PRESIDENTIAL AIDS ADVISORY PANEL REPORT
HIV TESTS AND THEIR ACCURACY
4.1 HIV testing
The ELISA, Western Blot and PCR viral load are the most frequently used tests to confirm HIV infections. The ELISA and Western Blot tests detect HIV antibodies in the serum of patients, whereas the PCR Viral Load test is a genetic test that detects small HIV nucleic acid fragments in whole blood. The veracity and reliability of these tests are key to the validity, reliability, quality and accuracy of epidemiological data used by any country. The ELISA test is mainly used to screen for HIV infection in blood donors and for general surveillance, whereas the Western Blot and PCR are generally used as confirmatory tests and in the context of research. All these tests, individually or in combination, are considered by the proponents of the HIV/AIDS theory as important indicators of infection by HIV. The CD4 count is an additional laboratory test used in combination with ELISA to make a diagnosis of AIDS; and with the Viral Load Test to determine the clinical progression of the AIDS disease and the monitoring of the effectiveness of anti-retroviral treatment. The Western Blot test is more expensive and requires a well-developed laboratory infrastructure; it is therefore not affordable for many developing countries. Dr Fiala also informed the panel that the Western Blot test is not accepted in the United Kingdom as a confirmatory test for HIV infection due to its unreliability.
Co-culturing of virus is used to isolate the virus from the blood of infected AIDS patients. This method is generally used as a research tool as it is too expensive and time-consuming to conduct as a routine surveillance and screening method. It also requires highly specialised staff and infrastructure.
According to Dr Lane, a substantial number of research publications indicate the value of HIV testing and making a diagnosis of HIV infection on the basis of the antibody tests. AIDS surveillance was first based on the clinical case definition, and AIDS was initially regarded as just a cluster of diseases. The subsequent discovery of HIV led to the incorporation in the definition of the disease of various clinical and immunological patterns, and, as recently as 1993, the CDC AIDS definition was widened to include a wider spectrum of clinical disease also utilising CD4 counts. Again, Dr Fiala informed the panel that European countries do not include the CD4 count in the definition of AIDS.
However, Dr Gayle pointed out that there is a general lack of standardisation of the definition of AIDS throughout the world. This, she argued, arose because it is possible to diagnose HIV infection by means of non-standardised laboratory tests, as well as by the verification of the presence of clinical symptoms. Since data are compared across countries, there is a need to standardise the definition of AIDS.
A particular concern raised by some members of the panel was that after 15 years of research, there is the lack of a 'gold standard' against which to measure the accuracy and reliability of the data generated from the commonly used methods to diagnose HIV infection.
4.1.1 ELISA test
The ELISA test is the most commonly used test for screening blood from donors. Its specificity and high sensitivity make it widely acceptable since it is able to detect all the possible HIV infections. It has also been found to be useful in surveillance. It is generally accepted that a single test cannot be regarded as proof of HIV infection. However, in order to improve the reliability and validity of ELISA, the CDC testing guidelines state that "a test for HIV antibody is considered positive when a sequence of tests, starting with a repeatedly reactive enzyme immunoassay (EIA) and including an additional, more specific assay, such as a Western Blot, are consistently reactive". Similarly, the WHO testing guidelines require confirmation of samples that are repeatedly reactive by ELISA using the same blood sample but a different ELISA kit. Both testing regimes call for repeated ELISA testing of a single blood sample rather than ELISA testing of more than one blood sample. However, the UNAIDS–WHO recommendations state: "An additional blood sample should be obtained and tested from all persons newly diagnosed as seropositive on the basis of their first sample. This will help eliminate any possible technical or clerical error". Major concerns surrounding the ELISA test, however, include its specificity, reliability and reproducibility, as well as the lack of a comparative 'gold standard'.
Dr Giraldo claimed that the HIV ELISA tests are not specific for HIV. He cited the fact that four repeat ELISA tests plus a Western Blot are required for a diagnosis of HIV disease in the USA - an example of how unsure the manufacturers are of its specificity. Furthermore, he elucidated that there is no standard by which to establish the specificity and sensitivity of the ELISA test. He also pointed out that the Roche Amplicor PCR kit specifically states that it must not be used as a screening test for HIV nor to diagnose HIV infection (see section 9.10.3).
