Determinants of Low birthweight and Preterm Delivery in the Volta Region of Ghana: Evidence from birth records

Low birthweight and preterm delivery are birth outcomes that can predict the survival, development and long-term health outcomes of newborns. Though advances in medical research have improved birth outcomes these birth outcomes still remain issues of public health concern. This study assessed the determinants of low birthweight and preterm delivery in the Volta region of Ghana. This retrospective cross-sectional study analyzed data from 680 birth records of all deliveries between October and December, 2018 the in the Ghana. Univariate and multivariate logistic regression models were used to predict determinants of low birthweight and preterm delivery using Stata MP/16.0.


Background
Birthweight is an important variable which in uences the survival and development of newborns. It refers to the rst weight of a neonate taken in the rst few hours of its birth. Therefore, normal birthweight is crucial for neonatal survival, optimal child development and a healthier life in adulthood (1). The World Health Organization (WHO) stipulates that the normal weight of a baby that reaches full term is between 2.7 and 4.1 kilograms with an average weight of 3.5 kilograms. In that regard, the organization de nes low birthweight (LBW) as weight at birth of less than 2.5 kilograms (2). This classi cation was in uenced by epidemiological data that newborns who weigh less than 2.5 kilograms are at increased risk of neonatal mortality compared to their counterparts who weigh heavier (3).
Preterm delivery (PTD) is a term used to describe babies born alive before 37 weeks of pregnancy are completed (4). Preterm babies are at a greater risk of neonatal infections and as such may require more complex care resulting in prolonged hospitalizations, increased costs and high mortality (5). According to WHO, the main reason for low birthweight is preterm delivery though its aetiology still remains a mystery.
However, some researchers argue morbidities and infections such as malaria, hypertension, syphilis and HIV can cause preterm delivery (6,7). Others argue preterm delivery is caused by multiple aetiologies such as individual and environmental factors which makes its prediction and prevention a di cult process during antenatal care (8,9).
Though advances in medical research have improved birth outcomes (10), these birth outcomes still remain issues of public health concern particularly in low and middle income countries as major contributors to morbidity and mortality during neonatal, infancy and childhood stages (11). Different studies have reported different determinants of low birthweight. For instance, a survey in China identi ed young maternal age, educational level, a history of adverse pregnancy outcomes and maternal morbidities such as hypertensive disorders and gestational diabetes to be associated with low birthweight (12). Several risk factors have also been reported to in uence PTD. These include age, socioeconomic status, gestational diabetes, pre-eclampsia and foetal distress (13,14).
In Sub-Saharan Africa, maternal malaria and HIV infections are major determinants of adverse birth outcomes. In Nigeria, a study found that 14.1% of babies born to the HIV positive women had LBW compared to 1.0% in women with no HIV infection (15). A randomized controlled trial in Malawi reported preterm babies were born to 36.4% of women who had malaria compared to 28.5% without malaria (6).
With regard to maternal age as a determinant of these birth outcomes, there are inconsistent results. For instance some studies have reported that risk of delivering low birthweight babies among teenage mothers is higher than that of older women (16,17). However, other studies older mothers are more likely to experience these birth outcomes (18,19).
In Ghana, the 2014 Ghana Demographic and Health Survey (GDHS) reported 10% of newborns in Ghana had LBW (20). However, the prevalence of LBW in parts of the country are higher. For instance, a prevalence of 26% was reported in a study in Northern Ghana in 2015 (21) and 21% in Ashanti region in 2013 (22). There is no national data on the prevalence of PTD. That notwithstanding, some studies have been conducted in the country to identify determinants of LBW and PTD. A study in the Greater Accra region found premature rapture of membrane and pre-eclampsia/eclampsia were associated with increased risk of preterm delivery while four or more antenatal care visits was protective (23). A similar study on LBW reported anaemia, preterm delivery, education and not taking iron supplements during pregnancy to be signi cantly associated with LBW (11).
The Volta region of Ghana is reported to have above average health indicators regarding antenatal care and hospital-based deliveries. This is corroborated by ndings from the 2014 GDHS that suggested the region improved health indicators compared to other regions (20). However, the region is known to have one of the highest prevalences of teenage pregnancy which is a known risk factor for LBW and PTD (24). Based on the different proportions of low birthweight reported by previous research and dearth of research on this subject matter in the region, it is important to study the determinants of these birth outcomes in the region.
Adequate knowledge on determinants of these birth outcomes is crucial for identifying them and providing appropriate care and attention to pregnant women at risk. However, in the Volta region, little is known about determinants that predispose a pregnant woman to deliver low birthweight or preterm babies. Consequently, identifying these determinants will considerably contribute to current efforts to address these issues of public health concern which can jeopardize the future of newborns. In that regard, a decline in the incidence of low birthweight and preterm babies will signi cantly reduce costs associated with catering for such babies, lessen the burden on the health system and reduce the occurrence of neonatal and childhood morbidity and mortality in the region and the country as a whole. This study assessed the determinants of low birthweight and preterm delivery in the Volta region of Ghana.

