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


Albumin measurements in a random urine sample accurately screened microalbuminuria and macroalbuminuria in NIDDM

ACP J Club. 1997 Nov-Dec;127:76. doi:10.7326/ACPJC-1997-127-3-076

Source Citation

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 Apr;20:516-9.



To assess the accuracy of urinary albumin concentration (UAC) and urinary albumin-to-creatinine ratio (UACR) in a random urine sample (RUS) as a screening test for microalbuminuria and macroalbuminuria in patients with non-insulin-dependent diabetes mellitus (NIDDM).


Comparison of UAC and UACR with 24-hour urinary albumin excretion rate (UAER) using receiver-operating characteristics (ROC) curves.


Diabetes clinic in a tertiary care center in Brazil.


95 patients (mean age 61 y, 52% women) who had NIDDM (mean duration 11 y). Exclusion criteria were cardiac failure or renal tract disease other than diabetic nephropathy.

Description of tests and diagnostic standard

All patients completed a 24-hour urine collection for UAER. The next morning they came to the clinic and an RUS was taken for UAC and UACR measurements. Urinary albumin was measured in duplicate by immunoturbidimetry. One hundred twenty-three 24-hour urine collections with creatinine measures between 700 and 1800 mg were used as the diagnostic standard.

Main outcome measures

ROC curves were constructed to analyze the performance of RUS measurements (UAC and UACR) as screening tests for microalbuminuria (UAER 20 to 200 µg/min) and macroalbuminuria (UAER > 200 µg/min). 2 cut points were determined: the first point with a sensitivity of 100%, and the RUS value that maximized both sensitivity and specificity.

Main results

The areas under the curve for microalbuminuria were 0.98 for UAC and 0.97 for UACR; for macroalbuminuria the areas were 0.99 for UAC and 0.96 for UACR. The UAC and UACR areas for microalbuminuria and macroalbuminuria did not differ (P > 0.05). A UAC value of 16.9 mg/L and a UACR value of 15.0 mg/g were 100% sensitive for microalbuminuria; for macroalbuminuria a UAC value of 174.0 mg/L and a UACR value of 116.0 mg/g were 100% sensitive. Greater than 88% sensitivity and specificity for microalbuminuria were achieved with UAC and UACR values of 33.6 mg/L and 26.8 mg/g, respectively; greater than 90% sensitivity and specificity for macroalbuminuria were achieved with UAC and UACR values of 296.2 mg/L, and 334.3 mg/g, respectively.


Measures of urinary albumin concentration and urinary albumin-to-creatinine ratio in a random urine sample were sound screening tests for microalbuminuria and macroalbuminuria in patients with non-insulin-dependent diabetes mellitus.

Sources of funding: Fundação de Amparo à Pesquisa do Rio Grande do Sul and Hospital de Clínicas de Porto Alegre.

For article reprint: Dr. M.J. de Azevedo, Serviço de Endocrinologia, Hospital de Clínicas de Porto Alegre, Rua Ramiro Barcellos 2350/635, 90035-003, Porto Alegre, RS, Brazil. FAX 55-51-332-83125.


Detecting mild increases in UAER in persons with diabetes is a good idea because early diabetic nephropathy is treatable (1). Screening tests, including the UACR and UAC on an RUS, can be used before resorting to the more inconvenient, costly, but definitive UAER on a timed sample. Most, but not all, agencies recommended only the UACR because the UAC does not account for urine volume and, therefore, may be less reliable at very low or high volumes (1).

This well-done study by Zelmanovitz and colleagues is distinguished by a moderately large sample and a "head-to-head" comparison of the UACR and UAC on an RUS to identify microalbuminuria by an accepted diagnostic standard. No difference in overall diagnostic accuracy between the 2 screening tests was found in the study. Further, cut points were defined for both that yielded 100% sensitivities (i.e., cut points below which no patients with potentially treatable microalbuminuria would be missed) with acceptable, but unavoidably lower (about 75%), specificities (i.e., about 25% of patients without microalbuminuria would go on to have a UAER). However, some caveats are noted. First, the range of urine volumes from which the UAC was determined was not provided; second, it was not clear how many eligible patients did not participate and, if they did not, for what reasons; and third, the RUSs were not truly random because they seemed to have been collected in the morning and not at other times. Nevertheless, these results suggest that the simpler spot UAC on a morning RUS may be as valid a screening test as the UACR.

For now, I will continue to use the UACR on an RUS in judging which patients need to have their UAER measured and which can be left for another year. However, confirmation of the results of this study would make me seriously reconsider the UAC as the preferred screening test.

Jeffrey Mahon, MD
University of Western OntarioLondon, Ontario, Canada


1. Mogensen CE, Keane WF, Bennett PH, et al. Prevention of diabetic renal disease with special reference to microalbuminuria. Lancet. 1995;346:1080-4.