CLINICAL RESEARCH: INTERVENTIONAL CARDIOLOGY
Comparison of Mayo Clinic risk score and American College of Cardiology/American Heart Association lesion classification in the prediction of adverse cardiovascular outcome following percutaneous coronary interventions
Mandeep Singh, MD*,*,
Charanjit S. Rihal, MD*,
Ryan J. Lennon, MS ,
Kirk N. Garratt, MD* and
David R. Holmes, Jr, MD*
* Division of Cardiovascular Diseases and Internal Medicine, Rochester, Minnesota, USA
Division of Biostatistics, Mayo Clinic College of Medicine, Rochester, Minnesota, USA
Manuscript received January 8, 2004;
revised manuscript received March 18, 2004,
accepted March 22, 2004.
* Reprint requests and correspondence: Dr. Mandeep Singh, 200 2nd Street SW, Mayo Clinic, Rochester, Minnesota 55905, USA.
singh.mandeep{at}mayo.edu
 |
Abstract
|
|---|
OBJECTIVES: We compared American College of Cardiology/American Heart Association (ACC/AHA) lesion classification with the recently proposed Mayo Clinic risk score to predict complications following percutaneous coronary intervention (PCI).
BACKGROUND: The ability of the ACC/AHA classification system to predict complications following PCI has been modest. With the inclusion of patient demographics, acuity of presentation, and measure of left ventricular function, models with better discriminatory accuracy are presently available.
METHODS: The Mayo Clinic risk score is constructed by adding integer scores for the presence of eight variables. We mapped the lesion-specific risk levels to a patient level by counting the number of lesions in each class (A, B1, B2, C, and unknown).
RESULTS: In 5,064 PCIs, 183 patients (4%) had the primary end point (death, Q-wave myocardial infarction, stroke, emergency coronary artery bypass graft). Of the 7,632 treated lesions, 891 (12%) were unsuccessfully treated with PCI (residual stenosis >20%). The discriminatory ability of the Mayo Clinic risk score model for prediction of the primary end point, as measured by the c-statistic, was 0.78 (95% confidence interval [CI] 0.74 to 0.81). The Mayo Clinic risk score offered significantly better risk stratification than the ACC/AHA lesion classification counts (95% CI for c-statistic difference: 0.05 to 0.15). Regarding angiographic success, the ACC/AHA lesion classification was a better system (95% CI for c-statistic difference: –0.08 to –0.03 favoring ACC/AHA classification), although its absolute ability was modest (c = 0.58).
CONCLUSIONS: Mayo Clinic risk score offers significantly better prediction for cardiovascular complications than the ACC/AHA classification. However, lesion classification by ACC/AHA classification is a better predictor for angiographic success.
|
Abbreviations and Acronyms
| | ACC | = American College of Cardiology | | AHA | = American Heart Association | | CABG | = coronary artery bypass graft | | CI | = confidence interval | | MACE | = major adverse cardiovascular events | | MI | = myocardial infarction | | NHLBI | = National Heart, Lung, and Blood Institute | | NYHA | = New York Heart Association | | PCI | = percutaneous coronary intervention | | SCAI | = Society for Cardiac Angiography and Interventions |
|
The American College of Cardiology/American Heart Association (ACC/AHA) lesion classification scheme was initially proposed in 1986 (1) and subsequently modified in 1990 (2). These classifications required evaluation of 11 angiographic variables for each lesion and were introduced to predict results based on the anticipated difficulty of the required percutaneous coronary intervention (PCI) procedure. Major limitations of this classification scheme are subjective angiographic lesion assessment with considerable interobserver variability (3) and exclusion of important patient (4,5), left ventricular function (5), and acuity of presentation variables (6) that have considerable influence on the outcome of a PCI procedure. We developed a risk score comprising five clinical and three angiographic variables to predict death, Q-wave myocardial infarction (MI), stroke, and the need for emergency coronary artery bypass graft surgery (CABG) (7). This model was initially internally validated and more recently externally validated using the National Heart, Lung, and Blood Institute (NHLBI) Dynamic Registry, with excellent discriminatory accuracy to predict complications following percutaneous coronary procedures. The information on the lesion complexity is routinely recorded in the Mayo Clinic PCIs registry. The utility of the ACC/AHA lesion classification score is not clear with the current advancement in the procedure techniques, operator experience, and availability of stents. With this background, we compared the discriminatory ability of the Mayo Clinic risk score with the ACC/AHA lesion classification for prediction of both complications and angiographic failure.
|
Materials and methods
|
|---|
Study population.
