We conducted a retrospective Spanish multicenter study involving 11 centers with extensive experience in LT. These centers accounted for 43% of the total number of LTs performed in Spain in 2020 [5].

Urgent LT performed in patients >18 years due to ALF between 2001–2020 were included. Criteria for ALF were a severe acute liver injury lasting less than 26 weeks, with jaundice, liver synthetic failure (INR ≥1.5 or prothrombin rate <40%), and hepatic encephalopathy (HE) in a patient without known chronic liver disease.

Exclusion criteria comprised patients under 18 years old and patients with pre-existing liver disease. The acute manifestation of certain chronic liver diseases (Wilson’s disease, HBV reactivation in a non-cirrhotic liver, acute Budd-Chiari, and autoimmune hepatitis) was included as an exception. Patients with prior LT and acute liver injury due to primary graft nonfunction or other causes were excluded.

Data were acquired from each LT center through a review of medical records.

In Spain, when a patient experiences ALF, the criteria to indicate an urgent LT are based on either fulfilling King`s College Criteria (KCC) [19], Clichy criteria [20] or presence of HE. A national urgent code is activated, enabling the allocation of the first suitable organ available within the country to the ALF patient. The median time until a liver is offered is approximately 40 h, and around 50% of patients receive a LT within 24 h [21].

This study was approved by the Ethics Committee of Clinical Research of La Fe Universitari and Politécnic Hospital (ref number: 2021-096-1) and was conducted according to the standards of Good Clinical Practice, adhering to the ethical principles outlined in the 1975 Declaration of Helsinki.

An exemption from the requirement for informed consent was granted due to the retrospective nature of the study. Some patients had been relocated or were no longer reachable during the study period.

To ensure confidentiality, patient information included in the database was anonymized and identified by a numerical code, in compliance with data protection legislation.

The recorded variables included donor and recipient demographic features, epidemiological information, clinical and biochemical data before and after LT, clinical post-LT outcomes, patient and graft survival and variables associated with mortality.

Recipient variables (demographics, co-morbidities and toxics abuse)

The diagnosis of cryptogenic ALF was reached after excluding any other aetiology through an exhaustive pre-LT differential diagnosis and the explant biopsy. Patients who received a LT in the context of an AI hepatitis fulfilled the criteria for ALF. No evidence of liver cirrhosis was found in the explants. The temporal classification of ALF (hyper acute, acute and sub-acute) was defined according to the interval between the onset of jaundice and the development of hepatic encephalopathy (published by O’Grady JG in 1993) [22].

Regarding pre- and post-LT complications, acute kidney injury (AKI) and chronic kidney disease (CKD) were established following KDIGO criteria [23, 24]. Renal replacement therapy (RRT) included both intermittent haemodialysis and continuous RRT. Infections were confirmed with positive culture or resolution after antibiotic treatment. Respiratory failure was defined as the necessity for mechanical ventilation, rather than in the context of HE. Early graft dysfunction was based on the definition proposed by Olthoff et al. [25], and acute liver allograft rejection was categorized following the Banff classification [26]. Finally, graft steatosis was assessed by biopsy.

A descriptive analysis was conducted for all the studied variables. Continuous variables are described as means or medians with standard deviation (SD) or quartiles 1 (Q1) and 3 (Q3) as appropriate, and qualitative variables as absolute and relative frequencies.

The normal distribution of outcome variables was confirmed using the Kolmogorov-Smirnov test. Chi-square and Fisher’s exact test were used to assess the degree of association between categorical variables, Student’s t and ANOVA model to compare quantitative variables, and non-parametric Mann-Whitney and Kruskal-Wallis tests to analyse the distribution of at least ordinal variables in 2 or more independent groups.

Graft and patient survival analyses were performed with Kaplan-Meier survival curves.

Variables associated with mortality and re-transplantation were determined using univariate and multivariate Cox regression tests and expressed by hazard ratio (HR) and 95% confidence interval (CI). The initial multivariate model included the variables with a p value < 0.10 in the univariate analysis. Variables with a p value above this threshold could be included if considered clinically relevant by the investigators.

A p-value of <0.05 was considered significant for all analyses.

Data analysis was performed using SPSS version 22.0 (IBM, Chicago, USA).

