Neurodevelopmental Outcomes of HIV/Anti-retroviral Drug Perinatally Exposed Uninfected Children Aged 3–6 Years

Abstract

Objective:

Given the roll out of maternal antiretroviral therapy (ART) for prevention-of-perinatal-HIV-transmission, increasing numbers of children are perinatally HIV/ART exposed but uninfected (CAHEU). Some studies suggest CAHEU may be at increased risk for neurodevelopmental (ND) deficits. We aimed to assess ND performance among preschool CAHEU.

Design:

This cross-sectional study assessed neurodevelopmental (ND) outcomes among 3–6 year-old CAHEU at entry into a multi-country cohort study.

Methods:

We used the Mullen Scales of Early Learning (MSEL) and Kaufman Assessment Battery for Children (KABC-II) to assess ND status among 3–6 year-old CAHEU at entry into the PROMISE Ongoing Treatment Evaluation (PROMOTE) study conducted in Uganda, Malawi, Zimbabwe and South Africa. Statistical analyses (Stata 16.1) was used to generate group means for ND composite scores and subscale scores, compared to standardized test score means. We used multivariable analysis to adjust for known developmental risk factors including maternal clinical/socioeconomic variables, child sex, growth-for-age measurements, and country.

Results:

1647 children aged 3–6 years had baseline ND testing in PROMOTE; group-mean unadjusted Cognitive Composite scores on the MSEL were 85.8 (SD 18.2) and KABC-II were 79.5 (SD 13.2). Composite score group-mean differences were noted by country, with South African and Zimbabwean children having higher scores. In KABC-II multivariable analyses, maternal age > 40 years, lower education, male sex, and stunting were associated with lower composite scores.

Conclusions:

Among a large cohort of 3–6 year old CAHEU from eastern/southern Africa, group-mean composite ND scores averaged within the low-normal range; with differences noted by country, maternal clinical and socioeconomic factors.

Introduction

Tremendous progress has been made globally in prevention of perinatal HIV transmission of HIV-1. With use of maternal triple anti-retrovirals (ARVs) during pregnancy and breastfeeding, the overall health of the mother has improved and perinatal transmission rates of less than 1% are achievable.^1,2 During 2020, UNAIDS estimated that 1.3million women living with HIV were pregnant of whom about 80% received antepartum ART; and that there were over 15 million children who were perinatally ARV-HIV exposed but uninfected ((CAHEU).³ Some studies have raised concerns that potent ARV regimens used for both maternal treatment as well as prevention- of- perinatal- HIV-transmission may place these children at risk for late sequelae from in utero as well as breastmilk exposure during a period of rapid brain growth and neurodevelopment.^4–5

Over the past two decades, perinatal HIV researchers in both resource-rich as well as resource-limited settings have studied potential untoward effects of perinatal exposure to maternal HIV and ARVs, on early developmental outcomes of young CAHEU including both neurodevelopment as well as physical growth. A number but not all studies of neurodevelopment^6–11 among CAHEU compared to unexposed children born to women without HIV (HUU), found that CAHEU demonstrated mild decrements in measures of overall cognitive functioning and/or specific deficits including receptive language and visual motor function when measured during the first two-three years of life. Some studies assessed potential perinatal exposure to specific ARV regimens on CAHEU’s neurodevelopment (ND), and reported some ND testing differences by specific ARVs or ARV class.^12–14 Other studies also found greater risk of stunting; along with potential effects on immune functioning with increased susceptibility to infections for AHEU compared to HUU children.^15–16 Biologic mechanisms postulated to be related to these findings include fetal HIV and ARV exposure in utero through cross-placental passage; and/or during breastfeeding. Based on animal studies, specific ARVs have been implicated as potentially causing toxicities affecting CNS and muscle mitochondrial function as well as direct toxicities to glial cells.^4–5 However, the majority of ND studies of CAHEU only presented ND findings through the first 12–24 months of age; without longer follow up into later preschool/early primary school years. Longer term assessment of neurodevelopmental outcomes is necessary to determine if early ND findings identified in some studies among young CAHEU will persist or resolve by the late preschool to early years.

The ongoing five year PEPFAR PROMISE Ongoing Treatment Evaluation (PROMOTE) Cohort study¹⁷ in eastern and southern Africa provided a unique opportunity to gather longer term ND and growth follow up data as the CAHEU from an earlier multi-site perinatal trial (PROMISE 1077BF) aged into late preschool and early school years. The PROMOTE Cohort study enrolled and followed up approximately 1900 mothers living with HIV-1 on standard of care ART and their children in Uganda, Malawi, Zimbabwe, and South Africa. The purpose of this analysis is to present ND findings among PROMOTE CAHEU at study entry.

Methods:

The PROMOTE Cohort study was a five year follow up study of mothers living with HIV (LWHIV) on lifelong ART and their CAHEU. It was conducted at eight research sites (one in Uganda; two in Malawi; three in Zimbabwe; and two in South Africa). PROMOTE followed up mothers living with HIV (LWHIV) and their CAHEU who were originally enrolled in the PROMISE 1077 multi-site perinatal trial conducted between 2011 and 2016. The mothers consented to be in this five years follow up study with their children and attended study visits every 6 months. For this analyses we included children enrolled in PROMOTE who were ARV and HIV perinatally exposed and uninfected, and 36 months or older at time of PROMOTE baseline neurodevelopmental (ND) testing. The study entry pediatric assessment included age-appropriate baseline ND testing described below as well as collection of growth measurements and clinical data. Relevant ethics committees and institutional review boards approved this study at all sites.

