Background

Preterm birth is the delivery of the fetus before 37 weeks of gestation. According to estimates from the World Health Organization, 15 million (or one in ten) babies are born early each year [1]. African and South Asian countries account for more than 80% of these preterm births [2]. On average, 12% of babies are born prematurely in low-income nations, compared to 9% in high-income countries [3] with differences in incidence among countries. According to different studies, this incidence is reported to vary from 5.0% in Sweden [4], 16.8% in Nigeria [5], 18.3% in Kenya [6], and between 4.4 and 25.9% in Ethiopia [7,8,9,10].

Preterm birth is the primary cause of perinatal illness and mortality worldwide [2]. It is one of the biggest healthcare challenges as it is associated with long-term disability and financial strain from the costs of care especially in underdeveloped nations [11]. According to Wagura et al. [6], it accounts for around one-third of all neonatal deaths as a result of increased risk of infection [12]. It also has long-term adverse effects, such as poor neurodevelopment leading to learning disabilities, cerebral palsy, and vision abnormalities among others [13].

Studies have found a number of risk factors for preterm birth, which can be categorized as: (1) Maternal risk factors, such as maternal age, education level, low socioeconomic status, marital status, maternal malnutrition, substance use, inadequate antenatal care (ANC), and stress, among others [14,15,16,17], (2) gynecological risk factors such as intrauterine infection, urinary tract infection, pregnancy induced hypertension, sexually transmitted infections, premature rupture of membranes, uterine malformations, and uterine adhesions [18,19,20,21] and (3) Obstetric risk factors include: primiparity, short birth interval, antenatal care, multiparity, history of preterm birth, history of abortion, history of stillbirth or miscarriage, and multiple pregnancies [22,23,24]. However, the factors from various studies are uneven, with some factors showing direct association in some studies but having an inverse association or no association in others.

In Rwanda, despite the improvement of maternal and child health services, at least 35,000 babies are born too soon each year, and 2600 children under five die due to direct preterm complications [25]. The factors causing increased preterm birth in Rwanda are currently unknown. Some small-scale studies representing specific locality or hospital have been done. For instance, Nwankwo et al. demonstrated preterm birth prevalence of 17.5% with husband’s smoking, low ANC attendance and low maternal MUAC as independent predictors of preterm birth [26]. This was a single center facility based cross-sectional study whose findings cannot be generalized. However, as far as we are aware, there is no publication of a longitudinal study that attempt to investigate risk factors associated with preterm birth at a national level. Hence, this study aimed at determining the maternal, obstetric, and gynecological factors associated with preterm birth in Rwanda.

Methods

Study design and setting

Rwanda is a landlocked country located in central-eastern part of Africa. It is a low-income country with a population of about 13 million people. The country ‘health system is a pyramid with the top of pyramid, being the Ministry of health responsible for sector coordination and oversight, and setting health policies and strategies; Rwanda Biomedical Centre responsible of implementation of health programs such as Tuberculosis, Malaria, HIV and Maternal and child health; Rwanda Food and Drugs Authority for regulation of human and veterinary medicines, vaccines and other biological products, processed food, poisons, medical cosmetics, medical devices, and other products; and Rwanda Medical Supply with the mandate of ensuring availability of medicines, medical supplies and consumables to the health facilities. Currently, Rwanda has 565 public health facilities including 504 Health centers, 7 Medicalised Health Centers, 40 District, 4 Provincial, 4 Specialized and 8 Referral hospitals. Health posts are entities working at lowest level and operating under public-private partnership models; there are 1222 Health Posts (HPs) in Rwanda. Private health facilities are 317 and distributed as follow: 115 private dispensaries, 113 general clinics, 33 specialized clinics, 28 polyclinics, 9 dental clinics, 6 nutrition cabinets, 4 Private Hospitals, 4 Specialized hospitals, 3 Psychology clinics, and 2 laboratories [27].

A health facility follow-up study using a longitudinal single baseline cohort design among first-trimester pregnant women was conducted from July 2020 to July 2021. The study was conducted in 30 health facilities including hospitals and health centers located in 10 districts around the country. The sampling unit was hospital and then convenience sampling was used to select one health center in urban area and one health center in rural area under each selected hospital. Therefore, pregnant women in the first trimester from 10 hospitals and 20 health centers were recruited in the study.

