Advanced heart failure in critical patients (INTERMACS 1 and 2 levels): ventricular assist devices or emergency transplantation?

Abstract

OBJECTIVE

For patients in advanced heart failure, emergency transplantation or ventricular assist devices (VADs) are possible strategies. The aim of this single-centre, retrospective study was to evaluate early and long-term results for these two strategies.

METHODS

From 2005 to 2011, we analysed 49 INTERMACS level 1 and 2 patients, who were divided into the following two groups: group A comprised 26 patients on the waiting list for heart transplantation with urgent conditions; and group B comprised 23 patients who underwent VAD implantation as a bridge to candidacy.

RESULTS

In group A, 25 patients underwent transplantation. In group B, 19 patients were supported with left VAD and four with biventricular VAD. Of these 23 patients, 13 underwent transplantation (mean time 279 ± 196 days). The 30 day mortality was 42.3 and 4.3% in group A and B, respectively. Survival at 6 and 12 months was significantly better in group B than in group A (87 vs 53%, P = 0.018 at 6 months; and 77 vs 48%, P = 0.045 at 12 months).

CONCLUSION

Improved outcomes may justify the use of mechanical assistance devices as a bridge to candidacy or bridge to transplantation in INTERMACS 1 and 2 patients in order to avoid high-risk transplants. Evaluation of long-term multicentre outcomes is needed to assess future strategies.

INTRODUCTION

Heart transplantation is the solution of choice for advanced heart failure unresponsive to medical treatment. In selected cases, it improves both survival and quality of life [1]. The prognosis of these patients may, however, be poor because of the small number of donors and the consequently long waiting list. In Italy, in 2005, the reduced number of heart transplants led the National Transplant Organization (CNT) to create a ranking of priorities among patients on the waiting list based on clinical conditions, as had been done in the USA (with United Network for Organ Sharing status) in 1989 and in Germany in 2000 [2–4]. At present, in Italy, in order to access the urgent transplant programme, patients must be on inotropic support, ventilated and with a mechanical circulatory support; alternatively, they have to be on biventricular assistance or with complicated long-term circulatory support [5]. A patient who matches these criteria is eligible the first compatible organ and is suitable for transplantation available in any Italian intesive care unit (ICU).

Clearly, the opportunity of being enrolled into the emergency programme gives a chance of survival to critically ill patients; however, the low number of transplants performed has led to a progressive increase of heart transplants (HTx) performed as urgent procedures in patients often in poor clinical condition. It is well known that urgent transplantation may be associated with slightly worse results compared with elective transplantation [6, 7] and with a higher incidence of severe complications, such as primary graft failure, need for intra-aortic balloon pump, more prolonged mechanical ventilation time and renal replacement therapy, as well as a greater number of bacterial infections and a significantly longer ICU stay [8–10].

The use of ventricular assist devices (VADs) as a bridge to transplantation (BTT) or to candidacy (BTC) has greatly influenced the management of emergency programmes. Many centres, rather than directly accessing the urgent programme, use VADs to stabilize critically ill patients and to provide more time to select the best treatment, especially for ‘unknown’ patients. Patients on VAD support are not included in the group of critically ill patients who need urgent HTx until the occurrence of device-related complications. In BTT patients, VADs improve organ function, optimizing the general condition of the patients at the time of transplantation. Whether these factors may have positive effects on post-transplant outcomes remains controversial [11–13].

In the present study, we analysed data from critically ill patients who were implanted with a VAD as a BTC and compared them with those of patients who underwent emergency transplantation.

PATIENTS AND METHODS

Between 2005 and 2011 at our institution, 55 patients with advanced heart failure [Interagency Registry for Mechanically Assisted Circulatory Support (INTERMACS)-level 1 or 2] were admitted in our ICU. Six of them were not included in this study because of severe neurological status (glasgow coma scale ≤3), irreversible renal or hepatic dysfunction (total bilirubin >7 mg/dl or creatinine >4 mg/dl) or sepsis (Fig. 1). The remaining 49 patients [36 (73.5%) were male] were considered for the study. The age of patients included in the study ranged from 14 to 70 years, with a mean of 51 ± 14 years.

