Day-to-Day Variability in Spot Urine Protein-Creatinine Ratio Measurements

Original Investigation Day-to-Day Variability in Spot Urine Protein-Creatinine Ratio Measurements Chetana N. Naresh, FRACP, MMed (Clin Epi),1,2 Andrew Hayen, PhD,3 Jonathan C. Craig, FRACP, PhD,3,4 and Steven J. Chadban, FRACP, PhD1,2 Background: Accurate measurement of proteinuria is important in the diagnosis and management of chronic kidney disease (CKD). The reference standard test, 24-hour urinary protein excretion, is inconvenient and vulnerable to collection errors. Spot urine protein-creatinine ratio (PCR) is a convenient alternative and is in widespread use. However, day-to-day variability in PCR measurements has not been evaluated. Study Design: Prospective cohort study of day-to-day variability in spot urine PCR measurement. Setting & Participants: Clinically stable outpatients with CKD (n ⫽ 145) attending a university hospital CKD clinic in Australia between July 2007 and April 2010. Index Test: Spot urine PCR. Outcomes: Spot PCR variability was assessed and repeatability limits were determined using fractional polynomials. Measurements: Spot PCRs were measured from urine samples collected at 9:00 AM on consecutive days and 24-hour urinary protein excretion was collected concurrently. Results: Paired results were analyzed from 145 patients: median age, 56 years; 59% men; and median 24-hour urinary protein excretion, 0.7 (range, 0.06-35.7) g/d. Day-to-day variability was substantial and increased in absolute terms, but decreased in relative terms with increasing baseline PCR. For patients with a low baseline PCR (20 mg/mmol [177 mg/g]), a change greater than ⫾160% (repeatability limits, 0-52 mg/mmol [0-460 mg/g]) is required to indicate a real change in proteinuria status with 95% certainty, whereas for those with a high baseline PCR (200 mg/mmol [1,768 mg/g]), a change of ⫾50% (decrease to ⬍100 mg/mmol [⬍884 mg/g] or increase to ⬎300 mg/mmol [⬎2,652 mg/g]) represents significant change. Limitations: These study results need to be replicated in other ethnic groups. Conclusions: Changes in PCR observed in patients with CKD, ranging from complete resolution to doubling of PCR values, could be due to inherent biological variation and may not indicate a change in disease status. This should be borne in mind when using PCR in the diagnosis and management of CKD. Am J Kidney Dis. 60(4):561-566. © 2012 by the National Kidney Foundation, Inc. Published by Elsevier Inc. All rights reserved. INDEX WORDS: Chronic kidney disease (CKD); diagnostic test; protein-creatinine ratio; variability.

roteinuria (protein excretion ⬎0.15 g/d) is a hallmark of chronic kidney disease (CKD)1 and a marker of increased cardiovascular risk.2 The magnitude of protein excretion is associated linearly with subsequent decrease in glomerular filtration rate (GFR) and risk of end-stage kidney disease3,4 and therefore is an important indicator of prognosis and response to therapy. Consequently, reliable measurement of proteinuria is an important aspect of clinical practice. The optimal method for detecting proteinuria in the clinic is yet to be defined. Point-of-care tests such as dipsticks are semiquantitative, and although potentially useful as a screening tool, suboptimal sensitivity and specificity limit their usefulness for informing prognosis and monitoring therapy in the clinic.5 A quantitative 24-hour urine collection for total protein excretion is the reference standard test to quantify proteinuria. However, this is cumbersome and subject to collection errors.6 Measurement of spot urinary protein-creatinine ratio (PCR) is convenient to the patient and is recommended in US guidelines.1,6

P

Am J Kidney Dis. 2012;60(4):561-566

Spot PCR has been shown to correlate well with 24-hour protein excretion in many studies, although there are differences in correlation levels at different magnitudes of protein excretion,6,7 and to be a superior predictor of disease progression in one longitudinal cohort.7 However, existing studies are largely retrospective, used various analytical techniques to

