Since the discovery of HIV in the 1980s, there is still no cure for HIV.1 HIV research has resulted in effective antiretroviral therapy (ART) in which a combination of drugs successfully controls the virus. ART reduces HIV replication2,3 and allows for the recovery of target cells (largely CD4+ T cells).4 Once HIV infects a CD4+ T cell, it either uses the cell for active replication or inserts viral DNA into the host's DNA and forms a latent reservoir. Since the virus is quiescent in the latent reservoir, ART fails to affect these established reservoirs, and once treatment stops, latent HIV reservoirs can get reactivated leading to the production of new HIV particles.5 Latent HIV reservoirs are the main hurdle in achieving HIV cure.
Current HIV research aims to prevent the establishment of viral reservoir and to reduce its size and diversity. An HIV reservoir is established during the early stages of an HIV infection.6 To keep the reservoir small and homogenous, it seems essential to keep the viral exposure time to a minimum by an early starting of ART.7 An ART initiation soon after contracting the infection can be accomplished in perinatally HIV-infected infants who are tested and diagnosed soon after birth if the HIV status of the mother is known.8 Thus, perinatally HIV-infected infants are the ideal target group for an early-ART initiation,9 resulting in the reduction of the viral reservoir size in these infants.8,10
Although early ART typically reduces the viral load (VL), the time to viral suppression (TTS) varies in infants. In 23 reviewed infant studies, the percentage of infants achieving viral suppression ranged from 19% to 81% with a median TTS of 6 months.11 Significant predictors of a fast TTS in infants have been found to be age, VL, and CD4 percentage (CD4%) at start of ART (baseline).12 The mechanisms underlying these differences in viral suppression in infants are unknown. As a follow-up to previous epidemiological analyses,11,12 we aimed to understand how the baseline factors of age, VL, and CD4% contribute to the observed differences in TTS among infants on ART. Therefore, we investigated individual VL decay dynamics from the beginning of treatment until viral suppression was established in perinatally HIV-infected infants receiving ART within the first 6 months of life. These longitudinal measurements were collected by the European Pregnancy and Paediatric HIV Cohort Collaboration (EPPICC).12,13
MATERIAL AND METHODS
Study Inclusion Criteria
We received virological, immunological, sociodemographic, and treatment-related data of 499 infants born between 1997 and 2013 from the EPPICC.12,13 These infants were perinatally infected with HIV-1, and 469 started standard ART within 6 months of age. The standard ART regime was composed of either a boosted protease inhibitor or non-nucleoside reverse transcriptase inhibitor and 2 or 3 nucleoside reverse transcriptase inhibitors.12 To study VL dynamics shortly after the initiation of ART, we required at least one VL measurement within 90 days after the first ART initiation. This requirement was fulfilled by a subset of 312 infants (see Figure S1, Supplemental Digital Content, http://links.lww.com/QAI/B428).
Defining Baseline Values, Viral Suppression, and Decay Dynamics
For each infant, we determined age, VL, and CD4+ T-cells as a percentage of total lymphocytes (CD4%) at baseline. As the EPPICC was an observational study, baseline VL and CD4% measurements were not available for most infants. To enrich the baseline information, we extrapolated missing measurements of baseline VL and baseline CD4% from the closest measurement within 10 days before ART initiation (see Table S1, Supplemental Digital Content, http://links.lww.com/QAI/B428). We chose a short period of 10 days because the VL changes rapidly during acute infections, making more extended extrapolations unreliable.
We restricted our analysis of VL dynamics to the time points before viral suppression. Viral suppression is defined as 2 consecutive VL measurements below 400 HIV RNA copies/mL.12 We defined the “time to viral suppression” (TTS) as the time to the first measurement <400 copies/mL and considered the second measurement <400 copies/mL to be the endpoint of our investigation. We further characterized 3 VL decay patterns describing the VL dynamics from ART initiation to TTS: “immediate” VL suppression reached TTS at the first measurement after ART initiation, “monotonic” VL suppression reached TTS with strictly decreasing VL measurements over all consecutive time points, and “erratic” VL suppression reached TTS in a nonmonotonic manner, including increasing VL measurements that increased between consecutive time points. Infants showing an immediate or monotonic VL decay pattern are referred to as the “clean” subset.
