Abnormal Liver Function at the Time of Tuberculosis Diagnosis in South India

Abstract

Background and objective:

Assessing liver function before antituberculosis treatment (ATT) is not routinely recommended for all tuberculosis (TB) patients in India. This study aims to determine the proportion of adult drug-sensitive TB patients with abnormal liver function (ALF) at diagnosis and its associated factors.

Materials and Methods:

A cross-sectional study was conducted among adult TB patients notified from January 2021 to January 2023 at a tertiary care institute in Puducherry, India. Liver function tests and alcohol use assessments (using AUDIT-C) were performed. ALF was defined as aminotransferase levels more than thrice the upper normal limit (>150 mg/dL for males, >105 mg/dL for females) or total bilirubin more than twice the upper reference value (>2.0 mg/dL).

Results:

Of 1809 TB patients enrolled, 63% were males, and 14% were aged over 60 years. Microbiologically confirmed cases were 49%, and newly diagnosed cases were 89%. Alcohol use was reported by 33%, and 5% were HIV positive. ALF was present in 137 (7.5%; 95%CI: 6.3–8.7%) patients, and the proportion was higher in patients with hazardous alcohol use (11%), age >60 years (10%), HIV seropositivity (14%), and severe anemia (16%).

Conclusion:

One out of 13 TB patients had baseline ALF. Liver function tests may be recommended for selected high-risk patients before initiating ATT.

INTRODUCTION

Hepatotoxicity during tuberculosis (TB) treatment poses a critical challenge in the treatment cascade. Hepatotoxicity affects around 2–28% of TB patients started on antituberculosis therapy (ATT).[1] Various risk factors including advanced age, harmful alcohol use, malnutrition, human immunodeficiency virus (HIV) infection, hepatitis infection, and extensive disease can contribute to this issue.[1,2] Baseline derangement of liver function test (LFT) and pre-existing liver diseases can impact the efficacy of antitubercular treatment.[1] Derangement of LFT parameters, such as aspartate aminotransferase (AST), alanine aminotransferase (ALT), alkaline phosphatase (ALP), gamma-glutamyl transpeptidase (GGT), and total bilirubin, may provide evidence of acute or chronic liver disease in an individual.[3] Globally, abnormal liver function (ALF) is found in 9–22% of general population.[4] However, there is paucity of data with respect to prevalence of ALF particularly among TB patients before the initiation of ATT. Few studies conducted across the globe to assess the risk factors of ATT induced hepatotoxicity have reported the proportion of individuals with abnormal baseline liver enzymes to be between 3 and 19%, correlating with two to three folds’ higher risk of ATT-induced hepatotoxicity.[5,6] This risk is further aggravated by the risk factors like chronic alcoholism, hepatitis-B or C and chronic liver disease.[7]

ALF increases the risk of developing hepatotoxicity. As three of the four first-line antitubercular treatment (ATT) has an adverse effect of hepatotoxicity, monitoring liver function in patients with abnormal LFT at the time of diagnosis and during the treatment allows early detection of any hepatic injury and thus permits withdrawal of potentially hepatotoxic drugs before serious life-threatening manifestations develop.[5]

Currently, the National Tuberculosis Elimination Program (NTEP) does not mandate liver function testing for all TB patients prior to ATT initiation. The American Thoracic Society advises baseline alanine transaminase (ALT) testing for all patients beginning TB treatment, with periodic monitoring throughout treatment for patients with liver disease, abnormal LFTs, alcohol use disorder, HIV or hepatitis, or those using any hepatotoxic medications.[8] The Taiwanese Centre for Disease Control also advise baseline and routine monitoring for all TB patients, whereas the World Health Organization (WHO) urge routine liver monitoring in only high-risk patients with pre-existing liver disease, baseline derangement of LFT, chronic alcoholism, hepatitis, or prior ATT induced hepatotoxicity.[2,9] The high prevalence of alcohol use (41.6%) in Puducherry, India, further necessitates the need to know the burden of liver derangements in TB patients as it has an impact on the TB treatment outcomes.[10]

Though there are studies regarding incidence of hepatotoxicity due to antitubercular drug; we do not have much studies that focus on abnormal liver function before TB treatment initiation. Therefore, the current study was undertaken to describe the proportion of patients with abnormal liver function and its associated factors in a cohort of adult drug-sensitive TB patients before treatment initiation.

