|Home | About | Journals | Submit | Contact Us | Français|
Acute liver failure (ALF) is a rare but challenging clinical syndrome with multiple causes; a specific etiology cannot be identified in 15% of adult and 50% of pediatric cases. The course of ALF is variable and the mortality rate is high. Liver transplantation is the only therapy of proven benefit, but the rapidity of progression and the variable course of ALF limit its use. Currently in the United States, spontaneous survival occurs in approximately 45%, liver transplantation in 25%, and death without transplantation in 30% of adults with ALF. Higher rates of spontaneous recovery (56%) and transplantation (31%) with lower rates of death (13%) occur in children. The outcome of ALF varies by etiology, favorable prognoses being found with acetaminophen overdose, hepatitis A, and ischemia (≈60% spontaneous survival), and poor prognoses with drug-induced ALF, hepatitis B, and indeterminate cases (≈25% spontaneous survival). Excellent intensive care is critical in management of patients with ALF. Nonspecific therapies are of unproven benefit. Future possible therapeutic approaches include N-acetylcysteine, hypothermia, liver assist devices, and hepatocyte transplantation. Advances in stem cell research may allow provision of cells for bioartificial liver support. ALF presents many challenging opportunities in both clinical and basic research.
Acute liver failure (ALF) is a dramatic but rare clinical syndrome marked by the sudden loss of hepatic function in a person with no prior history of liver disease. The causes of ALF include viral hepatitis, drug-induced and toxin-induced liver disease, metabolic errors, ischemia, and miscellaneous rare causes. Currently, there are no specific therapies of proven benefit except for emergency liver transplantation. The many challenges in understanding and management of ALF led to the organization of a 2-day research workshop held on December 4–5, 2006 in Bethesda, Maryland, sponsored by the National Institute of Diabetes and Digestive and Kidney Diseases (NIDDK) with support from the National Institute of Biomedical Imaging and Bioengineering (NIBIB) and the Office of Rare Diseases of the National Institutes of Health (NIH). This manuscript summarizes the presentations at that meeting and the recommendations arising therefrom.
The sudden loss of hepatic function in a person without preexisting liver disease defines ALF.1,2 The most reliable signs of severe acute liver injury are the presence of coagulopathy (international normalized ratio [INR] ≥ 1.5) and any degree of hepatic encephalopathy, the length of illness being considered anything ≤ 24 weeks. Many patients evolve to coma within 1 week or less. The term “acute liver failure” is preferable to fulminant hepatic failure1 or acute hepatic necrosis.2 ALF is rare and represents a syndrome rather than a specific disease, having multiple causes that vary in course and outcome.
The rarity of ALF and its unpredictable, severe course make it a challenging entity for prospective studies. ALF is difficult to identify in its early stages, resulting in frequent delays in initiation of treatment. Deliberate decision-making is often impossible. For these reasons, the U.S. ALF Study Group was established with pilot funding from the Food and Drug Administration (FDA) and full support from the NIH.3–6 Its two aims were to utilize a database for prospective, standardized collection of clinical information, serum, and DNA and tissue samples and to develop controlled trials of innovative therapies.
Between January 1998 and July 2007, the adult ALF Study Group enrolled 1,147 patients at 23 clinical sites. Overall, 67% of cases were women and the average age was 38 years (range 17–79 years). The most frequent single cause of ALF was acetaminophen overdose (Fig. 1) accounting for 46% of cases. Over the 9-year period, there was a gradual increase in the proportion of cases of ALF due to acetaminophen overdose, and somewhat decreasing rates of hepatitis A and B. The distribution of etiologies differs around the world. In Africa and Asia, viral hepatitis is the leading cause of ALF, and cases of hepatitis E as well as hepatitis A and B are common.7
In the U.S. ALF Study Group database, the overall outcomes included spontaneous recovery without need for transplantation in 45% of the patients. Of the 44% of patients who were listed for transplantation, 25% of the overall group received a graft; of the remaining 19%, 10% died on the waiting list while 9% recovered. Overall, 30% died. Outcome rates are far better than in the era before transplants when mortality rates were 80% or higher. Changing patterns of etiology accounted for some outcome differences (Table 1). Acetaminophen, shock, and hepatitis A have favorable outcomes, with spontaneous recovery rates of 58% to 64%, compared to drug-induced, autoimmune, and indeterminate ALF (≈20% to 25%). The median time from listing to transplant was typically one day (range <1–4 days) compared to a time from listing to death without transplantation of 3 days (range <1–6 days), suggesting that organ shortages limit transplantation of ALF patients. While overall outcomes have improved since the era before transplants, ALF remains a challenging syndrome with high mortality and a major burden to the medical care system.
