Liver disease is a growing burden. Transplant organs are scarce. Extracorporeal liver support systems (ELSS) are a bridge to transplantation for eligible patients. For transplant-ineligible patients the objective becomes liver recovery.

We review seven decades of non-cell-based ELSS research in humans. Where possible, we emphasize randomized controlled trials (RCTs). When RCTs are not available, we describe the available human clinical data.

There are three broad cell-free approaches to remove protein-bound toxins (PBTs) and treat liver failure. The first is a dialysate binder suspension. A material that binds the PBT (the binder) is added to the dialysate. Binders include albumin, charcoal, and polystyrene sulfonate sodium. The unbound fraction of the PBT crosses the dialyzer membrane along a chemical gradient and binds to the binder. The second approach is using grains of sorbent fixed in a plastic housing to remove PBTs. Toxin-laden blood or plasma flows directly through the column. Toxins are removed by binding to the sorbent. The third approach is exchanging toxin-laden blood, or fractions of blood, for a healthy donor blood product. Most systems lack widespread acceptance, but plasma exchange (PE) is recommended in many guidelines. The large donor plasma requirement of PE creates demand for systems to complement or replace it.

Now that PE has become recommended in some, but not all, jurisdictions, we discuss the importance of reporting precise PE protocols and dose. Our work provides an overview of promising new systems and lessons from old technologies to enable ELSS improvement.

Renal replacement therapy (RRT) is frequently required to manage critically ill patients with acute kidney injury (AKI). There is limited evidence to support the current practice of RRT in intensive care units (ICUs). Recently published randomized control trials (RCTs) have further questioned our understanding of RRT in critical care. The optimal timing and dosing continues to be debatable; however, current evidence suggests delayed strategy with less intensive dosing when utilising RRT. Various modes of RRT are complementary to each other with no definite benefits to mortality or renal function preservation. Choice of anticoagulation remains regional citrate anticoagulation in continuous renal replacement therapy (CRRT) with lower bleeding risk when compared with heparin. RRT can be used to support resistant cardiac failure, but evolving therapies such as haemoperfusion are currently not recommended in sepsis.

Liver failure is associated with a high morbidity and mortality rate and is the seventh leading cause of death worldwide. Orthotopic liver transplantation remains the definitive treatment; however, because of the limited number of available organs many patients expire while on the transplant list. Currently, there are no established means for providing liver support as a means of bridging patients to transplantation or allowing for recovery from liver injury. Analogous to the clinical situation of renal failure, there is great interest in developing liver support systems that replace the metabolic and waste removal functions of the liver. These support systems are of two general types: artificial and bioartificial livers. In this review, based on a presentation from the 57th American Society of Artificial Internal Organs Annual Meeting (Washington, D.C., June 2011), we review the current status of liver support systems.

Extracorporeal support has been advocated for patients with acute and chronic liver failure. Patients with acute liver failure and those with decompensated cirrhosis can be broadly divided into two groups. The first group comprises those with acute liver failure and ongoing hepatic necrosis, and the second, those with long-standing chronic decompensation admitted with one or more complications of liver failure, such as encephalopathy without any evidence of a precipitating factor or accompanying acute deterioration of liver function. This second group includes patients with acute liver failure, where the insult causing hepatic necrosis has been resolved, and those patients with chronic decompensation who suffer another insult to the liver, such as acute infection or variceal hemorrhage that causes further liver injury in the setting of multiorgan failure. These two groups are likely to have different outcomes and may need to be managed differently. In the first group, liver transplantation is the only possible long-term therapeutic option, whereas in the second group, other possibilities such as extracorporeal liver support systems and/or medical therapy may allow these patients to return to their previous state before the acute insult. Over time extracorporeal support has expanded from simple peritoneal dialysis and hemodialysis, to the development of circuits designed primarily to remove both water and lipid-soluble toxins and, in addition, bioartificial devices to provide replacement synthetic hepatic function. Because many of the patients with an acute liver insult have ongoing chronic liver disease and develop hepatorenal syndrome, this group of patients has been targeted by several groups to study the role of liver support systems.

