|
| HOME | SUBSCRIPTIONS | CURRENT ISSUE | PAST ISSUES | CARDIOSOURCE | SEARCH | HELP | FEEDBACK |
|
|
|
|||||||||
|
|||||||||||
|
J Am Coll Cardiol, 2002; 40:387-393 © 2002 by the American College of Cardiology Foundation |
* Division of Cardiovascular Diseases and Internal MedicineRochester, Minnesota, USA
Section of Biostatistics, Mayo Clinic, Rochester, Minnesota, USA
Manuscript received September 6, 2001; revised manuscript received March 25, 2002, accepted April 30, 2002.
* Reprint requests and correspondence: Dr. Charanjit S. Rihal, Mayo Clinic, 200 First Street SW, Rochester, Minnesota 55905, USA.
rihal{at}mayo.edu
 |
Abstract |
|---|
 Top Abstract MethodsresultsDiscussionReferences |
|---|
BACKGROUND: Both clinical and angiographic features influence risk of PCIs.
METHODS: Percutaneous coronary interventions performed between January 1, 1996, and December 31, 1999, were analyzed. Logistic regression and bootstrap methods were used to create an integer risk score for estimating the risk of procedural complications using baseline, angiographic and procedural characteristics. The risk score was tested in a validation-set consisting of all procedures performed in the year 2000.
RESULTS: Among 5,463 procedures, 5 clinical and 3 angiographic variables were significantly correlated with procedural complications: cardiogenic shock, left main coronary artery disease, severe renal disease, urgent or emergent procedure, congestive heart failure class III or higher, thrombus, multivessel disease and older age. In the validation-set, the model fitted the data adequately; the average receiver operating characteristic curve area was 0.782 (standard deviation, 0.018).
CONCLUSIONS: Eight variables were combined into a convenient bedside risk scoring system that estimates the risk of complications after PCIs.
| ||||||||||||||||
The goals of the present study were: 1) to identify clinical and angiographic risk factors associated with major cardiovascular complications in a consecutive series of PCIs performed over a recent four-year period; 2) to construct a simple risk score (based on those risk factors) for identifying subgroups within various risk strata; and 3) to validate internally the risk score using a consecutive one-year sample of PCI procedures.
|
Methods |
|---|
|
TopAbstractMethodsresultsDiscussionReferences |
|---|
For data analysis and risk score construction, we included PCIs that were performed between January 1, 1996, and December 31, 1999 (study-set). Internal validation used PCIs performed during 2000 (validation-set). The period from 1996 to 1999 was chosen because: 1) stent deployment became routine rather than being used only as a bail-out treatment; 2) a thienopyridine was routinely added to aspirin therapy after stent deployment; and 3) parenteral GP IIb/IIIa receptor antagonists became available. Only first-time procedures were included.
End point
The outcome of interest was major complications, defined as one or more of the following: 1) in-hospital death; 2) Q-wave myocardial infarction (MI); 3) urgent or emergent coronary artery bypass graft surgery (CABG) during the index hospitalization; and 4) cerebrovascular accident. Myocardial infarction was diagnosed in the presence of two of the following three criteria: 1) chest pain for at least 20 min; 2) elevation of creatine kinase (or the MB fraction) >2 times normal; and 3) new Q waves on electrocardiography. Other procedural complications, such as non–Q-wave MI and vascular access site problems, were not included in the present analysis. Patients who underwent elective bypass surgery during hospitalization (n = 20) for severe residual disease were not included.
Baseline clinical characteristics
Baseline clinical characteristics included age, gender, body mass index, history of hypertension, diabetes, hypercholesterolemia (serum cholesterol >6.21 mmol/l [240 mg/dl]), peripheral vascular disease, prior MI, prior CABG, Canadian Cardiovascular Society angina class, New York Heart Association (NYHA) heart failure class, smoking status and severe renal disease (patients with serum creatinine >265 µmol/l [3 mg/dl] or patients receiving dialysis).
Indications for PCI
Indications for PCI were classified as stable angina or unstable angina (defined as rest pain or post-MI angina), acute MI or cardiogenic shock (defined as systolic blood pressure <95 mm Hg or <110 mm Hg with inotropic or intra-aortic balloon pump support, unresponsive to fluid challenge or vasopressor therapy). Indications for PCI were subclassified as elective or emergent (patient with acute coronary syndrome brought into the cardiac catheterization laboratory because of ongoing chest pain or hemodynamic compromise or both).
