Trends and Outcomes of TAVI and SAVR in Cancer and Noncancer Patients

Background Patients with severe aortic stenosis and cancer are often denied surgical aortic valve replacement (SAVR) due to a prohibitive risk of perioperative mortality. Objectives The purpose of this study was to determine the safety of transcatheter aortic valve implantation (TAVI) in patients with severe aortic stenosis and cancer. Methods The Nationwide Inpatient Sample database (2002-2018) was used to study the outcomes of TAVI vs SAVR in patients with active or prior history of prostate, lung, colorectal, breast, and renal cancer. A propensity score-matched analysis to calculate adjusted odds ratios (aORs) for major adverse cardiovascular events (MACEs) and its components. Results A total of 1,505,995 crude population and a subset of 345,413 noncancer and 33,565 cancer patients were selected on propensity score-matched analysis. The yearly trend showed a steep increase in the utilization of TAVI. Compared with SAVR, TAVI had a lower risk of in-hospital mortality in prostate cancer, while there was no difference among other cancer types. Patients with lung (aOR: 0.65; 95% CI: 0.43-0.97) and prostate cancer (aOR: 0.79; 95% CI: 0.66-0.96) had lower, while colorectal cancer (aOR: 1.43; 95% CI: 1.08-1.90) had higher odds of MACE with TAVI. The incidence of major bleeding was lower with TAVI (except for lung cancer), while the risk of stroke was similar (except for colorectal cancer) between TAVI and SAVR. Conclusions TAVI in patients with prostate, breast, lung, and renal cancer appears to be a reasonable alternative to SAVR with lower or similar risks of mortality and MACE.

W ith early diagnosis and contin- uous advancement in cancer therapies, there has been a significant improvement in the life expectancy of oncology patients. 1This has paradoxically increased the prevalence of cardiovascular diseases in cancer-surviving adults.A growing body of literature has demonstrated a mechanistic overlap between cancer and heart diseases. 2Cardiovascular disease manifesting after cancer may be due to the shared risk factors profile, adverse effects of cancer therapies, and severe inflammatory state associated with malignancies.Specifically, the extension of potentially cardiotoxic chemotherapies to elderly populations with preexisting aortic stenosis (AS), and radiotherapy-induced damage to the aortic valve has resulted in the simultaneous presence of both severe AS and cancer.Recent reports show that the prevalence of malignancy in patients with severe AS is >26%. 3The improved survival following cancer treatment, coupled with the toxic impacts of cancer therapies on AS have made the latter an important determinant of morbidity and mortality in this patient cohort. 4Currently, there has been no consensus on the appropriate management of these patients who have severe AS and a history of cancer.][7][8][9][10] Hence, we sought to determine the relative merits of TAVI vs SAVR in patients with the most common cancers.STUDY OUTCOMES.The primary end point was a composite of stroke and in-hospital mortality termed as major adverse cardiovascular events (MACEs).
Secondary end points consisted of individual components of MACE, major bleeding, postprocedural bleeding, and the need for permanent pacemaker (PPM) implantation.A detailed explanation of outcomes is given in Supplemental Table 2.
STATISTICAL ANALYSIS.NIS discharge weights were used to obtain national estimates.Baseline comorbidities were compared using descriptive statistics.
A chi-squared test was used to compare the percentages of categorical variables.The Cochran-Mantel Haenszel test was used to compute unadjusted odds ratios (ORs) for crude data.To compute adjusted odds ratios (aORs), a 2-step approach was adopted for each subgroup (no-cancer, CRC, prostate, lung, kidney,  3).The complete list of matching variables is given in Supplemental Table 1.A sensitivity analysis based on the sequential exclusion of potential covariates such as males, females, older adults, and those with nonactive and metastatic cancer was also performed to assess its impact on pooled outcomes.Using a multivariate interaction analysis, the intergroup estimates of major outcomes for TAVI vs SAVR groups were also calculated across males vs females and younger vs older populations.
Linear regression analysis was used to assess the annual trends of the utilization of procedures.
The trend analysis of annual percentage change (APC) for the overall cohort was also performed using the joinpoint regression method.The calculated effect sizes were analyzed using point estimates with its 95% CI and a type I error of P # 0.05 was chosen as a cutoff for statistical significance.All analyses were performed using SPSS v24 (IBM Corp), R 3.2, and STATA v16 (StataCorp).