The Pediatric ALF Study Group5 developed in 2000 comprises 19 clinical centers in the United States (n = 16), Canada (n = 1), and the United Kingdom (n = 2); it maintains a similar registry of ALF in children and is conducting a randomized controlled trial of N-acetylcysteine (NAC) for non–acetaminophen-related ALF.
In children, signs of hepatic encephalopathy can be subtle and appear late in the course of injury.5,8 Accordingly, modified criteria for ALF are used: either encephalopathy with coagulopathy (INR > 1.5) or more severe coagulopathy alone (INR > 2.0) in a child with evidence of acute liver injury, thus allowing enrollment to occur at an earlier point in the illness.
Currently, the Pediatric ALF database has enrolled 548 cases, 38% of which occurred in children less than 3 years of age. Etiologies were: acetaminophen overdose in 12% of cases, metabolic causes in 10% of cases, viral hepatitis in 6% of cases, ischemia in 4% of cases, non-acetaminophen drug-induced liver injury in 4% of cases, and other causes in 15% of cases; 49% were considered indeterminate. The distribution of etiologies differed from adults and varied by age (Figs. 2, ,33).5,8,9 In older children, acetaminophen toxicity and Wilson disease were more frequent. Unusual viral causes predominated. There was no hepatitis B, 1 hepatitis C, and 5 hepatitis A cases, but 11 herpes simplex virus, 7 Epstein-Barr virus, and 1 instance each of adenovirus, enterovirus, paramyxovirus, and influenza A infections. There were 21 cases of idiosyncratic drug-induced liver disease.
Outcomes of ALF were more favorable in children than in adults.10 Overall, 56% survived short-term without transplantation, 31% received a graft, and 13% died. Importantly, 16% of children with ALF who never developed hepatic encephalopathy either died or required liver transplantation, providing further support to use of the modified ALF criteria. Diagnoses associated with high rates of survival included acetaminophen toxicity (93%)5,11 and hepatitis A (100%).
The pediatric trial of NAC for non–acetaminophen-induced ALF is ongoing. Other areas of research focus include investigation of the cause of indeterminate cases of ALF5 and better delineation of the metabolic causes of liver injury in children.12
The leading causes of death in ALF are cerebral edema and sepsis.4,13 Respiratory distress requiring mechanical ventilation and/or acute renal failure are also common as are bacterial and fungal infections.14 Coagulopathy itself rarely is life threatening and can be corrected.
In the ALF Study Group patients, 44% were listed for transplantation, but only 25% of the entire group received a transplant. The short-term posttransplant outcome was good, with 90% survival at 1 month and 70% survival at 1 year.4,13 Long-term outcome after ALF is not well defined.15,16 Concerns include neurological damage, complications of intensive care, and persistent or recurrent liver disease following spontaneous recovery or liver transplantation. Early detection of cerebral edema is difficult, because clinical signs such as posturing and pupillary dilatation are signs of advanced cerebral edema. Long-term neurological impairment due to cerebral edema is well recognized and includes developmental delays and impaired cognition of varying degrees. If evidence of cerebral edema is present in children or adults at the time of transplant, developmental delay is frequent but probably not inevitable.15–18
Recurrence of ALF after spontaneous recovery occurs with repeated acetaminophen overdoses.19,20 Other forms of liver disease rarely recur, except for posttransplant autoimmune hepatitis.21 More rigorous long-term follow up studies are needed to better guide management of cerebral edema and the timing of liver transplantation in ALF.22
ALF requires an expert, multidisciplinary, collaborative effort among hepatologists, transplant surgeons, intensive care physicians and nurses, nephrologists, neurosurgeons, and transplant coordinators.23 Patients should be rapidly evaluated for etiology and severity of liver injury and an urgent assessment made regarding suitability for liver transplantation.
