Growth of HIV-exposed and HIV–unexposed children in South Africa. Anthropometric nutritional status and growth rates

Background and objectives South Africa is characterised by colliding epidemics of non-communicable and communicable diseases. Additionally, the prevalence of stunting, which affects almost a third of children under-five, has remained unchanged over the past decade. Regarding communicable diseases, infection with Human Immunodeficiency Virus (HIV) is endemic as South Africa is home to 1 in 5 people living with HIV globally, and approximately one in three pregnant women is HIV positive. Despite this, the success of the Prevention of the Mother to Child HIV Transmission (PMTCT) programme, particularly in the era of widespread use of triple antiretroviral treatment (ART) for PMTCT and maintenance of maternal health, has made it possible for most HIV-exposed children to remain uninfected. Thishas led to an emergence of a large population of HIV exposed uninfected (HEU) children for whom there are limited data on health outcomes. Notwithstanding the benefits of ART for PMTCT, in-utero exposure to ART may have risks that require further investigation. The significance of various postnatal growth velocities also needs further exploration as it is a research area that is not well understood and international growth velocity standards were not available prior to 2009. Thus very few studies have applied these standards. Given this background, this thesis sought to describe child growth and its determinants in South Africa, a country with high malnutrition and HIV prevalences, under different PMTCT policies between 2002 and 2013, using the latest WHO growth standards. Specifically the thesis aimed to assess the effect of early infant feeding practices on growth velocity in the first 6 months, and the effect of both on body mass index-for-age-age Z-score (BMI-for-age Z-score) at 2 years of age (Paper I); to compare the longitudinal growth of HIV-exposed and -unexposed children in South Africa using the WHO growth velocity standards (Paper II) and to study the effect of infant in-utero HIV and ART exposure on preterm delivery (PTD), low birth weight (LBW) and small-for-gestational age (SGA) at birth, and underweight-for-age (UFA) at six weeks postpartum in children attending primary health care facilities in South Africa during established implementation of WHO PMTCT Option A policy (Paper III). Methods: Data from 3 studies conducted in South Africa were utilised. For Paper I HIV-negative women were recruited in pregnancy and followed-up at 3, 6, 12 and 24 weeks and 2 years postpartum with their children. For Paper II, the majority of the HIV-positive and - negative women were also recruited antenatally, with only a few recruited postnatally while still in hospital, and followed up with their children at 3, 24 and 36 weeks postpartum. For Paper III, data from a cross-sectional survey of HIV-exposed and -unexposed children attending their 6-week immunisation clinic visit were utilised. Exposures: Only self-reported HIV-negative women and their children were considered in Paper I. The main exposures in this paper were infant feeding practices and growth velocity in the first few months of life. The child’s HIV infection and exposure status, and the additional antiretroviral drug exposure status for children born to HIV-positive mothers, were the main exposures in Papers II and III. In these papers the child’s HIV infection status was ascertained by the polymerase chain reaction test (PCR) using dried blood spot specimens obtained from the children by trained data collectors. In each primary study questionnaires were used to also collect information on key factors related to child health from mothers/caregivers, including child feeding practices established through 24-hour and seven-day recall questions. Outcomes: Child weights and recumbent lengths/standing heights were measured by well-trained data collectors in Papers I and II and extracted from the patient-held road to health booklets in Paper III. Birthweights and gestational ages (based on fundal height and/or last menstrual period) were extracted from the road to health booklets in all three studies. Anthropometric scoring was done using the 2006 WHO attained growth standards (used to estimate weightfor- age Z-scores (WAZ), length-for-age Z-scores (LAZ), weight-for-length Z-scores (WLZ) and body mass index(BMI)-for-age Z-score), 2009 WHO growth velocity standards (used to estimate weight and length velocity Z-scores (WVZ and LVZ respectively)) and recent Intergrowth-21st standards for assessing new born size for term and preterm infants (used to estimate birthweight-for-gestational-age Z-scores). Birth outcomes included low birth weight (LBW), preterm delivery (PTD) and small-for-gestational age (SGA). Analysis: Simultaneous quantile regression was used to assess the effect of 1) early infant feeding practices on growth velocity in the first 6 months, and 2) both on BMI-for-age Z-score at 2 years of age in Paper I. Mixed effect regression was used to compare the mean growth velocity Z-scores of HIV-exposed and -unexposed children in Paper II. Logistic regression was used to study the effect of in-utero HIV and ART exposure on PTD, LBW and SGA at birth, and UFA at six weeks postpartum in Paper III. Results: Paper I demonstrated that children who were not breastfed at 12 weeks had higher mean WVZ between 12 and 24 weeks, higher BMI-for-age Z-scores at 2 years and were more likely to be overweight or obese. Although most of the children were initiated on breastmilk early, the proportion of breastfed children decreased in the first 12 weeks of life while the frequency of formula feeding increased. The early introduction of solids such as cereals was also common. Paper II demonstrated that HIV-infected children were not only more underweight, wasted and stunted compared to HEU and HIV unexposed uninfected (HUU) children, but also had poorer growth velocity in the first few of months. The data also showed that maternal viral load, LBW and Nevirapine use were independent influencers of growth velocity in HIVexposed children and that HIV-positive children had increased infectious morbidity compared to HUU children. Paper II also demonstrated that HEU children had similar attained growth and growth velocities compared to HUU children in the absence of maternal ART. Unlike in Paper II, data from Paper III demonstrated that HEU children had poorer birth and early attained growth outcomes than HUU children. The results in Paper III also demonstrated that HEU children born to women with unmanaged HIV infection had an increased odds of being born preterm than children born to women on ART, and that children whose mothers initiated ART before conception had an increased odds of PTD than children whose mothers started ART after conception, but no increased odds for other outcomes. Conclusion: This thesis highlights the importance of not only addressing the double burden of malnutrition in South Africa, but also the burden of colliding disease epidemic of communicable and noncommunicable diseases. The strong association observed between infant HIV infection and poor growth highlights the importance of addressing the unfinished agenda of combating the HIV epidemic. While early initiation of HIV-positive women on ART is important for preventing MTCT and maintenance of maternal health, data does signal that pre-conception ART initiation may have an adverse effect on PTD. As access to ART increases routine surveillance system should be set up to monitor adverse outcomes. The emergence of a large population of HEU for whom there are limited data also warrants the urgent need for the close follow-up, through surveillance systems and in-depth cohort studies, of this sub-population of children. There is also an urgent need to tackle persistent undernutrition in both HIVexposed and -unexposed South African children while curbing the concomitant rise of overnutrition, possibly thorough facility and community based support programmes