Dr Turner pointed out that WHO, in its description of the 34 HIV antibody ELISA tests on the market, uses one antibody test as a gold standard for another. In his opinion mycobacterial and fungal antigens can cross react with HIV ELISAs causing false positives.
Prof Montagnier argued that no test is perfect and, moreover, that the current generation of ELISA tests is much more sensitive and more specific than in 1984. The current ELISA test uses recombinant proteins made from clones and consequently minimises cross-reactivity by other proteins from the plasma. This view was echoed by Dr Makgoba, who repeatedly pointed out that all immunoassays used in South Africa are based on recombinant viral proteins, which are more specific than tests used previously and are similar to those used in other countries. All HIV diagnosis in South Africa is supported by laboratory tests "and HIV screening is always followed by two confirmatory tests. All immunoassays in South Africa are designed, calibrated, optimised and standardised to the level of their discriminating power – which is the power to discriminate between negative and positive cases." According to Makgoba, the tests used in South Africa have high confidence results of 99.9%. In a low risk population with more than 36 000 regular blood donors in South Africa, the results showed that 99.9% of the tests produced negative results. Furthermore, comparison of data from three laboratories in South Africa shows a low false positive rate. The rates of false positive diagnoses in over 2000 cases for various South African HIV testing institutions were: 1.6% false positive reports at the National Institute of Virology (NIV), 0.23% at the South African Institute of Medical Research (SAIMR) and about 0.33% at a Stellenbosch laboratory. Dr Magkoba pointed out that the ELISA test has a predictive rate of over 99% in South Africa. Both the predictive rate and the false positive rate of ELISA tests in South Africa compare very well with similar results obtained in the United Kingdom. Screening tests used in South Africa are as good as those used anywhere else in the world. The tests are highly specific, sensitive and reliable.
There were, however, a number of opposing views in the panel about the specificity, reliability and validity of ELISA.
Dr Turner argues that the data used as a basis for the AIDS pandemic in South Africa are generated from the antibody ELISA test, which is fundamentally inaccurate, unreliable and invalid. The antibody ELISA test is based on the reaction between the unique viral protein (the p24 proteins 'supposedly' from HIV) and serum antibodies from a blood sample. Independent data show that p24 proteins, the basis for the ELISA antibody test, have been found to cross react with a wide variety of uninfected human tissue and blood samples from other disease states. For example, antibodies to candidiasis and mycobacterium infections cross react with p24. Furthermore a warning in the manufacturer's inserts suggests that the ELISA should not be used on its own for HIV diagnosis. According to Dr Giraldo, many other disease conditions – such as leprosy, malaria, leishmaniasis and other viral infections – give rise to false positive results in the ELISA test without the concomitant HIV infection. Furthermore, many of the conditions that cause a false positive result in the ELISA test are conditions that are also prevalent in many of the recognised AIDS risk groups. A great deal of scientific data indicate widespread non-specific interactions between what are considered retroviral antigens and unrelated antibodies. A positive HIV ELISA test may also indicate previous antigenic stimulation by other retroviral infection. Another concern is that there is no precedent for the diagnostic utilisation of the ELISA test for other viral diseases. In general, the presence of antibodies specific to a particular disease is a major indicator of potential immune protection by the body, which is not the case with HIV infection, since antibodies to HIV fail to confer any immunoprotection against HIV. The ELISA test may therefore not be a true indicator of infection but an artefact arising from cross reactivity of other naturally occurring viral proteins.
The lack of standardisation of ELISA results, both within the country and across countries, was a source of major concern to some panellists. Results of ELISA tests may be interpreted differently within a single laboratory, between laboratories within one country, and between countries. This may mean that a person that tests positive at one laboratory in South Africa may test negative at a different laboratory in the same country. Moreover, the lack of standardisation across countries could result in an individual's testing positive in one country and negative in another.
4.1.2 Western Blot
The Western Blot is an antibody test, which, according to Dr Makgoba, is one of the tests used to confirm the diagnosis of HIV infection in South Africa. Dr Sonnabend reported that he regularly uses the Western Blot as a confirmatory test for HIV infection in the USA. A positive Western Blot result is synonymous with HIV infection and the attendant risk of developing AIDS. He was satisfied that there was general agreement on the correlation between Western Blot and AIDS and patients that were suffering from AIDS always reacted positively to the Western Blot test.
