Preterm birth determinants among newborns at six public hospitals in Northeast Amhara, Ethiopia: unmatched case-control study

Background Preterm birth is a global public health problem. Worldwide, every year about 15 million newborns are born preterm and about 1.1 million of them die due to complications related to prematurity. More than 81% of preterm births were in Asia and Sub-Saharan Africa. Nowadays, preterm birth is the rst leading cause of under-ve mortality and is one of the limiting factors for achieving a reduction in under-ve mortality rate. Therefore, study identify the of Method Hospital-based unmatched case-control study was conducted at public hospitals in Northeast Amhara, Ethiopia from March 15 th to April 25 th , 2019. Data were collected using interviewer-administered structured pre-tested tool and document review for cases and controls. Data were entered in to Epi Info version 7.2.2.6 and exported to SPSS version 23 software for data analysis. Bi-variable and multivariable binary logistic regression were used to measure the association between preterm birth and risk factors. Adjusted Odds ratio (AOR) at 95% condence interval (CI) was calculated and P value ≤ 0.05 declared statistically signicant.

found in Sub-Saharan African except Pakistan and Indonesia. Besides, poor countries have a higher prevalence of preterm births than rich countries (4). In Ethiopia, the overall PTB rate is 12% and the prevalence of PTB in different parts of the country ranges from 4.4-25.9% (2,5,6) For the rst time in history, PTB is the leading cause of death in children younger than 5 years worldwide covering 17% of under-ve and 35% of neonatal deaths (7)(8)(9). Premature infants are particularly vulnerable to complications due to impaired respiration, di culty in feeding, poor body temperature regulation and high risk of infection. Babies who are born before they are physically ready to face the world encounter greater risks of serious health problems, including cerebral palsy, intellectual impairment, chronic lung disease, and vision and hearing loss (4,10). Premature babies also face drastically different odds of death depending on where they are born. From those extremely preterm babies (< 28 weeks) born in developing countries, greater than 90% of them die, while greater than 90% of extremely preterm babies in developed countries survive (4).
Among the 10 countries with the highest numbers of babies dying from PTB complications, India is the rst followed by Nigeria. Ethiopia takes the 8th place with 24,400 child deaths per year (9). Different interventions have been implemented to prevent and improve the survival of PTB which demands huge nancial capacity imposing a huge economic burden for the family, the community and the nation at large (4,11).
Every Women Every Child initiative set a target for 50% reduction in PTB deaths by 2025 which needs governments and the global community to take direct action to address deaths due to preterm birth, otherwise, the global total of preterm deaths will not reduce signi cantly, with around 900,000 premature babies continuing to die every year by 2025 (12).
Several risk factors for PTB have been identi ed across the world. The potential risk factors for PTB includes, but not limited, maternal risk factors like past adverse birth outcomes(PTB, stillbirth, LBW), obstetric complications(PIH, PROM), maternal infections(HIV, malaria, UTI) and no ANC follow up while having ≥ 4 ANC checkup, IPI 24-36 months, maternal age 20-35 were found to be protective for PTB (13)(14)(15). However, risk factors for PTB are not always global; risk assessment should be carried out progressively at local, regional, and then global levels(16).
In poorer countries, like Ethiopia, where fewer resources are destined for healthcare, the direct, indirect, immediate and later repercussions of prematurity constitute a relevant public health issue. However, the great majority of scienti c evidence accumulated on this subject is generated in developed countries. Besides, the potential risk factors are varied in different locations. Therefore, this study will have its own contribution to avert PTB and its consequences by identifying risk factors.

Method
Study design and period A facility-based unmatched case-control study was conducted at public hospitals in Northeast Amhara, Ethiopia from March 15th to April 25th, 2019.

Study Area
The study was conducted in Northeast part of Amhara region, Ethiopia. The study was employed in six randomly selected governmental hospitals (Debre Birhan Referral hospital, Tarmaber Primary Hospital, Kemisie General Hospital, Dessie Referral Hospital, Borumeda Hospital, and Woldia general Hospital).

Source Population
All newborns who were born in public hospitals of Northeast Amhara, Ethiopia.

Study Population
All newborns who were born in selected public hospitals of Northeast Amhara, during data collection time.

Inclusion And Exclusion Criteria
Inclusion criteria Cases All alive singleton preterm newborns (< 37 completed weeks of gestation) who was born in study hospitals of northeast Amhara.

Controls
All alive singleton term newborns (37 completed weeks to 42 weeks) at the same hospitals.

Exclusion Criteria
Newborns whose mothers did not remember their LMP.
Newborns whose mothers had a serious illness or with di culty of communication during the study period.