There were 6,023 coronary angioplasty procedures performed on 5,193 unique patients at Saint Mary's Hospital at Mayo Clinic from January 1, 2000, through July 31, 2003; 129 patients refused to allow use of their records for research and hence were excluded. We selected the earliest PCI from the remaining 5,064 patients for study. The Mayo Clinic's Institutional Review Board approved the study.
Classification measures.
The Mayo Clinic risk score for PCI complications has been previously described (7). In brief, this score is constructed by adding integer scores for the presence of eight possible risk factors. Five variables are patient/procedural characteristics: age, congestive heart failure, New York Heart Association (NYHA) functional class  III, urgent/emergent PCI, chronic renal disease, and preprocedural cardiogenic shock. The other three are related to angiographic characteristics: left main stenosis 70%, multivessel disease (more than 70% in more than 1 major epicardial coronary artery, or in one of their major branches), and presence of thrombus in any lesion. Some risk factors were unrecorded for some patients. In the presence of missing data (<10% for any component), the risk was assumed to have not been present for calculation of the score. The score was also grouped into five categories of risk: very low, low, moderate, high, and very high.
The ACC/AHA lesion classification orders lesions into four groups: A, B1, B2, and C. This classification scheme was not adopted for patients presenting with acute MI, who was therefore treated as not available. We mapped the lesion-specific risk levels to a patient level by two methods. First, we counted the number of lesions in each class (A, B1, B2, C, and unknown). Thus, five integer variables defined the patient-level ACC/AHA classification. Second, we simply recorded the worst lesion classification. Here, we had four indicators for the worst lesion classification (A, B1, B2, or C) plus an indicator for a missing measure on any lesions. The first method contains more information, especially regarding the number of lesions treated. The second method is simple and commonly used.
Primary end point.
The primary end point was post-PCI in-hospital major adverse cardiac event (death, Q-wave MI, emergency CABG, or stroke). The secondary end point was lesion-specific angiographic failure defined as residual stenosis >20% post-PCI.
Statistical analysis.
Patient, procedural, and angiographic characteristics are summarized as mean ± SD for continuous measures unless noted otherwise, and frequency (percentage) for discrete risk factors. The c-statistic (the area under the receiver operator curve) was used to estimate the discriminatory ability of the risk stratification systems. This reflects the classification system's ability to correctly order patients in sets of increasing risk. Angiographic failure is a lesion-specific end point; thus each lesion was an individual observation for the analysis of angiographic failure. The Mayo Clinic risk score was attributed to each lesion by adding the patient-level risk integers as usual, but only adding left main stenosis and thrombus to the relevant lesions. A total of 1,000 bootstrap samples were used to estimate 95% confidence intervals for the c-statistics. Differences in the c-statistic for different classification systems were calculated for each bootstrap sample. Bootstrapping for the angiographic failure end point, however, did not resample individual lesions, but rather resampled each patient to maintain the within-patient lesion correlations.
|
Results
|
|---|
The primary end point of death, Q-wave MI, stroke, or emergency CABG was seen in 183 patients (4%). Of the 7,632 treated lesions, 891 (12%) were unsuccessfully treated with PCI (residual stenosis >20%). The patient characteristics are listed in Table 1. The mean age was 66.9 ± 12.0 years. There were 19% of patients with MI within 24 h of the procedure, 4% had cardiogenic shock, 66% had nonelective PCI, 7% had NYHA functional class III or higher heart failure, and 4% had moderate to severe renal disease. The median Mayo Clinic risk score was 6. We divided the patients into five risk groups based on the previous risk score model. Patients were considered very low risk to very high risk depending upon the number of clinical and angiographic variables present in a particular patient. There were 19% patients in the moderate-to-high risk group and 2% belonged to the very-high-risk group.
