You can't trust everything you read on the Internet 22 March 2004
Previous Rapid Response Next Rapid Response Top
Brian T Foley,
HIV Researcher
Los Alamos National Lab, Los Alamos, NM 87545

Send response to journal:
Re: You can't trust everything you read on the Internet

Mark Bartlett claims that it "would seem" that most infants infected with HIV in South Africa were infected by dirty needles, transfusions or other mechanisms because he claims that Rodney Richards claims that only 65,000 of some 765,000 infected infants could possibly have been infected by mother-to-infant transmission.

I have no idea where Mark got his information from. Has Rodney Richards published his study? In at least a few hospitals in South Africa, the HIV-1 seroprevalence rate for antenatal women is near 20%. If 25% of them pass the virus to their newborns due to the lack of even the simplest and cheapest drug regimens (such as single-dose nevirapine), then the rate among the infants would be expected to be 5%.

Only 2,348 children were tested in the Nelson Mandella/HSRC 2002 study, so I would caution against attempting to extrapolate the results from such a small number to the entire population of South Africa. The true seroprevalence rate among all children in South Africa is probably not exactly 5.6%, this is just the percentage found among those sampled in the study. It is not adjusted for any biases in the sampling that was done in the study.

I am very curious to know how Rodney Richards computed that only 65,000 children could be infected in South Africa by mother to infant transmission, most estimates I have heard have put the figure at closer to 50,000 infants per year in the 2000-2004 time period.

Competing interests: None declared