Figure 1:

The flow chart shows the surgical treatments that were carried out on two groups of patients. Group A comprised 26 patients including in the waiting list for heart transplantation with urgent conditions. Group B comprised 23 patients who underwent VAD implantation as a bridge to candidacy with urgent conditions. BiVAD: biventricular assist device; BTR: bridge to recovery; BTT: bridge to transplantation; DT: destination therapy.

Patients were divided into two groups: group A comprised 26 (53%) patients who were included in the urgent heart transplant waiting list; and group B comprised 23 (47%) patients who underwent VAD implantation on an emergency basis (Fig. 1).

Preoperative characteristics of both groups are listed in Table 1. In group A, there was more severe dysfunction of the right ventricle, with a lower right ventricular stroke work index (5.20 vs 7.52 g m/m²/beat; P = 0.03) and higher pulmonary vascular resistance (2.99 vs 2.25 Wood units; P = 0.05). In patients on extracorporeal membrane oxygenation (ECMO), the right ventricular function was evaluated with a transoesophageal echocardiographic examination and Swan-Ganz catheter pressure monitoring, after reducing the ECMO assistance to a minimum of 0.5 l/min for 30 min. The serum creatinine level was significantly higher in group A (2.19 vs 1.46 mg/dl; P = 0.05).

Table 1:

Preoperative characteristics of patients

	Group A (urgent transplant waiting list)	Group B (urgent VAD implant)
Baseline characteristics	(mean ± SD or %)	(mean ± SD or %)	P value
Age (years)	52.6 ± 11.3	50.4 ± 14.0	0.10
Body mass index (kg/m2)	24.79 ± 4.44	24.84 ± 4.83	0.97
Male	(18/26) 69.2%	(19/23) 82.6%	0.30
INTERMACS level 1	(18/26) 69.2%	(12/23) 52.2%	0.24
Dialysis	(8/26) 30.8%	(4/23) 17.4%	0.30
Mechanical ventilation	(19/26) 73.1%	(14/23) 60.8%	0.38
Haemodynamic profiles
LVEF (%)	20.71 ± 9.05	15.61 ± 5.39	0.04
Cardiac output (l/min)a	3.36 ± 0.96	3.46 ± 0.74	0.71
Cardiac index (l/min/m2)a	1.90 ± 0.43	2.03 ± 0.40	0.31
Wedge pressure (mmHg)	18.63 ± 6.34	21.44 ± 5.41	0.13
SVO2	63.29 ± 12.60	61.22 ± 7.77	0.53
Pulmonary artery pressure (mmHg)
Systolic	38.11 ± 8.08	37.18 ± 7.89	0.70
Diastolic	22.11 ± 7.54	23.35 ± 5.62	0.55
Mean	28.26 ± 7.01	28.76 ± 5.91	0.80
Pulmonary vascular resistance (Wood's Units)	2.99 ± 1.31	2.25 ± 1.05	0.05
Transpulmonary gradient (mmHg)	9.36 ± 3.88	7.29 ± 3.53	0.07
Central venous pressure (mmHg)	14.43 ± 4.90	12.89 ± 3.69	0.26
RVSWI (g m/m2/beat)	5.20 ± 4.11	7.52 ± 2.01	0.03
Tricuspid regurgitation ≥ 3+	(8/26) 30.8%	(5/23) 22.5%	0.73
TAPSE	13.95 ± 3.15	13.72 ± 3.71	0.82
Haemodynamics supported with device
IABP	(10/26) 38.4%	(8/23) 34.8%	0.80
ECMO	(12/26) 46.1%	(5/23) 21.7%	0.08
Albumin serum (g/dl)	2.76 ± 0.42	2.87 ± 0.47	0.41
AST	448.73 ± 832.75	150.17 ± 220.81	0.14
ALT	395.77 ± 718.76	293.22 ± 525.64	0.6
Total bilirubin (mg/dl)	2.96 ± 3.17	1.73 ± 0.94	0.12
BUN (mg/dl)	106.00 ± 68.11	76.11 ± 38.80	0.10
Creatinine serum (mg/dl)	2.19 ± 1.48	1.46 ± 0.60	0.05
INR	1.73 ± 0.81	1.46 ± 0.33	0.19
LDH (mg/dl)	1979.64 ± 2206.64	974.83 ± 556.32	0.07
Haemoglobin (g/dl)	10.75 ± 1.83	10.74 ± 1.92	0.99
Haematocrit (%)	32.50 ± 5.69	33.21 ± 6.19	0.69
Leukocytes (×1000)/ml	12.90 ± 4.74	13.32 ± 3.98	0.76
Platelets (×1000/ml)	156.50 ± 119.59	160.72 ± 117.65	0.91

