Drug-induced hepatotoxicity (DIH) constitutes a clinically relevant complication in the treatment of Human Immunodeficiency Virus (HIV) infection and tuberculosis (TB). Both conditions require prolonged and combined therapeutic regimens, which increase exposure to multiple drugs and consequently the risk of liver injury [1, 2]. The liver, as the central organ responsible for the biotransformation of antiretroviral and antituberculosis drugs, is particularly vulnerable to injury. Hepatic damage may manifest across a wide clinical spectrum, ranging from asymptomatic elevations of liver enzymes to severe forms of hepatic failure, complicating timely recognition and appropriate management [3, 4].

In the context of HIV–TB coinfection, this risk is further amplified by polypharmacy, drug–drug interactions, and the presence of comorbidities [5]. In this setting, pharmacogenetics has emerged as a valuable tool to explain interindividual variability in treatment response and susceptibility to liver injury. However, its integration with systematic clinical and biochemical assessment of hepatotoxicity remains limited. This gap underscores the need for approaches that consider both components in a complementary manner [1].

A literature search was conducted in PubMed, Embase, and Google Scholar to identify relevant publications addressing hepatotoxicity, HIV, tuberculosis, and pharmacogenetics [6]. Priority was given to studies published within the past six years in order to capture the most recent advances in pharmacogenetics and drug-induced hepatotoxicity research.

Eligible publications included clinical studies, observational studies, systematic reviews, meta-analyses, and clinical practice guidelines that provided relevant clinical or mechanistic evidence related to hepatotoxicity in HIV and tuberculosis treatment. Studies not directly related to hepatotoxicity, pharmacogenetics, HIV, or tuberculosis, as well as articles lacking clinical relevance, were excluded. Seminal earlier references were included when necessary to provide historical or conceptual context.

The objective of this narrative review is to analyze the relevance of hepatotoxicity assessment in the context of HIV and tuberculosis pharmacogenetics, integrating the pathophysiological, clinical, and genetic aspects of this condition, and to discuss its implications for clinical practice and the development of personalized medicine strategies.

DIH results from a complex interaction between hepatic drug metabolism and host cellular and immunological responses (Fig. 1). Many medications require hepatic biotransformation for elimination through phase I and/or phase II metabolic pathways, processes that may generate reactive metabolites capable of inducing cellular stress, mitochondrial dysfunction, and direct hepatocellular injury [7]. These events can activate cell death pathways, such as apoptosis or necrosis, and trigger an inflammatory response that amplifies liver damage [2].

In addition to direct toxicity, immune-mediated mechanisms play a key role in DIH. In these cases, drug metabolites may bind to hepatic proteins, forming neoantigens that activate the adaptive immune response. T-lymphocyte activation and the release of proinflammatory cytokines contribute to the progression of liver injury, helping to explain the marked interindividual variability in clinical presentation and disease duration [3].

At the molecular level, genetic factors significantly influence susceptibility to hepatotoxicity. Variants in genes encoding drug-metabolizing enzymes and proteins involved in oxidative stress, endoplasmic reticulum stress responses, and mitochondrial function may modify the liver’s capacity to detoxify drugs [9]. This provides a mechanistic basis for the relevance of pharmacogenetics in identifying individuals at increased risk of hepatotoxicity, particularly in prolonged and combined treatment regimens, such as those used for HIV and tuberculosis.

At the cellular level, aminotransferases, particularly ALT and aspartate aminotransferase (AST), are enzymes predominantly localized in the cytosol and, in the case of AST, also in the mitochondria of hepatocytes. During drug-induced hepatotoxicity, cellular injury mechanisms (including mitochondrial dysfunction, oxidative stress, and activation of cell death pathways) disrupt hepatocellular membrane integrity, facilitating the release of these enzymes into the systemic circulation [7]. Consequently, elevated serum aminotransferase levels serve as an indirect reflection of hepatocellular damage. In particular, predominant ALT elevation is closely associated with the hepatocellular pattern of injury, given its greater hepatic specificity, which explains its widespread use as a key biochemical marker for the identification and characterization of DIH [13].

DILI is a specific diagnostic entity that encompasses liver injury attributable to the use of medications administered at therapeutic doses and is considered a diagnosis of exclusion [2]. Its identification requires the systematic exclusion of other causes of liver injury, such as viral hepatitis, autoimmune diseases, metabolic liver disorders, or biliary obstruction. DILI may occur after variable drug exposure and exhibits marked clinical and biochemical heterogeneity, which complicates its timely recognition in clinical practice [3]. In this sense, DILI represents a clinically defined form of drug-induced hepatotoxicity, characterized by the application of more stringent biochemical, temporal, and causality criteria. Therefore, while all cases of DILI involve drug-induced hepatotoxicity, not all manifestations of drug-induced hepatotoxicity meet the clinical and biochemical criteria required to be classified as DILI [1].

