CLINICAL INVESTIGATION
Urinary Protein/Creatinine Ratio Weighted by Estimated Urinary Creatinine Improves the Accuracy of Predicting Daily Proteinuria Chun-Fan Chen, MD, Wu-Chang Yang, MD, Chih-Yu Yang, MD, Szu-Yuan Li, MD, Shuo-Ming Ou, MD, Yung-Tai Chen, MD, Chia-Jen Shih, MD, Chih-Chiang Chien, MD, Min-Chi Chen, PhD, Yu-Jen Wang, RN and Chih-Ching Lin, MD, PhD
Abstract: Background: The spot urine protein/creatinine ratio (UPCR) is proposed to be a substitute for 24-hour urinary protein (24h-UP). This study is aimed to determine whether the predictive accuracy of 24h-UP using UPCR can be improved by simply multiplying estimated daily urine creatinine excretion (eUCr) and UPCR together. Methods: This study enrolled 120 participants to investigate the correlation between spot UPCR and 24h-UP. Three sets of spot urine samples were randomly collected throughout the day and night, along with the first morning void. UPCR was weighted by eUCr to investigate the improvement of accuracy in using spot urine samples to predict 24hUP. Results: There were strong correlation and concordance between UPCR and 24h-UP irrespective of the time of spot urine sampling, and the correlation, concordance and agreement were improved after multiplying the UPCR value by the eUCr. Greater improvement was found in the subgroups with measured daily urine creatinine excretion #0.8 g/d and $1.2 g/d. Conclusions: This investigation demonstrated that multiplying UPCR by eUCr can improve the accuracy of only using UPCR to predict 24h-UP. Key Indexing Terms: Spot urine protein-to-creatinine ratio; Urine PCR; UPCR; Daily urine protein; 24-hour urine protein. [Am J Med Sci 2015;349(6):477–487.]
U
rine biochemistry is one of the most important tests in the field of nephrology, and urine biochemical assessments can facilitate accurate diagnosis of kidney diseases. Urine protein, a major characteristic of most kidney diseases, is highly associated with the rate of kidney function decline.1 Accurate quan-
From the School of Medicine (C-FC, W-CY, C-YY, S-YL, S-MO, Y-TC, C-JS, Y-JW, C-CL), National Yang-Ming University, Taipei, Taiwan; Division of Nephrology (C-FC), Department of Medicine, National Yang-Ming University Hospital (C-FC), Yilan, Taiwan; Division of Nephrology (W-CY, C-YY, S-YL, S-MO, Y-JW, C-CL), Department of Medicine, Taipei Veterans General Hospital, Taipei, Taiwan; Division of Nephrology (Y-TC), Department of Medicine, Taipei City Hospital, He-Ping Branch, Taipei, Taiwan; Division of Nephrology (C-JS), Department of Medicine, Yuan-Shan Branch, Taipei Veterans General Hospital, Yilan, Taiwan; Department of Nephrology (C-CC), Department of Medical Research, Chi-Mei Medical Center, Tainan, Taiwan; Department of Food Nutrition (C-CC), Chung Hwa University of Medical Technology, Tainan, Taiwan; and Department of Public Health (M-CC), Biostatistics Consulting Center, School of Medicine, Chang Gung University, Taoyuan, Taiwan. Submitted November 17, 2014; accepted in revised form March 25, 2015. The authors have no financial or other conflicts of interest to disclose. Supported by intramural Grants (V99A-142, V101C-188, V102C-060, V103C-043, V104C-026) and Grants for the Integrated Genome Project (V102E2-001) from Taipei Veterans General Hospital, and the National Science Council (NSC101–2314-B010–024-MY3) in Taiwan. Correspondence: Chih-Ching Lin, MD, PhD, Division of Nephrology, Department of Medicine, Taipei Veterans General Hospital, No. 201, Section 2, Shih-Pai Road, Taipei, Taiwan 112, China (E-mail: lincc2@ vghtpe.gov.tw).
The American Journal of the Medical Sciences
tification of urine protein can help doctors to diagnose kidney disease, make therapeutic decision and monitor therapeutic effects. Traditionally, doctors have used 24-hour timed urine collection to determine the amount of urine protein excretion, thus avoiding the potential fluctuations caused by daily activity. However, this procedure, especially in outpatients, is timeconsuming, cumbersome and frequently unreliable because of missed collection of urine samples. The spot urine protein-tocreatinine ratio (UPCR) has been used as a substitute for timed urine protein collection on the basis of the assumption that the urine protein excretion rate, which is proportional to the urine creatinine excretion, is relatively constant throughout the day.2 Several studies have confirmed the high correlation between the UPCR and 24-hour urine protein (24h-UP) with different regression coefficients (r 5 0.92–0.97).2–5 The National Kidney Foundation has also recommended that untimed (spot) urine samples should be used to detect and monitor proteinuria in adults or children and that it is not necessary to perform timed urine collection (overnight or 24-hour).6 However, the amount of urine creatinine excretion per day has been found to be highly variable between individuals, depending on age, gender and body weight,7 and the differences may influence the accuracy of the UPCR in predicting 24h-UP. In patients with extremely high or low daily urine creatinine excretion, the UPCR may be even double or half of the 24h-UP. As a result, the UPCR should not be used for initial risk stratification in glomerular diseases such as membranous nephropathy. In addition, the time of spot urine sampling and the use of medications may potentially contribute to differences in the values for UPCR and 24h-UP. This study aims primarily to determine whether there is a high correlation, concordance and agreement between UPCR and 24h-UP values and to examine the circadian fluctuations of the UPCR in patients with chronic kidney disease. A second aim is to investigate whether the accuracy of using UPCR in predicting 24h-UP would be improved after multiplying the UPCR value by the estimated daily urine creatinine excretion (eUCr) based on the Cockcroft-Gault equation,7 especially in patients with extremely high or low daily urine creatinine excretion. Finally, the authors investigate whether the differences in the amount of 24h-UP, renal function and use of reninangiotensin-aldosterone system (RAAS) blockers affect the correlation between the UPCR and 24h-UP.
MATERIALS AND METHODS Patient Selection The study group consisted of 120 participants who met the following inclusion criteria: (1) chronic kidney disease with stable renal function for 3 months before the commencement of the study and (2) daily urine protein level higher than 100 mg or
Volume 349, Number 6, June 2015
477
Chen et al
UPCR higher than 0.1 (g/g). These participants were regularly monitored by nephrology outpatient services in Taipei Veterans General Hospital. All enrolled participants were asked to maintain their regular dietary intake and medications for underlying diseases. Participants who were unsure of accurate compliance in collecting urine samples or who had collected 24h-UP with less than 100 mg per day were excluded from the analysis. Furthermore, participants with the ratio of measured daily urine creatinine excretion (mUCr) and eUCr more than 1.5 or less than 0.5 were also excluded. After obtaining informed consent from every participant in this study, urine samples were collected according to the following protocol. Study Protocol for Urine Collection and Testing A plastic bucket with cover was provided to each participant participating in this study. Participants had to void before the study began, and 24-hour urine collection began at 7:00 AM on the first morning of the study. Three sets of spot urine samples, each containing 10 mL of urine, were collected at random times throughout the day (7:00 AM to 3:00 PM; labeled as “Random daytime”) and night (3:00 PM to 11:00 PM; labeled as “Random nighttime”) and at 7:00 AM the next morning (labeled as “First morning”). These urine samples were then put in the plastic bucket and returned to the laboratory together with the 3 spot urine samples immediately after the urine collection is complete. In the meantime, confirmation of successful and complete collection of all urine samples was sought from the participants. The 24-hour urine samples and 3 sets of spot urine samples were analyzed for creatinine and total protein levels in the morning completing urine collection. Urine protein was measured by the pyrogallol red-molybdate method, and urine creatinine was measured by the autoanalyzer method with the Jaffe reaction. The general data included age, gender, body weight (at the time of study), serum creatinine concentration (SCr), body mass index and the coincidence of diabetes mellitus and hypertension were recorded. Medical records were reviewed, and the use or nonuse of RAAS blockers was confirmed with the participants. The estimated glomerular filtration rate (eGFR) was calculated by simplified MDRD formula, and the staging of chronic kidney disease was determined according to the K/DOQI guideline.6 Statistical Analyses Three statistical approaches were used to compare UPCR from different times of collection and 24h-UP. First, linear regression analysis was used to yield a regression equation: y 5 a + bx, where UPCR served as “x” variable and 24h-UP served as “y” variable. Both the slope (b) and correlation coefficient (r2) between individual UPCR and 24h-UP measurements were evaluated. The “b” value of the regression equation represents the accuracy of different methods used to predict 24h-UP. Any method with a “b” value close to 1 has a high accuracy to predict 24h-UP. Second, concordance correlation coefficient (CCC)8 was used to evaluate the agreement between UPCR and 24h-UP. A CCC close to 1 shows strong concordance. Third, calibration plot9 was used to verify the agreement between UPCR and 24h-UP. The x axis shows the value of 24h-UP, whereas the y axis represents the ratio of UPCR and 24h-UP (UPCR/24h-UP) in this plot. A mean value of ratio close to 1 represents high accuracy of using UPCR to predict 24h-UP, and smaller
478
TABLE 1. Baseline demographic and clinical characteristics of the study group Variables Values Age (yrs; mean 6 SD) Gender Male, n (%) Female, n (%) Body weight (kg; mean 6 SD) SCr (mg/dL; mean 6 SD) Body mass index (kg/m2; mean 6 SD) CKD, n (%) Stage 1 (eGFR $90) Stage 2 (eGFR: 60–89) Stage 3 (eGFR: 30–59) Stage 4 (eGFR: 15–29) Stage 5 (eGFR #15) Diabetes mellitus, n (%) Hypertension, n (%) RAAS blockers use, n (%)
67 6 14.83 57 (63) 33 (37) 65.57 6 12.91 2.84 6 1.93 24.94 6 4.77 7 7 29 29 18 38 75 56
(7.8) (7.8) (32.2) (32.2) (20.0) (42.2) (83.3) (62.2)
Staging of CKD is based on eGFR by MDRD formula: GFR (mL$min21$1.73 m22) 5 175 3 (SCr)21.154 3 (age)20.203 3 (0.742, if female) 3 (1.212, if African American). CKD, chronic kidney disease; eGFR, estimated glomerular filtration rate; MDRD, modification of diet in renal disease; RAAS, reninangiotensin-aldosterone axis; SCr, serum creatinine.
range between upper and lower limits of agreement (61.96 SD) shows better agreement. To eliminate the effect of variable urine creatinine excretion between individuals on the analysis, UPCR was multiplied by the eUCr value obtained by the CockcroftGault equation7: 24-hour urine creatinine excretion (g) 5 {[140 2 age (years)] 3 weight (kg)} / 5000 (30.85 in females). The eUCr-weighted UPCR (eUPCR) was then used instead of UPCR in analyzing all participants and subgroups with mUCr #0.8 g/d, mUCr $1.2 g/d, mUCr .0.8 g/d and ,1.2 g/d. Subgroup analysis according to the amount of urine protein (24h-UP .3g/d and #3g/d) and renal
TABLE 2. Regression parameters and CCC of the study group UPCR vs. 24h-UP eUPCR vs. 24h-UP
All patients (n 5 90) Random daytime Random nighttime First morning
b
r2
CCC
b
r2
CCC
0.848
0.857
0.918
0.942
0.944
0.971
0.866
0.882
0.933
0.957
0.966
0.982
0.869
0.846
0.916
0.970
0.935
0.967
Regression equation: 24h-UP 5 a + (b 3 UPCR [or eUPCR]). For the regression coefficient, all P , 0.001. CCCs, concordance correlation coefficients; 24h-UP, timed 24-hour urine protein; UPCR, urine protein-to-creatinine ratio; eUPCR, estimated daily urine creatinine excretion–weighted UPCR; random daytime, spot urine between 7:00 AM and 3:00 PM; random nighttime, spot urine between 3:00 AM and 11:00 PM; first morning, first voided at 7:00 AM the next morning.
