Catch-up growth and nutritional status of preterms born at a Ugandan tertiary hospital: An Observational study.

Introduction Advances in neonatal care have led to the increasing survival of preterm/low birth weight infants worldwide. However, there is limited data on growth outcomes following preterm births especially in low and middle income countries. We assessed the catch-up growth, nutritional status and associated factors of Ugandan children who were born preterm/ low birth weight at Mulago National Referral Hospital. Methods: We enrolled children aged 22-38 months who had been born as preterm with low birth weight and their mothers. Participants were identied and recruited from the follow up clinic for preterm babies discharged from the neonatal unit of Mulago Hospital. Anthropometric measurements for mothers and children were taken. The children’s weight for height z-scores, height for age z-scores, weight for age z-scores, head circumference and mid upper arm circumference (MUAC) were taken and the mothers’ body mass index were generated based on the World Health Organization standards. Results: Of the 251 children and mother pairs recruited, 129 children (51.4%) were male, the mean age was 29.7 months (SD±4.5) and the mean maternal age was 29.9 (SD±5.3). 101(40.1%) of the children enrolled had attained catch up growth on the normal anthropometric measurements for their age. Among the participants, the prevalence of wasting, underweight and stunting was: 8 (3.2%), 36 (14.4%) and 106 (42.2%) respectively. Factors associated with stunting were male sex (AOR 2.36, 95% CI 1.42 to 3.95); p=0.001), maternal age ≤ 25years: AOR 2.27 (95%CI 1.13, 4.52); p= 0.020, and maternal height <150cm: AOR 5.57(95%CI 1.90, 16.94); P=0.002. Associations with underweight in the children were; birth weight ≤ 1500 gms: AOR 2.38(95%CI 1.14, 4.95); p=0.020 and post-natal hospitalization of more than 14 days: AOR


Background
Prematurity is on the rise globally, contributing about 15 million births annually [1]. More than 60% of preterm with low birth weight (LBW) babies are born in Asia and Sub Saharan Africa [2]. Prematurity causes an estimated one million neonatal deaths worldwide and it is a signi cant contributor to morbidities among survivors extending to adulthood [3,4].
Uganda is ranked 13th of the 184 countries in the world with the highest number of babies born prematurely; with approximately 14% of babies born preterm [1,5]. Although mortality has been high among preterm/LBW infants, advances in neonatal care have led to better survival of preterm and low birth weight infants worldwide [2,6]. Therefore, children born preterm should be followed up with long term growth monitoring with the aim of achieving catch-up growth as early as possible to combat health consequences like undernutrition later in life [7,8]. Growth status is de ned by normal anthropometric measurements for age and sex with weight and height z scores > -2 SDs of the reference population [9,10]. At two years of age, children born preterm are expected to have caught up on the normal growth curve with their appropriate to gestation age (AGA) counterparts [11][12][13]. However, some studies have reported that at this age some children are still stunted and underweight, and this growth restriction persists into adulthood [14][15][16][17]. Survivors are at risk of long-term consequences and require follow up beyond the neonatal period [18,19]. It is therefore important that early childhood follow up identi es children poorly progressing in growth to ensure that appropriate interventions are instituted at the earliest opportunity [15,20,21].
Compared to high income countries, in low and middle income countries (LMICs) there is limited literature on post-hospital discharge growth status of preterm children [7,8]. We therefore sought to evaluate the catch up growth, nutritional status and associated factors for children aged at 22-38 months born preterm/low birth weight in a low resource setting.

