Longer-Term Data Still Worrisome for Impella-Treated AMICS Patients

Patients undergoing PCI for acute MI complicated by cardiogenic shock (AMICS) had worse outcomes—including more mortality and bleeding—when their hemodynamic support during that hospitalization was provided by the Impella (Abiomed) left ventricular assist device (LVAD) as compared with an intra-aortic balloon pump (IABP), not just in the short term but also at 30 days and 1 year, a new retrospective analysis suggests.

Moreover, use of the intravascular microaxial LVAD continues to be associated with more kidney dysfunction and higher mean costs a year after the device has been removed, researchers reported online in JAMA Internal Medicine.

“Our analysis certainly suggests that the use of intravascular LVADs might be associated with higher rates of adverse events and increased cost,” P. Elliott Miller, MD (Yale School of Medicine, New Haven, CT), lead author of the new study, told TCTMD. “Clinicians should take this into account, as well other studies in the literature, and their experience to manage patients.”

The new findings extend the time frame covered by two earlier retrospective reports of short-term outcomes, first released at the 2019 American Heart Association meeting, that also attempted to overcome confounding variables in observational studies by propensity-matching patients treated with LVAD or IABP. Much debate followed over what role the Impella should play in this setting while the field awaits results from randomized controlled trials.

Providers need to apply the evidence and their expertise to the patients they see. P. Elliott Miller

Such trials are notoriously hard to conduct in AMICS. The first to offer any randomized insights will be DanGer, which is still enrolling patients.

“Cardiogenic shock is a difficult topic to study in a randomized fashion given the critical nature of the disease. However, an RCT is this space will be crucial to move the field forward, and most importantly for our patients,” said Miller.

The question, then, is how these acutely ill patients should be managed in the meantime. For his part, Miller said that “providers need to apply the evidence and their expertise to the patients they see.”

Alexander Papolos, MD (MedStar Washington Hospital Center, Washington, DC), to TCTMD, praised the investigators for undertaking this work in an area urgently needing answers, but cautioned that such retrospective data can only be considered hypothesis-generating.

“Randomized data will be needed to influence behavior,” Papolos predicted. For instance, one limitation here is that propensity-matching “cannot accurately correct for the bias of patient illness-severity as it relates to physician-driven MCS platform selection.”

In a similar vein, Maya Guglin, MD, PhD (Indiana University School of Medicine, Indianapolis), also speaking with TCTMD, noted that administrative claims data should be “taken with a grain of salt,” since analyses that rely on these data aren’t, for example, able to determine which iteration of Impella was used—the 2.5, CP, 5.0, or 5.5—and whether the varying degrees of hemodynamic support they provide might affect outcomes.

Having said that, at least for now, “retrospective studies with propensity matching are the best we have, unfortunately. . . . They all point in the same direction. This is of serious concern,” she added.

Renal Dysfunction More Common

For the study, Miller and colleagues searched a database of commercially insured patients spanning 14 US states that includes details on medical and pharmacy claims as well as lab results. Between January 2015 and April 2020, there were 3,077 patients undergoing PCI for AMICS (mean age 65.2 years; 29% women), of whom 32% experienced cardiac arrest.

Overall, 72% of patients received IABP support and 28% received an Impella LVAD. In early 2015, just 10.7% got an LVAD. By 2019, that proportion had risen to 39.1%.

Randomized data will be needed to influence behavior. Alexander Papolos

Patients in the LVAD group were more likely than the IABP-treated patients to have hypertension, diabetes, mellitus, heart failure, atrial fibrillation, peripheral vascular disease, chronic lung disease, chronic kidney disease, and prior cardiac arrest. They also had higher healthcare utilization in the 6 months leading up to their index hospitalization.

The researchers propensity matched patients for age, sex, race/ethnicity, region, Deyo-Charlson Comorbidity Index score, comorbidities, and prior healthcare utilization/costs, arriving at 817 pairs.

During the index hospital stay, at 30 days, and at 1 year, LVAD use was associated with higher likelihood of mortality and severe bleeding compared with IABP use. By 30 days and 1 year, there also were increases in kidney replacement therapy (KRT) with LVAD use.

Outcomes of 817 Propensity-Matched Pairs

*One-year outcomes were calculated as hazard ratios (not odds ratios)

Mean costs, too, were higher with LVAD versus IABP support, coming in at $60,279 more during the index hospital stay, $51,680 more at 30 days, and $46,609 more at 1 year.

One possible explanation for the increased risk among LVAD-treated patients relates to kidney injury, Miller et all suggest. “Among numerous potential complications of cardiogenic shock, kidney failure often represents severe end-organ hypoperfusion and portends a particularly poor prognosis,” they write. “Importantly, studies assessing associations of device selection and subsequent kidney failure are limited and often have mixed results depending on the included population.”

In this analysis, the more-frequent need for kidney replacement therapy in conjunction with LVAD use “may partly explain” the higher mortality, the researchers say. “Of note, the higher cardiac output ascribed to percutaneous circulatory support devices did not appear to provide kidney protection in our analysis; rather, use of the devices was associated with increased rates of kidney replacement therapy.”

Guglin said that the accumulation of evidence on Impella may lead to less enthusiasm for the device.

“How do you get away from this current impression that there is no benefit and there are actually well-described hazards? The hazards are real. We clinicians see them when we deal with Impella on a day-to-day basis, because [patients] do develop hemolysis and they do develop bleeding. In this analysis, profound renal dysfunction was more common, as well,” she said.

Nor has a benefit to IABP been entirely ruled out, Guglin observed, despite trial data to the contrary.

In her experience, “it is surprising, but balloon pumps sometimes really work. It is difficult to explain,” she said, since IABPs provide “minimal hemodynamic support” amounting to around 0.5 L of circulation per hour. “In cardiogenic shock, when you need almost complete replacement of the circulation [this] is almost nothing. . . . Cardiologists tend to think in physiological terms. In physiological terms, Impella sounds much better than a balloon pump. It ought to deliver better results, but it doesn’t and that’s what’s so puzzling to us.”

Thus, whatever the results of RCTs on Impella show, there’s likely to be ongoing discussion among clinicians about what approach is best in AMICS.

Gene Hu, MD, Anand R. Habib, MD, and Rita F. Redberg, MD (all from University of California San Francisco), writing in an accompanying editorial, urge caution for the time being.

“The lack of data from RCTs to support a mortality benefit from the intravascular microaxial LVAD and the available observational data of increased mortality and serious patient harms are troubling,” they write, adding that both guideline-writing committees and the US Food and Drug Administration should take note.

Until those results arrive, use of these LVADs for AMICS “should be restricted to patients enrolled in an RCT.”

They conclude: “Although it is tempting for clinicians to want to be able to provide patients with more-advanced therapies for diseases as fatal as myocardial infarction with cardiogenic shock, the medical community must ensure that interventions should ‘do no harm.’ It is crucial that clinicians only administer therapies that have proven safety and efficacy.”