Significance

In this study, a positive influence of healthier habits and higher education on predominant breastfeeding (PBF) initiation and duration has been demonstrated. The decrease in PBF duration rates in Spain can be interpreted partly as a consequence of the return to work of women. This study will facilitate the revision and amendment of the guidelines and recommendations for the protection and promotion of PBF by health services. The findings breastfeeding imply the need to review existing measures to reconcile work and family life.

Introduction

Breastfeeding is one of the most important contributions, not only to a child’s health and development (Der et al. 2006) but also to maternal health (Salone et al. 2013; Stuebe and Schwarz 2010). International guidelines recommend exclusive breastfeeding (EB) from birth up to 6 months of life, and then continuing breastfeeding with complementary foods until at least, 2 years of age (World Health Organization and UNICEF 2007). Despite the well-documented benefits of EB, few women are meeting the recommendations at 6 months postpartum, particularly in industrialized countries.

In Spain, 66.2% of infants were exclusively breastfed during 6 weeks, 53.6% for the first 3 months and 28.5% during 6 months (INE 2011). In industrialized countries, the highest percentages of initiation of exclusive breastfeeding were in Norway, Denmark, and Japan with 99, 98.7 and 98.3% respectively. The lowest percentages were in the United Kingdom, the United States, and France with, 70.0, 69.5 and 62.6% respectively (Ibanez et al. 2012). Successful breastfeeding depends on multiple factors such as socio-demographic characteristics (Grjibovski et al. 2005), healthcare-related or psychosocial influences (Sunyer et al. 2010) and those associated factors with a supportive environment (Giovannini et al. 2005; Swanson and Power 2005). In a review of the literature on developed countries, the key positive factors include older maternal age, higher level of education, higher socioeconomic status and healthy maternal habits (Lande et al. 2003; Thulier and Mercer 2009). Other studies also confirm a strong and consistent association of breastfeeding initiation and duration with maternal intention to breastfeed prior to the birth (Kronborg and Vaeth 2004). Further, returning to work is a social factor that may influence women’s decisions to discontinue breastfeeding. In fact, maternal employment is often linked to premature weaning (Johnston and Esposito 2007). In Spain the maternity leave (the responsibility of the Ministry of Labor and Immigration) lasts 16 weeks. Regarding protection, promotion, and support of breastfeeding in Spain, currently national plans that include general objectives and recommended strategies from the European Plan of Protection, promotion, and support of breastfeeding have been developed. The Spanish Association of Pediatrics in collaboration with the Health Ministry has implemented actions and recommendations to promote the breastfeeding such as the baby friendly hospital initiative (BFHI) or the plan to support breastfeeding in the workplace, conducted in collaboration with the business sector.

Variations in the definition of breastfeeding are likely to underlie mixed results when comparing breastfed and artificially-fed infants (Ibanez et al. 2012; Kramer and Kakuma 2012). As the rates of EB, only breast milk, no other liquids or infant formula, are extremely low in our setting, we decided to study PBF defined by the WHO as receiving breast milk only, but allowing supplementations of non-milk liquids (e.g., water or water-based drinks such as sweetened and flavored water, teas, infusions), fruit juice, oral rehydration salts solution, and drop or syrup forms of vitamins, minerals, and medicines (World Health Organization et al. 2008) taking into account that this definition has a widespread use in the literature (Ibanez et al. 2012; Santini et al. 2008; Scott et al. 2006) (Appendix 1).

The identification of variables related to breastfeeding duration offers many implications for clinical and public health practice. Here, we present information on breastfeeding among women from the INfancia y Medio Ambiente, the Spanish for Childhood and the Environment (INMA) project, a multicenter cohort study which aims to investigate the effect of environmental exposures and diet during pregnancy on foetal and child development (Guxens et al. 2012). The objectives of this research were to (i) assess PBF prevalence in women participating in the INMA project, (ii) analyze the association of non-initiation of PBF and cessation with social and lifestyle factors among these women, and (iii) identify profiles of increased susceptibility to PBF cessation or a shorter duration of breastfeeding.

Methods

Study Population

The study population comprised pregnant women from four cohorts of the INMA project. Specifically, women were recruited in the first trimester of pregnancy in four regions in Spain: two northern regions, Asturias (N = 494) and Gipuzkoa (N = 638; Basque Country), and two Mediterranean, Sabadell (N = 657) and Valencia (N = 855) (Guxens et al. 2012). Recruitment took place between 2003 and 2008. Inclusion criteria were: maternal age of 16 years or older, singleton pregnancy, enrollment at 10–13 weeks of gestation, unassisted conception, delivery scheduled at the reference hospital, and no limitations in communication. Women participating in the study signed an informed consent form and the Ethics Committees of the centres involved in the study approved the research protocol. Of the 2506 women who continued in the study at birth, 2315 women complemented questionnaire regarding breastfeeding. Women who did not complete both breastfeeding questionnaires at 6 and 14 months after delivery (N = 120) were excluded. Overall, data of 2195 women was analyzed (Fig. 1).

