HIV is the strongest risk factor for developing tuberculosis (TB) disease in those with latent or new mycobacterium tuberculosis infection. The risk of developing TB is between 20 to 37 times greater in people living with HIV (PLHIV) than among those who do not have HIV infection [1]. Morbidity and mortality from drug-sensitive and drug-resistant TB among PLHIV are unacceptably high [2,3].

Highly active antiretroviral treatment (HAART) significantly but not entirely reduces the risk of TB disease [4]. In HIV infected patients, isoniazid preventive therapy (IPT) reduces reactivation of latent TB infection, both in industrialized countries [5–7] as well as in developing countries [8-10]. In 1993, the World Health Organization (WHO) first issued a policy statement that recognized the efficacy of TB preventive therapy and recommended IPT for PLHIV [11-13]. The WHO recommended regimen for TB preventive therapy in PLHIV is isoniazid for at least 6 months [14].

The IPT program has since been implemented worldwide. The outcome of IPT has been predominantly determined in clinical trials and observational studies in African countries. Data on feasibility, efficacy, safety and completion rates of IPT among PLHIV in Asia are limited. There are specific concerns about adverse events of isoniazid in those with concurrent HAART.

In this study we assessed the demographics and compliance of PLHIV taking IPT in two hospital-based HIV clinics in Penang, Malaysia. We also aimed to determine the rates and contributory factors of adverse events, as well as the incidence of active TB within 1 year after initiation of IPT.

Study design and population

This retrospective cohort study was conducted at two HIV specialist clinics in Penang General Hospital and Seberang Jaya Hospital respectively, where IPT has been implemented since 2011. Subjects included HIV patients who were commenced on IPT between 2011 and 2014, with or without concurrent HAART.

Interventions

All subjects underwent TB symptom screening (fever, cough, weight loss and night sweats for the past 4 weeks) and further investigations (including chest X-ray) if indicated. Chest X-ray was not a mandatory screening before initiating IPT. Tuberculin skin test (Mantoux) was not required for IPT eligibility. After ruling out active TB, IPT consisting of isoniazid 300 mg and pyridoxine 20 mg daily was started for 6 month duration in accordance with WHO recommendations. HAART was initiated before, during or after IPT based on indication and clinical appropriateness. Follow-up visits during IPT were monthly to three monthly. Patients were monitored for TB symptoms, compliance and adverse events including elevated alanine transaminase (ALT). Upon completion of IPT, patients would resume routine HIV follow-up visits with regular screening of TB symptoms.

Outcome

The outcome of interest in this study was the IPT completion rate, incidence of adverse events and occurrence of active TB infection at 1 year follow-up. Patients who were lost to follow-up or expired (due to causes not related to IPT) were excluded from the adverse events and TB occurrence analysis. Patients who were transferred to other hospitals were excluded from all outcome analysis.

Hepatotoxicity was defined as ALT level of at least 2 times the upper limit of normal (ULN) [15]. The severity of hepatotoxicity was graded based on a system defined by AIDS Clinical Trials Group (ACTG) (Table 1).

Table 1: AIDS Clinical Trials Group (ACTG) Severity Grading of Drug Induced Liver Injury.

Data collection and analysis

Information on patient demographics, baseline characteristics (CD4 count, HAART regimen, ALT level, viral hepatitis co-infection etc.) and outcome was retrieved from patients’ clinic records and data sheets. Data integrity was assessed and descriptive statistics were generated. Relative risks (RR) were determined for the cohort analysis of IPT non-completion and adverse events.

We started IPT for 242 patients, of whom 204 (84.3%) were males and their mean age was 37.4 ± 10.9 years. Their median baseline CD4 count was 315.5 cells/uL (IQR, 327); 31.4% had a CD4 count less than 200 cells/uL and 72.7% were receiving HAART during IPT. Baseline chest X-rays were done for 85 (35.1%) patients. Majority of patients (229/242, 94.6%) had normal baseline ALT level (< ULN), while 12 (5%) patients had slightly higher ALT level (1-1.9 x ULN) at the start of IPT. Only 1 patient had baseline ALT more than 2 times the ULN. Twenty nine (12%) patients were co-infected with viral hepatitis (10 had Hepatitis B, 18 had Hepatitis C, and 1 had both Hepatitis B and C) (Table 2).

Table 2: The socio-demographic and baseline characteristics of study cohort

HAART regimens that patients received consisted of Tenofovir- Emtricitabine, Efavirenz (67.1%), Zidovudine- Lamivudine, Efavirenz (14.2%), Zidovudine-Lamivudine, Nevirapine (7.4%), Tenofovir- Emtricitabine, Raltegravir (4.6%) and others (Table 3).

