https://orcid.org/0000-0001-5947-265X
Abstract
Antiretroviral therapy (ART) initiation during acute and early human immunodeficiency virus infection (AEHI) limits HIV reservoir formation and may facilitate post-ART control but is logistically challenging. We evaluated the performance of AEHI diagnostic criteria from a prospective study of early ART initiation.
AIDS Clinical Trials Group A 5354 enrolled adults at 30 sites in the Americas, Africa, and Asia who met any 1 of 6 criteria based on combinations of results of HIV RNA, HIV antibody, Western blot or Geenius assay, and/or the signal-to-cutoff (S/CO) ratio of the ARCHITECT HIV Ag/Ab Combo or GS HIV Combo Ag/Ab EIA. HIV status and Fiebig stage were confirmed by centralized testing.
From 2017 through 2019, 195 participants were enrolled with median age of 27 years (interquartile range, 23–39). Thirty (15.4%) were female. ART was started by 171 (87.7%) on the day of enrollment and 24 (12.3%) the next day. AEHI was confirmed in 188 (96.4%) participants after centralized testing, 4 (2.0%) participants were found to have chronic infection, and 3 (1.5%) found not to have HIV discontinued ART and were withdrawn. Retrospectively, a nonreactive or indeterminate HIV antibody on the Geenius assay combined with ARCHITECT S/CO ≥10 correctly identified 99 of 122 (81.2%) Fiebig II–IV AEHI cases with no false-positive results.
Novel AEHI criteria that incorporate ARCHITECT S/CO facilitated rapid and efficient ART initiation without waiting for an HIV RNA result. These criteria may facilitate AEHI diagnosis, staging, and immediate ART initiation in future research studies and clinical practice.
NCT02859558.
Early antiretroviral therapy (ART) is associated with decreased risk of human immunodeficiency virus (HIV)-related illnesses, other serious illnesses, and death compared with deferred ART initiation [1, 2]. Immediate ART initiation upon HIV diagnosis is now recommended by most international guidelines [3–5]. Despite this, HIV is routinely diagnosed during the chronic phase of infection, after establishment of HIV reservoirs that represent the major barrier to HIV cure [6]. In contrast, ART initiation during acute and early HIV infection (AEHI) limits the initial establishment of HIV reservoirs [7], enhances reservoir decay [8], restricts viral genetic diversification [9], may facilitate posttreatment control [10, 11], and reduces the risk of onward transmission [12]. For these reasons, individuals who initiate ART during AEHI constitute an ideal population in which to test novel strategies to achieve durable HIV suppression without ART.
Although there is a strong rationale for initiating ART during AEHI, early diagnosis and treatment are challenging and therefore mostly limited to research settings [13]. The window of opportunity is small, with rapid evolution of diagnostic markers that include detectability of HIV RNA (Fiebig I), followed by p24 antigen (Fiebig II), HIV immunoglobulin (Ig) M antibody (Fiebig III), indeterminate Western blot or Geenius HIV-1/2 antibody assay (Fiebig IV), and positive Western blot or Geenius HIV-1/2 antibody assay with negative p31 band (Fiebig V) [14]. Fiebig stages I–IV each last about 3–5 days or less [14–16].
Diagnosis is complicated by the fact that AEHI may cause no or nonspecific symptoms [17–21]. Antibody-based assays may not detect HIV during its earliest stages, and rapid diagnostic tests that detect the p24 antigen may perform poorly when used on whole blood in field conditions [22–24]. In resource-limited settings, AEHI diagnoses may be missed because testing algorithms only include these types of tests for reasons of cost and feasibility [25]. Where included in testing algorithms, confirmatory tests such as Western blots or nucleic acid testing for HIV RNA can delay diagnosis and ART initiation, particularly when testing is performed off site or in batches [15]. Delays of only a few days may allow progression through Fiebig stages and expansion of HIV reservoirs [26]. Therefore, there is a need to improve upon current diagnostic strategies for AEHI that leverage compatible clinical syndromes to identify candidates for intensive testing and generally require detectable HIV RNA as a prerequisite for diagnosis [27].
Fourth-generation immunoassays are commonly used screening tests that simultaneously detect HIV antigens and antibodies. Though results are often dichotomized as reactive or nonreactive, the assays actually report a signal-to-cutoff (S/CO) ratio that, when high, is strongly predictive of confirmatory HIV test results and can be rapidly communicated to clinicians without adding another step to diagnostic algorithms [28, 29]. The magnitude of the S/CO ratio may also be useful for distinguishing AEHI from chronic HIV infection, particularly in combination with other markers of recent infection such as an undetectable HIV antibody [30]. S/CO-based criteria from fourth-generation HIV immunoassays may therefore facilitate rapid AEHI diagnosis and ART initiation without a need for confirmatory HIV RNA or other off-site testing.
We evaluated the performance of new AEHI diagnostic criteria from an ongoing multinational prospective study of ART initiation during AEHI.
