Abstract
Introduction
A 6-month course of isoniazid, 300 mg daily, was programmatically introduced in Eritrea in 2014 as tuberculosis preventive therapy in people living with human immunodeficiency virus (PLHIV). The rollout of isoniazid preventive therapy (IPT) in PLHIV was successful in the first 2–3 years. After 2016, rumours based on rare but real incidents of liver injuries following use of IPT spread widely across the country and created concerns amongst healthcare professionals and consumers, that ultimately caused dramatic decline in the rollout of the intervention. Decision makers have been demanding better evidence as previously conducted local studies had inherent methodological limitations. This real-world observational study was conducted to evaluate the risk of liver injury associated with IPT among PLHIV attending Halibet national referral hospital, Asmara, Eritrea.
Methods
A prospective cohort study, that consecutively enrolled PLHIV attending Halibet hospital, was conducted between 1 March and 30 October 2021. Those exposed to anti-retroviral therapy (ART) plus IPT were considered as exposed and those taking only ART were considered as unexposed. Both groups were prospectively followed up for 4–5 months with monthly liver function tests (LFTs). A Cox proportional hazard model was used to explore whether there was increased risk of drug-induced liver injury (DILI) associated with IPT. Probability of survival without DILI was also estimated using Kaplan–Meier curves.
Results
A total of 552 patients, 284 exposed and 268 unexposed, completed the study, with a mean follow-up time of 3.97 (SD 0.675) months for the exposed and 4.06 (SD 0.675) months for the unexposed. Twelve patients developed drug-induced liver injury (DILI), with a median time-to-onset of 35 days (interquartile range: 26.8, 60 days). All cases were from the exposed group and all except two cases were asymptomatic. The incidence rate of DILI in the exposed group was 10.6 cases per 1000 person–months and zero for the unexposed group (p = 0.002).
Conclusion
DILI in PLHIV taking IPT was common; therefore, liver function should be closely monitored to safely administer the product. Despite high levels of deranged liver enzymes, the majority had no symptoms of DILI, emphasising the importance of close laboratory monitoring, especially during the first 3 months of treatment.
Key Points
| Different institutions have different guidelines on the need for routine liver function monitoring while administering isoniazid preventive therapy in people living with human immunodeficiency virus (PLHIV). |
| This study revealed that PLHIV on isoniazid preventive therapy (IPT) were at higher risk of liver injury compared with those unexposed to IPT. |
| Roll out of IPT in PLHIV necessities regular laboratory monitoring of liver function tests, especially in the first 3 months of treatment to ensure safer use of the product, which urges the need for capacitating the existing laboratory setups. |
Introduction
Isoniazid (INH) has been one of the preferred tuberculosis (TB) preventive therapies in people living with human immunodeficiency virus (PLHIV) despite its risks of liver injury [1, 2]. The mechanism for isoniazid (INH)-induced liver injury is yet not well understood. Several pathways such as metabolite-mediated toxicity, mitochondrial damage, oxidative stress, disruption of endobiotic homeostasis and immune-mediated toxicity have been proposed as possible causes of INH-induced liver injury [3–6]. A plausible biochemical mechanism is that the metabolism of INH, a hydrazide, produces reactive or toxic metabolites that bind and damage cellular macromolecules in the liver. Three INH metabolites namely ‘acetyl hydrazine, hydrazine, and a radical metabolite resulting from the bioactivation of INH itself’ have been proposed to be responsible for INH-induced liver injury [2, 7].
The World Health Organization (WHO) assessed the risk of INH-induced liver injury and concluded that it was generally safe [8]. The risk of hepatic injuries following use of isoniazid preventive therapy (IPT) in PLHIV significantly varies from study to study. In some studies, IPT-related liver injury was relatively low (0.07–2.5% [9–14]) while others have reported higher incidences (6.4–22% [15–18]).
When IPT for PLHIV was programmatically introduced in Eritrea, in 2014, increased risks of isoniazid-associated hepatic injuries, including some fatalities, were reported [17, 19]. These studies indicated that many of the cases, while on only highly active anti-retroviral therapy (HAART), had no history of drug-induced liver injury (DILI) for several years. Due to a lack of comparators or controls, the aforementioned studies could not rule out the effect of anti-retroviral therapy (ART) on DILI.
Moreover, rumours on the rarely reported cases of IPT-related hepatic injuries spread widely in the country and created concerns among healthcare professionals and PLHIV. Consequently, some healthcare professionals showed resistance against implementation of IPT that possibly caused the enormous decline in the rollout of the intervention since 2017 [20]. This put the communicable disease control (CDC) division of the Ministry of Health of Eritrea at a crossroad. Accordingly, decision makers have been demanding stronger evidence for decision making as the previously conducted studies had methodological limitations; namely they only consisted of a case series and a prospective study with active surveillance but without a comparator group [17, 19]. This real-world prospective cohort study was conducted to evaluate the risk of IPT among PLHIV attending Halibet national referral hospital located in Asmara, Eritrea.
Methods
Study Design and Setting
This was a prospective cohort study that consecutively enrolled PLHIV. PLHIV taking a 6 months IPT course were considered as exposed, and those without IPT as unexposed. To avoid the challenges of loss to follow-up and other competing risks, the cohort was open/dynamic, allowing patients to be recruited at different times and change their exposure status if deemed necessary by the HIV care clinic prescribers. Moreover, patients were informed on the benefits of monthly follow-ups and only those residing in the city, Asmara, were recruited to make the monthly follow-up practical. In case patients discontinued their IPT for different reasons, were lost to follow-up, died or changed their follow-up to other HIV care clinics in the country, patients were censored and only the time-period that contributed to the study was considered. The study was conducted in Halibet national referral hospital between 1 March and 30 October 2021. Halibet national referral hospital is based in the capital city of Eritrea and serves more than 1200 PLHIV; the second highest hospital in the country in serving PLHIV.
