Advertisement

Redefining “low risk”: Outcomes of surgical aortic valve replacement in low-risk patients in the transcatheter aortic valve replacement era

      Abstract

      Objectives

      Guidelines suggest aortic valve replacement (AVR) for low-risk asymptomatic patients. Indications for transcatheter AVR now include low-risk patients, making it imperative to understand state-of-the-art surgical AVR (SAVR) in this population. Therefore, we compared SAVR outcomes in low-risk patients with those expected from Society of Thoracic Surgeons (STS) models and assessed their intermediate-term survival.

      Methods

      From January 2005 to January 2017, 3493 isolated SAVRs were performed in 3474 patients with STS predicted risk of mortality <4%. Observed operative mortality and composite major morbidity or mortality were compared with STS-expected outcomes according to calendar year of surgery. Logistic regression analysis was used to identify risk factors for these outcomes. Patients were followed for time-related mortality.

      Results

      With 15 observed operative deaths (0.43%) compared with 55 expected (1.6%), the observed:expected ratio was 0.27 for mortality (95% confidence interval [CI], 0.14-0.42), stroke 0.65 (95% CI, 0.41-0.89), and reoperation 0.50 (95% CI, 0.42-0.60). Major morbidity or mortality steadily declined, with probabilities of 8.6%, 6.7%, and 5.2% in 2006, 2011, and 2016, respectively, while STS-expected risk remained at approximately 12%. Mitral valve regurgitation, ventricular hypertrophy, pulmonary, renal, and hepatic failure, coronary artery disease, and earlier surgery date were residual risk factors. Survival was 98%, 91%, and 82% at 1, 5, and 9 years, respectively, superior to that predicted for the US age-race-sex–matched population.

      Conclusions

      STS risk models overestimate contemporary SAVR risk at a high-volume center, supporting efforts to create a more agile quality assessment program. SAVR in low-risk patients provides durable survival benefit, supporting early surgery and providing a benchmark for transcatheter AVR.

      Graphical abstract

      Key Words

      Abbreviations and Acronyms:

      AVR (aortic valve replacement), CI (confidence interval), O:E (observed:expected ratio), PROM (predicted risk of mortality), SAVR (surgical aortic valve replacement), STS (Society of Thoracic Surgeons), TAVR (transcatheter aortic valve replacement)
      Figure thumbnail fx2
      Temporal trend of observed and expected results of major morbidity or mortality after SAVR.
      Patients who undergo surgical aortic valve replacement in a center of excellence are at low risk with durable survival benefit, supporting early surgery. This risk is overestimated by STS models.
      Outcomes of SAVR in low-risk (STS PROM <4%) patients have improved over time, and expected risk no longer reflects current outcomes. SAVR in these patients is a very low-risk intervention in an experienced center, with durable survival benefit. More agile real-time risk modeling is needed to evaluate the evolving role of transcatheter procedures in these patients.
      See Commentaries on pages 605, 606, and 607.
      American College of Cardiology/American Heart Association 2017 guidelines for treating aortic valve disease suggest that surgical aortic valve replacement (SAVR) might be indicated in selected patients at low risk for aortic valve replacement (AVR), whereas transcatheter AVR (TAVR) is indicated only for symptomatic patients.
      • Nishimura R.A.
      • Otto C.M.
      • Bonow R.O.
      • Carabello B.A.
      • Erwin III, J.P.
      • Fleisher L.A.
      • et al.
      2017 AHA/ACC focused update of the 2014 AHA/ACC guideline for the management of patients with valvular heart disease: a report of the American College of Cardiology/American Heart Association task force on clinical practice guidelines.
      Recently, 2 randomized trials showed favorable outcomes for TAVR in low-risk patients.
      • Mack M.J.
      • Leon M.B.
      • Thourani V.H.
      • Makkar R.
      • Kodali S.K.
      • Russo M.
      • et al.
      Transcatheter aortic-valve replacement with a balloon-expandable valve in low-risk patients.
      ,
      • Popma J.J.
      • Deeb G.M.
      • Yakubov S.J.
      • Mumtaz M.
      • Gada H.
      • O'Hair D.
      • et al.
      Transcatheter aortic-valve replacement with a self-expanding valve in low-risk patients.
      Two simultaneous trends—evolution in device technology
      • Chiam P.T.
      • Ruiz C.E.
      Percutaneous transcatheter aortic valve implantation: evolution of the technology.
      and improvements in echocardiographic imaging
      • Zamorano J.L.
      • Badano L.P.
      • Bruce C.
      • Chan K.L.
      • Gonçalves A.
      • Hahn R.T.
      • et al.
      EAE/ASE recommendations for the use of echocardiography in new transcatheter interventions for valvular heart disease.
      —allow more accurate decision-making regarding timing of TAVR in asymptomatic patients, combining to accommodate a rapidly shifting landscape in aortic valve disease.
      • Vahl T.P.
      • Kodali S.K.
      • Leon M.B.
      Transcatheter aortic valve replacement 2016: a modern-day “Through the Looking-Glass” adventure.
      What is often lost in this discussion is that SAVR has also undergone refinements in technique and safety. Previous studies have shown that for coronary artery bypass grafting, risk prediction models, such as those of the Society of Thoracic Surgeons (STS), often overestimate observed risk of operative mortality and morbidity, especially for experienced centers performing a high volume of operations.
      • Nilsson J.
      • Algotsson L.
      • Höglund P.
      • Lührs C.
      • Brandt J.
      Early mortality in coronary bypass surgery: the EuroSCORE versus the Society of Thoracic Surgeons risk algorithm.
      To provide a picture of the current state of SAVR in a single high-volume center, which might provide guidance for patients, cardiologists, and surgeons in comparing real-world outcomes with emerging technology, we studied outcomes after isolated SAVR in patients with STS predicted risk of mortality (PROM) <4%. This reflects in part the definition of “low risk” used in the 2 TAVR trials, with notable exceptions such as bicuspid aortic valve disease. We evaluated operative mortality and morbidity outcomes according to STS definitions, residual risk factors for adverse outcomes over and above those in STS models, and post-discharge survival and reoperation.

