Timing of initiation of intra‐aortic balloon pump in patients with acute myocardial infarction complicated by cardiogenic shock: A meta‐analysis

Abstract Background For patients with acute myocardial infarction (AMI) complicated by cardiogenic shock (CS) undergoing primary percutaneous coronary intervention (PCI), the optimal timing of the initiation of intra‐aortic balloon pump (IABP) therapy remains unclear. Therefore, we performed the first meta‐analysis to compare the outcomes of IABP insertion before vs after primary PCI in this population. Methods Electronic databases of PubMed, EMBASE, and Cochrane Library were comprehensively searched from inception to April 1, 2019, to identify the eligible studies. The main outcomes were short‐term (in‐hospital or 30 days) and long‐term (≥ 6 months) mortality. In addition, pooled analysis of risk‐adjusted data were also performed to control for confounding factors. Results Seven observational studies and two sub‐analysis of randomized controlled trials involving 1348 patients were included. Compared to patients inserted IABP after PCI, patients who received IABP therapy before primary PCI had similar risks of short‐term (odds ratio [OR] 0.88, 95% CI 0.49 to 1.59) and long‐term (OR 0.99, 95% CI 0.58 to 1.68) all‐cause mortality. Moreover, a pooled analysis of risk‐adjusted data also found similar effects of the two therapies on short‐term (OR 0.65, 95% CI 0.34 to 1.25) and long‐term (OR 0.68, 95% CI 0.17 to 2.72) mortality. Besides, no significant difference was found between the two groups with respect to reinfarction, repeat revascularization, stroke, renal failure, and major bleeding. Conclusions The timing of the initiation of IABP therapy does not appear to impact short‐term and long‐term survival in patients with AMI complicated by CS undergoing primary PCI.


| INTRODUCTION
In patients with acute myocardial infarction (AMI), 6%-9% can be affected by cardiogenic shock (CS) and the mortality rate is close to 50% during hospitalization. 1 Despite adoption of early revascularization strategies, CS remains the leading cause of death in this population. 2 Furthermore, supportive drug treatments with inotropes and vasopressors bring no benefit to patients. Cardiologists hope that mechanical circulatory support will improve clinical outcomes in this population. 3,4 The intra-aortic balloon pump (IABP) becomes the first and most widely used device due to its ability to reduce afterload and improve coronary blood flow. [4][5][6][7] However, recent meta-analyses and the landmark randomized controlled trial did not show a beneficial effect of IABP in patients with AMI complicated by CS. [8][9][10][11] Thus, the recommendations of IABP therapy have been reduced both in the American and European guidelines. 12,13 Nonetheless, the lack of efficacy of IABP usage might be partly influenced by the timing of initiation of IABP therapy, that is, before or after primary percutaneous coronary intervention (PCI). Nevertheless, almost all studies comparing the sequence of IABP and primary PCI are of a small scale, and current trials have shown conflicting results. Thus, we conducted this meta-analysis to identify the optimal timing of the initiation of IABP in patients with AMI complicated by CS undergoing primary PCI.

| METHODS
This study was performed based on the preferred reporting items for systematic reviews and meta-analyses (PRISMA) and meta-analysis of observational studies in epidemiology (MOOSE) statements. 14,15

| Search strategy
Two independent investigators (Lingxiao Chen and Kuo Zhou) searched the electronic databases of PubMed, EMBASE, and Cochrane Library from inception to April 1, 2019, to identify the pertinent English articles regarding the IABP inserted before vs after primary PCI for the treatment of AMI complicated by CS. The following medical subject headings and search terms were used: "acute myocardial infarction," "cardiogenic shock," "before primary percutaneous coronary intervention," "after primary percutaneous coronary intervention," and "timing." In addition, the references of the identified articles and relevant reviews were examined to include other potentially eligible studies.

