Review of the European Coronary Surgery Study
Long-Term Results of Prospective Randomised Study of Coronary Artery Bypass Surgery in Stable Angina Pectoris
Background: The first coronary artery bypass graft surgery (CABG) was performed in 1964 and by the 1970’s it was commonly used for relief of angina. However, whether it improved survival was unknown. The European Coronary Surgery Study (ECSS) sought to test the hypothesis that CABG compared to medical therapy improved survival at 5 years.
Note to readers: Several preliminary reports of ECSS results were published at earlier time points (2 and 3-5 years). We are reporting the 5-year results since this was the prespecified hypothesis the investigators sought to test.
Patients: Men under 65 years of age with angina pectoris of more than 3 months duration, a left ventricular ejection fraction >50%, and angiographic obstruction of >/=50% in at least 2 major coronary vessels with at least 1 vessel suitable for grafting. Patients with severe angina that could not be controlled with medical therapy were excluded.
Baseline characteristics: No information is provided in the main paper on the number or characteristics of individuals screened to enrolled. There were 768 patients enrolled in the study. They were recruited from September, 1973 to March, 1976. The average age of patients was 50 years and the left ventricular ejection fraction was 65%. Approximately, 46% had a previous heart attack, 43% smoked, 35% had a high cholesterol, 15% had hypertension and 6% had diabetes. In terms of coronary anatomy, 53% had 3-vessel disease, 40% had 2-vessel disease, and 7% had left main disease.
Procedures: Patients were randomly assigned to receive medical or surgical treatment. Medical measures varied based on location. The authors reported that strict standardization was not felt to be practical or necessary. Surgical treatment was either with saphenous-vein graft or internal mammary artery and was performed as soon as possible following randomization. The average time from randomization to surgery was approximately 4 months.
Follow-up evaluations were performed 6 months after randomization and annually thereafter. Graft angiography was planned at 6 and 12 months after operation.
Endpoints: The primary endpoint was all-cause death. The prespecified minimum follow-up time, set at the start of the trial, for all patients was 5 years. At the time of this report, some patients had been followed up to 8 years. A strict hypothesis was not tested (i.e., CABG would reduce death by X% compared to medical therapy). The primary analysis was a traditional intention to treat analysis and medical patients who crossed over to surgery and surgical patients who died prior to receiving surgery or refused surgery after randomization were retained within their original groups.
Results: There were 767 patients included in the final analysis; 373 patients in the medical group and 394 in the surgical group (1 patient was lost from the surgical group immediately following randomization and was not counted in the group). At 5 years, 90 patients (24%) of the medical group had crossed over to surgery and 26 (7%) of the surgical patients were not operated on. An average of 1.9 grafts per patient were performed in the 2-vessel disease subgroup and 2.4 grafts per patient in the 3-vessel disease subgroup. The graft patency rate was 90% within 9 months and 77% between 9 and 18 months.
Compared to medical therapy, surgery significantly reduced death at 5 years by 53% (7.6% vs 16.4%; p=0.00025). Operative (in-hospital) mortality was 3.6% for a total of 494 operations and 7.7% for 26 reoperations.
Seven of 27 prespecified variables recorded at the time of randomization were found to be associated with significant treatment effect heterogeneity. They included: (i) extent of disease; (ii) location of lesion(s) in the proximal third of the left anterior descending artery (proximal LAD); (iii) resting ECG suggestive of previous possible or probable myocardial infarction and/or with other specified abnormalities (iv) ischemic ST-segment response predominantly in lead V5 during maximum level of a multistage symptom/sign-limited bicycle exercise test; (v) history of peripheral arterial disease; (vi) age; and (vii) mode of treatment.
Subgroup analysis on the basis of coronary anatomy supported a significant advantage of surgery for patients with left main disease (14.3% vs 32.1%; p=0.11) and 3-vessel disease (6% vs 17.6%; p=0.003) but not in 2-vessel disease (8.8% vs 11.8%; p>0.20).
The left main subgroup could have had 2- or 3-vessel disease and the p-value was insignificant due to the small sample size.
Surgery significantly reduced death in patients with proximal LAD disease (7.3% vs 18%; p=0.0004) but not in those without it (6.7% vs 7.9%; p>0.20). In the subset of patients with 2-vessel disease and without proximal LAD disease, surgery caused a numerical increase in death at 5 years, attributed to operative mortality.
Surgery significantly reduced death in patients with >/= 1.5mm exertional ST depression on bicycle testing (8.3% vs 21%; p=0.003) but not in those without it (5.1% vs 9.7%; p>0.20).
Angina and exercise performance were significantly improved in the surgery group compared to medicine.
Conclusions: Compared to medical therapy, bypass surgery using internal mammary arteries and saphenous vein grafts significantly reduced mortality at 5 years in men under 65 years of age with normal left ventricular function. Approximately 11 men would need to be treated with CABG to prevent 1 death. This represents a large benefit for bypass surgery in well-selected patients at the time the study was undertaken.
Contemporary caveats to this interpretation include improvements in medical therapy since the publication of ECSS mainly involving aspirin and cholesterol lowering drugs for patients with CAD as well as an improvement in the general management of conditions like hypertension and diabetes. Also, smoking rates have significantly declined at the population level.
Despite the impressive benefit of bypass surgery seen in this study, important treatment effect heterogeneity was identified for certain lower risk patient groups including those with 2-vessel disease, absence of proximal LAD disease, and minimal ST depression on symptom limited bicycle testing. Theoretically, such patients would be expected to benefit from bypass surgery even less today given the improvements in medical therapy mentioned above.
Finally, it is worth pointing out the difference in treatment effects seen in this study compared to the Veterans Administration Cooperative Study that we reviewed earlier this week, which was a negative trial. In the Veterans Administration Cooperative Study, coronary bypass was performed primarily with saphenous vein grafting whereas ECSS used internal mammary arteries and saphenous vein grafts. Internal mammary arteries are superior conduits compared to vein grafts. They have improved long-term patency rates, which is attributed to their striking resistance to the development of atherosclerosis. Furthermore, they are used almost exclusively on the LAD, which is the most important vessel.
In conclusion, ECSS demonstrated that CABG surgery dramatically reduced death at 5 years compared to medical treatment; however, we should be aware of the caveats mentioned above and appreciate that the trial was limited to highly selected male patients under the age of 65.