Background: No study has yet investigated the validity of prescreening by albumin measurements in a spot morning urine sample to identify in the general population subjects with microalbuminuria. We therefore tested the diagnostic performance of urinary albumin concentration (UAC) and albumin-creatinine ratio (ACR), measured in a spot morning urine sample, in predicting a urinary albumin excretion (UAE) > or =30 mg in subsequent 24-hour urines (microalbuminuria).
Methods: Subjects (2527) participating in the PREVEND study, a representative sample from the general population, collected a spot morning urine sample and, on average, 77 days later, two 24-hour urine collections.
Results: The ROC curve of UAC in predicting microalbuminuria has an area-under-the-curve of 0.92 with a discriminator value of 11.2 mg/L. Using this cut-off value for UAC, sensitivity in predicting microalbuminuria is 85.0%, and specificity 85.0%. For ACR these values are, respectively: area-under-the-curve 0.93, discriminator value 9.9 mg/g, sensitivity 87.6%, and specificity 87.5%. Sensitivity for UAC in predicting microalbuminuria does not differ significantly from the sensitivity for ACR, whereas the difference between the specificities of UAC and ACR reaches statistical significance, but is numerically very small. In various subgroups characterized by differences in urinary creatinine excretion, the area-under-the-ROC curve, sensitivity, as well as specificity, do not increase relevantly compared to the results in the overall study population. This holds true for ACR as well as UAC.
Conclusion: The diagnostic performance of measuring UAC in a spot morning urine sample in predicting microalbuminuria in subsequent 24-hour urine collections is satisfactory, and, moreover, comparable to that of measuring ACR. In order to keep the burden and costs involved in population screening for microalbuminuria as low as possible, we therefore propose prescreening by measuring UAC in a spot morning urine sample. Those subjects with a UAC above a certain predefined level (e.g., 11 mg/L) should be asked to collect timed urine samples.