From the 1Department of Renal Medicine, The Royal Prince Alfred Hospital, Camperdown; 2Sydney Medical School, University of Sydney; 3Screening and Test Evaluation Program, School of Public Health, University of Sydney, Sydney; and 4Department of Nephrology, Children’s Hospital at Westmead, Westmead, NSW, Australia. Received October 26, 2011. Accepted in revised form April 11, 2012. Originally published online May 17, 2012. Address correspondence to Steven J. Chadban, FRACP, PhD, Level 9 Renal Transplantation, The Royal Prince Alfred Hospital, Camperdown, NSW 2050, Australia. E-mail: steve.chadban@ sswahs.nsw.gov.au © 2012 by the National Kidney Foundation, Inc. Published by Elsevier Inc. All rights reserved. 0272-6386/$36.00 http://dx.doi.org/10.1053/j.ajkd.2012.04.010 561

Naresh et al

measure protein, and have determined correlation rather than agreement between tests.6,7 Diurnal variation in proteinuria occurs in health and in CKD; thus, timing of spot urine collections may be important.8 The extent of day-to-day variability in PCRs has not been documented in the existing literature, but may be important in determining whether a change in the quantity of proteinuria over time indicates a change in disease status, response to therapy, or simple test variability. We performed a prospective cohort study to quantify day-to-day variability in spot urine PCR.

METHODS Study Design We performed a study between July 2007 and April 2010 at a metropolitan tertiary-care teaching hospital in Sydney, Australia, which was designed and reported using the STARD (Standards for Reporting of Diagnostic Accuracy) guidelines.9 The Sydney South West Area Health Service Ethics Review Committee approved this study, protocol no. X06-0196.

Specimen Assay The 24-hour specimens were assessed for adequacy. Any specimen with creatinine excretion ⬍15 mg/kg/d in men and ⬍12 mg/kg/d in women was regarded as incomplete and excluded from the study analysis. Spot urine samples were not routinely cultured to detect bacteriuria because there is no convincing evidence that the presence of asymptomatic urinary tract infection significantly alters protein excretion rates.11 The spot specimens were analyzed for protein (grams per liter) and creatinine (millimoles per liter), and PCR was derived by dividing the protein concentration by the creatinine concentration. The ratio was expressed as milligrams per millimoles. Urine protein was measured by immunoturbidimetry using a Roche Hitachi modular analyzer (Roche Diagnostics, www.rocheaustralia.com). The analytical detection sensitivity limit for the urine protein assay was 0.04-2 g/L. Urine protein concentrations ⬎2 g/L were diluted before measurement. The laboratory withinrun and between-run coefficients of variation for urine protein were 5.2% and 3.8%, respectively. Urine creatinine was measured by the kinetic Jaffé method on a Roche Hitachi modular analyzer. The detection sensitivity limit for urine creatinine was 360-57,500 mmol/L. For urine creatinine at concentrations of 5.39 mmol/L, the laboratory within-run and between-run coefficients of variation were 1.1% and 1.2%, respectively.

Patient Recruitment and Consent Patients were recruited from the hospital’s CKD and kidney transplant clinics. Eligible individuals identified from an electronic database were adults (aged ⱖ18 years) with albuminuria (albumincreatinine ratio ⬎3.5 mg/mmol) or proteinuria (24-hour urine total protein ⬎150 mg/d) with stable kidney function (outpatients with less than ⫾15% variation in proteinuria and estimated GFR [eGFR] during the preceding 3 months). Patients were excluded if they were on dialysis therapy, known to be pregnant or less than 3 months post partum, had symptomatic urinary tract infection, were treated for sepsis or hospitalized within the past 2 weeks, had overt cardiac failure, were menstruating, or were unable to provide informed consent. Participants provided written consent and no financial incentives were provided.

Specimen Collection and Storage Participants were advised to continue their usual lifestyle, diet, and medications during the study period without changes, restrictions, or exclusions in accordance with usual clinical practice. They were given a urine collection kit containing 2 spot containers, a 24-hour urine collection (5 L) bottle, a sterile 10-mL plastic syringe, and written instructions for urine collection and storage. Participants voided urine into a clean container at 9:00 AM and, using a syringe, collected a 10-mL aliquot of this urine into a spot container and stored it at 1°C-4°C. On the next day at 9:00 AM, another spot urine collection was performed and stored using the same methods. The spot collections at 9:00 AM on both days were not first morning voids. All urine passed in the intervening 24 hours was collected into the 24-hour bottle. Specimens were returned to the hospital that day and analyzed in the hospital’s accredited centralized laboratory within 48 hours. No specimen was frozen. Participants underwent a blood test for hemoglobin, urea, and creatinine when urine specimens were returned; blood pressure, height, weight, medication, and relevant medical history were recorded; and standard demographic information was collected from all participants. eGFR was derived using the isotope-dilution mass spectrometry–traceable 4-variable MDRD (Modification of Diet in Renal Disease) Study equation.10 The data were deidentified before analysis and 10% of the entered data was randomly audited for accuracy of data entry. 562

