The benefits of breast-feeding for the health and well-being of infants have been well recognized(1–Reference Ip, Chung, Raman, Chew, Magula, DeVine, Trikalinos and Lau4). However in most industrialized countries, including Greece, the percentage of mothers following the recommended breast-feeding practices is far from that desired(Reference Theofilogiannakou, Skouroliakou, Gounaris, Panagiotakos and Markantonis5–Reference Cattaneo, Yngve, Koletzko and Guzman9). As recent studies in local cohorts in Greece have shown, although more than 90 % of women initiate breast-feeding(Reference Theofilogiannakou, Skouroliakou, Gounaris, Panagiotakos and Markantonis5, Reference Pechlivani, Vassilakou, Sarafidou, Zachou, Anastasiou and Sidossis10), only 16·0 % of mothers provide any breast-feeding at 6 months postpartum(Reference Theofilogiannakou, Skouroliakou, Gounaris, Panagiotakos and Markantonis5) and only 7·3 % breast-feed for more than 1 year(Reference Antoniou, Daglas, Iatrakis, Kourounis and Greatsas11). However, these were small-scale studies conducted in urban areas and therefore cannot be representative of the Greek population(Reference Theofilogiannakou, Skouroliakou, Gounaris, Panagiotakos and Markantonis5, Reference Pechlivani, Vassilakou, Sarafidou, Zachou, Anastasiou and Sidossis10). Several characteristics of the child or mother as well as environmental factors have been associated with poor breast-feeding practices in studies conducted in other European countries and the USA. Included among these risk factors are maternal age, level of education, smoking, parity, infant health, hospital practices and social support(Reference Dennis12, Reference Scott and Binns13). Some of these factors have been studied locally in large urban areas of Greece but not in a representative sample of the population(Reference Theofilogiannakou, Skouroliakou, Gounaris, Panagiotakos and Markantonis5, Reference Pechlivani, Vassilakou, Sarafidou, Zachou, Anastasiou and Sidossis10).
Another important factor that has been shown to be associated with failure to breast-feed is maternal obesity. Recent studies have shown that initiation and duration of breast-feeding are poor among mothers who are classified as overweight or obese according to their BMI (kg/m2) before pregnancy(Reference Li, Ogden, Ballew, Gillespie and Grummer-Strawn7, Reference Li, Jewell and Grummer-Strawn14, Reference Donath and Amir15). Similar results are found for the effect of gestational weight gain on breast-feeding initiation and duration(Reference Li, Jewell and Grummer-Strawn14, Reference Hilson, Rasmussen and Kjolhede16).
The aim of the current paper is to provide information on the characteristics of mothers who fail to initiate breast-feeding as well as on the factors that contribute to longer duration of breast-feeding in Greece. Additionally, the paper aims to identify how maternal obesity is associated with both initiation and duration of breast-feeding in the Greek population.
Methods
Sampling
The study design has been presented in detail elsewhere(Reference Manios17) and only a brief description of the sampling and the methods used are presented here. This cross-sectional study involved Greek preschool children aged 12 to 60 months participating in the GENESIS (Growth, Exercise and Nutrition Epidemiological Study In preSchoolers) study, which was carried out from April 2003 to July 2004. A representative number of randomly selected public and private nurseries, including day-care centres, within municipalities in five counties (Attica, Aitoloakarnania, Thessalonica, Halkidiki and Helia) were invited to participate in the study. All nurseries invited to participate responded positively. The sampling of the nurseries was random, multistage and stratified by the total population of children, according to data provided by the National Statistical Service of Greece (Census 1999). Furthermore, an extended letter explaining the aims of the current study and a consent form were provided to each parent or guardian having a child in these nurseries. Those parents agreeing to participate in the study had to sign the consent form and provide their contact details. Approval to conduct the study was granted by the Ethical Committee of Harokopio University of Athens and by all municipalities invited to participate in the study.