      Methods

      Patients

      From January 2005 to January 2017, 3474 adults with an STS PROM score <4% underwent 3493 isolated SAVRs at Cleveland Clinic's main campus. Mean age at surgery was 65 ± 13 years, and 1231 (35%) of the patients were female (Table 1). Mean preoperative aortic valve area was 0.77 ± 0.38 cm2 and mean preoperative gradient 47 ± 20 mm Hg. Age younger than 18 years, etiology of endocarditis, and those with concomitant cardiac procedures were exclusion criteria.
      Table 1Characteristics and surgical approach in aortic valve replacement patients (n = 3493)
      Variablen
      Patients with data available.
      No. (%) or mean ± SD
      Demographics
       Age, years349364 ± 13
       Female34931231 (35)
       Body mass index, kg/m2347030 ± 6.2
       Race3481
      Black77 (2.2)
      White3229 (93)
      Other161 (4.6)
      STS score
       PROM, %34931.6 ± 0.92
      Symptoms
       NYHA functional class2868
      I825 (29)
      II1550 (54)
      III473 (16)
      IV20 (0.70)
       Emergency operation34920 (0)
      Valve morphology and pathology
       Bicuspid aortic valve34581349 (39)
       Stenosis34932784 (80)
       Effective aortic valve area, cm226120.77 ± 0.38
       Mean gradient, mm Hg309647 ± 20
       Aortic regurgitation grade3484
      None1470 (42)
      Mild553 (16)
      Moderate636 (18)
      Severe825 (24)
       Other valve regurgitation3493
      Mitral1088 (31)
      Tricuspid713 (20)
      Cardiac comorbidity
       Previous cardiac operation3490686 (20)
       Coronary artery systems diseased >50%3493
      02901 (83)
      1332 (9.5)
      2118 (3.4)
      3142 (4.1)
       Previous myocardial infarction3493264 (7.6)
       Left ventricular ejection fraction, %346257 ± 8.9
       Atrial fibrillation3459174 (5.0)
       Complete heart block/pacer345746 (1.3)
      Noncardiac comorbidity
       Peripheral arterial disease3493184 (5.3)
       Previous stroke3480187 (5.4)
       Hypertension34932659 (76)
       Chronic obstructive pulmonary disease3493515 (15)
       Diabetes3482
      Pharmacologically treated626 (18)
      Insulin treated188 (5.4)
       History of smoking34841713 (49)
       Creatinine, mg/dL34480.98 ± 0.37
       Renal dialysis349314 (0.40)
       Bilirubin, mg/dL34240.61 ± 0.38
      Surgical approach
       Full incision34881578 (45)
       Less invasive34881910 (55)
      SD, Standard deviation; STS, Society of Thoracic Surgeons; PROM, predicted risk of mortality; NYHA, New York Heart Association.
      Patients with data available.

      Surgical Technique

      Isolated SAVR was performed using the standard technique of aortic clamping, arrested heart with cardioplegia, valve excision, and prosthesis implant. Operative approach and prosthesis choice were at the discretion of the surgeon, with 3163 (92%) being a bioprosthesis. Among the approaches (full sternotomy, upper hemisternotomy J-incision, right minithoracotomy), 1910 (55%) were less invasive incisions (Table 1). All prostheses were included with the exception of porcine or allograft roots implanted via a full root technique with reimplantation of coronary buttons, because these fall outside of the STS definition of isolated SAVR.

      End Points

      Study end points included operative mortality and 6 in-hospital major morbidity outcomes: (1) permanent stroke, (2) renal failure, (3) prolonged ventilation, (4) deep sternal wound infection, (5) reoperation, and (6) the combined end point of major morbidity or mortality (composite operative mortality and any of the mentioned morbidities). In addition, prolonged postoperative length of stay (>14 days) was assessed. All of these end points were defined according to the STS national cardiac database (https://www.sts.org/registries-research-center/sts-national-database/sts-adult-cardiac-surgery-database).
      Time-related outcomes included mortality from time of operation and aortic valve reoperation. For these, systematic follow-up was conducted at 2, 5, and 10 years with patient consent, yielding 5298 patient-years of data.
      Patient, operative, and follow-up variables were obtained from the Cleveland Clinic Cardiovascular Information Registry. These data are approved for use in research by the Institutional Review Board, with patient consent waived (approval number 4826, approved annually through December 31, 2021).

      Data Analysis

      All analyses were performed using SAS version 9.4 (SAS Institute Inc, Cary, NC).

      STS benchmark scores

      STS-defined operative mortality and morbidity outcomes were compared with contemporaneous STS benchmarks generated from national surgical experience from 2002 through 2006.
      • O'Brien S.M.
      • Shahian D.M.
      • Filardo G.
      • Ferraris V.A.
      • Haan C.K.
      • Rich J.B.
      • et al.
      The Society of Thoracic Surgeons 2008 cardiac surgery risk models: part 2–isolated valve surgery.
      ,
      • Shahian D.M.
      • O'Brien S.M.
      • Filardo G.
      • Ferraris V.A.
      • Haan C.K.
      • Rich J.B.
      • et al.
      The Society of Thoracic Surgeons 2008 cardiac surgery risk models: part 3–valve plus coronary artery bypass grafting surgery.
      STS-certified equations were solved for expected risks and compared with observed outcomes using STS-defined variables. Confidence intervals (CIs) for observed:expected ratios (O:Es) were estimated using the bootstrap percentile method
      • Efron B.
      • Tibshirani R.J.
      An Introduction to the Bootstrap.
      with 1000 bootstrap samples, and P values for comparisons between observed and expected frequencies were obtained using the χ2 goodness of fit test.

      Temporal trends of major morbidity or mortality

      Locally weighted regression (locally estimated scatterplot smoothing) curves were used to smooth temporal trends of observed and expected major morbidity or mortality.
      • Cleveland W.S.
      • Devlin S.J.
      Locally weighted regression: an approach to regression analysis by local fitting.
      The Corrected Akaike Information Criterion was used to determine an optimal smoothing parameter for both models; 68% confidence limits are presented.
      • Hurvich C.M.
      Smoothing parameter selection in nonparametric regression using an improved Akaike Information Criterion.

      Characteristics associated with major morbidity or mortality

      Multivariable models of major morbidity or mortality were developed with and without STS risk score forced into the models; the latter analyses identified residual risk factors not accounted for by the risk score or identified those under- or overvalued in magnitude at our institution.
      • Raza S.
      • Sabik III, J.F.
      • Rajeswaran J.
      • Idrees J.J.
      • Trezzi M.
      • Riaz H.
      • et al.
      Enhancing the value of population-based risk scores for institutional-level use.
      These factors were identified using machine learning for variable selection.
      • Efron B.
      • Tibshirani R.J.
      An Introduction to the Bootstrap.
      ,
      • Rajeswaran J.
      • Blackstone E.H.
      Identifying risk factors: challenges of separating signal from noise.
      In brief, using 1000 bootstrap data sets, variable selection from numerous possible risk factors (Appendix E1) were selected using automated forward stepwise selection. Those identified in at least 50% of models (aggregation step) were included in the final models.

      Missing values

      For multivariable model development, we used multiple imputation
      • Rubin D.B.
      Multiple Imputation for Non-Response in Surveys.
      via a Markov chain Monte Carlo technique to impute missing values for any variables missing <30% of values. Five data sets of imputed missing values were analyzed, and thereafter a final model was estimated by combining these results.
      • Rubin D.B.
      Multiple Imputation for Non-Response in Surveys.

      Time-related analyses

      Time-related survival and aortic valve reoperation for durability of AVR were estimated. Time from operation to death or reoperation was estimated nonparametrically using the Kaplan–Meier estimator and parametrically using a multiphase hazard model.
      • Blackstone E.H.
      • Naftel D.C.
      • Turner Jr., M.E.
      The decomposition of time-varying hazard into phases, each incorporating a separate stream of concomitant information.
      The parametric model was used to resolve a number of phases of instantaneous risk of events (hazard functions) and to estimate shaping parameters. Number of hazard phases and their shaping parameters were completely data-driven. Survival was referenced to the age-race-sex–matched US population based on 2008 census data.