Statistical Analyses The statistical significance of the mean difference between day-1 and day-2 PCRs was determined using paired t tests, with 95% confidence intervals (CIs) and significance level at 0.05. Data were not geometrically transformed. Correlation between day-1 and day-2 PCRs was measured using Spearman ␳. We constructed Bland Altman plots in which the difference in measurements is plotted against the average of measurements. We then calculated repeatability limits; that is, lower and upper limits in which 95% of the differences between 2 measurements on the same person should lie, using the methods described by Bland and Altman.12,13 First, we performed a regression of the absolute difference (D) between measurements against the average (A) of the methods. There was a small number of observations (n ⫽ 6) with an average PCR ⬎600 mg/mmol (all in the range of 6261,341 mg/mmol [5,534-11,855 mg/g]). Because of the lack of data in this range, we excluded these observations from the regression models. Thus, we restricted analysis to the 139 observations with PCR ⱕ600 mg/mmol [ⱕ5,304 mg/g]. Because the absolute difference between measurements depended on the level of measurement in a nonlinear manner, we used fractional polynomials in the regression. We fitted a fractional polynomial with 2 fractional polynomial terms, but this had only marginally better fit than a model with a single term (P ⫽ 0.826). Therefore, we used the simpler model for ease of explanation. The fractional polynomial model was |D| ⫽ ⫺7.095 ⫹ 3.441兹A. However, to avoid problems with negative standard deviations (SDs) at small values of the measurement, we refitted the model without the constant term, which gave |D| ⫽ 2.868兹A (this model had almost identical fit). We also bootstrapped the final regression model using 10,000 replicates to obtain estimates of uncertainty around the regression coefficient. These gave a 95% CI for the regression coefficient of 2.270-3.509. The SD of the differences is then given by SD ⫽ 兹␲⁄2 ⫻ 2.868兹A ⫽ 3.594兹A. The 95% repeatability limits are then given by ⫾1.96 ⫻ 3.594 ⫻ 兹A ⫽ 7.045兹A. In other words, 95% of repeated measurements should lie within 7.0兹A of the original measurements. We tested whether the repeatability limits differed by a fixed amount across the levels of age (⬍55 vs ⱖ55 years), sex, and Am J Kidney Dis. 2012;60(4):561-566

Day-to-Day Variability in Spot PCR eGFR category (⬍30, 30-⬍60, and ⱖ60 mL/min/1.73 m2) by including a term for each of these variables in the regression equations. The repeatability limits for test results were statistically extrapolated at different baseline PCR thresholds, if 2 or 3 repeated test results were available. Data were analyzed using Stata, version 12.0 (www.stata.com).

RESULTS Patient Characteristics Of 570 patients who were invited to participate in this study, 270 consented and were enrolled in the study. There were 125 patients who were excluded because they failed to complete the study (specimens not provided or were incomplete). Paired PCR results were analyzed from 145 patients who completed the study, and their demographics are listed in Table 1. Baseline characteristics of those who failed to complete the study were not significantly different from those who completed the study. Median age of participants was 56 (range, 20-86) years and 59% were men. Median body mass index of participants was 27.9 (range, 17.8-50.9) kg/m2, history of hypertension was present in 79% (n ⫽ 115), and 26% (n ⫽ 37) had a functioning kidney transplant. Median 24-hour protein excretion was 0.7 (range, 0.06-35.7) g/d, with 42% of patients with protein excretion ⬍0.5 g/d; 46%, 0.5-3 g/d; and 12%, ⬎3 g/d. Seventy-five percent of patients were using either an angiotensin-converting enzyme inhibitor or angiotensin receptor blocker. PCR Measurements Mean day-1 PCR was 149.9 ⫾ 232 (SD) mg/mmol (1,325 ⫾ 2051 mg/g) and mean day-2 PCR was 145.9 ⫾ 214 mg/mmol (1,290 ⫾ 1,892 mg/g). The difference in mean values between day-1 and day-2 PCRs was not significant (4.0 mg/mmol; 95% CI, ⫺5.6 to 13.7 [35; 95% CI, ⫺50 to 121 mg/g]; P ⫽ 0.4). As shown in Fig 1, there was a high correlation between day-1 and day-2 PCR measurements (Spearman ␳ ⫽ 0.92). Repeatability Limits of PCR The lower and upper limits within which 95% of repeated measurements for the same person in a clinically stable state should lie were ⫾ 1.96 ⫻ 3.594 ⫻ 兹A ⫽ 7.045兹A, where A is the first measurement (Fig 2). The magnitude of the baseline PCR determines the magnitude of the absolute difference in a repeated test measurement, as shown in Fig 2. The absolute difference or the repeatability coefficient between paired serial measurements is expected to lie within 1.96 SD from the baseline measurement for 95% of paired measurements.12 Therefore, in a clinically stable patient, at any baseline PCR, a repeated test result can be expected to lie within the repeatability limits (which are the upper and lower boundaries Am J Kidney Dis. 2012;60(4):561-566