RESULTS
SELECTION OF CASES.On the initial search, 38,349,112 weighted hospitalizations of aortic valve disease were identified.Of these, 1,505,995 patients underwent isolated AVR.Among these, 1,373,817 had no history of cancer, while 132,177 had a prior history of cancer or active cancer.From the latter, 104,003 patients with the most common cancer types were selected.Of these, 14,593 patients had active cancer and 4,412 had metastatic cancer (Supplemental Table 4).A weighted sample of 345,413 noncancer and 33,565 most common cancer propensity-matched populations were selected for adjusted analysis (Supplemental Figure 3).The detailed frequency of major cancer types undergoing TAVI vs SAVR is given in Table 1.BASELINE CHARACTERISTICS.Detailed baseline comorbidities for the overall population on crude unadjusted analysis are presented in Supplemental Table 5.The mean age of patients undergoing TAVI was significantly higher than those having SAVR in all cancer groups.The majority of patients (>70%) in both groups were Caucasian.Patients undergoing TAVI had a significantly higher burden of traditional cardiovascular comorbidities irrespective of the type of cancer.More than 80% of the procedures were nonemergent and elective over the weekdays.Using PSM, the baseline covariates were evenly distributed in the matched groups (TAVI vs SAVR) across all cancer types.(Figure 2, Supplemental Figure 4, Tables 2 and 3).6 and 7, Figure 3).There was a decline in the utilization of SAVR, (APC: À0.48%; 95% CI: À1.1% to 0.2%), and a significant increase in the use of TAVI (APC: 4.79%; 95% CI: 1.5%-11.5%)during the study period (2011-2018).Regarding the major outcomes, there was a decreasing trend observed in the percentage of MACE over the years in all cancer types.
The risk of major bleeding and stroke was variable, while the need for PPM implantation remained significantly higher in TAVI across all timepoints regardless of the history and type of cancer (Supplemental Tables 8 and 9).CRUDE UNADJUSTED ANALYSIS.On crude unadjusted analysis, the odds of MACE, mortality, and other procedure-related complications (stroke and major bleeding) were significantly lower with TAVI in patients with no history of cancer and those having prostate cancer.Patients with breast cancer had a higher, while those with lung and colorectal cancer had a similar risk of in-hospital mortality with TAVI compared with SAVR.The need for PPM remained significantly higher among TAVI groups irrespective of the type of malignancy.The detailed estimates of all unadjusted outcomes are given in Supplemental Tables 10 to 15.
Contrary to the net CRC odds, restricting the analysis to men only showed no significant difference in the rate of MACE and stroke with TAVI vs SAVR.
Restricting the analysis to older adults with breast cancer, showed a higher rate of MACE and mortality with TAVI compared with SAVR (Supplemental Tables 17 and 18).expertise, and widespread availability of the TAVI.It also implies that the off-label use of TAVI was preferred in low-risk patients with cancer. 11The findings of our study echo the observation by the

STS-ACC TVT (Society of Thoracic Surgeons-American College of Cardiology Transcatheter Valve
Therapy) registry that also indicated a gradual decline in the use of SAVR, while the portion of cancer patients among TAVI recipients increased over the years. 12 management in cancer patients is very complex due to nonspecific and overlapping symptoms, higher periprocedural risks of complications, and concerns for futility in advanced stages of cancer with lower life expectancy.Previous smaller studies suggested that TAVI in cancer might be safe, with a similar risk of short-term mortality and procedural complications compared with those without cancer, however, the relative merits of TAVI vs SAVR in the full spectrum of different cancer types and different baseline characteristics were not evaluated. 13,14We believe that the safety of TAVI in cancer patients is contingent upon the cancer type, demographics, and underlying comorbidities.Our findings suggest that compared with SAVR, TAVI significantly reduced the overall risk of composite ischemic events in patients with cancer except those with CRC.
In the cancer patient, the highest benefits with comes for these patients treated with TAVI. 15C, on the other hand, had a higher in-hospital MACE with TAVI compared with SAVR, but these findings were driven by higher odds of stroke, as there was no difference in mortality between the 2 procedures.Also, the higher MACE risk was only observed in patients >65 years, indicating that older .9 0 0 0 0.9 0.8 adults, being frail, have less cardiac reserve and thus might benefit less from TAVI.Additionally, high stages of malignancy coupled with a lower life expectancy of CRC patients might be contributing to these outcomes.This contrasts with the limited data available from the case reports, which showed favorable outcomes with TAVI in patients with CRC. 16,17Similar trends were observed in patients with breast cancer, where older adults had a significantly higher risk of MACE and mortality compared with both younger patients undergoing TAVI, as well as those of the same age group undergoing SAVR.This could hypothetically be explained by the relatively higher incidence of prior exposure to therapeutic radiation in older adults, causing significant fibrosis and dystrophic calcifications of the aortic valve. 18,19Nonetheless, overall TAVI in patients with breast cancer patients had a similar risk of MACE and mortality as SAVR.
In the current study, irrespective of the type, active status, and metastasis of cancer, most of the procedure-related complications were lower with TAVI (except the need for PPM) due to the lessinvasive nature of the procedure, shorter hospital stay, and lower need for general anesthesia.We observed a numerically lower rate of postprocedural and major bleeding with TAVI across all cancers except lung cancer, highlighting the safety and feasibility of the TAVI procedure in the highly coagulopathic environment of cancer.The higher bleeding risk in lung cancer could be due to the antiangiogenic properties of the chemotherapeutic agents and the innate nature of lung cancers to form cavitary lesions. 20Nonetheless, this did not translate into hard clinical outcomes as indicated by a lower or similar risk of MACE and in-hospital mortality between TAVI and SAVR in lung cancer patients.
These findings contrast the results of the study by Landes et al 13 that showed higher 1-year mortality and increased bleeding (any bleeding 14.4% vs 6.9%; major bleeding 8.6% vs 3.1%) with TAVI among cancer vs noncancer patients, respectively.However, the higher mortality was mostly attributed to advanced cancer-related deaths from noncardiovascular causes,    postprocedure events, and events that can be linked with TAVI and SAVR (bleeding, PPM), we could not determine the actual timing (preprocedure or postprocedure) of these events.Data are subject to inherent selection bias as only eligible patients with cancer and severe AS were chosen for TAVI.Similarly, the influence of differences in disease assessment, documentation, and center-specific expertise on the measured outcomes could not be assessed.Lack of information regarding the echocardiographic measures, individual valvular anatomy, STS score, stages of cancer, and hemodynamics precluded further individualized risk assessment.Although all codes were validated from reliable sources, the possibility of an inadvertent error in coding cannot be entirely excluded.Due to the nature of the data, we were unable to retrieve information regarding the severity of AS, duration of intervention, type of valve used, stage of cancer, and medications used by the patients.
Similarly, due to the inclusion of repeated observations from patients who underwent >1 AVR (TAVI or SAVR), a distinction between index and repeat events was not possible.Although some of the unquantifiable potential confounders such as operators' skills and device type could not be assessed, the yearly trend analysis was performed as a surrogate marker for these covariates.A larger-scale study of a randomized population followed up on the prospective scale is needed to validate our findings.

