Individualized Assessment for Percutaneous or Surgical-Based Revascularization

Patient Comorbidities

The demographics of patients presenting to tertiary care services in need of revascularization are constantly evolving. This has been largely the consequence of increased longevity of the general population, a lower threshold to investigate patients who present with symptoms suggestive of obstructive CAD, and increased resources that have made revascularization via percutaneous coronary intervention (PCI) or coronary artery bypass grafting (CABG) more accessible. Together with increased age, patients in need of revascularization are currently more likely to have comorbidities such as diabetes, hypertension, and hyerlipidemia. These factors are all implicated in accelerating the progression of CAD, and consequently patients are more likely to present with more extensive CAD. The Arterial Revascularization Therapies Studies (ARTS) parts I and II were separated by a period of 5 years, and despite both studies having the same inclusion criteria, patients in ARTS-II had a significantly greater incidence of risk factors and overall increased disease complexity ( Table 1.1 ).

Patient comorbidities must be taken into consideration when assessing patients for revascularization because they have the potential to significantly influence patient outcomes; moreover, they may have a different impact depending on the underlying revascularization strategy selected. Notably in patients enrolled in the ARTS-I and II studies, patient age was shown to be a significant independent correlate of major adverse cardiovascular and cerebrovascular events (MACCEs) who were treated with CABG. More recently, in the randomized all-comers SYNTAX trial, increasing age was shown to favor PCI over CABG when adjustments were made for other anatomic and clinical factors. In addition, other anatomic and clinical factors were shown to have an impact on long-term mortality, and thereby decision making between CABG and PCI (SYNTAX score II ), and this topic is discussed later under “SYNTAX-Based Clinical Tools.”

In a collaborative patient-level analysis of 10 randomized trials of patients with multivessel disease (MVD) treated with PCI using bare-metal stenting (BMS) and CABG, Hlatky and coworkers demonstrated comparable rates of 5-year mortality between both treatment groups in patients without diabetes. Notably, when patients with diabetes were viewed as a whole, mortality was significantly higher in those treated with PCI, even after multivariate adjustment ( Fig. 1.1 ). In the Future Revascularization Evaluation in Patients With Diabetes Mellitus: Optimal Management of Multivessel Disease (FREEDOM) trial, it was shown that in patients with diabetes and advanced CAD, CABG was superior to PCI in that it significantly reduced rates of death and myocardial infarction (MI) but at the expense of a higher rate of stroke ( Fig. 1.2 ). In addition, using the American College of Cardiology Foundation (ACCF) National Cardiovascular Data Registry (NCDR) and the Society of Thoracic Surgeons (STS) Adult Cardiac Surgery Database, Weintraub and colleagues found that subjects who had elective intervention for MVD had a long-term survival advantage among patients who underwent CABG compared with PCI ( Fig. 1.3 ). Findings have been corroborated in the randomized, all-comers SYNTAX trial, as discussed later.

In addition, within the randomized Evaluation of the Xience Everolimus-Eluting Stent Versus Coronary Artery Bypass Surgery for Effectiveness of Left Main Revascularization (EXCEL) trial, in appropriately selected patients with left main (LM) CAD (low-intermediate anatomic SYNTAX scores), PCI with everolimus-eluting stents was shown to be noninferior to CABG with respect to the rate of the composite end point of death, stroke, or MI at 3 years ( Fig. 1.4A ). Moreover, a substantially greater early benefit in quality of life (QOL) was evident with PCI at 1 month, with a similar QOL improvement at 36 months with both PCI and CABG (see Fig. 1.4B ). This corroborates findings originally made in the original landmark SYNTAX trial. Conversely, in the randomized NOBLE (Nordic-Baltic-British left main revascularisation study) and BEST (The Randomized Comparison of Coronary Artery Bypass Surgery and Everolimus-Eluting Stent Implantation in the Treatment of Patients with Multivessel Coronary Artery Disease) trials investigating patients with unprotected LM and MVD, respectively, CABG was shown to offer superior long-term clinical outcomes compared with PCI with contemporary drug-eluting stents (DESs); notably in both trials, patients were not appropriately risk stratified and appropriately selected as occurred in EXCEL. Consequently, an urgent need exists for clinical tools that account for both anatomic and clinical factors and comorbidity to assist the heart team in decision making in regard to the most appropriate revascularization modality in patients with complex CAD.