METHODS
Study Design and Participants
The AIDS Clinical Trials Group A5354 study (Early ART to Limit Infection and Establishment of Reservoir [EARLIER]) is an open-label study to evaluate the impact of early ART initiation on the establishment of HIV reservoirs and development of HIV-specific immune responses. Participants aged ≥18 years with AEHI were enrolled at 30 sites in the Americas, Africa, and Southeast Asia. Local strategies for identifying AEHI varied by site and included serial HIV testing of behaviorally vulnerable individuals; outreach to providers who administer preexposure and postexposure prophylaxis; performance of HIV testing soon after suspected exposures and in patients with symptoms consistent with acute retroviral syndrome; and use of AEHI screening algorithms with combinations of antigen/antibody (Ag/Ab) immunoassays, antibody-only immunoassays, and/or nucleic acid testing.
Participants presented with community-based testing results and had to satisfy 1 of 6 study-specific AEHI criteria prior to enrollment (Table 1). These criteria were designed to support HIV diagnosis in the early stages of infection, establish relative recency of infection, accommodate differences in testing procedures at various study sites, and estimate AEHI stage according to protocol-defined groupings. HIV RNA results had to be reported from a US Food and Drug Administration (FDA)-approved assay. Where available, S/CO ratios from the fourth-generation ARCHITECT HIV Ag/Ab Combo chemiluminescent microparticle immunoassay (CMIA) (Abbott Diagnostics, Chicago, IL) or GS HIV Combo Ag/Ab enzyme immunoassay (EIA) (Bio-Rad Laboratories, Redmond, WA) could be used to support recency of infection [30]. Both assays report a single S/CO without distinguishing HIV p24 antigen from antibody reactivity; an S/CO value ≥1 is considered reactive according to manufacturer instructions.
Diagnostic Criteria for Acute and Early Human Immunodeficiency Virus Infection Used at Clinical Research Sites for Enrollment of Participants in the A5354/EARLIER (Early Antiretroviral Therapy to Limit Infection and Establishment of Reservoir) Study
| Inclusion AEHI Criterion | Criterion to Diagnose HIV Infection | Criterion to Establish Recency | Estimated Fiebig Stage Group | |
|---|---|---|---|---|
| A | Detectable HIV RNA within 28 days | and | Nonreactive HIV antibody within 7 days | I/II |
| B | Detectable HIV RNA or reactive HIV antibody within 28 days | and | Negative/indeterminate WB or Geenius HIV-1/2 Supplemental Assay within 7 days | III/IV |
| C | Reactive HIV antibody, positive WB that is negative for p31 band, or positive Geenius HIV-1/2 Supplemental Assay that is negative for p31 band within 7 days | and | Nonreactive HIV antibody or undetectable HIV RNA within 90 days prior to study entry | V |
| D | ARCHITECT or GS HIV Combo S/CO ≥10 within 7 days | and | Nonreactive HIV antibody within 7 days | I/II |
| E | ARCHITECT or GS HIV Combo S/CO ≥ 1 within 7 days and prior S/CO <0.5 within 90 days | and | Nonreactive HIV antibody within 7 days | I/II |
| F | Detectable HIV RNA within 7 days | and | Nonreactive HIV antibody within 7 days and ARCHITECT or GS HIV Combo S/CO 0.5–9.9 within 7 days | I/II |
| Inclusion AEHI Criterion | Criterion to Diagnose HIV Infection | Criterion to Establish Recency | Estimated Fiebig Stage Group | |
|---|---|---|---|---|
| A | Detectable HIV RNA within 28 days | and | Nonreactive HIV antibody within 7 days | I/II |
| B | Detectable HIV RNA or reactive HIV antibody within 28 days | and | Negative/indeterminate WB or Geenius HIV-1/2 Supplemental Assay within 7 days | III/IV |
| C | Reactive HIV antibody, positive WB that is negative for p31 band, or positive Geenius HIV-1/2 Supplemental Assay that is negative for p31 band within 7 days | and | Nonreactive HIV antibody or undetectable HIV RNA within 90 days prior to study entry | V |
| D | ARCHITECT or GS HIV Combo S/CO ≥10 within 7 days | and | Nonreactive HIV antibody within 7 days | I/II |
| E | ARCHITECT or GS HIV Combo S/CO ≥ 1 within 7 days and prior S/CO <0.5 within 90 days | and | Nonreactive HIV antibody within 7 days | I/II |
| F | Detectable HIV RNA within 7 days | and | Nonreactive HIV antibody within 7 days and ARCHITECT or GS HIV Combo S/CO 0.5–9.9 within 7 days | I/II |
Eligibility for enrollment was assessed using available test results to satisfy 1 of 6 study-specific criteria for acute and early HIV infection, each of which was mapped to an estimated Fiebig stage group (defined by protocol as I/II, III/IV, or V). Criteria were designed to support HIV diagnosis in early stages of infection and establish relative recency of infection (including minimizing the risk of progression to Fiebig VI by the time of enrollment) while accommodating differences in testing procedures at study sites. Time periods refer to the number of days that a test could be collected prior to screening for study entry. Criteria D–F were designed to capture participants in the earliest stages of human immunodeficiency virus infection, with low-level viremia and before development of antibody responses to HIV. The S/CO was used as a surrogate for HIV RNA based on the understanding that p24 antigen detection at an S/CO of 0.5–10 (and a negative immunoglobulin M/immunoglobulin G immunoassay) corresponds roughly to an HIV RNA of 4.0–5.5 log10 copies/mL but is also the most common range for false-positive tests [30]. A recent increase of S/CO from below to within this range was considered indicative of new infection (criterion E). To minimize false-positive diagnoses, participants with this range of S/CO but no prior S/CO assessment required HIV RNA testing prior to enrollment (criterion F). The concomitant requirement for a nonreactive HIV antibody test within the preceding 7 days was designed a priori to exclude participants in Fiebig stages III–V.
