Current issues of ACP Journal Club are published in Annals of Internal Medicine


Review: CABG leads to less angina and less reintervention than PTCA, but mortality and subsequent MI are the same

ACP J Club. 1996 Mar-April;124:43. doi:10.7326/ACPJC-1996-124-2-043

Related Content in this Issue
• Companion Abstract and Commentary: CABG led to less angina, reintervention, and medication than did PTCA, but mortality did not differ

Related Content in the Archives
• Correction: CABG led to less angina, reintervention, and medication than did PTCA, but mortality did not differ

Source Citation

Pocock SJ, Henderson RA, Rickards AF, et al. Meta-analysis of randomised trials comparing coronary angioplasty with bypass surgery. Lancet. 1995 Nov 4;346:1184-9. [PubMed ID: 7475657]



To compare, using meta-analysis, the effectiveness of revascularization done using coronary artery bypass grafting (CABG) with that of revascularization done using percutaneous transluminal coronary angioplasty (PTCA) in patients with angina.

Data Sources

Studies were identified through shared communication at large annual cardiology meetings and by contacting experts.

Study Selection

Studies were selected if they were randomized controlled trials comparing initial CABG and PTCA in patients with coronary artery disease who were suitable for either treatment method.

Data Extraction

To seek consistent reporting of information, a standard form was sent to every principal investigator. The form asked for the following information for each treatment group: number of patients randomized; median length of follow-up; number of patients who received the randomized procedure; distribution of angina grade at 1 and 3 years; number of deaths classified as cardiac, other cardiovascular, and noncardiovascular; number of nonfatal myocardial infarctions (MIs); and additional CABG and PTCA procedures.

Main Results

8 trials that included 3371 patients met the selection criteria. 1 additional trial, the Bypass Angioplasty Revascularization Investigation (BARI), was unable to supply data. For all available follow-up, 73 deaths occurred in the CABG group and 79 occurred in the PTCA group {P = 0.75}*. 127 patients in the CABG group compared with 135 patients in the PTCA group had MI or cardiac death in the first year of follow-up (P = 0.49). All trials had a lower prevalence rate of angina in the CABG group at 1 year (11% vs 18%, P < 0.001). {This absolute risk reduction of 7% means that 16 patients would need to be treated with CABG to prevent 1 case of angina at 1 year, 95% CI 12 to 26; the relative risk reduction was 43%, CI 29% to 53%.}* By 3-year follow-up, this difference had diminished (13% vs 16%) {P = 0.06}*. When all trials were combined, 55 patients (3%) allocated to CABG compared with 577 patients (34%) allocated to PTCA required at least 1 additional revascularization procedure {P < 0.001}*. Separate analyses for single-vessel and multivessel disease showed that the rates of mortality, additional intervention, and prevalent angina were slightly lower in single-vessel disease.


The risk for death and myocardial infarction is not different between patients with coronary artery disease assigned to coronary artery bypass surgery and those assigned to percutaneous transluminal coronary angioplasty. Patients allocated to percutaneous transluminal coronary angioplasty have higher rates of subsequent revascularization and angina symptoms.

Source of funding: Not stated.

For article reprint: Prof. S.J. Pocock, London School of Hygiene and Tropical Medicine, London WC1E 7HT, England, UK. FAX 44-207-637-2853.

*Numbers calculated from data in article.


CABG and PTCA are effective, frequently done revascularization techniques that use a substantial proportion of health care resources. Although both techniques increase coronary blood flow, they have fundamental differences beyond initial cost and inconvenience and discomfort to patients. PTCA reconstructs the coronary lumen and improves flow while maintaining future revascularization options. Progression of atherosclerosis favors proximal vessels, whereas CABG places conduits to distal locations that are mostly immune from advancing disease. Venous conduits, however, do deteriorate substantially with time. In a specific patient, technical or clinical factors may preclude 1 or both procedures. Where both are alternatives, solid trial data based on natural history observations have not been available to guide the selection process.

The report from CABRI, amplified by the meta-analysis by Pocock and colleagues, is an important contribution to the process of defining the relative roles of CABG and PTCA in patients with stable coronary artery disease. When the primary clinical end points of death and MI were analyzed, the 2 procedures appeared to have equal benefit. Randomized patients were highly selected, however, because only 4.6% of all patients who needed revascularization were actually randomized. This implies that the process of assessing a patient's suitability for each procedure is an important first step in clinical practice and that the investigators' conclusion of equivalence applies only when both options exist.

Despite this equivalence for survival and MI, there is the expected but nonetheless disquieting news that the need for further revascularization procedures and the level of angina were significantly greater in patients who received coronary artery reconstruction with PTCA than in patients who had CABG. Additional procedures are needed because of the elastic recoil and proliferative response that can be identified angiographically in 40% of patients treated with PTCA. Although most patients treated with PTCA remain asymptomatic and do not require further procedures, restenosis has substantial clinical and economic importance. Patients should be made aware of this fact when options are being discussed. At present, the initial cost advantage of PTCA is considerably reduced by the need for repeated procedures.

Two important qualifications should be made on the repeat procedure issue. First, several randomized trials required follow-up angiography whether the patients were symptomatic or not. Restenosis is common angiographically but is of clinical concern in only 20% of patients. Although there is no justification for redilating asymptomatic lesions, repeated PTCA and a crossover to CABG in response to angiographic restenosis without symptoms occurred; this finding magnifies the apparent clinical importance of restenosis.

Second, the PTCA technique has undergone major improvements since trial completion, in particular, the introduction of intracoronary stents. These devices buttress the artery after balloon dilation, largely eliminate elastic recoil, and reduce restenosis. In both the Benestent Study Group (1) and the Stent Restenosis Study (STRESS) (2), trials that compared stenting with ordinary balloon PTCA, stenting reduced the incidence of clinical restenosis by 50%. Further reductions seem certain with improved delivery and anticoagulation strategies. Because stenting was not used in CABRI and other trials referred to in the subject meta-analysis, there is reason to believe that the large repeated-procedure differential noted is greater than that reported in the current “stent era.”

Although the reviewed trials are helpful in the procedure selection process, careful consultation with surgeons and interventional cardiologists as to the suitability of the procedures will continue to be an essential part of decision making when the medical management of patients with coronary artery disease has failed.

Merril L. Knudtson, MD
Foothills HospitalCalgary, Alberta, Canada


1. Serruys PW, de Jaegere P, Kiemeneij F, et al. A comparison of balloon-expandable-stent implantation with balloon angioplasty in patients with coronary artery disease. Benestent Study Group. Benestent Study Group. N Engl J Med. 1994;331:489-95.

2. Fischman DL, Leon MB, Bairn DS, et al. A randomized comparison of coronary-stent placement and balloon angioplasty in the treatment of coronary artery disease. Stent Restenosis Study Investigators. N Engl J Med. 1994;31:486-501.