Specific therapies for unique causes of ALF are few: NAC for acetaminophen overdose and prompt delivery for pregnancy-related ALF. Children with inborn errors of metabolism appear to benefit by removal of galactose-containing formula in children with galactosemia, intravenous glucose and avoidance of fasting for children with inherited defects in fatty acid oxidation, and prompt administration of 2-(2-nitro-4-3 trifluoro-methylbenzoyl)-1,3-cyclohexanedione (NTBC) in children with hereditary tyrosinemia type 1 24. Other therapies in use but of unproven benefit include activated charcoal and high-dose intravenous penicillin for mushroom poisoning;25 corticosteroids for autoimmune hepatitis;2 copper chelation, plasmapheresis, and antioxidant therapy for Wilson disease;27 lamivudine or entecavir for acute hepatitis B;28 acyclovir for herpes simplex virus infection;29 hemodynamic support for shock or ischemic liver injury; and decompressive surgery or transjugular intrahepatic portosystemic shunts (TIPS) for acute Budd-Chiari syndrome.30 Currently, the only effective therapy of ALF is liver transplantation. Transfer to a transplant center is appropriate for any patient with severe coagulopathy during an acute hepatic illness, regardless of whether hepatic encephalopathy is present. Although a number of prognostic scores have been developed to predict outcome without transplantation, none is considered standard at present. Current United Network for Organ Sharing (UNOS) criteria for Status 1 listing are (1) onset of any degree of hepatic encephalopathy within 8 weeks of onset of acute liver injury; (2) absence of pre-existing liver disease; (3) life expectancy of less than 7 days; and (4) in the intensive care unit (ICU) requiring either mechanical ventilation, renal dialysis, or with severe coagulopathy (INR > 2.0). When an organ becomes available, the difficult but critical decision is made. Bad outcomes are possible: performing transplantation on someone who might have survived, failing to perform transplantation on someone who ultimately succumbs, or worse, performing transplantation on someone too late and losing both the patient and the donor organ.
Special dilemmas related to ALF management are many: when liver ultrasound suggests cirrhosis, this may be false positive due to a nodular necrotic liver. Likewise, patients who are known hepatitis B carriers are sometimes inappropriately excluded from listing because of pre-existing liver disease.31 Intracranial pressure monitoring remains controversial, because it is invasive and complications occur.22 Evidence for efficacy of NAC outside of early acetaminophen poisoning awaits results from the ALF Study Group trial.32 Nonspecific therapies of promise that deserve future evaluation include inhibitors of liver injury and apoptotic pathways (caspase or specific kinase inhibitors) and promoters of liver regeneration (growth factors).33
Transplant candidacy must be determined rapidly once the diagnosis of ALF is made. Underlying medical or psychosocial conditions that may preclude consideration include alcohol or drug abuse, repeated suicide attempts, or inadequate family support. Because of the rapid progression of ALF and persistent organ shortages, improved predictive models and more donor organs are sorely needed.34 Outcome data from the European experience have been presented recently.35 Between 1995 and 2005, 90,445 candidates for liver transplantation were listed in the United States. During this time, ALF accounted for 3.9% of overall listings.36
Rates of survival after liver transplantation were similar between patients with ALF or end-stage liver disease (Fig. 4). Although survival was decreased during the first year after transplantation, at 4 years, the ALF cohort had improved survival compared to patients with chronic liver disease. Among transplant recipients, those with acetaminophen and viral hepatitis-related ALF had better post-transplant survival than patients with ALF of other viral causes as well as autoimmune and drug-induced liver disease (Fig. 5).
There is relatively meager experience worldwide with living donation for ALF.37–39 Although well accepted for children where left lobes can be used, its use in adults, which requires right-lobe donation, remains controversial. The decision to embark on living donor liver transplantation for ALF needs to take into account the time involved evaluating the donor candidate versus the time spent waiting for a deceased donor organ. Hastening the donor evaluation raises many issues, including whether the necessary elements in donor candidate evaluation have been adequately completed, ethical considerations, and the eventual outcome for both recipient and donor.
Preliminary results on outcome of transplantation for ALF are available from the Adult-to-Adult Living Donor Liver Transplantation Cohort Study (A2ALL), a NIH-supported consortium of 9 liver transplant centers that collect data on both recipients and donors. From 1998 to 2006, 1,079 potential living donor recipients were enrolled in the cohort, among whom 11 had ALF. Eight received a living donor liver transplant; 2 received a deceased donor transplant, and 1 recovered spontaneously. The average length of stay after donor surgery was 5 days, and 4 donors suffered complications including bile leak, abscess, pleural effusion, bacterial infection, and neuropraxia. Eight of 10 patients who underwent transplantation and all donors survived. Thus, living donor liver transplantation is a viable but challenging approach to management of ALF in adults.
