Beyond KCH selection and options in acute liver failure

Beyond KCH selection and options in acute liver failure The King’s prognostic criteria for patients with acute liver failure (ALF) introduced in 1989 have been used worldwide. This distinguished for the first time cases with ‘hyper-acute’ course (characteristically paracetamol overdose) where there is a better chance of recovery with medical supportive care alone from those etiologies with a less acute course and paradoxically lower chances of ‘spontaneous’ recovery. Ongoing use showed the limited sensitivity of the criteria to constitute a significant practical limitation. Subsequent models including the MELD score and composite ones with markers of necrosis, an apoptotic liver cell death, proposed to improve sensitivity did not have the required high specificity. Two recent models utilizing new availability of web- and app-based computing delivering outcome predication through sophisticated algorithms are described. The first is a dynamic model described for paracetamol-induced ALF based upon admission findings and sequential variables over the first 2 days. The new model of the US Acute Liver Failure group was devised to cover all etiologies of ALF for predicting ‘transplant-free’ survival and accurately predicated spontaneous survival in two-thirds of cases. Improved survival results with medical management, particularly in hyper-acute cases, now approach those obtained with successful liver transplant and have raised the question of transplant benefit. Also considered in the review are new non-transplant approaches to treatment including the use of plasma exchange and based on successful results in acute-on-chronic liver failure, agents to modulate and improve hepatic regeneration. Keywords Acute liver failure  King’s criteria  Dynamic model  ALFSG  Plasma exchange  Liver transplant Introduction and historical context would be transplanted who would otherwise have recov- ered with medical management alone; variously termed Difficulties in determining prognosis in patients with acute ‘transplant-free’ or ‘spontaneous’ survival. liver failure (ALF) from paracetamol hepatotoxicity have The description of the King’s College Criteria (KCC) in been apparent for many years. Initial issues related to the 1989 represented a major step forward in identifying can- difficulties in identifying those patients with paracetamol didates for liver transplantation (Text Box 1) [2]. It stres- intoxication who would benefit from antidotal therapy [1]. sed the importance of etiology and mode of presentation in The need for accurate predictive tests of patient outcome the outcome of ALF, recognizing the chances of recovery became even more apparent when the life-saving procedure with medical supportive care alone with rapidly evolving of liver transplant for patients with ALF was introduced in paracetamol hepatotoxicity as being much higher than with the early 1980. Concerns were raised that some patients other etiologies of ALF, particularly those with illness of more gradual onset. Derived and validated in a cohort of 763 patients managed before the introduction of trans- & Roger Williams plantation for ALF, the high specificity of the KCC meant r.williams@researchinliver.org.uk that relatively few cases fulfilling criteria would be 1 unnecessarily transplanted. However, their sensitivity was The Institute of Liver Studies, Cheyne Wing, King’s College lower, indicating that a significant number of cases not Hospital, Denmark Hill, London SE5 9RS, UK fulfilling criteria would progress and die without earlier The Institute of Hepatology London and Foundation for Liver identification and consideration for possible liver trans- Research, 111 Coldharbour Lane, London SE5 9NT, UK plantation. For the non-paracetamol cases, both sensitivity Faculty of Life Sciences & Medicine, King’s College, and specificity were high but within this overall group, London, UK 123 Hepatology International (2018) 12:204–213 205 certain etiologies with rapid illness onset—specifically These changes in the spectrum of complications and ischemic hepatitis, acute fatty liver of pregnancy and improvement in survival have since been documented in hepatitis A—had similar survival rates with medical man- other large historical series [8–10]. agement alone to that of paracetamol-induced cases. It was In this review, we will consider: first, the basis of in those cases with an indolent presentation and ‘sub-acute’ prognostic assessment systems in ALF and the practical phenotype and often ‘indeterminate’ etiology that had the difficulties in their application, with focus upon the KCC as worst outcome without transplantation [3]. an example of a system in current use. Second, the improved results currently being obtained with liver Text Box 1 transplantation in these very sick patients are described and third, how new approaches to treatment may impact upon The original Kings College Criteria for poor prognosis in future care and outcome. Acute Liver Failure. Paracetamol: pH<7.3 (irrespective of grade of encephalopathy) Improving prediction of outcome or Prothrombin time >100 s and serum creatinine >300 µmol/L In patients with grade III or IV encephalopathy. A number of different systems are in current use for the evaluation of prognosis in patients with ALF, with clinical Non-paracetamol: experience of their application and operational perfor- Prothrombin time >100 s (irrespective of grade of encephalopathy) mance reported for the KCC and criteria from France and Or Japan [2, 10–12]. Though details of the models used vary, Any 3 of the following variable (irrespective of grade of they share common features (Table 1). All recognize the encephalopathy) key importance of the development of encephalopathy as a Age <10 or > 40 years Etiology –non A, non B hepatitis, halothane hepatitis, marker of critically impaired liver function and retain it as idiosyncratic drug reactions a central criterion of poor prognosis. All also utilize lab- Duration of jaundice before onset of encephalopathy >7 days oratory measures of liver function to quantify severity of Prothrombin time >50 s Serum Bilirubin >300 µmol/L liver injury, most common measures of coagulation status. Source: O’Grady et al. Patient Age is also recognized by all as being of prognostic importance—likely reflecting the compromise of physio- Current UK ALF Transplant wait listing criterial found at http://odt.nhs.uk/pdf/liver_selection_policy.pdf logic reserve with increasing age—key to surviving any critical illness, and a possible parallel impairment of hep- atic regenerative ability. An important difference of the KCC from the other established criteria is the distinction Decision making in relation to the use of transplantation between paracetamol and non-paracetamol etiologies, in ALF became even more difficult as improvements in reflecting the high volume of paracetamol cases seen in the intensive liver care led to higher survival rates with med- United Kingdom and differences seen in clinical course. ical management alone. In the King’s College Hospital Their clinical simplicity has, however, served them well (KCH) experience of over 3000 patients treated between over time, relying not on complex calculation or use of 1973 and 2008, transplant-free survival had increased to specialized laboratory tests but rather standard bedside 48%, with a near fourfold increase in overall hospital survival rates—including cases transplanted from 16.7 to Table 1 Comparison of elements of prognostic scoring systems in 62.2% [4].The increase in ‘transplant-free’ survival was current widespread use most apparent in paracetamol cases and associated with a substantial fall in the occurrence of cerebral edema and Factor Criteria intracranial hypertension, which in the past was often the b c d Kings Clichy Japanese cause of death in cases with a ‘hyper-acute’ presentation. Age ?? ? An important contribution to this improvement is likely Etiology ? –– through control of circulating levels of ammonia—the Encephalopathy ?? ? principal neurotoxin in this setting—through interventions Bilirubin -/? – ? that include the earlier and more widespread use of con- tinuous hemofiltration [5, 6]. In paracetamol-induced ALF, Coagulopathy ?? ? the narrowing in the gap between survival with medical ? Factor included in criteria. - Factor not included in criteria management alone and survival with transplantation has a b c d Factor common to all criteria. References: [2], [56], [57] raised doubts about the transplant benefit in such cases [7]. 123 206 Hepatology International (2018) 12:204–213 assessment and readily available routine laboratory both paracetamol and non-paracetamol etiologies sug- measures. gested an overall sensitivity of 59% and specificity of 79%, Published data illustrate their limitations in current and with better performance values for prediction of non- clinical practice, particularly in relation to hyper-acute survival in paracetamol than in non-paracetamol etiologies disease. Here the initial intensity of critical illness may be [22]. severe but paradoxically the potential for native liver Indications are, therefore, of particular aspects where regeneration is high [4, 9, 11]. In early case series of hyper- diagnostic performance of selection criteria could be acute patients, the dominant clinical issues were those of improved beyond existing systems. First, they should either a failure to identify all patients with a poor prognosis consider etiology of illness and reflect the differential in for consideration of transplantation—limited sensitivity— current outcomes of medical care alone. Second, they in concert with that of rapid deterioration of those who would perform sequential rather than single time point fulfilled KCC and high risk of death on the transplant estimations of prognosis, a process identified on meta- waitlist prior to grafts becoming available, late identifica- analysis as having additional diagnostic value and reflect- tion [13–15]. Efforts were, therefore, initially focussed on ing the rapidly changing clinical condition seen particularly refining the KCC by improving criteria sensitivity enabling in paracetamol and other hyper-acute etiologies and earlier identification of patients with a poor prognosis already utilized in other novel prognostic criteria [23]. through integration of supplemental prognostic markers. Recent studies suggest that in these cases waitlist Initial studies explored the inclusion of arterial blood improvement rather than deterioration may now principally lactate measurements determined using point of care test- confound decision making [9, 11, 12, 15]. Additionally, ing, as hyperlactatemia reflects both decreased hepatic any new or supplemental criteria should ideally retain the clearance by a damaged liver and global illness severity clinical simplicity of existing