The New England Journal of Medicine
e-mail icon  FREE NEJM E-TOC    HOME   |   SUBSCRIBE   |   CURRENT ISSUE   |   PAST ISSUES   |   COLLECTIONS   |    Advanced Search
Sign in | Get NEJM's E-Mail Table of Contents — Free | Subscribe
 
Original Article
PreviousPrevious
Volume 352:2174-2183 May 26, 2005 Number 21
NextNext

Long-Term Outcomes of Coronary-Artery Bypass Grafting versus Stent Implantation
Edward L. Hannan, Ph.D., Michael J. Racz, Ph.D., Gary Walford, M.D., Robert H. Jones, M.D., Thomas J. Ryan, M.D., Edward Bennett, M.D., Alfred T. Culliford, M.D., O. Wayne Isom, M.D., Jeffrey P. Gold, M.D., and Eric A. Rose, M.D.

 

This Article
-Abstract
- PDF
-PDA Full Text
-PowerPoint Slide Set

Commentary
-Editorial
 by Gersh, B. J.
-Letters

Tools and Services
-Add to Personal Archive
-Add to Citation Manager
-Notify a Friend
-E-mail When Cited
-E-mail When Letters Appear

More Information
-PubMed Citation
ABSTRACT

Background Several studies have compared outcomes for coronary-artery bypass grafting (CABG) and percutaneous coronary intervention (PCI), but most were done before the availability of stenting, which has revolutionized the latter approach.

Methods We used New York's cardiac registries to identify 37,212 patients with multivessel disease who underwent CABG and 22,102 patients with multivessel disease who underwent PCI from January 1, 1997, to December 31, 2000. We determined the rates of death and subsequent revascularization within three years after the procedure in various groups of patients according to the number of diseased vessels and the presence or absence of involvement of the left anterior descending coronary artery. The rates of adverse outcomes were adjusted by means of proportional-hazards methods to account for differences in patients' severity of illness before revascularization.

Results Risk-adjusted survival rates were significantly higher among patients who underwent CABG than among those who received a stent in all of the anatomical subgroups studied. For example, the adjusted hazard ratio for the long-term risk of death after CABG relative to stent implantation was 0.64 (95 percent confidence interval, 0.56 to 0.74) for patients with three-vessel disease with involvement of the proximal left anterior descending coronary artery and 0.76 (95 percent confidence interval, 0.60 to 0.96) for patients with two-vessel disease with involvement of the nonproximal left anterior descending coronary artery. Also, the three-year rates of revascularization were considerably higher in the stenting group than in the CABG group (7.8 percent vs. 0.3 percent for subsequent CABG and 27.3 percent vs. 4.6 percent for subsequent PCI).

Conclusions For patients with two or more diseased coronary arteries, CABG is associated with higher adjusted rates of long-term survival than stenting.


Coronary-artery bypass grafting (CABG) and percutaneous coronary intervention (PCI) have long been the definitive aggressive options for treating patients with coronary artery disease. In the past few years, several randomized clinical trials1,2,3,4,5,6,7,8,9,10,11,12,13,14,15 and observational studies16,17,18,19,20 have examined the relative long-term benefits of these interventions. However, with few exceptions, these studies were conducted before the availability of stenting.8,11,13,15 We used observational data from a very large registry to compare short-term and long-term outcomes among patients with multivessel disease who underwent CABG or stenting in New York State hospitals.

Methods

Databases

The two main databases used in the study were New York's two cardiac registries, the Cardiac Surgery Reporting System (CSRS) and the Percutaneous Coronary Intervention Reporting System (PCIRS). These databases include information about numerous risk factors; admission, surgery, and discharge dates; and discharge status for all patients undergoing CABG and PCI in nonfederal hospitals in New York State.

To ensure that these registries are complete, data are matched to New York State's acute care hospital-discharge data set, the Statewide Planning and Research Cooperative System (SPARCS), which is used for reimbursement. Also, reporting of in-hospital deaths is matched against SPARCS data, and hospitals are asked to resolve discrepancies. Accuracy of risk-factor reporting is ascertained by having the Department of Health's utilization-review agent audit hospitals' medical records. To identify deaths that did not occur in hospitals, we used patients' Social Security numbers to link New York's Vital Statistics Death File to the two registries. The institutional review board of the University of Albany approved the use of these data for this study.