Volume 349, Number 6, June 2015
Urine Protein/Creatinine Ratio and 24-Hour Urinary Protein
FIGURE 1. The relationship between the ratio of 24h-UP/UPCR (eUPCR) and 24h-UP in different times of urine collection. After using eUPCR (right panel) instead of UPCR (left panel), the mean value of ratio and limit of agreement were improved in all 3 spot urine samples. The eUPCR in nighttime urine sample yielded the best mean value of ratio and limit of agreement. eUPCR, estimated daily urine creatinine excretion–weighted UPCR; UPCR, urine protein-to-creatinine ratio; 24h-UP, 24-hour urinary protein.
function (eGFR $60 mL$min21$1.73 m22 and ,60 mL$min21$1.73 m22) were also performed. Finally, participants were divided into 2 subgroups, depending on whether they were receiving RAAS blockers, such as angiotensinCopyright © 2015 by the Southern Society for Clinical Investigation.
converting enzyme inhibitors, angiotensin type 1 receptor antagonists or spironolactone during the urine sampling period. The correlation, concordance and agreement between the UPCR and 24h-UP were then assessed in both
479
Chen et al
TABLE 3. Regression parameters and CCCs in subgroups according to mUCr UPCR vs. 24h-UP b mUCr #0.8 (n 5 31) Random daytime Random nighttime First morning mUCr $1.2 (n 5 16) Random daytime Random nighttime First morning mUCr ,1.2 and .0.8 (n 5 43) Random daytime Random nighttime First morning
r
2
eUPCR vs. 24h-UP
CCC
b
r2
CCC
0.688 0.709 0.730
0.902 0.917 0.888
0.855 0.876 0.871
0.916 0.954 0.972
0.930 0.950 0.909
0.960 0.973 0.952
1.250 1.199 1.164
0.980 0.978 0.938
0.940 0.953 0.922
0.991 0.966 0.927
0.975 0.984 0.937
0.987 0.991 0.966
0.875 0.888 0.888
0.950 0.961 0.921
0.967 0.975 0.956
0.914 0.953 1.003
0.931 0.963 0.957
0.963 0.979 0.977
Regression equation: 24h-UP 5 a + (b 3 UPCR [or eUPCR]). For the regression coefficient, all P , 0.001. CCCs, concordance correlation coefficients; mUCr, measured daily urine creatinine excretion (g/d); UPCR, urine protein-to-creatinine ratio; 24h-UP, timed 24-hour urine protein; eUPCR, estimated daily urine creatinine excretion–weighted UPCR; random daytime, spot urine during 7:00 AM to 3:00 PM; random nighttime, spot urine between 3:00 AM and 11:00 PM; first morning, first voided at 7:00 AM the next morning.
subgroups. All statistical tests were two-tailed, and significance was set at P , 0.05.
RESULTS Patient Characteristics One hundred and twenty participants examined by the nephrology outpatient services in Taipei Veterans General Hospital were initially enrolled in this study. Of these, 9 participants were unsure of accuracy in urine collection, 5 participants were sure of accuracy in urine collection but the 24h-UP was less than 100 mg per day and for 16 participants the ratio of mUCr to eUCr was above 1.5 or below 0.5. After excluding the above mentioned participants from the statistical analysis, 90 participants completed the protocol for urine collection and these participants’ baseline characteristics were listed in Table 1. Statistical Analysis There were both high accuracy of using UPCR in predicting 24h-UP and strong correlation between UPCR and 24h-UP. The correlation between UPCR and 24h-UP was slightly higher in random nighttime UPCR compared with random daytime UPCR and first morning UPCR. However, the first morning UPCR had the highest b value, representing the best accuracy in predicting 24h-UP. When the eUCr, calculated by the Cockcroft-Gault formula, was multiplied to the UPCR (eUPCR), the b value for all 3 spot urine samples increased significantly, as did the r2. The agreement between UPCR and 24h-UP, expressed by CCC, showed strong concordance and was the highest in the random nighttime UPCR. The concordance between eUPCR and 24h-UP was improved irrespective of the time of urine sampling (Table 2). In the calibration plot analysis, mean value of the ratio eUPCR/24h-UP was closer to 1 compared with that of UPCR/24h-UP. Besides, smaller range of limit of agreement (LoA) and higher percentage of individuals with UPCR within 15%, 20% and 30% of 24h-UP were noted in the analysis of eUPCR and 24h-UP, as illustrated in Figure 1.
480
When participants were divided into 3 subgroups on the basis of their mUCr levels (#0.8, $1.2, .0.8 and ,1.2 g/d), both the r 2 of the UPCR and eUPCR with 24h-UP were comparably high. The b value increased in the subgroups with mUCr #0.8 g/d and mUCr .0.8 and ,1.2 g/d but decreased in the subgroup with mUCr $1.2 g/d. Analysis of CCC showed stronger concordance of using eUPCR instead of UPCR in all 3 subgroups (Table 3). In calibration plot, mean value of the ratio eUPCR/24h-UP was much closer to 1 compared with that of UPCR/24h-UP in all 3 subgroups. The range of LoA was also narrower and the percentage of individuals with UPCR within 15%, 20% and 30% of 24h-UP was higher using eUPCR instead of UPCR except in the subgroup of mUCr $1.2 g/d, as illustrated in Figure 2. Subgroup analysis according to the amount of urine protein (24h-UP .3g/d and #3 g/d) showed relatively poor b value, r 2 and CCC between UPCR and 24h-UP. Using eUPCR instead of UPCR enhanced the b value, r 2 and CCC regardless the amount of 24h-UP was more or less than 3 g/d (Table 4). In the subgroup of 24h-UP .3 g/d, the mean value of ratio was closer to 1 and the range of LoA was narrower compared with the subgroup of 24h-UP #3g/d. Similarly, there were better results of mean value of ratio, the range of LoA and higher percentage of individuals with UPCR within 15%, 20% and 30% of 24h-UP if using eUPCR instead of UPCR (Figure 3). Subgroup analysis according to eGFR $45 mL$min21$1.73 m22 or ,45 mL$min21$1.73 m22 was performed, but there were inconsistencies of different times of urine collection in both subgroup. As the authors expected, the b value, r2 and CCC were improved between eUPCR and 24h-UP in both subgroups (Table 5). In the subgroup of eGFR $45 mL$min21$1.73 m22, the mean value of ratio was closer to 1 with narrower range of LoA in calibration plot. Using eUPCR instead of UPCR resulted in narrower range of LoA but inferior mean value of ratio in this subgroup. The percentage of individuals with UPCR within 15%, 20% and 30% of 24h-UP was higher in eUPCR analysis (Figure 4). Volume 349, Number 6, June 2015
Urine Protein/Creatinine Ratio and 24-Hour Urinary Protein
FIGURE 2. The relationship between the ratio of 24h-UP/UPCR (eUPCR) and 24h-UP in the samples of first-voided urine next morning, subgroup by mUCr #0.8 g/d, $1.2 g/d, and .0.8 and ,1.2 g/d. After using eUPCR (right panel) instead of UPCR (left panel), the mean value of ratio was closer to 1, especially in subgroups with mUCr #0.8 g/d and mUCr $1.2 g/d. The range between limit of agreement was narrower by eUPCR except in the subgroup of mUCr $1.2 g/d. eUPCR, estimated daily urine creatinine excretion–weighted UPCR; UPCR, urine protein-tocreatinine ratio; 24h-UP, 24-hour urinary protein; mUCr, measured daily urine creatinine excretion.
Copyright © 2015 by the Southern Society for Clinical Investigation.
481
Chen et al
TABLE 4. Regression parameters and CCCs in subgroups according to measured daily urine protein UPCR vs. 24h-UP eUPCR vs. 24h-UP 24h-UP #3 g/d (n 5 69) Random daytime Random nighttime First morning 24h-UP .3 g/d (n 5 21) Random daytime Random nighttime First morning
b
r2
CCC
b
r2
CCC
0.665 0.706 0.678
0.786 0.755 0.699
0.824 0.836 0.796
0.767 0.869 0.887
0.784 0.795 0.780
0.876 0.890 0.882
0.698 0.739 0.685
0.618 0.687 0.614
0.772 0.816 0.776
0.899 0.929 0.886
0.858 0.921 0.832
0.925 0.959 0.911
Regression formula: 24h-UP 5 a + (b 3 UPCR [or eUPCR]). For the regression coefficient, all P , 0.001. CCCs, concordance correlation coefficients; UPCR, urine protein-to-creatinine ratio; 24h-UP, timed 24-hour urine protein; eUPCR, estimated daily urine creatinine excretion–weighted UPCR; random daytime, spot urine during 7:00 AM to 3:00 PM; random nighttime, spot urine between 3:00 AM and 11:00 PM; first morning, first voided at 7:00 AM the next morning.
FIGURE 3. The relationship between the ratio of 24h-UP/UPCR (eUPCR) and 24h-UP in the samples of first-voided urine next morning, subgroup by 24h-UP of #3 g/d and .3 g/d. After using eUPCR (right panel) instead of UPCR (left panel), the mean value of ratio and limit of agreement were improved in both subgroups. The range between limit of agreement in the subgroup by 24h-UP of #3 g/d was wider than another subgroup. eUPCR, estimated daily urine creatinine excretion–weighted UPCR; UPCR, urine protein-to-creatinine ratio; 24h-UP, 24-hour urinary protein.