Study Population
Halibet national referral hospital has been delivering IPT to adult populations of PLHIV since November 2014. In Eritrea, once patients complete a 6 months IPT course, they are expected to take another round after 5 years of the first dose. PLHIV who had never received IPT or who had received IPT before 5 years were consecutively enrolled to the exposed group. Those who refused to be exposed to IPT or were ineligible to retake IPT (exposed to 6 months of IPT within 5 years of the commencement of the study) were also consecutively enrolled as unexposed. PLHIV who had previous history of INH-induced liver injury were excluded from the study. The study enrolled patients who had no active liver disease or elevated liver function enzymes at baseline. Thus, patients who tested positive for hepatitis B and C, had deranged liver function enzymes and a history of regular alcohol consumption were excluded from the study.
Exposure Definition and Measurement
The exposure of interest in this study was a 6 months isoniazid 300 mg tablet daily used as tuberculosis preventive therapy. PLHIV on IPT were considered as exposed and those without isoniazid as unexposed. All patients (whether exposed or not) were on ART. To minimise neurological adverse events related to isoniazid, the exposed group were also taking vitamin B6 tablets as supportive medication. Whenever required by clinicians, the study participants were free to shift their regimens of ART and/or IPT.
Outcome Definition and Measurement
The primary outcome or endpoint of this study was development of drug-induced liver injury (DILI) in the exposed and unexposed patients. Severity and classes of clinical patterns of DILI were considered as secondary outcomes.
Diagnostic criteria of DILI and its severity and clinical pattern classification were based on a consensus paper developed by a international DILI expert working group of clinicians and scientists [21]. Accordingly, DILI was diagnosed based on one of the following thresholds:
Serum alanine aminotransferase (ALT) value ≥ 5× the upper limit of normal (ULN), or
Alkaline phosphatase (ALP) value ≥ 2× ULN or
ALT value ≥ 3× ULN with simultaneous elevation of bilirubin concentration to ≥ 2× ULN.
The clinical pattern of liver injury was classified according to the liver enzyme elevations that qualified as DILI [21]. Pattern of liver injury was defined using the R-value; where R = (ALT/ULN)/(ALP/ULN). Estimation of ALT and ALP was made from the same serum sample. Accordingly, DILI was classified as hepatocellular pattern if R ≥ 5, mixed pattern if R > 2 and < 5, and cholestatic pattern if R ≤ 2.
As per the above-mentioned consensus paper, severity of DILI was classified as follows:
Mild: Elevated ALT/ALP concentration reaching criteria for DILI but bilirubin concentration < 2× ULN.
Moderate: Elevated ALT/ALP concentration reaching criteria for DILI and bilirubin concentration ≥ 2× ULN, or symptomatic hepatitis.
- Severe: Elevated ALT/ALP concentration reaching criteria for DILI, bilirubin concentration ≥ 2× ULN, and one of the following:
- International normalised ratio ≥ 1.5, and/or
- Ascites and/or encephalopathy, disease duration < 26 weeks, and absence of underlying cirrhosis, and/or
- Other organ failure considered to be due to DILI.
Fatal or transplantation: Death or transplantation due to DILI.
Patient Recruitment, Follow-Up and Study Instrument
PLHIV taking HAART were assessed for eligibility by the HIV care clinic prescriber, a medical doctor. Those eligible, as described above, and who consented to participate were subjected to baseline laboratory investigations. Patients with normal liver function tests (LFTs) and free of hepatitis B and C were consecutively enrolled to IPT. While those willing to participate but without liver diseases or deranged liver enzymes were consecutively enrolled as the comparator group (unexposed) (Fig. 1). Eligible patients were forwarded to data collectors (two registered nurse practitioners and a pharmacist working at the HIV care clinic) for enrollment. The data collectors were responsible for enrollment and follow-up of patients under the terms of the study protocol. Patients consented to participate were consecutively recruited to the exposed and unexposed groups until the required sample size was obtained. In the routine practice, patients were followed-up every 3 months, even for those taking IPT. Thus, recruitment continued for 3 months to give all patients an equal chance of recruitment.
Fig. 1.
Eligibility criteria for enrollment and approach of PLHIV attending Halibet national referral hospital, 1 March–30 October 2021, Asmara, Eritrea. IPT: Isoniazid preventive therapy; LFT: Liver function test; HAART: Highly active anti-retroviral drugs; ULN: Upper limit of normal
Liver function tests such as alanine transaminase (ALT), alkaline phosphatase (ALP) and total bilirubin were performed for all study participants, during enrollment and monthly afterwards until completion of the study. In case ALT was not available, aspartate transaminase (AST) was considered. Moreover, body weight, height and CD4 cell count were taken for all patients in both groups. Drug and medical history were also assessed prior to recruitment and during follow-up. Study participants were informed on the signs and symptoms of hepatic injury such as nausea, vomiting, abdominal pain, unexplained fatigue, jaundice and dark urine, and advised to visit the health facility or call the research team in case they experienced similar events. All follow-ups were made through hospital visits.