Mathematical Model and Statistics
For infants with a “clean” VL decay pattern, we applied parametric fitting methods to estimate baseline VL and to mathematically determine the TTS for each infant. We used Monolix (2018R1, www.lixoft.com) to perform a nonlinear mixed-effects analysis and fitted a phenomenological model of exponential decay to the logarithmically transformed (log10) VL measurements:where V0 is the baseline log10(VL), and γ is the decay rate of log10(VL). This model, adapted from Ásbjörnsdóttir et al, 14 can capture the typical “biphasic” decline of the HIV VL during treatment.15 We applied random effects on both parameters and obtained for each infant a mean estimate and SD for each parameter. These individual parameter sets were used to calculate an expected TTS by solving Equation 1 for the time at which the threshold of log10(400 copies/mL) is reached. We opted for this phenomenological model because the data for some infants were too sparse to fit the classical mechanistic biexponential model.15
Statistical analyses were performed with R (Version 3.4.4, R Core Team 2018, https://www.r-project.org). We used linear regression and the Spearman correlation test to examine the associations between the different parameters. We quote Spearman correlation coefficients ρ with a significance level of 5%. We assessed cross-sectional time courses with locally estimated scatterplot smoothing (LOESS regressions). Differences in decay patterns were compared using two-sided Wilcoxon-tests. Pearson correlation tests were used to compare estimates derived from fits to experimental measurements. In these cases, we quote the coefficients of determination, R2, with a significance level of 5%. Finally, we performed multivariable linear models with the inclusion of intercepts, by first performing univariate analyses, and subsequently considering all univariably significant variables in a multivariable model. We then performed backward stepwise regression, selecting on the basis of the Akaike information criterion. For all cases, we quote β-coefficients and P values with a significance level of 5%.
Early-Treated Infants Differ in Their TTS
From 312 infants of the EPPICC cohorts fulfilling our inclusion criteria (see Methods), 276 infants showed viral suppression (see Figure S1, Supplemental Digital Content, http://links.lww.com/QAI/B428). These infants started standard ART at a median age of 82 days [interquartile range, IQR = (34–121)] with median baseline values of 5.3 log10(VL) [IQR = (4.2–5.9), n = 128] and 33 CD4% [IQR = (22.5–42.5), n = 99], and were virally suppressed within a median 132 days [IQR = (64–283), Fig. 1A]. Their distribution of baseline values and sociodemographic characteristics was indistinguishable from the 36 infants who showed no viral suppression in the data available (age: P = 0.105, VL: P = 0.047, CD4%: P = 0.671, open circles in Fig. 1B–D and see Table S1, Supplemental Digital Content, http://links.lww.com/QAI/B428). We excluded these 36 infants from the further analysis, as our objective was to study the VL dynamics from first treatment initiation until viral suppression. Thus, the majority of infants (88%) in the EPPICC cohorts starting early ART suppressed their VLs, but they did not suppress VLs equally rapidly. We aimed to understand these differences.
Baseline Conditions and VL Decay Pattern Lead to Differences in TTS
To understand the wide range of observed TTS, we assessed differences among the 276 viral suppressors. First, we examined correlations between observed TTS and baseline measurements. Infants with a shorter TTS started treatment at a younger age (P < 9 × 10−9, Fig. 1B), a lower baseline VL (P < 2 × 10−5, Fig. 1C), and a higher baseline CD4% (P < 7 × 10−4, Fig. 1D). All these associations are in agreement with a previous analysis by the EPIICAL consortium, using a somewhat larger selection of infants from the same cohorts.12 Thus, the age, VL, and CD4% at which infants start treatment are clearly correlated with differences in TTS.