Table 2 summarizes the procedural and angiographic variables. Among these variables, 2% had significant left main disease, 52% had multivessel disease, and 30% had angiographic thrombus. Glycoprotein IIb/IIIa inhibitors were used in 63%. The ACC/AHA lesion type A was seen in 4% of patients, types B1 and B2 in 52% of patients, and type C in 36% of patients. The lesion classification was not available in 8% of patients.
Table 3 outlines the in-hospital complications. There were 86 deaths (1.7%), 67 Q-wave MI (1.3%), and 28 patients (0.6%) were referred for emergency CABG. Angiographic failures were seen in 891 lesions (11.7%).
Table 4 demonstrates the increase in major adverse cardiovascular events (MACE) with increasing Mayo Clinic risk score, and also with worsening ACC/AHA lesion classification when all treated lesions are classified. The risk increased from 1.1% for MACE in patients with a very low Mayo Clinic risk score to 33.3% in the very-high-risk category. The risk for MACE in ACC/AHA type A and B1 lesion category was very low, and similar to Mayo Clinic risk score for the very-low risk category; however, the MACE rates for ACC/AHA type B2 and C were 2.2% and 4.9%, respectively. The risk when some lesion classes are unknown is 8.0%.
Table 5 compares the discriminatory ability of the Mayo Clinic risk score with the ACC/AHA lesion score. The Mayo integer score and group score gave similar results for MACE prediction. The c-statistic for association with MACE for the Mayo Clinic score was 0.78 (95% confidence interval [CI]: 0.74 to 0.81). It was 0.77 (95% CI: 0.73 to 0.80) when the Mayo Clinic risk score was applied to stratify patients into five groups of risk. This was significantly better than ACC/AHA lesion classification counts (c-statistic 0.67, 95% CI: 0.64 to 0.72) and c-statistic of 0.66 (95% CI: 0.63 to 0.70) for worst lesion scheme, as the 95% confidence interval (0.05 to 0.15) for the difference in c-statistics did not include 0. The angiographic failure prediction with ACC/AHA lesion classification was modest. The c-statistic was 0.58 (95% CI: 0.56 to 0.60), which was superior to the Mayo Clinic risk score (c-statistic 0.53; 95% CI: 0.50 to 0.55). The difference was statistically significant (95% CI for difference: –0.08 to –0.03). We also tested this model using 50% residual stenosis as the definition for angiographic failure (data not shown). The conclusions are similar, with slightly better discriminatory ability for both scores (Mayo score: c = 0.55, ACC/AHA score: c = 0.65).
|
Discussion
|
|---|
This study demonstrated that in the risk stratification for MACE (death, Q-wave MI, stroke, emergency CABG), the Mayo Clinic risk score fared significantly better than ACC/AHA lesion classification. For prediction of angiographic failure, the ACC/AHA lesion classification has only modest ability to stratify patients into increasing levels of risk, but is superior to the Mayo Clinic risk score.
ACC/AHA lesion classification.
There has been a rapid evolution of percutaneous angioplasty techniques since the original proposal of this classification in 1986 with subsequent modification in 1990 (1,2). It is still the most widely used system to assess patients' risk from PCI, and is also used to scorecard the individual operator and the institution. More recently in a new classification system, two variables (non-chronic total occlusion and degenerated saphenous vein grafts) were significantly correlated with death, non–Q-wave MI, and the need for emergency CABG (8). Krone et al. (9) applied the ACC/AHA lesion classification to 61,926 patients undergoing single-vessel angioplasty in the ACC-National Cardiovascular Data Registry (NCDR) and found ACC/AHA less accurate in predicting major complications compared with Society for Cardiac Angiography and Interventions (SCAI) lesion classification. The c-statistic for major complications following PCI, however, was modest (0.599 for original ACC/AHA, 0.624 for modified ACC/AHA, and 0.665 for SCAI lesion classification). On the other hand, the predictive accuracy of the lesion classification, whether assessed by ACC/AHA or SCAI classification, was better for lesion success (0.69, 0.708, 0.75 for original, modified ACC/AHA lesion classification, and SCAI classification, respectively). In other studies, procedural success is adversely affected by presence of type C lesions (10,11). Among the type C lesions, chronic total occlusion and extreme tortuosity continue as significant lesion variables affecting procedure outcome (12). Type C lesion was also one of the significant variables predictive of the need for emergency CABG (13). In a recent analysis of angiographic variables from the Do Tirofiban and ReoPro Give Similar Efficacy Outcomes? (TARGET) trial, patients with a combination of lesion eccentricity, lesion length 20 mm, and presence of angiographic thrombus had 21.4% composite of death, MI, or urgent target vessel revascularization at 30 days, compared with 4.2% in those patients without these high-risk features (14). High-risk patient subgroups, namely acute MI, shock, left main, and decompensated heart failure, were excluded from this study. In the risk adjustment model derived from ACC-NCDR from 1998 to 2000 for mortality prediction following PCI, significant difference were noted in patients presenting with and without acute MI within 24 h of PCI (15). The worse SCAI lesion classification was significantly correlated with mortality in patients not presenting with acute MI. In contrast, the correlation of these lesion variables was weak in patients presenting with acute MI.
