Maternal breastfeeding and education impact infant growth and development more than in utero HIV/ART exposure in context of universal ART: a prospective study

Abstract

Background:

Exposure to HIV and antiretroviral therapy (ART) in utero may influence infant growth and development. Most available evidence predates adoption of universal ART (Option B+ ART regimens). In a recent cohort, we compared growth and development in HIV-exposed uninfected (HEU) to HIV-unexposed (HUU) infants.

Design:

Prospective cohort study: data from Impact of Maternal HIV on Mycobacterium Tuberculosis Infection among Peripartum Women and their Infants (MiTIPS) in Western Kenya.

Methods:

Women were enrolled during pregnancy. Mother-infant pairs were followed until 24 months postpartum. We used multivariable linear mixed-effects models to compare growth rates (weight-for-age z-score [WAZ] and height-for-age z-score [HAZ]) and multivariable linear regression to compare overall development between HEU and HUU children.

Results:

About 51.8% (184/355) of the infants were HEU, 3.9% low birthweight (<2.5 kg), and 8.5% preterm (<37 gestational weeks). During pregnancy, all mothers of HEU received ART; 67.9% started ART pre-pregnancy, and 87.3% received 3TC/FTC,TDF,EFV. In longitudinal analyses, HEU children did not differ significantly from HUU in growth or development (p>0.05 for all). In the combined HEU/HUU cohort, higher maternal education was associated with significantly better growth and development: WAZ (β=0.18 [95% CI:0.01, 0.34]), HAZ (β=0.26 [95% CI:0.04, 0.48], and development (β=0.24 [95% CI:0.02, 0.46]). Breastfeeding was associated with significantly better HAZ (β=0.42 [95% CI:0.19, 0.66]) and development (β=0.31 [95% CI:0.08, 0.53]).

Conclusion:

HEU children in the setting of universal maternal ART had a similar growth trajectory and development to HUU children. Breastfeeding and maternal education improved children’s weight, height, and overall development irrespective of maternal HIV status.

Background

Expanded access to combined antiretroviral therapy (ART) and prevention of mother-to-child transmission of HIV services (PMTCT) have enabled the vast majority of pregnant women living with HIV (WLWH) to give birth to HIV-uninfected children (1,2). The population of HIV-exposed uninfected (HEU) children is growing (3). HIV exposure and antiretroviral therapy (ART) may independently affect infant growth (4–7). Evidence before universal maternal ART indicates that HEU children have a higher risk of morbidity (8,9), mortality (10–14), and neurodevelopment delay (15).

HIV exposure in utero is associated with preterm birth (<37 weeks of gestation), low birth weight (LBW, <2.5 kg) and length, small for gestational age (SGA), and stillbirth (4,5). Studies comparing HEU and HUU in the era of short course antiretrovirals for PMTCT observed that HEU children had slower growth than HUU children (16,17). More recent data suggest that even with combination ART in pregnancy, growth deficits persist in HEU (18–21). In a longitudinal study in Botswana, ART-exposed children were significantly smaller and shorter at 24 months of age than those exposed to zidovudine (ZDV) monotherapy during pregnancy (6). Together, these studies suggest that exposure to both HIV and ART in utero may affect infants’ growth. However, evidence regarding growth deficits in HEU has been conflicting – with some studies noting no differences between HEU and HUU (4,22). Similarly, while some studies showed poor language development in HEU (15), others found no differences in neurodevelopment delay between HEU and HUU (23). Most existing studies use data from before the adoption of Option B+, which involves lifelong combination ART, often starting pre-pregnancy. We used a recent longitudinal study (24,25) to compare the growth and development of HEU and HUU children and to identify predictors of growth and development in the setting of universal material ART.

Methods

Study design

We used data from Impact of Maternal HIV on Mycobacterium Tuberculosis Infection among Peripartum Women and their Infants (MiTIPS), an observational, prospective study of parallel longitudinal cohorts of pregnant women and their infants in Western Kenya from April 2016 to March 2021 (24). Details of the study have been previously summarized. Participants in the study included 200 WLWH and 200 HIV-uninfected pregnant women seeking antenatal care services at three public hospitals in Nyanza Province, Kenya, as well as their infants. Women who had TB disease in the past year or at enrollment and those who lived outside the catchment area or planned to move there were excluded from the parent study. Children born to WLWH (HEU) were compared with children born to HIV-uninfected mothers (HUU) (24,25). This analysis included 355 infants with growth measurements from at least two visits.

