Acute liver failure from anti-tuberculosis drug-induced liver injury: An update

Abstract

Tuberculosis (TB) is still a major public health issue in developing countries, where it causes a heavy disease burden. Although current anti-TB treatment regimens demonstrate high efficacy, the hepatotoxic potential of first-line anti-TB drugs (ATDs) - particularly isoniazid, rifampicin, and pyrazinamide—poses a considerable risk, as these agents are associated with a significant incidence of ATD-induced liver injury (AT-DILI). The clinical presentation of AT-DILI can range from asymptomatic elevations in serum transaminases, which may resolve spontaneously due to hepatic adaptation, to acute liver failure (ALF), a potentially life-threatening condition. A recent meta-analysis reported a global incidence of AT-DILI of 11.5%, with rates varying from 2% to 28%. Approximately 7% of patients with AT-DILI progress to ALF, a condition characterized by a poor survival rate with medical therapy. ATD-induced ALF (AT-ALF) is clinically indistinguishable from ALF due to other causes and disproportionately affects young female patients, typically within eight weeks of treatment initiation. Emergency liver transplantation has become an effective therapeutic option for AT-ALF, although outcomes are generally poorer compared to elective transplantation. This minireview provides a comprehensive overview of AT-ALF, covering its epidemiology, risk factors, clinical presentation, prognosis, and treatment options.

Key Words: Tuberculosis; Anti-tuberculosis drugs; Hepatotoxicity; Drug-induced liver injury; Acute liver failure

Core Tip: Tuberculosis (TB) remains a significant public health concern in developing countries. First-line anti-TB drugs (ATDs) pose a risk of drug-induced liver injury (DILI), which can progress to acute liver failure (ALF), a life-threatening condition. The global incidence of ATD-induced DILI is approximately 11.5%, with around 7% of cases progressing to ALF within two months of treatment initiation. Notably, young female patients are disproportionately affected by this condition, which is clinically indistinguishable from other causes of ALF and is associated with a high mortality rate, necessitating emergency liver transplantation. This minireview provides a comprehensive overview of ATD-induced ALF, covering epidemiology, risk factors, and treatment options.

INTRODUCTION

Tuberculosis (TB) remains a leading cause of death from infectious diseases globally. In 2021, there were 10.6 million new TB cases[1]. Southeast Asia and Africa bear 68% of the global TB burden, with India accounting for the largest share of TB cases worldwide at 28[2,3]. The primary treatment for TB involves first-line anti-TB drugs (ATDs), including rifampicin (RIF), isoniazid (INH), Ethambutol (EMB), and pyrazinamide (PYZ). However, three of these four key drugs—INH, RIF, and PYZ—carry a risk of hepatotoxicity, which escalates when used in combination, making hepatotoxicity the most significant adverse effect of TB treatment. The reported incidence of ATD-induced liver injury (AT-DILI) varies widely, ranging from 2% to 28%[4-7]. This significant variation in prevalence may be attributed to differences in cohort characteristics, diagnostic criteria, ATD regimens, nutritional status, co-morbidities, and genetic predisposition. The incidence of AT-DILI also varies geographically. In a recent systematic review and meta-analysis, the pooled incidence of AT-DILI varied from 1.13% in Italy to 35.07% in Uganda[8]. Meta-analysis indicates a lower risk of DILI with a thrice-weekly ATD regimen (3.5%) compared to a daily regimen (16.5%), although intermittent regimens are no longer recommended[4].

Although most patients with AT-DILI exhibit either asymptomatic elevation of serum transaminases or symptomatic hepatitis, a small proportion may progress to acute liver failure (ALF), a life-threatening condition. Historically, two cases of fatal ALF caused by INH were documented in the 1970s[9]. Since then, multiple cases of ALF attributed to one or more ATDs have been documented[10-13]. Notably, ATDs are the leading cause of drug-induced ALF in India, whereas acetaminophen is the most common cause in the Western world. As ALF is a devastating and often fatal condition without liver transplantation (LT), timely recognition of ATD-induced ALF (AT-ALF) is crucial[14]. Given the paucity and patchiness of data on AT-ALF in existing literature, our study aims to provide a comprehensive and updated overview of this rare but significant condition. Furthermore, we highlight the knowledge gaps in this area, to stimulate further research and advance our understanding of AT-ALF.

