The Evidence for Intra-aortic Counterpulsation in High-risk PCI
▪ Registry data
The applicability and validity of registry data remains open to debate. When studying a patient cohort that is notoriously difficult to enroll in randomized trials, registry data can certainly add value by giving an all-comers, real-world, viewpoint on management trends and, since registries can be easily maintained, subsequent longer-term outcomes. Clearly there is the scepter of selection bias overhanging any registry, but with adjustment for baseline differences and the utilization of propensity score matching, the data accrued can be robust and can be used to evaluate the safety and efficacy of a particular intervention in everyday practice. Furthermore the larger registries allow investigators to study relatively rare outcomes and generate statistically significant inferences from them. This has certainly been the case for studying trends pertaining to IABP usage in high-risk clinical scenarios.
Brodie et al. looked at a registry of 1490 consecutive patients admitted to a single center with acute MI treated with primary PCI (the majority with balloon angioplasty but also coronary stenting in the last 3 years) without prior thrombolytic therapy from 1984 to 1997. An IABP was implanted in 213 (14.2%) of these patients, 133 had CS, 80 were hemodynamically stable but high-risk, 108 received counterpulsation before intervention and 105 after intervention. In 119 of the CS patients, the prophylactic use of IABC before intervention resulted in significantly lower rates of catheterization laboratory events (defined as ventricular fibrillation/tachycardia, cardiopulmonary arrest and prolonged hypotension) compared with no IABC or IABC after intervention. In 119 high-risk patients (defined as congestive heart failure or left ventricular ejection fraction (LVEF) ≤30%) preintervention IABC was also associated with a nonsignificant reduction in laboratory events. IABC led to an increased propensity for major bleeding although the investigators did recognize the potential adverse effect of larger sheaths, high dose heparin anticoagulation and longer in-dwelling of the sheath, could have had on this outcome.
The SHOCK Trial Registry concurrently enrolled 1190 patients with suspected CS from 36 participating centers during the SHOCK Trial recruitment period. Of these, 856 patients with CS secondary to acute MI were available for analysis. Those patients supported with an IABP were more likely to proceed to coronary angiography (p < 0.001) and as such, IABC use was associated with a lower mortality rate compared with no IABC (50 vs 72%; p < 0.0001). Patients proceeding to revascularization (PCI or CABG) following thrombolysis and IABP insertion had a significantly lower mortality than those who remained on a conservative strategy (39 vs 78%; p < 0.0001). A subgroup analysis of patients undergoing PCI was also undertaken to examine the effect of diastolic augmentation. Interestingly, there was no difference in in-hospital mortality for patients with an IABP in situ prior to PCI (n = 98, 47% mortality) versus those receiving counterpulsation after PCI (n = 95, 47% mortality) versus patients undergoing PCI without IABP support altogether (n = 56, 46% morality), although this result could be attributed to differences in patient selection, a well recognized flaw of all registries.
In the National Registry of Myocardial Infarction-2 (NRMI-2), investigators studied the outcomes of 23,180 participants presenting with CS or in whom CS developed during their hospital admission. IABC was used in 7268 (31%) patients. In those that received thrombolytic therapy, supplemental IABC conferred a significant mortality advantage (thrombolysis + IABP 49% vs thrombolysis alone 67%), perhaps reflecting the synergistic effect of both therapies working in tandem to establish patency of the infarct-related artery. High-risk primary PCI in this setting produced the lowest mortality rate, although insertion of an IABP did not lead to any further advantage (primary PCI 42% vs primary PCI + IABP 47%) and, in fact, was associated with higher hospital mortality rates. Unlike thrombolysis, PCI does not rely on coronary perfusion pressure to establish patency. It should also be borne in mind that the specific timing of IABP insertion was not available for this analysis.