      Results

      Overview

      STS-defined morbidity, operative mortality, and combined major morbidity or mortality occurred similarly to STS-expected outcomes early in the study period, but occurred markedly less than STS-expected outcomes thereafter with the exception of stroke, which remained approximately constant. Operative mortality and morbidity were less than expected across the spectrum of risk in this low-risk population (PROM 0%-4%). There were relatively few intermediate-term reoperations. Long-term survival in this cohort was slightly superior to that of the age-race-sex–matched US population.

      Operative Mortality

      There were 15 operative deaths (0.43%), compared with 55 expected (1.6%), yielding an O:E for mortality of 0.27 (95% CI, 0.14-0.42; Table 2). Operative mortality in 2005 was similar to STS-expected mortality, but it declined to less than half of STS-expected risk by 2010. There were no operative deaths in the last 4 years of the study period. STS-expected operative mortality remained relatively constant over the same time frame (Figure 1, A). Observed operative mortality was lower than expected across all deciles of predicted risk, from STS PROM 0% to 4%, with the most pronounced difference occurring in patients with a higher STS PROM (Figure 2, A).
      Table 2Observed versus expected outcomes after surgical aortic valve replacement in low-risk patients with aortic stenosis, based on contemporary STS benchmarks
      STS outcomen
      Patients with data available.
      Observed, No. (%)Expected, No. (%)Observed/Expected (95% CI)
      Estimated via the bootstrap percentile method using 1000 bootstrap samples.
      P value
      Operative mortality349315 (0.43)55 (1.6)0.27 (0.14-0.42)<.0001
      Permanent stroke349326 (0.74)40 (1.2)0.65 (0.41-0.89).02
      Renal failure347952 (1.5)97 (2.8)0.54 (0.40-0.69)<.0001
      Prolonged ventilation (>24 h)3493149 (4.3)249 (7.1)0.60 (0.50-0.69)<.0001
      Deep sternal wound infection34936 (0.17)8 (0.23)0.75 (0.25-1.4).5
      Reoperation3493116 (3.3)230 (6.6)0.50 (0.42-0.60)<.0001
      Major morbidity or mortality (composite adverse event)3493278 (8.0)449 (13)0.62 (0.55-0.69)<.0001
      Prolonged length of stay (>14 d)3493129 (3.7)165 (4.7)0.78 (0.67-0.91).004
      STS, Society of Thoracic Surgeons; CI, confidence interval.
      Patients with data available.
      Estimated via the bootstrap percentile method using 1000 bootstrap samples.
      Figure thumbnail gr1af
      Figure 1Yearly trend of observed and expected Society of Thoracic Surgeons outcomes. Note that mortality and morbidity that was commensurate with Society of Thoracic Surgeons expected probabilities in 2005 became better through quality improvement initiatives that are duplicatable. Dots represent yearly data points, solid red line enclosed within a 95% confidence band represents a locally estimated scatterplot smoothing fit, and solid blue line represents a locally estimated scatterplot smoothing fit to yearly expected estimates. A, Operative mortality. B, Permanent stroke. C, Renal failure. D, Prolonged ventilation. E, Deep sternal wound infection. F, Reoperation. G, Major morbidity or mortality (composite of preceding). H, Prolonged length of stay.
      Figure thumbnail gr1gh
      Figure 1Yearly trend of observed and expected Society of Thoracic Surgeons outcomes. Note that mortality and morbidity that was commensurate with Society of Thoracic Surgeons expected probabilities in 2005 became better through quality improvement initiatives that are duplicatable. Dots represent yearly data points, solid red line enclosed within a 95% confidence band represents a locally estimated scatterplot smoothing fit, and solid blue line represents a locally estimated scatterplot smoothing fit to yearly expected estimates. A, Operative mortality. B, Permanent stroke. C, Renal failure. D, Prolonged ventilation. E, Deep sternal wound infection. F, Reoperation. G, Major morbidity or mortality (composite of preceding). H, Prolonged length of stay.
      Figure thumbnail gr2
      Figure 2Observed versus expected outcomes. Note that red dots generally fall below the dashed line of identity, indicating that observed occurrence of major morbidity or mortality was better than expected. Red dots represent deciles of observed events, dashed diagonal lines are lines of identity, and blue lines are a hanging distribution of number of patients at each expected value. A, Operative mortality. B, Permanent stroke. C, Renal failure. D, Prolonged ventilation. E, Deep sternal wound infection. F, Reoperation. G, Major morbidity or mortality. H, Prolonged length of stay.

      Morbidity

      There were 26 (0.74%) postoperative strokes, 52 (1.5%) cases of renal failure, 149 (4.3%) instances of prolonged ventilation, 116 (3.3%) early reoperations, and 129 (3.7%) prolonged hospital stays. O:Es were consistently <1.0 (Table 2). Only 6 (0.17%) deep sternal wound infections occurred, limiting meaningful comparisons between observed and expected events. Although fewer strokes occurred than were expected (O:E of 0.65; 95% CI, 0.41-0.89), there was no discernable change in observed stroke rate over the study period, in contrast to the pattern observed for operative mortality and other morbidity (Figure 1, B). With the exception of stroke, observed morbidity declined over the study period while STS-expected outcomes remained relatively constant (Figure 1, C-G). Observed morbidity was generally lower than expected across all deciles of risk with the exception of deep sternal wound infection, where observed numbers were too low to be meaningful (Figure 2, B-F).

      Combined Major Morbidity or Mortality

      Two hundred seventy-eight patients (8%) experienced at least 1 major morbidity or died. Major morbidity or mortality steadily declined over time, with probabilities of 8.6%, 6.7%, and 5.2% in 2006, 2011, and 2016, respectively, while STS-expected major morbidity or mortality remained relatively constant at approximately 12% over that period, yielding an O:E of 0.62 (95% CI, 0.55-0.69; Figure 1, G).

