Effect of dolutegravir-based antiretroviral therapy transition on body mass index in adolescents living with HIV: secondary analysis of data from the ANRS 12225 Pediacam III cohort in Cameroon Authors: Thierry Serge Joël Noumsi, Arsène Brunelle Sandie, Francis Ateba Ndongo, Suzie Tetang Ndiang, Georgette Guemkam, Verlaine Bolyse Mbouchong, Mireille Fossi, Ida Calixte Penda, Jules Brice Tchatchueng Mbougua, Albert Faye, Josiane Warszawski, Mathurin Cyrille Tejiokem Published: 2026 ABSTRACT AbstractBackgroundDolutegravir (DTG) is widely used in people living with HIV, but its potential association with weight gain remains insufficiently documented in adolescents. This study assessed the effect of transitioning to a DTG-based regimen on body mass index (BMI) in adolescents living with HIV (ALHIV) in Cameroon.MethodsData prospectively collected between 2020 and 2024 from the ANRS 12,225-Pediacam III cohort were retrospectively analyzed. The analysis included 105 ALHIV who switched to DTG (HI-DTG), 18 ALHIV who did not (HI-nDTG), and 180 HIV-uninfected adolescents born to HIV-positive mothers (HEU, n = 79) or to HIV-negative mothers (HUU, n = 101). Body mass index (BMI) trajectories and levels were compared across groups using linear mixed-effects models, as well as before and after the DTG transition within the HI-DTG group.ResultsThe median age was 11.5 years (IQR:11.0–12.5), which significant differences between groups. Overall, 51.6% were female. The median follow-up on DTG was 2.6 years (IQR:2.0–3.2). BMI trajectories differed across groups (p = 0.045) after adjustment for sex, age, baseline BMI z-score, household income, and interaction terms. Compared with HI-nDTG as the reference group, the HI-DTG group showed a modest but significant greater annual increase in BMI (0.32 kg/m²/year; 95% CI: 0.03–0.62; p = 0.03), with no differences observed for HEU or HUU. Baseline BMI z-score, household income, and female sex were more strongly associated with BMI over time than DTG use. Within the HI-DTG group, a mean annual increase of 0.19 kg/m²/year was observed after switching to DTG.ConclusionSwitching to DTG was associated with an increase in BMI among ALHIV. These findings underscore the need for close metabolic monitoring to prevent long-term cardiometabolic complications.Keywords: HIV, Dolutegravir, Weight gain, Body mass index, Adolescents Introduction In 2019, the World Health Organization (WHO) recommended the inclusion of dolutegravir (DTG) in first- and second-line combination antiretroviral therapy (cART) regimens [1]. DTG offers several advantages over older antiretrovirals, including high efficacy, good tolerability, and a high genetic barrier to resistance [2, 3]. Due to its demonstrated superiority over Efavirenz-based regimen, and in alignment with international guidelines and emerging evidence, Cameroon, along with 82 other low- and middle-income countries rapidly adopted DTG in its national treatment guidelines in 2019 [4, 5]. However, DTG has been associated with some adverse effects, most notably neuropsychiatric complaints, which were among the first reported [6]. Moreover, several studies have suggested that DTG, like other antiretrovirals such as darunavir, bictegravir, and tenofovir alafenamide, may be associated with significant weight gain [7, 8], particularly among adults in sub-Saharan African countries, including Ethiopia [9], Cameroon, as shown in the NAMSAL randomized controlled trial [10, 11], Kenya, Nigeria, Tanzania, and Uganda [12]. Among children and adolescents living with HIV, evidence of DTG-associated weight gain is limited and inconsistent [13]. A cohort study in Eswatini found significant weight gain after DTG initiation among adolescents [14], whereas a study conducted in Italy reported no association between DTG use and weight gain [15]. In Cameroon and other sub-Saharan African countries, data on this