Five-year results of patients supported by HeartMate II: outcomes and adverse events

Abstract

OBJECTIVES

Improved outcomes over the past decade have increased confidence of physicians and patients in extended duration of left ventricular assist device (LVAD) support. This single-centre cohort study reports 5-year outcomes with the HeartMate II (HMII) LVAD.

METHODS

We describe a cohort of 89 patients who received a HMII LVAD between February 2004 and December 2010. The causes of death and adverse events were assessed by examination of medical records. A total of 202.74 patient-years were analysed.

RESULTS

After 5 years, of the 89 patients, 15 patients remained on device therapy, 39 patients died, 28 patients underwent heart transplantation and 7 patients underwent explantation of the HMII for recovery. One year after the HMII implantation, there was a survival of 71% in the study cohort. In the following years, the survival rate was 65% in the 2nd year, 63% in the 3rd year, 56% in the 4th year and 54% after 5 years of LVAD support. Ten LVAD exchanges were performed in 8 (11%) patients. Currently (March 2017), 12 patients still remain on their original device. The longest ongoing patient on the HMII has been supported for over 11 years (4097 days). The most common adverse events were bleeding (68%; 1.5837 events per patient-year) and LVAD infection [49%; 1.0666 events per patient-year]. Seven cases of pump thrombosis were described (8%; 0.1131 events per patient-year).

CONCLUSIONS

This is the first single-cohort study to describe a 5-year survival of HMII patients on extended duration of support. A 5-year survival of 54% was observed in this single-centre cohort.

INTRODUCTION

Mechanical circulatory support (MCS) has become a widely accepted therapeutic option in medical therapy refractory patients. The biggest advantage of MCS compared with cardiac transplantation is the readily available option of ‘off-the-shelf’ ventricular assist device (VAD) implantation without delaying haemodynamic support, which is particularly important in high urgency and critically ill patients [1]. Also, devoid of the biological limitation of an exceedingly scarce resource, the indications for MCS therapy have become more liberal over time and patients ineligible for heart transplantation, e.g. because of age or comorbidities, are now considered potential candidates for MCS. This evolution has resulted in the number of assist devices implanted worldwide exceeding the number of heart transplantations performed [1–3].

Despite these medical successes, the devices still have several major inherent limitations, such as risk for infection, stroke, pump thrombosis, haemorrhagic complications and right ventricular failure, not to mention major lifestyle changes for patients and caregivers, which need to be carefully considered prior to implantation [4–11]. Therefore, heart transplantation remains the gold standard for eligible patients with end-stage heart failure.

The HeartMate II (HMII) (St. Jude Medical Corporation, St. Paul, MN, USA) has been the most frequent commercially implanted left ventricular assist device (LVAD) with over 22.000 patients worldwide. Safer long-term support has been achieved increasing physician experience, improved medical management and reductions in major complication rates. Eligible patients for transplantation are able to remain waitlisted for extended durations, improving end-organ function and physical capacity [11–13]. This study reports our single-centre experience of 5-year outcomes with the HMII LVAD.

METHODS

Study organization

This study was conducted at a single, high-volume VAD centre, was investigator initiated and was not supported by industry or any grant funding. The authors of this study vouch for the completeness and accuracy of the data and analyses. The data collection and analysis were approved by the institutional review board.

Study patients

Between February 2004 and December 2015, more than 500 LVADs were implanted at our institution. Of these, we retrospectively studied a cohort of 89 patients who received an HMII assist device at a single institution between February 2004 and December 2010. The first 8 patients of the study cohort were part of the initial HMII Conformité Européenne (CE) mark trial. All patients underwent LVAD implantation via conventional full sternotomy. Biventricular assist devices and other types of assist devices were excluded from the study.

Study design

This study is a retrospective, single centre study. Data were collected by electronic medical record review. The end-points of the study were death, device explantation or heart transplantation. The secondary end-point was 5 years on device.