Infant growth characterization

Mother-infant pairs were followed up to 24 months postpartum. Infants were enrolled at six weeks of age and their birth weight and length were obtained from maternal report and review of maternal child health (MCH) booklets. Study team members were trained by CDC-Kenya mentors on growth measurement. Weight and length of children were measured at 6 weeks and 6, 12, 18, and 24 months of age. Growth was measured twice at each visit, and we used the average value rounded to the nearest 0.1 kg and 0.1 cm for weight and length, respectively.

Weight-for-age z-score [WAZ], weight-for-length z-score [WLZ], and length-for-age z-score [LAZ]) were defined using WHO child growth standards (26). Growth faltering was defined as underweight WAZ <−2, wasting WLZ <−2, and stunting LAZ <−2 (26).

The short caregiver-reported early development instruments (CREDI) were used to measure Early Childhood Development (ECD) (27). We used the 24–35 month short CREDI questionnaire. Due to COVID-19, some study visits were delayed. Data from children who attended study visits from 24–35 months were included for early childhood development analysis. We generated age and sex-specific z-score standards for development.

Maternal age, family income, educational status, marital status, body mass index (BMI), infant sex, and time-varying breastfeeding were assessed as relevant covariates.

Data analyses

Mean and standard deviation (SD) were used to describe normally distributed continuous variables, median and interquartile range (IQR) were used to describe skewed distributions, and frequency and percentage to describe categorical variables.

Multivariable linear mixed-effects models (LMEMs) with autoregression correlation structure, random intercept for subjects, and random slope for follow-up time were used to compare growth (WAZ, WHZ, HAZ) between HEU and HUU from enrollment to 24 months of age and multivariable linear regression was used to compare development of children at 24–35 months of age.

Approach to data missingness

To address missing data, multiple imputations by chained equations (MICE), including a series of regression models for each missing variable conditional upon other specified variables, were used to manage missing data. Baseline infant and maternal characteristics (infant’s age and sex, breastfeeding, maternal age, education, employment, income, BMI, marital status, and HIV infection) were used in the MICE analyses. We imputed the data 25 times. Pooled parameter estimates and their standard errors were calculated according to Rubin’s rules to account for the between- and within-imputation variance (28). We employed one single imputation model to obtain imputed values for outcomes (WAZ, HAZ, and WHZ) missed at each visit during follow-up. In the imputation models, we specified the appropriate distributions for each of the variables in the model. In the presence of missing data, the multiple imputation approach should yield unbiased estimates assuming data are missing at random (MAR).

In all analyses, we used coefficient (95% CI) and p-value <0.05 to determine the statistical significance of associations. R version 3.5.1 and STATA version 16 statistical software were used in the analyses.

Results

Maternal and infant baseline characteristics

Overall, 355 children, 184 HEU and 171 HUU were included in this analysis. The mean (SD) age of mothers was 26.0 (5.4) years, mean (SD) BMI was 24.05 (4.2), and median (IQR) monthly income in 1000 KSH was 10 (5–20), approximately 72 USD. Overall, 81.4% (289/355) of mothers were currently married, 44.2% (157/355) completed primary school or less, and 46.5% (165/355) were unemployed. Among mothers of HEU children, 15.2% (28/184) used isoniazid preventive therapy (IPT) during pregnancy, 67.9% (125/184) were on ART pre-pregnancy and 87.3% (158/181) received 3TC/FTC,TDF,EFV regimens.

Median infant (IQR) birthweight was 3.2 kg (3.0–3.6), 4.0% (14/353) had LBW, and 8.5% (26/307) were preterm. Of all infants, 51.4% (182/354) were females, and 96.1% (324/337) of infants were breastfed at enrollment. The frequency of LBW, preterm, female births, and breastfeeding did not differ significantly between HEU and HUU infants (Table 1).