A total of 217 adult patients received urgent LT due to ALF between January 2001 and December 2020. Among them, 134 were women (61.8%). The overall median age was 41 years old (IQR 32–53). Baseline clinical variables and pre-LT management are shown in Table 1, and analytical data on the LT Day in Supplementary Table 1.

A small number of patients had concomitant diseases or toxic habits. The prevalence of arterial hypertension (AHT), diabetes and dyslipidaemia were 11.6%, 3.7% and 8.3%, respectively. Regarding toxic substances, the smoking rate was 27.8%, 15% consumed alcohol regularly and 7.5% were drug users. A concomitant autoimmune non-liver disease was present in 14.4%, and 12.1% reported a psychiatric disease.

The predominant aetiologies of ALF were cryptogenic (26.7%) and autoimmune (26.3%). Viral aetiologies accounted for less than 25%, with hepatitis B (HBV) being the most common (17.1%). Drug-induced liver injury (DILI) represented 17% of LT indications. Only 4.1% of patients who underwent LT due to DILI-ALF did so in the context of acetaminophen intake.

In terms of temporality, most cases were acute (40.6%) and hyperacute (31.3%). The most frequent complications were ascites (42.8%) and AKI (39%), while infections and haemorrhagic complications were uncommon. RRT was used in 22.1%. The median MELD (Model for end-stage Liver Disease) score on the day of LT was 25.

Before LT, antibiotic prophylaxis was widely implemented (73.3%). The use of NAC and especially MARS had little relevance in our cohort of patients.

All the patients were transplanted with a national urgent priority, resulting in a median time on the WL of only 1 day (IQR: 1–2 days).

Compliance with the KCC and Clichy criteria was 91.7% and 55.9%, respectively. Of note, only a limited number of patients (n = 59) had Factor V determination performed, especially during the early years.

The cohort was subdivided into two 10-year periods (2001–2010 and 2011–2020). The number of ALF-LT remained stable overtime: 113 patients (3.6%) in 2001–2010 and 114 patients (3.1%) in 2011–2020.

Cryptogenic and autoimmune were the most common aetiologies of ALF-LT overall. Autoimmune (22.3% vs. 29.8%) and DILI (13.6% vs. 21.1%) aetiologies increased with time while HBV showed a decline (23.3% vs. 11.4%), although without reaching statistical significance (p 0.115). Despite the increase in DILI, Acetaminophen toxicity was not particularly prevalent and even decreased with time (8.7% and 3.5%). Cryptogenic ALF remained stable. Other viruses, such as HAV (1% and 1.8%) or HEV (1% and 0.9%) were extremely uncommon in both periods (Figure 1).

We first conducted an analysis of LT characteristics and post-LT evolution of the whole cohort (Supplementary Table 2). Most grafts were total (98.6%), with AB0 compatibility (isogroup 62.5% and compatible 37%), and showed minimal steatosis (<10% in 91.3% of grafts) and most donors were brain dead. In the early post-LT period, the main complications were infections (60.7%) and AKI (61%), while in the late post-LT period, AHT (30.3%), biliary complications (27.4%) and CKD (19.7%) predominated. The mortality rate was 27.2%, with infections (41.5%) and liver-related complications (20.8%) being the leading causes of death. The survival rates at 1, 5, and 10 years were 82%, 78% and 72%, respectively (Supplementary Figure 1).

Regarding differences over time in pre-LT characteristics and management (Table 2), there was a trend towards an increase of women (55.3% vs. 67.3%) approaching statistical significance (p 0.065) and a decline in the rate of Caucasian race (90.3% vs. 77.2%; p 0.001). Alcohol consumption was reported less frequently in recent years (21.6% vs. 8.9%; p 0.01). Antibiotic prophylaxis and, notably, the use of NAC significantly increased (65.4% vs. 78.9%; p 0.04% and 4% vs. 30.1%; p < 0.001).