Sample Size Determination
Sample size was determined by using EPI info version 7.2.2.6 by a double-population proportion difference formula. Different variables that have a high contribution for PTB was used and the variable that yields the highest sample size was selected. The sample size was calculated considering the following assumptions: r = 3 (ratio of controls to cases), 80% power and 95% con dence level. ANC < 4 with P Control = 42%, P Case = 60.7% was selected. Based on this, the sample size was 323. By considering a 10% non-response rate, the nal sample size was 356 with 89 cases and 267 controls.

Sampling Technique
The total sample size was allocated proportionately according to their caseloads for selected hospitals. Cases were preterm newborns (gestational age < 37 completed weeks) and controls were term newborns (gestational age of ≥ 37 weeks to 42 weeks). Systematic sampling technique was employed to select controls and cases every fth intervals and every two intervals respectively. The rst participant was selected by lottery method. Passive smoking living with someone who smoked during the mother's pregnancy

Data Collection Tools And Procedures
Data was collected using interviewer-administered, structured pretested questionnaires and record review. The questionnaires include data on socio-demographic characteristics, obstetrics characteristics, infant characteristics, maternal anthropometrics, medical disorders, maternal infections, and behavioral factors. The questionnaires were prepared by reviewing different kinds of literature from related studies and EDHS 2016, (14,15,(21)(22)(23). Con rmation of Gestational age was taken from documentation and for nonrecorded data, information was taken from mother's report of LNMP. The data was collected by six trained midwifery graduates and supervised by MSc nurses.
MUAC was measured by using exible non-stretched tape taken at mid-point of left upper arm at relaxed position, without any clothing and with optimal tape tension between acromion process on shoulder blade and tip of olecranon process of ulna. Height was also measured using height Stadiometer with sliding head bar without shoes, standing erect, with shoulder blades, buttocks and heels touching the measuring board, looking straight ahead, shoulders relaxed, arms at sides, legs straight and knees together, feet at and with heels together. Two height measurements were taken and average was used if readings are within 1 cm each other and record to the nearest 1 cm.

Data Quality Control
Data collectors and supervisor were trained for half a day on objective and methods of the study, study tools, how to approach the patient, and take anthropometric measures. In order to assess the validity and reliability of the tool and clarity of the variable, pre-test was done on 5% of the sample size newborns who were delivered out of the study hospitals two weeks ahead of data collection. During the data collection process, regular monitoring and supervision of the overall activity was done by the supervisors and principal investigator to check for completeness and ensure the quality of data. Besides, data cleaning and coding was carried out before analysis. The measurement tool was prepared in English and translated into local language (Amharic).

Data Processing And Analysis
The collected data were checked for completeness and entered into EPI info version 7.2.2.6 and exported to SPSS version 23 for analysis. Data were summarized using text, tables, and graphs. Model goodness of t test was checked by Hosmer-Lemeshow test (P-value = 0.224) and tted. Multicollinearity between independent variables was assessed by using VIF (1.1-1.27) and no problem detected. Binary logistic regression model was used to measure the association between preterm birth and risk factors. Variables that had P-value ≤ 0.2 on bi-variable regression analysis and had association with preterm birth in previous studies were tted into the multivariable logistic regression model to assess their association with the outcome variable. Adjusted Odds ratio (AOR) with 95% con dence interval (CI) was calculated and P value ≤ 0.05 declared statistically signi cant.

Socio-demographic characteristics of respondents
In this study, 356 participants were included. From these participants, 89 were preterm newborns (cases) and 267 were term newborns as controls with 100% response rate. The median age of mothers was 25

Determinants Associated With Preterm Birth
In bi-variable logistic regression analysis at 95% level of con dence, ANC follow up, pregnancy status, PIH, PROM, Congenital malformation, Iron/folic acid supplementation, Anemia and hemoglobin level during current pregnancy, Height of mothers, and MUAC were statistically signi cant with PTB. In multivariable logistic regression analysis at P-value ≤ 0.05, ANC follow up, PIH, Congenital malformation, Iron/folic acid consumption, and Anemia during current pregnancy remained signi cantly associated determinants of PTB.
The chance of being preterm among newborns whose mothers had had fewer than four ANC follow up was about 2 times more than newborns whose mothers had had four or more ANC follow

