Reactivity of routine HIV antibody tests in children who initiated antiretroviral therapy in early infancy as part of the Children with HIV Early Antiretroviral Therapy (CHER) trial: a retrospective analysis

Abstract

Background: Early antiretroviral therapy (ART) and virological suppression can affect evolving antibody responses to HIV infection. We aimed to assess frequency and predictors of seronegativity in infants starting early ART.

Methods: We compared HIV antibody results between two of three treatment groups of the Children with HIV Early Antiretroviral Therapy (CHER) trial, done from July, 2005, until July, 2011, in which infants with HIV infection aged 5·7-12·0 weeks with a percentage of CD4-positive T lymphocytes of at least 25% were randomly assigned to immediate ART for 96 weeks (ART-96W) or deferred ART until clinical or immunological progression (ART-Def). We measured antibody from all available stored samples for ART-96W and ART-Def at trial week 84 using three assays: fourth-generation enzyme immunoassay HIV antigen-antibody combination, HIV-1 and HIV-2 rapid antibody test, and quantitative anti-gp120 IgG ELISA. We also assessed odds of seropositivity with respect to age of ART initiation and cumulative viral load. The CHER trial was registered with ClinicalTrials.gov, number NCT00102960.

Findings: The median age of the infants from when samples were taken (184 samples from 268 infants) was 92 weeks (IQR 90·6-93·4). More specimens from the ART-96W group were seronegative than from the ART-Def group by enzyme immunoassay (ART-96W 49 [46%] of 107 vs ART-Def eight [11%] of 75; p<0·0001) and rapid antibody test (54 [53%] of 101 vs eight [11%] of 74; p<0·0001). Median anti-gp120 IgG concentration was lower in the ART-96W group (230 μg/μL [IQR 133-13 129]) than in the ART-Def group (6870 μg/μL [1706-53 645]; p<0·0001). If ART was started between 12 and 24 weeks of age, odds of seropositivity were increased 13·7 times (95% CI 3·1-60·2; p=0·001) compared with starting it between 0 and 12 weeks. All children starting ART aged older than 24 weeks were seropositive. Cumulative viral load to week 84 correlated with anti-gp120 IgG concentrations (coefficient 0·54; p<0·0001) and increased odds of seropositivity (odds ratio 1·59 [95% CI 1·1-2·3]) adjusted for ART initiation age.

Interpretation: About half of children starting ART before 12 weeks of age were HIV seronegative by almost 2 years of age. HIV antibody tests cannot be used to reconfirm HIV diagnosis in children starting early ART. Long-term effects of seronegativity need further study. Clear guidelines are needed for retesting alongside improved diagnostic tests.

Funding: Wellcome Trust, Medical Research Council, and National Institutes of Health.

Figure 1

Overview of CHER trial treatment strategies and median duration on ART within arms. ART-Def = ART deferred until clinical/immunological progression; ART-96W = ART initiated at median 7 weeks of age for 96 weeks. From trial enrolment, median time to ART initiation for the 75 children examined in this study from the deferred arm, ART-Def, was 15 weeks (range 1-253). Median duration of ART-interruption for the 109 children from ART-96W was 74 weeks (range 0-102). Red = deferred or interrupted ART, Green = Start or re-start ART, dotted blue line = HIV-antibody assays performed at 84 weeks since randomization/enrolment of the CHER trial (approximately 92 weeks of age).

Figure 2

Box-whisker plot comparing the distribution of HIV-specific antibody (Log10 anti-gp120 IgG μg/μl) from the quantitative anti-gp120 antibody ELISA between ART-Def and ART-96W (p=0.04) at 84 weeks of the CHER trial. ART-Def = ART deferred until clinical/immunological progression; ART-96W = ART initiated at median 7 weeks of age for 96 weeks. Anti-gp120 IgG was detectable in all children.

Figure 3

A: Frequency of HIV-1 antibody seropositivity by automated serology at 84 weeks of the CHER trial (∼2 years of age) according to age of ART initation. The bar chart demonstrates the percentage of children who were seropositive at 2 years depending on whether commencing ART at 0-12, 12-24 or after 24 weeks of age (Blue = seronegative, Red = seropositive; here, a weakly reactive serology was interpreted as seropositive). B: Estimated probability of HIV seropositivity at ∼2 years of age derived from a logistic regression model fitting antibody response on age at ART initiation as linear in a logit scale. Here, the red data represents where weakly reactive serology was interpreted as seropositive, and the black data represents weakly reactive serology being seronegative; the circles represent the proportion of seropositivity in equally-sized groups for age at ART initiation; individual data-points are represented by the red or black dots at 0.0 (seronegative) or 1.0 (seropositive) on the probability y-axis. The inset table gives the odds or odds ratio and 95% confidence intervals of the two logistic regression models.

Figure 4

Relationship between quantitative anti-gp120 IgG at trial week 84 and cumulative viral load from enrolment until trial week 84. Spearman's rank correlation p<0.0001, correlation coefficient=0.54.