NDT Advance Access published online on November 5, 2008
Nephrology Dialysis Transplantation, doi:10.1093/ndt/gfn612
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Use of a first-line urine protein-to-creatinine ratio strip test on random urines to rule out proteinuria in patients with chronic kidney disease
1 Department of Clinical Biochemistry, Salford Royal NHS Foundation Trust, Salford, UK 2 Highover Park, Amersham, Bucks, UK 3 Department of Renal Medicine, Salford Royal NHS Foundation Trust, Salford, UK 4 Department of Clinical Biochemistry, University of Oxford, Oxford, UK
Correspondence and offprint requests to: Mark Guy, Dept Clinical Biochemistry, Salford Royal NHS Foundation Trust, Salford, UK. Tel: +44-(0)1612064968; Fax: +44-(0)1617887443; E-mail: mark.guy{at}srft.nhs.uk
 |
Abstract |
|---|
 Top Abstract IntroductionSubjects and methodsResultsDiscussionReferences |
|---|
Background. Urine protein strip tests are often used in the ward or clinic as first-line measures of proteinuria. The ability of a semi-quantitative meter-read strip test for the protein:creatinine ratio, Multistix* PRO® 10LS (Siemens Medical Solutions, Tarrytown, USA), was assessed as a first-line test to exclude significant proteinuria in the monitoring of patients with established chronic kidney disease.
Methods. Eighty-six patients attending a hospital renal outpatient clinic collected three random urine samples during a 24-h period. Random urine protein:creatinine ratios measured by the strip test were compared to the laboratory estimation of 24-h protein excretion on that same day.
Results. At significant protein excretion of 0.3 g/24 h, the strips elicited negative predictive values in the range of 91.2–94.1% and negative likelihood ratios of 0.01–0.12, using all the random urines. Receiver–operator characteristic curve analysis also demonstrated good performance with all samples.
Conclusions. The strip test allows the physician to rule out significant proteinuria at the patient consultation on a random urine sample, obviating the need for specially collected samples, and with the added benefit of reducing the need for a lengthy and costly quantitative laboratory follow-up by 
40–48%.
Keywords: chronic kidney disease; protein:creatinine ratio; proteinuria; urine strip tests
|
Introduction |
|---|
|
TopAbstractIntroductionSubjects and methodsResultsDiscussionReferences |
|---|
Urine strip tests have been used for many years in the physician's office and hospital clinic as first-line measures of the presence of proteinuria, in order to detect chronic kidney disease (CKD), while proteinuria is also an associated independent indicator of risk for cardiovascular disease. Thereafter, laboratory tests are generally performed to confirm and quantitate the increased excretion of protein or albumin, in accordance with the recognized classification and guidelines for CKD [1–3], so that appropriate patient management can then be initiated.
These quantitative laboratory tests have traditionally entailed timed 24-h urine collections, which are notoriously cumbersome and unreliable, and it has been suggested that they could be replaced by the protein- or albumin-to-creatinine ratio from random urines [1,3,4]. Although it is advocated to use creatinine ratios to compensate for variations in the urine flow rate [5], the accuracy of this process may be reduced in individuals with high or low creatinine excretion, due to high or low muscle mass. However, inaccuracies in the collection of 24-h urines are often far greater. A systematic review of the literature on the relationship of the random urine protein:creatinine ratio (P:C ratio) showed that if the ratio was <30 mg/mmol, it provided a reliable rule-out of significant proteinuria, with a cumulative negative likelihood ratio (LR) of 0.14 (95% CI 0.09–0.24) based on 10 studies in women presenting with symptoms of pre-eclampsia [4]. This would reduce the unnecessary collection of 24-h urines, with its associated inconvenience, unreliability, time delays and costs by routinely utilizing urinary P:C ratio testing. It would reduce the number of confirmatory tests performed from true negative results, and lead to only a small number of false positive tests that would require further quantitation. The question then arises as to whether a first-line reagent strip test can also perform this ‘rule-out’ function, in the physician's office, allowing the provision of immediate advice or reassurance to the patient at the consultation.
A relatively new semi-quantitative urine strip test, Multistix PRO® 10LS (Siemens Medical Solutions, Tarrytown, USA), incorporates dual protein reagent pads—‘Protein-Low’ and ‘Protein-High’—alongside the usual, more traditional, tests including, for example, glucose, blood and ketones. The Protein-Low pad is more specific for albumin and augments the traditional protein pad, while still being reported as protein. The strip test is also unique in that it includes a test pad for creatinine allowing for the semi-quantitative estimation of the P:C ratio, which demonstrates effective performance in detecting proteinuria [6,7], especially when the strips are read using an automated strip reader.