DEFINITION OF AT-ALF

The most crucial requirement for defining AT-ALF is the occurrence of ALF after the initiation of ATDs. However, there is currently no consensus on the precise definition of ALF. Trey and Davidson made the initial recognition of ALF as fulminant hepatic failure in the early 1970s, defining it as a severe but potentially reversible liver injury with the onset of hepatic encephalopathy (HE) within eight weeks of the first symptoms in the absence of pre-existing liver disease[15]. Since then, numerous revised definitions have been proposed, but no agreement has been reached. A systematic review revealed 41 different definitions of ALF used in 81 studies[16]. Presently, the most recognized definition of ALF is the occurrence of severe acute liver injury leading to HE (any grade) and coagulopathy (INR of 1.5 or greater) in a patient with a period of illness of < 26 weeks and no pre-existing liver disease[17]. Nevertheless, the Indian subcontinent still uses the icterus-encephalopathy period of less than four weeks to define ALF[18].

Similarly, there is no consensus-based definition for AT-ALF. In a large study by Kumar et al[19], AT-ALF was diagnosed in patients with ALF who had a history of taking at least two of the three first-line hepatotoxic medications for a week, with no other known cause of severe acute liver injury. However, since INH monotherapy has been shown to produce ALF, consuming two hepatotoxic ATDs is not necessary[9-12]. Moreover, AT-ALF has been reported as early as three days[11] and five days[10] after starting ATDs, indicating that a minimum treatment duration of one week should not be a prerequisite for diagnosis. Furthermore, diffuse hepatic involvement from TB might also result in ALF, posing a diagnostic conundrum[20]. Although hepatic involvement in patients with active TB is rare (approximately 1%), it has been reported in as many as 50%-80% of patients with disseminated TB from endemic areas[21]. Thus, there is an unmet need for AT-ALF to be further refined and defined in a way that is robust, inclusive, and universally acceptable. Until a broader consensus is reached, we propose diagnosing AT-ALF when liver failure, defined by any grade of HE and coagulopathy (INR ≥ 1.5), occurs after starting one or more hepatotoxic ATDs in patients with normal baseline liver function and the absence of alternative causes.

MAGNITUDE OF AT-ALF

The proportion of AT-ALF cases shows considerable variability, mirroring the variations in the reported incidence of AT-DILI[4,22-24] (Table 1). In a recent meta-analysis of cohort studies from India, approximately 7% of AT-DILI cases progressed to ALF[4]. Similarly, in a large cohort study from China, only 5.2% of AT-DILI patients progressed to ALF, while most patients with transaminitis (> 3 times) resolved without serious consequences[25]. Consequently, the anticipated risk of AT-ALF would be 0.58% or lower when extrapolating the 12% incidence of AT-DILI in TB patients taking first-line ATDs. In contrast, a study by Devarbhavi et al[26] in India reported ALF in 25.7% (69/269) of AT-DILI patients, and Wang et al[27] in China observed ALF in 35.4% (55/155) of AT-DILI patients. Such large variations may be attributed to the characteristics of the selected cohort, varying diagnostic criteria for AT-ALF, drug regimens, and other co-factors. In a large study involving 1223 ALF patients from India, 5.7% (n = 70) of cases were due to AT-DILI[19]. A French study found AT-DILI responsible for 2.8% of ALF cases (566 patients) between 1986 and 2008[28]. In a retrospective study of patients from the United Network for Organ Sharing registry, ATDs accounted for 8% (50/661) of drug-induced ALF and 0.07% of total LT due to ALF (50/73,977). The majority (48, 96%) of AT-ALF patients were due to INH alone, with only 2 receiving a combination of INH, PYZ, and RIF[29]. Although the proportion of ALF cases attributed to ATDs is relatively low, it remains a significant concern in South Asia and Africa due to the high prevalence of TB and large population sizes.