The ND testing included the Mullen Scales of Early Learning (MSEL)¹⁸ which was used for ND testing of the PROMOTE children aged 36–48 months at baseline; and the Luria model of the Kaufman Assessment Battery for Children, Second Edition (KABC-II) with a summary Mental Processing Index (MPI),¹⁹ which was used for the baseline testing of the PROMOTE children who were 48 month-60 months of age. These standardized ND tests with overall cognitive and motor scores (mean 100) and subscale scores (see Tables 2–4 for specific subscales) had previously been validated²⁰ and adapted²¹ in African settings including Uganda; and were administered by trained ND testers at the research sites. The mothers were also administered questionnaires concerning their child’s health (excellent through poor), development and behavior, as well as a standardized Patient Health questionnaire (PHQ-9) on maternal depression,²² and questions on ART adherence. At study entry, the mothers had HIV viral load and CD4 testing; as well as WHO HIV clinical staging. PROMOTE study staff measured physical growth (weight, height, head circumference, body mass index [BMI]) of the children using standardized height stadiometer, and weight scale instruments. These anthropometric measurements were conducted by trained nursing staff.

Table 2A:

Baseline mean Mullen Scales of Early Learning (MSEL) composite and component T-scores overall and by PROMOTE Child Characteristics (n=812)

Characteristic	Mean Mullens composite (std)	SD	Gross motor	SD	Fine motor	SD	Visual reception	SD	Expr language	SD	Recep language	SD
OVERALL	85.8	18.2	44.6	10.8	41.0	13.2	39.8	17.0	45.3	11.5	41.1	9.5
Sex
Male	84.2	17.9	46.9	10.5	40.6	12.6	38.5	20.8	44.3	11.7	40.6	9.2
Female	87.3	18.3	42.5	10.8	41.4	13.9	41.1	12.0	46.4	11.3	41.7	9.8
Country
Uganda	77.9	13.4	46.7	7.2	38.4	12.6	35.8	11.3	40.7	9.1	37.5	7.5
Malawi	79.6	15.3	43.0	11.9	36.4	11.2	38.5	25.1	42.8	10.8	37.7	9.2
Zimbabwe	92.8	18.2	46.6	13.1	44.7	14.0	41.3	11.6	47.4	11.3	45.2	9.3
South Africa	92.6	19.7	40.9	6.3	45.3	12.8	43.2	10.9	50.6	12.3	44.1	8.7
Current health status
Excellent	87.3	18.8	44.2	10.4	42.5	13.8	40.5	22.7	45.3	11.4
Very good	84.6	17.4	46.0	11.3	39.9	12.9	39.5	11.9	45.1	11.4	40.3	9.3
Good	84.9	18.0	43.3	11.1	40.0	12.6	39.2	11.4	45.6	12.2	41.6	8.7
Fair	88.6	18.0	-	-	45.6	10.5	36.9	18.4	46.9	10.2	42.0	9.4
Poor	-	-	-	-	-	-	-	-	-	-	-	-
Birthweight
≥2500 g	85.7	18.3	45.1	11.0	41.3	13.3	39.3	11.8	45.2	11.7	41.0	9.5
<2500 g	85.9	17.2	51.8	9.3	39.3	12.7	42.8	36.0	46.0	10.2	41.7	9.5
Gestational age at birth
≥37 weeks	86.2	18.1	44.9	11.0	39.4	11.7	41.5	13.3	45.3	11.5	41.2	9.5
<37 weeks	82.8	18.2	41.8	7.5	42.1	37.2	37.3	12.0	45.7	11.6	40.4	9.2
Anthropometry
Not stunted (HAZ ≥ −2.0)	87.6	18.7	45.1	9.9	42.5	13.5	40.4	11.7	46.1	11.8	41.8	9.4
Stunted (HAZ < −2.0)	81.4	16.1	43.8	12.7	37.5	11.8	38.3	25.5	43.6	10.8	39.6	9.6
Not underweight (WAZ ≥ −2.0)	86.0	18.1	45.3	10.6	41.2	13.2	39.6	11.6	45.5	11.5	41.1	9.4
Underweight (WAZ <−2.0)	81.6	19.0	37.9	11.5	37.6	13.2	42.4	51.3	42.4	11.7	41.2	10.0
Hemoglobin
≥10.9 g/dl	86.1	17.7	45.5	10.4	41.5	13.4	37.5	11.9	44.1	11.7	41.3	9.4
<10.9 g/dl	84.1	20.2	41.6	12.1	38.7	12.3	40.3	18.0	45.6	11.5	40.5	9.7

Abbreviations: g, grams; WAZ, weight-for-age z-score; HAZ, height-for-age z-score; Hb, hemoglobin; dl, deciliter; SD, standard deviation

¹Current health status of child was self-reported by mother at baseline interview.

Table 4 A.

Change in mean baseline Mullen Scales of Early Learning (MSEL) composite score associated with PROMOTE Child and Maternal Characteristics (linear regression, with variance clustered by site)

