Articles

1556 www.thelancet.com Vol 386 October 17, 2015

First clinical use of a bioprosthetic total artifi cial heart: report of two casesAlain Carpentier, Christian Latrémouille, Bernard Cholley, David M Smadja, Jean-Christian Roussel, Elodie Boissier, Jean-Noël Trochu, Jean-Pierre Gueff et, Michèle Treillot, Philippe Bizouarn, Denis Méléard, Marie-Fazia Boughenou, Olivier Ponzio, Marc Grimmé, Antoine Capel, Piet Jansen, Albert Hagège, Michel Desnos, Jean-Noël Fabiani, Daniel Duveau

SummaryBackground The development of artifi cial hearts in patients with end-stage heart disease have been confronted with the major issues of thromboembolism or haemorrhage. Since valvular bioprostheses are associated with a low incidence of these complications, we decided to use bioprosthetic materials in the construction of a novel artifi cial heart (C-TAH). We report here the device characteristics and its fi rst clinical applications in two patients with end-stage dilated cardiomyopathy. The aim of the study was to evaluate safety and feasibility of the CARMAT TAH for patients at imminent risk of death from biventricular heart failure and not eligible for transplant.

Methods The C-TAH is an implantable electro-hydraulically actuated pulsatile biventricular pump. All components, batteries excepted, are embodied in a single device positioned in the pericardial sac after excision of the native ventricles. We selected patients admitted to hospital who were at imminent risk of death, having irreversible biventricular failure, and not eligible for heart transplantation, from three cardiac surgery centres in France.

Findings The C-TAH was implanted in two male patients. Patient 1, aged 76 years, had the C-TAH implantation on Dec 18, 2013; patient 2, aged 68 years, had the implantation on Aug 5, 2014. The cardiopulmonary bypass times for C-TAH implantation were 170 min for patient 1 and 157 min for patient 2. Both patients were extubated within the fi rst 12 postoperative hours and had a rapid recovery of their respiratory and circulatory functions as well as a normal mental status. Patient 1 presented with a tamponade on day 23 requiring re-intervention. Postoperative bleeding disorders prompted anticoagulant discontinuation. The C-TAH functioned well with a cardiac output of 4·8–5·8 L/min. On day 74, the patient died due to a device failure. Autopsy did not detect any relevant thrombus formation within the bioprosthesis nor the diff erent organs, despite a 50-day anticoagulant-free period. Patient 2 experienced a transient period of renal failure and a pericardial eff usion requiring drainage, but otherwise uneventful postoperative course. He was discharged from the hospital on day 150 after surgery with a wearable system without technical assistance. After 4 months at home, the patient suff ered low cardiac output. A change of C-TAH was attempted but the patient died of multiorgan failure.

Interpretation This preliminary experience could represent an important contribution to the development of total artifi cial hearts using bioprosthetic materials.

Funding CARMAT SA.

IntroductionIn 1969, Cooley reported the fi rst use of a total mechanical heart substitute as a bridge-to-transplantation procedure in a 47-year-old man who survived 39 h after the transplantation.1 In 1986, Copeland and associates2 reported two cases of successful bridge to trans plantation with the Jarvik orthotopic heart. The same year, we reported3 a successful bridge to transplantation using two heterotopic ventricles in a patient still living today. Large series of long-term survivors of transplantation using the Cardiowest Total Artifi cial Heart (TAH, SynCardia Systems; Tucson, AZ, USA) as a bridge to transplantation were subsequently reported with a 60% survival 5 years after transplantation at and 41% survival at 10 years.4–6 Although these remarkable results were marred by strokes (8%), and haemorrhage (52%),5 they were encouraging enough to open a new era in cardiac substitution—ie, the development of TAH. This development was pioneered by DeVries and Dowling,

however thromboembolic events remained a major issue.7,8 Other concerns included the noise related to pneumatic actuation, the infections associated with the percutaneous drivelines, and minimal adaptation to variations in venous return. These side-eff ects stimulated our endeavour to design the CARMAT TAH (C-TAH) with special emphasis on the reduction of thromboembolic and haemorrhagic complications. In this regard, the main novelty with the new device is the use of bioprosthetic materials. We aimed to see if, by analogy to bioprosthetic valves, a bioprosthetic heart might help reduce the throm boembolic or haemorrhagic complications.9

