Urinary C-peptide analysis in an intervention study: experience from the DEFEND-2 otelixizumab trial

To demonstrate that analysis of urinary C-peptide across multiple study sites in the context of an intervention trial (DEFEND-2) is a viable alternative to mixed meal testing and delivers results that correlate to mixed meal testing estimation of endogenous insulin production.

Second morning void urine was collected for analysis and was available from 161 subjects at baseline (55 placebo, 106 otelixizumab), and 146 subjects (47 placebo, 99 otelixizumab) at month 12. Urinary C-peptide concentration was corrected for urinary creatinine [urinary C-peptide/creatinine ratio (UCPCR)] and serum C-peptide from the mixed meal tolerance test was calculated using area under the plasma concentration-time curve (AUC) normalized over 120 min. The correlation between mixed meal stimulated C-peptide AUC (mmol/l/min) and UCPCR (nmol/mmol), as well as the correlation between insulin use (IU/kg), and HbA_1c (%) with UCPCR, was determined.

UCPCR and mixed meal testing C-peptide AUC were correlated, with a correlation coefficient of 0.4172. UCPCR was not correlated with exogenous insulin use (r = -0.089) or with HbA_1c (r = -0.032).

Urinary C-peptide estimation should be considered as a measure of endogenous insulin production in future Type 1 diabetes mellitus outcome trials. A change in the timing for urine collection (to 120 min post standard meal) may provide a tighter correlation to C-peptide measured via a traditional mixed meal test.