Evolution of Surgical Aortic Valve Replacement in the Era of Transcatheter Valve Technology

Transcatheter aortic valve replacement (TAVR) has shown similar clinical benefit and symptomatic recovery to surgical aortic valve replacement (SAVR) for high- and intermediate-risk patients with severe aortic stenosis in the PARTNER I and II trials.^1,2 We hypothesized the adoption of TAVR would have significantly impacted resource use and postoperative outcomes after SAVR.

A total of 173 108 adult patients undergoing isolated aortic valve replacement between January 1, 2004, and December 31, 2013, were identified from the National Inpatient Sample’s Healthcare Cost and Utilization Project using the International Classification of Diseases, Ninth Revision procedural codes for SAVR (35.21 and 35.22) and TAVR (35.05 and 35.06). The National Inpatient Sample is a 20% stratified sample of discharges from more than 4000 community hospitals modeling over 35 million US hospitalizations annually. Patients undergoing redo or concomitant cardiac operations were excluded. This study was exempted from review by the institutional review board at the University of California, Los Angeles because the National Inpatient Sample is a publicly available deidentified database sponsored by the Agency for Healthcare Research and Quality. A data use agreement with the Healthcare Cost and Utilization Project was completed.

Cost, length of stay, and mortality were estimated using Healthcare Cost and Utilization Project survey weights. Costs were standardized to the 2013 US gross domestic product using US Department of Commerce consumer price indices and adjusted for diagnosis related group–based severity. The Elixhauser Comorbidity Index, which identifies 31 common comorbidities, was used to estimate disease severity. Cardiovascular comorbidities and complications were further identified using peer-reviewed International Classification of Diseases, Ninth Revision codes for cardiac surgery.³ Mortality, length of stay, and log-transformed costs were modeled using hierarchical multivariable logistic, Poisson distribution, and linear distribution, respectively, controlling for patient demographics, comorbidities, complications, and hospital characteristics. A piecewise multivariable regression model was used to compare trends in SAVR cost before and after 2011. Statistical analyses were performed using Stata version 14 (StataCorp LLC), with P < .001 considered significant after Bonferroni correction.

Demographics, comorbidities, and outcomes were estimated for each of the 3 cohorts of patients undergoing isolated SAVR from 2004 to 2010 (early cohort), SAVR from 2011 to 2013 (late cohort), and TAVR from 2011 to 2013 (Table).

Although the mean unadjusted cost of the late cohort was $3093 greater than that of the early cohort (95% CI, 730-5456; P < .001), the annual cost of SAVR has decreased by 4.92% (95% CI, 3.26-6.54; P < .001) since 2011. In SAVR patients, the Elixhauser Comorbidity Index was significantly associated with increased cost (β = 0.066 [95% CI, 0.062-0.070]; P < .001) and stabilized after the advent of TAVR (Figure). Complications associated with increased cost and length of stay, including infection and stroke, also remained stable after 2011. Mortality after SAVR decreased throughout the decade (odds ratio, 0.75; 95% CI, 0.63-0.89; P < .001).

Since US Food and Drug Administration approval, use of TAVR has rapidly increased from 1164 to 13 525 cases annually. The mean cost of TAVR has risen from $51 008 to $55 136 (P < .001) despite no significant change in Elixhauser Comorbidity Index score, proportions of individual comorbidities, or rates of postoperative complications. There was no difference in adjusted mortality (odds ratio, 0.81; 95% CI, 0.62-1.05; P = .11) or rate of postoperative neurologic complications (odds ratio, 0.90; 95% CI, 0.55-1.47; P = .68) between TAVR and late cohort SAVR patients. Yet, TAVR was 8.38% (95% CI, 5.98-10.84; P < .001) more expensive than SAVR after multivariable adjustment.

Our study demonstrates the impact of newly introduced TAVR technology on resource use in SAVR. The reduction in cost of SAVR and stabilization of disease severity reflect more efficient allocation of resources between SAVR and TAVR. However, the cost of TAVR is increasing. Previous analyses have recommended reductions in the initial cost of TAVR to ensure its cost-effectiveness in practice and implicated the higher fixed cost of the valve.^4,5 Our data show that this discrepancy remains despite the development of new generations of valves and increased competition in transcatheter technology.⁶ Further research is necessary to elucidate whether this increase reflects a learning curve as TAVR programs become established. As the indication for TAVR expands to medium- and low-risk cohorts, legislation may be necessary to ensure its cost-effectiveness.

Corresponding Author: Peyman Benharash, MD, Division of Cardiac Surgery, David Geffen School of Medicine at UCLA, 10833 Le Conte Ave, 62-249 Center for Health Sciences, Los Angeles, CA 90095 (pbenharash@mednet.ucla.edu).

Accepted for Publication: April 24, 2017.

Published Online: July 19, 2017. doi:10.1001/jamasurg.2017.2344

Author Contributions: Mr Mantha and Dr Benharash had full access to all of the data in the study and take responsibility for the integrity of the data and the accuracy of the data analysis.

Study concept and design: Mantha, Juo, Shemin, Benharash.

Acquisition, analysis, or interpretation of data: Mantha, Morchi, Ebrahimi, Ziaeian, Benharash.

Drafting of the manuscript: Mantha, Ziaeian, Shemin, Benharash.

Critical revision of the manuscript for important intellectual content: Mantha, Juo, Morchi, Ebrahimi, Ziaeian, Benharash.

Statistical analysis: Mantha, Shemin.

Administrative, technical, or material support: Mantha, Juo.

Study supervision: Juo, Morchi, Ebrahimi, Ziaeian, Benharash.

Conflict of Interest Disclosures: None reported.