Table 1:

Maternal and Infant Demographics and Clinical Characteristics

Characteristics	HEU* (n=184)	HUU¥ (n=171)	P-value
Maternal characteristics
Mean age (SD) – years	27.9 (5.4)	24.1 (4.7)	<0.001
Maternal BMI	25.3 (4.1)	24.8 (4.2)	0.250
Mother completed primary school or less	89 (48.4)	68 (39.8)	0.126
Currently married mothers – no. (%)	152 (82.6)	137 (80.1)	0.641
Unemployed mothers – no. (%)	77 (41.8)	88 (51.5)	0.088
Monthly income in KSH – no. (%)
Less than 5,000	38 (45.8)	44 (54.2)	0.041
5,000 to < 10,000	41 (50.0)	41 (50.0)
10,000 to <20,000	53 (65.4)	28 (34.6)
20,000 or above	51 (47.2)	57 (52.8)
Maternal ARV started before pregnancy – no. (%)	125 (67.9)	-	-
Current ART regimen – no. (%)
3TC/FTC,TDF,EFV	158 (87.3)	-	-
3TC/FTC,ZDV/TDF,NVP	16 (8.8)	-	-
AZT + 3TC + EFV	3 (1.7)	-	-
TDF/AZT + 3TC + LPV/r	4 (2.2)	-	-
IPT during pregnancy – no. (%)	87 (47.3)	-	-
Viral load (>40 HIV copies /ml) – no. (%)	53 (31.4)
Infant characteristics
Female infants – no. (%)	99 (53.8)	83 (48.5)	0.376
Breastfeeding at enrollment – no. (%)	168/175 (96.0)	156/162 (96.3)	0.999
Mean birth weight in kg (SD)	3.3 (0.6)	3.3 (0.5)	0.851
Low birth weight (<2.5 kg) ¶ – no./total no. (%)	11/184 (5.9)	3/169 (1.8)	0.080
Mean birth length in cm (SD)	49.3 (3.2)	49.6 (3.2)	0.595
Preterm birth (<37 gestational weeks) – no./total no. (%)	19/184 (10.3)	7/123 (5.7)	0.222

^*HEU – HIV-exposed uninfected

^¥HUU – HIV-uninfected unexposed.

^ⴃBody mass index is the weight in kilograms divided by the square of height in meters.

^¶Low birth weight is an infant born weighing 5.5 pounds (2.5 kilograms) or less.

Effect of HIV-ART exposure on growth – WAZ, HAZ, and WHZ

In univariate analyses, HEU children did not differ from HUU children in WAZ, HAZ or WHZ during 1- and 2-year follow-up. Imputation was conducted to address missing data (WAZ [3.3%], HAZ [14.5%], and WHZ [14.5%]. Adjusted for mother’s age, BMI, educational level, marital status, income and infant sex and breastfeeding, HEU children had no statistically significant differences in WAZ (β = −0.08 [95% CI: −0.25, 0.09]), HAZ (β = 0.14 [95% CI: −0.09, 0.36]), and WHZ (β = −0.18 [95% CI: −0.42, 0.06]) when compared to HUU children (Table 2). There was no interaction by sex of infants in all associations, p-value>0.05. There was no difference in the rate of the monthly change in WAZ, HAZ, and WHZ between HEU and HUU children in the first 24 months, p-value >0.05 (Figure 1).

Table 2:

Comparison of WAZ, HAZ, and WHZ over a 2-years follow-up between HEU and HUU and evaluation of other cofactors

