Seven RCTs (n=1,009 patients with ST-segment elevation myocardial infarction) and ten cohort studies (n=10, 529 patients with ST-segment elevation myocardial infarction and cardiogenic shock) were included.
RCTs in high-risk ST-segment elevation myocardial infarction (STEMI) patients: Five RCTs reported randomisation methods, four used intention-to-treat analysis; in five the outcome assessors were blinded. There was no statistically significant difference between intra-aortic balloon pump and no intra-aortic balloon pump in 30-day mortality or change in left ventricular ejection fraction. Intra-aortic balloon pump was associated with a statistically significant increased risk of stroke, risk difference 2% (95% CI 0 to 4) and bleeding, risk difference 6% (95% CI 1 to 11). Results were similar when trials were analysed according to type of reperfusion therapy. No significant heterogeneity was found for any of the analyses. Funnel plots showed no evidence of publication bias.
Cohort studies in STEMI patients with cardiogenic shock: None of the studies properly controlled for confounders. Patients in the intra-aortic balloon pump groups were younger and more commonly male. Significant heterogeneity was found for the analysis of all cohort studies (I2=94%). For thrombolysis studies, adjunctive intra-aortic balloon pump treatment was associated with a statistically significant decrease in 30-day mortality, risk difference 18% (95% CI 16 to 20). For primary percutaneous coronary intervention studies, adjunctive intra-aortic balloon pump treatment was associated with a statistically significant increase in 30-day mortality, risk difference 6% (95% CI 3 to 10). Funnel plots showed no evidence of publication bias. Revascularisation rates (rescue percutaneous coronary intervention) were significantly higher in intra-aortic balloon pump compared to control patients, relative risk 4.0 (95% CI 3.6 to 4.5).