      Risk factors

      Higher grades of mitral valve regurgitation, thicker preoperative intraventricular septum, higher preoperative bilirubin levels, lower creatinine clearance values, history of chronic obstructive pulmonary disease, left anterior descending coronary artery disease ≥70% stenosis, and operations performed in the earlier years of the study were associated with increased odds of major morbidity or mortality (Table 3). Variables either not accounted for or incompletely accounted for by the STS model included higher bilirubin level, history of chronic obstructive pulmonary disease, mitral valve regurgitation, and earlier surgery date (Table E1).
      Table 3Multivariable logistic regression estimates for risk factors associated with STS major morbidity or mortality
      Risk factorCoefficient ± SEP valueReliability, %
      Percent of times variable appeared in 1000 bootstrap models.
      Preoperative echocardiogram
       Mitral valve regurgitation (yes vs no)0.38 ± 0.13.00469
       Larger intraventricular septal thickness
      (Intraventricular septal thickness)2, squared transformation.
      0.15 ± 0.06.0150
      Preoperative laboratory values
       Higher bilirubin
      (Bilirubin)2, squared transformation.
      0.16 ± 0.04.000192
       Lower creatinine clearance
      Ln (creatinine), natural logarithmic transformation.
      −0.62 ± 0.17.000376
      Noncardiac comorbidity
       History of COPD (yes vs no)0.83 ± 0.15<.000199
      Coronary anatomy
       LAD system disease (≥70% stenosis)0.68 ± 0.19.000266
      Experience
       Earlier date of surgery−0.08 ± 0.02<.000193
      SE, Standard error; COPD, chronic obstructive pulmonary disease; LAD, left anterior descending coronary artery.
      Percent of times variable appeared in 1000 bootstrap models.
      (Intraventricular septal thickness)2, squared transformation.
      (Bilirubin)2, squared transformation.
      § Ln (creatinine), natural logarithmic transformation.

      Survival

      There were 253 deaths by the end of follow-up. Parametric survival estimates were 98%, 91%, and 82% at 1, 5, and 9 years, respectively (Figure 3, A). The hazard function for death showed an early decreasing phase, lasting approximately 5.5 months after surgery, before merging into an increasing hazard phase (Figure 3, B). Survival was superior to that of the US age-race-sex–matched population.
      Figure thumbnail gr3
      Figure 3Vital status after surgical aortic valve replacement in low-risk patients. For reference, vital status in an age-race-sex–matched US population is superimposed. Note that after an early phase of elevated risk, survival was greater than (and mortality fell beneath) that expected of the matched US population. A, Survival (inset: mortality). Each symbol represents a death, vertical bars are 68% confidence limits equivalent to ±1 standard error, and solid blue lines represent parametric estimates enclosed within 68% confidence bands. For reference, survival and mortality for the matched US population is shown by the red dash-dot-dash lines. Numbers below the horizontal axis are patients at risk at each listed time. B, Instantaneous risk of death. The solid blue line represents parametric estimates enclosed within a 68% confidence band, equivalent to ±1 standard error. For reference, hazard for the matched US population is shown by the red dash-dot-dash line.

      Reoperation

      A total of 51 aortic valve reoperations for any reason were performed by the end of follow-up. Freedom from reoperation was 99%, 96%, and 93% at 1, 5, and 7 years, respectively (Figure 4, A). The hazard function for reoperation showed a phase of transiently higher risk beginning soon after operation, primarily related to endocarditis, which then merged into one of modestly accelerating risk after approximately 2 years (Figure 4, B). These estimates deviate by about 1% at 7 years because of the competing risk of death (Figure E1).
      Figure thumbnail gr4
      Figure 4Aortic valve reoperation in low-risk patients after surgical aortic valve replacement. A, Freedom from reoperation (inset: time to reoperation). Each symbol represents a reoperation, vertical bars are 68% confidence limits equivalent to ±1 standard error, and solid blue lines represent parametric estimates enclosed within a 68% confidence band. Numbers below the horizontal axis are patients at risk at each listed time. B, Instantaneous risk of reoperation. Solid blue line represents parametric estimates enclosed within a 68% confidence band, equivalent to ±1 standard error.

      Variables Associated With Other Specific STS-Defined Outcomes

      Prolonged ventilation

      Prolonged ventilation occurred in 149 patients (4.3%). Higher New York Heart Association functional class, higher grade of mitral valve disease, higher bilirubin level, coronary artery disease, and earlier surgery date were among the variables associated with increased likelihood of prolonged ventilation over and above risk accounted for by the STS model (Table E2).

      Prolonged length of hospital stay

      Hospital stay was prolonged in 129 patients (3.7%), but occurrence decreased over the study period (Figure 1, H) and was lower than expected (Figure 2, H). Higher mitral valve regurgitation grade, higher bilirubin level, ventricular arrhythmia, and earlier surgery date were associated with increased risk of prolonged length of stay over and above that accounted for by the STS model (Table E3).

      Discussion

      Principal Findings

      For low-risk patients with STS PROM <4%, observed operative mortality and morbidity occurred substantially less often than expected (Figure 5). This held true for the lowest risk patients with STS PROM <1%, as well as for those with PROM of 3% to 4%. Over the decade of the study, major morbidity or mortality declined steadily to less than half of expected, while expected risk remained relatively constant, despite the increasing use of TAVR at our institution. Survival after SAVR was superior to that of the US matched population, and aortic valve reoperations were commensurate with that expected for a large bioprostheses cohort.
      Figure thumbnail gr5
      Figure 5For a low-risk patient with Society of Thoracic Surgeons (STS) predicted risk of mortality (PROM) <4% undergoing a surgical aortic valve replacement (SAVR), across the spectrum of risk, the combined end point of major morbidity or mortality improved over the 11 years of the study period, while expected risk remained stable. For all deciles from 0% to 4%, observed operative mortality (red dots) was superior to STS expected outcomes. That survival in this patient group exceeded that of the reference US population (red dot-dash line) suggests that the benefit of SAVR is durable. These data support recommendations for early surgery, should be used as a benchmark for evolving transcatheter aortic valve replacement (TAVR) technologies, and support the effort to develop agile risk models that more accurately reflect variability in contemporary practice.

      Risk Versus Reality

      Surgical risk models provide an important tool set to aid patients, cardiologists, and surgeons in decision-making regarding the timing of intervention as well as choice of SAVR versus TAVR. National guidelines recommend intervention based on risk categories, and although randomized trial data, which support the guidelines, use STS risk stratification, the guidelines do not recommend a particular risk stratification tool. It has been well demonstrated that risk stratification tools might be effective at stratifying low- versus high-risk patients; however, their accuracy in predicting outcomes for an individual patient is limited and might not be reflective of the magnitude of actual risk.
      • Shahian D.M.
      • Blackstone E.H.
      • Edwards F.H.
      • Grover F.L.
      • Grunkemeier G.L.
      • Naftel D.C.
      • et al.
      Cardiac surgery risk models: a position article.
      In our study, the observed risk was less than expected risk for operative mortality and all categories of STS-defined morbidity. Additionally, this held true across the spectrum of predicted risk, from <1% to 4%. Indeed, the risk of operative mortality was most overestimated for patients at the higher end of the range.
      The discrepancy between expected outcomes and real-world observations has been documented in studies of TAVR as well.
      • Henn M.C.
      • Zajarias A.
      • Quader N.
      • Sintek M.
      • Lasala J.M.
      • Koogler K.
      • et al.
      Observed to expected 30-day mortality as a benchmark for transcatheter aortic valve replacement.
      • Brennan J.M.
      • Holmes D.R.
      • Sherwood M.W.
      • Edwards F.H.
      • Carroll J.D.
      • Grover F.L.
      • et al.
      The association of transcatheter aortic valve replacement availability and hospital aortic valve replacement volume and mortality in the United States.
      • Winkel M.G.
      • Stortecky S.
      • Wenaweser P.
      Transcatheter aortic valve implantation current indications and future directions.
      The Transcatheter Valve Therapy registry
      Society of Thoracic Surgeons
      STS/ACC TVT registry.
      reported an O:E of 1:0 for TAVR, whereas another large study in 2018 by Henn and colleagues
      • Henn M.C.
      • Zajarias A.
      • Quader N.
      • Sintek M.
      • Lasala J.M.
      • Koogler K.
      • et al.
      Observed to expected 30-day mortality as a benchmark for transcatheter aortic valve replacement.
      of early TAVR showed an O:E for 30-day mortality of 0.0 (CI, 0.0-2.02) in low-risk patients and 0.4 (CI, 0.25-0.63; P < .001) for the entire study group.
      These results for SAVR and TAVR suggest that there is substantial variability in clinical outcomes among institutions and operators. It is important to differentiate average outcome probabilities across the nation from those that are achievable in centers of excellence. With emphasis on large trials and multi-institutional registries to drive guidelines and practice using average treatment effect, we must not lose sight of studies that set a high bar for procedure safety and help define best, rather than average, practice.

