- Email: [email protected]
Prediction of Complications Following Nonemergency Percutaneous Coronary Interventions
Prediction of Complications Following Nonemergency Percutaneous Coronary Interventions Mandeep Singh, MDa,*, Charanjit S. Rihal, MDa, Ryan J. Lennon, MSb, Kirk N. Garratt, MDa, Verghese Mathew, MDa, and David R. Holmes, Jr., MDa Previous models for prediction of complications after percutaneous coronary interventions (PCIs) have included in-hospital mortality and major in-hospital complications. In general, these models have excluded elevated cardiac biomarkers as a complication. We sought to determine whether a risk model could predict complications, including biomarker elevation, in patients undergoing nonemergency PCI. We examined the outcomes of nonemergency PCI performed on patients at Mayo Clinic from 2000 to 2003. The primary end point was in-hospital complications of death, myocardial infarction (MI) (Q-wave MI, or post-PCI creatine kinase-MB elevation >3 times the upper limit of normal), emergency coronary artery bypass grafting, or stroke. We used the Hosmer-Lemeshow test to demonstrate the adequacy of the model fit, and the c-index for discriminatory ability of the model. Of 2,894 nonemergency PCIs, the end point was noted in 232 (8%). The final prediction model included vein graft intervention (odds ratio [OR] 2.19), angiographic thrombus (OR 2.12), preprocedure stenosis of a minor (OR 1.98) or major (OR 1.62) side branch, and type C lesion (OR 1.48). The model had modest ability to discriminate between event and nonevent patients (c ⴝ 0.641). In the 500 bootstrap samples for internal validation, the c-index was 0.642 ⴞ 0.020, indicating only fair discriminatory ability. The average number of observed events was 232.0 ⴞ 14.7 compared with 232.1 ⴞ 2.5 expected events (average difference ⴚ0.06 ⴞ 14.5). In conclusion, the 5 risk variables associated with an increased risk of complications in patients undergoing elective PCI included vein graft intervention, presence of angiographic thrombus, stenosis of a major or minor side branch, and type C lesion; however, the discriminatory ability of the model derived from the variables was only modest. © 2005 Elsevier Inc. All rights reserved. (Am J Cardiol 2005;96:907–912)
The available models for the prediction of complications after percutaneous coronary interventions (PCIs) include mortality and other major adverse cardiovascular events (i.e. Q-wave myocardial infarction [MI], need for emergency coronary artery bypass grafting [CABG], and stroke).1– 8 Notably, the definition of MI has evolved, and the current American College of Cardiology/American Heart Association guidelines include biochemical evidence of myocardial necrosis in the definition of periprocedure MI.9 The markers of myocardial necrosis, as determined by elevated cardiac enzymes after PCI, are associated with a higher risk of death, subsequent MI, and the need for repeat revascularization procedures.10 –12 This important measure of complications is not available in the current risk models. Additionally, inclusion of patients with high-risk variables, such as cardiogenic shock, heart failure, and advancing age, likely dilute the adverse influence of cardiac biomarkers.13,14 No a Divisions of Cardiovascular Diseases and Internal Medicine and bBiostatistics, Mayo Clinic College of Medicine, Rochester, Minnesota. Manuscript received March 18, 2005; revised manuscript received and accepted May 12, 2005. * Corresponding author: Tel: 507-255-6092; fax: 507-255-2550. E-mail address: [email protected] (M. Singh).
0002-9149/05/$ – see front matter © 2005 Elsevier Inc. All rights reserved. doi:10.1016/j.amjcard.2005.05.045
studies have been done to date that have addressed the issue of the prediction of complications in patients undergoing nonemergency PCI. The aim of the present study, therefore, was to determine whether a risk model for procedural complications, including elevated cardiac biomarkers, can be developed for low- to intermediate-risk patients treated with conventional PCI techniques.
Methods Under a protocol approved by the institutional review board, a prospective interventional database has been maintained at the Mayo Clinic since 1979; it includes demographic, clinical, angiographic, and procedural information. The study group included all patients who underwent PCI from August 1, 2000 to October 31, 2003 at the Mayo Clinic. Patients with a normal baseline creatine kinase-MB fraction (CK-MB) were included for this analysis. The period from August 2000 was chosen because the CK-MB isoenzyme test became available at our institution. Patients with elevated baseline CK-MB, unknown baseline CK-MB, MI within 12 hours before PCI, including cardiogenic www.AJConline.org
908
The American Journal of Cardiology (www.AJConline.org)
Table 1 Clinical and angiographic characteristics Variable
Overall (n ⫽ 2,894)
Age (yrs), median (Q1, Q3) Women Myocardial infarction 12–24 h before PCI Unstable angina Procedure type Elective Urgent MI (1–7 d) MI (⬎7 d) Congestive heart failure on presentation NYHA class ⱖIII Moderate/severe renal disease Diabetes mellitus Hypertension Hyperlipidemia Current smoker Peripheral vascular disease Previous PCI Previous CABG Current/former smoker Multivessel disease Left main stenosis 50–69% ⱖ70% ACC/AHA lesion complexity A/B1 B2 C Thrombus in any lesion Calcium in any stenosis Ulcer in any lesion Vein graft intervention Proximal left anterior descending intervention Stent use Median No. of stents placed (Q1, Q3) Glycoprotein IIb/IIIa use Median No. of segments treated (Q1, Q3) Total No. vessels treated 1 2 3
68 (59, 76) 842 (29%) 37 (1%) 1,846 (64%) 1,353 (47%) 1,541 (53%) 293 (10%) 920 (33%) 293 (10%) 115 (4%) 116 (4%) 773 (27%) 2,082 (75%) 2,301 (86%) 389 (14%) 295 (11%) 895 (31%) 696 (24%) 1,850 (65%) 1,418 (52%) 65 (2%) 34 (1%) 677 (24%) 1,060 (38%) 1,072 (38%) 365 (14%) 983 (37%) 194 (7%) 245 (8%) 487 (17%) 2,585 (89%) 1.0 (1.0, 2.0) 1,604 (55%) 1.0 (1.0, 2.0) 2,333 (81%) 516 (18%) 45 (2%)
ACC/AHA ⫽ American College of Cardiology/American Heart Association; NYHA ⫽ New York Heart Association; Q ⫽ quartile.
