Survival Analysis of Liver Transplants in Patients with Acute Liver Failure from Acetaminophen and Mushroom Toxicity

Introduction

Acute liver failure (ALF) is a condition defined as severe, rapid onset injury to hepatocytes, which leads to altered coagulation and mentation in absence of chronic liver disease. ALF accounts for approximately 8% of liver transplantations (LT). Implementation of this treatment modality decreased overall mortality from nearly 80% to 33%. Despite that improvement, the prognosis in most severe ALF cases is still unfavorable and is associated with significant financial expenses resulting from rising costs of intensive care therapies [1].

The most common ALF cause is acetaminophen (paracetamol) intoxication, responsible for about 50% of cases. Its hepatotoxicity is mainly attributed to production of heightened levels of toxic metabolite N-acetyl-p-benzoquinone imine (NAPQI) exceeding liver detoxification capacity, leading to irreversible hepatocyte necrosis [1,2].

Acetaminophen poisoning begins with non-specific symptoms, with first manifestations of liver injury being noticeable 12–36 hours after ingestion [3]. The overall mortality is less than 1% but can reach over 50% in the highest-risk subgroup [2,4].

In addition to various synthetic toxicants, liver injury can be caused by naturally occurring agents, including substances found in Amanita phalloides fungi; ingestion leads to hepatic cell necrosis resulting in up to 30% mortality [5–7].

In A. phalloides poisoning, toxicity is mediated mainly by amatoxins that are responsible for hepato- and nephrotoxicity by binding to RNA polymerase II, and impaired protein synthesis results in cell death. Presentation is similar to acetaminophen poisoning, with non-specific symptoms proceeding to acute liver injury [8].

Despite having readily available tools for assessment of ALF patients (eg, King’s College Criteria), any strict criteria cannot be viewed as a sole point of reference in deciding whether a patient is an appropriate candidate for salvage LT, especially while considering rare LT indications that have no standardized prognostic scales [9].

Without timely qualification for emergency LT in cases of irreversible hepatocyte damage, the chances of patient death increase tremendously [10]. That, in addition to the ever-growing organ shortage, makes accurate selection of suitable candidates for organ transplantation and optimization of LT outcomes a matter of the highest importance [11].

Unfortunately, given the rarity of the most severe cases of toxic ALF, the data on qualification process and postoperative results are still scarce. Therefore, there is a need to determine the factors contributing to unfavorable treatment outcomes, especially when LT is considered in non-routine indications.

The aim of this study was to assess outcomes of patients undergoing LT due to acetaminophen or A. phalloides poisoning and to define factors affecting short-term (90-day) survival, thus improving decision-making in ALF cases.

Material and Methods

This study was reviewed and approved by the Bioethics Committee of Medical University of Warsaw, Poland, No. AKBE/82/2024. We retrospectively selected patients using case-by-case analysis of the transplant database of the Department of General, Transplant and Liver Surgery of the Medical University of Warsaw out of all LTs performed between 1994 and 2021 among patients whose indication for LT was ALF without preexisting hepatic disease. All cases of acute-on-chronic liver failure and retransplantations were excluded from the analysis.

Among selected patients, those who had confirmed history of either acetaminophen or A. phalloides poisoning were enrolled. Demographic characteristics, laboratory results at transplantation, and course of perioperative period were retrieved from patients’ archival records. Donor’s characteristics were retrieved from organ procurement reports. Admission, progress, and follow-up notes were screened to confirm the accuracy of patient selection and to obtain information on somatic and mental comorbidities and postoperative outcomes. Cold ischemia time and transfusion volume were recovered from intraoperative charts.

Statistical analyses were performed using Statistica 13.3 software. Group characteristics described in quantitative and qualitative variables are presented as medians (interquartile range [IR] and numbers (percentages), respectively. Basic characteristics of acetaminophen and A. phalloides groups were compared by applying the Mann-Whitney U test and chi-square test to identify any significant differences between groups.

Importance of 90-day mortality on overall treatment results was shown in survival analysis. The 1-, 3-, and 5-year survival rates of patient and graft were calculated using Kaplan-Meier method. Patient survival was defined as time between LT and date of death. All deaths were confirmed in the National Registry. Graft survival was defined as time between LT and either retransplantation or patient death. Time of censoring is the last recorded visit in an outpatient clinic, which were scheduled in at least 6-month intervals.

Logistic regression was used to define risk factors of short-term treatment failure. Primary endpoint of all models was 90-day patient survival. All clinically significant patient and donor characteristics were included in univariable regression analysis. Factors that turned out to be significant in univariable analysis were taken to build a multivariable model using backward elimination with Wald statistic 0.3 as threshold for step elimination. Odds ratios (OR) are presented with 95% confidence intervals (CI). For further clarification of regression results, Spearman correlation analysis of explanatory variables was used.

