Currently, the standard medical treatment (SMT) for pediatric liver failure incorporates optimal supportive care with hemodynamic and respiratory support, infection control, and avoidance of gastrointestinal bleeding, aiming to bridge patients to recovery or to liver transplantation (LT). LT is the only definitive treatment for liver failure. Since the 1980s, emerging literature, within the adult setting, has studied the efficacy of extracorporeal liver support system (ELSS) in addition to SMT in liver failure as a means to improve biochemical and clinical parameters while a patient is being bridged to LT or recovery. ELSS can be broadly divided into (1) artificial liver support (ALS) and (2) bioartificial liver support (BALS). There are limited reports of ALS and no reports of BALS in pediatric literature. Biological support, such as hepatocyte transplantation, is a modality being studied as a bridge to LT and recovery in pediatric acute liver failure, although outside the remit of this chapter. There is much debate regarding the value of ELSS in liver failure, especially with improved median times for super-urgent LT and improved graft and survival outcomes post-LT. The routine use of ELSS has currently not been advocated in any pediatric guideline.
Indications for an Extracorporeal Liver Support System
Literature on ELSS is mainly focused on adult liver failure, so caution needs to be taken when extrapolating evidence to pediatric liver failure because key differences exist. The definition of pediatric acute liver failure (PALF), unlike in adults, does not necessarily require the presence of hepatic encephalopathy (HE); the cause is often age-dependent, with metabolic causes more common in neonates and infants, and infectious, drug, or autoimmune-induced pathology in older children. This is in contrast to adult ALF, where paracetamol (PCM) toxicity is the most common cause in the developed world. Despite these differences, both pediatric and adult ALF have a high mortality rate, reaching 70% in children without transplantation, which necessitates super-urgent LT in 10% to 15% and 10% of pediatric and adult cases, respectively. Spontaneous recovery has been reported in one-third of pediatric and 40% of adult ALF cases.
ELSS has been most studied in acute-on-chronic liver failure (AoCLF), which is a well-described clinical entity in adults. It is defined as an acute deterioration of preexisting chronic liver disease, usually related to a precipitating event and associated with an increased 3-month mortality resulting from multisystem organ failure. It is associated with a mortality rate between 50% and 90%, and commonly follows an acute trigger to underlying alcoholic liver disease. However, in pediatrics, data on AoCLF are scarce, and no specific definition has been established. Documented experience for pediatric AoCLF is mainly from non-transplantation centers in India, where hepatotropic viruses and drugs are the most common triggers. Only a minority of patients are diagnosed with chronic liver disease before acute presentation—namely, autoimmune hepatitis and Wilson disease (WD). Mortality is lower in pediatric AoCLF compared with adults, albeit from limited data.
The ideal aim of an effective ELSS would be to bridge patients with ALF or AoCLF to recovery or LT, reducing wait-list mortality and improving clinical outcomes postrecovery or LT (e.g., neurological). However, short median waiting times for super-urgent LT in PALF and the inability to predict spontaneous recovery in PALF accurately lead to challenges when assessing the effectiveness of ELSS.
Artificial Liver Support
In liver failure, hemodialysis (HD) is highly effective for the removal of small and medium-sized water-soluble, unbound molecules (e.g., ammonia, urea) from the circulation across a semipermeable membrane into hypotonic dialysate fluid along their concentration gradients. However, larger albumin-bound molecules (e.g., bilirubin, bile acids), which accumulate and contribute to the progression of ALF and AoCLF, are not removed via HD. Subsequently, ALS should incorporate devices that can facilitate the removal of albumin-bound molecules to reduce liver injury and support bridging to recovery or LT.
Types of Artificial Liver Support
Albumin dialysis devices constitute the main form of ALS. There are various adaptations of these devices, based on filtration and adsorption. The two commercially available devices ( Fig. 31.1 A, B ), which have been evaluated in the clinical setting, are MARS (Molecular Adsorbents Recirculating System; Baxter, Sweden) and Prometheus (FPSA-fractionated plasma separation and adsorption system; Fresenius Medical Care, Germany).