CONCLUSIONS
Overall, TAVI appears to be a safer alternative to SAVR in patients with breast, prostate, renal, and lung cancers due to a similar or lower risk of MACE and mortality.Females and patients >65 years old might have higher MACE with TAVI in colorectal cancer; however, there was no difference in mortality irrespective of age and sex.Due to variable outcomes, informed multifaceted decisions by the heart team in conjunction with the oncologists should be used to choose appropriate candidates for TAVI.

FUNDING SUPPORT AND AUTHOR DISCLOSURES
The authors have reported that they have no relationships relevant to the contents of this paper to disclose.

ADDRESS FOR CORRESPONDENCE
METHODSDATA SOURCE.Nationwide Inpatient Sample (NIS) is the largest United States payer database of hospitalized patients that is publicly available through the Agency for Healthcare Research and Quality.It is managed and monitored by the Healthcare Cost and Utilization Project.It contains de-identified in-hospital data of weighted discharges for more than 15 million hospitalizations per year.Due to the anonymized nature of data, NIS is exempted from the approval of the institutional review board.STUDY DESIGN AND POPULATION.The current study included all adult patients (>18 years of age) with a diagnosis of AS undergoing aortic valve replacement (AVR) between 2002 and 2018.The included sample was divided into 2 groups: patients undergoing TAVI vs SAVR.All comparisons were stratified based on the prior history of or active presence of one of the 5 major cancer types (breast, prostate, kidney, colorectal cancer (CRC), and lung).The standard International Classification of Disease, Clinical Modifications codes were used to identify different cancers, baseline comorbidities, and outcomes of interest (other than mortality).The detailed list of International Classification of Disease-10 codes is provided in Supplemental Figures1 and 2).For PSM, more than 30 variables

FIGURE 1
FIGURE 1 Propensity Matched Analysis Showing the Standardized Mean Differences of Major Comorbidities Showing No Deviation Beyond the Allowable Threshold (0.1 SMD and 0.05 KSS) TEMPORAL TREND ANALYSIS.The yearly trend showed a steep increase in the utilization of TAVI from 2011 to 2018 irrespective of the type of cancer (Supplemental Figure5).By the year 2015, the proportion of TAVI procedures among all cancer patients had surpassed the corresponding proportion of SAVR procedures.TAVI in lung cancer observed this transition relatively earlier during 2012 (Supplemental Tables

FIGURE 2
FIGURE 2 Baseline Comorbidities Among Patients With No Cancer and Different Types of Cancers Undergoing SAVR vs TAVI on Propensity-Matched Analysis

FIGURE 3
FIGURE 3 Yearly Trend of SAVR and TAVI Procedures Performed for Patients With No Cancer, Breast, Colorectal, Lung, Prostate, and Renal Cancer

