The predictability of periodontal treatment as measured by tooth loss: a retrospective study

OBJECTIVE: To analyse the validity of the Micraltest and Microbumintest semi-quantitative methods for microalbuminuria screening in type 2 diabetes mellitus in primary care clinics. DESIGN: Crossover study to validate diagnostic tests. SETTING: Three general practices at an urban health centre. PATIENTS: Consecutive sample of 64 diabetics not dependent on insulin. MEASUREMENTS AND MAIN RESULTS: Each patient's own doctor performed in the consultation the semi-quantitative determinations in a simple sample of the morning's first urine. As standard, a rate of albumin excretion above 20 micrograms/min, determined by immunonephelometry in a 2-hour controlled time sample, was used. The prevalence of Microalbuminuria was 25% (C.I. 95%, 14.4-35.6). The sensitivity of Micraltest was 69% (CI, 42-88), and its specificity 52% (CI, 37-67), corresponding to the cut-off point of 10 mg/L. The sensitivity of Microbumintest was 63% (CI, 36-84), and its specificity 67% (CI, 52-79). The combination of both tests in one sample had 75% sensitivity (CI, 47-92) and 44% specificity (CI, 30-59). Negative predictive values ranged between 84 and 85%. CONCLUSIONS: The semi-quantitative microalbuminuria detection tests are of limited use in the primary care clinic. Their sensitivity and the negative predictive value obtained in an isolated sample do not seem acceptable for a screening method.