association among adolescents are scarce. Weight gain associated with DTG use in adolescents living with HIV may have distinct psychological and clinical consequences, underscoring the need for targeted monitoring strategies in this population. In this study, we assessed the effect of transitioning to a DTG-based regimen on body mass index (BMI) among ALHIV, followed from the first week of life, in a prospective cohort in Cameroon. Methods Ethics statement The ANRS 12,225-PEDIACAM III cohort study was approved by Ethics Committees in Cameroon (N°038/CNE/DNM/07 in 2007, with renewal in 2010, 2017, and 2021) and France (Pasteur Institute, N°XN/LT/CT/2007.021). Administrative authorization was provided by the Cameroon Ministry of Public health. Parental or guardian written consent was obtained before enrolment, and assent was collected from children aged ≥ 11 years, HIV-infected children assenting only after disclosure. Data sources The data analyzed in this study originate from the ANRS 12,225-PEDIACAM III cohort, an ongoing prospective study involving adolescents living with perinatal HIV infection (ALHIV) and uninfected controls. Recruitment was conducted between November 2007 and October 2011 in three referral hospitals in Cameroon (Mother and Child Center of the Chantal Biya Foundation and Essos Hospital Center in Yaoundé, and Laquintinie Hospital in Douala) under the coordination of the Center Pasteur du Cameroun. Follow-up data collected from 2020 to 2024 was used in this analysis. A detailed description of the PEDIACAM study design, objectives, and data collection procedures of the Pediacam study has been published elsewhere [16, 17]. Study participants Four groups were defined: ALHIV who switched to DTG (HI-DTG), ALHIV not on DTG (HI-nDTG), uninfected adolescents born to mothers living with HIV (HIV-exposed uninfected [HEU]), and HIV-negative mothers (HIV-uninfected [HUU]). Participants with more than six months follow-up and documented weight and height data were included. ALHIV who briefly switched to DTG but subsequently reverted to their previous regimen were excluded. Figure 1 presents the flow of participants from the PEDIACAM cohort included in this study. Flowchart of the patients included in the study according to study group, ANRS 12,225 Pediacam Cohort, Cameroon Transitioning to dolutegravir-based regimen Therapeutic management adhered to national recommendations [4, 5]. Transitions to alternative cART regimen occurred in response to guideline updates, drug stock-outs, and treatment failure. Initially, DTG-based regimens were prescribed according to body weight. During the study period, adolescents eligible for DTG received weight-adjusted combinations: ABC/3TC (600/300) + DTG (50 mg) for those weighing between 25 and 34.9 kg and TDF/3TC/DTG (TLD) for those weighing ≥ 35 kg. These weight-based restrictions were later relaxed. Main outcome and covariables The main outcome was body mass index (BMI) calculated from weight (kg) and height (cm) at each six-month visit using WHO growth standards [18]. Although BMI z-scores support age- and sex-standardized comparisons at a single time point, they are suboptimal for longitudinal analyses because they rely on static reference populations and may misrepresent within-person weight changes [19]. For this reason, absolute BMI values were used to assess individual weight trajectories and isolate the effect of switching to DTG-based cART. The main covariate was the adolescent group defined by HIV status and transition to DTG. Additional covariates included follow-up site, sex, age, parental cohabitation, BMI for age at baseline, viral load, CD4 levels, family income, and household access to running water. Statistical analysis Participants’ characteristics were compared across the four groups using chi-square, Fisher’s exact, or the Kruskal-Wallis tests as appropriate. Association between DTG