Table 1 Magnitude of anti-tuberculosis drug-induced acute liver failure in various published studies.

Ref.	Study design	n (diagnosis)	n (AT-DILI)	n (AT-ALF)	Proportion of AT-ALF
Anand et al[44], 2006	Observational	152 (TB)	22 (13.8%)	09	13% of prospective cohort (6/69)
Pande et al[22], 1996	Prospective, case-control	492 (TB)	86	05	5.8% of AT-DILI
Devarbhavi et al[26], 2013	Retro-prospective cohort	269 (AT-DILI)	269	69	25.7% of AT-DILI
Dawra et al[24], 2029	Retrospective cohort	141 (TB)	10	01	10% of AT-DILI
Latief et al[23], 2017	Prospective cohort	200 (TB)	16	01	6.2% of AT-DILI
Raj Mani et al[13], 2021	Prospective cohort	397 (TB)	38	04	10.5% of AT-DILI
Kumar et al[19], 2010	Retrospective analysis	1223 (ALF)	-	70	5.7% of ALF
Wang et al[27], 2020	Retrospective analysis	155 (AT-DILI)	155	55	35.4% of AT-DILI
Ichai et al[28], 2010	Retrospective analysis	566 (ALF)	-	14	2.47% of ALF
Mindikoglu et al[29], 2009	Retrospective analysis	661 (drug-induced ALF)	-	50	7.5% of drug-induced ALF

TB: Tuberculosis; ALF: Acute liver failure; AT-ALF: Anti-tuberculosis drug-induced acute liver failure; AT-DILI: Anti-tuberculosis drug-induced liver injury.

PATHOGENESIS AND RISK FACTORS

The liver is crucial in metabolizing and detoxifying drugs, making it vulnerable to injury. Since most ATDs are lipophilic and need to be bio-transformed into more water-soluble compounds for elimination, toxic metabolites are probably responsible for AT-DILI and subsequent effects. However, the precise mechanisms underlying AT-DILI remain incompletely understood. The pathogenic mechanisms underlying early-onset and late-onset ALF are likely to differ. Early-onset ALF may be attributed to drugs such as INH, with or without RIF, which causes idiosyncratic DILI. The metabolism of INH by N-acetyl transferase (NAT2) produces metabolites like acetyl diazine and reactive acetyl free radicals, which contribute to liver injury[30,31]. Additionally, INH can be directly converted into INH hydrazine, leading to hepatocellular damage. Notably, co-administration with RIF can significantly amplify this pathway[32]. PZA is metabolized to 5-hydroxy pyrazinoic acid and pyrazinoic acid, both of which contribute to hepatotoxicity[33]. A recent study has shed light on the role of ferroptosis and lipid peroxidation in AT-DILI. This novel form of cell death is triggered by iron-dependent lipid peroxidation and regulated by the HIF-1α/SLC7A11/GPx4 pathway. Interestingly, glutathione replenishment can prevent ferroptosis, while iron supplementation can potentially worsen it[34,35]. Figure 1 depicts the possible mechanisms leading to AT-ALF based on existing knowledge.

Figure 1 Depicts the possible mechanisms leading to anti-tuberculosis drug-induced acute liver failure. The precise mechanisms underlying anti-tuberculosis drug-induced liver injury remain incompletely understood. The metabolism of isoniazid (INH) by NAT2 produces metabolites like acetyl diazine and reactive acetyl free radicals, which cause liver injury. Additionally, INH can be directly converted into INH hydrazine, causing hepatocellular damage. Notably, co-administration with rifampicin can significantly amplify this pathway. Pyrazinamide is metabolized to 5-hydroxy pyrazinoic acid and pyrazinoic acid, which are responsible for the hepatotoxicity. RIF: Rifampicin; INH: Isoniazid; PYZ: Pyrazinamide; ER: Endoplasmic reticulum; ROS: Reactive oxygen species; AT-DILI: Anti-tuberculosis drug-induced liver injury; AT-ALF: Anti-tuberculosis drug induced acute liver failure.

