Potential for False-Positive Results with Serological Assays for HIV in Central Africa: Implications for the HIV Serodiagnosis Algorithm According to the 2015 Consolidated WHO Recommendations for Resource-Constrained Countries

Abstract

Editor: A high frequency of false-positive results with “alternative” testing algorithms using certain HIV rapid tests, as proposed by the WHO, has been observed in Central Africa.^1,2 Inaccurately diagnosing an HIV infection may induce significant psychological distress in individuals and may result in the unnecessary prescription of antiretroviral therapy. The recent WHO recommendations for HIV testing in resource-limited settings, updated in 2012 and consolidated in 2015,³ use combinations of two to three HIV tests in series and basically avoid the so-called tie-breaker strategy for final results in case of discordance between tests results. To maximize the accuracy of results, choosing the optimal combination of HIV tests is of paramount importance. A field evaluation was conducted to examine the proportion of discordance among different combinations of HIV tests, including rapid tests, used on sera from individuals from Cameroon, an area of high HIV prevalence and broad HIV genetic diversity.⁴

A total of 464 sera from HIV-negative individuals were prospectively collected in Cameroon and were further tested in parallel by two to five HIV serological assays, as previously reported.² The number and proportions of sera classified as “inconclusive” (neither positive nor negative), that is, giving discrepant tests results by a given combination of 2, 3, 4, or 5 HIV tests,² carried out in parallel, are shown in Figure 1. The total number of “inconclusive” combinations was 227 with the two-test combination, 319 with the three-test combination, 211 with the four-test combination, and 52 with the five-test combination (Fig. 1A). The percentage of individuals with discordant test results increased significantly and proportionally as more tests were used, ranging from an average of 4.8% of individuals when two tests were used, 6.8% of individuals when three tests were used, 9.4% of individuals when four tests were used, and 11.2% of individuals when five tests were combined (p < .001) (Fig. 1B). Finally, the number of “inconclusive” combinations with all tests of the combination being either all positive or inconclusive decreased significantly with the number of tests in the combination: 16 (7.1%) with the two-test combination, 10 (3.1%) with the three-test combination, 5 (2.4) with the four-test combination, and 1 (1.9%) with the five-test combination (Fig. 1C) (p < .04). All blood samples from individuals with inclusive results (whatever the combination) were all negative for HIV-1 LTR PCR in dried blood spot collected on using Whatman 903 filter paper.

FIG. 1.

Absolute number (n) (A) and average percentages (%) (B) of discordant or “inconclusive” combinations of two to five tests, and percentage of combinations with all tests being either positive, “doubtful,” or indeterminate, thus subjected to misinterpretation as false-positive results (C). Results are expressed as a function of number of tests performed in combination (2–5) among five serological screening tests for HIV (four rapid tests and one enzyme-linked immune assay, as described²), yielding “inconclusive” results from a panel of 464 serum samples from adults living in Cameroon. The number N of combinations of p HIV tests is given by the formula N = C_n ^p = n!/p! (n − p)! (N is equal to 10 for two and three combinations and equal to 5 for four combinations of HIV tests). The variations in the number of inconclusive combinations ranged from 0.6% to 11.8% with the two-test combination, from 1.1% to 10.6% with the three-test combination, and from 3.5% to 11.2% with the four-test combination. The sera were tested in parallel by the following commercially available HIV serological assays: Alere Determine™ HIV-1/2 (Alere, Inc., Waltham, MA), Alere Determine Test HIV-1/2 Ag/Ab Combo (Alere, Inc.), ImmunoComb^® II HIV 1 and 2 CombFirm Kit (Alere, Inc.), SD Bioline HIV 1/2 3.0 (Standard Diagnostics, Inc., Kyonggi-do, South Korea), and Genscreen Plus HIVAg-Ab (Bio-Rad, Marnes la Coquette, France).

We further applied the sequential 2015 consolidated WHO testing algorithm in high prevalence settings (>5%) to the three tests providing the lowest number of inconclusive sera when tested in parallel,³ including the following HIV serological assays: Alere Determine™ HIV-1/2 [T1], SD Bioline HIV 1/2 3.0 [T2], and Genscreen Plus HIVAg-Ab [T3]. The latter combination provided 10 inconclusive sera, none of which being positive or indeterminate by all three tests. The results are shown in Table 1. The number of inconclusive (I) sera was less important when the tests are used in series (I: 1–9) than in parallel (I = 10). Certain combinations gave rise to higher numbers of inconclusive sera than others. Finally, it was possible to find and validate with field sera serological combination of three tests used sequentially giving rise to very low numbers of inconclusive sera, and harboring 100% sensitivity and very high specificity, such as T2 → T5 → T1 or T5 → T2 → T1 (Table 1).