MethodsDescription of the deviceThe C-TAH is a biventricular pulsatile, electrically powered, hydraulically actuated fl ow pump with all components embodied in a single device mimicking the

Lancet 2015; 386: 1556–63

Published OnlineJuly 29, 2015

http://dx.doi.org/10.1016/S0140-6736(15)60511-6

See Comment page 1517

Department of Cardiovascular Surgery (Prof A Carpentier MD,

Prof C Latrémouille MD, O Ponzio MD, Prof J-N Fabiani),

Department of Anaesthesiology and Intensive

Care (Prof B Cholley MD, D Méléard MD,

M-F Boughenou MD), Department of Cardiology

(Prof M Desnos MD, Prof A Hagège MD), and

Department of Haematology (Prof D M Smadja MD), Hôpital Européen Georges Pompidou,

Assistance Publique-Hôpitaux de Paris, Université Paris

Descartes-Sorbonne Paris Cité, Paris, France; Institut du

Thorax, Hôpital Guillaume et René Laënnec, Université de

Nantes, France (Prof D Duveau MD, Prof J-C Roussel MD,

E Boissier MD, Prof J-N Trochu MD, Prof J-P Gueff et MD,

Prof M Treillot MD, P Bizouarn MD); and CARMAT

SA, Vélizy-Villacoublay, France (M Grimmé MEng, A Capel MSc,

P Jansen MD)

Correspondence to:Prof Alain Carpentier,

Department of Cardiovascular Surgery, Hôpital Européen

Georges Pompidou, Université Paris Descartes, 75015 Paris,

Francealain.carpentier@egp.aphp.fr

Articles

www.thelancet.com Vol 386 October 17, 2015 1557

natural heart implanted in the pericardial sac (fi gure 1). Its most original feature is the use of bioprosthetic materials, similar to those used for bioprosthetic heart valves to potentially reduce the need for anticoagulation.9 Each of the two artifi cial ventricles consists of two compartments separated by a hybrid membrane. One compartment contains silicone oil activated by a rotary pump, which deploys the hybrid membrane back and forth in a systolic and diastolic motion. The other compartment, which contains the blood, has its static surface covered with expanded polytetrafl uoroethylene, a blood compatible material currently used in vascular surgery. The hybrid membrane is made of two layers glued together. One layer, in contact with the silicone oil, is made of polyurethane, the other layer in contact with the blood, is made from bovine pericardial tissue chemically treated by glutaraldehyde to achieve long-term tolerance and haemocompatibility.10 Carpentier-Edwards bioprosthetic valves (Edwards Lifesciences; Irvine, CA, USA) and connecting atrial fl anges, treated with the same chemical process, are positioned at the infl ow and outfl ow orifi ces of each blood compartment to direct forward fl ow. The pumps are regulated so as to mimic the viscous elastic contractility of the natural heart, thus providing physiological pressure curves (fi gure 2), which preserves valve durability and optimises blood circulation. Electronics and micro processors that drive the system are incorporated into the C-TAH. Sensors in the ventricles allow instantaneous monitoring of pressure providing continuous assessment of right and left ventricular preload and afterload. The position of the membrane is detected in situ by ultrasound

transducers. A control algorithm responds to changes in preload and afterload by adjusting heart rate (from 35 to 150 beats per min) and stroke volume (30–65 mL) while assuring the ejection of the total blood mass contained in the ventricle to prevent stasis. The resulting pulsatile blood fl ow ranges from 2 L/min to 9 L/min with automated adjustment to balance right-side and left-side fi lling, protecting the lungs from high pressure in the left atrium. All the components of the prosthesis are surrounded by a polyurethane sac containing silicone oil that serves as a compliance chamber when stroke volume varies. Energy is provided by lithium ion batteries carried in a wearable bag or belt connected to the C-TAH by a single transcutaneous highly fl exible driveline, 8 mm in diameter (appendix).

