Reply to the Editor

Article Outline

• References
• Copyright

CTSNet classification: 23.1

We thank Takagi and Umemoto¹ for their interest in our recent systematic review of propensity score (PS) analyses in off-pump versus on-pump coronary artery bypass grafting.² As Takagi and Umemoto correctly point out, our estimate for the effect of off-pump coronary artery bypass grafting on short-term mortality (odds ratio [OR], 0.69; 95% confidence interval [CI], 0.60–0.75) differs from the corresponding estimate from randomized controlled trials (RCTs). In their large meta-analysis, Møller and colleagues³ found a relative risk of 0.98 (95% CI, 0.66–1.44). Takagi and Umemoto¹ calculated a relative risk of 1.06 (95% CI, 0.67–1.67) by adding the results of the ROOBY trial⁴ to Møller and colleagues' results.

Setting aside the subtlety of equating ORs with relative risks, we would like to point out 2 facts. First, even after including the ROOBY results to the data of Møller and colleagues,³ the CI for the effect still includes our PS effect estimate of 0.69. That means that even data from nearly 100 randomized trials on the off-pump/on-pump issue still leave considerable uncertainty (as reflected by the large CI) on the size of the true effect. In particular, ware still far from achieving equivalent short-term mortality, which Takagi and Umemoto¹ deduce from the current data. This lack of information is also emphasized by Møller and colleagues³ in their trial sequential analysis: The authors state that demonstrating equivalence or a minimal clinically relevant effect of the off-pump technique on mortality would require more than 240,000 patients.

Second, and this is also pointed out correctly by Takagi and Umemoto,¹ we do not expect the effect estimates from PS analyses and RCTs to be equal because the underlying study populations usually differ. Patients in RCTs are, in general, younger and healthier than the average patient. In a study currently under review, we reviewed 28 PS analyses and 51 RCTs that compared off-pump and on-pump coronary artery bypass grafting.⁵ We found an average age of 65.8 years and an average left ventricular ejection fraction of 58.8% in the PS analyses, compared with an average age of 63.1 years and a mean left ventricular ejection fraction of 62.7% in the RCTs, confirming that patients in the PS analyses are older and in poorer health. After generating similar study populations from PS analyses and RCTs by a meta-matching algorithm (resulting in a meta-matched sample of 10 PS analyses and 29 RCTs), we found an OR for short-term mortality of 0.53 (95% CI, 0.43–0.66) in the PS analyses and 0.58 (95% CI, 0.24–1.39) from the RCTs, resulting in an OR difference of −0.05 (95% CI, −0.56 to 0.47). Unfortunately, this difference is also associated with a large CI, but there is some evidence (supported by other clinical outcome data not shown) that the treatments effects are similar, provided the underlying populations in PS analyses and RCTs are similar.

Finally, we completely agree with Takagi and Umemoto¹ on the importance of long-term mortality. Sufficient long-term mortality figures were available in 7 PS analyses in our data set, initially accounting for 6813 patients. By restricting the analysis to information provided only in the text (and not, eg, extracted from Kaplan–Meier curves) and considering only the effect of the longest time of follow-up from each study, we calculated an OR of 0.82 (95% CI, 0.67–1.02; P = .07). This shows that in PS analyses the superiority of the off-pump approach applies not only to short-term mortality but also to long-term mortality, albeit to a smaller degree.

Eventually, the results of current randomized studies on risk groups that mirror today's typical patient populations will elucidate the truth about the differences between the on- and off-pump approaches.

Back to Article Outline

References 

1. Takagi H, Umemoto T. To pump, or not to pump, that is the question. J Thorac Cardiovasc Surg. 2010;140:485–486
   • View In Article
   • Full Text
   • Full-Text PDF (48 KB)
   • CrossRef
2. Kuss O, von Salviati B, Börgermann J. Off-pump versus on-pump coronary artery bypass grafting: a systematic review and meta-analysis of propensity score analyses. J Thorac Cardiovasc Surg. 2010 Feb 16;[Epub ahead of print]
   • View In Article
3. Møller CH, Penninga L, Wetterslev J, Steinbruchel DA, Gluud C. Clinical outcomes in randomized trials of off- vs. on-pump coronary artery bypass surgery: systematic review with meta-analyses and trial sequential analyses. Eur Heart J. 2008;29:2601–2616
   • View In Article
   • CrossRef
4. Shroyer AL, Grover FL, Hattler B, Collins JF, McDonald GO, Kozora E, et al. On-pump versus off-pump coronary-artery bypass surgery. N Engl J Med. 2009;361:1827–1837
   • View In Article
   • CrossRef
5. Kuss O, Legler T, Börgermann J. Do treatments effects differ between randomized trials and propensity score analyses in similar populations? Evidence from a meta-propensity score analysis in off-pump versus on-pump coronary artery bypass surgery. 2010 (submitted).
   • View In Article