Missing data were analyzed and appeared randomly within the cohort; the main reason was incomplete medical records. The significance level was set at 0.05.

Results

Out of 2464 LTs performed in the Department of General Transplant and Liver Surgery of the Medical University of Warsaw, we identified 110 LTs performed due ALF without preexisting hepatic disorder. Among those, ALF etiology was unknown in 38 cases, followed by A. phalloides (n=21) and acetaminophen (n=18) poisoning.

Data of 39 patients undergoing LT due either acetaminophen or A. phalloides poisoning were set under analysis. There were 24 females and 15 males. The median age was 41years (IR 21–50; standard deviation 16.6).

All patients received full-organ grafts procured from donors after brain death (DBD). As compared to patients with ALF in the course of A. phalloides intoxication, those after acetaminophen overdose were younger (P<0.001) and required fewer red blood cell (RBC) (P=0.003) and fresh frozen plasma (FFP) (P=0.001) transfusions intraoperatively. The 90-day mortality rate was 22.20% (4 of 18) and 38.10% (8 of 21) in LT recipients with acetaminophen and A. phalloides poisoning, respectively (P=0.399; Table 1).

Screening of patients’ records for history of confirmed mental disorders revealed 3 cases of affective disorders and 1 case of schizophrenia in the acetaminophen group and none in the A. phalloides group.

In multivariate analysis, the only independent factor affecting 90-day mortality was higher number of transfused RBC units (OR 1.609; 95% CI 1.119–2.314; P=0.010).

Risk factors for 90-day mortality that were significant only in univariate regression analysis included occurrence of renal failure requiring hemodialysis in perioperative period (OR 13.50; 95% CI 1.473–123.748; P=0.021), lower concentration of serum albumin at transplantation (OR 4.281 per 1g/dL decrease; 95% CI 1.273–14.397; P=0.019), lower activity of serum alanine transferase at transplantation (ALT) (OR 1.029 per 100U/L decrease; 95% CI 1.003–1.056; P=0.029), more days spent on a transplantation wait-list (OR 2.289 per 1 day increase; 95% CI 1.189–4.406; P=0.013), and higher number of FFP (OR 1.346 per 1 unit; 95% CI 1.104–1.640; P=0.003) units transfused intraoperatively (Table 2). Higher serum ALT activity at transplantation was correlated with shorter time spent on a transplant wait-list (Spearman’s R=−0.560; P<0.001).

Independent factors associated with graft loss were renal failure requiring dialysis in the perioperative period (OR 10.912; 95% CI 1.611–73.904; P=0.014) and number of days on a transplant wait-list (OR 3.456; 95% CI 1.170–10.208; P=0.025) (Table 3).

Risk factors significantly associated with graft loss only in univariate regression analysis were lower serum albumin concentration at transplantation (OR 3.823 per 1 g/dL decrease; 95% CI 1.273–11.483; P=0.017), lower serum ALT activity at transplantation (OR 1.025 per 100 U/L decrease; 95% CI 1.003–1.048; P=0.027), higher number of RBC units (OR 1.308 per 1 unit; 95% CI 1.046–1.635; P=0.018), and FFP units transfused intraoperatively (OR 1.239 per 1 unit; 95% CI 1.049–1.262; P=0.012).

The 1-, 3-, and 5-year survival rates were 66.6%, 63.5%, and 59.6%, respectively (Figure 1). The main cause of death was multiorgan dysfunction syndrome (MODS), which occurred in 9 patients, followed by septic shock in 2 patients. One patient developed alcoholic liver disease and died 6 years after LT of complications related to liver cirrhosis.

The 1-, 3-, and 5-year graft survival rates were 58.9%, 55.9%, and 48.2% (Figure 2). Most graft losses were a result of patient death. Others causes were primary non-function (3 cases), hepatic artery thrombosis (2 cases), and portal vein thrombosis (1 case).

Discussion

Liver transplantation in ALF is associated with high risk of postoperative complications and mortality, with 33.4% 1-year mortality in an analyzed cohort [12]. The rarity of LT in toxin-induced ALF justifies merging data of 2 clinically similar groups of patients undergoing LT for acetaminophen and A. phalloides poisoning to provide more comprehensive data for assessment of potential risk factors impairing survival.