Devices for support of patients with liver failure are of two types: bioartificial livers and artificial livers. Bioartificial livers include hepatocytes in bioreactors to provide both excretory and synthetic liver functions. Artificial livers use nonliving components to remove toxins of liver failure, supply nutrients and macromolecules. Current artificial liver devices use columns or suspensions of sorbents (including adsorbents and absorbents) to selectively remove toxins and regenerate dialysate, albumin-containing dialysate, plasma filtrate or plasma. This article reviews three artificial liver devices. Liver Dialysis uses a suspension of charcoal and cation exchangers to regenerate dialysate. MARS uses charcoal and an anion exchanger to regenerate dialysate with albumin. Prometheus uses neutral and anion exchange resins to regenerate a plasma filtrate containing albumin and small globulins. We review the operating principles, chemical effects, clinical effects and complications of use of each type of artificial liver. These devices clearly improve the clinical condition of patients with acute or acute-on-chronic liver failure. Further randomized outcome studies are necessary to prove clinical outcome benefit of the artificial liver support devices, and define what types of patients appear most amenable to therapy.

Enthusiasm for liver support devices, particularly cell-based biological systems and albumin dialysis, increased over the last decade and there has been considerable clinical activity both within and without the construct of clinical trials. Most data have been generated on patients with acute liver failure or in patients with decompensation of chronic liver disease, often referred to as acute-on-chronic liver failure. In acute liver failure liver, liver support devices are more realistically being used as a 'bridge' to liver transplantation rather than to transplant-free survival. In acute-on-chronic liver failure the clinical objective of attaining clinical stability with treatment appears more achievable. The so-called bioartificial liver device, based on porcine hepatocytes, is the most extensively evaluated biological device. A sizeable clinical trial failed to demonstrate efficacy, but secondary analyses suggest it would be unwise to assume futility had been established with this device. Molecular adsorbent recirculating system leads the way in the non-biological category in terms of the number of patients treated, but data from large clinical trials are not yet available. One of the strongest conclusions of this review is that the amount of high-quality data available on liver support devices dramatically understates the effort and money that have been expended in their assessment. It is very clear that randomized controlled trials are mandatory to establish clinical efficacy, but it is less clear how the ideal trial should be constructed.

Extracorporeal devices have had limited effectiveness in liver failure due to consequences of inadequate anticoagulation. The purpose of this study was to evaluate an anticoagulation protocol developed for Liver Dialysis Unit (LDU) treatments. Twelve patients underwent 19 LDU treatments for acetaminophen overdose (n = 1), subacute liver failure (n = 1), and refractory encephalopathy in cirrhosis (n = 10). The initial 6 patients (group 1) were treated according to the manufacturer's recommendations. The subsequent 6 patients (group 2) were treated using a formal heparin anticoagulation protocol that included 2000 units in prime solution and 30 units/kg induction, activated clotting time (ACT) measurements, and heparin administered by dose-response curve to maintain 300-second ACT. Protamine reversal was used. Treatments were well tolerated and equally effective in both groups. Adequate (ACT > or = 250 seconds) and therapeutic (ACT 250-350 seconds) anticoagulation was maintained more consistently in group 2 (90.9% vs 50.0%, p < 0.0001; and 77.4% vs 45.8%, p < 0.001). There was less consumption of fibrinogen (12.1% vs 43.3%, p < 0.005) and platelets (13.8% vs 29.9%, p < 0.05) and a trend for less blood product used (8.3 vs 15.8 units/patient, p = 0.13) and fewer complications (16.7% vs 66.7%, p = 0.15) in group 2. Anticoagulation using ACT monitoring, heparin dose-response curves, and a target ACT of 300 seconds improves the safety of LDU treatments.

Hepatic failure is a significant medical problem which has been unsuccessfully treated by hemodialysis. However, similar therapies using recirculated dialysate regenerated by sorbents in place of single-pass dialysate have been beneficial in treating acute-on-chronic liver failure. The advantages of sorbent-based treatments include some selectivity of toxin removal and improved removal of protein-bound toxins. Activated carbon has been extensively used in detoxification systems, but has often had insufficient toxin capacity. Powdered activated carbon, because of its large surface area, can provide greater binding capacity for bilirubin and other toxins than granular carbon commonly used in detoxifying columns. Methods of using powdered carbon in extracorporeal blood treatment devices are reviewed in the present paper, including liver dialysis and a new sorbent suspension reactor (SSR); and the abilities and limitations of the SSR and columns to process protein solutions are discussed.

