Re: The "HIV" antibody tests are not diagnostic 15 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

The Perth Group responded to my points about TB with a smattering of papers I was well aware of. I can understand their basic premise (which is that if one protein can cross react, then multiple different proteins may mimick a positive western blot) but I have found very little evidence at all for it occuring in practise.

HIV seropositivity is NOT the be-all and end-all of diagnosis. Often, diagnosis is performed based on clinical suspicion and supported by secondary clinical and laboratory findings. I would be entirely unconcerned about a positive HIV ELISA (without confirmatory WB) in a middle-aged unmarried, asymptomatic woman. In a young man with widespread oral candidiasis and generalised lymphadenopathy then the positive ELISA alone would make me extremely concerned he had HIV even in the absence of a WB (which would be performed anyway). In many ways, getting the diagnosis right for HIV is far more important to the clinician and patient than for many other diseases. A false-positive HSV PCR for example doesn't really matter, since giving Acyclovir to protect against herpes encephalitis has limited risk and carries no associated long-term consequences. CD4 count and HIV viral load will go hand-in-hand with the serology and clinical judgement: as some of the Perth Group are apparently practising clinicians they ought to be aware that 80% of diagnosis is from history and examination: screening and specific tests contribute to the rest. If an elderly lady FOOSH'ed and had tenderness in the anatomical snuffbox, but X-rays of the wrist (including scaphoid views) were clear, would they assume she hadn't broken her wrist?

A measure of the fact that the antigen cross-reactivity isn't a significant problem is that, as an example, in one series of 250 TB cases in children, only 5 were seropositive for HIV [1]. It is obvious that there are many different western blot systems out there, and the various criteria of interpretation will vary depending which test is used. Some may have a greater or lesser sensitivity/specificity for the individual bands, hence different interpretations [2]. The Perth Group latch onto the fact that antibodies occasionally cross-react, and reject the whole idea of using antibodies. Antibodies evolved to act as _specific_ immune protective mechanisms, and have functioned as such for probably at least 500 million years. If they were inherently unreliable surely the mechanism would be long-defunct?

HIV/AIDS experts do NOT advocate antiretrovirals rather than specific drugs tailored to the current infections - this statement alone should alert the intelligent reader to the fact that the Perth Group are putting words into people's mouths, the Strawman argument all over again. In fact, the standard of care as I am aware of it is to CEASE the antiretrovirals in very severe cases, and in all situations treat specifically for the current opportunistic infection. Once the immediate danger is over, retrovirals are restarted if they were stopped. The HAART drugs buy time in terms of months and years, whereas antibiotics can saves lives in days: avoiding adverse side effects and drug interactions is a sensible approach. A developing exception is KS, where once the therapy would have been anti-cancer. With the realisation that it is an immune- deficiency related tumour caused by a virus (HHV8) that has no specific pharmacological treatment, treating the HIV appears to cause the KS to regress (as I have pointed out before).

It is very clear that TB is more aggressive and more frequent in those infected with HIV. It is also more likely to BE spread from an infected person, interestingly enough. I can fully appreciate that a severe case of TB in an HIV-negative person is going to be clinically similar to a mild case of TB in an HIV-positive person: and coupled with the potential cross-reactivity might pose a diagnostic dilemma. However, I am not aware of any studies where this has been the case, and dual TB/HIV infection displays characteristics different from TB infection alone (more likely to have a false-negative TB skin test, lack of sputum mycobacteria, more aggressive disease). Going back to our famous Swiss Cohort study, they showed a low rate of antibodies to lipoarabinomannan (12% versus 67% in normal TB cases) even in confirmed cases of TB in HIV+ people [3]. 0% of non-TB HIV-infected patients showed antibodies to lipoarabinomannan. Once again the armchair theorising from the Perth Group fails to stand up to scientific scrutiny. Additionally, it seems as if the concerns of the Group have already been, once again, met by the current literature. One wonders why they persist in thinking otherwise.

As regards the post regarding childhood diagnostics, the references the Perth Group used were specific to adults and clearly showed the diagnostic superiority of serology to nucleotide based tests. The fact that they choose to interpret the literature otherwise is not my concern - the spread of misinformation is. I did NOT say that "serology was useless" - as my quote clearly states I say "neonatal serology is useless". If one test is 0% effective and another is (say) 80% effective, I'd take the 80% any day. The Perth Group would have us do nothing!

The Perth Group end by telling us what "HIV experts" are doing. Respectfully, as one of the Experts I would do or say very little that the Perth Group claim. Can the Perth Group find a paper where DNA-PCR of neonates is clearly stated to be not just useless but actually detrimental, as stated by "HIV experts"? In fact the Experts also say that HIV-antibodies from vertical transmission may persist for far longer than 9 months, and to claim that a disappearing serology implies "clearance of HIV infection" throws out decades of immunology research as well as 500 million years of evolution!!! (I make no apologies for the use of multiple exclamation marks, despite Terry Pratchett's views on them). In fact:

"amplification of the integrated viral genome by PCR has been the preferred method for the diagnosis of HIV infection in children for many years." [4-5]

I fully appreciate the implications of the Perth Group's arguments: some years ago it became clear to me that if the Dissident views were true then we really should be re-appraising the current thinking. Unfortunately for the dissidents, their interpretations of the primary literature flew so violently in the face of current epidemiology, molecular biology, virology and pharmacology that I had no option but to reject their theories. Investigating the Dissident claims actually made me more certain that the current understanding of HIV and AIDS was correct: and the Perth Group reaffirm that certainty on a regular basis!

Nick Bennett njb35@cantab.net

1. T. Shahab et al. Prevalence of Human Immunodeficiency Virus Infection in Children with Tuberculosis INDIAN PEDIATRICS 599 VOLUME 41: JUNE 17, 2004

2. B Weber et al. J Clin Microbiol. 1992 March; 30 (3): 691697 Multicenter evaluation of the novel ABN Western blot (immunoblot) system in comparison with an enzyme-linked immunosorbent assay and a different Western blot.

3. Boggian et al. J Clin Microbiol. 1996 Jul;34(7):1854-5. Infrequent detection of lipoarabinomannan antibodies in human immunodeficiency virus-associated mycobacterial disease. Swiss HIV Cohort Study.

4.A. Fischer et al J Clin Microbiol. 2004 January; 42 (1): 1620 DOI: Simple DNA Extraction Method for Dried Blood Spots and Comparison of Two PCR Assays for Diagnosis of Vertical Human Immunodeficiency Virus Type 1 Transmission in Rwanda

5. Katrien et al. Molecular and Cellular Probes Volume 8, Issue 4 , August 1994, Pages 317-322 Design and evaluation of new, highly sensitive and specific primers for polymerase chain reaction detection of HIV-1 infected primary lymphocytes.

Competing interests: None declared