SYNTAX Trial

The landmark SYNTAX trial represents the largest (and only) assessment of revascularization with PCI or CABG in all-comers with complex CAD. SYNTAX aimed to supply evidence to support the somewhat established but non–evidence-based practice of performing PCI in patients with complex CAD. In addition, SYNTAX also sought to identify which patients should be treated with CABG only. Through an all-comers design, SYNTAX addressed the limitations of the earlier CABG versus PCI trials, which were plagued by profound selection bias as previously discussed (see Table 1.3 ), and in doing so it was anticipated that the results would be more relevant to contemporary routine clinical practice. Specifically:

• To ensure results were applicable to routine practice, the study was designed as an all-comers trial such that there were no specific inclusion criteria other than the need to have revascularization of de novo 3VD or unprotected LM CAD (in isolation or with CAD). Exclusion criteria were limited to prior revascularization, ongoing MI, and patients requiring concomitant cardiac surgery. In contrast to the earlier studies, 70.9% of eligible patients were enrolled.

• The previously indicated problem of reporting outcomes from all patients with complex CAD together, irrespective of disease severity, was addressed in the SYNTAX trial through the use of the anatomic SYNTAX score ( www.syntaxscore.com ; Fig. 1.5 ), which enabled CAD complexity to be objectively and prospectively quantified.

  

• To ensure assessment of patients on an individual level, all patients eligible for enrollment were discussed by the heart team. An interventional cardiologist and cardiac surgeon carried out a careful and comprehensive review of the patient in terms of their anginal status, comorbidities, and coronary anatomy using the respective Braunwald score, European System for Cardiac Operative Risk Evaluation (EuroSCORE), and SYNTAX score (discussed under “SYNTAX-Based Clinical Tools”). The consensus reached from this meeting was subsequently used to allocate the patient into one of the three arms of the trial. In total, 3075 patients were enrolled into one of the following:

  • 1.

    Randomized group ( n = 1800 [58.5%]; 897 CABG, 903 PCI): These patients had CAD and were equally suitable for revascularization with PCI or CABG. The mean SYNTAX score for this group was 26.1 and 28.8 in patients treated with CABG and PCI, respectively.

  • 2.

    Nested CABG registry ( n = 1077 [35.0%]): These patients had CAD that was considered unsuitable for PCI, clearly reflected in the high mean SYNTAX score (37.8) for this group.

  • 3.

    Nested PCI registry ( n = 198 [6.4%]): These patients were deemed unsuitable for CABG. The commonest reason for this decision was the presence of multiple comorbidities reflected in the mean EuroSCORE, which was 2 points higher in this group than the mean in the randomized group (5.8 vs. 3.8).

Overall, SYNTAX failed to meet the prespecified primary end point of noninferiority in terms of 12-month MACCEs, a composite of death, stroke, MI, and repeat revascularization (17.8% vs. 12.4%, P = .002). Final 5-year reporting of SYNTAX demonstrated significantly higher incidence of MACCE with PCI compared with CABG (26.9% vs. 37.3%, P < .0001; Fig. 1.6 ).

As indicated earlier, analyses of all patients irrespective of disease severity does not provide adequate information for clinicians who are faced daily with patients who display a wide variety of CAD complexity. To address this limitation of earlier studies, patient outcomes in SYNTAX were stratified according to tertiles of the anatomic SYNTAX score. As shown in Fig. 1.7 , clinical outcomes between patients treated with PCI and CABG in SYNTAX differed according to the presence of 3VD or unprotected LM CAD. With 3VD, a low SYNTAX score (<23) allowed for similar outcomes between CABG and PCI, whereas higher SYNTAX scores (particularly in the high SYNTAX score [>32] group) clearly favored CABG. With unprotected LM CAD, a low-intermediate SYNTAX score (<33) allowed for similar outcomes between CABG and PCI, whereas a high SYNTAX score (>32) clearly favored CABG. Furthermore, the SYNTAX score II, essentially the anatomic SYNTAX score augmented with clinical variables shown directly to affect decision making between CABG and PCI, was developed in the randomized, all-comers SYNTAX trial and allowed for the identification of higher- and lower-risk subjects in all tertiles of the anatomic SYNTAX score who had a long-term mortality that favored either CABG, PCI, or both revascularization modalities (discussed under “SYNTAX-Based Clinical Tools”).

