Author + information
- Keith A.A. Fox, BSc, MBChB, FMedSci⁎ (, )
- Tim C. Clayton, BSc, MSc,
- Peter Damman, MD,
- Stuart J. Pocock, BSc, MSc, PhD,
- Robbert J. de Winter, MD, PhD,
- Jan G.P. Tijssen, PhD,
- Bo Lagerqvist, MD, PhD,
- Lars Wallentin, MD, PhD,
- FIR Collaboration
- ↵⁎Centre for Cardiovascular Science, University of Edinburgh, Chancellor's Building, 49 Little France Crescent, Edinburgh EH16 4SB, United Kingdom
We thank Dr. Sanchis and colleagues for their interest in our paper and for their comments.
The first point raised was that there was heterogeneity in the definition of the selective invasive strategy, in the 3 studies comprising the meta-analysis. In our paper, we discuss factors that may account for the trial-to-trial differences (page 2,441), and these not only include the differences in the rate of revascularization in the “selective invasive arm” but also differences in inclusion criteria. As we point out, the FRISC (Fast Revascularization During Instability in Coronary Artery Disease) II trial and the RITA (Randomized Intervention Trial of Unstable Angina) 3 trial had a wide separation in the frequency of revascularization rates between the 2 arms of the respective trials compared with a modest difference in the ICTUS (Invasive Versus Conservative Treatment in Unstable Coronary Syndromes) trial.
However, despite these differences, there is clear evidence of superior outcome at 5 years with the routine invasive strategy. The heterogeneity would tend to minimize the statistically significant differences. In short, we agree that the substantial difference in revascularization rates between strategies in the FRISC II study and the RITA 3 study are likely to have contributed to the observed superior outcomes of those respective studies. We would also like to point out, as shown in Figure 2 of our paper, that the confidence intervals for the respective trials overlap, and that is true for cardiovascular death or myocardial infarction, cardiovascular death alone, and for myocardial infarction alone. Hence, the overall result is consistent with the findings from the individual trials in view of the overlap of the confidence intervals.
We agree that more work needs to be done in the selection of patients with most benefit from revascularization, and that is the rationale for the presentation of the simple “integer score.” Although patients in the highest risk group were demonstrated to have most absolute gain, this group was the smallest numerically. Nevertheless, there was evidence of benefit in the remaining groups and even in the lowest risk group the absolute benefit was about 2%, and this is greater than that seen in many pharmacological studies.
The second point made by Dr. Sanchis and colleagues is that “pooling the 3 studies implies the assumption that all the strategies in the conservative groups were equivalent… .” In contrast, one assumes that the trials have sufficient common ground to be meta-analyzed. It is inevitable in any meta-analysis that there will be differences in specific rates of treatment. As pointed out, this heterogeneity will contribute to “noise” and tend to minimize the chance of revealing a statistically robust effect.
We disagree with the conclusion of Dr. Sanchis and colleagues that the “superiority of the invasive strategy versus the true selective invasive strategy is still unresolved.” As demonstrated in this long-term outcome study, and despite the differences in strategy from trial to trial, there is nevertheless a net treatment effect observable at 5 years, and it is of a magnitude greater than seen in many trials of pharmacological therapy (1).
- American College of Cardiology Foundation