Evidence for the use of IABP in MI and cardiogenic shock

This level of depth has never been interrogated by the college examiners in the written paper. However, given that this large pulsatile helium-filled object gets shoved up into people quite often, the author felt an acute need to familiarise himself with the available evidence for and against this practice. A great thanks is due to LITFL; their IABP page formed the core of this brief literature summary. Generally, the breadth of human experience with the IABP in cardiogenic shock can be divided into different historical periods in cardiology. There is the pre-thrombolysis dark age (characterised by medieval peasant levels of mortality from cardiogenic shock), the Renaissance which followed the wide acceptance of thrombolytic therapy, the Industrial Revolution associated with enthusiastic use of early stenting, and the cynical Post-Modern era characterised by the bleak nihilism of its largely unfavourable meta-analysis findings.

Pre-thrombolysis era

O'Rourke et al (1981)

• Small early study: 30 patients with early STEMI and signs of acute heart failure
• All received standard therapy for this historical period (i.e. heparin aspirin and GTN)
• No improvement in hospital mortality or functional outcome at 36 months

Flaherty et al (1985)

• Small early study; 20 patients of whom 10 received IABP for 5 days
• No improvement in LVEF or infarct size after the pump was removed.
• Tiny sample means whatever they concluded about mortality probably didn't mean much. However, the authors remarked that "combined afterload-lowering effects of intraaortic balloon pumping and nitroglycerin did appear to prevent dilation or remodeling of noninfarcted segments during the first 2 weeks after acute myocardial infarction".

Thrombolysis and balloon angioplasty era

Randomised IABP study group trial (1994)

• Moderate size trial; 182 patients were randomised within 24 hrs of infarction.
• Primary intervention was balloon angioplasty and/or intra-arterial thrombolysis; in order to get the pump you needed TIMI 2 or 3 flow restoration.
• IABP group had significantly less reocclusion of the infarct-related artery during follow-up (median 5 days) compared with control patients (8% versus 21%).
• However, cardiogenic shock patients were stupidly excluded, as was anybody in pulmonary oedema. Some might argue that this would have been the most interesting group to investigate.

Kovack et al (1997)

• Finally somebody looked at cardiogenic shock.
• Unfortunately, they looked at it through the lens of a retrospective chart audit.
• 335 cardiogenic shock cases were reviwed (1985-1995); 46 were included because they underwent thrombolysis, and of these lucky few 27 underwent IABP.
• The pumped patients appeared to die later (about 24 hours down the track, rather than only 6-7 hours for the non-IABP cohort) which was viewed as something of a victory.
• Even more positively, survival for the IABP cohort appeared to be more than doubled when compared to the non-IABP group ( 93% vs 37% in-hospital survival and 67% vs 32% at one year follow-up).
• The authors made strong recommendations regarding the use of IABP, particularly where it leads to a cardiogenic shock patient surviving long enough to be transferred from a regional hospital to a center where revascularisation is possible. These days, as every corner pub and brothel seems to have an angio suite, the latter point is perhaps less relevant.

GUSTO-I trial (1997)

• "Global Utilization of Streptokinase and Tissue Plasminogen Activator for Occluded Coronary Arteries" was a large-scale trial looking not at IABP but at the use of thrombolysis for AMI. The linked article is actually a sub-group analysis of those GUSTO-I patients who had cardiogenic shock.
• Early IABP use was associated with a trend toward lower 30-day mortality (47% vs 60%) but of course the study was not powered to detect this difference.
• However, of the total massive  GUSTO-I cohort (n = 41,021) only 310 patients had cardiogenic shock, and of these only 62 patients were blessed with an IABP.
• In spite of more bleeding complications being detected in the IABP group and the weaknesses of statistical analysis in such small numbers, the authors somehow concluded that IABP was being "underutilised in patients presenting with cardiogenic shock, both within and outside the United States."

SHOCK (2000)

• The linked article is a discussion of the impact of IABP together with thrombolysis in the SHOCK trial ("Should We Emergently Revascularize Occluded Coronary Arteries for Cardiogenic Shock"). As he name suggests, all patients had cardiogenic shock.
• In this group (of 1,126 patients) there was statistically significant lower in-hospital mortality among patients who received IABP verses those who did not (50% vs 72%).
• However,  the groups were markedly unequal: the IABP group got revascularised much more frequently, which naturally resulted in far more survivors. The effect of intervention diminished when the authors made an attempt to correct for this inequality, but the remaining in-hospital mortality improvement (47% vs 63%) was still encouraging.

NRMI-2 (2001)

• The NRMI-2 was a prospective observational cohort which started colelcting data in 1994. The number of patients with cardiogenic shock in this cohort was massive (n = 23,180).
• Weirdly, IABP use improved mortality in patients who received thrombolysis (67% vs 49%) but was not in those who were treated with primary angioplasty (45% vs 47%).
• Overall, IABP use together with thrombolysis resulted in a decrease in the 3-day odds of death by 18%.
• Critics of this study point to the research methodology; the poorer the quality of the study, the greater the benefit it demonstrates.

TACTICS trial (2005)

• Small trial, 57 patients; thrombolysis alone vs  thrombolysis plus IABP.
• There was no additional survival benefit in the IABP group.
• Killip Class III or IV (i.e. those patients with acute pulmonary oedema or in cardiogenic shock, who might have expected a mortality of 38% to 81% in 1967) demonstrated a trend towards improved mortality, but with such small numbers the study could not reach any firm conclusions about this group.

Modern Percutaneous Reperfusion Era

IABP-SHOCK II trial (2012)

• The link above points to the even-more-interesting 1-year follow-up of the IABP-SHOCK-II patients.
• The trial randomised post-angio patients (n= 600) to either IABP or no IABP.
• There was neither improvement in 30-day mortality, nor 1-year mortality, nor in any of the secondary outcomes (eg. stroke or functional capacity)
• Critics have pointed to the inadequate power to detect a 12% difference in survival at 30 days.  Additionally, the "shocked" patients were not "severely shocked" (i.e. their mortality rate of 40% was lower than the average of 42-48%).
• Overall, one may draw the conclusion that the real life-saver in these situations was the early revascularisation, and that with open vessels the IABP adds little even in cardiogenic shock.

Conclusions of a modern meta-analysis (Cochrane, 2015)

• From the seven eligble studies, a total of 790 patients was scraped together, of whom the vast majority still came from the IABP-SHOCK II trial.
• Unsurprisingly, there was still no mortality benefit.
• The Cochrane analyst concluded that the IABP "may have a beneficial effect on some haemodynamic parameters", which does not seem to translate into a mortality benefit in cardiogenic shock.