Abbreviations: AEHI, acute and early HIV infection; HIV, human immunodeficiency virus; S/CO, signal-to-cutoff ratio (from ARCHITECT HIV Antigen/Antibody Combo assay or GS HIV Combo Antigen/Antibody enzyme immunoassay); WB, Western blot.
Diagnostic Criteria for Acute and Early Human Immunodeficiency Virus Infection Used at Clinical Research Sites for Enrollment of Participants in the A5354/EARLIER (Early Antiretroviral Therapy to Limit Infection and Establishment of Reservoir) Study
| Inclusion AEHI Criterion | Criterion to Diagnose HIV Infection | Criterion to Establish Recency | Estimated Fiebig Stage Group | |
|---|---|---|---|---|
| A | Detectable HIV RNA within 28 days | and | Nonreactive HIV antibody within 7 days | I/II |
| B | Detectable HIV RNA or reactive HIV antibody within 28 days | and | Negative/indeterminate WB or Geenius HIV-1/2 Supplemental Assay within 7 days | III/IV |
| C | Reactive HIV antibody, positive WB that is negative for p31 band, or positive Geenius HIV-1/2 Supplemental Assay that is negative for p31 band within 7 days | and | Nonreactive HIV antibody or undetectable HIV RNA within 90 days prior to study entry | V |
| D | ARCHITECT or GS HIV Combo S/CO ≥10 within 7 days | and | Nonreactive HIV antibody within 7 days | I/II |
| E | ARCHITECT or GS HIV Combo S/CO ≥ 1 within 7 days and prior S/CO <0.5 within 90 days | and | Nonreactive HIV antibody within 7 days | I/II |
| F | Detectable HIV RNA within 7 days | and | Nonreactive HIV antibody within 7 days and ARCHITECT or GS HIV Combo S/CO 0.5–9.9 within 7 days | I/II |
| Inclusion AEHI Criterion | Criterion to Diagnose HIV Infection | Criterion to Establish Recency | Estimated Fiebig Stage Group | |
|---|---|---|---|---|
| A | Detectable HIV RNA within 28 days | and | Nonreactive HIV antibody within 7 days | I/II |
| B | Detectable HIV RNA or reactive HIV antibody within 28 days | and | Negative/indeterminate WB or Geenius HIV-1/2 Supplemental Assay within 7 days | III/IV |
| C | Reactive HIV antibody, positive WB that is negative for p31 band, or positive Geenius HIV-1/2 Supplemental Assay that is negative for p31 band within 7 days | and | Nonreactive HIV antibody or undetectable HIV RNA within 90 days prior to study entry | V |
| D | ARCHITECT or GS HIV Combo S/CO ≥10 within 7 days | and | Nonreactive HIV antibody within 7 days | I/II |
| E | ARCHITECT or GS HIV Combo S/CO ≥ 1 within 7 days and prior S/CO <0.5 within 90 days | and | Nonreactive HIV antibody within 7 days | I/II |
| F | Detectable HIV RNA within 7 days | and | Nonreactive HIV antibody within 7 days and ARCHITECT or GS HIV Combo S/CO 0.5–9.9 within 7 days | I/II |
Eligibility for enrollment was assessed using available test results to satisfy 1 of 6 study-specific criteria for acute and early HIV infection, each of which was mapped to an estimated Fiebig stage group (defined by protocol as I/II, III/IV, or V). Criteria were designed to support HIV diagnosis in early stages of infection and establish relative recency of infection (including minimizing the risk of progression to Fiebig VI by the time of enrollment) while accommodating differences in testing procedures at study sites. Time periods refer to the number of days that a test could be collected prior to screening for study entry. Criteria D–F were designed to capture participants in the earliest stages of human immunodeficiency virus infection, with low-level viremia and before development of antibody responses to HIV. The S/CO was used as a surrogate for HIV RNA based on the understanding that p24 antigen detection at an S/CO of 0.5–10 (and a negative immunoglobulin M/immunoglobulin G immunoassay) corresponds roughly to an HIV RNA of 4.0–5.5 log10 copies/mL but is also the most common range for false-positive tests [30]. A recent increase of S/CO from below to within this range was considered indicative of new infection (criterion E). To minimize false-positive diagnoses, participants with this range of S/CO but no prior S/CO assessment required HIV RNA testing prior to enrollment (criterion F). The concomitant requirement for a nonreactive HIV antibody test within the preceding 7 days was designed a priori to exclude participants in Fiebig stages III–V.