Table 1. Demographic and Baseline Characteristics of the Study Population 24-h Protein Excretion (g/d) Characteristic

Overall

<0.5

0.5-3

>3

145 (100) 69 (48) 86 (59)

61 (42) 33 (54) 33 (54)

67 (46) 28 (42) 41 (61)

17 (12) 8 (47) 12 (71)

3 (2) 37 (26) 48 (33) 57 (39)

— 16 (26) 24 (39) 21 (34)

3 (4.5) 18 (27) 17 (25) 29 (43)

— 3 (18) 7 (41) 7 (41)

116 (80) 16 (11) 6 (4) 7 (5)

47 (77) 7 (12) 5 (8) 2 (3)

56 (84) 8 (12) 1 (1) 2 (3)

13 (76) 1 (6) — 3 (18)

Systolic BP ⬍120 mm Hg 120-⬍140 mm Hg 140⬍160 mm Hg ⱖ160 mm Hg

31 (21) 63 (43) 39 (27) 12 (8)

17 (28) 25 (41) 15 (25) 4 (7)

11 (16) 33 (49) 17 (25) 6 (9)

3 (18) 5 (29) 7 (41) 2 (12)

Diastolic BP ⬍80 mm Hg 80-⬍90 mm Hg 90-⬍100 mm Hg ⱖ100 mm Hg

59 (41) 62 (43) 16 (11) 8 (6)

30 (49) 22 (36) 6 (10) 3 (5)

22 (33) 32 (48) 9 (13) 4 (6)

7 (41) 8 (47) 1 (6) 1 (6)

Diabetes present Cause of CKD GN Diabetes Other Not available

50 (34)

17 (28)

26 (39)

7 (41)

52 (36) 24 (17) 56 (39) 13 (9)

21 (34) 8 (13) 26 (43) 6 (10)

23 (34) 12 (18) 26 (39) 6 (9)

8 (47) 4 (24) 4 (24) 1 (6)

Kidney transplant eGFR ⱖ60 mL/min/1.73 m2 30-⬍60 mL/min/ 1.73 m2 15-⬍30 mL/min/ 1.73 m2 ⬍15 mL/min/1.73 m2

37 (26)

15 (25)

20 (30)

2 (12)

46 (32) 64 (44)

29 (48) 24 (39)

13 (19) 34 (51)

4 (24) 6 (32)

32 (22)

7 (11)

20 (30)

5 (29)

3 (2)

1 (2)

—

2 (12)

No. of participants Age ⬍55 y Male sex Body mass index ⬍18.5 kg/m2 18.5-24.9 kg/m2 25-29.9 kg/m2 ⱖ30 kg/m2 Ethnicity White Asian Indian Other

Note: Values expressed as number (percentage). Conversion factor for eGFR in mL/min/1.73 m2 to mL/s/1.73 m2, ⫻0.01667. Abbreviations: BP, blood pressure; CKD, chronic kidney disease; eGFR, estimated glomerular filtration rate; GN, glomerulonephritis.

for the absolute difference in PCRs) with 95% certainty. Consequently, a repeated measurement of PCR that lies outside the 95% repeatability limits is statistically significant and thereby is likely to be indicative of a true change in disease status rather than measurement error. Of the 139 differences in PCR in our sample, 133 (95.7%) were within the repeatability limits. 563

0

500 1000 Day 1 PCR (mg/mmol)

1500

Figure 1. Scatter plot shows spot protein-creatinine ratio (PCR) measurements from 145 patients with chronic kidney disease collected at 9:00 AM on 2 consecutive days. Spearman correlation ␳ ⫽ 0.92.

