Detoxification as therapeutic goal in liver failure
Cerebral oedema and hepatic encephalopathy (HE) are the most deleterious complications of acute liver failure (ALF), a life-threatening multisystem disease following severe hepatic injury with jaundice, coagulopathy and HE (1). Until recently, many researchers were convinced that in ALF, a lack of protective factors synthesised by a healthy liver caused or significantly contributed to HE. This stimulated the development of extracorporeal systems using a perfusion of fresh-cut liver slices, isolated hepatocytes or hepatoblastoma cells. Unfortunately, the strategy failed clinically.
Already the earliest examples of artificial liver support have reported some improvement of neurological function in patients treated with extracorporeal detoxification (2). Consequently, the treatment of ALF has increasingly adopted strategies of extracorporeal detoxification, including high-volume plasmapheresis, high-flux haemodialysis, albumin dialysis, haemofiltration and single-pass albumin filtration (3). The close temporal link between elevated plasma ammonia and the subsequent evolution of cerebral oedema suggests an endogenous intoxication with ammonia to be the single most important cause of HE and cerebral oedema in ALF. Most recently, a large multicentre trial has demonstrated significant improvement of transplant-free survival when patients with (hyper)acute liver failure were treated with three sessions of high-volume (10 l) plasma exchange (Larsen and colleagues, AASLD 2010 Annual Meeting, #114).
Acute liver failure: detoxification as a bridge to regeneration
Patients with hyperacute liver failure (typically younger patients with paracetamol intoxication or viral hepatitis) are at a high risk of cerebral herniation in spite of their considerable potential for liver regeneration. Metabolic brain dysfunction with astrocyte swelling, cerebral vasodilation and intracranial hypertension is an important cause of death in these patients. There is now increasing evidence to favour a ‘toxin hypothesis’ of HE, with blood ammonia, endotoxin and intracellular glutamine as the most deleterious toxins. Arterial ammonia levels correlate with both intracerebral glutamine concentrations and intracranial pressure recordings (4). Despite massive coagulopathy, patients with ALF usually tolerate extracorporeal treatment without developing significant complications and are likely to benefit from extracorporeal liver support as a ‘bridge to regeneration’. Ideally, emergent liver transplantation, which is currently decided using the King's College criteria, could be avoided in a proportion of patients.
Acute-on-chronic liver failure: biocompatibility as a critical issue
The term ‘acute-on-chronic liver failure’ (AOCLF), has been used to describe episodes of acute decompensation in patients with pre-existing chronic liver disease. Decompensation is usually triggered by bacterial infection, variceal bleeding, overdose of diuretics or other causes. Short-term mortality is excessive, and emergent transplantation is impossible. The goal of extracorporeal support in AOCLF is to reverse encephalopathy, ameliorate multi-organ dysfunction and provide temporary renal support in patients not responding to conventional treatment. Unfortunately, patients with AOCLF usually have some degree of portal hypertension, hyperdynamic circulation and activation of coagulation, all increasing the risk of treatment-related complications. The results of current trials using state-of-the art liver support technologies will help to clarify the future role of extracorporeal detoxification in AOCLF.
The ‘toxin hypothesis’ of cerebral hyperaemia and brain oedema
For more than a century, ammonia, bypassing the failing liver, has been considered the major neurotoxin in HE. Only recently has it been demonstrated that high arterial ammonia levels preceded cerebral herniation and death in ALF (5). Ammonia levels remained high in patients developing intracranial hypertension and declined in those who did not, with young age, vasopressor therapy and renal replacement therapy identified as additional independent risk factors (6). Importantly, the relation between plasma ammonia and severity of HE is modified by the astrocytic handling of glutamine. The typical delay of 1–2 days between the emergence of massive hyperammonaemia and cerebral herniation represents a window of opportunity for performing extracorporeal detoxification before the advent of severe HE. This strategy, however, still has to be confirmed by clinical data.
In ALF, studies of the cerebral circulation consistently show impaired cerebral autoregulation and cerebral hyperperfusion in advanced stages of HE. These changes can be reproduced in the absence of hepatic necrosis by ammonia infusion in portocaval-shunted animals and appear to be linked to ammonia/glutamine toxicity. Glutamine-induced astrocytic water accumulation acts as an integrative trigger for the development of intracranial hypertension, which is further aggravated by inflammatory changes. Moderate hypothermia decreases cerebral ammonia uptake and glutamine synthesis, which could explain some of the effects of liver support systems. Interestingly, a recent study found no clinical benefit of hypothermia in patients with refractory intracranial hypertension (Larsen FS, EASL 2011).
Methods for detoxification in artificial and bioartificial liver support systems
Artificial liver support systems
The success of haemodialysis in the treatment of renal failure introduced by the Dutch physiologist Kolff in 1943 has been a constant stimulus for the development of extracorporeal liver support systems. Small-molecular toxins including ammonia and its equivalent glutamine are effectively removed by standard haemodialysis, and up to 63% of patients with ALF treated by high-permeability haemodialysis improved neurologically (7). In the late 1970s, however, liver transplantation was still experimental, and cuprophane haemodialysis (showing very limited biocompatibility) failed to improve survival in ALF. Treatment goals subsequently shifted to the removal of large-molecular and protein-bound toxins using charcoal or resin haemoperfusion and sorbent dialysis. None of them improved the outcome, questioning the entire concept of large-molecular toxins in liver failure. Plasma exchange, widely performed in the 1980s, has recently re-emerged as a possible treatment for ALF and severe HE. As patient plasma is exchanged by donor plasma in this procedure, treatment could be regarded as a bridge between artificial and bioartificial therapy in patients with ALF.