exposure and BMI trajectories were assessed using linear mixed-effects models with subject-specific random intercepts to account for intra-individual correlation. Random slopes were not included due to the small sample size of the HI-nDTG subgroup (n = 18), the heterogeneous number of repeated measurements (range: 2–10 per participant), and the risk of model over-parameterization. Clinically relevant interactions, such as time × sex (to assess potential sex-specific effects), time × group (the main interaction of interest to determine whether BMI slopes differed by DTG exposure), and time × sex × group (to explore potential sex-related differences in DTG-associated weight changes reported in previous studies [20, 21]), were tested. Covariates with p < 0.20 in univariate analyses were entered into the multivariable model and retained through backward elimination. The linearity of BMI change was confirmed graphically, and model diagnostics supported adequate fit. A within-subject before-and-after analysis was conducted among HI-DTG adolescents using data from three years before and after DTG initiation., incorporating a time × period interaction to compare slope. All analyses were conducted in R version 4.4.2 (lme4, lmerTest, and simr). Results Characteristics of study population A total of 335 adolescents completed follow-up visits between September 2020 and December 2024: 131 ALHIV on cART, 90 HEU, and 114 HUU. After exclusions, the final analytic sample included 105 HI-DTG, 18 HI-nDTG, 79 HEU, and 101 HUU participants. The median age was 11.5 years (IQR: 11.0-12.5), with significant differences across groups. Additional group differences were observed for household water supply, income, and parental cohabitation (Table 1). The sample was 51.6% female. Participants contributed an average of 6.5 repeated measurements, with a median follow-up duration of 2.6 years (HI-DTG), 3.9 years (HI-nDTG), 3.6 years (HEU), and 3.7 years (HUU). Table 1Characteristics of adolescents at baseline according to study group, ANRS 12,225 Pediacam Cohort, CameroonAdolescents characteristics at baselineHIV-exposed uninfected (HEU)N = 79n (%)HIV-unexposedUninfected (HUU)N = 101n (%)Did not switch to TLD (HI-nDTG)N = 18n (%)Switched to TLD(HI-DTG)N = 105n (%)p-valueSite0.003 CME43 (54.4)54 (53.5)9 (50.0)52 (49.5) HLD12 (15.2)8 (7.9)8 (44.4)19 (18.1 CHE24 (30.4)39 (38.6)1 (5.6)34 (32.4)Sex0.306 Male37 (46.8)56 (55.4)7 (38.9)46 (43.8) Female42 (53.2)45 (44.6)11 (61.1)59 (56.2)Age at baseline (yr)< 0·001 Median [Q1, Q3]11·1 [11·0, 11.6]11·2 [11·0, 11.7]11.5 [11·0, 12.9]12·5 [12·0, 13·5] Min - Max10·5–13.910.3–13.210.5–13.010.0–16.4Weight (wt)0·130 Median [Q1, Q3]39.0 [33·0, 45·0]41.0 [35·0, 50·0]37·0 [32.0,47.2]38.0 [34.0, 46.0] Min – Max27·0–95·027·0–78·013·0–54·025·0–78·0BMI z-score0.004 <- 2 (Thinness)1 (1.3)0 (0.0)0 (0.0)6 (5.7) [-2, 1] (Normal)57 (72.2)66 (65.3)13 (76.5)84 (80.0) >1 (overweight/obese)21 (26.6)35 (34.7)4 (23.5)15 (14.3)Viral load (copies/mL) ≤ 1 00011 (64.7)87 (87.9) > 1 0006 (35.3)12 (12.1)%CD4 ≥254 (30.8)62 (70.8) < 259 (69.2)26 (29.2)Parent cohabitation< 0.001 Yes36 (45.6)65 (64.4)7 (38.9)17 (16.2) No43 (54.4)36 (35.6)11 (61.1)88 (83.8)Income (thousands XAF) (ref=”<100”)< 0.001 ≤ 10039 (49.4)16 (15.8)11 (61.1)52 (49.5) 100–15018 (22.8)36 (35.6)5 (27.8)22 (21.0) > 15022 (27.8)49 (48.5)2 (11.1)31 (29.5)Running water in house0·034 No20 (25.3)16 (15.8)2 (11.1)33 (31.4) Yes59 (74.7)85 (84.2)16 (88.9)72 (68.6) Characteristics of adolescents at baseline according to study group, ANRS 12,225 Pediacam Cohort, Cameroon Among ALHIV who transitioned to DTG (HI-DTG), the median weight at switch was 38.0 kg, with nearly one-quarter weighing < 35 kg. At baseline, 12.1% had unsuppressed viral load (VL > 1000 copies/mL) and 29% had CD4 levels < 25%; at the time of evaluation, these proportions were 14.1% and 22.2%, respectively. Overweight prevalence at baseline (BMI > + 1 SD) was: HEU, 26.6%; HUU, 34.7%; HI-nDTG, 23.5%; and HI-DTG, 14.3%. By the end of follow-up, overweight declined in HEU (− 2.5%), HUU (-10.9%), and HI-nDTG (-6.8%), but increased in HI-DTG (+ 3.8%). Influence of transition to DTG-based regimen on BMI trajectories In the fully adjusted model (including baseline BMI z-score, sex, family income, and interactions terms, Table 2), the estimated annual BMI increase, corresponding to the effect of time in the reference category (male, HI-nDTG) was + 0.40 kg/m²/year (95% CI: 0.14–0.65; p = 0.003). The overall group effect was not significant (χ² = 3.19; p = 0.36), but the time x group interaction was significant (χ² = 8.04; p = 0.045). Adolescents on DTG exhibited a modest but significant BMI increase than the reference group (+ 0.32 kg/m²/year; 95% CI: 0.03–0.62; p = 0.031), whereas HEU and HUU did not differ. Table 2Initial and adjusted models of BMI change based on certain characteristics, ANRS 12,225 Pediacam Cohort, CameroonInitial model and fully adjusted models of BMI change over time(a)VariableInitial modelFully adjusted modelEstimate (CI 95%)p valueEstimate (CI 95%)p-valueA – baseline characteristics (effects on BMI level) Sex (female vs. male)1.18 (0.42, 1.92)0.0020.66 (0.11, 1.21)0.020 Group (Ref: HI-nDTG)0.2190.362 HI-DTG0.13 (−1.55, 1.81)0.62 (−0.59, 1.83) HEU0.47 (−1.25, 2.19)0.18 (−1.06, 1.41) HUU1.05 (−0.64, 2.74)0.04 (−1.20, 1.28)BMI for age zscore (ref: [-2, 1] (Normal))< 0001 < -2 SD (Thinness)−3.64 (−5.38, −1.90)*** >+1 SD (overweight/obese)4.89 (4.26, 5.52)***Family income (thousands XAF, ref: <100)0.025 100–1500.38 (−0.30, 1.06) > 1500.91 (0.26, 1.57)**B – interaction with time (effects on BMI slope) Time0.40 (0.14, 0.66)0.0020.40 (0.14, 0.65)***0.002 Time x Sex0.43 (0.10, 0.75)0.0100.43 (0.12, 0.75)**0.007 Time x Group0.0590.045Time : HI-DTG0.31 (0.02, 0.61)0.32 (0.03, 0.62)* Time : HEU0.09 (−0.20, 0.37)0.09 (−0.19, 0.37) Time : HUU0.15 (−0.13,0.42)0.15 (−0.12, 0.43)Time x Sex x Group0.0110.006 Time : HI-DTG : female−0.10 (−0.47, 0.27)−0.12 (−0.48, 0.25) Time : HEU : female0.08 (−0.27, 0.44)0.08 (−0.26, 0.44) Time : HUU : female−0.27 (−0.62, 0.07)−0.28 (−0.63, 0.06)*: 0.05 < p < 0.01; **: 0.01 < p < 0.001; ***:p < 0,001Ref: reference; (a): an initial model includes time, group, sex, and their interactions based on a priori hypotheses, and the fully adjusted model additionally accounts for family income and BMI for age zscore at baseline. Time represents the annual change in BMI in the reference categories of variables included in interaction with time (i.e., male adolescents in the HI-nDTG group). Initial and adjusted models of BMI change based on certain characteristics, ANRS 12,225 Pediacam Cohort, Cameroon *: 0.05 < p < 0.01; **: 0.01 < p < 0.001; ***:p < 0,001 Ref: reference; (a): an initial model includes time, group, sex, and their interactions based on a priori hypotheses, and the fully adjusted model additionally accounts for family income and BMI for age zscore at baseline. Time represents the annual change in BMI in the reference categories of variables included in interaction with time (i.e., male adolescents in the HI-nDTG group). Girls had higher baseline BMI levels (β = +0.66; p = 0.020) and steeper trajectories (+ 0.43 kg/m²/year; p = 0.008), with pairwise contrasts showing steeper slopes than boys across all exposure groups. However, the effect of sex on BMI trajectories was not modified by DTG treatment (time × DTG × sex interaction, p = 0.53). Baseline BMI z-score strongly predicted BMI levels during follow-up: compared to adolescents with a normal baseline BMI z-score (between − 2 and + 1), those with a baseline BMI z-score > + 1 maintained significantly higher BMI throughout follow-up (β = +4.89 kg/m²; 95% CI: 4.26–5.52; p < 0.001), while those with a baseline BMI z-score < − 2 had lower BMI throughout follow-up (β = −3.64 kg/m²; 95% CI: −5.40 to − 1.88; p < 0.001). Family income was also associated with BMI (χ² = 7.35; p = 0.025), with adolescents from higher-income households (> 150,000 FCFA/month) presenting higher BMI levels (β = +0.91; p = 0.007). In the pre–post DTG analysis (Fig. 2; Table 3), the predicted BMI from the model was already rising before the switch, with an adjusted annual increase of 0.45 kg/m²/year (p < 0.001). After the transition to DTG, the predicted trajectory became significantly steeper, as indicated by an additional + 0.19 kg/m²/year in the post-DTG slope (p = 0.007). In addition, a baseline BMI z-score > + 1 was strongly associated with a consistently higher mean BMI level throughout follow-up (β = +2.72, p < 0.001). Changes in predicted BMI trajectories before and after switching to a DTG-based regimen in the HI-DTG group, with confidence intervals highlighting an increased post-switch slope: ANRS 12,225 Pediacam Cohort, Cameroon Table 3Initial and adjusted models of BMI change based on certain characteristics, ANRS 12,225 Pediacam Cohort, CameroonInitial model and fully adjusted models of BMI change over time(a)VariableInitial modelFully adjusted modelEstimate (CI 95%)p valueEstimate (CI 95%)p-valueA – Baseline characteristics (effects on BMI level) Sex (female vs. male)1.36 (0.41, 2.31)**0.0050.78 (0.08, 1.49)*0.032BMI for age z score (ref: [-2, 1] (Normal))< 0001 < -2 SD (Thinness)-2.92 (-4.13, -1.71)*** >+1 SD (overweight/obese)2.72 (1.76, 3.68)***Family income (thousands XAF, ref: <100)0.163 100–1500.44 (-0.30, 1.19) > 1500.65 (-0.04, 1.33)B – Interaction with time (effects on BMI slope) Time0.45 (0.30, 0.60)***< 0.0010.45 (0.30, 0.59)***< 0.001 Time x Period (DTG effect)0.19 (0.05, 0.33)**0.0080.19 (0.05, 0.33)**0.008 Time x sex0.43 (0.26, 0.60)***< 0.0010.43 (0.26, 0.60)***< 0.001*: 0.05 < p < 0.01; **: 0.01 < p < 0.001; ***:p < 0,001Ref: reference, (a): An initial model includes time, group, sex, and their interactions based on a priori hypotheses. The fully adjusted model additionally accounts for socioeconomic status and baseline.Time represents the annual change in BMI in the reference categories of variables included in interaction with time (i.e., male adolescents before switching to DTG). Initial and adjusted models of BMI change based on certain characteristics, ANRS 12,225 Pediacam Cohort, Cameroon *: 0.05 < p < 0.01; **: 0.01 < p < 0.001; ***:p < 0,001 Ref: reference, (a): An initial model includes time, group, sex, and their interactions based on a priori hypotheses. The fully adjusted model additionally accounts for socioeconomic status and baseline. Time represents the annual change in BMI in the reference categories of variables included in interaction with time (i.e., male adolescents before switching to DTG). Discussion Key results and interpretation This study is among the few to assess the effect of DTG-based antiretroviral regimens on BMI trajectories in adolescents living with perinatal HIV (ALHIV) in sub-Saharan Africa and is the first conducted in Cameroon. Overall, our results showed that DTG was associated with a modest but statistically significant increase in BMI. However, this effect reached statistical significance only after adjustment for time, age, sex, baseline BMI z-score, and baseline family income in linear mixed-effects models. Notably, these covariates, were more strongly associated with BMI trajectories than DTG use itself. Consistent with these findings, data from this cohort study showed a statistically significant increase in BMI after the switch to DTG-based therapy, particularly among adolescents who were overweight at baseline. This pattern was consistent across two analytical approaches: (i) a between-group comparison including ALHIV on non-DTG regimens (HI-nDTG) and HIV-uninfected groups (HEU and HUU) and (ii) a within-subject pre–post transition analysis. In the between-group comparison, adolescents in the HI-DTG