shock, were excluded. Patients with an emergency indication for PCI were also excluded. The outcome of interest was major in-hospital complications defined as ⱖ1 of the following: death, Q-wave MI, urgent or emergency CABG during the index hospitalization, cerebrovascular accident or transient ischemic attack, and postprocedure CK-MB elevation. A CK-MB elevation was defined as CK-MB ⱖ3 times the upper limit of normal. CK-MB levels were measured at 8 and 16 hours after PCI. Some patients had ⬎2 measurements. In these patients, only the measurements closest to 8 and 16 hours were used. MI was considered to have occurred if either of the 2 measurements was elevated. Any in-hospital death occurring after PCI was considered related to the procedure. Lesion success was defined as achievement of a residual luminal diameter
Table 2 In-hospital complications after percutaneous coronary interventions (PCIs) Variable
Overall (n ⫽ 2,894)
Death Q-wave MI Emergency/urgent CABG Cerebrovascular accident/transient ischemic attacks Any of above 8-h Elevation 16-h elevation Any CK-MB elevation PCI complication
8 (0.3%) 36 (1%) 7 (0.2%) 7 (0.2%) 53 (2%) 71 (2%) 202 (7%) 209 (7%) 232 (8%)
stenosis of ⬍50%, including a ⱖ20% improvement by visual estimation. Procedural success was defined as ⱖ1 successful lesion without in-hospital death, Q-wave MI, stroke, emergency CABG, or CK-MB elevation, as defined previously. The lesions were classified according to the American College of Cardiology/American Heart Association lesion classification into types A, B1, B2, and C. Single-vessel disease was defined as ⬎70% luminal diameter stenosis in 1 major epicardial vessel and multivessel disease as ⬎70% stenosis in 1 major epicardial vessel and ⬎50% stenosis in ⱖ1 other major epicardial vessel. Statistical analysis: Continuous variables are summarized as means ⫾ SDs. Discrete variables are summarized as frequencies and percentages. Logistic regression analysis was used to estimate the odds ratios (ORs) and associated p values for the primary end point. All analyses were conducted using Statistical Analysis Systems, version 8.2, software (SAS Institute, Cary, North Carolina). Bootstrap data reduction was used to choose risk factors for the final model. Twenty-two variables (age with linear and quadratic components, gender, MI within 12 to 24 hours of PCI, urgent procedure, MI within 1 to 7 days before PCI, use of intra-aortic balloon pump, congestive heart failure on presentation, chronic renal disease, history of hyperlipidemia, history of MI ⬎7 days before PCI, multivessel disease, left main disease, type B2 or C lesion, type C lesion, presence of thrombus, calcified stenosis, ulcerated lesion, stenosis involving minor branches, stenosis involving major branches, vein graft intervention, and previous CABG) were considered candidates for the final model. We created 500 bootstrap samples from the study sample, and logistic regression analysis with backward variable selection at a 0.05 significance level was used to create models for the composite end point in each bootstrap sample. Variables included in ⱖ70% of the bootstrap models were chosen for the final model.15 Multicollinearity of explanatory variables was initially assessed by generalized variance inflation factors. After the bootstrap selection process, variables selected in ⬍70% of the samples were inspected to determine whether related variables might have competed for selection. If ⱖ70% of the sample analyses selected ⱖ1 of 2
Coronary Artery Disease/Prediction of Complications After Coronary Interventions
909
Table 3 Univariate predictors of procedural complications after percutaneous coronary interventions (PCIs) Variable
n
Complication (%)
OR
95% CI
Clinical Age (yrs) ⬍50 50–59 60–69 70–79 ⱖ80
232 567 803 919 373
7.8 6.9 7.3 8.2 11.0
1.00 0.88 0.94 1.06 1.47
Reference 0.49–1.57 0.54–1.63 0.62–1.81 0.82–2.62
2052 842 37 293 293 115 116
7.7 8.8 13.5 13.0 10.6 10.4 11.2
1.00 1.16 1.81 1.85 1.41 1.40 1.49
Reference 0.87–1.54 0.70–4.69 1.28–2.68 0.95–2.11 0.76–2.59 0.82–2.69
2053 296 55 15 52 773 2082 2301 389 295 920 895 696 1850
7.9 10.1 9.1 20.0 7.7 8.3 8.3 7.6 9.5 9.8 9.1 7.6 8.0 8.4
1.00 1.31 1.16 2.90 0.97 1.05 1.19 0.77 1.24 1.25 1.21 0.92 1.01 1.13
Reference 0.87–1.97 0.46–2.95 0.81–10.4 0.34–2.71 0.78–1.42 0.86–1.66 0.53–1.12 0.86–1.79 0.83–1.89 0.91–1.60 0.69–1.23 0.74–1.38 0.85–1.51
0.228 — — — — — 0.331 — — 0.255 0.002 0.101 0.303 0.210 0.455 — — — — — 0.739 0.291 0.176 0.268 0.292 0.189 0.577 0.948 0.392
1353 1541 12
6.8 9.1 25.0
1.00 1.37 3.86
Reference 1.04–1.80 1.04–14.4
0.023 — — 0.075
1418
9.7
1.58
1.19–2.10
2795 65 34
7.8 10.8 17.6
1.00 1.42 2.52
Reference 0.64–3.15 1.03–6.16
677 1060 1072 365 1096 554 983 194 245 487
5.0 7.0 11.0 15.1 10.0 9.6 9.6 13.4 14.7 8.4
1.00 1.42 2.34 2.43 1.48 1.29 1.48 1.90 2.16 1.07
Reference 0.93–2.16 1.58–3.47 1.75–3.37 1.13–1.95 0.94–1.79 1.11–1.97 1.23–2.95 1.47–3.16 0.75–1.52
Men Women MI 12–24 h before PCI MI 1–7 d before procedure Congestive heart failure on presentation NYHA class ⱖ3 Moderate/severe renal disease Creatinine (mg/dl) ⬍1.5 1.5–1.9 2.0–2.4 2.5–3.0 ⱖ3.0 Diabetes mellitus Hypertension Hyperlipidemia Current smoker Peripheral vascular disease MI ⬎7 d before Previous PCI Previous CABG Current/former smoker Procedural Procedure type Elective Urgent Prophylactic intra-aortic balloon pump Angiographic Multivessel disease (70/70) Left main stenosis ⬍50% 50–69% 70%⫹ ACC/AHA lesion complexity A/B1 B2 C Thrombus in any lesion Minor branches (any lesion) Major branches (any lesion) Calcium in any stenosis Ulcer in any lesion Vein graft intervention Proximal left anterior artery intervention
p Value
0.001 0.135 — — — ⬍0.001 — — — ⬍0.001 0.005 0.127 0.008 0.007 ⬍0.001 0.722
Abbreviations as in Table 1.
related variables, the variable that was selected most often would also be added to the model. Logistic regression analysis was used to fit the final model. The Hosmer-Lemeshow test was used to test the adequacy of the model fit to the data. The c-index was used to assess the discriminatory ability of the model. The c-index reflects the model’s ability
to identify the higher risk patient from a random pair of patients. Thus a c-index of 0.75 means that the model correctly identified the higher risk patient from a pair of patients 75% of the time. A c-index of 0.50 is equivalent to randomly (e.g., flipping a coin) assigning a higher risk to 1 patient instead of another. To create a simple additive scor-
910
The American Journal of Cardiology (www.AJConline.org)
Table 4 Multivariate predictors of procedural complications in patients undergoing nonemergency percutaneous coronary interventions (PCIs) Variable Vein graft treated Thrombus Stenosis of minor branches Stenosis of major branches ACC/AHA type C Intercept
Parameter Estimate
OR (95% CI)
p Value
Integer Score
0.7845 0.7526 0.6825 0.4820 0.3930 ⫺3.2511
2.19 (1.41–3.40) 2.12 (1.51–2.99) 1.98 (1.45–2.69) 1.62 (1.13–2.32) 1.48 (1.11–1.98)
⬍0.001 ⬍0.001 ⬍0.001 0.009 0.008
5 5 4 3 2
Overall model chi-square 57.2 on 4 degree of freedom; Hosmer-Lemeshow goodness-of-fit p ⫽ 0.97. Abbreviations as in Table 1.