	Unadjusted			Adjusted1
Characteristic	Change in mean Mullen composite	95% CI	p	Change in mean Mullen composite	95% CI	p
CHILD
Sex
Male	REF			REF
Female	+3.1	1.6–4.6	0.002	3.4	1.2– 5.5	0.016
Country
Uganda	REF			REF
Malawi	1.8	−3.5 – 7.1	0.454	2.5	−0.5 – 4.4	0.028
Zimbabwe	15.0	9.4–20.5	<0.001	+14.1	10.5 – 17.5	0.001
South Africa	14.7	0.1–29.4	0.049	+12.2	9.2 – 15.2	0.001
Anthropometry
Not stunted (HAZ ≥ −2.0)	REF			REF
Stunted (HAZ < −2.0)	−6.2	−10.9 – −1.5	0.016	−3.3	−10.9 – 4.4	0.260
Not underweight (WAZ ≥ −2.0)	REF			REF
Underweight (WAZ <−2.0)	−4.4	−11.2 – 2.4	0.172	−−2.9	−−10.0–4.3	0.285
Gestational age at birth (per week)	0.44	−1.2 – 2.1	0.542	0.12	−0.76 – 1.0	0.695
Birthweight (per 100g)	−0.10	−0.30 – 0.10	0.283	−0.05	−0.2–0.4	0.155
Hemoglobin
≥ 10.9 g/dl	REF
< 10.9 g/dl	−1.8	−6.6 – 3.1	0.414	−1.1	−5.0 – 2.8	0.424
MOM
Maternal age
21–29 years	REF			REF
30–39 years	1.4	−3.6 – 6.5	0.522	−0.02	−8.1– 8.1	0.995
40+ years	−3.3	−13.4 – 6.7	0.460	−5.5	−26.4–15.3	0.460
Marital status
Regular partner/married	REF			REF
No regular partner	0.38	−8.5 – 9.2	0.923	−2.1	−7.1 – 3.1	0.368
Separated/widowed/divorced	−5.6	−11.1– −0.05	0.048	−1.7	−6.2 – 3.0	0.385
Maternal education
Less than secondary	REF			REF
Any Secondary or more	9.2	3.6 – 14.9	0.006	2.7	−0.1 – 6.4	0.102
Maternal employment
Formal	REF			REF
Informal	−5.7	−17.9 – 6.5	0.204	−2.1	−7.9 – 3.7	0.303
Not employed	−1.5	−9.6 – 6.4	0.669	−0.9	−6.5 – 2.6	0.478
Parity (per additional birth)	−1.7	−3.4 – 0.07	0.058	−0.5	−1.7 – 0.4	0.484
Viral load
Suppressed (VL<200 copies/ml)	REF			REF
Unsuppressed (VL≥200 copies/ml)	−6.7	−12.8– −0.5	0.037	−4.5	−5.4 – 0.15	0.057
Per 100 CD4 cells/mm 3	0.41	−0.49 – 1.3	0.319

Abbreviations: IQR, interquartile range; SD, standard deviation; VL, viral load; ml, milliliter; g, grams; WAZ, weight-for-age z-score; HAZ, height-for-age z-score; Hb, hemoglobin; dl, deciliter

¹Multivariable model adjusted for child sex, country, stunting and underweight status, gestational age at birth, maternal age, maternal marital status, maternal level of education, maternal employment, maternal parity and maternal viral load

Statistical Methods:

The cognitive composite scores of the PROMOTE CAHEU’s baseline Mullen Scales of Early Learning¹⁸ and the global Mental Processing Index (MPI) score of the Kaufman KABC II¹⁹ were analyzed unadjusted; and by multivariable linear regression analyses controlling for demographic and clinical factors previously shown to be associated with developmental delays in the general and HIV research literature, including maternal age, clinical and socioeconomic factors, maternal viral load, maternal depression; and infant growth measures (e.g. birthweight, height-for- age, weight-for-age)) and anemia.^23–30 For the multivariable analyses, adjustment was made for maternal age, parity, country of residence, maternal HIV viral load. Pediatric growth measures were plotted on WHO growth norm charts³¹ and height-for- age, weight-for- age Z scores were calculated and compared to the WHO 50^th percentile mean as well as BMI. The Stata 16.1 Statistical package was used for these analyses.³²

Funding and financial statements:

Funding was provided by the US government (OGAC and NIAID/NIH). The authors have no financial or other conflicts of interest.

Results:

A total of 1786 children from the PROMISE randomized trial were enrolled into PROMOTE of whom 1750 (98%) were ARV/HIV exposed but uninfected (AHEU). Among the AHEU children, 1647 (94.1%) had ND MSEL or KABC II testing done at baseline and were over 36 months of age. These included 395 (24.0%) aged 36– 48 months who had only MSEL testing done; 417 (25.3%) children who had both the KABC II and MSEL testing done; and 811 (49.2%) aged 48–60 months who had only the KABC II testing done (See Figure 1).

Figure 1: Consort Schema for PROMOTE Cohort ND Analyses.

Consort Formation of analysis cohort by tests performed.

Abbreviations:

KABC-II: Kaufman Assessment Battery for Children II;

Mullen: Mullen Scales of Early Learning.

Baseline characteristics of the PROMOTE children with ND testing (Table 1A) were as follows: 835 51% male, median age was 50 months (range 39–58 months of age); median birth weight 2930 gram with 241 (15%) born low birthweight. Among the children analyzed 242 (15%) were born low birthweight (LBW <2500 grams); of which 17 (1%) were born very low birthweight (VLBW, <1500grams). The median gestational age at birth for PROMOTE infants was 39 weeks based primarily on Ballard³³ newborn testing done in the PROMISE trial; with 241 (15%) born preterm (PT <37weeks) of which 17 (2%) were born very preterm (VPT at 34 weeks or less). Close to 99% (1609) CAHEU were judged to be in “good, very good or excellent health” by their mother or primary caretaker. Using WHO growth standards, 75 (5%) were < 2 SD below the mean for weight-for-age; whereas 351 (22%) were stunted (> 2 SD below the WHO 50^th percentile mean Z scores for height-for-age.

Table 1A:

Baseline characteristics of PROMOTE Children (n=1623)

Characteristic	n	% or Median(IQR)
Sex
Male	835	51%
Female	788	49%
Age(months)	1623	50 (39–58)
Country
Uganda	300	18%
Malawi	518	32%
Zimbabwe	396	24%
South Africa	409	25%
Gestational age at birth
Weeks	241	39 (38–40)
<37 weeks	209	13%
<34 weeks	32	2%
Birthweight
≥2500 g	1622	2930 (2600–3250)
1500 g– <2500 g	225	14%
<1500 g	17	1%
Current health status 1
Excellent	705	43%
Very good	593	37%
Good	304	19%
Fair	20	1%
Poor	1	<1%
Anthropometry
Underweight (WAZ < −2.0)	75	5%
Stunted (HAZ < −2.0)	351	22%
WAZ	1373	−0.53 (−1.18– 0.11)
HAZ	1373	−1.29 (−2.03 – −0.53)
Hb g/dl	1605	12.0 (11.3–12.7)

Abbreviations: IQR, interquartile range; g, grams; WAZ, weight-for-age z-score; HAZ, height-for-age z-score; Hb, hemoglobin; dl, deciliter

¹Current health status of child was self-reported by mother at baseline interview.