482
Volume 349, Number 6, June 2015
Urine Protein/Creatinine Ratio and 24-Hour Urinary Protein
TABLE 5. Regression parameters and CCCs in subgroups according to eGFR UPCR vs. 24h-UP b 21
r
2
eUPCR vs. 24h-UP CCC
b
r2
CCC
22
eGFR $45 mL$min $1.73 m (n 5 23) Random daytime Random nighttime First morning eGFR ,45 mL$min21$1.73 m22 (n 5 67) Random daytime Random nighttime First morning
0.796 0.891 1.173
0.758 0.843 0.788
0.866 0.918 0.848
0.920 0.947 1.225
0.935 0.971 0.899
0.965 0.984 0.904
0.863 0.863 0.851
0.879 0.889 0.876
0.928 0.936 0.925
0.950 0.960 0.951
0.945 0.964 0.957
0.972 0.981 0.977
Regression formula: 24h-UP 5 a + (b 3 UPCR [or eUPCR]). For the regression coefficient, all P , 0.001. eGFR is calculated by MDRD formula: GFR (mL$min21$1.73 m22) 5 175 3 (SCr)21.154 3 (age)20.203 3 (0.742, if female) 3 (1.212, if African American). CCCs, concordance correlation coefficients; eGFR, estimated glomerular filtration rate; MDRD, modification of diet in renal disease; SCr, serum creatinine; UPCR, urine protein-to-creatinine ratio; 24h-UP, timed 24-hour urine protein; eUPCR, estimated daily urine creatinine excretion–weighted UPCR; random daytime, spot urine during 7:00 AM to 3:00 PM; random nighttime, spot urine between 3:00 AM and 11:00 PM; first morning, first voided at 7:00 AM the next morning.
When participants were divided into 2 subgroups on the basis of the use of RAAS blockers, r2 between the UPCR and 24h-UP was higher in the subgroup taking RAAS blockers than in the subgroup not taking RAAS blockers, as did b value and CCC. Irrespective of the use of RAAS blockers, the correlation and concordance were improved after using the eUPCR instead of the UPCR (Table 6). Similarly, using the eUPCR improved the accuracy and agreement in both subgroups in the calibration plot (Figure 5).
DISCUSSION This study was conducted to investigate the accuracy of using UPCR as a substitute of 24h-UP and the correlation between UPCR and 24h-UP. Traditionally, 24-hour urine protein collection has been the standard for assessing and quantifying proteinuria in patients with kidney diseases. A strong correlation between the UPCR and 24h-UP has been demonstrated in many studies assessing various etiologies of renal dysfunction, such as primary glomerulonephritis, diabetic nephropathy, systemic lupus erythematosus and renal transplantation in adults and children.2–5,10–17 Some studies have concluded that the best approach to measuring UPCR is to use the first daily voided urine,12,16 but some suggest obtaining spot urine readings during normal daily activity.2,18,19 In this study, there was very high correlation and concordance between the UPCR of all 3 spot urine samples and 24h-UP. Besides, the random nighttime UPCR seemed to be the best prediction of 24h-UP. The variability in UPCR throughout a 24-hour period mainly comes from the change of urine protein excretion rather than the change of urine creatinine excretion in each individual patient.20 The high degree of variability in spot UPCR can lead to error in the management of the individual patient. The authors assume that the concentration of urine protein may be altered by activity and positional change, and most urine is produced in daytime and nighttime rather than in sleep. As a result, the random daytime or nighttime UPCR has better accuracy and correlation in predicting 24h-UP compared with the first morning void UPCR. The ratio of UPCR and 24h-UP was widely distributed in the calibration plot and the mean value of ratio was 1.17 to Copyright © 2015 by the Southern Society for Clinical Investigation.
1.24, indicating potential bias of 17% to 24% introduced by UPCR. Daily urine creatinine excretion was widely variable among different patients and may lower the accuracy of using UPCR to predict the value of 24h-UP between cohorts. However, the above mentioned errors could be offset by adequate randomization and appropriate correction by multiplying by eUCr, which thus resulted in a higher correlation of 24h-UP with eUPCR than with spot UPCR. The bias may have great impact if using spot UPCR in establishing the diagnosis of nephrotic syndrome, deciding whether renal biopsy is needed and choosing the recommended therapeutic regimen classified by 24h-UP. Multiplying UPCR by eUCr (estimated urine creatinine excretion) attenuated the bias without the need to collect 24-hour urine in this study. The ratio of mUCr to eUCr has been used to validate the reliability of urine collection to confirm successful 24-hour urine collection in previous studies. Samples with measured-toestimated ratios in the range 0.9 to 1.1 are considered complete collections.2,21 However, these studies did not emphasize the bias introduced by variable daily urine creatinine excretion between individuals, which might attenuate the accuracy of the UPCR to predict 24h-UP. In patients with extremely higher or lower daily urine creatinine excretion, using UPCR may lead to overt estimation or underestimation of 24h-UP compared with patients with urine creatinine excretion close to 1 g/d. Based on the understanding that difference in age, gender and body weight may affect daily urine creatinine excretion and hence the accuracy of the UPCR, the authors tried to use eUCr-weighted UPCR (eUPCR) instead of simple UPCR to investigate the correlation and determined the accuracy of eUPCR in predicting 24h-UP. There have been several equations derived to estimate creatinine excretion rate (CER), and the bias on average, precision and accuracy between these equations is assessed by Ix et al.22 Among the estimating equations compared in the study, the Cockcroft-Gault equation had statistically significantly greater bias and tended to overestimate CER in individuals with the highest CER values. As a result, the eUCr obtained by the Cockcroft-Gault equation may have potential errors and thus lead to incorrect estimation of daily urine protein, however, which could be improved by replacing the actual body weight with estimated lean body weight in obese subjects. Furthermore, the
483
Chen et al
FIGURE 4. The relationship between the ratio of 24h-UP/UPCR (eUPCR) and 24h-UP in the samples of first-voided urine next morning, subgroup by eGFR $45 mL$min21$1.73 m22 and ,45 mL$min21$1.73 m22. After using eUPCR (right panel) instead of UPCR (left panel), the limit of agreement was improved in both subgroups. However, the mean value of ratio was inferior by eUPCR in subgroup with eGFR $45 mL$min21$1.73 m22. eGFR, estimated glomerular filtration rate; eUPCR, estimated daily urine creatinine excretion– weighted UPCR; UPCR, urine protein-to-creatinine ratio; 24h-UP, 24-hour urinary protein.
Cockcroft-Gault equation is easier to be calculated and used in clinical approach, avoiding the process of complex computing work. Using the weighted “eUPCR,” the authors obtained a stronger correlation and concordance in all 3 of the spot urine samples no matter what time the spot urine was collected. Besides, the mean value of ratio was closer to 1 and the range between LoA was narrower in calibration plot analysis after using the eUPCR instead of traditional UPCR. Similar concept was published in the study (the PREVEND study) of using estimated CER (eCER) multiplied to urine albumin-to-creatinine ratio (ACR) for assessment of albuminuria.23 Various eCER computed from the Ix,22 Ellam24 and Walser25 equations were used, and the study concluded that ACR multiplied by eCER was more accurate than ACR alone in assessing albuminuria. Multiplying UPCR of 24hour urine collection by eUCr to estimate the magnitude of 24hUP was also recommended by Glassock et al.26 Although they did not recommend using random spot or first morning UPCR to determine 24h-UP, the authors demonstrated improvement of
484
accuracy using eUPCR instead of UPCR to predict 24h-UP in spot urine sample. When participants were divided into subgroups according to mUCr values, the authors observed a greater increase in accuracy in subgroups with mUCr values #0.8 g/d or $1.2 g/d after utilization of eUPCR instead of UPCR. This finding has not been emphasized in previous studies. The difference between UPCR and 24h-UP in the subgroup with mUCr #0.8 g/d may be counteracted by the difference in the subgroup with mUCr $1.2 g/d and mask a potential bias in predicting 24h-UP. As a result, using eUPCR can improve the correlation and the accuracy between UPCR and 24h-UP, particularly in patients with daily urine creatinine excretion higher than 1.2 g/d or lower than 0.8 g/d. UPCR was believed to be unreliable for estimating 24hUP over subnephrotic range (24h-UP #3g/d) in lupus nephritis.27–29 The authors investigated if this occurred in general chronic kidney patients and thus divided the participants into Volume 349, Number 6, June 2015
Urine Protein/Creatinine Ratio and 24-Hour Urinary Protein
TABLE 6. Regression parameters and CCCs in subgroups according to use of RAAS blockers UPCR vs. 24h-UP b RAAS blockers (+) (n 5 56) Random daytime Random nighttime First morning RAAS blockers (2) (n 5 34) Random daytime Random nighttime First morning
r
2
eUPCR vs. 24h-UP
CCC
b
r2
CCC
0.884 0.887 0.892
0.871 0.912 0.897
0.931 0.951 0.944
1.006 0.991 0.975
0.951 0.978 0.945
0.974 0.998 0.971
0.794 0.825 0.825
0.840 0.825 0.746
0.894 0.898 0.858
0.858 0.902 0.965
0.956 0.950 0.916
0.965 0.971 0.957
Regression formula: 24h-UP 5 a + (b 3 UPCR [or eUPCR]). For the regression coefficient, all P , 0.001. CCCs, concordance correlation coefficients; 24h-UP, timed 24-hour urine protein; UPCR, urine protein-to-creatinine ratio; eUPCR, estimated daily urine creatinine excretion–weighted UPCR; random daytime, spot urine between 7:00 AM and 3:00 PM; random nighttime, spot urine between 3:00 AM and 11:00 PM; first morning, first voided at 7:00 AM the next morning; RAAS, renin-angiotensin-aldosterone axis.
FIGURE 5. The relationship between the ratio of 24h-UP/UPCR (eUPCR) and 24h-UP in the samples of first-voided urine next morning, subgroup use or nonuse of RAAS blockers. Both the mean value of ratio and limit of agreement in the subgroup with RAAS blockers were superior to another subgroup. Using eUPCR enhanced the accuracy and agreement in both subgroups. eUPCR, estimated daily urine creatinine excretion–weighted UPCR; UPCR, urine protein-to-creatinine ratio; 24h-UP, 24-hour urinary protein; RAAS, renin-angiotensin-aldosterone axis. Copyright © 2015 by the Southern Society for Clinical Investigation.
