Comprehensive Analysis of Stroke in the Long-Term Cohort of the MOMENTUM 3 Study: A Randomized Controlled Trial of the HeartMate 3 Versus the HeartMate II Cardiac Pump

Paolo C. Colombo, Mandeep R. Mehra, Daniel J. Goldstein, Jerry D. Estep, Christopher Salerno, Ulrich P. Jorde, Jennifer A. Cowger, Joseph C. Cleveland, Nir Uriel, Gabriel Sayer, Eric R. Skipper, Francis X. Downey, Masahiro Ono, Robert Hooker, Anelechi C. Anyanwu, Michael M. Givertz, Claudius Mahr, Ia Topuria, Sami I. Somo, Daniel L. CrandallDouglas A. Horstmanshof

Research output: Contribution to journalArticlepeer-review

104 Scopus citations


Background: The MOMENTUM 3 study (Multicenter Study of MagLev Technology in Patients Undergoing Mechanical Circulatory Support Therapy With HeartMate 3) has demonstrated that the HeartMate 3 (HM3) pump is associated with reduced strokes compared with the HeartMate II (HMII) device. We now perform a comprehensive analysis of stroke events to evaluate their longitudinal occurrence, clinical correlates, patterns, and impact on outcome across the 2-year duration of support. Methods: MOMENTUM 3 is a randomized controlled trial of the HM3 centrifugal-flow pump versus the HMII axial-flow pump in patients with advanced heart failure, regardless of the intended goal of support (bridge to transplantation or destination therapy). Baseline and postimplantation clinical correlates of stroke events were assessed with multivariable analyses. Longitudinal patterns, including device association, type of stroke (hemorrhagic versus ischemic), changing severity of impairment assessed with the modified Rankin Scale (disabling [modified Rankin Scale score >3] versus nondisabling [modified Rankin Scale score ≤3]) over time, and association with outcome, were determined. Results: In 361 patients with the intended implant (189 HM3 and 172 HMII), 65 strokes (40 ischemic strokes and 25 hemorrhagic strokes) occurred in 52 patients at a median of 131 (range, 1-733) days. No difference in stroke rate was noted between 0 and 180 days of follow-up between devices. However, stroke incidence in the long-term period (181-730 days after left ventricular assist device) was 3.3 times lower for the HM3 group (HM3: 0.04 versus HMII: 0.13 events per patient-year; odds ratio, 0.23; 95% CI, 0.08-0.63; P=0.01). Treatment with the HM3 pump was the only independent predictor of lower stroke events. We found no direct association of blood pressure or antithrombotic regimens with observed stroke rates. A stroke event significantly lowered 2-year postimplantation survival regardless of subtype or initial severity of neurological impairment compared with patients without a stroke (43±12% for hemorrhagic stroke, 57±9% for ischemic stroke, 51±11% for disabling, and 51±11% for nondisabling compared with 85±2% 2-year survival for patients without stroke). Conclusions: The HM3 pump is associated with a marked reduction in stroke rates compared with the HMII device, with benefits observed in the long-term period (>6 months). The occurrence of stroke of any type (hemorrhagic and ischemic) or of any functional severity (disabling and nondisabling) is predictive of a poor 2-year clinical outcome.

Original languageEnglish (US)
Pages (from-to)155-168
Number of pages14
Issue number2
StatePublished - Jan 8 2019


  • fibrinolytic agents
  • heart failure
  • heart-assist device
  • stroke
  • treatment outcome

ASJC Scopus subject areas

  • Cardiology and Cardiovascular Medicine
  • Physiology (medical)


Dive into the research topics of 'Comprehensive Analysis of Stroke in the Long-Term Cohort of the MOMENTUM 3 Study: A Randomized Controlled Trial of the HeartMate 3 Versus the HeartMate II Cardiac Pump'. Together they form a unique fingerprint.

Cite this