We then studied how individual VL decay dynamics lead to viral suppression. We classified the data into 3 decay patterns (Fig. 2A): 47 infants showed immediate VL suppression after ART initiation (immediate), 141 infants had a VL declining in a strictly monotonic manner (monotonic), and 88 infants showed an “erratic” VL with irregular and intermittent increases in the VL (erratic). As expected, infants with a “clean” (immediate or monotonic) decay pattern suppressed the virus in a significantly shorter time than infants with an erratic decay pattern (P < 2.2 × 10−16, Fig. 2B). Infants with erratic decay patterns had, according to their medical records, more changes in treatment (mean = 2.89, SD = 2.26, P < 2.2 × 10−16) and accumulated treatment interruptions (mean = 0.28, SD = 0.55), suggesting treatment-related challenges such as poor adherence, drug resistance, toxicity, or irregular drug administration. Thus, long TTS in infants with erratic VL declines are at least partly due to treatment complications. Consequently, the VL dynamics of infants with erratic VL decay patterns provide no reliable information regarding the factors associated with VL suppression after first ART initiation because we cannot exclude the possibility that the observed TTS might be a response to later initiated ART. Therefore, we removed these infants from the analysis.
Erratic VL Patterns are Associated With Significantly Lower Baseline CD4%
Next, we asked whether the exclusion of infants with erratic VL decay dynamics would introduce a bias in our analysis regarding the baseline conditions. To address this issue, we compared the different decay patterns by the age and the conditions at which infants started treatment. Although the age at start of treatment (P = 0.213, Fig. 2C) and baseline VL (P = 0.465, Fig. 2D) was comparable between infants with erratic and clean VL declines, infants with an erratic decay pattern started treatment with a significantly lower baseline CD4% (P = 0.032, Fig. 2E). We found no significant differences in baseline CD4% between immediate suppressors and infants with a monotonic decay pattern (P = 0.065, Fig. 2E). Infants with an immediate suppression started treatment significantly earlier (P = 0.015, Fig. 2C) and with a significantly lower baseline VL (P < 7 × 10−5, Fig. 2D) than monotonic decliners. Thus, we acknowledge that infants with an erratic decay pattern are likely biased towards significantly lower baseline CD4% than infants with an immediate or monotonic VL suppression.
Mathematical Modeling Estimates Missing Baseline VL and Determines TTS More Precisely
To further determine with more precision which factors determine TTS during a successful initial ART, we focused on the “clean” subsets of infants with an immediate or monotonic decay pattern. To estimate additional parameters describing these VL decay dynamics, and to more precisely determine their TTS rather than defining it as the time of the first measurement of VL <400 copies/mL, we performed mathematical modeling (see Figure S2, Supplemental Digital Content, http://links.lww.com/QAI/B428). By fitting the phenomenological model (Equation 1), we inferred the baseline VL for all 188 infants with a “clean” VL decay kinetic (population estimate V0, fixed ± SD of the random effect: 4.52 ± 0.196); these estimated VLs were in good correspondence with the available measured values (R2 = 0.85, P < 2.2 × 10−16, n = 87, see Figure S3A, Supplemental Digital Content, http://links.lww.com/QAI/B428). We also estimated the slope parameter γ, which is the readout for the VL decline, showing huge individual variability (population estimate γ, fixed ± SD of the random effect: 0.01 ± 0.705). The residuals of both parameter estimates were uncorrelated (see Figure S3B, Supplemental Digital Content, http://links.lww.com/QAI/B428). From these parameters, we calculated a TTS; these were (as expected) shorter than those derived from the measurement times (R2 = 0.52, P < 2.2 × 10−16, see Figure S3C, Supplemental Digital Content, http://links.lww.com/QAI/B428), as viral suppression occurs before an undetectable measurement. The median estimated TTS for the clean subset was 54 days [IQR = (32–82), see Figure S3D, Supplemental Digital Content, http://links.lww.com/QAI/B428]. We confirmed the previously observed correlations of age (P = 4.793 × 10−12, Figure S3A), estimated baseline VL (P < 2.2 × 10−16, see Figure S4B, Supplemental Digital Content, http://links.lww.com/QAI/B428), and measured baseline CD4% (P = 0.005, n = 69, see Figure S4C, Supplemental Digital Content, http://links.lww.com/QAI/B428) with the TTS estimated by data fitting.12 In addition, TTS negatively correlated with the slope parameter (P < 2.2 × 10−16, see Figure S4D, Supplemental Digital Content, http://links.lww.com/QAI/B428). All these parameters are highly correlated with each other (see Figure S4–S6, Supplemental Digital Content, http://links.lww.com/QAI/B428). Thus, mathematical modeling provides additional information in infants with monotonic VL decay dynamics, allowing us to identify factors driving the TTS.