FIGURE 4
FIGURE 4 Proportion of MACE, Stroke, Mortality, Major Bleed, and PPM Implantation Among Patients With No Cancer and Different Types of Cancers Undergoing SAVR vs TAVI on Propensity-Matched Analysis

FIGURE 5
FIGURE 5 Forest Plot Showing Adjusted OR and its 95% CI for Major Outcomes Across All Outcomes rather than TAVI-related complications.A recent meta-analysis by Murphy et al 21 showed no difference in both short-and long-term complications and mortality between cancer and noncancer patients with TAVI reinforcing its role in cancer patients.Overall, our analysis revealed an increased need for PPM implantation with TAVI, which was consistently high over the years from 2002 to 2018.This implies that factors other than operator skills and expertise such as inherent procedural and prosthesis-related factors might be playing a predominant role here. 22STUDY LIMITATIONS.The findings of our study should be interpreted with the acknowledgment of certain limitations.Being a cross-sectional study, we could only report a temporal association; no definitive conclusions regarding the causality and longterm outcomes could be made.Also, it is important to note that despite using codes for possible CENTRAL ILLUSTRATION Showing Major Outcomes Among All Cancer Patients Undergoing TAVI vs SAVR Ullah W, et al.JACC Adv.2023;2(1):100167.¼ means no significant difference in the estimates.CRC ¼ colorectal cancer; MACE ¼ major adverse cardiovascular events; PPM ¼ permanent pacemaker; SAVR ¼ surgical aortic valve replacement; TAVI ¼ transcatheter aortic valve implantation.

TABLE 1
Sample Size of Unmatched and Propensity-matched Populations Across No Cancer and Different Types of Cancer Undergoing SAVR ¼ surgical aortic valve replacement; TAVI ¼ transcatheter aortic valve implantation.Central Illustration).The intergroup (TAVI vs SAVR)odds of stroke and in-hospital mortality were similar across all cancer types, except that CRC showed a higher rate of stroke (aOR: 2.01; 95% CI: 1.43-2.84),whileprostatecancerhadalowerrisk of in-hospital mortality (aOR: 0.65; 95% CI: 0.46-0.91)withTAVI.The bivariate analysis of major components of MACE is given in Supplemental Figure6.Among all cancer types, TAVI was associated with a lower risk of major bleeding and a higher rate of PPM implantation except in lung cancer, which showed a higher risk of major bleed (aOR: 3.08; 95% CI: 2.07-4.59)andalowerneedfor PPM (aOR: 0.62; 95% CI: 0.45-0.84)comparedwithSAVR.The risk of procedural complications was mostly similar or lower with TAVI in most of the cancer types (Table4).Patients with CRC had a higher mean length of stay compared with other cancer types (Supplemental Figure7).

TABLE 2
Baseline Crude Demographics of Patients Without Cancer and With Cancer ACH ¼ acute care health; NT ¼ nonteaching; SAVR ¼ surgical aortic valve replacement; TAVI ¼ transcatheter aortic valve implantation.

TABLE 3
Propensity Matched Baseline Comorbidities of Patients Underwent TAVI and SAVR Without Cancer and With Different Cancers

:
Dr Waqas Ullah, Department of Cardiology, Thomas Jefferson University Hospital, 111 S 11th Street, Philadelphia, Pennsylvania 19107, USA.E-mail: waqasullah.dr@gmail.com.17.Tanaka T, Yahagi K, Okuno T, et al.Transcatheter aortic valve implantation in a patient with severe aortic valve stenosis, colon cancer, and obstructive ileus: a case report.J Cardiol Cases.Donnellan E, Masri A, Johnston DR, et al.Long-term outcomes of patients with mediastinal radiation-associated severe aortic stenosis and subsequent surgical aortic valve replacement: a matched cohort study.J Am Heart Assoc.2017;6(5):e005396.20.Topkan E, Selek U, Ozdemir Y, et al.Risk factors for fatal pulmonary hemorrhage following concurrent chemoradiotherapy in stage 3B/C squamous-cell lung carcinoma patients.J Oncol.2018;2018:4518935.21.Murphy AC, Koshy AN, Cameron W, et al.Transcatheter aortic valve replacement in patients with a history of cancer: periprocedural and longterm outcomes.Catheter Cardiovasc Interv.2021;97(1):157-164.22. Ullah W, Zahid S, Zaidi SR, et al.Predictors of permanent pacemaker implantation in patients undergoing transcatheter aortic valve replacement-a systematic review and meta-analysis.J Am Heart Assoc.2021;10(14):e020906.KEY WORDS cancer, severe aortic stenosis, surgical aortic valve replacement, transcatheter aortic valve implantation APPENDIX For supplemental tables and figures, please see the online version of this paper.