MARS consists of a primary circuit, whereby the patient’s blood passes through a specialized hemofilter with a size selection threshold (< 60 kDa) and a secondary circuit containing albumin-enriched dialysate, which passes the filter in a counterdirectional flow. Hemofiltration is required simultaneously to control the blood and dialysate circuits. This system enables albumin-bound toxins, but not albumin itself, to dissociate, passing through the membrane to attach to albumin in the secondary circuit. The secondary circuit then undergoes dialysis to remove the water-soluble toxins, and the albumin dialysate is regenerated by removing toxins via two adsorbers (activated charcoal and an anion resin), ready for use again. MARS has two filters available based on weight—above 25 kg, the adult filter (2.1 m; fill volume, 152 mL), and below 25 kg, the MARS minifilter (0.6 m; fill volume, 57 mL) suitable for pediatric use.
Prometheus consists of a primary circuit, whereby the patient’s blood passes through a large-pore (250–300 kDa), albumin-permeable membrane, enabling the patient’s albumin fraction to pass into a secondary circuit where albumin-bound toxins are removed via two adsorber columns. The cleansed dialysate can then reenter the primary circuit. Hence, unlike MARS, albumin is detoxified in vitro by temporarily separating albumin from blood. To date, Prometheus has been designed for the adult population only.
Single-pass albumin dialysis (SPAD) therapy, based on the concept of MARS, uses a standard hemofiltration filter to dialyze the patient’s blood against an albumin solution, removing small, protein-bound molecules. The albumin dialysate is discarded once used. This select plasma exchange therapy device is based on the Prometheus device, but instead the plasma is discarded and replaced with human albumin solution (HAS).
Another approach to albumin cleansing is simply to remove and replace it. Plasma exchange or plasmapheresis (see Fig. 3.1 C) therapy involves separating plasma from cells in an extracorporeal device, which uses a filter or centrifuge. The plasma is then replaced with HAS, fresh-frozen plasma, or a combination of the two. It effectively removes albumin and albumin-bound toxins, cytokines, and antibodies and improves coagulopathy. There are variations of this procedure, including low- or high-volume plasma exchange (> 10 L/day) and plasma filtration, where filtered plasma is not discarded but is detoxified using adsorption column(s) and returned to the patient.
Efficacy of Albumin Dialysis on Laboratory and Clinical Outcomes
Until recently, the effects of albumin dialysis were limited by small uncontrolled studies. However, data from three prospective, multicenter, randomized controlled trials (RCTs), evaluating the clinical efficacy of albumin dialysis in both adult ALF and AoCLF, are now available. The RELIEF trial and FULMAR trial compared MARS with SMT in AoCLF and ALF patients, respectively. The HELIOS trial compared Prometheus with SMT in AoCLF.
Small uncontrolled and controlled studies applying MARS or Prometheus in AoCLF and ALF have consistently reported short-term improvement in bilirubin, which has been corroborated by larger multicenter RCTs. However, the sustainability of bilirubin clearance has not been demonstrated. Small studies comparing MARS and Prometheus against each other suggest a superiority of Prometheus in bilirubin clearance; however, further larger studies would be needed to verify this data. Renal function and hyperammonia will often improve after albumin dialysis therapy; however, HD will play a significant role in their elimination, which is not accounted for in the control patients in the RCTs, hence increasing the treatment effect in the therapy arm.
A meta-analysis has demonstrated a positive effect of MARS on HE, such that the FDA (US Food and Drug Administration) has approved its use in this clinical scenario. Prometheus therapy in AoCLF has demonstrated an improvement in HE to a lesser extent. However, Hassanein et al., in their large prospective RCT ( n = 70), found a nonsignificant improvement in HE in patients with AoCLF and HE grade III or IV (West Haven criteria) following 5 days of MARS therapy, compared with SMT (MARS, 62% vs. SMT, 40%; P = . 076). This was corroborated by results from the RELIEF trial. Fewer studies exist on neurological outcomes after albumin dialysis in ALF. Small noncontrolled trials using MARS therapy in ALF have shown inconsistent results for HE improvement, and the FULMAR trial showed no significant clinical improvement in patients with mild HE. Similarly, for Prometheus, few case reports or series have revealed HE improvement in ALF. An improvement in HE is not always associated with a concomitant reduction in ammonia, suggesting the role of other neurotoxins, or a different mechanism whereby albumin dialysis affects the complex pathophysiological process of HE. The difficulty of HE grading in ventilated patients makes this a challenging outcome measure for albumin dialysis.