MATERIALS AND METHODS

Study design and duration

A cross-sectional analytical study was conducted on notified TB patients at a tertiary hospital in Puducherry, India, between January 2021 and January 2023.

Study setting

Annually, about 1200 patients are referred from this tertiary care hospital in Puducherry, India. The NTEP recommends random blood sugar (RBS) and HIV testing for all patients diagnosed with TB. Additionally, LFT is recommended before ATT initiation for selected patient groups like those with pre-existing liver disease, HIV seropositivity, and drug-resistant TB.[11] If there are no abnormalities in the patient’s liver function status, drug-susceptible TB patients are initiated on the standard ATT regimen, which includes a combination of Isoniazid, Rifampicin, Ethambutol, and Pyrazinamide for 6 months (4 HRZE+2HRE). However, patients exhibiting symptoms of liver disease or abnormal liver function test results are started on a modified regimen as per NTEP guidelines.[12]

Sample size and study participants

Assuming the proportion of TB patients with abnormal liver function tests (LFT) as 3% (conservative estimate based on a previous study from Mumbai), a relative precision of 25%, and an alpha error of 5%, the sample size determined was 1984 (OpenEpi version 3.01).[13] However, we included all the TB patients aged above 18 years notified from the tertiary care center. Patients diagnosed with multidrug-resistant TB were excluded.

Data sources and data entry

Information on age, gender, place of residence, type of TB, and site of TB was obtained from the TB notification register. Patients were asked about alcohol use in the year before TB diagnosis. Alcohol usage was categorized into hazardous and nonhazardous using the Alcohol Use Disorder Identification Test (AUDIT-C). As alcohol use was not reported among female TB patients in our study, a uniform AUDIT-C cut-off score of ≥4 (validated for males) was considered as hazardous use.[14]

The biochemical investigations were run in an automated blood chemical analyzer by Beckman Coulter (Model–AU 5811) to ascertain the serum levels of AST, ALT, ALP, and total bilirubin. The reports of liver function, hemogram, and hepatitis-B were accessed through the hospital information system. The elevation of AST or ALT more than thrice the upper normal limit (ULN: >150 mg/dL for males and >105 mg/dL for females) or total bilirubin more than twice (>2.0 mg/dL) the upper reference value was considered abnormal liver function.[15]

For assessing the anemic status of patients, a cut-off hemoglobin values of less than 12.0 g/dL were taken for females, and less than 13.0 g/dL was considered for males.[16] Hemoglobin level of <7 g/dL was considered severe anemia.[17] All these details were entered into REDCap, and data were imported as an Excel sheet to STATA v14 for further analysis.

Statistical analysis

The presence of abnormal liver function was summarized as frequency and percentage with a 95% confidence interval (CI). Chi-squared test was performed to assess the association between ALF and selected factors (sociodemographic and clinical) among eligible TB patients, and an adjusted prevalence ratio with 95% CIs was calculated from log-binomial regression. Variables with P < 0.2 in the unadjusted analysis were included in the regression model. All the analysis was carried out at 5% significance level using STATA version 14 (StataCorp, Texas, USA).

Ethics

Approval from the Institute Ethics Committee was obtained before the start of the study, and informed consent was obtained from all TB patients during the alcohol use assessment. The results of LFTs were reported to the treating physician for further management or decision regarding the chosen regimen.

RESULTS

Of the total 1879 adult drug-susceptible TB patients notified, 70 patients did not have LFT values, and hence, 1809 patients (96%) were included. The mean (standard deviation) age of the study participants was 43 (15) years; 1138 (63%) were males, and 1080 (60%) belonged to rural areas. Alcohol use was reported by 590 patients (33%); of them, 340 (58%) had hazardous alcohol use. Ninety-three patients (5%) had HIV; 358 patients (20%) had diabetes, and severe anemia was seen in 115 patients (6%). One patient had hepatitis C infection, and 33 (2%) had hepatitis B infection [Table 1]. Of the total, 877 (48%) had microbiologically confirmed TB, 199 (11%) were previously treated patients, and 889 (49%) had pulmonary TB.