For the performance of this retrospective study, the digital database of a single centre was used. The baseline characteristics were obtained for all patients prior to HMII implantation. All causes of death and adverse events were determined through the examination of medical records. After the HMII implantation, survival results and adverse events were recorded. All causes of death and adverse events were evaluated by examination of medical records.

Statistical methods

All statistical analyses were performed using Software SAS 9.3 as well as SAS Enterprise Guide 4.3 and SPSS 22 (IBM^® SPSS^® Statistics, version 22). Categorical variables were reported in frequencies, and continuous variables were reported as mean ± standard deviation or as median with range, if applicable.

RESULTS

Baseline characteristics

Between February 2004 and December 2015, more than 500 patients were implanted with a mechanical assist device at Hannover Medical School. Of these patients, 89 patients received an HMII device between February 2004 and December 2010. Other implanted LVADs [e.g. HeartMate 3 (HM3); HeartWare left ventricular assist device (HVAD), HeartWare Inc., Framingham, MA, USA; MVAD, HeartWare; HeartAssist 5 (HA5), ReliantHeart; and other types] and biventricular assist devices were excluded from the study. Detailed baseline characteristics of the study cohort are presented in Table 1. A total of 202.74 patient-years (74 001 days) were analysed.

The majority of patients were INTERMACS Level I (55%) and Level II (30%). Preoperatively, 12 patients were supported by extracorporeal membrane oxygenation and 15 patients were supported by intra-aortic balloon pump. Seven patients required ongoing extracorporeal membrane oxygenation support after LVAD implantation because of initial postoperative right ventricular failure. Ten patients were classified as bridge-to-destination candidates, whereas 79 patients were grouped in the bridge-to-transplantation category. The average predischarge haemolysis values show a lactate dehydrogenase level of 1154.62 U/l and a free haemoglobin level of 70.26 mg/l.

Five-year outcomes of the cohort

The results of the survival analysis are displayed in Fig. 1. In the first 3 months after LVAD implantation, the Kaplan–Meier analysis shows a steep decrease. One year after the HMII implantation, there is a survival of 71% in the study cohort. In the following years, the survival rate is 65% in the 2nd year, 63% in the 3rd year, 56% in the 4th year and 54% after 5 years of LVAD support.

After 5 years, of the total of 89 patients, 15 patients remained on LVAD support, 28 patients underwent heart transplantation and 7 patients underwent HMII explantation for recovery of systolic function (Table 2). Today, there are 12 patients still on their original device. As of 1 February 2017, the longest supported patient has remained on the device for over 11 years (4097 days). The mean support of the patients who are still on the device is 2977 days (range 2243–4097 days).

Causes of death

Survival rate at 5 years was 54%; however, of the 89 patients in the study group, 39 (44%) patients died during the study period. Of these 39 patients, 11 (28%) patients died due to septic multisystem organ failure. Eight patients died because of circulatory failure (20%), 3 because of right heart failure (8%) and 3 because of respiratory failure (8%). Four patients committed suicide in the study group (5%). One patient died via strangulation. The others disconnected the batteries of the VAD. Two (5%) patients died of haemorrhagic stroke, and 1 (3%) patient died due to bleeding complications (Table 3).

Complications

The documented complications are listed in Tables 4 and 5. The most frequent complication was LVAD-associated infection in 46 patients [49%; 0.2269 events per patient-year (ppy)] with 66 events. This includes driveline as well as combined driveline and systemic device infections.

Right heart failure was documented in 8 patients during the 1st month after LVAD implantation (10%). In total, 10 (11%) patients experienced right heart failure. Eight patients experienced it in the 1st month after LVAD implantation, 1 patient 2 years after implantation and 1 patient was diagnosed with right heart failure 3 years after implantation.

Fifty-five patients (68%; 1.5837 events ppy) had bleeding complications such as postoperative bleeding, epistaxis or gastrointestinal bleeding. Of these, 36 patients required surgery and 14 patients had a gastrointestinal bleeding. Sixty-five (81%) patients were diagnosed with acquired von-Willebrand syndrome.