Risk factors and prediction of AT-ALF

Although many risk factors for AT-DILI have been identified, those specific to AT-ALF remain unclear. Key risk factors for AT-DILI include older age, female gender, poor nutritional status, hypoalbuminemia, concomitant alcohol consumption or viral hepatitis, non-alcoholic fatty liver disease, and some genetic polymorphisms—including NAT2, cytochrome P450 2E1, and glutathione S-transferases mu 1[4,36,37]. Recent studies have linked hypoxia-inducible factor-1α gene polymorphism and CpG island methylation to an increased risk of AT-DILI[38]. The concomitant presence of multiple risk factors for AT-DILI might put the patient at risk of developing ALF[4,39]. In a study from China, elevated serum bilirubin, aspartate aminotransferase (AST), white blood cell count, pre-existing hepatitis, and low platelet count were identified as independent predictors for the development of AT-ALF among AT-DILI patients. Moreover, patients with ALF experienced a considerably longer median time to DILI onset (51 days vs 24 days) than those without ALF. However, selection bias, retrospective design, and the lack of a standardized definition of AT-ALF continue to limit its applicability in real-world situations[27]. Once AT-DILI occurs, predicting which patients will develop ALF is challenging. Identifying patients at high risk for ALF could enable closer monitoring, hospitalization, or timely referral to a transplant center. Hy's Law, which indicates the presence of jaundice (total bilirubin > 2 times) with hepatocellular injury [alanine aminotransferase (ALT) > 3 times], can be used to predict progression to ALF[40]. Robles-Diaz et al[41] have proposed a modified Hy's Law that includes total bilirubin > 2 mg/dL and (ALT or AST, whichever is highest/upper limit normal (ULN)/(alkaline phosphatase/ULN) of 5 or above to predict ALF in the DILI cohort. This law offers a better balance between sensitivity (90%) and specificity (63%). Furthermore, a recent prospective cohort study identified malnutrition and anemia as significant independent risk factors for moderate to severe AT-DILI; however, ALF was not specifically mentioned[42]. According to a metabolomic study, alterations in fatty acid and bile acid metabolism play a crucial role in the progression of liver injury in patients with AT-DILI[43]. Continuation of ATDs after the development of jaundice increases the risk of progression to ALF[44,45]. In a study by Agal et al[45], among patients who directly presented with AT-DILI, 75% were icteric, and 12.5% had ALF. In contrast, among patients who developed AT-DILI in a prospective cohort, only one-third had mild symptoms, and none progressed to ALF.

CLINICAL CHARACTERISTICS

AT-ALF typically manifests within 2 months of starting therapy, though it can occur up to one year[12,19,27,46-48]. In the largest cohort of AT-ALF patients, the median duration between the initiation of ATDs and ALF was 30 days. Furthermore, the interval between the onset of icterus and the development of encephalopathy was less than seven days, indicating hyper-acute progression to liver failure[19]. Demographic analysis of AT-ALF reveals a predilection for younger individuals and females. In a study by Kumar et al[19], the mean age of patients was 32.87 years, with 70% of patients being under 35 years. Additionally, females comprised 70% of the patient population. Similarly, a study by Martino et al[49] found that all eight AT-ALF patients were female, with a median age of 39 years. Interestingly, the female preponderance in ALF is not unique to AT-ALF, as evidenced by a US-based ALF study. Prospective data collected from 23 US sites between 1998 and 2007 revealed that 67% of 1,147 ALF patients were women, regardless of the etiology[50]. Women are also more susceptible to AT-DILI. This sex disparity in DILI susceptibility has been attributed to various factors, including pharmacokinetic and pharmacodynamic differences, hormonal influences, and interactions with immunomodulating agents or signalling molecules[51]. Nevertheless, studies by Wang et al[27] and Huh et al[52] did not observe significant age or sex differences between AT-DILI patients, indicating that further clarification is needed regarding age and gender susceptibility for AT-ALF. The clinical presentation of AT-ALF is similar to that of ALF caused by other etiologies. The proportion of advanced HE, cerebral edema, seizures, gastrointestinal bleeding, and acute kidney injury has not been reported to differ significantly between patients with AT-ALF and ALF resulting from other etiologies[19]. Nonetheless, studies have shown that AT-ALF patients exhibit significantly lower transaminase elevations compared to ALF due to other causes[19].

