If you don't remember your password, you can reset it by entering your email address and clicking the Reset Password button. You will then receive an email that contains a secure link for resetting your password
If the address matches a valid account an email will be sent to __email__ with instructions for resetting your password
Address for reprints: Douglas R. Johnston, MD, Department of Thoracic and Cardiovascular Surgery, Cleveland Clinic, 9500 Euclid Ave/Desk J4-1, Cleveland, OH 44195.
Department of Thoracic and Cardiovascular Surgery, Heart, Vascular, and Thoracic Institute, Cleveland Clinic, Cleveland, OhioThe Aortic Valve Center, Heart, Vascular, and Thoracic Institute, Cleveland Clinic, Cleveland, Ohio
Department of Thoracic and Cardiovascular Surgery, Heart, Vascular, and Thoracic Institute, Cleveland Clinic, Cleveland, OhioThe Aortic Valve Center, Heart, Vascular, and Thoracic Institute, Cleveland Clinic, Cleveland, Ohio
Department of Thoracic and Cardiovascular Surgery, Heart, Vascular, and Thoracic Institute, Cleveland Clinic, Cleveland, OhioThe Aortic Valve Center, Heart, Vascular, and Thoracic Institute, Cleveland Clinic, Cleveland, Ohio
Department of Thoracic and Cardiovascular Surgery, Heart, Vascular, and Thoracic Institute, Cleveland Clinic, Cleveland, OhioThe Aortic Valve Center, Heart, Vascular, and Thoracic Institute, Cleveland Clinic, Cleveland, OhioDepartment of Quantitative Health Sciences, Lerner Research Institute, Cleveland Clinic, Cleveland, Ohio
Department of Thoracic and Cardiovascular Surgery, Heart, Vascular, and Thoracic Institute, Cleveland Clinic, Cleveland, OhioThe Aortic Valve Center, Heart, Vascular, and Thoracic Institute, Cleveland Clinic, Cleveland, Ohio
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.
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.
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.
—allow more accurate decision-making regarding timing of TAVR in asymptomatic patients, combining to accommodate a rapidly shifting landscape in aortic valve disease.
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.
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)
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.
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
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.
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.
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
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.
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.
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
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 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 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.
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
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 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.
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 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 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.
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.
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.
The association of transcatheter aortic valve replacement availability and hospital aortic valve replacement volume and mortality in the United States.
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.
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,
limiting the use of retrograde coronary sinus cannulation for isolated valves. A pre–chest-closure checklist was developed to reduce postoperative bleeding.
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.
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.
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.
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.
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.
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.
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.
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
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.
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).
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.
The association of transcatheter aortic valve replacement availability and hospital aortic valve replacement volume and mortality in the United States.
Funding provided by the Advanced Heart Valve Therapy Fund, the Delos M. Cosgrove M.D. Chair for Heart Disease Research, and the Donna and Ken Lewis Endowed Chair in Cardiothoracic Surgery.
Cleveland Clinic Aortic Valve Center Collaborators: Mona Kakavand, MD, A. Marc Gillinov, MD, Samir Kapadia, MD, Milind Y. Desai, MD, Daniel Burns, MD, MPhil, Patrick R. Vargo, MD, Shinya Unai, MD, Gösta B. Pettersson, MD, PhD, Aaron Weiss, MD, Haytham Elgharably, MD, Rishi Puri, MD, PhD, Grant W. Reed, MD, Zoran B. Popovic, MD, PhD, Wael Jaber, MD, Suma A. Thomas, MD, MBA, Faisal G. Bakaeen, MD, Tara Karamlou, MD, Hani Najm, MD, Brian Griffin, MD, Amar Krishnaswamy, MD, Kenneth R. McCurry, MD, L. Leonardo Rodriguez, MD, Nicholas G. Smedira, MD, MBA, Michael Zhen-Yu Tong, MD, MBA, Per Wierup, MD, PhD, and James Yun, MD.
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).
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.
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.
00571-7&id=fx1.jpg){kind=link}
00571-7&id=fx2.jpg){kind=link}
00571-7&id=gr1af.jpg){kind=link}
00571-7&id=gr1gh.jpg){kind=link}
00571-7&id=gr2.jpg){kind=link}
00571-7&id=gr3.jpg){kind=link}
00571-7&id=gr4.jpg){kind=link}
00571-7&id=gr5.jpg){kind=link}
00571-7&id=fx3.jpg){kind=link}
00571-7&id=fx5.jpg){kind=link}