The results of SYNTAX reiterate the importance of assessing patients when selecting a revascularization strategy. SYNTAX was able to identify those patients in whom either CABG or PCI was appropriate or in whom CABG or PCI was the optimal treatment. Considering the distribution of CAD in SYNTAX, overall one-third of patients with 3VD/LM disease were deemed to have CAD that could be treated safely and effectively with PCI or CABG, whereas in the remaining two-thirds, CABG remained the standard of care. Although these results helped further delineate the boundaries between a percutaneous and surgical revascularization approach in patients with complex CAD, the validation of the anatomic SYNTAX score and development of the SYNTAX score II notably facilitated a more objective assessment of patients by the heart team as discussed later under “SYNTAX-Based Clinical Tools.”

Qualitative Versus Quantitative Risk Assessment

Qualitative risk stratification is subjective and relies on a clinician’s experience. This subjective qualitative assessment also allows risk to be calculated and tailored to the expertise of the physician performing the procedure, as opposed to a clinician in another region who may use different techniques and who may have different equipment available. In addition, assessments of patient frailty can be made that are frequently not captured by conventional risk-scoring systems. This assessment does not require a calculator or computer and can be “computed” subconsciously very quickly. The major disadvantages of this method of risk assessment are its dependence on the operator’s prior experience, potential personal bias to undertake or withhold potential revascularization, and its high interobserver variability. In addition, influences of local practice often dominate clinical decision making, irrespective of the revascularization guidelines.

Quantitative risk stratification can be performed using a variety of risk scores that frequently incorporate clinical variables sourced from large patient registries, with the exception of the SYNTAX score II, which was developed in the all-comers randomized SYNTAX trial to reduce unavoidable (but often appropriate) selection bias inherent to all registries no matter their size. These risk scores largely incorporate objective variables to ensure adequate reproducibility of the score; however, those risk scores—such as the American College of Cardiology/American Heart Association (ACC/AHA) lesion score or the anatomic SYNTAX score/newly developed SYNTAX score II, which include angiographic variables—continue to have documented intraobserver and interobserver variability. However, these tools do provide a more objective assessment of the patient risk and suitability for the most appropriate revascularization modality, which may be modified by the heart team consensus. In addition to their role in the risk stratification of individual patients, these quantitative risk scores have increasing use in the wider context of overall health care. They can provide a vital measure of overall patient care and can help to identify future directions to further improve outcomes. Clinical governance and the increasing requirement to publicly report clinical performance and complications have also propelled the need to risk stratify patients, thereby allowing a useful comparison of performance to be made between clinicians and institutions against the standards dictated by regulatory authorities. In addition, calculation of risk using accepted risk scores may aid clinicians faced with an increasing need to be able to justify their clinical decisions to peers, regulatory bodies, and patients.

In comparison with the qualitative risk scores, the use of a finite number of variables results in these risk scores lacking the sensitivity to accurately predict risk in an individual, such that they are more apt at predicting risk for a population of patients with similar comorbidities. The number of variables included in the score must strike a balance between sufficient numbers to enable a meaningful prediction of risk to be calculated; however, the number must not be excessive so as to prevent use in routine practice. In addition, a minimal number of variables reduces the chances of colinearity between independent variables, which can result in redundant information being collected while also increasing the chances of “overfitting” the score and thereby reducing the overall applicability and accuracy of the results to conventional clinical practice.

The applicability of a risk score to contemporary practice must also take into consideration the time when the score was developed. Risk scores rely on large patient databases to derive appropriate weighting factors for variables in the score to enable the final calculation of risk. It follows that they are developed using retrospective information that may no longer be relevant in the era when the risk score is being used. For example, the EuroSCORE was developed in 1999; however, there have been calls for its recalibration because repeated evaluations indicate that it overestimates risk by a factor of 2 to 3, which has largely been attributed to improvements in surgical techniques and lower perioperative mortality in the decade following its construction. The updated EuroSCORE II currently addresses many of the limitations of the original EuroSCORE. The STS score is also derived from a large patient database and is periodically recalibrated to ensure its results are applicable to contemporary practice.

