Scientific Articles

Extracorporeal Liver Support Devices: Where we are, and where we’re going

Ryder Damen, MEng


The treatment of acute liver failure (ALF) remains to be a difficult endeavour.2 Orthotopic Liver Transplant (OLT), the gold standard treatment, still maintains a high mortality rate, even excluding those patients that die on the waiting list.3 To combat this, the concept of the extracorporeal liver was created, with solutions arising with both biological and non-biological methods. While non-biological solutions have demonstrated good efficacy in treatment, it is likely a complete solution will be found in the form of a Bioartificial Liver (BAL), or a combination of non-biological and BAL technologies. This is simply because non-biological systems don’t have the technology to fully replicate the many functions of adult hepatocytes in vivo. The non-biological field is well established and currently undergoing tests to determine efficacy in clinical environments. The much newer biological field however is still in the process of pre-clinical work with the immortalized hepatocyte line based Academic Medical Centre Bioartificial Liver (AMC-BAL) and the primary porcine hepatocyte based Spheroid Reservoir Bioartificial Liver (SRBAL). With their function proven, the most important aspect of getting bioartificial livers to human patients in the near future will not only be establishing safety and efficacy in a clinical model, but solving problems associated with scale up, and portability.


Few organs in the human body perform as many complicated tasks as the liver.2 Responsible for ammonia detoxification, lactate elimination, drug metabolism, immunomodulation, protein synthesis, as well as lipid and carbohydrate metabolism to name a few, hepatocytes – the functional cells of the organ – have their work cut out for them.20 A complicated structure in both morphology and metabolism, the replication of the functional subunit of the liver has long been a challenge in the field of tissue engineering, and cellular therapies. While the liver does maintain a partial ability to regenerate following injury, many worldwide find themselves in need of a partial or complete orthotopic liver transplant (OLT) every year, from a variety of reasons. 2

Acute Liver Failure

A life-threatening condition, acute liver failure (ALF) is a rare but serious complication for many. Characterized by severe hepatic dysfunction, encephalopathy, and abnormal liver biochemistry, this condition most commonly strikes otherwise healthy adults with no history of liver disease.3 While causes vary (and many are unknown) the most common cause of ALF in the developing world is viral hepatitis.2 In the developed world, causing nearly 40% of cases in the United States, is acetaminophen (TylenolTM) toxicity.2 These two major causes in addition to several others place thousands of patients on transplant lists a year, many of whom don’t survive to see a transplant surgery.2 Those lucky enough to receive a donor liver may still be at a substantial risk for future failure, as patient mortality rates range from 20-50% within the first year post-procedure.3

Other Sources of Liver Failure

Acute Liver Failure is not the only condition that can benefit from various treatment options. Acute-on-chronic liver failure is characterized by a sudden onset of ALF in patients with pre-existing liver conditions (chronic liver failure).2 Patients with an acute-on-chronic diagnosis, due to their previous conditions, are often not eligible for transplants.12 The need for an external liver support device is perhaps the greatest in this population, as many do not have a transplantation option for their already diminished liver capacity from their pre-existing condition.12

External Liver Support Devices

While a complete or partial liver transplant remains the gold standards for patients with severe liver damage, an external liver support device (ELSD), or artificial liver, can elongate a patient’s life on the donor list in a similar manner to renal dialysis.18, 21 For patients with an advanced state of ALF, the goal of the support device is to bridge them to a transplant.18 In addition to bridging, an ELSD is also hypothesized to improve transplant outcome by normalizing liver biochemistry before the new organ is transplanted, potentially reducing shock on the transplanted tissue.18 If, however a medical team is able to intervene at an early stage of ALF progression, the ELSD can be used to relieve pressure or workload on the patient’s damaged organ.21 One of the features of the human liver is a substantial ability for regeneration as compared to other tissues.2 It therefore remains the hope in these early treatments of ALF that the extracorporeal device will take over or offset the function of the patient’s liver, allowing it to partially or completely regenerate, altogether negating the need for transplant.21

There exist three classifications of extracorporeal liver assist devices; non-biological, biological, and a combination of the two (modular). Non-biological therapies often operate with similar methods: the removal of albumin-bound or water-soluble toxic metabolites in an attempt to normalize a patient’s biochemical profile and reverse any corresponding side effects. Reviewed here are the biological, and non-biological systems that continue to show the most promise in the field.

