Decreased Seroreactivity in Individuals Initiating Antiretroviral Therapy during Acute HIV Infection

Antiretroviral therapy (ART) during acute HIV infection (AHI) interrupts viral dynamics and may delay the emergence of serological markers targeted by current HIV screening and confirmatory assays, thus creating challenges for correctly classifying HIV infection status.

ABSTRACT

Antiretroviral therapy (ART) during acute HIV infection (AHI) interrupts viral dynamics and may delay the emergence of serological markers targeted by current HIV screening and confirmatory assays, thus creating challenges for correctly classifying HIV infection status. The performance of three HIV antigen/antibody combination (HIV Ag/Ab Combo) assays (the Bio-Rad GS, Abbott Architect, and Bio-Rad BioPlex 2200 assays) was evaluated with samples collected from RV254/South East Asia Research Collaboration in HIV 010 (RV254/SEARCH010) study (Bangkok, Thailand) participants at weeks 12 and 24 following the initiation of ART at Fiebig stage I (FI) (n = 23), FII (n = 39), or FIII/IV (n = 22). Supplemental, confirmatory testing was performed by the Geenius HIV 1/2 and HIV-1 Western blot assays (Bio-Rad). Samples from 30 untreated, HIV-1-infected individuals demonstrated robust HIV Ag/Ab Combo assay reactivity with well-developed HIV-1 Western blotting profiles by 24 weeks after infection. In contrast, 52.2% of samples from individuals initiating ART at FI, 7.7% of samples from individuals initiating ART at FII, and 4.5% of samples from individuals initiating ART at FIII/IV were nonreactive by the HIV Ag/Ab Combo assays, with 36.4 to 39.1% of samples having low signal-to-cutoff (S/CO) results by the Architect and BioPlex assays (S/CO < 10). Seroreversion from a reactive to a nonreactive status was observed in 10 individuals initiating ART at FII and 3 individuals initiating ART at FIII/IV. The Geenius and HIV-1 Western blot assay results were negative or indeterminate for 73.9% and 69.6% of individuals, respectively, treated at FI; 50.0% and 26.3% of individuals, respectively, treated at FII; and 54.5% and 40.9% of individuals, respectively, treated at FIII/IV. Virologic suppression of HIV-1 by ART during AHI impedes seroconversion to biomarkers of infection, limiting the utility of HIV Ag/Ab Combo and supplemental, confirmatory assays for infection status determination.

INTRODUCTION

Current guidelines recommend the initiation of antiretroviral therapy (ART) as soon as possible after the diagnosis of HIV infection to minimize the risk of forward transmission, interrupt the seeding of latent reservoirs, delay disease progression, and ultimately, enhance overall health outcomes (1–7). However, the reduction or elimination of viral expansion by ART in acute HIV infection (AHI) to below the threshold required for evolution of an anti-HIV immune response may block or delay the emergence of HIV-specific antigen (Ag) and/or antibody (Ab) targets for detection of HIV infection (8–11). Negative results on standard HIV Ab tests are common among HIV-1-infected children who initiate ART at an early age (12–15). At 2 years of age, about half of all HIV-1-infected infants who initiated ART within 2 weeks of birth were seronegative by HIV immunoassays and rapid device tests, whereas 11% were seronegative if ART was initiated at 12 and 24 weeks of age. All children initiating ART later than 24 weeks of age were seropositive by HIV immunoassays at 2 years of age (6, 16). In the absence of ongoing antigenic stimulation, seroconversion, as determined by second-generation immunoassay Ab tests, also failed to occur among adult participants treated during AHI (9, 15, 17, 18). Several studies reported seroreversion from reactive to nonreactive results by HIV Ag/Ab combination (Combo) assays in HIV-infected participants who initiated ART during AHI (8, 14, 15, 17–20). A delay in the appearance of serological markers and, in some cases, reversion to nonreactive results were observed in participants who initiated ART during AHI prior to the development of a full Ab response (RV254/South East Asia Research Collaboration in HIV 010 study [the RV254/SEARCH010 study], Thai Red Cross AIDS Research Centre, Bangkok, Thailand) (18). The present study expands upon the aforementioned RV254/SEARCH010 study and assesses the ability of current HIV screening and supplemental, confirmatory assays to detect HIV biomarkers of infection in this early-treated cohort.