Study site and design
We conducted a retrospective cross-sectional study at the Ho Teaching Hospital (HTH) in the Volta region of Ghana which serves as the major referral hospital in the region. The hospital was established in December, 2000 and was upgraded to a teaching hospital in April, 2019. HTH is located in the Ho Municipality, one of the 17 districts in the Volta region. The Ho Municipality has a total land area of 2,361 square kilometers. The Municipality has about 49 health facilities that provide health services to Municipality population of 177,281. The 300-bed capacity hospital is strategically located to render specialised health services to indigenes of the Volta region and beyond. The hospital is also patronized by clients from the Republics of Togo and Benin as well as the Federal Republic of Nigeria. The HTH has approximately 103,964 annual Outpatient attendance and provides about forty-one (41) essential services including maternal and child health services. Pregnancy-related complications, anaemia and malaria remain the top three causes of all hospital admissions (25). Professionally trained nurses and midwives at the maternity unit of the HTH record pregnancy and birth outcome information of expectant mothers as well as demographic information of these mothers in delivery registers. This study examined birth records of all mothers who delivered live babies at the HTH between October and December, 2018 to identify determinants of low birthweight and preterm delivery.

Data Extraction And Sample Size
A pre-tested data extraction sheet was used to extract data about maternal and newborn characteristics from delivery registers. Four trained midwives extracted the data from the registers. These midwives were oriented on the eligibility criteria for inclusion into the study as well as the data extraction process. Data collection was supervised daily to ensure consistency, completeness and accuracy of the extracted data. All live births within the study period were considered for this study. However, analysis was conducted on 680 birth records after stillbirths, multiple births, babies born with congenital abnormalities and birth records with missing information were excluded. Birth records of babies born with congenital abnormalities and multiple births were excluded based on the assumption that they had different risk and aetiology for low birthweight and preterm delivery. Stillbirths were also excluded because they were few and majority had missing information on birth weight and gestational age.

Study Variables
The outcome variables of this study were low birthweight and preterm delivery. Birthweight was dichotomized as a binary variable; 0 "normal weight" when birthweight ≥ 2.5 kilograms and 1 "LBW" when birthweight was < 2.5 kilograms. Similarly, gestational age was also dichotomized into: 0 "carried to term" when gestational age at delivery was ≥ 37 weeks and "PTD" when gestational age at delivery was < 36 weeks.

Data Management And Analysis
The data extraction sheets were checked for accuracy and completeness before passed on for data entry.
Data entry was done using EpiData Data Entry Client (v4.4.3.1) and then exported to STATA MP/16.0 (College Station, Texas, USA) for analysis. Descriptive statistics was used for frequencies. Percentages were reported for categorical variables. Means and standard deviations were also determined for continuous variables. Univariate and multivariate logistic regression analyses were performed for outcome variables. Strengths of associations between outcome variables and explanatory variables were determined using crude odds ratios. Explanatory variables with p-values < 0.05 in the univariate analysis were considered for a stepwise multivariate logistic regression model. Adjusted odds ratios and their con dence intervals were then computed with p-values < 0.05 considered as statistically signi cant.

Socio-demographic characteristics of women who delivered
Six hundred and eighty (680) women were involved in this study with a mean age of 27.5 ± 6.18. Women aged 20-34 years formed the majority of the participants. Few women 96(14.1%) had no formal education while the remaining 85.9% had varying forms of education ranging from primary school 82(12.1%) to tertiary education 116(17.1%). More 431(63.4%) women were employed in the private sector and more than half 445(65.4%) were married. The dominant religion was Christianity 603(88.7%) while Ewe was the dominant tribe (Table 1).