Variable	WAZ¶			HAZ¥			WHZꬰ
	cCoefficient* (95% CI)	aCoefficient¤ (95% CI)	P-value	cCoefficient* (95% CI)	aCoefficient¤ (95% CI)	P-value	cCoefficient* (95% CI)	aCoefficient¤ (95% CI)	P-value
HEU (ref: HUU)ꞩ	−0.05 (−0.22, 0.12)	−0.08 (−0.25, 0.09)	0.370	0.10 (−0.11, 0.31)	0.14 (−0.09, 0.36)	0.248	−0.17 (−.36, 0.06)	−0.18 (−0.42, 0.06)	0.134
Age of mothers in years	0.00 (−0.01, 0.02)	−0.01 (−0.02, 0.01)	0.531	0.01 (−0.01, 0.03)	0.00 (−0.02, 0.03)	0.804	−0.01 (−0.03, 0.01)	−0.01 (−0.03, 0.01)	0.454
Mean maternal BMIⴃ	0.05 (0.03, 0.06)	0.04 (0.03, 0.06)	<0.001	0.04 (0.02, 0.07)	0.03 (0.01, 0.06)	0.008	0.03 (0.00, 0.05)	0.03 (0.00, 0.06)	0.033
Secondary school or above (ref: primary school or less)	0.21 (0.05, 0.38)	0.18 (0.01, 0.34)	0.038	0.26 (0.05, 0.48)	0.26 (0.04, 0.48)	0.022	0.12 (−0.09, 0.33)	0.04 (−0.20, 0.27)	0.763
Currently married mothers (ref: not married)	−0.13 (−0.35, 0.08)	−0.05 (−0.27, 0.17)	0.683	−0.15 (−0.42, 0.12)	−0.07 (−0.35, 0.22)	0.659	−0.04 (−0.32, 0.23)	0.03 (−0.26, 0.33)	0.827
Income in KSH
Less than 5,000	Referent	Referent		Referent	Referent		Referent	Referent
5,000 to < 10,000	0.01 (−0.22, 0.25)	−0.05 (−0.28, 0.18)	0.669	−0.08 (−0.38, 0.23)	−0.17 (−0.47, 0.13)	0.273	0.05 (−0.25, 0.35)	0.05 (−0.26, 0.35)	0.765
10,000 to <20,000	0.37 (0.13, 0.61)	0.34 (0.11, 0.58)	0.004	0.16 (−0.14, 0.47)	0.07 (−0.24, 0.38)	0.662	0.33 (0.02, 0.64)	0.35 (0.03, 0.68)	0.034
20,000 or above	0.26 (0.04, 0.49)	0.12 (−0.11, 0.35)	0.299	0.15 (−0.14, 0.43)	−0.03 (−0.33, 0.27)	0.838	0.29 (0.00, 0.58)	0.24 (−0.07, 0.56)	0.130
Female infants	0.46 (0.26, 0.67)	0.24 (0.08, 0.39)	0.004	0.46 (0.26, 0.67)	0.45 (0.24, 0.66)	<0.001	−0.01 (−0.22, 0.21)	−0.03 (−0.24, 0.18)	0.771
Breastfeeding**	0.42 (0.19, 0.65)	0.11 (−0.03, 0.25)	0.137	0.42 (0.19, 0.65)	0.42 (0.19, 0.66)	<0.001	0.04 (−0.20, 0.28)	0.00 (−0.25, 0.025)	0.995

^¶WAZ – weight-for-age.

^¥HAZ – height-for-age.

^ꬰWHZ – weight-for-height z-score.

^ꞩHEU – HIV-exposed uninfected, and HUU – HIV-uninfected unexposed.

^ⴃBody mass index is the weight in kilograms divided by the square of height in meters.

^*cCoefficient – crude (unadjusted) coefficient.

^¤aCoefficient – coefficients adjusted for other cofactors (HIV exposure status, maternal age, education, marital status, household income, infant’s sex, and breastfeeding).

^**Breastfeeding – was a time-varying factor

Figure 1.

presents scatter plots illustrating the change in observed and adjusted weight-for-age z-score (WAZ), height-for-age z-score (HAZ), and weight-for-height z-score (WHZ) over time, categorized by HIV exposure. The scatter plots at the top depict the change in observed WAZ, HAZ, and WHZ over time for each HIV exposure group, while the Lowess curves represent the average change in growth from birth to 24 months of age. The curves at the bottom display the change in adjusted WAZ, HAZ, and WHZ.