ing algorithm, several simplified integer scores were investigated. For each, an integer was assigned to risk factors from the final model. A patients risk score is then the sum of the integers of the risk factors carried by the patient. The final integer system was selected such that the integers were proportional to the parameter estimates from the final model and that the potential summations could be easily conducted without the aid of a calculator. Logistic regression with the integer score of the study sample was used to estimate the expected risk for each score. Five hundred separate bootstrap samples were used to internally validate the simplified scoring system. The model was applied to the samples and the c-index, and observed and expected event rates were recorded. Results A total of 5,421 PCIs were performed in 4,690 patients treated at Mayo Clinic from August 1, 2000 to October 31, 2003. Of the 4,690 patients, 118 denied the use of their medical records for research, and were excluded. Of the remaining 4,572 patients, only the first PCI within the period was included in the study. Patients with elevated baseline CK-MB (n ⫽ 936), unknown baseline CK-MB (n ⫽ 267), MI within 12 hours before PCI, or the same day but time unknown (n ⫽ 287), emergency PCI (n ⫽ 160), or cardiogenic shock (n ⫽ 28) were excluded. The remaining 2,894 PCIs were included in the present analysis. Baseline clinical characteristics: The median patient age was 68 years (interquartile range 59 to 76), 71% were men, 64% had unstable angina, 293 (10%) had had an MI 1 to 7 days before the procedure, and 920 (33%) had had an MI ⬎7 days before PCI. The prevalence of other risk factors is listed in Table 1. Procedural outcome: The composite end point (i.e., any major complication) occurred in 232 procedures (8%) and included 8 deaths (0.3%), 36 Q-wave MIs (1%), 7 patients with stroke (0.2%), 7 (0.2%) with urgent/emergency CABG, and 209 (7%) with CK-MB elevation after PCI (Table 2). Clinical and angiographic correlates of procedural complications: Univariate associations between the baseline demographic characteristics and adverse outcomes, with ORs
and confidence intervals (CIs), are listed in Table 3 and 4. The significant correlates included urgent PCI and MI within 1 to 7 days. Angiographic variables significantly associated with procedural complications included multivessel disease, type C lesion, ulcerated or calcified lesions, and intervention on a vein graft lesion. Multivariable correlates: Only 4 variables were selected in ⱖ70% of the multivariate analyses of the bootstrap samples. The related variables, type C and type B2 or C lesions, were selected in 54.6% and 23.4% of the samples, respectively, but only simultaneously in 3.6%. Thus, ⱖ1 of these variables was present in 74% of the bootstrap sample models. A type C lesion was selected most often and, therefore, was included in the final model. Multicollinearity did not appear to affect the model building process for any other variables. In the original data, no explanatory variable had a variance inflation factor of ⬎3.0 (well below the threshold of concern of 10.0). The final model (fitted to the original data) contained the following variables (and estimates): vein graft intervention (OR 2.19, 95% CI 1.41 to 3.40), thrombus (OR 2.12, 95% CI 1.51 to 2.99), stenosis of a minor (OR 1.98, 95% CI 1.45 to 2.69) or major (OR 1.62, 95% CI 1.13 to 2.32) side branch before the procedure, and ⱖ1 type C lesion (OR 1.48, 95% CI 1.11 to 1.98) (Table 4). The model appeared to fit the data adequately (HosmerLemeshow test statistic 1.83 on 6 degrees of freedom, p ⫽ 0.93). It had only modest ability to discriminate between event and nonevent patients (c ⫽ 0.641). Figure 1 presents a simple scoring algorithm and relation between an increasing integer score and the expected complication risk from PCI. In the 500 bootstrap samples for internal validation, the mean c-index was 0.642 ⫾ 0.020, indicating less than the desired discriminatory ability. The average number of observed events was 232.0 ⫾ 14.7 compared with 232.1 ⫾ 2.5 expected events, with an average of difference of ⫺0.06 ⫾ 14.5. Discussion The principal findings from this study were that major adverse cardiovascular events in elective low- to intermediate-risk PCI are mainly driven by the elevation of cardiac biomarkers, with extremely low rates of other major adverse
Coronary Artery Disease/Prediction of Complications After Coronary Interventions
Figure 1. Expected complication rates for value of integer score.
cardiac events. The 5 risk variables associated with an increased risk of procedural complications included vein graft intervention, the presence of angiographic thrombus, stenosis of a major or minor side branch, and a type C lesion. However, the discriminatory ability of this model for the prediction of complications was modest. Currently, the drivers for major cardiovascular complications after PCI are largely the patient’s clinical characteristics, acuity of presentation, and left ventricular function.3,4,13,16 The incidence of in-hospital mortality and other major complications in the recent trials that excluded patients with acute MI was very low.17–20 The incidence of postprocedural MI (with the inclusion of cardiac biomarkers) was higher compared with that in previously published studies. The incidence of 8% was almost twice the 28-day MI rates observed in the Clopidogrel for Reduction of Events During Observation (CREDO) trial and was twice the 4% incidence of 30-day MI seen in the Intracoronary Stenting and Antithrombotic Regimen-Rapid Early Action for Coronary Treatment (ISAR-REACT) study.17,18 The lower incidence in these 2 studies was probably a result of the enrollment of lower risk patients with exclusion of patients with recent MI (2 weeks in the ISAR-REACT study and 24-hour ST elevation in the CREDO trial). A much higher incidence (17.2%) in CK-MB elevation was seen in a single-center experience with patients with nonacute MI with successful PCI and no Q-wave MI or need for emergency CABG.10 The CK-MB elevation (1 to 5 times normal) was defined differently in that study. The performance of directional atherectomy and the occurrence of coronary thromboembolism, such as is encountered more often in revascularization of saphenous vein grafts, are known strong predictors of CK-MB elevation after a successful procedure.11,21–24 Other variables included a history of recent infarction, the occurrence of minor procedural complications (e.g., transient in-laboratory closure, side-branch compromise, large dissections, hypotension re-
911
quiring intravenous pressors or intra-aortic balloon counterpulsation), a multivessel procedure, higher residual stenosis, complex lesions, and severe initial stenosis. Despite the positive influence of glycoprotein receptor inhibitors on postprocedure cardiac biomarkers, they were not considered a candidate variable because we wanted the model to use only clinical and angiographic variables. Additionally, we could not adjust for the confounding influence of the physician’s preference and the timing of these agents with regard to the PCI procedure (before, during, or after). The results of our study support the conclusion of previous studies in that we found similar predictors for the elevation of cardiac biomarkers. However, the age group in the present study was older. The limitations of our study were that the prediction of the elevation of these biomarkers from the available demographic and angiographic variables is modest. Second, because of the low number of patients, this study was underpowered to predict major adverse cardiovascular events after elective PCI. This is in contrast to the availability of excellent predictive models for in-hospital mortality and major adverse in-hospital cardiovascular events that include patients with acute coronary syndromes with