The serological and nucleic acid tests for HIV-1 and HIV-2 are as good as or better than the tests for any other infection. 18 March 2004
Previous Rapid Response Next Rapid Response Top
Brian T Foley,
HIV Researcher
Los Alamos National Lab, Los Alamos, NM 87545

Send response to journal:
Re: The serological and nucleic acid tests for HIV-1 and HIV-2 are as good as or better than the tests for any other infection.

Chris Tyler asks:

”What is the evidence that the antibody tests are specific? “

There have been several hundred papers published in peer reviewed scientific journals over the past 20 years on this topic. They conclude that the dozens of antibody tests that have been developed for detection and classification of lentiviral infections are as good or better than the antibody tests developed for any other infection or non-infectious condition (such as the PSA tests for detection of prostate cancer).

The better question really, is: “What is the evidence that any FDA-approved HIV-1 EIA or Western Blot test is not specific?” The various denialist web sites are full of claims that things such as pregnancy cause false positive HIV antibody test results. But if you take the time to trace these claims back to the actual data, you will find that such false-positives are very rare, certainly not all pregnant women get a positive EIA result, and that they are easily distinguished as false and not true positive results by confirmatory western blots or other tests.

Chris Tyler asks:

”Why is there no gold standard for the antibody tests? How does Christopher Noble tell the difference between a 'true positive' and a 'false positive'? That is, when an antibody reacts with the antigen in the test kit, how does he know it is an 'HIV antibody' and not simply one of many different antibodies which are known to be reactive in the 'HIV' tests? "

There are many “gold standards” for antibody tests. Infectious molecular clones of lentiviruses can be used to produce pure virus proteins from cell cultures (mammalian, insect or bacterial). The clones can also be sequenced so that predicted peptides can be manufactured by in vitro chemical synthetic methods. It is not Chistopher Noble’s job to distinguish true infections from false positives. There are dozens of different methods for distinguishing true infection from false positive results. Indeed the goal of serological detection of viral infections (not just HIV but any virus) is to generate test kits that reproducibly have a very high sensitivity and a very low false positive rate. Increases in sensitivity most often come at the expense of an increase in the false positive rate. It is for this very reason that countries which can afford to do so use a very sensitive test such as EIA for initial screening, followed by a more extensive test such as western blot on those samples which react positive to the initial EIA. With two tests, the initial test can be set at a lower threshold for increased sensitivity (low rate of false negatives) because the second test can then be used to decrease the rate of false positives.

Chris Tyler asks:

”Why is there global variation in the interpretive criteria for 'HIV' infection (how can a person be 'positive' in one location and not in another?). “

Each country has its own medical councils and governing bodies, comparable to the USA’s CDC, NIH and FDA. The World Health Organization can make recommendations, but in the end each country is free to define its own standards for both diagnosis of infections (whether it be HIV or TB or Giardia) and for treatment of people who are diagnosed as having those infections.

Again, if you look at the differences in criteria for diagnosis of HIV infection, and you look at actual EIA and western blot results from thousands of people, you will find that a very small percentage of actual test results fall into the range where they would be labeled as “indeterminate” in one country and “infected” in another. It would be foolish to assume that any test that is attempting to give a binary “yes or no” answer to a question that is based on an analog signal (After infection, most infected hosts begin to produce antibodies, and the level of antibody production rises over time. It does not instantaneously jump from zero antibody to a high level.) to NOT have a rather arbitrarily set point where values below are labeled as “no” and values above are labeled as “yes”.

Most EIA and western blot results that would be labeled as “indeterminate” in the USA are transient. In many cases, the patient is recently HIV-infected and just beginning to produce antibodies to a few HIV proteins. In those cases, samples taken at later time points show that the patient has fully seroconverted and is now producing high levels of antibodies to many HIV proteins (typically more than 4 bands on the western blot are strongly reactive). In other cases, the “indeterminate” result was caused by pregnancy, infection with another virus that transiently raised overall antibody levels, or some other condition that then goes away. In these cases, samples taken from the same person at later time points do not become fully seropositive.

Chris Tyler asks:

”Why is it that a test, the PCR, whose basis is claimed to be recognition of nucleotide sequences specific to an exogenous retrovirus HIV, "should NOT be used" even as a screening test in adults and adolescents and infants infected via blood transfusion yet is recommended and approved to prove perinatal transmission?

The answer is that while PCR has a very low false-negative rate, it can have an unacceptably high false positive rate. In the case of newborns, the infant has maternal antibodies in its bloodstream which make it difficult to distinguish infected infants by antibody methods.

Chris Tyler asks:

”What is the evidence that the proteins used in the test kits are structural components of a retroviral particle, 'HIV'? That is, what evidence is there that 'HIV' has been isolated from everything else not 'HIV', such that its structural components have been shown to be these proteins?

Infectous molecular clones of dozens of different isolates of HIV-1, HIV-2 and many different SIVs have been created and used in research. HIV-1 has thereby been shown to be a group or family of related viruses most closely related to the lentiviruses isolated from Chimpanzees. HIV-2 has thereby been shown to be a group or family of viruses most closely related to the lentiviruses isolated from Sooty mangabeys (and macaques infected in captivity when housed with Sooty mangabeys). Dozens of other primate and non-primate lentiviruses such as feline and bovine immunodeficiency viruses have also been fully characterized. The characterization of lentiviruses leaves no doubt whatsoever that they are exogenous. No complex retrovirus has ever been found in an endogenous state. All endogenous retroviruses (such as HERVs) are simple, containing only gag pol and env genes and lacking regulatory genes such as vif, vpr, vpu and nef.

Anyone wishing to determine whether or not HIV-1 and HIV-2 have been "isolated" would be wise to look at the processes used for isolating, cloning, and studying all other viruses, and compare the methods used on HIV, SIV and other lentiviruses to those methods. No virus has ever been "isolated" by the exact criteria that the Perth group describes. Accurate classification of viruses cannot be done by electron microscopy, EIAV looks identical to HIV-1, and a mixture or EIAV and HIV-1 cannot be distinguished from pure HIV-1 by electron microscopy. Serological methods are much better, but the true "gold standard" is cloning and sequencing infectious molecular clones.

It is true that back in the 1970s, before molecular genetics techniques were available, and when immunological techniques were in their infancy, that electron microscopy of gradient-centrifugation-enriched fractions was a standard method. It remains a useful method. But it is not the be-all and end-all of virology.

Competing interests: None declared