Preclinical testingWe tested performance and durability of the components of the C-TAH in a mock circulation and durability bench test. Numeric computer-assisted conception and simulation were extensively used to minimise the need for implantations in animals.11 The C-TAH implantation in 12 calves allowed to validate numeric conception, device performances, and physiological response.12 These analyses demonstrated the eff ectiveness of the C-TAH in vivo since animals recovered their vital functions within 48 h after the operation. We noted no haemolysis, nor thrombus formation in the kidney, the brain, or inside the device itself. Haemocompatibility was further confi rmed by in-vitro studies exposing a large surface of the bioprosthetic materials to circulating blood.13 We assessed anatomical fi t by virtual implantation in a 3D visualisation

Research in context

Evidence before this studyWe searched PubMed with the terms “total artifi cial heart”, “biventricular assist device”, and “destination therapy” to fi nd relevant articles in any language published up to February, 2015. We found that up to 30% of patients with end-stage heart failure experienced biventricular failure that requires biventricular support. For these patients, three options are available: (1) biventricular assist devices that use bulky paracorporeal pumps, (2) the use of two centrifugal pumps implanted in the left and right ventricles, and (3) total artifi cial heart (TAH) replacing the native ventricles. All these options enable only a limited quality of life and carry on a signifi cant risk of thrombus formation within the devices.

Added value of this studyWith the aim of improving existing devices, we took advantage of our long-term experience with bioprosthetic materials (Carpentier-Edwards valvular bioprosthesis) that reduce the risk of thrombus formation to initiate the development of a bioprosthetic TAH intended for destination therapy. To our knowledge, this is the fi rst report describing the clinical

experience with bioprosthetic material used within an artifi cial heart. In addition to haemocompatibility, the new CARMAT TAH (C-TAH) has other promising features. The use of an electro-hydraulically driven TAH instead of the pneumatically-driven TAH has resulted in a signifi cant noise reduction, contributing to an improved quality of life. The precise anatomic compatibility is associated with a complete lack of perception of the prosthesis by the patients. A single driveline with an 8 mm diameter comes out of the patient’s skin. The pulsatile fl ow is delivered through a viscoelastic pattern mimicking the natural contraction of the heart and resulting in physiological pressure curves and absence of haemolysis.

Implications of all the available evidenceOur fi ndings in two patients supported for 74 and 270 days, respectively, show that the acquired haemocompatibility provided by the bioprosthetic materials holds good also in devices with much larger surfaces of these materials exposed to blood circulation. They open a promising new avenue in the development of life-saving long-term circulatory assist devices and artifi cial hearts for destination therapy.

See Online for appendix

Articles

1558 www.thelancet.com Vol 386 October 17, 2015

of the chest (appendix). CT scan analysis of 110 patients showed that the available model of the prosthesis would fi t in 84% of men and 16% of women. The Agence Nationale de Sécurité du Médicament (ANSM) authorised a clinical trial for feasibility and safety in four patients at high risk of death from end-stage biventricular failure and not eligible for transplant.

Patient selectionThe panel shows the inclusion and exclusion criteria. Three cardiac surgery centres in France participated in patients screening and enrolment: Hôpital Européen Georges Pompidou, AP-HP, Paris; Hôpital Laënnec, Nantes; and Centre Médico-Chirurgical Marie Lannelongue, Le Plessis Robinson. 22 patients were

Figure 1: The CARMAT bioprosthetic total artifi cial heartAll components, except batteries, are embodied in a single device mimicking a normal heart. The polyurethane sac serves as a compliant chamber. (A) External view. (B) Internal view. Electrical rotary pumps (1) activate silicone oil (2) deploying back and forth hybrid membranes. Electronic components are located in the interventricular septum (3).

A B

1

2

3

Electricalrotary pumps

Left ventricularhybrid membrane

Figure 2: Arterial pressure and fi nger plethysmographic recordings obtained on day 74, shortly before cardiac arrest in patient 1 Waveforms had normal physiological shape resulting from the viscoelastic type of ventricular contraction and relaxation. Values indicated a mean arterial pressure of 84 mm Hg, and peripheral oxygen saturation of 100%. PA=arterial pressure (in mm Hg). PLETH=arterial plethysmographic waveform. SpO2=peripheral oxygen saturation.

SpO2 100

PA

PLETH

16:23:10 (02/03/2014)

0·0/75·0/150·0

PA 159/59 (84) Pulse 94

Articles

www.thelancet.com Vol 386 October 17, 2015 1559

screened between Oct 6, 2013, and Aug 3, 2014, by an independent group of cardiologists at each centre (appendix).