Risk Scores in Contemporary Practice

Numerous risk scores are available to assist clinicians in stratifying risk among patients undergoing revascularization. Some scores are appropriate for patients prior to the selection of a revascularization strategy, whereas some have been validated only in patients undergoing one form of revascularization. Nevertheless, the various risk scores can largely be categorized according to the variables—clinical, angiographic, or a combination of both—used in the overall estimation of risk. Tables 1.4 and 1.5 summarize the different risk scores used in contemporary CABG and PCI practice (excluding SYNTAX-based tools), and Table 1.6 summarizes SYNTAX-based clinical tools. A selection of these is described in more detail later.

Clinical Scores

These risk scores incorporate only clinical variables and do not require any data from the angiogram. They offer the advantage of being able to be computed relatively quickly, usually at the bedside, and principally include variables that are not subject to user interpretation, thereby ensuring excellent reproducibility.

EuroSCORE and EuroSCORE II

The EuroSCORE is an established risk score that uses 17 clinical variables used in cardiothoracic surgical practice for predicting operative mortality, and it has been validated in many populations around the world. In use since 1999, the score was derived from almost 20,000 consecutive patients from 128 hospitals in eight European countries. The additive EuroSCORE assigns an individual score to 17 clinical variables ( Table 1.7 ) with a low-risk tertile that ranges from 1 to 2, an intermediate-risk tertile from 3 to 5, and a high-risk tertile of 6 and higher. However, early validation studies suggested that the additive EuroSCORE underestimated risk in those at highest risk; this led to the development of the logistic EuroSCORE, which uses the same clinical variables and requires use of an online calculator (available at www.euroscore.org ) to quantify risk. However, the logistic EuroSCORE has been shown to potentially overestimate observed mortality, and its accuracy at predicting risk varies in different surgical subgroups.

TABLE 1.7

Components of the EuroSCORE and Relevant Weighting Factors of the Additive and Logistic EuroSCOREs

From Singh M, Rihal CS, Lennon RJ, et al. Bedside estimation of risk from percutaneous coronary intervention: the new Mayo Clinic Risk Scores. Mayo Clinic Proc . 2007;82:701–708; and Singh M, Peterson ED, Milford-Beland S, et al. Validation of the Mayo Clinic risk score for in-hospital mortality after percutaneous coronary interventions using the National Cardiovascular Data Registry. Circ Cardiovasc Interv . 2008;1:36–44.

Patient Characteristics	Additive		Logistic β Coefficient
Age	Per 5 years or part thereof over the age of 60 years	1	0.07
Sex	Female	1	0.33
Chronic pulmonary disease	Long-term use of bronchodilators or steroids for respiratory disease	1	0.49
Peripheral arteriopathy	Claudication, carotid stenosis >50%, previous or planned intervention on the abdominal aorta, limb arteries, or carotids a	2	0.66
Neurologic dysfunction	Severely affected mobility or day-to-day function	2	0.84
Previous cardiac surgery	Previous opening of the pericardium	3	1.00
Serum creatinine	Preoperatively greater than 200 μmol/L	2	0.65
Active endocarditis	Antibiotic therapy at time of surgery	3	1.10
Critical preoperative state	Preoperative cardiac arrest, ventilation, renal failure, inotropic support, intraaortic balloon pump use, ventricular arrhythmia a	3	0.91
Cardiac-Related Factors
Unstable angina	Rest pain that requires IV nitrates	2	0.57
Left ventricular function	Moderate (30%–50%) Poor (<30%)	1 3	0.42 1.09
Recent MI	Within 90 days	2	0.55
Pulmonary hypertension	Systolic pulmonary pressure greater than 60 mm Hg	2	0.77
Operation-Related Factors
Emergency	Operation performed before the start of next working day	2	0.71
Other than isolated CABG	Major cardiac procedure other than or in addition to CABG	2	0.54
Surgery on thoracic aorta		3	1.16
Postinfarct septal rupture		4	1.46
Constant β 0			−4.79

The logistic EuroSCORE can be calculated at www.euroscore.org .

CABG, Coronary artery bypass grafting; IV, intravenous; MI, myocardial infarction.

a Any of these.