| Study selection
Studies satisfying the following criteria were eligible: (a) patients who were diagnosed with CS from AMI; (b) studies that compared the strategy of IABP insertion before vs after primary PCI; and (c) studies that assessed the endpoints of interest. The selection was conducted by scanning titles and/or abstracts, and full-text reviews were performed for further analysis. When several reports overlapped, we selected the largest and the latest one. The studies were reviewed by two independent investigators (Jinfan Tian and Yunfeng Yan) to determine whether they met the inclusion criteria. Any disagreements were resolved through discussion with a third investigator (Dongfeng Zhang).

| Data extraction and quality assessment
For each eligible study, three authors (Fei Yuan, Mingduo Zhang, and Wei Wang) independently extracted the following data through a standardized form: first author, year of publication, study design, quality indicators, baseline as well as procedural characteristics, and clinical outcomes. Discrepancies were resolved by consensus. The primary endpoint was short-term mortality (in-hospital or 30 days). Long-term mortality (≥ 6 months), reinfarction, stroke, repeat revascularization, acute renal failure, and major bleeding were the secondary outcomes.
Deaths were classified as either cardiac or noncardiac, and classifications of other outcomes were in agreement with the included studies.
The methodological quality of the observational studies was assessed using the Newcastle Ottawa Scale. 16 Studies with a Newcastle-Ottawa score ≥ 6 (maximum, 9) were considered high quality. In addition, the quality of randomized controlled trials (RCTs) were assessed using the Cochrane risk of bias tool. 17

| Statistical analysis
The present study used Review Manager 5.3 (The Cochrane Collaboration, The Nordic Cochrane Centre, Copenhagen, Denmark) and Stata/SE12.0 (StataCorp, College Station, Texas) for data analysis. All results were presented as odds ratios (ORs) and 95% confidence intervals (CIs). Potential heterogeneity was evaluated with the I 2 statistic, and a value >50% was defined as statistical heterogeneity. For all comparisons, the DerSimonian and Lair random-effects model was used to account for the wide range of methodological variability across the studies.
Pooled analysis of risk-adjusted data were performed to control for confounding factors, and to test the sensitivity of the short-term and long-term mortality. The adjusted variables are listed in Table S1.
In addition, sensitivity analysis was conducted by reanalyzing the results of studies that enrolled patients presented with ST-segment elevation myocardial infarction (STEMI) or published in full text. In case of significant heterogeneity, sensitivity analysis was also conducted by omitting one study in each turn to test the influence of single trial. Meta-regression analysis was carried out to assess patient characteristics with the primary endpoint, that is, male, current smoker, diabetes mellitus, hypertension, and culprit vessel of left anterior descending coronary artery. The risk of potential publication bias was assessed by the Begg's and Egger's tests. 18,19 When there was an indication of publication bias from the statistical tests, we used the trim and fill method to evaluate the influence of potentially unpublished studies on the summary estimates. All statistical tests were two-sided and were considered to be statistically significant at P < .05.  Figure S1).
As presented in Table 2, baseline characteristics of the patients were similar between the two treatment strategies, except that dyslipidemia was more common in patients who received IABP insertion before primary PCI than the control group (48.3% vs 38.7%).

| Primary endpoint
In summary, short-term death occurred in 149 patients (36.3%) in the IABP inserted before primary PCI group compared with 264 patients

| Secondary endpoints
In the pooled estimate, the initiation of IABP therapy before primary PCI had similar risk of long-term mortality compared to that of inserted after primary PCI based on both unadjusted data (OR 0.99, 95% CI 0.58 to 1.68, P = .96, I 2 = 57%) ( Figure 2A) and risk-adjusted data (OR 0.68, 95% CI 0.17 to 2.72, P = .59, I 2 = 94%) ( Figure 2B). After removing the study