Both exposed and unexposed groups, once enrolled, were expected to be followed-up for a period of 6 months because IPT was prescribed for 6 months. Due to a shortage of laboratory supplies impacted by the COVID-19 pandemic, the majority of the study participants completed their follow-up in 4 months. During follow-up, liver function tests (LFTs) were taken on a monthly basis for both groups. All study participants, regardless of their exposure status, were actively and equally monitored by the treating physician and data collectors for the risk of hepatic injury until the onset of hepatic injury or death, discontinuation of IPT or completion of the study, whichever came first. Patients who developed liver damage during the study period were managed according to the national treatment protocol. In all cases who developed hepatotoxicity, possible alternative explanations such as hepatitis B and C infections, alcohol consumption and intake of other drugs were assessed and ruled out.
A structured questionnaire adopted from a similar study conducted previously [17] was used to collect data. The questionnaire had four main sections: patient demographic characteristics, drug details, baseline and follow-up test results, and reaction (liver injury) details. Data were collected by the data collectors working in the HIV care clinic and supervised by one of the authors (M.R.).
Sample Size Determination
For this cohort study, sample size was calculated to be 502 (251 from each group) with Epi Info version 7.2 [22] with the following assumptions: Margin of error: α = 0.05; statistical power: 90%; exposed to unexposed ratio: 1 and assumed relative risk: 2.5. Besides, as there were no local studies that evaluate the incidence of hepatic injury associated with HAART medicines, a median expected cumulative incidence of hepatic injury in the unexposed population (HIV patients without IPT) was taken from the literature as 8.4% [23]. Including 10% for non-response, the sample size was finally calculated to be 552 (276 from each group).
Data Analysis
Data were entered using a program census and survey processing system (CSPro) version 7.4 and analysed using R 4.0.4 [24] and SPSS 20.0. Descriptive statistics including percentages, frequencies, number needed to harm, mean (standard deviation, SD) and median (interquartile range, IQR) were computed as appropriate. As a measure of frequency, incidence rate of DILI was computed using descriptive statistics. Chi-square tests of independence and bivariate logistic regression were used to identify factors associated with DILI. A Cox proportional hazard model was used to explore whether there was increased risk of drug-induced liver injury (DILI) associated with IPT and compare the time-to-event onset of DILI in both groups (time-to-event analysis). To compare the incidence rate of DILI in the exposed versus the unexposed, the rate ratio (RR) or the hazard ratio with 95% confidence interval (CI) was computed, as appropriate. Calculating the RR/HR to compare risk of DILI in the exposed and unexposed was not applicable or was undefined as none of the unexposed group could develop at least one case of DILI. Even though it is unknown how many cases from the exposed group would develop DILI until at least one case was observed in the unexposed, a sensitivity analysis was carried out to calculate the lowest possible RR by adding one case to the unexposed group. A two-tailed test with a 0.05 alpha level of significance was used for all statistical tests.
Results
Background and Follow-Up Information of Study Participants
A total of 578 PLHIV on ART, 302 (51%) exposed and 276 (49%) unexposed, were enrolled into the study (Fig. 2). The exposed group had a median age of 48 years [interquartile range (IQR): 43, 53 years] and for the unexposed group it was 47 years (IQR: 41, 52 years). About two-thirds (67.2%) of the study participants were females (Table 1). The exposed and unexposed groups had no differences in baseline characteristics except in history of comorbidities (p = 0.015) and type of ART which was taken during the study period (p < 0.001). The majority of the study participants were on efavirenz-based (58.3%) and 40.2% were on a dolutegravir-based ART regimen. Of those exposed to IPT, 65.5% and 32.7% were on efavirenz-based and dolutegravir-based ART regimens, respectively, while in the unexposed group the proportion was 50.7% for efavirenz-based regimen and 48.1% for dolutegravir-based regimen.
Fig. 2.
Enrollment and follow-up information of PLHIV attending Halibet national Referral hospital, 1 March–30 October 2021, Asmara, Eritrea
Table 1.
Sociodemographic characteristics of PLHIV, both exposed and unexposed cohorts to isoniazid preventive therapy, attending Halibet national referral hospital, Asmara, Eritrea, 2021 (N = 552)
| Variables (measured at baseline) |
Category | Cohort group | p value | |
|---|---|---|---|---|
| Exposed n (%) n = 284 |
Unexposed n (%) n = 268 |
|||
| Age group (years) | ||||
| 15–59 | 266 (93.7) | 244 (91.1) | 0.246 | |
| 60 and above | 18 (6.3) | 24 (8.9) | ||
| Sex | ||||
| Male | 94 (33.1) | 87 (32.5) | 0.874 | |
| Female | 190 (66.9) | 181 (67.5) | ||
| Body weight (kg) | ||||
| < 50 | 60 (21.1) | 74 (27.6) | 0.076 | |
| 50 and above | 224 (78.9) | 194 (72.4) | ||
| BMI | ||||
| < 18.5 | 50 (17.6) | 54 (20.2) | 0.526 | |
| 18.5 to < 25 | 179 (63.0) | 167 (62.3) | ||
| 25 to < 30 | 40 (14.1) | 38 (10.3) | ||
| 30 and above | 15 (5.3) | 8 (3.0) | ||
| Co-morbidity | ||||
| Yes | 18 (6.3) | 33 (12.3) | 0.015 | |
| No | 266 (93.7) | 235 (87.7) | ||
| CD4 cell count (cell/µL) (just before or at enrollment)α | ||||
| < 200 | 15 (5.3) | 28 (10.4) | 0.039 | |
| > 200 | 259 (91.2) | 234 (87.3) | ||
| Type of ART* | ||||
| Nevirapine based | 5 (1.8) | 1 (0.4) | < 0.001 | |
| Efavirenz based | 186 (65.5) | 136 (50.7) | ||
| Dolutegravir based | 93 (32.7) | 129 (48.1) | ||
PLHIV: People living with HIV; IPT: Isoniazid preventive therapy; SD: Standard deviation; IQR: Interquartile range; ART: Anti-retroviral therapy
*Few patients on other ART regimens are removed from this analysis as they were few in number
αFor some patients CD4 cell count was missing
There was no difference in follow-up time between groups (p = 0.109). The mean follow-up time for the exposed group was 3.97 (SD 0.68) months, and 4.06 (SD 0.68) months for the unexposed (Table 2).
