# Estimating 24-hour urine phosphate excretion from spot urine

## TL;DR

This study develops a new equation to estimate 24-hour urine phosphate excretion from a single urine sample, which could simplify clinical assessments.

## Contribution

A novel predictive equation for 24hUrP using spot urine parameters and clinical data in individuals with normal kidney function.

## Key findings

- Previously published equations poorly predicted 24hUrP from spot urine.
- A new model using log-transformed uPi, uCr, age, sex, and BMI improved prediction accuracy (r2 = 0.1820).

## Abstract

24-hour urinary phosphate excretion (24hUrP) is indicative of intestinal phosphate absorption in steady-state conditions. Nevertheless, 24-hour urine collections are cumbersome and error-prone. Previous studies suggested that spot urine phosphate (uPi) could serve as a practical substitute to predict 24hUrP, however, these data originated only from patients with chronic kidney disease. Here, we investigated the validity of predictive equations using spot urine parameters to assess 24hUrP in a cohort with normal kidney function (eGFR >60 ml/min per 1.73 m2) including 761 kidney stone patients and 207 non-kidney stone formers as assessed by low-dose CT scans, the Swiss Kidney Stone Cohort (SKSC).

Published equations for 24hUrP were tested in our cohort and a novel predictive equation was developed. Pearson correlation coefficients and Bland–Altman plots were used to assess the relationship between spot uPi and spot urine creatinine (uCr) and 24hUrP. Additionally, forward multivariate analysis was performed to predict uPi excretion.

Previously published equations provided less accurate prediction of 24hUrP from spot urine. Log-transformed 24hUrP with log-transformed spot uPi and creatinine yielded the best model fit. In addition, inclusion of age, sex, and BMI significantly improved prediction of 24hUrP. Compared with spot uPi and uCr alone (r2 = 0.0561, P < .001) the new equation predicted 24hUrP (r2 = 0.1820, P < .001) more accurately.

Here, we present a new equation for predicting 24hUrP from spot urine samples of individuals with normal kidney function. This model has a moderate ability to explain 24hUrP variance but has the strength to use only parameters routinely collected in clinical settings such as spot urinary phosphate and creatinine, sex, BMI, and age.