Non-Biological Liver Support Devices

MARS (Molecular Adsorbent Recirculating System)

MARS therapy, first developed in the late 90’s, obtained FDA approval in 2005 for the treatment of drug overdoses.10, 11 It has since been amended for treatment of hepatic encephalopathy, and is pending approval for treatment of long term hepatic disorders.11 Short for molecular adsorbent recirculating system, it functions in a similar manner to extracorporeal membrane oxygenation (ECMO) or renal dialysis (hemodialysis).11 In the case of ALF, albumin (a toxic metabolite binding protein) in a patient’s blood is quickly oversaturated and these saturating toxic metabolites (bilirubin, bile acids, ammonia, etc.) are free floating in high quantities.10 The overwhelming quantities of these metabolites quickly cause systemic dysfunction, including increasing damage to the patient’s liver.10 MARS therapy removes these albumin-bound waste products with high efficiency, and is often paired with hemodialysis to support a critically ill ALF patient.10 MARS accomplishes this feat by exposing a patient’s blood with the toxic metabolites to a counter-current flow membrane with a 16% albumin dialysate circuit on the other side.11 Water-soluble, and protein-bound metabolites dissociate, and diffuse across the membrane and are bound by albumin proteins with free binding sites in the secondary ( 16% albumin dialysate) circuit.10 In the second circuit, albumin proteins are then stripped of metabolites by albumin deligandizing agents (low-flux dialysis), and recirculated for a continued use; a key aspect of the therapy.10, 12 A commercially available system, MARS has recently been tested in level 1 trauma centers and intensive care units in the United States, often in conjunction with standard hemodialysis.11 While recent studies performed have not entirely eliminated selection and other biases, data is promising.11 MARS is however a stopgap; even upon initial publishing, authors and creators acknowledged that the therapy would need to be replaced, or offset with a biological solution.10 While efficacious in the removal of albumin-bound and free-floating toxins, MARS, like other non-biological methods simply isn’t a complete solution.

SPAD (Single Pass Albumin Dialysis)

Similar to MARS therapy, single pass albumin dialysis (SPAD) exposes a patient’s blood to a membrane, with an albumin dialysate solution on the other side.12 A major difference however, is that dialysate is only used once.12 When exposed to the patient’s blood, this dialysate is discarded, as compared to the recirculation via low-flux dialysis as seen in MARS.12 Preceding this therapy was the complete replacement of a patient’s plasma. Total plasma replacement and SPAD however have fallen out of favour due to more efficient innovations like MARS and Prometheus.

Prometheus (Fractionated Plasma Separation, Adsorption and Dialysis Device)

Combining high-flux hemodialysis and fractionated plasma separation and adsorption, Prometheus therapy takes a slightly different approach than MARS and SPAD to metabolite filtration.12 Whereas with MARS and SPAD, where albumin molecules never leave the patient circuit, a large pore membrane (cut-off approximately 250kD) is used to remove albumin in a process known as fractionated plasma separation and adsorption.12, 13 Once removed from a patient’s blood, albumin, its bound toxic metabolites, and other filtrates are circulated in a secondary circuit containing an isotonic NaCl solution.12 This solution is passed over two filters (Prometh01 and Prometh02) with specific selectivity for albumin-bound metabolites.12 Once unbound from their metabolites, albumin proteins and the solution is returned to the blood. Finally, blood is passed through a standard hemodialyzer before being returned to the patient, to remove any water-soluble metabolites.12 This solution while achieving similar results to MARS, is quite simply another way to achieve the same results. Currently, there exist no definitive comparative clinical studies between the two therapies.

Hepa Wash

Most recently introduced, Hepa Wash treatment also takes advantage of an albumin dialysate. Composed of three separate circuits (blood, dialysate, and Hepa Wash), the technology is similar in dialysis to aforementioned solutions; Hepa Wash however uses a significantly lower 2% albumin dialysate.17 Filtrate albumin in the dialysate solution is then passed through two parallel circuits and filters, each with an acidic and basic environment.17 Cationic and anionic metabolites are then removed respectively, and the filtered, unbound albumin is returned to the blood circuit via the initial dialyzer membrane.17 While similar in operation to MARS, Hepa Wash’s employment of acidic and basic environments is a key difference with similar efficacy and safety to MARS or Prometheus treatment.17  In recent clinical testing, Hepa Wash was administered to 14 patients with acute-on-chronic liver failure (n=9), or secondary liver failure.15 While no measurements of serum ammonia levels were made, Hepa Wash treatment resulted in significant declines of serum creatinine, bilirubin, and blood urea nitrogen (BUN), with no indicated adverse effects.15, 1  Similar results were observed in a porcine model with induced ischemic and cholestatic liver injury showing significant declines in serum ammonia and bilirubin, and associated cardiovascular and cerebral conditions.16

