From Wiki Journal Club
Jump to: navigation, search
Mehra MR, et al. "Two-year Outcomes with a Magnetically Levitated Cardiac Pump in Heart Failure". The New England Journal of Medicine. 2018. 378(15):1386-95.
PubMedFull textPDF

Clinical Question

In patients with advanced heart failure refractory to medical management referred for durable left ventricular assist device (LVAD) placement, is a magnetically levitated centrifugal-flow device superior to an axial-flow device with regard to reoperation for device malfunction or disabling stroke?

Bottom Line

In patients with advanced heart failure refractory to medical management referred for durable left ventricular assist device (LVAD) placement, a magnetically levitated centrifugal-flow device was associated with 19.3% absolute reduction in reoperation for device malfunction or disabling stroke at 2 years. This was driven primarily by a 15.4% absolute reduction in the need for reoperation for device malfunction, although there was also a 9.1% absolute reduction in overall stroke (not disabling stroke).

Major Points

LVADs are proven effective both for providing temporary support to patients awaiting cardiac transplant (termed "bridge to transplant") as well as durably improving survival and quality of life in patients who are not transplant candidates (termed "destination therapy", see REMATCH and HEARTMATE II). Since the original REMATCH trial, LVAD technology has improved, with transition from first-generation pulsatile-flow devices to second-generation continuous-flow devices, which were shown in HEARTMATE II to be associated with fewer mechanical failures and disabling strokes. Nevertheless, even second-generation continuous flow devices are susceptible to mechanical complications such as pump thrombosis, which often results in need for reoperation. Magnetically levitated centrifugal-flow devices may improve upon the earlier axial-flow design by using a magnetic induction motor to generate flow centrifugally, allowing for wide blood-flow paths and intrinsic pulsatility, factors that may reduce pump thrombosis rates. In a 6-month analysis of the HeartMate III centrifugal-flow device, there was an absence of pump thrombosis compared to a 10.1% rate observed with second-generation axial-flow LVADs.[1] Whether, this difference persists in longer-term follow up or is associated with clinical events remained unproven.

The 2018 Multicenter Study of MagLev Technology in Patients Undergoing Mechanical Circulatory Support Therapy with HeartMate 3 (MOMENTUM 3) trial randomized 366 patients with advanced heart failure requiring LVAD for either bridge or destination therapy (LVEF ≤ 25%, NYHA III-IV, inotrope dependent or cardiac index < 2.2) to the HeartMate III centrifugal-flow LVAD or a second-generation axial-flow LVAD. At 2 years, the HeartMate III device was associated with a 19.3% absolute reduction in reoperation for device malfunction or disabling stroke. This was driven primarily by a 15.4% absolute reduction in the need for reoperation for device malfunction, which was in turn driven by 14.6% absolute reduction in pump thrombosis rates with the HeartMate III. Rates of disabling stroke and death were similar in both groups, although there was a statistically significant 9.1% absolute reduction in overall stroke. Measures of heart failure symptoms, quality of life, and functional status improved similarly in both groups. Adverse events other than stroke or reoperation were similar with both devices, including rates of drive-line infection, right heart failure, and bleeding.

The findings of MOMENTUM 3 provide strong evidence that centrifugal-flow pumps represent yet another advance in LVAD technology, with a robust signal for increased durability with markedly less pump thrombosis as well as a suggestion of fewer overall strokes. On the basis of this evidence, these third-generation devices are likely to gain widespread adoptance among patients referred for LVAD placement.


As of May 2018, no guidelines have been published that reflect the results of this trial.


  • Multicenter, open-label, randomized, controlled trial
  • N=366
    • HeartMate III centrifugal-flow LVAD (N=190)
    • HeartMate II axial-flow LVAD (N=176)
  • Setting: 52 sites in the US
  • Enrollment: September 2014 - November 2015
  • Duration of follow-up: 2 years
  • Analysis: Intention-to-treat
  • Primary Outcome: Reoperation for device malfunction or disabling stroke


Inclusion Criteria

  • Age ≥ 18
  • Body surface area ≥ 1.2 m2
  • NYHA class III-IV
  • LVEF ≤ 25%
  • Advanced heart failure defined as either of:
    • Inotrope dependent
    • Cardiac index < 2.2 without inotropes and either of:
      • On optimal medical therapy for at least 45 out of 60 days
      • Advanced heart failure for at least 14 days and dependent on intra-aortic balloon pump for at least 7 days

