Background The gold standard to assess proteinuria is the protein content of 24h urine collection. Random spot urine protein/creatinine (P/C) ratio seems to have a good correlation with 24h protein and has been used to simplify urine collection. However, it has been reported that it is unreliable for monitoring proteinuria in lupus nephritis (LN) patients.
Objectives To evaluate the agreement between random spot urine P/C ratio (mg/g) and protein measured by 24h urine collection (g/24h) in LN patients.
Methods Cross-sectional study was performed. Patients with biopsy proven LN underwent 24h urine collection (for P and C) and random urine (for P/C ratio), sequentially. Correlation (Spearman rank correlation analysis) and limits of agreement between the two methods (Bland-Altman plot analysis) were evaluated. The discriminant cut-off values for spot urine P/C ratio in predicting 24h protein threshold excretion >300, 500, 1000 and 1500 mg were determined using receiver operating characteristics curve (ROC).
Results Sixty-one paired (total 122) spot and 24h urine collection were evaluated. The mean glomerular filtration rate (assessed by CKD-E equation) was 97mL/min/1.73m2 (SD 10.9). Strong correlation (r=0.921, p<0.001) was found between the two measures. Stratifying by protein degrees, less strong correlation was found between the two methods (r=0.722, p<0.001 when 24h protein <500mg and r=0.699, p<0,001 when 24h protein was between 500 and 1000 mg).
Bland-Altman analysis showed wide limits of agreement (-496.1 to 891.3 mg) and the limits became wider as the protein excretion increased. In low level of protein excretion (<500mg) the two tests had acceptable limits of agreement (-122.3 to 194.7 mg) but in subnephrotic range (500-3400mg), Bland-Altman plot showed clinically inacceptable wide limits of concordance (-438.1 to 1198.8 mg).
Using ROC curves, the spot urine P/C ratio discriminant values of 270.3 mg/g (sensitivity (S) 89.2%, specificity (E) 91.7%, area under the curve (AUC) 0.976), 336.5 mg/g (S 93.8%, E 96.6%, AUC 0.990), 655 mg/g (S 87%, E 92.1%, AUC 0.953) and 931 mg/g (S 100%, E 91.7%, AUC 0.958) predicted 24h protein equivalent “thresholds” of >300, 500, 1000 and 1500mg, respectively.
Conclusions Spot urine P/C ratio and 24h protein have good agreement in low protein levels (<500 mg). In subnephrotic protein range (500-3400), even though we found a good correlation between the two methods, the P/C ratio seems to be a weak predictor of 24h protein showing unacceptable limits of agreement.
Disclosure of Interest None declared