The Mayo Clinic risk score was initially developed and internally validated from a combination of eight clinical and angiographic variables (7). More recently, it was externally validated in the NHLBI Dynamic Registry with excellent prediction for in-hospital complications following PCI. This risk score, whether applied as initially described or applied with patients stratified into different groups based on the risk profile, fared significantly better than ACC/AHA lesion classification for prediction of complications following PCI. The disadvantages of the lesion score are its modest predictive accuracy, poor interobserver correlation of the variables, lack of demographic, left ventricular function, acuity of presentation variables, improvement in the equipment, operator experience, and stent techniques such that the lesion variables currently may have lesser influence on outcome. Also, lesion variables with varying results on the outcomes are lumped together in the same category. For example, chronic total occlusion has higher angiographic failure rates than longer lesions (>20 mm); however, both are included in ACC/AHA type C category.
The current role of the lesion variables in the procedural outcome after PCI is ill defined. The outcome is driven more by patient-related conditions (advancing age, renal failure, peripheral vascular disease), acuity of presentation (acute MI, shock), or angiographic or procedure-related characteristics (multivessel disease, intra-aortic balloon pump use, thrombus) (6,16–18). A recent simplified scoring system found four variables: age above 65 years, renal failure, multivessel disease, and MI <14 days. These four variables significantly predicted mortality following PCI, and none were lesion-related (19). The influence of lesion variables is more pronounced in elective cases and in studies where MI was defined as elevation of biomarkers, as noted in the study by Ellis et al. (8). Similar conclusions can be drawn from other studies (14,15). Further studies are needed to develop modified risk scores that would integrate variables included in the Mayo Clinic risk score with the inclusion of elevation of biomarkers as one of the major complications.
Limitations.
Experienced angiographers performed the assessment of lesion morphology; however, no core-lab confirmation was obtained. The fact that some lesions had some missing values for the ACC/AHA classification complicated the analysis, and the system may have fared slightly better if complete data were available. However, <10% of data were missing, and the unknown class is reflective of actual practice. Important postprocedure variables, such as elevation of cardiac biomarkers, were not included in the analysis. Significant elevation of these markers after a PCI procedure may have important prognostic implications (20). The question of the applicability of the Mayo Clinic model to non-referral, low-volume centers and to low-volume interventionists cannot be evaluated from the present study. Operative volume has been found to be significantly associated with adverse events following PCI (21–23). This variable was not addressed in the present study. Although this report was accurate in predicting procedural complications, no predictive model can overcome the effect of chance and uncertainty inherent in invasive treatments.
Conclusions.
The present study demonstrated that the current Mayo Clinic risk score has significantly better discriminatory ability to risk-stratify patients for complications following PCI than does the ACC/AHA lesion classification. Stratification by ACC/AHA class for angiographic failure was modestly successful at best, but superior to the Mayo Clinic risk score.