Fig. 1
figure 1

Selection of study population (n = 2195), INMA birth cohort (2003–2008)

Full size image

Questionnaires

Women filled in questionnaires during the first and third trimester of pregnancy and 6 and 14 months post-delivery. The questionnaires were administered by trained interviewers and focused on breastfeeding, socio-demographic and environmental factors and lifestyle information related to pregnancy. The maternal covariates used in this study were country of birth, maternal age, complications in the current pregnancy, miscarriages, parity, type of delivery, educational level, working situation during week 32 of pregnancy and at 14 months post-delivery, social class, physical activity, pre-pregnancy body mass index (BMI) and smoking during pregnancy. Educational level was based on the total years of education completed by the women, grouped into three categories: primary education (< 11 years of education), secondary education (12–15 years), and university or other higher education (> 16 years). Five categories of occupational class were considered according to the Spanish adaptation of the British Register General’s Social Class classification (Domingo-Salvany et al. 2000). In order to increase the statistical power, the five levels were grouped into two: manual or lower class (IV and V) and non-manual/skilled or higher class (I–III), respectively.

Breastfeeding

Data on breastfeeding were based on questionnaires administered to the women at 6 and 14 months after delivery, and only women who fully completed these questionnaires were included in the analysis. Women were considered to have initiated breastfeeding if they continued to breastfeed at day 3 or 4 after the birth, corresponding with the time of discharge from hospital in most cases.

PBF is defined by the WHO as receiving breast milk only, but allowing supplementations of non-milk liquids (e.g., water or water-based drinks such as sweetened and flavored water, teas, infusions), fruit juice, oral rehydration salts solution, and drop or syrup forms of vitamins, minerals, and medicines (World Health Organization et al. 2008). The date of PBF cessation was defined as the minimum value of the date of introduction of infant formula or non-human milk and the date of introduction of food. These dates, and the specific food items introduced, were collected in the questionnaires administered to the women at 6 and 14 months.

Statistical Analysis

Descriptive statistics were generated for the whole cohort and data were expressed as mean and standard deviation (SD) for continuous variables. The association between categorical variables was performed using (χ2). P value less than 0.05 was considered statistically significant, and all p values were based on two-tailed tests. Data analysis was performed at fixed time points: at discharge, 3 months (week 13), 6 months (week 26), and 14 months post-delivery.

Binary and multivariate logistic analysis was used to examine the relationship of non-initiation and cessation of PBF with maternal characteristics in whole sample (N = 2195). All significant variables (p < 0.05) in binary analysis were included in multivariate models. The final multivariate logistic model was performed by stepwise backward regression. Binary and multivariate Cox proportional hazards regression were used to estimate the association between stopping PBF and maternal characteristics in those that started PBF (N = 1874). All significant variables (p < 0.05) in binary analysis were included in multivariate models, the final multivariate Cox proportional hazards model was performed by stepwise backward regression. All models were adjusted for study cohort. The Kaplan–Meier method was employed to assess rates of PBF on the sample of women that initiated PBF. Infants who were continuing to breastfeed at the time of data collection were entered as censored data. All statistical analyses were conducted using SPSS (version 21) and R-project (version 3.2.0). Statistical significance was established in p < 0.05.

Results

Overall, 2195 women completed the questionnaires, identifying 1874 mothers who initiated PBF and 321 (14.6%) who never breastfed (Table 1). The duration of PBF was 17.50 (median) weeks. The prevalence of PBF at hospital discharge was 85.3%, at 13 weeks 53.4%, at 16 weeks 46.1% and at 26 weeks 7.2% (Table 2). Only two women take PBF at 12 months and none at 14 months.

Table 1 Description of the study sample
Full size table
Table 2 Rates of PBF over the study period as a function of maternal characteristics
Full size table

Most participants were Spanish (92%) and, at the time of enrollment, they had an average age of 30.73 ± 4.14 years, over half were nulliparous (56.8%), almost a third (31.8%) reported having complications in the current pregnancy, and 22.9% had had previous miscarriages. Just over half (51%) were categorized as low social class (unqualified and/or manual workers), a percentage slightly lower than the figure for the general population (Instituto Nacional de Estadística 2014). Employment status recorded from delivery to 14 months post-partum showed that 71% of the women worked at least until week 32 of pregnancy and 67.3% returned to work by 26 weeks post-delivery.