group had a significantly greater annual BMI gain (+ 0.32 kg/m² per year) than those in the HI-nDTG group, while no significant differences were observed between the HI-nDTG and either the HEU or HUU groups. In the within-group analysis, the rate of BMI increase accelerated after switching to DTG-based therapy, with an additional gain of 0.19 kg/m² per year compared to the pre-transition period. Our findings are consistent with those of Thivalapill et al., who reported a mean BMI increase of 1.2 kg/m² per year after transition to DTG-based therapy in adolescents in Eswatini [14]. In contrast, studies from Italy and Sweden did not observe significant weight gain with DTG [15, 20], likely reflecting contextual and racial differences. Our cohort consisted exclusively of Black adolescents, whereas European studies were predominantly White, an important distinction given evidence that Black individuals may be more susceptible to INSTI-associated weight gain [7, 22, 23]. Although DTG is frequently implicated, weight gain is not unique to this agent. Increases have also been documented with darunavir (DRV) [24], as well as with raltegravir (RAL), tenofovir alafenamide (TAF), and bictegravir (BIC) [25, 26]. Our findings indicate an increase in BMI after transition to the DTG-based regimen, but this should be interpreted within the broader metabolic effects of various antiretroviral classes. Notably, agents such as DRV, RAL, TAF, and BIC are not yet routinely used in our cohort (Fig. 1), reinforcing the likelihood that the observed changes are related to dolutegravir. Given that uncontrolled HIV replication can impair growth in children and adolescents [27], some BMI variation may also reflect virological status. In our cohort, 35% of ALHIV in the non-DTG group and 12% in the DTG group had unsuppressed viral load at baseline. Viral load was included as a covariate, with no evidence of confounding in the association between DTG use and BMI change. Increases in BMI among HIV-infected adolescents on DTG may reflect either catch-up growth or a increased risk of obesity and cardiometabolic complications [28]. In our cohort, a higher proportion of underweight adolescents was observed at baseline in the HI-DTG group (5.7% vs. 0-1.3% in other groups, Table 1), suggesting possible catch-up growth. However, all analyses were adjusted for baseline BMI z-score, and the association between DTG exposure and BMI increase remained significant. This indicates that the observed weight gain is unlikely to be explained solely by catch-up growth and may reflect a DTG-related effect. Nonetheless, residual confounding cannot be excluded, and longer-term studies are needed to confirm these findings. ALHIV on DTG-based regimens may therefore be vulnerable to excess weight gain and related comorbidities, although adolescent-specific evidence, particularly in sub-Saharan Africa, remains limited. Adult studies in the region, including those by Shamu et al. [29], Lartey et al. [30], and Brennan et al. [31], have shown increases in blood pressure after DTG transition, and a meta-analysis by Jemal et al. supports links between DTG use, weight gain, and cardiometabolic outcomes [32]. These adult findings raise concern about long-term risks as ALHIV enter adulthood. Psychosocial effects also merit attention, as visible DTG-associated weight changes may worsen stigma, affect body image and self-esteem, and hinder social integration [33]. Strengths and limitations This study has notable strengths and some methodological limitations. A key strength is its dual analytical approach, combining within-subject pre–post comparisons with a between-group analysis, which enhanced internal validity. Multivariable mixed-effects models incorporating key clinical and sociodemographic variables likely mitigated residual confounding. Limitations include the relatively short follow-up after DTG