Discussion
This study sought to identify the determinants of LBW and PTD among women in the Volta Region of Ghana. Overall, the study found the prevalence of LBW to be 12.9%. This proportion is higher than the 10% national prevalence which was reported in the 2014 by GDHS. The reported prevalence in this study is also higher than the 6% recorded for the Volta region in the same 2014 by DHS (20). The prevalence of LBW in this study is also higher than what was reported in the United Arab Emirates. That study reported a prevalence of 9.4% (27). This could be attributed to geographical differences in study sites.
This study found that women aged 20-34 years were less likely to have LBW babies compared to those aged less than 20 years. This is consistent with data published by Althabe and colleagues, Alemu, and Taha and friends These studies reported teenage mothers have an increased risk of delivering LBW babies compared to older women (16,17,27). However, this nding was inconsistent with those found by (19) and (18) who reported older women have an increased risk of LBW as compared to younger women.
These results still point to the fact that inconsistent results still exist about maternal age and adverse birth outcomes, particularly LBW. Teenage mothers are most likely to be rst timers with little or no experience with management of pregnancies. This could be a contributory factor to why women aged 20-34 years had less odds of LBW compared to teenage mothers in this study. Additionally, teenage mothers may not be physically and emotionally mature. Thus, their bodies may be unable to deal with the stress of pregnancy (24). Coupled with this, good maternal nutrition, socio-economic status and adequate ANC attendance could have made the 20-34-year-old women less likely to experience LBW.
Low birthweight babies were less likely to be born to multiparous women. This is consistent with a recent study conducted in India that reported that increased parity of a mother increased the mean birthweight of babies (28). Similarly, another study in Bangladesh also found that increasing parity increases birthweight leading to reduction in the occurrence of LBW (29). A plausible explanation for this observation might be that increased parity might lead to increased experience with pregnancy and childcare, ANC attendance, nutritional status and health seeking behaviour. However, this nding was incongruent with a similar study conducted in the Brong-Ahafo region of Ghana. That literature suggested that increasing parity signi cantly increased the odds of low birthweight (30). Another study in Ethiopia reported similar ndings to those of Mohammed and colleagues (31).
In this study, the odds of delivering a LBW baby was signi cantly high among women who delivered their babies through CS compared to those with vaginal deliveries. This nding resonates with a studies conducted in the United Arab Emirates (27) and China (32). Some studies have reported an epidemic of CS which these studies have found to increase adverse birth outcomes such as LBW and preterm delivery (33,34). A plausible explanation for this nding in this study could be attributed to the abuse of planned CS. A phenomenon which has been documented in an earlier study in Brazil where it was reported that CS was wrongfully associated with LBW particularly among private hospitals (35). In that regard, there is the need to adhere to WHO's recommendations that CS birth should not be planned before 39 completed weeks of gestation unless it is medically indicated for the bene t of either the foetus or mother or both (36). There are inconsistent results regarding the association between LBW and CS. This is because some studies have reported that CS is protective against low birth weight (37,38) while others have shown that it increases the likelihood of LBW (39) which is similar to the current ndings.
The WHO recommends pregnant women take three or more doses of Sulphadoxine Pyrimethamine for intermittent prevention of malaria in pregnancy (SP/IPTp) in moderate to high malaria transmission areas (40). We found that more than 3 doses of SP/IPTp signi cantly reduced the odds of LBW. This is in conformity with several studies conducted in Tanzania (41,42), Cameroon (43), Nigeria (44) and Ghana (45). The protective nature of SP against LBW could be explained by its therapeutic effect against both malaria and non-malaria infections. This is supported by evidence from a Zambian study which reported that the bacterial and parasitic effects of SP signi cantly improved the birthweight of neonates born to women who took more doses of SP during pregnancy. The Sulphadoxine component of SP provides a broad spectrum of anti-parasitic and bacterial activities (46). Thus, constant exposure via monthly update of SP could reduce microbial density and immunological reactions leading to adverse birth outcomes such as LBW (46)(47)(48).
Our ndings further indicate that the likelihood of LBW was signi cantly higher among hypertensive women compared to their normotensive counterparts. This is consistent with literature in China (49), Ethiopia (50), Brazil (51) and Haiti (52). Some studies have linked the association between pregnancy induced hypertension and LBW to intrauterine growth restriction as a result of the placenta not receiving enough nutrients. This occurs as a result of poor perfusion of blood containing nutrients via the placenta.
The placenta provides blood and essential nutrients from the mother to the foetus for optimal growth and development (50). Thus, pregnancy induced hypertension increases the risk of poor foetal nutrition hence poor foetal growth leading to LBW (49,53).
The current study also identi ed determinants of PTD and found that women who took more than three doses of SP had reduced odds of having preterm birth. This resonates with data published in an earlier study in Northern Ghana where it was reported that high uptake was signi cantly associated with delivery at term (54). The uptake of more doses of SP is known to reduce prevalence and intensity of placenta malaria as well as placenta parasitemia which are signi cant risk factors for preterm delivery (44,55,56). This nding provides useful information on the effectiveness of SP particularly in malaria endemic settings. Additionally, some researchers have reported that SP may have some secondary effects on bacterial and fungal infections which promotes maternal and foetal health thereby reducing the occurrence of preterm delivery (57)(58)(59).
The importance of Antenatal care in the prevention, detection and treatment of pregnancy-related conditions cannot be underestimated. With this, WHO recommends all pregnant women go for their rst ANC visits in the rst trimester of their pregnancy. This recommendation will allow for early diagnosis and management of health conditions as well as identifying risk factors which can negatively affect the progress and outcomes of pregnancy (60). Our study found that women with more than four ANC visits had reduced odds of delivering preterm. In Ghana, ANC services assessed by pregnant women are in consonance with WHO's recommendations including counselling on healthy diet and good nutrition, tobacco and substance use and physical activity, HIV and malaria prevention, tetanus vaccination, foetal measurements and advice for dealing with physiological pregnancy symptoms such as nausea, back pain and constipation (60). This comprehensive ANC package could have helped identify high risk pregnancies among the women with more ANC visits. Subsequently, measures would have been put in place to reduce these high-risk pregnancies and thus the lower risk of preterm delivery. This assertion is corroborated by different studies by Turienzo and Cunningham who emphasized on how the content and type of ANC packages help reduce adverse birth outcomes (61,62).
We found that babies born through caesarean section were twice as likely to be born preterm compared to those born through vaginal delivery. One plausible explanation for this could be that preterm babies were delivered through CS due to foetal compromises such as foetal distress (63). It could also mean that the CS was planned due to history of previous CS or as a result of pregnancy induced hypertension. There are inconsistent results on the impact of CS on preterm delivery. This is because, some earlier studies reported CS improves the outcomes of preterm babies (64,65), while others suggest vaginal birth is protective against preterm delivery (66).
A major limitation to this study is that, data collected for routine healthcare services and not primarily for research was used for analysis. With this, there is possibility of errors occurring during documentation.
However, routine healthcare services data can be used for planning, monitoring and evaluation of public health interventions.