Table 3: Concurrent HAART during IPT.

Excluding patients who were transferred to other hospitals, 193 (81.1%) patients completed IPT (Table 4). Major reasons for noncompletion of IPT were adverse events (46.7%), defaulting from followup (37.8%) and poor compliance (6.7%). One patient passed away during 2nd month of IPT due to septic shock, which was not related to TB or IPT. Another patient had interruption of IPT medications due to prescription error. One patient stopped IPT due to pill burden (he was receiving concomitant HAART) and another was worried about side effects (Figure 1).

Table 4: Summaryof study outcomes.

Figure 1: Reasons for non-completion of PTI.

Completion rate was higher among patients with concurrent HAART (149/173, 86.1%) compared to HAART naïve patients (44/65,67.7%) (P=0.001). Co-administration of HAART significantly reduced the incidence of non- completion of IPT (RR, 0.43; 95% confidence interval [CI], 0.3-0.7; P< 0.005). Age greater than 35 years, male gender and CD4 count less than 200 cells/uL were not associated with non-completion of IPT (Table 5).

Table 5: Risk factors associated with non-completion of IPT.

Adverse Events

Forty patients (18%) developed adverse events during IPT (Table 4). There were 19 incidences of hepatotoxicity, 12 rashes, 6 gastrointestinal (GI) symptoms (diarrhea, nausea), 5 peripheral neuropathy and 2 central nervous system (CNS) symptoms (sleepiness, lightheadedness) (Figure 2).

Figure 2: Adverse events 7occurred during IPT.

The incidence rate of hepatotoxicity was 8.56%. Significant risk factors for hepatotoxicity during IPT were Hepatitis B and/or C coinfection (RR, 3.99; 95% CI, 1.7-9.4; P< 0.01) and baseline ALT level = ULN (RR, 8.85; 95% CI, 4.1-19.0; P< 0.001). Concurrent HAART, age greater than 35 years, male gender and CD4 count less than 200 cells/ uL were not associated with the occurrence of hepatotoxicity (Table 6).

Table 6: Risk factors associated with hepatotoxicity during IPT.

The severity of liver injury sustained was either Grade 1 (6 patients) or Grade 2 (13 patients), all of which were asymptomatic and reversible. Ten patients (52.6%) out of those who developed hepatotoxicity managed to complete IPT without complication despite significant elevation of ALT level (Table 7).

Table 7: Characteristics of study cohort who developed hepatotoxicity during IPT (N=19).

Skin rash occurred at a rate of 5.41%. There was no significant predisposing factor for the development of rashes (Table 8). No correlation analysis was done for other adverse events as the incidence was low.

Table 8: Risk factors associated with skin rash during IPT.

TB Occurrence

None of the patients who were exposed to IPT (including those who stopped IPT prematurely) developed active TB infection at 1 year follow-up (Table 4).

Sub-analysis on cohort who were lost to follow-up

There were 15 (6.2%) patients who lost to follow-up during IPT. This was significantly associated with patients co-infected with Hepatitis C (RR, 3.73; 95% CI, 1.4-9.8; P< 0.05). Patients with concurrent HAART were less likely to be lost to follow-up (RR, 0.25; 95% CI, 0.1-0.6; P< 0.01). There was no clear association with age, gender or baseline ALT (Table 9).

Table 9: Risk factors associated with patients lost to follow-up during IPT.

Studies have consistently shown early initiation of HAART and immune reconstitution prolongs life expectancy of HIV patients, to the extent of approaching that of general population [16]. The START (Strategic Timing of Anti- Retroviral Treatment) study, a large-scale randomized clinical trial has established that early antiretroviral treatment at CD4 count above 500 cells/uL significantly reduces the rates of serious AIDS and non-AIDS events [17]. Many HIV guidelines including the WHO have been revised to recommend early initiation of HAART regardless of presenting CD4 cell count or clinical stage [18].

In our study, 72.7% of our patients were started on HAART before or during IPT. As majority of HIV patients are receiving HAART, it is important to study the outcome of co-administration of IPT with HAART in terms of compliance and adverse events.

The overall completion rate of IPT in our study was acceptable and comparable to previous studies. Despite the increase in pill burden, IPT was well tolerated with concurrent HAART. In fact, there was a significantly higher completion rate in patients with concurrent HAART. We attributed this observation to the pre-HAART counseling which emphasized on strict drug adherence. Most of these patients were mentally prepared and committed to be on long term medications.

