Time-dependent C-peptide decline in 4411 patients with recent-onset type 1 diabetes followed for up to 10 years: a meta-analysis

Background and aims: C-peptide is a valuable indicator of residual beta cell function in patients with type 1 diabetes (T1D). Critical factors associated with decline of C-peptide after diagnosis include age at onset, autoantibody and HLA status, severity of initial metabolic derangement, insulin use and quality of glucose control. Whilst recent studies suggest that significant beta cell function may be preserved for several years after diagnosis, C-peptide data must be corrected for these variables and robust data is lacking. To this aim, we collected retrospectively data on C-peptide from T1D patients from 8 different European centres followed-up from diagnosis and up to 10 years. Materials and methods: We performed a meta-analysis of 4411 T1D patients in total (57.2% males; mean age at onset: 18.5 yrs. ± 11.7 SD, age range 1-60). Records were extracted for age at diagnosis, fasting and stimulated C-peptide, Body Mass Index, HbA1c and insulin requirement at diagnosis, and from 1 to 10 years follow-up. Results: Data on C-peptide (nmol/l) at diagnosis were available from 3648 patients who were subdivided according to age at diagnosis in three groups: Group A n=1356 (0-11 yrs.); Group B n=647 (12-18 yrs.); Group C n=1645 (>18 yrs). There was a significant trend (P<0.001) age-dependent increase in fasting C-peptide at diagnosis (0.20±0.20 in Group A, 0.28±0.26 in Group B, and 0.30±0.35 in Group C), not observed for stimulated C-peptide (0.46±0.43 in Group A, 0.47±0.32 in Group B and 0.44±0.38 in Group C). In Group A the percentage decline of baseline C-peptide at 1, 2,5 and 10 years was 25, 70, 75, 90 percent, respectively; in Group B, 7,60,67 and 85 percent, respectively; in Group C, 23,43,75 and 77 percent, respectively (p for trend <0.001). Conclusions: This study (the largest of this kind) reveals an inverse correlation between age at T1D onset and basal C-peptide at diagnosis with a more rapid decline in beta-cell function in the very young patients. The data will be useful for the calculation of treatment effects and sample sizes in trials aiming to preserve insulin secretion with C-peptide as end- point