Clinical outcomes after ABO-incompatible renal transplantation: a systematic review and meta-analysis

Abstract

ABO-incompatible renal transplantation (ABOi-rTx) is increasingly used to overcome organ shortage. Evidence about its non-inferiority in comparison with ABO-compatible renal transplantation (ABOc-rTx) needs to be analysed at early and late timepoints. We aimed to investigate differences in outcome after ABOi-rTX and ABOc-rTX. We did a systematic review and meta-analysis of observational studies published up until Dec 31, 2017, that reported outcome data (≥1 year of follow-up) after ABOi-rTx and included an ABO-compatible control group, by searching the Cochrane Central Register of Controlled Trials (CENTRAL), Embase Ovid, MEDLINE Ovid, and PubMed. Trials on recipients of ABOi-rTx were assessed, if an ABO-compatible control group was included and if outcome data on at least graft or recipient survival with 1 year or more of follow-up were available. Exclusion criteria included case reports, editorials, reviews and letters, animal studies, meeting papers, studies unable to extract data, non-renal solid organ and bone-marrow transplant studies, and deceased donor ABOc-rTx. Data were extracted from published reports. Primary endpoints were all-cause mortality and graft survival at 1, 3, 5, and more than 8 years after transplantation. In the meta-analysis, we used a fixed-effects model if the I2 value was 0, and both a fixed-effects and random-effects model if I2 was more than 0. This study is registered with PROSPERO, number CRD42018094550. 1264 studies were screened and 40 studies including 49 patient groups were identified. 65 063 patients were eligible for analysis, 7098 of whom had undergone ABOi-rTx. Compared with ABOc-rTx, ABOi-rTx was associated with significantly higher 1-year mortality (odds ratio [OR] 2·17 [95% CI 1·63-2·90], p<0·0001; I2=37%), 3 years (OR 1·89 [1·46-2·45], p<0·0001; I2=29%), and 5 years (OR 1·47 [1·08-2·00], p=0·010; I2=68%) following transplantation. Death-censored graft survival was lower with ABOi-rTx than with ABOc-rTx at 1 year (OR 2·52 [1·80-3·54], p<0·0001; I2=61%) and 3 years (OR 1·59 [1·15-2·18], p=0·0040; I2=58%) only. Graft losses were equivalent to that of ABOc-rTx after 5 years and patient survival after 8 years. No publication bias was detected and the results were robust to trial sequential analysis until 5 years after transplantation; thereafter, data became futile or inconclusive. Despite progress in desensitisation protocols and optimisation of ABOi-rTx procedures, excess mortality and loss of kidney grafts was found compared with ABOc-rTx within the first 3 years after transplantation. Only long-term outcomes after 5 years yielded equivalent survival rates and organ function. Awareness of the increased risks of infection, organ rejection, and bleeding could improve care of patients and promote efforts towards paired kidney exchange programmes. None.