Predictors and trends of Caesarean section and breastfeeding in the Eastern Mediterranean region: Data from the cross-sectional Cyprus Women's Health Research (COHERE) Initiative

Caesarean section (C-section) can be a life-saving intervention for both mother and child when medically necessary. In 2015 [ 1 ], the World Health Organization (WHO) released their statement on C-section stating that while maternal and neonatal mortality decreased as C-section rates increased towards 10%, C-section rates above 10% were not found to be associated with any reductions in maternal and newborn mortality and could have detrimental short- and long-term effects on the health of the mother and child [ 2 , 3 ]. Despite WHO’s guidance [ 4 ] on potential non-clinical interventions that could be used to help reduce unnecessary C-sections, recent trend analyses of 169 countries has shown that that between 2010–2018, 21% of women gave birth by C-section worldwide, with an average annual increase of 4% [ 5 ]. Cyprus is the third largest Mediterranean island with approximately 300,000 Turkish Cypriot and 700,000 Greek Cypriot residents ( Box 1 ). There is a lack of population-level health data from the Turkish Cypriot population in Northern Cyprus which is an emerging region in Europe [ 6 ], and currently very limited data are available on C-section prevalence. C-section rates have been recently estimated as 55.3% in the Republic of Cyprus [ 5 ], but this excludes the northern part of the island where midwifery care is particularly limited; according to the Ministry of Health of Northern Cyprus [ 7 ], there were a total of 26 midwives employed in public hospitals across the region and approximately 3.53 nurses per 1,000 people. This low number is partly due to the fact that midwives are required to have an accredited midwifery degree, but there are no universities or institutions that offer this degree in Northern Cyprus [ 8 ]. In addition to this, the legal framework in Northern Cyprus means that deliveries can only be performed by clinicians in hospitals. Located in the Eastern Mediterranean region, Cyprus is the third largest island in the Mediterranean. The island is divided into two parts; the internationally recognised Republic of Cyprus and the Turkish Republic of Northern Cyprus (Northern Cyprus), a de facto state only recognised by Türkiye. In 2004, the Republic of Cyprus became a member of the European Union and economically prospered. Conversely, Northern Cyprus has remained under economic sanctions for the past 40 years. There are approximately 300,000 people living in Northern Cyprus, the majority are Turkish Cypriot. There are four potential pathways to accessing healthcare in Northern Cyprus [ 6 ]: 1) The public healthcare system–this is a heavily discounted service providing individuals with social security insurance, which is mandatory for everyone in the workforce, their partners and those under 18. Services in accident and emergency departments are free of charge for everyone. 2) The private healthcare system–although the proportions of individuals purchasing voluntary private health insurance has increased in recent years, it is not widespread so there are high out-of-pocket health care costs. 3) Public services in Türkiye–the Northern Cyprus government has a formal agreement with Türkiye whereby individuals can be sent to Türkiye free of charge for specialist healthcare if the required services are not available within the public sector. 4) Public services in the Republic of Cyprus–since Turkish Cypriots are eligible for citizenship from the Republic of Cyprus, some choose to cross the border and receive healthcare from the public services in the South. It is thought only a small percentage of Turkish Cypriots choose to do this. Similarly, patients can access gynaecology and obstetrics services in public healthcare for free or opt to access private healthcare via their private insurance coverage and at their own expenses. Due to the fragmented healthcare seeking behaviours, there is a lack of data on health needs and behaviours of the population of Northern Cyprus. There is mixed evidence on whether C-section influences breastfeeding success. A systematic review and meta-analysis [ 9 ] showed a negative association between C-section and breastfeeding initiation but when breastfeeding was initiated, then C-section had no effect on the proportion of mothers continuing to do so up to 6 months. A study that took place in the Republic of Cyprus showed that although 84.3% of mothers initiated breastfeeding before being discharged, this figure decreased to 32.4% at 6 months, with mothers that gave birth vaginally being three times more likely to initiate breastfeeding (OR = 3.1; 95%CI 1.7,5.4) compared to those who gave birth by C-section [ 10 ]. In Northern Cyprus, there have been various campaigns [ 11 ] surrounding breastfeeding awareness, but in the absence of any systematically collected data, there is no way of knowing what the prevalence of breastfeeding is, or whether these campaigns have made a difference to rates in the population. In addition, if the C-section rate is similarly high to the Republic of Cyprus, then this could have negative consequences on breastfeeding initiation and the subsequent length a woman chooses to breastfeed for. Therefore, this study aimed to estimate the prevalence, trends and predictors of C-section and breastfeeding in Northern Cyprus and explore any associations between the two.

