Analysis of HIV incidence is the most direct approach for measuring the efficacy of interventions for HIV prevention.1 HIV incidence estimates can be obtained through repeated testing of individuals in longitudinal cohorts.2 However, the longitudinal cohorts may be difficult to establish and costly to maintain.3 They may also suffer from bias related to loss of follow-up.4 An alternative approach for HIV incidence estimation relies on tests that distinguish recent from nonrecent infection in a cross-sectional survey.5 HIV incidence can be estimated from cross-sectional surveys by measuring biomarkers that evolve during the course of HIV infection.6 Many cross-sectional incidence assays measure antibody maturation as a marker of duration of HIV infection (reviewed by Murphy and Parry7 and Guy et al8).
One limitation of using serologic assays for cross-sectional HIV incidence estimation is that some individuals have immature-appearing antibody a year or more after infection. Many factors are associated with “false-recent” misclassification, including low HIV viral load, low CD4 cell count, and long-term use of antiretroviral therapy (ART).9–13 We previously reported that the frequency of false-recent misclassification varies in different regions of Africa.14 Particularly, high rates of false-recent misclassification were observed using the BED capture enzyme immunoassay (BED-CEIA)15 or an antibody avidity assay16 in Eastern Africa, where subtypes A and D predominate.17 The frequency of false-recent misclassification is higher in those infected with subtype D HIV, compared with those with subtype A infection.18 Other studies have also noted subtype-based differences in cross-sectional incidence assay performance.19–21 In Uganda, women with subtype D infection were more likely to have low BED-CEIA results and lower antibody avidity than women with subtype A infection.18,22 Subtype D HIV has been shown to be more pathogenic than subtype A HIV.23 It was not clear whether the high frequency of false-recent misclassification in subtype D-infected individuals was because of faster disease progression (eg, faster progression to AIDS, with a waning antibody response24) or some other mechanism associated with a weak initial humoral response to HIV infection that was sustained over time. In this study, we used the BED-CEIA and the avidity assay to analyze the humoral response to HIV infection in adult women with subtypes A and D HIV infection.
The Genital Shedding and Disease Progression (GS) Study evaluated the use of hormonal contraceptives, genital shedding of HIV, and HIV disease progression among 303 Ugandan and Zimbabwean women with known dates of seroconversion.25 We analyzed 2614 samples from a subgroup of Ugandan women, aged 18–45 years, who were infected with HIV subtype A (N = 84) or subtype D (N = 34) and who had samples available from at least 3 study visits after HIV seroconversion, including at least 1 sample collected a year or more after seroconversion (2001–2009). The median number of samples per woman was 23 (range, 3–41) and the median follow-up was 6.56 years (range, 0.13–9.19 years). During the course of follow-up, 38 women initiated ART. CD4 cell count, viral load, and HIV subtype data were determined previously.25,26 Date of HIV seroconversion was defined as either the midpoint between the last negative HIV antibody test and the first positive HIV antibody test or 15 days after acute infection was documented (defined as HIV RNA positive/HIV antibody negative).25
The BED-CEIA was performed according to the manufacturer’s directions (Calypte Biomedical Corporation, Lake Oswego, OR), with 1 exception: samples were run in duplicate and the average result was reported; results were reported as normalized optical density units (OD-n).14,18 The BED-CEIA measures the proportion of total IgG that binds to a branched synthetic tripeptide that contains three 18-amino acid components derived from an immundominant region of gp41 (region corresponding to positions 590–607 of HXB2 gp160 in HIV subtypes B, E, and D).27 The avidity assay was performed using the Genetic Systems HIV-1/HIV-2 PLUS O EIA (Bio-Rad Laboratories, Redmond, WA) with the following modifications: plate covers were used rather than sealers; an incubation time of 30 minutes rather than of 60 minutes was used after adding cold diluent to each sample; and deionized water was used rather than wash buffer solution when preparing the diethylamine solution.16 The Genetic Systems HIV-1/HIV-2 PLUS O EIA is an enzyme immunoassay that uses a direct antibody sandwich technique to detect anti-HIV antibodies. The percent avidity [Avidity Index (AI)] is calculated for each sample by dividing the optical density of the diethylamine-treated well by the optical density of the nontreated well for the same sample and multiplying by 100.16 Maturation of the immune response to HIV infection assessed within the first year and within the first 2 years after seroconversion; for each time interval, we assessed whether the BED-CEIA result rose above the cutoff of 0.8 OD-n and whether the AI rose above the cutoffs of 40%14,16,28,29 or 90%. Antibody maturation was classified as “normal” if the BED-CEIA value was >0.8 OD-n and the AI was >40% by 2 years after seroconversion. Antibody regression was defined as decline in the BED-CEIA or the AI result of at least 20% compared with the maximum value measured for the same women at an earlier time point.
