Item #13: Tropical Diseases and False Positive HIV Tests

Even when HIV tests are performed, many diseases that are endemic to Africa, such as malaria and TB, are known to cause false positives.

Footnote 4 further states “many diseases that are endemic to Africa, such as malaria and TB, are known to give false positives.” Farber fails to supply a reference.

The risk of a false positive HIV test in Africa, as elsewhere, is very small if the correct protocol is followed. Some HIV antibody tests have been tested in Africa and found to be very accurate. These are the ones generally used. For example, the Abbott Determine rapid test used widely in South Africa has a specificity of at least 98% (and in some studies has achieved close to 100%). When this test is combined with a second rapid test or an ELISA test to determine HIV status, the risk of a false positive is negligible. The contribution of TB and malaria to false positives on today's tests is also negligible.

For examples of trials of HIV tests used in Africa and Brazil, see Sauer et al. (2000)29 , Phili et al. (2002)30 , Ferreira et al. (2005)31 , Koblavi-Dîme et al. (2001)32 and Foglia et al. (2004)33 .

There are several papers indicating that diseases common in tropical areas, including Africa, can cause false positive HIV tests, although not enough work has been done in this area:

• Leprosy, or contact with Leprosy patients: “Sera from 63.6% of leprosy patients and 23% of their contacts was repeatedly positive by HIV ELISA [which in Africa and England is enough for a diagnosis of HIV infection]” [1]
• Malaria: “The prevalence of malarial parasitaemia was 13% [in a serological survey of 250 outpatients in rural Zaire]. However, 72% of patients had antibodies against P falciparum [the malaria parasite]…The proportions of subjects positive in the ELISA test for HTLV-I, HTLV-II and HTLV-III [HIV] were 14%, 25% and 12%, respectively...When the titer [level] of antibodies against P falciparum was considered...this single factor dominated all others...If the human retrovirus reactivity observed in the ELISA tests is frequently non-specific among Africans, the causes of the non-specificity need to be clarified.” [2]
• Malaria: “We evaluated serum samples from a group of 12 patients with acute Plasmodium vivax infection [malaria] who lived in…southwestern Venezuela…Our second group, 12 patients with Plasmodium falciparum infection…None of the patients were receiving antimalarial drugs…and none belonged to any of the recognized AIDS risk groups or had any AIDS-associated disorders…3 of the patients with P. falciparum infection (25%) and 5 with P. vivax (41%) were found to be positive for HTLV-III/LAV [HIV] antivodies by [3 antibody test types]” [3]
• Unspecified: “the proportion of HTLV-III [HIV] seropositive patients was notably high among those with idiopathic splenomegaly [enlarged spleen with no known cause] and schistosomiasis [parasitic infection]…Among Kenyans, the Turkana have the highest and the [closely related] Masai have the lowest prevalence of antibodies against both viruses [HTLV-I and HTLV-III(HIV)]…the Turkana inhabit a desolate area of desert and scrub-brush and are the poorest and most isolated of the groups that we studied…the Turkana have many other parasitic and viral infections [apart from seasonal malaria], including one of the highest rates of hydatid [tapeworm] disease in the world and a high prevalence of hepatitis markers…In view of the high prevalence of HTLV-III antibody, AIDS illness might have been expected to be frequent, yet no cases have been documented in Kenya [to 1985]. We are aware of only 3 cases suspected of being AIDS-related…it is unlikely that AIDS-illnesses of the recognized varieties are occurring at a rate commensurate with the prevalance of antibody against HTLV-III” [4]
• Several Diseases: “The currently used laboratory based and rapid HIV-tests show intolerably high numbers of false positive results when used in developing countries. This is mainly due to interfering effects of diseases virtually absent in developed countries.” [5]

Even if the error rate of these HIV tests was just 2% this would mean that, in a population where 10% of people are testing positive, it would mean that 8% were true positives, and 2% false, with no way to tell the difference.

1. Kashala O et al. Infection with human immunodeficiency virus type 1 (HIV-1) and human T cell lymphotropic viruses among leprosy patients and contacts: correlation between HIV-1 cross-reactivity and antibodies to lipoarabinomannan. J Infect Dis. 1994 Feb; 169: 296-304.
2. Biggar RJ et al. ELISA HTLV retrovirus antibody reactivity associated with malaria and immune complexes in healthy Africans. Lancet. 1985 Sep 7; 2(8454): 520-3.
3. Volsky DJ et al. Antibodies to HTLV-III/LAV in Venezuelan patients with acute malarial infections. N Engl J Med. 1986 Mar 6; 314(10): 647-8.
4. Biggar RJ et al. Regional variation in prevalence of antibody against human T-lymphotropic virus types I and III in Kenya, East Africa. Int J Cancer. 1985 Jun 15; 35(6): 763-7.
5. gaifar.com. 2004 Sep [accessed]. www.gaifar.com.