Association of repeated HLA mismatches with graft survival in kidney transplantation: data from the Dutch transplant registry

---

J.W. van der Heijden, T. Hoekstra, C. Ranzijn, C. Konijn, N.M. Lardy, F. van Ittersum, A. Nurmohamed

Chair(s): Prof. dr. L.B. Hilbrands & dr. H.M. Kwakkel-van Erp

Wednesday 8 march 2017

17:30 - 17:42h at Willem Pijperzaal

Categories: Parallel - Klinisch

Parallel session: Parallelsessie VI - Klinisch 1 - HLA-antilichamen en immunomodulatie

---

Background:
Kidney re-transplantation is a risk factor for decreased allograft survival. Among other factors, repeated mismatched HLA antigens potentially trigger an alloimmune memory response against the graft resulting in antibody mediated rejection or chronic allograft nephropathy. After the introduction of calcineurin inhibitors as standard immunosuppressive therapy, the question rises whether transplantation with a repeated HLA mismatch (RMM) is still a risk for decreased kidney allograft survival. Literature on this subject is inconclusive. Furthermore, the policy in transplant centers in the Netherlands for accepting a RMM differ substantially. 

Objective:
To evaluate the risk of a RMM (on the A, B or DR locus) on kidney allograft survival in the Dutch transplant registry (NOTR).

Methods:
Between 1994 and 2014 records of 1698 re-transplants were found in the Dutch transplant registry (deceased donors and living donors). Of 160 transplantations no clinical data were available and of 517 transplantations it could not be determined whether there was a RMM, due to missing HLA data of a (previous) transplantation leaving 1,021 transplantations available for the current analyses. Patients were followed up for graft failure and mortality. The primary end point was graft failure with a potential immunological cause. Cox regression analysis was performed to calculate hazard ratios (HR) in RMM transplantations. Adjustments were performed for donor and recipient characteristics, year of transplant and PRA. Multiple imputation with 5 repetitions was performed to account for missing covariates. 

Results:
919 transplantations (90%) were performed with a kidney without a RMM and 102 transplantations (10%) had a RMM on the A,B or DR locus (or both). Baseline characteristics were comparable between the groups, except for a higher percentage of living donation in the RMM group (57% versus 31%) and a subsequent shorter cold-ischemic time. 299 death-censored graft failures were registered of which 192 (64%) were classified as graft failures with a potential immunological cause. Of these graft failures, 19 occurred in the RMM group. After a median follow-up of 5.9 years, a significant decreased death-censored graft survival was observed for a DR RMM. Further analysis revealed an adjusted hazard ratio for immunological graft failure of 2.12 (1.09-4.09) for a HLA-DR RMM and 1.01 (0.49-2.06) for a HLA-A or HLA-B RMM. There was correlation of a RMM with patient survival. 

Conclusion:
A HLA-DR, but not HLA-A or -B RMM confers a substantial increased risk for graft failure. The risk of a HLA-DR RMM has to be weighed against the risk of staying on dialysis.