Further iterations of the Benchmark Registry also provide a noteworthy insight in to IABP usage patterns in acute MI. Of the 22,663 patients that received an IABP at 250 centers between June 1996 and August 2001, 5495 (24%) were documented as having an acute MI. CS (n = 1498, 27.3%) and high-risk catheterization and angioplasty (n = 1495, 27.2%) were the principal indications for IABC emanating from the analysis. Unfortunately the investigators do not specifically state what constitutes 'high risk' in the paper, but with 59% of those undergoing cardiac catheterization having triple vessel disease and 16% LMCA involvement in addition to a mean ejection fraction of 36.5 ± 14.3% within the cohort, we can confidently surmise the type of characteristics that may have been used. As with the first Benchmark Registry analysis, rates of major IABP-related complications remained low and were similar between patients undergoing PCI (2.8%), surgery (2.6%) or conservative management with or without angiography (2.7%). Indeed only three (0.05%) of the 5495 patients died as a direct result of IABP placement. Furthermore, the process of IABP insertion was itself shown to be safe, with only 2.2% of patients suffering a failed procedure as a consequence of balloon leak, poor inflation, poor augmentation or insertion difficulty. And finally, in-hospital mortality of those acute MI patients requiring IABC support for high-risk catheterization/PCI was proportionately low (9.6%) despite it being the primary indication for hemodynamic assistance in over a quarter of the entire cohort (see Figure 10).
In-hospital mortality of 5495 patients with acute myocardial infarction requiring intra-aortic balloon pump counterpulsation, stratified by principal usage indication from the Benchmark Registry.
AMI: Acute myocardial infarction; PCI: Percutaneous coronary intervention.
Reproduced with permission from .
Results from an analysis of the National Cardiovascular Data Registry CathPCI Registry, in which the use and effectiveness of IABC for high-risk PCI was examined, give the most contemporaneous and provocative look at recent trends. Importantly, PCI was designated as high-risk if one of the following factors were present:
Unprotected LMCA as the target vessel;
Severely depressed LV function (<30%);
A total of 181,599 high-risk patients undergoing PCI at 681 hospitals between 1 January 2005 and 31 December 2007 were included in the analysis. The primary outcome measure was in-hospital mortality. Of the high-risk PCIs, 144,190 (79.4%) presented with STEMI, 21,259 (11.7%) had CS, 3592 (2.0%) had unprotected LMCA PCI and 37,394 (20.6%) had significant LV systolic dysfunction. An IABP was implanted in 44.4% of CS patients, 10.3% of STEMI patients, 28.1% receiving LMCA PCI and 13.9% of those with depressed LV function. Overall the perceived hemodynamic benefits of IABP were only deemed necessary for 10.5% of all high-risk cases, a relatively low figure given the widespread availability of the device. The investigators went on to categorize hospitals according to their proportional IABP use and grouped them in to corresponding quartiles. They found significant variation in IABP use across all participating centers, with a median odds ratio of 1.93 for a hospital effect, indicating a substantial influence held by actual location on whether a patient would receive IABC or not. Most poignantly, and after adjustment for multiple variables, in-hospital mortality or complication rates did not vary across hospital quartiles, despite differences in the rate of IABP implantation. Furthermore, a meticulous analysis of subgroups found no particular subset of patients benefitted from an increased frequency of IABP use. A fundamental limitation of this analysis was a lack of information on the timing of IABP insertion, one that has been recognized by the investigators. Nevertheless, they conclude there is little evidence to support the increased use of IABP at hospitals which were more selective on implementing this adjunct to high-risk PCI. Based on these results, it appears the use of IABP in a high-risk setting is not only low for such a commonly available device but also appears discretionary and is influenced more by local expertise, familiarity and protocols as opposed to definitive evidence from clinical trials and direction from guidelines.
There have been several randomized trials on the use of IABC in high-risk PCI, results of which, much like the registry data above, have produced conflicting results that serve to convolute the matter, rather than provide a uniform statement on the clinical effectiveness of IABP. These are summarized in Table 4. Three landmark trials are, however, worthy of closer scrutiny.