485
Chen et al
subgroups with 24h-UP .3g/d or #3g/d. The correlation between UPCR and 24h-UP was better in the subgroup of 24h-UP #3g/d, but the accuracy and agreement showed the opposite results. Overall, subgroup analysis showed less optimal results compared with the results over whole range of proteinuria. Whether the 24h-UP was higher or less than 3 g/d, eUPCR improved the accuracy in predicting 24h-UP together with better correlation and agreement. In the above study using eCER multiplied to urine ACR for assessment of albuminuria,23 the participants were community-living individuals from general population with mean eGFR of 84 6 15 mL$min21$1.73 m22. In contrast, this study enrolled participants with chronic kidney disease and a lower mean eGFR of 35.01 6 25.15 mL$min21$1.73 m22. The authors divided the participants to subgroups with eGFR ≧45 mL$min21$1.73 m22 or ,45 mL$min21$1.73 m22 in investigation of whether lower eGFR will influence the association between UPCR and 24h-UP or not. Inconsistent results were noted with higher correlation but lower accuracy and agreement of using UPCR to predict 24h-UP in the subgroup with relatively poor renal function. Similarly, the eUPCR was a better predictor of 24h-UP irrespective of preserved or impaired renal function. RAAS blockers (including angiotensin-converting enzyme inhibitors, angiotensin type 1 receptor antagonists and spironolactone) are widely used in patients with chronic kidney disease for both slowing down the renal function decline and decreasing the degree of proteinuria.30–33 In this study, the authors investigated if the use of RAAS blockers could influence the reliability of UPCR in predicting 24h-UP. Among participants using RAAS blockers, the correlation and accuracy were higher than in participants who did not take RAAS blockers. RAAS blockers can decrease glomerular hyperfiltration and cytokine secretion, preserving the integrity of glomerular structure and the function of the basement membrane. As a result, RAAS blockers may provide a relatively constant proteinto-creatinine filtration fraction by this mechanism. Further investigation is needed to fully understand the mechanisms underlying the authors’ findings. There are 4 limitations to this study. First, most participants (n 5 51) had a ratio of mUCr to eUCr higher than 1.1 or lower than 0.9, despite the fact that all the finally enrolled participants insisted they collected the 24-hour urine samples completely. It was2,10 considered to indicate overt collection or undercollection of urine samples in previous studies.2,21 But under certain conditions, urine creatinine excretion can significantly decrease (low meat diet, muscle wasting or limb amputation) or increase (high meat diet, increased muscle mass, creatine supplement or fenofibrate therapy).34 As a result, a ratio of mUCr to eUCr higher than 1.1 or lower than 0.9 might suggest either that errors occurred during 24-hour urine collection or just the influence of urine creatinine variation. The authors excluded participants with a ratio of mUCr to eUCr higher than 1.5 or lower than 0.5 before analysis exactly for the reason that these participants showed higher likelihood of error in urine collection. Second, the “spot” urine sample in the study was a urine collection of variable duration in urinary bladder rather than a “snapshot” urine sample. Participants were not obligated to have a snapshot urine sample or fixed time urine collection in bladder to simulate the situation in clinical practice. The variable duration of “spot” urine collection between participants together with the diurnal variations in the PC ratio may influence the accuracy of using UPCR in predicting 24h-UP. The closer the “spot” urine collection approaches 24 hours, the closer the UPCR of “spot” urine collection approaches that of
486
a complete 24-hour urine collection.21 However, the advantage of using eUPCR was existed and not vanished by these variations. Third, the urine samples were not refrigerated during the time of experiment. Despite the urine specimens being returned and analyzed in laboratory immediately after completing urine collection, differences in degradation of albumin may introduce bias in the analysis of UPCR and 24h-UP. Lastly, the number of participants in each subgroup was not equal and might lead to bias in statistics. A prospective study with larger number of participants specifically designed for subgroup analysis will further demonstrate whether the amount of 24h-UP, renal function and use of RAAS blockers will influence the relationship between UPCR and 24h-UP.
CONCLUSIONS UPCR is a valid substitute for timed urine protein collection, and the result is less affected by the time of the spot urine sampling. However, UPCR may lead to bias in predicting 24h-UP because of variable urine creatinine excretion between individuals. Using eUPCR overcomes the shortcoming and is more accurate than UPCR in assessment of 24h-UP. Accordingly, the authors recommend the use of eUPCR instead of UPCR in clinical approach to attenuate the bias introduced by variable creatinine excretion between individuals. REFERENCES 1. Keane WF. Proteinuria: its clinical importance and role in progressive renal disease. Am J Kidney Dis 2000;35:S97–105. 2. Ginsberg JM, Chang BS, Matarese RA, et al. Use of single voided urine samples to estimate quantitative proteinuria. N Engl J Med 1983; 309:1543–6. 3. Schwab SJ, Christensen RL, Dougherty K, et al. Quantitation of proteinuria by the use of protein-to-creatinine ratios in single urine samples. Arch Intern Med 1987;147:943–4. 4. Lane C, Brown M, Dunsmuir W, et al. Can spot urine protein/creatinine ratio replace 24 h urine protein in usual clinical nephrology? Nephrology 2006;11:245–9. 5. Ruggenenti P, Gaspari F, Perna A, et al. Cross sectional longitudinal study of spot morning urine protein:creatinine ratio, 24 hour urine protein excretion rate, glomerular filtration rate, and end stage renal failure in chronic renal disease in patients without diabetes. BMJ 1998;316:504–9. 6. National Kidney Foundation. K/DOQI clinical practice guidelines for chronic kidney disease: evaluation, classification, and stratification. Am J Kidney Dis 2002;39:S1–266. 7. Cockcroft DW, Gault MH. Prediction of creatinine clearance from serum creatinine. Nephron 1976;16:31–41. 8. Lin LI. A concordance correlation coefficient to evaluate reproducibility. Biometrics 1989;45:255–68. 9. van Stralen KJ, Jager KJ, Zoccali C, et al. Agreement between methods. Kidney Int 2008;74:1116–20. 10. Morgenstern BZ, Butani L, Wollan P, et al. Validity of proteinosmolality versus protein-creatinine ratios in the estimation of quantitative proteinuria from random samples of urine in children. Am J Kidney Dis 2003;41:760–6. 11. Salesi M, Karimifar M, Farajzadegan Z, et al. The protein-creatinine ratio in spot morning urine samples and 24-h urinary protein excretion in patients with systemic lupus erythematosus. Rheumatol Int 2009;29:503–7. 12. Xin G, Wang M, Jiao LL, et al. Protein-to-creatinine ratio in spot urine samples as a predictor of quantitation of proteinuria. Clin Chim Acta 2004;350:35–9.
Volume 349, Number 6, June 2015
Urine Protein/Creatinine Ratio and 24-Hour Urinary Protein
13. Rodrigo E, Pinera C, Ruiz JC, et al. Quantitation of 24-hour urine protein excretion in kidney transplant patients by the use of protein to creatinine ratio. Transplant Proc 2003;35:702. 14. Wahbeh AM, Ewais MH, Elsharif ME. Comparison of 24-hour urinary protein and protein-to-creatinine ratio in the assessment of proteinuria. Saudi J Kidney Dis Transpl 2009;20:443–7. 15. Antunes VV, Veronese FJ, Morales JV. Diagnostic accuracy of the protein/creatinine ratio in urine samples to estimate 24-h proteinuria in patients with primary glomerulopathies: a longitudinal study. Nephrol Dial Transplant 2008;23:2242–6. 16. Nagasako H, Kiyoshi Y, Ohkawa T, et al. Estimation of 24-hour urine protein quantity by the morning-urine protein/creatinine ratio. Clin Exp Nephrol 2007;11:142–6. 17. Torng S, Rigatto C, Rush DN, et al. The urine protein to creatinine ratio (P/C) as a predictor of 24-hour urine protein excretion in renal transplant patients. Transplantation 2001;72:1453–6. 18. Koopman MG, Krediet RT, Koomen GC, et al. Circadian rhythm of proteinuria: consequences of the use of urinary protein:creatinine ratios. Nephrol Dial Transplant 1989;4:9–14. 19. Dyson EH, Will EJ, Davison AM, et al. Use of the urinary protein creatinine index to assess proteinuria in renal transplant patients. Nephrol Dial Transplant 1992;7:450–2. 20. Koopman MG, Krediet RT, Zuyderhoudt FJ, et al. A circadian rhythm of proteinuria in patients with a nephrotic syndrome. Clin Sci (Lond) 1985;69:395–401. 21. Shidham G, Hebert LA. Timed urine collections are not needed to measure urine protein excretion in clinical practice. Am J Kidney Dis 2006;47:8–14. 22. Ix JH, Wassel CL, Stevens LA, et al. Equations to estimate creatinine excretion rate: the CKD epidemiology collaboration. Clin J Am Soc Nephrol 2011;6:184–91. 23. Abdelmalek JA, Gansevoort RT, Lambers Heerspink HJ, et al. Estimated albumin excretion rate versus urine albumin-creatinine ratio for the assessment of albuminuria: a diagnostic test study from the Prevention of Renal and Vascular Endstage Disease (PREVEND) Study. Am J Kidney Dis 2014;63:415–21.
Copyright © 2015 by the Southern Society for Clinical Investigation.
24. Fotheringham J, Campbell MJ, Fogarty DG, et al. Estimated albumin excretion rate versus urine albumin-creatinine ratio for the estimation of measured albumin excretion rate: derivation and validation of an estimated albumin excretion rate equation. Am J Kidney Dis 2014;63:405–14. 25. Walser M. Creatinine excretion as a measure of protein nutrition in adults of varying age. JPEN J Parenter Enteral Nutr 1987;11:73S–8S. 26. Glassock RJ, Fervenza FC, Hebert L, et al. Nephrotic syndrome redux. Nephrol Dial Transplant 2014;0:1–6. 27. Hebert LA, Birmingham DJ, Shidham G, et al. Random spot urine protein/creatinine ratio is unreliable for estimating 24-hour proteinuria in individual systemic lupus erythematosus nephritis patients. Nephron 2009;113:c177–82. 28. Birmingham DJ, Rovin BH, Shidham G, et al. Spot urine protein/creatinine ratios are unreliable estimates of 24 h proteinuria in most systemic lupus erythematosus nephritis flares. Kidney Int 2007;72: 865–70. 29. Birmingham DJ, Shidham G, Perna A, et al. Spot PC ratio estimates of 24-hour proteinuria are more unreliable in lupus nephritis than in other forms of chronic glomerular disease. Ann Rheum Dis 2014;73:475–6. 30. Lewis EJ, Hunsicker LG, Bain RP, et al. The effect of angiotensinconverting-enzyme inhibition on diabetic nephropathy. The Collaborative Study Group. N Engl J Med 1993;329:1456–62. 31. Andersen S, Tarnow L, Rossing P, et al. Renoprotective effects of angiotensin II receptor blockade in type 1 diabetic patients with diabetic nephropathy. Kidney Int 2000;57:601–6. 32. Brenner BM, Cooper ME, de Zeeuw D, et al. Effects of losartan on renal and cardiovascular outcomes in patients with type 2 diabetes and nephropathy. N Engl J Med 2001;345:861–9. 33. Bianchi S, Bigazzi R, Campese VM. Long-term effects of spironolactone on proteinuria and kidney function in patients with chronic kidney disease. Kidney Int 2006;70:2116–23. 34. Wilmer WA, Rovin BH, Hebert CJ, et al. Management of glomerular proteinuria: a commentary. J Am Soc Nephrol 2003; 14:3217–32.