There have been no studies to date that have evaluated the diagnostic accuracy of point-of-care P:C ratio strip tests in ruling in or ruling out proteinuria. The objective of this study, therefore, was to assess whether this P:C ratio strip test, used as a first-line screen, could effectively rule out proteinuria in random urine samples from hospital renal outpatients, by comparison with quantitative laboratory results for 24-h protein excretion rates at defined cut-off values. In addition, the most appropriate random urine sample to use, if any, was also assessed, for application in the physician's office or hospital clinic.
|
Subjects and methods |
|---|
|
TopAbstractIntroductionSubjects and methodsResultsDiscussionReferences |
|---|
A total of 86 patients requiring 24-h urine collections as part of the investigation or management of CKD were recruited from renal outpatient clinics at Hope Hospital, Salford. Patient consent was first obtained after explaining the project objectives with the aid of a patient information sheet. Patients were requested to collect urine for 24 h in four portions. After discarding the first sample of the morning, the following samples were collected: (a) second morning, (b) third or random, (c) the remaining urine passed that day and (d) the early-morning urine (EMU) the following day.
The urine samples (a), (b) and (d) were tested with Multistix PRO 10LS reagent strips and read semi-quantitatively on the Clinitek Status® automated strip reader (Siemens Medical Solutions), for the P:C ratio. The P:C ratio was reported as negative, negative/dilute (requires repeating) and positive at 17, 33.9 and 
56.6 mg protein/mmol creatinine. The results equating to 17, 33.9 and 56.6 were all combined and reported as ‘positive’ for data analysis.
For the purposes of this study, P:C ratio analyser strip readings of ‘negative/dilute’ were treated as ‘normal’ (together with those reported as ‘normal’), since these patients would not be returning to the hospital clinic until their next scheduled appointment. In the primary care setting, these patients would have been requested to provide a repeat specimen, preferably early-morning, to obtain a more concentrated sample which could be more quantitatively analysed.
The strip testing was performed by two operators (MG and JB), who checked strip quality control using Chek Stix® controls (Siemens Medical Solutions). The test strips and strip reader are simple to use, with the latter requiring minimal maintenance and no calibration as this is performed automatically. The equipment is portable and specifications state that the strips are stable until the expiry date printed on the bottle, provided that they are stored at 15–30°C.
Following strip testing, all urine aliquots were combined to give the full 24-h collection, and urine total protein excretion measured in the laboratory using pyrogallol red on the Roche Integra 800 analyser [8]. The operators of the strip and automated tests were unaware of the results from the other procedures when the respective analyses were being performed.
Statistical analyses on semi-quantitative data were performed by cross-tabulation in 2 x 2 tables, and the calculation of sensitivity, specificity, predictive values and LRs were performed to assess the diagnostic accuracy for the absence or presence of proteinuria, for all three urine aliquots and for each of the 24-h urine excretion cut-off values, while Spearman correlation coefficients and scatter plots were also produced, using Analyse-It for Excel (www.analyse-it.com). Receiver–operator characteristic (ROC) curves were constructed (Analyse-It) to assess whether the nature of the curve data added to the initial analyses and revealed superiority for any of the three urine aliquots.
The study was approved by Greater Manchester Local Ethics Committees with signed written consent being obtained from each patient.
|
Results |
|---|
|
TopAbstractIntroductionSubjects and methodsResultsDiscussionReferences |
|---|
The study was conducted and patients were recruited between September and December 2005, but with samples collected until December 2006. The 86 patients (26 females, 60 males; median age 67 years, age range 28– 86 years) had established CKD, with the following CKD stage distribution: stage 1 = 1.2%, stage 2 = 12.8%, stage 3 = 41.9%, stage 4 = 33.7% and stage 5 = 10.5%. The median eGFR was 36 ml/min/1.73 m2, with a range of 6–102. Sixteen patients (18.6%) had diabetic nephropathy.
The protein excretion cut-off levels chosen for the method comparisons were 0.15, 0.20 and 0.30 g/day, with the first two representing normal/borderline, and the latter significant, proteinuria.