Biological Liver Support Devices

The most promising types of treatment are those that incorporate live hepatocytes into the design, as they have the potential to offset the full spectrum of tasks hepatic tissue regularly performs. Known simply as a Bioartificial Liver (BAL), these devices at their most basic are bioreactors with affluent and effluent media exchanged with a patient’s circulatory system either directly, via plasmaphoresis, or another filtration mechanism. While there have been numerous advances and options in bioartificial livers, ones discussed in this review are selected for their overall efficacy, safety, or promise for future applications. Most devices employ two circuits: a patient and a bioreactor circuit. The patient circuit retrieves blood from a patient’s circulation, and either as whole blood or via plasmaphoresis, plasma; passes it through a filter similar to non-biological solutions. Fluid and molecules in the bioreactor circuit are transferred via this filter, and circulated through the bioreactor for the hepatocytes to metabolize.1, 19 These primary filters, in most cases, allow for an exchange of toxic metabolites and proteins, without exposing a patient directly to a xenogeneic, or allogeneic source. Individual differences arise in technologies based on hepatocyte sources, methods of filtration, efficacies, and safety.

Spheroid Reservoir Bioartificial Liver (SRBAL)

The spheroid reservoir bioartificial liver (SRBAL) is a tissue engineering solution that has shown tremendous promise in the treatment of ALF.1 While still pre-clinical, this solution will likely soon be ready to enter human testing. SRBAL devices employ primary porcine hepatocytes.1 Extracted and digested from donor animals, they are cultured in vitro forming spheroids (cell clusters). Spheroids maintain the ability to be cryopreserved with approximately 65% viability, and show at least 24-hour efficacy during treatment.1 In late 2015, a major landmark in SRBAL technology was reached when researchers demonstrated significant efficacy of treatment in an induced-ALF porcine model.1 Subjects in the experimental arm showed substantial reductions in circulating ammonia, effectively reversing their hyperammonemia-induced encephalopathy, and other complications.1 More recently, SRBAL technology was tested in rhesus monkeys with similar effects.14 Once again using primary porcine hepatocytes as the device workhorse, rhesus monkeys with induced ALF showed significant improvement in ALF-associated complications.14


Originally designed for primary porcine hepatocytes, the Academic Medical Centre of Amsterdam Bioartificial Liver (AMC-BAL) in its most recent iteration is showing good efficacy with the HepaRG-progenitor hepatocyte line.18, 19 Contrary to most BAL devices, the AMC-BAL puts patient plasma in direct contact with hepatocytes, a much more dangerous, but also more efficient practice.18, 19 Patient plasma is filtered from blood via a plasmaphoresis device, and cycled directly through the bioreactor as affluent media.18 Effluent media is then returned to plasmaphoresis circuit, and eventually the patient. Within the bioreactor, hepatocytes are oxygenated directly by a series of capillaries with a free flowing oxygen-carbon dioxide gas mixture, and plasma is re-heated to maintain a constant temperature in the extracorporeal environment.18 The switch from primary hepatic to immortalized human hepatic lines was made due to regulatory crackdowns in Europe on xenotransplantation in recent years, including extracorporeal medical devices employing xenogeneic tissues.18 With efficacy demonstrated almost immediately in this technology, recent studies are now focused on identifying scale up challenges with the HepaRG-progenitor (HRGP) implementation of the technology.18


It is extremely likely that future solutions to ALF will be based on biological external liver support devices or a combination of biological and non-biological devices. Some clinicians suggest employing a modular support system, to be tailored to specific patients, to reduce load on BAL hepatocytes.9 Simply because in vivo adult hepatocytes perform so many functions, it is unlikely in the near future that a human liver will be replaced by a completely non-biological system.20 Moving forward with Biological solutions however does have its challenges. Setting aside challenges with demonstrating safety and efficacy (both of which are not insignificant), an almost insurmountable challenge with current technology is scaling the device for human and widespread use. Noted in this section are particular challenges and iterations of upcoming biological technologies in the ELSD field.