Study Group and End Points

The study included New York residents with multivessel coronary artery disease (defined as stenosis of at least 70 percent in at least two of the three main coronary arteries) who underwent isolated CABG or stent implantation for at least one lesion between January 1, 1997, and December 31, 2000, in New York. Patients who had previously undergone revascularization, those with disease of the left main coronary artery (defined as stenosis of more than 50 percent), and those who had an acute myocardial infarction within 24 hours before revascularization were excluded. The study was limited to New York residents because the New York Vital Statistics Death File applies only to residents of the state. The study group comprised 37,212 patients who underwent CABG at the 34 hospitals in the state in which CABG was performed and 22,102 patients who underwent stenting at the 35 hospitals in the state in which PCI was performed.

End points included death and death or revascularization (CABG or PCI) at any time before December 31, 2000. Subsequent revascularizations were identified by matching patients' Social Security numbers with CABG and PCI procedures listed in CSRS and PCIRS.

Statistical Analysis

The primary purpose of the data analyses was to determine whether long-term mortality differed significantly between patients undergoing CABG and those undergoing PCI, after controlling for differences in patients' preprocedural risk. A secondary purpose was to determine whether there were treatment-related differences in long-term mortality among patients with diabetes, patients with compromised ventricular function (defined by a left ventricular ejection fraction of less than 40 percent), and patients without compromised ventricular function. We planned to accomplish this by first identifying factors that were associated with a risk of death in univariate analyses and then using a multivariate (Cox proportional-hazards) model that controlled for significant risk factors while testing for significant differences in long-term mortality between patients undergoing CABG and those undergoing PCI. Because the number of diseased vessels, the presence or absence of involvement of the left anterior descending coronary (LAD) artery, and the presence or absence of disease in the proximal portion of the LAD artery are major factors in the determination of which procedure to perform, the analyses were performed separately for each of five anatomical groups among patients with two- or three-vessel disease (patients with single-vessel disease were omitted because the vast majority who undergo revascularization do so by means of PCI).

We compared the prevalences of potential risk factors for each procedure using chi-square and Fisher's exact (for binary risk factors) tests. Factors included the number of diseased vessels, the patient's age and sex, the presence or absence of a variety of coexisting conditions, and measures of the patient's hemodynamic state and ventricular function. Kaplan–Meier estimates were used to plot the percentage of patients in each group who underwent subsequent revascularization; data on patients who died before subsequent revascularization were censored.

Each patient was placed in one of five anatomical groups according to whether two or three vessels were diseased, whether there was clinically significant disease in the LAD artery, and if there was, whether it was in the proximal region. A backward stepwise Cox model was used to identify risk factors in each subgroup that were significantly related to long-term mortality. A P value of less than 0.05 was used to indicate statistical significance. The type of revascularization (CABG vs. PCI) was then added as an independent binary variable, with "1" denoting CABG.

For each of the five groups of patients, point estimates and confidence intervals were calculated for hazard ratios.21 Also, unadjusted survival curves were generated for each intervention for each anatomical group with the use of Kaplan–Meier estimates, and adjusted survival curves were generated with the use of the Cox model in conjunction with methods described by Ghali et al.22 The log-rank test was used to identify significant differences in unadjusted survival rates. As prespecified in the protocol, survival rates in all anatomical subgroups were evaluated to identify differences in the subgroups of patients with diabetes, patients with compromised ventricular function, and those without such compromise.

Selection bias was examined with the use of a propensity model.23,24 Covariates that were considered for inclusion in the models included all variables presented in Table 1. Variables that were significant predictors of the type of revascularization were identified by fitting a logistic-regression model with a binary dependent variable representing CABG. For each anatomical group, the propensity score was subdivided into quintiles and hazard ratios were compared across quintiles. All analyses were conducted with the use of SAS software (version 8.2), and all reported P values are two-sided.

View this table:
[in this window]
[in a new window]
 
Table 1. Baseline Characteristics of the Patients.