Acetaminophen toxicity causes about half of all ALF cases, but most patients are able to recover without LT [1]. Given the common occurrence of acetaminophen-related liver injury, it is interesting that in the cases we analyzed there were more patients transplanted due to A. phalloides ALF (16.6% for acetaminophen and 19.1% for A. phalloides). This unique distribution is most likely due to the tradition of mushroom foraging, which is still widely popular in Poland, and often leads to accidental ingestion of Amanita mistaken for edible fungi.

One of the biggest studies investigating A. phalloides poisoning is a review article by Tan et al, including results of 131 studies published between 1975 and 2020, that described 76 cases of patients undergoing LT [13], which makes the incidence of A. phalloides LT in our center even more unique.

Demographics of the analyzed subgroups differed in median age of recipient, with the acetaminophen group being significantly younger, with an evident, although not statistically significant, trend towards females in comparison with the A. phalloides group. That discrepancy might be explained by acetaminophen poisoning being most prevalent in females and in young adults [14].

There were no differences in severity of patients’ preoperative condition, as we found no differences in MELD (Model for End-Stage Liver Disease) score or in fulfilment of King’s College Criteria at transplant. There were also no differences found in any laboratory parameters’ values at transplant.

Nevertheless, the A. phalloides group received more RBC and FFP transfusions, perhaps because older patients presented more often with various, especially cardiac, comorbidities, which can cause higher risk of intraoperative circulatory instability, requiring larger intraoperative transfusion volumes.

In ALF treatment, before LT is considered, there are various modalities of conservative therapy that should be taken into account. Acetaminophen poisoning can be successfully managed by administration of N-acetylcysteine (NAC), which is reported to be used in A. phalloides intoxication as well, but data on the effectiveness of this treatment in A. phalloides poisoning is discouraging and it did not improve patient survival [3,8,13]. In our center, all eligible patients received NAC. Unfortunately, due to the retrospective nature of this study, we were not able to obtain information about the administered dosage and duration of treatment.

Additionally, in the preliminary stage of treatment, gastric lavage and oral charcoal administration can aid in removal of toxicants [15]. Nevertheless, those strategies are justified only in early stages of the disease. In acetaminophen poisoning, where the toxic factor is easier to define, mechanical detoxication can be implemented with higher efficacy than in patients poisoned with rare substances [4]. Nearly all our patients were transferred to a reference center only when their clinical condition significantly deteriorated. Therefore, we were rarely able to implement any modalities of mechanical detoxification.

Patients developing ALF are prone to development of acute renal failure (ARF) [1]. About a half of analyzed patients developed ARF requiring hemodialysis in the perioperative period, which was found to be one of the factors affecting 90-day mortality and graft loss. Additionally, renal failure proved to be an independent factor correlated with graft loss. Those results correspond with the literature reporting that acute renal failure increases post-LT mortality by up to 90% [16].

Assessment of ALF oftentimes relies on the King’s College criteria first described by O’Grady et al in 1989 [9]. Unfortunately, its predictive value is limited in non-acetaminophen-induced ALF [17]. Patients in the present study were qualified for emergency LT based on the King’s College criteria in association with severity of clinical condition and a lack of response to supportive treatment. King’s College criteria at transplantation were met in only 66.66% of A. phalloides- and 57.14% of acetaminophen-poisoned patients. Fulfilment of King’s College criteria and MELD score were not found to be risk factors for 90-day mortality or graft loss.

Out of all analyzed laboratory parameters, lower serum albumin concentration was correlated with higher mortality and graft loss, confirming that albumin concentration reflects the efficiency of the liver’s synthetic function [1]. Interestingly, higher ALT activity was correlated with lower 90-day mortality, conflicting with common knowledge of ALT being a marker of hepatocyte damage [18]. This surprising outcome might be explained by higher activity of ALT at transplantation correlating with a shorter time spent on a transplant wait-list. The decision to qualify patients for urgent LT and graft allocation might have been influenced by the observed surge in aminotransferase activity, thereby improving patient outcomes.

In a light of the ever-growing organ shortage, transplant centers are forced to extend donor acceptance criteria [19,20]. Higher donor risk index (DRI) was reported to not be associated with 90-day mortality or the occurrence of graft loss, contrary to days spent on a transplant wait-list [21]. Longer time on a wait-list also turned out to be an independent risk factor for graft loss. These results suggest that for ALF patients graft quality is not as important as accurate identification of patients with low chances of spontaneous recovery.

Prompt qualification of non-improving patients to urgent LT before life-threatening consequences develop can improve patient survival, as reported by Tan et al, who found a noticeable, although non-significant, survival benefit in patients transplanted due to A. phalloides ALF versus patients receiving supportive care only, with a 75% overall survival for LT patients and 59% for non-LT patients [13].

