The value of screening for diabetes in patients with skin sepsis

Four hundred and eighty-two patients with spontaneous skin and superficial sepsis and 291 controls of similar age and sex underwent random capillary blood glucose measurements in order to assess whether screening for diabetes in patients presenting with skin sepsis to an Accident & Emergency Department detects a greater number of cases than that present in the background population. All subjects with a concentration > 7.8 mmol/l were subsequently followed up with a 75 g oral glucose tolerance test. Forty-two (8.7%) of the 482 skin sepsis patients had a capillary blood glucose > 7.8 mmol/l compared to eight (2.7%) of the 291 without sepsis (chi 2 = 9.71, P < 0.002). Of these, 26 of the skin sepsis group and 7 of the control group attended for follow up. Of those who attended, 13 of the skin sepsis group had an abnormal glucose tolerance test (seven diabetes, six impaired glucose tolerance-IGT) compared to two (one diabetes, one IGT) of the control group (chi 2 = 2.87, P < 0.1). The difference in cases of frank diabetes between the two groups was not statistically significant. Of the total eight diabetic cases identified, five (on direct questioning) had symptoms of hyperglycaemia (thirst, polyuria and/or weight loss) and two of the others were obese, one of whom had documented ischaemic heart disease. Thus, while most cases of diabetes in patients with skin sepsis could be detected by specifically asking about hyperglycaemic symptoms and performing a blood glucose estimation when these are present, we suggest that the screening of patients with skin sepsis over 40 years of age provides an opportunistic method of screening. This strategy should yield clinically significant numbers of abnormal cases.