For samples collected >2 years after seroconversion, Fisher exact test or the χ2 test were used to identify factors associated with false-recent misclassification using the BED-CEIA and the avidity assay. These factors included age, HIV viral load, CD4 cell count, duration of ART at the time of sample collection, and the year of sample collection. Logistic regression using general estimating equations was performed to determine the odds of false-recent misclassification for all factors analyzed.30 Factors that were associated with false-recent misclassification in the univariate analysis with P < 0.15 were included in a multivariate logistic regression model. All analyses were performed using STATA v11 (StataCorp, College Station, TX).
Subtype Distribution in the Study Cohort
The women included in this study had either subtype A or subtype D infection (overall, 72% and 28%, respectively). The relative proportions of these subtypes were not significantly different for most of the subgroups included in the analysis (Table 1). The following exceptions were noted: (1) for women on ART < 1 year: 56.5% (87/154) samples were subtype A and 43.5% (67/154) samples were subtype D (P < 0.05), (2) for women on ART > 1 year: 27.0% (55/204) samples were subtype A and 73.0% (149/204) samples were subtype D (P < 0.01). Most of the samples were collected >2 years after seroconversion: 68.5% (1256/1833) for subtype A and 71.5% (558/781) for subtype D. There were 385 samples collected between 1 and 2 years after seroconversion, none of which were from individuals on ART.
Antibody Maturation in Women With Subtypes A and D Infection
The proportion of women whose BED-CEIA results remained <0.8 OD-n was similar for subtype A (Fig. 1A) and subtype D (Fig. 1B) during the first year after seroconversion [48% (39/82) vs. 50% (16/32), P = 0.83] and during the first 2 years after seroconversion [21% (17/82) vs. 16% (5/32), P = 0.60, Table 2]. Over the entire follow-up period (median follow-up, 5.9 years), 10% (8/82) of women with subtype A and 9% (3/32) of women with subtype D infection never attained a BED-CEIA result above 0.8 OD-n.
The proportion of women whose avidity assay result remained <40% was 0% (0/82) for subtype A (Fig. 2A) and 6% (2/32) for subtype D infection (Fig. 2B) during the first year after seroconversion (P = 0.084); the avidity assay result rose above 40% for all women by 2 years after seroconversion (Table 2). The proportion of women whose avidity result did not rise above 90% after seroconversion was lower for women with subtype A than for women with subtype D [within 1 year: 17% (14/82) vs. 56% (18/32), P < 0.01; within 2 years: 0% (0/82) vs. 34% (11/32), P < 0.01]. Over the entire follow-up period (median follow-up, 5.9 years), none of the women with subtype A and 19% (6/32) of women with subtype D infection never attained an avidity result >90%.
Antibody Regression in Women With Subtypes A and D Infection
Regression in antibody response was defined by a decrease of >20% in the BED-CEIA or avidity results compared with the maximum value obtained for the same women at an earlier study visit. Antibody regression was observed for 40 women using the BED-CEIA and for 3 women using the avidity assay. For both the assays, regression was more common for the subtype D-infected women for the BED-CEIA: 33% (27/82) for subtype A vs. 41% (13/32) for subtype D, P = 0.51; and for the avidity assay: 11% (1/82) for subtype A vs. 6% (2/32) for subtype D, P = 0.19 (Table 2). Antibody regression was associated with ART initiation. In this cohort of women, 24 women with subtype A and 14 women with subtype D infection initiated ART during the follow-up. Among the women with subtype A who initiated ART, 63% (15/24) saw regression in the BED-CEIA values, whereas none of these women saw a decrease in avidity values. For the 14 women with subtype D initiating ART, 79% (11/14) had regression in their BED-CEA and 21% (3/14) had regression in their avidity values. The timing of ART initiation was also associated with the antibody regression. The proportion of women who had antibody regression measured with the BED-CEIA was higher for those who started ART >3 years after seroconversion [78% (22/26)] than for those who started ART earlier [15% (4/26), (P < 0.01)]. In contrast, there was no significant difference in antibody regression measured with the avidity assay among women who started ART less than vs. more than 3 years after seroconversion [50% (1/2) for both the groups]. Some women with subtype A infection [8% (2/24)] had increasing BED-CEIA values after starting ART; in contrast, none of the women had increasing avidity values after starting ART [0% (0/24)].
