- Email: [email protected]
Outcomes in African Americans and whites after percutaneous coronary intervention
The American Journal of Medicine (2005) 118, 1019-1025
CLINICAL RESEARCH STUDY
Outcomes in African Americans and whites after percutaneous coronary intervention Michael S. Chen, MD,a Deepak L. Bhatt, MD,a Derek P. Chew, MBBS, MPH,b David J. Moliterno, MD,c Stephen G. Ellis, MD,a Eric J. Topol, MDa a
Department of Cardiovascular Medicine, The Cleveland Clinic Foundation, Cleveland, Ohio; Flinders Medical Centre, Adelaide, Australia; and c Gill Heart Institute and Division of Cardiovascular Medicine, University of Kentucky. b
KEYWORDS: Percutaneous coronary intervention; Revascularization; African American; Race; Coronary artery disease
ABSTRACT PURPOSE: We aimed to determine whether African Americans and whites have different outcomes after percutaneous coronary intervention (PCI). METHODS: We prospectively selected 8832 patients (707 African Americans) for long-term follow-up after PCI at our institution from 1992 to 2002. The primary outcome studied was death or myocardial infarction at 1 year. Propensity adjustment was performed to account for baseline differences between African Americans and whites. RESULTS: African Americans had higher rates of diabetes and less prior revascularization. Percutaneous coronary interventions in African Americans were more often urgent. Stent use was similar. Procedural success rates were similar, as were periprocedural and 30-day composite rates of death or myocardial infarction. In 1-year unadjusted outcomes, African Americans had a higher rate of death or myocardial infarction (18.0% vs 14.5%; hazard ratio (HR) ⫽ 1.25; 95% confidence interval (CI): 1.04 to 1.50; P ⫽ 0.017), but the difference was no longer significant after propensity adjustment (HR ⫽ 1.18; 95% CI: 0.98 to 1.43, P ⫽ 0.087). African Americans had a higher risk for periprocedural bleeding that persisted after propensity adjustment (adjusted odds ratio ⫽ 1.45; 95% CI: 1.14 to 1.84, P ⫽ 0.002). CONCLUSIONS: After PCI, African Americans have similar short-term rates of death or myocardial infarction when compared with whites but have a nonsignificant trend toward worse long-term outcomes. Our findings, when interpreted in the context of reportedly lower revascularization rates among African Americans, suggest that continued efforts to optimize the appropriate use of coronary revascularization among African Americans are warranted. © 2005 Elsevier Inc. All rights reserved.
Background African Americans are reported to have lower rates of diagnostic cardiac catheterization compared with whites,1-4 even in the setting of unstable angina5 or acute myocardial Requests for reprints should be addressed to Deepak L. Bhatt, MD, Director, Interventional Cardiology Fellowship Program, Cleveland Clinic Foundation, Department of Cardiovascular Medicine, 9500 Euclid Avenue, Desk F25, Cleveland, Ohio 44195. E-mail address: [email protected].
0002-9343/$ -see front matter © 2005 Elsevier Inc. All rights reserved. doi:10.1016/j.amjmed.2004.12.035
infarction.6,7 Furthermore, African Americans have lower adjusted rates of revascularization after myocardial infarction or coronary angiography1,2,6,8-13 in most but not all studies.5 Less well characterized is whether race affects outcomes after revascularization. Prior studies generally report similar outcomes after percutaneous coronary intervention (PCI) in African Americans compared with whites,14-19 although higher adjusted 1-year20 and 2-year mortality rates have been reported in African Americans.21 Prior studies are limited by lack of
1020
The American Journal of Medicine, Vol 118, No 9, September 2005
periprocedural outcomes,15 lack of long-term follow-up,17 small sample size of African Americans,14,16 or a homogenous population consisting primarily of male veterans.18 The objective of this study was to provide a comprehensive analysis of PCI outcomes in African Americans and whites.
Methods Study sample The Cleveland Clinic PCI database was queried for all PCIs performed between 1992 and 2002, with race denoted as African American or white. To avoid double counting of patients who had multiple PCIs, only the first PCI for any patient was included, resulting in 16 607 PCIs. Staged procedures were then excluded, resulting in 16 202 PCIs. From 1993 to 1999, due to variations in annual research resources, the first 500 to 1000 PCIs in each year performed with abciximab, as well as the first 500 to 1000 PCIs in each year performed not using abciximab, were prospectively selected for follow-up. The number of abciximab and non-abciximab cases was approximately equal. Case selection from 1993 to 1999 was on a consecutive basis. Since 2000, all patients have been prospectively followed for 1 year after PCI. We included only patients prospectively selected for 1-year follow-up, resulting in 8832 patients (707 African American and 8125 white) in our final cohort. Research nurses reviewed each patient’s chart after the initial PCI to corroborate database entries with the medical record. Access to the interventional registry and data analysis were approved by the institutional review board.
troponin T or I above the limit of myocardial infarction within 24 hours after the clinical event, creatine kinase-MB (CK-MB) greater than 2 times the upper limit of normal on one occasion during the first several hours after the clinical event or CK-MB greater than the upper limit of normal on 2 successive samples, or CK greater than 2 times the upper limit of normal. Troponin and CK-MB were the preferred biochemical methods of diagnosis of myocardial infarction, but CK could be used in the absence of troponin or CK-MB assays. Periprocedural myocardial infarctions were not included in 30-day and 1-year outcomes. Periprocedural and 30-day outcomes were secondary endpoints. All patients had electrocardiograms (ECGs) on the day after PCI and routine serial monitoring of cardiac enzymes. In accordance with American College of Cardiology National Cardiovascular Data Registry definitions, periprocedural myocardial infarction was defined as electrocardiographic changes (evolving ST-segment elevation, new or presumably new left bundle branch block or new Q waves on two or more contiguous leads) or CK-MB greater than 3 times the upper limit of normal.22 After PCI, the decision to perform stress testing or repeat coronary angiography was made by the patient’s cardiologist. Procedural success was defined as procedures in which all attempts resulted in post-intervention stenosis of ⱕ50% of the arterial diameter, grade 3 flow Thrombolysis In Myocardial Infarction (TIMI) was present, and there was at least a 20% decrease in percentage stenosis by visual estimation. Chronic renal insufficiency was defined as serum creatinine greater than 2 mg/dL preprocedure, and acute renal insufficiency was defined as an increase in creatinine of at least 1 mg/dL afterward. Bleeding was defined as blood transfusion, hematoma, or retroperitoneal bleed. Blood transfusions were ordered at the discretion of the attending physician.