487
Urinary Protein/Creatinine Ratio Weighted by Estimated Urinary Creatinine Improves the Accuracy of Predicting Daily Proteinuria Chun-Fan Chen, MD, Wu-Chang Yang, MD, Chih-Yu Yang, MD, Szu-Yuan Li, MD, Shuo-Ming Ou, MD, Yung-Tai Chen, MD, Chia-Jen Shih, MD, Chih-Chiang Chien, MD, Min-Chi Chen, PhD, Yu-Jen Wang, RN and Chih-Ching Lin, MD, PhD
Abstract: Background: The spot urine protein/creatinine ratio (UPCR) is proposed to be a substitute for 24-hour urinary protein (24h-UP). This study is aimed to determine whether the predictive accuracy of 24h-UP using UPCR can be improved by simply multiplying estimated daily urine creatinine excretion (eUCr) and UPCR together. Methods: This study enrolled 120 participants to investigate the correlation between spot UPCR and 24h-UP. Three sets of spot urine samples were randomly collected throughout the day and night, along with the first morning void. UPCR was weighted by eUCr to investigate the improvement of accuracy in using spot urine samples to predict 24hUP. Results: There were strong correlation and concordance between UPCR and 24h-UP irrespective of the time of spot urine sampling, and the correlation, concordance and agreement were improved after multiplying the UPCR value by the eUCr. Greater improvement was found in the subgroups with measured daily urine creatinine excretion #0.8 g/d and $1.2 g/d. Conclusions: This investigation demonstrated that multiplying UPCR by eUCr can improve the accuracy of only using UPCR to predict 24h-UP. Key Indexing Terms: Spot urine protein-to-creatinine ratio; Urine PCR; UPCR; Daily urine protein; 24-hour urine protein. [Am J Med Sci 2015;349(6):477–487.]
U
rine biochemistry is one of the most important tests in the field of nephrology, and urine biochemical assessments can facilitate accurate diagnosis of kidney diseases. Urine protein, a major characteristic of most kidney diseases, is highly associated with the rate of kidney function decline.1 Accurate quan-
From the School of Medicine (C-FC, W-CY, C-YY, S-YL, S-MO, Y-TC, C-JS, Y-JW, C-CL), National Yang-Ming University, Taipei, Taiwan; Division of Nephrology (C-FC), Department of Medicine, National Yang-Ming University Hospital (C-FC), Yilan, Taiwan; Division of Nephrology (W-CY, C-YY, S-YL, S-MO, Y-JW, C-CL), Department of Medicine, Taipei Veterans General Hospital, Taipei, Taiwan; Division of Nephrology (Y-TC), Department of Medicine, Taipei City Hospital, He-Ping Branch, Taipei, Taiwan; Division of Nephrology (C-JS), Department of Medicine, Yuan-Shan Branch, Taipei Veterans General Hospital, Yilan, Taiwan; Department of Nephrology (C-CC), Department of Medical Research, Chi-Mei Medical Center, Tainan, Taiwan; Department of Food Nutrition (C-CC), Chung Hwa University of Medical Technology, Tainan, Taiwan; and Department of Public Health (M-CC), Biostatistics Consulting Center, School of Medicine, Chang Gung University, Taoyuan, Taiwan. Submitted November 17, 2014; accepted in revised form March 25, 2015. The authors have no financial or other conflicts of interest to disclose. Supported by intramural Grants (V99A-142, V101C-188, V102C-060, V103C-043, V104C-026) and Grants for the Integrated Genome Project (V102E2-001) from Taipei Veterans General Hospital, and the National Science Council (NSC101–2314-B010–024-MY3) in Taiwan. Correspondence: Chih-Ching Lin, MD, PhD, Division of Nephrology, Department of Medicine, Taipei Veterans General Hospital, No. 201, Section 2, Shih-Pai Road, Taipei, Taiwan 112, China (E-mail: lincc2@ vghtpe.gov.tw).
The American Journal of the Medical Sciences
tification of urine protein can help doctors to diagnose kidney disease, make therapeutic decision and monitor therapeutic effects. Traditionally, doctors have used 24-hour timed urine collection to determine the amount of urine protein excretion, thus avoiding the potential fluctuations caused by daily activity. However, this procedure, especially in outpatients, is timeconsuming, cumbersome and frequently unreliable because of missed collection of urine samples. The spot urine protein-tocreatinine ratio (UPCR) has been used as a substitute for timed urine protein collection on the basis of the assumption that the urine protein excretion rate, which is proportional to the urine creatinine excretion, is relatively constant throughout the day.2 Several studies have confirmed the high correlation between the UPCR and 24-hour urine protein (24h-UP) with different regression coefficients (r 5 0.92–0.97).2–5 The National Kidney Foundation has also recommended that untimed (spot) urine samples should be used to detect and monitor proteinuria in adults or children and that it is not necessary to perform timed urine collection (overnight or 24-hour).6 However, the amount of urine creatinine excretion per day has been found to be highly variable between individuals, depending on age, gender and body weight,7 and the differences may influence the accuracy of the UPCR in predicting 24h-UP. In patients with extremely high or low daily urine creatinine excretion, the UPCR may be even double or half of the 24h-UP. As a result, the UPCR should not be used for initial risk stratification in glomerular diseases such as membranous nephropathy. In addition, the time of spot urine sampling and the use of medications may potentially contribute to differences in the values for UPCR and 24h-UP. This study aims primarily to determine whether there is a high correlation, concordance and agreement between UPCR and 24h-UP values and to examine the circadian fluctuations of the UPCR in patients with chronic kidney disease. A second aim is to investigate whether the accuracy of using UPCR in predicting 24h-UP would be improved after multiplying the UPCR value by the estimated daily urine creatinine excretion (eUCr) based on the Cockcroft-Gault equation,7 especially in patients with extremely high or low daily urine creatinine excretion. Finally, the authors investigate whether the differences in the amount of 24h-UP, renal function and use of reninangiotensin-aldosterone system (RAAS) blockers affect the correlation between the UPCR and 24h-UP.
MATERIALS AND METHODS Patient Selection The study group consisted of 120 participants who met the following inclusion criteria: (1) chronic kidney disease with stable renal function for 3 months before the commencement of the study and (2) daily urine protein level higher than 100 mg or
Volume 349, Number 6, June 2015
477
Chen et al
UPCR higher than 0.1 (g/g). These participants were regularly monitored by nephrology outpatient services in Taipei Veterans General Hospital. All enrolled participants were asked to maintain their regular dietary intake and medications for underlying diseases. Participants who were unsure of accurate compliance in collecting urine samples or who had collected 24h-UP with less than 100 mg per day were excluded from the analysis. Furthermore, participants with the ratio of measured daily urine creatinine excretion (mUCr) and eUCr more than 1.5 or less than 0.5 were also excluded. After obtaining informed consent from every participant in this study, urine samples were collected according to the following protocol. Study Protocol for Urine Collection and Testing A plastic bucket with cover was provided to each participant participating in this study. Participants had to void before the study began, and 24-hour urine collection began at 7:00 AM on the first morning of the study. Three sets of spot urine samples, each containing 10 mL of urine, were collected at random times throughout the day (7:00 AM to 3:00 PM; labeled as “Random daytime”) and night (3:00 PM to 11:00 PM; labeled as “Random nighttime”) and at 7:00 AM the next morning (labeled as “First morning”). These urine samples were then put in the plastic bucket and returned to the laboratory together with the 3 spot urine samples immediately after the urine collection is complete. In the meantime, confirmation of successful and complete collection of all urine samples was sought from the participants. The 24-hour urine samples and 3 sets of spot urine samples were analyzed for creatinine and total protein levels in the morning completing urine collection. Urine protein was measured by the pyrogallol red-molybdate method, and urine creatinine was measured by the autoanalyzer method with the Jaffe reaction. The general data included age, gender, body weight (at the time of study), serum creatinine concentration (SCr), body mass index and the coincidence of diabetes mellitus and hypertension were recorded. Medical records were reviewed, and the use or nonuse of RAAS blockers was confirmed with the participants. The estimated glomerular filtration rate (eGFR) was calculated by simplified MDRD formula, and the staging of chronic kidney disease was determined according to the K/DOQI guideline.6 Statistical Analyses Three statistical approaches were used to compare UPCR from different times of collection and 24h-UP. First, linear regression analysis was used to yield a regression equation: y 5 a + bx, where UPCR served as “x” variable and 24h-UP served as “y” variable. Both the slope (b) and correlation coefficient (r2) between individual UPCR and 24h-UP measurements were evaluated. The “b” value of the regression equation represents the accuracy of different methods used to predict 24h-UP. Any method with a “b” value close to 1 has a high accuracy to predict 24h-UP. Second, concordance correlation coefficient (CCC)8 was used to evaluate the agreement between UPCR and 24h-UP. A CCC close to 1 shows strong concordance. Third, calibration plot9 was used to verify the agreement between UPCR and 24h-UP. The x axis shows the value of 24h-UP, whereas the y axis represents the ratio of UPCR and 24h-UP (UPCR/24h-UP) in this plot. A mean value of ratio close to 1 represents high accuracy of using UPCR to predict 24h-UP, and smaller
478
TABLE 1. Baseline demographic and clinical characteristics of the study group Variables Values Age (yrs; mean 6 SD) Gender Male, n (%) Female, n (%) Body weight (kg; mean 6 SD) SCr (mg/dL; mean 6 SD) Body mass index (kg/m2; mean 6 SD) CKD, n (%) Stage 1 (eGFR $90) Stage 2 (eGFR: 60–89) Stage 3 (eGFR: 30–59) Stage 4 (eGFR: 15–29) Stage 5 (eGFR #15) Diabetes mellitus, n (%) Hypertension, n (%) RAAS blockers use, n (%)
67 6 14.83 57 (63) 33 (37) 65.57 6 12.91 2.84 6 1.93 24.94 6 4.77 7 7 29 29 18 38 75 56
(7.8) (7.8) (32.2) (32.2) (20.0) (42.2) (83.3) (62.2)
Staging of CKD is based on eGFR by MDRD formula: GFR (mL$min21$1.73 m22) 5 175 3 (SCr)21.154 3 (age)20.203 3 (0.742, if female) 3 (1.212, if African American). CKD, chronic kidney disease; eGFR, estimated glomerular filtration rate; MDRD, modification of diet in renal disease; RAAS, reninangiotensin-aldosterone axis; SCr, serum creatinine.
range between upper and lower limits of agreement (61.96 SD) shows better agreement. To eliminate the effect of variable urine creatinine excretion between individuals on the analysis, UPCR was multiplied by the eUCr value obtained by the CockcroftGault equation7: 24-hour urine creatinine excretion (g) 5 {[140 2 age (years)] 3 weight (kg)} / 5000 (30.85 in females). The eUCr-weighted UPCR (eUPCR) was then used instead of UPCR in analyzing all participants and subgroups with mUCr #0.8 g/d, mUCr $1.2 g/d, mUCr .0.8 g/d and ,1.2 g/d. Subgroup analysis according to the amount of urine protein (24h-UP .3g/d and #3g/d) and renal
TABLE 2. Regression parameters and CCC of the study group UPCR vs. 24h-UP eUPCR vs. 24h-UP
All patients (n 5 90) Random daytime Random nighttime First morning
b
r2
CCC
b
r2
CCC
0.848
0.857
0.918
0.942
0.944
0.971
0.866
0.882
0.933
0.957
0.966
0.982
0.869
0.846
0.916
0.970
0.935
0.967
Regression equation: 24h-UP 5 a + (b 3 UPCR [or eUPCR]). For the regression coefficient, all P , 0.001. CCCs, concordance correlation coefficients; 24h-UP, timed 24-hour urine protein; UPCR, urine protein-to-creatinine ratio; eUPCR, estimated daily urine creatinine excretion–weighted UPCR; random daytime, spot urine between 7:00 AM and 3:00 PM; random nighttime, spot urine between 3:00 AM and 11:00 PM; first morning, first voided at 7:00 AM the next morning.