higher event rates. The adverse events in these models occur as a result of the high-risk characteristics of the patients and their presentation. Consequently, the risk variables associated with high mortality are acute MI, advancing age, and left ventricular dysfunction, including cardiogenic shock. Once these high-risk characteristics were removed from the model, the variables linked to the elevation of cardiac biomarkers had only modest discriminatory ability to predict this complication. No risk models are yet available to address the prediction of complications in patients undergoing nonemergency PCI. The results of the present study are in line with the modest discriminatory ability of the previously published Mayo Clinic Risk Score model in an elective subgroup of patients.7,8 In the multivariate analysis reported by the American College of Cardiology-National Cardiovascular Data Registry (ACCNCDR), lesion variables were prominent in the analysis of lower risk patients not presenting with MI.16 In a recent comparison of the Mayo Clinic risk model with the American College of Cardiology/American Heart Association lesion classification, the latter predicted for angiographic failure better; however, the major procedural complications were better predicted by the Mayo model.25 More recently, in a new classification system, 2 variables (nonchronic total occlusion and degenerated saphenous vein grafts) correlated significantly with death, CK-MB elevation, and the need for emergency CABG.26 Third, our study had the limitations of a retrospective analysis. The current analysis was derived from the data set of a single referral center with high-volume operators, and broader applicability is open to question. Operator volume and technical details such as direct stenting were not considered in the present analysis. As PCI procedures become safer, the discriminatory ability of any model may be modest in the low-
912
The American Journal of Cardiology (www.AJConline.org)
to intermediate-risk population due to the extremely low number of events. 1. Holmes DR Jr, Berger PB, Garratt KN, Mathew V, Bell MR, Barsness GW, Higano ST, Grill DE, Hammes LN, Rihal CS. Application of the New York State PTCA mortality model in patients undergoing stent implantation. Circulation 2000;102:517–522. 2. Holmes DR, Selzer F, Johnston JM, Kelsey SF, Holubkov R, Cohen HA, Williams DO, Detre KM. Modeling and risk prediction in the current era of interventional cardiology: a report from the National Heart, Lung, and Blood Institute Dynamic Registry. Circulation 2003; 107:1871–1876. 3. Moscucci M, Kline-Rogers E, Share D, O’Donnell M, Maxwell-Eward A, Meengs WL, Kraft P, DeFranco AC, Chambers JL, Patel K, McGinnity JG, Eagle KA. Simple bedside additive tool for prediction of in-hospital mortality after percutaneous coronary interventions. Circulation 2001;104:263–268. 4. O’Connor GT, Malenka DJ, Quinton H, Robb JF, Kellett MA Jr, Shubrooks S, Bradley WA, Hearne MJ, Watkins MW, Wennberg DE, Hettleman B, et al, for the Northern New England Cardiovascular Disease Study Group. Multivariate prediction of in-hospital mortality after percutaneous coronary interventions in 1994 –1996. J Am Coll Cardiol 1999;34:681– 691. 5. Qureshi MA, Safian RD, Grines CL, Goldstein JA, Westveer DC, Glazier S, Balasubramanian M, O’Neill WW. Simplified scoring system for predicting mortality after percutaneous coronary intervention. J Am Coll Cardiol 2003;42:1890 –1895. 6. Resnic FS, Ohno-Machado L, Selwyn A, Simon DI, Popma JJ. Simplified risk score models accurately predict the risk of major inhospital complications following percutaneous coronary intervention. Am J Cardiol 2001;88:5–9. 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. 8. Singh M, Rihal CS, Selzer F, Kip KE, Detre K, Holmes DR. Validation of Mayo Clinic risk adjustment model for in-hospital complications after percutaneous coronary interventions, using the National Heart, Lung, and Blood Institute dynamic registry. J Am Coll Cardiol 2003;42:1722–1728. 9. Smith SC Jr, Dove JT, Jacobs AK, Kennedy JW, Kereiakes D, Kern MJ, Kuntz RE, Popma JJ, Schaff HV, Williams DO, et al. ACC/AHA guidelines of percutaneous coronary interventions (revision of the 1993 PTCA guidelines)— executive summary: a report of the American College of Cardiology/American Heart Association Task Force on Practice Guidelines (committee to revise the 1993 guidelines for percutaneous transluminal coronary angioplasty). J Am Coll Cardiol 2001;37:2215–2239. 10. 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. 11. Abdelmeguid AE, Topol EJ, Whitlow PL, Sapp SK, Ellis SG. Significance of mild transient release of creatine kinase-MB fraction after percutaneous coronary interventions. Circulation 1996;94:1528 –1536. 12. Harrington RA. Cardiac enzyme elevations after percutaneous coronary intervention: myonecrosis, the coronary microcirculation and mortality. J Am Coll Cardiol 2000;35:1142–1144.
13. Hochman JS, Sleeper LA, Webb JG, Sanborn TA, White HD, Talley JD, Buller CE, Jacobs AK, Slater JN, Col J, McKinlay SM, LeJemtel TH, for the SHOCK Investigators. Early revascularization in acute myocardial infarction complicated by cardiogenic shock: should we emergently revascularize occluded coronaries for cardiogenic shock. N Engl J Med 1999;341:625– 634. 14. Singh M, Mathew V, Garratt KN, Berger PB, Grill DE, Bell MR, Rihal CS, Holmes DR Jr. 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. 15. Sauerbrei W, Schumacher M. A bootstrap resampling procedure for model building: application to the Cox regression model. Stat Med 1992;11:2093–2109. 16. Shaw RE, Anderson HV, Brindis RG, Krone RJ, Klein LW, McKay CR, Block PC, Shaw LJ, Hewitt K, Weintraub WS. 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. 17. Kastrati A, Mehilli J, Schuhlen H, Dirschinger J, Dotzer F, Ten Berg JM, Neumann FJ, Bollwein H, Volmer C, Gawaz M, Berger PB, Schomig A. A clinical trial of abciximab in elective percutaneous coronary intervention after pretreatment with clopidogrel. N Engl J Med 2004;350:232–238. 18. Steinhubl SR, Berger PB, Mann JT III, Fry ET, DeLago A, Wilmer C, Topol EJ. Early and sustained dual oral antiplatelet therapy following percutaneous coronary intervention: a randomized controlled trial. JAMA 2002;288:2411–2420. 19. Moses JW, Leon MB, Popma JJ, Fitzgerald PJ, Holmes DR, O’Shaughnessy C, Caputo RP, Kereiakes DJ, Williams DO, Teirstein PS, Jaeger JL, Kuntz RE. Sirolimus-eluting stents versus standard stents in patients with stenosis in a native coronary artery. N Engl J Med 2003;349:1315–1323. 20. Stone GW, Ellis SG, Cox DA, Hermiller J, O’Shaughnessy C, Mann JT, Turco M, Caputo R, Bergin P, Greenberg J, Popma JJ, Russell ME. A polymer-based, paclitaxel-eluting stent in patients with coronary artery disease. N Engl J Med 2004;350:221–231. 21. Abdelmeguid AE, Whitlow PL, Sapp SK, Ellis SG, Topol EJ. Longterm outcome of transient, uncomplicated in-laboratory coronary artery closure. Circulation 1995;91:2733–2741. 22. Abdelmeguid AE, Topol EJ. The myth of the myocardial “infarctlet” during percutaneous coronary revascularization procedures. Circulation 1996;94:3369 –3375. 23. Abdelmeguid AE, Ellis SG, Sapp SK, Whitlow PL, Topol EJ. Defining the appropriate threshold of creatine kinase elevation after percutaneous coronary interventions. Am Heart J 1996;131:1097–1105. 24. Harrington RA, Lincoff AM, Califf RM, Holmes DR Jr, Berdan LG, O’Hanesian MA, Keeler GP, Garratt KN, Ohman EM, Mark DB, et al. Characteristics and consequences of myocardial infarction after percutaneous coronary intervention: insights from the Coronary Angioplasty Versus Excisional Atherectomy Trial (CAVEAT). J Am Coll Cardiol 1995;25:1693–1699. 25. Singh M, Rihal CS, Lennon RJ, Garratt KN, Holmes DR Jr. 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. J Am Coll Cardiol 2004;44:357–361. 26. 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.
