The MICRAL(R) test for diabetic microalbuminuria: Predictive values as a function of prevalence

The MICRAL(R) test is an immunospecific dipstick for detection of low concentrations of albumin in urine (microalbuminuria). The test is intended to be used for screening in an ambulatory setting. The utility of the test depends on its ability to accurately predict which patients will be classified as either microalbuminuric or normoalbuminuric by means of a standard laboratory method for determination of albumin in urine. We have analysed data from studies with a total of 2904 samples of urine from diabetic patients. The data are from our own study (190 samples) and from a selected set of 10 publications. The results from standard laboratory measurements of albumin in urine were used as ''gold standards''. The sensitivity and specificity as calculated from the pooled data were 83.2 and 92.3%, respectively. The predictive values were calculated using simulated changes in the prevalence of microalbuminuria (MAU). At a prevalence of MAU of 1% the predictive value of a negative test is 99.9% but that of a positive test only 9.8%. At a prevalence of MAU of 80% the positive predictive value is 97.7% and the negative predictive value 57.9%. In general, any change in the prevalence will lead to a change in the predictive values. Thus the prevalence of MAU in the given clinical setting is a decisive factor in determining the utility of the MICRAL(R) test.