Cofactors for growth in overall cohort (combined HEU and HUU)

In univariate and multivariate analyses, maternal BMI, educational level, and infant sex were each independently associated with WAZ and HAZ. Moreover, breastfeeding was significantly associated with WAZ in only univariate analysis, and with HAZ in both univariate and multivariate analyses. In multivariate analyses, for every 1 kg/m² increase in maternal BMI, infants WAZ increased by 0.04 (β = 0.04 [95% CI: 0.03, 0.06]) z score and HAZ increased by 0.03 (β = 0.03 [95% CI: 0.01, 0.06]) z-score. Infants whose mothers attained secondary school or above had 0.18 WAZ (β = 0.18 [95% CI: 0.01, 0.34]) and 0.26 HAZ (β = 0.26 [95% CI: 0.04, 0.48]) scores higher than infants born from mothers who completed primary school or less. Female infants had 0.24 WAZ (β = 0.24 [95% CI: 0.08, 0.39]) and 0.45 HAZ (β = 0.45 [95% CI: 0.24, 0.66]) scores higher than male infants. Infants who were breastfeeding had 0.42 HAZ (β = 0.42 [95% CI: 0.19, 0.66]) higher than infants who were not breastfeeding. Infants whose household income was 10,000 to <20,000 Kenyan shilling (KSH) had 0.34 WAZ (β = 0.34 [95% CI: 0.11, 0.58]) higher than infants whose household income was less than 5,000 KSH (Table 2).

Effect of HIV-ART exposure on overall development z-score

In univariate analyses, HEU and HUU children did not differ in development z-scores. Adjusted for all pre-specified factors, HEU children had no statistically significant differences in overall development z-score compared to HUU children, (β = −0.04 [95% CI: −0.27, 0.18]) (Table 3).

Table 3:

Comparison of HEU to HUU and effect of other cofactors on overall development z-score assessed at 24–35 months

Variable	Overall development z-score¥
	cCoefficient* (95% CI)	aCoefficient¤ (95% CI)	P-value
HEU children (ref: HUU children)*	−0.08 (−0.29, 0.12)	−0.04 (−0.27, 0.18)	0.721
Age of mothers in years	0.00 (−0.02, 0.02)	0.01 (−0.27, 0.18)	0.679
Mean maternal BMIⴃ	−0.02 (−0.04, 0.00)	−0.03 (−0.05, −0.01)	0.017
Secondary school or above (ref: primary school or less)	0.25 (0.04, 0.46)	0.24 (0.02, 0.46)	0.031
Currently married mothers (ref: not married)	−0.04 (−0.30, 0.21)	−0.05 (−0.34, 0.23)	0.707
Income in KSH
Less than 5,000	Referent	Referent
5,000 to < 10,000	0.07 (−0.24, 0.37)	0.11 (−0.20, 0.41)	0.495
10,000 to <20,000	0.03 (−0.29, 0.34)	−0.03 (−0.36, 0.29)	0.844
20,000 or above	0.24 (−0.03, 0.51)	0.17 (−0.11, 0.45)	0.239
Female infants	0.03 (−0.18, 0.23)	0.05 (−0.16, 0.26)	0.633
Breastfeeding at 24 months	−0.05 (−0.51, 0.40)	0.31 (0.08, 0.53)	0.008
Birth weight (kg)	0.05 (−0.14, 0.24)	0.08 (−0.11, 0.27)	0.831

^¥Overall development measured by Caregiver Reported Early Development Instruments (CREDI).

^*HEU – HIV-exposed uninfected, and HUU – HIV-uninfected unexposed.

^ⴃBody mass index is the weight in kilograms divided by the square of height in meters.

^*cCoefficient – crude (unadjusted) coefficient.

^¤aCoefficient – coefficients adjusted for other cofactors (HIV exposure status, maternal age, education, marital status, household income, infant’s sex, and breastfeeding)

Cofactors of development in the combined cohort (HEU and HUU)

In univariate analyses, maternal was associated with overall developmental score. In multivariate analyses, maternal BMI, educational level, and breastfeeding were independently associated with overall development score. For every 1 kg/m² increase in maternal BMI, infants’ overall development decreased by 0.03 (β = −0.03 [95% CI: −0.05, 0.01]) z score. Infants whose mothers attained secondary school or above had a 0.24 higher overall developmental z-score (β = 0.24 [95% CI: 0.02, 0.46]) than infants born from mothers who completed primary school or less. Infants who were breastfeeding at 24 months had a 0.31 higher overall development z-score (β = 0.31 [95% CI: 0.08, 0.53]) than infants who were not breastfeeding (Table 3).