|
References
|
|---|
1. Ryan TJ, Faxon DP, Gunnar RM, et al. Guidelines for percutaneous transluminal coronary angioplasty. A report of the American College of Cardiology/American Heart Association Task Force on Assessment of Diagnostic and Therapeutic Cardiovascular Procedures (Subcommittee on Percutaneous Transluminal Coronary Angioplasty). Circulation. 1988;78:486–502[Free Full Text]
2. Ellis SG, Vandormael MG, Cowley MJ, et al. Coronary morphologic and clinical determinants of procedural outcome with angioplasty for multivessel coronary disease. Implications for patient selection. Multivessel Angioplasty Prognosis Study Group. Circulation. 1990;82:1193–1202[Abstract/Free Full Text]
3. White CJ, Ramee SR, Collins TJ, et al. Coronary thrombi increase PTCA risk. Angioscopy as a clinical tool. Circulation. 1996;93:253–258[Abstract/Free Full Text]
4. Singh M, Mathew V, Garratt KN, et al. Effect of age on the outcome of angioplasty for acute myocardial infarction among patients treated at the Mayo Clinic. Am J Med. 2000;108:187–192[CrossRef][Medline]
5. Hochman JS, Sleeper LA, Webb JG, et al. Early revascularization in acute myocardial infarction complicated by cardiogenic shock. SHOCK Investigators. Should We Emergently Revascularize Occluded Coronaries for Cardiogenic Shock. N Engl J Med. 1999;341:625–634[Abstract/Free Full Text]
6. Moscucci M, Kline-Rogers E, Share D, et al. Simple bedside additive tool for prediction of in-hospital mortality after percutaneous coronary interventions. Circulation. 2001;104:263–268[Abstract/Free Full Text]
7. Singh M, Lennon RJ, Holmes DR Jr, Bell MR, Rihal CS. Correlates of procedural complications and a simple integer risk score for percutaneous coronary intervention. J Am Coll Cardiol. 2002;40:387–393[Abstract/Free Full Text]
8. Ellis SG, Guetta V, Miller D, Whitlow PL, Topol EJ. Relation between lesion characteristics and risk with percutaneous intervention in the stent and glycoprotein IIb/IIIa era: An analysis of results from 10,907 lesions and proposal for new classification scheme. Circulation. 1999;100:1971–1976[Abstract/Free Full Text]
9. Krone RJ, Shaw RE, Klein LW, et al. Evaluation of the American College of Cardiology/American Heart Association and the Society for Coronary Angiography and Interventions lesion classification system in the current "stent era" of coronary interventions (from the ACC-National Cardiovascular Data Registry). Am J Cardiol. 2003;92:389–394[CrossRef][Medline]
10. Block PC, Peterson ED, Krone R, et al. Identification of variables needed to risk adjust outcomes of coronary interventions: Evidence-based guidelines for efficient data collection. J Am Coll Cardiol. 1998;32:275–282[Abstract/Free Full Text]
11. Kastrati A, Schomig A, Elezi S, et al. Prognostic value of the modified American College of Cardiology/American Heart Association stenosis morphology classification for long-term angiographic and clinical outcome after coronary stent placement. Circulation. 1999;100:1285–1290[Abstract/Free Full Text]
12. Zaacks SM, Allen JE, Calvin JE, et al. Value of the American College of Cardiology/American Heart Association stenosis morphology classification for coronary interventions in the late 1990s. Am J Cardiol. 1998;82:43–49[CrossRef][Medline]
13. Seshadri N, Whitlow PL, Acharya N, Houghtaling P, Blackstone EH, Ellis SG. Emergency coronary artery bypass surgery in the contemporary percutaneous coronary intervention era. Circulation. 2002;106:2346–2350[Abstract/Free Full Text]
14. Ross MJ, Herrmann HC, Moliterno DJ, et al. Angiographic variables predict increased risk for adverse ischemic events after coronary stenting with glycoprotein IIb/IIIa inhibition: Results from the target trial. J Am Coll Cardiol. 2003;42:981–988[Abstract/Free Full Text]