As listed in Table 2, there is a wide range in variability of repeated test results at different PCR thresholds. At lower baseline PCRs (⬍50 mg/mmol [⬍442 mg/g]), the maximum range in variability for a repeated test result was relatively large and exceeded the baseline value; for example, at a baseline PCR of 20 mg/mmol (177 mg/g), a repeated test result could range from 0-52 mg/mmol (0-460 mg/g), a change of ⫾160%. However, at a higher baseline PCR of 200 mg/mmol (1,768 mg/g), a repeated test result could range from 100-300 mg/mmol (884-2,652 mg/g), a change of ⫾50%. Variability was numerically greater for those with higher baseline PCRs; however, variability as a percentage of baseline value decreased to less than ⫾50% for those with baseline PCR ⬎200 mg/ mmol. The repeatability limits did not differ by a fixed amount by age (⬍55 vs ⱖ55 years; P ⫽ 0.2), sex (P ⫽ 0.9), or eGFR category (P ⫽ 0.2). Statistical Extrapolation of PCR Test Reliability After Multiple Tests The reliability of serial PCR results improved when more test results were averaged and compared with the baseline PCR (Table 2). The range in PCR repeatability limits progressively decreased, and this decrease was inversely proportional to the number of repeated test results averaged and compared with the baseline PCR (Table 2). This was observed consistently at low, medium, and high baseline PCRs. For example, if the baseline PCR was 100 mg/mmol (884 mg/g), the range in PCR repeatability limits if one repeated test was available was 30-170 mg/mmol (265-1,503 mg/g), the average of 2 repeated tests decreased the range in repeatability to 39-161 mg/ mmol (345-1,423 mg/g) and further with the average of 3 repeated tests to 42-158 mg/mmol (371-1,397 mg/g). 564

DISCUSSION Measurement of proteinuria is of critical importance in the diagnosis and management of patients with CKD. Although several methods of measuring proteinuria are clinically available, all have their own limitations. Measurement of PCR in a spot urine sample is convenient for the patient, less prone to collection errors compared with 24-hour collections, predictive of kidney outcomes,6,14 and consequently recommended for routine use by key guideline groups, including NKF-KDOQI (National Kidney Foundation’s Kidney Disease Outcomes Quality Initiative) and CARI (Caring for Australasians With Renal Impairment).1,15 Our study shows that spot urine PCR is subject to substantial day-to-day variability. Such variability in repeated PCR measurements may have important implications in the risk stratification and clinical management of patients with CKD and in clinical research, when a spot PCR is used to indicate protein excretion levels or serial PCR results are used to monitor changes in disease status and response to therapy. Day-to-day variability may arise due to a finite set of variables. Proteinuria is known to show diurnal variation.8 We eliminated this source of variation from our study by collecting samples at 9:00 AM on both collection days. Disease progression, regression, or treatment may alter protein excretion. By restricting the study to clinically stable patients on stable medications and collecting samples on consecutive days, any risk of changing disease status was effectively eliminated. Laboratory storage or measurement errors may contribute to variability; however, we stored all samples at 1°C-4°C for no more than 48 Difference in PCR (mg/mmol) (Day 1 minus Day 2) -200 0 100 200 300 -100

0

Day 2 PCR (mg/mmol) 500 1000

1500

Naresh et al

0

200 400 Average of day 1 and day2 PCR (mg/mmol)

600

Figure 2. Variability in repeated spot protein-creatinine ratio (PCR) measurements from 139 patients with chronic kidney disease collected at 9:00 AM on 2 consecutive days. Fractional polynomial model shows the 95% repeatability limits of agreement between repeated spot urine PCR results. Regression was used to the 95% confidence interval within which repeated PCR measurements are expected to fall while the patient remains in a steady state. Note: this figure does not include participants with PCR ⬎600 mg/mmol (n ⫽ 6). Am J Kidney Dis. 2012;60(4):561-566