ALT: alanine aminotransferase; AST: aspartate aminotransferase; BUN: blood urea nitrogen; ECMO: extracorporeal membrane oxygenation; IABP: intra-aortic balloon pump; INR: international normalized ratio; LDH: lactate dehydrogenase; LVEF: left ventricular ejection fraction; RVSWI: right ventricular stroke work index; SVO2: oxygen venous mixed saturation; TAPSE: tricuspid annular plane systolic excursion; VAD: ventricular assist device.

^aNot significant in patients on ECMO.

Statistical analysis and limitations

This is a single-centre, retrospective, not blinded and not randomized study. Decision criteria used to assign patients to one or the other treatment (emergency transplantation or VAD implantation) were influenced by the growing experience in the centre regarding VAD implantation and management in critical patients in recent years. In the first phase of the study, patients were more frequently enrolled on the urgent transplant waiting list, while in more recent years they were treated preferentially with emergency long-term VAD implantation.

In every graph, 30 day mortality and in-hospital mortality were included in the overall mortality analysis.

RESULTS

Early results

In group A, one patient died before transplantation, while 25 patients underwent transplantation. The mean time on the waiting list for the patients who underwent transplantation was 2.73 ± 2.23 days (range 0.5–9 days). Mortality on the waiting list before receiving heart transplantation was 3.8%. No differences have been shown, in terms of characteristics of the donors, between transplants performed in patients on the urgent waiting list and transplants performed in patients on the ordinary waiting list (Table 2).

Table 2:

Baseline characteristics of donors

	Ordinary waiting list donors	Emergency donors
Baseline characteristics of donors	(mean value ± SD)	(mean value ± SD)	P value
Age (years)	39.97 ± 12.31	40.29 ± 14.22	0.91
Body mass index (kg/m2)	24.92 ± 3.88	25.73 ± 3.95	0.36
LVEF	63.55 ± 5.86	64.90 ± 6.07	0.32
Inotropic support
Noradrenaline (mcg/kg/min)	0.21 ± 0.21	0.15 ± 0.09	0.18
Dopamine (mcg/kg/min)	6.26 ± 2.67	5.00 ± 2.57	0.04
	[n (%)]	[n (%)]
Male	76/119 (64%)	15/23 (65%)	0.55
Cardiac arrest in donors	12/119 (10%)	5/23 (22%)	0.11
Cause of death
CVA	71/119 (60%)	13/23 (57%)	0.48
Trauma	43/119 (36%)	8/23 (35%)	0.55
Anoxia	4/119 (3%)	1/23 (4%)	0.59
CAD risk factors
Hypertension	27/119 (23%)	8/23 (35%)	0.17
Smoking	20/119 (17%)	5/23 (22%)	0.38

CAD: coronary artery disease; CVA: cerebral vascular accident; LVEF: left ventricular ejection fraction; mcg: micrograms.

In group B, four biventricular VADs (BiVADs; Excor Berlin Heart-Berlin, Germany) and 19 left ventricular assist devices (LVADs) (17 Incor Berlin Heart-Berlin, Germany; 2 HVAD Heartware) were implanted as a BTT or as a BTC.

In-hospital mortality was significantly higher in group A (42.3 vs 4.3%; P = 0.002). Considering only the group A patients who underwent transplantation, in-hospital mortality was 41.2%. Mean postoperative ICU stay (group A: 561.35 ± 641.15 h vs group B: 332.44 ± 276.97 h; P = 0.17) and mean time on mechanical ventilation (group A: 118.80 ± 107.96 h vs group B: 94.83 ± 110.55 h; P = 0.49) were not significantly higher in group A vs Group B.

