Patients With Acute Liver Failure Listed for Superurgent Liver Transplantation in France: Reevaluation of the Clichy-Villejuif Criteria

Transcription

1 LIVER TRANSPLANTATION 21: , 2015 ORIGINAL ARTICLE Patients With Acute Liver Failure Listed for Superurgent Liver Transplantation in France: Reevaluation of the Clichy-Villejuif Criteria Philippe Ichai, 1,2,3 Camille Legeai, 4 Claire Francoz, 5 Karim Boudjema, 6 Olivier Boillot, 7 Christian Ducerf, 8 Philippe Mathurin, 9 François-Rene Pruvot, 9 Bertrand Suc, 10 Philippe Wolf, 11 Olivier Soubrane, 12 Yves Patrice Le Treut, 13 Daniel Cherqui, 14 Laurent Hannoun, 15 Georges-Philippe Pageaux, 16 Jean Gugenheim, 17 Christian Letoublon, 18 Jean Saric, 19 Vincent Di Martino, 20 Armand Abergel, 21 Laurence Chiche, 22 Teresa Maria Antonini, 1,2,3 Christian Jacquelinet, 4 Denis Castaing, 1,2,3 Didier Samuel, 1,2,3 and the French Liver Transplant Teams 1 Liver Intensive Care Unit, Centre Hepato-Biliaire, H ^opital Paul Brousse, AP-HP, Villejuif, France; 2 Unite 785, Inserm, Villejuif, France; 3 Unit es Mixtes de Recherche en Sant e 785, Universit e Paris-Sud, Villejuif, France; 4 Agence de la Biom edecine, La Plaine Saint-Denis, France; 5 H ^opital Beaujon, AP-HP, Clichy, France; 6 H ^opital Pontchaillou, Centre Hospitalier Universitaire de Rennes, Rennes, France; 7 H ^opital Edouard Herriot; 8 H ^opital de la Croix-Rousse, Hospices Civils de Lyon, Lyon, France; 9 H ^opital Claude Huriez, Centre Hospitalier R egional Universitaire de Lille, Lille, France; 10 H ^opital de Rangueil, Centre Hospitalier Universitaire de Toulouse, Toulouse, France; 11 H ^opital de Hautepierre, Centre Hospitalier Universitaire de Strasbourg, Strasbourg, France; 12 H ^opital Saint-Antoine, AP-HP, Paris, France; 13 H ^opital de la Conception, AP-HM, Marseille, France; 14 H ^opital Henri Mondor, AP-HP, Cr eteil, France; 15 Groupe Hospitalier Piti e-salp etrière, AP- HP, Paris, France; 16 H ^opital Saint Eloi, Centre Hospitalier Universitaire de Montpellier, Montpellier, France; 17 H ^opital de l Archet, Centre Hospitalier Universitaire de Nice, Nice, France; 18 H ^opital Michallon, Centre Hospitalier Universitaire de Grenoble, Grenoble, France; 19 H ^opital Pellegrin, Centre Hospitalier Universitaire de Bordeaux, Bordeaux, France; 20 H ^opital Jean Minjoz, Centre Hospitalier Universitaire de Besanc on, Besanc on, France; 21 H ^opital de l Estaing, Centre Hospitalier Universitaire de Clermont-Ferrand, Clermont- Ferrand, France; and 22 H ^opital de la C ^ote de Nacre, Centre Hospitalier Universitaire de Caen, Caen, France In France, decisions regarding superurgent (SU) liver transplantation (LT) for patients with acute liver failure (ALF) are principally based on the Clichy-Villejuif (CV) criteria. The aims of the present study were to study the outcomes of patients registered for SU LT and the factors that were predictive of spontaneous improvement and to determine the usefulness of the CV criteria. All patients listed in France for SU LT between 1997 and 2010 who were 15 years old or older with ALF were included. In all, 808 patients were listed for SU transplantation: 22% with paracetamol-induced ALF and 78% with non Additional Supporting Information may be found in the online version of this article. Abbreviations: ABM, Agence de la Biomedecine; ALF, acute liver failure; AUC, area under the curve; CI, confidence interval; CV, Clichy-Villejuif; FH, fulminant hepatitis; HBV, hepatitis B virus; IQR, interquartile range; ITT, intention to treat; KCH, King s College Hospital; LT, liver transplantation; NPV, negative predictive value; OR, odds ratio; PPV, positive predictive value; PT, prothrombin time; SU, superurgent. Financial support: Nothing to report. Daniel Cherqui is currently affiliated with Centre Hepato-Biliaire, H^opital Paul Brousse, AP-HP, Villejuif, France. For the members of the French Liver Transplant Teams, see the supporting information. Address reprint requests to Philippe Ichai, Liver Intensive Care Unit, Centre Hepato-Biliaire, H^opital Paul Brousse, AP-HP, 14 Avenue Paul Vaillant Couturier, Villejuif Cedex, France. philippe.ichai@pbr.aphp.fr DOI /lt View this article online at wileyonlinelibrary.com. LIVER TRANSPLANTATION.DOI /lt. Published on behalf of the American Association for the Study of Liver Diseases VC 2015 American Association for the Study of Liver Diseases.