Patient descriptionPatient 1 was a man aged 76 years with severe biventricular failure. The fi rst clinical manifestation occurred in 2004 with chronic shortness of breath, lower limb oedema, and several episodes of pulmonary oedema. A diagnosis of dilated cardio myopathy was established. He had a left ventricular ejection fraction (LVEF) of 20%. Additional risk factors included atrial fi brillation, dyslipidaemia, type 2 diabetes, renal dysfunction, and chronic bronchitis. He underwent cardiac resynchronisation and was implanted with an internal cardioverter defi brillator (ICD) in 2011. Despite maximal medical therapy, the patient’s condition continued to worsen, requiring repeated hospital admissions. On Dec 12, 2013, he was referred to our tertiary centre with a new episode of decompensated heart failure and was classifi ed as stage 2 according to the Interagency Registry for Mechanically Assisted Circulatory Support (INTERMACS) classifi cation. Biventricular involvement and continuous worsening of the clinical status prompted us to discuss the indication for a total replacement of the heart. C-TAH implantation was performed on Dec 18, 2013.

Patient 2 was a man aged 68 years, with end-stage biventricular failure due to dilated cardiomyopathy with an LVEF of 15%. Despite resynchronisation and maximal medical therapy, his clinical status remained very unstable and he was admitted repeatedly to the hospital for decompensated heart failure during the 3 months preceding the current admission in July, 2014. Interrogation of his ICD showed four episodes of ventricular fi brillation. He was admitted on July 18, 2014 for cardiogenic shock requiring inotropic support and high-dose intravenous diuretics and was classifi ed in INTERMACS class 2. Additional risk factors included systemic hypertension, dyslipidaemia, and peripheral arteriopathy. Age over 65 years and biventricular involvement precluded a heart transplantation as well as the use of a left ventricular assist device. Therefore, implantation with a C-TAH was proposed and carried out on Aug 5, 2014.

For both patients, written informed consent and Ethical Committee’s agreement were obtained.

Surgical techniqueThe surgery was initiated with a median sternotomy and a midline vertical incision of the pericardial sac. Cardiopulmonary bypass (CPB) was established between the ascending aorta and the two venae cavae. The ventricles were excised leaving intact the right and left atrium and the atrioventricular junctions. The aorta and the pulmonary artery were transected right above the commissures. The sizing of the mitral and tricuspid orifi ces indicated a 35 mm diameter for both of them. The bioprosthetic

fl ange of the left atrial connecting ring was sutured to the left atrioventricular orifi ce. A similar suturing was completed for the right atrial connecting ring. The left and right rings were connected to a double orifi ce connecting device. As predicted by numeric simulation, the C-TAH could be easily positioned in the pericardial sac of both patients and then secured to the double orifi ce connecting device with clips. The pulmonary conduit of the C-TAH was sewn to the distal pulmonary artery and the aortic conduit to the distal aorta. At the end of the implantation, careful de-airing was achieved by CO2 insuffl ation, and aortic and pulmonary suction. After unclamping the aorta, the output of the C-TAH was progressively increased while weaning the patient from the CPB. The mean arterial pressure was 70 mm Hg (± 10 mmHg), the left atrial pressure was maintained below 15 mm Hg and the right atrial pressure below 10 mm Hg. The pericardial sac was closed without change in pressures. CPB times were 170 min for patient 1 and 157 min for patient 2.

Role of the funding sourceCARMAT SA (Vélizy-Villacoublay, France) provided the two implanted C-TAHs and engineering support during patient care. The funder did not participate in data analysis and writing of the present report. The corresponding author, cofounder of CARMAT SA wrote the report. All authors had access to all the data: CL, BC, DMS, and DD were responsible for the decision to submit and revise the manuscript for publication.