Abbreviations: AEHI, acute and early HIV infection; HIV, human immunodeficiency virus; S/CO, signal-to-cutoff ratio (from ARCHITECT HIV Antigen/Antibody Combo assay or GS HIV Combo Antigen/Antibody enzyme immunoassay); WB, Western blot.
Participants were encouraged to initiate ART on the day of enrollment, either with study-provided single-tablet elvitegravir/cobicistat/emtricitabine/tenofovir alafenamide fumarate (EVG/COBI/FTC/TAF) or another regimen at participant or provider discretion. Study-provided single-tablet bictegravir/emtricitabine/tenofovir alafenamide fumarate became available as a second option after enrollment was completed, and ART changes were allowed at participant or provider discretion. Participants were followed longitudinally with a primary objective of evaluating the impact of early ART on cell-associated HIV DNA in CD4+ T cells after 48 weeks.
All participants provided written informed consent prior to enrollment. The study was approved by ethics committees and institutional review boards at all participating institutions.
HIV Confirmatory Testing and Staging
Samples from the day of enrollment underwent retrospective, centralized HIV confirmatory testing and Fiebig staging [14]. Centralized testing included plasma HIV RNA quantification using the Abbott m2000rt RealTime HIV-1 RNA Viral Load assay (Abbott Molecular, Des Plaines, IL) as well as the ARCHITECT HIV Ag/Ab Combo CMIA, the GS HIV Combo Ag/Ab EIA, and the third-generation IgM-sensitive Bio-Rad Genetic Systems HIV-1/HIV-2 PLUS O EIA (Bio-Rad Laboratories, Redmond, WA). The Geenius HIV-1/2 Supplemental Assay (Bio-Rad Laboratories, Redmond, WA) was used for HIV-1/2 IgG antibody discrimination, facilitating Fiebig stage assignment through detection of antibodies against gp41, gp160, p24, and p31. All testing was conducted per manufacturer instructions.
Statistical Methods
Demographic and diagnostic data from the screening and enrollment visits for all enrolled participants were used for these analyses. Data were summarized using descriptive statistics, including median and interquartile range (IQR) for continuous measures. The rank-based Spearman correlation coefficient was used to assess associations between outcomes. The exact Wilcoxon rank sum test was used for comparisons between groups, and the Jonckheere-Terpstra test was used to assess trends across Fiebig stages. The Fisher exact test was used for comparisons of 2 proportions. All analyses were performed using SAS version 9.4 (SAS Institute Inc, Cary, NC).
RESULTS
Study Population and Antiretroviral Therapy Initiation
From January 2017 through December 2019, 195 participants enrolled and completed centralized confirmatory HIV testing, including 133 (68.2%) from sites in the United States and 62 (31.8%) from other countries (Supplementary Table 1). Enrolled participants had a median age of 27 years (IQR, 23–39) and median body mass index 24.0 kg/m2 (IQR, 21.2–28.0); 30 (15.4%) were female. The interval between the first specimen used for HIV screening and study enrollment was a median of 5 days (IQR, 3–7) for US sites and 6 days (IQR, 2–12) for other countries. ART was started by 171 (87.7%) participants on the day of enrollment and 24 (12.3%) the next day, mostly with study-provided EVG/COBI/FTC/TAF (n = 151; 77.4%) or dolutegravir/lamivudine/tenofovir disoproxil fumarate (n = 37; 19.0%). Same-day ART initiation was more common at US sites than other countries (92.5% vs 78.0%, P = .005).
Centralized testing identified 3 (1.5%) participants without HIV, all of whom were enrolled based on erroneous HIV RNA determinations due to local specimen processing errors. These participants received ART for 2, 13, and 35 days before discontinuation and study withdrawal. Centralized testing identified 4 (2.0%) participants with Fiebig VI (chronic) HIV who were followed for up to 24 weeks before study withdrawal and referral for standard local care.
Among the 192 participants confirmed to be living with HIV, 136 (70.8%) were symptomatic at diagnosis. Sixty-eight (35.4%) were enrolled while hospitalized.
HIV-Related Laboratory Characteristics
Median CD4+ T-cell count at enrollment was 403 cells/mm3 (IQR, 292–563), CD8+ T-cell count 617 cells/mm3 (IQR, 322–1030), and HIV RNA 6.3 log10 copies/mL (IQR, 5.3–6.8). HIV RNA was lower among female compared with male participants (median, 5.21 [IQR, 4.34–6.49] vs 6.36 [IQR, 5.58–6.88] log10copies/mL, P = .001) and among asymptomatic compared with symptomatic participants (median, 5.39 [IQR, 4.59–6.64] vs 6.41 [IQR 5.68–6.88] log10 copies/mL, P < .001). HIV RNA results were available at a median of 8 days (IQR, 4–11) after sample collection.
Two participants with HIV RNA <40 copies/mL at enrollment were in Fiebig stages V and VI with no history of ART use or participation in a candidate HIV vaccine trial. The participants were enrolled based on a reactive third-generation EIA on the day of enrollment combined with a nonreactive Alere Determine HIV-1/2 Ag/Ab Combo assay 8 and 34 days prior to enrollment, respectively (AEHI criterion C). Results from the Abbott ARCHITECT HIV Ag/Ab Combo assay and Bio-Rad Geenius HIV-1/2 assay, run retrospectively on samples from the day of enrollment at a centralized laboratory, confirmed both HIV diagnoses despite undetectable HIV RNA.