Table 2.
Follow-up outcome characteristics of PLHIV, both exposed and unexposed cohorts to IPT, attending Halibet national referral hospital, Eritrea, 2021(N = 552)
| Variables (Measured at baseline) |
Category | Cohort group | p value | |
|---|---|---|---|---|
| Exposed n (%) n = 284 |
Unexposed n (%) n = 268 |
|||
| Cumulative exposure to ART in years (before and during study) | ||||
| 2 and below | 99 (35.2) | 141 (52.8) | < 0.001 | |
| 2–5 | 10 (3.6) | 21 (7.9) | ||
| > 5 | 172 (61.2) | 105 (39.3) | ||
| Mean follow-up time (in months) | 3.97 | 4.06 | 0.109 | |
| Follow-up status | ||||
| Follow-up completed | 284 | 268 | 0.097* | |
| Follow-up defaulted with unknown reasons | 9 | 7 | ||
| Follow-up interrupted with travel | 7 | 1 | ||
| Follow-up interrupted with adverse event(s) | 2 | 0 | ||
PLHIV: People living with HIV; IPT: Isoniazid preventive therapy; SD: Standard deviation; IQR: Interquartile range; ART: Anti-retroviral therapy
*Fisher’s exact test
Incidence Rate of Hepatic Injury Associated with IPT
The cumulative follow-up time in the exposed and unexposed groups were 1128 and 1089 months, respectively. In this time period, a total of 12 study participants, all from the exposed group, developed the outcome of interest (DILI). None of the unexposed group developed DILI. The incidence rate of DILI in the exposed group was 10.6 cases per 1000 person–months and in the unexposed group it was zero (0) cases per 1000 person–months (Table 3). The exposed group were likely to develop DILI compared with the unexposed (p = 0.002), even though it was not tested at the multivariate level for reasons explained in the data analysis section.
Table 3.
Incidence rate of hepatic injury associated with isoniazid preventive therapy (IPT) and of associated factors in PLHIV attending Halibet national referral hospital in 2021 (N = 552)
| Variables | Cases of hepatic injury n (%) |
Person–time (months) | Incidence rate per 1000 person–months | Crude rate ratio (95% CI) | p-value | |
|---|---|---|---|---|---|---|
| Exposure to *IPT | ||||||
| Exposed | 12 (4.93) | 1128 | 10.60 | Not applicable¥ | – | |
| Unexposed | 0 (0.00) | 1089 | 0.00 | |||
| Age (years)* | ||||||
| ≤ 45 | 4 | 802 | 4.99 | Ref | 0.099 | |
| > 45 | 8 | 1415 | 5.62 | 1.33 (0.303, 5.144) | ||
| Sex | ||||||
| Female | 10 | 1491 | 6.71 | 2.67 (0.58, 24.86) | 0.244 | |
| Male | 2 | 726 | 2.75 | Ref | ||
| BMI | ||||||
| < 18.5 | 4 | 418 | 9.57 | 0.87 (0.08, 42.64) | 0.998 | |
| 18.5 to < 25 | 5 | 1396 | 3.58 | 0.32 (0.04, 15.33) | 0.314 | |
| 25 to < 30 | 2 | 309 | 6.48 | 0.59 (0.03, 34.59) | 0.538 | |
| 30 and above | 1 | 90.5 | 11.0 | Ref | ||
| Total number of drugs taken* | ||||||
| ≤ 2 | 0 | 1062 | 0.00 | Ref | ||
| 3–4 | 10 | 1105 | 9.05 | 10.57 (1.52, 455.05) | 0.006 | |
| > 4 | 2 | 50.1 | 39.9 | 42.39 (2.21, 501.19) | 0.006 | |
| CD4 cell count (cell/µL)* | ||||||
| < 200 | 0 | 177 | 0.00 | 1.01 (0.02, 6.96) | 0.999 | |
| > 200 | 10 | 1974 | 5.07 | Ref | ||
| Type of ART regimen | ||||||
| Nevirapine based | 1 | 22.7 | 44.1 | 8.14 (0.17, 72.75) | 0.135 | |
| Efavirenz based | 7 | 1262 | 4.75 | 0.88 (0.22, 3.64) | 0.998 | |
| Dolutegravir based | 6 | 924 | 5.41 | Ref | ||
PLHIV: People living with HIV; IPT: Isoniazid preventive therapy; Ref: Reference category; ART: anti-retroviral therapy; CI: Confidence interval
¥Crude rate ratio could not be calculated as the incidence rate in the unexposed was zero
*For two patients, CD4 cell count was not available
Sensitivity Analysis
Rate ratio was calculated by adding one case to the unexposed group to estimate the lowest relative risk of DILI and was found to be: (RR: 11.32; 95% CI: 1.68, 484.07).
Description of the DILI Cases
Cases of DILI occurred in ten females and two males. Their median ALT and ALP activity was 7.3 (IQR: 5.4, 13.8) and 0.85 (IQR: 0.63, 1.38), respectively. The median R-value of the cases was 8.8 (IQR: 6.3, 14.2). In all cases except one, the type of pattern of liver injury was hepatocellular. Severity of the DILI was mild in ten cases and in the other two it was moderate (Table 4). All but two cases were asymptomatic. The time to onset, the time between initiation of INH and first diagnosis of DILI, was 35 days (IQR: 26.8, 60 days). The association of DILI with IPT was probable in all cases. A symptomatic treatment, a vitamin B complex tablet, was provided to two patients with DILI, and none of the other cases received treatment. In all cases, DILI was reported as either recovered (n = 9) or recovering (n = 3). As part of management of DILI, only IPT was withdrawn in all the cases.
