorus did not relate to net phosphorus absorption, but negatively related to whole-body phosphorus retention in CKD patients. Correlations between (A) 24-hour urine phosphorus and whole-body phosphorus retention and (B) 24-hour urine phosphorus and net phosphorus absorption. The solid line is the regression fit, the shaded area shows the 95% confidence limits, and the dotted lines indicate the 95% prediction limits. The regression equation for (A) is: Retention (mg/d)=833–1.088×(24-hour urine phosphorus, mg/d). Note that this equation could be used to estimate phosphorus retention from 24-hour urine phosphorus (ideally two or more replicates) with the caveat that subjects are consuming a similar phosphorus intake of approximately 1500 mg/d as those in this study. P, phosphorus.

Figure 4.

Twenty four-hour urine calcium was generally low and showed high percent coefficients of variation, but low absolute variation in CKD patients. Daily variation in subjects in (A) 24-hour urine calcium (absolute values) and (B) 24-hour urine calcium (% variation above and below the mean [set at zero] for each subject); (C) 24-hour urine calcium-to-creatinine ratio (absolute values) and (D) 24-hour urine calcium-to-creatinine (% variation above and below the mean [set at zero] for each subject). In (B and D), the mean for each subject is set at zero and the percentage fluctuation each day above or below the mean is shown; zero is indicated by a horizontal black dashed line. In all panels, different color lines represent individual subjects. Ca, calcium; Ca/Cr, calcium-to-creatinine.