Twelve patients suffered a stroke (14%; 0.1939 events ppy). The timing of stroke was 563 days after surgery (mean), maximum 1605 days and minimum 3 days after LVAD implantation. Of the observed 12 strokes, 6 were classified as ischaemic and 6 as haemorrhagic strokes.

Pump thrombosis was observed in 7 patients (8%; 0.1131 events ppy). One (1%) patient was successfully treated via medical thrombolysis. Six (7%) patients had a pump thrombosis, which was treated by LVAD exchange.

A total of 10 LVAD exchanges were performed in the study group (11%). Of this group, 2 patients underwent 2 consecutive LVAD exchanges. A total of 6 exchanges were caused by pump thrombosis, 3 by infection and 3 by device malfunction (2 driveline damages and 1 battery switch). The mean timing of the procedure was 1002 days after LVAD implantation (minimum 28 days and maximum 2467 days after LVAD implantation).

Twenty (23%) technical failures were observed in the study group. Three device malfunctions were treated by LVAD exchange.

DISCUSSION

This long-term single-cohort study describes the survival of a large HMII study group on extended duration of support. Our group described a case series of successful long-term LVAD therapy in 4 HMII patients in a previous publication, which demonstrated that mechanical support beyond 5 years is feasible [11]. We now present the results of LVAD therapy with the HMII after implantation in our centre from 2004 to 2010.

Slaughter et al. [2] reported a 2-year survival of a study group with continuous-flow LVAD of 62% in 2009. In this study, we observed a 2-year survival of 65%. Medical treatment of terminal heart failure showed a survival of less than 10% after 2 years [10].

The post-approval study for destination therapy published by Jorde et al. [10] followed the first 247 HMII patients for 2 years. The survival of this cohort was 62% after 2 years of LVAD support, which is comparable to our study cohort. A recent publication by Kirklin et al. [4], which analyses INTERMACS data of 8988 patients, reports survival rates of 80% at 1 year and 69% at 2 years after implantation. Tsiouris et al. [12] describe outcomes from a mixed group of HMII and HVAD patients. The authors report a survival rate for their LVAD patients at 30 days, 6 months, 12 months, 2 years, 3 years and 4 years of 94%, 86%, 78%, 71%, 62% and 45%, respectively. Takeda et al. [13] report an estimated on-device survival at 1, 3 and 5 years of 83%, 75% and 61%, respectively, for a cohort supported by 4 different kinds of assist devices including the HMII.

Pump thrombosis has been extensively discussed in the past few years [4, 5, 14, 15]. Our 5-year cohort showed a relatively low incidence of pump thrombosis [7.5% (0.0296 events ppy)]. Kirklin et al. [4] report that pump exchange or death due to pump thrombosis increased during 2011 and 2012, but the absolute increase remained small. A risk factor analysis suggests that a number of patient-related factors contribute to the risk of thrombosis. Smedira et al. [14] analysed a total of 995 thrombosed pumps and also reported an increase in pump thrombosis in 2010, which reached a maximum in 2012, and then plateaued at a level that was reportedly 3 times higher than pre-2010. Anticoagulation management is a major key to prevention of pump thrombosis. Yet too strict regiments increase the risk for bleeding complications such as gastrointestinal bleeding or stroke. The TRACE study questioned the need for double anticoagulation in HMII patients to reduce the risk of pump thrombosis. The 1-year results of this study revealed that reducing antithrombotic therapies in response to bleeding among HMII patients was achievable but may be associated with a higher risk for device thrombosis [16]. The recently published 2-year results from the European TRACE study showed that managing HMII patients with a vitamin K antagonist with a target international normalized ratio of 2.3 without antiplatelet therapy may help to reduce the incidence of major bleeding without increasing the risk for thromboembolic events, including ischaemic stroke and pump thrombosis [17].