Of the 7,646 women recruited into COHERE, 3,684 had a first singleton pregnancy, and 77.0% (n = 2,836) had complete data and were included in the study ( Fig 1 ). The women included in the study and women excluded due to missing data were broadly similar except that women excluded tended to have lower educational attainment compared to those included in the study ( S1 Table ). Table 1 shows the characteristics of the study population overall as well as the prevalence of mode of delivery, whether they breastfed and if so, how long they breastfed for. Of the 2,836 births between 1981 and 2018, 55.3% (n = 1,568) were delivered by C-section. Women who gave birth after 2005 had a higher C-section rate (68.6%) than those who gave birth before 1995 (21.4%). The C-section rate was also higher in women who were older than 24 when they gave birth and had a higher educational achievement. Turkish Cypriot women had the highest C-section rate at 57.4%, compared to Turkish women (48.9%) and Other/Mixed women (54.3%). The C-section rate was also higher in women who had a preterm birth (69.4%) compared with a term birth (54.3%), and in those who had fertility treatment prior to their pregnancy (80.6%) compared with those who did not (54.0%). Only includes women >25 years of age at time of recruitment. a Percentages in the first column (All) are column percentages, the rest are row percentages ^Includes all first births regardless of whether it was c-section or vaginal *Only includes women who breastfed C-section rate increased dramatically from 11.1% in 1981 to 72.5% in 2017, with C-section births overtaking vaginal births in 1999 ( Fig 2 ). Table 2 shows the crude RR of mode of birth by time, as well as after adjustment for sociodemographic and pregnancy related factors. Before any adjustment, women who gave birth between 1995–2005 (RR: 2.82; 95% CI 2.34 to 3.40) and those who gave birth after 2005 (RR: 3.20; 95% CI 2.67 to 3.84) had approximately a three-fold increased risk of having a Caesarean birth compared to women who gave birth before 1995. Adjustment for sociodemographic characteristics and pregnancy-related factors attenuated these RRs to 2.6 (95% CI 2.14 to 3.15) and 2.7 (95% CI 2.19 to 3.25), respectively. After full adjustment, risk of giving birth by C-section increased with age with women who gave birth over the age of 35 having a 1.41 (95% CI 1.08 to 1.84) times greater risk compared to women who gave birth under 25 years of age. Women with a postgraduate degree had a 1.35 (95% CI 1.08 to 1.69) times greater risk of C-section compared to women with a primary/middle school degree after fully adjusting for sociodemographic and maternal medical and pregnancy-related factors. Only includes women >25 years of age at time of recruitment. ***p<0.001, **p<0.01, *p<0.05 a Does not include variables that were not associated with c-section or breastfeeding in univariable analysis (BMI, gestational diabetes, gender of baby, hypertension, hyperemesis gravidarum) For all births, breastfeeding was initiated by 88.7% (n = 2,515) of women with prevalence rising from 85.4% in pregnancies before 1995 to 92.9% in pregnancies after 2005 ( Table 1 ). Turkish Cypriot women had a slightly lower prevalence of breastfeeding (87.6% vs 91.2% in Turkish women and 91.2% in Other/Mixed women) and women with a postgraduate degree had the highest breastfeeding prevalence (94.7%). Prevalence of a post-graduate degree in Turkish women was 7.1% compared to 17.8% of Turkish Cypriot women and Turkish women had lower levels of employment (68.2%) compared to Turkish Cypriot (88.0%) and Mixed/Other (88.7%) women, p<0.001 (data not shown). Breastfeeding prevalence was slightly lower in mothers who had pre-term births (81.7% vs 89.2%) and those who used fertility treatments (86.1% vs 88.7%). Of the women who breastfed, 68.6% (n = 1,726) did so for over 12 weeks and the proportion of women doing so