2 LIVER TRANSPLANTATION, Vol. 21, No. 4, 2015 ICHAI ET AL. 513 paracetamol-induced ALF. Of these 808 patients, 112 improved spontaneously, 587 underwent LT, and 109 died or left the waiting list because of a worsening condition. The 1-year survival rate according to an intention-to-treat analysis and the survival after LT were 66.3% [interquartile range (IQR), 62.7%-69.7%] and 74.2% (IQR, 70.5%-77.6%), respectively. The factors that were predictive of a spontaneous recovery with ALF-related paracetamol hepatotoxicity were as follows: hepatic encephalopathy grade 0, 1, or 2 [odds ratio (OR), 4.8; 95% confidence interval (CI), ]; creatinine clearance 60 ml/minute/1.73 m 2 (OR, 4.77; 95% CI, ), a bilirubin level < 200 mmol/l (OR, 21.64; 95% CI, ); and a factor V level > 20% (OR, 5.79; 95% CI, ). For ALF-related nonparacetamol hepatotoxicity, the factor that was predictive of a spontaneous recovery was a bilirubin level < 200 mmol/l (OR, 10.38; 95% CI, ). The sensitivity, specificity, and positive and negative predictive values for the CV criteria were 75%, 56%, 50%, and 79%, respectively, for ALF due to paracetamol and 69%, 50%, 64%, and 55%, respectively, for ALF not related to paracetamol. The performance of current criteria for SU transplantation could be improved if paracetamol-induced ALF and non paracetamol-induced ALF were split and 2 other items were included in this model: the bilirubin level and creatinine clearance. Liver Transpl 21: , VC 2015 AASLD. Received August 4, 2014; accepted December 14, Emergency liver transplantation (LT) has transformed the prognosis of acute liver failure (ALF). The 1-year survival rate after LT now reaches 77%. Assessing the prognosis of fulminant hepatitis (FH) is vital when one is making decisions about the need for and timing of LT. The King s College Hospital (KCH) and Clichy- Villejuif (CV) criteria remain the tools most widely used worldwide to determine a prognosis for ALF or FH. 1 The KCH criteria differ in patients with paracetamol- and non paracetamol-related disease. They include clinical and biological findings [age, time elapsing between the onset of jaundice and encephalopathy, bilirubin, prothrombin time (PT), arterial ph, and serum creatinine] and display good sensitivity and specificity. The inclusion in this score of other parameters such as arterial lactate has increased its sensitivity and specificity. 2 The CV criteria are mainly used in France, and they include grade 3 or 4 hepatic encephalopathy and factor V levels < 20% in patients < 30 years of age and < 30% in patients 30 years of age. 3 They were compiled on the basis of a cohort of patients with FH of a viral origin [mainly due to hepatitis B virus (HBV)]. The purpose of these scores is to enable the accurate identification, early in the course of the illness, not only of patients who will die without transplantation but also of those patients who will survive with medical treatment. Until now, the outcomes of French patients listed on the basis of these criteria had not been reported, nor had the prognostic predictions of the CV criteria been evaluated. The objectives of the present study were, therefore, to focus on a larger population in France to study the outcomes of patients registered for superurgent (SU) LT and the predictive factors for spontaneous improvement and to determine the usefulness of the CV criteria. PATIENTS AND METHODS Study Design All patients waiting for LT are registered in the national information system run by the French Biomedicine Agency [Agence de la Biomedecine (ABM)]. Demographic, clinical, and biological data are collected at the time of listing, at the time of transplantation, and annually thereafter. Withdrawals from the waiting list and the deaths of patients are updated prospectively. Between 1997 and 2010, 1700 requests for SU LT from 19 centers were thus registered, and 1649 were for patients in France (Fig. 1). Only patients who were 15 years old or older and were presenting with FH were included in our study (n ). The reasons for requesting SU registration were as follows: ALF (n 5 808; 58.6%), retransplantation (n 5 477; 34.6%) due to primary or secondary nonfunction of the graft, and rapid worsening of the initial disease (n 5 95; 6.9%). All requests had been validated by 1 or 2 experts. The CV criteria were applied in most cases, and they were indicators of the presence of confusion or coma (ie, encephalopathy grades 3 and 4) associated with a factor V level lower than 20% in patients who were 30 years old or younger or with a factor V level lower than 30% in patients who were more than 30 years old. 1,4-6 However, a minority were determined with the KCH criteria. When a graft was proposed, the clinical situation was reassessed, and the indication for LT was validated. For the study, the principal characteristics of patients were obtained from the electronic medical records held by ABM and the transplantation centers. Supplementary clinical and biological data gathered at registration were retrieved from the medical records held by the transplantation centers. Statistical Analysis For the overall analysis, patients who were withdrawn from the waiting list because of a worsening in their condition were considered to have died. Continuous variables were reported as medians and 25th to 75th percentiles [interquartile range (IQR)] and were compared between the groups with nonparametric Wilcoxon or Kruskal-Wallis tests. Categorical variables were reported as percentages and were compared with the v 2 test or Fisher s exact test. Characteristics associated with the risk of spontaneous recovery were identified with univariate and multivariate logistic