 
Results

The median follow-up was 706.0 days in the CABG group (interquartile range, 328 to 1089) and 585.0 days in the stenting group (interquartile range, 265 to 948). Patients who underwent CABG were significantly older than patients who received stents, although more patients in the latter group were 80 years of age or older. Also, patients who underwent CABG were more likely to be white men and less likely to be Hispanic. Patients who underwent CABG had significantly lower median ejection fractions and were less likely than patients who received stents to have had a myocardial infarction in the week before the procedure. Patients who underwent CABG also had a significantly higher prevalence of numerous coexisting conditions and were significantly more likely to have three-vessel disease (Table 1).

In the stenting group, 7.8 percent underwent subsequent CABG and 27.3 percent underwent repeated PCI in the three years after the initial procedure. Also, 0.3 percent of the CABG group underwent CABG and 4.6 percent underwent PCI in the ensuing three-year period (Figure 1). The overall rates of revascularization were significantly lower in the CABG group than in the stenting group (P<0.001).


View larger version (7K):
[in this window]
[in a new window]
 
Figure 1. Percentage of Patients Undergoing a Second Revascularization Procedure within Three Years.

 
The observed (unadjusted) in-hospital mortality rate among the patients in the CABG group was significantly higher than the rate among patients who received a stent (1.75 percent [650 deaths] vs. 0.68 percent [150 deaths], P<0.001). The respective adjusted hazard ratios ranged from 0.76 (P=0.02) for patients with two-vessel disease and involvement of the nonproximal LAD artery to 0.64 (P<0.001) for patients with three-vessel disease and involvement of the proximal LAD artery (Table 2). Significant risk factors in the statistical models included a lower ejection fraction; the presence of diabetes, congestive heart failure, chronic obstructive pulmonary disease, carotid-artery disease, aortoiliac disease, shock, renal failure, femoral or popliteal disease, and stroke; advanced age; and male sex. Figure 2 presents unadjusted survival curves and Figure 3 presents adjusted survival curves for three of the anatomical groups. For the unadjusted curves, the survival rate was significantly higher after stent placement than after CABG among patients who had two-vessel disease without involvement of the LAD artery (P=0.03), whereas the opposite was true among patients who had three-vessel disease with involvement of the proximal LAD artery (P<0.001); there were no significant treatment-related differences in survival in the other anatomical groups.

View this table:
[in this window]
[in a new window]
 
Table 2. Hazard Ratios for Death after CABG as Compared with after Stenting in Various Subgroups.

 

View larger version (11K):
[in this window]
[in a new window]
 
Figure 2. Unadjusted Kaplan–Meier Survival Curves among Patients with Two-Vessel Disease without Involvement of the LAD Artery (Panel A), Patients with Two-Vessel Disease with Involvement of the Proximal LAD Artery (Panel B), and Patients with Three-Vessel Disease with Involvement of the Proximal LAD Artery (Panel C).

Values are percentages at one, two, and three years.

 

View larger version (12K):
[in this window]
[in a new window]
 
Figure 3. Adjusted Survival Curves among Patients with Two-Vessel Disease without Involvement of the LAD Artery (Panel A), Patients with Two-Vessel Disease with Involvement of the Proximal LAD Artery (Panel B), and Patients with Three-Vessel Disease with Involvement of the Proximal LAD Artery (Panel C).

Values are percentages at one, two, and three years; they were adjusted for the ejection fraction; the presence or absence of diabetes, congestive heart failure, chronic obstructive pulmonary disease, carotid-artery disease, aortoiliac disease, shock, renal failure, femoral or popliteal disease, and stroke; age; and sex.

 
After adjustment for the severity of illness before revascularization, CABG was associated with a significantly higher likelihood of survival in all anatomical groups, as indicated in Table 2. The finding that unadjusted survival estimates favored PCI and adjusted survival estimates favored CABG for patients with two-vessel disease without LAD-artery involvement can be explained by the fact that the prevalences of several important risk factors were much higher in the CABG group than in the stenting group (e.g., 3.4 percent vs. 2.2 percent for renal failure, 16.4 percent vs. 5.9 percent for chronic obstructive pulmonary disease, and 12.3 percent vs. 7.0 percent for congestive heart failure).