Because ART treatment can confound the impact of subtype on the regression of response as measured by the BED-CEIA and avidity assay, we performed an analysis that was limited to those who were never treated. Individuals who were never treated displayed similar progression trends as those presented in Table 2. One individual with subtype A infection and 1 individual with subtype D infection demonstrated antibody regression by the avidity assay (with AI < 40% and AI < 90%, respectively). The proportion of individuals who had a regression of BED-CEIA values was higher for individuals with subtype A compared with subtype D infection; however, this difference was not statistically significant (12/58 = 21% vs. 2/18 = 11%, respectively; P = 0.36). Among individuals who had a regression in BED-CEIA values (both on ART and not on ART), a higher proportion of nontreated individuals with subtype A had regression compared with those with subtype D infection; this difference was of borderline significance (12/27 = 44% vs. 2/13 = 15%, respectively; P = 0.07). When this analysis was restricted to individuals on ART, there was no significant difference in regression of the BED-CEIA values obtained for those with subtype A vs. D infection (15/24 = 63% vs. 11/14 = 79%, respectively, P = 0.30). In contrast, a difference of borderline statistical significance was observed for the avidity assay for these 2 groups (0.24 = 0% vs. 2/14 = 14%, respectively; P = 0.057).
Analysis of Factors Associated With False-Recent Misclassification
Overall, 17.6% (319/1814) of the samples from women more than 2 years after seroconversion had BED-CEIA values <0.8 OD-n; in contrast, only 3.0% (55/1814) had avidity assays <40% and only 3.9% (70/1814) had avidity values <90% in this time frame.
The following factors were associated with false-recent misclassification using the BED-CEIA in univariate analysis (Table 3 and Table S1, Supplemental Digital Content,https://links.lww.com/QAI/A472): CD4 cell count < 200 cells per microliter (P < 0.05); missing CD4 data (P < 0.05); viral load < 400 copies per milliliter (P < 0.01); and on ART < 1 year (P < 0.01). In a multivariate model, false-recent misclassification was associated with missing CD4 cell count data (P < 0.05) and on ART < 1 year (P < 0.01) and on ART > 1 year (P < 0.01). In a multivariate model, false-recent misclassification using the avidity assay was associated with the duration of infection (48–72 months, P < 0.01) and HIV subtype (P < 0.05). We compared results obtained using the BED-CEIA and avidity assays. False-recent misclassification with the BED-CEIA was seen in 17.3% (305/1759) samples that had avidity assay results >40% and in 25.5% (14/55) samples with avidity assay results <40%; this difference was not statistically significant.
For samples collected more than 2 years after seroconversion and stratified by subtype, 0.0% (0/1256) samples from individuals with subtype A infection were misclassified by the avidity assay, compared with 9.9% (55/558) samples from individuals with subtype D infection (P < 0.001; Table S1, Supplemental Digital Content,https://links.lww.com/QAI/A472). Similar differences were observed for the BED-CEIA 15.5% (195/1256) samples from individuals with subtype A infection were misclassified compared with 22.2% (124/558) samples from individuals with subtype D infection (P < 0.001).
