Background: Type 1 diabetes (T1D) results from T-cell mediated destruction of beta cells. Preclinical studies and pilot clinical trials suggest that antithymocyte globulin (ATG) may blunt this response. Our goal was to evaluate ATG in subjects with recent-onset T1D.
Methods: START is an NIH-sponsored phase 2, randomized, placebo-controlled, blinded multicenter trial. Eligible patients had recent-onset T1D, aged 12-35 years, and with peak C-peptide ≥ 0.4 nM on mixed meal tolerance test (MMTT). 58 participants were randomly assigned 2:1 to receive 6.5 mg/kg ATG or placebo. The primary endpoint was change in C-peptide response to MMTT from baseline to 12 months, with follow-up to 24 months.
Findings: The study did not meet its primary endpoint. The ATG group had a mean change in C-peptide of -0.195 pmol/mL vs. -0.239 pmol/mL in the placebo group (p=0.591). The rate of C-peptide decline in the ATG group was greater during months 0-6 vs. months 6-12 (p=0.002); the month 0-6 C-peptide decline in the ATG group was limited to subjects < 22 years of age. All except one ATG-treated subject experienced both cytokine release syndrome and serum sickness, which was associated with a transient rise in IL-6 and acute-phase proteins. ATG-treated subjects had substantial depletion of naïve, central memory and regulatory T cells, whereas effector memory T cells were not depleted. Reconstitution rates of naïve and memory T cells were more rapid in ATG-treated subjects ≤ 21 years vs. subjects > 21 years (p<0.05). Month 24 data is now being analyzed, and will be presented.
Interpretation: A brief course of ATG does not preserve beta cell function 12 months later in new-onset T1D. Several issues may have limited efficacy, including cytokine release, Treg depletion, and effector memory persistence. Month 24 data will determine durability of response for those whose beta cell function appeared to stabilize.