Volume 349, Number 6, June 2015
Urine Protein/Creatinine Ratio and 24-Hour Urinary Protein
FIGURE 1. The relationship between the ratio of 24h-UP/UPCR (eUPCR) and 24h-UP in different times of urine collection. After using eUPCR (right panel) instead of UPCR (left panel), the mean value of ratio and limit of agreement were improved in all 3 spot urine samples. The eUPCR in nighttime urine sample yielded the best mean value of ratio and limit of agreement. eUPCR, estimated daily urine creatinine excretion–weighted UPCR; UPCR, urine protein-to-creatinine ratio; 24h-UP, 24-hour urinary protein.
function (eGFR $60 mL$min21$1.73 m22 and ,60 mL$min21$1.73 m22) were also performed. Finally, participants were divided into 2 subgroups, depending on whether they were receiving RAAS blockers, such as angiotensinCopyright © 2015 by the Southern Society for Clinical Investigation.
converting enzyme inhibitors, angiotensin type 1 receptor antagonists or spironolactone during the urine sampling period. The correlation, concordance and agreement between the UPCR and 24h-UP were then assessed in both
479
Chen et al
TABLE 3. Regression parameters and CCCs in subgroups according to mUCr UPCR vs. 24h-UP b mUCr #0.8 (n 5 31) Random daytime Random nighttime First morning mUCr $1.2 (n 5 16) Random daytime Random nighttime First morning mUCr ,1.2 and .0.8 (n 5 43) Random daytime Random nighttime First morning
r
2
eUPCR vs. 24h-UP
CCC
b
r2
CCC
0.688 0.709 0.730
0.902 0.917 0.888
0.855 0.876 0.871
0.916 0.954 0.972
0.930 0.950 0.909
0.960 0.973 0.952
1.250 1.199 1.164
0.980 0.978 0.938
0.940 0.953 0.922
0.991 0.966 0.927
0.975 0.984 0.937
0.987 0.991 0.966
0.875 0.888 0.888
0.950 0.961 0.921
0.967 0.975 0.956
0.914 0.953 1.003
0.931 0.963 0.957
0.963 0.979 0.977
Regression equation: 24h-UP 5 a + (b 3 UPCR [or eUPCR]). For the regression coefficient, all P , 0.001. CCCs, concordance correlation coefficients; mUCr, measured daily urine creatinine excretion (g/d); UPCR, urine protein-to-creatinine ratio; 24h-UP, timed 24-hour urine protein; eUPCR, estimated daily urine creatinine excretion–weighted UPCR; random daytime, spot urine during 7:00 AM to 3:00 PM; random nighttime, spot urine between 3:00 AM and 11:00 PM; first morning, first voided at 7:00 AM the next morning.
subgroups. All statistical tests were two-tailed, and significance was set at P , 0.05.
RESULTS Patient Characteristics One hundred and twenty participants examined by the nephrology outpatient services in Taipei Veterans General Hospital were initially enrolled in this study. Of these, 9 participants were unsure of accuracy in urine collection, 5 participants were sure of accuracy in urine collection but the 24h-UP was less than 100 mg per day and for 16 participants the ratio of mUCr to eUCr was above 1.5 or below 0.5. After excluding the above mentioned participants from the statistical analysis, 90 participants completed the protocol for urine collection and these participants’ baseline characteristics were listed in Table 1. Statistical Analysis There were both high accuracy of using UPCR in predicting 24h-UP and strong correlation between UPCR and 24h-UP. The correlation between UPCR and 24h-UP was slightly higher in random nighttime UPCR compared with random daytime UPCR and first morning UPCR. However, the first morning UPCR had the highest b value, representing the best accuracy in predicting 24h-UP. When the eUCr, calculated by the Cockcroft-Gault formula, was multiplied to the UPCR (eUPCR), the b value for all 3 spot urine samples increased significantly, as did the r2. The agreement between UPCR and 24h-UP, expressed by CCC, showed strong concordance and was the highest in the random nighttime UPCR. The concordance between eUPCR and 24h-UP was improved irrespective of the time of urine sampling (Table 2). In the calibration plot analysis, mean value of the ratio eUPCR/24h-UP was closer to 1 compared with that of UPCR/24h-UP. Besides, smaller range of limit of agreement (LoA) and higher percentage of individuals with UPCR within 15%, 20% and 30% of 24h-UP were noted in the analysis of eUPCR and 24h-UP, as illustrated in Figure 1.
480
When participants were divided into 3 subgroups on the basis of their mUCr levels (#0.8, $1.2, .0.8 and ,1.2 g/d), both the r 2 of the UPCR and eUPCR with 24h-UP were comparably high. The b value increased in the subgroups with mUCr #0.8 g/d and mUCr .0.8 and ,1.2 g/d but decreased in the subgroup with mUCr $1.2 g/d. Analysis of CCC showed stronger concordance of using eUPCR instead of UPCR in all 3 subgroups (Table 3). In calibration plot, mean value of the ratio eUPCR/24h-UP was much closer to 1 compared with that of UPCR/24h-UP in all 3 subgroups. The range of LoA was also narrower and the percentage of individuals with UPCR within 15%, 20% and 30% of 24h-UP was higher using eUPCR instead of UPCR except in the subgroup of mUCr $1.2 g/d, as illustrated in Figure 2. Subgroup analysis according to the amount of urine protein (24h-UP .3g/d and #3 g/d) showed relatively poor b value, r 2 and CCC between UPCR and 24h-UP. Using eUPCR instead of UPCR enhanced the b value, r 2 and CCC regardless the amount of 24h-UP was more or less than 3 g/d (Table 4). In the subgroup of 24h-UP .3 g/d, the mean value of ratio was closer to 1 and the range of LoA was narrower compared with the subgroup of 24h-UP #3g/d. Similarly, there were better results of mean value of ratio, the range of LoA and higher percentage of individuals with UPCR within 15%, 20% and 30% of 24h-UP if using eUPCR instead of UPCR (Figure 3). Subgroup analysis according to eGFR $45 mL$min21$1.73 m22 or ,45 mL$min21$1.73 m22 was performed, but there were inconsistencies of different times of urine collection in both subgroup. As the authors expected, the b value, r2 and CCC were improved between eUPCR and 24h-UP in both subgroups (Table 5). In the subgroup of eGFR $45 mL$min21$1.73 m22, the mean value of ratio was closer to 1 with narrower range of LoA in calibration plot. Using eUPCR instead of UPCR resulted in narrower range of LoA but inferior mean value of ratio in this subgroup. The percentage of individuals with UPCR within 15%, 20% and 30% of 24h-UP was higher in eUPCR analysis (Figure 4). Volume 349, Number 6, June 2015
Urine Protein/Creatinine Ratio and 24-Hour Urinary Protein
FIGURE 2. The relationship between the ratio of 24h-UP/UPCR (eUPCR) and 24h-UP in the samples of first-voided urine next morning, subgroup by mUCr #0.8 g/d, $1.2 g/d, and .0.8 and ,1.2 g/d. After using eUPCR (right panel) instead of UPCR (left panel), the mean value of ratio was closer to 1, especially in subgroups with mUCr #0.8 g/d and mUCr $1.2 g/d. The range between limit of agreement was narrower by eUPCR except in the subgroup of mUCr $1.2 g/d. eUPCR, estimated daily urine creatinine excretion–weighted UPCR; UPCR, urine protein-tocreatinine ratio; 24h-UP, 24-hour urinary protein; mUCr, measured daily urine creatinine excretion.
Copyright © 2015 by the Southern Society for Clinical Investigation.
481
Chen et al
TABLE 4. Regression parameters and CCCs in subgroups according to measured daily urine protein UPCR vs. 24h-UP eUPCR vs. 24h-UP 24h-UP #3 g/d (n 5 69) Random daytime Random nighttime First morning 24h-UP .3 g/d (n 5 21) Random daytime Random nighttime First morning
b
r2
CCC
b
r2
CCC
0.665 0.706 0.678
0.786 0.755 0.699
0.824 0.836 0.796
0.767 0.869 0.887
0.784 0.795 0.780
0.876 0.890 0.882
0.698 0.739 0.685
0.618 0.687 0.614
0.772 0.816 0.776
0.899 0.929 0.886
0.858 0.921 0.832
0.925 0.959 0.911
Regression formula: 24h-UP 5 a + (b 3 UPCR [or eUPCR]). For the regression coefficient, all P , 0.001. CCCs, concordance correlation coefficients; UPCR, urine protein-to-creatinine ratio; 24h-UP, timed 24-hour urine protein; eUPCR, estimated daily urine creatinine excretion–weighted UPCR; random daytime, spot urine during 7:00 AM to 3:00 PM; random nighttime, spot urine between 3:00 AM and 11:00 PM; first morning, first voided at 7:00 AM the next morning.
FIGURE 3. The relationship between the ratio of 24h-UP/UPCR (eUPCR) and 24h-UP in the samples of first-voided urine next morning, subgroup by 24h-UP of #3 g/d and .3 g/d. After using eUPCR (right panel) instead of UPCR (left panel), the mean value of ratio and limit of agreement were improved in both subgroups. The range between limit of agreement in the subgroup by 24h-UP of #3 g/d was wider than another subgroup. eUPCR, estimated daily urine creatinine excretion–weighted UPCR; UPCR, urine protein-to-creatinine ratio; 24h-UP, 24-hour urinary protein.