Estimated and Actual Fiebig Stages
AEHI was confirmed in 188 (96.4%) participants with a variety of Fiebig stages (Table 2). The most commonly used criterion by sites to enroll participants was a combination of detectable HIV RNA and nonreactive HIV antibody (n = 113, 57.9%). Criteria that incorporated the S/CO ratio were used for prospective enrollment of 26 participants (13.3%), all of whom were confirmed to be in Fiebig stages I–V via retrospective centralized testing.
Actual Fiebig Stage of Participants by Inclusion Criterion for Acute and Early Human Immunodeficiency Virus Infection
| Actual Fiebig Stage | ||||||||
|---|---|---|---|---|---|---|---|---|
| Inclusion AEHI Criterion | Uninfected (n = 3) | I (n = 6) | II (n = 43) | III (n = 56) | IV (n = 23) | V (n = 60) | VI (n = 4) | Total (N = 195) |
| A (I/II) | 2 | 5 | 27 | 47 | 13 | 18 | 1 | 113 (57.9%) |
| B (III/IV) | 1 | 1 | 4 | 3 | 4 | 6 | 0 | 19 (9.7%) |
| C (V) | 0 | 0 | 0 | 0 | 2 | 32 | 3 | 37 (19.0%) |
| D (I/II) | 0 | 0 | 12 | 4 | 4 | 4 | 0 | 24 (12.3%) |
| E (I/II) | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 (0%) |
| F (I/II) | 0 | 0 | 0 | 2 | 0 | 0 | 0 | 2 (1.0%) |
| Actual Fiebig Stage | ||||||||
|---|---|---|---|---|---|---|---|---|
| Inclusion AEHI Criterion | Uninfected (n = 3) | I (n = 6) | II (n = 43) | III (n = 56) | IV (n = 23) | V (n = 60) | VI (n = 4) | Total (N = 195) |
| A (I/II) | 2 | 5 | 27 | 47 | 13 | 18 | 1 | 113 (57.9%) |
| B (III/IV) | 1 | 1 | 4 | 3 | 4 | 6 | 0 | 19 (9.7%) |
| C (V) | 0 | 0 | 0 | 0 | 2 | 32 | 3 | 37 (19.0%) |
| D (I/II) | 0 | 0 | 12 | 4 | 4 | 4 | 0 | 24 (12.3%) |
| E (I/II) | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 (0%) |
| F (I/II) | 0 | 0 | 0 | 2 | 0 | 0 | 0 | 2 (1.0%) |
Eligibility for enrollment was assessed using available test results to satisfy 1 of 6 study-specific criteria for acute and early human immunodeficiency virus (HIV) infection, each of which was mapped to an estimated protocol-defined Fiebig stage group (I/II, III/IV, or V). Actual Fiebig stage was determined using centralized testing of samples collected on the day of enrollment. See Table 1 for definitions of inclusion criteria for acute and early HIV infection. Abbreviation: AEHI, acute and early HIV infection.
Actual Fiebig Stage of Participants by Inclusion Criterion for Acute and Early Human Immunodeficiency Virus Infection
| Actual Fiebig Stage | ||||||||
|---|---|---|---|---|---|---|---|---|
| Inclusion AEHI Criterion | Uninfected (n = 3) | I (n = 6) | II (n = 43) | III (n = 56) | IV (n = 23) | V (n = 60) | VI (n = 4) | Total (N = 195) |
| A (I/II) | 2 | 5 | 27 | 47 | 13 | 18 | 1 | 113 (57.9%) |
| B (III/IV) | 1 | 1 | 4 | 3 | 4 | 6 | 0 | 19 (9.7%) |
| C (V) | 0 | 0 | 0 | 0 | 2 | 32 | 3 | 37 (19.0%) |
| D (I/II) | 0 | 0 | 12 | 4 | 4 | 4 | 0 | 24 (12.3%) |
| E (I/II) | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 (0%) |
| F (I/II) | 0 | 0 | 0 | 2 | 0 | 0 | 0 | 2 (1.0%) |
| Actual Fiebig Stage | ||||||||
|---|---|---|---|---|---|---|---|---|
| Inclusion AEHI Criterion | Uninfected (n = 3) | I (n = 6) | II (n = 43) | III (n = 56) | IV (n = 23) | V (n = 60) | VI (n = 4) | Total (N = 195) |
| A (I/II) | 2 | 5 | 27 | 47 | 13 | 18 | 1 | 113 (57.9%) |
| B (III/IV) | 1 | 1 | 4 | 3 | 4 | 6 | 0 | 19 (9.7%) |
| C (V) | 0 | 0 | 0 | 0 | 2 | 32 | 3 | 37 (19.0%) |
| D (I/II) | 0 | 0 | 12 | 4 | 4 | 4 | 0 | 24 (12.3%) |
| E (I/II) | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 (0%) |
| F (I/II) | 0 | 0 | 0 | 2 | 0 | 0 | 0 | 2 (1.0%) |
Eligibility for enrollment was assessed using available test results to satisfy 1 of 6 study-specific criteria for acute and early human immunodeficiency virus (HIV) infection, each of which was mapped to an estimated protocol-defined Fiebig stage group (I/II, III/IV, or V). Actual Fiebig stage was determined using centralized testing of samples collected on the day of enrollment. See Table 1 for definitions of inclusion criteria for acute and early HIV infection. Abbreviation: AEHI, acute and early HIV infection.