The decision to transplant is more difficult in children than in adults because they will face an entire lifetime needing immunosuppression. Likewise, children with ALF fare poorly compared to those with chronic end-stage liver disease,8,40 with an average 6-month survival from the time of listing of 60% versus 90% for children with chronic liver disease. Even after liver grafting, differences persist, with ALF carrying a 6-month posttransplant survival rate of 80%, versus 92% among children with chronic liver disease. Improving prognostic accuracy in children who present with ALF is an important goal.41,42
Acetaminophen poisoning currently represents the most common cause of ALF in North America and Europe.43 Patterns of presentation differ between countries, with more unintentional cases reported from the United States than Europe. Between 400 and 500 deaths occur annually in the United States due to acetaminophen-related liver failure.44
Acetaminophen is a dose-related toxin, with higher dosing generating the highly reactive intermediate, N-para-aminoquinonimine (NAPQI). When glutathione is depleted,45 NAPQI binds to cell proteins to yield cysteine adducts.46 A recently developed high-pressure liquid chromatography with electrochemical detection (HPLC-EC) assay has allowed the detection of adducts in tissue and serum samples.47 In patients with acetaminophen toxicity enrolled in the ALF Study Group registry, acetaminophen-cysteine adduct levels were present in virtually all patients and persisted for an average of 7 days, declining in parallel with aminotransferase levels.32 Importantly, significant levels were detected in 7 of 36 (19%) indeterminate cases of ALF,32 suggesting that these cases represented unrecognized instances of acetaminophen toxicity. Similar results were seen in children.11 Future studies are needed to define the absolute level of adducts associated with significant hepatic toxicity.
A suicidal overdose involves a single time-point ingestion of a large amount of acetaminophen with suicidal intent, while an unintentional case involves ingestion of lesser multiple doses over several days usually for pain control and without suicidal intent, hence the term “therapeutic misadventure”.48,49 Among 275 acetaminophen-related cases reviewed, intentional and unintentional cases were similar in number, whereas 8% could not be categorized.50
Acetaminophen liver injury is characterized by a “hyperacute” biochemical pattern, with markedly high levels of alanine aminotransferase (ALT; median 4,149 U/L and as high as 26,600 U/L) and only modest elevations in bilirubin (median 4.5 mg/dL). Women predominate (74%), with a median age of 36 years. Despite better survival (Table 1), acetaminophen-related cases make up the largest number of deaths. Comparing intentional and unintentional cases revealed few differences (Table 2). Total dosages taken and severity of illness were similar. The unintentional cases were more likely to consume narcotics and to have used multiple acetaminophen-containing preparations, possibly reflecting tolerance to acetaminophen in patients addicted to the narcotic-acetaminophen combinations.51 Alcohol use was more commonly associated with unintentional cases, but its role in worsening the liver injury is unclear.
How often ALF develops with therapeutic doses of acetaminophen is uncertain.50 However, patients reporting ingestion of lower doses were more often alcohol abusers. Patients with acute viral hepatitis who ingest acetaminophen appear more likely to develop ALF.52,53 Low levels of acetaminophen adducts have been identified in some cases of ALF due to viral hepatitis, but the effect on outcome is uncertain.54
A recent Cochrane Systematic Review concluded that data supporting most established interventions for acetaminophen overdose except for NAC are weak.55 Although efforts to limit gastrointestinal absorption or to dialyze off the parent compound may be used, detoxification of the reactive intermediate is the most effective and accepted therapy.57,58
NAC is a specific antidote that supplies glutathione, a sulfhydryl donor, to limit NAPQI formation yielding instead the readily excreted, nontoxic mercapturic acid. When given within 24 hours of overdose, NAC clearly decreases the risk of liver injury.59–61 Administration of NAC more than 24 hours after overdose is not supported by prospective studies.59 Because side effects are minimal, NAC is often given up to 72 hours after ingestion.
Plasma drug levels can be helpful in determining severity of injury if the time of ingestion is known.57 Intravenous NAC is used if encephalopathy or nausea are present.63 Allergic reactions (hives, bronchospasm), seen particularly with the intravenous form, are generally mild, responding to antihistamines and/or cessation of the infusion. NAC has been associated with exacerbation of asthma, and one case of fatal bronchospasm has been reported. Recommendations as to duration of NAC therapy vary. Plasma acetaminophen levels are not useful for this purpose.
