The gastrointestinal bleeding (GIB) is one of the most common complications and commonest cause of readmission after implantation of long-term ventricular assist devices (VAD) ranging between 5 and 30%.1,2 It has been reported that the incidence of the GIB might be reduced with intermittent opening of the aortic valve (AV) or higher arterial pulsatility.3,4 The European version of the HeartWare VAD (HeartWare Inc., Framingham, MA) has an intermittent low speed software (lavare cycle) that promote ventricular/aortic root washout and may facilitate intermittent AV opening.5 This lavare cycle is activated at our institution in all of our HeartWare patients roughly 2 weeks after VAD implantation. The aim of this study was to investigate the incidence of the GIB events in HeartWare patients with activated lavare cycle. We hypothesize that lavare cycle activation may reduce the incidence of GIB in HeartWare patients.
The study protocol was approved by the local Research Ethics Board. Data of patients receiving HeartWare VAD at our institution between January 2010 and August 2016 were retrospectively analyzed. A GIB event was defined as any documented bleeding event from upper or lower gastrointestinal tract requiring blood transfusion or hospital admission.
A total of 105 patients with mean age of 56 ± 12 years were supported with HeartWare VAD for a median duration of 216 days (IQR: 47–478 days). Thirty-two GIB events were documented in 22 patients (21%) with mean age of 58 ± 11 years after median duration of 22 days (IQR: 11–157 days) accounting for 0.34 events per patient’s years (ppy). Fifteen patients (68%) had 1 event, 6 patients (27%) had 2 events, and 1 patient (5%) had 5 events. Table 1 summarizes the pre- and postoperative characteristics of patients with or without GIB. Multivariate logistic regression analysis demonstrated preoperative veno-arterial extracorporeal membrane oxygenation (VA-ECMO) (OR, 3.61; 95% CI, 1.16–12.22, p = 0.027) and history of peripheral vascular disease (OR, 5.63; 95% CI, 1.13–27.99, p = 0.034) as independent predictors for any postoperative GIB event.
Interestingly, we observed that 13 out of 22 patients developed their first GIB event at the time of initial hospitalization. Table 2 shows the main differences between patients who had their first GIB at the time of first hospitalization (early bleeders) compared with patient who were readmitted for GIB event after discharge from hospital (late bleeders). Notably, early bleeders were sicker patient population with higher incidence of preoperative VA-ECMO.
As the causes of GIB in the early bleeders were not typical “VAD-associated” lesions and the possible effect of lavare cycle is not to be expected early after left ventricular assist device (LVAD) implantation, we thought to analyze the data of the late bleeders and compare these “late bleeders” cohort with the patient population, who were supported with HeartWare VAD and included in the pivotal bridge to transplant (BTT) trial and under the continuous access protocol (CAP).7 Patients included in that study were supported with HeartWare VAD in United States, where the lavare cycle is not approved. In our study, a total of 17 GIB events in 9 patients (8.6%) (late bleeders group) was observed after a median support duration of 174 days, accounting for 0.18 events ppy. This rate is comparable with the GIB events reported in the pivotal BTT trial and under the CAP (with deactivated lavare cycle) of 15.4% and 0.27 events ppy.
Considering the association of GIB events to opening status of the AV; in the late bleeders cohort (n = 9), 67% had continuous opening of the AV, and 33% had intermittent opening of the AV. Meanwhile, 68% of patients without GIB (n = 83) had documented continuous opening of the AV, 19% had intermittent opening of the AV, and 13% had continuous closure of the AV. The average pump flow, index, and speed in patients with GIB versus those without GIB were comparable.
Although the GIB events significantly increased the rate of readmission after LVAD implantation, these GIB events did not negatively impact the survival rate after assist device support (log-rank test, p = 0.849).
This study shows that GIB is a frequent complication in patients with HeartWare VAD regardless of the activation of lavare cycle. Further, a significant difference in patient characteristics was observed between early and late bleeders. Early bleeders are sicker patients with higher rate of preoperative VA-ECMO and postoperative right VAD support.