The SHOCK Trial randomized, in a 1:1 fashion, 302 patients presenting with acute MI complicated by CS precipitated by LV failure to emergency revascularization versus initial medical stabilization. Although an IABP was inserted in 86% of the patients in both groups, no clear benefit of IABC could be delineated for either treatment strategy. Indeed the high uptake of IABC in the conservative treatment arm, along with the frequent use of thrombolytic therapy, could well have led to a convergence in overall 30-day mortality, which was not significantly reduced by early revascularization (46.7% revascularization vs 56.0% medical therapy, p = 0.11). By 6 months, however, overall mortality was significantly lower in the revascularization arm (50.3 vs 63.1%; p = 0.027), a trend that persisted to improved 1-year survival rates in favor of revascularization (46.7 vs 33.6%; p < 0.03). Although not a direct investigation of IABC, this seminal study highlighted the frequent use of IABC in the context of revascularization for CS secondary to acute MI. By virtue of the improved 1-year survival rate, but by no means a statistically valid inference, we might surmise that the use of IABC during PCI in this high-risk cohort of patients appeared to be beneficial. The results of this landmark study have been used by both the ESC/EACTS and the ACCF/AHA/SCAI to justify their class I recommendation for IABC in acute MI complicated by CS. The survival benefit of early revascularization persisted to 6-year follow-up (32.8 vs 19.6%), irrespective of the emergency mode of revascularization used (i.e., PCI or CABG).
The BCIS-1 study was the first prospective, open, multicenter, randomized controlled trial to determine whether elective IABP insertion prior to high-risk single-vessel or multivessel PCI was able to reduce MACCE (i.e., death, acute MI, cerebrovascular event or the need for further revascularization by either PCI or CABG at hospital discharge capped at 28 days). High-risk in this instance was characterized by the following:
Depressed LV function (ejection fraction ≤30%);
A BCIS-1 Jeopardy Score of ≥8 (a modification of the Duke Jeopardy Score);
A target vessel that provided collateral supply to an occluded second vessel that in turn supplied more than 40% of myocardium.
Those patients with pre-existing class I or II indications for IABP insertion (i.e., CS, acute MI within the last 48 h, and complications of acute MI) were excluded. Elective IABP insertion took place before coronary intervention. Bailout IABC was permitted in the no planned IABP group if clinical circumstances warranted it. Overall 301 patients with multivessel disease and LV systolic dysfunction were randomized in a 1:1 fashion between December 2005 and January 2009. There were similar rates of the primary end point of MACCE between both arms of the trial (15.2% elective IABP vs 16.0% no planned IABP; p = 0.85). There was no significant difference in the secondary end points of 6-month mortality (it should be noted that the trial was not sufficiently powered to detect a difference in mortality at 6 months) or overall rates of bleeding although, when broken down, there were significantly more minor bleeds in the elective IABP arm (15.9% elective IABP vs 7.3% no planned IABP; p = 0.02). This was perhaps tempered by more periprocedural complications occurring in the no planned IABP arm, predominantly due to procedural hypotension (1.3 vs 10.7% in favor of elective IABP; p < 0.001), which might explain the need for rescue/bailout IABC in 18 (12%) patients overall.
Overall the study did not support the use of elective/prophylactic IABP insertion prior to high-risk PCI in terms of reducing the incidence of MACCE. Despite this, evidence from the study helped to justify the current class IIb recommendations assigned by the new ACCF/AHA/SCAI PCI guidelines on the elective use of hemodynamic support devices before performing high-risk PCI. Given the fact that 12% of the no planned IABP arm required rescue/bailout IABC during their procedure, does suggest, however, that an initial strategy of standby IABC for PCI in those patients with compromised LV functional reserve and extensive CAD could attenuate the delay in gaining arterial access and therefore prevent patients from entering a cascade of irreversible hemodynamic collapse. Interestingly, those patients requiring rescue IABC had a significantly higher BCIS-1 Jeopardy Score than those not requiring salvage in the no planned IABP group (Jeopardy Score: 11.2 vs 10.2; p = 0.02) further emphasizing the potential need for provisional hemodynamic support in those at the extreme end of the risk spectrum.
The CRISP AMI trial was a prospective, multicenter, open, randomized controlled trial undertaken to determine whether prophylactic IABP insertion within 6 h of pain onset and planned primary PCI for acute anterior STEMI (without CS) was able to reduce infarct size, as measured by cardiac MRI between 3 and 5 days postintervention, when compared with standard care. As with BCIS-1, the insertion of an IABP in the primary PCI alone group was at the operator's discretion for indications such as persistent hypotension or overt CS, malignant arrhythmias, or acute MI complications such as mitral regurgitation or ventricular septal defect. Of note, 15 (8.5%) patients initially receiving standard care crossed over to receive IABC.