Among the total number of nursery schools studied (n 115), sixty-three were in Attica, ten were in Thessalonica, twelve were in Halkidiki, twenty-two were in Aitoloakarnania and eight were in Helia. The selected counties are widely scattered over the Greek dominion while their overall local population comprises about 70 % of the total Greek population (Census 1999). Signed parental consent forms were collected for 2518 children, aged 1 to 5 years (response rate 75 %). From the total number of positive responses complete data became available for 2374 children with participation rate varying from 54 % to 95 %, reaching the highest rates in rural areas and the lowest ones in urban areas. After adjusting for parental age and educational level of the population agreeing to participate in the study, no significant differences were observed between the overall population characteristics and the study sample within the counties according to data provided by the National Statistical Service of Greece (Census 1999). Thus, the final sample is representative of 70 % of the Greek population with the proportion of children aged 12–23, 25–35, 36–47 and 48–60 months being 8·7 %, 21·7 %, 38·8 % and 31·4 %, respectively.
Additional information obtained from parents
During the morning interview at the nursery, additional information was obtained from the guardians with respect to: (i) parental demographic characteristics, such as age and educational level; (ii) parental anthropometric data, such as stature and body weight; (iii) gestational age of the infant in weeks; (iv) multiparity or uniparity; and (v) smoking and alcohol consumption during pregnancy. Furthermore, parents were asked to bring with them the medical record of their child, from which information on the child’s weight and length for the first 12 months of life was copied. Data on maternal weight status before conception, before and after childbirth, as well as on the child’s feeding patterns from birth to 6 months of age, were collected from the mothers of the children during either a face-to-face or telephone interview.
Classification of maternal weight before pregnancy and gestational weight gain
Self-reported weight before pregnancy and height were used to calculate BMI before pregnancy; it was then categorized as underweight (<19·8 kg/m2), normal weight (19·8–26·0 kg/m2), overweight (>26·0–29·0 kg/m2) or obese (>29·0 kg/m2) according to the guidelines from the Institute of Medicine (IOM)(18). Gestational weight gain was self-reported by the mother and categorized according to the IOM recommendations(18) as: (i) below IOM (<12·5, <11·5, <7·0 and <6·0 kg for underweight, normal-weight, overweight and obese women, respectively); (ii) above IOM (>18·0, >16·0, >11·5 and >9·1 kg for underweight, normal-weight, overweight and obese women, respectively); and (iii) within IOM (weight gain between the cut-off values for below and above IOM recommendation).
Definition of children’s gestational weight
The Nutstat module of the EpiInfo Database and Statistics Software for Public Health Professionals (Centers for Disease Control and Prevention (CDC), Atlanta, GA, USA) was used to determine the age- and sex-specific percentiles for weight according to the CDC 2000 Growth Charts(19). Using the CDC weight-for-age growth charts children were classified according to their birth weight as small for gestational age (≤10th percentile), appropriate for gestational age (10th–90th percentile) and large for gestational age (≥90th percentile).
Statistical analyses
All variables are reported categorically. To test the effect of the factors under investigation on failure to breast-feed, univariate logistic regression analyses were used and data were modelled using multiple logistic regression analysis. Odds ratios with 95 % confidence intervals were computed from the results of the logistic regression analyses. Linear regression analyses were used to estimate the associations of the different factors with duration of breast-feeding for mothers who initiated, after adjusting for possible cofactors. Gestational weight gain correlated significantly in univariate analyses with maternal weight status before pregnancy and was examined separately both in the logistic and the linear regression models because model diagnostics with these two parameters in the models indicated that the regression estimates were highly collinear. The two models created, in both the logistic and linear regression analyses, were adjusted for birth weight for gestational age, gestational age, parity, maternal age, maternal education and maternal smoking habits during the third trimester of pregnancy. Hypothesized interactions of variables in the models were not significant. All P values reported are two-tailed. Statistical significance was set at 0·05 and analyses were conducted using the STATA statistical software package version 8·0 (StataCorp, College Station, TX, USA).
Results
The characteristics of the study population are shown in Table 1. In the total population almost two-thirds (67·5 %) had a normal BMI before pregnancy, and 68·2 % did not meet the recommended weight gain during pregnancy (37·1 % gained less and 31·1 % gained more than the IOM recommended amount). Similar were the results for women included in the multiple logistic regressions. The multiple linear regression analyses on breast-feeding duration included the women who initiated breast-feeding and had complete data on breast-feeding duration (n 853). The distribution of maternal obesity in this subgroup was similar to that of the total population.