Maternal PROMOTE characteristics at baseline (N=1609) are shown in Table 1B. By self-report, average maternal age at entry into PROMOTE was 31 years (range 27–35 years); 1301 (81%) of mothers had a regular partner or were married; 126 (8%) had no current male partner and 182 (11%) were separated, widowed or divorced; 1066 (66%) had completed some or all secondary school and 93(6%) had completed some or all university education. Maternal median entry CD4 cells/mm³ was 838 (658–1047); 1428 (90%) of the PROMOTE women had viral load (<200 RNA copies/ml) at baseline entry into PROMOTE; and 191 (12%) of mothers scored in the moderate to severe depression range using the PHQ-9 Depression Scales.²²

Table 1B.

Baseline characteristics of PROMOTE Mothers (n=1609

Characteristic	n	% or Median(IQR)
Median age, years	1609	31 (27–35)
Age category
18–29 years	619	39%
30–39 years	858	53%
40 years	128	8%
Marital status
No regular partner	126	8%
Primary regular partner	413	26%
Married	888	55%
Separated/widowed/divorced	182	11%
Education
None	61	4%
Some primary	263	16%
Completed primary	122	8%
Some secondary	547	34%
Completed secondary	519	32%
University	93	6%
Work status
Yes formal	337	21%
Yes informal	518	32%
No	748	47%
Moderate-Severe depression 1	191	12%
CD4 count, cells/mm 3	1588	838 (658–1051)
HIV viral load VL≥ 200 copies/ml (unsuppressed)	159	10%
VL <200 copies/ml (suppressed)	1428	90%

Abbreviations: IQR, interquartile range; SD, standard deviation; VL, viral load; ml, milliliter

¹Depression was measured by self-report at baseline interview using the Patient Health Questionnaire-9 (PHQ-9). Responses were scored, and those scoring ≥10 were considered to have moderate to severe depression. Ref: Kroenke K, Spitzer R, Williams JBW. The PHQ-9: Validity of a brief depression severity measure. J Gen Intern Med. 2001 Sep; 16(9): 606–613.²⁰

Factors Associated with MSEL and KABC II Developmental Scores.

Table 2: Mean MSEL Cognitive composite scores by pediatric and maternal sociodemographic/clinical factors.

MSEL Cognitive Composite Scores

MSEL Cognitive Composite Scores were on average 3 points higher for female compared to male children; and children from Zimbabwe and South Africa on average had scores that were 15 points higher than CAHEU from Uganda and 13 points higher than CAHEU from Malawi. Lower growth parameters were also associated with lower MSEL scores: CAHEU with height- for- age (HAZ) scores more than 2 SD below the mean had MSEL scores that averaged 6.2 points lower than children with higher HAZ scores; and those with weight-for-age (WAZ) Z-scores more than 2 SD below the WHO mean, scored an average of 4.4 points lower on MSEL scores compared to those with higher WAZ scores.

For MSEL Gross Motor Scores

For MSEL Gross Motor Scores, males scored an average of 4.4 points higher than females; children from Uganda and Zimbabwe had the highest Gross Motor Component Scores; those children who were underweight (WAZ more than 2 Standard deviations below the WHO mean) had average Gross Motor scores 7.4 points lower than children who were not underweight; and children with hemoglobin <10.9 g/dl averaged gross motor skills 3.9 points lower than children with higher hemoglobin.

For Fine Motor Component Scores

For Fine Motor Component Scores, mean scores were similar for females and male children; and children from Zimbabwe and S. Africa scored about 7 points higher on average than children from Malawi and Uganda. Those children with HAZ and WAZ scores that were more than 2 SD below the mean had Fine Motor Skills that averaged 5 points lower than children with higher HAZ and WAZ scores; and those with lower hemoglobin (<10.9 g/dl) had scores that averaged 2.8 points lower.

For Visual Reception Component Scores

For Visual Reception Component Scores, females scored an average of 2.6 points higher than males; those from S. Africa had the highest mean scores (average 7.4 points higher) than children from Uganda who had the lowest average visual reception scores. In contrast, CAHEU with hemoglobin scores < 10.9 g/dl averaged 2.8 points higher on visual perception scores than CAHEU with higher hemoglobin.

In terms of the Receptive and Expressive Language Component Scores

In terms of the Receptive and Expressive Language Component Scores, male and female children had similar scores irrespective of site; and children from Zimbabwe and S. Africa had higher average Receptive and Express language scores compared to children from Uganda and Malawi. Receptive and Expressive Language Component scores were similar across for HAZ and WAZ score categories; and by hemoglobin status.

By maternal variables, CAHEU born to younger mothers scored 3.9 points higher on the MSEL cognitive composite score than those born to mothers over age 40 years; those children whose mothers had a secondary or more education had average scores that were 9.3 points higher; those whose mothers were para 1 scored 9.5 points higher than those who were para 5 or above; and those children whose mothers had viral load < 200 copies/ml had average scores that were 6.7 points higher. However, there were no statistical differences by maternal depression scores on the PHQ-9. Some minor MSEL score differences were also noted for receptive language by maternal education level; and for gross motor scores with higher scores for CAHEU whose mothers were widowed or divorced compared to those who were married or had a regular partner.