      Temporal Trend in Outcomes

      We observed improvements in combined major morbidity or mortality over the 12 years of the study, while the expected risk remained stable. It should be noted that by 2005, the study institution had a substantial and diverse valve practice dating back decades. Thus, the results presented should not be interpreted as a learning curve, as observed with TAVR.
      • Suri R.M.
      • Minha S.
      • Alli O.
      • Waksman R.
      • Rihal C.S.
      • Satler L.P.
      • et al.
      Learning curves for transapical transcatheter aortic valve replacement in the PARTNER-I trial: technical performance, success, and safety.
      ,
      • Carroll J.D.
      • Vemulapalli S.
      • Dai D.
      • Matsouaka R.
      • Blackstone E.
      • Edwards F.
      • et al.
      Procedural experience for transcatheter aortic valve replacement and relation to outcomes: the STS/ACC TVT Registry.
      Rather, observed risk in 2005 was similar to STS expected risk. In the following years, a number of iterative changes in practice were made. Cleveland Clinic began producing annual outcomes reports for cardiac surgery available to referring physicians in 1998. By 2004, these were made publicly available and are now published online. By 2008, the Department of Thoracic and Cardiovascular Surgery was reorganized as part of the Heart and Vascular Institute, together with the Department of Cardiovascular Medicine. All STS outcomes, and additional relevant quality, cost, and efficiency data, are reviewed monthly at a quality-centric staff meeting at which surgeon-specific data are transparently available. This is in addition to regular morbidity and mortality review of major events and root-cause analyses where appropriate.
      This frequent, rigorous discussion of reported outcomes generated a number of iterative care interventions: preoperative screening and imaging protocols were standardized according to case type rather than surgeon. Intraoperative pump and cardioplegia protocols were refined. del Nido cardioplegia was introduced as routine for isolated valve cases, backed by an internal randomized trial,
      • Mick S.L.
      • Robich M.P.
      • Houghtaling P.L.
      • Gillinov A.M.
      • Soltesz E.G.
      • Johnston D.R.
      • et al.
      del Nido versus Buckberg cardioplegia in adult isolated valve surgery.
      limiting the use of retrograde coronary sinus cannulation for isolated valves. A pre–chest-closure checklist was developed to reduce postoperative bleeding.
      • Loor G.
      • Vivacqua A.
      • Sabik III, J.F.
      • Li L.
      • Hixson E.D.
      • Blackstone E.H.
      • et al.
      Process improvement in cardiac surgery: development and implementation of a reoperation for bleeding checklist.
      Early extubation protocols were developed and refined. Separate postoperative care teams were developed for intensive care and step-down units. A Cardiac Medical Emergency Team was instituted to respond to early patient decompensation events, with ability of family members to activate the team.
      Increase in transparency of data for all outcomes, including cost and efficiency, rather than a pure mortality focus, has led to consistency and standardization of care. More recently, our focus has been on improving the “leading indicators” of patient safety by emphasizing teamwork and a culture of safety in the operating room.
      In short, the improvement in outcomes likely reflects an intentional process of continuous improvement with frequent, transparent outcome review. It is also possible that the trend toward earlier surgery, changes in patient selection, and the growth of TAVR played a role, although a change in expected risk would be anticipated if patient selection were a primary driver. That observed outcomes at a single center improved to this extent compared with expected outcomes provides support for the concept of a more agile “real-time” risk tool, such as that being developed by the American Association for Thoracic Surgery Quality Gateway.
      • Blackstone E.H.
      • Swain J.
      • McCardle K.
      • Adams D.H.
      A comprehensive American Association for Thoracic Surgery quality program for the 21st century.
      Continuously updated risk models that reflect the state-of-the-art in quality outcomes are necessary if we are to provide patients with the best recommendations for appropriate valve therapy.

      Stroke

      Although observed occurrence of stroke was less than expected, we did not experience a significant decrease over time. This is despite a concerted focus within the institution on stroke prevention, including increasing use of preoperative imaging to identify aortic calcium. It is possible that these efforts might be more effective in higher-risk populations or resulted in aortic interventions excluded in the data set. These results highlight, however, that stroke remains an important issue in aortic valve intervention even in low-risk patients, and the long-term effect on patients should not be discounted despite its low occurrence.

      Generalizability

      These results differ from other recent series of low-risk SAVR. L.E. Johnston and colleagues observed outcomes similar to STS-expected outcomes across 18 institutions in a statewide database.
      • Johnston L.E.
      • Downs E.A.
      • Hawkins R.B.
      • Quader M.A.
      • Speir A.M.
      • Rich J.B.
      • et al.
      Outcomes for low-risk surgical aortic valve replacement: a benchmark for aortic valve technology.
      That results at a single high-volume valve center differ from large clinical trials or multi-institutional databases should not be surprising in light of the large body of data supporting a volume–outcome relationship.
      • Hirji S.A.
      • McCarthy E.
      • Kim D.
      • McGurk S.
      • Ejiofor J.
      • Ramirez-Del Val F.
      • et al.
      Relationship between hospital surgical aortic valve replacement volume and transcatheter aortic valve replacement outcomes.
      Our intention is not, however, to suggest that volume is the only contributor to outcome. It should be evident from the previous discussion that the multiple improvements made in the care of patients with aortic valve disease over the study period are not specific to individual surgeon or surgeon volume. In our study, 33 surgeons treated patients over 12 years, with significant turnover from more experienced to less experienced surgeons.
      Conversely, we should not expect the interventions that positively affect risk to be evenly distributed among institutions contributing to the STS risk model. A distinction should be made between generalizable results—meaning achievable at most institutions performing SAVR—and repeatable results—meaning achievable at institutions willing to make alterations to the organization, practice patterns, and institutional culture aimed at rigorous incremental reductions in morbidity and mortality. The authors suggest that most of the interventions made in our institution can be, and many have been, adopted by other centers.
      It has been shown for both SAVR and TAVR at multiple institutions that very low morbidity and mortality is achievable.
      • Ram E.
      • Amunts S.
      • Zuroff E.
      • Peled Y.
      • Kogan A.
      • Raanani E.
      • et al.
      Outcomes of isolated surgical aortic valve replacement in the era of transcatheter aortic valve implantation.
      ,
      • Forrest J.K.
      • Kaple R.K.
      • Tang G.H.L.
      • Yakubov S.J.
      • Nazif T.M.
      • Williams M.R.
      • et al.
      Three generations of self-expanding transcatheter aortic valves: a report from the STS/ACC TVT Registry.
      As payment for surgical procedures evolves and payors consider regionalization of care, these data provide a lens by which to examine the application of TAVR and SAVR across the spectrum of practice. Patients, in particular, deserve to understand not only what constitutes an average outcome, but what is potentially achievable in minimizing their risk.