A Low Baseline VL and a High Baseline CD4% Lead to a Short TTS
The slope parameter γ has the strongest correlation with TTS (see Figure S5, Supplemental Digital Content, http://links.lww.com/QAI/B428) and is a phenomenological parameter describing the “quality” of suppression. We used linear regression analysis to assess which biological parameters (age, VL, or CD4%) were the best predictors of the slope parameter γ. Although a univariable model confirmed the Spearman correlations from Figure S5, Supplemental Digital Content, http://links.lww.com/QAI/B428, the results of a multivariable model highlighted a significant effect of the estimated baseline VL on the slope parameter γ (Table 1). This significance further increased in a subsequent stepwise model. Here, the factor age dropped out, whereas the estimated baseline VL and measured baseline CD4% remained as biological factors explaining the slope parameter γ. We then performed a similar linear regression analysis for the estimated TTS (Table 1). Again age, baseline VL, and baseline CD4% were significantly associated with TTS in a univariable model. In a multivariable model, baseline VL was the only remaining significant factor. This remained the case in a subsequent stepwise model, in which the CD4% was also retained. Thus, TTS depends mainly on the baseline VL and CD4%, and the effect of age might be indirect as it affects baseline VL and baseline CD4%.
Perinatal HIV Infection Progresses Rapidly in Untreated Infants
To study how age affects VL and CD4%, we considered the natural course of HIV infection before the onset of treatment. We used all pretreatment measurements of infants from the EPPICC cohorts to form a cross-sectional time-course of VL and CD4% (Fig. 3). In general, infants have very low or undetectable VLs at birth. As they grow older, their HIV infection progresses in the absence of treatment. VL measurements increase within a few weeks to reach a plateau around 6 log10(VL) after approximately 50 days (Fig. 3A). In contrast to adults, infants maintain this high setpoint VL for at least half a year.8,16 The measured CD4% then exhibits a constant decline as HIV infection progresses (Fig. 3B). We used the raw baseline CD4% measurements and did not normalize them to correct for the natural CD4% decline because the natural CD4% barely shows a decline during the first 6 months of life (see dashed band in Fig. 3B).17 The observed decline in CD4% is much steeper than the natural decline in infants (Fig. 3B). Thus, age markedly affects the VL and CD4%, as infants starting treatment late are expected to have a high VL and a low CD4%. Consequently, these infants are expected to have a long TTS (Table 1). Note that the effect of age on the VL vanishes after 6–8 weeks, when most infants have approached a plateau level, while the effect of CD4% persists (Fig. 3). Summarizing, we find that baseline VL and CD4% are the main drivers of TTS and that age operates indirectly by increasing VL and decreasing CD4%. Although at an early stage of an HIV infection age increases TTS by increasing VL and decreasing CD4%, at a later stage, the effect of the VL stabilizes and the deteriorated immune system reflected by declining CD4% determined prolonged TTS (see also Fig. 4A).
Baseline VL and Baseline CD4% Predict TTS
Finally, we checked how well the results of the stepwise linear model (Table 1), ie,
could be used to predict TTS. This model (Equation 2), including an intercept (P = 0.085), predicts the TTS estimated by the model of Equation 1 reasonably well (Fig. 4B, P < 2.2 × 10−16), using the estimated baseline VL (V0, P < 7 × 10−5) and the measured baseline CD4% (CD40, P = 0.101) as input values. We decided for the estimated VL baseline values as we had more data available, but the results with the measured VL baseline values performed equally good. The TTS could be predicted well with Equation 2, if the baseline VL was within the increasing VL phase of a natural HIV infection course (Fig. 3A) because the baseline VL is the major predictor (Table 1). Once the VL plateau is reached, the model (Equation 2) underestimates the TTS. These skewed model predictions for late times to suppression suggest that the effect of CD4% is nonlinear and might play a more dominant role in predicting TTS at a later stage of an HIV infection. This prediction, therefore, provides a lower bound for the expected TTS at the start of treatment. Equation 2 could be used as a diagnostic tool for early treated infants that are responding poorly to their treatment.
