The Western blot assay, the most commonly used confirmatory test for human immunodeficiency virus (HIV) antibody detection, is expensive and may give indeterminate results. The Joint United Nations Programme on HIV/AIDS and World Health Organization therefore recommend three testing strategies for alternative HIV confirmation to maximize accuracy while minimizing cost. All three strategies (Strategy I, II, and III) have algorithms involving the use of one to three enzyme-linked immunosorbent assays (ELISA) and/or simple/rapid assays. Which strategy is most appropriate depends on the objective of the test, the prevalence of HIV in the population, and whether or not the patient is symptomatic (1).
Strategy III involves the use of three immunoassays, each using different antigen preparations and/or different test principles. Sera are considered HIV antibody positive when all three tests are reactive. This most stringent of the alternative strategies is recommended for the diagnosis of HIV in asymptomatic persons in low-prevalence populations (≤10%) (1).
We report six cases tested during 1994 to 1997 where the use of Strategy III would have caused a false-positive result. These patients were asymptomatic and were tested for HIV antibodies for work permit applications (five) and antenatal screening (one). The prevalence of HIV in this group is <0.1%.
The following three tests were used in order. (i) The Abbott HIV-1/HIV-2 3rd Generation Plus EIA, a double antigen sandwich assay. Antigens used are recombinant HIV-1 env and gag proteins, HIV type 2 (HIV-2) env proteins, and HIV-1 synthetic env peptides. (ii) The Genelavia Mixt (Sanofi Diagnostic Pasteur), an indirect binding assay detecting both immunoglobulin M (IgM) and IgG. Antigens used are gp160 recombinant protein and peptides mimicking the immunodominant epitopes of the HIV-1 and HIV-2 envelope glycoprotein. (iii) The Serodia Particle Agglutination Assay (Fujirebio), which is essentially a double antigen sandwich assay using gelatin particles as the solid phase. Antigens used are viral lysates of HIV-1 and HIV-2. The selection of the three assays fulfilled the criteria for Strategy III.
All six patients had reactive results for the three assays. Supplementary tests were carried out, because at least one of the two ELISA tests had a low OD/CO (sample optical density/cutoff) value. The four patients with follow-up samples showed no seroprogression, and negative PCR (HIV-1 proviral DNA) by both Amplicor HIV 1 (Roche) and in-house assays gave further weight to the HIV-negative status. The Western blot (HIV Blot 2.2; Genelabs Diagnostics) assays performed produced indeterminate results. This was, however, more acceptable than erroneous HIV antibody-positive results, as it points to the need for further evaluation of the patient (Table 1).
TABLE 1.
Serology and PCR results of the six patientsa
| Patient no. | Date (day/mo/yr) | First EIA (OD/CO) | Second EIA (OD/CO) | Particle agglutination for:
|
Antigen(s) used for Western blot | PCR HIV-1 | |
|---|---|---|---|---|---|---|---|
| HIV-1 | HIV-2 | ||||||
| 1 | 03/06/94 | Reactive (>19.13) | Nonreactive (0.9) | 1+ | Not done | gp160 | Not done |
| 09/06/97 | Reactive (>15.17) | Reactive (2.95) | 2+ | Negative | gp160 | Not done | |
| 17/06/97 | Reactive (>15.17) | Reactive (1.39) | 2+ | Negative | gp160 | Negative | |
| 02/07/97 | Reactive (>15.83) | Reactive (2.21) | 2+ | 2+ | gp160 | Negative | |
| 14/07/97 | Reactive (>17.60) | Reactive (1.20) | 2+ | Negative | p66, gp160 | Not done | |
| 18/12/98 | Reactive (>15.28) | Reactive (1.93) | 2+ | Negative | gp160 | Not done | |
| 2 | 04/03/94 | Reactive (13.77) | Reactive (1.10) | 1+ | Not done | p24, p55, gp160 | Not done |
| 04/05/95 | Reactive (3.62) | Nonreactive (0.17) | Negative | Not done | p24, p51, p55, gp160 | Negative | |
| 3 | 28/04/98 | Reactive (1.42) | Reactive (1.75) | 2+ | Negative | gp120, gp160 | Negative |
| 19/05/98 | Reactive (1.39) | Insufficient serum | 2+ | Negative | gp120, gp160 | Not done | |
| 4 | 15/08/97 | Reactive (1.34) | Reactive (7.58) | 2+ | Negative | p24, gp160 | Not done |
| 21/08/97 | Reactive (1.15) | Reactive (2.40) | 2+ | Negative | p24, gp160 | Not done | |
| 01/09/97 | Reactive (1.35) | Reactive (3.01) | 2+ | Negative | p24, gp160 | Negative | |
| 16/02/98 | Reactive (1.19) | Reactive (2.38) | 2+ | Negative | p24, gp160 | Negative | |
| 5 | 18/09/97 | Reactive (14.60) | Reactive (1.78) | 2+ | Negative | gp160 | Negative |
| 6 | 08/09/97 | Reactive (>17.32) | Reactive (3.56) | 2+ | Negative | gp160 | Negative |
Patient number 4 was an antenatal patient. EIA, enzyme immunoassay.
From March 1994 to September 1997, 2,256 samples were submitted to our laboratory for the confirmation of HIV antibodies. The samples, referred by screening laboratories, had either reactive or grey zone HIV screening test results and were then subjected to Strategy III. Altogether, 61 samples had reactive results for all three immunoassays but had at least one low OD/CO result. Western blot testing confirmed 31 (50.8%) to be positive, 2 (3.3%) negative, and 28 (45.9%) indeterminate. Almost half of such samples could have been mislabeled as HIV positive according to the Strategy III criteria. The six samples described were selected because PCR results as well as serial serology were available.
Hence, even the most stringent alternative HIV confirmatory strategy requires judicious use. Samples giving low positive OD/CO values should be subjected to other supplementary tests, such as the Western blot assay and/or PCR. Guidelines should be developed in each laboratory as to which cases require further supplementary tests. Clinical risk assessment is also necessary to decide if asymptomatic patients require follow-up serology.
REFERENCE
- 1.World Health Organization. 1999. Operational characteristics of commercially available assays to determine antibodies to HIV-1 and/or HIV-2 in human sera. Report 11, WHO/BTS/99.1, UNAIDS/99.5. Joint United Nations Program on HIV/AIDS, Geneva, Switzerland.