When the subgroup of patients with diabetes was analyzed, the adjusted hazard ratios were lower after CABG than after stenting in all anatomical subgroups except patients who had three-vessel disease with involvement of the proximal LAD artery (0.69; 95 percent confidence interval, 0.55 to 0.86). For patients with a left ventricular ejection fraction below 40 percent, the hazard ratios were significantly in favor of CABG, as compared with stenting, for patients with three-vessel disease and patients with two-vessel disease with involvement of the proximal LAD artery, but the hazard ratios were not significant for the other patients with two-vessel disease. For patients with no left ventricular dysfunction (as defined by an ejection fraction of at least 40 percent), CABG was associated with significantly lower adjusted hazard ratios in all anatomical groups (Table 2).

The majority of patients with two-vessel disease underwent stenting, whereas the majority of patients with three-vessel disease underwent CABG. For example, 90.6 percent of patients who had three-vessel disease with involvement of the proximal LAD artery underwent CABG, and patients in this group who underwent stent implantation were more likely than patients who underwent CABG to be in shock (0.4 percent vs. 0.2 percent) or to have received cardiopulmonary resuscitation (0.2 percent vs. 0.04 percent).

In the propensity analyses, significant covariates included age, sex, race, left ventricular ejection fraction, and presence or absence of a previous myocardial infarction and a variety of coexisting conditions, including diabetes, chronic obstructive pulmonary disease, carotid disease, femoral or popliteal disease, and renal failure with a need for dialysis. Distributions of patients who underwent CABG ranged from about 10 percent to 40 percent across quintiles for the anatomical groups with two-vessel disease and no disease of the proximal LAD artery, from 82 percent to 97 percent for patients who had three-vessel disease with disease of the proximal LAD artery, and from about 45 percent to 85 percent for the other two groups of patients with multivessel disease. Hazard ratios consistently favored CABG surgery over stent implantation, with 22 of the 25 anatomical-subgroup quintiles having hazard ratios between 0.44 and 0.84.

Discussion

Our observational study, which included 37,212 patients who underwent CABG and 22,102 patients who underwent stenting, found that the adjusted hazard ratios for the long-term risk of death after CABG relative to stent implantation ranged from 0.76 (95 percent confidence interval, 0.60 to 0.96) for patients with two-vessel disease with involvement of the nonproximal LAD artery to 0.64 (95 percent confidence interval, 0.56 to 0.74) for three-vessel disease with involvement of the proximal LAD artery. Also, revascularization rates were considerably higher after stenting than after CABG (7.8 percent vs. 0.3 percent for subsequent CABG surgery and 27.3 percent vs. 4.6 percent for subsequent PCI).

Most randomized, controlled trials predating the stenting era that compared CABG with PCI did not find significant differences in long-term mortality between the two interventions. This finding may have been due in part to insufficient statistical power. For example, in the Bypass Angioplasty Revascularization Investigation study, the five-year mortality rate was 10.7 percent after CABG, as compared with 13.7 percent after PCI, but this difference was not significant (P=0.19), at least partially because of the relatively small numbers of patients in the study (914 and 915, respectively).5 A meta-analysis of randomized, controlled trials that compared CABG with PCI found that for patients with multivessel disease, CABG provided survival advantages of 2.3 percent (P=0.03) at five years and 3.4 percent (P=0.03) at eight years.19

Prior information on relative long-term outcomes for CABG and stent implantation has been limited to four randomized, controlled trials with relatively small numbers of patients (123 to 1205 patients in each study).8,11,13,15 During follow-up periods ranging from one to three years, one of these studies found significantly lower mortality rates after CABG,8 one found significantly higher mortality rates after CABG,13 and the other two, one of which included patients with single-vessel disease, found no significant difference in mortality rates between the two interventions.11,15 A meta-analysis of these studies found no significant differences in one-year or three-year mortality rates between the two interventions19 but did find that patients who underwent CABG had significantly fewer subsequent revascularizations (15 percent fewer at one year and 15 percent fewer at three years) than patients who underwent stenting.19 As evidenced by the small numbers of patients in the studies of long-term survival after stenting, as compared with CABG, very little information is currently available from randomized, controlled trials.