A total of 102 (53.1%) participants had a later actual Fiebig stage at enrollment than had been estimated based on the AEHI criterion used for enrollment. For these cases, a median of 5 days (IQR, 4–7) had elapsed between the initial specimen collection for HIV screening and study enrollment. When stratified by estimated Fiebig stage group, discrepancies were directly correlated with time between first HIV screening test and study enrollment (Figure 1).
Associations between timing of enrollment, initial HIV screening test, and actual Fiebig stage. Actual Fiebig stage at enrollment was plotted against the number of days elapsed since initial HIV testing for all participants (panel A) and separately for participants who enrolled with estimated Fiebig I/II (panel B), estimated Fiebig III/IV (Panel C), and estimated Fiebig V (panel D). The left and right edges of each box in the figure indicate the interquartile range. The line inside the box indicates the median. Points plotted on panel A using the log-scale were binned in order to accurately present the number of data points. Abbreviation: HIV, human immunodeficiency virus.
Architect Signal-to-Cutoff Ratio
Centralized retrospective testing showed that the ARCHITECT S/CO ratio increased overall with actual Fiebig stage, with participants in Fiebig I (n = 6) having median 0.4 (IQR, 0.4–0.6), Fiebig II (n = 43) 41.9 (IQR, 11.0–114.9), Fiebig III (n = 56) 35.7 (IQR, 17.0–108.3), Fiebig IV (n = 23) 44.0 (IQR, 18.3–116.4), Fiebig V (n = 60) 58.6 (IQR, 22.0–90.1), and Fiebig VI (n = 4) 147.2 (IQR, 63.5–346.5) (P = .033). ARCHITECT S/CO was directly correlated with enrollment HIV RNA (unadjusted Spearman correlation coefficient 0.62, P < .001; adjusted for Fiebig stage 0.71, P < .001). Retrospectively, ARCHITECT S/CO ≥10 combined with nonreactive or indeterminate HIV antibody on the Geenius assay at enrollment confirmed 99 of 122 (81.2%) Fiebig II–IV AEHI cases (Figure 2). By definition, Fiebig I was associated with S/CO <1 and Fiebig V with a reactive HIV antibody [14].
ARCHITECT S/CO ratios of participants with negative/indeterminate Geenius HIV-1/2 antibody results by Fiebig stage. ARCHITECT and Geenius HIV-1/2 antibody assays were run using samples from the day of enrollment for all participants. The bottom and top edges of each box indicate the interquartile range. The line inside the box indicates the median. Filled circles represent actual data points. *Two Fiebig II participants had an indeterminate Geenius with cross-reacting immunoglobulin G antibody to gp140/p31 and a negative third-generation HIV-1/2 result. Points plotted were binned in order to accurately present the number of data points. Abbreviation: HIV, human immunodeficiency virus; S/CO, signal-to-cutoff.
DISCUSSION
In this study, we demonstrated the feasibility of diagnosing HIV and initiating ART during AEHI across a wide variety of clinical research sites, including both resource-rich and resource-limited settings. Criteria that can be broadly applied in this way are necessary to undertake research that reflects the heterogeneity of both the virus and affected populations, including geographic diversity of HIV subtypes that may affect disease pathophysiology and outcomes [31, 32].
Importantly, the AEHI criteria in this study allowed for ART initiation before receipt of HIV RNA or other confirmatory HIV testing. Despite recent scale-up of HIV RNA testing worldwide, transportation of specimens to centralized laboratories and batch testing on automated machinery introduce delays in the process of getting HIV RNA results to healthcare providers and research participants [33–36]. Point-of-care HIV RNA testing could eliminate such delays and facilitate same-day ART initiation but is not widely available [37]. ART initiation on the day of HIV diagnosis improves engagement in care in many settings [38–42], produces rapid and durable viral suppression for most individuals [43, 44], and decreases HIV transmission [45]. In our study, enrolling participants without HIV RNA results available to confirm their HIV diagnosis facilitated rapid ART initiation. Despite potential concern for false-positive HIV diagnoses, we found that the few cases without HIV were due to erroneous results from local HIV RNA tests and not failures of the AEHI criteria. These 3 participants were exposed to days or weeks of ART with a well-established and favorable safety profile. Risks and benefits of potentially exposing individuals without HIV to investigational agents would need to be considered if these criteria are used in future studies of novel interventions during AEHI.