Discussion
This study was intended to identify determinants for PTB in order to tackle the burden of the disease and its associated problems. The study has tried to look at the determinants of PTB by incorporating numerous risk factors as possible.
This study revealed that newborns whose mothers followed ANC less than four times were about two times more likely to be preterm as compared to their counterparts. This nding is consistent with facilitybased case-control studies done in Ethiopia; Jimma (AOR: 4.8), and Central Tigray (AOR: 2.15), North-Eastern Tanzania (AOR 3.25), and Nigeria (AOR 6.13). This might be due to that lack of/ inadequate ANC follow up decreases the chance of identifying risks of PTB and providing appropriate interventions for its prevention (14,21).
In this study, current PIH showed statically signi cant association with PTB. PIH increased the odds of PTB by four and half folds in newborns of mothers having PIH than newborns of mothers without PIH. Similarly, signi cant associations have been reported from studies done in Ethiopia, Gondar (AOR = 5.36), Jimma (AOR 6), and Central Tigray (AOR 3.24), and other studies from Kenya (7.81), and Iran (AOR 7.30). This might be due to hypertension increases the resistance of blood to uterine vessels reducing the uteroplacental blood ow that may affect the exchange of nutrients and oxygen between mother and fetus which in turn results intrauterine growth restriction that causes PTB. Additionally, complications of PIH may cause vascular damage to the placenta, which induces the oxytocin receptors, leads to preterm labor and delivery (14,(24)(25)(26). Even though this may not be causal in nature, PIH is a common reason for indicated preterm deliveries and this may explain its association with PTB.
This study also showed that anemia during current pregnancy were another independent determinant for PTB. Newborns whose mothers had anemia during current pregnancy were at increased risk to be born preterm compared to those newborns of mothers without anemia. This nding is in line with the studies conducted in Ethiopia, Jimma (AOR 2.7), Systematic review and meta-analysis in low-and middle-income countries (RR: 1.63) and Korea (AOR 1.53). This might be due to that anemia and iron de ciency by causing hypoxia and maternal infections induce maternal and fetal stress which leads to the formation of corticotrophin-releasing hormone. Excess corticotrophin-releasing hormone is the major risk factor for preterm labor, PIH, PROM and inhibits fetal growth by stimulating cortisol production. Another possible explanation is that iron de ciency may cause oxidative damage to the erythrocytes and the feto-placental unit affecting placental function (27,28).
The present study revealed that the chance of being preterm among neonates having birth defect was about six times more signi cant compared to those neonates without any visible birth defect. This nding agrees with the studies done in Tigray, Ethiopia (AOR 3.20) and Missouri U. S. A (AOR 3.20). There are plausible but need to be proven explanations that birth defects may result PTB. Among these, certain birth defects associated with abnormal connective tissue may increase the risk of preterm delivery, PTB and birth defects may have shared a common risk factors, and even not observed in our case, prenatal diagnosis of birth defect may lead to deliberately planned delivery through induction or cesarean section before 37 weeks of gestation (29,30). As stated by the study done in Ethiopia, this might be due to the interaction of environmental and genetic risk factors that contribute to PTB (15). As observed in this study, it may be due to that most mothers with birth defect newborns had no iron/folic acid supplementation during current pregnancy.
Lack of iron/folic acid supplementation during current pregnancy also another signi cant factor for PTB. It increases the chance of PTB by four folds. A study from Kenya also showed 80% protective effect of prescribed iron/folic acid during pregnancy with PTB (AOR OR = 0.20). Another study from Sudan reported that folic acid level was associated with a lower risk of PTB (OR = 0.64) and there was a signi cant positive correlation between gestational age and folic acid level. This nding also supported by ndings from China. The reason, in part, may be due to iron/folic acid supplementation prevents anemia and its effect. However, this nding contradicts the report from Bangladesh that iron/folic acid supplementation was a risk factor for PTB. We cannot de nitively elucidate the disagreements between the studies in the association between maternal iron/folic acid supplementation and PTB, but it has been reported that the effects of hemoglobin concentration on birth outcomes could differ by ethnicity (31). It needs further interventional study.

Conclusion
The study indicated that ANC follow up fewer than four visits, PIH, no iron/folic acid supplementation during current pregnancy, being anemic and birth defect were signi cantly associated determinants of preterm birth.
Health education and awareness creation about the bene t of adequate ANC follow, iron/folic acid consumption, and early prevention of pregnancy-related complications are very important during pregnancy. It is also advisable to prevent and treat maternal anemia during pregnancy.
Strengthening patient tracing mechanism through phone call or home visit when mothers interrupt their follow up is advisable. Monitoring and evaluation of the effectiveness of maternal and child health/MCH services is also better. The Ethical clearance was obtained from the Ethical Review Committee of School of Nursing on behalf of University of Gondar Ethical Review Board. A permission letter was also obtained from the study Hospitals' quality control o ce. After explanation had given about the purpose of the study, written consent was obtained from parent or legal guardian of the children. For parents or legal guardian who could not read and write, a thumbprint was used in place of the participant's signature. Anonymous questioner was used to protect the identity and con dentiality of the information obtained. Besides, once the information was gathered from the patient, it was put under a locked cupboard and in a password protected computer.