Predictors of low birth weight and 24-hour perinatal outcomes at Muhimbili National Hospital in Dar es Salaam, Tanzania: a five-year retrospective analysis of obstetric records

Introduction The global prevalence of low birth weight (LBW) is 16%, representing more than 20 million infants worldwide, of which 96% are born in low-income countries. This study aimed to determine the prevalence, predictors and perinatal outcomes of LBW newborns. Methods We conducted a retrospective analysis of data obtained from the hospital's obstetric and neonatal database. Descriptive statistics and multivariate logistic regression were performed with 95% confidence intervals (CI). Results The prevalence of LBW was 21% (n = 8,011) and two-thirds of these were delivered at term. Seven percent of newborns were stillbirths and 2% died within 24hrs after birth. Logistic regression revealed that primigravida and grand multiparity were associated with LBW (OR: 1.25, 95%CI: 1.15-1.37; and OR: 1.21, 95%CI: 1.01-1.25, respectively). Having <4 antenatal care (ANC) visits was associated with increased odds of LBW (OR: 1.74, 95%CI: 1.59-1.87). Regression models revealed an independent association between LBW and increased odds of stillbirths (OR = 7.20, 95%CI 6.71-7.90), low Apgar score (OR = 3.42, 95%CI: 3.12-3.76) and early neonatal deaths (OR = 1.82, 95%CI: 1.51-2.19). Conclusion The prevalence of LBW was high and was associated with extreme maternal age groups, grand multiparity, low maternal education, low number of ANC visits and obstetrics risks factors and complications. Both LBW and prematurity were independently associated with poor perinatal outcome. Future interventions should focus on improving the quality of ANC and integrating peripartum emergency obstetric and neonatal care.


Introduction
The World Health Organization (WHO) defines low birth weight (LBW) as a newborn weighing is less than 2,500 grams, with the measurement taken within the first hours of life, before significant postnatal weight loss has occurred [1]. Globally, the prevalence of LBW is 14% and, according to the WHO and the World Bank, more than 96% of LBW newborns are born in developing countries [1,2].
Hospital-based studies in Sub-Saharan Africa indicated a wide variation in inter-country prevalence of LBW, from 9% in Rwanda [3,4] to 31% in Sudan [2]. In Tanzania, despite regional variation and an increase in population from 34.4 million in 2002 to 44.9 million in 2012, the prevalence of LBW has been in decline; from 13% to 7% over nearly the same period (from 2000 to 2010) [5].
LBW has been strongly associated with poor health outcomes during the neonatal period, infancy and adulthood [2,6], regardless of race and socioeconomic status [7,8]. For example: a study conducted in the USA investigating births among different races showed that LBW substantially increased the risk of stillbirth, irrespective of race [8].
Furthermore, approximately every ten seconds, an infant in a developing country dies as a result of a disease or infection that can be attributed to LBW [7]. A recent epidemiological study [9] addressed the association between LBW and medical conditions, including metabolic syndrome (a cluster of conditions that include high blood pressure, high blood sugar, excessive body fat around the waist) and abnormal cholesterol or triglyceride levels. A metaanalysis of the literature also found that LBW is associated with an increased risk of heart disease, stroke and diabetes in adulthood [10,11].
Having LBW imposes a 20-times higher likelihood of death during infancy compared to newborns who had normal birth weight [1]; therefore, LBW is associated with high infant mortality and low life expectancy and also predicts prevalence of disability and educational achievement in childhood and adulthood [9,10]. As a health indicator, LBW has been used as a good summary measure to indicate a wide range of multifaceted public health problems that include health inequality, long-term maternal malnutrition, ill health, hard physical work and poor care during pregnancy [9]. Female newborns and first-born infants are generally lighter than their counterparts [12]. Evidence from various studies and the WHO report highlighted modifiable and non-modifiable predictors of risk factors for LBW, including female sex, primiparity, multiple pregnancies, low socio-economic status (SES), nutritional insufficiency and short stature [12][13][14]. Other identified factors were higher education, single motherhood, maternal complications and lack of antenatal care (ANC) [6]. The most recent national estimates are drawn from TDHS 2010 [5] as the 2015 full report has not yet been released [15]. The absence of recent estimates of LBW hinders adequate assessment of the demands and appropriate allocation of health resource for managing LBW newborns in hospital settings.
Local needs assessment enables the successful implementation of interventions for achieving Sustainable Development Goals that are essential for a healthy start for a newborn, from the pregnancy period to safe childbirth. Prevention and adequate management of LBW also make an important contribution to the third Sustainable Development Goal (SDG 3), which focuses on the reduction of neonatal and under-5 mortality [16]. Therefore, this study aimed to determine the prevalence and predictors of LBW and its 24- weeks. Varying management protocols for mothers and babies have been described in previous papers in this setting [18]. Qualified data entry clerks have been digitally entering data, including maternal and neonatal outcomes, prospectively from midwifery books since Page number not for citation purposes 3 1998 [18,19]. Maternal and neonatal diagnoses are based on doctors' records in the case notes. Neonatal deaths reflect only babies who were born at MNH and were admitted to the neonatal unit.