MATERIALS AND METHODS

Clinical specimens.

The plasma from HIV-infected participants used in this evaluation was selected from two studies in which very early HIV infection was identified by screening of individuals at high risk of HIV infection with the Hologic Aptima HIV-1 RNA qualitative assay (the Aptima assay). The evolution of serological markers in 30 untreated participants was examined by the use of samples obtained from the RV217 Early HIV Cohort study conducted in Kenya, Uganda, Tanzania, and Thailand (21, 22) and was compared to that in participants enrolled in the RV254/SEARCH010 study in Bangkok, Thailand, who initiated ART prior to the establishment of fully evolved HIV-1 Ab responses (10, 13, 18, 23). The stage of HIV infection at the time of ART initiation was assigned on the basis of the Fiebig staging system: Fiebig stage I (FI; HIV RNA positive, p24 Ag negative, HIV Ab negative; n = 23), Fiebig stage II (FII; HIV RNA positive, p24 Ag positive, HIV Ab negative; n = 39), Fiebig stage III (FIII; HIV Ab positive, Western blot [WB] assay negative; n = 13), and Fiebig stage IV (FIV; HIV Ab positive, HIV-1 WB assay indeterminate; n = 9) (18, 24).

Assays.

Serial specimens from untreated individuals and from those who initiated therapy at Fiebig stage I to IV of infection were tested at the HIV Diagnostic Reference Laboratory (HDRL) in Silver Spring, MD, by the Bio-Rad GS HIV Combo Ag/Ab Combo enzyme immunoassay (EIA) (the BRC assay), the Bio-Rad GS HIV-1/HIV-2 Plus O EIA (the 1/2/O assay), and the Bio-Rad GS HIV-1 Western blot (HIV-1 WB) and the Geenius HIV 1/2 (Geenius) supplemental assays. The Bio-Rad Genscreen HIV-1 Ag assay (the p24 Ag assay) was used to determine the concentration of serum HIV-1 p24 antigen. Testing of untreated individuals (from the RV217 cohort) was conducted at frequent intervals for up to 4 to 7 weeks after the first Aptima reactivity by the 1/2/O and BRC assays. HIV-1 WB testing for 15 of 30 individuals was conducted on serial samples from the first Aptima-reactive sample to 24 weeks of treatment. Individuals treated at Fiebig stages I to IV (the RV254/SEARCH010 cohort) were tested by the 1/2/O, BRC, WB, and Geenius assays at weeks 1, 2, 12, and 24. Testing by the Abbott Architect HIV Ag/Ab Combo (ARC) assay and assignment of the Fiebig stages of the samples from the RV254/SEARCH010 cohort were performed at the Thai Red Cross AIDS Research Centre, Bangkok, Thailand, as previously described (18). Testing by the Bio-Rad BioPlex 2200 HIV Ag-Ab Combo (BPX) assay was performed at Bio-Rad, Inc., in Benicia, CA.

Statistical analysis.

Calculations of correlations, linear regression, and paired t test parametric analysis were conducted in Prism (version 7) software (GraphPad Software, La Jolla, CA).

Regulatory approval.

The clinical specimens used in this study were collected between 2009 and 2014 under human subjects research protocols approved by the institutional review boards at the respective collecting institutions and at the Walter Reed Army Institute of Research (WRAIR). Specimens were provided to the laboratory with no accompanying private health or personal identifying information. The investigator(s) adhered to the policies regarding the protection of human subjects prescribed by the Code of Federal Regulations (CFR), title 45, vol 1, part 46 (25); CFR, title 32, chapter 1, part 219 (26); CFR, title 21, chapter 1, part 50 (protection of human subjects) (27), and Army Regulation 70-25 (28).