Table 4.
Overview of the cases who developed isoniazid-induced hepatic injury in PLHIV attending Halibet national referral hospital, followed-up between 1 March and 30 October 2021 (N = 552)
| Case no. | Age | Sex | Exposure status | ALT | ALP | Total bilirubin | INR | ALT ULN | ALP ULN | ALT activity | ALP activity | R-value | TTO | Symptomatic? | Pattern of liver injury | Severity |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| 1 | 57 | F | Exposed | 583 | 389 | 1 | 1.078 | 41 | 153 | 14.22 | 2.54 | 5.59 | 90 | No | Hepatocellular pattern | Mild |
| 2 | 38 | F | Exposed | 262 | 80 | 0.8 | N/A | 41 | 153 | 6.39 | 0.52 | 12.22 | 60 | No | Hepatocellular pattern | Mild |
| 3 | 51 | F | Exposed | 291 | 121 | 0.7 | N/A | 41 | 153 | 7.10 | 0.79 | 8.97 | 60 | No | Hepatocellular pattern | Mild |
| 4 | 54 | M | Exposed | 480 | 251 | 1 | N/A | 41 | 153 | 11.71 | 1.64 | 7.14 | 39 | No | Hepatocellular pattern | Mild |
| 5 | 41 | F | Exposed | 206 | 90 | 0.5 | N/A | 41 | 153 | 5.02 | 0.59 | 8.54 | 34 | No | Hepatocellular pattern | Mild |
| 6 | 42 | F | Exposed | 515 | 132 | 0.6 | N/A | 41 | 153 | 12.56 | 0.86 | 14.56 | 60 | No | Hepatocellular pattern | Mild |
| 7 | 49 | F | Exposed | 1,256 | 226 | 3.5 | N/A | 41 | 153 | 30.63 | 1.48 | 20.74 | 22 | Yes | Hepatocellular pattern | Moderate |
| 8 | 47 | F | Exposed | 207 | 90 | 0.5 | 1.019 | 41 | 153 | 4.93 | 0.59 | 8.38 | 26 | No | Hepatocellular pattern | Mild |
| 9 | 59 | M | Exposed | 839 | 161 | 0.8 | 1.136 | 41 | 153 | 20.46 | 1.05 | 19.45 | 36 | No | Hepatocellular pattern | Mild |
| 10 | 38 | F | Exposed | 1,115 | 165 | 1 | N/A | 41 | 153 | 27.20 | 1.08 | 25.22 | 27 | Yes | Hepatocellular pattern | Moderate |
| 11 | 59 | F | Exposed | 277 | 246 | 0.8 | N/A | 41 | 153 | 6.76 | 1.61 | 4.20 | 33 | No | Mixed pattern | Mild |
| 12 | 54 | F | Exposed | 309 | 89 | 0.4 | N/A | 41 | 153 | 7.54 | 0.58 | 12.96 | 27 | No | Hepatocellular pattern | Mild |
ALT: Alanine aminotransferase; ALP: Alkaline phosphatase; INR: International normalized ratio; ULN: Upper limit of normal; ALT activity: ALT/ULN; ALP activity: ALP/ULN; R-value: ALT activity/ALP activity; TTO: Time to onset (in days)
Survival Analysis
Kaplan–Meier estimation shows that the unexposed group had higher survival probability without DILI compared with the exposed (Fig. 3). At about five months follow-up, the probability of survival without DILI for the unexposed was 100% while for the exposed group, it was estimated to be around 93%.
Fig. 3.
Kaplan Meier curve showing probability of survival without drug-induced liver injury in PLHIV attending Halibet national referral hospital, 1 March–30 October 2021, Asmara, Eritrea. The figure shows that none of the unexposed group had developed drug-induced liver injury (DILI) during the study period
Discussion
In this real-world observational cohort study, a higher risk of DILI was observed in the exposed group but its statistical significance at a multivariate level could not be explored for reasons reported in the data analysis of the methods section. Nevertheless, considering the life-threatening nature of the outcome, observing 12 cases (4.2%) only in the exposed group is clinically important and of concern. All possible causes that could be potentially hepatotoxic were assessed and no alternative explanation was found for all cases. However, we acknowledge that other unmeasured confounders may still exist, which is an inherent limitation of such studies. Comparing this finding with studies reported elsewhere is a challenge due to variability of the results. A recent meta-analysis and systematic review of individual participant data from randomised controlled trials reported a 2.5% (range: 0.83–9.68%) incidence rate of hepatic injuries following the use of IPT in PLHIV [9]. While different observational studies that assessed the risk of IPT-induced hepatic injury in PLHIV reported relatively higher but variable incidences (4.9–22% [12, 16, 17]). The difference in cumulative incidences of DILI associated with IPT in PLHIV might be related to the diagnostic criteria used, study designs employed, length of use of IPT, duration and strictness of follow-up, other concomitant drugs used, patients’ clinical conditions and so on. Reporting inconsistent results of DILI due to differences in diagnostic criteria have been of concern [25], which could result in inappropriate decision making.