Table 1.

Sociodemographic and clinical characteristics of adult tuberculosis patients notified during 2021–2023 at a tertiary care center, Puducherry, India (n=1809)

Characteristic	Frequency (%)
Age group (years)
18–30	447 (24.7)
31–45	562 (31.1)
46–60	550 (30.4)
>60	250 (13.8)
Gender
Male	1138 (62.9)
Female	671 (37.1)
Area of residence
Urban	729 (40.3)
Rural	1080 (59.7)
Alcohol use*
Hazardous	340 (18.8)
Nonhazardous	250 (13.8)
Nonusers	954 (52.7)
Unknown†	265 (14.7)
HIV status
Positive	93 (5.1)
Negative	1702 (94.1)
Unknown	14 (0.8)
Diabetes mellitus
Yes	358 (19.8)
No	1432 (79.2)
Unknown	19 (1.0)
Hemoglobin (g/dL)
<7	115 (6.4)
≥7	1633 (90.3)
Unknown	61 (3.3)
Hepatitis B
Yes	33 (1.8)
No	1651 (91.3)
Unknown	125 (6.9)

*Alcohol use was assessed for the year prior to the diagnosis of tuberculosis using the AUDIT-C questionnaire. Scores of 4 or more, 1–3, and 0 were categorized as hazardous use, nonhazardous use, and no use of alcohol, respectively. ^†Patients were not reachable to obtain alcohol history

At the time of TB diagnosis, 137 patients (7.5%, 95% CI: 6.3%−8.7%) had abnormal liver function; 42 (31%) had elevated AST, 33 (24%) had elevated ALT, while both AST and ALT were elevated in 14 (10%) patients. Total bilirubin was deranged in 91 (66%) patients, and all three parameters were elevated in two (1%) patients.

In adjusted analysis, hazardous alcohol use, age over 60 years, HIV positivity, and severe anemia were significantly associated with abnormal liver function [Table 2]. Hazardous alcohol users compared to nonusers had a higher prevalence of ALF (11% vs 7%; aPR: 1.8; 95% CI: 1.0–3.2). ALF was higher among the elderly (>60 years) than patients aged 18–30 years (10% vs 5%; aPR: 2.2; 95% CI: 1.2–4.1). TB patients with HIV had a higher prevalence of ALF as compared to HIV negative patients (14% vs 7%; aPR: 1.8; 95% CI: 1.1–3.3). Those patients with severe anemia had twice the prevalence of ACF compared to those with hemoglobin levels more than 7 g/dL (16% vs 7%; aPR: 2.2; 95% CI: 1.3–3.4).

Table 2.

Association of sociodemographic and clinical characteristics with baseline liver function in adult tuberculosis patients notified during 2021–2023 in a tertiary care center, Puducherry, India (n=1809)