3 514 ICHAI ET AL. LIVER TRANSPLANTATION, April 2015 Figure 1. Outcomes of patients registered for SU transplants in France. regression models. All variables associated with the probability of spontaneous recovery with a P value lower than 20% under the univariate analysis were included in the multivariate model. A backward stepwise selection process of variables with P < 0.10 was employed to determine the final multivariate logistic model. In a sensitivity analysis, transplant patients were excluded in order to assess the probability of spontaneous recovery directly in relation to the risk of death. Missing data were not imputed, and patients for whom at least 1 datum was missing were excluded from the multivariate analysis. In a sensitivity analysis, missing data for a given categorical variable were replaced by the value of the class at lower risk, and the multivariate analysis was repeated with these imputed variables on the whole sample. Similar analyses were performed by the replacement of the missing data with the value of the class at higher risk. Poor prognostic prediction of the CV criteria was assessed by the calculation of the sensitivity, specificity, positive predictive value (PPV), and negative predictive value (NPV) for all patients and separately for patients with paracetamol-induced ALF and patients with non paracetamol-induced ALF. Sensitivity and specificity were compared between ALF etiology groups with the v 2 test or Fisher s exact test. The threshold for each continuous variable was defined after the consideration of all of the following points: clinical relevance, the increase in the risk of spontaneous recovery according to the value of the continuous variable, and the optimum threshold obtained from the receiver operating characteristic curve illustrating the performance of each value for a variable in discriminating between patients who will recover spontaneously and those who will receive a graft or die on the transplant waiting list. The prognostic predictive accuracy of CV criteria was also assessed after the exclusion of transplant patients. The Kaplan-Meier method and the log-rank test were used to estimate and compare the posttransplantation survival rates of patients. The median waiting time in days for LT was also assessed with the Kaplan-Meier method. The median period in hours elapsing between SU registration and the proposal of a graft was measured only for patients to whom a graft had been proposed; a Kaplan-Meier median time estimation was not performed because of a lack of precision in the measurement of the time occurrence of some events. All analyses were 2-sided, and P < 0.05 indicated statistical significance. Statistical

4 LIVER TRANSPLANTATION, Vol. 21, No. 4, 2015 ICHAI ET AL. 515 analyses were performed with SAS 9.3 (SAS Institute, Cary, NC). RESULTS Patients Listed for Emergency LT (n 5 808) Clinical and Demographic Characteristics of SU Patients In all, we examined 808 patients who were listed for SU LT in France between 1997 and The principal characteristics of ALF patients at the time of listing are shown in Table 1. The median age of the patients was 40.3 years. More than a third of the patients (35.5%) were 30 to 45 years old. The principal known causes of FH were as follows: paracetamol (n 5 179; 22%), viruses (n 5 134; 16.6%), nonparacetamol drug induced (n 5 76; 9.4%), toxins (n 5 38; 5%), and other causes (n 5 157; 19.5%). The most common virus was hepatitis B (acute hepatitis B, 78; reactivation of chronic hepatitis B, 24), which was followed by hepatitis A (n 5 16). Other causes were mainly autoimmune hepatitis (n 5 69), Wilson s disease (n 5 29), acute Budd-Chiari syndrome (n 5 23), and hypoxic hepatitis (n 5 11). In 222 patients (27.5%), no cause could be determined. Out of the 171 patients (25.6%) with grade 0, 1, or 2 encephalopathy, the breakdown of causes of hepatitis was as follows: viruses (15.7%), paracetamol (26.3%), nonparacetamol drugs (9.4%), toxins (mainly mushroom poisoning, 20/ 21; 12.3%), miscellaneous causes (16.4%), and indeterminate etiologies (19.9%). FH due to paracetamol poisoning was more common after 2002 than before that year (27.8% versus 11.1%, P < 0.001). Outcome of Listed Patients (Fig. 1, Table 1) Out of the total of 808 patients, 587 (72.6%) underwent LT, 112 (13.9%) improved spontaneously, and 109 (13.5%) died on the waiting list. Mortality on the Waiting List (n 5 109) Patients who died on the waiting list more frequently suffered from grade 3 or 4 encephalopathy (P ), had a higher bilirubin level (P < 0.001), had a lower factor V level (P ), and had poorer creatinine clearance (P < 0.001) at listing than transplant patients. Cases of paracetamol-related ALF were more frequent among patients who died without LT versus those who underwent LT (26.9% versus 16.9%, P ). Thirteen patients with no encephalopathy or encephalopathy grade 1 or 2 died on the waiting list. Among the latter, FH was due to HBV (n 5 3), paracetamol (n 5 5), other medications (n 5 2), Budd- Chiari syndrome (n 5 1), autoimmune hepatitis (n 5 1), or an undetermined cause (n 5 1). Patients Alive Without LT (n 5 112) Fifty-seven patients (59.5%) presented with grade 3 or 4 hepatic encephalopathy. The principal causes of hepatitis in these 57 patients were paracetamol (n 5 24; 42.1%), nonparacetamol drugs or toxins (n 5 7; 12.3%), viruses (n 5 8; 14.0%), miscellaneous causes (n 5 7; 12.3%), and undetermined hepatitis (n 5 11; 19.3%). The number of patients withdrawn from the list because of a spontaneous recovery was higher among those with a paracetamol overdose versus those with non paracetamol-related ALF (28.5% versus 9.7%, P < 0.001). PT, factor V, and creatinine clearance values did not differ significantly among patients who lived without LT and those who died or underwent transplantation, whereas the bilirubin level was lower (76 versus 265 mmol/l, P < 0.001), and grade 3 to 4 hepatic encephalopathy was less frequent (59.4% of grade 3 to 4 hepatic encephalopathy versus 77.0%, P < 0.001). Transplant Patients (n 5 587; Table 1) The causes of hepatitis in these 587 patients were as follows: paracetamol (16.9%), HBV (14.2%), other virus (3.1%), toxin (14.7%), undetermined causes (30%), autoimmune disease (9.9%), and other causes (11.3%). One hundred nineteen patients (24.7%) with lower than grade 3 hepatic encephalopathy were considered for transplantation. Among these, LT rates ranged from 81.1% (toxic hepatitis) to 85.3% (indeterminate hepatitis); paracetamol was excluded as a cause because this accounted for only 33.3% of cases. Two patients who were initially registered as SU underwent transplantation otherwise: one electively and the other as a regional emergency (the former on day 30 and the latter 12 days after SU registration). Of the 347 patients in blood groups A, B, and AB, 172 (49.6%) received a group O graft; for 107 of them (62.2%), the recipient belonged to group A. Nine patients (1.5%) received an incompatible graft. The median waiting time between emergency listing and the offer of a graft that was subsequently accepted was 7.5 hours (IQR, hours). The median waiting time for a group A recipient (4.8 hours; IQR, hours) was significantly shorter than that for a group B (11.9 hours; IQR, hours) or O recipient (12.0 hours; IQR, hours) (P < 0.01). The median waiting times for blood groups A and B were similar (P ). Waiting times thus varied as a function of the blood group (P < 0.001). Offers of a subsequently accepted group A graft for a group A recipient or a subsequently accepted group O graft for a group O recipient occurred 4.5 hours (IQR, hours) and 11.2 hours (IQR, hours), respectively, after SU registration. Predictive Factors for a Spontaneous Improvement (Tables 2 and 3) For patients with ALF related to paracetamol hepatotoxicity, multivariate analysis showed that hepatic encephalopathy grade 0, 1, or 2 [odds ratio (OR), 4.8; 95% confidence interval (CI), ], creatinine

5 516 ICHAI ET AL. LIVER TRANSPLANTATION, April 2015 TABLE 1. Principal Characteristics of Patients at Listing for SU Transplantation Patients Who Improved Spontaneously All (n 5 808) (n 5 112) Patients Who Received Grafts (n 5 587) Patients Who Died or Left List for Worsening Condition (n 5 109) P Value* P Value P Value Sex Female 470 (58.2) 69 (61.6) 353 (60.1) 48 (44) 0.77 <10 22 < Male 338 (41.8) 43 (38.4) 234 (39.9) 61 (56) Age, years 40.3 ( ) 36.1 ( ) 40.6 ( ) 45.3 ( ) <10 22 < Etiology of FH, n Paracetamol 179 (22.2) 51 (45.5) 99 (16.9) 29 (26.9) < HBV 102 (12.7) 5 (4.5) 83 (14.2) 14 (13) Other virus 32 (4.0) 6 (5.4) 18 (3.1) 8 (7.4) Toxin and non paracetamol drug 114 (14.1) 13 (11.6) 86 (14.7) 15 (13.9) Undetermined 222 (27.5) 25 (22.3) 176 (30) 21 (19.4) Autoimmune disease 69 (8.6) 2 (1.8) 58 (9.9) 9 (8.3) Other 88 (10.9) 10 (8.9) 66 (11.3) 12 (11.1) Hepatic encephalopathy, n Grade 0 31 (4.6) 9 (9.4) 17 (3.5) 5 (5.4) <10 22 < Grade 1 or (20.9) 30 (31.2) 102 (21.2) 8 (8.7) Grade 3 or (74.4) 57 (59.4) 362 (75.3) 79 (85.9) Creatinine clearance, ml/minute/ 51.2 ( ) 62.0 ( ) 55.1 ( ) 26.5 ( ) 0.27 <10 23 < m 2,n5 732 Bilirubin, mmol/l, n ( ) 76 (42-122) 285 (97-468) ( ) <10 23 <10 23 <10 23 Factor V (%), n (10-25) 16 (11-22) 18 (10-26) 13 (5-22) PT (%), n (9-20) 13 (10-19) 13 (9-20) 13 (5-20) Blood group of recipient A 319 (39.5) 38 (33.9) 248 (42.2) 33 (30.3) AB 43 (5.3) 4 (3.6) 34 (5.8) 5 (4.6) B 91 (11.3) 13 (11.6) 65 (11.1) 13 (11.9) O 355 (43.9) 57 (50.9) 240 (40.9) 58 (53.2) NOTE: The data are presented as medians (IQR) or populations (%). P values were found with a v 2 test or Fisher s test for qualitative variables and with the Kruskal- Wallis test for quantitative variables. *Patients who improved spontaneously versus patients who underwent transplantation. Patients who improved spontaneously versus patients who died or withdrew from the list because of worsening. Patients who received a graft versus patients who died or withdrew from the list because of worsening.

6 LIVER TRANSPLANTATION, Vol. 21, No. 4, 2015 ICHAI ET AL. 517 TABLE 2. Predictive Factors for Spontaneous Recovery in Patients With Paracetamol ALF (n 5 179) Univariate Analysis (n 5 179) Multivariate Analysis (n 5 152), AUC Number of Improvements n on Waiting List OR 95% CI P OR 95% CI P Sex Female Male Age at request for SU registration <50 years years Hepatic encephalopathy Grade 0, 1, or < < Grade 3 or Creatinine clearance <60 ml/minute/1.73 m < < ml/minute/1.73 m Bilirubin <200 mmol/l mmol/l Factor V 20% < >20% PT 20% >20% NOTE: Odds ratios, 95% CI, and P values were estimated by logistic regression models. clearance 60 ml/minute/1.73 m 2 (OR, 4.77; 95% CI, ), a bilirubin level < 200 mmol/l (OR, 21.64; 95% CI, ), and a factor V level > 20% (OR, 5.79; 95% CI, ) were predictive of a spontaneous recovery rather than death or transplantation. For patients with ALF unrelated to paracetamol hepatotoxicity, multivariate analysis showed that a bilirubin level < 200 mmol/l (OR, 10.38; 95% CI, ) was predictive of a spontaneous recovery. An age < 50 years (OR, 2.14; 95% CI, ) tended to be associated with a spontaneous recovery. The predictive ability [area under the curve (AUC)] of the model was 0.78 for paracetamol ALF and 0.79 for nonparacetamol ALF. Outcome of SU Patients According to CV Criteria (Table 4) Data enabling the classification of patients according to the CV criteria (encephalopathy grade, factor V level, and age) were available for 622 patients. Patients Meeting CV Criteria at Listing (n 5 364; 58.5%) These patients included 223 women (61.3%) and 141 men (38.7%) with a median age of 41.9 years. At the time of listing, PT, factor V, bilirubin, and creatinine clearance values were 11%, 14%, 139 mmol/l, and 43.6 ml/mn/1.73 m 2, respectively. All patients were encephalopathic. Of these 364 patients, 42 (11.5%) recovered spontaneously, 59 (16%) died, and 263 (72.3%) underwent LT. Patients with paracetamolrelated ALF improved