Annals of Clinical & Laboratory Science 31:99-102 (2001)
© 2001 Association of Clinical Scientists
Decreasing the Variability Observed in Urine Analysis
Z.K. Shihabi1,
R.P. Schwartz2 and
M.J. Pugia3
1 Department of Pathology, and 2 Department of Pediatrics and Brenner Children’s Hospital, Wake Forest University Baptist Medical Center, Winston-Salem, North Carolina;
3 Bayer Corporation, Elkhart, Indiana
Address correspondence to Zakariya K. Shihabi, PhD, Department of Pathology, Wake Forest University School of Medicine, Winston-Salem, NC 27157, USA; tel 336 716 2639; fax 336 716 9944; e-mail zshihabi{at}wfubmc.edu.
 |
Abstract
|
|---|
Urine analysis is affected significantly by biological variability. The objective of this study was to study the feasibility of reducing the biological variability of excretion of various analytes in urine, especially albumin in children with diabetes, by mixing small volumes of early morning samples. Twenty-two male children with type 1 diabetes collected early morning aliquots of approximately 10 ml of urine on 3 consecutive days and kept them refrigerated in sealed containers. The urine collection was repeated every 4–6 months in the diabetic children. Ten normal children and 10 normal adults participated as controls. The specimens were analyzed individually and as mixed samples for each subject. Mixing the 3 urine samples before analysis decreased the biological variability of all urine assays (albumin, glucose, creatinine, total protein, potassium). The diabetic children had 3 times higher variability of urine albumin (as a ratio to creatinine) compared to normal children, when the urine samples were collected individually (61% vs 19%, respectively). The variability in the diabetic children decreased when the 3 specimens were analyzed as a single sample after mixing, especially when urine albumin was expressed as a ratio to creatinine. Blood glycated hemoglobin levels correlated better with urine glucose levels when 3 urine samples were mixed before analysis.
(received 26 July 2000; accepted 15 September 2000)
Keywords: Microalbuminuria, urine glucose, urine collection, diabetes mellitus
|
Introduction
|
|---|
Urine analysis is affected by the great biological variability observed for most analytes, which exceeds 50% in many instances. However, the analytical or instrumental variability is below 5%. This high biological variability diminishes the clinical significance and utilization of urine analysis in medical practice.
Two approaches are usually adopted to decrease the variability: collecting a large volume of urine over a long period of time (eg, 24 hr) or collecting random urine samples and reporting the results as a ratio to creatinine. Both approaches have limitations. Among the various tests, two are of special concern in diabetics: urine glucose and albumin (microalbuminuria).
Microalbuminuria (MA), ie, the urinary excretion of albumin between 30–300 mg/day, is a predictor of diabetic complications in general, and specifically of nephropathy in type 1 diabetes and coronary disease in type 2 diabetes [1]. A major problem with MA is the high biological variability in individual patients, yielding a coefficient of variation (CV) of 30–100% [2–4] or a correlation coefficient (r) of 0.43 between 2 samples 1 yr apart [4]. This variability decreases the usefulness of MA testing when following a particular patient over time. Investigators have tried to decrease this variability by expressing MA results as a ratio to creatinine [5–10], or a timed excretion rate (µg/min) [11,12]. Smulders et al [9] found great variability of MA in adults regardless of the methods of collection or expression.
Many investigators find that overnight urine samples have the least variability [2,9,11,13]. However, the variability of overnight samples for MA still remains relatively high (35–50%) [3,9]. It is impractical to obtain timed samples or 24-hr urine collections in outpatients, especially children. Smulders et al [9] suggested decreasing the variability by using the median of multiple urine collections. This approach is costly, since it entails the analysis of several urine samples.
