Multiple pre-transplant scores aimed at stratifying the risk of death or allograft loss after heart transplantation have been developed over time but their use in clinical practice remains limited.
We aimed to evaluate the statistical performance of all existing pre-transplant risk scores in a contemporary United Network for Organ Sharing (UNOS) cohort.
We applied 16 different scoring systems developed between 1999 and 2019 to a contemporary UNOS cohort of heart transplant recipients. We selected first non-combined adult heart transplantations performed between January 1st, 2014 to December 31st, 2017. The primary endpoint was all-cause mortality or retransplantation during the first-year post-transplant. We analyzed the association between scores and outcomes, sensitivity, specificity, likelihood ratios (LR) and discrimination (c-statistics and overlap of individual scores according to the occurrence of the primary endpoint).
The cohort included 9396 patients. Recipients were mostly males (74%) and mean age was 52.9±13.5. All scores were significantly associated with the primary outcome (P<0.001 for all scores). The LR of the scores were poor (positive LR<2, negative LR>0.5). The discriminative performance of all scores was limited, with c-statistics ranging from 0.548 to 0.652 (median 0.600, Fig. 1A) and an important overlap of individual scores between patients with or without the primary endpoint (Fig. 1B). Subgroup analyses revealed that discrimination may vary greatly according to donor age, recipient age and the type of assist device used at transplant.
In a contemporary cohort, none of the 16 post-transplant risk scores were shown to accurately stratify the risk of early death or retransplantation after heart transplantation. our findings raise concerns concerning the use of the currently available prognosis scores in the clinical files and emphasize the need for a robust and reliable score.Le texte complet de cet article est disponible en PDF.