Acute Kidney Injury and Renin-Angiotensin System Inhibition Following Transcatheter Aortic Valve Replacement

Objective. To identify renin-angiotensin system (RAS) inhibition utilization and discontinuation after transcatheter aortic valve replacement (TAVR) and identify predictors of use and discontinuation. Background. RAS inhibition after TAVR has been associated with lower cardiac mortality and heart failure readmissions. Methods. We analyzed 735 consecutive TAVR patients (2014-2019) who survived to hospital discharge at a high-volume TAVR center to determine the utilization and discontinuation of RAS inhibition after TAVR and identify predictors of use and discontinuation. Clinical characteristics, procedural variables, and hospital outcomes were compared between patients receiving vs not receiving discharge RAS inhibitors. Data were compared using t-test and Chi-square test. Multivariable analysis was used to determine independent clinical predictors. Results. Of the 735 patients, 41.9% were discharged with at least 1 RAS inhibitor. In TAVR patients with heart failure with reduced ejection fraction (HFrEF), defined as EF ≤40%, the utilization of RAS inhibitors at discharge was 51.1%. Patients receiving discharge RAS inhibitors had lower incidences of acute kidney injury (AKI) post procedure (8.1% vs 17.8%; P<.01). Discontinuation of RAS inhibition was observed in approximately 1 in 3 patients and was associated with AKI and pacemaker requirement. Three predictors of RAS inhibitor utilization were higher systolic blood pressure, RAS inhibitor use prior to TAVR, and HFrEF. Conversely, new pacemaker and AKI were associated with less utilization of RAS inhibitors; patients developing AKI were 74% less likely to receive RAS inhibitors than those without AKI. Conclusion. Decreased RAS inhibition provides a potential mechanism for worse outcomes in TAVR patients who develop AKI.

J INVASIVE CARDIOL 2021;33(8):E662-E669.

Key words: medical therapy, RAS inhibition, TAVR

Transcatheter aortic valve replacement (TAVR) has positive effects on left ventricular (LV) function, including improvement in LV ejection fraction (EF), regression of LV hypertrophy, and reduction in myocardial fibrosis.^1,2 However, LV improvement is not observed in up to one-third of patients post TAVR and persistent LV dysfunction has been associated with worse outcomes.^2,3 Renin-angiotensin system (RAS) inhibitors have been extensively studied in patients with heart failure (HF) with reduced EF (HFrEF), but whether they are indicated post TAVR is unclear as prior RAS inhibitor trials did not include patients with aortic stenosis or TAVR procedures.⁴ Animal and human studies have shown that RAS inhibition in severe aortic stenosis can improve LV hypertrophy and myocardial fibrosis.^5,6 Recent registry data suggest that patients undergoing TAVR may benefit from routine utilization of RAS inhibitors, with a significant decrease in 3-year mortality after TAVR associated with RAS inhibition.⁷ In another registry analysis of 15,896 TAVR patients, decreased HF readmission rates were associated with RAS inhibition.⁸ Barriers to the routine use of RAS inhibition in TAVR patients warrant further study. In the current study, we investigated the predictors of RAS inhibitor utilization at discharge after TAVR and in particular hypothesized that discharge use of RAS inhibition may be influenced by acute kidney injury (AKI) post TAVR.

We performed a retrospective registry analysis of 735 consecutive patients with severe symptomatic aortic stenosis who were discharged after undergoing TAVR between January 2014 and March 2019 at the University of Vermont Medical Center (UVMMC). Selected patients were defined as being at low, intermediate, high, or extreme risk for surgical aortic valve replacement based on the Society for Thoracic Surgeons (STS) mortality risk score as well as an independent evaluation by the UVMMC TAVR team.

All patients undergoing TAVR received either a balloon-expandable Edwards Sapien valve (Edwards Lifesciences) or a self-expanding Medtronic CoreValve (Medtronic). Prosthesis sizes were determined using preprocedural echocardiographic and multislice computed tomography angiographic (CTA) images. The TAVR devices were delivered via femoral, apical, subclavian, or transaortic approaches.