Descriptive characteristics of the study population in multiple logistic regression (n 1897) and multiple linear regression (n 853) analyses: cross-sectional GENESIS (Growth, Exercise and Nutrition Epidemiological Study In preSchoolers) study involving 2374 Greek preschool children aged 12 to 60 months, April 2003 to July 2004
*Categorized according to guidelines from the Institute of Medicine (IOM): below IOM, weight gain <12·5, <11·5, <7·0 and <6·0 kg for underweight, normal-weight, overweight and obese women, respectively; above IOM, weight gain >18·0, >16·0, >11·5 and >9·1 kg for underweight, normal-weight, overweight and obese women, respectively; within IOM, weight gain between the cut-off values for below and above IOM.
Figure 1 shows the percentage of infants who were never breast-fed by maternal gestational weight gain categorized according to IOM guidelines(18). The percentage of mothers who had gestational weight gain above the IOM recommendation and never breast-fed their children was significantly higher than that of mothers who had gestational weight gain within the recommended range and never breast-fed their children (53·8 % v. 45·1 %, P = 0·018).
Percentage of infants who were never breast-fed by maternal gestational weight gain categorized according to guidelines from the Institute of Medicine (IOM) (below IOM: weight gain <12·5, <11·5, <7·0 and <6·0 kg for underweight, normal-weight, overweight and obese women, respectively; above IOM: weight gain >18·0, >16·0, >11·5 and >9·1 kg for underweight, normal-weight, overweight and obese women, respectively; within IOM: weight gain between the cut-off values for below and above IOM). Data from the cross-sectional GENESIS (Growth, Exercise and Nutrition Epidemiological Study In preSchoolers) study involving 2374 Greek preschool children aged 12 to 60 months, April 2003 to July 2004. Value was significantly different from that of the reference group: *P = 0·018
Figure 2 shows the percentage of infants who were never breast-fed by maternal BMI values before pregnancy categorized according to IOM guidelines(18). The percentage of mothers who were overweight and obese before pregnancy and never breast-fed their children was higher than that of mothers who were of normal weight before pregnancy and never breast-fed their children (58·6 % and 79·2 % v. 45·9 %, P = 0·017 and P < 0·001, respectively).
Percentage of infants who were never breast-fed by maternal BMI value before pregnancy categorized according to guidelines from Institute of Medicine (<19·8 kg/m2, underweight; 19·8–26·0 kg/m2, normal weight; >26·0–29·0 kg/m2, overweight; >29·0 kg/m2, obese). Data from the cross-sectional GENESIS (Growth, Exercise and Nutrition Epidemiological Study In preSchoolers) study involving 2374 Greek preschool children aged 12 to 60 months, April 2003 to July 2004. Value was significantly different from that of the reference group: *P = 0·017, ‡P < 0·001
The adjusted odds ratios in multiple logistic regression analyses of failure to initiate breast-feeding for different characteristics are shown in Table 2. Obese mothers were shown to be 2·86 times more likely to fail in initiating breast-feeding. As revealed in the analysis, preterm delivery, multiparity, smoking during the third trimester of pregnancy and higher maternal education also decreased the likelihood of breast-feeding initiation.
Logistic regression models showing crude and adjusted* odds (with 95 % confidence intervals) of failure to initiate breast-feeding: cross-sectional GENESIS (Growth, Exercise and Nutrition Epidemiological Study In preSchoolers) study involving 2374 Greek preschool children aged 12 to 60 months, April 2003 to July 2004
*Adjusted for all other variables presented in the table.
†OR of 1·000 indicates the reference category of each variable.
‡Categorized according to guidelines from the Institute of Medicine (IOM): below IOM, weight gain <12·5, <11·5, <7·0 and <6·0 kg for underweight, normal-weight, overweight and obese women, respectively; above IOM, weight gain >18·0, >16·0, >11·5 and >9·1 kg for underweight, normal-weight, overweight and obese women, respectively; within IOM, weight gain between the cut-off values for below and above IOM.