However, a number of concerns were raised around the specificity, reliability and reproducibility of the Western Blot test.
Dr Turner believed that the Western Blot should not be used to confirm and validate the results of the ELISA test since the Western Blot and ELISA tests are based on the same antibody reaction mechanism. As with the ELISA test, another concern over the use of the Western Blot test is its non-specific positive reaction to a number of diseases (including tuberculosis, a variety of parasitic infections and other viral infections) in the absence of HIV infection. The antigens used in the Western Blot test may be similar or identical to other human proteins, and hence the results of the Western Blot may thus not provide an indication of HIV infection.
Dr Papadopoulos-Eleopoulos presented a transparency showing the results of a Western Blot test with a number of samples from leprosy, TB and AIDS patients. It appeared that the Western Blot results from the different samples were indistinguishable from one another, showing the Western Blot test to be non-specific and unreliable. All the samples tested positive, even those from leprosy and TB patients. In further deliberations, Dr Mark Smith pointed out that Max Essex had already demonstrated the indeterminate results from Western Blot results in 1994. The above underlines the fact that the Western Blot test cannot be used as a determinate diagnostic tool.
4.1.3 PCR test for viral load
The PCR viral load test is also used as a confirmatory test. It is based on the amplification of tiny HIV viral particles that are supposed to originate from HIV in the blood. This test is virus specific and specifically detects HIV RNA. It is used to determine the level of viral load in the blood. It is mainly used in the tracking of the clinical progression of advanced HIV infection to AIDS disease, the monitoring of the effect of anti-retroviral treatment and the monitoring of mother-to-child transmission. There is a high correlation between clinical disease progression and the viral load. A high viral load is associated with an increased risk of transmission and the clinical progression to AIDS. Also the level of virus in the blood is directly related to the degree of risk of transmission to uninfected individuals. People with undetectable levels of virus in their blood do not transmit to uninfected partners. Mothers with high viral loads had the highest chance of transmitting the virus to their infants.
Arguments against the use of PCR are that this test is characterised by high variability and lack of reproducibility. In addition, the very wide variability may lead to the erroneous interpretation of results, thus compromising the accuracy and validity of the PCR results. Dr Bialy pointed out that the PCR viral load test might not be a legitimate measure of infectious virus. It demonstrates a high level of fluctuation, and the viral load can be increased non-specifically by other viral and bacterial infections (opportunistic infections may also increase viral load). Research results indicate that the viral load test may not always be an indicator for the clinical progression of HIV to AIDS.
Another point of concern that was raised was the fact that the PCR test was developed for the non-C-clade virus, whereas the clade-C virus is the most prevalent in South Africa.
4.1.4 CD4 count
The CD4 count is a determination of the concentration of CD4 T-lymphocytes in the blood. The associated immune deficiency leading to infection by opportunistic infections is ascribed mainly to the depletion of CD4 T-cells. The CD4 count can therefore be regarded as an accurate determination of the robustness and functionality of the immune capability and status to effectively protect the body against general infections. HIV infects and destroys CD4 cells (though some dispute this), rendering the immune system incapable of protecting the body against general infections, hence the resultant immunodeficiency in HIV infection and AIDS. This immunological test is used to monitor the progression of HIV infection to clinical AIDS disease and to monitor the effectiveness of anti-retroviral therapy. The CD4 count can be inversely correlated with the viral load. The higher the viral load, the lower the CD4 count will be. Intermediate progressors (patients who take longer than 10 years to progress from HIV infection to AIDS) consistently maintain the concentration of CD4 within normal range. When the CD4 count drops, it predicts the onset of opportunistic infections. In rapid progressors (those who developed AIDS within 2-4 years after infection), the CD4 drops precipitously, coinciding with the onset of infections and clinical progression to AIDS.
The improvement of the concentration of CD4 during anti-retroviral therapy is used as a surrogate marker for the effectiveness of the treatment.