Follow-up and endpoints Statistical analysis The primary outcome studied was death or myocardial infarction at 1 year. Research nurses contacted patients at 1 year to ascertain their status. Follow-up was performed by mailings and telephone calls to patients by dedicated research nurses. Of patients prospectively selected, follow-up at 1 year was 97.3% complete for African Americans and 98.3% complete for whites. Of the subjects lost to annual follow-up, none of the African Americans and only 5 whites were lost prior to 30 days. The primary reason that 1-year follow-up could not be obtained was inability to contact the patient, despite mailings and telephone calls. Because rates of follow-up were similar in African Americans and whites, and because follow-up was ⬎97% complete overall, bias from informative censoring was not likely. The myocardial infarction endpoint was based on the American College of Cardiology National Cardiovascular Data Registry, Cardiac Cath Lab Module version 2.0c definitions.22 Myocardial infarction required biomarker elevation and either electrocardiographic changes or ischemic symptoms. Biomarker elevation was defined as elevation of
Continuous variables are expressed as either mean ⫾ standard deviation or median with interquartile range (25% to 75%) and compared by the Mann-Whitney U test. Categorical variables were compared by chi-squared analysis. Propensity analysis23 adjusted for the likelihood of being African American. First, a logistic regression model was constructed for the likelihood of being African American compared with white, with a P value of ⱕ0.25 used for inclusion in the model. Variables included in the propensity score were sex, prior myocardial infarction, acute myocardial infarction, unstable angina, recent myocardial infarction, diabetes on insulin, diabetes not on insulin, history of multiple PCIs, hypertension, glycoprotein IIb/IIIa use, current smoker, calcium channel blocker use, systolic blood pressure, number of diseased vessels, left ventricular ejection fraction, and rotablator use. From this model, the predicted probability for being African American was generated (propensity score). Age was not significant early in the stepwise model and thus was excluded from the propensity
Chen et al
PCI outcomes in African Americans and whites
analysis. African Americans and whites were then divided into blocks of patients based on their propensity score. The optimal number of blocks was 11, which ensured that the mean propensity score was not different between African Americans and whites in each propensity score block. Unadjusted event rates for 30-day ischemia, bleeding, and postprocedural elevation in creatinine were compared by chi-squared analysis, as were the proportion of patients experiencing death or myocardial infarction by 1 year. The hazard ratio for risk associated with being African American was calculated by Cox proportional hazards modeling, with the “propensity for race” used to adjust this estimate for baseline differences between the races. Also, an estimate of relative influence of each baseline characteristic on the adjusted hazard ratio for race was explored by comparing the effect of race on death or myocardial infarction at 1 year in an unadjusted model and contrasting it with a model including race and one other baseline variable. A KaplanMeier plot of survival free of death or myocardial infarction, stratified by race and adjusted for propensity category (propensity score category set to zero), was created. For composite endpoints, the time to the first event was used. Stata 8.0 (Stata Corporation, College Station, Tex) was used for statistical analysis.
Results Baseline, angiographic, and PCI characteristics At baseline, African Americans were younger, and a higher percentage of the cohort was female (Table 1). The African American cohort had significantly higher rates of cardiovascular risk factors, including diabetes mellitus, dyslipidemia, hypertension, and current smoking. African Americans were also more likely to have renal insufficiency. However, African Americans were significantly less likely to have had prior myocardial infarction or prior revascularization. Within the 24 hours prior to PCI, African Americans were significantly less likely to have received aspirin or statins but significantly more likely to have received angiotensinconverting enzyme (ACE) inhibitors, intravenous heparin, and intravenous glycoprotein IIb/IIIa inhibitors. African Americans were significantly more likely to present with acute myocardial infarctions (19.8% vs 6.6%; P ⬍ 0.001) and had slightly lower mean left ventricular ejection fraction. Stent use was similar.
Procedural and short-term outcomes Procedural success rates were similar in both races, as were the rates of composite in-hospital death or myocardial infarction (Table 2). African Americans had a significantly higher rate of periprocedural bleeding, both prior to and after propensity adjustment (Figure 1). African Americans also had a significantly higher unadjusted rate of postpro-
1021 cedural renal insufficiency. After propensity score adjustment, rates of postprocedural renal insufficiency were similar, likely because African Americans had higher rates of diabetes and preexisting renal insufficiency (Figure 1). At 30 days, the rate of death or myocardial infarction was similar.
Outcomes at 1 year African Americans had a higher unadjusted composite endpoint of death or myocardial infarction than did whites at 1 year (Table 2) (18.0% vs 14.5%, P ⫽ 0.017), but after propensity adjustment, the difference was no longer significant (hazard ratio [HR] ⫽ 1.18; 95% CI: 0.98 to 1.43, P ⫽ 0.087) (Figures 1 and 2). Among patients who received stents (n ⫽ 5362; 441 African Americans), risk for the composite endpoint of death or myocardial infarction (unadjusted HR for African Americans ⫽ 1.21; 95% CI: 0.96 to 1.53; P ⫽ 0.10; propensity-adjusted HR ⫽ 1.26; 95% CI: 0.99 to 1.61; P ⫽ 0.058) was similar to the relation seen in our entire cohort. Because the unadjusted death or myocardial infarction hazard ratio for African Americans, as compared with whites, was 1.25 but the propensity-adjusted value was no longer significant at 1.18, 28% of the unadjusted race effect was explained by covariates. Higher rates of presentation with acute myocardial infarction and insulin-treated diabetes, as well as lower left ventricular ejection fraction, were the factors most responsible for the differences in 1-year outcomes between African Americans and whites (Table 3).
Discussion Primary finding and comparison with prior studies We conclude that after PCI, African Americans’ procedural success rates and short-term outcomes are similar to whites but have a nonsignificant trend toward a higher 1-year composite death or myocardial infarction hazard ratio after propensity adjustment. Prior studies have shown similar procedural,14,16,17,19 6-month,15 and 1-year16 outcomes after PCI in African Americans and whites, although higher 1-year20 and 2-year event rates in African Americans have been reported.18,21 Because clinical outcomes can be more directly attributed to an intervention shortly after the intervention as compared with several months afterward, it appears that African Americans fare just as well after PCI as whites. Our findings, when interpreted in the context of lower revascularization rates among African Americans,1,2,6,8-13 suggest that continued efforts to optimize the appropriate use of coronary revascularization among African Americans are warranted.24 Differences in long-term outcomes after PCI are likely to be multifactorial in etiology, potentially due to differences
1022 Table 1
The American Journal of Medicine, Vol 118, No 9, September 2005 Baseline, angiographic, and PCI characteristics Percentage, median (interquartile range), or mean ⫾ SD