Table 2 presents the rates of PBF over the study period as a function of maternal characteristics. Initiation of PBF was most strongly associated with higher levels of maternal education, have a country of birth different from Spain, being a first-time mother, and those who worked in a non-manual occupation. Healthier habits, such as a higher level of physical activity, not smoking, and having a healthy BMI, were also positively associated to PBF initiation. Additionally, these lifestyle-related variables were factors strongly related to PBF continuation until week 26 after delivery. Women with higher levels of education were significantly more likely to breastfeed for a longer period (p < 0.001). On the other hand, smoking and working status were found to be significantly related to a shorter PBF duration (p = 0.001). Further in each of the working periods studied, in week 16 and onwards, the rate of PBF was significantly higher among women who did not work than those who worked (p < 0.05). Nevertheless, at discharge, the working mothers were more likely to breastfeed that the non-working ones (p < 0.05). Employment appears to have a less deleterious effect on initiation of PBF.

Figure 2 shows the Kaplan–Meier survival rates of PBF (%) during the infants first 26 weeks of life. Overall the 1874 women that initiated PBF, during the first week 10.1% stopped. At 13 and at 16 weeks of age, 62.7 and 53.9% of the infants were receiving breast milk respectively. And at week 26, the prevalence of PBF fell down to 8.4%.

Fig. 2
figure 2

Kaplan–Meier survival rates of PBF (%) during the infants’ first 26 weeks of life (N = 1874)

Full size image

Factors associated with the risk of no initiation and cessation PBF are compiled in Table 3, which represents the unadjusted and adjusted models as a function of maternal characteristics. Among the whole sample, women (N = 2195), with university education and non-manual occupational class, with normal weight, nulliparous, physically active > 3 h week and those who did not smoke during pregnancy were more likely to initiation of PBF at delivery, even after adjusting for potential confounders. In the adjusted model the same variables except occupational class explained the initiation of PBF. Moreover, smokers had a higher risk of no initiation throughout the whole sample as unadjusted and adjusted OR (95% CI) were; 1.80 (1.34–2.41), and 1.74 (1.27–2.34) respectively.

Table 3 Unadjusted and adjusted Odds and hazard ratios for the likelihood of non-initiation and cessation of PBF as a function of various socio-demographic and lifestyle factors in the whole sample (N = 2195) and those that started PBF (N = 1874)
Full size table

In women that started PBF (N = 1874), the cessation of breastfeeding was higher in young women, who had had complications during the pregnancy, nulliparous, obese, who had lower levels of education and manual occupational class, with physically active > 1 h week, and those who did smoke during pregnancy. In the adjusted model, age, complications during the pregnancy, educational level of the women and obesity were the variable that explained the risk of PBF cessation. The employment status of the women, in week 32 HR (95% CI) were; 1.16 (1.04–1.30), and 1.19 (1.04–1.36) of pregnancy and also in month 14 post-delivery OR (95% CI) were; 1.15 (1.04–1.28), and 1.16 (1.03–1.32); both determined the likelihood of PBF cessation in women who started PBF, even after adjusting for potential confounders. Overall, the results imply that the later the return to work, the longer the mean duration of PBF.

Discussion

The results of our study suggest that women over 25 years, with university education, non-manual occupational class, and with healthy lifestyles have a positive influence on both the initiation and duration of PBF. Evidence in the literature suggests that greater duration and intensity of PBF, as well as exclusivity, have important beneficial effects on maternal and infant health, including long-term outcomes (Duijts et al. 2010; Stuebe et al. 2011).

In Spain 68.4% of newborns are given PBF at 3 months and 24.72% at 6 months. This percentage varies depending on the region. With regards to our study areas, at 3 months of life the percentages of PBF are 28.47, 64.12, 47.74 and 55.88 in Asturias, Basque Country, Valencia and Catalonia respectively. The prevalence decreases at 6 months, with 14.68, 39.19, 25.61, 28.89% respectively. According to social class, the percentage is 54.1% for non-manual (I–III) and 53.1% for manual (IV–V) workers at 3 months and 37.3 and 27.9% to 14 months (INE 2011). Taking into account that the 18 autonomous communities of Spain are included in the Spanish Health Survey, PBF ranges are between 13.75 and 74.37% at 3 months and between 1.59 and 39.19% at 6 months.

The INMA Project is a population-based study. However, it might not be fully representative of all children in the study areas due to refusals and losses to follow up. The internal validity regarding the association of PBF with the studied variables is not necessarily affected. Nevertheless, although attrition rates were low, non-participants were more likely to have a lower socio-economic status (Guxens et al. 2012). This fact could lead to a mild overestimation of the PBF prevalence in the Spanish population. Moreover, of the 2315 women participating in the study at birth, 120 women did not complete the questionnaires at 6 or 14 months. Among those who do not respond, A higher proportion of women of low social class and educational level was observed. The losses represent only 5% of the total, so the habit of breastfeeding in these women would have to be very different from the rest to interfere with the study results.

