| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Ann Thorac Surg 2005;79:383
© 2005 The Society of Thoracic Surgeons
Health Technology Assessment Group, ECRI, 5200 Butler Pike, Plymouth Meeting, PA 19462
To the Editor:
We thank Dr Bainbridge and coauthors for their comments regarding our article [1], and we will address the points they raised.
In our study, we note that we did examine baseline patient characteristics. We observed a tendency toward fewer grafts per patient in the off-pump groups, even among some of the randomized controlled trials (RCTs), although it appeared more often in nonrandomized studies. However, even in studies where the number of grafts per patient was significantly lower in the group undergoing off-pump coronary artery bypass grafting (OPCABG), that group often had a higher average surgical risk score than the CABG group. We compared the results of studies potentially biased toward better results in the OPCABG group and studies potentially biased toward better results in the CABG group. For two of 11 outcomes (myocardial infarction and reoperation for bleeding), initial heterogeneity in the meta-analyses was eliminated by removing studies with bias favoring better outcomes among patients undergoing OPCABG.
We recognize that the low power of the comparisons means that equivalence of RCTs and nonrandomized studies cannot be proven in this instance. However, if bias is present in the nonrandomized studies, it most often appears to favor better results for conventional CABG. For the majority of outcomes, the nonrandomized studies showed a result that was less favorable for OPCABG than did the RCTs. This was true even for the one outcome (atrial fibrillation) where the difference between RCTs and nonrandomized studies was significant. Sensitivity analyses supported the conclusion that OPCABG has some short-term benefits.
Our mention of ten RCTs referred to the total number of publications. We were aware that references 9, 14, and 17 were parts of the same study (BHACAS-1), which is why each reference was used in different meta-analyses. Reference 14 was the first report on the short-term results of the complete patient group enrolled in BHACAS-1 and was therefore our primary reference for the trial. References 9 and 17 contained specific outcome data not reported in reference 14, and they were used in separate meta-analyses. Although reference 1 in our article reiterates the BHACAS-1 short-term results, this reference also reports separate data from a different study (BHACAS-2) in which enrollment criteria were not the same as in BHACAS-1. We treated these as separate trials in our meta-analyses, thus bringing the total number of RCTs to eight.
In reference to the statement that it is difficult to conclude that "OPCABG is superior to conventional CABG for short-term and midterm results," we did not make such a claim for the midterm results and acknowledge that the relative effectiveness of OPCABG and CABG cannot be determined with certainty until more midterm to long-term data are available. However, our work in progress incorporating data from RCTs alone still shows a significant short-term benefit for OPCABG. Our intention was not to diminish the importance of RCTs, but simply to suggest that evidence from nonrandomized studies can be used in certain instances when few RCTs are available for analysis.
References
Related Article
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| ANN THORAC SURG | ASIAN CARDIOVASC THORAC ANN | EUR J CARDIOTHORAC SURG |
| J THORAC CARDIOVASC SURG | ICVTS | ALL CTSNet JOURNALS |