AoCLF includes a systemic hyperdynamic circulation-reduced systemic vascular resistance index (SVRI), reduced mean arterial pressure (MAP), increased cardiac index (CI), increased heart rate, and portal hypertension. Several small studies in AoCLF have demonstrated an increase in MAP after MARS therapy. Laleman et al. found a significant improvement in CI, SVRI, and MAP 1 hour after MARS therapy in their small RCT ( n = 12; MARS, 6, vs. SMT, 6) directed at alcoholic liver disease (ALD). In ALF, two small noncontrolled studies (hepatitis B, n = 10; PCM, n = 10) applying MARS did show an improvement in MAP. The clearance of endogenous vasoactive substances, including nitric oxide, is postulated as a potential mechanism, but has not been clearly demonstrated. In addition, albumin dialysis provides a cooling effect, which could benefit hemodynamics. No positive effect on hemodynamics has been observed with Prometheus. Larger studies accounting for confounding factors (e.g., HD, temperature) are needed to understand the effect of albumin dialysis on hemodynamics better.
Efficacy of Albumin Dialysis on Laboratory Parameters and Survival Outcomes
The RELIEF trial 13 ( n = 189) compared the survival benefit at days 28 and 90 between MARS and SMT ( n 1 = 95) and SMT alone ( n 2 = 94) in a heterogenous group of AoCLF patients, dominated by ALD, across 19 European centers. Despite short-lived improvement in biochemistry and neurology, no survival benefit was demonstrated on day 28 (60.7% MARS vs. 58.9% SMT; P = . 79) or day 90 (46.1% MARS vs. 42.2 % SMT; P = .71), even on subgroup analyses (model for end-stage liver disease [MELD] > 20, hepatorenal syndrome at admission, severe HE, serum bilirubin [SBR] > 20 mg/dL). Three patients in each group underwent LT; however, no other information on listed patients or post-transplantation outcomes was available.
The FULMAR trial randomized a heterogenous group of ALF patients (~ 50% PCM toxicity) eligible for LT to receive SMT ( n 1 = 49) or MARS + SMT ( n 2 = 53). No significant difference in overall survival at 6 months (75.5% SMT vs. 84.9% MARS; P = . 28) or 1 year (75.5 SMT vs. 83% MARS; P = . 35) or transplant-free survival at 6 months (27% SMT vs. 19% MARS; P = . 35) was seen. One-year post-LT the survival was also similar between both groups, but no other post-transplantation outcomes (e.g., neurology) were formally assessed. One major limitation was that super-urgent LT occurred in 64.7%, with median time to transplantation being 16.2 hours (range, 11.4–28.2 hours), questioning the effect of albumin dialysis in such a short time period.
The HELIOS trial compared Prometheus + SMT and SMT in a heterogenous group of AoCLF patients (consisting of > 50% ALD) in their multicenter RCT and only revealed survival benefit in a subgroup of patients with MELD > 30. There are no large studies on the efficacy of Prometheus in ALF.
Safety of Albumin Dialysis
Thrombocytopenia is the main complication reported after albumin dialysis therapy and has been demonstrated in both multicenter RCTs for MARS in ALF and AoCLF. However, the number and type of adverse effects, including bleeding and infection, were similar between dialysis and non-dialysis groups in all three multicenter RCTs. Catheter-related complications have been reported, which have not led to significant problems. However, the large studies have excluded unstable patients, such as those with severe coagulopathy and thrombocytopenia, in whom albumin dialysis could have deleterious effects. Hence, although albumin dialysis appears safe, careful monitoring of hemodynamic, infectious, and laboratory parameters, and patient selection, are essential.
