Abstract
Cortisol to creatinine ratios in overnight urine samples, urinary glucose excretion, and plasma glucose concentrations were determined in 43 diabetic inpatients. All initially had normal cortisol to creatinine ratios (less than 55 x 10(-6)) and were initially treated by increasing their long-acting insulin component. Nine patients in whom this ratio became raised then had their long-acting insulin component reduced until their fasting plasma glucose concentration was 4-7 mol/l (72-126 mg/100 ml). The 34 patients who had never had a raised ratio were treated by increasing their long-acting insulin component until their fasting plasma glucose concentration was in the range 4-7 mmol/l. All the raised cortisol to creatinine ratios were clearly separate from the other values. A mean reduction in total insulin dose of 23% and in long-acting insulin dose of 53% was achieved, abolishing presumptive nocturnal hypoglycaemia by reducing the ratio to normal and dramatically improving diabetic control. Although there was no definite evidence that the patients who had raised cortisol to creatinine ratios had suffered from nocturnal hypoglycemia, these results strongly support the view that a raised ratio indicates an otherwise unrecognised episode of this condition.