Outcomes of HIV-exposed but Uninfected Children in South Africa over 5 Years

Expansion of services to prevent mother-to-child transmission of HIV substantially reduced infections among children.^[1] However, high maternal HIV prevalence in many countries, such as South Africa (30%), contributes to growing numbers of children exposed to HIV but uninfected (CHEU).^[2] Understanding the health status of CHEU is a critical research priority.

Previous studies found CHEU have higher risks for mortality and morbidities compared to unexposed children (CHUU).^[3] A recent study in AIDS found infant mortality among CHEU was higher than the national rate and growth rates over 18 months were poor.^[4] Given that maternal HIV likely has broad impacts on children’s well-being, we conducted a study with an extended examination of growth, and cognitive and behavioural functioning of CHEU over five years, compared to CHUU in the same communities.

We conducted a retrospective cohort study using data from the Philani+ trial in Cape Town, South Africa. Philani+ methods have been previously described.^[5] Between 2009 and 2010, recruiters enrolled pregnant women from 24 neighbourhoods, which were matched and randomized to either an intervention or control group. The intervention group received home visits by paraprofessional community health workers. Health outcomes were collected during pregnancy, at two weeks, and six, 18, 36, and 60 months post-birth. Findings stratified by intervention have been previously reported.^[6] There were no significant intervention effects based on maternal HIV status on children’s outcomes, thus we did not include intervention status in these analyses.

Children’s height and weight were measured at each assessment. Growth measures were converted to z-scores.^[7] Child Behavior was assessed at 36 months using the Strengths and Difficulties Questionnaire^[8] and Child Behavior Checklist (CBCL).^[9] Cognitive functioning was assessed at 18 months using the Bayley Scale of Infant Development,^{[10, 11]} 36 and 60 months with the Peabody Picture Vocabulary Test,^[12] and 60 months through the Kaufman Assessment Battery for Children.^[13] We documented maternal ART use, CHEU NVP receipt, breastfeeding status, child hospitalizations, respiratory disorders, and diarrhoea.

We examined child outcomes by the CHEU and CHUU groups using t-tests and chi-square tests. We used mixed-effects linear regression models with restricted maximum likelihood estimation for continuous outcomes, and mixed-effects logistic and mixed-effects negative binomial regression models for discrete and count outcomes. Models included fixed effects for mother’s HIV status, time point, and an interaction between children’s HIV exposure and time point; and adjustments controlling for children’s age and any breastfeeding.

Table 1 summarizes the differences between CHUE and CHUU children. Our sample included 289 CHEU and 596 CHUU. Among MLH, 78.8% were ever on ART, and among CHUE, 90.9% received NVP, and 95.9% received a PCR HIV test at 6 months. Over five years, 22 mothers and 104 infants died, and mortality did not differ by maternal HIV status. CHEU had similar odds of being admitted to the hospital and acquiring respiratory illnesses or pneumonia compared to CHUU. Diarrhea was more common among CHEU at 6 months, but was more common among CHUU at 5 years. Immunizations were similar.

Table1.

Child outcomes by Children HIV exposed, but uninfected (CHEU) (N=289) and children who are HIV unexposed and uninfected (CHUU) (N=596) up to 5 years

	HEU		HUU		Estimated Odds Ratio HEU vs. HUU
	n	%	n	%	OR1	95% CI
Growth measures
Low birth weight infant (< 2500 grams)3	34	11.8	53	8.9	0.06a	−0.01 – 0.13
Stunted (HAZ < −2 SD)	20	6.9	41	6.9	0.81a,c	0.31 – 2.10
Malnourished (WAZ < −2 SD)	7	2.4	13	2.2	2.40a,c	0.35 – 16.28
Hospitalizations
Admitted to hospital	32	11.1	83	14.1	0.71a,c	0.39 – 1.31
					Estimated Mean Difference HEU vs. HUU
	Mean	SD	Mean	SD	Difference2	95% CI
Growth measures
Height-for-age Z-score (HAZ)	−0.56	0.93	−0.51	0.98	−0.04a,b	−0.23 – 0.15
Weight-for-age Z-score	−0.11	1.00	0.02	0.97	−0.14a,b	−0.32 – 0.04
Behavior Outcomes
Strength and Difficulties Questionnaire5	9.58	4.67	9.63	4.65	0.23a	−0.87 – 1.33
Child Behavior Checklist
Aggressive Behavior Subscale Score	10.51	8.19	10.77	8.11	−0.64a,c	−2.16 – 0.89
Cognitive Outcomes
Bayley Scale of Infant Development5
Cognitive Composite Score	100.64	14.58	99.65	13.84	−1.12a	−6.46 – 4.23
Motor Composite Score	103.7	13.53	103.4	12.11	1.46a	−3.35 – 6.26
Peabody Picture Vocabulary Test	60.30	16.38	62.63	17.92	−2.47a,c	−6.30 – 1.36
Kaufman Assessment
Global Scale Index (MPI) Standard Score**	81.65	10.96	84.12	11.34	−2.12a	−4.70 – 0.46
Hospitalizations
Number of times admitted to hospital	0.15	0.49	0.24	0.74	−0.11a,c	−0.24 – 0.03

^*p ≤ 0.05

^**p ≤ 0.01 (t-tests or χ² tests);

⁺p < 0.05 (regression estimates) adjusting for any breastfeeding, up to 18 months

¹Mixed-effects logistic regression for binary outcomes

²Mixed-effects linear regression for continuous outcomes; Mixed-effects negative binomial regression for count outcomes

³Measured at post-birth only, N=1001;

⁴Measured at 18 Months only, N=503

⁵Measured 3-years only, N=896

^aRandom-intercept for neighborhood and child (neighborhood random-intercept only for LBW, Immunizations, Bayley, and Kaufman assessment; child random-intercept only for WAZ < −2 SD)

^bRandom-slope for time using unstructured covariance structure;

^cRandom-slope for time using identity covariance structure

CHEU had lower weight-for-age z-scores (WAZ) than CHUU at post-birth and 18-months, but differences were not significant at other time points. Height-for-age z-score(HAZ) differences were observed at 6 and 18 months, but not at 3 or 5 years. In adjusted analyses, CHEU and CHUU were similar in terms of HAZ, WAZ, low birth weight (<2500mg), stunting (HAZ < −2 SD), and malnourishment (WAZ < −2 SD). There were no significant differences on the Child Behaviour Checklist, Strengths and Difficulties Questionnaire, Bayley Scale of Infant Development, Peabody Picture Vocabulary Test, or Kaufman Assessment.

We found that CHEU were similar to CHUU on health, growth, and cognitive and behavioural functioning over the first five years of life in Cape Town, South Africa. While CHEU experienced lower weight and height at post-birth and 18-month, they were similar to CHUU at 3 and 5-years.

Our findings are consistent with other longer longitudinal studies indicating that differences in growth present in the first year of life and then decrease over time.^{[14, 15]} Further, a South African study found breastfed CHEU had transient increases in morbidity risk between 8 days and 3 months, which were no longer present after 3 months; and infection-related hospitalizations in CHEU with mothers on ART were similar to CHUU.^[16]

Despite the challenges faced by mothers living with HIV, health and developmental trajectories of CHEU were similar to CHUU after accounting for ART receipt and breastfeeding. While these data suggest time-limited impact of in-utero HIV exposure, questions remain in terms of impacts in later developmental periods, such as adolescence.