Methods. Statistical analysis

Transcription

1 Gender differences in adverse outcomes after contemporary percutaneous coronary intervention: An analysis from the Blue Cross Blue Shield of Michigan Cardiovascular Consortium (BMC2) percutaneous coronary intervention registry Claire S. Duvernoy, MD, a Dean E. Smith, PhD, MPH, a Prerana Manohar, MD, b Ann Schaefer, MS, RN, b Eva Kline-Rogers, MS, RNP, a David Share, MD, MPH, a Richard McNamara, MD, b Hitinder S. Gurm, MD, a and Mauro Moscucci, MD, MBA a,c Ann Arbor, MI Background Prior studies have shown a relationship between female gender and adverse outcomes after percutaneous coronary interventions (PCIs). Whether this relationship still exists with contemporary PCI remains to be determined. Methods We evaluated gender differences in clinical outcomes in a large registry of contemporary PCI. Data were prospectively collected from 22,725 consecutive PCIs in a multicenter regional consortium (Blue Cross Blue Shield of Michigan Cardiovascular Consortium) between January 2002 and December The primary end point was in-hospital all-cause mortality; other clinical outcomes evaluated included in-hospital death, vascular complications, transfusion, postprocedure myocardial infarction, stroke, and a combined major cardiovascular adverse event (MACE) end point including myocardial infarction, death, stroke, emergency coronary artery bypass grafting, and repeated PCI at the same site. Independent predictors of adverse outcomes were identified using multivariate logistic regression analysis. Results Compared with men, women were older, had a higher prevalence of comorbidities, and had a significantly higher frequency of adverse outcomes after PCI. After adjustment for baseline demographics, comorbidities, clinical presentation, and lesion characteristics, female gender was associated with an increased risk of in-hospital death, vascular complication, blood transfusion, stroke, and MACE. The relationship between female gender and increased risk of death and MACE was no longer present after further adjustment for kidney function and low body surface area. Conclusions Differences in mortality rates between men and women no longer exist after PCI. However, our data suggest that technological advancements have not completely offset the relationship between gender and adverse outcomes after PCI. (Am Heart J 2010;159: e1.) From the a University of Michigan Medical Center, Veterans Affairs Ann Arbor Healthcare System, Ann Arbor, MI, b Meijer Heart Center, and c University of Miami Miller School of Medicine. Submitted July 1, 2009; accepted December 14, Reprint requests: Claire S. Duvernoy, MD, FACC; University of Michigan Health System, Veterans Affairs Ann Arbor Healthcare System, 2215 Fuller Rd, Box 111a, Ann Arbor, MI duvernoy@umich.edu /$ - see front matter 2010, Mosby, Inc. All rights reserved. doi: /j.ahj Cardiovascular disease remains the biggest killer in the United States, and population studies indicate that more women than men die of cardiovascular disease in this country. Nonetheless, women make up only approximately one third of all percutaneous coronary interventions (PCIs) performed in the United States. 1 One possible explanation for the discrepancy in numbers of female patients referred for PCI versus male patients may be the widespread belief that women do not do as well as men after invasive coronary procedures. Nearly all published data concerning clinical outcomes after PCI in women compared with men suggest that, in the aggregate, women have worse outcomes. 2,3 Many investigators have attributed at least some of the gender-associated excess risk to greater age at presentation, greater presence of comorbid conditions, and smaller size. 3 Furthermore, others have suggested that, with the advent of adjunctive pharmacotherapy, smallercaliber catheters, and widespread use of intracoronary stents, the gender gap in outcomes has been eliminated. 4 Using data from a multicenter regional consortium, we