Prespecified clinical and laboratory data were collected prospectively. Baseline data, including demographics, medical history, and baseline medications, and procedural data, including total intravenous contrast administered and intra- and periprocedural complications, were obtained by review of stored images and chart review. Imaging data, including baseline echocardiographic and CTA findings, were recorded in all patients. Procedural angiograms were analyzed for vascular complications. Follow-up serum creatinine and estimated glomerular filtration rate (eGFR) were obtained in all patients. eGFR was calculated using the Chronic Kidney Disease Epidemiology Collaboration (CKD-EPI) equation.⁹AKI was defined as an increase in creatinine ≥0.30 mg/dL at any time between TAVR and discharge. HFrEF was defined as EF ≤40%.¹⁰RAS inhibition was defined as any dose of an angiotensin-converting enzyme (ACE) inhibitor, angiotensin-receptor blocker (ARB), or aldosterone antagonist prescribed at discharge.

All baseline clinical and procedural variables, clinical endpoints, and complications were reported as defined by the Society of Thoracic Surgeons/American College of Cardiology/Transcatheter Valve Therapy (STS/ACC/TVT) registry, and verified by review of patient-level data submitted to the national registry. Procedural data, including type and size of TAVR bioprosthesis, vascular access route, and amount of contrast, were tabulated based on chart review. Vascular complications were defined as femoral dissection, occlusion, or rupture requiring percutaneous or surgical intervention. Major bleeding was defined as bleeding that met at least one of these criteria: (1) hemoglobin decrease ≥3 g/dL; (2) any transfusion of whole blood or packed red blood cells; or (3) procedural intervention or surgery to reverse/stop or correct bleeding. Patient records were reviewed through hospital discharge and adverse events were documented, including major bleeding complications, new pacemaker requirement, or stroke/transient ischemic attack (TIA).

The two co-primary study endpoints were rates of RAS inhibition utilization at discharge and RAS inhibitor discontinuation prior to discharge. Predictors of these two endpoints were then identified by univariate and multivariable analyses. Variables were summarized using means, standard deviations, or percentages. Overall, group comparisons of baseline demographics and medical histories, baseline laboratory and medications, echocardiographic characteristics, and procedural characteristics were conducted. These measures were compared between patients with RAS inhibition vs those without RAS inhibition at discharge as well as between those with RAS inhibitor continuation vs discontinuation. Categorical variables were reported as percentages and compared using the Chi-square test for independence; continuous variables were reported as mean values and compared using the Wilcoxon rank-sum test. Multivariable models were generated using clinical and procedural variables with a P-value <.10.

Separate models were created for each of the two primary endpoints. A multivariable model was generated from the entire study population (n = 735) for predictors of RAS inhibition at discharge using univariate predictors with P-value <.10. A second model was developed for the subgroup of patients who were receiving RAS inhibition at index hospital admission (n = 388; 52.8% of the entire cohort); predictors of RAS discontinuation at discharge were identified from univariate predictors with P-value <.10. Both models were performed first utilizing only univariate baseline clinical and procedural characteristics that predicted discharge RAS inhibition or discontinuation of RAS inhibition. The models were then redone to identify the additional impact of major adverse outcomes after TAVR on discharge RAS inhibition or discontinuation. Major adverse outcomes after TAVR were defined according to ACC/STS/TVT registry, including new pacemaker implantation, major bleeding complications, AKI, and stroke. Adjusted odds ratios (ORs) were derived using multivariable logistic regression models. For all variables, a P-value <.05 was considered to be statistically significant. All P-values were two-sided. Data were analyzed using the statistical software package Stata, version 14.2 (StataCorp).

Between January 2014 to March 2019, a total of 735 consecutive TAVR patients who survived to hospital discharge were included in the study. Of these patients, 52.8% were prescribed RAS inhibitors at the time of admission for TAVR and 41.9% were prescribed RAS inhibitors at discharge. In general, the TAVR population had multiple elements of increased mortality risk; approximately one-half were >75 years old, 40% had diabetes mellitus, and 15% had a prior stroke/TIA. HFrEF was present in 14% of the cohort and the STS risk score was 6.4 ± 5.1%. Transfemoral access route was utilized in approximately 95% of cases. Edwards balloon-expandable valves were used in 55% of cases and Medtronic CoreValves were used in 45% of cases. The baseline clinical and echocardiographic characteristics, laboratory data, and procedural information of patients are summarized in Table 1.