Variable	Liver Function*			Unadjusted PR† (95% CI)	Adjusted PR‡ (95% CI)	P
	Total	Abnormal n (%)¶	Normal n (%)¶
	1809	137 (7.5)	1672 (92.5)
Age groups (years)
18–30	447	23 (5.1)	424 (94.9)	ref	ref	–
31–45	562	49 (8.7)	517 (91.3)	1.7 (1.0–2.7)	1.5 (0.9–2.5)	0.082
46–60	550	39 (7.1)	516 (92.9)	1.3 (0.8–2.2)	1.4 (0.8–2.3)	0.160
>60	250	26 (10.4)	225 (89.6)	2.0 (1.1–3.4)	2.2 (1.2–4.1)	0.006
Gender
Male	1138	86 (7.6)	1052 (92.4)	ref	ref	–
Female	671	51 (7.6)	620 (92.4)	1.0 (0.7–1.4)	1.3 (0.7–2.3)	0.286
Site of TB
Pulmonary	889	60 (6.8)	829 (93.2)	ref	ref	–
Extra pulmonary	920	77 (8.4)	843 (91.6)	1.2 (0.8–1.7)	1.1 (0.7–1.7)	0.577
Type of patient
New	1610	125 (7.7)	1485 (92.2)	ref	–	–
Previously treated	199	12 (6.0)	187 (93.9)	0.7 (0.4–1.3)	–	–
Type of TB diagnosis
Microbiologically confirmed	877	57 (6.5)	820 (93.5)	ref	ref	–
Clinically diagnosed	932	80 (8.5)	852 (91.4)	1.3 (0.9–1.8)	1.2 (0.8–1.8)	0.368
HIV status
Positive	93	13 (13.9)	80 (86.1)	1.9 (1.1–3.2)	1.8 (1.1–3.3)	0.026
Negative	1702	124 (7.3)	1578 (92.7)	ref	ref	ref
Unknown	14	0 (0)	14 (100)	–	–	–
Diabetes mellitus
Yes	358	24 (6.7)	334 (93.3)	0.8 (0.5–1.3)	–	–
No	1432	111 (7.8)	1321 (92.2)	ref	–	–
Unknown	19	2 (10.5)	17 (89.5)	1.3 (0.3–5.0)	–	–
Hemoglobin (g/dL)
<7	115	18 (15.6)	97 (84.4)	2.2 (1.3–3.4)	2.2 (1.3–3.4)	0.001
≥7	1633	116 (7.1)	1517 (92.9)	ref	ref	–
Unknown	61	3 (4.9)	58 (95.1)	0.7 (0.2–2.1)	1.0 (0.3–3.1)	0.940
Hepatitis B
Yes	33	2 (6.1)	31 (93.9)	0.7 (0.2–2.9)	0.7 (0.2–2.8)	0.745
No	1651	131 (7.9)	1520 (92.1)	ref	ref	–
Unknown	125	3 (2.4)	122 (97.6)	0.4 (0.1–1.0)	0.4 (0.1–1.1)	0.096
Alcohol use§
Hazardous drinking	340	36 (10.6)	304 (89.4)	1.5 (1.0–2.2)	1.8 (1.0–3.2)	0.042
Nonhazardous drinking	250	23 (9.2)	227 (90.8)	1.3 (0.8–2.1)	1.6 (0.9–3.0)	0.107
Nonusers	954	66 (6.9)	888 (93.1)	ref	ref	–
Unknown	265	12 (4.5)	253 (95.5)	0.6 (0.3–1.2)	0.8 (0.4–1.7)	0.618

*Elevation of any one of the liver function enzymes (alanine transaminase or aspartate transaminase) more than thrice the upper normal limit or total bilirubin more than twice the upper normal limit was defined as abnormal liver function. ^†Prevalence ratio. ^‡Multivariable analysis (log binomial regression) was done. Exposure variables with P<0.2 in the unadjusted analysis were included in the multivariable analysis. Model fit was assessed using BIC (–12,524), AIC (0.53), and log-likelihood (–465.2). ^§Alcohol use was assessed for the year prior to the diagnosis of tuberculosis using the AUDIT-C questionnaire. Scores of 4 or more, 1–3, and 0 were categorized as hazardous use, nonhazardous use, and no use of alcohol. ^¶Percentages in parentheses represent row-wise proportions within each category of the variable.

DISCUSSION

In our setting, one out of every 13 adult TB patients had abnormal liver function at the time of TB diagnosis. Hazardous alcohol use, age over 60 years, HIV-positivity, and severe anemia were associated with abnormal liver function.