RESULTS

The time course of evolution of serological markers of HIV-1 infection in untreated individuals was evaluated by the use of serial samples collected from participants in the RV217 study who were identified during AHI. Figure 1A shows the results of the BRC assay for 30 individuals whose first Aptima-reactive sample was detected within 3 to 4 days of the last nonreactive result. The BRC assay yielded a reactive result, on average, at 6.5 ± 2.8 days (range, 2 to 11 days) after the first detection of RNA, rose rapidly to a saturated signal-to-cutoff (S/CO) value of the assay (S/CO, 12 to 16) by 3 to 5 weeks, and remained high thereafter. The evolution of serological reactivity from the time of first RNA detection is shown in Fig. 1B. The timing of the earliest BRC assay reactivity corresponds to that of positive results by the p24 Ag assay and represents reactivity to the HIV-1 p24 Ag component of the BRC assay. The 1/2/O assay, which detects only IgM/IgG Ab, was reactive, on average, by day 15.0 ± 4.8 (range, days 6 to 22) after the day of detection of the first Aptima RNA-reactive sample. Both the 1/2/O and BRC assays reached maximum assay S/CO values of 10 to 13 and 12 to 16, respectively, by weeks 2 to 3, and the values remained high throughout subsequent testing. HIV-1 WB assay positivity was shown by the percentage of participants with a positive result. All participants were HIV-1 WB assay positive by 4 to 5 weeks and remained positive at subsequent time points, as expected.

FIG 1.

Time course of evolution of serological markers in untreated HIV-1-infected individuals (RV217 cohort). (A) Reactivity of 30 untreated HIV-1-infected individuals by the Bio-Rad GS Ag/Ab Combo EIA (BRC assay), which measures both p24 antigen and anti-HIV antibody. (B) Average S/CO values for the BRC and 1/2/O assays, the concentration of p24 Ag measured by the Bio-Rad p24 Ag assay, and the percentage of individuals with HIV-1 Western blot (WB) assay-positive samples at each time point.

The evolution of p24 antigenicity and HIV-1 seroreactivity after ART initiation for each individual (RV254/SEARCH010 cohort) was evaluated by the p24 Ag, BPX, and ARC assays. All samples from individuals at FII to IV prior to ART initiation contained quantifiable p24 Ag by the Bio-Rad Ag assay and detectable p24 Ag by the BPX assay, but p24 Ag was nondetectable within 1 week of therapy and remained nondetectable at weeks 12 and 24 on ART (not shown). Figure 2 shows the evolution of Ag/Ab Combo results by the BPX (solid lines) or by the ARC (dotted lines) assay in cases in which insufficient sample was available to be tested by the BPX assay prior to treatment initiation. For those cases where both assays were performed on the same sample, the BPX and ARC assays yielded similar results. Samples from all 23 FI individuals were initially nonreactive by Ag/Ab Combo assays at week 0; samples from 5 (21.7%) were weakly reactive (S/CO < 10) and samples from 2 (8.8%) were strongly reactive (S/CO ≥10) by the BPX or ARC assay by week 1. Samples from 12 FI individuals (52.2%) remained nonreactive throughout the 24-week period. All FII individuals were Ag/Ab Combo reactive at ART initiation (week 0); these included 15 of 39 (38.5%) individuals with strong BPX assay S/CO values which declined after 1 week of therapy. Samples from several individuals showed a transient increase in S/CO at week 1, followed by a subsequent decline. Seroreversion from Ag/Ab Combo reactive to nonreactive was observed in 10 of 39 (25.6%) individuals who initiated therapy at Fiebig stage II and 3 of 22 (15.6%) individuals who initiated therapy at FIII/IV. By week 24, samples from 3 (7.7%) FII individuals remained nonreactive, those from 15 (38.5%) were weakly reactive, and those from 21 (53.8%) were strongly reactive. Samples from individuals who initiated ART at FIII/IV also showed an initial decline (BPX assay), with samples from 1 (4.5%) individual remaining nonreactive, those from 8 (36.4%) remaining weakly reactive, and those from 13 (59.1%) remaining strongly reactive after 24 weeks of treatment.

FIG 2.

Evolution of BioPlex (BPX) or Architect (ARC) assay reactivity of individuals initiating ART during acute HIV infection. Solid lines represent BPX assay results, and trends very similar to those seen by the BPX assay were seen by the ARC assay (dotted lines). ARC assay results are shown for individuals that did not have sufficient sample volume for BPX testing at week 0. Individuals for whom the results are depicted in green were nonreactive by the ARC or BPX assay by week 24, while those for whom the results are depicted in blue demonstrated low reactivity (S/CO < 10) and those for whom the results are depicted in red showed strong reactivity (S/CO ≥ 10). Very few samples demonstrated the very strong reactive response (S/CO > 100) seen in established infections in untreated individuals by 24 weeks of infection.