Long-term results

The mean follow-up time was 21.7 ± 26.3 months (range 1 day to 6.4 years) for group A and 9.36 ± 3.6 months (range 16 days to 2.4 years) for group B.

Long-term survival for patients who underwent VAD implantation and were not censored for heart transplantation (23 patients) was 87% at 6 months, 77% at 12 months, 66% at 18 months and 55% at 24 months. Survival at 6 and 12 months of VAD-implanted patients (not censored for heart transplantation) was significantly higher than in patients who were on the urgent waiting list (26 patients; 53% at 6 months, P = 0.02; and 48% at 12 months, P = 0.04) and in patients on the urgent waiting list who underwent transplantation (25 of 26 patients; 55% at 6 months P = 0.02; and 50% at 12 months, P = 0.05). No significant difference was observed at 18 and 24 months (P = 0.14; P = 0.34 respectively; Fig. 2).

Figure 2:

Actuarial survival curves showed a significantly better survival at 6 and 12 months for group B than group A (P = 0.02 and P = 0.04, respectively). Patients in group B were not censored at the time of heart transplantation.

Kaplan–Meier survival curves showed a lower survival of group A at 6, 12 and 18 months with respect to group B censored at the time of transplantation. This difference was significant only at 6 months (P = 0.02). (Fig. 3).

Figure 3:

Kaplan–Meier survival curves showed a lower survival of group A at 6, 12 and 18 months with respect to group B (patients censored at the time of transplantation). This difference was significant only at 6 months (P = 0.02).

In our series, there was no recovery of the native heart function under VAD support.

Causes and incidences of hospitalization in both groups are summarized in Table 3. Morbidity and in-hospital re-admission were more frequent in group B. Main causes of hospitalization were: major driveline infections, VAD controller replacement and ischaemic or haemorrhagic neurological complications in group B; and infections, rejections and tumours in group A (Table 3).

Table 3:

Causes of hospitalization in both groups

Causes of hospitalization	Incidence (patients/year)
Group A
Heart failure	0.03
Infection	0.05
Colon cancer resection	0.05
PM implant	0.03
Angina	0.03
Rejection	0.05
LV study	0.05
Group B
Driveline infection	0.61
INR altered	0.47
VAD replacement parts	0.27
Minor CVA	0.14
Haemorrhagic stroke	0.14
Right heart failure	0.14
Haematemesis	0.14
Major depression	0.07
Bladder cancer resection	0.07
Anaemia	0.07
AICD implant	0.07
Arrhythmia	0.07
Pulmonary embolism post-transplant	0.07
Seizure disorder post-transplant	0.07

AICD: automatic implantable cardiac defibrillator; CAD: coronary artery disease; CVA: cerebral vascular accident; INR: international normalized ratio; LV: left ventricular; PM: pacemaker; VAD: ventricular assist device.

In group B, five patients were switched from BTC to destination therapy because of inelegibility for heart transplantation.

Thirteen (56.5%) patients supported by VAD underwent transplantation after a mean time of 9.1 ± 7.2 months. In five patients (38.5%), the transplant was done on an emergency basis for complications related to the assistance device (in three cases because of major local wound infection, in one case because ischaemic stroke occurred in a 16-year-old patient, and in one case because of left inflow cannula malpositioning). Four of these five (80%) patients were supported by a BiVAD. Elective heart transplantation was carried out in the remaining eight patients (61.5%), with a mean time on the waiting list of 11.1 ± 6.3 months (range 2.5–20.8 months). None of these eight patients transplanted electively died in hospital. In-hospital mortality after heart transplantation in group B (13 patients) was 15.4%.

In group B, good long-term results were limited by serious adverse events, such as major driveline infections and cerebral vascular accident (CVA), i.e. ischaemic stroke and haemorrhagic stroke. Kaplan–Meier curves of freedom from these major events are shown in Fig. 4.

Figure 4:

Kaplan–Meier curves of freedom from major cerebral vascular accident (CVA), major driveline infection and these combined two events with major haemorrhagic episodes requiring haemotransfusions in group B.

