#5565 USEFULNESS OF URINARY BIOCHEMISTRY IN THE FOLLOW-UP OF PATIENTS WITH CHRONIC KIDNEY DISEASE STAGES 3B-5NOD

Abstract

Abstract Background and Aims Ion and protein determinations in a simple urine sample are simple and inexpensive laboratory tests that reflect well ion balance and proteinuria. However, they are underused. Evidence of their usefulness in estimating 24-hour urine biochemistry in patients with chronic kidney disease (CKD) would allow widespread use of these tests and would help to optimize therapies. It would also avoid 24-hour urine collections. To evaluate In CKD stages G3b-5noD the accuracy of equations that estimate, through simple urine sampling, creatinine, protein, and ions in 24-hour urines. Method Retrospective observational study in incident adult patients with CKD G3b-5noD from the nephrology outpatient clinic in a Madrid university hospital, with a follow-up of at least two years and with at least one simultaneous determination of biochemistry in a simple urine sample and in 24-hour urine. The 24-hour urine analyses were restricted to those that were estimated to be well collected because urinary creatinine was between 15–25 mg/kg in men and 10–20 mg/kg in women. Correlation, bias (mean differences between estimated and measured analyte), accuracy (percentage of samples estimated at 15–30 or 50% of the measured value - P15, P30, P50) and agreement by the Bland-Altman method were analyzed. CKD progression was defined by the mean annual and percentage decrease in eGFR and the renal event at follow-up as loss of more than 40% of eGFR or entry into renal replacement therapy or conservative-palliative renal treatment. Kawasaki, Tanaka and CKD-epi (Table 1) equations were used. Results The analyzed cohort was made up of 300 patients with a median follow-up of 72 (25) months. Mean eGFR was 32.8 (9.2) ml/min/1.73m2 and mean protein/creatinine ratio was 1,180 (1,740) mg/g. During follow-up 30.7% started renal replacement therapy or conservative and palliative treatment. Twenty-one patients (7%) died. The estimation of 24-hour urine creatinine by CKD-EPI equation showed excellent correlation and accuracy with measured urinary creatinine (r = 0.887; P30: 96.1% in 464 urines). Accuracy was maintained regardless of age and CKD stage. Formulas that, using single urine samples, estimate 24-hour urine elimination of Na and K correlated well with 24-hour urine measurements, but with high dispersion, especially with Na. For Na, the Tanaka and CKD-EPI formulas had the highest performance (r = 0.540; 0.532; P30: 52.6; 54.5%; bias: -14.4; -16.1 mmol, Tanaka and CKD-EPI respectively for 464 urines). These values make them useful for population studies but not for individualized analysis. For K, the Kawasaki and CKD-EPI formulas were the most suitable (r = 0.712; 0.612; P30 = 74.6; 72.4%; bias: 4.2; 5.5 mmol; Kawasaki and CKD-EPI respectively for 464 urines). These results make them useful for both population-based studies and individualized analyses. Patients with Nao < 87 mmol/day (salt-poor diet) estimated by CKD-EPI, Kawasaki or Tanaka and/or Na/K ratio < 2 directly measured had less eGFR loss at follow-up and fewer renal events. Urinary K elimination measured or estimated did not influence renal prognosis. Proteinuria measured in 24-hour urine correlated much better with proteinuria estimated by CKD-EPI in a single urine sample than with the protein/creatinine ratio (r: 0.900 vs. 0.801; mean bias: 92.2 vs. 486 mg; P30: 72.0 vs. 42.7%; CKD-EPI vs. Proto/Cro in 464 24-hour urine samples). Conclusion Urinary biochemistry in a single urine sample in patients with advanced CKD (stages G3b-5noD) is a useful tool for estimating 24-hour urine biochemistry and to identify scenarios of increased progression of renal disease.