2 678 Duvernoy et al American Heart Journal April 2010 Table I. Baseline demographic and clinical characteristics Women Men Type Variable n = 7877 % n = % P value Demographic Age, y (mean [SD]) 66.9 (14.8) 61.9 (12.0) b.0001 Age y Age y b.0001 Age 80+ y b.0001 Lean (BMI b25) b.0001 Overweight (25 BMI b 30) b.0001 Obese (BMI 30) b.0001 BSA b1.6 m b.0001 BSA m b.0001 BSA 2.0 m b.0001 Current smoking b.0001 Historical Hypertension b.0001 Diabetes mellitus b.0001 Congestive heart failure b.0001 Extracardiac vascular disease b.0001 Renal failure with dialysis Significant valve disease b.0001 Recent gastrointestinal bleeding Chronic obstructive pulmonary disease b.0001 Atrial fibrillation Cardiac arrest MI b.0001 PCI b.0001 CABG b.0001 Indication for PCI Emergent PCI b.0001 Unstable angina b.0001 Cardiac arrest With prevalent MI MI present (b7 d) Acute MI (b24 h) b.0001 Cardiogenic shock Ventricular tachycardia/fibrillation Rescue PTCA BMI, Body mass index; PTCA, percutaneous transluminal coronary angioplasty. sought to determine whether women continue to be at greater risk than men for adverse events after PCI in the contemporary era. Methods Between January 2002 and December 2003, we prospectively collected data on 22,725 PCIs performed at 17 hospitals in Michigan that participate in the Blue Cross Blue Shield of Michigan Cardiovascular Consortium (BMC2), a regional consortium. The population of the state of Michigan numbers just N10 million people, about 3% of the US population. 5 A standardized data collection form was used, and the collection of data was approved by the University of Michigan's institutional review board as well as the local review board of each participating center. Funding for the study was provided in part by the BMC2; no other extramural funding was obtained. The authors are solely responsible for the design and conduct of this study, all study analyses, the drafting and editing of the paper, and its final contents. Details concerning standardization and validation of the data have been previously described. 6,7 Clinical outcomes evaluated included the primary end point of in-hospital death as well as vascular complications, transfusion rates, gastrointestinal bleeding, contrast-induced nephropathy, infection and/or sepsis, repeated revascularization, periprocedural myocardial infarctions (MIs), stroke, and a combined major cardiovascular adverse event (MACE) that included MI, death, stroke, emergency coronary artery bypass grafting (CABG), and repeated PCI at the same lesion site. 8 Outcomes were defined by the BMC2 Data Dictionary (provided to every site) as listed in Appendix A (online). Statistical analysis Differences in discrete variables between groups were evaluated by the χ 2 test and Fisher exact test, whereas continuous variables were analyzed using the t test and Wilcoxon rank sum test. Multivariate fixed effects logistic regression modeling was used to estimate adjusted odds of postprocedural events in association with gender. Separate models were developed for each end point. Candidate models covariates included prior history of: stroke or peripheral vascular disease, hypertension, diabetes, congestive heart failure, renal failure requiring dialysis, recent gastrointestinal bleeding, atrial fibrillation, cardiac arrest, PCI, CABG, chronic

3 American Heart Journal Volume 159, Number 4 Duvernoy et al 679 Table II. Clinical and angiographic characteristics Women Men Type Variable n = 7877 % n = % P value Laboratory WHO anemia b.0001 Creatinine 1.5+ (mg/dl) Creatinine 2.0+ (mg/dl) Precreatinine (mean [SD]) (mg/dl) 1.12 (1.02) 1.22 (1.53) b.0001 CrCl b.0001 CrCl CrCl b.0001 CrCl b b.0001 Left ventricular ejection fraction b40% b.0001 Left ventricular ejection fraction b50% b.0001 Left ventricular ejection fraction (mean [SD]) 53.3 (11.8) 51.3 (11.6) b.0001 Total contrast (ml) (SD) 203 (91.8) 219 (98.9) b.0001 Fluoroscopy time (min) (SD) 14.3 (10.5) 15.0 (11.6) b.0001 Medications Aspirin b.0001 Statins b.0001 ACE inhibitors b β-blockers Calcium-channel blockers b.0001 Diuretics b.0001 GP IIB/IIIA b.0001 Intraortic balloon pump Postprocedure IV heparin Postprocedure LMW heparin Ticlopidine/clopidogrel Procedural variables MACD exceeded Single-vessel disease (N70%) b.0001 Double-vessel disease (N70%) Triple-vessel disease (N70%) b Lesion Lesions Lesions Vessel with PCI Vessels with PCI Vessels with PCI Left main stenosis (N70%) b.0001 Thrombus b.0001 Calcification b.0001 Any stenting WHO, World Health Organization; ACE, Angiotensin-converting enzyme; MACD, maximum allowable contrast dose; IV, Intravenous; LMW, low molecular weight. Medications noted are given for entire hospital course, that is, pre-, during, and postprocedure use. Exceeding a weight- and creatinine-adjusted contrast dose (5 ml*body weight [in kilograms])/serum creatinine (in milligrams per deciliter). obstructive airway disease, and MI. Additional risk variables evaluated included age, serum creatinine 1.5 mg/dl, creatinine clearance (CrCl), anemia (defined as per the World Health Organization definition), left ventricular ejection fraction b50%, periprocedural cardiac arrest, cardiogenic shock, and ventricular tachycardia/fibrillation. Angiographic variables included 3 or more lesions, left main stenosis N70%, 3-vessel disease (N70%), 3 or more vessels treated with PCI, visible thrombus, calcification, chronic total occlusion, and restenotic lesions. Likelihood ratio testing was used to ascertain whether the addition of CrCl and body surface area (BSA) category variables (CrCl categories 90, 60-89, 30-59, and b30 ml/min; BSA categories b1.6 m 2 and 1.6 BSA b2.0 m 2 ) to fully adjusted models explained the higher complication rates in women. The level of significance for tests was.05. Statistical analysis was performed using SAS 8.2 (SAS Institute; Cary, NC). Continuous variables are expressed as mean ± SD, and discrete variables are expressed as frequency counts and percentages. Results There were 22,725 patients in the study period. Slightly more than one third (7,877 or 34.7%) were female. Females had more existing risk factors such as hypertension, diabetes, and congestive heart failure than males. Men smoked more frequently and had more previous history of PCI and CABG than women, despite the fact that men tended to be younger than women (Table I). Women were more likely to be on periprocedural calcium-channel blockers and diuretics than men, but less likely to be on aspirin, statins, angiotensin-

4 680 Duvernoy et al American Heart Journal April 2010 Table III. In-hospital outcomes Women Men Variable n = 7877 % n = % P value Procedural variables Postprocedure TIMI 3 success Postprocedure stenosis success Major outcomes All CABG Emergent CABG Vascular complications b.0001 Contrast-induced nephropathy b.0001 Nephropathy requiring dialysis Postprocedure transfusion b.0001 Gastrointestinal bleeding b.0001 Infection and/or sepsis b.0001 Repeated revascularization (same site) Stroke b.0001 MI Death b.0001 MACE TIMI, Thrombolysis in Myocardial Infarction. For peak minus baseline creatinine 0.5 mg/dl; patients with history of renal failure on dialysis are excluded. All CABG, stroke, MI, death, or repeated PCI (same site). Table IV. Parsimonious fully adjusted models for 13 outcomes estimating the odds ratio for female versus male risk Not BSA or CrCl adjusted OR CrCl adjusted OR BSA + CrCl adjusted OR Outcome OR 95% CI P value OR 95% CI P value OR 95% CI P value Any CABG Emergency CABG Vascular complication b b b.0001 Contrast-induced nephropathy b b b.0001 Nephropathy requiring dialysis Postprocedure transfusion b b b.0001 Gastrointestinal bleeding Infection and/or sepsis b b Repeated revascularization Stroke or TIA Postprocedure MI Death MACE OR, Odds ratio. Defined as peak minus baseline creatinine 0.5 mg/dl; patients with a history of renal failure on dialysis are excluded. Repeated revascularization in the same site. Any CABG, stroke or TIA, MI, death, or repeated PCI (same site). converting enzyme inhibitors, and β-blockers. Women were more likely to have moderate (CrCl ml/min) and severe (CrCl b30 ml/min) chronic kidney disease. In addition, they tended to have more single-vessel disease (N70%) than men and a lower incidence of 3-vessel disease (Table II). Coronary stents were used in N80% of the patients, and use of stents was not different for men versus women. Six-French sheaths were used in the majority of patients (57%); 7F sheaths in just over one third, and 8F sheaths in just N7%. There were no gender differences in sheath sizes used. Access site was femoral in N99% of cases and again did not differ by gender. No closure device was used in more than three fourths of patients; again, likelihood of use or nonuse did not differ by gender. Glycoprotein (GP) IIb/IIIa inhibitors were used less commonly in women. Women had significantly worse outcomes than men for several in-hospital events. Women had 3 times as many vascular complications, nearly twice as much contrastinduced nephropathy, and more than twice as many transfusions as men (8.84% vs 3.87%, P b.0001). In addition, gastrointestinal bleeding, infection and/or sepsis, and stroke/transient ischemic attack (TIA) were more often found in women. Finally, the main outcome of

5 American Heart Journal Volume 159, Number 4 Duvernoy et al 681 Figure 1 including vascular complication, contrast-induced nephropathy, transfusion, gastrointestinal bleeding, infection/sepsis, and stroke/tia, female gender remained an independent predictor of adverse outcomes even after adjustment for impaired kidney function and low BSA. Outcomes by categories of BSA (A) and categories of CrCl (B). Percentage of death for each category is stratified by gender. No significant differences in death rates for each category was detected between men and women; however, there were more women than men in the low-bsa categories and in the low-crcl categories. death occurred significantly more frequently in women than in men (1.69% vs 1.02%, P b.0001) (Table III). Crude death rates for men and women by hospital showed that in all but 3 of the 17 participating hospitals, death rates for women were higher than those for men. For the main outcome of death (Table IV), full adjustment for age and for historical, observational, and diagnostic risk factors did not account for the higher risk of death in women. However, when CrCl and BSA adjustments were made in addition to the full adjustment described above, the odds ratio in the new model became null (P =.20), thus implying that traditional risk factors that incorporate adjustment for kidney function plus low BSA can explain the higher risk of death in women. Figure 1 provides the percentage of death for men and women as stratified by categories of BSA (Figure 1, A) and CrCl (Figure 1, B). There were no significant differences in death between men and women within each category for either variable. In contrast, for secondary outcomes Discussion In this study, we analyzed the relationship between gender and outcomes in a large, quality-controlled multicenter registry of contemporary PCIs. This data set included a higher use of coronary stents than any genderspecific data previously reported and the use of GP IIb/ IIIa inhibitors in almost three fourths of the patients. We found that, compared with men, women were older, had a higher prevalence of comorbidities, and had a significantly higher frequency of adverse outcomes after PCI. After adjusting for baseline demographics, comorbidities, clinical presentation, and lesion characteristics, we found that female gender was associated with an increased risk of vascular complications, contrast nephropathy, gastrointestinal bleeding, infection/sepsis, blood transfusions, stroke, MACE, and death in the hospital. However, after further adjustment for chronic kidney disease and low BSA, the odds ratios of MACE and death for women were no longer statistically significant. Our findings differ in some important aspects from the findings reported by the Northern New England PCI registry, another large, quality-controlled regional database. Malenka et al 4 reported changes in gender-related outcomes from that database between 1994 and 1999 and concluded that, unlike in earlier years, there were no longer significant differences in any outcome by gender. This finding contrasted with their report from 1989 to 1993, when the authors did find a significantly higher rate of death and adverse outcomes for women undergoing PCI versus men. 9 Our findings, although more contemporary than the 5-year period of the late 1990s, continue to indicate gender-related adverse outcomes for several important adverse events. Several other groups have underscored this finding. Peterson et al 3 reported findings from the National Cardiovascular Network Database between 1994 and 1998, including the fact that women were at significantly higher risk than men for stroke, vascular complications, and repeated in-hospital revascularization. However, once the authors corrected for BSA, they found that female gender was no longer associated with higher procedural mortality. 3 Subsequently, Ashby et al 10 published a comparison of PCI outcomes in men and women limited to small (b3.0 mm) arteries. The authors reported that although vascular complications, bleeding, and transfusion rates remained higher in women than in men, there was no significant difference between genders in rates of MACE, MI, or death. 10 The authors concluded that adverse PCI results

6 682 Duvernoy et al American Heart Journal April 2010 reported for women were in fact the consequence of smaller vessel size in women. Other groups have advocated that the widespread use of GP IIb/IIIa inhibitors would decrease the gender gap in PCI outcomes. Iakovou et al 11 reported single-center outcomes between 1996 and 2000 for 1,215 patients (322 women) who received GP IIb/IIIa inhibitors during PCI with stent placement and found that there were no significant differences in bleeding, vascular complications, transfusion rates, MI, or MACE. No deaths occurred in this relatively small analysis. 11 Lansky et al 12 subsequently reported gender differences in PCI outcomes from the Controlled Abciximab and Device Investigation to Lower Late Angioplasty Complications trial. The investigators found that female gender was an independent predictor of MACE and of bleeding complications in the trial, but that comorbid risk factors and BSA, not gender, predicted 1-year death rates. 12 Recently, the American Heart Association and the American College of Cardiology published a joint statement on PCI and adjunctive pharmacotherapy in women. 13 In it, the authors discuss the available genderspecific data with respect to each of the interventional strategies available and in various clinical scenarios. They conclude that, although in many instances no genderspecific data are available, in most instances evidence favors using the same procedures and adjunctive therapies in women that have been shown to be beneficial in men. The authors readily acknowledge that despite procedural advances such as smaller sheath sizes, less aggressive anticoagulation regimens, and weightadjusted heparin dosing, vascular complications continue to be more frequent in women than in men. They conclude, however, that adjustment for comorbidities and BSA has largely eradicated sex-specific mortality differences between genders. Members of our group have recently reported that larger guide catheter size is associated with significantly higher complication rates, including higher death and MACE rates. 14 Many groups have reported that chronic kidney disease is associated with adverse outcomes in patients undergoing PCI, including higher rates of bleeding complications and higher rates of death and MI Latif et al 15 recently reported that patients with impaired renal function were more likely to be female and older. Patients with impaired kidney function are also less likely to receive evidencebased medications, a finding that we observed as well in our female patients. Our analysis provides new clues into the observed gender advantage enjoyed by men over women for PCI. Two main factors kidney disease and smaller body size appear to be key to the higher death rates and major adverse outcomes seen in women. Other investigators have also reported that excess bleeding particularly bleeding associated with IIb/IIIa use occurs much more frequently in women than in men, and that using creatinine values instead of CrCl estimates frequently overestimates renal function, leading to overdosing of anticoagulation regimens in women especially. 19 The use of CrCl estimates for calculation of medication administration is relatively new, and likely was not used in our consortium during the study period. Similarly, adjustment of dosing and other treatment calculations by BSA would help reduce remaining gender disparities because BSA is likely a crude surrogate for the most critical factor determining outcomes, namely, vessel size. With respect to other procedural complications, movement toward ever smaller French sizes for catheters and sheaths may help reduce the gender disparity that still remains. Study limitations This study has several limitations. We analyzed data from 17 hospitals in 1 state; and thus, our results may not be generalizable to the entire country. However, these institutions represent a wide spectrum of practice patterns, from small private hospitals to large universitybased practices. We did not collect socioeconomic information and thus cannot comment on these potential confounders. We did not collect data on drug dosing; it is thus possible that adverse outcomes related to overdosing were not tracked by our analysis. Our data are observational in nature; and thus, we cannot fully control for practice patterns and other possible confounders. Conclusions We conclude that sex differences for all-cause mortality no longer exist between men and women after PCI. Impaired kidney function and smaller size in female patients in the consortium appear to largely explain the worse outcomes seen in women. However, our data suggest that technological advances have not completely offset the relationship between gender and adverse outcomes after PCI. Consideration should be given to quality improvement efforts focused on decreasing complication rates in women, such as using smaller catheters and devices, and medication dosing by CrCl and BSA. References 1. American Heart Association. Statistical fact sheet populations: women and cardiovascular diseases statistics. American Heart Association; Mehilli J, Kastrati A, Dirschinger J, et al. Differences in prognostic factors and outcomes between women and men undergoing coronary artery stenting. JAMA 2000;284: Peterson ED, Lansky AJ, Kramer J, et al. Effect of gender on the outcomes of contemporary percutaneous coronary intervention. Am J Cardiol 2001;88: Malenka DJ, Wennberg DE, Quinton HA, et al. Northern New England Cardiovascular Disease Study Group: gender-related changes in the practice and outcomes of percutaneous coronary

7 American Heart Journal Volume 159, Number 4 Duvernoy et al 683 interventions in northern New England from 1994 to J Am Coll Cardiol 2002;40: QuickFacts USCBSaC: population, 2007 estimate. 6. Moscucci M, Share D, Kline-Rogers E, et al. The Blue Cross Blue Shield of Michigan Cardiovascular Consortium (BMC2) collaborative quality improvement initiative in percutaneous coronary interventions. J Interv Cardiol 2002;15: Moscucci M, Share D, Smith D, et al. Relationship between operator volume and adverse outcome in contemporary percutaneous coronary intervention practice: an analysis of a quality-controlled multicenter percutaneous coronary intervention clinical database. J Am Coll Cardiol 2005;46: Moscucci M, Rogers EK, Montoye C, et al. Association of a continuous quality improvement initiative with practice and outcome variations of contemporary percutaneous coronary interventions. Circulation 2006;113: Malenka DJ, O'Connor GT, Quinton H, et al. Differences in outcomes between women and men associated with percutaneous transluminal coronary angioplasty. A regional prospective study of 13,061 procedures. Northern New England Cardiovascular Disease Study Group. Circulation 1996;94:II99-II Ashby DT, Mehran R, Aymong EA, et al., Comparison of outcomes in men versus women having percutaneous coronary interventions in small coronary arteries. Am J Cardiol 91: , A974, A977, Iakovou I, Dangas G, Mehran R, et al. Gender differences in clinical outcome after coronary artery stenting with use of glycoprotein IIb/ IIIa inhibitors. Am J Cardiol 2002;89: Lansky AJ, Pietras C, Costa RA, et al. Gender differences in outcomes after primary angioplasty versus primary stenting with and without abciximab for acute myocardial infarction: results of the Controlled Abciximab and Device Investigation to Lower Late Angioplasty Complications (CADILLAC) trial. Circulation 2005;111: Lansky AJ, Hochman JS, Ward PA, et al. Percutaneous coronary intervention and adjunctive pharmacotherapy in women: a statement for healthcare professionals from the American Heart Association. Circulation 2005;111: Grossman PM, Gurm HS, McNamara R, et al. Percutaneous coronary intervention complications and guide catheter size: bigger is not better. JACC Cardiovasc Interv 2009;2: Latif F, Kleiman NS, Cohen DJ, et al, on behalf of the EVENT Investigators. In-hospital and 1-year outcomes among percutaneous coronary intervention patients with chronic kidney disease in the era of drug-eluting stents: a report from the EVENT (Evaluation of Drug Eluting Stents and Ischemic Events) Registry. J Am Coll Cardiol Intv 2009;2: Rubenstein MH, Harrell LC, Sheynberg BV, et al. Are patients with renal failure good candidates for percutaneous coronary revascularization in the new device era? Circulation 2000;102: Best PJ, Lennon R, Ting HH, et al. The impact of renal insufficiency on clinical outcomes in patients undergoing percutaneous coronary interventions. J Am Coll Cardiol 2002;39: Kahn JK, Rutherford BD, McConahay DR, et al. Short- and long-term outcome of percutaneous transluminal coronary angioplasty in chronic dialysis patients. Am Heart J 1990;119: Alexander KP, Chen AY, Newby LK, et al, for the CRUSADE Investigators. Sex differences in major bleeding with glycoprotein IIb/ IIIa inhibitors: results from the CRUSADE (Can Rapid Risk Stratification of Unstable Angina Patients Suppress Adverse Outcomes With Early Implementation of the ACC/AHA Guidelines) Initiative. Circulation 2006;114:

8 American Heart Journal April Year Duvernoy et al 683.e1 Appendix A. BMC2 Data Dictionary 1. Emergency CABG: In the opinion of the operator or the responsible physician, the patient needed to be moved directly to surgery from the catheterization laboratory or hospital ward, typically because of indications such as ongoing ischemia, rest angina despite maximal treatment, pulmonary edema requiring intubation, or shock. For example, Severe dissection resulting in transferring from catheterization laboratory to operating room with guidewire in place. This does not include patients that go to recovery area or coronary care unit before emergent CABG. 2. Vascular complications: any vascular complication including pseudoaneurysm, arteriovenous fistula, femoral neuropathy, retroperitoneal hematoma, hematoma at the access site requiring transfusion/ prolonged hospital stay/causing a drop in hemoglobin N3.0 g/dl, or any access site complication requiring surgical repair. 3. Contrast-induced nephropathy: peak minus baseline creatinine 0.5 mg/dl; patients with a history of renal failure requiring dialysis are excluded. 4. Nephropathy requiring dialysis: decrease in renal function requiring peritoneal dialysis, hemodialysis, or continuous veno-venous hemofiltration postprocedure to discharge. 5. Postprocedure transfusion: any transfusion for any reason including packed red blood cells, platelets, fresh frozen plasma, whole blood. Includes all transfusions even if transfusion was given during major surgery in association with vascular repair or bypass surgery. 6. Gastrointestinal bleeding: hematemesis, grossly positive stool, or melena associated with a decrease in hematocrit and hemoglobin. 7. Infection and/or sepsis: positive blood, urine, or sputum cultures requiring treatment with antibiotics. 8. Repeated revascularization: repeated PCI of the same lesion during the same hospitalization 9. Stroke or TIA: loss of neurologic function caused by an ischemic event that is severe enough to leave a persistent deficit for N24 hours, or abrupt loss of neurologic function with complete return of function within 24 hours. 10. Postprocedure MI: any rise in creatine phosphokinase myocardial band (CPK-MB) fraction above normal (in your institution) and/or troponins in the positive range without any new Q waves on electrocardiogram, and/or an MI associated with the development of new Q waves that are.03 seconds in width and/or more than one third of the total QRS complex in contiguous leads and as evidenced by subsequent enzyme level elevation (specifically CPK-MB rise to 3 times the baseline value just before intervention with a N5% MB subfraction over baseline and/or troponins in the positive range). 11. Death: The patient died in association with this procedure while in the laboratory. 12. MACE: any CABG, stroke or TIA, MI, death, or repeated revascularization. Appendix B. Statistical model used for death Table of odds ratios for a parsimonious death model Risk factor Model OR OR 95% CI P value Gender female Age y Age y Age 80+ y (reference b60 y) CrCl CrCl CrCl b30 (reference CrCl 90+) BSA m BSA b1.6 m 2 (reference BSA 2.0+ m 2 ) Extracardiac vascular disease History of renal failure with dialysis b.0001 History of atrial fibrillation History of coronary artery bypass grafting b.0001 History of chronic obstructive pulmonary disease WHO anemia Left ventricular ejection fraction b50% b.0001 Presenting cardiac arrest b.0001 Presenting cardiac shock b.0001 Left main stenosis (N70%) Vessel disease Visible thrombus OR, Odds ratio; WHO, World Health Organization.