Panel: Eligibility criteria

Inclusion criteria• Age ≥18 years• INTERMACS profi le 1–2• LVEF ≤30%• Optimised medical treatment (European Society of

Cardiology, American Heart Association recommendations)• On intravenous inotropes ≥7 days• BSA ≥1·7 m²• Anatomic compatibility verifi ed by 3D modelling• Signed informed consent

Exclusion criteria• Technical obstacle, high surgical risk (judged by the

investigator)• Platelet count <150 000 cells per μL or international

normalised ratio ≥1·5 without anticoagulant therapy• Haemorrhagic stroke <6 weeks• Active uncontrolled bloodstream infection• Haemodynamically signifi cant peripheral vascular disease• Malignant neoplasm with life expectancy <6 months• On corticosteroids medication equivalent to prednisone

7·5 mg per day• Irreversible cognitive dysfunction, psychosocial issues, or

psychiatric disease that are likely to impair compliance

Articles

1560 www.thelancet.com Vol 386 October 17, 2015

ResultsPatient 1 was extubated 12 h postoperatively. Continuous veno-venous haemodiafi ltration was instituted from day 2 because of worsening in renal function. Hyper glycaemia was initially controlled using continuous insulin infusion followed by iterative subcutaneous injections. Infusion of unfractionated heparin was progressively restored to achieve an anti-Xa activity of 0·20 IU/mL. The INTERMACS defi nes haemolysis as a concentration of plasma free haemoglobin higher than 40 mg per 100 mL in the clinical follow-up after implantation.14 No relevant

haemolysis was detected at any time (fi gure 3 and appendix), with only two non-consecutive values (8·7 mg per 100 mL at day 3 and 7 mg per 100 mL at day 21) above the physiological upper range values (5 mg per 100 mL). On postoperative day 14, the patient was able to walk a few steps and started a rehabilitation programme. On day 23, he presented with respiratory distress requiring mechanical ventilation. A 10 mm Hg pressure rise in the compliance bag suggested a tamponade. A CT scan showed the presence of periprosthetic clots responsible for left atrial and pulmonary veins compression. The patient was operated on. A large thrombus was found surrounding the C-TAH but no source of bleeding could be identifi ed. Blood tests showed a low platelet count and increased d-dimers due to the periprosthetic thrombus. These markers normalised after thrombus removal. Heparin infusion was stopped on day 24 because of persisting bleeding in the chest tubes. On day 27, pneumonia required ventilatory support. On day 41, we noted gastrointestinal bleeding due to antral gastritis and angiodysplasia, which were treated with omeprazole (8 mg/h) and cauterisation during gastroscopy. On day 47, a gastric ulcer was discovered and successfully treated by endoscopic clipping. However, gastrointestinal bleeding persisted, requiring iterative blood transfusion. Con-comitantly, the patient developed haemoptysis due to alveolar bleeding diagnosed by bronchoscopy. Because of these complications, neither anticoagulant nor antiplatelet treatments were given. The hypothesis of an excess cardiac output was raised and on day 51 the fl ow of the artifi cial heart was reduced from 5·8 L/min to 5·0 L/min. Both alveolar and gastrointestinal bleeding stopped following this manoeuver, as confi rmed by video capsule endoscopy on day 56. In the meantime, three tracheal extubations were attempted and failed. Reasons for ventilator-weaning failure included bronchial infection with Stenotrophomonas maltophilia, and global weakness due to prolonged stay in the intensive care unit (ICU). A tracheostomy was done on day 47, which greatly improved the comfort of the patient. On day 74, while enjoying a conversation with his family and drinking a soft drink, the patient died suddenly. The cause of the prosthesis arrest was identifi ed by two independent groups of experts, as being a very rare and preventable failure of an electronic component. Expert’s analysis revealed a manufacturing defect of a bulk capacitor from an electronic sensor. The analysis performed by the technical team also concluded that every mechanical component of the prosthesis was fully operational, meaning that the capability of the device to deliver the blood fl ow was kept intact.

Patient 2 was extubated 6 h after ICU admission. He spent 13 days in ICU because of postoperative acute renal failure requiring transient haemodialysis. A single episode of bronchial infection was successfully treated by targeted antibiotic therapy. A CT scan done at day 8 did not show any focus of lung infection or atelectasis due to

Figure 3: Plasma free haemoglobin for both patients over timePlasma free haemoglobin values for patients 1 and 2, indicating the absence of relevant haemolysis during the entire clinical follow-up. Day 0 for patient 1 corresponds to preoperative value; all other values correspond to postoperative days.

0 1 2 3 4 5 6 7 9 11 13 14 15 16 17 18 19 20 21 22 27 34 41 48 50 55 64 69 73 76 83 90 97104 111

1180

1

2

3

4

5

6

7

8

9

10

Plas

ma

free

haem

oglo

bin

(mg

per 1

00 m

L)

Postoperative days

Patient 1Patient 2

Figure 4: CT scan obtained at postoperative day 8 showing the C-TAH implanted in patient 2Note the absence of lung atelectasis in the vicinity of the prosthesis. C-TAH=CARMAT total artifi cial heart.