The available models for the prediction of complications after percutaneous coronary interventions (PCIs) include mortality and other major adverse cardiovascular events (i.e. Q-wave myocardial infarction [MI], need for emergency coronary artery bypass grafting [CABG], and stroke).1– 8 Notably, the definition of MI has evolved, and the current American College of Cardiology/American Heart Association guidelines include biochemical evidence of myocardial necrosis in the definition of periprocedure MI.9 The markers of myocardial necrosis, as determined by elevated cardiac enzymes after PCI, are associated with a higher risk of death, subsequent MI, and the need for repeat revascularization procedures.10 –12 This important measure of complications is not available in the current risk models. Additionally, inclusion of patients with high-risk variables, such as cardiogenic shock, heart failure, and advancing age, likely dilute the adverse influence of cardiac biomarkers.13,14 No a Divisions of Cardiovascular Diseases and Internal Medicine and bBiostatistics, Mayo Clinic College of Medicine, Rochester, Minnesota. Manuscript received March 18, 2005; revised manuscript received and accepted May 12, 2005. * Corresponding author: Tel: 507-255-6092; fax: 507-255-2550. E-mail address: [email protected] (M. Singh).
0002-9149/05/$ – see front matter © 2005 Elsevier Inc. All rights reserved. doi:10.1016/j.amjcard.2005.05.045
studies have been done to date that have addressed the issue of the prediction of complications in patients undergoing nonemergency PCI. The aim of the present study, therefore, was to determine whether a risk model for procedural complications, including elevated cardiac biomarkers, can be developed for low- to intermediate-risk patients treated with conventional PCI techniques.
Methods Under a protocol approved by the institutional review board, a prospective interventional database has been maintained at the Mayo Clinic since 1979; it includes demographic, clinical, angiographic, and procedural information. The study group included all patients who underwent PCI from August 1, 2000 to October 31, 2003 at the Mayo Clinic. Patients with a normal baseline creatine kinase-MB fraction (CK-MB) were included for this analysis. The period from August 2000 was chosen because the CK-MB isoenzyme test became available at our institution. Patients with elevated baseline CK-MB, unknown baseline CK-MB, MI within 12 hours before PCI, including cardiogenic www.AJConline.org
908
The American Journal of Cardiology (www.AJConline.org)
Table 1 Clinical and angiographic characteristics Variable
Overall (n ⫽ 2,894)
Age (yrs), median (Q1, Q3) Women Myocardial infarction 12–24 h before PCI Unstable angina Procedure type Elective Urgent MI (1–7 d) MI (⬎7 d) Congestive heart failure on presentation NYHA class ⱖIII Moderate/severe renal disease Diabetes mellitus Hypertension Hyperlipidemia Current smoker Peripheral vascular disease Previous PCI Previous CABG Current/former smoker Multivessel disease Left main stenosis 50–69% ⱖ70% ACC/AHA lesion complexity A/B1 B2 C Thrombus in any lesion Calcium in any stenosis Ulcer in any lesion Vein graft intervention Proximal left anterior descending intervention Stent use Median No. of stents placed (Q1, Q3) Glycoprotein IIb/IIIa use Median No. of segments treated (Q1, Q3) Total No. vessels treated 1 2 3
68 (59, 76) 842 (29%) 37 (1%) 1,846 (64%) 1,353 (47%) 1,541 (53%) 293 (10%) 920 (33%) 293 (10%) 115 (4%) 116 (4%) 773 (27%) 2,082 (75%) 2,301 (86%) 389 (14%) 295 (11%) 895 (31%) 696 (24%) 1,850 (65%) 1,418 (52%) 65 (2%) 34 (1%) 677 (24%) 1,060 (38%) 1,072 (38%) 365 (14%) 983 (37%) 194 (7%) 245 (8%) 487 (17%) 2,585 (89%) 1.0 (1.0, 2.0) 1,604 (55%) 1.0 (1.0, 2.0) 2,333 (81%) 516 (18%) 45 (2%)
ACC/AHA ⫽ American College of Cardiology/American Heart Association; NYHA ⫽ New York Heart Association; Q ⫽ quartile.