The percentage of samples from individuals grouped as nonreactive (S/CO < 1.0), weakly reactive (S/CO < 10), or strongly reactive (S/CO ≥ 10) by the BPX or ARC assay by week 12 and week 24 on ART is summarized in Table 1. After 12 weeks of therapy initiated in early infection, 69.6% of samples from individuals who initiated ART at Fiebig stage I, 12.8% of samples from individuals who initiated ART at Fiebig stage II, and 9.1% of samples from individuals who initiated ART at Fiebig stage III/IV were nonreactive. Of those samples that were reactive in the assays, 17.4% (FI), 56.4% (FII), and 40.9% (FIII/FIV) exhibited low S/CO (<10) responses in the BPX and ARC assays. By week 24, samples from 52.2%, 7.7%, and 4.5% of individuals initiating therapy at FI, FII, and FIII/IV, respectively, remained nonreactive. Samples from 10 of the 84 (11.9%) individuals initiating ART at FI to FIV in our study achieved the very strong reactive response (S/CO > 100), as is observed in individuals with untreated HIV-1 infection by week 24.

TABLE 1.

Percentage of samples from individuals with nonreactive, weakly reactive, or strongly reactive results by the BPX or ARC assay at weeks 12 and 24 after initiation of therapy at FI to FIV

Assay resulta	% of samples
	Wk 12			Wk 24
	FI	FII	FIII/IV	FI	FII	FIII/IV
NR	69.6	12.8	9.1	52.2	7.7	4.5
S/CO < 10	17.4	56.4	40.9	34.8	38.5	36.4
S/CO ≥ 10	13.0	30.8	50.0	13.0	53.8	59.1
Total	100	100	100	100	100	100

^aNR, nonreactive (S/CO < 1.0); S/CO < 10, weakly reactive; S/CO ≥ 10, strongly reactive.

The results of the BRC, ARC, and BPX Ag/Ab Combo assays with samples from RV254/SEARCH010 participants who initiated ART at FI to FIV and who were tested after 24 weeks on therapy are shown in Fig. 3. The three assays demonstrated overall good agreement at an S/CO of <10, while at higher reactivities with S/CO values up to >200, the chemiluminescent ARC assay and the fluorometric BPX assay yielded a broader dynamic range than the colorimetry-based BRC, which saturates at a maximum S/CO of 12 to 16. The ARC and BPX assay results were highly correlated to each other, with an R² value of 0.8404. The BPX assay, which, in addition to HIV-1/2 reactivity, also provides independent measures of p24 Ag and anti-HIV Ab detection for each sample, showed that reactivity, as expected, is due to the anti-HIV Ab component alone, with no p24 Ag being detected in any samples at week 12 or 24.

FIG 3.

Comparison of reactivity of the Architect (ARC), Bio-Rad (BRC), and BioPlex (BPX) Ag/Ab Combo assays in individuals who initiated ART at FI to FIV and who were tested after 24 weeks on ART.

A more detailed picture of the relative performance of the BRC, ARC, and BPX assays at 24 weeks after ART initiation stratified by those who initiated therapy at FI, FII, or FIII/IV is shown in Fig. 4. By 2 to 4 weeks of infection, samples from untreated HIV-infected individuals consistently showed S/CO values near the maximum range of the BRC assay (S/CO = 12 to 16) and 1/2/O assay (S/CO = 10 to 13) and remained high thereafter (Fig. 1). In contrast, more than half (12 of 23) of the samples from individuals initiating ART at Fiebig stage I remained nonreactive (S/CO < 1) by all three Ag/Ab Combo assays at week 24. An additional 8 samples demonstrated very low S/CO reactivity (S/CO < 10) by the ARC and BPX assays, with those from only two individuals demonstrating an S/CO of >100. Samples from all individuals treated at Fiebig stages II and III/IV were reactive by all three assays prior to treatment initiation (at week 0), but by 24 weeks on ART, samples from 3 of the 39 FII individuals and 1 of the 22 FIII/IV individuals were nonreactive by all three assays.

