0916 Threshold for initiation of Antiretroviral Therapy (ART)

Given the burden of maternal HIV infection in South Africa today, the significant risk of transmission of the virus, the number of infected children and the body of evidence showing physical and neurodevelopmental impact f HIV/AIDS, one would expect fairly strong support for any studies that argue in favour of earlier rather than later initiation of drug treatment with anti-retrovirals. One such study, which appeared in a recent edition of the Lancet, reviewed 18 observational cohort studies that involved some 45000 patients from two ‘era’s’; approximately 50% of subjects before the availability of combination treatment and 50% after. The research question was whether there were differences in AIDS and mortality if treatment was started later (CD4 counts 251-351) vs at counts of 351-400cells per μl. Using Kaplan-Meier estimates of cumulative probability of progression to AIDS and death, and Cox regression analysis to estimate Hazard Ratios the authors found that delayed initiation of ART was associated with more AIDS and mortality and they therefore suggested that a CD4 count of 350 should be the minimum threshold for initiation of ART. This is in keeping with the South African sentiment that initiation of ART should be much sooner than at a CD4 count of 400 and it is therefore perhaps a little surprising to see the accompanying editorial written by a South African that seems to caution somewhat against the findings. While making the statement that local data show unusually high rates of AIDS and death with CD4 counts in the 200-350 range and that the range of ARTs is more restrictive, the editorial also cautions against relying on observational data that might focus on AIDS and death while ignoring other factors such as quality of life and cardiovascular, renal and hepatic complications of (earlier) treatment. It will be interesting to see whether those involved in HIV/AIDS research will go the route of accumulating sufficient data on the basis of randomized controlled trials vs accepting the combination of newer, less toxic drugs plus data from 18 well-analysed observational studies.

Read more:
Lancet 2009; 373: 1352-63 and 1314-6
J Infect Dis 2008; 197: 1133-44
Lancet 2006; 368: 1254-9