Studies from India and other countries reported prevalence of ALF ranging from 6 to 20%.[18,19] The difference in the prevalence may partly be explained by the variations in the definitions used for ALF. Studies that defined ALF as values above ULN but less than three times the ULN reported higher prevalence. A study from Mumbai, India, reported a lower prevalence of 6.4% as they used five times the ULN as the cut-off.[13] Estimates of ALF could also be influenced by other factors like hepatitis B and C prevalence, alcohol use, and HIV prevalence. In addition to these clinical factors, the study setting itself may have influenced the findings. As the study was conducted in a tertiary care institute, some patients may have been referred from peripheral facilities for comorbidities or suspected liver dysfunction. This may have contributed to a higher prevalence of ALF. However, referral status was not captured, which limits further interpretation. Although there are guidelines for assessing liver injury after ATT initiation, defining ALF at baseline remains unclear in programmatic settings. Modifications in ATT regimens are considered usually based on symptoms, etiology and LFT parameters.

One-third of our TB patients reported alcohol use, with more than half exhibiting hazardous use. This finding was consistent with other studies from India.[20,21] Our study shows a two-fold higher proportion of hazardous alcohol users having abnormal LFT compared to nonusers, which is consistent with other findings.[22] Excessive alcohol consumption is associated with elevated liver enzymes, leading to alcoholic liver disorder.[23,24] Alcohol dependence is also higher among those with hazardous use.[25] Hence, evaluating LFT in these patients before TB treatment initiation can reduce ATT-induced hepatotoxicity.

In our study, elderly TB patients had a higher prevalence of ALF, similar to another study from China.[26] Increasing age and medication for comorbidities in the elderly can affect liver function. Additionally, the incidence of ATT-induced adverse events (24%) and unfavorable TB treatment outcomes (9%) was reported to be higher in this group.[27] In our setting, 5% of TB patients were HIV positive, and these patients had a higher prevalence of ALF. All these patients undergo LFT as part of the national TB guidelines. We also found an association between severe anemia and ALF. Severe anemia is a marker for severe TB disease, and studies have shown the co-existence of anemia and liver disease.[28]

Currently in India, the NTEP does not recommend routine LFT testing for all TB patients before treatment initiation. Based on our findings, we recommend that policymakers under NTEP consider targeted baseline LFT screening for high-risk subgroups such as those with hazardous alcohol use, severe anemia, or advanced age particularly in settings with adequate laboratory capacity. Monitoring liver function in these patients during treatment may enable early detection and management of ATT-induced hepatotoxicity, potentially preventing treatment interruptions and adverse treatment outcomes.[13] However, there is limited evidence from programmatic settings on whether baseline LFT testing reduces hepatotoxicity and improves treatment outcomes, highlighting the need for further research.

Although we did not perform a formal economic evaluation, implementing targeted baseline LFT under NTEP would involve incremental program costs for laboratory consumables, equipment, and staff time, which should be weighed against potential savings from reduced hospitalizations, fewer treatment interruptions, and improved treatment completion. From a program point of view, implementing such a policy could present challenges like potential laboratory overload, increased cost, and limited testing facilities at the primary healthcare level. However, expanding the testing network through a hub and spoke model and public-private partnership presents an opportunity. Targeted testing for selected groups and in regions with high alcohol use may be beneficial. Future programmatic studies should therefore evaluate both clinical effectiveness and cost-effectiveness in real-world settings to guide NTEP decision-making.

The study has a few strengths. We could perform LFT in more than 95% of TB patients in the study setting. Information on hepatitis B and C was available for nearly all patients. We used the AUDIT-C tool to capture alcohol use, though desirability bias cannot be ruled out. Information on etiology in patients with abnormal liver function tests was not available. However, there is limited evidence from programmatic settings on whether baseline LFT testing reduces hepatotoxicity and improves treatment outcomes, highlighting the need for further research.

CONCLUSION

One out of every 13 patients had abnormal liver function, with the prevalence being two times higher among patients with hazardous alcohol use. Elderly patients, HIV-positive individuals, and those with severe anemia had a higher prevalence of abnormal liver function at baseline. The above patient groups may be considered for LFT in settings where laboratory resources are available.

Ethical statement

Approval from the Institute Ethics Committee was obtained before the start of the study. (IEC approval number – JIP/IEC/2022/133) and informed consent was obtained from all TB patients during the alcohol use assessment. The results of the LFT were informed to the treating physician for further management or decision regarding the regimen to be chosen.

Conflicts of interest

There are no conflicts of interest.

Funding Statement

Nil.