Day-to-Day Variability in Spot PCR Table 2. Day-to-Day Variability in Spot PCR by PCR Threshold

Baseline Range of Proteinuria

Low Medium High

Approximate 24-h Protein Excretion (mg/d)

Baseline PCR (mg/mmol)

1 Test Before and 1 Test After

1 Test Before and Average of 2 Tests After

1 Test Before and Average of 3 Tests After

Lowera

Uppera

% Changeb

Lowera

Uppera

% Changeb

Lowera

Uppera

% Changeb

50 200 500 1,000 2,500 3,000 5,000

5 20 50 100 200 300 500

0 0 0 30 100 178 342

21 52 100 170 300 422 658

320 160 100 70 50 41 32

0 0 7 39 114 194 364

19 47 93 161 286 406 636

280 135 86 61 43 35 27

0 0 9 42 119 200 371

18 46 91 158 281 400 629

260 130 82 58 41 33 26

Note: The range of variability at all protein excretion thresholds decreases with repeated test measurements, improving the spot PCR test’s reliability. Conversion factor for PCR in mg/mmol to mg/g, ⫻8.84. Abbreviation: PCR, protein-creatinine ratio. a Repeatability limits for PCR (in mg/mmol) or the upper and lower range within which 95% of repeated PCR measurements should be present depend on the number of repeated tests. b From baseline.

hours and ran samples on a single analyzer, including paired samples in single runs to minimize any risk of this. The coefficient of variation of each laboratory measure of urinary protein and creatinine was ⬍5%. Thus, we are confident that the day-to-day variation we reported represents inherent test variability under “ideal” circumstances. Test variability is likely to be of at least this magnitude in the clinic. The range in PCR test variability differs substantially with the magnitude of proteinuria: absolute variability in PCR increases with the magnitude of proteinuria. However, as a percentage of baseline PCR, variability decreases. Our data show that for a patient with baseline PCR of 20 mg/mmol (177 mg/ g), on repeated testing, PCR will be 0-52 mg/mmol (0-460 mg/g) with 95% certainty (range in variability, ⫾160%). For a patient with baseline PCR of 200 mg/mmol (1,768 mg/g), on repeated testing, PCR will be 100-300 mg/mmol (884-2,652 mg/g) with 95% certainty (range in variability, ⫾50%). Thus, a 100% change in PCR for a patient with low-grade proteinuria (PCR ⬍50 mg/mmol [⬍442 mg/g]) is likely to reflect test variability rather than a change in disease status, whereas a 100% decrease in PCR for a patient with high-grade proteinuria is very likely to indicate a true reduction in proteinuria. This is of importance to the clinician because variability in serial PCR measurements can influence diagnostic thresholds and management decisions, while not necessarily reflecting a true biological change in patient status. To facilitate decision making in clinical practice, we have provided a series of reference ranges for repeated PCR measurements, such that repeated values that lie outside the repeatability limits indicate a change in disease status, either improvement or progression, with ⬎95% probability (Table 2). Am J Kidney Dis. 2012;60(4):561-566

To ascertain whether PCR test reliability could be improved by repeated testing, we statistically extrapolated and compared the mean of 2 or 3 repeated test results with the baseline PCR (because it was not feasible in this study to conduct multiple repeated tests). Although PCR test reliability was improved by doing so, such improvement was modest (Table 2). Our study has several limitations. By design, we included only patients with stable kidney function and because patients encountered in clinical practice may or may not have stable kidney function, this may affect the generalizability of our results. Although this study population otherwise was representative of a typical CKD clinic, further prospective studies are needed to replicate our findings in other racial and ethnic groups. Because spot PCR increasingly is being used to stratify kidney and cardiovascular risk, it behooves further studies to evaluate the relationship between the threshold change in serial spot PCR measurements and the development of kidney and cardiovascular outcomes, after allowing for day-today variability in protein excretion. Further research also is needed to differentiate whether similar day-today variation exists in the measurement of albumin and nonalbumin proteins excreted in urine. Spot urinary PCR commonly is used in research settings. Our study found substantial day-to-day variability at an individual level, but no significant difference between the mean values of repeated measurements for the entire cohort. This suggests that urine PCR may be suitable for comparing mean values between groups or at different times for a study population. However, caution should be taken in using PCR in studies with small numbers and studies in which the number of participants exceeding a thresh565

Naresh et al

old is used as an outcome measure, such as change in CKD stage. In conclusion, measurement of PCR from a spot urine sample is convenient; however, the test is subject to substantial day-to-day variability. Although absolute variability varies in proportion to the magnitude of proteinuria, percentage of variability bears an inverse relationship. Such variability may limit the usefulness of spot PCR and should be borne in mind when using spot PCR to monitor and manage patients with CKD. We provide tabulated repeatability limits to enable clinicians to determine whether changes in serial PCR measurements are likely to reflect a change in clinical condition or simply test variability.