FIG 4.

Distribution of S/CO values at week 24 after initiation of ART at Fiebig stage I, II or III/IV by the Bio-Rad (BRC), Architect (ARC), or BioPlex (BPX) HIV Ag/Ab Combo assays. The horizontal bars show the geometric mean and standard deviation for each set.

The ability of the supplemental, confirmatory assays to confirm HIV reactivity in samples from individuals who initiated ART during AHI was evaluated using the Geenius and HIV-1 WB assays. The HIV-1 WB assay profiles of all 15 of 30 untreated individuals tested demonstrated typical HIV-1 WB assay seroconversion profiles. An example for an untreated individual is shown in Fig. 5. The reactivity to each antigen increased with time, yielding a mature, florid HIV-1 WB assay profile at weeks 11 and 21. In contrast to untreated HIV-1-infected individuals, who seroconverted to HIV-1 WB assay positivity by 7 weeks after the first Aptima RNA reactivity, individuals who initiated ART at FI to FIV showed a much slower evolution of HIV-1 WB assay antigen reactivity. The HIV-1 WB assay demonstrated only faint antigen reactivity by 24 weeks on ART, which in many cases was insufficient for interpretation as a positive result. As shown by the examples in Fig. 5, a number of individuals initiating ART during early stages of AHI were either negative or indeterminate by an HIV-1 WB assay after 12 and 24 weeks on ART. Although specific HIV-1 WB assay patterns were variable among individuals, even those who scored HIV-1 WB assay positive failed to develop the expected florid antigen reactivity typically observed in untreated individuals.

FIG 5.

(Left) Evolution of HIV-1 Western blot (WB) assay profiles of an untreated HIV-1-infected individual with time after the first Aptima-reactive RNA sample. (Right) HIV-1 WB assay profiles for selected individuals at weeks 12 and 24 following initiation of ART at Fiebig stages I to IV of HIV infection. MW, molecular weight marker.

The reactivity to the HIV p24, p31, gp41, and gp160 antigens, as detected by the Geenius and WB assays at 24 weeks after ART initiation in early infection, is shown in Fig. 6. The reactivity of all antigens was significantly inhibited when ART was initiated at Fiebig stage I. The reactivity to gp41 in samples from individuals initiating therapy at FI to FIV was comparable by the Geenius and HIV-1 WB assays, while detection of gp160 (P = 0.005) and p24 (P < 0.0001) at each Fiebig stage was significantly lower by the Geenius assay than by the HIV-1 WB assay. Reactivity to p31, a late-appearing antigen, was either low or negative by both the Geenius and HIV-1 WB assays at 24 weeks after early ART initiation.

FIG 6.

Reactivity of HIV antigens by the Geenius (GN) and HIV-1 Western blot (WB) assays at 24 weeks after initiation of therapy at FI (green), F2 (yellow), and FIII/IV (blue).

The decreased antigen reactivity translates to a high percentage of early-treated individuals remaining negative or indeterminate by the Geenius and HIV-1 WB supplemental, confirmatory tests after 12 and 24 weeks of treatment (Table 2). At 12 weeks of ART, the Geenius assay was positive for only 25% of samples from FI individuals, 44.7% of samples from FII individuals, and 59.1% of samples from FIII/IV individuals, while the samples from the remaining individuals were negative or indeterminate. Similar results were observed at week 24, with samples from 26.1%, 50.0%, and 45.5% of FI, FII, and FIII/IV individuals, respectively, being positive by the Geenius assay. HIV-1 WB assay detection sensitivity was somewhat higher, with 30.4% (FI), 73.7% (FII), and 59.1% (FIII/IV) samples yielding a positive result at week 24.

TABLE 2.