Participants were also enrolled before AEHI staging could be determined. The novel AEHI criteria used in this study were mapped to estimated Fiebig stages, with actual Fiebig stages only determined retrospectively via centralized testing. This strategy, paired with study-provided ART options, facilitated early identification and treatment of AEHI. Understanding that progression through Fiebig stages can occur rapidly [14, 15], it is not surprising that discrepancies between estimated and actual Fiebig stages in this study were directly associated with increased duration between initial HIV testing and study enrollment. This was particularly true among participants whose initial testing indicated the earliest stages of HIV. Delays between initial testing and enrollment were due to multiple factors such as turnaround time for initial test results, need for follow-up testing to satisfy criteria, and site-specific administrative issues. Our findings underscore the urgency to identify AEHI so that ART can be initiated as early as possible in order to achieve its potential maximal benefit.
Of the 6 AEHI criteria allowed by the study protocol, by far the most used was a combination of detectable HIV RNA by polymerase chain reaction (PCR) with a nonreactive HIV antibody. A number of prior and ongoing studies of AEHI have used this strategy successfully, sometimes leveraging pooled PCR to decrease costs [46–49]. While S/CO-based criteria were used relatively less frequently for enrollment into this study, the retrospectively evaluated combination of an ARCHITECT Ag/Ab Combo assay S/CO ≥10 with a nonreactive or indeterminate HIV antibody by the Geenius assay performed well in identifying AEHI during Fiebig stages II–IV. By definition, Fiebig I was associated with S/CO <1 and Fiebig V with detectable HIV antibody. In resource-limited settings, transportation of specimens to centralized laboratories and the need for trained technicians may create barriers to using the S/CO ratio in AEHI diagnostic algorithms [50]. These barriers mirror those observed with centralized HIV RNA testing, but Ag/Ab assays return results faster than high-throughput HIV RNA assay systems and may be useful in the setting of failed preexposure prophylaxis that can suppress viremia [51, 52]. As these testing platforms become more commonplace, S/CO may prove a useful tool for identifying AEHI in the absence of HIV RNA testing, thereby facilitating rapid ART initiation in individuals with a high pretest probability of AEHI.
Most participants in this study were symptomatic at the time of AEHI diagnosis, and approximately one-third were hospitalized. While prior studies have reported a similar prevalence of symptoms during AEHI [17–19], this may be, at least in part, an artifact of symptomatic patients being more likely to seek care and diagnostic testing. When participants are screened prospectively for AEHI with serial HIV testing, clinical manifestations tend to be less common at diagnosis [20, 21]. Symptoms were not considered in the AEHI criteria for this study, which sought to capture a wide range of testing practices at participating research sites. However, symptom-based scoring systems have been proposed to identify individuals at highest risk for AEHI in order to optimize screening yields [53, 54]; some participating sites may have considered symptoms when deciding who to test for HIV.
These analyses leveraged centralized laboratory testing from a global network of participating sites to evaluate the performance of novel criteria for AEHI diagnosis. By design, strategies for identifying candidates for HIV screening varied by site and enrollment into this study only occurred after the initial diagnosis was made. Consequently, we are unable to assess diagnostic yield of screening strategies at participating sites. Similarly, the relative frequency of each of the 6 criteria used for enrollment largely reflects site-level differences in procedures and does not necessarily reflect the relative usefulness of each criterion. It should be noted that the Abbott m2000rt platform, used for centralized viral load quantification in this study, is not FDA approved for HIV diagnostic testing. Although the S/CO ratio is an intrinsic readout from fourth-generation HIV immunoassays, use of the ratio to diagnose AEHI is not an FDA-approved approach nor is it included in manufacturer instructions. Some participating sites reported difficulty obtaining this result from commercial laboratories, but our findings suggest a potential value of routinely reporting S/CO ratio. In lower prevalence populations, S/CO-based criteria for AEHI may yield false-positive diagnoses that were not observed in this study [30]. Symptom-based scores [53, 54] or other risk-stratification methods could be considered to focus S/CO assessment on individuals with a high pretest probability of AEHI. The Geenius assay used for Fiebig staging may not correlate perfectly with the standard Western blot due to differences such as the gp160 band detected by each assay and timelines for positivity of other bands. However, there is evidence that the p31 band critical for determining progression out of Fiebig V becomes positive at the same time or earlier on the Geenius assay [55]. The Ag/Ab assays used in this study did not differentiate the reactive component per se; however, testing platforms that do differentiate Ag/Ab may have added value in evaluating AEHI and should be further evaluated for this purpose. The study was carried out at experienced clinical research centers, and there may be additional barriers to using these criteria in routine clinical care centers.
In summary, novel AEHI criteria that incorporate the ARCHITECT S/CO into diagnostic algorithms facilitated rapid ART initiation in individuals with a high pretest probability of AEHI. ART initiation during AEHI was feasible even in situations where HIV RNA testing remained pending, with minimal ART exposure observed for the few participants in whom centralized testing did not confirm HIV. Importantly, retrospectively applied S/CO-based criteria yielded no false-positive HIV diagnoses. These new criteria may facilitate AEHI diagnosis, staging, and immediate ART initiation in future research studies that will test novel interventions to achieve durable HIV suppression with products administered during AEHI or after ART initiation during AEHI [56].
Supplementary Data
Supplementary materials are available at Clinical Infectious Diseases online. Consisting of data provided by the authors to benefit the reader, the posted materials are not copyedited and are the sole responsibility of the authors, so questions or comments should be addressed to the corresponding author.