ACKNOWLEDGEMENTS We thank Ms Georgia Whitman, Associate Professor Josette Eris, Dr Paul Snelling, Associate Professor Adrian Gillin, Dr Vicki Levidiotis, Dr Kate Wyburn, and Professor David Harris for assistance with patient recruitment. Support: None. Financial Disclosure: The authors declare that they have no relevant financial interests.

REFERENCES 1. National Kidney Foundation. K/DOQI Clinical Practice Guidelines for Chronic Kidney Disease: evaluation, classification and stratification. Am J Kidney Dis. 2002;39(2)(suppl 1):S1-S266. 2. Go AS, Chertow GM, Fan D, McCulloch CE, Hsu C. Chronic kidney disease and the risks of death, cardiovascular events, and hospitalization. N Engl J Med. 2004;351(13):12961305. 3. Halbesma N, Kuiken DS, Brantsma AH, et al; for the PREVEND Study Group. J Am Soc Nephrol. 2006;17(9):25822590. 4. Iseki K, Ikemiya Y, Iseki C, Takishita S. Proteinuria and the risk of developing end stage renal disease, Okinawa, Japan. Kidney Int. 2003;63(4):1468-1474.

566

5. White SL, Yu R, Craig JC, Polkinghorne KR, Atkins RC, Chadban SJ. Diagnostic accuracy of urine dipsticks for detection of albuminuria in the general community. Am J Kidney Dis. 2011;58(1):19-28. 6. Ginsberg JM, Chang BS, Matarese RA, Garella S. Use of single voided urine samples to estimate quantitative proteinuria. N Engl J Med. 1983;309:1543-1546. 7. Ruggeneti P, Gaspari F, Perna A, Remuzzi G. 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 without diabetes. BMJ. 1998;316(7130):504-509. 8. Koopman MG, Krediet RT, Koomen GCM, Strackee J, Arisz L. Circadian rhythm of proteinuria: consequences of the use of protein:creatinine ratios. Nephrol Dial Transplant. 1989;4(1):9-14. 9. Bossuyt PM, Reitsma JB, Bruns DE, et al. Towards complete and accurate reporting of studies of diagnostic accuracy: the STARD Initiative. Standards for Reporting of Diagnostic Accuracy. Ann Intern Med. 2003;138(1):40-44. 10. Levey AS, Coresh J, Greene T, et al; for the Chronic Kidney Disease Epidemiology Collaboration. Using standardized serum creatinine values in the Modification of Diet in Renal Disease Study equation for estimating glomerular filtration rate. Ann Intern Med. 2006;145:247-254. 11. Carter JL, Thompson CRV, Stevens P, Lamb EJ. Does urinary tract infection cause proteinuria or microalbuminuria? A systematic review. Nephrol Dial Transplant. 2006;21:3031-3037. 12. Bland JM, Altman DG. Measuring agreement in method comparison studies: Stat Methods Med Res. 1999;8:135-160. 13. Sevrukov AB, Bland JM, Kondos GT. Serial electron beam CT measurements of coronary artery calcium: has your patient’s calcium score actually changed? AJR Am J Roentgenol. 2005;185: 1546-1553. 14. Chronic Kidney Disease Prognosis Consortium. Association of estimated glomerular filtration rate and albuminuria with all-cause and cardiovascular mortality in general population cohorts: a collaborative meta-analysis. Lancet. 2010;375(9731):20732081. 15. Caring for Australians With Renal Impairment (CARI). The CARI guidelines. Urine protein as diagnostic test: performance characteristics of tests used in the initial evaluation of patients at risk of renal disease. Nephrology (Carlton). 2004;9(suppl 3):S8S14 Nephrol 43:110-115.

Am J Kidney Dis. 2012;60(4):561-566