spontaneously in 18% of cases, whereas patients with hepatitis B had only 5.6% of spontaneous recoveries. Creatinine clearance in patients who recovered spontaneously was significantly higher than that in patients who died (39.6 versus 24.5 ml/minute/1.73 m 2, P ). A multivariate analysis of factors for a poor prognosis in ALF patients meeting the CV criteria showed that bilirubin > 200 mmol/l was a factor independent of death (OR, 5.68; 95% CI, ). After the exclusion of transplant patients, creatinine clearance and bilirubin level were both independent predictive factors for spontaneous recovery in a multivariate analysis (data not shown). Patients Who Did Not Meet the CV Criteria at Listing (n 5 258; 41.5%) This category included 153 women (59.3%) and 105 men (40.7%) with a median age of 39.4 years (IQR, years). PT, factor V, bilirubin, and creatinine clearance values were significantly higher than those in patients meeting the CV criteria (17%, 22%,

7 518 ICHAI ET AL. LIVER TRANSPLANTATION, April 2015 TABLE 3. Predictive Factors for Spontaneous Recovery in Patients With Nonparacetamol ALF (n 5 627) Univariate Analysis (n 5 179) Multivariate Analysis (n 5 152), AUC Number of Improvements n on Waiting List OR 95% CI P OR 95% CI P Sex Female Male Age at request for SU registration <50 years years Hepatic encephalopathy Grade 0, 1, or Grade 3 or Creatinine clearance <60 ml/minute/1.73 m ml/minute/1.73 m Bilirubin <200 mmol/l < < mmol/l Factor V 20% >20% PT 20% >20% NOTE: Odds ratios, 95% CI, and P values were estimated by logistic regression models. 295 mmol/l, and 69.1 ml/minute/1.73 m 2, respectively; all P < 0.001). Ninety-three patients (36%) presented with grade 3 or 4 hepatic encephalopathy (but with factor V > 20% or 30% according to age). Of the total of 258 patients in this category, 186 (72.1%) underwent LT, 48 (18.6%) recovered, and 24 (9.3%) died. Among the latter, bilirubin levels were higher, creatinine clearance values were lower, the male/ female ratio was higher, and paracetamol-induced ALF was less common in comparison with patients who recovered (all P < 0.05). Prognostic Value of the CV Criteria Table 5 shows the sensitivity, specificity, PPVs, and NPVs of the CV criteria for a poor outcome in patients presenting with ALF due to paracetamol or not. The sensitivity of CV criteria for detecting patients with a poor prognosis (death or transplantation) was significantly higher among paracetamol-induced ALF patients versus the non paracetamol-induced ALF group (81% versus 55%, P < 0.001), even though the specificity was slightly the same (56% versus 50%, P ). When transplant patients were not included in the analysis, the sensitivity for detecting patients who died was still higher in the paracetamol-induced ALF group, but the difference was not statistically significant (75% versus 69%, P ). In the paracetamol group, death or transplantation was more frequent among patients with the CV criteria versus patients without CV criteria (81.9% versus 46.0%, P < 0.001) whereas there was no difference in prognosis in the nonparacetamol group between patients with or without CV criteria (89.9% versus 91.5%, P ). As for HBV-induced ALF, there was also no difference in prognosis between patients with and without the CV criteria (94.4% versus 96.1%, P 5 1). The specificity of the CV criteria was low in both HBV-induced ALF patients and non HBVinduced ALF patients (25.0% versus 55%, P ). The sensitivity did not significantly differ between these 2 groups (67% in HBV-induced ALF patients versus 59%, P ). The sensitivity of CV criteria was significantly better in the paracetamol group versus the HBV group (81% versus 67.1%, P ). Survival The 1-year Kaplan-Meier survival estimate for the 808 patients under an intention-to-treat (ITT) analysis was 66.3% (95% CI, 62.7%-69.7%). Survival was reduced among patients presenting with FH due to a virus or paracetamol in comparison with those with other causes (medicinal products other than paracetamol, toxins, other causes, and undetermined causes; P < 0.01; Fig. 2A). After LT, the estimated 1-year

8 LIVER TRANSPLANTATION, Vol. 21, No. 4, 2015 ICHAI ET AL. 519 TABLE 4. Characteristics and Outcomes of 622 Patients as a Function of Compliance With CV Criteria Patients With Criteria Patients Without Criteria Spontaneous recovery, n n 5 42 (11.5) Died, n 5 59 (16) LT, n (72.3) n Spontaneous recovery, n 5 48 (18.6) Died, n 5 24 (9.3) LT, n (72) Sex Female (10.8) 30 (13.4) 169 (75.8) (20)* 8 (5) 114 (75) Male (12.8) 29 (20.6) 94 (66.6) (16) 16 (15) 72 (69) Age, years ( ) 45.3 ( ) 41.1 ( ) ( )* 43.1 ( ) 42.7 ( ) Etiology of FH Paracetamol (18) 21 (18) 74 (63.8) (54)* 7 (14) 16 (32) HBV 54 3 (5.6) 6 (11.1) 45 (83.3) 26 1 (4) 3 (11.5) 22 (84.5) Other virus 10 2 (20) 3 (30) 5 (50) 12 3 (25) 1 (8) 8 (67) Toxin 38 3 (7.9) 8 (21) 27 (71) 43 7 (16) 2 (5) 34 (79) Undetermined 87 8 (9) 13 (14.9) 66 (76) 63 6 (9.5) 2 (3) 55 (87) Autoimmune disease 26 1 (3.8) 3 (11.5) 22 (84.6) 36 1 (3) 5 (14) 30 (83) Other 33 4 (12) 5 (15) 24 (72.7) 28 3 (11) 4 (14) 21 (75) Hepatic encephalopathy Grade (30) 5 (17) 16 (53) Grade 1 or (20) 8 (6) 100 (74) Grade 3 or (12) 59 (16) 263 (72) (13) 11 (12) 70 (75) Creatinine clearance ( )* 24.5 ( ) 49.2 ( ) ( )* 31.1 ( ) 68.8 ( ) (ml/minute/1.73 m 2 ) Bilirubin (mmol/l) (37-139) # 90 (50-240) 200 ( ) (49-109) k# ( )** ( ) Factor V (%) (11-19) 10 (5-18) 16 (8-20) (9-26) 23.5 (9.5-35) 24 (13-32) PT (%) (10-16) 9 (1-16)** 11 (6-17) ( ) 22 (9.5-28) 16 (11-22) NOTE: The data are presented as medians (IQR) or populations (%). *P < 0.05, improvement versus died. P < 0.01, died versus LT. P < 0.01, improvement versus LT. P < 0.001, improvement versus LT. P < 0.001, died versus LT. P < 0.05, improvement versus LT. k P < 0.001, improvement versus died. # P < 0.001, improvement versus LT. **P < 0.05, died versus LT. P < 0.01, improvement versus died.