In addition to the EuroSCORE’s assessment and validation in patients undergoing surgical revascularization, Kim and colleagues first demonstrated that the high-risk tertile of the additive EuroSCORE was an independent predictor of death/MI after unprotected LM intervention with sirolimus-eluting stents. Subsequently, Romagnoli and coworkers applied the additive EuroSCORE to predict in-hospital mortality in 1173 consecutive patients undergoing PCI in a single high-volume center and correlated the higher-risk tertiles of the EuroSCORE with in-hospital mortality; the study population also included patients who had undergone unprotected LM PCI. The EuroSCORE has since been evaluated in numerous studies of patients undergoing PCI, the majority of which specifically enrolled patients with LM disease. Notably, all studies, irrespective of disease severity, have demonstrated the EuroSCORE to be an independent predictor of mortality and/or MACCE at follow-up ranging from 1 to 3 years. Importantly, those studies that also included a surgical control group—such as the SYNTAX study, the Revascularization for Unprotected Left Main Coronary Artery Stenosis: Comparison of Percutaneous Coronary Angioplasty Versus Surgical Revascularization (MAIN-COMPARE) study, and the registry by Rodés-Cabau and colleagues —also demonstrated the EuroSCORE to be an independent predictor of MACCE in surgical patients. Only one study has examined the logistic EuroSCORE in PCI patients; however, little differences were found in stratifying risk when compared with the additive EuroSCORE.

Specifically in the SYNTAX trial, which represents the only randomized study to assess the EuroSCORE, the additive EuroSCORE was shown to be an independent predictor of MACCE at 1-year follow-up irrespective of the method of revascularization (odds ratio [OR]: 1.21; 95% confidence interval [CI]: 1.12 to 1.32; P < .001) in 705 patients undergoing LM revascularization. Similarly, at intermediate follow-up of 23 months, Rodés-Cabau and colleagues identified a EuroSCORE of 9 or higher as the best predictor of MACCE after PCI and CABG among 249 octogenarians with LM disease. In the MAIN-COMPARE registry, which enrolled more than 1500 patients with LM disease followed up for a median of 3.1 years, the EuroSCORE has been identified as an independent predictor of death, MI, and stroke irrespective of revascularization strategy. In addition, in the same registry, a EuroSCORE of 6 or higher has been shown to be an independent predictor of mortality following either PCI or CABG.

More recently, the EuroSCORE II ( Table 1.8 ) was developed to improve the risk prediction of the original EuroSCORE. EuroSCORE II was developed on newer data to reflect more contemporary surgical practice given that cardiac surgical mortality has decreased significantly in the last 15 years, despite patients being older and sicker, and that the previous additive and logistic EuroSCOREs were suggested to be representative of outdated surgical practices.

TABLE 1.8

Final Risk Factors by Multivariate Regression for the EuroSCORE II

From Nashef SA, Roques F, Sharples LD, et al. EuroSCORE II. Eur J Cardiothorac Surg . 2012;41(4):734–744.