      Durable Outcomes

      Decision-making in younger, healthier patients who make up much of this low-risk cohort, in which mean age was 64 years, depends not only on the early success of the operation but on intermediate- and long-term outcomes, survival most importantly. That survival in this patient group exceeded that of the US-matched reference population suggests that the benefit of SAVR is durable. This does not imply that SAVR somehow improves survival; as more data accumulate to support surgery in asymptomatic patients with aortic valve disease based on imaging characteristics such as strain, these findings lend support to the decision for early intervention, in which long-term success is critical. Reoperation was uncommon—<7% at 7 years—even in this young patient group. These data will form an important comparison group as information accrues regarding TAVR durability in young patients.

      Limitations

      This is a relatively large single-institution report. Intraoperative conduct and technical details, such as valve choice and operative technique, play a role in clinical decision-making and might affect outcomes; however, these nuances are not adequately reflected in the data. Our inferences are limited by effective sample size for morbidity and operative mortality, which is proportional to the number of events observed, not the number of patients. These results might not be generalizable across institutions or geographic area, although they represent a valve practice with a national referral base.

      Conclusions

      The current STS risk model, although effective at stratifying lower versus higher risk for SAVR, substantially overestimates actual SAVR risk in a high-volume center. Observed outcomes improved over time while expected risk remained static, supporting individual institutions' efforts to make continuous iterative improvements in practice even for “mature” operations such as SAVR. In addition, these results support the effort to develop agile risk models that more accurately reflect the variability in contemporary practice. SAVR in low-risk patients resulted in survival superior to that of US 2008 census age-race-sex–matched controls, supporting early surgery recommendations.

      Webcast

      Conflict of Interest Statement

      The authors reported no conflicts of interest.
      The Journal policy requires editors and reviewers to disclose conflicts of interest and to decline handling or reviewing manuscripts for which they may have a conflict of interest. The editors and reviewers of this article have no conflicts of interest.

      Supplementary Data

      • Video 1

        Results from current practice show that with gradual enhancements in technique and planning that have occurred over the decade of this study, low-risk patients undergoing surgical aortic valve replacement achieved outcomes superior to those expected by Society of Thoracic Surgeons risk models. Postoperative survival was also better than that of the US matched population. These results can serve as an indication for what is widely achievable and a benchmark for evolving technologies. Video available at: https://www.jtcvs.org/article/S0022-5223(21)00571-7/fulltext.

      Appendix E1

      Table E1Additional or refined patient variables associated with higher likelihood of STS major morbidity or mortality
      VariableCoefficient ± SEP valueReliability, %
      Percent of times variable appeared in 1000 bootstrap models.
      Higher STS risk score
      Logit (predicted STS major morbidity or mortality score), logit transformation.
      1.2 ± 0.16<.0001
      Forced into all 1000 models.
      Younger age
      (Age/50)2, inverse transformation.
      0.18 ± 0.10.0858
      Higher bilirubin level
      (Bilirubin)2, squared transformation.
      0.13 ± 0.04.00293
      COPD0.51 ± 0.16.00188
      Mitral valve regurgitation0.32 ± 0.13.0263
      Earlier date of surgery−0.10 ± 0.02<.000197
      SE, Standard error; STS, Society of Thoracic Surgeons; COPD, chronic obstructive pulmonary disease.
      Percent of times variable appeared in 1000 bootstrap models.
      Logit (predicted STS major morbidity or mortality score), logit transformation.
      Forced into all 1000 models.
      § (Age/50)2, inverse transformation.
      (Bilirubin)2, squared transformation.
      Table E2Residual or refined patient variables associated with higher likelihood of STS prolonged ventilation
      VariableCoefficient ± SEP valueReliability, %
      Percent of times variable appeared in 1000 bootstrap models.
      Higher STS risk score
      Logit (predicted STS prolonged ventilation score), logit transformation.
      1.3 ± 0.20<.0001
      Forced into all 1000 models.
      Higher NYHA functional class0.16 ± 0.11.1677
      COPD0.44 ± 0.21.0449
      Higher grade of mitral valve regurgitation0.23 ± 0.10.0261
      Mitral valve stenosis0.89 ± 0.32.00568
      Higher bilirubin level
      (Bilirubin)2, squared transformation.
      0.13 ± 0.04.00493
      LCx system disease <50% stenosis−0.97 ± 0.37.00848
      RCA system disease ≥50% stenosis0.72 ± 0.28.0154
      Larger valve area
      (Area)2, squared transformation.
      0.06 ± 0.05.260
      Earlier date of surgery−0.11 ± 0.03<.000190
      SE, Standard error; STS, Society of Thoracic Surgeons; NYHA, New York Heart Association; COPD, chronic obstructive pulmonary disease; LCx, left circumflex coronary artery; RCA, right coronary artery.
      Percent of times variable appeared in 1000 bootstrap models.
      Logit (predicted STS prolonged ventilation score), logit transformation.
      Forced into all 1000 models.
      § (Bilirubin)2, squared transformation.
      (Area)2, squared transformation.
      Table E3Residual or refined patient variables associated with higher likelihood of STS prolonged hospital stay
      VariableCoefficient ± SEP valueReliability, %
      Percent of times variable appeared in 1000 bootstrap models.
      Higher STS risk score
      Logit (predicted STS prolonged hospital stay score), logit transformation.
      1.3 ± 0.18<.0001
      Forced into all 1000 models.
      Mitral valve regurgitation0.56 ± 0.19.00386
      Higher bilirubin level
      (Bilirubin)2, squared transformation.
      0.41 ± 0.17.0148
      Ventricular arrhythmia0.87 ± 0.34.0156
      RCA system disease ≤50% stenosis0.54 ± 0.24.0250
      Earlier date of surgery
      (1/date of surgery), inverse transformation.
      −0.09 ± 0.03.00484
      SE, Standard error; STS, Society of Thoracic Surgeons; RCA, right coronary artery.
      Percent of times variable appeared in 1000 bootstrap models.
      Logit (predicted STS prolonged hospital stay score), logit transformation.
      Forced into all 1000 models.
      § (Bilirubin)2, squared transformation.
      (1/date of surgery), inverse transformation.
      Figure thumbnail fx5
      Figure E1Competing risk analysis of reoperation with death as a competing risk. The usual probability-of-event curve (KM), the cumulative incidence that estimates the lower likelihood of a reoperation for patients because of the competing risk of death (CIF), and the conditional probability (CP) that quantifies what reoperation is expected to be if the competing risk of death were removed (an estimate of prosthesis durability).