15. Shaw RE, Anderson HV, Brindis RG, et al. Development of a risk adjustment mortality model using the American College of Cardiology-National Cardiovascular Data Registry (ACC-NCDR) experience: 1998–2000. J Am Coll Cardiol. 2002;39:1104–1112[Abstract/Free Full Text]
16. Holmes DR Jr, Berger PB, Garratt KN, et al. Application of the New York State PTCA mortality model in patients undergoing stent implantation. Circulation. 2000;102:517–522[Abstract/Free Full Text]
17. Maynard C, Goss JR, Malenka DJ, Reisman M. Adjusting for patient differences in predicting hospital mortality for percutaneous coronary interventions in the Clinical Outcomes Assessment Program. Am Heart J. 2003;145:658–664[CrossRef][Medline]
18. O'Connor GT, Malenka DJ, Quinton H, et al. Multivariate prediction of in-hospital mortality after percutaneous coronary interventions in 1994–1996. Northern New England Cardiovascular Disease Study Group. J Am Coll Cardiol. 1999;34:681–691[Abstract/Free Full Text]
19. Qureshi MA, Safian RD, Grines CL, et al. Simplified scoring system for predicting mortality after percutaneous coronary intervention. J Am Coll Cardiol. 2003;42:1890–1895[Abstract/Free Full Text]
20. Ellis SG, Chew D, Chan A, Whitlow PL, Schneider JP, Topol EJ. Death following creatine kinase-MB elevation after coronary intervention: Identification of an early risk period: importance of creatine kinase-mb level, completeness of revascularization, ventricular function, and probable benefit of statin therapy. Circulation. 2002;106:1205–1210[Abstract/Free Full Text]
21. Ellis SG, Weintraub W, Holmes D, Shaw R, Block PC, King SB 3rd. Relation of operator volume and experience to procedural outcome of percutaneous coronary revascularization at hospitals with high interventional volumes. Circulation. 1997;95:2479–2484[Abstract/Free Full Text]
22. Hannan EL, Racz M, Ryan TJ, et al. Coronary angioplasty volume-outcome relationships for hospitals and cardiologists. JAMA. 1997;277:892–898[Abstract/Free Full Text]
23. Malenka DJ, McGrath PD, Wennberg DE, et al. The relationship between operator volume and outcomes after percutaneous coronary interventions in high volume hospitals in 1994–1996: The northern New England experience. Northern New England Cardiovascular Disease Study Group. J Am Coll Cardiol. 1999;34:1471–1480[Abstract/Free Full Text]
This article has been cited by other articles:
|
|
|
|
 
M. Singh, B. J. Gersh, R. J. Lennon, H. H. Ting, D. R. Holmes Jr, B. J. Doyle, and C. S. Rihal
Outcomes of a System-Wide Protocol for Elective and Nonelective Coronary Angioplasty at Sites Without On-site Surgery: The Mayo Clinic Experience
Mayo Clin. Proc.,
June 1, 2009;
84(6):
501 - 508.
[Abstract]
[Full Text]
[PDF]
|
|
|
|
|
|
|
E Romagnoli, F Burzotta, C Trani, M Siviglia, G G L Biondi-Zoccai, G Niccoli, A M Leone, I Porto, M A Mazzari, R Mongiardo, et al.
EuroSCORE as predictor of in-hospital mortality after percutaneous coronary intervention
Heart,
January 1, 2009;
95(1):
43 - 48.
[Abstract]
[Full Text]
[PDF]
|
|
|
|
|
|
|
M. Singh, E. D. Peterson, S. Milford-Beland, J. S. Rumsfeld, and J. A. Spertus
Validation of the Mayo Clinic Risk Score for In-Hospital Mortality After Percutaneous Coronary Interventions Using the National Cardiovascular Data Registry
Circ Cardiovasc Intervent,
August 1, 2008;
1(1):
36 - 44.
[Abstract]
[Full Text]
[PDF]
|
|
|
|
|
|
|
C. E. Kurtz, Y. Gerber, S. A. Weston, M. M. Redfield, S. J. Jacobsen, and V. L. Roger
Use of Ejection Fraction Tests and Coronary Angiography in Patients With Heart Failure
Mayo Clin. Proc.,
July 1, 2006;
81(7):
906 - 913.
[Abstract]
[Full Text]
[PDF]
|
|
|
|
|
|
|
W. W. O'Neill, S. R. Dixon, and C. L. Grines
The year in interventional cardiology
J. Am. Coll. Cardiol.,
April 5, 2005;
45(7):
1117 - 1134.
[Full Text]
[PDF]
|
|
|
|
Top
Abstract
Materials and methods