The previous approaches for urine collection or expression of results have not been widely accepted. A more practical and precise method for collection of urine samples would aid in diagnosing, monitoring, and treating patients with diabetes. We studied the feasibility of reducing the variability of several urine tests by collecting 3 small volumes (~10 ml) of urine on successive mornings and mixing them before analysis. We found the variability of several tests was decreased by this simple approach. The results were calculated as concentrations (eg, mg/L) and as ratios to urine creatinine (eg, mg/g Cr).
|
Materials and Methods
|
|---|
Subjects.
Twenty-two male children with diabetes (mean age 12.9 yr, range 4–17 yr), seen during routine clinic visits, were invited to participate in this study after the consent of their parents was obtained. The study was approved by the Institutional Review Board of Wake Forest University School of Medicine. Ten normal children and 10 normal adults were controls. Early on 3 consecutive mornings prior to a scheduled clinic visit, the child voided ~10 ml of urine into a small plastic-capped bottle (20 ml volume). The child’s parents stored the 3 samples in a refrigerator inside a large plastic-capped container (200 ml volume). The 3 urine samples were brought to the clinic on the day that the last sample was collected. Urine collections were repeated every 4–6 months in diabetic children.
Chemical Analysis.
The urine samples were analyzed individually and as mixed samples (equal volumes of 1 ml) for each patient. When the samples arrived in the laboratory, microalbuminuria (MA) was measured by nephelometry ("Array" analyzer, Beckman Instruments, Brea, CA) [14]. The other urine analytes were measured after the samples were stored at -20° C. Urine glucose was measured by the glucose oxidase method with a specific electrode [15] and urine potassium by an ion-specific electrode [15] (Beckman "CX 3" analyzer). Urine creatinine was measured by the Jaffé reaction [15]. A blood specimen collected on the day of the last urine collection was assayed for glycated hemoglobin (glycated Hb) by HPLC using an affinity column (Primus Corporation, Kansas City, MO) [16]. All tests had an analytical variability of <5%.
Statistical Analysis.
Microsoft Excel software was used to calculate values for the mean, SD, CV, and correlation coefficient (r) of data sets. Trend analysis was performed according to Conover [17].
|
Results
|
|---|
None of the urine analytes showed a statistically significant trend with samples from successive days.
The average individual variability of all urine analytes was decreased greatly by mixing the 3 samples and analyzing the composite sample (Table 1
). Glucose had an average variability of 106% based on analyses of single samples; the variability was decreased to 20% when 3 samples were mixed. Creatinine had the lowest variability (39%) based on analyses of single samples; it decreased to 2% when 3 samples were mixed.
View this table:
[in this window]
[in a new window]
|
Table 1. Average individual variability of the analytical results for selected tests, using single urine samples vs mixed urine samples from diabetic children.
|
|
The mean value for MA was 2 times higher in the diabetic children, compared to normal children (Table 2). The diabetic children had 3 times higher variability of MA in single samples compared to normal children when expressed as mg/L (89% vs. 33%, respectively). The variability of MA in single samples was decreased when expressed as mg/g of creatinine (61% vs 19%), but the difference between the 2 groups remained large.
View this table:
[in this window]
[in a new window]
|
Table 2. Average individual variability of the analytical results for microalbuminuria (MA), using single urine samples vs mixed urine samples.
|
|
Analyzing the composite samples, compared to separate samples, reduced the variability of MA in diabetic children, especially when the results were based on creatinine (from 61% to 13%). Reduced variability was also evident when diabetic patients were followed up over a 10-month period, and was also seen in the normal children and adults (Table 2
).
The correlation coefficients between the glycated Hb levels in blood and the glucose concentrations in individual urine samples collected on days 1, 2, and 3 were 0.34, 0.63, and 0.64, respectively. The correlation of glycated Hb levels in blood vs urine glucose concentrations was improved (r, 0.71), when 3 urine samples were mixed prior to glucose analysis.
|
Discussion
|
|---|
We found that mixing an aliquot of 3 successive morning urine samples decreased the variability for all measured analytes, including microalbumin. This simple procedure, even more than the correction based on creatinine excretion, results in lower coefficients of variation. The procedure for sample collection is practical for both children and adults, and less expensive than analyzing each sample separately.