Efficacy of Albumin Dialysis in Pediatrics
Despite large emerging trials evaluating albumin dialysis in adult liver failure, pediatric sources are few. There are only case reports and case series for the use of MARS in PALF, which have been summarized in Table 31.1 . Some evidence exists for SPAD in pediatric liver failure. There are no studies available for pediatric AoCLF, and there have been no reports published to date using Prometheus in older children because of challenges with large extracorporeal volumes (700–750 mL). An adapted Prometheus protocol for children from a group in Belgium was presented as an abstract; however, the full report is still pending.
|Study, Year, Location of Study||No. of Patients |
|Cause of Liver Failure||MARS Treatment Regimen||Pre-MARS Parameters Or Mean MELD-PELD-CTP Score||Biochemical, Coagulation, Neurological Parameters Post-MARS||Outcome||Adverse Effects Secondary to MARS|
|Hommann et al., |
|(1, 25)||ALF, PNF||4 days, |
3 or 4 hours daily
|SBR, 41 mg/dL||↓ SBR, ↑ NH 3 ; ↔ INR, albumin||Bridged to re-LT||Hypotension ➔ resolved with MARS- mini|
|Prokurat et al., |
|(3, 5)||ALF, copper-chromium intoxication||Two sessions (18, 17 hours)||SBR, 13.9 mg/dL; |
INR, 3.3; Cr, 1.4 mg/dL
|↓ SBR; ↓ NH 3 ; ↔ INR, albumin||Death||No complications|
|Covic et al., |
|6 (7, 10, 10, 12, 14, 16)||ALF, |
mushroom poisoning (6)
|Two 6-hour sessions||HE—I (4), II (1), IV (1); CTP, 6.7||↓ SBR; ↓ INR; ↓ HE grade (4); ↔ HE, INR, SBR (2)||Bridged to recovery (4); death (2)||No complications|
|Rustom et al., |
|4 (7, 10, 13, 15)||ALF, WD (4)||Two to four sessions||WI > 11 (12, 18, 14, 16)||↓ SBR; ↔/↓INR (3); ↓ NH 3 (2); ↑GCS 2 points (4)||Bridged to LT (4)||Catheter- related hemorrhage; resolved |
|Auth et al., |
|2 (6, 14)||ALF, WD (2)||Two 6-hour sessions||PELD, 23, 24; |
HE, III or IV (2)
|↓ SBR (2); ↓ NH 3 (1); ↑ NH 3 (1); no HE (2) at 3 weeks||Bridged to LT (2)||Thrombocytopenia|
|Novelli et al., 2008, |
|6 (3, 9, 11, 12, 14, 15)||ALF, PCM (2), HBV (1), HZV (1), unknown (2)||Continuous||PELD, 39.3 (mean)||↓ SBR; ↓ NH 3 ↓ Cr; ↔ INR (6); ↓ ICP (4/4)||Bridged to LT (3); bridged to recovery (2); death (1)||No complications|
|Bourgoin et al., 2014, |
|6 (0.08, 0.08, 0.25, 0.92, 15, 16)||ALF—GAD (2), VOD (1) |
|One to six sessions; |
6- to 8-hour sessions
|MELD-PELD, 42.1 (mean)||↓ SBR; ↓ NH 3 ; ↔ INR (3); |
↑SBR, ↑ NH 3 (3; all MARS-mini)
|Death (3); bridged to LT (2); |
bridged to recovery (1)
|Hypotension, PTX, clotting events|
|Lexmond et al., 2015, |
|20 (7.1 ± 5.6; mean ± SD)||ALF (17), |
|One or two sessions (until LT); 8-hour sessions||MELD-PELD, 47, 6 (mean); HE grade 3,4 ± 0.6||HE ↓ (6); |
↓SBR, ↓ NH 3
|Death (3); bridged to LT (16); bridged to recovery (1)||Thrombocytopenia; serious bleeding (5)|