We studied “clean” VL decay dynamics of early treated perinatally HIV-infected infants from a heterogeneous database (EPPICC) to identify the major factors underlying TTS during the first treatment. We identified low baseline VL and high baseline CD4% as the major factors contributing to fast TTS. Because VL might be regarded as a trivial factor influencing TTS, the contribution of CD4% at start of ART is our most interesting and surprising result, although the association with TTS is not significant at the 5% level (Table 1). In addition, high CD4 levels are likely to be protective because the CD4% was also significantly lower in infants experiencing an erratic response. The exact mechanisms by which CD4 T-cell levels affect viral suppression remain elusive, but our results suggest that the status of the deteriorating immune system at the onset of treatment influences the TTS. Further immunological studies, considering different immune compartments, eg, levels of CD8 T cells or innate immune cells, are required to investigate and clarify the contribution of the immune system to viral suppression.
In contrast to a previous analysis of a somewhat larger subset of the same EPPICC cohort,12 we cannot confirm an independent effect of age on TTS. A major difference is that we selected for a subset of infants with “clean” VL decay dynamics to ensure that our analysis was not confounded by inefficient or ineffective treatment. In the previous studies,11,12 the TTS need not be the response to the initial treatment and could also be the response to a later initiated treatment because of subsequent changes in treatment. Note that we have not generated an age bias by excluding infants with an “erratic” VL decay dynamics (see Table S1, Supplemental Digital Content, http://links.lww.com/QAI/B428). Thus, we find that age describes the TTS in an indirect manner (via baseline VL and CD4%).
The age at start of treatment need not reflect the time period of progression of an HIV infection. With improving diagnosis and treatment of pregnant women, increasing proportions of infants are in-utero infected as opposed to intrapartum infected, and although infants usually start ART soon after birth, they might have progressed already in their HIV infection. In-utero infections might, therefore, confound the age effect on TTS, and baseline VL and baseline CD4% are more appropriate covariates in this setting. Unfortunately, we lacked any information about time of infection (in-utero or intrapartum). To determine whether a separate age contribution affects TTS, comparisons between in-utero versus intrapartum-infected infants are necessary, which will require more detailed studies addressing virological and immunological dynamics in young infants (eg, the EARTH study by EPIICAL).
We based our analyses on a simple phenomenological mathematical model describing the decay in the VL during treatment. Typically, a mechanistic model is used to capture the multiphase decay dynamic of HIV,15 but the sparseness of virological measurements did not allow proper identification of all parameters of this mechanistic model. The EPPICC database is a collection of observational studies and was not specifically designed to investigate TTS. Available virological and immunological measurements were sparse, and baseline measurements, and regular measurements during the initial VL decay, were often missing. The sparse datum is a generic problem in studies with infants. According to the WHO recommendations, only a limited volume of blood can be extracted from infants, which reduces the variety of medical tests that can be performed. To counter these issues, we used an exponential fit to the log-transformed VL data to predict missing baseline VLs and to estimate decay rates reflecting multiphase VL decay dynamics phenomenologically. In addition, we used a nonlinear mixed-effect model approach to compensate for missing data with a population mean. By combining phenomenological and mixed-effect approaches, we dealt with the sparseness of the data and nevertheless obtained individual estimates for infants in a heterogeneous data set allowing us to precisely determine individual TTS.