      Appendix E1. Variables Considered in the Analyses

      Demographics

      Age (years), sex, race (white, black, other), height (cm), weight (kg), body mass index (kg/m2), and body surface area (m2).

      Symptoms

      New York Heart Association functional class (I-IV), Canadian Angina Class (I-III), and previous myocardial infarction.

      Ventricular Function

      Left ventricular ejection fraction (%), left ventricular systolic dysfunction (%), and fractional shortening (%).

      Valve Pathology

      Aortic valve

      Regurgitation, regurgitation degree/severity (none, mild, moderate, severe), stenosis, and stenosis degree/severity (none, mild, moderate, severe).

      Mitral valve

      Regurgitation, regurgitation degree/severity (none, mild, moderate, severe), stenosis, and stenosis degree/severity (none, mild, moderate, severe).

      Tricuspid valve

      Regurgitation, regurgitation degree/severity (none, mild, moderate, severe), stenosis, and stenosis degree/severity (none, mild, moderate, severe).

      Pulmonary valve

      Regurgitation, regurgitation degree/severity (none, mild, moderate, severe), stenosis, and stenosis degree/severity (none, mild, moderate, severe).

      Hemodynamics

      Valve area (cm2), mean gradient (mm Hg), and peak gradient (mm Hg).

      Left Ventricular Structure

      Inner diastolic diameter (cm), inner diastolic volume (mL), inner diastolic volume index (mL/m2), inner systolic diameter (mm), inner systolic volume (mL), inner systolic volume index (mL/m2), and relative wall thickness (mm).

      Left Ventricular Mass

      Posterior wall thickness (mm), intraventricular septal thickness (mm), mass (g), and mass index (g/m2).

      Left Atrial Dimensions

      Diameter (cm), volume (mL), and volume index (mL/m2).

      Cardiac Comorbidities

      Atrial fibrillation or flutter, complete heart block/pacer, ventricular arrhythmia, history and number of previous cardiovascular operations, heart failure, and endocarditis.

      Noncardiac Comorbidities

      Peripheral arterial disease, hypertension (mm Hg), diabetes (treated, insulin-dependent, non-insulin-dependent), history of chronic obstructive pulmonary disease, history of smoking, renal dialysis, stroke, bilirubin (mg/dL), creatinine (mg/dL), creatinine clearance (Cockcroft-Gault; mL/min), estimated glomerular filtration rate (modification of diet in renal disease; mL/min/1.73 m2), blood urea nitrogen (mg/dL), total cholesterol (mg/dL), high-density lipoprotein cholesterol (mg/dL), low-density lipoprotein cholesterol (mg/dL), triglycerides (mg/dL), and hematocrit (%).

      Coronary Anatomy

      Number of systems diseased (≥50% stenosis), left main trunk stenosis (≥50%, ≥70%), left anterior descending stenosis (≥50%, ≥70%), left circumflex stenosis (≥50%, ≥70%), and right coronary artery stenosis (≥50%, ≥70%).

      Procedural Characteristics

      Surgical approach (full incision, less invasive), aortic clamp time (min), date of surgery (fractional years from 2005).