From a clinical perspective, DILI is classified as intrinsic or idiosyncratic [2]. The intrinsic form is generally predictable, dose-dependent, and characterized by a short latency period, whereas the idiosyncratic form is unpredictable, not dose-dependent, and exhibits variable latency. The latter is the most common in clinical settings and is influenced by individual factors, including genetic determinants and immune-mediated responses, which explains the marked interindividual variability in its presentation and clinical course [9].

In the context of prolonged and combined treatment regimens, such as those used for HIV and tuberculosis, the identification of cases fulfilling DILI criteria is of particular clinical relevance due to the difficulty in determining the causative drug and the therapeutic implications of discontinuing or modifying essential treatment regimens [5]. In these scenarios, DILI represents a diagnostic and therapeutic challenge that requires systematic and standardized evaluation to minimize the risk of progression to severe liver injury and to optimize treatment continuity.

DIH is determined by the presence of liver injury associated with exposure to one or more drugs, identified through biochemical abnormalities, without the need for formal establishment of causality [1]. Unlike DILI, which requires attribution of liver injury to a specific drug, hepatotoxicity may occur across multiple clinical contexts [14, 15]. In clinical studies, hepatotoxicity has historically been defined using operational biochemical criteria. In this regard, the Division of Acquired Immunodeficiency Syndrome (DAIDS) of the National Institutes of Health (NIH) developed standardized criteria for the identification and classification of hepatotoxicity in clinical trials, with the aim of harmonizing the assessment of liver injury severity [16].

Initial evaluation includes measurement of ALT, AST, ALP, gamma-glutamyl transferase (GGT), and bilirubin levels in order to detect and characterize liver injury at an early stage [13, 15]. In populations exposed to polypharmacy, substance use, comorbidities, and coinfections it is often difficult or impossible to establish causality (DILI); therefore, assessment of drug-induced hepatotoxicity becomes essential for the detection, classification, and prevention of liver injury [1].

According to DAIDS criteria, elevations exceeding 1.25 times the ULN constitute the threshold for defining hepatotoxicity, primarily using aminotransferases (particularly ALT) as initial markers of hepatocellular injury [1, 15, 16]. The severity of liver injury can be graded based on the magnitude of enzyme elevation, allowing classification into grade I (mild), grade II (moderate), grade III (severe), and grade IV (potentially life-threatening) toxicity [13, 16]. In addition, the R ratio enables classification of the pattern of liver injury as hepatocellular, cholestatic, or mixed, providing a standardized framework for biochemical interpretation of hepatic damage [13].

To establish a causal relationship between a suspected drug and liver injury, particularly in cases of clinically significant damage, the use of structured methods such as the Roussel Uclaf Causality Assessment Method (RUCAM) is recommended [2]. This scale, specifically developed for the evaluation of DILI, integrates clinical, biochemical, and temporal variables, including the course after drug withdrawal, exclusion of alternative causes, and prior evidence of hepatotoxicity associated with the specific drug [17]. Therefore, RUCAM constitutes a key tool for differentiating nonspecific biochemical elevations from cases that fulfill formal DILI criteria.

Hepatotoxicity in patients with HIV is a frequent and multifactorial event, associated both with the use of antiretroviral drugs and with host-related factors. Several antiretroviral agents may induce elevations in liver enzymes through direct or idiosyncratic mechanisms, particularly during the initial phases of treatment or following regimen changes [11]. In addition, factors such as HIV-associated chronic inflammation, the presence of viral hepatitis coinfections, alcohol consumption, and other metabolic comorbidities increase susceptibility to liver injury. In this context, hepatotoxicity may compromise adherence to and continuity of antiretroviral therapy, with clinically relevant implications for patient outcomes [18].

In tuberculosis, DIH related to antituberculosis therapy represents one of the most important adverse drug reactions. Isoniazid, rifampicin, and pyrazinamide are the main drugs implicated, and hepatotoxicity typically occurs during the first weeks or months of treatment [19]. The incidence and severity vary widely across populations and are influenced by factors such as age, nutritional status, alcohol consumption, and genetic predisposition. In particular, interindividual variability in isoniazid metabolism partially explains differences in the risk of developing hepatotoxicity during antituberculosis treatment [20].