Patients who were discharged with RAS inhibitors were over 3 times more likely to be prescribed RAS inhibition prior to TAVR (90.6% vs 25.5%; P<.01). The average length of stay from TAVR to discharge date in patients discharged with RAS inhibition vs no RAS inhibition was shorter (2.93 ± 3.02 vs 3.48 ± 2.49; P<.01). In addition, hypertension (93.5% vs 85.7%; P<.01), diabetes mellitus (44.5% vs 36.5%; P=.03), prior pacemaker (14.7 vs 8.9%; P=.02), and HFrEF (17.2% vs 11.5%; P=.03) were more prevalent in patients receiving RAS inhibitors at discharge (Table 1). Patients discharged with RAS inhibitors had a lower incidence of AKI post procedure (8.1% vs 17.8%; P<.01) (Figure 1) and had a higher incidence of HFrEF (12.0% vs 7.0%; P=.02). Those who were not discharged with RAS inhibition had a higher prevalence of moderate to severe lung disease (30.0% vs 16.9%; P<.01), higher STS risk scores (6.9 ± 5.8 vs 5.7 ± 3.7; P<.01), and a higher incidence of new pacemaker placement (11.7% vs 5.8%; P=.01) (Table 2). In patients who developed postprocedure AKI, there was an increased incidence of new pacemaker vs non-AKI patients (14.9% vs 8.2%, respectively; P=.03).

Multivariate analysis identified several independent predictors of RAS inhibition at discharge. The most potent predictor was RAS inhibition prior to the TAVR procedure (OR, 28.48; 95% confidence interval [CI], 18.12-44.75; P<.001). The other baseline predictors included HFrEF (OR, 2.09; 95% CI, 1.17-3.76; P=.01), and higher systolic blood pressure (OR, 1.02; 95% CI, 1.01-1.03; P<.01) (Figure 2). In a secondary model including post TAVR outcomes, similar baseline clinical variables remained potent predictors. In addition, patients who developed AKI after TAVR (OR, 0.26; 95% CI, 0.14-0.47; P<.01) or required a new pacemaker (OR, 0.33; 95% CI, 0.16-0.69; P<.01) were much less likely to be discharged with RAS inhibition (Figure 3).

Discontinuation of RAS inhibition was observed in approximately 1 in 3 post-TAVR patients who were on RAS inhibitors at the time of index hospitalization. In general, patients with RAS inhibitor discontinuation were sicker (Table 3) and more frequently elderly, had lower baseline blood pressures, and had higher STS risk scores and worse baseline renal function. Clinical outcomes were worse among patients with RAS inhibitor discontinuation; the incidence of AKI (28.4% vs 6.8%; P<.01) and new pacemaker placement (13.8% vs 6.5%; P=.02) was increased among patients who had their RAS inhibitors stopped at the time of discharge (Table 4). Multivariable analysis identified 3 independent baseline predictors of RAS discontinuation, ie, age >75 years (OR, 1.71; 95% CI, 1.02-2.86; P=.04), lower baseline GFR (OR, 0.98; 95% CI, 0.97-1.00; P<.01), and lower systolic blood pressure (OR, 0.98; 95% CI, 0.97-0.99; P<.01). A secondary multivariable analysis including both baseline predictors and major adverse outcomes identified the same 3 baseline clinical variables as independent predictors of RAS inhibitor discontinuation. In addition, patients who developed AKI after TAVR (OR, 5.18; 95% CI, 2.65-10.11; P<.001) or required a pacemaker (OR, 2.87; 95% CI, 1.29-6.39; P=.01) were more likely to have RAS inhibition discontinued. Notably, patients who developed AKI after TAVR were 5 times more likely to have RAS inhibitors discontinued prior to discharge.

This single-center registry has 3 main observations: (1) only 42% of TAVR patients were discharged with RAS inhibitors; (2) the main predictors of RAS inhibitor utilization at discharge were RAS inhibitor use prior to TAVR, HFrEF, and higher systolic blood pressure — notably, AKI and new pacemaker requirement predicted lack of RAS inhibitor utilization; and (3) the main predictors of RAS inhibitor discontinuation were generally similar and again included AKI after TAVR as a major predictor of RAS discontinuation.

To our knowledge, this is the first study to analyze the barriers to routine use of RAS inhibitors among patients who undergo TAVR. Previous studies in acute myocardial infarction patients, especially those >75 years old, have demonstrated that the initiation of guideline-directed medical therapy prior to discharge is the best predictor of their use in the outpatient setting¹¹ and has been associated with decreased mortality.¹² Despite the recommendation of dual-antiplatelet therapy (DAPT) after percutaneous coronary intervention (PCI),¹³ the clinical dilemma of whether to continue DAPT (and for how long) in the setting of bleeding continues. In a recent study by Sorrentino et al, patients from the PARIS (Patterns of Non-adherence to Antiplatelet Regimens in Stented Patients) registry were stratified according to bleeding risk and were found to have increased risks for adverse cardiac events with DAPT disruption irrespective of bleeding risk.¹⁴ In addition to quantifying the relationship between post-TAVR AKI and absence of RAS inhibition, our study provides a potential mechanistic insight into the relationship between AKI and adverse long-term outcomes. This concept is unique and analogous to prior work related to acute bleeding, discontinuation of antiplatelet therapy, and adverse 1 year outcomes after percutaneous coronary interventions.