PCI angiographic and procedural variables
The PCI angiographic and procedural variables included: location of the lesion; presence of multivessel disease (stenosis diameter 
70% of vessel diameter in two or more epicardial coronary arteries or their major branches); multivessel angioplasty; type of lesion defined by the operator before the intervention (American College of Cardiology/American Heart Association [ACC/AHA] type A, B1, B2, or C); graft or native vessel angioplasty; left main coronary artery disease (stenosis diameter 70% of vessel diameter); intervention of unprotected left main coronary artery; intracoronary thrombus or presence of calcium at the lesion; PCI on a moderate (45° to 90°) or severe (>90°) bend; left ventricular function (
0.40); and use of intracoronary stents and GP IIb/IIIa inhibitors.
Statistical methods
Logistic regression was used to model the incidence of procedural complications and estimate odds ratios. All p values were two-tailed. Characteristics significant in univariate analysis were combined into an initial model, and the bootstrap method was used to remove variables to avoid overfitting the data (10–12). Two hundred bootstrap samples were selected. Backward selection at the 0.05 significance level was used to eliminate extraneous variables in each sample. Variables that were selected in at least 140 of the samples (70%) were included in the final multivariate model.
To develop a simple risk prediction score, the risk factors identified through multivariate modeling were assigned an integer coefficient. Integers were chosen to be approximately proportional to the estimated continuous coefficients from the logistic model. The score starts at 0, and each risk factor’s corresponding coefficient is added. The final score is typically between 0 and 25. The patient population was classified into five risk categories: 1) very low, 0 to 5; 2) low, 6 to 8; 3) moderate, 9 to 11; 4) high, 12 to 14; and 5) very high, 15. Patients were ordered by the predictive risk and separated into eight groups of similar size. Observed and expected numbers of events were calculated within each group. Model adequacy of the scoring system was then evaluated with the Hosmer-Lemeshow goodness-of-fit test (13,14). Within the study-set, discriminatory ability of the score was assessed with another 200 bootstrap samples. The risk score was calculated for each patient, and the area under the receiver operating characteristic (ROC) curve was determined for each sample. To study the performance of the prediction rule on patient subgroups, the data were stratified according to various clinical, procedural, and angiographic variables.
For the validation-set of procedures performed during 2000, the predicted probabilities of in-hospital death, Q-wave MI, stroke or need for emergent CABG were calculated from the integer risk score. Model discrimination was assessed by ROC curve analysis, and goodness-of-fit was tested with the Hosmer-Lemeshow statistic.
|
results |
|---|
|
TopAbstractMethodsresultsDiscussionReferences |
|---|
Baseline clinical characteristics. Mean age (± SD) of the patients was 65.8 ± 12 years, 70.4% were men, 43.1% presented with unstable angina and 14.6% presented with acute MI (Table 1). At the time of the procedure, 4.2% of the patients were in cardiogenic shock. The prevalence of risk factors is shown in Table 1: 60.6% had hypertension, 22% had diabetes, 11.3% had peripheral vascular disease and 3.4% had severe renal disease.
|
III, congestive heart failure on presentation, severe renal disease, treatment of acute MI, peripheral vascular disease and age (by decade). The following variables were not significantly associated with procedural complications: unstable angina, diabetes, hypertension, hypercholesterolemia and current smoking status. Elective procedures and male gender were associated with reduced rates of complications.
|
|
class III, thrombus, multivessel disease and age (Table 4). The data did not deviate significantly from the logistic model, as indicated by the nonsignificant Hosmer-Lemeshow test result (p = 0.93) (Fig. 1). The mean area (± SD) under the ROC curve of the bootstrap samples was 0.782 (±0.018), indicating a good ability to discriminate between patients who had complications during the index hospitalization and those who did not.
|
|
|
2%) for very low-risk; 3.0% (>2% to 5%), low-risk; 6.2% (>5% to 10%), moderate-risk; 19.5% (>10% to 25%), high-risk; and 35.0% (>25%), very high-risk (Fig. 2).
|
26 = –4.25), indicating that the model fitted the data well and typically correctly ranked patient risk (Fig. 3). Generally, patients who died had higher risk scores than those who did not (c statistic, 0.88), although this did not hold as well for patients who had a Q-wave MI, stroke or emergency CABG (c statistic, 0.684, 0.719 and 0.641, respectively).
|
|
Discussion |
|---|
|
TopAbstractMethodsresultsDiscussionReferences |
|---|
Comparison with previous studies. Ellis et al. (8) proposed an angiographic risk assessment scheme and found a strong correlation between complications after PCI and intervention in degenerated vein grafts and recent total occlusions. Our goal was to develop a risk score incorporating clinical and angiographic evaluations that are commonly available to clinicians before PCI. For risk prediction we found that clinical variables were as important as angiographic variables. In particular, age, the presence of shock or severe congestive heart failure, severe renal disease and procedural urgency were correlated with procedural complications, along with angiographic variables such as stenosis of the left main coronary artery, thrombus and presence of multivessel disease.
