Six assays for C-peptide were evaluated. Five assays were radioimmunoassays; one was a chemiluminescence assay. The within-run coefficient of variation ranged between 2.1 and 22.9%. The methods correlated well with each other (r = 0.925 to 0.978). There were, however, proportional biases between the assays, in particular in the higher concentration range. All assays were linear up to at least 4.5 ng/ml C-peptide concentration. All assays were easy to perform. However, the systematic bias between the assays demonstrates that further effort is needed to standardize the measurement of C-peptide. In the mean-time, we recommend that laboratories conducting C-peptide measurements establish local, method-specific reference ranges.