The rate of RAS inhibitor use after TAVR (41.9%) in our study is consistent with several other studies, which reported rates between 39.7%-58%.^7,8,15 This utilization rate is lower than among patients with non-TAVR HFrEF; ACE inhibitors/ARBs are prescribed in 79% of non-TAVR HFrEF patients, while aldosterone antagonists are prescribed in 36.4% of patients.¹⁶ Even among HFrEF patients who present for TAVR, only 52.8% of the patients were discharged with RAS inhibitors. In non-TAVR HFrEF patients, older age and lung disease have been demonstrated to be negative predictors of RAS inhibitor use.^16-18 Older age was not a potent predictor of discharge utilization of RAS inhibition in our TAVR registry, suggesting that other factors guide the discharge medication decisions in this particular high-risk population. The most predictive factor for RAS inhibitor use after TAVR was RAS inhibitor use prior to TAVR.

We hypothesized that utilization of RAS inhibition after TAVR would interact with changes in renal function; we observed that patients who developed AKI after TAVR were significantly less likely to be discharged with RAS inhibitor use. Patients who were discharged without RAS inhibition had a longer length of stay, which likely reflected a sicker population. Of note, longer length of stay did not translate into increased initiation of RAS inhibition. The incidence of AKI post TAVR has been estimated to be up to 20% and has been associated with increased mortality.^15,19-21 The mechanism of this increased death risk is unknown, as AKI is usually a transient phenomena. Our observations regarding the interaction between AKI post TAVR and discharge medication utilization provide a potential hypothesis as to a mechanism for increased mortality risk.

The predictors of RAS inhibitor discontinuation were generally similar to the predictors of RAS inhibitor use at discharge; ie, AKI, low blood pressure, and new pacemaker requirement all predicted RAS inhibitor discontinuation. Our results are concordant with prior studies of RAS inhibitor use in non-TAVR populations: AKI and blood pressure have previously been shown to be predictors of RAS inhibitor discontinuation in non-TAVR HF patients.²² In addition to the interaction with AKI, we observed an interaction of new pacemaker placement with discontinuation of RAS inhibition; there are conflicting data regarding the association of new pacemaker implantation after TAVR with long-term mortality.^23-25 Given the potential benefits of RAS inhibition on post-TAVR LV remodeling,⁷ the negative interaction of post-TAVR complications (AKI and pacemaker) with appropriate post-TAVR medical therapy may be associated with worse outcomes.

Study limitations. This study is limited by the single-center registry design; on the other hand, we were able to enroll all consecutive TAVR patients over a 5-year period to maximize generalizability. Second, as with all registry-based associations, there may be other confounding variables that were not measured and may impact the decision to use or discontinue RAS inhibitors. Our study is also not powered to compare outcomes among patients with RAS inhibition vs those without RAS inhibition; our focus is on the predictors of utilization of these agents. Finally, we have limited our analysis to in-hospital outcomes. We cannot comment on long-term patient outcomes or whether or not patients resumed RAS inhibition at 30 days or 1 year after TAVR. However, we do note that among patients who develop AKI after coronary angiography, <4% of patients progress to dialysis.²⁶ Thus, we do not believe that 30-day analysis would have significantly altered the primary observation of our study.

RAS inhibitors are used in approximately 40% of TAVR patients at hospital discharge. RAS inhibition prior to TAVR is the strongest predictor of RAS inhibitor use at discharge, and 2 TAVR-related complications (AKI and pacemaker requirement) predict both lack of discharge RAS inhibition and RAS inhibitor discontinuation. Mainly, AKI is associated with decreased prescriptions of guideline-directed medical therapy, and this may be an explanation for the important association of AKI with poor 1-year outcomes. As a result, the routine use of RAS inhibition after TAVR warrants further study and the interaction of AKI with optimal post-TAVR medical therapy suggests a target for future studies.