Table 1.

Characteristics and Youden' J Index of the Six Sequential Test Combinations According to the 2015 Consolidated WHO HIV Testing Algorithm of the Three Tests T1, T2, and T3, Providing Less Number of Inconclusive Sera When Tested in Parallel

Sequential combination ^a	Incon-clusive sera (I)	N′ = N − I	True positive	True negative	False positive	Sensitivity ^b [95% CI]	Specificity ^b [95% CI]	Youden' J index
T1 → T2 → T5	9	455	26	427	2	100% [99.6–100.0]	99.5% [98.8–100.0]	1
T1 → T5 → T2	2	462	26	433	3	100% [99.6–100.0]	99.3% [98.5–100.0%]	0.99
T2 → T1 → T5	1	463	26	435	2	100% [99.6–100.0]	99.5% [98.9–100.0]	1
T2 → T5 → T1	2	462	26	435	1	100% [99.6–100.0]	99.8% [99.3–100.0]	1
T5 → T1 → T2	1	463	26	434	3	100% [99.6–100.0]	99.3% [98.5–100.0]	0.99
T5 → T2 → T1	3	461	26	434	1	100% [99.6–100.0]	99.8% [99.3–100.0]	1

T1: Alere Determine™ HIV-1/2; T2: SD Bioline HIV 1/2 3.0; T3: Genscreen Plus HIVAg-Ab.

The sensitivity and specificity of a given sequential combination of the three tests T1, T2, and T3 according to the 2015 consolidated WHO testing algorithm in high-prevalence settings (>5%³) were estimated by taking into account for calculation N′, corresponding to the size of the study population (N) deleted of the number of inconclusive sera according to the combination used.

CI, confidence interval; I, number of inclusive sera with a given sequential combination; N, effective of the study population (N = 464); N′: N − I; Youden's J index = sensitivity + specificity – 1.

Inconclusive serological diagnosis of HIV infection is particularly frequent in Central Equatorial Africa,² especially for some HIV tests and HIV test combinations.¹ Our observations demonstrate that the frequency of inconclusive sera depends on the nature of the serological tests used in a given combination and that it is also directly proportional to the number of HIV tests used in parallel: the more the serological tests used, the higher the frequency of inconclusive sera. The decline in inconclusive results is the reason behind the WHO recommendation of using a series of tests. The 2015 consolidated WHO alternative strategy suggests using highly sensitive first-line assays, followed by highly specific second- and third-line assays.³ Our results show that even with three tests used in combination, as proposed in the 2015 WHO recommendations,³ the risk of false HIV-positive reactions may persist, even with the use of fourth generation combo enzyme-linked immune assay. However, the risk may be very limited by certain combinations of HIV tests. Taken together, these findings confirm that the best tests and test combinations must absolutely be carefully validated on the field with local sera panel corresponding to given environmental and human biotopes.

Footnotes

Acknowledgment

M.-A.J. is the holder of the Canada research chair tier 2 in immunovirology.

Author Disclosure Statement

No competing financial interests exist. C.T.C is a FRQ-S Junior 1 awardee.

References

Klarkowski

, O'Brien

, Shanks

, Singh

: Causes of false-positive HIV rapid diagnostic test results. Expert Rev Anti Infect Ther, 2014; 12:49–62.

Mbopi-Keou

, Ndjoyi-Mbiguino

, Talla

, et al.: Association of inconclusive sera for human immunodeficiency virus infection with malaria and Epstein-Barr virus infection in Central Africa. J Clin Microbiol, 2014; 52:660–662.

World Health Organization: Consolidated guidelines on HIV testing services. Available at www.who.int/hiv/pub/guidelines/hiv-testing-services/en (2015), accessed March 3, 2017 .

Aghokeng

, Mpoudi-Ngole

, Dimodi

, et al.: Inaccurate diagnosis of HIV-1 group M and O is a key challenge for ongoing universal access to antiretroviral treatment and HIV prevention in Cameroon. PLoS One, 2009; 4:e7702.