Screening for non-insulin-dependent diabetes mellitus (NIDDM): how often should it be performed?

To examine whether a screening programme for diabetes repeated after an interval of 30 months is worthwhile both in terms of yield of new cases and continued high response rate.

Self testing for postprandial glycosuria was used as it has been shown to have a good response rate, a good yield of cases of diabetes, and a sensitivity and specificity which compares favourably with more expensive and invasive screening methods. A total of 3231 subjects aged 45-70 years in one practice were screened on two occasions 30 months apart. Eighty seven subjects known to have diabetes were excluded. This number included five new patients who had presented since the initial screen; two of whom had moved into the practice, one who screened positive at the initial screen but failed to attend for the oral glucose tolerance test (OGTT), and two subjects who had screened negative, the first presenting 24 months after screening.

At repeat screening the return rate was lower than at the first screening (72.5% v 79.2%, P < 0.0001). Glycosuria was detected in 52 subjects (2.3%), at the repeat screening, similar to that at the initial screen. Attendance for the OGTT after a positive screening test was 93.2%. Only six subjects of the 24 with glycosuria but normal glucose tolerance at initial screen were found to have glycosuria again; repeat OGTTs were all normal. Of the remaining 46 subjects, 10 had non-insulin-dependent diabetes mellitus (NIDDM) and five impaired glucose tolerance (IGT). The number of subjects with diabetes was not significantly different from that at the initial screening (0.44% v 0.72%, P = 0.2).

Repeat screening after 30 months has a high response rate, similar rate of detection of glycosuria, and a further yield of 0.44% of newly diagnosed cases of diabetes. A screening programme detecting postprandial glycosuria identifies additional diabetic subjects 30 months after a previous screening programme.