Current issues of ACP Journal Club are published in Annals of Internal Medicine


Albumin-to-creatinine ratio in a timed overnight urine sample was accurate for screening for microalbuminuria in diabetes mellitus

ACP J Club. 1999 Sept-Oct;131:47. doi:10.7326/ACPJC-1999-131-2-047

Related Content in the Archives
• Correction: Albumin-to-creatinine ratio in a timed overnight urine sample was accurate for screening for microalbuminuria in diabetes mellitus

Source Citation

Bakker AJ. Detection of microalbuminuria. Receiver operating characteristic curve analysis favors albu-min-to-creatinine ratio over albumin concentration. Diabetes Care. 1999 Feb;22:307-13. [PubMed ID: 99266492]



What is the relative accuracy of a first morning urinary albumin-to-creatinine ratio (UACR) and urinary albumin concentration (UAC) for detecting microalbuminuria in patients with diabetes mellitus?


Comparison of UACR and UAC with a timed overnight urinary albumin excretion rate (UAER) using receiver-operating characteristic (ROC) curves.


{A regional health care center in Leeuwarden, the Netherlands.}*


2394 patients ({mean age 57 y}*, 51% women) with diabetes who had routine screening for microalbuminuria. Patients with urine samples that were subject to timing errors were excluded.

Description of tests and diagnostic standard

All patients completed a timed overnight urine collection, and the UAER, UACR, and UAC were determined. The UAER was used as the diagnostic standard. Patients with UAER ≤ 20 µg/min were characterized as having normoalbuminuria.

Main outcome measures

Sex- and age-specific ROC curves were constructed for analysis of the sensitivity and specificity of UACR and UAC measurements as screening tests for microalbuminuria. The cut points that maximized both sensitivity and specificity were determined.

Main results

37% of men and 28% of women had microalbuminuria. The maximum cut point for UACR differed between men and women (1.8 and 2.5 g/mol, respectively), but the maximum cut point for UAC was 15 mg/L for both men and women; the point estimates for sensitivity and specificity were higher for UACR than UAC (Table). The maximum cut point increased with increasing age for UACR but did not vary with age for UAC.


The point estimates for the urinary albumin-to-creatinine ratio were more accurate than the urinary albumin concentration as a screening test for microalbuminuria in patients with diabetes mellitus. The optimal diagnostic ratio increased with age and differed by sex.

Source of funding: None.

For correspondence: Dr. A.J. Bakker, Department of Clinical Chemistry, Klinisch Chemisch Laboratorium, P.O. Box 850, 8901 BR, Leeuwarden, The Netherlands. FAX 31-582-882-227.

*Information supplied by author.

Table. Sex-specific urinary albumin-to-creatinine ratio (UACR) and urinary albumin concentration (UAC) test features at maximum cut points for the detection of microalbuminuria‡

Test Sex Cut point Sensitivity (95% CI) Specificity (CI) +LR -LR
UACR Men 1.8 g/mol 94% (92 to 97) 93.1% (91 to 95) 13.4 0.06
UACR Women 2.5 g/mol 94% (92 to 96) 92.3% (91 to 94) 11.8 0.07
UAC Men 15 mg/L 90% (87 to 93) 88.8% (86 to 91) 8.2 0.1
UAC Women 15 mg/L 89% (85 to 92) 90.4% (88 to 92) 8.0 0.1

‡Abbreviations defined in Glossary; LRs calculated from data in article.


There has been controversy over whether UACR is more accurate than UAC for detecting microalbuminuria; an earlier study suggested that they were equivalent (1). In the study by Bakker, which had greater statistical power than similar previous studies, UACR seemed to be more sensitive and more specific than UAC for predicting microalbuminuria when appropriate age- and sex-specific cut points were used. Therefore, it is preferable to measure UACR in addition to UAC because the additional cost is small.

This study has several minor limitations in terms of generalizability to clinical practice. First, UACR and UAC were measured on the same timed, overnight sample, which was also used for the diagnostic standard measurement. Use of untimed, spot urine specimens offers advantages in terms of convenience and therefore compliance. Warram and colleagues (2) found that the UACR in a spot specimen closely correlated with UAER in a subsequent timed collection (R2 = 0.935). It would have been more clinically relevant if Bakker had measured UACR and UAC on a separate "spot" sample collected later that morning. Second, this study did not take into account the role of repeated testing. Microalbuminuria is often transient, and at least 2 of 3 samples in a 3- to 6-month period should be positive before persistent microalbuminuria is diagnosed.

This article provides additional support for the use of UACR in screening for microalbuminuria. It unfortunately does not test the accuracy of an untimed, spot UACR for this purpose.

Ronald J. Sigal, MD, MPH
Loeb Health Research UnitOttawa Hospital, Civic CampusOttawa, Ontario, Canada


1. Zelmanovitz T, Gross JL, Oliveira JR, et al. The receiver operating characteristics curve in the evaluation of a random urine specimen as a screening test for diabetic nephropathy. Diabetes Care. 1997;20:516-9.

2. Warram JH, Gearin G, Laffel L, Krolewski AS. Effect of duration of type I diabetes on the prevalence of stages of diabetic nephropathy defined by urinary albumin/creatinine ratio. J Am Soc Nephrol. 1996;7:930-7.