4.2 Virus isolation or co-culturing
It is standard practice in virology to use virus isolation or co-culturing to confirm the presence of a virus in any sample (a requirement stated by Prof Montagnier's co-workers a decade ago). Two laboratories have been credited with the successful isolation of the virus from AIDS-infected patients. These are the Koch Institute (work published by Hans Gelderblom) and the US National Cancer Institute. In both of these cases, the authors admitted the possibility of contamination from other viruses in the blood. Once the virus has been cultured, it can be visualised by electron microscopy. While the difficulty of visualising HIV in blood culture is generally accepted, it is possible to see the virus in lymph nodes, which is the major site for replication. Dr Morris pointed out that the large amounts of plasma proteins complicate the visualisation, so that, even when the viral load is high, it is difficult to see any virus by electron microscopy. Transmission electron microscopy, as well as in situ hybridisation, demonstrating that HIV resides on the surface of follicular dendritic cells, are satisfactory means for the visualisation, she said.
Dr Turner supported a proposal that will enable comparison of the antibody titre with the isolation of the virus. However, HIV isolation is difficult and expensive to undertake.
4.3 Moratorium on HIV testing
Dr Rasnick, supported by Prof Mhlongo, recommended that the South African government should consider terminating HIV testing by blood banks and for general surveillance since the results of all the tests are unreliable and non-specific and hence give wrong information. He argued further that AIDS would disappear instantaneously if all HIV testing were outlawed. The basic question was whether hundreds of people in South Africa are dying of AIDS or of TB, malaria, Pneumocystis carinii pneumonia and parasitic infections. The absence of data indicating the rate of deaths due to AIDS should receive urgent attention.
Prof Mhlongo spoke of the preoccupation with biomedicine in an attempt to be scientific even at the expense of the wider distressing situation of poverty, poor housing, lack of sanitation and a multitude of diseases associated with the deprivation and urban squalor that characterise the reality of the majority of black South Africans. He presented evidence to show that mortality in England and Wales due to tuberculosis declined precipitously from 1838 with hardly any people dying of TB in the United Kingdom by 1960 (McKeown). This occurred despite the fact that TB chemotherapy only started in 1945 and the BCG vaccine was only developed in 1956. This decline in mortality due to TB had nothing to do with medical interventions but was rather a consequence of improved sanitation, improved housing, improved nutrition and improved environmental conditions that occurred in the United Kingdom during the course of the period under review. Prof Mhlongo went on to present evidence that a similar picture was demonstrated in the USA from 1920 to 1991 (Seale and Stephen).
4.4 Recommendations on HIV testing
4.4.1 Proposed studies and experiments
The key issue that came under focus was the reliability of the ELISA testing in terms of delivering time infection data. As a diagnosis of HIV infection has such a profound effect on a person’s life and future, it was considered of utmost importance that the tests are unimpeachably reliable. Since all epidemiological predictions concerning HIV/AIDS in South Africa are based on the results of such tests, their absolute reliability was declared to be of utmost importance.
A major recommendation arising from the two meetings was to apply a series of HIV tests of increasing stringency in order to establish the validity, veracity, rigour, reliability and concordance of ELISA, PCR and viral isolation. Details on some of these experiments are presented in Chapter 9 of this report.
The experiment will consist of the following series of steps:
- ELISAs as they are currently employed
- The same tests using a pre-absorption protocol to remove cross-reacting antibodies such as those against Mycobacteria which some panellists asserted frequently confound ELISA tests
- A variety of nucleic acid based protocols
- The application of the classical gold standard of virus isolation, including electron microscopy
These tests will be performed on cohorts chosen to represent the full spectrum of the South African population.
4.4.2 Recommendation on future HIV testing
The panellists who refuted the causal linkage between HIV and AIDS recommended the suspension of all HIV testing until its relevance is proved, especially in the African context, given the evidence of false results in a tropical setting and the fact that most assumptions and predictions on AIDS in Africa are based on HIV testing.
4.4.3 General recommendations on testing
- The case definition of AIDS to be standardised for clinical practice in South Africa.
- Any positive HIV ELISA result to be repeated with at least two additional blood samples before an HIV diagnosis is confirmed in order to improve the reliability and validity of ELISA.
- Apply a series of HIV tests of increasing stringency in order to establish the validity, veracity, rigour, reliability and concordance of ELISA, PCR and viral isolation.
Contents | Continue