The Northern New England Cardiovascular Disease Study Group performed a multivariate analysis of in-hospital mortality (but not other major complications) in a data set of 15,331 patients who had PCI performed between 1994 and 1996 (6). Since 1996, however, PCI practice has changed considerably, with higher utilization rates of stents and glycoprotein IIb/IIIa inhibitors. Nonetheless, factors associated with in-hospital mortality were similar to our findings; they included old age, lower ejection fraction, cardiogenic shock, treatment of an acute MI, urgent and emergent priority, type C lesions and use of a preprocedure intra-aortic balloon pump.
In two prior studies from our laboratory, we documented the potential of the New York State multivariate risk score for predicting overall procedural mortality—but not other clinically important complications (9,15). The current study further demonstrates the utility of simple clinical prediction rules for complications after PCI. Several other studies have analyzed the risk factors associated with PCI, but they were performed in an era antedating current practice (16–18).
Clinical application
When considering treatment options, the potential for benefit must be weighed against the potential for harm for each option. For invasive procedures, in general, long-term benefit can be partially offset by procedural risk. Our risk score allows rapid bedside stratification of patients into five risk strata. It should be recognized that no model can predict which individual patients will experience complications (the "crystal ball approach"); however, the model can estimate the likelihood of a major PCI complication. Insofar as probabilistic estimates of risk and benefit are useful in clinical decision making, clinicians may find the model contributory.
Models such as ours, if externally validated, may also be useful as benchmarking mechanisms for hospitals and operators. Over large numbers of procedures, expected rates of complications can be calculated and compared with observed rates.
Study limitations
Although no specific subsets of patients were excluded, the current analysis is derived from the data set of a single referral center, and broader applicability is open to question. Operator volume was not considered in the current analysis (19). Independent validation in other data sets is needed. Perhaps the greatest limitation of any model in prediction of PCI complications is that discriminatory ability for individual elective patients is limited because of the overall low probability of an event. Over 95% of the elective interventions in the validation-set had predicted complication rates under 4%. As PCI procedures become safer, discriminatory ability will inevitably erode.
Conclusions
Clinical and angiographic characteristics are equally important in determining procedural risk with PCI. We developed and internally validated a simple integer risk score for prediction of major in-hospital complications after PCI. Five clinical variables (age, cardiogenic shock, congestive heart failure, renal failure and urgent or emergent PCI) and three angiographic variables (thrombus, multivessel disease and left main disease) can be used to accurately risk-stratify subgroups of patients undergoing PCI. This score was developed from a data set with low rates of missing values and with high rates of stent use and parenteral antiplatelet-agent use. This model may help clinicians assess procedural risk.