A caveat of our study is that it is observational, and not a randomized, controlled trial. As such, the choice of treatment was left to the physician, and one of the treatment options may have been contraindicated in some patients. Also, the analysis of survival does not include patients who died before they could undergo one of the scheduled revascularization procedures. Consequently, if there is a longer wait for one of the procedures and this results in additional adverse outcomes, this fact would not have been reflected by our findings.

We attempted to minimize the selection bias by adjusting for differences among patients in demographic characteristics, coexisting conditions, ventricular function, and hemodynamic state. Furthermore, propensity analyses demonstrated that the survival advantage for CABG persisted regardless of the chance of patients' being referred for stent implantation rather than CABG. In addition, comprehensive studies that have compared the results of randomized, controlled trials and observational studies in assessing two competing interventions have concluded that the treatment effects are not qualitatively different and that the observational studies do not overestimate the magnitude of the treatment effects.25,26 Nevertheless, there remains the possibility that observational studies may fail to identify all confounders, and it is also true that propensity analyses cannot account for selection bias related to unmeasured characteristics.

Another important caveat is that although this study is based on data as recent as those used by any other study comparing these two interventions, the state of the art for both treatments is rapidly changing. The most notable of these changes is probably the availability of drug-eluting stents, but the use of CABG without cardiopulmonary bypass and cardiac arrest ("off pump") has also become much more common in the past few years and is underrepresented in our study. In particular, future studies will be needed to compare long-term outcomes for drug-eluting stents with those for CABG in patients with various types of risk factors before the procedure.27,28,29

There are many considerations when one is choosing an intervention for patients with ischemic heart disease. Stenting is far less invasive than CABG, and many patients may prefer to have one or more stents implanted in the hope of avoiding CABG. As noted earlier, stent implantation has a much lower in-hospital mortality rate than CABG. Depending on the patient's life expectancy and attitude about the relative merits of short-term and long-term risk, stent implantation may be preferred even for a patient in an anatomical group for which CABG has significantly higher risk-adjusted long-term survival. Howard et al. provide an excellent comprehensive discussion of these trade-offs.30

As with our earlier comparison of CABG with PCI in the infancy of stent implantation,16 we restricted ourselves to the study of New York State residents because we were unable to determine when patients who lived in other states died after discharge. However, we had no way of knowing whether our patients moved to another state and died there or underwent revascularization there, since we used only the New York State death file. Nevertheless, we do not believe that any bias was introduced by this limitation, because we would expect roughly the same percentage of patients in each intervention group to be lost to follow-up.16

We are indebted to Kenneth Shine, M.D., the chair of New York State's Cardiac Advisory Committee, and to the other members of the committee for their encouragement and support of this study; and to Paula Wauselauskas, Casey Joseph, Kimberly S. Cozzens, Rosemary Lombardo, and the cardiac-surgery departments and cardiac-catheterization laboratories of the 35 participating hospitals for their tireless efforts to ensure the timeliness, completeness, and accuracy of the registry data.


Source Information

From the University at Albany, State University of New York, Albany (E.L.H., M.J.R.); St. Joseph's Hospital, Syracuse, N.Y. (G.W.); Duke University Medical Center, Durham, N.C. (R.H.J.); Boston University School of Medicine, Boston (T.J.R.); St. Peter's Hospital, Albany, N.Y. (E.B.); New York University Medical Center, New York (A.T.C.); New York Hospital–Weill Cornell Medical Center, New York (O.W.I.); Montefiore Medical Center, Bronx, N.Y. (J.P.G.); and Columbia–Presbyterian Medical Center, New York (E.A.R.).

