Serial Dried Blood Spot C-Peptide Sampling, but Not Urine C-Peptide–to–Creatinine Ratio, Detects Early Preservation of β-Cell Function in New-Onset Type 1 Diabetes: Experience From the USTEKID Trial

OBJECTIVE To determine whether less invasive C-peptide tests, urine C-peptide–to–creatinine ratio (UCPCR) and dried blood spot (DBS), could replace the mixed-meal tolerance test (MMTT) in the context of an interventional clinical trial (USTEKID). RESEARCH DESIGN AND METHODS C-peptide was assessed at screening and weeks 28 and 52 using a 2-h MMTT. UCPCR samples were taken after each MMTT. Fasting and 60-min postmeal DBS samples were collected weekly to week 28 and then monthly to week 52. UCPCR and glucose-adjusted DBS results were compared with MMTT results. Weekly DBS averages were calculated and differences between treatment groups assessed using a mixed linear model, bootstrap one-sample t tests, and autoregressive integrated moving averages. Six months of weekly DBS data were used to predict 12-month C-peptide levels. RESULTS Unlike MMTT C-peptide, UCPCR did not change in the first 6 months, before decreasing by 12 months. The DBS area under the curve from 0 to 60 min declined steadily over 12 months. The DBS C-peptide decline in the intervention group was significantly less than that in the control group by week 20 (P 0.05). This was not apparent with UCPCR and did not become evident until 52 weeks with MMTT. The slope from 6 months of DBS could predict 12-month MMTT C-peptide in the control but not in the intervention group, consistent with the increasing impact of the intervention from 6 to 12 months indicated by the MMTT data. CONCLUSIONS Frequently sampled glucose-adjusted DBS was more sensitive to early change than MMTT and could serve as a home-based marker of β-cell function, whereas UCPCR was not sensitive to change in the early period.