Characteristic Age, years Male sex Insulin-dependent diabetes mellitus Non-insulin-dependent diabetes mellitus Renal insufficiency Chronic obstructive pulmonary disease Peripheral vascular disease Elevated or treated total cholesterol Low HDL cholesterol Family history of myocardial infarction Hypertension Current smoker Remote myocardial infarction Prior coronary artery bypass graft surgery Prior multiple PCIs Medications 24 hours prior to PCI ACE inhibitor Aspirin Beta-blocker HMG CoA reductase inhibitor Intravenous nitrates Intravenous heparin Thrombolytic therapy (within past 48 hours) Glycoprotein IIB/IIIA inhibitor Preprocedural characteristics Acute myocardial infarction Unstable angina Clinical shock Intra-aortic balloon pump in prior 48 hours Left ventricular ejection fraction (%) Hematocrit level* (mg/dL) Angiographic and PCI characteristics Number of diseased vessels Number of treated sites Longest lesion Worst ACC lesion score† Reference diameter (mm) Bifurcation lesion Ostial lesion In-stent restenosis Rotablator Stent Balloon angioplasty only Site of intervention Proximal left anterior descending artery Left anterior descending artery Right coronary artery Circumflex artery Saphenous vein graft Arterial graft Multivessel PCI
African Americans (n ⫽ 707)
Whites (n ⫽ 8125)
P value
63 (55-71) 60.1 18.4 20.8 7.5 7.5 9.6 26.4 2.3 30.9 77.7 30.3 38.5 14.6 2.3
65 (56-73) 71.8 10.1 16.1 3.9 7.5 8.3 31.6 7.4 36.7 62.3 19.4 43.4 29.7 5.6
⬍0.001 ⬍0.001 ⬍0.001 0.001 ⬍0.001 0.96 0.21 0.0043 ⬍0.001 0.0024 ⬍0.001 ⬍0.001 0.011 ⬍0.001 ⬍0.001
24.6 90.5 44.1 10.9 31.8 68.6 4.2 63.1
20.4 93.6 41.0 14.5 18.4 56.6 3.1 50.4
0.0080 ⬍0.001 0.10 0.0078 ⬍0.001 ⬍0.001 0.082 ⬍0.001
19.8 56.2 0.7 6.2 51 ⫾ 13 38.4 (35.0-41.6)
6.6 61.2 1.2 5.0 53 ⫾ 12 40.5 (37.0-43.4)
⬍0.001 0.0083 0.25 0.14 0.0024 ⬍0.001
2 1 9 3 2.9 19.0 15.7 2.1 4.7 62.4 31.3
2 1 9 3 2.9 19.3 19.8 2.4 12.3 60.6 30.7
⬍0.001 ⬍0.001 0.22 0.97 0.97 0.80 0.0077 0.59 ⬍0.001 0.35 0.78
13.6 41.9 35.6 37.1 4.5 1.7 14.9
(1-2) (1-2) (6-13) (2-4) (2.6-3.3)
14.3 41.4 39.2 35.6 10.2 1.5 16.2
(1-3) (1-2) (6-14) (2-3) (2.6-3.3)
0.61 0.80 0.06 0.44 ⬍0.001 0.74 0.35
ACE ⫽ angiotensin-converting enzyme; HDL ⫽ high-density lipoprotein; HMG CoA ⫽ 3-hydroxy-3-methylglutaryl coenzyme A; PCI ⫽ percutaneous coronary intervention. *n ⫽ 334 for African Americans and 3090 for whites. Prior to 1999, hematocrit was not routinely collected in the PCI database. †For American College of Cardiology (ACC) score, class A lesion ⫽ 1, B1 ⫽ 2, B2 ⫽ 3, and C ⫽ 4. ACC lesion score refers to the worst ACC score of any lesion with intervention.
Chen et al Table 2
PCI outcomes in African Americans and whites
1023
Procedural and 1-year unadjusted outcomes Percentage
Procedural outcome Procedural success In-hospital death or myocardial infarction In-hospital blood transfusion Postprocedural retroperitoneal bleed Postprocedural arteriovenous fistula Postprocedural pseudoaneurysm Postprocedural rise in creatinine (⬎1 mg/dL) 1-year unadjusted outcome Death Myocardial infarction Death or myocardial infarction
in access to and quality of health care24 for African Americans. For instance, African Americans receive less preventive health services than do whites,25 and the physicians treating them have been reported to be less well trained clinically.26 Due to these discrepancies in health care delivery, a lower percentage of African Americans may be taking cardiac medications of proven benefit such as statins, clopidogrel, or aspirin. If such a discrepancy in medical regimens exists, resulting differences in clinical outcomes may become more obvious in longer- rather than short-term follow-up. Explanations for the trend toward worse longterm outcomes after PCI in African Americans merits further investigation.
Periprocedural complications African Americans had higher unadjusted bleeding rates. However, important baseline and treatment differences existed between African Americans and whites. In African
African Americans (n ⫽ 707)
Whites (n ⫽ 8125)
P value
91.1 7.2 12.1 1.3 1.0 0.4 3.7
91.4 6.9 6.3 0.3 1.6 0.6 2.0
0.77 0.72 ⬍0.001 ⬍0.001 0.53 0.22 0.0036
7.9 11.0 18.0
6.1 9.1 14.5
0.050 0.095 0.017
Americans, PCIs were more likely to be urgent procedures. African Americans had a significantly higher rate of use of intravenous platelet glycoprotein IIb/IIIa inhibitors and a slightly higher, but nonsignificant, rate of thrombolytic therapy prior to PCI. Higher rates of bleeding in African Americans may also be due to problems with arterial access, women have smaller arteries than do men and may be more subject to access site complications,27,28 and there was a higher percentage of women in the African American cohort than in the white cohort. After propensity analysis to adjust for potential confounders, however, the significantly higher risk for bleeding complications persisted. Differences in baseline hematocrit may account for some of the disparity in blood transfusion rates, and African Americans did have a significantly lower hematocrit level (Table 1). Finally, differences in responses to periprocedural antithrombotic regimens are also possible. The higher rate of contrast nephropathy in African Americans is likely due to their higher rates of diabetes,
Figure 1 Unadjusted and propensity-adjusted outcomes by race. Hazard ratios are depicted for the endpoints except for periprocedural bleeding and periprocedural rise in creatinine, which are odds ratios. After propensity adjustment, African Americans had an increased risk of periprocedural bleeding (P ⫽ 0.002). CI ⫽ confidence interval; MI ⫽ myocardial infarction.
1024
The American Journal of Medicine, Vol 118, No 9, September 2005
renal insufficiency prior to PCI, and hypertension. After propensity analysis adjusting for potential confounders, African Americans no longer were significantly likely to develop periprocedural renal insufficiency (odds ratio [OR] ⫽ 1.27; 95% CI: 0.82 to 1.97; P ⫽ 0.29).
Study limitations Numerous differences in baseline characteristics existed between African Americans and whites. African American patients were younger, had more cardiovascular comorbid conditions, and had fewer prior coronary revascularizations. However, such baseline differences are in concordance with the differences noted in prior studies, including comparisons of angiography or revascularization rates.16-18,29 The baseline differences that we observed in part reflect the nature of our institution as a tertiary referral center. African Americans comprised a smaller proportion of patients referred from outside the immediate Cleveland area. Our referrals were often patients with extensive coronary artery disease referred for complex intervention. This referral bias resulted in African Americans being less likely to have had prior revascularization or myocardial infarction. In addition, African Americans had lower rates of use of aspirin and statins. Such differences potentially could reflect substandard care received by African Americans. Alternatively, such differences simply may reflect the lower prevalence of known coronary artery disease in African Americans prior to diagnostic heart catheterization. Another important difference between African American and white patients in this study is that a higher proportion of African Americans presented with acute myocardial infarction. This difference occurred because a higher proportion of African Americans were local patients, whereas a higher percentage of white patients lived outside the immediate Cleveland Clinic vicinity and were referred for elective PCI. Propensity analysis should adjust, at least in part, for these baseline differences.
Figure 2 Kaplan-Meier death or myocardial infarction at 1 year, stratified by race. Propensity score-controlled Kaplan-Meier rates of death or myocardial infarction at 1 year in African Americans and whites were similar. CI ⫽ confidence interval; PCI ⫽ percutaneous coronary intervention.