There were 157 samples (95 subtype A and 65 subtype D) that did not have CD4 cell count data associated with them. Ninety-two (62 subtype A and 30 subtype D) of those samples were collected at a time point early in infection (median 110 days after seroconversion) from women who were not on ART. Eighty-six percent [79 (56 subtype A and 23 subtype D)/92] samples was misclassified by the BED-CEIA, 46% [42 (31 subtype A and 11 subtype D)/92] misclassified by the avidity assay, and 95% (40/42) of samples that misclassified by the avidity assay was misclassified by the BED-CEIA. Two samples both subtype D infected of the remaining 65 samples were collected early in infection (<2 years) and on ART. Both the samples were misclassified by the BED-CEIA and neither misclassified by an AI cutoff of 40%. The remaining 63 (33 subtype A and 30 subtype D) samples were collected later in infection (>2 years), and 76% (48/63) were recently placed on ART (median, 28 days). Eight percent (5/63) was misclassified by the BED-CEIA all subtype D and 10% (6/63) misclassified by an AI cutoff <40% all of which were subtype D. No sample that was misclassified by the avidity assay was misclassified by the BED-CEIA.
Previous studies in other cohorts have demonstrated that individuals with long-standing subtype D HIV infection often have low levels of anti-HIV antibodies detected with serologic HIV incidence assays.14,18 This type of false-recent misclassification lowers the precision of these assays for HIV incidence estimation in populations where subtype D is prevalent. This report investigated the biologic basis for false-recent misclassification in adults with subtype D infection by evaluating antibody maturation and antibody regression in a cohort of women with subtypes A and D infection. We found that differences in the performance of the BED-CEIA and avidity assays in these 2 subtypes reflect a weaker initial antibody response to HIV infection in those with subtype D infection. We did not see a significant difference in antibody regression (decline in BED-CEIA and avidity assay results) in these 2 subtypes. For both the subtypes, antibody regression was associated with the initiation of ART and duration of treatment.
A longer time on ART was associated with lower OD-n by the BED-CEIA but not a lower AI by the avidity assay because of the nature of the assay. The BED-CEIA measures the proportion of IgG antibody that is directed to an immunodominant region of gp41. When the level of circulating HIV antigens decreases (with viral suppression) antibody production may be downregulated, and the titers of anti-HIV antibodies may fall. This decrease in antibody titers may result in lower BED-CEIA values.13,14,31 In contrast, the avidity assay measures the strength of antibody binding to target antigens in the presence of a chaotropic agent. A decrease in the amount of circulating antigen would not necessarily impact antibody avidity and is, therefore, less likely to impact the results obtained with an avidity assay. A previous study demonstrated a lack of impact of decreasing antigen (as measured by decreasing HIV viral load) on the avidity assay used in this report.31
In a previous study of individuals from the Rakai Community Cohort Study (RCCS), we observed more frequent false-recent misclassification among women with subtype D compared with women with subtype A; this association was not observed for men with these subtypes.18 In this report, we extended that study by showing that 9% women with subtype D infection did not attain a mature antibody response even after many years of infection; this failure of antibody maturation was uncommon in women with subtype A infection. This study also extends our previous study of the RCCS, because the GS cohort analyzed in this report included women who initiated ART during the study follow-up. In the GS cohort, women with subtype D were less likely to attain BED-CEIA values above the assay cutoff; this indicates that they maintained low levels of HIV-specific IgG in their serum, even years after seroconversion.22 BED-CEIA values also declined over time in more women with subtype D than in those with subtype A; this effect, which we refer to as antibody regression, was associated with ART use.
The findings in this report support the use of multiassay algorithms (MAAs) for cross-sectional HIV incidence estimation.22 The performance of MAAs is likely to be improved when they include 2 serologic assays that measure different features of the immune response to HIV infection; in this study, false-recent misclassification by the BED-CEIA and avidity assay was not associated.22 In this report, low CD4 cell count and duration of time on ART were associated with misclassification by the BED-CEIA. Use of ART for more than 1 year doubled the frequency of false-recent misclassification by the BED-CEIA; the duration of ART was also associated with false-recent misclassification by the BED-CEIA in previous studies.14,32 Inclusion of nonserologic biomarkers in MAAs, such as CD4 cell count and HIV viral load, may further reduce the frequency of false-recent misclassification.14 For both the subtypes A and D, the performance of the avidity assay was superior to that of the BED-CEIA.