Positivity of HIV-1 Western blot and Geenius assays at 12 and 24 weeks after initiation of ART at FI, FII, or FIII/IV^a

Result	% of samples with the indicated result
	Wk 12						Wk 24
	FI		FII		FIII/IV		FI		FII		FIII/IV
	WB	GN	WB	GN	WB	GN	WB	GN	WB	GN	WB	GN
POS	25.0	25.0	65.8	44.7	72.7	59.1	30.4	26.1	73.7	50.0	59.1	45.5
IND	37.5	20.8	31.6	52.6	27.3	40.9	21.7	21.7	21.1	44.7	36.4	45.5
NEG	37.5	54.2	2.6	2.6	0.0	0.0	47.8	52.2	5.3	5.3	4.5	9.1
Total	100	100	100	100	100	100	100	100	100	100	100	100

^aA high percentage of samples tested at 12 and 24 weeks of ART remained negative or indeterminate when therapy was initiated early. WB, Western blot assay; GN, Geenius assay. POS, positive; IND, indeterminate; NEG, negative.

DISCUSSION

Previous results from the RV254/SEARCH010 study showed a marked delay in the appearance of or the absence of serological markers of HIV-1 infection in serial samples from individuals who initiated ART during AHI (18). The present study extends the evaluation of samples from this cohort to examine the impact of early ART on the reactivity of newer HIV Ag/Ab Combo assays and the supplemental, confirmatory assays used to generate laboratory evidence of HIV infection. The detection of serological markers of HIV-1 infection in samples from ART-treated individuals was contrasted to the detection of those from untreated individuals from the RV217 Early HIV Cohort study (21, 22).

As demonstrated in the RV217 cohort of individuals with early HIV infection, the first antigen marker appears very early in infection with the emergence of the p24 Ag, which is detected by Ag/Ab Combo assays. The antigen concentration measured by the Bio-Rad p24 Ag assay reaches a maximum concentration at approximately 2 weeks after the first RNA detection and then rapidly declines following the emergence of anti-HIV-1 Ab. The p24 Ag in all samples from individuals from the RV254/SEARCH010 cohort at FII to FIV was detectable prior to ART but was nondetectable within 1 week on ART and remained nondetectable thereafter. Since the p24 Ag is no longer detectable in untreated individuals after 4 to 5 weeks of infection or after 1 week of therapy, the high S/CO values observed by the Combo assays is due to the Ab component only, as exhibited by the 1/2/O assay or the Ab signal in the BPX assay.

Plasma from untreated individuals generated an S/CO at the maximum range for both the BRC and 1/2/O assays within 3 to 4 weeks of the first Aptima RNA reactivity. HIV-1 WB assay antigen profiles continued to increase in intensity with time, yielding a mature, florid HIV-1 WB assay pattern by week 24. Although testing by the BPX or ARC assay was not performed on these samples at week 24, results from other studies have shown that the intensity of the BPX and ARC assay results continues to increase over this time frame (29–31).

Previous studies examining the ability of HIV Ag/Ab Combo assays to detect AHI showed comparable performances among BRC, BPX, and ARC assays. The limit of detection of the p24 Ag by these assays ranged from 2 to 7 pg, which corresponds to 4.4 to 4.9 log copies of RNA per ml. The assays detect HIV infection from 7 to 11 days earlier than the anti-HIV IgM/IgG immunoassays (32–35). In our study, the results of the BRC, ARC, and BPX assays from samples collected at 24 weeks after the initiation of ART in AHI also showed generally good agreement at S/CO results below 10. Whereas the BRC assay optical density readout saturates quickly at an S/CO of 12 to 16, the ARC and BPX assays provided a useful measure of the impact of early treatment on the evolution of Ab reactivity. We show that the high S/CO values typically seen in the ARC (S/CO > 700) and BPX (S/CO > 200) assays by 12 to 24 weeks in untreated individuals (29–31) were greatly suppressed by early ART. Although the most dramatic inhibition of serologic reactivity at weeks 12 and 24 postinitiation of ART was observed when treatment was initiated at FI, a significant number of samples from individuals treated at FII to FIV also demonstrated lower reactivity by the ARC and BPX assays.

A decrease in Ab detection after early treatment was also observed in other studies, which showed increased frequencies of false-negative results, weakly reactive HIV rapid test results, and increased numbers of indeterminate HIV-1 WB assay results (11, 36). Of note, S/CO values of less than 10 by the ARC and BPX assays have frequently been associated with false-positive results (31, 37–39). The lower rate of Ab detection in ART-treated individuals also led to increased estimates of recent infection when tested by Ab avidity-based HIV incidence assays, such as the limiting antigen (LAg), Bio-Rad avidity, and HIV-1 multiplex assays, which rely on the semiquantitative detection of the intensity of the immune response (29, 40).