Driveline and device infection were the most common adverse events in our study group (0.2268 events ppy), with sepsis and multisystem organ failure as the most common causes of death. Jorde et al. [18] also showed 0.22 events ppy for device-related infection. Similar results consistent with previous studies emphasize once more the essential nature of adherence with driveline hygiene in VAD therapy [18, 19]. A recent study suggests that obese patients have a higher risk for developing driveline infections. Nutrition management should therefore be recommended to all VAD patients [20].

Four (4.5%) patients committed suicide in our study group. One patient died via strangulation and the other 3 patients disconnected the batteries of the VAD. Several studies investigated depression before and after VAD implantation. Most studies show that depression improves in the 1st year after surgery [21–23]. Nevertheless, Eshelman et al. [23] state that psychiatric symptoms are widely underdiagnosed and undertreated in the LVAD patient population, and LVAD patients show a trend for quality of life and psychological functioning to be poorer than for transplanted and explanted patients. Postoperative psychological supervision should therefore be considered for patients at risk [21–23].

The current trend towards miniaturization and less invasive procedures has led to the development of smaller LVADs, e.g. HVAD (HeartWare Inc.) or the HM3 (St. Jude Medical Corporation) [1, 24, 25]. The HM3 is a novel compact LVAD with a fully magnetically levitated pump rotor without mechanical or hydrodynamic bearings [26]. It has obtained its CE Mark approval in 2015 and has been commercially available in Europe since October 2015. This next-generation LVAD design showed excellent survival and a significant reduction of adverse events in its CE Mark study [27]. Compared with its predecessor, the HM3 offers a smaller design, improved software features, such as artificial pulsatility, as well as an improved ‘click-in’ mechanism to facilitate easier surgical implantation [26–28]. However, clinical trials are still needed to compare the long-term outcomes between the 2 devices. The MOMENTUM trial compared the HMII and HM3 in a randomized, prospective study. The 6-month survival was 86.2% for HM3 vs 76.8% for HMII. No significant differences between the 2 groups in the rates of death or disabling stroke were observed. The authors attribute the difference between the pumps to the low rate of reoperation, e.g. for pump thrombosis for the HM3 [28].

The ENDURANCE trial compared the HMII in a 2:1 randomized study to the HVAD with a primary end-point of 2-year survival. The 2-year survival analysis showed a survival of 55.4% for HVAD and 59.1% for HMII [29]. Our patient cohort showed a survival of 65.2% at 2 years. The authors found the HVAD device non-inferior to the HMII with respect to stroke rate and device replacement. However, the HMII showed a higher number of device malfunctions and a higher need for surgical intervention [29].

In summary, the new generation of VADs is promising, but long-term results are still missing, and further clinical trials should be performed to prove the inferiority of these new devices.

Limitations

This study has some inherent limitations. The data are retrospectively collected and analysed and thus potentially affected by limitations associated with retrospective studies. Furthermore, limitations apply because of the single-centre characteristic of this study. The results of surgical studies are prone to learning curves and single centre’s specific characteristics. Moreover, the study period began several years ago. Therapeutic strategies and increased clinical experience might have improved contemporary outcomes.

Additionally, the number of patients on the device for more than 5 years was small, which reduces the statistical power. As such, larger studies are required to further study VAD patients on extended support duration. Furthermore, this study is limited to a single LVAD. Larger, randomized controlled studies should be performed to shed light on the comparative efficacy of contemporary devices.

CONCLUSION

This is the first study to describe the long-term survival of HMII patients in a single-centre cohort. It is possible to attain a 5-year survival of 54%, which one would expect to increase with a contemporary strategy.

Funding

This work was supported by the German Research Foundation [Deutsche Forschungsgemeinschaft (DFG)] [KFO 311] (Principal Investigators: J. Bauersachs, M. Hoeper and J.D. Schmitto).

Conflict of interest: Jan D. Schmitto received consultation fees as well as grant support from St. Jude Medical Corporation. Jan D. Schmitto, Günes Dogan and Sebastian V. Rojas are consultants for St. Jude Medical Corporation as well as HeartWare Corporation. The other authors have no disclosures.

REFERENCES

Author notes