Letter by Boden Regarding Article, “Effects of Percutaneous Coronary Intervention on Death and Myocardial Infarction Stratified by Stable and Unstable Coronary Artery Disease: A Meta-Analysis of Randomized Controlled Trials”

To the Editor:

I am writing to clarify some irregularities in the recent publication by Chacko et al¹ of their meta-analysis of the effects of revascularization on outcomes derived from prior published trials that have been stratified according to acute coronary syndrome (ACS) versus stable coronary artery disease. In this article, the authors purport to show that there is a statistically significant reduction in all-cause mortality of 16%, in cardiac mortality of 31%, and in myocardial infarction (MI) of 26% with percutaneous coronary intervention (PCI) in patients with ACS and non–ST-segment MI (NSTEMI) as compared to stable coronary artery disease patients where they accurately reported no death/MI benefit. I question the validity of the data in their analysis of outcomes attributed to revascularization (principally with PCI) in these patients with ACS/NSTEMI.

In critically reading the Chacko article, there are multiple critical flaws—the most important and egregious of which is their omission of our VANQWISH trial (Veterans Affairs Non–Wave Infarction Strategies In-Hospital) publication² in New England Journal of Medicine from 1998. This is remarkably similar to a similar flawed meta-analysis from 2006 published by Bavry et al³ in JACC, in which the VANQWISH trial was selectively omitted. VANQWISH clearly showed that there was an increase in all-cause mortality with the invasive strategy at hospital discharge, at 1 month, and at 1 year, as follows: the number of patients with one of the components of the primary end point (death or nonfatal myocardial infarction) and the number who died were significantly higher in the invasive-strategy group at hospital discharge (36 versus 15 patients, P=0.004, for the primary end point; 21 versus 6, P=0.007, for death), at 1 month (48 versus 26, P=0.012; 23 versus 9, P=0.021), and at 1 year (111 versus 85, P=0.05; 58 versus 36, P=0.025). Overall mortality during follow-up did not differ significantly between patients assigned to the conservative-strategy group and those assigned to the invasive-strategy group (hazard ratio, 0.72; 95% CI, 0.51–1.01).

Obviously, it is unclear why VANQWISH could have been excluded as it was a prominent trial published in New England Journal of Medicine and would have appeared in any PubMed or MEDLINE search. Was this merely an oversight? Or was this a purposeful exclusion to make a point that PCI is superior to a conservative strategy in ACS patients? To be clear, I am not making accusations of intellectual dishonesty here, but this failure to include VANQWISH clearly requires an explanation.

In addition, as VANQWISH was a trial contemporaneous with The Framingham and Fast Revascularization During Instability in Coronary Artery Disease (FRISC)-2 trial⁴, it is not clear why this was selectively excluded from the current meta-analysis while FRISC-2 was prominently included in their meta-analysis. Based on the data published in the VANQWISH trial in the 1998 New England Journal of Medicine paper, I believe a re-analysis of the Chacko meta-analysis that includes the VANQWISH trial would result in a significant change in the reported RR (95% CI) for all-cause mortality of 0.84 (0.72–0.97) for NSTEMI and risk ratio of 0.84 (0.75–0.93) for all unstable coronary artery disease trials and that the outcome of all-cause mortality would no longer be significant for PCI + OMT versus OMT alone as shown in Figure 2 of their paper. In support of this, please find a reconfigured analysis of the Chacko meta-analysis that now includes VANQWISH, and which now does not show a significant risk ratio/95% CI (Figure). Moreover, in Figure 3, Chacko et al show that the odds ratio for cardiac death is 0.69 (0.53–0.90) for unstable angina coronary artery disease trials yet, unless I am reading this figure incorrectly, only 2 trials are shown in this cardiac death analysis (Second Randomized Intervention Treatment of Angina-2⁸ and the Savonitto study⁷ of only 213 patients), whereas for NSTEMI the risk ratio is 0.80 (0.59–1.08). VANQWISH did not include an analysis of cardiac mortality, so it cannot further inform this particular analysis. But, in addition to the exclusion of VANQWISH, it is also troubling that the authors chose to include several small trials, such as VINO (Value of First Day Angiography/angioplasty in Evolving Non-ST-elevation Myocardial Infarction trial)^9,10, where the validity of including such studies with small sample sizes may raise issues about generalizability.