Reevaluation of urine C-peptide as measure of insulin secretion.

Urine C-peptide (UCP) has been proposed as a measure of insulin secretion, because insulin and C-peptide are cosecreted in equimolar concentrations by the pancreatic β-cell. The validity of this approach was tested by comparing insulin secretion rates, calculated by application of a two-compartmental analysis of peripheral C-peptide concentrations, with UCP excretion rates. Insulin secretion and UCP excretion with subjects on a mixed diet were simultaneously measured over a 24-h period in 13 patients with non-insulin-dependent diabetes mellitus and in 14 matched nondiabetic control subjects. The fraction of secreted C-peptide that was excreted in the urine (fractional C-peptide excretion) showed considerable intersubject variability in the diabetic (11.3 ± 1.6%, range 3.9–20.8) and control (8.0 ± 1.7%, range 1.1–27.9, P = .07) subjects (means ± SE). UCP clearance demonstrated a similar degree of variability and was not significantly different ( P = .07) between diabetic (23.8 ± 3.0 ml/min) and control (16.5 ± 2.7 ml/min) subjects. In control subjects, the 24-h insulin secretion rate correlated more closely with the fasting insulin secretion rate ( r = .97, P = .0001), fasting C-peptide ( r = .81, P = .0005), and fasting insulin ( r = .80, P = .0005) concentrations than with the 24-h UCP excretion rate ( r = .62, P = .02). Similar results were obtained in the diabetic patients. The mean coefficient of variation of fractional UCP excretion in 7 nondiabetic control subjects who were studied on a mixed diet over a 24-h period on two occasions was 28.4 ± 10.5%, that of UCP clearance was 28.9 ± 8.6%, and that of simultaneously measured creatinine clearance was 7.8 ± 3.5%. In summary, the fraction of secreted C-peptide that appears in the urine varies considerably between subjects and in the same subject studied repeatedly. UCP excretion does not correlate as well with 24-h insulin secretion as does the fasting insulin secretion rate or the fasting C-peptide or fasting insulin concentration. We conclude that, because the fraction of secreted C-peptide that is excreted in the urine varies considerably between subjects and in the same subject studied on different occasions, UCP is of only limited value as a quantitative measure of endogenous insulin secretion.