In-Hospital Mortality Hasn’t Budged Despite Changes in MCS for Cardiogenic Shock

Observational US data on more than 400,000 patients speak to the need for continued efforts at better care.

In-Hospital Mortality Hasn’t Budged Despite Changes in MCS for Cardiogenic Shock

SAN FRANCISCO, CA—That newer forms of mechanical circulatory support (MCS) are on the rise and intra-aortic balloon pump (IABP) therapy is on the decline won’t come as a surprise to anyone watching trends in cardiogenic shock care. Less intuitively, though, the mortality rate has stayed stubbornly high despite these shifts, according to an analysis of US data on more than 400,000 patients treated between 2012 and 2015.

Patterns were similar for cardiogenic shock regardless of whether it occurred in the setting of acute MI, researchers found.

IABP therapy, once the gold standard for MCS, did not show any short-term survival benefit in the IABP-SHOCK II trial, which led to its fall from grace. European guidelines advise against routine IABP use (class III recommendation), but US guidelines are more neutral (class IIb recommendation, level of evidence B). Numerous MCS technologies, from percutaneous ventricular assist devices (pVADs) to extracorporeal membrane oxygenation (ECMO), have emerged as options for treating these complex, challenging patients.

Yet practice changes have been slow, in part because of the lack of randomized trial data to date, and it’s unclear whether MCS has—on a large scale—had an impact on the outcomes of cardiogenic shock.

“Despite significant advances in cardiovascular medicine, mortality from cardiogenic shock is still approximately 40% to 50%,” said Ankur Kalra, MD (Cleveland Clinic, OH), when presenting the results here last week at TCT 2019. Kalra served as senior author of the study, published simultaneously online in Structural Heart, along with Srihari S. Naidu, MD (New York Medical College/Westchester Medical Center, Valhalla, New York).

However, some attendees at the TCT session still held out hope for a survival benefit with MCS, noting that the numbers of patients receiving newer cardiogenic-shock therapies might have been too small and the length of follow-up too short to detect a difference.

National Inpatient Sample

Lead investigator Muhammad Siyab Panhwar, MD (Tulane University School of Medicine, New Orleans, LA), and colleagues queried the National Inpatient Sample from 2012 to 2015 to identify 422,575 patients who had been diagnosed with cardiogenic shock. Among these patients, 44.7% had shock in the context of acute MI. Patients in the acute-MI subgroup were older (mean 68.8 vs 65.8 years) and more likely to be white (69.8% vs 63.5%), but less likely to be women (37.8% vs 39.6%, P < 0.001 for all).

Across the years, the cardiogenic shock patients became higher risk in general. Length of stay decreased slightly but significantly (mean 6.75 days in 2012 vs 6.71 days in 2015; P = 0.001), while average hospital costs generally remained similar at around $28,000.

From 2012 to 2015, both MCS overall and IABP therapy declined in use, as did that of nonpercutaneous MCS devices. Meanwhile, pVADs and ECMO—though still employed in the minority of patients—grew more popular. In-hospital mortality (primary endpoint) held steady over the 4-year study period, whether or not patients had acute MI.

MCS Use and Adjusted Mortality for Cardiogenic Shock: NIS




P for Trend

Overall MCS



< 0.001




< 0.001

Nonpercutaneous VAD




Percutaneous VAD



< 0.001





In-Hospital Mortality

    With Acute MI

    Without Acute MI











Amid these shifts, adjusted rates of acute kidney injury requiring dialysis and bleeding did not change. However, among patients receiving MCS, vascular complications increased from 4.8% in 2012 to 6.4% in 2015 (P for trend = 0.04). Specifically for patients receiving ECMO, bleeding rose over time (from 13.0% in 2012 to 21.7% in 2015; P for trend = 0.01).

Independent predictors for higher in-hospital mortality included: age (OR 1.03; 95% CI 1.02-1.04), female sex (OR 1.16; 95% CI 1.12-1.20), cardiac arrest (OR 2.81; 95% CI 2.70-2.92), and mechanical ventilation (OR 2.97; 95% CI 2.86-3.08). On the flip side, the presence of acute MI was associated with lower in-hospital mortality (OR 0.94; 95% CI 0.91-0.98).

Kalra stressed that cardiogenic shock patients are a heterogeneous group. “Cardiogenic shock exists along the spectrum of severity, and it’s difficult to establish a one-size-fits-all treatment,” he said. This is why the recent attempt by the Society for Cardiovascular Angiography and Interventions to standardize the “verbiage of how we define shock” is so important, Kalra noted.

The data from NIS should inspire “early recognition of impending cardiogenic shock, and institution of appropriate and aggressive measures in an effort to improve survival,” he concluded. “Our results also highlight the continued challenges faced in management of cardiogenic shock in an era of mechanical circulatory support. Further research is needed to determine optimal utilization to improve outcomes.”

Some Aren’t Convinced

In discussion following Kalra’s presentation, moderator James M. McCabe, MD (University of Washington Medical Center, Seattle), pointed out that the rise in pVAD amounted to only 1%, meaning that it would be difficult to observe any changes in mortality associated with these devices. Though he praised the decision to separate out acute MI, McCabe questioned how these cases would be managed differently as compared to other presentations of cardiogenic shock.

Kalra noted that one reason that acute MI was associated with lower in-hospital mortality is that the strategies for treating these patients are relatively well known. “We revascularize and support those patients,” he said, adding that even without acute MI, early recognition and treatment are key.

For panelist Richard Smalling, MD, PhD (Memorial Hermann Heart and Vascular Institute, Houston, TX), the way to detect a benefit of MCS is straightforward: look long term.

“If you’re going to help somebody with acute MI, you have to unload within 2 hours of onset of symptoms to show a reduction in infarct size. That’s because it reduces endothelial dysfunction and calcium overload,” he explained. “What unloading also does is it prevents apoptosis, which has a latent manifestation. So if you use in-hospital mortality, you won’t find a signal. If you look at 30-day or 6-month mortality, you’ll find a signal.” Patients who are unloaded prior to reperfusion will do better, Smalling predicted.

Caitlin E. Cox is News Editor of TCTMD and Associate Director, Editorial Content at the Cardiovascular Research Foundation. She produces the…

Read Full Bio
  • Panhwar, Kalra, and Naidu report no relevant conflicts of interest.