Estimating the risk of valvular procedures

INTRODUCTION — A variety of factors affect mortality and morbidity associated with surgical or transcatheter interventions for valvular heart disease. These factors can be grouped as patient related, procedure related, and operator and medical center related [1-5]. An expert consensus statement from several professional societies recommended evaluation and treatment at a primary or comprehensive valve center to improve patient outcomes, with choice of center based on the patient’s risk and planned procedure [6].

The most effective way of stratifying operative risk related to patient factors is by using one of the available validated risk stratification models. The risk models are helpful in the evaluation and counseling of individual patients and planning perioperative care but are subject to significant limitations. Models have been developed to estimate risk of in-hospital mortality and complications associated with valve surgery with or without coronary artery bypass graft surgery (CABG).

Risk stratification models for valvular surgery and transcatheter procedures will be reviewed here [7-12]. Other considerations in preoperative assessment of potential cardiac surgery candidates are discussed separately. (See "Operative mortality after coronary artery bypass graft surgery" and "Indications for valve replacement for high gradient aortic stenosis in adults" and "Natural history and management of chronic aortic regurgitation in adults" and "Surgical and investigational approaches to management of mitral stenosis" and "Surgical procedures for severe chronic mitral regurgitation".)

LIMITATIONS — Risk stratification models are subject to limitations that may reduce their prognostic value. The models lag behind current outcomes as new technologies and clinical practice evolve. Society of Thoracic Surgeons (STS) and EuroSCORE models have been updated infrequently in processes extending over years. The models estimate risk for only specified procedures (eg, the 2018 STS risk model does not include estimates for concomitant aortic and mitral valve surgery [13]). These include dependence upon incomplete or uncertain clinical variables. The models may not be generalizable to patient populations different from the ones in which they were formulated, and they require revision over time as surgical and perioperative care evolve and as the recognition and impact of clinical variables change.

In addition, current risk stratification models do not include several variables that may be clinically important. For example, a severely calcified ascending aorta (also known as porcelain aorta, a risk factor for patients undergoing cardiac surgery) and severe mitral annular calcification (a risk factor in patients undergoing mitral surgery (see "Management and prognosis of mitral annular calcification", section on 'Mitral valve surgery')) are each associated with high operative morbidity and mortality but are not included in standard risk models [14-18]. Other clinically important variables that are not included in risk models are difficult to measure, such as frailty, exercise capacity, and nutritional status, as well as other hemodynamic factors such as pulmonary pressures and right ventricular dysfunction. These additional factors are especially important when considering the choice between a surgical or transcatheter valve procedure. Some of these factors may be the determining factor in this decision for specific patients. (See "Choice of intervention for severe calcific aortic stenosis".)

In addition, risk models for cardiac surgery do not necessarily indicate the risk of newer transcatheter procedures, such as transcatheter aortic valve implantation, that increasingly are effective alternates to conventional surgery [19]. Current guidelines recommend evaluation of patients by a specialized heart valve clinic when valve intervention is needed so that risks and benefits of a surgical versus transcatheter procedure can be discussed by a multidisciplinary team [20,21].

RISK STRATIFICATION MODELS

Commonly used models — Risk stratification models for hospital mortality following cardiac surgery include the following calculators:

●The STS risk model is the most widely used algorithm and it is continuously updated. The 2018 Society of Thoracic Surgeons (STS) risk models for coronary artery bypass graft surgery (CABG), valve surgery, or both were developed and validated based on data on 774,881 patients operated on during 2002 to 2006 [13,22]. The STS registry includes data from nearly 90 percent of cardiac surgery providers in the United States [22]. The models provide estimates of risk of mortality as well as several nonfatal complications such as stroke, renal failure, and prolonged ventilation. A calculator is available online (www.sts.org/resources/risk-calculator).

●The first EuroSCORE model was developed based upon data from 1995 to stratify risk of in-hospital mortality [23]. This model has been replaced by the 2011 EuroSCORE II model, which was developed based on data from 22,381 patients in 43 countries operated on during May to July 2010 [24]. Overall mortality was 3.9 percent, which is lower than would have been predicted by old risk models (additive predicted 5.8 percent and logistic predicted 7.6 percent). A calculator for the EuroSCORE II is available online (euroscore.org).

