PRIORiTize-TAVI score: a novel clinical tool Predicting moRtalIty Or uRgent TAVI on waiting list

Abstract Background Waiting list (WL) for transcatheter aortic valve implantation (TAVI) has been increasing and prioritization strategies are lacking. We sought to derive a simple clinical score to predict increased risk of adverse outcomes in patients waiting for TAVI. Methods Single-center retrospective study of all consecutive ambulatory patients accepted for TAVI (Jan/2017 - Jun/2022). Patients were admitted to active WL after Heart Team meeting and waiting time was defined as the interval between the date of the meeting and the date of TAVI or the primary outcome. The primary outcome was a composite of all-cause mortality while on WL or urgent CV admission leading to TAVI. Demographic, clinical, echo and CT-angio variables were collected, including the Charlson Comorbidity Index which accounts for age, renal disease, and comorbidities. A raw risk score weighted on β-coefficients was developed after identifying independent predictors of the primary outcome at multivariate analysis (Cox Regression). The raw score was simplified to a point system weighted on the positive and negative predictive values of each variable. Discrimination ability was assessed by the area under the ROC curve (AUC). Internal validation was performed with bootstrapping (1000 samples). Kaplan-Meier (KM) survival analysis according to risk categories was performed. Results We identified 427 patients (83 ± 6 years; 56% female; ES II 4.4% [IQR 3.1 – 6.2%]). Median WL time was 44 days (IQR 26 – 76 days). While on active WL, 54 patients (12.6%) attained the primary endpoint (34 deaths and 20 urgently admitted for TAVI). Five independent predictors of the primary endpoint were identified: Charlson Comorbidity Index; NYHA class; NT-proBNP; LVEF and aortic mean gradient (see figure for adjusted hazard ratios). The simplified point system and its distribution across the cohort are depicted in the figure. Patients were stratified into 3 risk strata: low (<2 points; n = 83 [20%]; 0 events [0%]); intermediate (3–4 points; n=217 [51%]; 18 events [8.3%]) and high-risk patients (5–8 points; n = 127 [29%]; 36 events [28.3%]). There were no significant differences between the discriminative power of the raw score and the simplified model (AUC 0.77 [95%CI 0.71 – 0.83] vs. AUC 0.77 [95%CI 0.71 – 0.83]; p = 0.96 for comparison). KM survival curves showed a progressive survival disadvantage along strata: 0 events per 100 persons-month in the low-risk; 2 events (95%CI 1 – 3) per 100 persons-month in the intermediate-risk; and 5 events (95% CI 3 – 7) per 100 persons-month in the high-risk group. Similar results were obtained by restricting the analysis to all-cause mortality. Conclusions A score to predict the risk of adverse events in patients waiting for TAVI was developed from five easy to ascertain variables. Risk strata provided by PRIORiTize-TAVI model may inform medical decision-making for priority assignment of patients in waiting list.