Paracetamol, as acetaminophen is known in Europe, accounts for 200–500 deaths and 20–40 liver transplants annually in the United Kingdom alone.19, 65 Only 25% of cases are thought to be unintentional overdoses. Deaths from acetaminophen overdose have been associated with older age, male sex, and alcohol or drug use. A surge in the incidence of cases of ALF due to acetaminophen occurred in the mid-1990s with 37,000 hospitalizations and 562 deaths in England and Wales in 1997.66 Acetaminophen overdoses accounted for 37% of emergency liver transplants.
Legislation was adopted in the United Kingdom in 1998 aimed at restricting the quantities found in the home, limiting package size to 16 tablets for general sales outlets and 24 tablets for pharmacies. As a result, the number of hospital admissions related to acetaminophen overdoses, number of transplant listings, and deaths all decreased by varying amounts in the range of 27%-33% between 1995–1998 and 2001–2004. 67–70
Although autoimmune hepatitis (AIH) is a chronic disease, an acute presentation occurs in approximately 22% of patients,71 an even smaller number presenting with ALF. The diagnosis generally relies on clinical suspicion, the presence of serum autoantibodies, liver histology (if available), and response to corticosteroid therapy. Of 1,147 patients entered into the ALF Study Group registry between 1998–2007, 52 (~5%) were attributed to AIH.4 Histologic features of AIH-related ALF may differ from those features typically associated with a more chronic presentation.72
Randomized controlled trials have shown that corticosteroids provide no benefit overall in ALF.73,74 However, cases of ALF due to AIH might be an exception. In the ALF Study Group data, patients treated with corticosteroids had a higher rate of spontaneous recovery than patients with AIH who were untreated. Better criteria for diagnosis and prospective studies are needed before any conclusions about the role of immunosuppression in ALF associated with AIH can be made.
Idiosyncratic drug-induced liver disease accounted for 11% of cases in the ALF Study Group registry.75,76 The establishment of causality in these cases is challenging,77,78 Among 119 drug-induced cases from the ALF Study Group, women predominated (67%) with a median age of 43 years, peak ALT of 571 U/L, and peak bilirubin of 21.6 mg/dL. Implicated medications included antituberculous drugs in 19 cases, sulfonamides in 10 cases, phenytoin in 7 cases, disulfiram in 4 cases, troglitazone in 4 cases, propylthiouracil in 4 cases, bromfenac in 4 cases, and herbals and dietary supplements in 11 cases. Outcome of drug-induced ALF was poor (Table 1); of those patients who underwent transplantation, however, 96% survived in the short-term, giving an overall survival of 64%.
Although acute hepatitis A and B have declined over the last decade, they accounted for 7.3% and 3% of US ALF cases, respectively.79 Hepatitis C and E accounted for only 1 case each and Epstein-Barr virus only 4 cases of more than 1000 cases.80,81 Other previously unidentified viruses have been suspected of causing indeterminate ALF.7,82 However, searches for viral agents using sera from indeterminate patients found no cases of parvovirus B19, hepatitis E or SEN virus.83,84 Other rare causes of ALF include Dengue virus infection85 and herpes zoster and simplex. Finally, reactivation of hepatitis B or infection with hepatitis B virus (HBV) harboring precore and core-promoter gene mutations or with genotype Bj or D appear to be important causes of ALF in some areas of the world or in special situations.86,87
Patients with hepatitis A virus (HAV) ALF have a better spontaneous (58% versus 24%) and overall survival rate (87% versus 61%) than those with ALF due to HBV. Although it has been reported that patients with chronic hepatitis C who are superinfected with HAV have a high risk for ALF,88 no such cases were seen in the ALF Study Group database nor by others.89,90 Superinfection with hepatitis D in carriers of HBV is another risk factor for ALF but was uncommon in the ALF Study Group patients.91
Antiviral treatment may be considered for HBV ALF based on a retrospective German study.27 Other results are conflicting. Herpes simplex virus–related ALF is difficult to diagnose and has a poor prognosis unless treatment is initiated early.92–94 Although declining in incidence, viral infections remain important causes of ALF.
Since the 1950s, a variety of therapies have been evaluated. Typically, small controlled trials failed to show evidence of benefit with most of these approaches, including corticosteroids, heparin, insulin and glucagon, and prostaglandin E1.22,73,95–98 N-acetylcysteine, the standard antidote for acetaminophen overdose, has recently been suggested to be beneficial in other forms of ALF,18 possibly through its effects on hemodynamics in multiorgan failure and its renal protective actions.99–101 A prospective, randomized, double-blind, placebo-controlled trial of NAC for cases of ALF not due to acetaminophen has been completed with 173 patients enrolled at 25 study sites, who received placebo or intravenous NAC (Acetadote) over 3 days. Trial results will be available in Spring 2008.