References

  1. Coronary angioplasty versus coronary artery bypass surgery: the Randomized Intervention Treatment of Angina (RITA) trial. Lancet 1993;341:573-580. [CrossRef][ISI][Medline]
  2. King SB III, Kosinski AS, Guyton RA, Lembo NJ, Weintraub WS. Eight-year mortality in the Emory Angioplasty versus Surgery Trial (EAST). J Am Coll Cardiol 2000;35:1116-1121. [Free Full Text]
  3. King SB III, Lembo NJ, Weintraub WS, et al. A randomized trial comparing coronary angioplasty with coronary bypass surgery. N Engl J Med 1994;331:1044-1050. [Free Full Text]
  4. Hamm CW, Reimers J, Ischinger T, Rupprecht HJ, Berger J, Bleifeld W. A randomized study of coronary angioplasty compared with bypass surgery in patients with symptomatic multivessel coronary disease. N Engl J Med 1994;331:1037-1043. [Free Full Text]
  5. The Bypass Angioplasty Revascularization Investigation (BARI) Investigators. Comparison of coronary bypass surgery with angioplasty in patients with multivessel disease. N Engl J Med 1996;335:217-225. [Erratum, N Engl J Med 1997;336:147.] [Free Full Text]
  6. Chaitman BR, Rosen AD, Williams DO, et al. Myocardial infarction and cardiac mortality in the Bypass Angioplasty Revascularization Investigation (BARI) randomized trial. Circulation 1997;96:2162-2170. [Free Full Text]
  7. First-year results of CABRI (Coronary Angioplasty versus Bypass Revascularisation Investigation). Lancet 1995;346:1179-1184. [CrossRef][ISI][Medline]
  8. SoS Investigators. Coronary artery bypass surgery versus percutaneous coronary intervention with stent implantation in patients with multivessel coronary artery disease (the Stent or Surgery trial): a randomised controlled trial. Lancet 2002;360:965-970. [CrossRef][ISI][Medline]
  9. Carrie D, Elbaz M, Puel J, et al. Five-year outcome after coronary angioplasty versus bypass surgery in multivessel coronary artery disease: results from the French Monocentric Study. Circulation 1997;96:Suppl 9:II-1.
  10. Goy JJ, Eeckhout E, Burnand B, et al. Coronary angioplasty versus left internal mammary artery grafting for isolated proximal left anterior descending artery stenosis. Lancet 1994;343:1449-1453. [CrossRef][ISI][Medline]
  11. Goy JJ, Kaufmann U, Goy-Eggenberger D, et al. A prospective randomized trial comparing stenting to internal mammary artery grafting for proximal, isolated de novo left anterior coronary artery stenosis: the SIMA trial. Mayo Clin Proc 2000;75:1116-1123. [ISI][Medline]
  12. Hueb WA, Soares PR, Almeida De Oliveira S, et al. Five-year follow-up of the Medicine, Angioplasty, or Surgery Study (MASS): a prospective, randomized trial of medical therapy, balloon angioplasty, or bypass surgery for single proximal left anterior descending coronary artery stenosis. Circulation 1999;100:Suppl 19:II-107. 
  13. Rodriguez A, Bernardi V, Navia J, et al. Argentine randomized study: coronary angioplasty with stenting versus coronary bypass surgery in patients with multiple-vessel disease (ERACI II): 30-day and one-year follow-up results. J Am Coll Cardiol 2001;37:51-58. [Erratum, J Am Coll Cardiol 2001;37:973-4.] [Free Full Text]
  14. Rodriguez A, Boullon F, Perez-Balino N, Paviotti C, Liprandi MI, Palacios IF. Argentine randomized trial of percutaneous transluminal coronary angioplasty versus coronary artery bypass surgery in multivessel disease (ERACI): in-hospital results and 1-year follow-up. J Am Coll Cardiol 1993;22:1060-1067. [Abstract]
  15. Serruys PW, Unger F, Sousa JE, et al. Comparison of coronary-artery bypass surgery and stenting for the treatment of multivessel disease. N Engl J Med 2001;344:1117-1124. [Free Full Text]
  16. Hannan EL, Racz MJ, McCallister BD, et al. A comparison of three-year survival after coronary artery bypass graft surgery and percutaneous transluminal coronary angioplasty. J Am Coll Cardiol 1999;33:63-72. [Free Full Text]
