Re: Re: More representative HIV test literature 3 February 2005
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: Re: More representative HIV test literature

In reply to Anita Allen's specific points:

I hope that my concurrent reply to Mr Spagnoli which included references to the aforementioned tests helps somewhat. I did state where the test literature could be found, on the FDA website, and give the names of the specific tests however. I had thought that to be sufficient.

Firstly there is no reason to suppose that because an organism has been isolated and characterised at the genetic level (as has been done for HIV) that serological testing should be abandoned. It is easy, quick, and reliable. It is true that serology is not the same as isolation, and indeed for many acute infections actually signifies resolution of the illness. This isn't so for HIV, since it can be grown from seropositive individuals after many years.

The Piot review is just that, a review. Much like my reference to the FDA website, one assumes that the curious individual is going to read the paper and follow-up the references contained within. The Padian study brought up my Mr Russell for example rather neatly demonstrates heterosexual transmission between discordant couples (over 15% with an average followup of only 7 months per couple).

The references given for "90% isolation" pre-date those used often by myself showing 100% isolation. One group which wrote two of those papers (Jackson is senior author) previously reported only a 57% culture rate from seropositive patients. The improvements relate to laboratory techniques (science after all is not static).

There is no FDA distinction between HTLV-III and HIV, unless one looks far enough back in the literature before the name HIV was universally accepted. Perhaps Ms Allen refers to HTLV-I/II which are two different viruses. This may be slightly unclear from the text formatting at:

With regard to a test being used to diagnose anything, as far as I am aware no such test exists for _any_ illness. Diagnosis can only be done by a qualified doctor, but various tests can be used as an aid to diagnosis (this phrase is used liberally throughout the HIV test literature). Even a heart attack, a rather concrete clinical event one might imagine, has criteria by which a diagnosis may be made. ECG changes may not appear, cardiac enzymes may not appear (or may not be available), and the actual event may be asympomatic e.g. in diabetics. This doesn't make the individual tests invalid or suggest we shouldn't use them. Compared to an ECG, the HIV antibody tests are remarkably sensitive and specific.

Another randomly googled test, for RSV infection.

"The NOW® RSV Test is a rapid immunochromatographic assay for the qualitative detection of respiratory syncytial virus (RSV) fusion protein antigen in nasal wash and nasopharyngeal swab specimens from symptomatic patients. This test is intended for in vitro diagnostic use to aid in the diagnosis of RSV infections in neonatal and pediatric patients under the age of five."

One wonders perhaps what possible reason the manufacturers might have not not permitting the test to be used in those over the age of five... Perhaps it's simply because the population used in validating the kit was under the age of five. Also note the "aid in diagnosis" phrase.

The kit quoted by Ms Allen was first developed back in 1985 by Abbott, but modified several times (it was reapproved twice in 1992, including the rDNA test referred to). The methods used in forming sensitivity and specificity measures are standard in the scientific world for any diagnostic kit. HIV tests do not get any kind of special dispensation.

I politely suggest to Ms Allen that her interpretation of the wording used in the HIV test kits is mistaken. For better or for worse, there is only one "kit" for diagnosing anything, and it comes in the form of a medically qualified doctor! The person legally responsible for a misdiagnosis is him or her, providing they used appropriate "aids to diagnosis" and good judgement. Expecting a test to be able to give a definitive diagnosis is incorrect. I urge any critics of HIV science reading this discussion to first compare the procedures used in validating other non-HIV kits before coming to the conclusion that HIV science is somehow subject to a different set of rules from anything else. What may be "obvious" (e.g. expecting non-serological tests to be used after a genome is isolated) is in fact not the case at all.

Sadly the evidence was available long ago to put the anti-HIV claims to rest, but the arguments persist because certain individuals choose to ignore it or use their own particular brand of pseudoscientific logic. In my opinion at least. For example I wonder what Mr Russell thinks of the lack of cross-reactivity from all the conditions tested in the HIV test kit validations. Does he admit that, even if some of the many conditions he claims to cause false-positives actually do cause cross-reactions, it happens on an unusual basis and is unlikely to be clinically significant?

Nick Bennett

Competing interests: None declared