Re: Preoperative intra-aortic counterpulsation in high-risk patients undergoing cardiac surgery: a meta-analysis of randomized controlled trials

In the study by Pilarczyk et al. [1] on the use of intra-aortic counterpulsation in high-risk patients undergoing cardiac surgery, the authors performed a meta-analysis looking at nine randomized, controlled trials (RCTs). Among these nine RCTs, three of them were written by Christenson et al. [2–4] in 1997. Upon closer examination of these three articles, Christenson et al. [3] looked at a series of 52 patients in a single centre in Switzerland between June 1994 and March 1996. These were all high-risk patients with coronary artery disease (CAD) undergoing myocardial revascularization [coronary artery bypass grafting (CABG)]. In the second article, Christenson et al. [4] looked at 33 patients with CAD undergoing CABG in a single centre between June 1994 and March 1996. These patients had a preoperative left ventricular ejection fraction (LVEF) of ≤40% and established arterial hypertension with left-ventricular hypertrophy. In the third article, Christenson et al. [2] looked at 48 patients in a single centre between June 1994 and October 1996. These patients underwent redo CABG with an LVEF of ≤40%, unstable angina at the time of surgery, left main stem stenosis ≥70% or a combination thereof. It appears that there is a significant overlap of patients between these three articles. The second article [4] is a subgroup with an LVEF of ≤40% from the first article [3]. The third article [2] is also a subgroup with only redo CABG from the first article with the addition of possibly some patients from the period of March 1996 to October 1996. Therefore, I don't think that Pilarczyk et al. [1] should consider the three articles as separate populations in the meta-analysis. The population described by Christenson et al. [3] would be over-represented on the forest plot and skew the data. In Figure 2, Comparison 1 looked at the incidence of in-hospital mortality, the three studies by Christenson et al. skewed the data in favour of the intra-aortic balloon pump (IABP) group. In Figure 3, Comparison 1 looked at the incidence of low cardiac output state and the overall result favoured IABP. However, three out of the four studies that had a statistically significant odds ratio were based on Christenson et al.'s studies.