Notably, the GIB rate in our study (0.18 events ppy) is comparable with the GIB events reported in the pivotal BTT trial and under the CAP (with deactivated lavare cycle) of 15.4% and 0.27 events ppy. However, the study cohorts were significantly different; there were more male patients with GIB in our cohort (p = 0.016). Further, our patient population with GIB had higher rate of ischemic cardiomyopathy (p ≤ 0.001), lower incidence of diabetes mellitus (p < 0.001) and more importantly lower preoperative right ventricular function (p <0.001) as well as significantly lower INTERMACS profile compared with Goldstein study (p <0.001). The association of severe right ventricular dysfunction with the postoperative GIB events has been recently described.6 The negative impact of the preoperative VA-ECMO on survival and postoperative morbidity after VAD implantation has been also reported from our group previously.7–9
In this study, no association between opening status of the AV and GIB events was observed. This finding suggests that the activation of lavare cycle may induce slight opening of the AV but not to an extend allowing adequate arterial pulsatility. In recent years, a concept of improving pulsatility status by reducing pump speed has evolved as a prevention strategy.10 It is unclear whether the positive effect of reducing pump speed is because of reducing shear stress and von Willebrand Factor (vWF) degradation or increased arterial pulsatility. The finding of our study further confirms that the speculative benefits of augmented pulsatility with continuous flow pumps might be overrated and are still incompletely evaluated.
Limitations of the study include retrospective nature of the study. Further, as our study aims to demonstrate the potential benefits of lavare cycle, we believe that a stronger conclusion can be drawn if we would have compared two propensity matched group of patients with or without activation of lavare cycle. We compared our data with Goldstein study, which included HeartWare patients without activated lavare cycle. However, our patients were in general sicker and the Goldstein study was performed in several centers with different anticoagulation protocols.
In conclusion, the GIB events appear to be a frequent complication in patients with HeartWare VAD regardless of the lavare cycle. The opening status of the AV does not seem to correlate with the GIB events.
1. Crow S, John R, Boyle A, et al. Gastrointestinal bleeding
rates in recipients of nonpulsatile and pulsatile left ventricular assist devices. J Thorac Cardiovasc Surg 2009.137: 208–215.
2. Goldstein DJ, Aaronson KD, Tatooles AJ, et al. Gastrointestinal bleeding
in recipients of the HeartWare ventricular assist system. JACC Heart Fail 2015.3: 303–313.
3. Ising MS, Sobieski MA, Slaughter MS, Koenig SC, Giridharan GA. Feasibility of pump speed modulation for restoring vascular pulsatility with rotary blood pumps. ASAIO J 2015.61: 526–532.
4. Stern DR, Kazam J, Edwards P, et al. Increased incidence of gastrointestinal bleeding
following implantation of the HeartMate II LVAD. J Card Surg 2010.25: 352–356.
5. Saeed D, Westenfeld R, Maxhera B, et al. Prevalence of de novo aortic valve insufficiency in patients after HeartWare VAD implantation with an intermittent low-speed algorithm. ASAIO J 2016.62: 565–570.
6. Sparrow CT, Nassif ME, Raymer DS, Novak E, LaRue SJ, Schilling JD. Pre-operative right ventricular dysfunction is associated with gastrointestinal bleeding
in patients supported with continuous-flow left ventricular assist devices. JACC Heart Fail 2015.3: 956–964.
7. Aubin H, Petrov G, Dalyanoglu H, et al. A suprainstitutional network for remote extracorporeal life support: A retrospective cohort study. JACC Heart Fail 2016.4: 698–708.
8. Maxhera B, Albert A, Ansari E, Godehardt E, Lichtenberg A, Saeed D. Survival predictors in ventricular assist device patients with prior extracorporeal life support: Selecting appropriate candidates. Artif Organs 2014.38: 727–732.
9. Abdeen MS, Albert A, Maxhera B, et al. Implanting permanent left ventricular assist devices in patients on veno-arterial extracorporeal membrane oxygenation support: Do we really need a cardiopulmonary bypass machine? Eur J Cardiothorac Surg 2016.50: 542–547.
10. Bartoli CR, Atluri P. Do patients with a continuous-flow left ventricular assist device benefit from induced-pulsatility or are we just spinning our wheels? J Thorac Cardiovasc Surg 2015.150: 945–946.