IAS: High risk of death for South Africans in first six months of ART, isoniazid may reduce risk

South African miners receiving antiretroviral therapy (ART) are just as likely to die during the first six months of treatment as their untreated counterparts, researchers from the London School of Hygiene and Tropical Medicine reported last week at the Third International AIDS Society Conference on HIV Treatment and Pathogenesis in Rio de Janeiro, Brazil. The reduction in death and illness associated with ART only begins to kick in after six months of treatment, the researchers said.

Many African clinicians have reported a high rate of mortality in patients starting ART in their clinics. The observation has led some to argue that less sick patients should be prioritised for ART, and so the London School of Hygiene group also set out to determine whether there are factors that place individuals starting ART at higher risk of death during the early months of treatment.

Several factors have been suggested as especially problematic:

• Starting treatment very late, often with a CD4 cell count below 50 cells/mm³.
• Active opportunistic infections, especially tuberculosis.
• Immune reconstitution syndrome, particularly due to prior or current infection with tuberculosis (TB).

The research group carried out a retrospective analysis of all individuals who had received ART through a workplace health programme in a goldmining company in Kwazulu Natal. They compared the risk of death with historical data from the same workplace treatment programme prior to the introduction of ART.

The two cohorts were well matched, with 649 on treatment and 679 untreated. The median age was around 40 in both groups, and the median baseline CD4 cell count was 140 cells/mm³ in the treated group and 188 cells/mm³ in the untreated group. Median follow-up was approximately one year in each group.

There was no significant difference in the relative hazard of death during the first six months of treatment compared to the historical control group after adjustment for baseline CD4 cell count (HR 1.07), but the risk of death fell dramatically between months 6 and 12 (HR 0.29, 95% confidence interval 0.14 - 0.57) and after month 12 (HR 0.11, CI 0.03 – 0.37) in the treated group. The reduction in mortality rate per 100 person years of follow up was 1.8 deaths.

The only factor significantly associated with an increased risk of death in multivariate analysis was having a baseline CD4 cell count below 50 cells/mm³. Receiving isoniazid preventive treatment for TB showed a trend towards a protective effect however (HR 0.49, CI 0.22 – 1.07), whilst univariate analysis showed that prior TB was also associated with an increased risk of death.

In a related presentation the same group reported on virological responses in 449 individuals receiving ART in the same cohort. Paradoxically, they found that individuals with more advanced HIV disease showed a trend towards a better virological response, whether disease stage was defined by CD4 cell count below 50 cells/mm³ or World Health Organization (WHO) disease stage 4. However baseline viral load above 50,000 copies/ml significantly increased the risk of virological failure at six months (HR 1.80), as did any report of non-adherence during the follow-up period. Seventy-four per cent had viral load below 400 copies/ml at month 6.