Hepatotoxicity in patients with HIV–TB coinfection represents a clinical scenario of particular complexity due to simultaneous exposure to prolonged and potentially hepatotoxic treatment regimens, as well as to clinically relevant drug–drug interactions [5]. Rifampicin, a cornerstone of antituberculosis therapy, acts as a potent inducer of cytochrome P450 enzymes, including cytochrome P450 3A4 (CYP3A4) and cytochrome P450 2B6 (CYP2B6), which can alter the pharmacokinetics of several antiretroviral agents [21].

However, recent studies have demonstrated that enzyme induction does not always translate into increased antiretroviral clearance, as coadministration of isoniazid and genetic variability in drug-metabolizing enzymes may counterbalance these effects. The risk of hepatotoxicity in people living with HIV exposed to preventive or therapeutic tuberculosis regimens is influenced by both the antiretroviral regimen and patient-specific genetic factors. This suggests that drug metabolism in coinfected patients is determined by a complex interaction between pharmacological and genetic determinants, with direct implications for the monitoring and management of liver damage [22].

Drug-induced hepatotoxicity (DIH) exhibits marked interindividual variability that cannot be explained solely by clinical, environmental, or treatment-related factors, highlighting the central role of pharmacogenetics. Genetic variants in enzymes involved in hepatic biotransformation may modify drug metabolism rates, promote the accumulation of reactive metabolites, and impair detoxification mechanisms, thereby increasing susceptibility to hepatotoxicity [1, 23, 24]. In particular, cytochrome P450 (CYP450) enzymes play a fundamental role in phase I metabolism of a wide range of drugs used in clinical practice [25], and polymorphisms in genes encoding these enzymes have been shown to contribute significantly to variability in drug response and toxicity, especially in complex therapeutic regimens such as those used for HIV and tuberculosis [26]. Table 1 (Ref. [23, 26, 27, 28, 29, 30, 31]) summarizes the main pharmacogenetically relevant polymorphisms involved in HIV and tuberculosis treatment and their associated clinical effects.

In tuberculosis treatment, N-acetyltransferase 2 (NAT2) represents one of the most relevant pharmacogenetic determinants of hepatotoxicity (Fig. 2) [32]. Polymorphisms in the NAT2 gene define slow, intermediate, or rapid acetylator phenotypes, directly influencing isoniazid metabolism and, consequently, systemic exposure to the drug and its potentially hepatotoxic metabolites [33, 34]. Multiple studies have consistently identified that carriers of NAT2 variants associated with slow acetylation, particularly NAT2*5 (rs1801280, c.341T>C) and NAT2*6 (rs1799930, c.590G>A), have a higher risk of significant ALT elevations and clinically relevant hepatotoxicity during antituberculosis treatment [26, 27]. Accordingly, dose adjustment strategies based on patient-specific factors, including acetylator phenotype, have been proposed [35].

More recently, clinical evidence has expanded this paradigm by demonstrating that rapid acetylator phenotypes may also have adverse clinical implications [36]. In patients with tuberculosis treated with isoniazid-based regimens, rapid NAT2 acetylators have been associated with lower drug exposure and an increased risk of unfavorable outcomes, including higher one-year mortality, particularly in the context of isoniazid resistance (hazard ratio 1.7–4.6) [36]. Reduced plasma exposure to isoniazid in rapid acetylators, mostly in the presence of drug resistance, may contribute to higher mycobacterial burden and poorer clinical outcomes, including increased mortality.

In patients with HIV, and particularly in those with HIV–tuberculosis coinfection, pharmacogenetics acquires additional complexity due to the interaction of multiple metabolic pathways, enzyme induction and inhibition, and polypharmacy [5, 37, 38]. Among the most extensively studied determinants, cytochrome CYP2B6 stands out as a key regulator of exposure to efavirenz and nevirapine [39, 40]. Polymorphisms such as CYP2B6*6 (rs3745274, c.516G>T), especially when combined with other functional variants, have been consistently associated with elevated plasma efavirenz concentrations and an increased risk of toxicity, including hepatotoxicity [23, 29, 30]. Pharmacokinetic studies in coinfected populations have shown that slow metabolizers exhibit significantly higher efavirenz exposure, even during coadministration with rifampicin and isoniazid, due to inhibition of alternative metabolic pathways such as cytochrome P450 2A6 (CYP2A6) by isoniazid, which may partially counterbalance the enzyme induction exerted by rifampicin [30]. This places pharmacogenetic factors as key determinants of efavirenz exposure and toxicity in specific clinical scenarios [41].