shock, were excluded. Patients with an emergency indication for PCI were also excluded. The outcome of interest was major in-hospital complications defined as ⱖ1 of the following: death, Q-wave MI, urgent or emergency CABG during the index hospitalization, cerebrovascular accident or transient ischemic attack, and postprocedure CK-MB elevation. A CK-MB elevation was defined as CK-MB ⱖ3 times the upper limit of normal. CK-MB levels were measured at 8 and 16 hours after PCI. Some patients had ⬎2 measurements. In these patients, only the measurements closest to 8 and 16 hours were used. MI was considered to have occurred if either of the 2 measurements was elevated. Any in-hospital death occurring after PCI was considered related to the procedure. Lesion success was defined as achievement of a residual luminal diameter
Table 2 In-hospital complications after percutaneous coronary interventions (PCIs) Variable
Overall (n ⫽ 2,894)
Death Q-wave MI Emergency/urgent CABG Cerebrovascular accident/transient ischemic attacks Any of above 8-h Elevation 16-h elevation Any CK-MB elevation PCI complication
8 (0.3%) 36 (1%) 7 (0.2%) 7 (0.2%) 53 (2%) 71 (2%) 202 (7%) 209 (7%) 232 (8%)
stenosis of ⬍50%, including a ⱖ20% improvement by visual estimation. Procedural success was defined as ⱖ1 successful lesion without in-hospital death, Q-wave MI, stroke, emergency CABG, or CK-MB elevation, as defined previously. The lesions were classified according to the American College of Cardiology/American Heart Association lesion classification into types A, B1, B2, and C. Single-vessel disease was defined as ⬎70% luminal diameter stenosis in 1 major epicardial vessel and multivessel disease as ⬎70% stenosis in 1 major epicardial vessel and ⬎50% stenosis in ⱖ1 other major epicardial vessel. Statistical analysis: Continuous variables are summarized as means ⫾ SDs. Discrete variables are summarized as frequencies and percentages. Logistic regression analysis was used to estimate the odds ratios (ORs) and associated p values for the primary end point. All analyses were conducted using Statistical Analysis Systems, version 8.2, software (SAS Institute, Cary, North Carolina). Bootstrap data reduction was used to choose risk factors for the final model. Twenty-two variables (age with linear and quadratic components, gender, MI within 12 to 24 hours of PCI, urgent procedure, MI within 1 to 7 days before PCI, use of intra-aortic balloon pump, congestive heart failure on presentation, chronic renal disease, history of hyperlipidemia, history of MI ⬎7 days before PCI, multivessel disease, left main disease, type B2 or C lesion, type C lesion, presence of thrombus, calcified stenosis, ulcerated lesion, stenosis involving minor branches, stenosis involving major branches, vein graft intervention, and previous CABG) were considered candidates for the final model. We created 500 bootstrap samples from the study sample, and logistic regression analysis with backward variable selection at a 0.05 significance level was used to create models for the composite end point in each bootstrap sample. Variables included in ⱖ70% of the bootstrap models were chosen for the final model.15 Multicollinearity of explanatory variables was initially assessed by generalized variance inflation factors. After the bootstrap selection process, variables selected in ⬍70% of the samples were inspected to determine whether related variables might have competed for selection. If ⱖ70% of the sample analyses selected ⱖ1 of 2
Coronary Artery Disease/Prediction of Complications After Coronary Interventions
909
Table 3 Univariate predictors of procedural complications after percutaneous coronary interventions (PCIs) Variable
n
Complication (%)
OR
95% CI
Clinical Age (yrs) ⬍50 50–59 60–69 70–79 ⱖ80
232 567 803 919 373
7.8 6.9 7.3 8.2 11.0
1.00 0.88 0.94 1.06 1.47
Reference 0.49–1.57 0.54–1.63 0.62–1.81 0.82–2.62
2052 842 37 293 293 115 116
7.7 8.8 13.5 13.0 10.6 10.4 11.2
1.00 1.16 1.81 1.85 1.41 1.40 1.49
Reference 0.87–1.54 0.70–4.69 1.28–2.68 0.95–2.11 0.76–2.59 0.82–2.69
2053 296 55 15 52 773 2082 2301 389 295 920 895 696 1850
7.9 10.1 9.1 20.0 7.7 8.3 8.3 7.6 9.5 9.8 9.1 7.6 8.0 8.4
1.00 1.31 1.16 2.90 0.97 1.05 1.19 0.77 1.24 1.25 1.21 0.92 1.01 1.13
Reference 0.87–1.97 0.46–2.95 0.81–10.4 0.34–2.71 0.78–1.42 0.86–1.66 0.53–1.12 0.86–1.79 0.83–1.89 0.91–1.60 0.69–1.23 0.74–1.38 0.85–1.51
0.228 — — — — — 0.331 — — 0.255 0.002 0.101 0.303 0.210 0.455 — — — — — 0.739 0.291 0.176 0.268 0.292 0.189 0.577 0.948 0.392
1353 1541 12
6.8 9.1 25.0
1.00 1.37 3.86
Reference 1.04–1.80 1.04–14.4
0.023 — — 0.075
1418
9.7
1.58
1.19–2.10
2795 65 34
7.8 10.8 17.6
1.00 1.42 2.52
Reference 0.64–3.15 1.03–6.16
677 1060 1072 365 1096 554 983 194 245 487
5.0 7.0 11.0 15.1 10.0 9.6 9.6 13.4 14.7 8.4
1.00 1.42 2.34 2.43 1.48 1.29 1.48 1.90 2.16 1.07
Reference 0.93–2.16 1.58–3.47 1.75–3.37 1.13–1.95 0.94–1.79 1.11–1.97 1.23–2.95 1.47–3.16 0.75–1.52
Men Women MI 12–24 h before PCI MI 1–7 d before procedure Congestive heart failure on presentation NYHA class ⱖ3 Moderate/severe renal disease Creatinine (mg/dl) ⬍1.5 1.5–1.9 2.0–2.4 2.5–3.0 ⱖ3.0 Diabetes mellitus Hypertension Hyperlipidemia Current smoker Peripheral vascular disease MI ⬎7 d before Previous PCI Previous CABG Current/former smoker Procedural Procedure type Elective Urgent Prophylactic intra-aortic balloon pump Angiographic Multivessel disease (70/70) Left main stenosis ⬍50% 50–69% 70%⫹ ACC/AHA lesion complexity A/B1 B2 C Thrombus in any lesion Minor branches (any lesion) Major branches (any lesion) Calcium in any stenosis Ulcer in any lesion Vein graft intervention Proximal left anterior artery intervention
p Value
0.001 0.135 — — — ⬍0.001 — — — ⬍0.001 0.005 0.127 0.008 0.007 ⬍0.001 0.722
Abbreviations as in Table 1.
related variables, the variable that was selected most often would also be added to the model. Logistic regression analysis was used to fit the final model. The Hosmer-Lemeshow test was used to test the adequacy of the model fit to the data. The c-index was used to assess the discriminatory ability of the model. The c-index reflects the model’s ability
to identify the higher risk patient from a random pair of patients. Thus a c-index of 0.75 means that the model correctly identified the higher risk patient from a pair of patients 75% of the time. A c-index of 0.50 is equivalent to randomly (e.g., flipping a coin) assigning a higher risk to 1 patient instead of another. To create a simple additive scor-
910
The American Journal of Cardiology (www.AJConline.org)
Table 4 Multivariate predictors of procedural complications in patients undergoing nonemergency percutaneous coronary interventions (PCIs) Variable Vein graft treated Thrombus Stenosis of minor branches Stenosis of major branches ACC/AHA type C Intercept
Parameter Estimate
OR (95% CI)
p Value
Integer Score
0.7845 0.7526 0.6825 0.4820 0.3930 ⫺3.2511
2.19 (1.41–3.40) 2.12 (1.51–2.99) 1.98 (1.45–2.69) 1.62 (1.13–2.32) 1.48 (1.11–1.98)
⬍0.001 ⬍0.001 ⬍0.001 0.009 0.008
5 5 4 3 2
Overall model chi-square 57.2 on 4 degree of freedom; Hosmer-Lemeshow goodness-of-fit p ⫽ 0.97. Abbreviations as in Table 1.