The P:C ratio comparative data for the three urine aliquots at each of the three protein excretion cut-offs are presented in Table 1. The trend in true and false positives and true and false negatives for the Multistix PRO test results was as expected as the reference method cut-off values were increased. Negative predictive values ranged from 67.6% to 71.8% at a cut-off excretion of 0.15 g/day, from 65.7% to 78.9% at 0.2 g/day and from 91.2% to 94.1% at 0.3 g/day. There was some evidence of a trend in the opposite direction for the positive predictive values. Negative LRs were in the region of 0.13–0.29 at excretion rate cut-offs of 0.15 and 0.2 g/day, but improved further to between 0.01 and 0.07 at 0.3 g/day, in spite of more false positives. These were accompanied by positive LRs between 2.56 and 12.0, the latter for an EMU at a cut-off of 0.15 g/day.
|
These data demonstrate that the best diagnostic performance was in ruling out significant proteinuria of 0.3 g/ 24 h or greater; negative LRs were slightly lower for the EMU samples at cut-offs of 0.15 and 0.20, but not for 0.30 g/ 24 h.
The scatter plots and Spearman correlation coefficients displayed in Figure 1 again demonstrate that proteinuria could effectively be ruled out by analysis of any of the three urine aliquots (P < 0.0001).
|
Similar data were obtained using ROC analysis, which combine the attributes of test positivity and negativity, as summarized in Table 2. The area under the curve (AUC) (curve not shown) ranged between 0.801 and 0.889 at 0.15 g/day and 0.20 g/day and between 0.890 and 0.917 at 0.30 g/day (P < 0.0001 in all cases), and again demonstrates that the strip test is effective at ruling out significant proteinuria.
|
|
Discussion |
|---|
|
TopAbstractIntroductionSubjects and methodsResultsDiscussionReferences |
|---|
During the monitoring and management of patients with CKD, it must not be forgotten that results indicating the absence of significant proteinuria are important, giving the physician confidence to withhold further unnecessary lengthy and costly confirmatory testing [4,8]. The P:C ratio is a reliable indicator of protein excretion and is recommended as a screening tool for CKD [3,9,10], and which, if it were available as a strip test, would lend itself for this purpose using an appropriate urine sample. The scatter plots and Spearman correlation coefficients show that there is good agreement between Multistix PRO 10LS results and 24-h protein excretion (Figure 1). Analysis of data in this way, however, does not indicate the clinical performance of the test and whether it can be used to rule in or rule out proteinuria. This is best done using ROC curve analysis or calculation of LRs. In order to confidently ‘rule out’ proteinuria, such a test should demonstrate very few false negatives accompanied by few false positives, and have high negative predictive values, at a defined threshold for protein excretion. Negative LRs <0.2 and positive LRs >5 indicate strong evidence to ‘rule out’ and ‘rule in’ disease, respectively, while ratios <0.1 and >10 indicate convincing evidence [11].
At the significant proteinuria threshold of 0.3 g/day, Multistix PRO 10LS demonstrated negative LRs of 0.12 or lower on all three urine samples (second, third void and EMU), accompanied by positive LRs of 2.7 to 4.8; negative predictive values were all >90% (Table 1). These indicate that, at this threshold, any of these random urines could be used to rule out significant proteinuria, and that the performance is less good as a rule-in test. This is supported by the scatter plots and Spearman coefficients (P < 0.0001) in Figure 1, which demonstrate that any of the three urine aliquots may be used for the rule-out mode. The higher sensitivities observed compared to the specificities also suggest that the test is more suited to the rule-out function. In addition, the number of true negatives at this level suggests a workload reduction of samples sent to the laboratory of between 40% and 48%, at the expense of two or three false negatives, reducing laboratory costs, time and enabling a more efficient investigation of patients. Similar conclusions may also be drawn from the ROC curves, summarized in Table 2, which demonstrate a superior performance in terms of AUC at an excretion threshold of 0.3 g/day.
For patients with diabetes and hypertension, where the presence of microalbuminuria has prognostic significance, strip testing (or laboratory measured urine P:C ratio) that has ‘ruled out’ macro-proteinuria should be followed by testing for microalbuminuria, which can also be performed by point-of-care tests in the clinic or surgery [12,13]. Ideally, testing for microalbuminuria using the albumin:creatinine ratio should be the procedure of choice in CKD patients with diabetes or hypertension.