| Meta-regression analysis and publication bias
Meta-regression analysis showed significant association between patient characteristics of diabetes mellitus (regression coefficient In addition, the assessment of the funnel plot was performed, and no publication bias was found for the outcomes except for major F I G U R E 1 Forest plots comparing short-term mortality for patients with acute myocardial infarction complicated by cardiogenic shock undergoing IABP insertion before or after primary percutaneous coronary intervention. A, Unadjusted short-term mortality. B, Adjusted shortterm mortality. CI, confidence interval; IABP, intra-aortic balloon pump; PCI, percutaneous coronary intervention F I G U R E 2 Forest plots comparing long-term mortality for patients with acute myocardial infarction complicated by cardiogenic shock undergoing IABP insertion before or after primary percutaneous coronary intervention. A, Unadjusted short-term mortality. B, Adjusted shortterm mortality. CI, confidence interval; IABP, intra-aortic balloon pump; PCI, percutaneous coronary intervention bleeding (Egger's test, P = .03; Begg's test, P = .09). One study was added with the trim and fill method, and the risk of major bleeding remained similar between the two treatment strategies (OR 1.00, 95% CI 0.53 to 1.88) ( Figure S4).

| DISCUSSION
This is the first meta-analysis comparing the two treatment strategies of IABP inserted before and after primary PCI in patients with AMI complicated by CS. Our data suggest that the timing of initiation of IABP therapy does not have an effect on short-term and long-term survival in this population. Besides, the risks of reinfarction, repeat revascularization, stroke, acute renal failure, and major bleeding were similar between the two groups.
Since 1968, the IABP has been used for mechanical cardiac assistance in patients with CS. 29 In theory, the deflation during systole reduces ventricular afterload and helps the ventricle push blood into the aorta, while the inflation during diastole enhances coronary artery perfusion and promotes blood flow to systemic organs. 5 Based on pathophysiological considerations and benefits observed in nonrandomized studies in the pre-PCI era, previous American Heart F I G U R E 3 Forest plots comparing reinfarction, repeat revascularizaton, stroke, acute renal failure, and major bleeding for patients with acute myocardial infarction complicated by cardiogenic shock undergoing IABP insertion before or after primary percutaneous coronary intervention. CI, confidence interval; IABP, intra-aortic balloon pump; PCI, percutaneous coronary intervention found that IABP insertion before PCI was associated with a larger infarct size, and no difference was found between the two strategies regarding short-term and long-term mortality. 23 Considering the small sample size of the studies and the controversial results, pooled analysis of the individual data may be informative.
The principal finding of this study is that the timing of IABP insertion that is, before or after primary PCI does not have an effect on the short-term and long-term mortality in patients with AMI complicated by CS. It is believed that the early initiation of IABP therapy improves myocardial perfusion and results in significant myocardial salvage than reperfusion alone. 33 More importantly, hemodynamic stabilization in the setting of cardiogenic shock can prevent the relevant multi-organ dysfunction or failure. 34 One possible explanation is that the advantages of early initiation of IABP support are offset by the delay in revascularization associated with the time needed for IABP insertion. 23

| Limitations
Our meta-analysis presents several limitations that merit attention.
First, observational studies were mainly included in our meta-analysis due to lack of randomized data. This introduced intense heterogeneity and potential bias. Hence, the findings of the present study should be interpreted as hypothesis-generating only, and could not be overstated. The random-effect model was used to account for the heterogeneity. Although sensitivity analysis with multivariableadjusted data was performed, the potential bias cannot be completely eliminated. Furthermore, meta-regression analysis found that shortterm mortality might be interfered by baseline characteristics of diabetes mellitus and hypertension. Second, patients with STEMI and non-STEMI were both enrolled in our meta-analysis. In this case, sensitivity analysis was conducted by analyzing the results of studies that enrolled patients with STEMI exclusively, and the results were in line with the overall population. Third, data about the time needed for IABP insertion or door-to-balloon time were not available in most of the studies. Finally, most of the eligible studies reported in-hospital or 30-day mortality, and long-term data with more than 6 months were limited.

| CONCLUSIONS
In patients with AMI complicated by CS undergoing primary PCI, the timing of initiation of IABP therapy does not appear to impact shortterm and long-term clinical outcomes. However, this result should be interpreted with caution based on observational data. Appropriately, powered randomized trials are warranted to investigate the relative benefit of the two strategies, that is, IABP inserted before or after primary PCI in the future.

CONFLICT OF INTEREST
The authors declare no potential conflict of interests.