9 520 ICHAI ET AL. LIVER TRANSPLANTATION, April 2015 TABLE 5. Prediction of Poor Outcomes in Patients With ALF Based on CV Criteria Sensitivity (%) Specificity (%) PPV (%) NPV (%) Prediction of deaths versus recovery (transplant patients included)* All patients (n 5 622) Paracetamol-induced ALF Non paracetamol-induced ALF HBV-induced ALF Non HBV-induced ALF Prediction of deaths versus recovery (transplant patients not included) All patients (n 5 173) Paracetamol-induced ALF Non paracetamol-induced ALF NOTE: Values were found with the v 2 test. *Transplant patients were considered to have died if no LT was performed. Sensitivity of CV paracetamol criteria > sensitivity of nonparacetamol criteria (P < 0.001). NPV of CV paracetamol criteria > NPV of nonparacetamol criteria (P < 0.001). PPV of CV nonparacetamol criteria > PPV of paracetamol criteria (P < 0.01). NPV of CV paracetamol criteria > NPV of nonparacetamol criteria (P ) survival reached 74.2% (IQR, 70.5%-77.6%); the 5- year survival rate, estimated for the 398 patients undergoing transplantation until 2006, was equal to 65.8% (IQR, 60.9%-70.3%). Global and posttransplant 1-year survival did not significantly differ according to the period of listing (Fig. 2B). DISCUSSION This study constitutes the first exhaustive review of a recent series of French patients aged 15 years old and older presenting with FH and listed for SU LT. Several important findings deserve to be underlined: (1) paracetamol appeared to be the leading indication for SU LT in France since 2002, and this confirmed the shift in etiology for ALF in France 7,8 ; (2) 13.9% of listed patients withdrew from the waiting list because their state improved, and this concerned 28.5% of those whose FH was due to paracetamol; (3) the median time elapsing before the transplanted graft was obtained was 7.5 hours (IQR, hours); (4) the bilirubin level, creatinine clearance, grade of hepatic encephalopathy, factor V level for paracetamol ALF, and bilirubin level for nonparacetamol ALF could be seen as independent factors for a spontaneous improvement; (5) the specificity of the CV criteria was low (53%), and the sensitivity of these criteria was higher in patients presenting with ALF due to paracetamol versus patients with ALF not due to paracetamol; and (6) the sensitivity of the CV criteria was significantly better in the paracetamol group versus the HBV group (81% versus 67.1%, P ). In our patient series, 221 of 808 listed patients (27.4%) did not undergo a transplant, and 50.7% of these patients improved spontaneously (13.9% of the total). This rate of withdrawal from the waiting list contrasted with that seen in the series studied by Bernal et al., 9 in which only 2.25% of patients (7 patients) on the waiting list improved (of whom 5 were still alive). However, 22% of the patients died on the waiting list in their series versus 13.5% in ours. Thus, the selection of patients for LT appeared to comply with stricter criteria, but this led to a higher mortality rate on the waiting list. In contrast to Anglo-Saxon countries, the criteria applied most widely in France are the CV criteria, which were defined in 1991 and were based on patients presenting with viral ALF, mainly due to hepatitis B. 3,4,10 These criteria are defined by the presence of grade 3 or 4 hepatic encephalopathy associated with a factor V level < 20% (age < 30 years) or grade 3 or 4 hepatic encephalopathy associated with a factor V level < 30% (age 30 years). Survival reached 95% if these criteria were not present but 10% if 1 of these 2 sets of criteria one present. 4 The specificity of CV criteria in patients with paracetamol-induced ALF is low. With these criteria, a high rate of patients with paracetamol-induced ALF who were listed for LT (28%) withdrew from the SU list, whereas only 9.7% (61/627) of patients with ALF not related to paracetamol did. Therefore, many patients could undergo transplantation unnecessarily. Otherwise, the sensitivity and specificity of the CV criteria in these patients (excluding those who underwent transplantation) are similar to the values found in the series studied by Izumi et al. 11 (75% and 56% versus 69% and 61%). The prognostic prediction for their patients was better when the KCH paracetamol criteria were applied (sensitivity, specificity, and PPV of 69%, 96%, and 92%). Recently, Simpson et al. 12 confirmed the predictive value of KCH paracetamol criteria (88%) with a sensitivity, specificity, and PPV of 77.6%, 95.5%, and 95%, respectively. 12 In patients presenting with ALF not linked to paracetamol, the specificity and PPV of the CV criteria

10 LIVER TRANSPLANTATION, Vol. 21, No. 4, 2015 ICHAI ET AL. 521 Figure 2. (A) Survival of patients under an ITT analysis as a function of etiology and (B) posttransplant survival. were low in our series (50% and 64%, respectively), whereas those in the study by Izumi et al. 11 were 67% and 91%, respectively. However, the sensitivity was similar (69% and 71%, respectively). There was no difference between CV and KCH nonparacetamol criteria in terms of prediction in Izumi et al. s series. If the patients who died were excluded, then 13.8% (n 5 42) of the patients with the CV criteria in our study did not undergo transplantation. This suggests that the decision for LT is probably associated with criteria other than the CV criteria. The selection of patients for SU LT based on the CV criteria was associated with good survival under an ITT analysis and after LT. However, the performance of