Although the benefit of IPT among PLHIV is well established, there were concerns about adverse events of isoniazid, particularly when co-administered with HAART. Isoniazid and NNRTI (nonnucleoside reverse transcriptase inhibitor), a commonly used class of antiretroviral drug, shared similar side effect profiles. They are known to cause hepatotoxicity and rash. A prospective study in Botswana of 1,995 PLHIV on IPT observed 1.1% rate of severe isoniazid-associated hepatitis, which was similar to rates in HIV-uninfected populations. HAART was received by 480 participants but was not statistically associated with isoniazid induced hepatitis (RR, 1.56; 95% CI, 0.62- 3.9) [19]. A South African study of 818 men and 50 women reported no statistically significant association between Grade 3 or 4 hepatitis in PLWH receiving IPT and efavirenz based anti-retroviral therapy (hazard ratio,0.83, 95% CI, 0.20–3.4) [20]. These observations were consistent with our study finding that patients concomitantly taking HAART and IPT had no significantly increased risk of hepatotoxicity (RR, 0.73; 95% CI, 1.8-0.3), despite majority of them were on NNRTI (efavirenz or nevirapine) based HAART. In addition, all our reported hepatotoxicity was either mild or moderate (severity Grade 1 to 2) and reversible. More than half of our patients with elevated ALT managed to complete IPT without any complications. The incidence of skin rash was low and not associated with co-administration of HAART.

Patients with Hepatitis B/C co-infection or baseline ALT more than ULN had to be closely monitored for risk of hepatotoxicity as shown in our study. However, these patients should not be denied IPT as the benefit still outweighed the risk.

Majority of our cohort had baseline CD4 count less than 350 cells/ uL and HAART was strongly indicated. Thirty- one point four percent (31.4%) of them had CD4 count less than 200 cells/uL, which rendered them very susceptible to TB reactivation. A meta-analysis of randomized clinical trials that assigned 7619 HIV patients to IPT or placebo found an overall 35% of TB risk reduction (RR,0.65, 95% CI, 0.51-0.84) in all participants [21]. Our cohort who received IPT did not develop active TB infection up to 1 year of follow-up. However, we could not attribute this finding solely to IPT as this was not a comparative study against IPT naïve patients and most of our patients were co-administered with HAART. It is well accepted that immune reconstitution after HAART reduced the risk of reactivation of TB. In a cohort study in South Africa, the number of TB cases averted by HAART was greatest in patients with WHO stage 3 or 4 (18.8 averted cases per 100 patientyears, adjusted rate ratio 0.22 [0.09–0.41]; p=0.03) and in those with CD4 count of less than 200 cells/µL (14.2 averted cases per 100 patientyears, adjusted rate ratio 0.18 [0.07–0.47]; p< 0.0001) [4]. In addition, evidence from two retrospective studies that examined the benefit of TB prevention in PLHIV receiving HAART and IPT indicated that the protection conferred by the combination was additive [22,23].

Six point two percent (6.2%) of patients were lost to followup during IPT. A significant proportion of these patients were coinfected with Hepatitis C. They comprised of ex-intravenous drug users with poor social- economic background and in general have a tendency to default treatment. We also suspect a few of them may have developed adverse events, which led to self-discontinuation of IPT and subsequent were lost to follow-up. The phenomenon of patients with concurrent HAART having significant lower risk of defaulting follow-up likely reflects the effectiveness and importance of our pre-HAART counseling.

Although international guidelines on the timing of initiation of IPT and HAART are unsettled, we suggest that IPT can precede HAART initiation if a patient’s CD4 count is still high (more than 350 cells/uL). Commencing IPT first would allow HAART naïve patients to undergo pill training prior to initiation of HAART, which requires life-long drug adherence and compliance. Conversely, in patients with low CD4 count (less than 350 cells/uL) or have previous opportunistic infections, immune reconstitution should be the priority. Hence IPT should be deferred until stabilization on HAART.

IPT is feasible and relatively safe among our local PLHIV. IPT and HAART can be co-administered to give additive prophylaxis against reactivation of TB and has not been shown to compromise safety or compliance in our study. Larger and longer comparative studies are needed to determine the outcome of IPT in our local HIV population. Other potential risk factors of isoniazid-hepatotoxicity such as alcohol and bactrim usage should be evaluated. Future studies also need to address the risk of developing drug-resistant TB after exposure to IPT.

The authors declare that there is no conflict of interest regarding the publication of this manuscript.