Conclusion
Evidence from this study indicates that maternal age, parity, number of ANC visits, hypertension, SP/IPTp and caesarean section were independent determinants of LBW and PTD. These ndings add up to literature on the determinants of these adverse birth outcomes particularly in resource-limited environments. Furthermore, this study could serve as a foundation for further research in the study area and for developing public health interventions geared towards reducing the risk and complications of these birth outcomes. The study only began after ethical approval had been given by the REC. Permission was also sought from the Volta Regional Health Directorate (VRHD) and the medical director of HTH. This study involved the extraction of secondary data and thus no informed consent was used in this study. The information obtained from this study was kept con dential. Personal information and information on personal practices were anonymized during data analysis and dissemination ndings from the study. Data management, storage, analysis and reporting was done using codes which could not expose the details of participants. Information about participants was also stored in a le with a code. This le was accessible only to the investigator.

Consent for publication
Not applicable Availability of data and materials All data and materials are available upon reasonable request from the corresponding author.

Competing interests
The authors declare no competing interests.

Funding
This study received no funding.
Authors' contributions WKA, FNB and MK conceived the study. WKA, FNB and MK did the data analysis and wrote the methods section. WKA, FNB, and MK, were responsible for the initial draft of the manuscript. All authors reviewed and approved the nal version of the manuscript.

Figure 1
Prevalence of low birthweight and preterm delivery