the CV criteria could be improved by the further inclusion of bilirubin levels and creatinine clearance in order to predict the outcome of paracetamolinduced ALF and by the inclusion of bilirubin levels for non paracetamol-induced ALF. Indeed, bilirubin appeared to be strongly predictive of a spontaneous improvement in patients with paracetamol and non paracetamol-alf with a good predictive ability (AUC and AUC , respectively). However, all these findings need to be confirmed in the context of a future, prospective study because of the limitations of this study. The first limitation of this study was the lack of data on the number of patients presenting with ALF who were not listed for LT (for social, psychiatric, or medical reasons) or who did not meet the criteria for LT. When we took account of these patients, the ratio between those who were considered for LT and those listed for transplantation was lower and tended toward the British ratio (39%). However, the objective of our study was to focus on patients who were actually listed and to monitor their outcomes. The second limitation affecting this study was that it was impossible (or difficult) to determine in a certain number of patients the criteria that had been used to place them on the list (KCH or CV criteria or other criteria). The lack of standardization of requests for SU listing during this period, the retrospective and multicenter nature of the study, and the period that it covered are the principal explanations for this. As of 2013, ABM standardized the demands for SU transplantation in an electronic file based on the CV and KCH criteria. Censoring these patients from the study would have biased its results and would not have been representative of the period under study. In order to take account of this bias when predicting outcomes, we considered only patients for whom only the CV criteria were applied for listing. Some items in the KCH criteria were not available for a large number of patients, and this is why it was not possible to compare the performance of the KCH and CV criteria. In our study, 171 patients (21.2%) who presented with hepatic encephalopathy graded lower than 3 at the time of listing (31 patients without hepatic encephalopathy and 140 patients with hepatic encephalopathy grade 1 or 2), so they did not meet the CV criteria. Apart from paracetamol poisoning, where the spontaneous recovery rate reached 56%, only 11% of these patients 14 improved during the waiting period, 8 died (plus 5 with paracetamol poisoning), and all of the others underwent transplantation. Patients with non paracetamol-induced ALF were placed on the waiting list before the occurrence of hepatic encephalopathy grade 3 or 4 because they met LT criteria specific to the etiology without hepatic encephalopathy taken into account (toxic FH caused by mushrooms or paracetamol for 1 criterion) or because they presented with an unfavorable etiology (autoimmune hepatitis or ALF of unknown etiology) or a borderline clinical state. In all cases, their waiting time on the transplant list provided a further opportunity to reach a final decision before the surgical procedure. During the past 10 years, the incidences of different etiologies for FH have changed. Indeed, since 2002, paracetamol poisoning has become the leading indication for LT in patients with FH in France (11.1% before 2002 versus 27.8% after 2002, P < ), and this is the same situation found in the United Kingdom and the United States, although the

11 522 ICHAI ET AL. LIVER TRANSPLANTATION, April 2015 frequency of cases remains much lower than the frequency in those countries (22% in France versus 43% in the United Kingdom). 7 This reversal has also been observed in Spain. 21 In parallel, there has been a decline in the incidence of FH due to HBV in both France and the United Kingdom. 17 However, in the present study, we did not observe a significant decrease in the frequency of this indication since 2002, unlike the period between 1986 and One hundred eighty-nine patients (32.2%) received a graft from a donor with a nonidentical but compatible blood group, and 107 of these patients belonged to group A and were matched with a group O donor (43.2% of group A recipients and 18.2% of all transplant patients). This figure was higher than that seen in the series studied by Farmer et al., 22 in which 22% of transplant patients received a nonidentical but compatible graft after a median delay of 2 days (blood groups not specified). Clearly, group A patients benefit most in terms of the time required to find a graft because they accumulate the advantages of being able to receive an A or O graft and because these groups are a majority in the general population (45% and 43%, respectively, in France). Requests for blood group derogations at registration may, therefore, be a subject for debate, particularly because the median waiting time in France is short as shown recently in another series in France. 23 However, the rapid and unpredictable course of liver disease justifies such requests for nonidentical grafts. The clinical status and, in particular, any deterioration in the neurological, hemodynamic, respiratory, or renal status prevail in this decision-making process. Unfortunately, these data were not available for our study. In countries where the waiting time is longer, LT with a nonidentical graft seems to constitute a good alternative. This is the case in Nordic countries, where the waiting time for a graft from an identical blood group is 4.8 days, whereas that for a compatible graft is 2.2 days. This long delay may justify the fact that 34% of FH patients in these countries receive a compatible and nonidentical graft. 