Data on rates across European countries are limited by inconsistent monitoring methods and definitions of breastfeeding (Ibanez et al. 2012; Kramer and Kakuma 2012); the definitions have varied between studies and have often been based on convenience, determined by the data the researchers had available.

Our results match those of previous studies that have consistently found that women who breastfeed tend to be older and well educated, with healthy lifestyle habits (Callen and Pinelli 2004; Thulier and Mercer 2009). Indeed, educational level and social class were the two variables that most strongly determined the decision to delay formula feeding. On the other hand, Dubois and Girard (2003) have reported that high income level was negatively associated with initiation. Our data supports the idea that immigrant women are more likely to breastfeed than non-immigrants as reported in other European countries, where a higher prevalence of PBF has been also identified in immigrant populations (Merten et al. 2007; Ibanez et al. 2012). Likewise, it provides data concerning a southern European population, less well studied than those of Nordic countries and with its own distinct sociocultural characteristics (Santini et al. 2008). Compared to data reported for Scandinavian countries such as Norway, Denmark and Sweden (Flacking et al. 2007), with PBF initiation rates of 99, 98.7, and 98% respectively, breastfeeding rates in Spain are lower, but they are somewhat higher than figures from Italy, the United Kingdom, the United States, with rates of 70, 69.5, and 62.6% respectively for PBF initiation. Specifically, the data collected in this study positions Spain in the middle, with a PBF initiation prevalence of 85.3%. Our observation that social class, educational level, and healthier lifestyle habits were strongly related to PBF continuation underlines the importance of social environment in the decision to breastfeed (Julvez et al. 2014; Lange et al. 2007).

Similar factors have been detected as potential breastfeeding promoters in Europe and America where educational level and social class are determinant, in contrast with the data from a Middle-Eastern country indicating that women with only primary education breastfed for longer periods than those with greater levels of education (Radwan 2013).

The most pronounced decline in exclusive breast-feeding occurs between 16 and 26 weeks post-delivery, coinciding with most of the working women going back to work. This finding demonstrates the interest of being able to measure prevalence at these two different moments. These results are in agreement with other studies showing that the PBF declines sharply in women who return to work and that employment of mothers outside the home has a negative influence on PBF duration (Fein et al. 2008) and contrasts with the suggestion of Bakoula et al. (2007) that working mothers are more likely to breastfeed than housewives. Our findings support the hypothesis that the decrease in PBF duration rates in Spain can be interpreted in part as a consequence of women returning to work. In Spain the maternity leave is 16 weeks: 6 weeks are obligatory and must be taken following the birth, while the remaining 10 weeks can be taken before or after birth. Also, there are differences between women that work in the private and the public sector have been observed, especially with regards to the duration of full and/or partial salary during maternity leave and the provision of paid breast-feeding breaks. Additionally, the data presented could help to identify strategies to improve PBF rates in our setting. Firstly, regarding the importance of educational level in the PBF decision, the influence could be partially compensated for by antenatal education (Artieta-Pinedo et al. 2013). Secondly, returning to work is an obstacle to continuing PBF and indicate a need to develop policies and practical measures to support women at this stage (Fein et al. 2008). Finally, it is recommended to establish healthy habits, such as not smoking, physical activity, and have a healthy weight before pregnancy, due to its association with longer breastfeeding and improving maternal and child health.

Like the authors of other European studies, we conclude that PBF is strongly influenced by socio-demographic determinants and further research is required to enable public health professionals to target appropriate subgroups of women at key stages. On the other hand, an accurate comparison of the social factors associated with PBF between countries because of several methodological differences. The comparison between different studies is hindered by the different definitions and characteristics of breastfeeding employed in the literature. In this study, data regarding breastfeeding and socio-demographic and life style factors were collected by questionnaire previously validated and administered face to face by trained interviewers.

Taking into account the evidence that confirms a strong association of PBF initiation and duration with maternal intention to breastfeed prior to the birth (Kronborg and Vaeth 2004), medical counseling during first stages of pregnancy could improve PBF rates and women’s behavior, although returning to work seems to have a substantial impact in PBF continuation. The design and development of effective policies and other strategies to promote PBF should be adapted to the target population. Initiatives such as the Baby-Friendly Hospital Initiative (Merten et al. 2005) contribute to promoting and supporting breastfeeding. In view of our data, we believe that until employers in Spain develop maternity policies which do not discourage breastfeeding, the recommended 6 months of PBF is unlikely to be achieved by most working women.