rose from 60.9% of births before 1995 to 75.3% of births after 2005. Women aged 12 weeks (64.0%) as did Turkish Cypriot women (64.4% vs 75.6% in Turkish and 71.8% in Other/Mixed). Breastfeeding for the longer duration was lower after pre-term birth (60.5% vs 69.2%) and after births from women who had used fertility treatments (65.3% vs 68.8%). Rates were lowest in babies born underweight (57.0%). Prevalence of ever breastfeeding has remained high between 1981 and 2017, with prevalence decreasing from 100% in 1981 to 94% in 2016, before dropping to 82% in 2017 ( Fig 3a ). Prevalence was lowest at 72% in 1992. Of the women who breastfed, 12.2% breastfed for 1 month or less, 57.3% breastfed for 6 or more months and 28.9% breastfed for 12 or more months. Table 3 shows the crude and adjusted RR (aRR) of ever breastfed by year of pregnancy. After adjustment for both sociodemographic and maternal medical and pregnancy-related factors, women with pregnancies after 2005 had a 10% higher prevalence (aRR: 1.10; 95% CI 0.97 to 1.24) of breastfeeding compared to women with pregnancies before 1995, but this was not statistically significant. None of the sociodemographic or maternal medical and pregnancy-related factors examined had a significant effect on breastfeeding initiation. Denominator only includes women who ever breastfed. Only includes women >25 years of age at time of recruitment. ***p<0.001, **p<0.01, *p<0.05 ^Denominator only includes women who ever breastfed a Does not include variables that were not associated with c-section or breastfeeding in univariable analysis (BMI, gestational diabetes, gender of baby, hypertension, hyperemesis gravidarum) Fig 3b depicts the proportion of time women breastfed for (≤12/>12 weeks). The proportion of women breastfeeding for over 12 weeks increased from 44.4% in 1981 to 73.3% in 2019, with the proportion of women breastfeeding for ≤12 weeks decreasing from 55.6% in 1981 to 24.7% in 2017; women who gave birth after 2005 were 1.2 times more likely (RR: 1.24; 95% CI 1.09 to 1.41) to breastfeed for >12 weeks compared to women who gave birth before 1995 ( Table 3 ) and after full adjustment, the effect estimate did not attenuate. Compared to Turkish Cypriot women, Turkish women were more likely to breastfeed for longer and women with high school/post-secondary education were less likely to breastfeed compared to women with primary/middle school education. The proportion of women who breastfed was 87.8% in those who had a Caesarean section and 89.8% in those who had a vaginal birth. C-section had no association with whether a woman reported to ever breastfeed or not; after adjusting for sociodemographic and maternal medical and pregnancy-related factors, there was no significant relationship between mode of delivery and ever breastfeeding ( Table 4 ). There was also no significant relationship between mode of delivery and length of breastfeeding in women who initiated breastfeeding after adjusting for the same factors. Only includes women >25 years of age at time of recruitment. ***p<0.001, **p<0.01, *p<0.05 ^Only includes women who ever breastfed a Does not include variables that were not associated with c-section or breastfeeding in univariable analysis (BMI, gestational diabetes, gender of baby, hypertension, hyperemesis gravidarum)

For the first time we have described the C-section rate in Northern Cyprus and demonstrated that it has increased dramatically, and it is much higher than the recommended levels. Breastfeeding initiation was high and there has been an increase in length of breastfeeding over the years. There was no relationship between Caesarean section and breastfeeding initiation or length. Advocacy activities to change the legal framework to allow midwives to deliver babies as well as public awareness campaigns and health interventions to inform women of their choice during pregnancy should be considered.