  17. Jones RH, Kesler K, Phillips HR III, et al. Long-term survival benefits of coronary artery bypass grafting and percutaneous transluminal angioplasty in patients with coronary artery disease. J Thorac Cardiovasc Surg 1996;111:1013-1025. [Free Full Text]
  18. Mark DB, Nelson CL, Califf RM, et al. Continuing evolution of therapy for coronary artery disease: initial results from the era of coronary angioplasty. Circulation 1994;89:2015-2025. [Free Full Text]
  19. Hoffman SN, TenBrook JA, Wolf MP, Pauker SG, Salem DN, Wong JB. A meta-analysis of randomized controlled trials comparing coronary artery bypass graft with percutaneous transluminal coronary angioplasty: one- to eight-year outcomes. J Am Coll Cardiol 2003;41:1293-1304. [Free Full Text]
  20. Lauer MS, Topol EJ. Clinical trials -- multiple treatments, multiple end points, and multiple lessons. JAMA 2003;289:2575-2577. [Free Full Text]
  21. Allison P. Survival analysis using the SAS system. Cary, N.C.: SAS Institute, 1995.
  22. Ghali WA, Quan H, Brant R, et al. Comparison of two methods for calculating adjusted survival curves from proportional hazards. JAMA 2001;286:1494-1497. [Free Full Text]
  23. Rosenbaum PR, Rubin DB. Reducing bias in observational studies using subclassification on the propensity score. J Am Stat Assoc 1984;79:516-524. [CrossRef][ISI]
  24. Rosenbaum PR, Rubin DB. The central role of the propensity score in observational studies for causal effects. Biometrika 1983;70:41-55. [Free Full Text]
  25. Benson K, Hartz AJ. A comparison of observational studies and randomized, controlled trials. N Engl J Med 2000;342:1878-1886. [Free Full Text]
  26. Concato J, Shah N, Horwitz RI. Randomized, controlled trials, observational studies, and the hierarchy of research designs. N Engl J Med 2000;342:1887-1892. [Free Full Text]
  27. Regar E, Serruys PW, Bode C, et al. Angiographic findings of the multicenter Randomized Study With the Sirolimus-Eluting BX Velocity Balloon-Expandable Stent (RAVEL): sirolimus-eluting stents inhibit restenosis irrespective of the vessel size. Circulation 2002;106:1949-1956. [Free Full Text]
  28. Moses JW, Leon MB, Popma JJ, et al. Sirolimus-eluting stents versus standard stents in patients with stenosis in a native coronary artery. N Engl J Med 2003;349:1315-1323. [Free Full Text]
  29. Tanabe K, Serruys PW, Grube E, et al. TAXUS III Trial: in-stent restenosis treated with stent-based delivery of paclitaxel incorporated in a slow-release polymer formulation. Circulation 2003;107:559-564. [Free Full Text]
  30. Howard G, Chambless LE, Kronmal RA. Assessing differences in clinical trials comparing surgical vs nonsurgical therapy: using common (statistical) sense. JAMA 1997;278:1432-1436. [Erratum, JAMA 1998;279:580.] [Abstract]

 

This Article
-Abstract
- PDF
-PDA Full Text
-PowerPoint Slide Set

Commentary
-Editorial
 by Gersh, B. J.
-Letters

Tools and Services
-Add to Personal Archive
-Add to Citation Manager
-Notify a Friend
-E-mail When Cited
-E-mail When Letters Appear

More Information
-PubMed Citation

Related Letters:

Coronary-Artery Bypass Grafting versus Stent Implantation
Flaherty J. D., Davidson C. J., Crouse J. R. III, Rubin S. A., Mody F. V., Hannan E. L., Walford G., Jones R. H.
Extract | Full Text | PDF  
N Engl J Med 2005; 353:735-737, Aug 18, 2005. Correspondence

This article has been cited by other articles:



HOME  |  SUBSCRIBE  |  SEARCH  |  CURRENT ISSUE  |  PAST ISSUES  |  COLLECTIONS  |  PRIVACY  |  HELP  |  beta.nejm.org

Comments and questions? Please contact us.

The New England Journal of Medicine is owned, published, and copyrighted © 2008 Massachusetts Medical Society. All rights reserved.