      References

        • Nishimura R.A.
        • Otto C.M.
        • Bonow R.O.
        • Carabello B.A.
        • Erwin III, J.P.
        • Fleisher L.A.
        • et al.
        2017 AHA/ACC focused update of the 2014 AHA/ACC guideline for the management of patients with valvular heart disease: a report of the American College of Cardiology/American Heart Association task force on clinical practice guidelines.
        Circulation. 2017; 135: e1159-e1195
        • Mack M.J.
        • Leon M.B.
        • Thourani V.H.
        • Makkar R.
        • Kodali S.K.
        • Russo M.
        • et al.
        Transcatheter aortic-valve replacement with a balloon-expandable valve in low-risk patients.
        N Engl J Med. 2019; 380: 1695-1705
        • Popma J.J.
        • Deeb G.M.
        • Yakubov S.J.
        • Mumtaz M.
        • Gada H.
        • O'Hair D.
        • et al.
        Transcatheter aortic-valve replacement with a self-expanding valve in low-risk patients.
        N Engl J Med. 2019; 380: 1706-1715
        • Chiam P.T.
        • Ruiz C.E.
        Percutaneous transcatheter aortic valve implantation: evolution of the technology.
        Am Heart J. 2009; 157: 229-242
        • Zamorano J.L.
        • Badano L.P.
        • Bruce C.
        • Chan K.L.
        • Gonçalves A.
        • Hahn R.T.
        • et al.
        EAE/ASE recommendations for the use of echocardiography in new transcatheter interventions for valvular heart disease.
        Eur Heart J. 2011; 32: 2189-2214
        • Vahl T.P.
        • Kodali S.K.
        • Leon M.B.
        Transcatheter aortic valve replacement 2016: a modern-day “Through the Looking-Glass” adventure.
        J Am Coll Cardiol. 2016; 67: 1472-1487
        • Nilsson J.
        • Algotsson L.
        • Höglund P.
        • Lührs C.
        • Brandt J.
        Early mortality in coronary bypass surgery: the EuroSCORE versus the Society of Thoracic Surgeons risk algorithm.
        Ann Thorac Surg. 2004; 77: 1235-1239
        • O'Brien S.M.
        • Shahian D.M.
        • Filardo G.
        • Ferraris V.A.
        • Haan C.K.
        • Rich J.B.
        • et al.
        The Society of Thoracic Surgeons 2008 cardiac surgery risk models: part 2–isolated valve surgery.
        Ann Thorac Surg. 2009; 88: S23-S42
        • Shahian D.M.
        • O'Brien S.M.
        • Filardo G.
        • Ferraris V.A.
        • Haan C.K.
        • Rich J.B.
        • et al.
        The Society of Thoracic Surgeons 2008 cardiac surgery risk models: part 3–valve plus coronary artery bypass grafting surgery.
        Ann Thorac Surg. 2009; 88: S43-S62
        • Efron B.
        • Tibshirani R.J.
        An Introduction to the Bootstrap.
        Chapman and Hall/CRC, New York1994
        • Cleveland W.S.
        • Devlin S.J.
        Locally weighted regression: an approach to regression analysis by local fitting.
        J Am Stat Assoc. 1988; 83: 596-610
        • Hurvich C.M.
        Smoothing parameter selection in nonparametric regression using an improved Akaike Information Criterion.
        J R Stat Soc Series B Stat Methodol. 1998; 60: 271-293
        • Raza S.
        • Sabik III, J.F.
        • Rajeswaran J.
        • Idrees J.J.
        • Trezzi M.
        • Riaz H.
        • et al.
        Enhancing the value of population-based risk scores for institutional-level use.
        Ann Thorac Surg. 2016; 102: 70-77
        • Rajeswaran J.
        • Blackstone E.H.
        Identifying risk factors: challenges of separating signal from noise.
        J Thorac Cardiovasc Surg. 2017; 153: 1136-1138
        • Rubin D.B.
        Multiple Imputation for Non-Response in Surveys.
        Wiley, New York1987
        • Blackstone E.H.
        • Naftel D.C.
        • Turner Jr., M.E.
        The decomposition of time-varying hazard into phases, each incorporating a separate stream of concomitant information.
        J Am Stat Assoc. 1986; 81: 615-624
        • Shahian D.M.
        • Blackstone E.H.
        • Edwards F.H.
        • Grover F.L.
        • Grunkemeier G.L.
        • Naftel D.C.
        • et al.
        Cardiac surgery risk models: a position article.
        Ann Thorac Surg. 2004; 78: 1868-1877
        • Henn M.C.
        • Zajarias A.
        • Quader N.
        • Sintek M.
        • Lasala J.M.
        • Koogler K.
        • et al.
        Observed to expected 30-day mortality as a benchmark for transcatheter aortic valve replacement.
        J Thorac Cardiovasc Surg. 2019; 157: 874-882.e8
        • Brennan J.M.
        • Holmes D.R.
        • Sherwood M.W.
        • Edwards F.H.
        • Carroll J.D.
        • Grover F.L.
        • et al.
        The association of transcatheter aortic valve replacement availability and hospital aortic valve replacement volume and mortality in the United States.
        Ann Thorac Surg. 2014; 98: 2016-2022
        • Winkel M.G.
        • Stortecky S.
        • Wenaweser P.
        Transcatheter aortic valve implantation current indications and future directions.
        Front Cardiovasc Med. 2019; 6: 179
        • Society of Thoracic Surgeons
        STS/ACC TVT registry.
        (Available at:) (Accessed April 15, 2020)
        • Suri R.M.
        • Minha S.
        • Alli O.
        • Waksman R.
        • Rihal C.S.
        • Satler L.P.
        • et al.
        Learning curves for transapical transcatheter aortic valve replacement in the PARTNER-I trial: technical performance, success, and safety.
        J Thorac Cardiovasc Surg. 2016; 152: 773-780
        • Carroll J.D.
        • Vemulapalli S.
        • Dai D.
        • Matsouaka R.
        • Blackstone E.
        • Edwards F.
        • et al.
        Procedural experience for transcatheter aortic valve replacement and relation to outcomes: the STS/ACC TVT Registry.
        J Am Coll Cardiol. 2017; 70: 29-41
        • Mick S.L.
        • Robich M.P.
        • Houghtaling P.L.
        • Gillinov A.M.
        • Soltesz E.G.
        • Johnston D.R.
        • et al.
        del Nido versus Buckberg cardioplegia in adult isolated valve surgery.
        J Thorac Cardiovasc Surg. 2015; 149 (discussion: 34-6): 626-634
        • Loor G.
        • Vivacqua A.
        • Sabik III, J.F.
        • Li L.
        • Hixson E.D.
        • Blackstone E.H.
        • et al.
        Process improvement in cardiac surgery: development and implementation of a reoperation for bleeding checklist.
        J Thorac Cardiovasc Surg. 2013; 146: 1028-1032
        • Blackstone E.H.
        • Swain J.
        • McCardle K.
        • Adams D.H.
        A comprehensive American Association for Thoracic Surgery quality program for the 21st century.
        J Thorac Cardiovasc Surg. 2019; 158: 1120-1126
        • Johnston L.E.
        • Downs E.A.
        • Hawkins R.B.
        • Quader M.A.
        • Speir A.M.
        • Rich J.B.
        • et al.
        Outcomes for low-risk surgical aortic valve replacement: a benchmark for aortic valve technology.
        Ann Thorac Surg. 2017; 104: 1282-1288
        • Hirji S.A.
        • McCarthy E.
        • Kim D.
        • McGurk S.
        • Ejiofor J.
        • Ramirez-Del Val F.
        • et al.
        Relationship between hospital surgical aortic valve replacement volume and transcatheter aortic valve replacement outcomes.
        JACC Cardiovasc Interv. 2020; 13: 335-343
        • Ram E.
        • Amunts S.
        • Zuroff E.
        • Peled Y.
        • Kogan A.
        • Raanani E.
        • et al.
        Outcomes of isolated surgical aortic valve replacement in the era of transcatheter aortic valve implantation.
        J Card Surg. 2020; 35: 1452-1457
        • Forrest J.K.
        • Kaple R.K.
        • Tang G.H.L.
        • Yakubov S.J.
        • Nazif T.M.
        • Williams M.R.
        • et al.
        Three generations of self-expanding transcatheter aortic valves: a report from the STS/ACC TVT Registry.
        JACC Cardiovasc Interv. 2020; 13: 170-179

      Linked Article

      • Commentary: The reality of The Society of Thoracic Surgeons risk calculators at high volume centers
        The Journal of Thoracic and Cardiovascular SurgeryVol. 165Issue 2
        • Preview
          In a large single-center retrospective series, Johnston and colleagues1 report outcomes of low-risk surgical aortic valve replacement (SAVR) in 3474 patients between 2005 and 2017. The authors used Society of Thoracic Surgeons (STS) risk models and found observed-to-expected ratios of <1 for mortality (0.27; 95% confidence interval [CI], 0.14-0.42), stroke (0.65; 95% CI, 0.41-0.89), and reoperation (0.50; 95% CI, 0.42-0.60).
        • Full-Text
        • PDF
      • Commentary: Lake Wobegon guidelines reach Lake Erie
        The Journal of Thoracic and Cardiovascular SurgeryVol. 165Issue 2
        • Preview
          Johnstone and colleagues1 are justifiably proud of their outstanding results in low-risk surgical aortic valve replacement (SAVR). As we've come to expect from the Cleveland Clinic, the statistical analysis is probing and impeccable. They conclude that the Society of Thoracic Surgeons risk model “substantially overestimates actual SAVR risk in a high-volume center.” Their stated goal is to provide “a benchmark for transcatheter AVR” that links to their introductory sentence about guidelines for treatment of low-risk patients with aortic stenosis.
        • Full-Text
        • PDF
      • Commentary: Surgery for low-risk aortic valve replacement: The road to extinction
        The Journal of Thoracic and Cardiovascular SurgeryVol. 165Issue 2
        • Preview
          Aortic valve stenosis remains the most common valvular disease in elderly patients1 in developed countries. For several decades, surgical aortic valve replacement (SAVR) has been the only treatment for the more severe forms. However, during the past decade, new trials have changed the treatment landscape and recent evidence2,3 suggests that even low-risk patients with aortic valve stenosis can be treated by transcatheter aortic valve replacement (TAVR) after heart team discussion.
        • Full-Text
        • PDF