Growth and development in stratified analyses among HEU with mothers starting ART before versus during pregnancy and by ART regimen

Adjusted for maternal age, BMI, educational status, marital status, income, sex of children, breastfeeding, ART regimen, pregnancy IPT, and viral load, there was no association between the timing of ART initiation (pre-pregnancy vs. pregnancy initiation) and growth – WAZ (β=−0.04 [95% CI:−0.38, 0.31]), HAZ (β=−0.05 [95% CI: −0.45, 0.35]), and WHZ (β=−0.02 [95% CI:−0.38, 0.34]) – among HEU children, p-value >0.05 (Figure 2). Similarly, adjusted for the same factors, maternal ART regimen had no significant association with growth among HEU children, p-value >0.05 (supplemental Table 1). ART timing and ART regimen were not also associated with development score among HEU children (supplemental Table 2).

Figure 2.

presents the coefficients (95% confidence intervals) for all cofactors among HIV-exposed uninfected (HEU) children.

Discussion

In this longitudinal prospective cohort study, we found that HEU children had similar growth trajectories and development to HUU children under universal ART. Maternal BMI, maternal education, and breastfeeding were associated with significant impact on growth and development in the combined cohort of HEU and HUU.

We found no statistically significant difference in WAZ, HAZ, and WHZ in HEU compared to HUU children. Our findings are consistent with a study from Zambia (22) but differ from other studies in Nigeria (19), Malawi and Uganda (20), and Kenya (18) which found that HEU children had lower WAZ, HAZ, and WHZ. A study in Botswana also observed a higher risk of stunting in HEU than in HUU children (7). Most of these studies demonstrating growth deficits in HEU used data collected at or before 2014 (when Option B+ was rolled out) (18–20,22). Our study and the study in Zambia (22) that found no difference in growth between HEU and HUU children both used data after 2017. Care or treatment may have improved over time, which could be partly explained by the higher proportion of mothers (67.9%) who started ART pre-pregnancy. In the study from Botswana, 67% of stunting risk was related to LBW (7). We had a low proportion of infants with LBW (4.0%), but HEU children did have a relatively higher LBW prevalence than the HUU. Breastfeeding prevalence was high in both HEU and HUU children in our study, which may have attenuated differences between the groups. After breastfeeding adjustment, a study in Denmark (4) found that WAZ scores of HEU and HUU children were the same. In addition, it found that the difference in HAZ and WHZ between HEU and HUU disappeared after the first year when adjusted for breastfeeding. The study in Uganda and Malawi found significant growth deficits in HEU in Uganda, but not in Malawi, which may reflect differences in breastfeeding practices between HEU and HUU (20). In Uganda, formula feeding was promoted, and prevalence of breastfeeding was substantially lower among HEU than HUU (49.0% HEU Vs 86.2% HUU), in contrast to Malawi which had more similar breastfeeding prevalence between the groups (80.0% HEU vs 97.0% HUU). It is also possible that our study lacked sufficient sample size to detect small differences in growth or development between HEU and HUU. The study was able to detect a 0.3 or higher Z-score difference (assuming 1 SD) between HEU and HUU, a real difference below this threshold would have been missed. Other studies have noted differences of 0.28 or less (19).

We found no differences in development among HEU compared to HUU, consistent with studies in Botswana (23) and South Africa (29). Our study used a short form CREDI questionnaire to study the overall development of 24–35 months of age children. Both the Botswana and South Africa studies used more comprehensive assessment, with the Bayley Scales of Infant and Toddler Development III (BSID III) (23,29,30). The short CREDI questionnaire provides an overall developmental score and does not assess specific domains (27). It is possible that more sensitive assessment tools may have detected differences between HEU and HUU in our cohort. Studies in Cameroon (31) and a recent meta-analysis found that HEU children had deficits in gross motor function and poorer expressive language, but similar cognitive development, fine motor skills, and receptive language development compared to HUU (15). Among breastfed children in South Africa, HEU children were more likely to have cognitive and motor delays, but not language delays (32).