Animal and human studies have shown that reduction of body temperature between 32°C and 35°C results in a consistent lowering of intracranial pressure and thereby decreased likelihood of cerebral edema.102,103 The possible mechanisms responsible for this include normalization of cerebral blood flow, reduction in delivery of ammonia to the brain, amelioration of anaerobic glycolysis and oxidative stress in astrocytes, and decrease in brain extracellular glutamate.104,105
The role of hypothermia in clinical practice remains unclear. Hypothermia has been recommended for use in patients after cardiac arrest or after traumatic brain injury based on controlled trials but has been applied with varying frequency.106,107 Technical aspects are not complex: the patient with ALF shows remarkable vasodilatation so that an external cooling blanket is usually adequate. Intracaval cooling balloons usually are needed in settings after trauma or cardiac arrest where vasoconstriction predominates.
Complications of hypothermia remain a concern; systemic vascular resistance increases and cardiac output decreases. Patients shiver, requiring paralytics and intubation, thus limiting its use in subjects with early coma grades. Pancreatitis, hyperglycemia, increased risk of infection, and lowering of platelet counts have been reported.105,107 Although hypothermia might theoretically limit hepatic regeneration, a mouse model of hypothermia showed no decrease in this parameter.108 It remains unclear whether therapeutic or prophylactic cooling can improve outcomes and/or protect brain function in these critically ill patients.
Over the years, means of exchanging elements in plasma with normal components have been streamlined so that current systems are fully automated and can rapidly exchange 10 L of plasma, consistent with full replacement of the extracellular fluid compartment. The benefits of such a system include removal of toxins, including protein-bound substances, and repletion of factors synthesized by the liver. In small, somewhat heterogeneous studies, plasmapheresis decreased plasma ammonia levels and improved hepatic encephalopathy, generally within hours.109,110 Cerebral blood flow increased as well as cerebral perfusion pressure.111 Mean arterial pressure was improved, cardiac index decreased, and systemic vascular resistance increased during plasmapheresis.112 The effect of plasmapheresis on survival from ALF has been more difficult to determine.109,113 However, an interim analysis of the large, controlled study currently underway in Europe has shown promise of benefit in the plasmapheresis group.
Hepatic assist devices generally fall into two categories: (1) non– cell-based detoxification systems and (2) cell-based systems to provide metabolic support as well as detoxification (bioartificial liver).114 Non– cell-based systems include open-loop and closed-loop approaches. Open-loop approaches include the single-pass albumin dialysis system and plasma exchange.115 Closed-loop systems include the Prometheus albumin dialysis system116 and the molecular absorbents recirculation system (MARS).117 Both approaches have limitations including thermodynamic limitation on removal rate,118 nonselective losses with open-loop systems, and equilibrium limitations on total removal and effective perfusion time with closed-loop systems.119
Five cell-based systems have been assessed clinically, including the ELAD system (Vital Therapies),120 HepatAssist (Arbios, formerly Circe),121 MELS (Charite Medical Center),122 BLSS (Excorp Medical),123 and AMC BAL (Amsterdam Medical Center).124 The systems differ in cell source and mass, use of plasma or whole blood, the rate of perfusion, and duration of treatment (continuous versus intermittent). Cell masses vary from 100–500 g/cartridge and flow rates from 20–200 mL/ minute. Advantages and disadvantages exist for each cell-based therapy. All appear to be safe and have at least some biologic effect, but none has received FDA approval for use in the United States. A Cochrane review of artificial and bioartificial support systems concluded that results have yet to demonstrate evidence that these systems affect survival in ALF.125
An important issue in design of bioartificial liver systems is whether plasma or whole blood is used. Because of the separation process, the effective treatment rate for plasma perfusion is one-third to one-half that for whole blood. A further issue in device design is the interaction between perfusion rate, cell mass, and bioreactor efficacy. No current bioartificial liver device even approaches the perfusion rate or extraction efficiency of the native liver. Improvement in function of bioartificial livers should aim to increase efficiency of the hepatocytes within the system rather than adding more (relatively inefficient) hepatocytes. All three non–cell-based albumin-recirculating systems have proven to be safe and to efficiently remove suspected toxins. None have yet to be shown to improve survival in ALF.