Notes
Acknowledgments. We thank the study participants who made this research possible. We gratefully acknowledge the following members of the AIDS Clinical Trials Group (ACTG) A5354 Study Team: Oladapo Alli (DAIDS Pharmacist), Deborah Anisman-Posner (laboratory specialist), LuAnn Borowski (laboratory technologist), Benjamin Chi (investigator), Susan E. Cohn (investigator), Ann Collier (investigator), Morgan Gapara (international site specialist), Sonya Heath (investigator), Tydie Higgins (data manager), Brenda Hoagland (investigator), Andrew Kaytes (community scientific subcommittee representative), Dimas Kliemann (field representative), Eugène Kroon (investigator), Susan Little (investigator), Gonasagrie Nair (investigator), Deborah Persaud (investigator), Sharon Riddler (investigator), James Rooney (industry representative), Scott Sieg (immunologist), Magdalena Sobieszczyk (investigator), Jennifer Tiu (clinical trials specialist), and Kyle Whitson (laboratory data manager). The T A5354 study was conducted at the following clinical research sites: Alabama (Sonya Heath), Barranco (Javier Lama), Brigham and Women’s Hospital Therapeutics (Paul Sax), Chapel Hill (Cynthia Gay, Brittney Soderman, Mandy Tipton), Columbia Physicians and Surgeons (Magdalena E. Sobieszczyk), Cincinnati (Carl Fichtenbaum), Greensboro (Cornelius Van Dam), Harbor-UCLA (Eric Daar), Hospital Nossa Senhora da Conceição (Breno Santos), Houston AIDS Research Team (Roberto Arduino), Instituto de Pesquisa Clinica Evandro Chagas (Sandra Cardoso), Joint Clinical Research Centre/Kampala (Cissy Kityo), Kenya Medical Research Institute/Walter Reed Project (Deborah Langat), Malawi (McNeil Ngongondo, Mina Hosseinipour), Massachusetts General Hospital (Rajesh Gandhi, Amy Sbrolla, Teri Flynn), Milton Park (Wadzanai Samaneka, Rachel Mahachi), Miriam Hospital (Karen Tashima), New Jersey Medical School Clinical Research Center (Shobha Swaminathan), Northwestern University (Babafemi Taiwo, Johnny Perez, Shannon Galvin), Ohio State University (Susan Koletar), Penn Therapeutics (Pablo Tebas), Ponce de Leon Center (Jeffrey Lennox, Albert Anderson, Ericka Patrick), Rush University (Beverly Sha), San Miguel (John Thays), Thai Red Cross AIDS Research Centre (Kiat Ruxrungtham), Trinity Health and Wellness Center (Roger Bedimo), UCLA CARE Center (Raphael Landovitz), UCSD Antiviral Research Center (Constance Benson), University of Pittsburgh, University of Washington (Janine Maenza), Washington University Therapeutics (Rachel Presti), Weill Cornell Chelsea (Timothy Wilkin), and Whitman-Walker Health (Sarah Henn).
Disclaimer. The content is solely the responsibility of the authors and does not necessarily represent the official views of the National Institutes of Health (NIH), the US Army, the US Department of Defense, or the US Department of Health and Human Services. The investigators have adhered to the policies for protection of human subjects as prescribed in AR-70.
Financial support. This work was supported by the National Institute of Allergy and Infectious Diseases of the NIH (grants UM1 AI068634, UM1 AI068636, UM1 AI068618, UM1 AI106701, and P30 AI027757). Gilead Sciences donated the antiretroviral therapy for the study.
Potential conflicts of interest. J. J. E. reports research grants from ViiV Healthcare, Gilead Sciences, and Janssen and personal fees from ViiV Healthcare, Gilead Sciences, Janssen, and Merck. J. W. M. reports research grants from Gilead Sciences; consultant fees from Merck, Accelevir Diagnostics, and Xi’an Yufan Biotechnologies; and financial interests in Abound Bio and CoCrystal Pharmaceuticals. K. R. reports receipt of the Senior Research Scholarship from the Thailand Research Fund; honoraria or consultation fees from Merck, Roche, Jensen-Cilag, Tibotec, Mylan, and the Thai Governmental Pharmaceutical Organization (GPO); and speaker fees from Abbott, Gilead Sciences, Bristol-Myers Squibb, Merck, Roche, Jensen-Cilag, GlaxoSmithKline, and Thai GPO. J. A. reports honoraria from ViiV Healthcare, Gilead Sciences, Merck, Roche, and AbbVie. E. S. D. reports research grants from Gilead Sciences and Merck and consultant fees from Gilead Sciences, Merck, and Genentech. T. C. reports an ACTG Clinical Trials Unit Award from the Division of AIDS, NIH. All remaining authors: No reported conflicts of interest. All authors have submitted the ICMJE Form for Disclosure of Potential Conflicts of Interest. Conflicts that the editors consider relevant to the content of the manuscript have been disclosed.
Presented in part: 27th Conference on Retroviruses and Opportunistic Infections, Virtual, 8–11 March 2020, Abstract 974.
References
.