DIAGNOSTIC CHALLENGES

The AT-ALF diagnosis is primarily clinical, relying on the temporal association between drug exposure and the onset of liver injury, as well as the exclusion of alternative causes. In clinical practice, diagnosing AT-ALF may pose significant challenges due to the absence of specific biomarkers and a universally accepted definition. The variable latency period between drug exposure and the onset of liver injury complicates the establishment of a definitive temporal association, thereby hindering accurate diagnosis. Routine use of liver biopsy is not advocated due to its invasive nature and associated risks. Furthermore, histopathological examination lacks specificity for AT-ALF, as evidenced by Kumar et al[19]. In this study, post-mortem liver biopsies from 38 AT-ALF patients revealed non-specific features of acute hepatitis along with sub-massive to massive hepatocyte necrosis, which were indistinguishable from ALF caused by other etiologies. Therefore, the decision to perform a liver biopsy should be guided by clinical discretion, and this invasive procedure should be reserved for instances where the ALF diagnosis remains uncertain and additional diagnostic clarity is essential. Although the drug-induced lymphocyte stimulation test can aid in diagnosing DILI secondary to ATDs, its sensitivity is limited[53]. Moreover, its role in diagnosing AT-ALF remains undefined. The Roussel Uclaf causality assessment method is a widely utilized tool for assessing causality in DILI. However, this method has inherent limitations, including challenges in discerning the hepatotoxic effects of multiple concurrently administered drugs and a degree of subjectivity that can lead to inter- and intra-observer variability[54].

The hepatotoxic potential of three out of four first-line ATDs complicates the identification of a single culpable agent in ALF. INH is the most frequently implicated ATD in ALF when administered as monotherapy[9-12]. RIF is known to potentiate INH-induced hepatotoxicity, although it has not been identified as a sole causative agent of ALF[55-57]. PYZ-associated liver injury exhibits dose-dependent toxicity, with doses exceeding 40 mg/kg linked to fatal hepatitis[55]. Moreover, co-administration of PYZ has been associated with an increased mortality risk in patients developing ALF[58].

OUTCOMES AND PREDICTORS OF MORTALITY

Non-acetaminophen DILI-associated ALF generally has a poor prognosis. A recent study by the US ALF study group reported a transplant-free survival rate of only 31% for DILI-ALF patients[59]. Similarly, AT-ALF is associated with high mortality rates without LT[4,19]. Prospective studies have reported mortality rates ranging from 67% to 85.7%[19,52,60]. A meta-analysis of eight studies yielded a pooled mortality rate of 71.8% for AT-ALF patients, significantly higher than the 1.72% overall mortality rate observed for AT-DILI patients[4]. The median time of death from day of hospitalization varies from 3 to 15 days in AT-ALF patients[19,52]. In the study by Kumar et al[19] nearly half of the patients died within five days of hospitalization, emphasizing the importance of prompt consideration of LT options. While AT-ALF patients in India and the United States have reported high mortality rates, a recent Chinese study by Wang et al[27] found a significantly lower mortality rate of 9.68%. The authors asserted that the use of a non-bioartificial liver support system contributed to this improved outcome. However, this disparity raises concerns about the diagnostic accuracy of ALF in these patients.