Population, sample size and sampling technique

Participants in this study were pregnant women in their first trimester. Being in the first trimester of pregnancy (within the 13 weeks gestation), intending to stay in the study for the duration of the study, and being able to sign or thumbprint informed consent were the inclusion criteria. However, pregnant women with chronic diseases such as hypertension, heart disease, and renal disease were excluded.

A multi-stage sampling strategy was adopted, with the first stage consisting of the simple random selection of 30 health facilities (acting as clusters) from around the nation (Fig. 1). Then, pregnant women who attended the antenatal care department in the selected healthcare facilities and were determined to be in their first trimester were included in the study. Between July and December 2020, all eligible women from each selected health facility were enrolled in the study. A total of 1159 pregnant women in their first trimester took part in the study after taking into account all inclusion criteria and 817 pregnant women were included in the analysis (Fig. 1).

Fig. 1
figure 1

Flowchart displaying participant enrollment and follow-up

Full size image

Data collection procedure and tools

30 qualified nurses and midwives collected the data under the direct supervision of 10 team leaders from Mount Kenya University, and the study team. Team leaders, nurses and midwives received one-day training on the study protocol, inclusion criteria, sampling, and research tools, and data collection procedures. The questionnaire was pretested among 50 pregnant women who attended two health centers in Kigali City for ANC services. With the results of this pilot study, the study tool was adjusted as necessary. Pregnant women were told about the goal of the study during the data collection procedure, and they consented to participate.

Face-to-face interviews were used to gather the data using a structured questionnaire, and antenatal care records were examined to gather some additional medical information on maternal conditions. The questionnaire was adopted from the literature and contextualized according to Rwandan context. The questionnaire was initially prepared in English, and then translated into Kinyarwanda, then back into English. At the base line (first trimester) as well as in the end line (shortly after delivery) factors related to maternal characteristics (age, residence marital status, level of education, employment, socio-economic status, alcohol use and smoking during pregnancy and exposure to passive smoking), obstetric (parity, birth spacing, mode of delivery ANC use, preterm rupture of membrane, amniotic fluid volume HIV status Blood group and history of still birth, abortion and previous Cesarean section) and gynecologic (hospitalization during pregnancy, gestational diabetes, hypertension in pregnancy, malaria during pregnancy, hepatitis B, urethritis infection, cystitis infection, pyelonephritis infection and vaginitis infection) were collected.

Measurement of the primary outcome (pre-term birth)

Preterm delivery, which is defined as a live birth at less than 37 weeks of gestation, was the main outcome in this study. The latest menstrual period (LMP) date was used to estimate the gestational age, and an ultrasound examination was used to confirm this. Since ultrasounds are not typically performed at health centers, the pregnant women recruited from the health centers were directed to the nearest hospital for ultrasound examination. Ultrasound examination was done by the obstetrician or general practitioner at hospital.

Data analysis

Using IBM SPSS Statistics 25, data was entered and examined. Counts and percentages were used to describe the characteristics of respondents. Tables and text were used to present the results. For each categorical independent variable, a Chi-square assumption was made. Then, binary logistic regression using bivariate and multivariate analysis was used to identify the risk factors related to preterm birth. For the multivariable logistic regression analysis, variables having a p-value of less than 0.2 in the bivariate logistic regression were taken into account. The Adjusted Odds Ratio (AOR) with a 95% confidence interval was determined in the multivariable logistic regression analysis. Finally, factors were deemed significant if their p-value was less than 0.05. The Hosmer-Lemeshow test was used to gauge the model’s fitness.

Ethical consideration

With reference No. 131/RNEC/2020, the Rwanda National Ethics Committee granted its approval, and each of the chosen hospitals provided a letter of permission. Participants in the study were given a thorough description of the study’s objectives before providing their written informed consent. Legally authorized representatives of illiterate participants provided informed consent for the study. Informed consent was obtained from a parent and/or legal guardian for participants aged under 16.

Participants in the study had the freedom to decline or leave at any moment without incurring any penalties. Furthermore, the participants received assurances regarding the confidentiality of the data, and no personal identifying information was included on the questionnaires.

