Purpose: Decision-making on transcatheter aortic valve replacement (TAVR) in patients aged 75 years and older is complex. It could be facilitated by the identification of predictors of long-term mortality. This study aimed to identify predictors of 2-year mortality to develop a 2-year mortality risk score.
Methods: Cohort study of consecutive patients aged ≥ 75 years who underwent TAVR after a comprehensive geriatric assessment (CGA) at our university hospital between 2012 and 2019. Predictors of 2-year mortality were determined using multivariable Cox regression. A point-based predictive model was developed based on risk factors and subsequently internally validated by fivefold cross-validation.
Results: The 345 patients (median age 87 years, 54% women) were fit/vulnerable (32%), mildly frail (37%), or moderately/severely frail (31%). The overall 2-year mortality rate was 26%, predicted by atrial fibrillation, hemoglobin ≤ 10 g/dL, age ≥ 87 years, BMI ≤ 24, eGFR ≤ 50 ml/min, and moderate/severe frailty. The risk score (range 0-12), named OLD-TAVR score, for 2-year mortality showed good discriminative power (AUC 0.70) and remained consistent after fivefold cross-validation (cvAUC 0.69). A risk score ≥ 8 (prevalence 20%) predicted a 45% (95%CI: 34-58%) two-year mortality, with high specificity (86%) and good positive predictive power (+ LR 2.43).
Conclusion: A 2-year mortality risk score (OLD-TAVR score) for very old patients undergoing TAVR was developed based on six bio-clinical items. A score ≥ 8 identified patients in whom 2-year mortality was very high and thereby the TAVR futile.
Trial registration number and date of registration: Study protocol B403, 26/09/2022, retrospectively registered.
Keywords: Aortic stenosis; CFS; Frailty; Older adults; Risk score; TAVR.
© 2023. The Author(s), under exclusive licence to European Geriatric Medicine Society.