Because acute liver cell failure is associated with an exceedingly high mortality, liver support has been proposed since the 1950s to improve patient outcome. Early devices, including hemodialysis, hemofiltration, exchange transfusion, plasmapheresis, hemoperfusion, plasma and cross-hemodialysis or cross-circulation, appeared inefficient. Meanwhile, documented results of extracorporeal liver perfusion (ECLP) suggested its superiority over conventional treatment. These devices were abandoned with the development of liver transplantation (LT), which allowed a better outcome and longer survival rate. In the present day, the fact that patients die while waiting for LT because of organ shortage led to a renewed interest in liver support as bridge to LT or regeneration. These devices can be classified according to the presence or lack of hepatocytes, whereas biologic devices refers to the presence of cells or other organic and biochemical component. The absence of individual success of early models led to the development of combined hepatocyte free devices, or artificial liver, which are based upon the hemodiabsorption principle (Biologic-DT) or on the "albumin bound toxin hypothesis" (Molecular Adsorbents Recirculating System) with encouraging results. Meanwhile, hepatocyte based bioartificial liver devices (BLD) were conceived for a global "metabolic support." BLD were developed with the use of human hepatoma cell line (C3A) or primary or cryopreserved porcine hepatocytes. Preliminary experience gave promising results bridging patients to LT. Based upon the same principle of global hepatocyte metabolic support, ECLP regained interest, particularly with the development of transgenic pigs. Several concerns were raised about these devices. Artificial livers lacked any metabolic synthetic activity, the use of human liver for ECLP seems hardly acceptable because of organ shortage, and the accepted use of borderline livers for transplantation is pending trials for the use of xenogenic livers. For BLD, the concerns were the low hepatocyte mass, the absence of accessory liver cells, and the potential risk of seeding tumor cells into patient with the use of human hepatoma cell line. The use of porcine hepatocytes (BLD or ECLP) raised physiologic and immunologic concerns and particularly the fear of a possible transfer of porcine viral material. Although recent studies clearly demonstrate clinical improvement of patients with the use of recently developed liver support devices, most of reported prospective, controlled, or randomized trials had a small number of patients. To give the deciding vote and avoid previous pitfalls, trials need to be developed with a larger number of patients based upon statistically significant models with the following characteristics: 1) comprehensive understanding of the acute liver cell failure mechanisms, 2) world wide classification of conditions that require liver support, and 3) a clear definition of treatment success pending patients to LT or recovery without transplantation. There has not yet been conclusive evidence to support the benefits of extracorporeal liver support. We are still waiting for the deciding vote.

Episodic (acute) type C hepatic encephalopathy (AHE) fails to respond to 5 days of medical therapy in 10-30% of patients and carries a 10-30% mortality rate. We prospectively studied extracorporeal liver support for AHE failing to respond to medical therapy to assess its safety and efficacy and the role of anticoagulation.

A series of patients with cirrhosis and AHE failing to respond to at least 24 h of medical therapy underwent a maximum of three 6-h charcoal-based hemodiabsorption (Liver Dialysis Unit) treatments. A standard anticoagulation protocol, with heparin dosing based on activated clotting time (ACT) determinations, heparin dose-response curve, and target ACT of 275-300 s, was developed. Therapy was terminated if patients met a predetermined clinical response, deteriorated, or underwent transplantation.

Eighteen patients with grade 2-4 AHE despite 5.9 +/- 3.9 days of medical therapy underwent a mean of 1.6 treatments. In 2.6 +/- 1.9 days, 16 patients (88.9%) improved to less than grade 2 HE or achieved at least a 50% hepatic encephalopathy index (HEI) reduction. Median mental status (grade 2 vs 1, p < 0.05) and HEI (0.634 +/- 0.194 vs 0.363 +/- 0.263, p < 0.005) improved significantly. Survival was 94.4% and 72.2% at 5 and 30 days, respectively. Use of our developed anticoagulation protocol resulted in less platelet (14.2% +/- 2.8% vs 32.5% +/- 5.8%, p < 0.005) and fibrinogen consumption (12.1% +/- 3.5% vs 43.3% +/- 8.6%, p < 0.0005) and blood product use (6.2 +/- 1.8 vs 19.0 +/- 5.6 units, p < 0.05) compared with treatments according to manufacturer's guidelines.

Charcoal-based hemodiabsorption treatments in which a standardized anticoagulation protocol is used is safe and effective treatment for AHE not responding to standard medical therapy.