PARIS – Coronary artery bypass grafting surpassed percutaneous coronary intervention in a "real-world," registry of patients with three-vessel coronary disease, largely confirming the findings of the SYNTAX trial in patients with high SYNTAX scores and in those with low scores.
However, because the new data – obtained at 26 Japanese centers during 2005-2007 and involving nearly 3,000 patients – came from a nonrandomized registry, it may have been flawed by selection biases that skewed which patients underwent bypass surgery and which ones had percutaneous revascularization.
Coronary artery bypass grafting (CABG) "remains the standard treatment option for patients with triple-vessel disease, even when their SYNTAX scores are high," Dr. Hiroki Shiomi said at the annual congress of the European society of Cardiology. "Use of PCI [percutaneous coronary intervention] in patients with high SYNTAX scores should be seriously discouraged unless their operative risk is prohibitively high," said Dr. Shiomi, a cardiologist at Kyoto University Hospital, Japan. The registry results "suggested that the clinical outcome of PCI is not comparable with CABG even in patients with low syntax scores."
But the study’s findings also suggested that a selection bias occurred that even the risk-adjusted analysis used by Dr. Shiomi and his associates failed to adequately control, commented Dr. Uwe Zeymer, an interventional cardiologist and professor at the Institute for Myocardial Infarction Research in Ludwigshafen, Germany.
A major clue that selection bias came into play was that all-cause death during the 3-year follow-up of the study was 62% higher among the patients who underwent PCI, compared with those who underwent CABG, a statistically significant difference; in contrast, the rate of cardiac death was not significantly different between the CABG and PCI patients.
This discrepancy "says there was selection bias. Physicians had to decide what to do with patients who had a lot of comorbidities," and most of those patients probably underwent PCI, Dr. Zeymer said in an interview. "If the patients [undergoing PCI and CABG] were the same clinically, you’d expect that with PCI you would at least have the same result" for all-cause mortality. "There seems to have been a selection bias toward using PCI in patients with more comorbidities.
"What we can say is that in this real-world situation, cardiac mortality was the same" with both revascularization methods, "which is reassuring for the use of PCI," he added.
CREDO-Kyoto (Coronary Revascularization Demonstrating Outcome Study in Kyoto) enrolled more than 13,000 patients who underwent coronary revascularization at any of 26 Japanese centers during 2005-2007, a period when all PCIs used a drug-eluting coronary stent. The current analysis focused on 2,981 patients with triple vessel disease and no left main coronary disease or acute MI, and included 1,825 who were treated with CABG and 1,156 treated by PCI. The average age of the PCI patients was 70 years; the CABG patients averaged 68 years old.
After 1 year of follow-up, the combined rate of death, MI, or stroke was 23% higher in the patients treated with PCI, compared with those treated with CABG, a statistically significant difference in an unadjusted analysis for the study’s primary end point. After adjustment for baseline differences between the CABG and PCI patients, the rate of this combined end point was 47% higher among the PCI patients, also a statistically significant difference.
Analysis of several secondary end points showed a mixed pattern of differences between the two treatment groups in the adjusted analyses. A statistically significant difference in favor of fewer events with CABG occurred in the end point of all-cause death, which was 62% higher with PCI, as well as in the end point of MI, which was 2.39-fold higher in the PCI group. The protection against MI by CABG, compared with PCI, was "particularly remarkable," Dr. Shiomi said. But the end points of cardiac death and stroke showed no statistically significant differences between the two treatment groups in either the unadjusted or adjusted analyses.
Dr. Shiomi said that baseline SYNTAX scores were available for 94% of the patients. The mean score was 24 in the PCI patients and 30 for the CABG patients.
Dr. Shiomi said he had no disclosures. Dr. Zeymer said that the Institute of MI Research in Germany, where he works, has received research grant support from multiple cardiac-device companies.