Heart failure prevalence, medication prescription rates, and adverse outcomes in patients undergoing TAVI: the real-world data analysis

Abstract Background Despite advancements in transcatheter aortic valve interventions (TAVI), cardiovascular morbidity and mortality remain high after TAVI. There is a lack of detailed data on the prevalence of heart failure (HF) in patients undergoing TAVI, medication prescription rates, and predictors of HF worsening following the procedure. Purpose We aimed to phenotype TAVI patient population, evaluate background medical therapy, assess risks and key predictors of adverse outcomes after TAVI using real world data. Methods This analysis used Optum’s de-identified Clinformatics Data Mart (Optum CDM or Optum Clinformatics) derived from a database of administrative health claims for members of large commercial and Medicare Advantage health plans. Inclusion criteria were: TAVI record in inpatient settings since 2017; continuous information on diagnosis, procedures and medication prescription at least 1 year before and 1 year after TAVI. Machine learning was used to derive a risk score for HF hospitalisation (hHF) or mortality after TAVI based on demographics, recent recording of ICD-10 codes and medication prescriptions. Results From a total of 54226 patients with records of TAVI, 37490 were included in the final analysis (median age 80 (75-85) years; 44.7% female). HF diagnosis was recorded in 21283 (56.9%) of patients during a year before TAVI and in 32465 (86.7%) at TAVI hospitalisation. 6974 (18.6%) had acute HF episode(s) during the year before TAVI. 18841 (50.3%) had a prior record of type 2 diabetes, 14207 (37.9%) - atrial fibrillation, 32795 (88.0%) - coronary artery disease, 21238 (56.6%) - chronic kidney disease. Medication prescription rates 1 year before TAVI vs 1 year after TAVI were as follows: ACEi or ARB 45.9% vs 46.4%; beta-blockers 49% vs 54%, loop diuretics 34% vs 40.3%, MRA 6.1% vs 8.4%, SGLT2i 2.1% vs 3.4%. The risk of hHF in the whole TAVI cohort 1 and 2 years after TAVI was 8.8% (95%CI: 8.5-9.2) and 13.8% (13.4-14.2), respectively; risk of hHF or all-cause mortality – 15.6% (15.2-16.0) and 25.4% (24.8-25.9), respectively. Using a machine learning algorithm, population was divided into fifths based on risk score (Figure 1, Table 1). The most predictive factors for hHF after TAVI were prior HF diagnosis (HR 95% CI: 1.21 1.11-1.33), number of prior acute HF exacerbations (HR: 1.25 1.19-1.30), atrial fibrillation and chronic kidney disease. Conclusions Prevalence of HF pre- and post-TAVI is high; however, standard of care HF medication use is low in this patient population. Despite successful TAVI, there is a significant residual risk of hHF and/or mortality after procedure indicating an unmet medical need and an opportunity for drug development. Previous diagnosis of HF and HF worsening events before TAVI, atrial fibrillation and chronic kidney disease were identified as significant and independent predictors of adverse outcomes.Risk score for hHF after TAVI. Characteristics of risk groups for hHF