TB treatment and HBV: risk factors for liver toxicity in South Africans on antiretroviral therapy

Hepatotoxicity

Study participants were followed up for a median of 239 days on ART (a combined 519 person years) and had a median of eight liver enzyme assays over the study period. For those who began with ALT and AST values in the normal range, grade 2 (moderate) hepatotoxicity was defined as either an ALT or AST value greater than 2.6 times the upper limit of normal (ULN = 40 iu/l), and severe hepatotoxicity (grade 3 or 4) a value of greater than five times the upper limit of normal. However, for the 196 (23%) whose baseline values were already above normal at baseline, grade 2 toxicity comprised values 2.6 or more times the baseline value, and severe toxicity as values 3.6 times higher than baseline or greater. Study reporting was based on the first incident of severe toxicity.

By these definitions, hepatotoxicity of grade 2 or higher occurred in 97 participants (11%) after starting ART – an incidence rate of 19.7 episodes per 100 person-years (95% confidence interval [CI], 16.1 to 24.0). Severe (grade 3 or 4) toxicity was seen in 40 subjects (4.6%) for a rate of 7.7 events per 100 person-years (95% CI, 5.6 to 10.5). Severe episodes were first seen a median of 57 days after treatment initiation. The clinical impact of the severe episodes was described as “modest”; 32 of the 40 remained on unchanged therapy (even in the three who had repeated severe episodes) and only eight (20%) changed their first-line regimen as a result.

Nine (22%) of the 40 people with severe episodes were hospitalised (although the admissions were not necessarily for that reason). Two deaths occurred within 28 days of a severe episode; one was due to liver failure (in a person being treated for tuberculosis) and the other to kidney disease.

Notably, during the twelve months before ART initiation, liver enzymes greater than five times ULN were seen in 34 of the participants (4%). These elevations did not predict severe hepatotoxicity after ART began, and there was no statistical difference in the proportion of those with hepatotoxicity before and after ART initiation (p = 0.55).

Liver toxicity appears to be fairly infrequent and non-severe among HIV-positive individuals on antiretroviral therapy in South Africa. According to a retrospective cohort study of 868 South Africans, severe hepatotoxicity occurred at a rate of 7.7 episodes per 100 person years – comparable with findings of studies in other areas. Tuberculosis treatment and hepatitis B were the greatest risk factors for episodes of severe toxicity. The findings were published in the June 2007 issue of AIDS.

Toxicities and side-effects due to antiretroviral treatment have not been as widely studied in Africa and other developing countries as they have been in the developed world. This study, conducted by a South African research team with collaborators from the US and the UK, investigated liver toxicity among 868 participants in a workplace HIV care program in South Africa. All were on their first antiretroviral treatment (ART) regimen, initiated between November 2002 and December 2005. (Another 54 individuals in the programme who started treatment during this time either declined participation, or did not have liver enzyme data available.)

The median age of the study participants was 41 years; nearly all (94%) were male. The median nadir (lowest recorded) CD4 count was 136 cells/mm³, and the median viral load was 4.7 log₁₀ copies/ml. Importantly, the ART regimens were efavirenz-based in 825 (95%) and nevirapine-based (a drug associated with liver toxicities) in only 39 (5%).

Risk factors

By multivariate analysis, only two factors significantly increased the risk of severe hepatotoxicity: receiving TB treatment, and positive serology for hepatitis B antigen (HBsAg). A nadir CD4 cell count of less than 100 cells/mm³ was marginally associated; baseline ALT levels were mildly predictive in univariate but not in multivariate analysis.

Four hundred and ninety-five of the participants (57%) had a history of tuberculosis (TB) before beginning ART. During the first year of ART, 636 (74%) received TB prophylaxis (isoniazid or cotrimoxazole) and 214 (25%) received TB therapy. Receiving any form of TB treatment increased the risk of severe hepatotoxicity 8.5-fold (95% CI, 2.7 to 27; p < 0.001).

Serology for HBV was available for a sample of 387 participants. Of these, 80 (20.7%) were positive for HBV surface antigen (HBsAg); this was associated with a threefold increased risk of severe hepatotoxicity (95% CI, 1.3 to 7.0; p = 0.016).

A CD4 cell count nadir of below 100 cells/mm³ showed a borderline association, with a little under twofold risk (95% CI, 0.96 to 3.8; p=0.052).

This study was significantly limited by including almost exclusively employed, male participants taking efavirenz-based rather than nevirapine-based therapy, and only the first year of treatment was observed. However, at least among this population, the study team found “a low incidence of and minimal morbidity due to hepatotoxicity”, with limited morbidity “even among patients with [HBV] or on anti-tuberculosis therapy”, although these were significant risk factors for liver enzyme elevation. The researchers believe their findings “support an approach of treating patients with ART even if they also require tuberculosis therapy or have a positive HBsAg test.”