BALs – Sources of Hepatocytes

There is some argument to be had regarding the source of hepatocytes to be used within a bioartificial liver. Both primary hepatocytes and immortalized lines have been used with good efficacy and safety. Primary hepatocytes quite obviously remain the most efficacious for treatment of ALF, simply because they are the closest phenotypically to an in vivo liver.5

Primary Hepatocytes

In bioartificial liver systems, the gold standard source of primary hepatocytes is quite obviously human hepatocytes of an autogenic origin, as this would pose no infection or immunological risk to the patient, while providing them with an exact hepatocyte match and biochemical benefits that go with it. This approach however has not yet been studied in the context of extracorporeal bioreactors, as the technology to expand a primary population from a biopsied damaged liver to a suitable passaged number, while maintaining phenotype does not exist. Therefore, the true “gold standard” in use, is the isolation of primary allogeneic hepatocytes. There are few devices employing primary hepatocytes of an allogeneic origin; the only sources are discarded donor organs.7 These donor hepatocytes are often high in variability and viability, making them unsuitable for guaranteed device efficacy.8 Guaranteed efficacious allogeneic sources of hepatocytes in BALs are mostly limited to immortalized lines, with the possibility of employing cell expansion from healthy donor biopsies. Because of poor availability in allogeneic sources, the highly available xenogeneic (primarily porcine) primary hepatocyte approach has been used for many of these devices. Using primary cells, no matter the origin remains the most efficacious for treatment, since these hepatocytes most closely resemble the original phenotype and differential gene expression patterns of in vivo hepatocytes.7 They are however subjected to the same limitations as all tissue engineering therapies: loss of phenotype, dedifferentiation, and, if trying to expand from a small source, the Hayflick limit. Because of culturing limitations, and risks associated with xenogeneic cells, some models have opted to pursue an immortalized approach.

Immortalized Lines

Immortalized lines like the HepG2/C3A human hepatoblastoma line, while an infinite source of rapidly-expanding hepatocytes, are subject to the same limitations as other cultured human cells; dedifferentiation and poor representation of original phenotype.5 While these cells are still able to perform some aspects of normal tissue homeostasis, their ability to employ the urea cycle and remove ammonia is significantly reduced, much like all immortalized lines.5 A more promising candidate however is the HepaRG-progenitor immortalized line, as currently used in the AMC-BAL. This line shows considerable efficacy in performing regular human liver functions, with a diminished loss of original phenotype as compared to other hepatoblastoma lines.18

Risks of Infection & Immune Response In BALs

By using a xenogeneic model, there is inherent risk of a cross-species immune response. Even in a best case scenario where no live hepatocyte makes it across the membrane into the patient’s circulation, proteins produced by the BAL construct may enter circulation and trigger an immune response in the host.4 An even greater risk arises when using porcine hepatocytes in the form of porcine endogenous retroviruses (PERVs) or other zoonoses.4 While mostly inactive in humans, endogenous retroviruses in pigs pose a significant zoonotic risk as they maintain the ability to bud and infect cells of a different species – the patient.4 While a recent study of SRBAL treated rhesus monkeys showed no PERV infection in treated animals, these retroviruses remain a large risk to health, and barrier to clinical trials and regulatory approval.14, 18 Since SRBAL devices require the extraction of primary hepatocytes, pigs sacrificed for human treatments should first be Specified-Pathogen-Free (SPF), and subsequently be specifically tested for the presence of these PERVs, as they are genome incorporated.7

Not only is there an immune risk to the host, but also to BAL devices. While not as important as the health of the patient, the patient’s immune system has the potential to reduce efficacy in the device by launching a cytotoxic immune attack on the BAL hepatocytes.6 In a pre-clinical trial of a SRBAL, the size of the membrane between the patient’s blood circuit and the spheroid circuit significantly influenced the amount of cytotoxicity seen in porcine hepatocytes.6 After 3 hours of treatment in a healthy dog model, SRBAL hepatocytes of a large 200nm pore showed increased levels of dog IgG and IgM as compared to the small 400kDa model.6

Risks associated with using immortalized hepatocyte lines are those that are shared by the regenerative medicine and induced stem cell therapy communities; neoplasia. While unlikely in most models due to filter sizes, the potential does still exist for immortalized lines to transmigrate and enter patient circulation, potentially initiating a neoplastic event. This event remains a very real possibility in the case of the newest iteration of the AMC-BAL device; which puts patient plasma in direct contact to proliferative hepatocytes of human origin.18 In recent studies, the proliferative HepaRG-progenitor line was subcutaneously injected in immunocompromised mice, along with a HeLa line as a positive control.18 Histological examination revealed no neoplasia in the HepaRG-progenitor line.18 A concern however with this study is its statistical power, and length of follow-up, both of which were low.18