systems. Two recent attempts and multi-organ failure. Initial reports confirmed their to address these issues are discussed below, both of which prognostic value, and their inclusion was found to improve have utilized the recent availability of web- and app-based specificity, sensitivity and timeliness of the KCC [16]. computing power to deliver outcome predictions through Other studies have also examined arterial blood lactate in sophisticated algorithms. ALF, finding elevated levels to be strongly and indepen- dently associated with death or transplantation in both paracetamol and non-paracetamol-induced disease Dynamic outcome predictive model [17–19]. Analysis of the performance of the specific for paracetamol-induced ALF thresholds for arterial lactate introduced into the KCC for paracetamol-induced disease has been less consistent, with Prediction models in hyper-acute disease would be of most one report suggesting that early levels resulted in an benefit if they could be applied sequentially identifying and increase in sensitivity but at the expense of reduction in quantifying both improvement and deterioration [22, 23]. specificity [17]. However, two other studies found that The latest model from the King’s Liver Intensive Care Unit lactate measurements alone 12 h after admission to trans- is a dynamic outcome prediction model developed and plant centers had high predictive accuracy, with perfor- validated for use in patients with paracetamol-induced ALF mance greater than the KCC alone [18, 19]. Though [24]. It is based on prospective data including analysis of adopted into United Kingdom ALF wait-listing criteria, more than 20 daily variables sequentially assessed for meta-analysis of validation studies to date has failed to 3 days after ICU admission in 912 un-transplanted patients confirm an improvement in diagnostic test performance between 2000 and 2012. The variables included in the final through their inclusion [20]. models to predict death-included age, hepatic Two further meta-analyses of the performance of the encephalopathy, cardiovascular failure, INR, creatinine and KCC have further illustrated their potential shortcomings arterial pH on admission and dynamic variables of (Fig. 1). Review of studies of the KCC in predicting out- changing arterial blood lactate and INR. On validation in come of non-paracetamol-induced ALF comprising 1105 independent datasets from four transplant centers, the cases, showed overall sensitivity of 68% and specificity models showed good discrimination between survivors and 82% [21]. Specificity was highest (93%) in patients with non-survivors, improving with the inclusion of changes in high-grade HE and where the criteria were sequentially INR and Lactate over time (Fig. 2). Innovative in this determined through the clinical course of illness. Impor- approach was its access though a dedicated website and the tantly, this analysis also described a fall-off in diagnostic generation of continuous survival estimates rather than a performance in more recent studies, reflecting the increased binary survival outcome, with the intention that the model success of non-transplant approaches to care. A second, should act as a decision-support tool to support clinical larger meta-analysis examining criteria performance in 123 Hepatology International (2018) 12:204–213 207 Fig. 1 Pooled (a) Sensitivity and (b) Specificity of Kings College Criteria (Source Ref. [20]) 123 208 Hepatology International (2018) 12:204–213 Fig. 2 Discrimination and calibration of dynamic prediction model (AUROC) for day 1 model, b AUROC for day 2 model, c Calibration on day 1 and day 2 of admission in 150 patients with paracetamol- curve for day 1 model, d Calibration curve for day 2 model (Source induced ALF. a Area under receiver operating characteristic curve Ref. [21]) judgement rather than a sole arbiter as to proceeding with transplantation. Model of acute liver failure study group (ALFSG) Elements of this approach were also adopted in the recent analysis by the ALFSG of 1974 subjects enrolled prospectively from 28 academic centers across North America between 1998 and 2013, and managed with and without transplantation [25]. The aim was to devise a mathematical model for all etiologies of ALF to predict transplant-free survival at 21 days, rather than mortality as adopted in other models—making comparison with other Fig. 3 Comparison of the ALFSG model with the Kings College studies adopting more standard approaches more complex. Criteria and MELD score in predicting survival. ALFSG acute liver Clinical features and laboratory values were collected at failure study group, MELD model for end stage liver disease APAP study enrolment and recorded serially up to 7 days. Vari- acetaminophen (paracetamol). AUROC: ALFSG: 0.843, MELD: ables of prognostic value adopted in the predictive model 0.717, KCC APAP 0.560, KCC non-APAP 0.655 (Source Ref. [22]) included admission coma grade, etiology and vasopressor requirement, and admission bilirubin and INR values. with the model correctly predicting outcome of illness in 66.3% of subjects. Its performance was best in patients Arterial blood lactate was not explored. In this analyses of both paracetamol and non-paracetamol cases, sequential with unfavorable etiologies and high-grade encephalopathy and more limited in those with more favorable etiologies values of the INR did not add to prediction over that on admission. Using AUROC analysis, test discrimination and high-grade encephalopathy. appeared superior to the KCC and MELD scores (Fig. 3) 123 Hepatology International (2018) 12:204–213 209 Though sharing similarities in their use of web-based predictive algorithms, key differences exist between these two predictive systems. One is based on the prediction of mortality and the other survival, with resultant differences in test performance. Further, they are targeted at different patient groups—with the USALF criteria taking a broad approach to all etiologies of ALF, whilst the Dynamic Model is focused upon liver injury from paracetamol where the USALF performs least well. It may be that application of these and other models are complimentary and decision making is based upon the results generated from several prognostic systems [23–25]. Other systems evaluated A variety of other scoring systems and supplemental markers have been proposed to identify candidates for liver transplantation, though in general their performance has been evaluated in patient cohorts of limited size (Table 2). Scores used in this way include non-liver specific ones such as SOFA and APACHE II that are widely used to quantify severity of multi-organ failure in other forms of critical illness [17, 26, 27]. Reports suggest similar sensi- tivity and specificity to the King’s criteria in predicting death and they can be applied sequentially. However, though they may identify those patients at risk of death, the multi-organ failure they quantify may not be corrected by transplantation, if for example, it results for other ALF- associated complications such as severe sepsis or pancre- atitis. Consequently, they are not in widespread use to select candidates for transplant. The MELD score is more focused on severity of liver injury and has been assessed more widely in ALF and subjected to meta-analysis [22, 28, 29]. In its unmodified form, it shows most promise in non-paracetamol etiologies, with diagnostic performance close to that of the KCC [22]. It has also been combined with circulating blood levels of cytokeratin K18 (CK18) a cell death-associated marker measured using the M30 assay which principally reflects apoptotic cell death [30]. Also, this measurement in place of bilirubin to give a modified MELD Score showed superior sensitivity and specificity to the standard MELD Score and the King’s criteria in predicting outcome of ALF [31]. Other studies have also reported prognostic models combining standard clinical variables with non-standard analytes. By example, the ALFSG index predictive model combined Coma-grade, INR, bilirubin and phosphorus levels with that of blood levels of M30 CK18 [32]. In a 250-strong validation cohort, discrimination between sur- vivors and non-survivors was greater than with the KCC or MELD, though later external validation in a smaller Table 2 Features of illustrative systems for mortality prediction in acute liver failure Study Publication Criteria Era Etiology n Non- Sensitivity Specificity AUROC Comment year survivors McPhail et al. [20] 2016 KCC 2001–2015 All 2153 Not Given 0.59 0.79 0.76 Meta-analysis (0.56–0.62) (0.77–0.81) Cholongitas et al. 2012 SOFA 1993–2010 Paracetamol 125 58 (46%) 0.67 0.80 0.79 SOFA Score threshold 12 [25] McPhail et al. [20] 2016 MELD 2001–2015 All 2153 Not given 0.74 0.67 0.78 Meta-analysis (0.71–0.77) (0.64–0.69) Bechmann et al. [29] 2010 MELD/CK18 2006–2009 All 68 18 (27%) 0.85 (0.69–1.0) 0.76 (0.6–0.91) 0.94 Peak M65 fragment Koch et al. [22] 2016 ALFSG 1998–2013 All 1974 987 (50%) Not given Not given 0.84 Prediction of survival not death Rutherford et al. [30] 2012 ALFSG/CK18 1998–2011 All 500 251 (50%) 0.86 0.65 0.82 Admission M30 fragment Antoine et al. [32] 2012 Acetylated Not Given Paracetamol 78 27 (35%) Not given Not given 0.87 Admission values HMGB1 KCC Kings College Criteria, SOFA sequential organ failure assessment, MELD model for end-stage liver disease, CK18 cytokeratin 18, ALFGSG Acute liver failure study group, HMGB1 high- mobility group box-1 Non-survivors; cases who died or were transplanted. Figures in parentheses are 95% Confidence Intervals where given 210 Hepatology International (2018) 12:204–213 independent cohort showed no advantage over SOFA or APACHE scoring [32, 33]. However, the primary mechanism of liver cell injury may vary by etiology and necrosis rather than apoptosis may predominate. Antoine and colleagues assessed circu- lating levels of a panel of cell death markers in 78 patients with paracetamol-induced ALF [34]. They found that best prediction of non-survival was with acetylated HMGBI—a biomarker of hepatic necrosis and cellular immunological activation, and to lesser degree with biomarkers of apop- tosis including molecular forms of CK18. To date, vali- dation studies have not indicated prognostic advantage above the standard KCC in predicting survival [35]. Cell death biomarkers do appear to be exquisitely sensitive in predicting clinically significant liver injury very early after paracetamol overdose but If they are to be adopted as adjunctive measures to select liver transplant candidates with ALF, key practical issues will need to be overcome [36, 37]. Techniques for their rapid determination will need to be widely available, and it is likely that a panel of biomarkers will need to be assessed to cover etiology- specific differences in the mechanisms of liver cell injury and death. Other simpler clinical measures have also