Journal of Medical Screening

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 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.

To investigate prospectively the value of routine urine analysis in a university based, medical outpatient clinic providing primary (> 90%) and referral care (< 10%) in general internal medicine, as determined by whether it led to a new diagnosis requiring a change in medical management (new treatment; advice to patient; further follow up appointment).

A dipstick urine analysis was performed in 610 consecutive patients (mean (SD) age 41 (16) years) making their first clinic visit. A urinary sediment of the same specimen was examined in those patients with abnormal results for haemoglobin, leucocytes, or protein. A urine analysis was defined as routine when it was not considered indicated for diagnostic or management purposes by the resident in charge at the patients' initial clinic visit. The main outcome measures were the number and nature of new diagnoses leading to a change in patient management.

In 183 patients (30%, 95% confidence interval (CI) 27% to 34%) the urine analysis was performed as an indicated test and in 427 (70%; 95% CI 64% to 73%) as a routine test. Urine analysis was abnormal in 71 of the 427 patients (17%) in whom it had been performed as a routine test. Abnormal findings led to a change in management in three patients only (0·7%; 95% CI 0·2% to 2·2%). Critical assessment makes the value of routine urine analysis debatable even in these three patients.

In a middle aged medical outpatient group of a university based, teaching clinic providing predominantly primary care, urine analysis can be restricted to patients in whom it is clinically indicated.

To optimise radiological screening in von Hippel-Lindau disease (VHL) while minimising cost and morbidity.

A model of VHL was based on retrospective studies, and Bayes's theorem used to calculate the probability of the gene's presence and the likelihood of further lesions in affected families. A six year follow up was conducted to test the validity of the model.

Follow up confirmed the accuracy and validity of the model. Posterior fossa haemangioblastomas occur in 79·2% of VHL cases, supratentorial, retinal and spinal haemangioblastomas in 6·9%, 42·8%, and 22·0%, phaeochromocytomas in 5·2%, and renal carcinomas in 14·5%. Population incidences are 1:15 700 live births (posterior fossa), 1:780 000 (supratentorial), and 1:116 000 (spinal). The birth rate of subjects with VHL is 1:43 000; new mutations occur in 1:178 000 live births. Penetrance is 90%; 40% present with multiple lesions and 6·4% die within two years after diagnosis.

For most patients presenting with a VHL-type lesion, with sufficient clinical and pedigree data, the presence or absence of the VHL gene, and the probability of further lesions occurring, can be assessed with a high degree of accuracy using the method described in this paper. Those cases in the non-VHL group do not require long term radiological follow up, nor do their relatives require radiological screening. Subjects in the VHL group should be screened for renal carcinoma indefinitely from the age of 20 years, and all clinically unaffected relatives should be screened genetically for the VHL gene. (Those found negative for the gene do not require further screening, but those found positive should be screened.)