Background: Islet transplantation has been associated with better metabolic control and quality of life than insulin treatment alone, but direct evidence of its effect on hard clinical endpoints is scarce. We aimed to assess the effect of islet transplantation on patient-graft survival in kidney transplant recipients with type 1 diabetes.

Methods: In this retrospective cohort study, we enrolled all patients with type 1 diabetes who received a kidney graft in France during the study period, identified from the CRISTAL nationwide registry. Non-inclusion criteria included recipients from transplant centres that never proposed islet transplantation during the study period, recipients with a functional pancreas throughout the follow-up duration, recipients with more than two kidney transplants, HLA-sensitised recipients, recipients with less than 1 year of follow-up after kidney transplantation, misclassified recipients with type 2 diabetes, recipients aged over 65 years, recipients of kidney grafts from Donation after Circulatory Death donors, recipient with HIV or hepatitis, recipients with cancer, and recipients of combined liver-kidney transplants. Patients who also received islet-after-kidney (IAK) transplantation were compared with controls who received kidney transplantation alone according to a 1:2 matching method based on time-dependent propensity scores, ensuring patients comparability at the time of islet transplantation. The primary outcome was patient-graft survival, a composite outcome defined by death, re-transplantation, or return to dialysis.

Findings: Between Jan 1, 2000, and Dec 31, 2017, 2391 patients with type 1 diabetes were identified as having received a kidney transplant, 47 patients of whom also received an islet transplantation. 2002 patients were not eligible for islet transplantation and 62 were excluded due to missing data. 327 eligible recipients from 15 centres were included in the study dataset for the target trial emulation. 40 patients who received IAK transplantation were successfully matched to 80 patients who received kidney transplantation alone. 13 (33%) of 40 patients in the IAK transplantation group returned to dialysis or died, compared with 36 (45%) of 80 patients in the kidney transplantation alone group. We found a significant benefit of islet transplantation compared with kidney transplantation alone on patient-graft survival, with a hazard ratio (HR) of 0·44 (95% CI 0·23-0·88; p=0·022), mainly explained by a protective effect on the risk of death (HR 0·41, 0·13-0·91; p=0·042). There was no meaningful association between IAK and death-censored graft survival (0·73, 0·30-1·89; p=0·36).

Interpretation: In kidney transplant recipients with type 1 diabetes, IAK transplantation was associated with a significantly better patient-graft survival compared with kidney transplantation alone, mainly due to a protective effect on the risk of death. These results potentially serve as compelling grounds for advocating wider access to islet transplantation in patients with type 1 diabetes undergoing kidney transplant, as reimbursement of islet transplantation is provided in few countries worldwide.