24 It is important to consider this point because group O patients who are listed for elective surgery on account of chronic liver disease may find themselves at a disadvantage. In conclusion, this series, which grouped all adult patients presenting with ALF in France, showed that the criteria for emergency LT are currently insufficient. Paracetamol and nonparacetamol criteria for LT need to be separated. The performance of the CV criteria could be improved by the integration of bilirubin into the prediction of outcomes for non paracetamolinduced ALF and by the integration of bilirubin and creatinine clearance for nonparacetamol ALF. A prospective study, therefore, needs to be performed in order to reassess the criteria used in France with respect to emergency LT. ACKNOWLEDGEMENTS The authors thank the members and institutions participating in the French Liver Transplant Teams from 1997 to The authors further thank all the coordinators from the study sites who participated in this study. REFERENCES 1. Bernal W, Auzinger G, Dhawan A, Wendon J. Acute liver failure. Lancet 2010;376: Bernal W, Donaldson N, Wyncoll D, Wendon J. Blood lactate as an early predictor of outcome in paracetamolinduced acute liver failure: a cohort study. Lancet 2002; 359: Bismuth H, Samuel D, Castaing D, Adam R, Saliba F, Johann M, et al. Orthotopic liver transplantation in fulminant and subfulminant hepatitis. The Paul Brousse experience. Ann Surg 1995;222: Bernuau J, Samuel D, Durand F, Saliba F, Bourlière M, Adam R, Gugenheim J, et al. Criteria for emergency liver transplantation in patients with acute viral hepatitis and factor V (FV) below 50% of normal: a prospective study. Hepatology 1991;14(4):Abstract 49A. 5. Bismuth H, Samuel D, Gugenheim J, Castaing D, Bernuau J, Rueff B, Benhamou JP. Emergency liver transplantation for fulminant hepatitis. Ann Intern Med 1987;107: O Grady JG, Alexander GJ, Hayllar KM, Williams R. Early indicators of prognosis in fulminant hepatic failure. Gastroenterology 1989;97: Ichai P, Samuel D. Etiology and prognosis of fulminant hepatitis in adults. Liver Transpl 2008;14(suppl 2):S67- S Ichai P, Faria L, Saliba F, Azoulay D, Antonini T, Karam V, et al. Improvement of survival in fulminant hepatitis in the recent years. Hepatology 2006;44(N 4, Suppl):371 A. 9. Bernal W, Cross TJ, Auzinger G, Sizer E, Heneghan MA, Bowles M, et al. Outcome after wait-listing for emergency liver transplantation in acute liver failure: a single centre experience. J Hepatol 2009;50: Bernuau J, Goudeau A, Poynard T, Dubois F, Lesage G, Yvonnet B, et al. Multivariate analysis of prognostic factors in fulminant hepatitis B. Hepatology 1986;6: Izumi S, Langley PG, Wendon J, Ellis AJ, Pernambuco RB, Hughes RD, Williams R. Coagulation factor V levels as a prognostic indicator in fulminant hepatic failure. Hepatology 1996;23: Simpson KJ, Bates CM, Henderson NC, Wigmore SJ, Garden OJ, Lee A, et al. The utilization of liver transplantation in the management of acute liver failure: comparison between acetaminophen and non-acetaminophen etiologies. Liver Transpl 2009;15: Ichai P, Duclos-Vallee JC, Guettier C, Hamida SB, Antonini T, Delvart V, et al. Usefulness of corticosteroids for the treatment of severe and fulminant forms of autoimmune hepatitis. Liver Transpl 2007;13: Escudie L, Francoz C, Vinel JP, Moucari R, Cournot M, Paradis V, et al. Amanita phalloides poisoning: reassessment of prognostic factors and indications for emergency liver transplantation. J Hepatol 2007;46: Ganzert M, Felgenhauer N, Zilker T. Indication of liver transplantation following amatoxin intoxication. J Hepatol 2005;42: Ostapowicz G, Fontana RJ, Schiødt FV, Larson A, Davern TJ, Han SH, et al.; for U.S. Acute Liver Failure Study Group. Results of a prospective study of acute liver failure at 17 tertiary care centers in the United States. Ann Intern Med 2002;137: Bernal W, Hyyrylainen A, Gera A, Audimoolam VK, McPhail MJ, Auzinger G, et al. Lessons from look-back

12 LIVER TRANSPLANTATION, Vol. 21, No. 4, 2015 ICHAI ET AL. 523 in acute liver failure? A single centre experience of 3300 patients. J Hepatol 2013;59: Lee WM. Acute liver failure in the United States. Semin Liver Dis 2003;23: Lee WM. Etiologies of acute liver failure. Semin Liver Dis 2008;28: Larson AM, Polson J, Fontana RJ, Davern TJ, Lalani E, Hynan LS, et al.; for Acute Liver Failure Study Group. Acetaminophen-induced acute liver failure: results of a United States multicenter, prospective study. Hepatology 2005;42: Escorsell A, Mas A, de la Mata M; for Spanish Group for the Study of Acute Liver Failure. Acute liver failure in Spain: analysis of 267 cases. Liver Transpl 2007;13: Farmer DG, Anselmo DM, Ghobrial RM, Yersiz H, McDiarmid SV, Cao C, et al. Liver transplantation for fulminant hepatic failure: experience with more than 200 patients over a 17-year period. Ann Surg 2003;237: Saliba F, Camus C, Durand F, Mathurin P, Letierce A, Delafosse B, et al. Albumin dialysis with a noncell artificial liver support device in patients with acute liver failure: a randomized, controlled trial. Ann Intern Med 2013;159: Bjøro K, Ericzon BG, Kirkegaard P, H ockerstedt K, S oderdahl G, Olausson M, et al. Highly urgent liver transplantation: possible impact of donor-recipient ABO matching on the outcome after transplantation. Transplantation 2003;75: