Outcome of mechanical circulatory support at the University Medical Centre Utrecht

Data of all patients in whom the HM-II, HM 3 or the HVAD was implanted at the University Medical Centre Utrecht between March 2006 and January 2018 (end of study) were collected in a central database. The database included baseline clinical characteristics and all adverse events defined according to the INTERMACS definitions [15]. Adverse events were described for the total population. The institutional ethics board approved the study.

cf-LVADs were implanted via a median sternotomy using extracorporeal circulation on the beating heart. In the absence of bleeding, heparin was started within 48 h after surgery if drainage during 3 consecutive hours post-implant did not exceed 50 ml/h. A vitamin K antagonist was started after drain removal and heparin was stopped when the INR reached the lower limit of the therapeutic range as described for each cf-LVAD type below.

A thrombocyte aggregation inhibitor was started after 48 h, in general aspirin 100 mg/day. From March 2006 until August 2009, in HM-II patients warfarin was titrated to an INR of 2–2.5, 7 days after implantation. The INR range was reduced to 1.5–2.0 due to a substantial incidence of bleeding, as reported in the literature [16]. From December 2011 the INR range was increased again to 1.8–2.5 because of increased thrombo-embolic complications. For the HVAD and HM 3, target INR was 2.5–3.5 and 2.0–3.0, respectively, according to the manufacturer’s advice.

The primary outcome of the study was survival on cf-LVAD support until a pre-specified end-point, i.e. death, device explantation, heart transplantation or the end of the study. Secondary outcomes included all adverse events defined according to INTERMACS.

All adverse events were defined according to the INTERMACS definitions. Major bleeding was defined as suspected internal or external bleeding, resulting in death, re-operation, hospitalisation and/or transfusion of red blood cells (within the first 7 days after the implantation requiring transfusion ≥4 units of packed red blood cells, or any transfusion beyond 7 days postoperatively). Neurological complications included a transient ischaemic attack and ischaemic/haemorrhagic strokes. Major infection was defined as a clinical infection accompanied by pain, fever, drainage and/or leukocytosis, treated by antimicrobial agents (non-prophylactic).

Major haemolysis included biochemical signs of haemolysis (free plasma haemoglobin >200 mg/l or lactate dehydrogenase >625 U/l), accompanied by at least one of the following symptoms: haemoglobinuria, anaemia, hyperbilirubinaemia and/or pump malfunction. Minor haemolysis comprised asymptomatic biochemical abnormalities.

Major device malfunction included pump thrombosis, high-urgency transplantation, pump replacement, pump explantation, breach of driveline or death. Minor device malfunction included inadequately functioning external components which required repair or replacement. Right heart failure (RHF) was defined as symptoms and signs of persistent right ventricular dysfunction (central venous pressure >18 mm Hg with a cardiac index <2.3 l/min/m² in the absence of elevated left atrial/pulmonary capillary wedge pressure (>18 mm Hg), tamponade, ventricular arrhythmias or pneumothorax) requiring implantation of a right ventricular assist device, inhaled nitric oxide or inotropic therapy for >1 week at any time after cf-LVAD implantation.

From March 2006 until January 2018, 268 patients underwent cf-LVAD implantation (69% male, mean age 50 (±13) years). In 59% of patients HM-II was implanted, 98.5% of devices as a bridge to transplantation, the remaining 1.5% as destination therapy. Follow-up was completed for all 268 patients for a median period of 542 (interquartile range (IQR): 205–1044) days, resulting in a total experience of 510 patient years (mainly determined by 380 patient years for HM-II).

The clinical profile before cf-LVAD implantation was most often INTERMACS 2 (42%) or 3 (27%), implying a progressive decline on inotropic support and stable but inotrope dependent, respectively. Furthermore, 19% of the patients were on temporary MCS prior to cf-LVAD implantation, mostly by central or peripheral extracorporeal life support, so were originally INTERMACS 1 but stabilised on temporary MCS. Baseline characteristics for the total cohort and per device type are summarised in Tab. 1.

The INTERMACS profile prior to the cf-LVAD implantation has changed over time. Since 2013, patients in higher INTERMACS profiles received implants more frequently than in the first few years. Statistical analysis was performed to compare survival of patients receiving implants between 2006 and 2012 with those between 2013 and 2017. Patients in whom a cf-LVAD was implanted between 2006 and 2012 were significantly younger (p < 0.001) and more frequently in INTERMACS profile 2 (p = 0.008) than those receiving implants between 2013 and 2017 (Electronic Supplementary Material, Table 1).

Seventy-one (26%) patients (44 HM-II, 24 HVAD and 3 HM 3) died after cf-LVAD implantation after a median of 216 (IQR: 20-807) days. Death was most often caused by neurological complications (22 patients) or infections (14 patients) (Tab. 2). Device malfunction was the cause of death in 7 patients, pump thrombosis in 5 cases and technical failure in 2. Eighty-two (31%) patients underwent a heart transplantation, after a median duration of 674 (IQR: 394–1028) days on cf-LVAD-support. Explantation of the cf-LVAD was possible in 8 (3%) patients after a median period of 529 (IQR: 351–670) days. In these 8 patients, myocarditis and peri-partum cardiomyopathy were the most common aetiologies. In one patient with dilated cardiomyopathy in whom the cf-LVAD was initially explanted following the recovery of left ventricular function, a new device had to be implanted after 144 days due to recurrent HF.

Survival after 1, 3 and 5 years was 83%, 72% and 57%, respectively (Fig. 1). There was a trend towards worse survival in patients with INTERMACS profile 1 in comparison to INTERMACS profile 2 or 3, though not significantly different (Fig. 2, p = 0.24). Neither did survival differ significantly between implants in 2006–2012 in comparison to implants in 2013–2017 (p = 0.44). Because patients receiving implants between 2006 and 2012 were significantly younger and more frequently in INTERMACS 2 in comparison to 2013–2017, correlation between survival and these variables was analysed. Both age and INTERMACS profile 2 were not associated with mortality in this cohort (hazard ratio (HR) 0.98 (95% confidence interval, CI, 0.96–1.01), p = 0.123 and HR 0.91 (95% CI 0.55–1.50), p = 0.702, respectively).

Beside localised infections not specifically related to the MCS, such as urinary tract infections and pneumonias, the three most commonly encountered adverse events were major bleeding, ventricular tachycardia and minor haemolysis with corresponding event rates of 0.51, 0.35 and 0.26 per patient year, respectively, as shown in Tab. 3.

Strokes (haemorrhagic and/or ischaemic) occurred 0.15 times per patient year. RHF occurred 0.23 times per patient year, most often (65%) within the first month after implantation. In 29 patients, RHF developed beyond 30 days after implantation, of whom 8 (28%) also suffered from early RHF.

Previously, only a few smaller single-centre studies were performed regarding long-term results of cf-LVAD support. Takeda et al. presented their results in 140 patients, showing a survival rate of 83%, 75% and 61% after 1, 3 and 5 years, respectively [17]. We now confirmed these results in a larger population. In the most recent annual INTERMACS report, survival rates at 1, 3 and 5 years were 83%, 63% and 46%, respectively [18]. With regard to the pre-operative condition, it is known that patients in INTERMACS profiles 1–3 have worse survival rates, especially INTERMACS profile 1 [15, 18]. Our study confirmed the relationship between the initial poor state and the trend towards worse survival of patients in INTERMACS profile 1, in comparison to INTERMACS profile 2 or 3, despite prior stabilisation on short-term MCS, although this was not statistically significant.

Generally, in MCS patient selection is of utmost importance for the outcome. Stewart et al. studied the use of the INTERMACS classification to identify ambulatory patients with advanced HF who may benefit from a cf-LVAD. In that study, patients in INTERMACS profile 4 had a higher mortality rate and needed MCS more often compared to patients in INTERMACS profile 5–7 [19]. In addition, the ROADMAP trial concluded that patients in INTERMACS 4 have better survival, functional capacity and improved quality of life when treated with a cf-LVAD in comparison to optimal medical management [20].

Furthermore, prediction of RHF is important, because this is related to worse survival. Recently, the EUROMACS-RHF risk score was developed, which can be used to predict early RHF [21]. Unfortunately, little is known about risk factors for late RHF, which needs further research in the setting of long-term MCS.

Technical improvements in the HM 3, using a magnetically levitating environment, revealed fewer haemocompatibility-related adverse events (e.g. pump thrombosis) in comparison to HM-II at 2 years, as concluded in the MOMENTUM 3 trial. However, the rate of bleeding events was comparable in both groups [22‐24]. A personalised anticoagulation regimen could decrease the individual risk for bleeding and thrombosis. Furthermore, the risk for infection could be decreased by the use of cf-LVAD with smaller (or no) external components including the driveline, which is the most frequently encountered location for VAD-related infections.

This is the largest single-centre study reporting on 5‑year outcome in cf-LVAD patients in whom the device was implanted as a bridge to transplantation, reflecting the Dutch results of long-term MCS. Complications were recorded prospectively and systematically in a central database. Patient follow-up was complete in our own centre, minimising the risk of missing data.

However, the single-centre design may indicate that our results cannot be extrapolated directly to other centres. Furthermore, in nearly all patients the cf-LVAD was implanted as a bridge to transplantation. In general these patients appear to have a more favourable outcome in comparison to those receiving this device as destination therapy [15, 18, 25]. Finally it has to be realised that our results are mainly based on the HM-II LVAD, as almost 60% of patients received this device.