Re: ELISA 'HIV' tests are non-specific and inaccurate 26 November 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: ELISA 'HIV' tests are non-specific and inaccurate

Aside from cutting-and-pasting package inserts, I wonder if Mr Russell has considered the fact that the FDA is inherently cautious about making any kind of claim, or allowing anyone else to make claims? The small print at the bottom of some of the many "drug" adverts here in the US often makes for amusing reading: "This product is not intended to diagnose, treat, cure or prevent any disease." Which makes you wonder just what the heck it IS for!

There has been an explosion in diagnostic tests and test reagents in recent years, brought about by advances in molecular biology knowledge and equipment. However, up until recently the FDA required every analyte package insert to include the language:

"This product has not been approved by the FDA for diagnostic use."

Or something similar. This raised more than a few eyebrows when the clinicians using the tests were made aware of this fact. It was only relatively recently that the companies were allowed to add:

"Such approval is not necessary."

Which makes rather more sense, since approval never was necessary for these analytes. These waivers are now standard, and are provided if the test has been verified by the supplier or is intended to be verified by in -house diagnostic staff. Of course whole test kits generally do need to be approved, but this was I think a good example of beaurocracy gone mad.

Also, quoting out of context is poor show: the third quote is from an RNA-based PCR assay, which was never intended for initial screening. In fact, an RNA test does exist which was specifically approved by the FDA for screening blood donations, but of course this cannot be used for diagnostic screening since it wasn't evaluated for that purpose. [1] Conversely some rapid antibody tests, while apparently perfect for screening blood donations are not approved for this purpose, again because they were only evaluated for diagnosis. [2]

The additional information in the product package would make interesting reading to those claiming a high false-positive rate due to cross-reactivity. The Reveal test for example looked at samples containing Hep B s-antigen, Hep C antibody, Hep B antibody (c and s), HTLV -1, CMV, measles, HSV, rubella, varicella, Rhematoid Arthritis factor, C- reactive protein, EBV, antistreptolysin-O, mycoplasma and syphilis reagin antibodies. All 271 samples were negative for HIV, indicating that "none of the above conditions interfered with the Reveal™ Rapid HIV-1 Antibody Test."

One other consideration is that the Abbott test in question was the first developed, back in 1985: it's no surprise that "no recognized standard" existed. It _was_ the standard. Certainly I've found it impossible to find similar language in current HIV test kit package inserts.

Nick Bennett njb35@cantab.net

1. COBAS AmpliScreen HIV-1 Test, version 1.5 (Roche)

2. Reveal™ Rapid HIV -1 Antibody Test (MedMira)

Competing interests: None declared