Table 3
Effect of covariates on unadjusted race effect*
Race adjusted for variables Race alone (African American compared with white) Sex Prior myocardial infarction Acute myocardial infarction Diabetes on insulin Diabetes not on insulin Multiple percutaneous coronary interventions Hypertension Glycoprotein IIb/IIIa inhibitor use Current tobacco use Calcium channel blocker Systolic blood pressure Number of diseased vessels Ejection fraction Rotablator use Stent implantation Thrombolytics Longest lesion Vein graft intervention
Death or myocardial infarction at 1 year Hazard ratio (95% confidence interval) 1.25 1.21 1.26 1.15 1.20 1.25
(1.04-1.50) (1.00-1.45) (1.05-1.52) (0.96-1.39) (1.00-1.44) (1.04-1.50)
1.24 1.21 1.25 1.26 1.25 1.27 1.31 1.17 1.28 1.25 1.25 1.28 1.30
(1.03-1.48) (1.01-1.46) (1.04-1.50) (1.05-1.52) (1.04-1.50) (1.06-1.53) (1.09-1.57) (0.96-1.42) (1.07-1.54) (1.04-1.50) (1.04-1.50) (1.06-1.54) (1.08-1.56)
*This table analyzes how much of the effect of race is explained by covariates. For example, the hazard ratio after adjusting for acute myocardial infarction was 1.15 compared with 1.25 prior to adjustment. Thus, the hazard ratio for African Americans was lower after adjustment for acute myocardial infarction, implying that acute myocardial infarction partially explains the worse unadjusted outcomes in African Americans.
References 1. Giles WH, Anda RF, Casper ML, et al. Race and sex differences in rates of invasive cardiac procedures in US hospitals. Data from the National Hospital Discharge Survey. Arch Intern Med. 1995;155: 318 –324. 2. Kressin NR, Petersen LA. Racial differences in the use of invasive cardiovascular procedures: review of the literature and prescription for future research. Ann Intern Med. 2001;135:352–366. 3. LaVeist TA, Arthur M, Morgan A, et al. The cardiac access longitudinal study. A study of access to invasive cardiology among African American and white patients. J Am Coll Cardiol. 2003;41:1159 –1166. 4. Whittle J, Conigliaro J, Good CB, Lofgren RP. Racial differences in the use of invasive cardiovascular procedures in the Department of Veterans Affairs medical system. N Engl J Med. 1993;329:621– 627. 5. Scirica BM, Moliterno DJ, Every NR, et al. Racial differences in the management of unstable angina: results from the multicenter GUARANTEE registry. Am Heart J. 1999;138:1065–1072. 6. Peterson ED, Wright SM, Daley J, Thibault GE: Racial variation in cardiac procedure use and survival following acute myocardial infarction in the Department of Veterans Affairs. JAMA. 1994;271:1175– 1180. 7. Chen J, Rathore SS, Radford MJ, et al. Racial differences in the use of cardiac catheterization after acute myocardial infarction. N Engl J Med. 2001;344:1443–1449. 8. Barnhart JM, Fang J, Alderman MH. Differential use of coronary revascularization and hospital mortality following acute myocardial infarction. Arch Intern Med. 2003;163:461– 466.
Chen et al
PCI outcomes in African Americans and whites
9. Ayanian JZ, Udvarhelyi IS, Gatsonis CA, et al. Racial differences in the use of revascularization procedures after coronary angiography. JAMA. 1993;269:2642–2646. 10. Peterson ED, Shaw LK, DeLong ER, et al. Racial variation in the use of coronary-revascularization procedures. Are the differences real? Do they matter? N Engl J Med. 1997;336:480 – 486. 11. Conigliaro J, Whittle J, Good CB, et al. Understanding racial variation in the use of coronary revascularization procedures: the role of clinical factors. Arch Intern Med. 2000;160:1329 –1335. 12. Gillum RF, Gillum BS, Francis CK. Coronary revascularization and cardiac catheterization in the United States: trends in racial differences. J Am Coll Cardiol. 1997;29:1557–1562. 13. Schneider EC, Leape LL, Weissman JS, et al. Racial differences in cardiac revascularization rates: does “overuse” explain higher rates among white patients? Ann Intern Med. 2001;135:328 –337. 14. Garg M, Vacek JL, Hallas D. Coronary angioplasty in black and white patients: demographic characteristics and outcomes. Southern Med J. 2000;93:1187–1891. 15. Iqbal U, Pinnow EE, Lindsay J, Jr. Comparison of six-month outcomes after percutaneous coronary intervention for whites versus AfricanAmericans. Am J Cardiol. 2001;88:304 –305. 16. Marks DS, Mensah GA, Kennard ED, et al. Race, baseline characteristics, and clinical outcomes after coronary intervention: The New Approaches in Coronary Interventions (NACI) registry. Am Heart J. 2000;140:162–169. 17. Mastoor M, Iqbal U, Pinnow E, Lindsay J, Jr. Ethnicity does not affect outcomes of coronary angioplasty. Clin Cardiol. 2000;23:379 –382. 18. Maynard C, Wright SM, Every NR, Ritchie JL. Racial differences in outcomes of veterans undergoing percutaneous coronary interventions. Am Heart J. 2001;142:309 –313. 19. Scott NA, Kelsey SF, Detre K, et al. Percutaneous transluminal coronary angioplasty in African-American patients (the National Heart, Lung, and Blood Institute 1985-1986 Percutaneous Transluminal Coronary Angioplasty Registry). Am J Cardiol. 1994;73:1141–1146.
1025 20. Leborgne L, Cheneau E, Wolfram R, et al. Comparison of baseline characteristics and one-year outcomes between African-Americans and Caucasians undergoing percutaneous coronary intervention. Am J Cardiol. 2004;93:389 –393. 21. Slater J, Selzer F, Dorbala S, et al. Ethnic differences in the presentation, treatment strategy, and outcomes of percutaneous coronary intervention (a report from the National Heart, Lung, and Blood Institute Dynamic Registry). Am J Cardiol. 2003;92:773– 778. 22. ACC National Cardiovascular Data Registry. Cardiac Cath Lab Module Version 2.0c. Avaliable at: www.acc.org/ncdr. Accessed December 15, 2003. 23. Rubin DB: Estimating causal effects from large data sets using propensity scores. Ann Intern Med. 1997;127:757–763. 24. Lillie-Blanton M, Maddox TM, Rushing O, Mensah GA. Disparities in cardiac care: rising to the challenge of Healthy People 2010. J Am Coll Cardiol. 2004;44:503–508. 25. Gornick ME, Eggers PW, Reilly TW, et al. Effects of race and income on mortality and use of services among Medicare beneficiaries. N Engl J Med. 1996;335:791–799. 26. Bach PB, Pham HH, Schrag D, et al. Primary care physicians who treat blacks and whites. N Engl J Med. 2004;351:575–584. 27. Rozenman Y, Gilon D, Zelingher J, et al. Age- and gender-related differences in success, major and minor complication rates and the duration of hospitalization after percutaneous transluminal coronary angioplasty. Cardiology. 1996;87:396 – 401. 28. Robertson T, Kennard ED, Mehta S, et al. Influence of gender on in-hospital clinical and angiographic outcomes and on one-year follow-up in the New Approaches to Coronary Intervention (NACI) registry. Am J Cardiol. 1997;80:26K–39K. 29. Asher CR, Topol EJ, Moliterno DJ. Insights into the pathophysiology of atherosclerosis and prognosis of black Americans with acute coronary syndromes. Am Heart J. 1999;138:1073–1081.