In this study, in almost all cases, the pattern of liver injury was hepatocellular; none of them were cholestatic. Such a pattern is in-line with INH-related hepatotoxicity. Besides, the majority of the cases were asymptomatic despite high levels of ALT and all except one developed the outcome within 2 months, which is consistent with a previously conducted local study [17]. This reflects that, in a situation where there is no close laboratory monitoring, patients might come to health facilities with severely elevated liver enzymes before they show clinical manifestations of liver injury. This could be even worse if the follow-up time is every 2–3 months, which was a common practice in Eritrea before and during the study period. Thus, relying mainly on clinical monitoring, as specified in national, International Union against Tuberculosis and Lung Diseases and WHO guidelines on implementation of IPT [8] could make it challenging and emphasises the need for regular laboratory monitoring (LFT) especially during the first 3 months of follow-up. A recently conducted qualitative study, aimed at determining the factors that affected implementation of IPT in Eritrea, revealed that inadequate laboratory setup, especially at subnational levels, was among the root causes for limited deployment of IPT in Eritrea [20].
Findings of a systematic review performed to understand the best clinical approach used in monitoring toxicity during the use of tuberculosis preventive therapies revealed that the existing monitoring strategies are not standardised and largely dependent on availability of resources, expert opinion and clinical practices [26]. The American Thoracic Society (ATS) and the Centers of Disease Control and prevention (CDC) of the USA, as well as the national guidelines of Portugal, Ireland and Sweden recommend at least monthly LFT monitoring during follow-up for PLHIV taking TB preventive therapy [26]. This is in-line with the recommendation of market authorization holders of isoniazid 300 mg tablet that advises regular monitoring of liver function (weekly during the first month, and then monthly) at least for the first 3 months of treatment [27, 28]. The systematic review also highlighted that the decision on intervals of LFT monitoring among different institutions/countries varied, which emphasizes the need to standardize frequency of LFT monitoring based on availability of resources, risk factors and so on. The systematic review also reflected that France and Canada recommend monthly LFT monitoring, regardless of their HIV status, for people on TB preventive therapy aged more than 50 and 65 years, respectively [26]. The Médicins Sans Frontierès guidelines on tuberculosis also advises that baseline LFTs can be performed in PLHIV starting TB preventive therapy, depending on the availability of resources, while follow-up tests are required if patients develop signs of hepatic injury [29].
This study showed that in all cases, because of fear of complications, IPT was discontinued even though the hepatic injury was mild in the majority. That being said, in randomised controlled trials, withdrawal of IPT following DILI was reported to be low (2.9–4.7% [13, 15]). This could possibly be related to the nature of the controlled environment that enabled investigators to closely monitor and follow-up patients. In our case, to minimise unnecessary treatment interruptions that could lead to suboptimal treatment outcomes and antibiotic resistance, efforts should be made to improve adherence to local and/or international guidelines on when to stop a suspected offending drug. A recently published guideline highlights that DILI could be transient and asymptomatic, and thus, interruption or change of treatment is not required unless there are signs of anorexia, malaise, vomiting or clinically evident jaundice [30].
The results of this study could have the following programmatic and policy implications. Deployment of IPT in PLHIV necessities regular laboratory monitoring of LFTs, especially in the first 3 months of treatment to ensure safer use of the product. This urge ensuring the existing laboratory setup have the capacity, including those in the remote areas, to minimise IPT implementation practicality challenges. Taking into consideration that the cases were asymptomatic and all recovered following withdrawal of the offending drug (IPT), someone might argue on the need for or cost-effectiveness of introduction of routine laboratory monitoring. It should, however, be well noted that the cases were closely monitored and intervention was discontinued when the liver function markers were noticeably elevated. In routine practice, patients are followed-up every 3 months which exposes them to the risk of liver injury complications, including death. Though implementation of routine laboratory monitoring might necessitate a formal cost-effectiveness assessment, the authors believe that it is applicable for the following reasons: the number people living HIV on ART in Eritrea is low (around 9000); the national HIV/AIDS control programme is well funded which could be helpful in capacitating the health system; the requested routine laboratory monitoring is short (at least 3 months); and the majority of the HIV patients (> 75%) are attending in the national and regional referral hospitals where a chemistry laboratory is available.
The study had the following limitations. Even though sample size was reasonably calculated, the fact that all of the cases were from the exposed group made the calculation of RR/HR not possible. Accordingly, a sensitivity analysis was performed by adding one case to the unexposed group to predict the lowest possible RR. Besides, considering the risks of INH-induced liver injury, participants exposed to IPT might have been closely followed, rigorously investigated, and strictly advised, compared to the unexposed, which might have affected the diagnosis rate of the outcome of interest. Efforts were made to follow the same protocol, for both groups, but the data collectors might have intuitively introduced a diagnostic suspicion bias. Due to shortage of laboratory consumables, impacted by COVID-19 restrictions, monitoring LFTs for all patients for 6 months – until they complete their IPT – was a challenge. Taking the short time-to-onset of DILI, namely a median of 52 days [31, 32], the authors decided to follow-up patients for a median of 4 months, and thus, we might have missed cases with delayed time to onset. Rarely, in situations where ALT test was not available, other liver enzymes such as ALP and/or aspartate transaminase (AST) along with total bilirubin were monitored, which might have underestimated the risk. The fact that the study was conducted in one national referral hospital limits generalizability of the results to other treatment sites.
Conclusions
Close liver function laboratory monitoring by means of liver function tests, especially during the first 3 months of treatment, is important to safely administer the product. This is because DILI in PLHIV taking IPT was commonly reported, and it was asymptomatic in the majority of the cases despite high levels of ALT. To ensure this, existing laboratory setups need to be further capacitated to minimise challenges associated with IPT implementation.