Table 3 A–B: Mean KABC-II Scores by various pediatric and maternal clinical/sociodemographic factors.

Table 3A:

Baseline mean KABC-II standardized scores overall and by PROMOTE Child Characteristics (n=1228

Child Characteristic	MPI	SD	Sequential processing	SD	Simultaneous processing	SD	Learning	SD	Delayed recall	SD	Non verbal Index	SD
OVERALL	79.5	13.2	87.5	13.8	77.0	14.9	84.5	13.6	87.7	14.7	76.0	14.7
Sex
Male	78.6	12.8	86.6	13.8	75.9	14.7	84.3	13.9	87.1	15.5	75.0	14.3
Female	80.4	13.6	88.4	13.8	78.2	15.1	84.7	13.3	88.2	13.8	77.0	15.1
Age
3 years	78.4	14.4	86.7	14.9	76.3	15.5	83.7	13.7	87.3	13.8	74.2	14.7
4 years	80.6	12.5	87.7	13.3	78.5	14.5	85.5	13.6	87.0	15.0	77.6	14.6
5 years	78.6	12.9	86.5	13.5	76.3	14.6	83.6	13.2	88.9	15.5	75.1	14.6
Country
Uganda	75.8	12.0	86.3	10.9	72.8	14.2	82.5	12.1	87.5	12.7	72.0	13.7
Malawi	74.0	10.7	81.3	14.0	71.9	12.9	81.3	12.3	79.5	12.2	70.3	12.1
Zimbabwe	83.8	14.5	91.2	14.1	81.1	15.3	85.7	14.5	89.8	17.0	80.1	16.2
South Africa	85.0	12.1	91.1	12.7	82.8	14.2	88.8	14.0	91.8	12.5	82.2	13.6
Current health status
Excellent	81.7	13.4	89.2	13.6	79.6	14.7	85.7	13.8	89.4	15.3	78.7	14.8
Very good	77.5	12.9	85.5	13.7	75.3	14.6	83.0	14.0	86.2	14.1	73.8	14.1
Good	78.0	12.8	85.1	14.3	74.6	15.1	84.6	12.1	87.3	12.3	73.6	14.9
Fair	76.8	12.0	88.8	9.9	73.3	16.7	2.0	8.7	80.4	31.3	73.8	13.6
Poor	-	-	-	-	-	-	-	-	-	-	-	-
Anthropometry
Stunted (HAZ < −2.0)	74.8	12.6	83.1	14.8	71.7	13.1	81.6	13.8	85.7	15.0	71.3	13.1
Not stunted (HAZ ≥ −2.0)	80.5	13.2	88.1	13.5	78.2	15.1	85.1	13.4	88.1	14.6	77.0	14.9
Underweight (WAZ <−2.0	74.2	10.9	85.8	12.1	69.6	11.5	81.8	12.5	81.4	13.4	70.0	12.0
Not underweight (WAZ ≥ −2.0))	79.7	13.3	87.2	13.9	77.3	15.0	84.6	13.6	88.0	14.7	76.2	14.8
Hemoglobin
≥10.9 g/dl	79.6	13.1	84.6	17.7	75.3	14.7	82.9	14.1	88.1	7.0	74.4	14.4
<10.9 g/dl	78.4	14.3	87.5	13.1	77.3	14.9	84.8	13.5	87.6	14.3	76.2	14.8

Abbreviations: g, grams; WAZ, weight-for-age z-score; HAZ, height-for-age z-score; Hb, hemoglobin; dl, deciliter; SD, standard deviation

¹Current health status of child was self-reported by mother at baseline interview.

Table 3B:

Baseline mean KABC-II standardized scores overall and by PROMOTE Maternal Characteristics (n=1228)

Characteristic	MPI	SD	Sequential processing	SD	Simultaneous processing	SD	Learning	SD	Delayed recall	SD	Non verbal Index	SD
OVERALL	79.9	12.7	87.5	13.8	77.1	15.0	84.8	14.1	87.9	14.9	76.1	14.9
Maternal age
21–29 years	79.6	13.2	87.8	14.9	77.4	15.5	85.9	13.3	88.3	14.0	76.4	14.6
30–39 years	79.8	13.5	87.5	13.3	77.1	14.9	85.0	14.6	87.9	15.8	76.2	15.4
40+ years	77.4	11.5	86.1	12.1	75.8	12.6	83.0	13.6	86.8	12.1	74.6	13.5
Marital status
Regular partner/married	79.9	13.2	87.9	13.6	77.6	15.0	85.2	14.3	88.3	14.7	76.5	15.0
No regular partner	80.6	11.0	89.0	13.3	78.1	13.5	86.1	11.6	86.5	18.2	77.5	14.6
Separated/widowed/ divorced	75.9	14.0	83.6	14.8	72.8	15.3	80.8	13.6	86.8	13.8	72.6	14.5
Maternal education
Less than secondary	73.7	11.1	82.6	12.1	71.6	13.2	80.8	12.2	84.4	13.6	69.9	13.2
Any secondary or more	81.8	13.3	89.5	14.0	79.3	15.1	86.4	14.4	89.3	15.1	78.6	14.9
Maternal employment
Formal	80.8	13.3	88.6	14.0	78.9	14.9	87.0	14.2	88.5	15.2	78.1	14.8
Informal	78.0	12.6	86.9	12.8	75.3	14.3	83.7	14.0	86.4	16.2	74.5	15.1
Not employed	80.0	13.6	87.4	14.4	77.6	15.3	84.6	14.0	88.8	13.6	76.4	14.8
Parity
1 child	83.2	14.1	90.3	14.6	81.6	17.0	88.4	14.7	91.5	11.0	80.3	15.7
2 children	81.1	13.3	88.3	13.0	78.9	14.6	85.8	13.6	91.1	14.4	77.8	15.0
3 children	79.8	12.9	87.6	13.6	76.5	14.3	85.0	14.4	85.6	16.5	76.1	14.8
4 children	78.1	12.8	86.0	15.3	75.0	15.3	83.2	13.6	87.3	14.6	74.1	14.9
5+ children	75.0	12.5	85.4	12.2	74.0	13.5	81.6	13.9	85.0	14.0	72.5	13.3
Viral load
Suppressed (VL<200 copies/ml)	79.9	13.0	87.8	13.6	77.4	14.8	85.0	14.3	88.2	14.3	76.4	14.6
Unsuppressed (VL≥200 copies/ml)	76.0	14.8	84.4	13.1	74.6	16.1	82.8	11.9	85.5	18.4	74.1	16.7
Depression
None or mild	79.6	13.2	87.2	14.1	77.3	14.7	84.6	13.7	87.6	14.7	76.2	14.6
Moderate-severe	78.7	13.3	86.6	12.3	75.3	16.2	84.1	12.6	88.4	14.6	75.8	15.8