The selection of suitable liver transplant (LT) candidates among ALF patients is often hindered by a lack of robust prognostic models. Although various models have been proposed to predict ALF outcomes, they fall short in terms of accuracy and early applicability, thereby limiting their utility in clinical decision-making. Kumar et al[19] identified serum bilirubin (> 10.8 mg/dL), prothrombin time prolongation (> 26 seconds), and grade III/IV HE at presentation as independent predictors of mortality in AT-ALF patients. The presence of either of these predictors demonstrated good sensitivity (81%) and fair specificity (72%) for predicting mortality. However, further validation studies are necessary to confirm the generalizability of these results. Conversely, commonly used prognostic models, including the King's College Hospital criteria and Model for End-Stage Liver Disease score, exhibited lower sensitivity and specificity in AT-ALF, with values of 34.04% and 74%, and 73% and 67%, respectively[19]. A study by Durand et al[61] identified several predictors of mortality without LT in AT-ALF patients. These included PYZ co-administration, an ATD-jaundice interval > 15 days, grade III encephalopathy, and factor V levels below 20%. However, further validation is necessary to confirm the reliability of these predictors. Additional research is required to develop more accurate prognostic models for AT-ALF patients and to investigate the impact of underlying TB on patient outcomes. Until a reliable prognostic model is developed, each AT-ALF case should be considered individually when making the crucial transplant decision. An additional concern pertains to the optimal timing of LT in AT-ALF patients due to the dynamic nature of ALF. Delayed decision-making may cause forfeited opportunities for transplantation, whereas premature decisions may cause unnecessary LT. Longer pre-transplant waiting times (exceeding five days) have been associated with increased mortality rates[62]. Therefore, the window of opportunity for LT in ALF may be limited. Early application of prognostic models for listing and expedited donor evaluation are crucial to achieve better outcomes.

LT IN AT-ALF

Martino et al[49] reported a survival rate of 50% at one year in 7 patients with AT-ALF who underwent LT. Two patients died one and seven days post-LT, while one patient passed away two months after the procedure. The low survival rate was attributed to the critical condition of patients before LT, requiring ventilatory and circulatory support. Ichai et al[28] found a similar post-LT survival rate (50%) in a study that included six AT-ALF patients who had LT. In contrast, several single case reports have described favorable outcomes of LT in AT-ALF patients[11,63-66]. However, these survival rates are likely to be overstated because authors are typically less inclined to report instances with unfavorable outcomes. Huh et al[52] reported excellent outcomes of emergency adult living donor LT (LDLT) in six patients with AT-ALF. These six patients were chosen from a pool of 19 patients with ALF, and considering the hazards to the donor, they may have had better clinical parameters prior to LDLT. In another recent study, seven AT-ALF patients undergoing LDLT in India had a good survival rate of 71.4% with a median follow-up of 22 months[67]. Overall, the post-LT survival rate for ALF patients remains around 10% lower than that of other transplanted non-ALF patients[62]. However, AT-ALF patients do not have a lower post-LT survival rate than those with other ALF etiologies. Table 2 depicts spontaneous as well as post-LT survival rates of AT-ALF patients in various published studies.

Table 2 Outcomes of patients with anti-tuberculosis drug-induced acute liver failure.

Ref.	AT-ALF (n)	Demographic features	ATD-ALF interval	LT, n (%)	Post-LT survival (%)	Mortality (%)
Ichai et al[28], 2010	14	04 (28.5%) male, 10 (71.5%) female, median age – 37 years	3.5 ± 3.1 months	6 (42.8)	50	28.5
Kumar et al[19], 2010	70	21 (30%) male, 49 (70%) female, mean age 328 years	30 (7-350) days	None	-	67.1
Idilman et al[11], 2006	1	Female, 19 years	3 days	1 (100)	100	None
Mindikoglu et al[29], 2009	50	32 (64%) female, 18 (36%) male	-	50 (100)	82	18
Bavikatte et al[67], 2017	7	03 (42.8%) male, 04 (57.2%) female, median age – 32 (5-39) years	2 months	7 (100)	71.4	28
Huh et al[52], 2017	19	10 (52.6%) male, 9 (47.4%) female, mean age – 46 years	61 (40-113) days	6 (31.5)	100	52.6
Li et al[39], 2018	1	Male, 67 years	77 days	1 (100)	0	100
Martino et al[49], 2018	8	08 (100%) female, median age – 39 (17-56) years	21 to 155 days	8 (100)	50	50
Bartoletti et al[63], 2017	26	14 (54%) male, 12 (46%) female, median age – 38 (25-50) years	-	26 (100)	85	15
Wang et al[27], 2020	55	34 (61.8%) male, 21 (39.2%) female, mean age 50.3 ± 15.9 years	51 (7-330 days)	None	-	9.68
Smink et al[65], 2006	2	01 male and 01 female	139.5 days	2 (100)	2 (100)	None
Zhu et al[66], 2021	1	Female, 26 years	3 days	1 (100)	100	None
Farrell et al[46], 1994	2	49 years male and 60 years female	4 months and 6 weeks	2 (100)	100	None
Barcena et al[47], 2005	1	Male, 39 years	1 month	1 (100)	100	None
Pessayre et al[12], 1977	6	01 (17%) male, 05 (83%) female, mean age – 46 ± 24 years	7.3 ± 1.8 days	0	-	None
Mitchell et al[48], 1995	4	01 (25%) male, 03 (75%) female, mean age – 51.75 ± 13.98	3.5 ± 2.3 months	2 (50)	50	50
Cillo et al[64], 2005	1	Female, 10 years	3 months	1 (100)	100	None
Campos et al[10], 2004	1	Female,16 years	5 days	0	-	None