Acknowledgements
The authors sincerely thank Dr Kahsay Michael, Pharmacist Yohana Teklu, Sr. Bisrat Gebrelul and Sr. Alganesh Hadgu for their outstanding contribution in enrolment and monitoring of patients and data collection. The study participants are grateful for their time and willingness to participate. The authors would also like to acknowledge the staff of the National Medicines and Food Administration of Eritrea namely: Merhawi Bahta, Merhawi Debesai, Sirak Tesfamariam, Marta Amanuel, Kisanet Dawit and Feven Ghebreberhan for their assistance in the data management. Finally, the authors sincerely thank the staff of the HIV control programme (Sr Nigisty Tesfamicael and Tadesse Kidane) for their collaboration and review comments.
Declarations
Funding
This research work was self-funded by the lead author.
Conflict of Interest
None declared.
Availability of Data and Material
All information gathered is included in the manuscript and additional information can be obtained from the corresponding author via: satiswt@gmail.com or m.russomghebremedhin@erasmusmc.nl.
Ethical Considerations
Ethical approval to conduct the study was obtained from the Health Research Ethics and Protocol Review Committee of the Ministry of Health of the State of Eritrea (reference number: 7-18/2020). Written consent was obtained from the study participants and all ethical and professional considerations were followed to ensure patient confidentiality. Thus, only anonymized information is reported. The study was performed in accordance with the ethical standards as laid down in the latest version (2013) of the 1964 Declaration of Helsinki.
Consent for participate
Prior to enrollment, written consent was obtained from each study participant to take part in the study and their anonymised data to be reported in an international arena.
Consent for publication
Not required.
Code availability
Not required.
Author Contributions
The idea was conceived by MR, KV and BHS. All authors made a significant contribution in the design and follow-up of the study and interpretation of the results. Data was analysed by HGW. Data collection, supervision and data analysis processes were led by MR. The manuscript was drafted by MR and critically reviewed and edited by the rest of the authors. Finally, all the authors gave approval of the manuscript to be published in the journal and agreed to be accountable for all aspects of the work.
References
- 1.World Health Organization . Guidelines for intensified tuberculosis case-finding and isoniazid preventive therapy for people living with HIV in resource-constrained settings. Geneva: WHO; 2011. [Google Scholar]
- 2.Metushi I, Uetrecht J, Phillips E. Mechanism of isoniazid-induced hepatotoxicity: then and now. Br J Clin Pharmacol. 2016;81(6):1030–1036. doi: 10.1111/bcp.12885. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3.Meng X, Maggs JL, Usui T, et al. Auto-oxidation of isoniazid leads to isonicotinic-lysine adducts on human serum albumin. Chem Res Toxicol. 2015;28:51–58. doi: 10.1021/tx500285k. [DOI] [PubMed] [Google Scholar]
- 4.Metushi IG, Sanders C, Acute Liver Study G. Lee WM, Uetrecht J. Detection of anti-isoniazid and anti-cytochrome P450 antibodies in patients with isoniazid-induced liver failure. Hepatology. 2014;59:1084–1093. doi: 10.1002/hep.26564. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Ramappa V, Aithal GP. Hepatotoxicity related to anti-tuberculosis drugs: mechanisms and management. J Clin Exp Hepatol. 2013;3:37–49. doi: 10.1016/j.jceh.2012.12.001. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.Lee KK, Fujimoto K, Zhang C, et al. Isoniazid-induced cell death is precipitated by underlying mitochondrial complex I dysfunction in mouse hepatocytes. Free Radic Biol Med. 2013;65:584–594. doi: 10.1016/j.freeradbiomed.2013.07.038. [DOI] [PubMed] [Google Scholar]
- 7.Metushi IG, Cai P, Zhu X, Nakagawa T, Uetrecht JP. A fresh look at the mechanism of isoniazid-induced hepatotoxicity. Clin Pharmacol Ther. 2011;89:911–914. doi: 10.1038/clpt.2010.355. [DOI] [PubMed] [Google Scholar]
- 8.World Health Organization. Adverse events associated with the use of isoniazid preventive therapy among people living with HIV. WHO information note; 2015, Geneva, Switzerland.