Methods
This was a cross sectional survey which assessed the anthropometric measurements of children aged 22-38 months born preterm/low birth weight (LBW). The study was conducted at the preterm/LBW follow up clinic of Mulago National Referral Hospital, Kampala, Uganda. This clinic attends to any baby born with weight ≤ 2500gms and majority of these are usually pre-term babies. This is a follow up clinic for babies discharged from Mulago hospital neonatal unit. A total of 251 child and caretaker pairs of eligible children were recruited from November 2019 to February 2020. Eligible children were children aged 22-38 months, and weighed ≤ 2000 grams at birth and attended the preterm/LBW follow up clinic at Mulago Hospital. For inclusion in the study, hospital outpatient records were evaluated for children who were aged 22-38months at the time of enrolment (were attending and registered in the follow up clinic between January 2017 to February 2018 and had a birth weight of ≤ 2000 grams). A list of the eligible children was made and every second mother was given a telephone call by the research team and was requested to bring back the child to the clinic for participation in the study. However, using this systematic sampling of every second child, we were only able to reach 108 out of the 506 eligible participants. We therefore consecutively called up every caretaker. A large proportion of telephone calls to the caretakers were not picked after three attempts. There were also some incomplete or wrong telephone numbers and some numbers were not validated or were out of service. A few caretakers/mothers were not able to come back to the clinic despite responding to our call.

Study Measurements
Mother/caretaker-infant pairs came to the clinic on the appointed days. Research assistants who were nurses obtained both infant and maternal demographics (age, sex, birth weight and socioeconomic characteristics). Other data collected included: mode and type of delivery, estimated distance from hospital, duration in hospital, duration of exclusive breastfeeding and duration in the follow up clinic. Anthropometric measurements were taken for each child including: weight, height/length, head circumference and mid-upper arm circumference. Mother's weight, height and body mass index (BMI) were also measured. The weights were taken using a digital portable SECA® weighing scale. Weight was taken while child was wearing light clothing and bare feet. The values were recorded to one decimal point. Height was measured to the nearest 0.1 cm precision using a stadiometer. All measurements were taken once by a trained research assistant.

Analysis And Data Management
The WHO Anthropometrics software was used to convert height, weight and age measurements to heightfor-age z-scores (HAZ), weight-height z-scores (WHZ) and weight-for-age z-scores (WAZ) which were used to classify stunting, wasting and underweight respectively. Child was considered to have caught up or attained normal growth if their z-scores were above − 2SD of the reference population and to have undernutrition when their z-scores below − 2SD. To establish the association of infant and maternal factors with nutritional status for wasting, underweight and stunting, stepwise logistic regression analysis was performed. Factors that were found signi cant at bivariate analysis were further tted in model 1 and 2 at multivariate analysis. We used multivariable logistic regression analysis to estimate adjusted odds ratios (AORs) along with their 95% con dence intervals (CIs) for the associations between each infant and maternal characteristic with child anthropometric outcomes (below − 2 SD) as the reference category. For all statistical tests a p < 0.05 was considered statistically signi cant.
Data were analyzed using STATA version 15 statistical software (StataCorp. 2017 College Station, TX: StataCorp LLC). The severity of stunting, underweight and wasting by prevalence ranges among children were assessed using the WHO classi cation for assessing severity of malnutrition by prevalence ranges among children under-5 years of age [10].