Because of poor correlation with blood glucose concentrations, analysis of urine glucose has been largely discontinued [19–21], except for limited use in managing type 2 diabetics, or for diabetes screening in developing countries [20,21]. Bacterial contamination can affect urine glucose levels. However, urine glucose testing is noninvasive compared to blood glucose testing. In the present study, urine glucose showed greater variability than other analytes. Nonetheless, the glucose levels in the mixed urine samples gave a reasonable correlation (r, 0.71) with the corresponding glycated Hb levels in blood samples.
In patients with renal disorders, Newman et al [18] noted high variability for MA and urine total protein, which decreased when the results were expressed as a ratio to creatinine. Numerous studies have shown the importance of MA as a predictor of gross proteinuria and diabetic nephropathy in patients with type 1 diabetes [22–24]. Several drugs, such as angiotensin-converting enzyme inhibitors, can decrease MA and thus delay the progress of diabetic nephropathy [25–27]. Many investigators have tried to decrease the variability of urine analytes, especially MA, by different methods of sample collection or expression of results. The present study confirms the great variability in urine testing for MA, especially in diabetic children.
The coefficients of variation (CV) reported in this study should be viewed as relative indices of individual variability. From a statistical viewpoint, the CV should theoretically decrease by the square root of the number of samples assayed (1.73 for 3 samples). The observed values reflect this theoretical decrease. The authors note that the choice of the optimum number of samples depends on the variability of the test, the desired precision, and the convenience of the patient.
In summary, less variability was observed for MA, glucose, and other analytes when analyses were performed on mixed samples of urine voided on three successive mornings, rather than on single samples.
|
Acknowledgements
|
|---|
The authors gratefully acknowledge the assistance of Patricia Bobbitt, RN, and Mark Hinsdale.
|
References
|
|---|
- Konen CJ, Curtis LG, Shihabi ZK, Dignan MB. Microalbuminuria and diabetes mellitus. J Fam Pract 1990;31:505–510.[Medline]
- Howey JE, Browning MC, Fraser CJ. Selecting the optimum specimen for assessing slight albuminuria, and a strategy for clinical investigation: novel uses of data on biological variation.Clin Chem 1987;33:2034–2038.[Abstract/Free Full Text]
- Marre M, Claudel JP, Ciret P, Luis N, Suarez L, Passa P. Laser immunonephelometry for routine quantification of urinary albumin excretion. Clin Chem 1987;33: 209–213.[Abstract/Free Full Text]
- Lauritzen T, Christiansen JS, Brock A, Mogensen CE. Repeated screening for albumin-creatinine ratio in an unselected population. The Ebeltoft Health Promotion Study, a randomized, population-based intervention trial on health test and health conversations with general practitioners. J Diabetes Complicat 1994;8:146–149.[Medline]
- Bangstad H-J, Jorgensen KD, Kjaersgaard F, Mevald K, Hanssen KF. Urinary albumin excretion rate and puberty in non-diabetic children and adolescents. Acta Paediatr 1993;82:857–62.[Medline]
- Hutchinson AS, Paterson RK. Collecting urine for microalbumin assay. Diabet Med 1988;5:527–532.[Medline]
- Bakker AJ. Detection of microalbuminuria. Receiver operating characteristic curve analysis favors albumin-to-creatinine ratio over albumin concentration. Diabetes Care 1999;22:307–313.[Abstract/Free Full Text]
- Gross JL, Zelmnovitz T, Oliveira J, Azevedo MJ. Screening for diabetic nephropathy: Is the measurement of urinary albumin to creatinine ratio worthwhile? Diabetes Care 1999;22:1599.[Free Full Text]
- Smulders YM, Slaats EH, Rakic M, Smulders FT, Stehouwer CD, Silberbusch J. Short-term variability and sampling distribution of various parameters of urinary albumin excretion in patients with non-insulin-dependent diabetes mellitus. J Lab Clin Med 1998; 132:39–46.[Medline]
- Zelmanovitz T, Gross JL, Oliveira JR, Paggi A, Tatsch M, Azevedo MJ. The receiver operating characteristics curve in the evaluation of a random urine specimen as a screening test for diabetic nephropathy. Diabetes Care 1997;20:516–519.[Medline]
- Kodama K, Tomika M, Otani T, Shimizu S, Uchigata Y, Hirata Y. The range of albumin concentrations in the single-void first morning urine of 1090 healthy young children. Diabet Res Clin Pract 1990;9:55–58.[Medline]
- Rowe DJ, Baga M, Betts PB. Normal variations in rate of albumin excretion and albumin to creatinine ratios in overnight and daytime urine collections in non-diabetic children. Br Med J 1985;291:693–694.