- 9.Ross JM, Badje A, Rangaka MX, et al. Isoniazid preventive therapy plus antiretroviral therapy for the prevention of tuberculosis: a systematic review and meta-analysis of individual participant data. Lancet HIV. 2021;8(1):e8–e15. doi: 10.1016/S2352-3018(20)30299-X. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10.Akolo C, Bada F, Okpokoro E, Nwanne O, Iziduh S, Usoroh E, et al. Debunking the myths perpetuating low implementation of isoniazid preventive therapy amongst human immunodefciency virus-infected persons. World J Virol. 2015;4(2):105–112. doi: 10.5501/wjv.v4.i2.105. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.Grant AD, Mngadi KT, van Halsema CL, Luttig MM, Fielding KL, Churchyard GJ. Adverse events with isoniazid preventive therapy: experience from a large trial. AIDS. 2010;24(Suppl5):S29–36. doi: 10.1097/01.aids.0000391019.10661.66. [DOI] [PubMed] [Google Scholar]
- 12.Tedla Z, Nguyen ML, Sibanda T, Nyirenda S, Agizew TB, Girde S, et al. Isoniazid-associated hepatitis in adults infected with HIV receiving 36 months of isoniazid prophylaxis in Botswana. Chest. 2015;147(5):13761384. doi: 10.1378/chest.14-0215. [DOI] [PubMed] [Google Scholar]
- 13.Rangaka MX, Wilkinson RJ, Boulle A, Glynn JR, Fielding K, van Cutsem G, et al. Isoniazid plus antiretroviral therapy to prevent tuberculosis: a randomised double-blind placebo-controlled trial. Lancet. 2014;384(9944):682–690. doi: 10.1016/S0140-6736(14)60162-8. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14.Le Roux SM, Cotton MF, Myer L, le Roux DM, Schaaf HS, Lombard CJ, et al. Safety of long-term isoniazid preventive therapy in children with HIV: a comparison of two dosing schedules. Int J Tuberc Lung Dis. 2013;17(1):26–31. doi: 10.5588/ijtld.11.0820. [DOI] [PubMed] [Google Scholar]
- 15.Ngongondo M, Miyahara S, Hughes MD, et al. Hepatotoxicity during isoniazid preventive therapy and antiretroviral therapy in people living with hiv with severe immunosuppression: a secondary analysis of a multi-country open-label randomized controlled clinical trial. J Acquir Immune Defic Syndr. 2018;78(1):54–61. doi: 10.1097/QAI.0000000000001641. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16.Mudzviti T, Shamu T, Chimbetete C, Munengerwa T, Bote S, Pascoe M. Tolerability of isoniazid preventive therapy in an hiv-infected cohort of paediatric and adolescent patients on antiretroviral therapy from a resource-limited setting: a retrospective cohort study. Drugs Real World Outcomes. 2019;6:37–42. doi: 10.1007/s40801-019-0147-3. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17.Russom M, Berhane A, Debesai M, Andom H, Tesfai D, Zeremariam Z, et al. Challenges of hepatotoxicity associated with isoniazid preventive therapy among people living with HIV in Eritrea. Adv Pharmacoepidemiol Drug Saf. 2019;8(2):2167. [Google Scholar]
- 18.van Griensven J, Choun K, Chim B, Thai S, Lorent N, Lynen L. Implementation of isoniazid preventive therapy in an HIV clinic in Cambodia: high rates of discontinuation when combined with antiretroviral therapy. Trop Med Int Health. 2015;20(12):1823–1831. doi: 10.1111/tmi.12609. [DOI] [PubMed] [Google Scholar]
- 19.Russom M, Debesai M, Zeregabr M, Berhane A, Tekeste T, Teklesenbet T. Serious hepatotoxicity following use of isoniazid preventive therapy in HIV patients in Eritrea. Pharmacol Res Perspect. 2018;6:e00423. doi: 10.1002/prp2.423. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20.Russom M, Jeannetot DYB, Tesfamariam S, Stricker BH, Verhamme K. Perspectives of healthcare professionals on factors limiting implementation of isoniazid preventive therapy in people living with HIV in Eritrea: a qualitative study. Risk Manag Healthc Policy. 2022;15:1407–1419. doi: 10.2147/RMHP.S368315. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21.Aithal GP, Watkins PB, Andrade RJ, et al. Case definition and phenotype standardization in drug-induced liver injury. Clin Pharmacol Ther. 2011;89:806–815. doi: 10.1038/clpt.2011.58. [DOI] [PubMed] [Google Scholar]
- 22.Internet. https://www.cdc.gov/epiinfo/index.html. Accessed Mar 2021.
- 23.Shubber Z, Calmy A, Andrieux-Meyer I, et al. Adverse events associated with nevirapine and efavirenz-based first-line antiretroviral therapy: a systematic review and meta-analysis. AIDS. 2013;27(9):1403–1412. doi: 10.1097/QAD.0b013e32835f1db0. [DOI] [PubMed] [Google Scholar]
- 24.R Core Team. R: a language and environment for statistical computing. R Foundation for Statistical Computing, Vienna, Austria, 2020. https://www.Rproject.org/. Accessed Feb 2022.
- 25.Kunst H, Khan KS. Age-related risk of hepatotoxicity in the treatment of latent tuberculosis infection: a systematic review. Int J Tuberc Lung Dis. 2010;14(11):1374–1381. [PubMed] [Google Scholar]
- 26.Sotgiu G, Matteelli A, Getahun H, et al. Monitoring toxicity in individuals receiving treatment for latent tuberculosis infection: a systematic review versus expert opinion. European Respiratory Journal. 2015 45: 1170-1173. 10.1183/09031936.00216814. URL: http://ow.ly/HyjAH. [DOI] [PMC free article] [PubMed]
- 27.Antibiotice S. A. Summary of product characteristics of isoniazid 300 mg tablet (TB266). Updated June 2014. (Accessed on 25 Jul 2022)
- 28.Micro Labs Limited. Summary of product characteristics of isoniazid 300 mg tablet (TB174). WHOPAR part 4. Updated February 2019.
- 29.Internet. Medicines Sans Frontiers. Medical Guidelines: Tuberculosis. URL: https://medicalguidelines.msf.org/viewport/TUB/latest/chapter-16-treatment-of-latent-tuberculosis-infection-20322864.html (accessed 19 Apr 2022)
- 30.World Health Organization. WHO consolidated guidelines on tuberculosis. Module 1: prevention–tuberculosis preventive treatment. Geneva: 2020. [PubMed]
- 31.Wang W, Lie P, Guo M, He J. Risk of hepatotoxicity in cancer patients treated with immune checkpoint inhibitors: a systematic review and meta-analysis of published data. Int J Cancer. 2017;141:1018–1028. doi: 10.1002/ijc.30678. [DOI] [PubMed] [Google Scholar]
- 32.Huffman BM, Kottschade LA, Kamath PS, Markovic SN. Hepatotoxicity after immune checkpoint inhibitor therapy in melanoma: natural progression and management. Am J Clin Oncol. 2018;41:760–765. doi: 10.1097/COC.0000000000000374. [DOI] [PubMed] [Google Scholar]