Articles

www.thelancet.com Vol 386 October 17, 2015 1561

the prosthesis itself (fi gure 4). The patient was transferred to a step down unit close to the ICU on day 13. Anticoagulation was provided using heparin, and platelet antiaggregation was achieved using aspirin. At day 45, a pericardial eff usion was detected and treated by pericardial drainage followed by ICU admission. A minor ascites leakage persisted around the percutaneous orifi ce of the driving line and was eventually dried up by additional sutures. Heparin was replaced by warfarin sodium, but adequate and stable anticoagulation proved diffi cult to achieve. Therefore, tinzaparin, a low molecular weight heparin, was introduced (with a dose of 175 IU/kg per day), resulting in stabilisation in platelet count. Monitoring of plasma free haemoglobin (fi gure 3) was consistent with the absence of haemolysis. Overall, the postoperative recovery was uneventful and the patient was discharged again from ICU 2 days after the pericardial drainage at day 47. He was rehabilitated rapidly and resumed exercising (walk and exercise bike). He returned home on Jan 2, 2015, a few days after the portable system had been approved by ANSM. The patient was at home on May 1, 2015, when he suddenly had symptoms of low cardiac output. He was admitted to the hospital in emergency and diagnosed with device failure. He was placed under veno-arterial extracorporeal life support and operated on 12 h later to replace the C-TAH. After the procedure, he developed intractable haemorrhage and multiorgan failure. He died on May 2, 270 days after the initial C-TAH implantation.

DiscussionThere is a lack of long-term therapies for patients with end-stage biventricular failure who are not eligible for heart transplantation. These fi rst two cases represent the preliminary results of the proof-of-concept study aiming at demonstrating the feasibility of using a new approach to biventricular assist devices as destination therapy. This novel approach is based on the use of haemocompatible materials and more physiological behaviour of the prosthetic pump.11 Withdrawing anticoagulants in the fi rst patient did not result in thromboembolic events and autopsy did not reveal macroscopic thrombus formation (fi gure 5 A, B). In patient 2, who had no bleeding-related complications, anticoagulant and antiplatelet therapies were maintained in accordance with current practice for valvular bioprosthesis management.15,16 It is now well established that valvular bioprosthesis undergoes progressive trans-

Figure 5: Open view of left ventricular cavity at autopsy of patient 1 (postoperative day 74)

(A) Left ventricular (LV) cavity static surface coated with polytetrafl uoroethylene (PTFE). (B) LV hybrid membrane exposed to blood did not display any clot

formation, including near the periphery, at the junction with PTFE-coated LV. (C) Histological preparation (hemtoxilin-eosin-safran, magnifi cation ×10) of the

glutaraldehyde-processed pericardial membrane (Pe). The side facing the blood cavity (Bl) was covered with a protein layer (Pr), known to play a key role in

preventing clot formation. The polyurethane layer (Pu) of the hybrid membrane has been removed for the preparation.

A

B

CBl

Pu

Pr

Pe

Articles

1562 www.thelancet.com Vol 386 October 17, 2015

formation following implantation in patients. This process consists in the lining of the bioprosthetic material by endogenous protein deposit.9 This deposit creates a thromboresistant layer resulting in a fully haemocompatible prosthesis. This process takes several weeks or months depending on the extent of biomaterial to cover. Until this thromboresistant layer is completed, there is a need for anticoagulants or antiplatelet therapy, or a combination of both, to minimise the risk of thromboembolic events.15,16 The total area of haemo-compatible material within a C-TAH is 28 times greater than that of a single valvular bioprosthesis, justifying a prolonged anticoagulation and antiplatelet preventive treatment. The histological examination of the C-TAH of patient 1 showed the ongoing constitution of a protein lining on the bioprosthetic components (fi gure 5C). The duration of anticoagulation required to achieve complete “acquired” haemocompatibility in the setting of C-TAH remains to be established.