We found several sociodemographic predictors for growth and neurodevelopment in the combined cohort. An increase in maternal BMI is linked to higher WAZ and HAZ in infants, yet was associated with a decrease in their overall development. Prior studies have shown a link between increased maternal BMI and increased infant growth (33) as well as poorer brain development (34). It is unclear how maternal BMI affects the brain development of infants, but systemic inflammation and altered circulating hormones may play a role (35).

We found that infants born to mothers who had completed secondary education or higher had substantial and significantly better growth and developmental score points than infants born to mothers who had only completed primary school. Among the core socioeconomic factors (occupation, income, and education), maternal education has been shown to most strongly influence children’s cognitive development in the United States and the United Kingdom (36). Education provides women with a broad range of social and human capital, that influences healthy interaction with their children, parenting style, and knowledge of growth and development that could positively influence a child’s development (37).

We observed that breastfed infants had better growth and neurodevelopment in the combined cohort. Formula-fed infants gain weight more rapidly than breastfed ones (38), but formula feeding may lead to diarrhea and slowed growth in settings with poor sanitation (39). Our findings are aligned with a meta-analysis demonstrating better cognitive outcomes with breastfeeding (39).

Female infants had higher WAZ and HAZ z-scores but the same development score compared to male infants. This is in line with other studies (15). It is unclear why males are more likely to have growth deficits than females, which may reflect a higher risk of morbidity among male infants.

In this study, the observed inverse association between income and growth, contrary to prevailing evidence from numerous other studies indicating a positive link between higher income and improved growth and development (40). This may be due to the fact that almost all the study participants were from the same socioeconomic class, with notably low median income making meaningful impact differences difficult to distinguish. However, it is unclear why inverse associations were observed.

Our study possesses several notable strengths. Firstly, it was a prospective study that enrolled both HIV-positive and HIV-negative pregnant mothers, allowing for a comprehensive evaluation. The follow-up period of two years for both HEU and HUU children ensured a thorough assessment of outcomes over an extended duration. Weight and height/length measurements were collected twice, and their average values were used for analyses to minimize measurement errors. This approach significantly reduced the potential impact of measurement variability on our results, enhancing the accuracy of our findings. Furthermore, we employed multiple imputations to address missing data, ensuring robust data management and reducing the potential bias associated with missing information. We also conducted stratified analyses among HEU children, allowing us to examine potential differences in the growth and development of HEU children based on their mothers’ initiation of ART before versus during pregnancy and the specific ART regimen utilized. However, it is important to acknowledge the limitations of our study, notably the small sample size. Given the sample size provided, this study has the power to detect differences in z-scores of 0.3 or greater. Therefore, if the differences in WAZ, HAZ, or WHZ between HEU and HUU children were less than 0.3 z-score units, this study may not have had the statistical power to detect such small differences. Because of COVID-19 emergence, follow-up periods were delayed. Children who returned for assessments within 9 months after the scheduled last follow-up were included in the study. Although the instrument used to collect data could be used up to 35 months of age, we acknowledge that some children may have exhibited recovery, while others may have experienced regression in their development status during this nine-month period. This variability could have potentially introduced a non-differential misclassification bias, which may have attenuated our effect estimates.

Conclusion:

This longitudinal prospective cohort study found no significant differences in growth trajectory and development between HEU and HUU children. Breastfeeding and maternal education positively influenced both growth and development of children in the first 24 months of life. Maternal BMI was positively associated with growth but negatively associated with development. Our results suggest that these sociodemographic factors play a more influential role in growth and development of infants than HIV or ART exposure in the context of Option B+ implementation and high pre-pregnancy ART usage. However, larger studies and those with contemporaneous dolutegravir-based regimens and/or integrase strand inhibitor regimens will be important to understand growth and developmental impact of in utero HIV or ART exposure.