Risk Factor	Coefficient	Standard Error	z	P ≥ | z |	[95% Confidence Interval]
New York Hospital Association (NYHA)
II	0.1070545	0.1463849	0.73	0.465	[−0.1798547 to 0.3939637]
III	0.2958358	0.141466	2.09	0.037	[0.0185674 to 0.5731042]
IV	0.5597929	0.1697565	3.30	0.001	[0.2270763 to 0.8925095]
CCS4	0.2226147	0.1462888	1.52	0.128	[−0.0641061 to 0.5093356]
IDDM	0.3542749	0.145863	2.43	0.015	[0.0683887 to 0.6401611]
Age	0.0285181	0.0065954	4.32	0.000	[0.0155914 to 0.0414448]
Female	0.2196434	0.0953505	2.30	0.021	[0.0327599 to 0.4065269]
ECA	0.5360268	0.1106046	4.85	0.000	[0.3192458 to 0.7528079]
CPD	0.1886564	0.1232126	1.53	0.126	[−0.0528358 to 0.4301486]
N/M mob	0.2407181	0.1729494	1.39	0.164	[−0.0982564 to 0.5796927]
Redo	1.118599	0.1226272	9.12	0.000	[0.8782539 to 1.3589440]
Renal Dysfunction
On dialysis	0.6421508	0.3083468	2.08	0.037	[0.0378021 to 1.2464990]
CC ≤ 50	0.8592256	0.1446758	5.94	0.000	[0.5756663 to 1.1427850]
CC 50–85	0.303553	0.1240518	2.45	0.014	[0.0604159 to 0.5466901]
Active endocarditis	0.6194522	0.2046001	3.03	0.002	[0.2184433 to 1.0204610]
Critical	1.086517	0.147657	7.36	0.000	[0.797115 to 1.3759200]
Left Ventricular Function
Moderate	0.3150652	0.1036182	3.04	0.002	[0.1119773 to 0.5181530]
Poor	0.8084096	0.1498233	5.40	0.000	[0.5147614 to 1.1020580]
Very poor	0.9346919	0.2917754	3.20	0.001	[0.3628227 to 1.5065610]
Recent MI	0.1528943	0.136257	1.12	0.262	[−0.1141646 to 0.4199531]
Pulmonary Artery Systolic Pressure
31–55 mm Hg	0.1788899	0.1266713	1.41	0.158	[−0.0693812 to 0.4271611]
≥55	0.3491475	0.1676641	2.08	0.037	[0.0205318 to 0.6777632]
Urgency
Urgent	0.3174673	0.1174178	2.70	0.007	[0.0873326 to 0.5476020]
Emergency	0.7039121	0.1719835	4.09	0.000	[0.3668306 to 1.0409940]
Salvage	1.362947	0.33706	4.04	0.000	[0.7023221 to 2.0235730]
Weight of Procedure
1 non-CABG	0.0062118	0.1463574	0.04	0.966	[−0.2806434 to 0.2930670]
2	0.5521478	0.1268137	4.35	0.000	[0.3035975 to 0.8006980]
3+	0.9724533	0.1463969	6.64	0.000	[0.6855206 to 1.2593860]
Thoracic aorta	0.6527205	0.221183	2.95	0.003	[0.2192097 to 1.0862310]
Constant	−5.324537	0.1682446	−31.65	0.000	[−5.65429 to −4.9947830]

CABG, Coronary artery bypass grafting; CC, creatinine clearance; CCS, Canadian Cardiovascular Society; CPD, chronic pulmonary dysfunction; Critical, critical preoperative state; ECA, extracardiac arteriopathy; IDDM, insulin-dependent diabetes mellitus; N/M mob, neurologic or musculoskeletal dysfunction severely affecting mobility; MI, myocardial infarction; Redo, previous cardiac surgery.

Weight of procedure: 1, non-CABG, single major cardiac procedure (MCP) that is not isolated CABG; 2, two MCPs; 3+, three or more MCPs. For age, X _i = 1 if patient age is 60 or younger; X _i increases by one point per year thereafter (e.g., age 60 or less, X _i = 1; age 61, X _i = 2; age 62, X _i = 3, etc.).

The EuroSCORE II was shown to have improved calibration (actual mortality 4.18%, predicted 3.95%) compared with the original EuroSCORE (actual 3.9%, additive predicted 5.8%, logistic predicted 7.57%) while preserving discrimination (area under the receiver operating characteristic [ROC] curve of 0.8095). However, it should be noted that regular revalidation of the EuroSCORE II will need to be continued to identify calibration drift or clinical inconsistencies seen in previous versions.

In summary, while acknowledging that most of these studies have been nonrandomized observational studies, the findings do suggest that the EuroSCORE and EuroSCORE II are valuable tools in the individual assessment of risk prior to the selection of a revascularization strategy.

New Mayo Clinic Risk Score

The new Mayo Clinic Risk Score (MCRS) was designed to replace the original MCRS by predominantly excluding angiographic variables, namely the presence of LM or MVD, and a few of the interaction effects of specific clinical variables (see Table 1.5 ).

The new MCRS is based solely on baseline clinical and noninvasive assessments and incorporates seven preprocedural variables (age, serum creatinine, left ventricular ejection fraction [LVEF], MI within the past 24 hours, preprocedural shock, congestive heart failure, and peripheral vascular disease). The risk score had a C-statistic (area under ROC curve) of 0.74 and 0.89 for major adverse cardiovascular events (MACEs) and procedural death, respectively, in the population from whom the risk score was derived. The risk score has since been validated for in-hospital mortality in the NCDR ; however, it has not been validated for MACEs. The new MCRS has also been demonstrated to be predictive of in-hospital mortality after CABG surgery.