Limitations And Futures of Current Therapies


In devices that have undergone clinical testing, efforts are now being made to identify scale up procedures and challenges. The newest iteration of the AMC-BAL is looking to scale up the device to support a 70kg human.18 This is estimated to require 250-300g of tissue, or 2.5-50m2 of cells in monolayer.18 The current maximal production however is around 40g of cells in a 540mL bioreactor; approximately 5% of the mass of a functional liver.18 Current limitations however make the culturing of 300g of tissue an unfeasible goal for commercial application, so thought is now being put into modular therapies to potentially reduce the amount of live tissue needed, by offsetting it with non-biological solutions.


Preservation problems are especially seen in BAL’s that employ primary hepatocytes. Populations can be cryopreserved, but often with poor survival rates with standard cryopreservation procedure.18 Cryopreservation of immortalized lines is more promising; AMC-BAL researchers have shown that cryopreserved HRGP hepatocytes can be placed directly into the device without any significant loss of function.18 Furthermore, these cells maintain the ability to undergo two population doublings before any loss of efficacy.18

Device Transportation

Physical transportation of BAL devices remains another important problem to solve. At current states of technology, it is unrealistic to expect that all medical centers will culture hepatocytes on site. Therefore, a BAL device must have the ability to be transported not only within a hospital, but between them. As device-resident cells need to be kept alive, an autonomous solution is needed. In order to meet these transportation requirements, it was proposed a device must function autonomously for a period of 48 hours, while simultaneously logging crucial data about the health of the culture.18 Recent studies with the AMC-BAL have been simulating just that; the physical transport of a device in real conditions. Specialized bioreactors designed specifically for transport were subjected to harsh conditions for a period of 24 hours. These conditions included loading onto a vehicle, unloading, driving over uneven terrain, and being left unattended at low temperatures (4oC).18 Bioreactors operated on an independent power supply (batteries), with additional thermoregulation (Figure 1).18 Hepatic functions after simulated transportation did not differ significantly from before, indicating the device could likely be transported safely for a period of at least 24 hours.18



Significant advances in non-biological technologies have been made in recent years, but it is unlikely that non-biological solutions can provide the full function of a damaged liver. For this reason, future treatments and commercial solutions for ALF, chronic-on-acute liver failure, and other hepatotoxic events will likely rely on a BAL, or BAL-non-biological combinatory effort. Non-biological systems have been widely tested in clinical environments, and approved for commercial use by many regulatory bodies. Biological systems however are still in their infancy. The BAL field of tissue engineering is young, but shows tremendous promise for a medical problem that has yet to obtain a definitive solution. Efficacy in primary systems has been demonstrated time and time again. It is therefore likely that the field will in the next coming years focus on guaranteeing patient safety from infection, and pursuing methods for increasing efficacy of immortalized-line based allogeneic therapies.  With the BAL field now gearing towards challenges associated with scale up and transportation, it is hoped that these devices will soon be making a tangible difference to ALF patients in the next 15 years.