ing algorithm, several simplified integer scores were investigated. For each, an integer was assigned to risk factors from the final model. A patients risk score is then the sum of the integers of the risk factors carried by the patient. The final integer system was selected such that the integers were proportional to the parameter estimates from the final model and that the potential summations could be easily conducted without the aid of a calculator. Logistic regression with the integer score of the study sample was used to estimate the expected risk for each score. Five hundred separate bootstrap samples were used to internally validate the simplified scoring system. The model was applied to the samples and the c-index, and observed and expected event rates were recorded. Results A total of 5,421 PCIs were performed in 4,690 patients treated at Mayo Clinic from August 1, 2000 to October 31, 2003. Of the 4,690 patients, 118 denied the use of their medical records for research, and were excluded. Of the remaining 4,572 patients, only the first PCI within the period was included in the study. Patients with elevated baseline CK-MB (n ⫽ 936), unknown baseline CK-MB (n ⫽ 267), MI within 12 hours before PCI, or the same day but time unknown (n ⫽ 287), emergency PCI (n ⫽ 160), or cardiogenic shock (n ⫽ 28) were excluded. The remaining 2,894 PCIs were included in the present analysis. Baseline clinical characteristics: The median patient age was 68 years (interquartile range 59 to 76), 71% were men, 64% had unstable angina, 293 (10%) had had an MI 1 to 7 days before the procedure, and 920 (33%) had had an MI ⬎7 days before PCI. The prevalence of other risk factors is listed in Table 1. Procedural outcome: The composite end point (i.e., any major complication) occurred in 232 procedures (8%) and included 8 deaths (0.3%), 36 Q-wave MIs (1%), 7 patients with stroke (0.2%), 7 (0.2%) with urgent/emergency CABG, and 209 (7%) with CK-MB elevation after PCI (Table 2). Clinical and angiographic correlates of procedural complications: Univariate associations between the baseline demographic characteristics and adverse outcomes, with ORs
and confidence intervals (CIs), are listed in Table 3 and 4. The significant correlates included urgent PCI and MI within 1 to 7 days. Angiographic variables significantly associated with procedural complications included multivessel disease, type C lesion, ulcerated or calcified lesions, and intervention on a vein graft lesion. Multivariable correlates: Only 4 variables were selected in ⱖ70% of the multivariate analyses of the bootstrap samples. The related variables, type C and type B2 or C lesions, were selected in 54.6% and 23.4% of the samples, respectively, but only simultaneously in 3.6%. Thus, ⱖ1 of these variables was present in 74% of the bootstrap sample models. A type C lesion was selected most often and, therefore, was included in the final model. Multicollinearity did not appear to affect the model building process for any other variables. In the original data, no explanatory variable had a variance inflation factor of ⬎3.0 (well below the threshold of concern of 10.0). The final model (fitted to the original data) contained the following variables (and estimates): vein graft intervention (OR 2.19, 95% CI 1.41 to 3.40), thrombus (OR 2.12, 95% CI 1.51 to 2.99), stenosis of a minor (OR 1.98, 95% CI 1.45 to 2.69) or major (OR 1.62, 95% CI 1.13 to 2.32) side branch before the procedure, and ⱖ1 type C lesion (OR 1.48, 95% CI 1.11 to 1.98) (Table 4). The model appeared to fit the data adequately (HosmerLemeshow test statistic 1.83 on 6 degrees of freedom, p ⫽ 0.93). It had only modest ability to discriminate between event and nonevent patients (c ⫽ 0.641). Figure 1 presents a simple scoring algorithm and relation between an increasing integer score and the expected complication risk from PCI. In the 500 bootstrap samples for internal validation, the mean c-index was 0.642 ⫾ 0.020, indicating less than the desired discriminatory ability. The average number of observed events was 232.0 ⫾ 14.7 compared with 232.1 ⫾ 2.5 expected events, with an average of difference of ⫺0.06 ⫾ 14.5. Discussion The principal findings from this study were that major adverse cardiovascular events in elective low- to intermediate-risk PCI are mainly driven by the elevation of cardiac biomarkers, with extremely low rates of other major adverse
Coronary Artery Disease/Prediction of Complications After Coronary Interventions
Figure 1. Expected complication rates for value of integer score.
cardiac events. The 5 risk variables associated with an increased risk of procedural complications included vein graft intervention, the presence of angiographic thrombus, stenosis of a major or minor side branch, and a type C lesion. However, the discriminatory ability of this model for the prediction of complications was modest. Currently, the drivers for major cardiovascular complications after PCI are largely the patient’s clinical characteristics, acuity of presentation, and left ventricular function.3,4,13,16 The incidence of in-hospital mortality and other major complications in the recent trials that excluded patients with acute MI was very low.17–20 The incidence of postprocedural MI (with the inclusion of cardiac biomarkers) was higher compared with that in previously published studies. The incidence of 8% was almost twice the 28-day MI rates observed in the Clopidogrel for Reduction of Events During Observation (CREDO) trial and was twice the 4% incidence of 30-day MI seen in the Intracoronary Stenting and Antithrombotic Regimen-Rapid Early Action for Coronary Treatment (ISAR-REACT) study.17,18 The lower incidence in these 2 studies was probably a result of the enrollment of lower risk patients with exclusion of patients with recent MI (2 weeks in the ISAR-REACT study and 24-hour ST elevation in the CREDO trial). A much higher incidence (17.2%) in CK-MB elevation was seen in a single-center experience with patients with nonacute MI with successful PCI and no Q-wave MI or need for emergency CABG.10 The CK-MB elevation (1 to 5 times normal) was defined differently in that study. The performance of directional atherectomy and the occurrence of coronary thromboembolism, such as is encountered more often in revascularization of saphenous vein grafts, are known strong predictors of CK-MB elevation after a successful procedure.11,21–24 Other variables included a history of recent infarction, the occurrence of minor procedural complications (e.g., transient in-laboratory closure, side-branch compromise, large dissections, hypotension re-
911
quiring intravenous pressors or intra-aortic balloon counterpulsation), a multivessel procedure, higher residual stenosis, complex lesions, and severe initial stenosis. Despite the positive influence of glycoprotein receptor inhibitors on postprocedure cardiac biomarkers, they were not considered a candidate variable because we wanted the model to use only clinical and angiographic variables. Additionally, we could not adjust for the confounding influence of the physician’s preference and the timing of these agents with regard to the PCI procedure (before, during, or after). The results of our study support the conclusion of previous studies in that we found similar predictors for the elevation of cardiac biomarkers. However, the age group in the present study was older. The limitations of our study were that the prediction of the elevation of these biomarkers from the available demographic and angiographic variables is modest. Second, because of the low number of patients, this study was underpowered to predict major adverse cardiovascular events after elective PCI. This is in contrast to the availability of excellent predictive models for in-hospital mortality and major adverse in-hospital cardiovascular events that include patients with acute coronary syndromes with