482
Volume 349, Number 6, June 2015
Urine Protein/Creatinine Ratio and 24-Hour Urinary Protein
TABLE 5. Regression parameters and CCCs in subgroups according to eGFR UPCR vs. 24h-UP b 21
r
2
eUPCR vs. 24h-UP CCC
b
r2
CCC
22
eGFR $45 mL$min $1.73 m (n 5 23) Random daytime Random nighttime First morning eGFR ,45 mL$min21$1.73 m22 (n 5 67) Random daytime Random nighttime First morning
0.796 0.891 1.173
0.758 0.843 0.788
0.866 0.918 0.848
0.920 0.947 1.225
0.935 0.971 0.899
0.965 0.984 0.904
0.863 0.863 0.851
0.879 0.889 0.876
0.928 0.936 0.925
0.950 0.960 0.951
0.945 0.964 0.957
0.972 0.981 0.977
Regression formula: 24h-UP 5 a + (b 3 UPCR [or eUPCR]). For the regression coefficient, all P , 0.001. eGFR is calculated by MDRD formula: GFR (mL$min21$1.73 m22) 5 175 3 (SCr)21.154 3 (age)20.203 3 (0.742, if female) 3 (1.212, if African American). CCCs, concordance correlation coefficients; eGFR, estimated glomerular filtration rate; MDRD, modification of diet in renal disease; SCr, serum creatinine; UPCR, urine protein-to-creatinine ratio; 24h-UP, timed 24-hour urine protein; eUPCR, estimated daily urine creatinine excretion–weighted UPCR; random daytime, spot urine during 7:00 AM to 3:00 PM; random nighttime, spot urine between 3:00 AM and 11:00 PM; first morning, first voided at 7:00 AM the next morning.
When participants were divided into 2 subgroups on the basis of the use of RAAS blockers, r2 between the UPCR and 24h-UP was higher in the subgroup taking RAAS blockers than in the subgroup not taking RAAS blockers, as did b value and CCC. Irrespective of the use of RAAS blockers, the correlation and concordance were improved after using the eUPCR instead of the UPCR (Table 6). Similarly, using the eUPCR improved the accuracy and agreement in both subgroups in the calibration plot (Figure 5).
DISCUSSION This study was conducted to investigate the accuracy of using UPCR as a substitute of 24h-UP and the correlation between UPCR and 24h-UP. Traditionally, 24-hour urine protein collection has been the standard for assessing and quantifying proteinuria in patients with kidney diseases. A strong correlation between the UPCR and 24h-UP has been demonstrated in many studies assessing various etiologies of renal dysfunction, such as primary glomerulonephritis, diabetic nephropathy, systemic lupus erythematosus and renal transplantation in adults and children.2–5,10–17 Some studies have concluded that the best approach to measuring UPCR is to use the first daily voided urine,12,16 but some suggest obtaining spot urine readings during normal daily activity.2,18,19 In this study, there was very high correlation and concordance between the UPCR of all 3 spot urine samples and 24h-UP. Besides, the random nighttime UPCR seemed to be the best prediction of 24h-UP. The variability in UPCR throughout a 24-hour period mainly comes from the change of urine protein excretion rather than the change of urine creatinine excretion in each individual patient.20 The high degree of variability in spot UPCR can lead to error in the management of the individual patient. The authors assume that the concentration of urine protein may be altered by activity and positional change, and most urine is produced in daytime and nighttime rather than in sleep. As a result, the random daytime or nighttime UPCR has better accuracy and correlation in predicting 24h-UP compared with the first morning void UPCR. The ratio of UPCR and 24h-UP was widely distributed in the calibration plot and the mean value of ratio was 1.17 to Copyright © 2015 by the Southern Society for Clinical Investigation.
1.24, indicating potential bias of 17% to 24% introduced by UPCR. Daily urine creatinine excretion was widely variable among different patients and may lower the accuracy of using UPCR to predict the value of 24h-UP between cohorts. However, the above mentioned errors could be offset by adequate randomization and appropriate correction by multiplying by eUCr, which thus resulted in a higher correlation of 24h-UP with eUPCR than with spot UPCR. The bias may have great impact if using spot UPCR in establishing the diagnosis of nephrotic syndrome, deciding whether renal biopsy is needed and choosing the recommended therapeutic regimen classified by 24h-UP. Multiplying UPCR by eUCr (estimated urine creatinine excretion) attenuated the bias without the need to collect 24-hour urine in this study. The ratio of mUCr to eUCr has been used to validate the reliability of urine collection to confirm successful 24-hour urine collection in previous studies. Samples with measured-toestimated ratios in the range 0.9 to 1.1 are considered complete collections.2,21 However, these studies did not emphasize the bias introduced by variable daily urine creatinine excretion between individuals, which might attenuate the accuracy of the UPCR to predict 24h-UP. In patients with extremely higher or lower daily urine creatinine excretion, using UPCR may lead to overt estimation or underestimation of 24h-UP compared with patients with urine creatinine excretion close to 1 g/d. Based on the understanding that difference in age, gender and body weight may affect daily urine creatinine excretion and hence the accuracy of the UPCR, the authors tried to use eUCr-weighted UPCR (eUPCR) instead of simple UPCR to investigate the correlation and determined the accuracy of eUPCR in predicting 24h-UP. There have been several equations derived to estimate creatinine excretion rate (CER), and the bias on average, precision and accuracy between these equations is assessed by Ix et al.22 Among the estimating equations compared in the study, the Cockcroft-Gault equation had statistically significantly greater bias and tended to overestimate CER in individuals with the highest CER values. As a result, the eUCr obtained by the Cockcroft-Gault equation may have potential errors and thus lead to incorrect estimation of daily urine protein, however, which could be improved by replacing the actual body weight with estimated lean body weight in obese subjects. Furthermore, the
483
Chen et al
FIGURE 4. The relationship between the ratio of 24h-UP/UPCR (eUPCR) and 24h-UP in the samples of first-voided urine next morning, subgroup by eGFR $45 mL$min21$1.73 m22 and ,45 mL$min21$1.73 m22. After using eUPCR (right panel) instead of UPCR (left panel), the limit of agreement was improved in both subgroups. However, the mean value of ratio was inferior by eUPCR in subgroup with eGFR $45 mL$min21$1.73 m22. eGFR, estimated glomerular filtration rate; eUPCR, estimated daily urine creatinine excretion– weighted UPCR; UPCR, urine protein-to-creatinine ratio; 24h-UP, 24-hour urinary protein.
Cockcroft-Gault equation is easier to be calculated and used in clinical approach, avoiding the process of complex computing work. Using the weighted “eUPCR,” the authors obtained a stronger correlation and concordance in all 3 of the spot urine samples no matter what time the spot urine was collected. Besides, the mean value of ratio was closer to 1 and the range between LoA was narrower in calibration plot analysis after using the eUPCR instead of traditional UPCR. Similar concept was published in the study (the PREVEND study) of using estimated CER (eCER) multiplied to urine albumin-to-creatinine ratio (ACR) for assessment of albuminuria.23 Various eCER computed from the Ix,22 Ellam24 and Walser25 equations were used, and the study concluded that ACR multiplied by eCER was more accurate than ACR alone in assessing albuminuria. Multiplying UPCR of 24hour urine collection by eUCr to estimate the magnitude of 24hUP was also recommended by Glassock et al.26 Although they did not recommend using random spot or first morning UPCR to determine 24h-UP, the authors demonstrated improvement of
484
accuracy using eUPCR instead of UPCR to predict 24h-UP in spot urine sample. When participants were divided into subgroups according to mUCr values, the authors observed a greater increase in accuracy in subgroups with mUCr values #0.8 g/d or $1.2 g/d after utilization of eUPCR instead of UPCR. This finding has not been emphasized in previous studies. The difference between UPCR and 24h-UP in the subgroup with mUCr #0.8 g/d may be counteracted by the difference in the subgroup with mUCr $1.2 g/d and mask a potential bias in predicting 24h-UP. As a result, using eUPCR can improve the correlation and the accuracy between UPCR and 24h-UP, particularly in patients with daily urine creatinine excretion higher than 1.2 g/d or lower than 0.8 g/d. UPCR was believed to be unreliable for estimating 24hUP over subnephrotic range (24h-UP #3g/d) in lupus nephritis.27–29 The authors investigated if this occurred in general chronic kidney patients and thus divided the participants into Volume 349, Number 6, June 2015
Urine Protein/Creatinine Ratio and 24-Hour Urinary Protein
TABLE 6. Regression parameters and CCCs in subgroups according to use of RAAS blockers UPCR vs. 24h-UP b RAAS blockers (+) (n 5 56) Random daytime Random nighttime First morning RAAS blockers (2) (n 5 34) Random daytime Random nighttime First morning
r
2
eUPCR vs. 24h-UP
CCC
b
r2
CCC
0.884 0.887 0.892
0.871 0.912 0.897
0.931 0.951 0.944
1.006 0.991 0.975
0.951 0.978 0.945
0.974 0.998 0.971
0.794 0.825 0.825
0.840 0.825 0.746
0.894 0.898 0.858
0.858 0.902 0.965
0.956 0.950 0.916
0.965 0.971 0.957
Regression formula: 24h-UP 5 a + (b 3 UPCR [or eUPCR]). For the regression coefficient, all P , 0.001. CCCs, concordance correlation coefficients; 24h-UP, timed 24-hour urine protein; UPCR, urine protein-to-creatinine ratio; eUPCR, estimated daily urine creatinine excretion–weighted UPCR; random daytime, spot urine between 7:00 AM and 3:00 PM; random nighttime, spot urine between 3:00 AM and 11:00 PM; first morning, first voided at 7:00 AM the next morning; RAAS, renin-angiotensin-aldosterone axis.
FIGURE 5. The relationship between the ratio of 24h-UP/UPCR (eUPCR) and 24h-UP in the samples of first-voided urine next morning, subgroup use or nonuse of RAAS blockers. Both the mean value of ratio and limit of agreement in the subgroup with RAAS blockers were superior to another subgroup. Using eUPCR enhanced the accuracy and agreement in both subgroups. eUPCR, estimated daily urine creatinine excretion–weighted UPCR; UPCR, urine protein-to-creatinine ratio; 24h-UP, 24-hour urinary protein; RAAS, renin-angiotensin-aldosterone axis. Copyright © 2015 by the Southern Society for Clinical Investigation.