Bioartificial liver devices are meant to allow time for spontaneous recovery or for the retrieval of a liver graft (bridge to transplantation) and to manage the systemic manifestations of ALF. Surrogate, short-term endpoints include decreases in serum ammonia, bilirubin, and amino acid abnormalities as well as improvements in hepatic encephalopathy, cerebral blood flow, and renal and pulmonary function. In judging the efficacy of bioartifi-cial liver support, however, the critical endpoint is increased survival with or without transplantation.
The general principles of bioartificial liver therapy are shown in Fig. 6. With a closed loop design, the patient’s circulation is separated from isolated liver cells by a semi-permeable membrane, protecting the cells from damage by recipient immune cells and eliminating the need for immunosuppression. The membrane also protects the patient from exposure to the allogeneic or xenogeneic cells used in the device. Besides oxygen and nutrients, albumin-bound nonpolar waste molecules are also permeable to the membrane, allowing conjugation of bilirubin and other solutes and further detoxification by the liver cells. The membrane is also permeable to proteins synthesized by hepatocytes in the device.
Design and results of more than 30 bioartificial liver systems for treatment of ALF have been reported since the first publication in 1987.126 The ELAD uses hepatocytes from a hepatoblastoma cell line (C3A) attached to hollow fibers within the bioartificial liver and has been extensively studied.127 The HepatAssist 2000 system was studied in a phase 1 clinical trial in 1997128 followed by a large randomized trial of 170 patients enrolled at 20 sites.121 Though survival was improved with HepatAssist therapy, the differences did not achieve statistical significance.
The next generation of bioartificial liver systems employ a “hybrid” or multimodal approach, combining albumin dialysis, charcoal therapy, adsorbent therapy, and dialysis with hepatocyte cell-based therapy (Fig. 7). The extracorporeal circuit consists of the patient’s whole blood whereas the perfusate circuit is an albumin-based solution. In one system, porcine hepatocyte spheroids are used along with albumin dialysis and absorbent therapy.129 In preliminary studies using 200 g of hepatocytes, the porcine hepatocyte spheroid system provided cerebral protection not evident in the control group.
Bioreactors may be used for human liver progenitor cell expansion. Simple hepatocyte suspension cultures are limited by the anchorage-dependence of hepatocytes for long-term function and survival. More complex systems rely on anchorage-dependence, but require a support biomatrix to prevent apoptosis-mediated cell death (anoikis).130 Hepatocytes may grow in 2-dimensional (2-D) “sandwiches” or as aggregates where they provide their own extracellular matrix and supporting cellular structures.131–133 The macroenvironment offered by the culture system appears to be critical. A variety of biomatrix signals have been evaluated including fibronectin, vitronectin, laminin, collagen, fibrinogen, and proteoglycans. Physical signals that influence cell differentiation include temperature, pH, chemical gradients, pressure forces, and media perfusion.134–135
Currently, a 4-compartment hepatocyte bioreactor demonstrating spontaneous aggregation of hepatocytes and nonparenchymal cells into 3-dimensional sinusoids is used.136–140 The ultimate goal is to construct liver tissue in vitro for implantation.
After success in animal models, human hepatocyte transplantation for ALF was attempted in the late 1990s.141 Studies used cryopreserved human hepatocytes infused into a small number of children with limited success.142–143 These studies provided a proof-of-principle for cell transplantation therapy, but further progress in technical issues is needed before prospective trials can be initiated. Important issues include the optimal dose of hepatocytes, preferred site of transplantation (liver, spleen, elsewhere), better cryopreservation techniques, use of apoptosis inhibitors,144 optimal dose and duration of immunosuppression, improved means of monitoring presence and function of transplanted cells, and development of alternative sources for hepatocytes for transplantation such as from stem cells of hepatic, bone marrow, or fetal origin.145
Stem cells are defined by their capacity for self-renewal and differentiation into multiple cell types.146–147 Human embryonic stem cells have great potential and a remarkable ability to expand as well as a unique “cryopreservability”. Disadvantages are the limited number of approved human cell lines and the incomplete understanding of how to control the sequence of activation of genes that leads from undifferentiated stem cells to mature, functional hepatocytes. Undifferentiated embryonic cells themselves cannot be used, because they form tumors if transplanted into any site other than in utero.
