Why Liver Injury Happens and How Science is Personalizing Medicine
Tuberculosis (TB) remains one of humanity's most persistent infectious diseases, ranking among the top causes of death worldwide from a single infectious agent.
People infected with TB globally in 2021, representing a 5% increase from the previous year 1
Prevalence range of anti-tuberculosis drug-induced liver injury (AT-DILI) across different populations
For decades, physicians observed that some patients developed severe liver injury while others on the same regimen experienced no problems. This variability puzzled scientists until they began looking beyond clinical factors to something more fundamental: our genetic blueprint.
Recent research has revealed that individual genetic differences, particularly in enzymes that metabolize anti-TB drugs, play a crucial role in determining who develops liver injury and who remains protected.
The most significant genetic discoveries in AT-DILI research center on the N-acetyltransferase 2 (NAT2) enzyme, responsible for processing isoniazid, a cornerstone of TB treatment.
While NAT2 represents the most prominent genetic risk factor, it's not the only player. The cytochrome P450 2E1 (CYP2E1) enzyme also contributes to isoniazid metabolism.
The meta-analysis revealed that individuals with the CYP2E1 C1/C1 genotype had increased AT-DILI risk 1 .
Other genes under investigation include those encoding glutathione-S-transferases (GSTs), which facilitate detoxification processes, and various hepatic transporter proteins.
| Genotype Category | Specific Genotypes | Association with AT-DILI |
|---|---|---|
| Slow Acetylators | NAT2*6A/6A, NAT2*6A/7B, NAT2*7B/7B, NAT2*5B/7B | Significantly increased risk |
| Rapid Acetylators | NAT2*4/4, NAT2*4/7B | Protective effect |
| No Significant Association | NAT2*4/*6A, NAT2*4/*5B, NAT2*5B/*5B | Neutral |
Based on a comprehensive 2025 systematic review and meta-analysis examining 10 studies and over 3,300 individuals of Asian ancestry 1 .
The distribution of genetic variants differs dramatically across populations, explaining why AT-DILI rates vary geographically.
Higher prevalence of hepatotoxic alleles in AGBL4, GSTP1, and SLCO1B1
Protective alleles in NOS gene
Intermediate risk profile
While genetic predisposition sets the stage, AT-DILI risk is multifactorial. Several clinical and demographic factors interact with genetic profile to determine overall risk.
| Risk Factor | Adjusted Odds Ratio (AOR) | Significance |
|---|---|---|
| HIV positive status | AOR = 6.73 | P = 0.005 |
| Female gender | AOR = 2.57 | P = 0.027 |
| Increasing age | AOR = 1.05 (per year) | P = 0.019 |
| Low body mass index | AOR = 0.81 | P = 0.009 |
Based on a 2025 prospective cohort study in central Ethiopia that followed 219 TB patients 2 .
The timing of liver injury also follows a pattern, with most cases occurring within the first two weeks to two months of treatment 4 9 .
This predictable window offers an opportunity for targeted monitoring of high-risk patients.
Initial cases may appear
Peak period for AT-DILI occurrence
Risk decreases significantly
Thai research published in 2025 reported even more dramatic risk increases:
Odds ratio for slow NAT2 acetylator status
Odds ratio for advanced age (>70 years)
Odds ratio for underweight BMI
Based on 2025 Thai research 5 .
Systematic Review and Meta-Analysis Methodology
The journey to establishing robust genetic associations requires synthesizing evidence from multiple studies. The 2025 meta-analysis on genetic polymorphisms and AT-DILI risk followed rigorous methodology 1 :
Across PubMed, EMBASE, and Cochrane Libraries
Focusing on human studies with sufficient genetic data
Using the Newcastle-Ottawa Scale checklist
Including genotype distributions in cases and controls
Using random-effects models to calculate pooled odds ratios
Preregistered in PROSPERO (CRD42022300625) before commencement
While many studies contribute to our understanding, systematic reviews with meta-analysis represent some of the most compelling evidence in the field.
The study was preregistered in PROSPERO (CRD42022300625) before commencement, reducing reporting bias.
Using PRISMA guidelines, they searched three major databases from inception to January 2023 with carefully constructed search strings.
From initially identified studies, they applied strict inclusion criteria, ultimately retaining 10 high-quality studies for quantitative synthesis.
Researchers extracted genotype distributions and recalculated odds ratios to ensure consistent comparisons across studies.
They employed random-effects models using the I² statistic to quantify heterogeneity, acknowledging the inherent differences in study designs and populations.
The analysis yielded several important findings, with the NAT2 associations being most prominent.
"NAT2 slow acetylator genotypes or CYP2E1 C1/C1 are causally linked to AT-DILI risk," suggesting that "genetic variants in drug-metabolizing enzymes regulated by NAT2 and CYP2E1 are involved in developing drug-induced liver injury in users of anti-TB drugs" 1 .
The growing understanding of genetic risk factors is paving the way for more personalized approaches to TB treatment.
Identifying slow acetylators before initiating treatment could allow for dose adjustments or alternative regimens.
Research suggests that reducing isoniazid doses by 50% in slow metabolizers may significantly lower hepatotoxicity risk without compromising efficacy 7 .
High-risk patients could receive more frequent liver function monitoring, particularly during the critical first 8 weeks of therapy.
While a 2016 retrospective analysis found that conventional hepatoprotective drugs like ursodeoxycholic acid (UDCA) didn't significantly shorten recovery time 8 , understanding the genetic mechanisms may lead to more targeted interventions.
As the Peruvian study demonstrates 7 , different populations may benefit from tailored treatment approaches based on their unique genetic profiles.
The journey from observing unpredictable liver injury to understanding its genetic basis represents a triumph of modern pharmacogenetics. As research continues to unravel the complex interplay between our genes, environment, and medications, we move closer to a future where TB treatment is both highly effective and universally safe, tailored to the individual patient rather than the average population.
As we continue to decode the human genome, the promise of truly personalized medicine for tuberculosis and countless other conditions comes closer to reality, transforming our approach from reactive treatment to proactive, prediction-based care.
References will be placed here manually in the final version.