Results

Socio-demographic and lifestyle factors stratified by preterm

A total of 817 women and new-born pairs were included in the study to achieve the objectives. The prevalence of preterm birth was 13.83%, with a 95% confidence interval of 11.54% and 16.39%. The majority of new-born babies (63.2%) were female; however, this was not significantly associated with preterm birth. Similarly, there was no statistically significant association observed between risk of preterm birth and residence, marital status, level of education, occupation, and social class, as well as alcohol consumption and smoking during pregnancy. However, a significant association was observed between maternal age, partners’ smoking status, and the risk of preterm birth. The proportion of mothers who delivered preterm babies was significantly higher in the 35–49 age group (23% vs. 13.4%, p = 0.018) (Table 1). Mothers with preterm deliveries were more likely to be exposed to secondhand smoke compared to mothers with full-term births (p = 0.017) (Table 1).

Table 1 Socio-demographic and lifestyle factors stratified by preterm
Full size table

Obstetric factors associated with preterm birth

The distribution of maternal obstetric factors according to gestational age is summarized in Table 2. Mothers with preterm deliveries attended fewer than four ANC visits than mothers with full-term deliveries (p = 0.021). Other obstetric factors associated with preterm birth that were statistically significant were history of abortion (p = 0.018), history of cesarean section (p = 0.042), and preterm membrane rupture (p < 0.001).

Table 2 Obstetric factors associated with preterm birth
Full size table

Gynecological factors associated with preterm birth

Table 3 presents factors related to gynecological conditions according to the gestational age at delivery. Hospitalization during pregnancy was significantly higher among mothers with preterm birth compared to those mothers with full term birth (p = 0.002). Similarly, the risk of preterm birth was significantly higher among mothers with hypertension (p = 0.002) and mothers with vaginal infections (p = 0.029).

Table 3 Gynecological factors associated with preterm birth
Full size table

Bivariate and multivariable logistic regression analysis of factors predicting preterm birth

On bivariate logistic regression analysis, maternal age, secondhand smoking, ANC visits, history of abortion, preterm rupture of membranes, hospitalization, hypertension, and vaginal infections were significantly associated with preterm delivery at a p value less than 0.05. After considering all these variables together in multivariable logistic regression by specifying the backward conditional method, maternal age, second-hand smoking, history of abortion, preterm rupture of membrane, and hypertension were predicting preterm birth (Table 4).

Table 4 Bivariate and multivariable logistic regression analysis of factors predicting preterm birth
Full size table

Older mothers aged 35 to 49 years were 2 times more likely to deliver preterm babies compared to younger mothers aged 15 to 24 years (AOR = 2.00; 95%CI = 1.13–3.53). Preterm birth was 1.9 times more likely to occur among mothers exposed to secondhand smoke (AOR = 1.91; 95% CI = 1.04–3.51). Mothers who had previously had an abortion had a 1.9-fold increased risk of preterm birth (AOR = 1.89; 95% CI = 1.13–3.15; Preterm birth was about nine fold more common among mothers who experienced preterm membrane rupture (AOR = 9.30; 95CI = 3.18–27.16) and 4.4 times more common among hypertensive mothers (AOR = 4.40; 95% CI = 1.18–16.42).

Discussion

The purpose of this study was to establish maternal, obstetric, and gynecological factors associated with preterm birth in Rwanda. In our study, the prevalence of preterm birth was found to be 13.8%. After adjusting for potential confounders using a multivariable logistic regression model, the identified independent risk factors associated with preterm birth were advanced pregnancy age, secondhand smoke, abortion history, preterm membrane rupture, and pregnancy-induced hypertension.

The prevalence of preterm birth in Rwanda was higher than studies done in multi-center health facilities in Iran (5.1%) [28] and in Australia (6.8%) [29], as well as the WHO estimates for sub-Saharan Africa (9.5%) and lower-income countries (12.0%) [30]. However, it was similar to the pooled prevalence from metal-analysis in Ethiopia (11.4%) [31] as well as findings from multi-center study in Brazil (12.3%) [32] and from a referral hospital in Tanzania (14.2%) [33]. Further, it was slightly lower than a study conducted in Kenya (18.3%) [6] and in one District Hospital in Rwanda (17.5%) [26].

However, this prevalence is lower than recent cross-sectional studies done in teaching hospitals in Ghana, whereby it was reported at 37.3% [34], and 25.9% in Ethiopia [10]. These disparities in the prevalence of preterm birth among the studies could be attributed to small sample sizes in a small-scale study setting [9, 26, 34]. It could also be the method used to estimate the gestational age, as most studies depend on the last menstrual period reported retrospectively by the mothers, which could result in a biased prevalence [6, 26, 35]. Moreover, most studies recruited respondents from a single hospital, which could possibly overestimate the prevalence of preterm birth [9, 34, 36]. Our study, on the other hand, used a sufficient sample size from the community and the gold standard approach of ultrasonography to determine gestational age.