In conclusion, the data presented here support the fact that age at start of ART is less helpful than markers of infection such as VL and CD4% to determine TTS. Thus, an early ART initiation not only prevents disease progression but also shortens TTS. In addition to a low VL and a high CD4% at start of treatment, a “clean” VL decay pattern might predict an initially successful treatment responses resulting in fast TTS. Early ART initiation combined with a fast viral suppression may lead to the reduction of the early established latent viral reservoir by shortening the viral exposure time, which may allow for temporarily sustained VL suppression during treatment interruptions as seen in cases such as the Mississippi baby18 and the VISCONTI child.19 The opportunity to schedule treatment interruptions for perinatally HIV-infected infants has been investigated in the CHER trial.20,21 However, only a minority of early treated children achieved a prolonged viral suppression without ART,21,22 and novel therapeutic strategies are needed to obtain viral remission.23 Testing novel immunotherapeutic approaches in the early treated HIV-infected children is the challenge of the EPIICAL network, of which we are part of.9 Potential candidates for such treatment interruptions studies should be chosen on an individual basis. Our results provide a first step for such selection criteria, namely early ART initiation, low baseline VL, high baseline CD4%, and a “clean” VL decay pattern to suppression. As a next step, the effect of these criteria on sustained VL suppression has to be investigated.
The authors thank all members of Work Package 2 of the EPIICAL consortium for their open discussions and critical feedback throughout the development of this manuscript. The authors also thank the EPPICC collaborators for the allocation of their data (see Appendix, http://links.lww.com/QAI/B428 for details of all participating cohorts within EPPICC).
The EPIICAL Consortium study team: Nigel Klein, Diana Gibb, Sarah Watters, Man Chan, Laura McCoy, Abdel Babiker (University College London, United Kingdom); Anne-Genevieve Marcelin, Vincent Calvez (Université Pierre et Marie Curie, France); Maria Angeles Munoz (Servicio Madrileño de Salud-Hospital General Universitario Gregorio Marañon, Spain); Britta Wahren (Karolinska Institutet, Sweden); Caroline Foster (Imperial College Healthcare NHS Trust, London, United Kingdom); Mark Cotton (Stellenbosch University-Faculty of Medicine and Health Sciences, South Africa); Merlin Robb, Jintanat Ananworanich (The Henry M. Jackson Foundation for the Advancement of Military Medicine, Maryland); Polly Claiden (HIV i-Base, United Kingdom); Deenan Pillay (University of KwaZulu-Natal Africa Center, South Africa); Deborah Persaud (Johns Hopkins University); Rob J De Boer, Juliane Schröter, Anet J.N. Anelone (University of Utrecht, Netherlands); Thanyawee Puthanakit (Thai Red Cross AIDS-Research Centre, Thailand); Adriana Ceci, Viviana Giannuzzi (Consorzio per Valutazioni Biologiche e Farmacologiche, Italy); Kathrine Luzuriaga (University of Massachusetts Medical School, Worcester, Massachusetts); Nicolas Chomont (Centre de Recherche du Centre Hospitalier de l'Universitè de Montreal-University of Montreal, Canada); Mark Cameron (Case Western Reserve University, Cleveland, Ohio); Caterina Cancrini (Università degli Studi di Roma Tor Vergata, Italy); Andrew Yates, Louise Kuhn (Columbia University, New York); Avy Violari, Kennedy Otwombe (University of the Witwatersrand, Johannesburg [PHRU] South Africa); Ilaria Pepponi, Francesca Rocchi (Children's Hospital “Bambino Gesu,” Rome, Italy); Stefano Rinaldi (University of Miami, Miller School of Medicine, Florida); Alfredo Tagarro (Hospital 12 de Octubre, Universidad Complutense, Madrid, Spain); Maria Grazia Lain, Paula Vaz (Fundação Ariel Glaser contra o SIDA Pediátrico, Mozambique); Elisa Lopez, Tacita Nhampossa (Fundação Manhiça, Mozambique).
1. Barré-Sinoussi F, Chermann JC, Rey F, et al. Isolation of a T-Lymphotropic retrovirus from a patient at risk for acquired immune deficiency syndrome (AIDS). Science. 1983;220:868–871.
2. Ho DD, Neumann AU, Perelson AS, et al. Rapid turnover of plasma virions and CD4 lymphocytes in HIV-1 infection. Nature. 1995;373:123–126.
3. Wei X, Ghosh SK, Taylor ME, et al. Viral dynamics in human immunodeficiency virus type 1 infection. Nature. 1995;373:117–122.
