Re: The "HIV" antibody tests are not diagnostic 26 October 2004
Previous Rapid Response Next Rapid Response Top
Nicholas Bennett,
Infectious Disease Postdoc/Clinician
Department of Pediatrics, University Hospital, Syracuse NY

Send response to journal:
Re: Re: The "HIV" antibody tests are not diagnostic

I don't find the childhood figures so surprising based on the vertical transmission rate, but the study is surprisingly bad at attributing transmission risks. I disagree that such an investigation would be easy: unlike an anonymous HIV test, questions about sexual activities are far more likely to result in bias.

I agree that there is likely some unexplained HIV transmission in that age group. I do not think it's necessarily due to the ELISA though. I can think of several potential reasons why young kids might get HIV, whereas much older adults have fewer reasons: as Mr Bartlett states, childhood risks like vertical transmission and childhood vaccinations are hardly likely explanations. "Highly suspicious" rates are entirely inexplicable without knowing transmission routes: they may indeed be suspicious, or maybe expected. A falsely high background rate of "infection" doesn't really detract from the HIV/AIDS story, since the molecular and epidemiological evidence from the West is in my opinion more than enough to suffice. What it would mean though is that there might very well be an inherent background unreliability of the tests (developed perhaps in one situation) when applied in the developing world. As an example, if there were a 5-6% false positive rate, the childhood and elderly rates would practically disappear from the Mandela study whereas the young adult rates would be (say) around 20% instead of 26%. Does this undermine the entire premise, or just the population estimates?

Of course without more information, we're practising armchair science. The Mandela studies are a fantastic start and a huge population base, but almost necessarily lose the details. Transmission route is hard enough to get in the US, never mind rural SA.

The OraSure ELISA might not be approved in the US for that age group, but that doesn't necessarily apply everywhere. The age limits may purely and simply be due to the study population trialed. Relenza was rejected in the UK originally because the study population did not match the target population, even though it worked and was clinically useful in the population as a whole. The Flumist nazal flu vaccine is NOT approved in those under the age of five, even through it is effective and safe in those aged 5 through 49 - arguably better than the inactivated flu shot. Approval isn't everything ;-) I doubt a CDC-approved lab would make a serious diagnostic error by using a dodgy kit, and I can't find an age- restriction on the kit anyway.

I doubt that lab bias plays any role at all: the numbers are the means of diagnosis for the Orasure, and the cutoff is preset.

The CDC definition changes only really made a difference with the addition of the "CD4 <200" criterium, which after the initial year probably didn't make much of a difference (since those with CD4 counts lower than 200 would generally fulfill an OI criterium in the next year or so).

I'm not aware that there is any real difference in transmission rates for the HIV clades. They seem generally similar from my memory, once other confounding factors are removed or controlled for (concurrent STDs, multiple partners etc). I'm not expecting either a spread of SA-HIV into the developed world, or a massive spread into the developed world heterosexual population, providing current behaviours and transmission patterns persist. Similarly, I would have predicted the same thing back in the 1980's as the mainstream scientists did, that there WOULD be a spread into the hetero population. That didn't happen, so I don't expect that to happen now. Basically if the exposure doesn't occur, neither can the transmissions. Note that the more recent SA estimates seem to have plataued in recent years as well. Better late than never I guess...

Nick Bennett njb35@cantab.net

Competing interests: The author is a Sub-Investigator in an ongoing Flumist vaccine study in young infants. He recieves no specific financial recompense for that work.