Our study showed that advanced maternal age was associated with preterm birth. Several studies have also found that advanced maternal age is a risk factor for preterm birth [16, 17, 37,38,39]. This increased risk of preterm birth among women of advanced age (> 35 years) could be attributed to the fact that reproductive organs and fertility decrease after 35 years. Moreover, complications related to pregnancy are more common in this advanced age group. However, other research from the United Kingdom [40] failed to detect an association, which may be related to socio-economic differences.

It is acknowledged that the primary risk factor for preterm birth is premature membrane rupture [41]. According to our research, women who experienced premature membrane rupture had a 9.3 times higher risk of having a preterm birth. This is consistent with several other studies [6, 36, 42,43,44,45,46]. This may be as the result of endogenous prostaglandins released during membrane rupture, which start the uterine contraction and lead to preterm birth [31]. This implies that pregnant women who have a history of PROM should receive additional care and management.

In concordance with other previous studies, [34, 43, 47, 48], this current study observed that hypertension during pregnancy was an independent risk factor for preterm birth. In East Africa, studies conducted in Kenya and Tanzania found that pregnancy induced hypertension is a risk factor for preterm birth [6, 33]. The poor pregnancy outcomes linked to hypertensive diseases during pregnancy, including premature delivery, are plausibly explained by uteroplacental ischemia [49, 50], despite the biology of this illness still not being fully understood. Thus, early detection of pregnancy-induced hypertension during antenatal care visits and screening is critical for appropriate treatment and management.

The current study found that women with a history of abortion were significantly more likely to give birth prematurely. This is in agreement with other studies as well as a systematic review and a meta-analysis [9, 10, 45, 51, 52]. According to the literature, the risk of abortion in early pregnancy is associated with undesirable birth outcomes, including preterm birth. The risk of infection associated with recurrent abortion may be the biological mechanism causing this connection. It has been reported that women who have had abortions before are more likely to get intra-amniotic infections [53]. Preterm delivery has a known risk factor known as intra-amniotic infection [54]. Women and medical professionals (midwives and obstetricians) should also be made aware of the potential link between abortion and premature delivery.

Our study also revealed that pregnant women exposed to secondhand smoking had significantly higher odds of delivering premature babies. There has been evidence that pregnant women who smoke cigarettes are at increased risk of preterm birth [9, 10, 29, 55]. In our study, the number of women who smoked during pregnancy was not significant (p = 0.089). However, the link between secondhand smoke exposure and preterm birth is still up for debate. Two recent studies done in Vietnam and the USA depicted a significant association between preterm birth and secondhand smoking [56, 57]. The pathway and connection between secondhand smoke and preterm delivery must be better understood in order to guide policymakers in putting preventative initiatives in place.

Although UTI and ANC attendance were significant in bivariate analysis, they were not significant in multivariate analysis. This finding is similar to that of a study conducted in Kenya which reported no significant association between preterm birth and number of ANC visits [6]. In contrast, a systematic and meta-analyses study found that women who had less than four ANC visits had a higher risk of preterm birth [45], which could be due to a failure to identify preterm birth risk factors during ANC visits. Other studies and reviews have found that maternal urinary tract infection (UTI) is linked to preterm birth [6, 20, 45], with infections weakening the baby’s amniotic sac, leading to PROM and preterm birth [58].

One of the study’s strengths is its large representative sample drawn from 30 health facilities selected in 10 districts. The findings thus represent the prevalence of preterm births in Rwanda with reasonable certainty. This study does have some drawbacks, though. The many independent variables and preterm birth do not necessarily have a cause-and-effect connection. Additionally, there may be a chance of recollection bias because the data were obtained through interviews. Loss to follow-up (LTFU) among recruited persons was another constraint (24.97%), but after comparing the baseline characteristics of the LTFU and those of participants included in the data analysis, there were no significant discrepancies, ensuring the internal validity of the study.

Conclusion

The magnitude of preterm birth was slightly higher than in the most recent Rwandan demographic health survey. It was independently affected by advanced age during pregnancy, secondhand smoke, a history of abortion, preterm membrane rupture, and pregnancy-induced hypertension. Therefore, it is preferable to take into account pregnant women who are older, exposed to secondhand smoke, have had abortions before, and have hypertension when screening and intervening in order to avoid the short- and long-term effects of premature birth.