Value of Age, Creatinine, and Ejection Fraction Score

Ranucci and colleagues demonstrated in a relatively simple risk score consisting of only three clinical variables—age, preoperative serum creatinine value, and LVEF—a risk score for assessing operative mortality risk in elective cardiac operations. Notably, despite the simplicity of the score, the clinical performance of the Age, Creatinine, and Ejection Fraction (ACEF) score appeared to be comparable with either the additive or the logistic EuroSCORE.

The ACEF score is calculated using the following formula:

$\text{ACEF}=\left[\text{Age}/\text{LVEF}\left(\text{\%}\right)\right]+\left[1\left(\text{if}\text{creatinine>}2\text{mg}/\text{dL}\right)\right]$

From ACEF, a mortality risk can be calculated from a graphical relationship of ACEF with an operative risk or an equation ( Fig. 1.8 ). ACEF was developed from an initial dataset of 4557 patients and a subsequent validation series of 4091 patients from a single institution. The results demonstrated a similar accuracy and calibration for the prediction of in-hospital mortality with ACEF when compared with other more complicated surgical risk scores such as the EuroSCORE and the Cleveland Clinic Score. Subsequent validation studies have shown ACEF to have an accuracy level at least comparable with that of the EuroSCORE for operative mortality risk stratification in a series of 29,659 patients undergoing elective cardiac surgery.

In addition, ACEF was applied to PCI patients from the all-comers LEADERS population at 1-year follow-up. Despite ACEF being demonstrated to be superior to the SYNTAX score alone as a predictor of cardiac death and MI after PCI, ACEF was found to be inferior to the SYNTAX score at predicting overall MACE rates and the risk of repeat revascularization. This reflects the observation that anatomic and clinical variables appear to be necessary requirements for a comprehensive risk score in predicting clinical outcomes after PCI.

National Cardiovascular Database Registry CathPCI Risk-Prediction Score

The NCDR CathPCI risk-prediction score is the most contemporary clinical risk score currently available. It incorporates information from nine clinical variables ( Table 1.9 ), which are assigned appropriate weighted values and are then added together to give a final score that can be translated into risk of in-hospital mortality ( Fig. 1.9 ). The score was developed using data from more than 180,000 patients from the voluntary U.S. NCDR database and was validated in more than 400,000 patients from the same database who underwent PCI between March 2006 and March 2007. Notably, the C-statistic for the prediction of in-hospital mortality was consistently greater than 0.90 for in-hospital mortality, whereas a lower but nevertheless adequate C-statistic of 0.83 was seen for 30-day mortality.

TABLE 1.9

National Cardiovascular Database

Registry risk score from Peterson ED, Dai D, DeLong ER, et al. Contemporary mortality risk prediction for percutaneous coronary intervention: results from 588,398 procedures in the National Cardiovascular Data Registry. J Am Coll Cardiol . 2010;55:1923–1932.

Variable	Scoring Response Categories
Age	<60	≥60, <70	≥70, <80	≥80
Weighted score	0	4	8	14
Cardiogenic shock	No	Yes
Weighted score	0	25
Prior CHF	No	Yes
Weighted score	0	5
Peripheral vascular disease	No	Yes
Weighted score	0	5
Chronic lung disease	No	Yes
Weighted score	0	4
GFR (mL/min)	<30	30 to 60	60 to 90	>90
Weighted score	18	10	6	0
NYHA Class IV	No	Yes
Weighted score	0	4
PCI Status (STEMI)	Elective	Urgent	Emergent	Salvage
Weighted score	12	15	20	38
PCI Status (no STEMI)	Elective	Urgent	Emergent	Salvage
Weighted score	0	8	20	42

CHF, Congestive heart failure; GFR, glomerular filtration rate; NYHA, New York Heart Association; PCI, percutaneous coronary intervention; STEMI, ST elevation myocardial infarction.

The risk of in-hospital mortality is derived using Fig. 1.11 .