Secondary Logo

Journal Logo

Brief Communication

The Heartware Lavare Cycle: A Cautionary Tale

Sorensen, Erik N.*; Dees, Lynn M.; Kaczorowski, David J.; Feller, Erika D.

Author Information
doi: 10.1097/MAT.0000000000001161

Hematologic complications of left ventricular-assist device (LVAD) therapy, such as stroke, pump thrombosis, and gastrointestinal (GI) bleeding, remain prevalent despite excellent post-LVAD survival.1 Continuous-flow LVAD support alters intraventricular flow dynamics, leading to areas of stasis that can precipitate thrombus formation.2 To mitigate this potentially deleterious condition, some devices incorporate speed-modulation algorithms to decrease intraventricular stasis.

In the Heartware HVAD (Medtronic, Minneapolis, MN), the speed-modulation algorithm is the Lavare cycle. It decreases pump RPM by 200 from baseline for 2 seconds, then raises it by 200 above baseline for one second, then returns to baseline. This recurs every 60 seconds. The algorithm can be turned on and off by the LVAD clinician.3

After several years of availability in Europe, the Lavare cycle was introduced in the US in mid-2018. European data suggest potentially lower rates of stroke in patients with Lavare enabled.4 However, that study has a small comparison group (33 patients, versus 215 with Lavare on) with significantly more RV failure and sepsis. The manuscript offers no guidance on when to enable Lavare, although the conference discussion which follows suggests the immediate postoperative period. The HVAD Instructions for Use recommend that it be enabled once the patient is hemodynamically stable, although they caution that its safety and effectiveness “have not been evaluated clinically.”3

Given the paucity of guidance, our team opted for a cautious approach, wherein we would activate Lavare only in newly implanted patients once they were stable and ambulatory. However, over time, we activated Lavare in several stable ongoing patients. After an unexpected spike in strokes and pump thrombosis in patients with recent activation, we elected to deactivate Lavare in all patients and impose a 6-month moratorium on its use to determine if the complication rate decreased. The purpose of this study is to analyze hematologic complication rates in patients with the Lavare cycle on and off.


Data for all patients implanted at our center and supported with an HVAD for >30 days between June 2018 (when Lavare became available at our center) and December 2019 were abstracted from an IRB-approved, prospectively collected LVAD database. The “Lavare On” group included all patients who had the Lavare cycle activated after postoperative day 1. The “Lavare Off” group was all patients who never had the Lavare cycle activated. Hematologic complications (pump thrombosis, embolic and hemorrhagic stroke, and GI bleeding) followed Intermacs definitions. We excluded perioperative stroke, traumatic intracranial bleeds, and GI bleeds related to aggressive treatment of thrombosis. During the study period, all patients were treated with the same anticoagulation protocol.


Sixty patients with a median support duration of 402 days (interquartile range [IQR] 212–484) met study criteria. Of these, 22 had Lavare enabled (at median postoperative day 283), and 38 did not. The two groups had similar baseline characteristics (Table 1), with the exception of more males in the Lavare On group. Median duration of Lavare activation was 78 days.

Table 1. - Baseline Characteristics of Patients Who Had Lavare Cycle Enabled (“Lavare On”) Versus Those Who Never Had it Enabled (“Lavare Off”)
Lavare On (N = 22) Lavare Off (N = 38) p-value
Age (years) 49.4 ± 15.4 53.4 ± 11.6 0.26
Male, N (%) 21 (95%) 27 (71%) 0.04
White race, N (%) 11 (50%) 12 (32%) 0.31
Ischemic etiology, N (%) 6 (27%) 8 (21%) 0.76
Intermacs class, N (%) 0.18
 1 2 (9%) 11 (29%)
 2 6 (27%) 10 (26%)
 3 9 (41%) 13 (46%)
 ≥4 5 (23%) 3 (8%)
VAD RPM 2686 ± 173 (nearest to Lavare on) 2671 ± 234 (at last follow-up) 0.80
Aortic valve opening, N (%) 13 (59%) (nearest to Lavare on) 21 (55%) (at last follow-up) 0.79
On support before study period, N (%) 15 (68%) 20 (53%)
 Support days before study period, median (IQR) 340 (309–684) 388 (137–462) 0.07
Support days during study period, median (IQR) 426 (320–441) 324 (196–557) 0.78
Postop day of Lavare activation, median (IQR) 283 (35–483)
Duration of Lavare activation (days), median (IQR) 78 (43–90)
IQR, interquartile range.

Table 2 compares hematologic complication incidences and rates. Both incidence and rate of pump thrombosis were significantly higher for the Lavare On group. Percentages and rates of both types of stroke were numerically, but not significantly, higher for Lavare On as well. Incidence and rate of GI bleeding were similar between the groups.

Table 2. - Incidence and Rate of Hematologic Complications in Patients Who Had Lavare Cycle Enabled (“Lavare On”) Versus Those Who Never Had it Enabled (“Lavare Off”)
Incidence, N (%) p-value Rate, EPPY p-value
Lavare On (N = 22) Lavare Off (N = 38) Lavare On (N = 22) Lavare Off (N = 38)
Pump thrombosis 7 (31.8) 3 (7.9) 0.02 0.315 0.081 0.04
Embolic stroke 3 (13.6) 1 (2.6) 0.14 0.135 0.027 0.15
Hemorrhagic stroke 3 (13.6) 2 (5.3) 0.25 0.180 0.054 0.14
GI bleed 3 (13.6) 5 (13.2) 1.00 0.360 0.296 0.24
EPPY, events per patient-year; GI, gastrointestinal.