The cumulative actuarial survival in group B shows that at 1 year following VAD implantation more than 50% of patients were ongoing (patient alive with device in place). Mortality increased after 1 year of ventricular support from 12 to 22% (Fig. 5).

Figure 5:

The cumulative actuarial survival in group B shows that at 1 year following VAD implantation, more than 50% of patients were ongoing (patient alive with device in place). Mortality increased after 1 year of ventricular support from 12 to 22%.

DISCUSSION

In our experience, emergency VAD implantation has been a viable alternative to urgent heart transplantation, avoiding the related high risk of mortality. As this study has shown, haemodynamic stabilization of critically ill patients can enable them to undergo elective cardiac transplantation (gold standard therapy) with good results. Newer indications of VAD implantation include medium- and long-term support in order to avoid the premature assignment to transplantation with uncertain results (recipient–donor optimization).

The purpose of the use of VADs in INTERMACS 1 and 2 patients is to overcome the critical stage and perform the heart transplant as an elective procedure and, thus, in better conditions. The analysis of our data shows that the advantage gained by the implantation of VADs is significant only if patients are transplanted electively within 12 months. High morbidity (caused by major driveline infections, ischaemic and haemorrhagic stroke during the follow-up period) reduces the survival beyond 1 year and limits the efficacious long-term use of these devices.

Improvements in outcomes with mechanical circulatory support for patients requiring LVAD have been universally acknowledged during the past decade [14]. However, for patients requiring biventricular support (BiVAD), outcomes still remain far inferior. Data from INTERMACS show that the presence of a BiVAD has a strong adverse effect on outcomes, with these patients facing twice the mortality risk. The BiVAD patients represent the most critically ill patients who undergo VAD implantation, and they are more likely to die after mechanical circulatory support. Adverse events of bleeding and infections are more frequent after a BiVAD implantation compared with LVAD.

These results refer to the use of paracorporeal BiVAD (like Excor Berlin Heart or Medos Thoratec) or total artificial heart. It is likely that in the future BiVAD implantation may have better results thanks to the recent use of continous-flow LVAD adapted on the right side. The feasibility of implantation of two Heartware HVAD centrifugal pumps as a biventricular device has been reported since 2010 from many centres, but the long term-results of this non-standardized procedure are yet known.

In our experience, four patients were supported by a BiVAD, and in all cases, they were transplanted on an emergency basis for a serious complication of the VAD after a mean time of 73 ± 10.9 days (range 9–222 days). The BiVAD exposes patients to a high risk of a urgent transplant procedure.

The limitations of this retrospective single-centre study are related to the restricted number of patients, and the decision criteria used to assign patients to one or other treatment (emergency transplantation or VAD implantation) were influenced by the growing experience of the centre concerning VAD implantation and management, although in critical patients, in recent years. Prospective multicentre, randomized studies should be undertaken in order to confirm outcomes and to define more clear indications/guidelines in such groups of critically ill patients.

In conclusion, in INTERMACS 1 and 2 patients, in the case of severe left ventricular dysfunction, the use of a mechanical assistance device may be the best choice. In the case of severe biventricular dysfunction, in a patient with a clear contraindication to heart transplantation, a BiVAD using two centrifugal pumps as destination therapy may be a solution. Long-term results are uncertain in this specific condition. In young patients, urgent heart transplantation may still be the best choice or, alternatively, biventricular mechanical assistance could be used for a short period of time (not longer than 30–60 days) in order to allow heart transplantation in the best patient conditions and, in particular, without acute haemodynamic impairment.

APPENDIX. CONFERENCE DISCUSSION

Dr S. Westaby (Oxford, UK): I thought this was a very interesting study, but actually would have preferred to have seen a prospective randomized trial of emergency transplantation vs. bridge to transplantation.

I took away some important messages from your paper. The first one was that even with very good-quality donor hearts, your 1 year mortality for emergency cardiac transplantation was more than 50%. Secondly, patients that were supported and bridged with a VAD to reverse multiple organ failure before transplantation actually did much better, with hospital mortality of 15%. Thirdly, after 1 year of VAD support, your results started to deteriorate because of VAD-related complications. This was very disappointing, since you were using the INCOR and the HeartWare VADs that are destination therapy VADs, but you still had a lot of problems with them. The last point was that the need for biventricular support in every single series conveys very high mortality.