●The older Ambler risk model for aortic and/or mitral valve surgery with or without concomitant CABG was developed from data from the Society of Cardiothoracic Surgeons of Great Britain and Ireland on 32,839 patients who underwent surgery during 1995 to 2003 [25].

Use — Accurate perioperative risk estimation is required to provide appropriate recommendations and informed consent for individual patients. Comparison of treatment options such as aortic valve surgery versus transcatheter aortic valve implantation requires accurate estimation of expected outcomes. However, the available models have limited applicability since the data on which they are based lag behind rapidly evolving clinical practice. When considering risk estimates from models, the heart team should also consider its own local data for valve surgery and transcatheter valve therapy.

The prognostic value of the risk models depends upon the population to which the models are applied. Several studies had found that the first version of the EuroSCORE overestimated mortality for various cardiac surgery populations [26-28], including patients undergoing mitral valve surgery [29] and patients undergoing high-risk aortic valve replacement [30-34].

In patients undergoing aortic valve replacement, the STS model may provide more accurate perioperative risk estimation than the first version of the EuroSCORE [31,33] or the Ambler risk score [33]. In various studies, the first version of the EuroSCORE often overestimated perioperative risk, while the STS model has been observed to under- or overestimate early postoperative mortality [33,35]. The original EuroSCORE and STS score also correlate with long-term postoperative mortality risk, though the relative strengths of these associations vary among series, and other clinical factors (such as presence of renal failure) may have greater predictive value in some groups [33,35]. EuroSCORE II performed similarly to the STS score in a study of patients undergoing mitral valve surgery [36] and had discriminatory power similar to STS score in a study of patients undergoing isolated aortic valve surgery [37].

Since the 2018 STS and EuroSCORE II risk models provide risk stratification based upon large contemporary databases with adjustment for multiple clinical variables, we recommend their use to estimate operative risk in patients in settings similar to those represented by their respective databases [38].

Clinical implications — Risk scores provide a starting point for estimating the surgical mortality and morbidity of valve surgery in an individual patient. This risk estimate may then be adjusted upward or downward depending on additional clinical, anatomic, or hemodynamic factors that are not included in the risk models. Risk estimates are helpful in balancing the risk-benefit ratio of surgical intervention, in optimizing timing of intervention, and for choosing between different interventions (eg, surgical versus transcatheter). In addition, knowledge of specific components of the risk score may suggest ways to reduce risk in an individual patient; for example, performing percutaneous coronary intervention before planned valve surgery, stabilizing the patient with advanced therapies to allow elective (rather than emergency) valve surgery, or optimizing heart failure therapy and hemodynamics prior to intervention.

An additional limitation is that the risk models don’t account for potential benefits and risks of concomitant procedures, such as coronary artery bypass graft surgery performed at the time of valve surgery or percutaneous coronary intervention performed at the time of transcatheter aortic valve implantation. Other procedures that may be performed at the time of valve surgery include surgical left atrial appendage occlusion and surgical atrial fibrillation ablation.

In treating younger patients, it is important to consider risk and plans for future interventions (eg, valve in transcatheter heart valve or valve in surgical bioprosthetic valve), which should inform selection of the initial intervention. These issues are particularly important for patients with size or anatomy variations (eg, lower coronary arteries, small sinuses, small aortic annuli, smaller prostheses).

SOCIETY GUIDELINE LINKS — Links to society and government-sponsored guidelines from selected countries and regions around the world are provided separately. (See "Society guideline links: Cardiac valve disease".)

SUMMARY

●Models have been developed to estimate the risk of in-hospital mortality and complications associated with valve surgery with or without coronary artery bypass graft surgery (CABG). (See 'Introduction' above.)

●Risk stratification models are subject to limitations that may reduce their prognostic value. These include dependence upon incomplete or uncertain clinical variables. In addition, current risk stratification models do not include several variables that may be clinically important. (See 'Limitations' above.)

●Calculators are available online for the 2018 STS model, the EuroSCORE II model, and the Ambler risk model. (See 'Risk stratification models' above.)

●Since the 2018 STS and EuroSCORE II risk models provide risk stratification based upon large contemporary databases with adjustment for multiple clinical variables, we recommend their use to estimate operative risk in patients in settings similar to those represented by their respective databases. (See 'Use' above.)