demonstrated potential for use as prognostic markers: platelet count has been shown to be closely linked to outcome. In a recent study from the USALFSG, the evolution of thrombocy- Fig. 4 a Platelet Counts on days 1–7 after admission in 1598 patients topenia was closely associated with development of multi- with acute liver failure according to outcome at 21 days. SS organ failure and a poor outcome in ALF and linked to the spontaneous survivor, OLT orthotropic liver transplant. Three sym- development of a systemic inflammatory response [38] bols, p \ 0.001; two symbols, p \ 0.01; one symbol, p \ 0.05. *SS (Fig. 4a). Liver volume may be easily determined using vs LT, SS vs death, and LT vs death. SS, spontaneous survival. b Survival According to CT-derived liver volume on admission in 37 analysis of CT images and reflect the balance between the non-transplanted patients with non-acetaminophen etiologies. parallel processes of liver collapse with cell death and 3 3 \ 1000 cm n = 11, C 1000 cm n = 26, p \ 0.001 log-rank increase with regeneration. Recent studies suggest that in (Source Ref. [36, 38]) some non-paracetamol etiologies, loss of liver volume in adults to less than 1000 cm may indicate irreversible absolute arbiter, these models should support decision damage and serve as an early indicator of poor prognosis, making and the multifactorial team assessment. often in advance of the development of encephalopathy (Fig. 4b) [39–41]. Other studies have also sought to identify at a very early Current results of liver transplantation stage those patients with acute liver injury (ALI) in the absence of encephalopathy which will later progress to a Life-saving though it was soon seen to be after its intro- poor outcome of death or transplantation. In a study of US duction, overall results not surprisingly for these very sick patients hospitalized with ALI, progression to a poor out- patients were significantly lower than for elective trans- come was more common in non-paracetamol etiologies and plants carried out for those with chronic diseases. Several with higher levels of INR and bilirubin, and a prolonged series have now demonstrated a progressive incremental duration of illness [42]. improvement in survival over time, with patient survival Finally, it is important to emphasize that prognostic now approaching that of elective surgery. Recent reports models should be only part of the overall functional eval- for the US and UK are of 1-year patient survival rates of uation of the very sick patient with ALF and an experi- 85–90% [43]. These results reflect advances in peri- and enced multi-disciplinary team in an intensive care setting is intra-operative care and the interplay of several factors: the required for correct interpretation. Rather than providing an earlier and more accurate prediction of need for transplant 123 Hepatology International (2018) 12:204–213 211 and identification those patients who are too sick for sur- pre-existing sub-clinical non-cirrhotic liver disease upon gery, and a better understanding of the interplay between liver regeneration is currently poorly characterized, for graft and recipient factors in determining patient survival. example, from chronic alcohol use or non-alcoholic Analysis of the European experience of 4903 transplant steatohepatitis. recipients between 1998 and 2009 found death after sur- In cases where the liver injury has progressed ‘beyond gery to relate to both with independent association shown the point of no return’, novel medical therapies are unlikely with male recipients (adjusted OR 1.25), recipient to be of benefit. It would thus be of no surprise that most [ 50 years of age (1.26), incompatible ABO matching efficacy would be seen either in treatment at an early stage (1.93), donors [ 60 years (1.21), and reduced size graft of disease and/or in etiologies with greatest regenerative (1.54). Recipients [ 50 years, combined with donors capacity. By example, the FULMAR trial of the MARS [ 60 years had a 57% mortality/graft loss within the first extra-corporeal device in patients with ALF failed to show year after transplant [44]. Matching of graft to recipient is a survival improvement in the overall study cohort but key to ensuring optimal outcome [13]. Case series also now there were indications of potential benefit in patients with confirms the practicality and excellent outcomes obtained paracetamol-related disease [49]. for selected patients transplanted using living donors, and Such caveats may also apply to the use of therapeutic from auxiliary liver grafting in etiologies where native liver plasma exchange (TPE). Through mechanisms thought to regenerative potential exists—though globally, these tech- include removal of deleterious inhibitors of hepatic niques are applied to only a small proportion of recipients regeneration and a complex pattern of immunomodulation, [45, 46]. use of this therapy has recently been demonstrated to Patient survival after transplantation for ALF follows a deliver a survival benefit in non-transplanted patients with characteristic pattern of increased mortality early after ALF. In a multi-center randomized controlled trial on transplant, but in patients who survive this phase, subse- mixed etiologies of disease, three 5-l TPE sessions signif- quent survival parallels closely with that seen in elective icantly improved survival above that seen with standard transplantation [44]. In an early series, the high prevalence medical care, although the survival benefit was inferior to of psychiatric comorbidity and addiction issues seen in that seen with transplantation (Fig. 5)[50]. There is recipients of transplantation for paracetamol-induced ALF ongoing debate about the place of this therapy in patients was reflected in an increased risk of death due to suicide, with ALF, given that its use normalizes measures of trauma or non-adherence to immunosuppression with over coagulation disturbance and thus precludes their use in half the deaths occurring within 12 months of transplan- prognostic evaluation for later transplantation. Our practice tation [44]. Later series have not demonstrated such an is in accordance with recent EASL guidance, with its early increase in mortality, but it is clear that paracetamol use in patients with ALF with an expected poor prognosis recipients show increased levels of psychotic comorbidity without transplantation but who have clear medical or which may be reflected in poor compliance with medica- psychiatric contra-indications to surgery [51]. In addition, tion and follow-up [47]. In these patients, assessment of the use of TPE is associated with significant improvements severity of psychiatric illness is a key element of the in cardiovascular status and we also treat patients who are selection of potential recipients and in those transplanted waitlisted for liver transplantation but with worsening follow-up must include close ongoing psychiatric moni- toring and support. New novel non-transplant options A number of novel non-transplant interventions have potential to serve as alternatives to transplantation. A central premise to their use is that the injured liver retains regenerative ability and that this may be augmented by these interventions. A better practical understanding of the processes influencing both liver injury and regeneration is Fig. 5 Survival in 182 patients with ALF according to treatment with likely to be of clinical importance here, exemplified by the liver transplantation and high-volume plasma exchange. Two groups receiving SMT (standard medical treated group) with and without major clinical and laboratory differences between parac- emergency transplantation (HVP ? LTx vs ?HVPLTx) and the two etamol hepatotoxity from single time point or staggered group receiving SMT with and without emergency transplantation overdoses—and the markedly worse outcomes seen in the (HVPLTx vs. ?HVPLTx) (log rank: p = 0.0058) and Cox propor- latter group [48]. The potentially important influence of tional hazard: LTx: p \ 0.0001; HVP: p = 0.0076) (Source Ref. [48]) 123 212 Hepatology International (2018) 12:204–213 3. O’Grady JG, Schalm SW, Williams R. Acute liver failure: cardiovascular failure and vasopressor requirement while redefining the syndromes. Lancet 1993;342(8866):273–275. awaiting a graft [50]. 4. Bernal W, Hyyrylainen A, Gera A, et al. Lessons from look-back Other forms of cellular and immune-modulatory therapy in acute liver failure? A single centre experience of 3300 patients. show promise as future treatments for ALF. Whilst hepato- J Hepatol 2013;59(1):74–80. 5. Slack A, Auzinger G, Willars C, et al. Ammonia clearance with cyte transplantation retains considerable theoretical attrac- haemofiltration in adults with liver disease. Liver Int tions, it presents considerable logistic challenges and a 2014;34(1):42–48. requirement for immunosuppression, and its benefit is yet to 6. Cardoso FS, Gottfried M, Tujios S, Olson JC, Karvellas CJ. be established in adult patients with ALF [52, 53]. Other Group USALFS. Continuous renal replacement therapy is asso- ciated with reduced serum ammonia levels and mortality in acute interventions in patients with acute-on-chronic liver failure liver failure. Hepatology 2017;31:31. (ACLF) show promise but are yet to be tested in ALF. 7. O’Grady J. Timing and benefit of liver transplantation in acute Understanding of the pathophysiologic basis of ACLF is liver failure. J Hepatol 2014;60(3):663–670. rapidly increasing and the insights gained, particularly in 8. Donnelly MC, Davidson JS, Martin K, Baird A, Hayes PC, Simpson KJ. Acute liver failure in Scotland: changes in aetiology relation to modulation of regeneration and immunological and outcomes over time (the Scottish Look-Back Study). Aliment function, may well be transportable to ALF. In a randomized Pharmacol Ther 2017;45(6):833–843. controlled single-blind trial in 110 patients with Hepatitis 9. Reuben A, Tillman H, Fontana R, et al. Outcomes in Adults With B-related ACLF of whom less than half were thought to be Acute Liver Failure Between 1998 and 2013: an observational cohort study. Ann Intern Med 2016;164(11):724–732. cirrhotic, treatment with allogenic mesenchymal stem cells 10. Oketani M, Ido A, Nakayama N, et al. Etiology and prognosis of improved survival—[54]. Similarly, in a single-center dou- fulminant hepatitis and late-onset hepatic failure in Japan: sum- ble-blind trial in 55 patients with decompensated cirrhosis, mary of the annual nationwide survey between 2004 and 2009. treatment with granulocyte colony stimulating factor and Hepatol Res 2013;43(2):97–105. 