Urinary C-peptide creatinine ratio to detect absolute insulin deficiency in type 2 diabetes

BackgroundNational Institute for Health and Clinical Excellence guidelines (CG87) recommend neutral protamine hagedorn (NPH) insulin for the provision of basal insulin in type 2 diabetes, but use of analogue insulin is as much as 40%. Where residual endogenous insulin secretory capacity is present there is no evidence that analogue insulins provide any additional benefit over human insulins, and they come at an expensive premium. Anecdotally, however, there is a reluctance to switch people back to NPH insulin, partly because of a perceived risk of pancreatic failure and potential ketosis. Urinary C-peptide creatinine ratio (UCPCR) has been validated as a method for evaluating residual endogenous insulin secretion in type 1 and type 2 diabetes, with a UCPCR of no more than 0·2 nmol/mmol suggestive of absolute insulin deficiency. We aimed to evaluate the prevalence of true insulin deficiency among patients with type 2 diabetes with UCPCR, and confirm findings with the gold standard mixed meal tolerance test (MMTT).Methods191 insulin-treated patients with a clinical diagnosis of type 2 diabetes (diagnosed at or after age 45 years and who did not start insulin within the first year of diagnosis) collected a 2-h post-prandial urine sample for UCPCR measurement. Nine patients from two subgroups (UCPCR ≤0·2 nmol/mmol and UCPCR >0·2) completed a standard MMTT.Findings11 (5·8%) of 191 patients had two consistent UCPCRs of less than or equal to 0·2 nmol/mmol. Nine were able to do the MMTT, of whom five were confirmed to have absolute insulin deficiency (stimulated serum c-peptide <0·2 nmol/L). Three of these five patients were glutamic acid decarboxylase antibody-negative. Nine of nine patients with UCPCR of more than 0·2 nmol/L had confirmed endogenous insulin secretion in their MMTT. Those with insulin deficiency had a shorter time to starting insulin (median 2·5 years [IQR 1·5–3·0] vs 6·0 [3·0–10·75], p=0·005) and lower body-mass index (25 kg/m² vs 29, p=0·04) but no other significant differences in clinical characteristics.InterpretationWe have demonstrated a very low prevalence of true pancreatic failure in this population of insulin-treated patients with type 2 diabetes. This requires further exploration by comparison of a population being treated with NPH insulin with one on analogue insulin, and then determining whether UCPCR could act as a clinical decision support tool to safely switch from analogue insulin to NPH insulin.FundingNational Institute for Health Research.