A prospective survey was performed in 302 consecutive diabetic outpatients from 3 French diabetic centres to study the sensitivity and specificity of screening for microalbuminuria using Clinitek Microalbumin. Urinary samples with positive (at least one +) proteinuria, hematuria, leucocyturia, or nitrates using the Multistix strip were excluded from the study. Results obtained with Clinitek Microalbumin were compared to those observed with the reference method of the biological laboratory of the centre on the same urinary sample. A positive result was defined as an albumin-to-creatinine ratio > or =30 mg/g. Results were described in terms of sensitivity, specificity, positive and negative predictive values and likelihood ratio. Agreement rates were compared with the Kappa test. In the study population, 48 patients (17%) had a positive microalbuminuria with reference assay. However, different rates were found in each site (25%, 11%, and 15%, respectively, p<0.001). Using the Clinitek Microalbumin, a positive result was found among 86 patients (29%), (39%, 26%, and 23%, respectively). A good agreement was observed in the population as a whole (81%, K=0.47 +/- 0.06) and in each site (77%, 81%, 84%, respectively). Sensitivity was 79% (82%, 80%, 75%), specificity 81% (76%, 81%, 85%), positive predictive value 46% (53%, 35%, 46%), negative predictive value 95% (93%, 97%, 95%), and positive likelihood ratio 4.2 (3.4, 4.3, 5.0, respectively). Due to the excellent negative predictive value, these results suggest that Clinitek Microalbumin is a good screening test for microalbuminuria. Positive results should be confirmed using a reference assay.