We also examined the changes in post-transplant management and outcome between the first and second decade (Table 2). There was a higher use of triple immunosuppression (61.9%–79.1%; p < 0.001) and basiliximab (34.1%–67%; p 0.002) in the recent cohort. Some differences were found in post-LT complications. Infection rates increased overtime (52.5% vs. 66.2%; p 0.02) while respiratory insufficiency decreased (29.1% vs. 16.1%; p 0.022). In the long-term, there was a reduction in CKD (27.1% vs. 13.6%; p 0.022), cardiovascular events (10.6% vs. 1%; p 0.003) and mortality (37.6% vs. 17.9%; p 0.001) in recent years. One year mortality improved with time, not reaching statistical significance (19% in the first cohort vs. 16.1% in the latter, p 0.575). Evolution of patient survival rates between the two time periods is shown in Figure 2. (80.5%, 74% and 67% vs 84%, 82% and 82% at 1, 5, and 10 years, respectively).

Given the observed increase in AI/DILI aetiologies, we implemented an analysis to determine whether there were differences in management and outcome when comparing AI/DILI ALF group to the rest of aetiologies (Supplementary Table 4). A total of 95 patients were transplanted in the context of AI or DILI ALF, and 122 patients had other ALF aetiologies. Subacute presentations were more prevalent in AI/DILI aetiologies (34.7% vs. 21.7%; p 0.044), with a different trend for hyperacute presentations. AKI was significantly less common AI/DILI subgroup (30.1% vs. 45.8%; p 0.024). In terms of post-LT outcome, differences were observed in the IS management (higher use of triple IS, p 0.034) and in early complications (lower requirement for RRT, p 0.015; higher incidence of infections, with a lower rate of bacterial infections, p 0.017; and a decrease in bleeding and CV complications, p 0.020 and p 0.021). A significant finding in late post-LT outcome was the lower rate of de novo tumours in AI/DILI aetiologies (3.6% vs. 10.6%, p 0.034), as well as lower mortality at 1-year post-LT (12% vs. 21.7%, approaching statistical significance: p 0.065).

Given that most deaths occurred early post-LT, we determined variables independently associated with mortality at 1-year post-LT (Table 3).

Significant variables related to patient’s baseline characteristics, clinical presentation, pre-LT complications and laboratory data at LT predicted poor outcomes in the univariate logistic analysis. Obese patients were at a significantly higher risk of death than those with normal BMI (HR = 3.33; p = 0.037). AHT and dyslipidaemia significantly influenced survival time (HR = 2.75; p = 0.008 and HR = 2.75; p = 0.016). Acute presentation was related to lower mortality (HR = 0.37; p = 0.017). Among the complications detected prior to transplantation, AKI, respiratory insufficiency, infections and vasopressor use significantly worsened the prognosis (p < 0.05). Finally, an increase in creatinine, sodium, phosphorus, ammonium, and lactate levels and a decrease in Factor V were independently associated with death (p < 0.05). The remaining pre-LT variables were not statistically significant.

Multivariable logistic regression analysis was performed on selected baseline variables from the univariate analyses, including independent predictors with clinical relevance, that were previously identified as significant (p < 0.05) and that were available in a relevant number of patients. Obesity, AHT, AKI, infections and sodium level on LT-day were entered into the multivariable model, and AHT, AKI and sodium remained as independent risk factors (HR 4.002 p 0.022, HR 3.819 p 0.023 and HR 1.065 p 0.047, respectively).

There was a trend towards an increase of women over time (55.3% vs. 67.5%), although this difference was not statistically significant (p 0.065) (Table 2).

Autoimmune and cryptogenic aetiologies were more frequent in women (31% vs. 19% and 31% vs. 20%) while HBV was more common in men (29% vs. 10%) (p 0.007).

Before LT, men had a higher history of alcohol, tobacco and drug consumption (p < 0.05). AKI was more frequently observed in men (52% vs. 29%). Renal function, ALT levels, platelets count and MELD score pre-LT were worse in men (p < 0.05). No significant differences were found in other pre-LT characteristics.

Early post-LT complications such as AKI (73% vs. 53%) and haemorrhage (26% vs. 14%) were more frequent in men, while rejection was more common in women (11% vs. 22.5%) (p < 0.05). Later complications including AHT (36% vs. 27%), dyslipidaemia (25% vs. 11%), CKD (24% vs. 17%) and biliary complications (32% vs. 21%) were all more frequent in men but without reaching statistical significance. Causes of death, survival and re-LT were similar in both groups (Supplementary Table 3).