11. Ichai P, Legeai C, Francoz C, et al. Patients with acute liver erythropoietin was associated with improved survival in failure listed for superurgent liver transplantation in France: association with indications of improved hepatic regenera- reevaluation of the Clichy-Villejuif criteria. Liver Transpl tion through mobilization of progenitor cells [55]. These 2015;21(4):512–523. results are yet to be replicated or tested in patients with ALF 12. Simpson KJ, Bates CM, Henderson NC, et al. The utilization of liver transplantation in the management of acute liver failure: but demonstrate mechanistic approaches likely to be asses- comparison between acetaminophen and non-acetaminophen sed in the near future. As ever, the rarity and severity of ALF etiologies. Liver Transpl 2009;15(6):600–609. represent major limiting factors for performing randomized 13. Bernal W, Cross TJS, Auzinger G, et al. Outcome after wait- controlled trials of these or other agents and true advances in listing for emergency liver transplantation in acute liver failure: a single centre experience. J Hepatol 2009;50(2):306–313. therapeutic intervention are likely to result only from inter- 14. Brandsaeter B, Hockerstedt K, Friman S, et al. Fulminant hepatic national collaborative studies. failure: outcome after listing for highly urgent liver transplanta- tion-12 years experience in the nordic countries. Liver Transpl 2002;8(11):1055–1062. 15. Reddy KR, Ellerbe C, Schilsky M, et al. Determinants of out- Compliance with ethical standards come among patients with acute liver failure listed for liver transplantation in the United States. Liver Transpl 2016;22(4):505–515. Conflict of interest Professor Roger Williams and Professor William 16. Bernal W, Donaldson N, Wyncoll D, Wendon J. Blood lactate as Bernal have no potential conflict of interest. an early predictor of outcome in paracetamol-induced acute liver failure: a cohort study. Lancet 2002;359(9306):558–563. Open Access This article is distributed under the terms of the Creative 17. Schmidt LE, Larsen FS. Prognostic implications of hyperlac- Commons Attribution 4.0 International License (http://creative tatemia, multiple organ failure, and systemic inflammatory commons.org/licenses/by/4.0/), which permits unrestricted use, dis- response syndrome in patients with acetaminophen-induced acute tribution, and reproduction in any medium, provided you give liver failure. Crit Care Med 2006;34(2):337–343. appropriate credit to the original author(s) and the source, provide a 18. Macquillan GC, Seyam MS, Nightingale P, Neuberger JM, link to the Creative Commons license, and indicate if changes were Murphy N. Blood lactate but not serum phosphate levels can made. predict patient outcome in fulminant hepatic failure. Liver Transpl 2005;11(9):1073–1079. 19. Cholongitas E, O’Beirne J, Betrossian A, et al. Prognostic impact References of lactate in acute liver failure. Liver Transpl 2008;14(1):121–122 (author reply 3). 1. Gazzard BG, Widdop B, Davis M, Hughes RD, Goulding R, 20. Craig DGN, Ford AC, Hayes PC, Simpson KJ. Systematic Williams R. Early prediction of the outcome of a paracetamol review: prognostic tests of paracetamol-induced acute liver fail- overdose based on an analysis of 163 patients. Postgrad Med J ure. Aliment Pharmacol Ther 2010;31(10):1064–1076. 1977;53(619):243–247. 21. McPhail MJW, Wendon JA, Bernal W. Meta-analysis of per- 2. O’Grady JG, Alexander GJ, Hayllar KM, Williams R. Early formance of Kings’s College Hospital Criteria in prediction of indicators of prognosis in fulminant hepatic failure. Gastroen- outcome in non-paracetamol-induced acute liver failure. J Hepa- terology 1989;97(2):439–445. tol 2010;53(3):492–499. 123 Hepatology International (2018) 12:204–213 213 22. McPhail MJ, Farne H, Senvar N, Wendon JA, Bernal W. Ability 40. Zabron A, Quaglia A, Peddu P, et al. Clinical and prognostic of King’s College Criteria and model for end-stage liver disease associations of liver volume determined by computed tomogra- scores to predict mortality of patients with acute liver failure: a phy in acute liver failure. Liver Int 2018 (In Press). meta-analysis. Clin Gastroenterol Hepatol 41. Yamagishi Y, Saito H, Tada S, et al. Value of computed 2016;14(4):516–525.e5 (quiz e43–e45). tomography-derived estimated liver volume/standard liver vol- 23. Kumar R, Shalimar Sharma H, et al. Prospective derivation and ume ratio for predicting the prognosis of adult fulminant hepatic validation of early dynamic model for predicting outcome in failure in Japan. J Gastroenterol Hepatol 2005;20(12):1843–1849. patients with acute liver failure. Gut 2012;61(7):1068–1075. 42. Koch DG, Speiser JL, Durkalski V, et al. The natural history of 24. Bernal W, Wang Y, Maggs J, et al. Development and validation severe acute liver injury. Am J Gastroenterol of a dynamic outcome prediction model for paracetamol-induced 2017;112(9):1389–1396. acute liver failure: a cohort study. Lancet Gastroenterol Hepatol 43. Blood and Transplant N. ANNUAL REPORT ON LIVER 2016;1(3):217–225. TRANSPLANTATION. 2017. https://nhsbtdbe.blob.core.win 25. Koch DG, Tillman H, Durkalski V, Lee WM, Reuben A. dows.net/umbraco-assets-corp/5007/annual_liver_transplantation_ Development of a model to predict transplant-free survival of report_2017.pdf Accessed 19 Feb 2018. patients with acute liver failure. Clin Gastroenterol Hepatol 44. Germani G, Theocharidou E, Adam R, et al. Liver transplantation 2016;14(8):1199–1206.e2. for acute liver failure in Europe: outcomes over 20 years from the 26. Craig DG, Zafar S, Reid TW, et al. The sequential organ failure ELTR database. J Hepatol 2012;57(2):288–296. assessment (SOFA) score is an effective triage marker following 45. Yamashiki N, Sugawara Y, Tamura S, et al. Outcomes after staggered paracetamol (acetaminophen) overdose. Aliment living donor liver transplantation for acute liver failure in Japan: Pharmacol Ther 2012;35(12):1408–1415. results of a nationwide survey. Liver Transpl 27. Cholongitas E, Theocharidou E, Vasianopoulou P, et al. Com- 2012;18(9):1069–1077. parison of the sequential organ failure assessment score with the 46. Goldaracena N, Spetzler VN, Marquez M, et al. Live donor liver King’s College Hospital criteria and the model for end-stage liver transplantation: a valid alternative for critically ill patients suf- disease score for the prognosis of acetaminophen-induced acute fering from acute liver failure. Am J Transplant liver failure. Liver Transpl 2012;18(4):405–412. 2015;15(6):1591–1597. 28. Schmidt L, Larsen F. MELD score as a predictor of liver failure 47. Karvellas CJ, Safinia N, Auzinger G, et al. Medical and psychi- and death in patients with acetaminophen-induced liver injury. atric outcomes for patients transplanted for acetaminophen-in- Hepatology 2007;45:789–796. duced acute liver failure: a case-control study. Liver Int 29. Craig DGN, Reid TWDJ, Wright EC, et al. The sequential organ 2010;30(6):826–833. failure assessment (SOFA) score is prognostically superior to the 48. Craig DGN, Bates CM, Davidson JS, Martin KG, Hayes PC, model for end-stage liver disease (MELD) and MELD variants Simpson KJ. Staggered overdose pattern and delay to hospital following paracetamol (acetaminophen) overdose. Aliment presentation are associated with adverse outcomes following Pharmacol Ther 2012;35(6):705–713. paracetamol-induced hepatotoxicity. Br J Clin Pharmacol 30. Eguchi A, Wree A, Feldstein AE. Biomarkers of liver cell death. 2012;73(2):285–294. J Hepatol 2014;60(5):1063–1074. 49. Saliba F, Camus C, Durand F, et al. Albumin dialysis with a 31. Bechmann LP, Jochum C, Kocabayoglu P, et al. Cytokeratin noncell artificial liver support device in patients with acute liver 18-based modification of the MELD score improves prediction of failure: a randomized, controlled trial. Ann Intern Med spontaneous survival after acute liver injury. J Hepatol 2013;159(8):522–531. 2010;53(4):639–647. 50. Larsen FS, Schmidt LE, Bernsmeier C, et al. High-volume 32. Rutherford A, King LY, Hynan LS, et al. Development of an plasma exchange in patients with acute liver failure: an open accurate index for predicting outcomes of patients with acute randomised controlled trial. J Hepatol 2016;64(1):69–78. liver failure. Gastroenterology 2012;143(5):1237–1243. 51. European Association for the Study of the Liver. Electronic 33. Craig DG, Simpson KJ. Accuracy of the ALFSG index as a triage address eee, Clinical practice guidelines p, Wendon J, et al. marker in acute liver failure. Gastroenterology 2013;144(1):e25. EASL Clinical Practical Guidelines on the management of acute 34. Antoine DJ, Jenkins RE, Dear JW, et al. Molecular forms of (fulminant) liver failure. J Hepatol 2017;66(5):1047–1081. HMGB1 and keratin-18 as mechanistic biomarkers for mode of 52. Iansante V, Mitry RR, Filippi C, Fitzpatrick E, Dhawan A. cell death and prognosis during clinical acetaminophen hepato- Human hepatocyte transplantation for liver disease: current status toxicity. J Hepatol 2012;56(5):1070–1079. and future perspectives. Pediatr Res 2017;06:06. 35. Craig DGN, Lee P, Pryde EA, Masterton GS, Hayes PC, Simpson 53. Ibars EP, Cortes M, Tolosa L, et al. Hepatocyte transplantation KJ. Circulating apoptotic and necrotic cell death markers in program: lessons learned and future strategies. World J Gas- patients with acute liver injury. Liver Int 2011;31(8):1127–1136. troenterol 2016;22(2):874–886. 36. Dear JW, Clarke JI, Francis B, et al. Risk stratification after 54. Lin BL, Chen JF, Qiu WH, et al. Allogeneic bone marrow- paracetamol overdose using mechanistic biomarkers: results from derived mesenchymal stromal cells for hepatitis B virus-related two prospective cohort studies. Lancet Gastroenterol Hepatol acute-on-chronic liver failure: a randomized controlled trial. 2018;3(2):104–113. Hepatology 2017;66(1):209–219. 37. Antoine DJ, Dear JW, Lewis PS, et al. Mechanistic biomarkers 55. Kedarisetty CK, Anand L, Bhardwaj A, et al. Combination of provide early and sensitive detection of acetaminophen-induced granulocyte colony-stimulating factor and erythropoietin improves outcomes of patients with decompensated cirrhosis. acute liver injury at first presentation to hospital. Hepatology 2013;58(2):777–787. Gastroenterology 2015;148(7):1362–1370 (e7). 38. Stravitz RT, Ellerbe C, Durkalski V, et al. Thrombocytopenia Is 56. Bernuau J, Rueff B, Benhamou JP. Fulminant and subfulminant liver associated with multi-organ system failure in patients with acute failure: definitions and causes. Semin Liver Dis 1986;6(2):97–106. liver failure. Clin Gastroenterol Hepatol 2016;14(4):613–620. 57. Mochida S, Nakayama N, Matsui A, Nagoshi S, Fujiwara K. Re- 39. Shakil AO, Jones BC, Lee RG, Federle MP, Fung JJ, Rakela J. evaluation of the Guideline published by the Acute Liver Failure Prognostic value of abdominal CT scanning and hepatic Study Group of Japan in 1996 to determine the indications of histopathology in patients with acute liver failure. Dig Dis Sci liver transplantation in patients with fulminant hepatitis. Hepatol 2000;45(2):334–339. Res 2008;38(10):970–979. http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png Hepatology International Springer Journals

Beyond KCH selection and options in acute liver failure

Free
10 pages

Loading next page...
 
/lp/springer_journal/beyond-kch-selection-and-options-in-acute-liver-failure-LMdxexIoxs
Publisher
Springer Journals
Copyright
Copyright © 2018 by The Author(s)
Subject
Medicine & Public Health; Hepatology; Colorectal Surgery; Surgery
ISSN
1936-0533
eISSN
1936-0541
D.O.I.
10.1007/s12072-018-9869-7
Publisher site
See Article on Publisher Site

Abstract

The King’s prognostic criteria for patients with acute liver failure (ALF) introduced in 1989 have been used worldwide. This distinguished for the first time cases with ‘hyper-acute’ course (characteristically paracetamol overdose) where there is a better chance of recovery with medical supportive care alone from those etiologies with a less acute course and paradoxically lower chances of ‘spontaneous’ recovery. Ongoing use showed the limited sensitivity of the criteria to constitute a significant practical limitation. Subsequent models including the MELD score and composite ones with markers of necrosis, an apoptotic liver cell death, proposed to improve sensitivity did not have the required high specificity. Two recent models utilizing new availability of web- and app-based computing delivering outcome predication through sophisticated algorithms are described. The first is a dynamic model described for paracetamol-induced ALF based upon admission findings and sequential variables over the first 2 days. The new model of the US Acute Liver Failure group was devised to cover all etiologies of ALF for predicting ‘transplant-free’ survival and accurately predicated spontaneous survival in two-thirds of cases. Improved survival results with medical management, particularly in hyper-acute cases, now approach those obtained with successful liver transplant and have raised the question of transplant benefit. Also considered in the review are new non-transplant approaches to treatment including the use of plasma exchange and based on successful results in acute-on-chronic liver failure, agents to modulate and improve hepatic regeneration. Keywords Acute liver failure  King’s criteria  Dynamic model  ALFSG  Plasma exchange  Liver transplant Introduction and historical context would be transplanted who would otherwise have recov- ered with medical management alone; variously termed Difficulties in determining prognosis in patients with acute ‘transplant-free’ or ‘spontaneous’ survival. liver failure (ALF) from paracetamol hepatotoxicity have The description of the King’s College Criteria (KCC) in been apparent for many years. Initial issues related to the 1989 represented a major step forward in identifying can- difficulties in identifying those patients with paracetamol didates for liver transplantation (Text Box 1) [2]. It stres- intoxication who would benefit from antidotal therapy [1]. sed the importance of etiology and mode of presentation in The need for accurate predictive tests of patient outcome the outcome of ALF, recognizing the chances of recovery became even more apparent when the life-saving procedure with medical supportive care alone with rapidly evolving of liver transplant for patients with ALF was introduced in paracetamol hepatotoxicity as being much higher than with the early 1980. Concerns were raised that some patients other etiologies of ALF, particularly those with illness of more gradual onset. Derived and validated in a cohort of 763 patients managed before the introduction of trans- & Roger Williams plantation for ALF, the high specificity of the KCC meant r.williams@researchinliver.org.uk that relatively few cases fulfilling criteria would be 1 unnecessarily transplanted. However, their sensitivity was The Institute of Liver Studies, Cheyne Wing, King’s College lower, indicating that a significant number of cases not Hospital, Denmark Hill, London SE5 9RS, UK fulfilling criteria would progress and die without earlier The Institute of Hepatology London and Foundation for Liver identification and consideration for possible liver trans- Research, 111 Coldharbour Lane, London SE5 9NT, UK plantation. For the non-paracetamol cases, both sensitivity Faculty of Life Sciences & Medicine, King’s College, and specificity were high but within this overall group, London, UK 123 Hepatology International (2018) 12:204–213 205 certain etiologies with rapid illness onset—specifically These changes in the spectrum of complications and ischemic hepatitis, acute fatty liver of pregnancy and improvement in survival have since been documented in hepatitis A—had similar survival rates with medical man- other large historical series [8–10]. agement alone to that of paracetamol-induced cases. It was In this review, we will consider: first, the basis of in those cases with an indolent presentation and ‘sub-acute’ prognostic assessment systems in ALF and the practical phenotype and often ‘indeterminate’ etiology that had the difficulties in their application, with focus upon the KCC as worst outcome without transplantation [3]. an example of a system in current use. Second, the improved results currently being obtained with liver Text Box 1 transplantation in these very sick patients are described and third, how new approaches to treatment may impact upon The original Kings College Criteria for poor prognosis in future care and outcome. Acute Liver Failure. Paracetamol: pH<7.3 (irrespective of grade of encephalopathy) Improving prediction of outcome or Prothrombin time >100 s and serum creatinine >300 µmol/L In patients with grade III or IV encephalopathy. A number of different systems are in current use for the evaluation of prognosis in patients with ALF, with clinical Non-paracetamol: experience of their application and operational perfor- Prothrombin time >100 s (irrespective of grade of encephalopathy) mance reported for the KCC and criteria from France and Or Japan [2, 10–12]. Though details of the models used vary, Any 3 of the following variable (irrespective of grade of they share common features (Table 1). All recognize the encephalopathy) key importance of the development of encephalopathy as a Age <10 or > 40 years Etiology –non A, non B hepatitis, halothane hepatitis, marker of critically impaired liver function and retain it as idiosyncratic drug reactions a central criterion of poor prognosis. All also utilize lab- Duration of jaundice before onset of encephalopathy >7 days oratory measures of liver function to quantify severity of Prothrombin time >50 s Serum Bilirubin >300 µmol/L liver injury, most common measures of coagulation status. Source: O’Grady et al. Patient Age is also recognized by all as being of prognostic importance—likely reflecting the compromise of physio- Current UK ALF Transplant wait listing criterial found at http://odt.nhs.uk/pdf/liver_selection_policy.pdf logic reserve with increasing age—key to surviving any critical illness, and a possible parallel impairment of hep- atic regenerative ability. An important difference of the KCC from the other established criteria is the distinction Decision making in relation to the use of transplantation between paracetamol and non-paracetamol etiologies, in ALF became even more difficult as improvements in reflecting the high volume of paracetamol cases seen in the intensive liver care led to higher survival rates with med- United Kingdom and differences seen in clinical course. ical management alone. In the King’s College Hospital Their clinical simplicity has, however, served them well (KCH) experience of over 3000 patients treated between over time, relying not on complex calculation or use of 1973 and 2008, transplant-free survival had increased to specialized laboratory tests but rather standard bedside 48%, with a near fourfold increase in overall hospital survival rates—including cases transplanted from 16.7 to Table 1 Comparison of elements of prognostic scoring systems in 62.2% [4].The increase in ‘transplant-free’ survival was current widespread use most apparent in paracetamol cases and associated with a substantial fall in the occurrence of cerebral edema and Factor Criteria intracranial hypertension, which in the past was often the b c d Kings Clichy Japanese cause of death in cases with a ‘hyper-acute’ presentation. Age ?? ? An important contribution to this improvement is likely Etiology ? –– through control of circulating levels of ammonia—the Encephalopathy ?? ? principal neurotoxin in this setting—through interventions Bilirubin -/? – ? that include the earlier and more widespread use of con- tinuous hemofiltration [5, 6]. In paracetamol-induced ALF, Coagulopathy ?? ? the narrowing in the gap between survival with medical ? Factor included in criteria. - Factor not included in criteria management alone and survival with transplantation has a b c d Factor common to all criteria. References: [2], [56], [57] raised doubts about the transplant benefit in such cases [7]. 123 206 Hepatology International (2018) 12:204–213 assessment and readily available routine laboratory both paracetamol and non-paracetamol etiologies sug- measures. gested an overall sensitivity of 59% and specificity of 79%, Published data illustrate their limitations in current and with better performance values for prediction of non- clinical practice, particularly in relation to hyper-acute survival in paracetamol than in non-paracetamol etiologies disease. Here the initial intensity of critical illness may be [22]. severe but paradoxically the potential for native liver Indications are, therefore, of particular aspects where regeneration is high [4, 9, 11]. In early case series of hyper- diagnostic performance of selection criteria could be acute patients, the dominant clinical issues were those of improved beyond existing systems. First, they should either a failure to identify all patients with a poor prognosis consider etiology of illness and reflect the differential in for consideration of transplantation—limited sensitivity— current outcomes of medical care alone. Second, they in concert with that of rapid deterioration of those who would perform sequential rather than single time point fulfilled KCC and high risk of death on the transplant estimations of prognosis, a process identified on meta- waitlist prior to grafts becoming available, late identifica- analysis as having additional diagnostic value and reflect- tion [13–15]. Efforts were, therefore, initially focussed on ing the rapidly changing clinical condition seen particularly refining the KCC by improving criteria sensitivity enabling in paracetamol and other hyper-acute etiologies and earlier identification of patients with a poor prognosis already utilized in other novel prognostic criteria [23]. through integration of supplemental prognostic markers. Recent studies suggest that in these cases waitlist Initial studies explored the inclusion of arterial blood improvement rather than deterioration may now principally lactate measurements determined using point of care test- confound decision making [9, 11, 12, 15]. Additionally, ing, as hyperlactatemia reflects both decreased hepatic any new or supplemental criteria should ideally retain the clearance by a damaged liver and global illness severity clinical simplicity of existing systems. Two recent attempts and multi-organ failure. Initial reports confirmed their to address these issues are discussed below, both of which prognostic value, and their inclusion was found to improve have utilized the recent availability of web- and app-based specificity, sensitivity and timeliness of the KCC [16]. computing power to deliver outcome predictions through Other studies have also examined arterial blood lactate in sophisticated algorithms. ALF, finding elevated levels to be strongly and indepen- dently associated with death or transplantation in both paracetamol and non-paracetamol-induced disease Dynamic outcome predictive model [17–19]. Analysis of the performance of the specific for paracetamol-induced ALF thresholds for arterial lactate introduced into the KCC for paracetamol-induced disease has been less consistent, with Prediction models in hyper-acute disease would be of most one report suggesting that early levels resulted in an benefit if they could be applied sequentially identifying and increase in sensitivity but at the expense of reduction in quantifying both improvement and deterioration [22, 23]. specificity [17]. However, two other studies found that The latest model from the King’s Liver Intensive Care Unit lactate measurements alone 12 h after admission to trans- is a dynamic outcome prediction model developed and plant centers had high predictive accuracy, with perfor- validated for use in patients with paracetamol-induced ALF mance greater than the KCC alone [18, 19]. Though [24]. It is based on prospective data including analysis of adopted into United Kingdom ALF wait-listing criteria, more than 20 daily variables sequentially assessed for meta-analysis of validation studies to date has failed to 3 days after ICU admission in 912 un-transplanted patients confirm an improvement in diagnostic test performance between 2000 and 2012. The variables included in the final through their inclusion [20]. models to predict death-included age, hepatic Two further meta-analyses of the performance of the encephalopathy, cardiovascular failure, INR, creatinine and KCC have further illustrated their potential shortcomings arterial pH on admission and dynamic variables of (Fig. 1). Review of studies of the KCC in predicting out- changing arterial blood lactate and INR. On validation in come of non-paracetamol-induced ALF comprising 1105 independent datasets from four transplant centers, the cases, showed overall sensitivity of 68% and specificity models showed good discrimination between survivors and 82% [21]. Specificity was highest (93%) in patients with non-survivors, improving with the inclusion of changes in high-grade HE and where the criteria were sequentially INR and Lactate over time (Fig. 2). Innovative in this determined through the clinical course of illness. Impor- approach was its access though a dedicated website and the tantly, this analysis also described a fall-off in diagnostic generation of continuous survival estimates rather than a performance in more recent studies, reflecting the increased binary survival outcome, with the intention that the model success of non-transplant approaches to care. A second, should act as a decision-support tool to support clinical larger meta-analysis examining criteria performance in 123 Hepatology International (2018) 12:204–213 207 Fig. 1 Pooled (a) Sensitivity and (b) Specificity of Kings College Criteria (Source Ref. [20]) 123 208 Hepatology International (2018) 12:204–213 Fig. 2 Discrimination and calibration of dynamic prediction model (AUROC) for day 1 model, b AUROC for day 2 model, c Calibration on day 1 and day 2 of admission in 150 patients with paracetamol- curve for day 1 model, d Calibration curve for day 2 model (Source induced ALF. a Area under receiver operating characteristic curve Ref. [21]) judgement rather than a sole arbiter as to proceeding with transplantation. Model of acute liver failure study group (ALFSG) Elements of this approach were also adopted in the recent analysis by the ALFSG of 1974 subjects enrolled prospectively from 28 academic centers across North America between 1998 and 2013, and managed with and without transplantation [25]. The aim was to devise a mathematical model for all etiologies of ALF to predict transplant-free survival at 21 days, rather than mortality as adopted in other models—making comparison with other Fig. 3 Comparison of the ALFSG model with the Kings College studies adopting more standard approaches more complex. Criteria and MELD score in predicting survival. ALFSG acute liver Clinical features and laboratory values were collected at failure study group, MELD model for end stage liver disease APAP study enrolment and recorded serially up to 7 days. Vari- acetaminophen (paracetamol). AUROC: ALFSG: 0.843, MELD: ables of prognostic value adopted in the predictive model 0.717, KCC APAP 0.560, KCC non-APAP 0.655 (Source Ref. [22]) included admission coma grade, etiology and vasopressor requirement, and admission bilirubin and INR values. with the model correctly predicting outcome of illness in 66.3% of subjects. Its performance was best in patients Arterial blood lactate was not explored. In this analyses of both paracetamol and non-paracetamol cases, sequential with unfavorable etiologies and high-grade encephalopathy and more limited in those with more favorable etiologies values of the INR did not add to prediction over that on admission. Using AUROC analysis, test discrimination and high-grade encephalopathy. appeared superior to the KCC and MELD scores (Fig. 3) 123 Hepatology International (2018) 12:204–213 209 Though sharing similarities in their use of web-based predictive algorithms, key differences exist between these two predictive systems. One is based on the prediction of mortality and the other survival, with resultant differences in test performance. Further, they are targeted at different patient groups—with the USALF criteria taking a broad approach to all etiologies of ALF, whilst the Dynamic Model is focused upon liver injury from paracetamol where the USALF performs least well. It may be that application of these and other models are complimentary and decision making is based upon the results generated from several prognostic systems [23–25]. Other systems evaluated A variety of other scoring systems and supplemental markers have been proposed to identify candidates for liver transplantation, though in general their performance has been evaluated in patient cohorts of limited size (Table 2). Scores used in this way include non-liver specific ones such as SOFA and APACHE II that are widely used to quantify severity of multi-organ failure in other forms of critical illness [17, 26, 27]. Reports suggest similar sensi- tivity and specificity to the King’s criteria in predicting death and they can be applied sequentially. However, though they may identify those patients at risk of death, the multi-organ failure they quantify may not be corrected by transplantation, if for example, it results for other ALF- associated complications such as severe sepsis or pancre- atitis. Consequently, they are not in widespread use to select candidates for transplant. The MELD score is more focused on severity of liver injury and has been assessed more widely in ALF and subjected to meta-analysis [22, 28, 29]. In its unmodified form, it shows most promise in non-paracetamol etiologies, with diagnostic performance close to that of the KCC [22]. It has also been combined with circulating blood levels of cytokeratin K18 (CK18) a cell death-associated marker measured using the M30 assay which principally reflects apoptotic cell death [30]. Also, this measurement in place of bilirubin to give a modified MELD Score showed superior sensitivity and specificity to the standard MELD Score and the King’s criteria in predicting outcome of ALF [31]. Other studies have also reported prognostic models combining standard clinical variables with non-standard analytes. By example, the ALFSG index predictive model combined Coma-grade, INR, bilirubin and phosphorus levels with that of blood levels of M30 CK18 [32]. In a 250-strong validation cohort, discrimination between sur- vivors and non-survivors was greater than with the KCC or MELD, though later external validation in a smaller Table 2 Features of illustrative systems for mortality prediction in acute liver failure Study Publication Criteria Era Etiology n Non- Sensitivity Specificity AUROC Comment year survivors McPhail et al. [20] 2016 KCC 2001–2015 All 2153 Not Given 0.59 0.79 0.76 Meta-analysis (0.56–0.62) (0.77–0.81) Cholongitas et al. 2012 SOFA 1993–2010 Paracetamol 125 58 (46%) 0.67 0.80 0.79 SOFA Score threshold 12 [25] McPhail et al. [20] 2016 MELD 2001–2015 All 2153 Not given 0.74 0.67 0.78 Meta-analysis (0.71–0.77) (0.64–0.69) Bechmann et al. [29] 2010 MELD/CK18 2006–2009 All 68 18 (27%) 0.85 (0.69–1.0) 0.76 (0.6–0.91) 0.94 Peak M65 fragment Koch et al. [22] 2016 ALFSG 1998–2013 All 1974 987 (50%) Not given Not given 0.84 Prediction of survival not death Rutherford et al. [30] 2012 ALFSG/CK18 1998–2011 All 500 251 (50%) 0.86 0.65 0.82 Admission M30 fragment Antoine et al. [32] 2012 Acetylated Not Given Paracetamol 78 27 (35%) Not given Not given 0.87 Admission values HMGB1 KCC Kings College Criteria, SOFA sequential organ failure assessment, MELD model for end-stage liver disease, CK18 cytokeratin 18, ALFGSG Acute liver failure study group, HMGB1 high- mobility group box-1 Non-survivors; cases who died or were transplanted. Figures in parentheses are 95% Confidence Intervals where given 210 Hepatology International (2018) 12:204–213 independent cohort showed no advantage over SOFA or APACHE scoring [32, 33]. However, the primary mechanism of liver cell injury may vary by etiology and necrosis rather than apoptosis may predominate. Antoine and colleagues assessed circu- lating levels of a panel of cell death markers in 78 patients with paracetamol-induced ALF [34]. They found that best prediction of non-survival was with acetylated HMGBI—a biomarker of hepatic necrosis and cellular immunological activation, and to lesser degree with biomarkers of apop- tosis including molecular forms of CK18. To date, vali- dation studies have not indicated prognostic advantage above the standard KCC in predicting survival [35]. Cell death biomarkers do appear to be exquisitely sensitive in predicting clinically significant liver injury very early after paracetamol overdose but If they are to be adopted as adjunctive measures to select liver transplant candidates with ALF, key practical issues will need to be overcome [36, 37]. Techniques for their rapid determination will need to be widely available, and it is likely that a panel of biomarkers will need to be assessed to cover etiology- specific differences in the mechanisms of liver cell injury and death. Other simpler clinical measures have also demonstrated potential for use as prognostic markers: platelet count has been shown to be closely linked to outcome. In a recent study from the USALFSG, the evolution of thrombocy- Fig. 4 a Platelet Counts on days 1–7 after admission in 1598 patients topenia was closely associated with development of multi- with acute liver failure according to outcome at 21 days. SS organ failure and a poor outcome in ALF and linked to the spontaneous survivor, OLT orthotropic liver transplant. Three sym- development of a systemic inflammatory response [38] bols, p \ 0.001; two symbols, p \ 0.01; one symbol, p \ 0.05. *SS (Fig. 4a). Liver volume may be easily determined using vs LT, SS vs death, and LT vs death. SS, spontaneous survival. b Survival According to CT-derived liver volume on admission in 37 analysis of CT images and reflect the balance between the non-transplanted patients with non-acetaminophen etiologies. parallel processes of liver collapse with cell death and 3 3 \ 1000 cm n = 11, C 1000 cm n = 26, p \ 0.001 log-rank increase with regeneration. Recent studies suggest that in (Source Ref. [36, 38]) some non-paracetamol etiologies, loss of liver volume in adults to less than 1000 cm may indicate irreversible absolute arbiter, these models should support decision damage and serve as an early indicator of poor prognosis, making and the multifactorial team assessment. often in advance of the development of encephalopathy (Fig. 4b) [39–41]. Other studies have also sought to identify at a very early Current results of liver transplantation stage those patients with acute liver injury (ALI) in the absence of encephalopathy which will later progress to a Life-saving though it was soon seen to be after its intro- poor outcome of death or transplantation. In a study of US duction, overall results not surprisingly for these very sick patients hospitalized with ALI, progression to a poor out- patients were significantly lower than for elective trans- come was more common in non-paracetamol etiologies and plants carried out for those with chronic diseases. Several with higher levels of INR and bilirubin, and a prolonged series have now demonstrated a progressive incremental duration of illness [42]. improvement in survival over time, with patient survival Finally, it is important to emphasize that prognostic now approaching that of elective surgery. Recent reports models should be only part of the overall functional eval- for the US and UK are of 1-year patient survival rates of uation of the very sick patient with ALF and an experi- 85–90% [43]. These results reflect advances in peri- and enced multi-disciplinary team in an intensive care setting is intra-operative care and the interplay of several factors: the required for correct interpretation. Rather than providing an earlier and more accurate prediction of need for transplant 123 Hepatology International (2018) 12:204–213 211 and identification those patients who are too sick for sur- pre-existing sub-clinical non-cirrhotic liver disease upon gery, and a better understanding of the interplay between liver regeneration is currently poorly characterized, for graft and recipient factors in determining patient survival. example, from chronic alcohol use or non-alcoholic Analysis of the European experience of 4903 transplant steatohepatitis. recipients between 1998 and 2009 found death after sur- In cases where the liver injury has progressed ‘beyond gery to relate to both with independent association shown the point of no return’, novel medical therapies are unlikely with male recipients (adjusted OR 1.25), recipient to be of benefit. It would thus be of no surprise that most [ 50 years of age (1.26), incompatible ABO matching efficacy would be seen either in treatment at an early stage (1.93), donors [ 60 years (1.21), and reduced size graft of disease and/or in etiologies with greatest regenerative (1.54). Recipients [ 50 years, combined with donors capacity. By example, the FULMAR trial of the MARS [ 60 years had a 57% mortality/graft loss within the first extra-corporeal device in patients with ALF failed to show year after transplant [44]. Matching of graft to recipient is a survival improvement in the overall study cohort but key to ensuring optimal outcome [13]. Case series also now there were indications of potential benefit in patients with confirms the practicality and excellent outcomes obtained paracetamol-related disease [49]. for selected patients transplanted using living donors, and Such caveats may also apply to the use of therapeutic from auxiliary liver grafting in etiologies where native liver plasma exchange (TPE). Through mechanisms thought to regenerative potential exists—though globally, these tech- include removal of deleterious inhibitors of hepatic niques are applied to only a small proportion of recipients regeneration and a complex pattern of immunomodulation, [45, 46]. use of this therapy has recently been demonstrated to Patient survival after transplantation for ALF follows a deliver a survival benefit in non-transplanted patients with characteristic pattern of increased mortality early after ALF. In a multi-center randomized controlled trial on transplant, but in patients who survive this phase, subse- mixed etiologies of disease, three 5-l TPE sessions signif- quent survival parallels closely with that seen in elective icantly improved survival above that seen with standard transplantation [44]. In an early series, the high prevalence medical care, although the survival benefit was inferior to of psychiatric comorbidity and addiction issues seen in that seen with transplantation (Fig. 5)[50]. There is recipients of transplantation for paracetamol-induced ALF ongoing debate about the place of this therapy in patients was reflected in an increased risk of death due to suicide, with ALF, given that its use normalizes measures of trauma or non-adherence to immunosuppression with over coagulation disturbance and thus precludes their use in half the deaths occurring within 12 months of transplan- prognostic evaluation for later transplantation. Our practice tation [44]. Later series have not demonstrated such an is in accordance with recent EASL guidance, with its early increase in mortality, but it is clear that paracetamol use in patients with ALF with an expected poor prognosis recipients show increased levels of psychotic comorbidity without transplantation but who have clear medical or which may be reflected in poor compliance with medica- psychiatric contra-indications to surgery [51]. In addition, tion and follow-up [47]. In these patients, assessment of the use of TPE is associated with significant improvements severity of psychiatric illness is a key element of the in cardiovascular status and we also treat patients who are selection of potential recipients and in those transplanted waitlisted for liver transplantation but with worsening follow-up must include close ongoing psychiatric moni- toring and support. New novel non-transplant options A number of novel non-transplant interventions have potential to serve as alternatives to transplantation. A central premise to their use is that the injured liver retains regenerative ability and that this may be augmented by these interventions. A better practical understanding of the processes influencing both liver injury and regeneration is Fig. 5 Survival in 182 patients with ALF according to treatment with likely to be of clinical importance here, exemplified by the liver transplantation and high-volume plasma exchange. Two groups receiving SMT (standard medical treated group) with and without major clinical and laboratory differences between parac- emergency transplantation (HVP ? LTx vs ?HVPLTx) and the two etamol hepatotoxity from single time point or staggered group receiving SMT with and without emergency transplantation overdoses—and the markedly worse outcomes seen in the (HVPLTx vs. ?HVPLTx) (log rank: p = 0.0058) and Cox propor- latter group [48]. The potentially important influence of tional hazard: LTx: p \ 0.0001; HVP: p = 0.0076) (Source Ref. [48]) 123 212 Hepatology International (2018) 12:204–213 3. O’Grady JG, Schalm SW, Williams R. Acute liver failure: cardiovascular failure and vasopressor requirement while redefining the syndromes. Lancet 1993;342(8866):273–275. awaiting a graft [50]. 4. Bernal W, Hyyrylainen A, Gera A, et al. Lessons from look-back Other forms of cellular and immune-modulatory therapy in acute liver failure? A single centre experience of 3300 patients. show promise as future treatments for ALF. Whilst hepato- J Hepatol 2013;59(1):74–80. 5. Slack A, Auzinger G, Willars C, et al. Ammonia clearance with cyte transplantation retains considerable theoretical attrac- haemofiltration in adults with liver disease. Liver Int tions, it presents considerable logistic challenges and a 2014;34(1):42–48. requirement for immunosuppression, and its benefit is yet to 6. Cardoso FS, Gottfried M, Tujios S, Olson JC, Karvellas CJ. be established in adult patients with ALF [52, 53]. Other Group USALFS. Continuous renal replacement therapy is asso- ciated with reduced serum ammonia levels and mortality in acute interventions in patients with acute-on-chronic liver failure liver failure. Hepatology 2017;31:31. (ACLF) show promise but are yet to be tested in ALF. 7. O’Grady J. Timing and benefit of liver transplantation in acute Understanding of the pathophysiologic basis of ACLF is liver failure. J Hepatol 2014;60(3):663–670. rapidly increasing and the insights gained, particularly in 8. Donnelly MC, Davidson JS, Martin K, Baird A, Hayes PC, Simpson KJ. Acute liver failure in Scotland: changes in aetiology relation to modulation of regeneration and immunological and outcomes over time (the Scottish Look-Back Study). Aliment function, may well be transportable to ALF. In a randomized Pharmacol Ther 2017;45(6):833–843. controlled single-blind trial in 110 patients with Hepatitis 9. Reuben A, Tillman H, Fontana R, et al. Outcomes in Adults With B-related ACLF of whom less than half were thought to be Acute Liver Failure Between 1998 and 2013: an observational cohort study. Ann Intern Med 2016;164(11):724–732. cirrhotic, treatment with allogenic mesenchymal stem cells 10. Oketani M, Ido A, Nakayama N, et al. Etiology and prognosis of improved survival—[54]. Similarly, in a single-center dou- fulminant hepatitis and late-onset hepatic failure in Japan: sum- ble-blind trial in 55 patients with decompensated cirrhosis, mary of the annual nationwide survey between 2004 and 2009. treatment with granulocyte colony stimulating factor and Hepatol Res 2013;43(2):97–105. 11. Ichai P, Legeai C, Francoz C, et al. Patients with acute liver erythropoietin was associated with improved survival in failure listed for superurgent liver transplantation in France: association with indications of improved hepatic regenera- reevaluation of the Clichy-Villejuif criteria. Liver Transpl tion through mobilization of progenitor cells [55]. These 2015;21(4):512–523. results are yet to be replicated or tested in patients with ALF 12. Simpson KJ, Bates CM, Henderson NC, et al. The utilization of liver transplantation in the management of acute liver failure: but demonstrate mechanistic approaches likely to be asses- comparison between acetaminophen and non-acetaminophen sed in the near future. As ever, the rarity and severity of ALF etiologies. Liver Transpl 2009;15(6):600–609. represent major limiting factors for performing randomized 13. Bernal W, Cross TJS, Auzinger G, et al. Outcome after wait- controlled trials of these or other agents and true advances in listing for emergency liver transplantation in acute liver failure: a single centre experience. J Hepatol 2009;50(2):306–313. therapeutic intervention are likely to result only from inter- 14. Brandsaeter B, Hockerstedt K, Friman S, et al. Fulminant hepatic national collaborative studies. failure: outcome after listing for highly urgent liver transplanta- tion-12 years experience in the nordic countries. Liver Transpl 2002;8(11):1055–1062. 