- Vestbo E, Damsgaard EM, Froland A, Mogensen CE. Microalbuminuria in a population-based cohort: ways of expression of albuminuria (Fredericia Study, second generation). J Diabetes Complicat 1994;8:176–177.[Medline]
- Roberts WL, Calcote CB, Cook CB, Gordon DL, Moore ML, Moore S, Scheer WD, Snazelle BA. Comparison of four commercial urinary albumin (microalbumin) methods: implications for detecting diabetic nephropathy using random urine specimens. Clin Chim Acta 1998;273:21–33.[Medline]
- Peake MJ, Pejakovic M, White GH. Laboratory evaluation of the Beckman Synchron CX3 clinical chemistry analyzer. Clin Chem 1988,34:404–407.[Abstract/Free Full Text]
- Nuttal FQ. Comparison of percent total gHb with percent Hb A1c in people with and without known diabetes. Diabetes Care 1998;21:1475–1480.[Abstract/Free Full Text]
- Conover WJ. Practical Nonparametric Statistics, John Wiley & Sons, New York, 1971; pp 122–142.
- Newman DJ, Pugia MJ, Lott JAQ, Wallace JF., Hiar AM. Urinary protein and albumin excretion corrected by creatinine and specific gravity. Clin Chim Acta 2000; 294:139–155.[Medline]
- Gallichan M. Self monitoring of glucose by people with diabetes: evidence based practice. Brit Med J 1997; 314:964–967.[Free Full Text]
- Kabadi UM. Urine glucose testing: reliable backup for whole blood glucose monitoring. Family Pract 1992; 34:495–497.
- Feleke Y, Abdulkadir J. Urine glucose testing: another look at its relevance when blood glucose monitoring is unaffordable. Ethiop Med J 1998;36:93–99.[Medline]
- Messent JW, Elliott TG, Hill RD, Jarrett RJ, Keen H, Viberti GC.Prognostic significance of microalbuminuria in insulin-dependent diabetes mellitus: a twenty-three year follow-up study. Kidney Int 1992;41:836–839.[Medline]
- Viberti GC. Diabetic renal disease in type 1 diabetes: aetiology and prevention. Diabet Med 1991;8:S38–42.
- Nielsen S, Schmitz A, Rehling M, Mogensen CE. The clinical course of renal function in NIDDM patients with normo- and microalbuminuria. J Intern Med 1997;24:133–141.
- Mogensen CE. Drug treatment for hypertensive patients in special situations: diabetes and hypertension. Clin Exp Hypertens 1999;21:895–906.
- Mogensen CE. Optimal management of hypertensive patients with type 2 diabetes and microalbuminuria. Int J Clin Pract Suppl 1998;98:13–20.[Medline]
- Parving HH. Benefits and cost of antihypertensive treatment in incipient and overt diabetic nephropathy. J Hypertens Suppl 1998,16:S99–101.[Medline]
Top
Abstract
Introduction