Study variables:
The variables included were: maternal age in complete years; education level; marital status; ANC attendance; and pregnancy and delivery complications, all of which were taken from their ANC card on admission. Others were: parity; gestational age; and area of residence. Gestational age was estimated from Naegele's rule of last normal menstrual period (LNMP) and a first trimester ultrasound. Gestational age was then dichotomised into preterm and term. Birth weight was measured within the first hour of life and dichotomised as LBW and normal birth weight, according to the WHO definitions (if less than 2,500 grams). In this paper we considered LBW and normal birth weight against various predictors, while for gestational age we added a regression model to control for its potential confounding effect. Outcome measures were birth weight, stillbirth (a baby born with no signs of life at or after 28 weeks' gestation) and the status of the baby after 24 hours, recorded as normal, admitted to neonatal unit, or dead.

Maternal-related characteristics:
We analysed 39,099 deliveries that met the inclusion criteria. The majority (81.6%) of the participants were aged between 20-35 years. Ninety-seven percent of women were ever-married (Table 1). One-third of the mothers attended ANC less than four times.
Birth outcomes: The median birth weight was 3,000g, ranging from 700g to 5,900g and the mean (SD) was 2,940g (702g). The prevalence of LBW was 21% (n = 8,011). The prevalence of stillbirths (both fresh and macerated) and very early neonatal deaths was 6.9% and 1.8%, respectively, of the live births who later died, making the total number of deaths about 9%.

Mode of delivery: Caesarean section (CS) delivery was performed
in 55% (n = 21,504) of all included deliveries, and one-fifth of these were performed electively. Premature deliveries accounted for only 3.9% and 2.4% of all CS and elective deliveries, respectively.

Birth weights, gestational age and perinatal outcomes:
Gestational age (GA) at delivery ranged from 28 to 43 weeks with the mean and median GA of 38 and 39 weeks, respectively. About 93.6% were term, 6.4% were preterm, and less than 0.1% were post-term deliveries. Of the term pregnancies, 15.0% were small for gestational age (SGA), sometimes referred to as intrauterine growth restriction (IUGR), 82.4% were appropriate for gestational age (AGA) and only 2.6% were large for gestational age (LGA). In both term and preterm categories, LBW was associated with adverse perinatal outcomes, although the rates were much higher among preterm deliveries, as shown in Table 2 Table 4. Separate regression analysis showed that LBW among preterm deliveries was associated with low Apgar score at the first and fifth minute but not stillbirth and early neonatal death (as the rates were higher regardless of birth weight), whereas LBW was associated with increased likelihood of adverse perinatal outcomes among the term deliveries, as shown in Table 5 below.

Discussion
One out of five newborns at MNH had LBW. There was also a high incidence of stillbirths and early neonatal deaths within 24 hours.
LBW babies were more likely to die before birth, during labour and within 24 hours as compared to normal-weight babies. In this study, LBW was more prevalent in those mothers who lived in semi-urban areas. The urban residence could be a reflection (proxy indicator) of good economic status of the mothers, which is associated with normal birth weight, reflecting good maternal nutrition and wellbeing, both of which have been reported to lead to good birth weights [13]. Also, those residing in urban settings are more likely to have good physical accessibility to the health facilities affecting ANC attendances [20,21]. As reported in other studies [6,13,14], LBW was associated with women who had a lower level of education. This may be explained by an increased awareness among educated women in relation to available health services and their increased likelihood to attend ANC earlier and more frequently [21]. Women who attended less than 4 ANC visits had increased  [26].
Teenage pregnancies are also more likely to be primigravida, whereby the first pregnancy has higher odds of the mother delivering an LBW baby compared to subsequent pregnancies [21].
The findings from this study show that single mothers gave birth to more LBW babies than married mothers. Economic and psychosocial support could be a key factor contributing to good maternal health and birth outcomes [6]. In this study, mothers who had anaemia during pregnancy had nearly three times as high odds (OR: 2.9) of having LBW babies as those reported by Oladeinde et al in Nigeria [21]. As reported in other studies, anaemia is a proxy indicator of poor nutrition during pregnancy, which gives rise to IUGR [12,27].
In the regression model we found no effect of HIV infection on LBW, unlike other reports [22,28]. This could be due to a long-term

Conclusion
The prevalence of LBW was high and was independently associated

Competing interests
The authors declare no competig interests.

Authors' contributions
Benjamin All authors agreed on the final manuscript.

Acknowledgments
We acknowledge the MNH administration for granting permission to use the hospital data for this study. We also wish to acknowledge the work of the nurse midwives in recording details of all deliveries in the midwifery registry, which were then entered into the perinatal database.