485
Chen et al
subgroups with 24h-UP .3g/d or #3g/d. The correlation between UPCR and 24h-UP was better in the subgroup of 24h-UP #3g/d, but the accuracy and agreement showed the opposite results. Overall, subgroup analysis showed less optimal results compared with the results over whole range of proteinuria. Whether the 24h-UP was higher or less than 3 g/d, eUPCR improved the accuracy in predicting 24h-UP together with better correlation and agreement. In the above study using eCER multiplied to urine ACR for assessment of albuminuria,23 the participants were community-living individuals from general population with mean eGFR of 84 6 15 mL$min21$1.73 m22. In contrast, this study enrolled participants with chronic kidney disease and a lower mean eGFR of 35.01 6 25.15 mL$min21$1.73 m22. The authors divided the participants to subgroups with eGFR ≧45 mL$min21$1.73 m22 or ,45 mL$min21$1.73 m22 in investigation of whether lower eGFR will influence the association between UPCR and 24h-UP or not. Inconsistent results were noted with higher correlation but lower accuracy and agreement of using UPCR to predict 24h-UP in the subgroup with relatively poor renal function. Similarly, the eUPCR was a better predictor of 24h-UP irrespective of preserved or impaired renal function. RAAS blockers (including angiotensin-converting enzyme inhibitors, angiotensin type 1 receptor antagonists and spironolactone) are widely used in patients with chronic kidney disease for both slowing down the renal function decline and decreasing the degree of proteinuria.30–33 In this study, the authors investigated if the use of RAAS blockers could influence the reliability of UPCR in predicting 24h-UP. Among participants using RAAS blockers, the correlation and accuracy were higher than in participants who did not take RAAS blockers. RAAS blockers can decrease glomerular hyperfiltration and cytokine secretion, preserving the integrity of glomerular structure and the function of the basement membrane. As a result, RAAS blockers may provide a relatively constant proteinto-creatinine filtration fraction by this mechanism. Further investigation is needed to fully understand the mechanisms underlying the authors’ findings. There are 4 limitations to this study. First, most participants (n 5 51) had a ratio of mUCr to eUCr higher than 1.1 or lower than 0.9, despite the fact that all the finally enrolled participants insisted they collected the 24-hour urine samples completely. It was2,10 considered to indicate overt collection or undercollection of urine samples in previous studies.2,21 But under certain conditions, urine creatinine excretion can significantly decrease (low meat diet, muscle wasting or limb amputation) or increase (high meat diet, increased muscle mass, creatine supplement or fenofibrate therapy).34 As a result, a ratio of mUCr to eUCr higher than 1.1 or lower than 0.9 might suggest either that errors occurred during 24-hour urine collection or just the influence of urine creatinine variation. The authors excluded participants with a ratio of mUCr to eUCr higher than 1.5 or lower than 0.5 before analysis exactly for the reason that these participants showed higher likelihood of error in urine collection. Second, the “spot” urine sample in the study was a urine collection of variable duration in urinary bladder rather than a “snapshot” urine sample. Participants were not obligated to have a snapshot urine sample or fixed time urine collection in bladder to simulate the situation in clinical practice. The variable duration of “spot” urine collection between participants together with the diurnal variations in the PC ratio may influence the accuracy of using UPCR in predicting 24h-UP. The closer the “spot” urine collection approaches 24 hours, the closer the UPCR of “spot” urine collection approaches that of
486
a complete 24-hour urine collection.21 However, the advantage of using eUPCR was existed and not vanished by these variations. Third, the urine samples were not refrigerated during the time of experiment. Despite the urine specimens being returned and analyzed in laboratory immediately after completing urine collection, differences in degradation of albumin may introduce bias in the analysis of UPCR and 24h-UP. Lastly, the number of participants in each subgroup was not equal and might lead to bias in statistics. A prospective study with larger number of participants specifically designed for subgroup analysis will further demonstrate whether the amount of 24h-UP, renal function and use of RAAS blockers will influence the relationship between UPCR and 24h-UP.
CONCLUSIONS UPCR is a valid substitute for timed urine protein collection, and the result is less affected by the time of the spot urine sampling. However, UPCR may lead to bias in predicting 24h-UP because of variable urine creatinine excretion between individuals. Using eUPCR overcomes the shortcoming and is more accurate than UPCR in assessment of 24h-UP. Accordingly, the authors recommend the use of eUPCR instead of UPCR in clinical approach to attenuate the bias introduced by variable creatinine excretion between individuals. REFERENCES 1. Keane WF. Proteinuria: its clinical importance and role in progressive renal disease. Am J Kidney Dis 2000;35:S97–105. 2. Ginsberg JM, Chang BS, Matarese RA, et al. Use of single voided urine samples to estimate quantitative proteinuria. N Engl J Med 1983; 309:1543–6. 3. Schwab SJ, Christensen RL, Dougherty K, et al. Quantitation of proteinuria by the use of protein-to-creatinine ratios in single urine samples. Arch Intern Med 1987;147:943–4. 4. Lane C, Brown M, Dunsmuir W, et al. Can spot urine protein/creatinine ratio replace 24 h urine protein in usual clinical nephrology? Nephrology 2006;11:245–9. 5. Ruggenenti P, Gaspari F, Perna A, et al. Cross sectional longitudinal study of spot morning urine protein:creatinine ratio, 24 hour urine protein excretion rate, glomerular filtration rate, and end stage renal failure in chronic renal disease in patients without diabetes. BMJ 1998;316:504–9. 6. National Kidney Foundation. K/DOQI clinical practice guidelines for chronic kidney disease: evaluation, classification, and stratification. Am J Kidney Dis 2002;39:S1–266. 7. Cockcroft DW, Gault MH. Prediction of creatinine clearance from serum creatinine. Nephron 1976;16:31–41. 8. Lin LI. A concordance correlation coefficient to evaluate reproducibility. Biometrics 1989;45:255–68. 9. van Stralen KJ, Jager KJ, Zoccali C, et al. Agreement between methods. Kidney Int 2008;74:1116–20. 10. Morgenstern BZ, Butani L, Wollan P, et al. Validity of proteinosmolality versus protein-creatinine ratios in the estimation of quantitative proteinuria from random samples of urine in children. Am J Kidney Dis 2003;41:760–6. 11. Salesi M, Karimifar M, Farajzadegan Z, et al. The protein-creatinine ratio in spot morning urine samples and 24-h urinary protein excretion in patients with systemic lupus erythematosus. Rheumatol Int 2009;29:503–7. 12. Xin G, Wang M, Jiao LL, et al. Protein-to-creatinine ratio in spot urine samples as a predictor of quantitation of proteinuria. Clin Chim Acta 2004;350:35–9.
Volume 349, Number 6, June 2015
Urine Protein/Creatinine Ratio and 24-Hour Urinary Protein
13. Rodrigo E, Pinera C, Ruiz JC, et al. Quantitation of 24-hour urine protein excretion in kidney transplant patients by the use of protein to creatinine ratio. Transplant Proc 2003;35:702. 14. Wahbeh AM, Ewais MH, Elsharif ME. Comparison of 24-hour urinary protein and protein-to-creatinine ratio in the assessment of proteinuria. Saudi J Kidney Dis Transpl 2009;20:443–7. 15. Antunes VV, Veronese FJ, Morales JV. Diagnostic accuracy of the protein/creatinine ratio in urine samples to estimate 24-h proteinuria in patients with primary glomerulopathies: a longitudinal study. Nephrol Dial Transplant 2008;23:2242–6. 16. Nagasako H, Kiyoshi Y, Ohkawa T, et al. Estimation of 24-hour urine protein quantity by the morning-urine protein/creatinine ratio. Clin Exp Nephrol 2007;11:142–6. 17. Torng S, Rigatto C, Rush DN, et al. The urine protein to creatinine ratio (P/C) as a predictor of 24-hour urine protein excretion in renal transplant patients. Transplantation 2001;72:1453–6. 18. Koopman MG, Krediet RT, Koomen GC, et al. Circadian rhythm of proteinuria: consequences of the use of urinary protein:creatinine ratios. Nephrol Dial Transplant 1989;4:9–14. 19. Dyson EH, Will EJ, Davison AM, et al. Use of the urinary protein creatinine index to assess proteinuria in renal transplant patients. Nephrol Dial Transplant 1992;7:450–2. 20. Koopman MG, Krediet RT, Zuyderhoudt FJ, et al. A circadian rhythm of proteinuria in patients with a nephrotic syndrome. Clin Sci (Lond) 1985;69:395–401. 21. Shidham G, Hebert LA. Timed urine collections are not needed to measure urine protein excretion in clinical practice. Am J Kidney Dis 2006;47:8–14. 22. Ix JH, Wassel CL, Stevens LA, et al. Equations to estimate creatinine excretion rate: the CKD epidemiology collaboration. Clin J Am Soc Nephrol 2011;6:184–91. 23. Abdelmalek JA, Gansevoort RT, Lambers Heerspink HJ, et al. Estimated albumin excretion rate versus urine albumin-creatinine ratio for the assessment of albuminuria: a diagnostic test study from the Prevention of Renal and Vascular Endstage Disease (PREVEND) Study. Am J Kidney Dis 2014;63:415–21.
Copyright © 2015 by the Southern Society for Clinical Investigation.
24. Fotheringham J, Campbell MJ, Fogarty DG, et al. Estimated albumin excretion rate versus urine albumin-creatinine ratio for the estimation of measured albumin excretion rate: derivation and validation of an estimated albumin excretion rate equation. Am J Kidney Dis 2014;63:405–14. 25. Walser M. Creatinine excretion as a measure of protein nutrition in adults of varying age. JPEN J Parenter Enteral Nutr 1987;11:73S–8S. 26. Glassock RJ, Fervenza FC, Hebert L, et al. Nephrotic syndrome redux. Nephrol Dial Transplant 2014;0:1–6. 27. Hebert LA, Birmingham DJ, Shidham G, et al. Random spot urine protein/creatinine ratio is unreliable for estimating 24-hour proteinuria in individual systemic lupus erythematosus nephritis patients. Nephron 2009;113:c177–82. 28. Birmingham DJ, Rovin BH, Shidham G, et al. Spot urine protein/creatinine ratios are unreliable estimates of 24 h proteinuria in most systemic lupus erythematosus nephritis flares. Kidney Int 2007;72: 865–70. 29. Birmingham DJ, Shidham G, Perna A, et al. Spot PC ratio estimates of 24-hour proteinuria are more unreliable in lupus nephritis than in other forms of chronic glomerular disease. Ann Rheum Dis 2014;73:475–6. 30. Lewis EJ, Hunsicker LG, Bain RP, et al. The effect of angiotensinconverting-enzyme inhibition on diabetic nephropathy. The Collaborative Study Group. N Engl J Med 1993;329:1456–62. 31. Andersen S, Tarnow L, Rossing P, et al. Renoprotective effects of angiotensin II receptor blockade in type 1 diabetic patients with diabetic nephropathy. Kidney Int 2000;57:601–6. 32. Brenner BM, Cooper ME, de Zeeuw D, et al. Effects of losartan on renal and cardiovascular outcomes in patients with type 2 diabetes and nephropathy. N Engl J Med 2001;345:861–9. 33. Bianchi S, Bigazzi R, Campese VM. Long-term effects of spironolactone on proteinuria and kidney function in patients with chronic kidney disease. Kidney Int 2006;70:2116–23. 34. Wilmer WA, Rovin BH, Hebert CJ, et al. Management of glomerular proteinuria: a commentary. J Am Soc Nephrol 2003; 14:3217–32.
487