LT: Liver transplantation; ALF: Acute liver failure; AT-ALF: Anti-tuberculosis drug-induced acute liver failure; ATD: Anti-tuberculosis drug.

ADJUNCT THERAPIES

In ALF, adjunctive therapies provide supportive care for liver function and manage associated complications. These therapies encompass extracorporeal treatments, such as therapeutic plasma exchange (TPE), high-flow continuous hemodiafiltration, albumin hemodialysis, and pharmacological interventions, including N-acetylcysteine (NAC). Extracorporeal therapies have been employed in ALF management to remove toxins and inflammatory mediators, thereby supporting liver function and potentially improving outcomes. These interventions may provide the native liver an opportunity to rest and recover while also serving as a bridge to LT. A meta-analysis demonstrated significant survival benefits with TPE in ALF patients, with a 41% improvement in 30-day survival and 35% improvement in overall survival[68]. However, a recent multicenter cohort study from the United Kingdom failed to replicate these findings, showing no significant difference in survival between TPE and standard medical therapy (51.4% vs 62.6%, P = 0.12)[69]. For AT-ALF, the evidence is limited to case reports, which associate successful treatment outcomes with high-volume TPE[70]. Nevertheless, large-scale studies are needed to confirm these findings. Furthermore, knowledge on standardized TPE protocols, timing of initiation, and predictors of response is still inadequate. Albumin dialysis is an extracorporeal treatment strategy for the removal of protein-bound drugs, bilirubin, and bile acids[71]. While this modality may improve HE by removing inducers such as ammonia and tryptophan, its survival benefit is not well established. Moreover, specific data on the role of albumin dialysis or high-flow continuous hemodiafiltration in AT-ALF patients are lacking, highlighting the need for further research.

NAC is commonly employed as adjunctive treatment for ALF, yet its efficacy in non-acetaminophen-induced ALF remains poorly understood. While a randomized controlled trial (RCT) found no significant mortality benefit with intravenous NAC in hospitalized AT-DILI patients, a meta-analysis suggested potential benefits of NAC treatment in patients with non-acetaminophen drug-induced ALF[72,73]. However, methodological limitations inherent to the included studies preclude definitive conclusions, underscoring the need for further research to elucidate the therapeutic potential of NAC in this context. Moreover, corticosteroids are not a standard treatment for AT-ALF due to the lack of robust evidence supporting their efficacy. A retrospective analysis of drug-induced ALF patients revealed no significant improvement in overall survival with steroid treatment[74].