|
References |
|---|
|
TopAbstractMethodsresultsDiscussionReferences |
|---|
This article has been cited by other articles:
|
|
 |
|
  M. Singh, B. J. Gersh, S. Li, J. S. Rumsfeld, J. A. Spertus, S. M. O'Brien, R. M. Suri, and E. D. Peterson Mayo Clinic Risk Score for Percutaneous Coronary Intervention Predicts In-Hospital Mortality in Patients Undergoing Coronary Artery Bypass Graft Surgery Circulation, January 22, 2008; 117(3): 356 - 362. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 S. B. King III, T. Aversano, W. L. Ballard, R. H. Beekman III, M. J. Cowley, S. G. Ellis, D. P. Faxon, E. L. Hannan, J. W. Hirshfeld Jr, A. K. Jacobs, et al. ACCF/AHA/SCAI 2007 Update of the Clinical Competence Statement on Cardiac Interventional Procedures: A Report of the American College of Cardiology Foundation/American Heart Association/American College of Physicians Task Force on Clinical Competence and Training (Writing Committee to Update the 1998 Clinical Competence Statement on Recommendations for the Assessment and Maintenance of Proficiency in Coronary Interventional Procedures) J. Am. Coll. Cardiol., July 3, 2007; 50(1): 82 - 108. [Full Text] [PDF]
|
|
|
|
|
|
M. Singh, C. S. Rihal, B. J. Gersh, R. J. Lennon, A. Prasad, P. Sorajja, R. E. Gullerud, and D. R. Holmes Jr Twenty-Five-Year Trends in In-Hospital and Long-Term Outcome After Percutaneous Coronary Intervention: A Single-Institution Experience Circulation, June 5, 2007; 115(22): 2835 - 2841. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 N. Beohar, C. J. Davidson, K. E. Kip, L. Goodreau, H. A. Vlachos, S. N. Meyers, K. H. Benzuly, J. D. Flaherty, M. J. Ricciardi, C. L. Bennett, et al. Outcomes and Complications Associated With Off-Label and Untested Use of Drug-Eluting Stents JAMA, May 9, 2007; 297(18): 1992 - 2000. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 K. M Jacobson, K. Hall Long, E. K McMurtry, J. M Naessens, and C. S Rihal The economic burden of complications during percutaneous coronary intervention Qual. Saf. Health Care, April 1, 2007; 16(2): 154 - 159. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 A Siotia and J Gunn Risk scoring for percutaneous coronary intervention: let's do it! Heart, November 1, 2006; 92(11): 1539 - 1540. [Full Text] [PDF]
|
|
|
|
|
|
A D Grayson, R K Moore, M Jackson, S Rathore, S Sastry, T P Gray, I Schofield, A Chauhan, F F Ordoubadi, B Prendergast, et al. Multivariate prediction of major adverse cardiac events after 9914 percutaneous coronary interventions in the north west of England Heart, May 1, 2006; 92(5): 658 - 663. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
H. H. Ting, G. Raveendran, R. J. Lennon, K. H. Long, M. Singh, D. L. Wood, B. J. Gersh, C. S. Rihal, and D. R. Holmes Jr A Total of 1,007 Percutaneous Coronary Interventions Without Onsite Cardiac Surgery: Acute and Long-Term Outcomes J. Am. Coll. Cardiol., April 18, 2006; 47(8): 1713 - 1721. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
C. Wu, E. L. Hannan, G. Walford, J. A. Ambrose, D. R. Holmes Jr, S. B. King III, L. T. Clark, S. Katz, S. Sharma, and R. H. Jones A Risk Score to Predict In-Hospital Mortality for Percutaneous Coronary Interventions J. Am. Coll. Cardiol., February 7, 2006; 47(3): 654 - 660. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
D. R. Holmes Jr, P. Hodgson, and M. Singh Guidelines, Lighthouses, and a Toe in the Water Circulation, November 1, 2005; 112(18): 2754 - 2755. [Full Text] [PDF]
|
|
|
|
|
|
A. Halkin, M. Singh, E. Nikolsky, C. L. Grines, J. E. Tcheng, E. Garcia, D. A. Cox, M. Turco, T. D. Stuckey, Y. Na, et al. Prediction of Mortality After Primary Percutaneous Coronary Intervention for Acute Myocardial Infarction: The CADILLAC Risk Score J. Am. Coll. Cardiol., May 3, 2005; 45(9): 1397 - 1405. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
Abstracts Heart, May 1, 2005; 91(suppl_1): A5 - A72. [Full Text] [PDF]
|
|
|
|
|
|
M. Singh, C. S. Rihal, R. J. Lennon, K. N. Garratt, and D. R. Holmes 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., July 21, 2004; 44(2): 357 - 361. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
D. A. Halon, S. Adawi, I. Dobrecky-Mery, and B. S. Lewis Importance of increasing age on the presentation and outcome of acute coronary syndromes in elderly patients J. Am. Coll. Cardiol., February 4, 2004; 43(3): 346 - 352. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
M. Singh, C. S. Rihal, F. Selzer, K. E. Kip, K. Detre, and D. R. Holmes Jr 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., November 19, 2003; 42(10): 1722 - 1728. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
P. J. de Feyter and E. McFadden Risk score for percutaneous coronary intervention: forewarned is forearmed J. Am. Coll. Cardiol., November 19, 2003; 42(10): 1729 - 1730. [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | SUBSCRIPTIONS | CURRENT ISSUE | PAST ISSUES | CARDIOSOURCE | SEARCH | HELP | FEEDBACK |