4. Zhang ZQ, Schuler T, Zupancic M, et al. Sexual transmission and propagation of SIV and HIV in resting and activated CD4+ T cells. Science. 1999;286:1353–1357.
5. Finzi D, Blankson J, Siliciano JD, et al. Latent infection of CD4+ T cells provides a mechanism for lifelong persistence of HIV-1, even in patients on effective combination therapy. Nat Med. 1999;5:512–517.
6. Chun TW, Engel D, Berrey MM, et al. Early establishment of a pool of latently infected, resting CD4+ T cells during primary HIV-1 infection. Proc Natl Acad Sci. 1998;95:8869–8873.
7. Hecht FM, Wang L, Collier A, et al. A multicenter observational study of the potential benefits of initiating combination antiretroviral therapy during acute HIV infection. J Infect Dis. 2006;194:725–733.
8. Goulder PJ, Lewin SR, Leitman EM. Paediatric HIV infection: the potential for cure. Nat Rev Immunol. 2016;16:259–271.
9. Palma P, Foster C, Rojo P, et al. The EPIICAL project: an emerging global collaboration to investigate immunotherapeutic strategies in HIV-infected children. J Virus Erad. 2015;1:134–139.
10. Tagarro A, Chan M, Zangari P, et al. Early and highly suppressive antiretroviral therapy are main factors associated with low viral reservoir in European perinatally HIV-infected children. J Acquir Immune Defic Syndr. 2018;79:269–276.
11. Shiau S, Abrams EJ, Arpadi SM, et al. Early antiretroviral therapy in HIV-infected infants: can it lead to HIV remission? Lancet HIV. 2018;5:e250–e258.
12. EPIICAL. Predictors of faster virological suppression
in early treated infants with perinatal HIV
from Europe and Thailand. AIDS. 2019;33:1155–1165.
13. EPPICC. Response to early antiretroviral therapy in HIV-1 infected infants in Europe, 1996–2008. AIDS. 2011;25:2279–2287.
14. Ásbjörnsdóttir KH, Hughes JP, Wamalwa D, et al. Differences in virologic and immunologic response to antiretroviral therapy among HIV-1-infected infants and children. AIDS. 2016;30:2835–2843.
15. Perelson AS, Neumann AU, Markowitz M, et al. HIV-1 dynamics in vivo: virion clearance rate, infected cell life-span, and viral generation time. Science. 1996;271:1582–1586.
16. Shearer WT, Quinn TC, LaRussa P, et al. Viral load and disease progression in infants infected with human immunodeficiency virus type 1. Women and Infants Transmission Study Group. N Engl J Med. 1997;336:1337–1342.
17. Huenecke S, Behl M, Fadler C, et al. Age-matched lymphocyte subpopulation reference values in childhood and adolescence: application of exponential regression analysis. Eur J Haematol. 2008;80:532–539.
18. Persaud D, Gay H, Ziemniak C, et al. Absence of detectable HIV-1 viremia after treatment cessation in an infant. N Engl J Med. 2013;369:1828–1835.
19. Sáez-Cirión A, Bacchus C, Hocqueloux L, et al. Post-treatment HIV-1 controllers with a long-term virological remission after the interruption of early initiated antiretroviral therapy ANRS VISCONTI study. PLoS Pathog. 2013;9:e1003211.
20. Violari A, Cotton MF, Gibb DM, et al. Early antiretroviral therapy and mortality among HIV-infected infants. N Engl J Med. 2008;359:2233–2244.
21. Cotton MF, Violari A, Otwombe K, et al. Early limited antiretroviral therapy is superior to deferred therapy in HIV-infected South African infants: results from the CHER (children with HIV early antiRetroviral) randomized trial. Lancet. 2013;382:1555–1563.
22. Violari A, Cotton MF, Kuhn L, et al. A child with perinatal HIV
infection and long-term sustained virological control following antiretroviral treatment cessation. Nat Commun. 2019;10:1–11.
23. Palma P, Cotugno N, Rossi P, et al. The mission is remission: hope for controlling HIV replication without ART in early-treated perinatally HIV-infected children. Pediatr Infect Dis J. 2019;38:95–98.