The high prevalence of subtype D HIV infections in Uganda likely explains the high frequency of false-recent misclassification observed with serologic incidence assays in populations from Eastern Africa.14 Our results suggest that even higher frequencies of false-recent misclassification will be observed as ART use increases in those populations. Increasing ART use may limit the utility of serologic HIV incidence assays in populations where subtype D is prevalent, even where those assays are used in parallel or in serial testing algorithms.18 Although these findings reveal differences in the immune responses to HIV infection in individuals with subtypes A and D infection, further experiments are needed to determine if these differences are specific to the 2 incidence assays used in this study or are more general in nature. To evaluate whether the antibody responses are generally weaker in individuals with subtype D infection, it would be necessary to evaluate the antibody responses to a broader spectrum of HIV antigens and to evaluate other features of the humoral immune response to HIV infection (eg, antibody neutralization and antibody isotype).
1. Busch MP, Pilcher CD, Mastro TD, et al.. Beyond detuning: 10 years of progress and new challenges in the development and application of assays for HIV incidence estimation. AIDS. 2010;24:2763–2771.
2. Beyrer C, Nelson KE. Loss to follow-up effect in investigations of HIV-1 incidence. Lancet. 1997;349:649–650.
3. Brookmeyer R. Measuring the HIV/AIDS epidemic: approaches and challenges. Epidemiol Rev. 2010;32:26–37.
4. Brookmeyer R. Accounting for follow-up bias in estimation of human immunodeficiency virus incidence rates. J R Stat Soc. 1997;160:127–140.
5. Barnighausen T, McWalter TA, Rosner Z, et al.. HIV incidence estimation using the BED capture enzyme immunoassay: systematic review and sensitivity analysis. Epidemiology. 2010;21:685–697.
6. Brookmeyer R, Quinn TC. Estimation of current human immunodeficiency virus incidence rates from a cross-sectional survey using early diagnostic tests. Am J Epidemiol. 1995;141:166–172.
7. Murphy G, Parry JV. Assays for the detection of recent infections with human immunodeficiency virus type 1. Euro Surveill. 2008;13:pii: 18966.
8. Guy R, Gold J, Calleja JM, et al.. Accuracy of serological assays for detection of recent infection with HIV and estimation of population incidence: a systematic review. Lancet Infect Dis. 2009;9:747–759.
9. Barnighausen T, Wallrauch C, Welte A, et al.. HIV incidence in rural South Africa: comparison of estimates from longitudinal surveillance and cross-sectional cBED assay testing. PLoS One. 2008;3:e3640.
10. Karita E, Price M, Hunter E, et al.. Investigating the utility of the HIV-1 BED capture enzyme immunoassay using cross-sectional and longitudinal seroconverter specimens from Africa. AIDS. 2007;21:403–408.
11. Marinda ET, Hargrove J, Preiser W, et al.. Significantly diminished long-term specificity of the BED capture enzyme immunoassay among patients with HIV-1 with very low CD4 counts and those on antiretroviral therapy. J Acquir Immune Defic Syndr. 2010;53:496–499.
12. Hladik W, Olara D, Mermin J, et al.. Effect of CD4(+) T cell count and antiretroviral treatment on two serological HIV incidence assays. AIDS Res Hum Retroviruses. 2012;28:95–99.
13. Hayashida T, Gatanaga H, Tanuma J, et al.. Effects of low HIV type 1 load and antiretroviral treatment on IgG-capture BED-enzyme immunoassay. AIDS Res Hum Retroviruses. 2008;24:495–498.
14. Laeyendecker O, Brookmeyer R, Mullis C, et al.. Specificity of four laboratory approaches for cross-sectional HIV incidence determination: analysis of samples from adults with known non-recent HIV infection from five African countries. AIDS Res Hum Retroviruses. 2012;28:1177–1183.
15. Dobbs T, Kennedy S, Pau CP, et al.. Performance characteristics of the immunoglobulin G-capture BED-enzyme immunoassay, an assay to detect recent human immunodeficiency virus type 1 seroconversion. J Clin Microbiol. 2004;42:2623–2628.