Results
Baseline characteristics of the children Of the 923 children registered in the follow up clinic from January 2017 to February 2018; 506 met our inclusion criteria (Fig. 1). We enrolled 251 children and the mean age was 29.7 months (SD ± 4.5) and 51.4% were male. Children with birth weight ≤ 1500 gm were 104 (41.6%) and those > 1500 were 146 (58.4%). The median duration of hospital stay post-delivery was 12 days (IQR: 7-18).
The summary of the baseline characteristics for the study participants are summarized in Table 1 below.
The ratio of male to female was 0.9:1. Most of the participants were delivered by spontaneous vertex delivery 177 (70.5%) and majority were singletons 179 (71.3%).  The summary of growth status for the participants using anthropometric measurements by sex and age are shown in Fig. 2 and Table 3. Of the participants, 101 (40.2%) had caught up on the normal anthropometric measurements z scores > -2 SD of the reference population for their age and sex.
There were more boys stunted, 64 out of 106 participants compared to girls (p = 0.001 The z-scores for the weight for age, weight for height, height for age, and head circumference for the participants are summarized in Table 3.  After bivariate and multivariate analysis to establish the association between child and maternal characteristics with the nutrition status are summarized in Tables 4 & 5  respectively as shown in Table 4.   [13,14]. Mukhopadhyay et al in India showed no catch up growth beyond 2 years among extreme low birth weight infants [13]. A cross sectional study from 51 nations showed no catch-up growth from 2 to 5 years [17]. This showed a steep decline in children's height for age z-scores from birth to the rst two years of life with no deterioration thereafter. The anthropometric parameters at two years can therefore fairly predict the growth outcomes later in life [17].
We enrolled children who were born preterm/LBW; these are a high risk for undernutrition [22] and this may explain the high prevalence of stunting demonstrated by our results. The prevalence of stunting in the children under-5 years based on a Uganda national survey was 29% [23]. The same survey showed that stunting increased with age, peaking at 37 percent among children 18-35 months [24] in the general population.
In our study, stunting was associated with male sex, lower maternal age and lower maternal height. In our study, the boys 61% were more likely to be stunted compared to the girls 39%. This is similar to other studies [25], although further studies are needed to evaluate the mechanism in which sex may contribute to stunting. Similar to previous studies [26,27], our study demonstrated that young mothers had higher chances of having stunted children compared to older mothers. It is thought that older mothers are usually more knowledgeable in aspects of baby care compared to young and teenage mothers [26] hence the better outcome of their children. The maternal height could most probably be attributed to genetic and environmental factors. Studies have demonstrated mothers with short stature or those born with low birth weight were more likely to give birth to children with the same features [17,25]. Even though factors such as maternal education wealth quintile, maternal BMI have been described to be associated with stunting [24,25,28], our study did not nd these to be associated with stunting.
Underweight was observed among 14.4% of our participants a prevalence slightly higher than the Uganda prevalence of 11% according to UDHS 2016, [23]. Underweight was associated with low birth weight and long hospital stay. In general, children who are born with low birth weight are at high risk of being underweight in their childhood which occasionally persists in adulthood [29]. It is possible that di culties encountered in feeding low birth weight infants majority of whom are preterm babies, contributes to their poor growth in early childhood [30,31] .The nding that long hospital stay in uences the ability of the baby to add weight and reach appropriate z-scores could be a marker of di culties in feeding or generally ill-health which may hinder adequate feeding and growth. Only 3.2% of our study participants were wasted as compared to the 4% in the general population [24]. Wasting is an indicator of acute illness and we did not identify many children in whom recent acute illness was reported and this probably led to low prevalence.
Our ndings have several implications on growth monitoring of children born preterm which should not be limited to the neonatal period but extend to childhood and beyond. Child growth and nutritional status may be strongly linked to fetal life suggesting a need for interventional focus on nutrition during pregnancy and early childhood. To end all forms of malnutrition by 2030: Sustainable Development Goal 2 [32], a life-course approach of nutritional interventions are needed targeting at risk populations. These will in turn lead to child survival, educational achievements and overall well-being later in life. Secondly, collection and analysis of long-term data in former preterm and LBW children linked to nutritional strategies and growth parameters are still strongly recommended in our setting. Although pre-pregnancy and natal nutritional status was not assessed in our study, they have been linked to growth failures in early childhood in other studies [33]. It is equally important to mind the mother's nutritional status from pre-conception throughout pregnancy [17,34].
The strength of this study is we enrolled children who weighed less than 2000 grams at birth majority of whom will also be preterm. The predominant mode of feeding in our study population was exclusive breastfeeding with no modi cation for all the participants. The limitations included the cross sectional data does not re ect changes in growth of individual children overtime. Details of dietary intake and complimentary feeding practices were not collected in our study to minimize long term recall bias. Nevertheless, the results clearly indicate a need for rigorous growth monitoring for children born preterm/LBW beyond the neonatal period.

Conclusion
Of the children born preterm/LBW 40.1% of the participants had attained the expected catch up growth at 2 to 3 years of age, 42.2% were observed to be stunted and 14.4% were underweight both higher than the national levels. Targeted interventions are speci cally needed for children born with very low birth weight, those requiring long postnatal hospitalization, males and those born to short or young mothers.