The use of biological material requires optimal haemodynamic conditions to avoid or reduce shear stress, turbulences, haemolysis, and areas of stasis, all causes of tissue deterioration and thrombus formation. These potential complications have been addressed by the choice of a pulsatile system, viscoelastic type of ventricular contraction, well designed ventricles and ejection chambers, and optimised medical regulation to mimic the physiology of a native heart. As a result, no haemolysis was noted in either patient. To optimise the tolerance of the C-TAH, other aspects than biocompatibility of the materials are also of paramount importance. Among those, the haemodynamics gen-erated by the C-TAH must be as close as possible from the human physiology to minimise shear stress and turbulences. The pulsatile character of the ejection is very important to mimic a native heart. Several issues are related to non-pulsatile fl ows resulting from various pump systems. Rotor pumps or centrifugal pumps are associated with blood trauma and haemolysis.17,18 Under chronic circumstances, non-pulsatile assist devices have been deemed responsible for gastrointestinal bleeding due to a combination of various factors, including arteriovenous malformations due to low pulse pressure.19 In patient 1, the gastric bleeding was completely con trolled after reduction of the pump fl ow. Subsequently, the angiodysplasia noted on the initial gastroscopy disappeared at the video capsule exam-ination 15 days later. As previously underlined by DeVries,8 the ability to deliver high cardiac output should be applied with caution, and moderate outputs should be preferred. The viscoelastic characteristic of the C-TAH contraction, which generates physiological arterial pressure curves, is also important to minimise the mechanical stress on the valvular bioprosthesis. Finally, the anatomical similarities with the native heart are essential to minimise blood stasis and the subsequent thromboembolic risk. Eventually, the ability

of the prosthesis to respond to variations in venous return might also better suit the physiological needs of the patient. Additionally, the C-TAH has the advantage to reproduce normal anatomy and to be totally implantable within the native pericardial sac, which avoids com pression of surrounding vessels and organs.

This initial experience with C-TAH in two patients provides an optimistic appraisal of the value of using bioprosthetic materials in the construction of artifi cial hearts and devices. The fact that both patients did not display any thromboembolic complication, even after a 50-day anticoagulant-free period in the fi rst patient, reinforces our hypothesis that a large surface of this material should behave like the limited surface of a valvular bioprosthesis. We present the fi rst two patients of the feasibility study for C-TAH as a destination therapy in patients with end-stage biventricular failure. These are interim results of an ongoing investigation with an endpoint of four patient implantations. However, a greater number of patients with a longer follow-up is required to validate this hypothesis and to establish precisely the type and duration of anticoagulant and antiplatelet regimens. The delay necessary to achieve dependability of the protein layer, which renders these bioprosthesis truly non-thrombogenic remains to be determined. Analysis of the two devices revealed no mechanical failure. In both devices, the sources of the defective electronics were found and corrected in subsequent manufacturing procedures. Since this report has been accepted for publication, a third patient has been implanted successfully and is now being rehabilitated close to his home town, 104 days postoperatively. ContributorsACar is the conceptor of the CARMAT TAH; he enrolled the two patients into the study, and wrote the report. CL implanted the fi rst patient, assisted for the operation of the second patient, and participated in the drafting of the report. BC was the anaesthesiologist in charge of peri-operative care of patient 1 and drafted the report with ACar. DMS did all the coagulation analysis and participated in the drafting of the report. J-CR, J-NT, J-PG, MT, and PB participated in the recruitment and supervised the follow-up of the second patient. EB supervised the coagulation laboratory investigation for the second patient. DM, M-FB, and OP participated in the management of the fi rst patient. PJ, MG, and ACap produced fi gures and contributed to draft amendment. AH, MD, and J-NF participated in the selection of the fi rst patient. DD implanted the second patient, assisted for the operation of the fi rst patient, and supervised the postoperative follow-up of the second patient.

Declaration of interestsACar is cofounder and shareholder of CARMAT SA. CL, DMS, BC, DM, JCR, and DD received consulting fees from CARMAT. PJ, MG, and ACap are employed by CARMAT SA. EB, J-NT, J-PG, MT, PB, M-FB, OP, AH, MD, and JNF declare no competing interests.

AcknowledgmentsFunding for the study was provided by CARMAT SA. We thank Patrick Bruneval for performing the histopathological analysis. We thank the nursing staff of the cardiovascular intensive care unit and surgery departments of Paris and Nantes university hospitals for taking care of both patients during the follow-up. We thank Christiane Venezziani for expert assistance with manuscript preparation.