The so-called “toxic liver syndrome” (TLS) is a life-threatening phenomenon during the standby period before ALF LT. TLS can cause extensive liver necrosis followed by release of toxic metabolites, leading to profound shock [22].

The only available salvage TLS treatment is total hepatectomy with subsequent creation of a porto-caval shunt, even if there is no available organ for transplantation at the time of a procedure. This somewhat controversial approach can be an effective last-chance treatment in patients with irreversible liver damage [23,24]. We used this approach in 1 patient with A. phalloides poisoning. Unfortunately, the patient died due to acute respiratory distress syndrome (ARDS) while awaiting an available graft and thus was not included in the analysis. Nevertheless, during hepatectomy, immediately after the necrotic liver was removed, we observed circulatory improvement. Histopathological examination of the excised specimen (Figure 3A, 3B) revealed >90% hepatocyte necrosis.

In patients that survive long enough to undergo LT, performing a procedure can pose a challenge even for experienced transplant teams due to the recipient’s severe clinical condition and coagulopathy that can only be partially reversed by supportive care. These patients usually present without portal hypertension, which forces modification of standard operative technique by utilizing veno-venous bypass (VVB) or temporary porto-caval shunt to prevent intestinal circulatory obstruction. Despite being performed with implementation of VVB, end-to-end vena cava anastomosis did not affect 90-day mortality or graft loss [25].

We found that the number of intraoperatively transfused units of RBC and FFP was associated with increased 90-day mortality and graft loss. Additionally, the number of transfused RBC units was independently associated with higher 90-day mortality.

Unequivocally, blood loss cannot be attributed to surgical technique only, although it appears to be the only factor that surgeons have any direct influence on. Consequently, procedures on the most unstable patients should be performed by the most experienced specialists.

The 1-, 3-, and 5-year survival rates in patients undergoing LT due to acetaminophen- and A. phalloides-induced ALF were 66.6%, 63.5%, and 59.6%, respectively (Figure 1). Those results, although unsatisfactory, correspond with already published data that report 53–75% 5-year survival for patients undergoing LT due to A. phalloides toxicity and 70% 5-year survival for patients who underwent LT due to acetaminophen-induced ALF [13,26–28]. Most of the deaths occurred due to MODS, followed by 2 patients who died due to septic shock. These results correspond with data published by Barshes et al, who found ALF patients were at the highest risk of death in the first 3 months after LT, with causes of death being consistent with our results [29].

The reported 3-year survival of ALF patients is higher than for those transplanted for chronic liver disease, and is similar to other critically ill patients with high-MELD scores, at about 80% [30]. Our outcomes, being worse than that, might be a result of the long study period (27 years), during which available treatment modalities, operative technique, and center’s experience underwent immense improvement. Development of new treatment strategies over the years helped increase survival of ALF patients from 16.7% in 1973 to 62.2% in 2008, as reported by King’s College [31].

Another factor that could impair outcomes in patients transplanted due to acetaminophen toxicity is the high incidence of preexisting mental health issues, which increase the risk of non-compliance [14]. Among the patients included in the present study, 4 had confirmed history of psychiatric disorders, although this number might be greatly underestimated due to retrospective data collection. Notably, the only 2 recorded cases of non-adherence occurred in patients with known pre-transplantation psychiatric diagnoses.

Postoperative non-compliances can only be partially avoided. Many patients develop profound hepatic encephalopathy and are unable to give informed consent, and adequate psychiatric examination and proper mental preparation for LT is many times impossible. Consequently, patients undergoing salvage LT and their relatives should be provided with careful postoperative follow-up and psychological support immediately after transplantation, as providing patients with tools to develop healthy coping mechanisms and adequate social support systems can improve clinical outcomes of transplant recipients [32].

Acetaminophen- and A. phalloides-induced ALF are rare indications for LT. The severity of a patient’s preoperative condition is assessed by the presence of renal failure and number of transfused RBC and FFP units. To provide the best possible prognosis for these patients, and to shorten the time from diagnosis of ALF to LT, multidisciplinary teams of transplant surgeons, intensive care specialists, hepatologists, and psychologists are required, which requires prompt transfer of ALF patients to highly specialized clinical centers for further assessment.

Conclusions

LTs performed in management of A. phalloides and acetaminophen poisoning are high-risk procedures with high mortality. Treatment outcomes are mainly influenced by the severity of the patient’s preoperative condition. Prompt LT qualification of patients unresponsive to conservative therapy can lead to improved survival. Given the rarity of LTs performed due to acetaminophen and A. phalloides poisoning, further investigation of the topic is required to continuously increase the effectiveness of ALF treatment.