CLINICAL RESEARCH STUDY
Outcomes in African Americans and whites after percutaneous coronary intervention Michael S. Chen, MD,a Deepak L. Bhatt, MD,a Derek P. Chew, MBBS, MPH,b David J. Moliterno, MD,c Stephen G. Ellis, MD,a Eric J. Topol, MDa a
Department of Cardiovascular Medicine, The Cleveland Clinic Foundation, Cleveland, Ohio; Flinders Medical Centre, Adelaide, Australia; and c Gill Heart Institute and Division of Cardiovascular Medicine, University of Kentucky. b
KEYWORDS: Percutaneous coronary intervention; Revascularization; African American; Race; Coronary artery disease
ABSTRACT PURPOSE: We aimed to determine whether African Americans and whites have different outcomes after percutaneous coronary intervention (PCI). METHODS: We prospectively selected 8832 patients (707 African Americans) for long-term follow-up after PCI at our institution from 1992 to 2002. The primary outcome studied was death or myocardial infarction at 1 year. Propensity adjustment was performed to account for baseline differences between African Americans and whites. RESULTS: African Americans had higher rates of diabetes and less prior revascularization. Percutaneous coronary interventions in African Americans were more often urgent. Stent use was similar. Procedural success rates were similar, as were periprocedural and 30-day composite rates of death or myocardial infarction. In 1-year unadjusted outcomes, African Americans had a higher rate of death or myocardial infarction (18.0% vs 14.5%; hazard ratio (HR) ⫽ 1.25; 95% confidence interval (CI): 1.04 to 1.50; P ⫽ 0.017), but the difference was no longer significant after propensity adjustment (HR ⫽ 1.18; 95% CI: 0.98 to 1.43, P ⫽ 0.087). African Americans had a higher risk for periprocedural bleeding that persisted after propensity adjustment (adjusted odds ratio ⫽ 1.45; 95% CI: 1.14 to 1.84, P ⫽ 0.002). CONCLUSIONS: After PCI, African Americans have similar short-term rates of death or myocardial infarction when compared with whites but have a nonsignificant trend toward worse long-term outcomes. Our findings, when interpreted in the context of reportedly lower revascularization rates among African Americans, suggest that continued efforts to optimize the appropriate use of coronary revascularization among African Americans are warranted. © 2005 Elsevier Inc. All rights reserved.
Background African Americans are reported to have lower rates of diagnostic cardiac catheterization compared with whites,1-4 even in the setting of unstable angina5 or acute myocardial Requests for reprints should be addressed to Deepak L. Bhatt, MD, Director, Interventional Cardiology Fellowship Program, Cleveland Clinic Foundation, Department of Cardiovascular Medicine, 9500 Euclid Avenue, Desk F25, Cleveland, Ohio 44195. E-mail address: [email protected].
0002-9343/$ -see front matter © 2005 Elsevier Inc. All rights reserved. doi:10.1016/j.amjmed.2004.12.035
infarction.6,7 Furthermore, African Americans have lower adjusted rates of revascularization after myocardial infarction or coronary angiography1,2,6,8-13 in most but not all studies.5 Less well characterized is whether race affects outcomes after revascularization. Prior studies generally report similar outcomes after percutaneous coronary intervention (PCI) in African Americans compared with whites,14-19 although higher adjusted 1-year20 and 2-year mortality rates have been reported in African Americans.21 Prior studies are limited by lack of
1020
The American Journal of Medicine, Vol 118, No 9, September 2005
periprocedural outcomes,15 lack of long-term follow-up,17 small sample size of African Americans,14,16 or a homogenous population consisting primarily of male veterans.18 The objective of this study was to provide a comprehensive analysis of PCI outcomes in African Americans and whites.
Methods Study sample The Cleveland Clinic PCI database was queried for all PCIs performed between 1992 and 2002, with race denoted as African American or white. To avoid double counting of patients who had multiple PCIs, only the first PCI for any patient was included, resulting in 16 607 PCIs. Staged procedures were then excluded, resulting in 16 202 PCIs. From 1993 to 1999, due to variations in annual research resources, the first 500 to 1000 PCIs in each year performed with abciximab, as well as the first 500 to 1000 PCIs in each year performed not using abciximab, were prospectively selected for follow-up. The number of abciximab and non-abciximab cases was approximately equal. Case selection from 1993 to 1999 was on a consecutive basis. Since 2000, all patients have been prospectively followed for 1 year after PCI. We included only patients prospectively selected for 1-year follow-up, resulting in 8832 patients (707 African American and 8125 white) in our final cohort. Research nurses reviewed each patient’s chart after the initial PCI to corroborate database entries with the medical record. Access to the interventional registry and data analysis were approved by the institutional review board.
troponin T or I above the limit of myocardial infarction within 24 hours after the clinical event, creatine kinase-MB (CK-MB) greater than 2 times the upper limit of normal on one occasion during the first several hours after the clinical event or CK-MB greater than the upper limit of normal on 2 successive samples, or CK greater than 2 times the upper limit of normal. Troponin and CK-MB were the preferred biochemical methods of diagnosis of myocardial infarction, but CK could be used in the absence of troponin or CK-MB assays. Periprocedural myocardial infarctions were not included in 30-day and 1-year outcomes. Periprocedural and 30-day outcomes were secondary endpoints. All patients had electrocardiograms (ECGs) on the day after PCI and routine serial monitoring of cardiac enzymes. In accordance with American College of Cardiology National Cardiovascular Data Registry definitions, periprocedural myocardial infarction was defined as electrocardiographic changes (evolving ST-segment elevation, new or presumably new left bundle branch block or new Q waves on two or more contiguous leads) or CK-MB greater than 3 times the upper limit of normal.22 After PCI, the decision to perform stress testing or repeat coronary angiography was made by the patient’s cardiologist. Procedural success was defined as procedures in which all attempts resulted in post-intervention stenosis of ⱕ50% of the arterial diameter, grade 3 flow Thrombolysis In Myocardial Infarction (TIMI) was present, and there was at least a 20% decrease in percentage stenosis by visual estimation. Chronic renal insufficiency was defined as serum creatinine greater than 2 mg/dL preprocedure, and acute renal insufficiency was defined as an increase in creatinine of at least 1 mg/dL afterward. Bleeding was defined as blood transfusion, hematoma, or retroperitoneal bleed. Blood transfusions were ordered at the discretion of the attending physician.
Follow-up and endpoints Statistical analysis The primary outcome studied was death or myocardial infarction at 1 year. Research nurses contacted patients at 1 year to ascertain their status. Follow-up was performed by mailings and telephone calls to patients by dedicated research nurses. Of patients prospectively selected, follow-up at 1 year was 97.3% complete for African Americans and 98.3% complete for whites. Of the subjects lost to annual follow-up, none of the African Americans and only 5 whites were lost prior to 30 days. The primary reason that 1-year follow-up could not be obtained was inability to contact the patient, despite mailings and telephone calls. Because rates of follow-up were similar in African Americans and whites, and because follow-up was ⬎97% complete overall, bias from informative censoring was not likely. The myocardial infarction endpoint was based on the American College of Cardiology National Cardiovascular Data Registry, Cardiac Cath Lab Module version 2.0c definitions.22 Myocardial infarction required biomarker elevation and either electrocardiographic changes or ischemic symptoms. Biomarker elevation was defined as elevation of
Continuous variables are expressed as either mean ⫾ standard deviation or median with interquartile range (25% to 75%) and compared by the Mann-Whitney U test. Categorical variables were compared by chi-squared analysis. Propensity analysis23 adjusted for the likelihood of being African American. First, a logistic regression model was constructed for the likelihood of being African American compared with white, with a P value of ⱕ0.25 used for inclusion in the model. Variables included in the propensity score were sex, prior myocardial infarction, acute myocardial infarction, unstable angina, recent myocardial infarction, diabetes on insulin, diabetes not on insulin, history of multiple PCIs, hypertension, glycoprotein IIb/IIIa use, current smoker, calcium channel blocker use, systolic blood pressure, number of diseased vessels, left ventricular ejection fraction, and rotablator use. From this model, the predicted probability for being African American was generated (propensity score). Age was not significant early in the stepwise model and thus was excluded from the propensity
Chen et al
PCI outcomes in African Americans and whites
analysis. African Americans and whites were then divided into blocks of patients based on their propensity score. The optimal number of blocks was 11, which ensured that the mean propensity score was not different between African Americans and whites in each propensity score block. Unadjusted event rates for 30-day ischemia, bleeding, and postprocedural elevation in creatinine were compared by chi-squared analysis, as were the proportion of patients experiencing death or myocardial infarction by 1 year. The hazard ratio for risk associated with being African American was calculated by Cox proportional hazards modeling, with the “propensity for race” used to adjust this estimate for baseline differences between the races. Also, an estimate of relative influence of each baseline characteristic on the adjusted hazard ratio for race was explored by comparing the effect of race on death or myocardial infarction at 1 year in an unadjusted model and contrasting it with a model including race and one other baseline variable. A KaplanMeier plot of survival free of death or myocardial infarction, stratified by race and adjusted for propensity category (propensity score category set to zero), was created. For composite endpoints, the time to the first event was used. Stata 8.0 (Stata Corporation, College Station, Tex) was used for statistical analysis.
Results Baseline, angiographic, and PCI characteristics At baseline, African Americans were younger, and a higher percentage of the cohort was female (Table 1). The African American cohort had significantly higher rates of cardiovascular risk factors, including diabetes mellitus, dyslipidemia, hypertension, and current smoking. African Americans were also more likely to have renal insufficiency. However, African Americans were significantly less likely to have had prior myocardial infarction or prior revascularization. Within the 24 hours prior to PCI, African Americans were significantly less likely to have received aspirin or statins but significantly more likely to have received angiotensinconverting enzyme (ACE) inhibitors, intravenous heparin, and intravenous glycoprotein IIb/IIIa inhibitors. African Americans were significantly more likely to present with acute myocardial infarctions (19.8% vs 6.6%; P ⬍ 0.001) and had slightly lower mean left ventricular ejection fraction. Stent use was similar.
Procedural and short-term outcomes Procedural success rates were similar in both races, as were the rates of composite in-hospital death or myocardial infarction (Table 2). African Americans had a significantly higher rate of periprocedural bleeding, both prior to and after propensity adjustment (Figure 1). African Americans also had a significantly higher unadjusted rate of postpro-
1021 cedural renal insufficiency. After propensity score adjustment, rates of postprocedural renal insufficiency were similar, likely because African Americans had higher rates of diabetes and preexisting renal insufficiency (Figure 1). At 30 days, the rate of death or myocardial infarction was similar.
Outcomes at 1 year African Americans had a higher unadjusted composite endpoint of death or myocardial infarction than did whites at 1 year (Table 2) (18.0% vs 14.5%, P ⫽ 0.017), but after propensity adjustment, the difference was no longer significant (hazard ratio [HR] ⫽ 1.18; 95% CI: 0.98 to 1.43, P ⫽ 0.087) (Figures 1 and 2). Among patients who received stents (n ⫽ 5362; 441 African Americans), risk for the composite endpoint of death or myocardial infarction (unadjusted HR for African Americans ⫽ 1.21; 95% CI: 0.96 to 1.53; P ⫽ 0.10; propensity-adjusted HR ⫽ 1.26; 95% CI: 0.99 to 1.61; P ⫽ 0.058) was similar to the relation seen in our entire cohort. Because the unadjusted death or myocardial infarction hazard ratio for African Americans, as compared with whites, was 1.25 but the propensity-adjusted value was no longer significant at 1.18, 28% of the unadjusted race effect was explained by covariates. Higher rates of presentation with acute myocardial infarction and insulin-treated diabetes, as well as lower left ventricular ejection fraction, were the factors most responsible for the differences in 1-year outcomes between African Americans and whites (Table 3).
Discussion Primary finding and comparison with prior studies We conclude that after PCI, African Americans’ procedural success rates and short-term outcomes are similar to whites but have a nonsignificant trend toward a higher 1-year composite death or myocardial infarction hazard ratio after propensity adjustment. Prior studies have shown similar procedural,14,16,17,19 6-month,15 and 1-year16 outcomes after PCI in African Americans and whites, although higher 1-year20 and 2-year event rates in African Americans have been reported.18,21 Because clinical outcomes can be more directly attributed to an intervention shortly after the intervention as compared with several months afterward, it appears that African Americans fare just as well after PCI as whites. Our findings, when interpreted in the context of lower revascularization rates among African Americans,1,2,6,8-13 suggest that continued efforts to optimize the appropriate use of coronary revascularization among African Americans are warranted.24 Differences in long-term outcomes after PCI are likely to be multifactorial in etiology, potentially due to differences
1022 Table 1
The American Journal of Medicine, Vol 118, No 9, September 2005 Baseline, angiographic, and PCI characteristics Percentage, median (interquartile range), or mean ⫾ SD
Characteristic Age, years Male sex Insulin-dependent diabetes mellitus Non-insulin-dependent diabetes mellitus Renal insufficiency Chronic obstructive pulmonary disease Peripheral vascular disease Elevated or treated total cholesterol Low HDL cholesterol Family history of myocardial infarction Hypertension Current smoker Remote myocardial infarction Prior coronary artery bypass graft surgery Prior multiple PCIs Medications 24 hours prior to PCI ACE inhibitor Aspirin Beta-blocker HMG CoA reductase inhibitor Intravenous nitrates Intravenous heparin Thrombolytic therapy (within past 48 hours) Glycoprotein IIB/IIIA inhibitor Preprocedural characteristics Acute myocardial infarction Unstable angina Clinical shock Intra-aortic balloon pump in prior 48 hours Left ventricular ejection fraction (%) Hematocrit level* (mg/dL) Angiographic and PCI characteristics Number of diseased vessels Number of treated sites Longest lesion Worst ACC lesion score† Reference diameter (mm) Bifurcation lesion Ostial lesion In-stent restenosis Rotablator Stent Balloon angioplasty only Site of intervention Proximal left anterior descending artery Left anterior descending artery Right coronary artery Circumflex artery Saphenous vein graft Arterial graft Multivessel PCI
African Americans (n ⫽ 707)
Whites (n ⫽ 8125)
P value
63 (55-71) 60.1 18.4 20.8 7.5 7.5 9.6 26.4 2.3 30.9 77.7 30.3 38.5 14.6 2.3
65 (56-73) 71.8 10.1 16.1 3.9 7.5 8.3 31.6 7.4 36.7 62.3 19.4 43.4 29.7 5.6
⬍0.001 ⬍0.001 ⬍0.001 0.001 ⬍0.001 0.96 0.21 0.0043 ⬍0.001 0.0024 ⬍0.001 ⬍0.001 0.011 ⬍0.001 ⬍0.001
24.6 90.5 44.1 10.9 31.8 68.6 4.2 63.1
20.4 93.6 41.0 14.5 18.4 56.6 3.1 50.4
0.0080 ⬍0.001 0.10 0.0078 ⬍0.001 ⬍0.001 0.082 ⬍0.001
19.8 56.2 0.7 6.2 51 ⫾ 13 38.4 (35.0-41.6)
6.6 61.2 1.2 5.0 53 ⫾ 12 40.5 (37.0-43.4)
⬍0.001 0.0083 0.25 0.14 0.0024 ⬍0.001
2 1 9 3 2.9 19.0 15.7 2.1 4.7 62.4 31.3
2 1 9 3 2.9 19.3 19.8 2.4 12.3 60.6 30.7
⬍0.001 ⬍0.001 0.22 0.97 0.97 0.80 0.0077 0.59 ⬍0.001 0.35 0.78
13.6 41.9 35.6 37.1 4.5 1.7 14.9
(1-2) (1-2) (6-13) (2-4) (2.6-3.3)
14.3 41.4 39.2 35.6 10.2 1.5 16.2
(1-3) (1-2) (6-14) (2-3) (2.6-3.3)
0.61 0.80 0.06 0.44 ⬍0.001 0.74 0.35
ACE ⫽ angiotensin-converting enzyme; HDL ⫽ high-density lipoprotein; HMG CoA ⫽ 3-hydroxy-3-methylglutaryl coenzyme A; PCI ⫽ percutaneous coronary intervention. *n ⫽ 334 for African Americans and 3090 for whites. Prior to 1999, hematocrit was not routinely collected in the PCI database. †For American College of Cardiology (ACC) score, class A lesion ⫽ 1, B1 ⫽ 2, B2 ⫽ 3, and C ⫽ 4. ACC lesion score refers to the worst ACC score of any lesion with intervention.
Chen et al Table 2
PCI outcomes in African Americans and whites
1023
Procedural and 1-year unadjusted outcomes Percentage
Procedural outcome Procedural success In-hospital death or myocardial infarction In-hospital blood transfusion Postprocedural retroperitoneal bleed Postprocedural arteriovenous fistula Postprocedural pseudoaneurysm Postprocedural rise in creatinine (⬎1 mg/dL) 1-year unadjusted outcome Death Myocardial infarction Death or myocardial infarction
in access to and quality of health care24 for African Americans. For instance, African Americans receive less preventive health services than do whites,25 and the physicians treating them have been reported to be less well trained clinically.26 Due to these discrepancies in health care delivery, a lower percentage of African Americans may be taking cardiac medications of proven benefit such as statins, clopidogrel, or aspirin. If such a discrepancy in medical regimens exists, resulting differences in clinical outcomes may become more obvious in longer- rather than short-term follow-up. Explanations for the trend toward worse longterm outcomes after PCI in African Americans merits further investigation.
Periprocedural complications African Americans had higher unadjusted bleeding rates. However, important baseline and treatment differences existed between African Americans and whites. In African
African Americans (n ⫽ 707)
Whites (n ⫽ 8125)
P value
91.1 7.2 12.1 1.3 1.0 0.4 3.7
91.4 6.9 6.3 0.3 1.6 0.6 2.0
0.77 0.72 ⬍0.001 ⬍0.001 0.53 0.22 0.0036
7.9 11.0 18.0
6.1 9.1 14.5
0.050 0.095 0.017
Americans, PCIs were more likely to be urgent procedures. African Americans had a significantly higher rate of use of intravenous platelet glycoprotein IIb/IIIa inhibitors and a slightly higher, but nonsignificant, rate of thrombolytic therapy prior to PCI. Higher rates of bleeding in African Americans may also be due to problems with arterial access, women have smaller arteries than do men and may be more subject to access site complications,27,28 and there was a higher percentage of women in the African American cohort than in the white cohort. After propensity analysis to adjust for potential confounders, however, the significantly higher risk for bleeding complications persisted. Differences in baseline hematocrit may account for some of the disparity in blood transfusion rates, and African Americans did have a significantly lower hematocrit level (Table 1). Finally, differences in responses to periprocedural antithrombotic regimens are also possible. The higher rate of contrast nephropathy in African Americans is likely due to their higher rates of diabetes,
Figure 1 Unadjusted and propensity-adjusted outcomes by race. Hazard ratios are depicted for the endpoints except for periprocedural bleeding and periprocedural rise in creatinine, which are odds ratios. After propensity adjustment, African Americans had an increased risk of periprocedural bleeding (P ⫽ 0.002). CI ⫽ confidence interval; MI ⫽ myocardial infarction.
1024
The American Journal of Medicine, Vol 118, No 9, September 2005
renal insufficiency prior to PCI, and hypertension. After propensity analysis adjusting for potential confounders, African Americans no longer were significantly likely to develop periprocedural renal insufficiency (odds ratio [OR] ⫽ 1.27; 95% CI: 0.82 to 1.97; P ⫽ 0.29).
Study limitations Numerous differences in baseline characteristics existed between African Americans and whites. African American patients were younger, had more cardiovascular comorbid conditions, and had fewer prior coronary revascularizations. However, such baseline differences are in concordance with the differences noted in prior studies, including comparisons of angiography or revascularization rates.16-18,29 The baseline differences that we observed in part reflect the nature of our institution as a tertiary referral center. African Americans comprised a smaller proportion of patients referred from outside the immediate Cleveland area. Our referrals were often patients with extensive coronary artery disease referred for complex intervention. This referral bias resulted in African Americans being less likely to have had prior revascularization or myocardial infarction. In addition, African Americans had lower rates of use of aspirin and statins. Such differences potentially could reflect substandard care received by African Americans. Alternatively, such differences simply may reflect the lower prevalence of known coronary artery disease in African Americans prior to diagnostic heart catheterization. Another important difference between African American and white patients in this study is that a higher proportion of African Americans presented with acute myocardial infarction. This difference occurred because a higher proportion of African Americans were local patients, whereas a higher percentage of white patients lived outside the immediate Cleveland Clinic vicinity and were referred for elective PCI. Propensity analysis should adjust, at least in part, for these baseline differences.
Figure 2 Kaplan-Meier death or myocardial infarction at 1 year, stratified by race. Propensity score-controlled Kaplan-Meier rates of death or myocardial infarction at 1 year in African Americans and whites were similar. CI ⫽ confidence interval; PCI ⫽ percutaneous coronary intervention.