Articles

www.thelancet.com Vol 386 October 17, 2015 1563

References1 Cooley DA, Liotta D, Hallman GL, Bloodwell RD, Leachman RD,

Milam JD. Orthotopic cardiac prosthesis for two-staged cardiac replacement. Am J Cardiol 1969; 24: 723–30.

2 Copeland JG, Levinson MM, Smith R, et al. The total artifi cial heart as a bridge to transplantation. A report of two cases. JAMA 1986; 256: 2991–95.

3 Carpentier A, Brugger JP, Berthier B, et al. Heterotopic artifi cial heart as bridge to cardiac transplantation. Lancet 1986; 2: 97–98.

4 Leprince P, Bonnet N, Rama A, et al. Bridge to transplantation with the Jarvik-7 (CardioWest) total artifi cial heart: a single-center 15-year experience. J Heart Lung Transplant 2003; 22: 1296–303.

5 Roussel JC, Senage T, Baron O, et al. CardioWest (Jarvik) total artifi cial heart: a single-center experience with 42 patients. Ann Thorac Surg 2009; 87: 124–29.

6 Copeland JG, Copeland H, Gustafson M, et al. Experience with more than 100 total artifi cial heart implants. J Thorac Cardiovasc Surg 2012; 143: 727–34.

7 Dowling RD, Gray LA Jr, Etoch SW, et al. The AbioCor implantable replacement heart. Ann Thorac Surg 2003; 75: S93–99.

8 DeVries WC, Anderson JL, Joyce LD, et al. Clinical use of the total artifi cial heart. N Engl J Med 1984; 310: 273–78.

9 Carpentier A. Lasker Clinical Research Award. The surprising rise of nonthrombogenic valvular surgery. Nat Med 2007; 13: 1165–68.

10 Carpentier A, Deloche A, Relland J, et al. Six-year follow-up of glutaraldehyde-preserved heterografts. With particular reference to the treatment of congenital valve malformations. J Thorac Cardiovasc Surg 1974; 68: 771–82.

11 Carpentier A. Animal experimentation, computer simulation and surgical research. Bull Acad Nat Med 2009; 193: 1747–55 (in French).

12 Latrémouille C, Duveau D, Cholley B, et al. Animal studies with the Carmat bioprosthetic artifi cial heart. Eur J Cardiothorac Surg 2015; 47: e172–78.

13 Jansen P, van Oeveren W, Capel A, Carpentier A. In vitro haemocompatibility of a novel bioprosthetic total artifi cial heart. Eur J Cardiothorac Surg 2012; 41: e166–72.

14 Feldman D, Pamboukian SV, Teuteberg JJ, et al. The 2013 International Society for Heart and Lung Transplantation Guidelines for mechanical circulatory support: executive summary. J Heart Lung Transplant 2013; 32: 157–87.

15 Joint Task Force on the Management of Valvular Heart Disease of the European Society of Cardiology, European Association for Cardio-Thoracic Surgery, Vahanian A, et al. Guidelines on the management of valvular heart disease (version 2012). Eur Heart J 2012; 33: 2451–96.

16 Whitlock RP, Sun JC, Fremes SE, Rubens FD, Teoh KH, American College of Chest Physicians. Antithrombotic and thrombolytic therapy for valvular disease: Antithrombotic Therapy and Prevention of Thrombosis, 9th edn. American College of Chest Physicians Evidence-Based Clinical Practice Guidelines. Chest 2012; 141: e576S–600S.

17 Saczkowski R, Maklin M, Mesana T, Boodhwani M, Ruel M. Centrifugal pump and roller pump in adult cardiac surgery: a meta-analysis of randomized controlled trials. Artif Organs 2012; 36: 668–76.

18 Palanzo DA, El-Banayosy A, Stephenson E, Brehm C, Kunselman A, Pae WE. Comparison of hemolysis between CentriMag and RotaFlow rotary blood pumps during extracorporeal membrane oxygenation. Artif Organs 2013; 37: E162–66.

19 Harvey L, Holley CT, John R. Gastrointestinal bleed after left ventricular assist device implantation: incidence, management, and prevention. Ann Cardiothorac Surg 2014; 3: 475–79.