I have two questions. First, because donor hearts are scarce, can it still be justified to use them as primary treatment for status 1 patients with multiple organ failure? Second, the implantable VADs cost the same as a Porsche car. Why not use the well-tried and tested temporary VADs for bridge-to-transplantation, since you can get your donor hearts in a relatively short time? Why not reverse multiple organ failure with inexpensive VADs and then either go on to destination therapy or transplantation according to the outcome?

Dr Attisani: Our experience suggests that emergency transplantation in patients in cardiogenic shock with end-organ failure is a very challenging procedure with uncertain results. Probably we have to access the emergency program more carefully in the future. Our series show that emergency permanent VAD implantation could be a good alternative although in patients without a clear condition of transplant eligibility at that time. In our experience, patients undergoing emergency transplant were supported by an ECMO in about 46% of cases.

We have not any experience with short temporary VAD both on the right and the left side without the oxygenator, we use a particular strategy to switch from ECMO to long-term VAD using transapical cannulation of the left ventricle. This strategy allows us to evaluate in two separate and consecutive times the right ventricular function and the pulmonary function and to simplify, if it is possible, the circuit to a short term left ventricular assist device as a bridge to decision. I think, it could be a good strategy to switch from ECMO to a long-term device in order to avoid high risk transplants.

Probably in these patients who underwent immunosuppressive therapy, this mild to moderate end-organ dysfunction is a very high risk factor for death after transplantation. In the majority of cases, we lost patients after emergency transplant for infective complications.

Dr Westaby: Very briefly, with a 1 year mortality of 50% from emergency transplantation, should you not use the scarce donor organs electively?

Dr Attisani: I agree with you, It is difficult to justify this choice with this data.

Dr C. Schmitz (Munich, Germany): Can you tell us how you decide when to switch from bridge to transplantation to destination therapy? It seems to be a difficult decision.

Dr Attisani: Sometimes patients supported by emergency VADs are in their 60s and they present a severe end-organ dysfunction, Sometimes we are not able before the implantation of the VAD to know if there were the conditions of transplant eligibility at that time. But I think in the majority of the cases it was the age, patients near to 65 years. For critical patients in INTERMACS level 1 or 2 today we have to change the concept “of bridge to transplant” in the concept of “bridge to candidacy”.

Dr Loisance: Let me ask you a question. One patient is deteriorating. He is a good candidate for ECMO or VAD. A second patient is perfectly stable on the VAD system. And a third patient is on VAD or ECMO with complications. For you, who is the patient who should receive a heart first?

Dr Attisani: Probably for the third patients (VAD or ECMO with complications) transplant is the only possible choice, but we have also to consider the high risk of this procedure. Is it justified? or is it better to transplant the stable patient?

Dr Loisance: You will rule out the patient in cardiogenic shock as a priority patient for cardiac transplantation?

Dr Attisani: I think that in case of cardiogenic shock and multiorgan failure, it is very difficult to think to emergency transplant as a treatment with good results. And so I think that transplant could be reserved for patients with a good prognosis and good probability of long-term survival in order to optimize the transplant results; I refer to the concept of recipient–donor optimization.

Dr Loisance: Then you will select the patient perfectly stable on a VAD?

Dr Attisani: I think the choice should be made on a case by case.

Dr Loisance: Okay. It is a difficult issue. Question from the floor.

Dr M. De Bonis (Milan, Italy): If I have understood correctly, you reported a freedom from driveline infection of approximately 50% at 1 year. That is a relatively high rate compared to destination therapy series, and also recent data. Could you give us some more insight from that point of view? I mean, how did you define driveline infection, and why do you think the rate was so high compared to what we have been seeing in recent years, because that would be very discouraging for destination therapy.

Dr Attisani: In the majority of cases, we used devices with a quite rigid driveline of 18 mm or 16 mm diameter. More recently, some devices have flexible and silicone drivelines, and it may be that in the future this incidence of infection will be less with such devices. In critical patients, I think that the risk of driveline infection is increased, and so I think that the incidence of driveline infection in our patients is also due to their critical perioperative condition.