The latest developments in the field of artificial liver support are the Molecular Adsorbent Recirculating System (MARS®), the Prometheus fractionated plasma absorption and filtration system, and single-pass albumin dialysis, where albumin is added to a haemofiltration solution. These systems are all capable of removing small molecular toxins, while they also aim to absorb bilirubin and hypothetical large-molecular and albumin-bound toxins. The MARS® system has been used extensively since 1993, reducing bilirubin concentrations and stabilising blood pressure. A randomized study investigating MARS versus conventional therapy in hepatic encephalopathy has demonstrated a positive effect on the severity and duration of HE (10).
The Prometheus® system has been introduced more recently. Effects on survival are unclear, as results of randomised studies in patients with decompensated liver disease are still awaiting publication. A multicentre study in ALF involving 88 patients revealed a possible benefit of MARS® treatment in subgroups receiving more than 3 treatments (Saliba F, oral presentation, AASLD 2008 and personal communication). Smaller studies have also shown haemodynamic stabilisation in patients treated with MARS®, possibly related to the use of albumin.
Biological and bioartifical liver support systems
‘Bioartifical’ liver support systems were originally designed to combine artificial liver support with perfusion of extracorporeal hepatocytes, according to the now obsolete ‘critical mass hypothesis’. With designs using plasma separators and oxygenators to sustain the viability of isolated hepatocytes, bioartificial liver support came at the cost of increased complexity, excessive cost, limited biocompatibility and low mass transfer rates. Moreover, bioartificial systems usually contained far less than the minimum residual amount of cells required for survival after liver resection (∼400 g). The secretion of heterologous proteins by porcine hepatocytes could trigger allergic symptoms. Bioartifical systems were mostly studied by small uncontrolled trials with the goal of bridging to transplantation, and no long-term outcomes were reported. The only prospective international multicentre trial investigating a system containing porcine hepatocytes and charcoal adsorbers had no effect on survival (8).
Peritoneal dialysis – a biocompatible alternative for patients with portal hypertension?
Peritoneal dialysis (PD), which can be easily performed using an implanted catheter inserted over a guidewire, should be considered an alternative to blood detoxification in patients with ascites, where the poor biocompatibility of conventional extracorporeal systems aggravates coagulopathy, thrombocytopenia and hypotension. PD has several advantages in chronic liver disease with portal hypertension, as it does not increase bleeding risk, allows drainage of ascites on a regular basis and facilitates the direct antibiotic treatment of spontaneous bacterial peritonitis. Glucose-based PD solutions can additionally improve the energy balance. The experience reported so far showed improved biocompatibility and prolonged survival with PD.
Continuous or intermittent detoxification?
In contrast to haemofiltration and haemodialysis, most liver support systems use adsorbents that saturate over time, requiring intermittent treatment. No prospective study so far has compared continuous with intermittent detoxification. Patients with grade III/IV encephalopathy and cerebral oedema may benefit from continuous treatment, as intermittent haemodialysis, causing variations in plasma osmolality, can increase intracranial pressure and aggravate cerebral oedema. In selected cases, rapid correction of hyperammonaemia and complete reversal of massive cerebral oedema by continuous detoxification in patients fulfilling transplantation criteria have been demonstrated (Fig. 1).
Pathophysiological concepts relevant for extracorporeal detoxification in liver failure
The current concepts of HE and cerebral oedema have challenged many assumptions on which previous blood detoxification systems were based. Rather than protein-bound and lipophilic large-molecular toxins, an endogenous intoxication with ammonia, a small, water-soluble molecule, explains many dramatic symptoms of acute and chronic HE. This knowledge has to be integrated into the clinical management of ALF, which inevitably affects extracorporeal treatment. Ammonia can be effectively removed by haemodialysis, at a rate dependent on ammonia concentration, dialyser surface, blood flow and dialysate flow. Rather than starting treatment upon emergence of severe HE (a late and potentially disastrous complication), massive elevation of ammonia, glutamine and uraemic toxins refractory to conventional treatment should be the trigger for initiating extracorporeal treatment. Patients with cerebral oedema should be preferably treated by continuous modalities to prevent deleterious shifts in osmolality. Considering the important role of inflammatory mediators in the pathogenesis of ALF and HE, biocompatibility of extracorporeal systems is a critical factor. Endotoxin has been shown to induce various complications in patients with AOCLD, but no study has investigated whether liver support affects endotoxin levels.
Will liver support systems improve survival?
Extracorporeal systems aiming to improve encephalopathy and sustain liver recovery have been investigated for more than 50 years. So far, they were unable to significantly reduce mortality. The long-standing uncertainty regarding the pathophysiology of HE has been the most important limitation in devising effective extracorporeal detoxification systems. It is conceivable that the poor biocompatibility of historical liver support systems (which caused coagulopathy, thrombocytopenia and aggravated hypotension) has offset their potential benefits. Fortunately, the past decade has witnessed rapid advances in deciphering the pathophysiology of HE and, simultaneously, biocompatibility of extracorporeal detoxification has been improved. Thus, effective anticoagulation, use of biocompatible membranes and reduction of extracorporeal blood volumes will help to improve biocompatibility of clinical liver support and eventually reduce the need for transplantation in an age of increasing donor organ shortage.