This study uses data collected as part of the COHERE Initiative. The COHERE Initiative is a population based cross-sectional study that has recruited 7,646 consenting women between the ages of 18–55 in Northern Cyprus. The aim of COHERE Initiative is to determine the relative burden of women’s health conditions and related co-morbidities in women living in Northern Cyprus and establish a women’s health cohort for the future to investigate regional risk factors. At baseline, each participant completed a detailed culturally adapted questionnaire on women’s health expanding upon the standardised Endometriosis-Phenome-and-Biobanking-Harmonization-Project (EPHect) questions [ 12 ] to include questions on other women’s health conditions and reproductive health as well. In particular, the questionnaire included questions on pregnancies history and breastfeeding per live birth. Data were collected through a combination of household (16%, (n = 1,208)) and workplace (84% (6,438)) face-to-face visits (93% (7,128)) as well as through online (7% (518)) recruitment methods [ 13 ], between January 2018 and February 2020. Women aged between 18 and 55 at recruitment, who were either citizens of Northern Cyprus or had been residing there for the past 5 years and were able to give informed consent were eligible to participate in the study. Women were recruited into the study from the 6 main districts in Northern Cyprus (Nicosia, Kyrenia, Famagusta, Morphou, Trikomo and Lefke) with recruitment targets being set using geographic population densities. We compared the demographics of the COHERE Initiative sample with projected 2019 population figures for Northern Cyprus obtained from the Northern Cyprus Statistics Institution (available on request from: http://www.stat.gov.ct.tr/ ). Our sample was broadly representative of these projected values with the main differences being seen for age and education in particular regions [ 14 ]. The analysis was limited to first pregnancies only. Mothers with singleton live births between 1981 and 2017 who gave details on mode of delivery were included. We excluded all women under the age of 26 at the time of recruitment as the mean age of first pregnancy in our sample was 24.9, with the majority of women between 18–25 having not had any pregnancies yet. Women in this age group who had been pregnant were likely not representative of all women, so they were excluded to reduce the risk of biased estimates. We also excluded all women who had any missing data on any of the below variables examined. C-section data was collected based on a woman’s self-report to the COHERE survey question: ‘If this pregnancy related to a birth, was the delivery vaginal or via Caesarean section?’. The C-section rate was defined as the percentage of all live singleton births born by C-section and we created a binary variable (vaginal/caesarean) to model the data. Data on ever breastfed and length breastfed for was based on the survey question: ‘If this pregnancy resulted in a birth, for how long did you breastfeed?’ where women were required to write the number of months they breastfed for. A response of 0 or non-response was taken as the woman not breastfeeding and we created a binary variable to analyse data on ever breastfeeding (yes/no). For women who breastfed, an additional binary variable was created to assess length of breastfeeding - ≤12 weeks (early cessation) or >12 weeks. Demographic variables included: age at pregnancy (<25, 25–29, 30–34, 35+), ethnicity (Turkish Cypriot, Turkish, Mixed/Other), education (Primary/middle school, high-school, post-secondary, undergraduate degree, postgraduate degree) and residence (city/village). Pregnancy-related variables included: pre-term birth (yes/no), fertility-treatment used (yes/no), and weight of the baby (underweight <2.5kg, normal weight, 2.5kg-4.0kg, overweight ≥4.0kg). The primary exposures were year of pregnancy, which was calculated by summing a woman’s birth year with the age at which she was pregnant. This was categorized into three groups: 2005. For the analysis of breastfeeding, the secondary exposure was mode of delivery (vaginal/caesarean). Descriptive statistics were used to describe the overall characteristics of the sample. To investigate the association between the exposures and outcomes, modified Poisson regression was used to estimate risk ratios (RRs) and 95% confidence intervals (CIs). For all models, univariable associations were estimated (model A) and then in a hierarchical fashion for potential confounders determined a priori : model B adjusted for sociodemographic factors; model C additionally adjusted for maternal medical and pregnancy-related factors. Confounders were determined a priori based on pre-existing hypotheses or evidence [ 15 – 18 ]. Variables were only included if they were associated with C-section or breastfeeding in univariable analysis. We excluded women who had any missing data on any of the variables included and performed a complete case analysis for all models. Statistical analyses were carried out using Stata SE version 17.0 (StataCorp LP, College Station, Texas, USA) and figures were produced using R Studio. The study was approved by the Oxford Tropical Research Ethics Committee (OxTREC) of the University of Oxford (OxTREC reference: 37–17). The study also received local ethics approval from the Eastern Mediterranean University Ethics Committee (ETK00-2017-0240). Verbal and written consent was obtained from all participants during recruitment.

(DOCX) Click here for additional data file.