CHALLENGES IN TREATMENT OF TB POST-LT

There is a higher risk of graft failure, rejection, and death when active TB is present in post-LT patients[75]. Moreover, immunosuppressants increase the risk of extrapulmonary and disseminated TB, necessitating the institution of ATDs. There are many challenges in the treatment of TB post-LT. Due to the graft's function and interactions with immunosuppressive medications, ATD re-challenge carries an increased risk of DILI[76]. The choice of ATD regimen early after LT also remains controversial[28]. Therefore, it is necessary to balance the risk of graft rejection against the advantages of administering RIF. RIF, and to a lesser extent, INH increase the metabolism of immunosuppressive drugs, such as corticosteroids, tacrolimus, and cyclosporine, by inducing the cytochrome P450 pathway. As a result, solid organ transplant recipients receiving RIF-based regimens may face an increased risk of organ rejection and DILI. Nevertheless, RIF-containing regimens have been advocated for solid organ transplant recipients who have severe or disseminated TB. RIF-sparing therapy may be used to treat patients with non-severe forms of TB or those who are at high risk of graft rejection in the early post-transplant phases. Furthermore, RIF can be replaced by rifabutin to lower the risk of drug interactions. The doses of cyclosporine, tacrolimus, and sirolimus may be increased three to five times, and corticosteroids by 50% when a RIF-containing regimen is employed[75]. Use of non-hepatotoxic ATDs (EMB, streptomycin, ciprofloxacin, and cycloserine) has been suggested to control active TB in these patients; however, their effectiveness remains unproven.

PREVENTIVE STRATEGIES

Currently, there are no recommended medications for the prevention of AT-DILIs. Given that the pathogenesis of AT-DILI involves the increased production of free radicals and toxic metabolites, antioxidant and anti-inflammatory agents may hold potential as prophylactic measures[5]. Some studies suggest that oral NAC and ursodeoxycholic acid may help prevent AT-DILI; however, the evidence is currently insufficient to support this claim[76-79]. Similarly, hepatoprotective agents like silymarin and glycyrrhetinic acid have failed to demonstrate a significant reduction in the risk of AT-DILI in a large study from China[80]. A recent systematic review and network meta-analysis evaluated the efficacy and safety of herbal/alternative medicines in preventing AT-DILI. The analysis included 3,423 patients from 14 RCTs. The results showed that supplementation with turmeric plus Tinospora cordifolia (RR 0.07) and NA (RR 0.09) significantly reduced AT-DILI incidence compared to the placebo[81]. Nevertheless, these findings are based on a limited number of low-quality studies, highlighting the need for further research to confirm these results.

Continuation of ATDs following the development of DILI significantly increases the risk of progression to ALF[44,45]. This underscores the importance of regular monitoring of liver function tests (LFTs) after initiating ATD therapy to facilitate early detection of DILI and prompt withdrawal of the offending agent. However, despite timely discontinuation, progression to ALF may still occur[39]. Furthermore, the limited understanding of hepatic adaptation to ATDs poses a challenge in determining the threshold of LFT derangement at which ATDs should be discontinued, thus underscoring the need for further research in this area.

CONCLUSION

AT-ALF is a severe and potentially fatal complication of TB treatment, particularly in regions with high TB burdens. INH is the most implicated drug among ATDs. Although the incidence of AT-ALF appears to be much lower (< 0.58%) than that of AT-DILI (approximately 12%), it carries a high mortality rate, often requiring LT for survival (Figure 2). The pathogenesis of AT-ALF is complex, involving interactions between drug metabolism, genetic susceptibility, and immune responses. Early detection and timely cessation of hepatotoxic drugs may be crucial for better outcomes. Even though about one-third of AT-ALF patients can survive with medical care, treatment allocation is made more difficult by the lack of reliable prognostic models. LT remains the definitive treatment, with outcomes comparable to other causes of ALF. However, LT poses logistical and clinical challenges, particularly in resource-limited countries where TB is prevalent. Further advancements in the management of AT-ALF necessitate the establishment of standardized diagnostic criteria, refinement of prognostic models, and exploration of hepatoprotective therapeutic strategies. Addressing these knowledge gaps will ultimately enhance patient outcomes, optimize anti-TB treatment regimens, and mitigate the hepatotoxicity-associated burden of TB treatment.

Figure 2 Diagrammatic representation of the magnitude of anti-tuberculosis drug-induced liver injury, its progression to anti-tuberculosis drug-induced acute liver failure, and subsequent patients’ outcomes. RIF: Rifampicin; INH: Isoniazid; PYZ: Pyrazinamide; ATD: Anti-tuberculosis drug; AT-DILI: Anti-tuberculosis drug-induced liver injury; AT-ALF: Anti-tuberculosis drug induced acute liver failure; LT: Liver transplantation; HE: Hepatic encephalopathy.