16. Masciotra S, Dobbs T, Candal D, et al.. Antibody avidity-based assay for identifying recent HIV-1 infections based on Genetic Systems TM 1/2 plus O EIA. Paper presented at: 17th Conference on Retroviruses and Opportunistic Infections, Abstract 937; February 16–19, 2010; San Francisco, CA.
17. Hemelaar J, Gouws E, Ghys PD, et al.. Global trends in molecular epidemiology of HIV-1 during 2000-2007. AIDS. 2011;25:679–689.
18. Mullis CE, Munshaw S, Grabowski MK, et al.. Differential misclassification of HIV-1 cross-sectional incidence assays by subtype in Rakai, Uganda
. Paper presented at: 19th Conference on Retroviruses and Opportunistic Infections, Abstract 541; March 5-8, 2012; Seattle, WA.
19. Young CL, Hu DJ, Byers R, et al.. Evaluation of a sensitive/less sensitive testing algorithm using the bioMerieux Vironostika-LS assay for detecting recent HIV-1 subtype B' or E infection in Thailand. AIDS Res Hum Retroviruses. 2003;19:481–486.
20. Parekh BS, Hanson DL, Hargrove J, et al.. Determination of Mean recency period for estimation of HIV type 1 incidence with the BED-capture EIA in persons infected with diverse subtypes. AIDS Res Hum Retroviruses. 2011;27:265–273.
21. Grabowski M, Redd A, Mueller A, et al.. The natural selection of HIV subtypes due to differential infectivity and pathogenicity is accelerated by ART: Rakai, Uganda
. Paper presented at: 19th Conference on Retroviruses and Opportunistic Infections, Abstract 521; March 5-8, 2012; Seattle, WA.
22. Chen P-L, et al.. Dual testing algorithms of BED-CEIA and BioRad ½ + O EIA avidity assays in evaluating HIV-1 false-recent misclassification in a longitudinal sample of African women. Paper presented at: XIX International AIDS Conference, abstract TUPE207; July 22-27, 2012; Washington, DC.
23. Kiwanuka N, Laeyendecker O, Robb M, et al.. Effect of human immunodeficiency virus type 1 (HIV-1) subtype on disease progression in persons from Rakai, Uganda
, with incident HIV-1 infection. J Infect Dis. 2008;197:707–713.
24. Thomas HI, Wilson S, O’Toole CM, et al.. Differential maturation of avidity of IgG antibodies to gp41, p24 and p17 following infection with HIV-1. Clin Exp Immunol. 1996;103:185–191.
25. Morrison CS, Chen PL, Nankya I, et al.. Hormonal contraceptive use and HIV disease progression among women in Uganda
and Zimbabwe. J Acquir Immune Defic Syndr. 2011;57:157–164.
26. Morrison CS, Demers K, Kwok C, et al.. Plasma and cervical viral loads among Ugandan and Zimbabwean women during acute and early HIV-1 infection. AIDS. 2010;24:573–582.
27. Parekh BS, Kennedy MS, Dobbs T, et al.. Quantitative detection of increasing HIV type 1 antibodies after seroconversion: a simple assay for detecting recent HIV infection and estimating incidence. AIDS Res Hum Retroviruses. 2002;18:295–307.
28. Mullis CE, Munshaw S, Grabowski MK, et al.. Differential specificity of HIV incidence assays in HIV subtypes A and D-infected individuals from Rakai, Uganda
. AIDS Res Hum Retroviruses. 2013;29:1146–1150.
29. James MM, Laeyendecker O, Sun J, et al.. Antibody maturation and viral diversification in HIV-infected women. PLoS One. 2013;8:e57350.
30. Zeger SL, Liang KY. Longitudinal data analysis for discrete and continuous outcomes. Biometrics. 1986;42:121–130.
31. Wendel SK, Mullis CE, Eshleman SH, et al.. Effect of natural and ARV-induced viral suppression and viral breakthrough on anti-HIV antibody proportion and avidity in patients with HIV-1 subtype B infection. PLoS One. 2013;8:e55525.
32. Laeyendecker O, Brookmeyer R, Oliver AE, et al.. Factors associated with incorrect identification of recent HIV infection using the BED capture immunoassay. AIDS Res Hum Retroviruses. 2012;28:816–822.