In the Lavare On group, pump thrombosis occurred at median day 131 (IQR 30–253) after activation. Four of seven (56%) events occurred after Lavare had been deactivated for a median 110 (IQR 75–173) days. The three embolic strokes occurred days 9, 140, and 337 after Lavare activation, with one (33%) occurring 218 days after deactivation. The three hemorrhagic strokes occurred days 25, 27, and 101 after activation, with two (66%) occurring 15 and 124 days after deactivation. Finally, the first GI bleed occurred at post-Lavare days 12, 51, and 71; one (33%) patient’s first bleed occurred 61 days after deactivation.

We also analyzed other variables associated with pump thrombosis to see if Lavare activation was an independent risk factor (see Supplemental Material, Supplemental Digital Content 1, On univariate analysis, only Lavare and Intermacs profiles were significant. A bivariate model showed marginally improved significance (p = 0.047) when Intermacs profile was added to Lavare activation.


Our center experienced a significant increase in both incidence and rate of pump thrombosis in patients who had the Lavare cycle enabled later in their postoperative course (median day 283), when compared to a contemporaneous cohort who did not have Lavare activated. Although our study was underpowered for it, multivariate analysis also suggested that Lavare, along with Intermacs profile, was independently associated with pump thrombosis. We also saw higher rates of strokes, although not significant. Since patients had similar baseline characteristics (except gender) and support duration, and were all treated with the same anticoagulation protocol, we believe Lavare was a contributory element.

Our theory is that late activation of Lavare disturbs a stable but potentially prothrombotic pannus at the inflow cannula. Glass et al.5 analyzed 24 explanted HVADs from 22 patients on support for 71–1041 days and found thrombus in 23 (96%), with 27% of patients suffering an ischemic stroke. In most cases, thrombus is concentrated near and extends above the junction of sintered and smooth titanium on the cannula. It is possible that increased shear stress caused by Lavare-induced speed changes could disturb this pannus, leading to embolization or migration into the pump. The time course of formation of this pannus has not been studied, but the lower stroke rate seen in the European series4 could be due to early postoperative activation of Lavare, thus maintaining a consistent fluid dynamic milieu near the cannula during the early post-surgical phase, when the pannus is likely to initiate.

Based on this hypothesis and analysis of our data, our group elected to resume using Lavare after a 6-month moratorium, but to activate it immediately post-implant on arrival to the intensive care unit. We have done this in eight consecutive patients, and in limited follow-up (median 88d; IQR 23–112) there have been no hematologic complications.

A limitation of our study is short follow-up duration and a small patient population. We also did not randomize patients to Lavare activation, and late activations were based on judgment of individual providers. However, aside from male gender, the populations were balanced on baseline characteristics, including pump speed and aortic valve opening. In large series with the HVAD,6,7 gender has not been identified as a risk factor for stroke or pump thrombosis, nor was it associated with pump thrombosis in our analysis (see Supplemental Material, Supplemental Digital Content 1, Finally, we elected to include all hematologic complications occurring in the Lavare On group, regardless of whether they occurred while the cycle was activated, or after it had been deactivated. The rationale for this is that we do not know how for how long a Lavare-induced disruption of an intraventricular thrombus or inflow cannula pannus might persist. A potential small disturbance could be exacerbated by a change in coagulation status or hemodynamics, precipitating a late event.

These data, despite their limitations, may provide some guidance for those deciding whether or not to use Lavare. We saw no benefit, and, at least in terms of pump thrombosis, a signal of harm, in patients who had it activated later in the postoperative period. Our data, along with the long-term European experience, suggest that a randomized controlled trial may be useful to determine whether use of the Lavare cycle is beneficial to patients.


1. Kormos RL, Cowger J, Pagani FD, et al. The Society of Thoracic Surgeons Intermacs database annual report: Evolving indications, outcomes, and scientific partnerships. J Heart Lung Transplant 2019.38: 114–126.
2. Wong K, Samaroo G, Ling I, et al. Intraventricular flow patterns and stasis in the LVAD-assisted heart. J Biomech 2014.47: 1485–1494.
3. Heartware Inc.: Heartware HVAD System Instructions for Use. Document IFU00375 Rev05 03/18, 2018.
4. Zimpfer D, Strueber M, Aigner P, et al. Evaluation of the heartWare ventricular assist device Lavare cycle in a particle image velocimetry model and in clinical practice. Eur J CardioThorac Surg 2016.50: 839–848.
5. Glass CH, Christakis A, Fishbein GA, et al. Thrombus on the inflow cannula of the HeartWare HVAD: An update. Cardiovasc Pathol 2019.38: 14–20.
6. Najjar SS, Slaughter MS, Pagani FD, et al. An analysis of pump thrombus events in patients in the HeartWare ADVANCE bridge to transplant and continued access protocol trial. J Heart Lung Transpl 2014.33: 23–34.
7. Teuteberg JJ, Slaughter MS, Rogers JG, et al.; ADVANCE Trial Investigators: The HVAD left ventricular assist device: Risk factors for neurological events and risk mitigation strategies. JACC Heart Fail 2015.3: 818–828.

LVAD; thrombosis; stroke; complications

Supplemental Digital Content

Copyright © 2020 by the American Society for Artificial Internal Organs