Abbreviations: IQR, interquartile range; SD, standard deviation; VL, viral load; ml, milliliter; MPI (Mental Processing Index)

¹Depression was measured by self-report at baseline interview using the Patient Health Questionnaire-9 (PHQ-9). Responses were scored, and those scoring ≥10 were considered to have moderate to severe depression. Ref: Kroenke K, Spitzer R, Williams JBW. The PHQ-9: Validity of a brief depression severity measure. J Gen Intern Med. 2001 Sep; 16(9): 606–613.

KABC II composite Scores/MPI global score

KABC II composite Scores/MPI global score among 48–60 month-old children showed similar findings as were seen for MSEL composite scores among the younger children. In unadjusted analyses, female children had higher group mean scores than males; children from South Africa and Zimbabwe had composite KABC-II group-mean scores 9–10 points higher compared to children from Malawi and Uganda; and that those with excellent health status had composite group-mean scores 5 points higher than those children with only “fair” health status. Mean KABC II scores were 5.7 points lower for those children who were stunted (HAZ scores more than 2 SD below the mean); and KABC II mean scores averaged 5.5 points lower for underweight children whose WAZ scores were more than 2 SD below the mean. Maternal characteristics associated with lower scores included being widowed or divorced, age greater than 40 years, higher parity, lacking secondary education, and detectable viral load.

Table 4 presents differences in mean baselines MSEL and KABC-II scores associated with pediatric and maternal characteristics using unadjusted and multivariable linear regression models.

Results for the MSEL composite scores

Results for the MSEL composite scores show similar patterns to findings in Table 2. Higher scores were associated with female sex, residing in S. Africa and Zimbabwe, hemoglobin ≥ 10.9 g/dl, and height-for-age and weight-for-age within 2 SD of the WHO median HAZ and WAZ scores. Likewise, children scored higher if their mothers had a secondary or higher education, were employed, lower parity, and had undetectable viral load at baseline entry into PROMOTE. With multivariable analyses, findings were in the same direction but dampened in terms of the change seen by pediatric or maternal variable.

Results for the KABC-II unadjusted and multivariable changes in baseline composite scores

Similar patterns to the MSEL unadjusted ND testing results at 36–48 months were seen among children tested with the KABC-II. Female sex was associated with +1.8 points higher composite MPI scores; being from Zimbabwe and S. Africa with increased composite scores of + 8 and +9 points respectively. Stunting was associated with lower composite score of −5.7 points; and weight-for-age composite scores more than 2 SD below the WHO mean were associated with −5.4 change in unadjusted scores; with similar but dampened change patterns seen in multivariable analyses. Likewise, maternal age over 40 years was associated with a score change of −2.2 points; maternal secondary education was associated with scores 8.2 points higher than maternal primary education only; and maternal viral load >200 copies/ml was associated with a −3.9 point lower mean composite scores. With multivariate analyses, test result differences were in the same direction but the average differences tended to be smaller in magnitude.

Discussion

With the roll out of ART for both treatment and prevention-of-perinatal-transmission, increasing numbers of children born to mothers living with HIV have been exposed to ART, both in utero and during breastfeeding. Our analyses add to the current research on neurodevelopmental outcomes of CAHEU by presenting late preschool and early school neurodevelopmental findings among a large cohort of preschool Southern and Eastern African CAHEU aged 3–6 years, who had well documented perinatal exposure to various ARV regimens in a large multisite international (PROMISE 1077BF) clinical trial. The analyses were further strengthened by being able to adjust for a number of demographic and clinical factors including maternal age, parity, country-of-residence, child’s height-for-age and weight-for- age, and maternal HIV viral load. The MSEL neurodevelopmental test results showed overall age-adjusted global mean scores within the low normal range): 85.8 (SD 18.2) for the MSEL; 79.5 (SD 13.2) for the KABC-II; and lower MSEL mean visual-reception and fine motor subscale scores for age. The significance of these findings for the composite global composite scores as well as visual reception and fine motor skills is unclear; but potentially could have a negative effect on later academic performance.^34–37

Using bivariate and multivariable linear analyses, we noted important differences in composite cognitive scores based on demographic socioeconomic and country, as well as maternal and child clinical factors. Of note, older mothers and those with higher parity had children with lower composite cores than younger mothers of parity one; and CAHEU from Uganda and Malawi had significantly lower average composite scores compared to CAHEU from Zimbabwe and S. Africa. Likewise in terms of pediatric clinical factors, CAHEU who were stunted (i.e. whose height-for-age fell greater than 2SD below WHO growth means)³¹ had significantly lower scores on the composite KABC II than children whose height-for-age Z scores were within 2SD of the mean.