For dolutegravir, plasma concentrations are influenced not only by rifampicin coadministration but also by polymorphisms in enzymes such as UDP-glucuronosyltransferase 1A1 (UGT1A1, rs3064744, 7 TA repeats) and arylacetamide deacetylase (AADAC, rs1803155, c.841G>A), which have been associated with increased drug exposure [31]. In addition, variants in regulatory genes involved in drug metabolism, such as NR1I2 (pregnane X receptor, PXR) rs2472677 (g.24087C>T) and hepatic transporter genes, including SLCO1B1 (solute carrier organic anion transporter family member 1B1) rs4149032 (g.38664C>T), have been shown to independently or jointly modulate exposure to antituberculosis and antiretroviral drugs [28]. In African cohorts with HIV–TB coinfection, PXR polymorphisms have been associated with poorer clinical outcomes, including increased mortality, suggesting that transcriptional regulation of hepatic enzymes and transporters plays a relevant role in interindividual variability beyond classical cytochrome pathways [28].

Taken together, this body of evidence supports that the risk of hepatotoxicity in people living with HIV exposed to isoniazid and antiretroviral therapy is determined by a complex interaction between therapeutic regimens and pharmacogenetic factors, with direct implications for clinical monitoring, risk stratification, and the development of personalized medicine strategies [22].

In HIV–TB coinfection, the pharmacogenetic phenotypes of highest clinical relevance are those that determine increased or decreased exposure to key drugs and, consequently, a higher risk of toxicity, therapeutic failure, or the need for dose adjustments [42, 43]. In tuberculosis, the NAT2 slow acetylator phenotype, derived from allelic combinations of NAT2*5, *6, *7, and *14, is associated with increased isoniazid exposure and a higher incidence of hepatotoxicity, making it a priority phenotype for risk stratification and early biochemical monitoring [27, 33, 43]. Conversely, rapid acetylators represent a subgroup at risk of underexposure and unfavorable clinical outcomes, reinforcing the need to consider pharmacogenetics not only as a tool for toxicity prevention but also for optimization of therapeutic efficacy [36].

Within the antiretroviral component, slow metabolizer phenotypes for efavirenz and nevirapine, primarily associated with variants such as CYP2B6*6 (rs3745274, c.516G>T) and *18 (rs28399499, c.983T>C) are linked to elevated plasma concentrations and an increased likelihood of toxicity, particularly in coinfection settings and during coadministration with antituberculosis drugs [26, 29]. Additional phenotypes derived from variants in regulatory genes (e.g., PXR) or transporter genes (e.g., SLCO1B1) may further modulate exposure to antiretrovirals or rifamycins, generating complex drug–gene–drug interaction scenarios that account for part of the variability observed in clinical practice [28, 44]. Taken together, identification of these pharmacogenetic phenotypes enables conceptualization of clinical heterogeneity in patients with HIV–TB and supports a pharmacovigilance and individualized dosing approach targeted toward subgroups with increased genetic susceptibility [1, 22].

DIH may lead to a wide spectrum of clinical complications, ranging from transient biochemical abnormalities to severe, life-threatening liver damage, including clinically overt acute hepatitis, prolonged cholestasis, hepatic insufficiency, and, in rare cases, fulminant liver failure [3]. In addition, the occurrence of hepatotoxicity often necessitates the discontinuation or modification of essential therapeutic regimens, which may compromise treatment efficacy, promote disease relapse, and increase the risk of adverse clinical outcomes. These complications are particularly relevant in settings of prolonged or combined therapy, where interruption of the implicated drug represents a significant therapeutic challenge and may be associated with a poorer clinical prognosis [2].

Drug-induced hepatotoxicity remains a clinically relevant complication in the management of patients with HIV, tuberculosis, and HIV–tuberculosis coinfection. Current evidence supports that systematic clinical and biochemical monitoring continues to represent the cornerstone of hepatotoxicity assessment, through periodic evaluation of liver enzymes and the application of standardized hepatotoxicity criteria. Within this framework, pharmacogenetics emerges as a highly valuable complementary tool, enabling the identification of patient subgroups with increased susceptibility to liver injury, particularly in relation to polymorphisms in genes such as NAT2 and CYP2B6. In certain situations, such as resource-limited settings, pharmacogenetic testing may not be routinely available; however, clinical and biochemical monitoring strategies based on liver enzyme assessment remain essential for clinicians in the early detection and management of hepatotoxicity. The integrated evaluation of these elements provides a more precise and preventive approach to the detection and management of hepatotoxicity, with the potential to optimize therapeutic safety and advance personalized medicine strategies in vulnerable populations.