In summary, Multistix PRO 10LS may be employed to rule out significant proteinuria using a random urine sample. This is especially beneficial in the physician's office or clinic, where the patient can provide a random urine sample without the need for an early-morning or timed specimen, the latter being unreliable and inconvenient. Samples with an abnormal P:C ratio on strip testing would require subsequent laboratory quantitative testing for the P:C ratio; any false positives on first-line testing would be identified, while true positives would be followed up and managed according to the local protocol. This approach represents an aid to immediate clinical decision making in the management of CKD.
|
Acknowledgments |
|---|
The study was supported by a grant from Siemens Medical Solutions, USA, which also supplied the strip testing materials.
Conflict of interest statement. RN and CP were consultants to Bayer Corp., now Siemens Medical Solutions, at the time this project was initiated.
|
References |
|---|
|
TopAbstractIntroductionSubjects and methodsResultsDiscussionReferences |
|---|
- Levey AS, Coresh J, Balk E, et al. National Kidney Foundation practice guidelines for chronic kidney disease: evaluation, classification, and stratification. Ann Intern Med (2003) 139:137–147.
[Abstract/Free Full Text] - Levey AS, Eckardt K-U, Tsukamoto Y, et al. Definition and classification of chronic kidney disease: a position statement from Kidney Disease: Improving Global Outcomes (KDIGO). Kidney Int (2005) 67:2089–2100.[CrossRef][Web of Science][Medline]
- Joint Speciality Committee for Renal Disease. Royal College of Physicians of London and the Renal Association. Available at: http://www.renal.org/guidelines/print/CKDfinalMar07.pdf (accessed June 2007).
- Price CP, Newall RG, Boyd JC. Use of protein:creatinine ratio measurements on random urine samples for prediction of significant proteinuria: a systematic review. Clin Chem (2005) 51:1577–1586.
[Abstract/Free Full Text] - Newman DJ, Pugia MJ, Lott JA, et al. Urinary protein and albumin excretion corrected by creatinine and specific gravity. Clin Chim Acta (2000) 294:139–155.[CrossRef][Web of Science][Medline]
- Croal BL, Finlay D, Davidson E, et al. Evaluation of the Bayer Multistix PRO 10LS point-of-care urine test. Point of Care (2003) 2:144–148.
- Watanabe M, Funabiki K, Tsuge T, et al. Using protein/creatinine ratios in random urine. J Lab Clin Anal (2005) 19:160–166.[CrossRef]
- Watanabe N, Kamei S, Ohkubo A, et al. Urine protein measured with a pyrogallol red-molybdate complex, manually and in a Hitachi 726 autoanalyser. Clin Chem (1986) 32:1551–1554.
[Abstract/Free Full Text] - 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) 7:450–452.
[Abstract/Free Full Text] - National Kidney Foundation. K/DOQI clinical practice guidelines for chronic kidney disease: evaluation, classification and stratification. Am J Kidney Dis (2002) 39(Suppl_1):S1–266.[CrossRef][Web of Science][Medline]
- Boyd JC. Statistical analysis and presentation of date. In: Evidence-Based Laboratory Medicine—Price CP, Christenson RH, eds. (2007) 2nd edn. Washington: AACC Press. 113–140.
- Parsons M, Newman DJ, Pugia M, et al. Performance of a reagent strip device for quantitation of the urine albumin:creatinine ratio in a point of care setting. Clin Nephrol (1999) 5:220–227.
- Parsons MP, Newman DJ, Newall RG, et al. Validation of a point-of-care assay for the urinary albumin:creatinine ratio. Clin Chem (1999) 45:414–417.
[Free Full Text]
Accepted in revised form: 7.10.08
CiteULike 
Connotea 
Del.icio.us What's this?
This article has been cited by other articles:
|
|
 |
|
  E.-Y. Park and T.-Y. Kim How to interpret the protein/creatinine ratio in patients with low GFR Nephrol. Dial. Transplant., December 1, 2009; 24(12): 3892 - 3893. [Full Text] [PDF]
|
|
|
|
|
|
M. Guy, R. Newall, J. Borzomato, P. A. Kalra, and C. P. Price Reply Nephrol. Dial. Transplant., December 1, 2009; 24(12): 3893 - 3893. [Full Text] [PDF]
|
|
|
|
 |
|
 M. Guy, J. K Borzomato, R. G Newall, P. A Kalra, and C. P Price Protein and albumin-to-creatinine ratios in random urines accurately predict 24 h protein and albumin loss in patients with kidney disease Ann Clin Biochem, November 1, 2009; 46(6): 468 - 476. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||