Our PROMOTE baseline study results among AHEU exposed preschool and early school aged children in African settings, contribute and extend findings from a number of earlier studies among CAHEU. Based on a review by Le Doare et al,⁶ which included 31 ND studies of HIV infected as well as CAHEU conducted between 1990–2011, the majority of studies were from the U.S. and Europe with only 25% being carried out in resource limited settings such as Africa. They were also conducted during a time of perinatal exposure to less complex ARV regimens than the current triple-ARV (ART) regimens; while our PROMOTE study includes children with exposure to ART as well as other combination ARVs and extended follow-up through age 48–60 months of age. Previous findings were mixed with some studies^7–14 finding lower overall cognitive scores comparable toward ARV unexposed children from similar backgrounds, and also some studies showing language and/or fine motor development deficits. In a meta-analyses/review done by McHenry et al⁸ which included 12 studies conducted between 1995–2014 in the US (6 studies), Africa (5 studies) and 1 study done in South America, findings were that CAHEU had lower mental and motors scores on the Bayley Scales of Infant Development³⁸ compared to HUU children. More recently, Le Roux et al¹⁰ in South Africa assessed ND outcomes for 215 AHEU exposed to ART in utero/postpartum during breastfeeding compared to 30 HUU children: they reported a two-fold increased risk of ND and motor delays for CAHEU compared to HUU children from similar socioeconomic backgrounds. Likewise, Benki-Nugent et al³⁹ in Kenya in a study of 65 CAHEU and 65 HUU children aged 5–12 years, reported significantly lower global scores on the Kaufman-II as well as short- term and delayed memory, attention, and processing speed subscales for the CAHEU group.

Some studies have investigated the associations with various specific ARVS on ND outcomes.^{12–14, 27} In Botswana, Chaudhury et al⁴⁰ conducted developmental testing for 598 AHEU children in two different studies, Tshipidi (2010–2012) and Mma Bana studies (2006–2008) and compared in utero exposure to ART (N= 382) to in utero ZDV exposure (N=210). At 24 months, they found no differences in ND outcomes. Looking at exposure to specific ARVs, in a recent study by Cassidy et al¹³ of CAHEU exposed in utero to either efavirenz (EFV) based ART versus non EFV ART, the authors reported decreased subscale scores in receptive language and visual motor performance for the EFV-ART exposed group compared to non-EFV based ART regimens. In a large multisite study in Malawi and Uganda, Boivin et al²⁹ compared ND findings between 12 months to 48–60 months among AHEU exposed children to age and gender matched HUU children using the MSEL and KABC-II. They found that both AHEU and HUU groups had lower composite scores than the standardized MSEL and KABC-II test means; but reported no significant differences in ND results for CAHEU compared to HUU children by specific type of ARV in utero exposure. Our PROMOTE findings showed low average scores for the CAHEU children, similar to the Boivin study results. The country differences seen in group cognitive scores with Zimbabwe and S. African CAHEU children scoring higher than CAHEU from Uganda and Malawi, may reflect background cultural, maternal education or socioeconomic differences in these African settings.

We acknowledge several limitations to our analyses. First, we did not have a group of children born to HIV-negative mothers to compare their ND results to the PROMOTE AHEU cohort, so findings can only be assessed against MSEL and KABC-II standardized test norms. However, the findings are consistent with several prior ND studies among younger CAHEU. In addition, the study is based on cross-sectional data at study entry into PROMOTE during the late preschool/early school years and does not include follow up of later academic performance. Relative strengths of our study include the large sample size and geographic diversity across multiple countries in Eastern and Southern Africa, which supports the generalizability of the findings for low-middle-income African settings. Another strength is that this is one of the few multi-site African studies to provide ND outcome data among older (3–6 year- old) preschool CAHEU. In addition, the neurodevelopmental testing for the PROMOTE study was done by well-trained testers using ND tests which have been used and validated among African populations.

In conclusion, the PROMOTE ND results contribute our understanding of neurodevelopmental outcomes among 3–6 year old African CAHEU following perinatal HIV-ARV exposure. Our findings are generally reassuring in demonstrating no major deficits in overall cognitive performance for CAHEU on standardized cognitive tests. However, we note the CAHEU’s average scores were toward the lower end of the normal range of the standardized instrument norms, and that there significant differences in scores by country. Longer follow up of CAHEU’s academic performance through the primary school years and adolescence is needed; as well as targeting interventions for CAHEU identified as most-at-risk academically.

Table 2B:

Baseline mean Mullen Scales of Early Learning (MSEL) composite and component T-scores overall and by PROMOTE Maternal Characteristics (n=812)

Characteristic	Mean Mullen composite Score	SD	Gross motor	SD	Fine motor	SD	Visual reception	SD	Expr language	SD	Recep language	SD
OVERALL	85.8	18.2	44.6	10.8	41.0	13.2	39.8	17.0	45.3	11.5	41.1	9.5
Maternal age
21–29 years	85.2	18.6	44.7	9.3	40.5	12.9	39.0	12.1	44.9	11.9	40.4	9.4
30–39 years	86.7	17.8	44.6	12.2	41.8	13.4	40.9	20.8	45.9	11.2	42.1	9.6
40+ years	81.9	17.5	44.5	12.1	38.1	14.1	35.6	10.7	42.9	11.7	37.8	8.3
Marital status
Regular partner/married	86.7	18.7	43.9	9.7	41.3	13.2	40.2	17.9	45.6	11.6	41.3	9.5
No regular partner	86.4	18.2	43.5	13.5	42.4	12.2	39.4	11.5	46.9	12.0	41.3	10.4
Separated/widowed/divorced	80.8	17.1	51.8	16.4	38.1	13.5	36.7	11.6	42.7	10.2	39.6	9.3
Maternal education
Less than secondary	79.1	14.8	43.6	9.8	36.9	12.2	36.1	11.9	42.5	10.9	38.0	8.8
Any Secondary or more	88.4	18.7	44.9	11.1	42.6	13.3	41.2	18.5	46.4	11.6	42.3	9.5
Maternal employment
Formal	88.4	17.0	42.0	8.9	42.4	12.8	41.1	11.2	47.8	12.1	42.3	9.1
Informal	82.7	17.5	46.5	11.1	39.6	13.0	39.5	24.8	43.1	10.5	39.9	0.4
Not employed	86.9	18.8	44.1	11.0	41.5	13.5	39.5	11.1	46.0	11.7	41.6	9.7
Parity
1 child	89.0	20.3	41.8	10.9	42.9	13.3	40.9	10.6	46.9	12.5	41.5	9.2
2 children	86.8	18.6	46.6	10.0	41.8	13.1	40.0	12.5	45.3	11.7	41.6	9.9
3 children	85.9	16.6	44.2	11.6	40.7	13.1	40.7	25.7	45.9	10.8	41.3	9.8
4 children	85.3	18.0	42.4	11.2	41.4	14.1	39.0	11.7	45.5	11.9	41.2	8.9
5+ children	79.5	17.3	45.6	10.7	36.9	11.6	36.8	11.9	41.8	10.8	39.0	8.8
Viral load
Suppressed (VL<200 copies/ml)	86.4	18.6	45	12.7	41.4	13.4	40.0	17.7	45.5	11.8	41.3	9.6
Unsuppressed (VL≥200 copies/ml)	79.7	12.6	52.0	1.4	36.3	11.1	37.1	10.2	43.5	9.4	39.3	7.8
Depression
None or mild	85.8	18.3	44.2	11.2	41.0	13.1	40.1	17.7	45.4	11.4	41.2	9.5
Moderate-severe	85.3	17.1	48.3	7.0	40.8	14.3	37.7	10.9	45.3	12.3	40.3	9.1

Abbreviations: IQR, interquartile range; SD, standard deviation; VL, viral load; ml, milliliter; SD, standard deviation

¹Depression was measured by self-report at baseline interview using the Patient Health Questionnaire-9 (PHQ-9). Responses were scored, and those scoring ≥10 were considered to have moderate to severe depression. Ref: Kroenke K, Spitzer R, Williams JBW. The PHQ-9: Validity of a brief depression severity measure. J Gen Intern Med. 2001 Sep; 16(9): 606–613.²⁰

Table 4B:

Change in baseline KABC-II MPI score associated with PROMOTE Child and Maternal Characteristics (multivariable linear regression)

	Unadjusted			Adjusted
Characteristic	Change in mean KABC-II MPI scores	95% CI	p	Change in mean KABC-II MPI scores	95% CI	p
CHILD
Sex
Male	REF			REF
Female	1.8	0.58–3.1		1.4	0.06 – 2.7	0.042
Country
Uganda	REF			REF
Malawi	−1.8	−6.2 – 2.6	0.365	−1.4	−5.3 – 2.5	0.439
Zimbabwe	7.9	1.2 – 14.7	0.027	7.4	0.87 – 14.0	0.032
South Africa	9.2	2.4 – 16.0	0.015	8.7	2.4 – 15.1	0.014
Anthropometry
Not stunted (HAZ ≥ −2.0)	REF			REF
Stunted (HAZ < −2.0)	−5.7	−7.8 – −3.5	<0.001	−3.3	−5.4 – −1.1	0.009
Not underweight (WAZ ≥ −2.0)	REF			REF
Underweight (WAZ <−2.0)	−5.4	−9.1 – −2.8	0.002	−1.3	−3.7 – 0.96	0.205
Gestational age at birth (per week	0.18	−0.88 – 1.3	0.691	0.01	−0.67 – 0.68	0.980
Birthweight (per 100g)	0.13	−0.05 – 0.31	0.124	0.12	−0.08 – 4.6	0.697
Hemoglobin
≥ 10.9 g/dl	REF			REF
< 10.9 g/dl	−1.2	−5.7– 3.2	0.530	0.68	−3.3 – 4.6	0.697
MOM
Maternal age
21–29 years	REF			REF
30–39 years	0.14	−2.2 – 2.4	0.889	0.58	−1.8 – 2.9	0.584
40 years	−2.2	−5.3 – ).91	0.139	−1.7	−4.5 – 1.0	0.183
Marital status
Regular partner/married	REF			REF
No regular partner	0.69	−4.5 – 5.9	0.764	−2.4	−5.2 – 0.29	0.072
Separated/widowed/divorced	−4.0	−9.8 – 0.84	0.092	−1.3	−5.1 – 2.6	0.462
Maternal education
Less than < Secondary	REF			REF
Any Secondary or more	8.2	5.3– 11.1	<0.001	3.2	0.72 – 5.7	0.019
Maternal employment
Formal	REF			REF
Informal	−2.7	−9.3 – 3.9	0.363	1.5	−1.4 – 4.3	0.267
Not employed	−0.81	−4.7 – 3.1	0.637	1.0	−1.8 – 3.8	0.408
Parity (per additional birth)	−1.9	−3.4 – −0.40	0.020	−0.73	−1.6 – 0.10	0.077
Viral load
Suppressed (VL<200 copies/ml)	REF			REF
Unsuppressed (VL≥200 copies/ml)	−3.9	−6.5 – −1.3	0.009	−2.1	−6.0 – 1.8	0.250

Abbreviations: IQR, interquartile range; SD, standard deviation; VL, viral load; ml, milliliter; g, grams; WAZ, weight-for-age z-score; HAZ, height-for-age z-score; Hb, hemoglobin; dl, deciliter

¹Multivariable model adjusted for child sex, country, stunting and underweight status, gestational age at birth, birthweight, hemoglobin level, maternal age, maternal marital status, maternal education, maternal employment and maternal viral load.