To date, maturation of fully functional mature hepatocytes from embryonic stem cells has yet to be achieved efficiently using human embryonic stem cell lines. 146–150 Another source of hepatocytes, however, are pluripotent cells that have the capacity to proliferate and differentiate into hepatocytes and cholangiocytes.151–155 Methods for the identification, isolation, expansion, and cryopreservation of hepatic stem cells have been developed.156–158 Hepatic stem cells are found in the ductal plates of fetal and neonatal tissues159 as well as in the Canals of Hering in pediatric and adult liver.160–162 Growth of hepatic stem cells as hepatocytes into cords can be promoted by use of defined matrices and growth factors, making them a potential source of mature liver cells for cell-culture studies, cell transplantation, and bioartificial liver assist devices.
Most in vitro studies of liver function and hepatocyte pathways have used cells cultured in single (2-D) cell layers that do not recapitulate the unique structure of liver with its apical and basal lateral membranes, separate directions of plasma and bile flow, and complex interactions with nonparenchymal cells. Using perfusion microreactors that are seeded with primary rat hepatocytes,163 these cells can be maintained demonstrating greater functional activity than cells in conventional 2-D culture.164 When liver-derived endothelial cells are added, they proliferate and form microvessel-like networks within 3-D reactor cultures, maintaining sinusoidal endothelial cell-specific markers and typical fenestrations.165 Morphogenesis and proliferation of endothelial and stellate cells appear to be governed by perfusion rates.
For any complex liver culture system, a biodegradable scaffold is needed which provides microvascular and macrovascular networks and supports the function of cells.166 A covalently cross-linked synthetic extracellular matrix material, made of a complex of proteins, was developed for 3-D culture of cells in vitro and in vivo.162,167,168
Rat hepatocytes can be cultured in 3-D hydrogels for more than 3 weeks without loss of drug metabolizing activity, providing a tool for the in vitro evaluation of medications for their effects on drug metabolism and potential for hepatic injury.166,168–170
Human hepatocytes can also be cultured in hydrogel systems, which have been shown to be capable of rescuing nude mice after 90% liver resection by using the synthetic extracellular matrix material to implant either mature human hepatocytes or a mixture of human liver progenitor cells on the residual liver tissue.
Classical approaches to tissue engineering have used cells seeded onto resorbable matrices and have had modest success in replicating simple, relatively avascular structures, such as cartilage or bone, but have been unable to create more complex parenchymal organs, such as liver. Because of inadequate vascularization, the largest composite graft that will survive must be less than 1.5 cm3. A potential solution to this shortcoming is the use of a preexisting, microcirculatory bed as a scaffold to engineer a vascularized neo-organ.171–173 In this approach, entire vascular beds (such as omentum) are removed, cultivated in a bioreactor system where they are seeded with progenitor cells, and then reimplanted using microsurgical techniques.174 Recently, vascularized beds have been successfully maintained ex vivo for several days while they were seeded with multipotent adult progenitor cells, such as liver-derived oval cells.175 Ultimately, vascular beds may provide a means of organ engineering that might be applied to hepatic support in humans.
There have been many advances in understanding and management of ALF, but the syndrome remains challenging and a still-too-common cause of mortality. Important areas for research include:
We wish to acknowledge the sponsorship of the Acute Liver Failure Workshop by the National Institute of Diabetes and Digestive and Kidney Diseases with the support of the National Institute of Biomedical Imaging and Bioengineering and the Office of Rare Diseases of the National Institutes of Health. We also thank Drs. Fei Wang, Patricia Robuck, and Linda Griffith for the efforts in organizing meeting. The speakers at the workshop included William M. Lee, Robert H Squires, Jr, Robert J. Fontana, Timothy Davern, II, Lorenzo Rossaro, Robert Merion, James Trotter, John Bucuvalas, Michael Narkewicz, Laura James, Julie Polson, Anne Larson, William Bernal, R. Todd Stravitz, Adrian Reuben, Anna S.F. Lok, Andres T. Blei, Lars E. Schmidt, John F. Patzer, II, Tarek Hassanein, Scott Nybert, Jorg C. Gerlach, Humberto Soriano, Lola Reid, Linda G. Griffith, Glenn D. Prestwich, and Geoffrey Gurtner.
The Acute Liver Failure Study Group and the Pediatric Acute Liver Failure Study Group are supported by NIH grants U01-DK58639 and U01-DK-072146.
Potential conflict of interest: Nothing to report.