Previous studies on the performance of the Geenius assay showed that its performance was comparable to that of the HIV-1 WB assay for the detection of AHI (41, 42). In contrast, our results showed that the HIV-1 WB assay is more sensitive than the Geenius assay for the confirmation of infection in individuals treated during acute infection. Examination of the reactivity to individual antigens detected by the Geenius assay in ART-treated individuals showed that this assay is less sensitive than the HIV-1 WB assay for the detection of p24 and gp160 (Fig. 6), two antigens important for interpretation of a positive result with this assay. Our results show that antigen reactivity by both the Geenius and WB assays fails to develop after early ART treatment. A high percentage of individuals remain negative or nonreactive by these confirmatory assays at weeks 12 and 24. The decrease in HIV-1 WB assay reactivity observed at 24 weeks after the start of treatment at FII (73.7%) versus FIII/IV (59.1%) (Table 2) may be due to the small sample size. Statistical analysis showed no significant difference (P = 0.243) between the percentage of HIV-1 WB assay-positive individuals at FII and FIII/IV at week 12 or week 24. The differences in the percentages of Geenius assay-positive individuals were smaller and were also not significant.

These results are similar to those of a previous study which found that in chronic HIV infection, the p24 and p31 antigens are less frequently detected by the Geenius assay than by the HIV-1 WB or Inno-Lia assay and that the Geenius assay is less suitable for distinguishing between acute and chronic infections (43). Our finding that 73.9% of individuals who initiated therapy at FI and over 50% of those who initiated therapy at FII or FIII/IV were not confirmed as HIV-1 positive by Geenius assay suggests these individuals may be misclassified by current diagnostic algorithms, especially in individuals on ART with undetectable HIV-1 RNA.

The impact of a delayed evolution of a serological response, as manifested by weaker reactivity on screening and confirmatory assays, has important implications in evaluating the infection status of individuals who initiate ART in AHI or following suspected exposure to HIV. The high rates of false-negative results by rapid device tests for HIV-1-infected children on long-term ART can lead to misclassification, inadequate patient management, and/or the false perception of a lack of infection or cure (44). Frequent false-negative serological results were reported in HIV-1-infected infants who were breast-fed by mothers on ART or in postexposure prophylaxis (PEP) programs to prevent mother-to-child transmission (45–47). A significant reduction in HIV Ab levels, including a lack of HIV-1 WB assay reactivity and an increase in the rate of false-negative serological results following ART initiation, was previously reported in other studies (9, 48, 49). Individuals from the RV254/SEARCH010 cohort described in this study who initiated treatment at FI remained HIV-1 infected, despite the full suppression of HIV-1 plasma viremia with no detectable serological markers for over a 2-year period, as HIV RNA rebound occurred when ART was discontinued (50).

Despite the high rate of success and continued expansion of ART-based prevention and treatment programs, such as treatment as prevention, preexposure prophylaxis (PrEP), and postexposure prophylaxis (PEP) programs, HIV breakthrough infections can still occur due to nonadherence or the emergence of drug-resistant strains. The serological and nucleic acid assays used in current HIV diagnostic algorithms may not be sufficient to accurately identify true infection in settings where antiviral treatment is provided very early in HIV infection. Individuals who become infected while on PrEP have slower Ab avidity maturation, prolonged seroconversion profiles, and an increase in the time to full HIV-1 WB assay positivity (from 49 to 80 days) (51, 52). Only 27/57 (47%) of RNA-positive individuals who were infected while on PrEP were found to be reactive by the Bio-Rad GS HIV Combo Ag/Ab EIA, and none were HIV-1 positive by either the HIV-1 WB or the Geenius assay (53). Alternative approaches, such as testing for cell-associated HIV total nucleic acid (RNA and DNA), may be required to rule out HIV-1 infections in individuals with inconclusive HIV diagnostic test results and may be helpful to confirm infections in such cases (6, 54).