  1. Glorioso, J. M., Mao, S. A., Rodysill, B., Mounajjed, T., Kremers, W. K., Elgilani, F., … & Nyberg, S. L. (2015). Pivotal preclinical trial of the spheroid reservoir bioartificial liver. Journal of hepatology, 63(2), 388-398.
  2. Bernal, W., & Wendon, J. (2013). Acute liver failure. New England Journal of Medicine, 369(26), 2525-2534.
  3. Plevris, J. N., Schina, M., & Hayes, P. C. (1998). the management of acute liver failure. Alimentary pharmacology & therapeutics, 12(5), 405-418.
  4. Denner, J. (2016). How active are porcine endogenous retroviruses (PERVs)?. Viruses, 8(8), 215.
  5. Nyberg, S. L., Remmel, R. P., Mann, H. J., Peshwa, M. V., Hu, W. S., & Cerra, F. B. (1994). Primary hepatocytes outperform Hep G2 cells as the source of biotransformation functions in a bioartificial liver. Annals of surgery, 220(1), 59.
  6. Nyberg, S. L., Amiot, B., Hardin, J., Baskin-Bey, E., & Platt, J. L. (2004). Cytotoxic immune response to a xenogeneic bioartificial liver. Cell transplantation, 13(7-1), 783-791.
  7. van de Kerkhove, M. P., Chamuleau, R. A. F. M., & Van Gulik, T. M. (2003). Clinical application of bioartificial liver support systems. In Encephalopathy and nitrogen metabolism in liver failure (pp. 389-406). Springer Netherlands.
  8. Sauer, I. M., Zeilinger, K., Obermayer, N., Pless, G., Grünwald, A., Pascher, A., … & Mas, A. (2002). Primary human liver cells as source for modular extracorporeal liver support–a preliminary report. The International journal of artificial organs, 25(10), 1001-1005.
  9. Sauer, I. M., & Gerlach, J. C. (2002). Modular extracorporeal liver support. Artificial organs, 26(8), 703-706.
  10. Stange, J., Mitzner, S. R., Risler, T., Erley, C. M., Lauchart, W., Goehl, H., … & Lohr, M. (1999). Molecular adsorbent recycling system (MARS): clinical results of a new membrane-based blood purification system for bioartificial liver support. ARTIFICIAL ORGANS-OHIO-, 23, 319-330.
  11. Hanish, S. I., Stein, D. M., Scalea, J. R., Essien, E. O., Thurman, P., Hutson, W. R., … & Scalea, T. M. (2017). Molecular adsorbent recirculating system effectively replaces hepatic function in severe acute liver failure. Annals of Surgery, 266(4), 677-684.
  12. Rifai, K., Ernst, T., Kretschmer, U., Bahr, M. J., Schneider, A., Hafer, C., … & Fliser, D. (2003). Prometheus®–a new extracorporeal system for the treatment of liver failure. Journal of hepatology, 39(6), 984-990.
  13. Falkenhagen, D., Strobl, W., Vogt, G., Schrefl, A., Linsberger, I., Gerner, F. J., & Schoenhofen, M. (1999). Fractionated plasma separation and adsorption system: a novel system for blood purification to remove albumin bound substances. Artificial organs, 23(1), 81-86.
  14. Li, Y., Wang, Y., Wu, Q., Nyberg, S. L., Amiot, B., Bao, J., & Bu, H. (2017, October). Evaluation of spheroid reservoir bioartificial liver with porcine hepatocytes in rhesus monkey model of acute liver failure. In HEPATOLOGY (Vol. 66, pp. 147A-147A). 111 RIVER ST, HOBOKEN 07030-5774, NJ USA: WILEY.
  15. Henschel, B., Schmid, R., & Huber, W. (2015). First clinical experience with a new type of albumin dialysis: the HepaWash® system. Critical Care, 19(1), P383.
  16. Al-Chalabi, A., Matevossian, E., Preissel, A. K., Yan, H., Geiger, A., Nairz, E., … & Kreymann, B. (2010). Survival improvement in pigs with liver failure and superimposed sepsis by a new liver support system (Hepa Wash®). Critical Care, 14(1), P508.
  17. Al-Chalabi, A., Matevossian, E., v Thaden, A. K., Luppa, P., Neiss, A., Schuster, T., … & Perren, A. (2013). Evaluation of the Hepa Wash® treatment in pigs with acute liver failure. BMC gastroenterology, 13(1), 83.
  18. van Wenum, M., Treskes, P., Tang, C. Y., Coppens, E. J., Jansen, K., Hendriks, E. J., … & Hoekstra, R. (2017). Scaling-up of a HepaRG progenitor cell based bioartificial liver: optimization for clinical application and transport. Biofabrication, 9(3), 035001.
  19. P van de Kerkhove, M & Di Florio, E & Scuderi, Vincenzo & Mancini, A & Belli, Antonello & Bracco, A & Dauri, Mario & Tisone, Giuseppe & Di Nicuolo, G & Amoroso, P & Spadari, A & Lombardi, G & Hoekstra, Ruurdtje & Calise, Fulvio & Chamuleau, Robert. (2002). Phase I clinical trial with the AMC-bioartificial liver. The International journal of artificial organs. 25. 950-9.
  20. van Wenum, M., Chamuleau, R. A., van Gulik, T. M., Siliakus, A., Seppen, J., & Hoekstra, R. (2014). Bioartificial livers in vitro and in vivo: tailoring biocomponents to the expanding variety of applications. Expert opinion on biological therapy, 14(12), 1745-1760.
  21. Sussman, N. L., & Kelly, J. H. (1993). Improved liver function following treatment with an extracorporeal liver assist device. Artificial organs, 17(1), 27-30.