initiation and the small HI-nDTG subgroup (n = 18), which reduced statistical precision and limited more complex random-effects modeling. In addition, pubertal stage was not accounted for, despite being a potential confounder. Adolescence is characterized by rapid physiological changes during puberty that substantially influence BMI trajectories, with sex-specific patterns [34]. These limitations are unlikely to affect the main conclusions. Overall, our findings indicate that, although DTG contributes modestly but significantly to weight gain, classic determinants of BMI in adolescence, particularly sex, baseline nutritional status, and socioeconomic level remain predominant. However, in most African countries, including Cameroon, national HIV strategies do not currently include specific guidance for monitoring weight gain in ALHIV receiving DTG-based regimens [5]. Public health authorities should integrate routine monitoring of weight and metabolic parameters into adolescent HIV care, with particular attention to those already overweight, who face high cardiometabolic risk. Targeted measures, such as nutritional counselling, promotion of physical activity, and education on metabolic risk factors, are needed to mitigate potential adverse effects in this vulnerable group. Acknowledgements The authors thank all adolescents and their parents or guardians for participating in the ANRS-12225 PEDIACAM cohort study. We are grateful to the healthcare staff at the study sites (MCC-CBF, EHC, LH, and CPC) for their invaluable contributions to patient care and data collection. We also wish to thank the ANRS EID authorities and the teams of the Clinical Research and Vigilance Departments for their dedicated commitment. Author contributions TSJN: Contributed to the statistical analysis of the data, to the interpretation of the results, and to the writing of the manuscript. ABS: participated in the design of the study protocol, provided essential expertise for the interpretation of the results, gave final approval of the version to be published, and agrees to be accountable for all aspects of the work in ensuring that questions related to the accuracy or integrity of any part of the work are duly investigated and resolved. MF, ICP, FAN, STN, and GG: participated in the study design , in the recruitment and follow-up of infants, in the writing of the manuscript, contributed to improving the clarity and scientific rigor of the manuscript, and gave their approval. VBM: participated in the interpretation of the data, contributed to improving the clarity and scientific rigor of the manuscript, gave final approval, and agrees to be accountable for all aspects of the work. JBT: participated in the study design, in the analysis of the data, in the writing of the manuscript, and in its approval. AF: participated in the study design, in the validation of the study’s conclusions, and in its approval. JW: participated in the study design, in the validation of the study’s conclusions, and in its approval. MCT: participated in the study design, the interpretation of the results, in the writing of the manuscript, and in the coordination of the study. Funding The ANRS 12225 - PEDIACAM III cohort study was funded by the French Agency for Research on AIDS and Viral Hepatitis Emerging Infectious Diseases (ANRS MIE), an agency of the Institut National de la Santé et de la Recherche Médicale (Inserm). Data availability The datasets generated and/or analyzed during the current study are available from the corresponding author on reasonable request. Declarations Competing interests The authors declare no competing interests. Footnotes Publisher’s Note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. References Associated Data Data Availability Statement The datasets generated and/or analyzed during the current study are available from the corresponding author on reasonable request.