HIV diagnostics 1 February 2005
Previous Rapid Response Next Rapid Response Top
Peter J Flegg,
Physician
Blackpool UK FY3 8NR

Send response to journal:
Re: HIV diagnostics

I agree with Anita Allen that there is often confusion about “diagnostic” tests. However, this confusion is often in the minds of the HIV-dissidents who insist and expect that we require a single laboratory test be used as a gold standard to confirm or deny the diagnosis of HIV infection, only to then complain that because the test is not 100% specific and 100% sensitive that it is either: (a) useless, or: (b) HIV infection is a myth, or, more usually: (c) both.

As a physician, I am acutely aware that the vast majority of tests that we perform merely assist in the clinical diagnosis, rather than serve as a totally objective and independent arbiter of whether the disease is present or not. Diagnosis is an imperfect process, and I have alluded to problems inherent in reaching specific diagnoses in previous posts.

It is up to the clinician to reach a diagnosis, and consider the relative weight of the “diagnostic” tests that are available. If I wish to diagnose a myocardial infarction, I need to consider clues from the history and examination, and appropriately interpret the significance of a number of tests (ECG, cardiac enzymes) which in themselves may be of low specificity. Diagnosing someone with clinical and laboratory signs of immunodeficiency as having AIDS, or “HIV infection” is simpler than determining if someone is merely HIV-infected. HIV testing consists of multiple tests applied in a serial fashion, which maximises specificity and positive predictive value, and this strategy works extremely well in practice (and is ignored by HIV-dissidents, who assume we rely on isolated, uncorroborated single test results). The tests used are approved for their purpose, namely detection of HIV-antibodies, and not diagnosis of HIV infection. If there were a perfect test there would be no problem. We obviously do not demand that the kit manufacturers of cardiac troponin assays declare that their assay is approved for the diagnosis of myocardial infarction (and any disclaimers are there to safeguards the manufacturers from liability for incorrect application of their test, and also to safeguard patients). The same principles apply with HIV.

As Nicholas Bennett has described, PCR based technology is opening up a new era of diagnostics, but their use as a sole marker of HIV infection is limited by the possibility of false positivity.

Anita Allen is incorrect to say that viral load tests are used to determine CD4 counts. If someone is unable to differentiate between these basic tests it does not inspire confidence in their ability to comprehend the subtleties of HIV diagnosis. She asks who bears legal responsibility for “non-approved” application of the tests. It is the clinician who requests the tests; he will be the one who makes the eventual diagnosis and who accepts any responsibility for misdiagnosis. There is no difference here with HIV or any other infection or illness – HIV diagnosis follows the same pathways as most other conditions.

Bearing in mind the diagnostic process I am reminded of the following saying by Epicetus (2nd C):

“Appearances to the mind are four kinds.
Things either are what they appear to be;
Or they neither are nor appear to be;
Or they are and do not appear to be;
Or they are not, yet appear to be.
Rightly to aim in all these cases is the wise man’s task”

Competing interests: None declared