15. Reddy KR, Ellerbe C, Schilsky M, et al. Determinants of out- Compliance with ethical standards come among patients with acute liver failure listed for liver transplantation in the United States. Liver Transpl 2016;22(4):505–515. Conflict of interest Professor Roger Williams and Professor William 16. Bernal W, Donaldson N, Wyncoll D, Wendon J. Blood lactate as Bernal have no potential conflict of interest. an early predictor of outcome in paracetamol-induced acute liver failure: a cohort study. Lancet 2002;359(9306):558–563. Open Access This article is distributed under the terms of the Creative 17. Schmidt LE, Larsen FS. Prognostic implications of hyperlac- Commons Attribution 4.0 International License (http://creative tatemia, multiple organ failure, and systemic inflammatory commons.org/licenses/by/4.0/), which permits unrestricted use, dis- response syndrome in patients with acetaminophen-induced acute tribution, and reproduction in any medium, provided you give liver failure. Crit Care Med 2006;34(2):337–343. appropriate credit to the original author(s) and the source, provide a 18. Macquillan GC, Seyam MS, Nightingale P, Neuberger JM, link to the Creative Commons license, and indicate if changes were Murphy N. Blood lactate but not serum phosphate levels can made. predict patient outcome in fulminant hepatic failure. Liver Transpl 2005;11(9):1073–1079. 19. Cholongitas E, O’Beirne J, Betrossian A, et al. Prognostic impact References of lactate in acute liver failure. Liver Transpl 2008;14(1):121–122 (author reply 3). 1. Gazzard BG, Widdop B, Davis M, Hughes RD, Goulding R, 20. Craig DGN, Ford AC, Hayes PC, Simpson KJ. Systematic Williams R. Early prediction of the outcome of a paracetamol review: prognostic tests of paracetamol-induced acute liver fail- overdose based on an analysis of 163 patients. Postgrad Med J ure. Aliment Pharmacol Ther 2010;31(10):1064–1076. 1977;53(619):243–247. 21. McPhail MJW, Wendon JA, Bernal W. Meta-analysis of per- 2. O’Grady JG, Alexander GJ, Hayllar KM, Williams R. Early formance of Kings’s College Hospital Criteria in prediction of indicators of prognosis in fulminant hepatic failure. Gastroen- outcome in non-paracetamol-induced acute liver failure. J Hepa- terology 1989;97(2):439–445. tol 2010;53(3):492–499. 123 Hepatology International (2018) 12:204–213 213 22. McPhail MJ, Farne H, Senvar N, Wendon JA, Bernal W. Ability 40. Zabron A, Quaglia A, Peddu P, et al. Clinical and prognostic of King’s College Criteria and model for end-stage liver disease associations of liver volume determined by computed tomogra- scores to predict mortality of patients with acute liver failure: a phy in acute liver failure. Liver Int 2018 (In Press). meta-analysis. Clin Gastroenterol Hepatol 41. Yamagishi Y, Saito H, Tada S, et al. Value of computed 2016;14(4):516–525.e5 (quiz e43–e45). tomography-derived estimated liver volume/standard liver vol- 23. Kumar R, Shalimar Sharma H, et al. Prospective derivation and ume ratio for predicting the prognosis of adult fulminant hepatic validation of early dynamic model for predicting outcome in failure in Japan. J Gastroenterol Hepatol 2005;20(12):1843–1849. patients with acute liver failure. Gut 2012;61(7):1068–1075. 42. Koch DG, Speiser JL, Durkalski V, et al. The natural history of 24. Bernal W, Wang Y, Maggs J, et al. Development and validation severe acute liver injury. Am J Gastroenterol of a dynamic outcome prediction model for paracetamol-induced 2017;112(9):1389–1396. acute liver failure: a cohort study. Lancet Gastroenterol Hepatol 43. Blood and Transplant N. ANNUAL REPORT ON LIVER 2016;1(3):217–225. TRANSPLANTATION. 2017. https://nhsbtdbe.blob.core.win 25. Koch DG, Tillman H, Durkalski V, Lee WM, Reuben A. dows.net/umbraco-assets-corp/5007/annual_liver_transplantation_ Development of a model to predict transplant-free survival of report_2017.pdf Accessed 19 Feb 2018. patients with acute liver failure. Clin Gastroenterol Hepatol 44. Germani G, Theocharidou E, Adam R, et al. Liver transplantation 2016;14(8):1199–1206.e2. for acute liver failure in Europe: outcomes over 20 years from the 26. Craig DG, Zafar S, Reid TW, et al. The sequential organ failure ELTR database. J Hepatol 2012;57(2):288–296. assessment (SOFA) score is an effective triage marker following 45. Yamashiki N, Sugawara Y, Tamura S, et al. Outcomes after staggered paracetamol (acetaminophen) overdose. Aliment living donor liver transplantation for acute liver failure in Japan: Pharmacol Ther 2012;35(12):1408–1415. results of a nationwide survey. Liver Transpl 27. Cholongitas E, Theocharidou E, Vasianopoulou P, et al. Com- 2012;18(9):1069–1077. parison of the sequential organ failure assessment score with the 46. Goldaracena N, Spetzler VN, Marquez M, et al. Live donor liver King’s College Hospital criteria and the model for end-stage liver transplantation: a valid alternative for critically ill patients suf- disease score for the prognosis of acetaminophen-induced acute fering from acute liver failure. Am J Transplant liver failure. Liver Transpl 2012;18(4):405–412. 2015;15(6):1591–1597. 28. Schmidt L, Larsen F. MELD score as a predictor of liver failure 47. Karvellas CJ, Safinia N, Auzinger G, et al. Medical and psychi- and death in patients with acetaminophen-induced liver injury. atric outcomes for patients transplanted for acetaminophen-in- Hepatology 2007;45:789–796. duced acute liver failure: a case-control study. Liver Int 29. Craig DGN, Reid TWDJ, Wright EC, et al. The sequential organ 2010;30(6):826–833. failure assessment (SOFA) score is prognostically superior to the 48. Craig DGN, Bates CM, Davidson JS, Martin KG, Hayes PC, model for end-stage liver disease (MELD) and MELD variants Simpson KJ. Staggered overdose pattern and delay to hospital following paracetamol (acetaminophen) overdose. Aliment presentation are associated with adverse outcomes following Pharmacol Ther 2012;35(6):705–713. paracetamol-induced hepatotoxicity. Br J Clin Pharmacol 30. Eguchi A, Wree A, Feldstein AE. Biomarkers of liver cell death. 2012;73(2):285–294. J Hepatol 2014;60(5):1063–1074. 49. Saliba F, Camus C, Durand F, et al. Albumin dialysis with a 31. Bechmann LP, Jochum C, Kocabayoglu P, et al. Cytokeratin noncell artificial liver support device in patients with acute liver 18-based modification of the MELD score improves prediction of failure: a randomized, controlled trial. Ann Intern Med spontaneous survival after acute liver injury. J Hepatol 2013;159(8):522–531. 2010;53(4):639–647. 50. Larsen FS, Schmidt LE, Bernsmeier C, et al. High-volume 32. Rutherford A, King LY, Hynan LS, et al. Development of an plasma exchange in patients with acute liver failure: an open accurate index for predicting outcomes of patients with acute randomised controlled trial. J Hepatol 2016;64(1):69–78. liver failure. Gastroenterology 2012;143(5):1237–1243. 51. European Association for the Study of the Liver. Electronic 33. Craig DG, Simpson KJ. Accuracy of the ALFSG index as a triage address eee, Clinical practice guidelines p, Wendon J, et al. marker in acute liver failure. Gastroenterology 2013;144(1):e25. EASL Clinical Practical Guidelines on the management of acute 34. Antoine DJ, Jenkins RE, Dear JW, et al. Molecular forms of (fulminant) liver failure. J Hepatol 2017;66(5):1047–1081. HMGB1 and keratin-18 as mechanistic biomarkers for mode of 52. Iansante V, Mitry RR, Filippi C, Fitzpatrick E, Dhawan A. cell death and prognosis during clinical acetaminophen hepato- Human hepatocyte transplantation for liver disease: current status toxicity. J Hepatol 2012;56(5):1070–1079. and future perspectives. Pediatr Res 2017;06:06. 35. Craig DGN, Lee P, Pryde EA, Masterton GS, Hayes PC, Simpson 53. Ibars EP, Cortes M, Tolosa L, et al. Hepatocyte transplantation KJ. Circulating apoptotic and necrotic cell death markers in program: lessons learned and future strategies. World J Gas- patients with acute liver injury. Liver Int 2011;31(8):1127–1136. troenterol 2016;22(2):874–886. 36. Dear JW, Clarke JI, Francis B, et al. Risk stratification after 54. Lin BL, Chen JF, Qiu WH, et al. Allogeneic bone marrow- paracetamol overdose using mechanistic biomarkers: results from derived mesenchymal stromal cells for hepatitis B virus-related two prospective cohort studies. Lancet Gastroenterol Hepatol acute-on-chronic liver failure: a randomized controlled trial. 2018;3(2):104–113. Hepatology 2017;66(1):209–219. 37. Antoine DJ, Dear JW, Lewis PS, et al. Mechanistic biomarkers 55. Kedarisetty CK, Anand L, Bhardwaj A, et al. Combination of provide early and sensitive detection of acetaminophen-induced granulocyte colony-stimulating factor and erythropoietin improves outcomes of patients with decompensated cirrhosis. acute liver injury at first presentation to hospital. Hepatology 2013;58(2):777–787. Gastroenterology 2015;148(7):1362–1370 (e7). 38. Stravitz RT, Ellerbe C, Durkalski V, et al. Thrombocytopenia Is 56. Bernuau J, Rueff B, Benhamou JP. Fulminant and subfulminant liver associated with multi-organ system failure in patients with acute failure: definitions and causes. Semin Liver Dis 1986;6(2):97–106. liver failure. Clin Gastroenterol Hepatol 2016;14(4):613–620. 57. Mochida S, Nakayama N, Matsui A, Nagoshi S, Fujiwara K. Re- 39. Shakil AO, Jones BC, Lee RG, Federle MP, Fung JJ, Rakela J. evaluation of the Guideline published by the Acute Liver Failure Prognostic value of abdominal CT scanning and hepatic Study Group of Japan in 1996 to determine the indications of histopathology in patients with acute liver failure. Dig Dis Sci liver transplantation in patients with fulminant hepatitis. Hepatol 2000;45(2):334–339. Res 2008;38(10):970–979.

Journal

Hepatology InternationalSpringer Journals

Published: Jun 1, 2018

References

You’re reading a free preview. Subscribe to read the entire article.


DeepDyve is your
personal research library

It’s your single place to instantly
discover and read the research
that matters to you.

Enjoy affordable access to
over 18 million articles from more than
15,000 peer-reviewed journals.

All for just $49/month

Explore the DeepDyve Library

Search

Query the DeepDyve database, plus search all of PubMed and Google Scholar seamlessly

Organize

Save any article or search result from DeepDyve, PubMed, and Google Scholar... all in one place.

Access

Get unlimited, online access to over 18 million full-text articles from more than 15,000 scientific journals.

Your journals are on DeepDyve

Read from thousands of the leading scholarly journals from SpringerNature, Elsevier, Wiley-Blackwell, Oxford University Press and more.

All the latest content is available, no embargo periods.

See the journals in your area

DeepDyve

Freelancer

DeepDyve

Pro

Price

FREE

$49/month
$360/year

Save searches from
Google Scholar,
PubMed

Create lists to
organize your research

Export lists, citations

Read DeepDyve articles

Abstract access only

Unlimited access to over
18 million full-text articles

Print

20 pages / month

PDF Discount

20% off