Type 1 diabetes with immune checkpoint inhibitors: A systematic analysis of clinical trials and a pharmacovigilance study of postmarketing data.

Background: To estimate the risk of type 1 diabetes associated with immune checkpoint inhibitor (ICI-T1DM), and to describe its clinical features.

Methods: ICI-T1DM events in randomized clinical trials (RCTs) available in electronic databases were systematically reviewed. The primary outcome was the summary risk of T1DM related to ICIs, a meta-analysis was conducted to obtain Peto odds ratios (ORs) with 95 % CIs. In pharmacovigilance study, ICI-T1DM cases were extracted from FAERs. Disproportionality analyses were performed by calculating reporting odds ratio (ROR) and information components (IC).

Results: A total of 29 RCTs (20,234 patients) were included, treatment with ICIs significantly increased the risk of all-grade ICI-T1DM (OR: 4.54, 95 % CI: 2.66-7.72), and high-grade (grade 3 or above) ICI-T1DM (OR: 4.26, 95 % CI: 2.12-8.58). No significant differences among subgroup analyses were observed: ICIs treatment schedule, tumor type, case of events (T1DM vs F-T1DM), study design (double blind vs open label) or median PFS (PFS favours ICIs vs PFS favours Control). A total of 978 case reports form FAERS was extracted, treatment with ICIs significantly increased the reporting of ICI-T1DM (n = 978; ROR = 38.45, 95 %CI:35.70-41.41; IC = 4.77, 95 %CI:4.43-5.14). In cases with available data, the median latency period was 10.4 weeks, drug interruption was recorded in 82.3 % of cases, with a positive dechallenge in 76 % of cases, and death was recorded as outcome in 3.6 % of reports.

Conclusions: Both data from clinical trials and postmarketing suggested that ICIs was associated with increased risk of ICI-T1DM. As ICIs gain greater clinical use, practitioners must be aware of ICI-T1DM events.