Exclusion Criteria

  • Etiology of heart failure is uncorrected thyroid disease, obstructive cardiomyopathy, pericardial disease, amyloidosis, or restrictive cardiomyopathy
  • Technical factors leading to inordinately high surgical risk
  • Existing durable mechanical circulatory support device
  • Pregnant
  • Mechanical aortic valve that will not be either converted to a bioprothesis or oversewn at the time of LVAD placement
  • History of organ transplantation
  • Platelet count < 100,000
  • Psychiatric disease or disorder, irreversible cognitive impairment, or psychosocial issues likely to impair compliance with the study protocol and LVAD management
  • History of confirmed, untreated abdominal aortic aneursym > 5 cm in diameter within 6 months
  • Active, uncontrolled infection
  • Intolerance to anticoagulant or antiplatelet therapies or any other peri/post-operative therapy that the investigator will require based on the patient's health status
  • Presence of severe end-organ dysfunction or failure
  • Moderate or severe aortic insufficiency without plans for correction during LVAD placement
  • Pre-albumin < 150 mg/L or albumin < 30 g/L
  • Planned Bi-VAD support candidate
  • Known hypo- or hypercoagulable state
  • Predicted survival less than 24 months (outside of heart failure)

Baseline Characteristics

From the centrifugal-flow pump group.

  • Demographics: Age 61 years, male 79%, white 67%, black 27%
  • Comorbidities: BSA 2.1 m2, Cr 1.4
  • Heart failure: Ischemic 42%, LVEF 17%, PCWP 24, cardiac index 2.0, PVR 3.2, RAP 11, bridge to transplant 26%, bridge to transplant candidacy 16%, destination 58%
  • Other cardiac: AF 43%, stroke 8%, hx CABG 23%, hx valve surgery 10%


  • Patients randomized 1:1
    • HeartMate III centrifugal-flow pump (N=190)
    • HeartMate II axial-flow pump (N=176)
  • Randomization performed with the use of permuted blocks, stratification by study center
  • Data collected at 1 day, 1 week, discharge, 1 month, 3 months, then every 6 months until 2 years
  • Recommended antithrombotic management in both groups included ASA 81-325MG QD and warfarin for goal INR 2-3
  • NYHA classification and 6-minute walk test performed by a trained technician used to assess functional status
  • Primary endpoint assessed as intention-to-treat, but all secondary endpoints assessed as per-protocol


Comparisons are centrifugal-flow pump vs. axial-flow pump

Primary Outcomes

Survival free of reoperation for device failure or disabling stroke (2 years, noninferiority)
151 (79.5%) vs. 106 (60.2%); Absolute difference 19.2 [95% lower confidence bound 9.8, p < 0.001]
Survival free of reoperation for device failure or disabling stroke (2 years, superiority)
151 (79.5%) vs. 106 (60.2%); HR 0.46 [95% CI 0.31-0.69, p < 0.001]

Secondary Outcomes

Reoperation to replace or remove pump
3 (1.6%) vs. 30 (17.0%); HR 0.08 (95% CI 0.03-0.27); p < 0.001
Disabling stroke
11 (5.8%) vs. 7 (4.0%); HR 1.45 (95% CI 0.56-3.75); p = 0.42
22 (11.6%) vs. 26 (14.8%); HR 0.75 (95% CI 0.43-1.33); p = 0.37

Adverse Events

Suspected pump thrombosis
2 (1.1%) vs. 27 (15.7%); HR 0.06 [95% CI 0.01-0.26, p < 0.001]
Any bleeding
81 (42.9%) vs. 90 (52.3%); HR 0.71 [95% CI 0.53-0.96, p = 0.07]
LVAD drive line infection
45 (23.8%) vs. 34 (19.8%); HR 1.15 [95% CI 0.73-1.79, p = 0.37]
Right heart failure
60 (31.7%) vs. 48 (27.9%); HR 1.12 [95% CI 0.77-1.64, p = 0.49]
Ventricular arrhythmia
45 (23.8%) vs. 39 (22.7%); HR 1.04 [95% CI 0.67-1.59, p = 0.80]

Subgroup Analysis

  • No significant interaction between groups with regard to the primary endpoint was observed for the prespecified subgroups according to age, sex, race or ethnic group, intended goal of device therapy, or INTERMACS profile.


  • Open-label design allows for bias in outcomes assessment, which is particularly important with a clinically-driven outcome (e.g., reoperation). For example, it is possible that treating physicians were more likely to refer for reoperation in patients known to have the second-generation LVAD. However, given the magnitude of the decision to embark upon reoperation, the effect of this bias is likely modest if present.
  • Given only two year follow up period, the effect of possible longer-term complications of the third-generation device (e.g., outflow-graft twist or electrical failure, which have been described) cannot assessed with certainty.


  • Study supported by Abbott, maker of both the LVAD models studied.
  • Authors with multiple ties to industry, including study sponsor.
  • Sponsor provided devices studied and performed data analysis (with verification by an independent statistician).

Further Reading

  1. Mehra MR et al. A Fully Magnetically Levitated Circulatory Pump for Advanced Heart Failure. N. Engl. J. Med. 2017. 376:440-450.