Table 3
Effect of covariates on unadjusted race effect*
Race adjusted for variables Race alone (African American compared with white) Sex Prior myocardial infarction Acute myocardial infarction Diabetes on insulin Diabetes not on insulin Multiple percutaneous coronary interventions Hypertension Glycoprotein IIb/IIIa inhibitor use Current tobacco use Calcium channel blocker Systolic blood pressure Number of diseased vessels Ejection fraction Rotablator use Stent implantation Thrombolytics Longest lesion Vein graft intervention
Death or myocardial infarction at 1 year Hazard ratio (95% confidence interval) 1.25 1.21 1.26 1.15 1.20 1.25
(1.04-1.50) (1.00-1.45) (1.05-1.52) (0.96-1.39) (1.00-1.44) (1.04-1.50)
1.24 1.21 1.25 1.26 1.25 1.27 1.31 1.17 1.28 1.25 1.25 1.28 1.30
(1.03-1.48) (1.01-1.46) (1.04-1.50) (1.05-1.52) (1.04-1.50) (1.06-1.53) (1.09-1.57) (0.96-1.42) (1.07-1.54) (1.04-1.50) (1.04-1.50) (1.06-1.54) (1.08-1.56)
*This table analyzes how much of the effect of race is explained by covariates. For example, the hazard ratio after adjusting for acute myocardial infarction was 1.15 compared with 1.25 prior to adjustment. Thus, the hazard ratio for African Americans was lower after adjustment for acute myocardial infarction, implying that acute myocardial infarction partially explains the worse unadjusted outcomes in African Americans.
References 1. Giles WH, Anda RF, Casper ML, et al. Race and sex differences in rates of invasive cardiac procedures in US hospitals. Data from the National Hospital Discharge Survey. Arch Intern Med. 1995;155: 318 –324. 2. Kressin NR, Petersen LA. Racial differences in the use of invasive cardiovascular procedures: review of the literature and prescription for future research. Ann Intern Med. 2001;135:352–366. 3. LaVeist TA, Arthur M, Morgan A, et al. The cardiac access longitudinal study. A study of access to invasive cardiology among African American and white patients. J Am Coll Cardiol. 2003;41:1159 –1166. 4. Whittle J, Conigliaro J, Good CB, Lofgren RP. Racial differences in the use of invasive cardiovascular procedures in the Department of Veterans Affairs medical system. N Engl J Med. 1993;329:621– 627. 5. Scirica BM, Moliterno DJ, Every NR, et al. Racial differences in the management of unstable angina: results from the multicenter GUARANTEE registry. Am Heart J. 1999;138:1065–1072. 6. Peterson ED, Wright SM, Daley J, Thibault GE: Racial variation in cardiac procedure use and survival following acute myocardial infarction in the Department of Veterans Affairs. JAMA. 1994;271:1175– 1180. 7. Chen J, Rathore SS, Radford MJ, et al. Racial differences in the use of cardiac catheterization after acute myocardial infarction. N Engl J Med. 2001;344:1443–1449. 8. Barnhart JM, Fang J, Alderman MH. Differential use of coronary revascularization and hospital mortality following acute myocardial infarction. Arch Intern Med. 2003;163:461– 466.
Chen et al
PCI outcomes in African Americans and whites
9. Ayanian JZ, Udvarhelyi IS, Gatsonis CA, et al. Racial differences in the use of revascularization procedures after coronary angiography. JAMA. 1993;269:2642–2646. 10. Peterson ED, Shaw LK, DeLong ER, et al. Racial variation in the use of coronary-revascularization procedures. Are the differences real? Do they matter? N Engl J Med. 1997;336:480 – 486. 11. Conigliaro J, Whittle J, Good CB, et al. Understanding racial variation in the use of coronary revascularization procedures: the role of clinical factors. Arch Intern Med. 2000;160:1329 –1335. 12. Gillum RF, Gillum BS, Francis CK. Coronary revascularization and cardiac catheterization in the United States: trends in racial differences. J Am Coll Cardiol. 1997;29:1557–1562. 13. Schneider EC, Leape LL, Weissman JS, et al. Racial differences in cardiac revascularization rates: does “overuse” explain higher rates among white patients? Ann Intern Med. 2001;135:328 –337. 14. Garg M, Vacek JL, Hallas D. Coronary angioplasty in black and white patients: demographic characteristics and outcomes. Southern Med J. 2000;93:1187–1891. 15. Iqbal U, Pinnow EE, Lindsay J, Jr. Comparison of six-month outcomes after percutaneous coronary intervention for whites versus AfricanAmericans. Am J Cardiol. 2001;88:304 –305. 16. Marks DS, Mensah GA, Kennard ED, et al. Race, baseline characteristics, and clinical outcomes after coronary intervention: The New Approaches in Coronary Interventions (NACI) registry. Am Heart J. 2000;140:162–169. 17. Mastoor M, Iqbal U, Pinnow E, Lindsay J, Jr. Ethnicity does not affect outcomes of coronary angioplasty. Clin Cardiol. 2000;23:379 –382. 18. Maynard C, Wright SM, Every NR, Ritchie JL. Racial differences in outcomes of veterans undergoing percutaneous coronary interventions. Am Heart J. 2001;142:309 –313. 19. Scott NA, Kelsey SF, Detre K, et al. Percutaneous transluminal coronary angioplasty in African-American patients (the National Heart, Lung, and Blood Institute 1985-1986 Percutaneous Transluminal Coronary Angioplasty Registry). Am J Cardiol. 1994;73:1141–1146.
1025 20. Leborgne L, Cheneau E, Wolfram R, et al. Comparison of baseline characteristics and one-year outcomes between African-Americans and Caucasians undergoing percutaneous coronary intervention. Am J Cardiol. 2004;93:389 –393. 21. Slater J, Selzer F, Dorbala S, et al. Ethnic differences in the presentation, treatment strategy, and outcomes of percutaneous coronary intervention (a report from the National Heart, Lung, and Blood Institute Dynamic Registry). Am J Cardiol. 2003;92:773– 778. 22. ACC National Cardiovascular Data Registry. Cardiac Cath Lab Module Version 2.0c. Avaliable at: www.acc.org/ncdr. Accessed December 15, 2003. 23. Rubin DB: Estimating causal effects from large data sets using propensity scores. Ann Intern Med. 1997;127:757–763. 24. Lillie-Blanton M, Maddox TM, Rushing O, Mensah GA. Disparities in cardiac care: rising to the challenge of Healthy People 2010. J Am Coll Cardiol. 2004;44:503–508. 25. Gornick ME, Eggers PW, Reilly TW, et al. Effects of race and income on mortality and use of services among Medicare beneficiaries. N Engl J Med. 1996;335:791–799. 26. Bach PB, Pham HH, Schrag D, et al. Primary care physicians who treat blacks and whites. N Engl J Med. 2004;351:575–584. 27. Rozenman Y, Gilon D, Zelingher J, et al. Age- and gender-related differences in success, major and minor complication rates and the duration of hospitalization after percutaneous transluminal coronary angioplasty. Cardiology. 1996;87:396 – 401. 28. Robertson T, Kennard ED, Mehta S, et al. Influence of gender on in-hospital clinical and angiographic outcomes and on one-year follow-up in the New Approaches to Coronary Intervention (NACI) registry. Am J Cardiol. 1997;80:26K–39K. 29. Asher CR, Topol EJ, Moliterno DJ. Insights into the pathophysiology of atherosclerosis and prognosis of black Americans with acute coronary syndromes. Am Heart J. 1999;138:1073–1081.