higher event rates. The adverse events in these models occur as a result of the high-risk characteristics of the patients and their presentation. Consequently, the risk variables associated with high mortality are acute MI, advancing age, and left ventricular dysfunction, including cardiogenic shock. Once these high-risk characteristics were removed from the model, the variables linked to the elevation of cardiac biomarkers had only modest discriminatory ability to predict this complication. No risk models are yet available to address the prediction of complications in patients undergoing nonemergency PCI. The results of the present study are in line with the modest discriminatory ability of the previously published Mayo Clinic Risk Score model in an elective subgroup of patients.7,8 In the multivariate analysis reported by the American College of Cardiology-National Cardiovascular Data Registry (ACCNCDR), lesion variables were prominent in the analysis of lower risk patients not presenting with MI.16 In a recent comparison of the Mayo Clinic risk model with the American College of Cardiology/American Heart Association lesion classification, the latter predicted for angiographic failure better; however, the major procedural complications were better predicted by the Mayo model.25 More recently, in a new classification system, 2 variables (nonchronic total occlusion and degenerated saphenous vein grafts) correlated significantly with death, CK-MB elevation, and the need for emergency CABG.26 Third, our study had the limitations of a retrospective analysis. The current analysis was derived from the data set of a single referral center with high-volume operators, and broader applicability is open to question. Operator volume and technical details such as direct stenting were not considered in the present analysis. As PCI procedures become safer, the discriminatory ability of any model may be modest in the low-
912
The American Journal of Cardiology (www.AJConline.org)
to intermediate-risk population due to the extremely low number of events. 1. Holmes DR Jr, Berger PB, Garratt KN, Mathew V, Bell MR, Barsness GW, Higano ST, Grill DE, Hammes LN, Rihal CS. Application of the New York State PTCA mortality model in patients undergoing stent implantation. Circulation 2000;102:517–522. 2. Holmes DR, Selzer F, Johnston JM, Kelsey SF, Holubkov R, Cohen HA, Williams DO, Detre KM. Modeling and risk prediction in the current era of interventional cardiology: a report from the National Heart, Lung, and Blood Institute Dynamic Registry. Circulation 2003; 107:1871–1876. 3. Moscucci M, Kline-Rogers E, Share D, O’Donnell M, Maxwell-Eward A, Meengs WL, Kraft P, DeFranco AC, Chambers JL, Patel K, McGinnity JG, Eagle KA. Simple bedside additive tool for prediction of in-hospital mortality after percutaneous coronary interventions. Circulation 2001;104:263–268. 4. O’Connor GT, Malenka DJ, Quinton H, Robb JF, Kellett MA Jr, Shubrooks S, Bradley WA, Hearne MJ, Watkins MW, Wennberg DE, Hettleman B, et al, for the Northern New England Cardiovascular Disease Study Group. Multivariate prediction of in-hospital mortality after percutaneous coronary interventions in 1994 –1996. J Am Coll Cardiol 1999;34:681– 691. 5. Qureshi MA, Safian RD, Grines CL, Goldstein JA, Westveer DC, Glazier S, Balasubramanian M, O’Neill WW. Simplified scoring system for predicting mortality after percutaneous coronary intervention. J Am Coll Cardiol 2003;42:1890 –1895. 6. Resnic FS, Ohno-Machado L, Selwyn A, Simon DI, Popma JJ. Simplified risk score models accurately predict the risk of major inhospital complications following percutaneous coronary intervention. Am J Cardiol 2001;88:5–9. 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. 8. Singh M, Rihal CS, Selzer F, Kip KE, Detre K, Holmes DR. Validation of Mayo Clinic risk adjustment model for in-hospital complications after percutaneous coronary interventions, using the National Heart, Lung, and Blood Institute dynamic registry. J Am Coll Cardiol 2003;42:1722–1728. 9. Smith SC Jr, Dove JT, Jacobs AK, Kennedy JW, Kereiakes D, Kern MJ, Kuntz RE, Popma JJ, Schaff HV, Williams DO, et al. ACC/AHA guidelines of percutaneous coronary interventions (revision of the 1993 PTCA guidelines)— executive summary: a report of the American College of Cardiology/American Heart Association Task Force on Practice Guidelines (committee to revise the 1993 guidelines for percutaneous transluminal coronary angioplasty). J Am Coll Cardiol 2001;37:2215–2239. 10. 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. 11. Abdelmeguid AE, Topol EJ, Whitlow PL, Sapp SK, Ellis SG. Significance of mild transient release of creatine kinase-MB fraction after percutaneous coronary interventions. Circulation 1996;94:1528 –1536. 12. Harrington RA. Cardiac enzyme elevations after percutaneous coronary intervention: myonecrosis, the coronary microcirculation and mortality. J Am Coll Cardiol 2000;35:1142–1144.
13. Hochman JS, Sleeper LA, Webb JG, Sanborn TA, White HD, Talley JD, Buller CE, Jacobs AK, Slater JN, Col J, McKinlay SM, LeJemtel TH, for the SHOCK Investigators. Early revascularization in acute myocardial infarction complicated by cardiogenic shock: should we emergently revascularize occluded coronaries for cardiogenic shock. N Engl J Med 1999;341:625– 634. 14. Singh M, Mathew V, Garratt KN, Berger PB, Grill DE, Bell MR, Rihal CS, Holmes DR Jr. 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. 15. Sauerbrei W, Schumacher M. A bootstrap resampling procedure for model building: application to the Cox regression model. Stat Med 1992;11:2093–2109. 16. Shaw RE, Anderson HV, Brindis RG, Krone RJ, Klein LW, McKay CR, Block PC, Shaw LJ, Hewitt K, Weintraub WS. 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. 17. Kastrati A, Mehilli J, Schuhlen H, Dirschinger J, Dotzer F, Ten Berg JM, Neumann FJ, Bollwein H, Volmer C, Gawaz M, Berger PB, Schomig A. A clinical trial of abciximab in elective percutaneous coronary intervention after pretreatment with clopidogrel. N Engl J Med 2004;350:232–238. 18. Steinhubl SR, Berger PB, Mann JT III, Fry ET, DeLago A, Wilmer C, Topol EJ. Early and sustained dual oral antiplatelet therapy following percutaneous coronary intervention: a randomized controlled trial. JAMA 2002;288:2411–2420. 19. Moses JW, Leon MB, Popma JJ, Fitzgerald PJ, Holmes DR, O’Shaughnessy C, Caputo RP, Kereiakes DJ, Williams DO, Teirstein PS, Jaeger JL, Kuntz RE. Sirolimus-eluting stents versus standard stents in patients with stenosis in a native coronary artery. N Engl J Med 2003;349:1315–1323. 20. Stone GW, Ellis SG, Cox DA, Hermiller J, O’Shaughnessy C, Mann JT, Turco M, Caputo R, Bergin P, Greenberg J, Popma JJ, Russell ME. A polymer-based, paclitaxel-eluting stent in patients with coronary artery disease. N Engl J Med 2004;350:221–231. 21. Abdelmeguid AE, Whitlow PL, Sapp SK, Ellis SG, Topol EJ. Longterm outcome of transient, uncomplicated in-laboratory coronary artery closure. Circulation 1995;91:2733–2741. 22. Abdelmeguid AE, Topol EJ. The myth of the myocardial “infarctlet” during percutaneous coronary revascularization procedures. Circulation 1996;94:3369 –3375. 23. Abdelmeguid AE, Ellis SG, Sapp SK, Whitlow PL, Topol EJ. Defining the appropriate threshold of creatine kinase elevation after percutaneous coronary interventions. Am Heart J 1996;131:1097–1105. 24. Harrington RA, Lincoff AM, Califf RM, Holmes DR Jr, Berdan LG, O’Hanesian MA, Keeler GP, Garratt KN, Ohman EM, Mark DB, et al. Characteristics and consequences of myocardial infarction after percutaneous coronary intervention: insights from the Coronary Angioplasty Versus Excisional Atherectomy Trial (CAVEAT). J Am Coll Cardiol 1995;25:1693–1699. 25. Singh M, Rihal CS, Lennon RJ, Garratt KN, Holmes DR Jr. 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. J Am Coll Cardiol 2004;44:357–361. 26. 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.