The coefficients in Table 3 represent how many weeks less a mother with an abnormal BMI before pregnancy or an abnormal gestational weight gain breast-fed than did the reference group after all other factors were controlled for. The results show that among women who initiated breast-feeding, those with a pre-pregnancy BMI below 19·8 kg/m2 or those who smoked at the third trimester of pregnancy breast-fed their children for about 1·5 weeks less than their normal-weight or non-smoking counterparts, respectively. Similarly, multiparous women who initiated breast-feeding breast-fed their children for about 7 weeks less than women who were uniparous. In women who initiated breast-feeding, no significant differences in duration of breast-feeding were found between women of different gestational weight gains and women with different pre-pregnancy BMI status.
Multiple linear regression analyses for breast-feeding duration*: cross-sectional GENESIS (Growth, Exercise and Nutrition Epidemiological Study In preSchoolers) study involving 2374 Greek preschool children aged 12 to 60 months, April 2003 to July 2004
*Adjusted for all other variables presented in the table.
†– indicates the reference category of each variable.
‡Categorized according to guidelines from the Institute of Medicine (IOM): below IOM, weight gain <12·5, <11·5, <7·0 and <6·0 kg for underweight, normal-weight, overweight and obese women, respectively; above IOM, weight gain >18·0, >16·0, >11·5 and >9·1 kg for underweight, normal-weight, overweight and obese women, respectively; within IOM, weight gain between the cut-off values for below and above IOM.
Discussion
The current study has shown that almost two-thirds of women had a normal BMI before pregnancy while the percentage of women who gained inadequate and excessive weight during gestation reached 37·1 % and 31·1 %, respectively. The percentage of women with gestational weight gain within the IOM recommendations is comparable to those described in previous reports of gestational weight gain(Reference Brawarsky, Stotland, Jackson, Fuentes-Afflick, Escobar, Rubashkin and Haas20–Reference Strychar, Chabot, Champagne, Ghadirian, Leduc, Lemonnier and Raynauld22), but more recent studies depict a trend for even lower percentages of women to have gestational weight gain falling within the IOM recommendations along with higher percentages of women with excessive gestational weight gain(Reference Helms, Coulson and Galvin23). Among a longitudinal cohort of pregnant women in San Francisco, gestational weight gain was inadequate for 14 % and excessive for 53 %(Reference Brawarsky, Stotland, Jackson, Fuentes-Afflick, Escobar, Rubashkin and Haas20), while the results of a large retrospective study in the USA showed that 43·3 % of women had gestational weight gain above the IOM recommendations(Reference Stotland, Cheng, Hopkins and Caughey24).
In our study, 68·6 % of the mothers who provided data on breast-feeding duration had a recall period of less than 48 months. According to a recent review, maternal recall on initiation and duration of breast-feeding can provide accurate estimates especially when duration is recalled over a relatively short period of time (≤3 years)(Reference Li, Scanlon and Serdula25). However, it has been shown that retrospective data on breast-feeding duration for children who received exclusive or any breast-feeding are very strongly correlated (r = 0·94 and r = 0·95, respectively) with the actual data even 8 years after child’s birth(Reference Vobecky, Vobecky and Froda26). In this context, the retrospective data obtained in our study can reliably depict duration of breast-feeding in Greece.
Overall, our findings suggest that obese mothers are less likely to initiate breast-feeding than women with normal weight before pregnancy. The percentage of mothers with increased pre-pregnancy BMI who did not breast-feed was significantly higher than that of their normal-weight counterparts. Findings are similar in other studies, revealing that, of mothers who ever put their infants to the breast, those who were overweight or obese had less success initiating breast-feeding than did their normal-weight counterparts(Reference Li, Jewell and Grummer-Strawn14, Reference Donath and Amir15, Reference Hilson, Rasmussen and Kjolhede27, Reference Amir and Donath28).
Our study showed no significant differences in duration of breast-feeding between women with excessive gestational weight gain or high pre-pregnancy BMI status compared with those with a normal weight gain or normal BMI status before pregnancy. In those women who initiated breast-feeding, low body weight before pregnancy, smoking at the third trimester of pregnancy and multiparity were shown to negatively affect the duration of breast-feeding. The results of several large studies have shown a negative effect of obesity on breast-feeding initiation and duration, even after adjusting for possible confounding factors(Reference Hilson, Rasmussen and Kjolhede16, Reference Hilson, Rasmussen and Kjolhede27, Reference Oddy, Li, Landsborough, Kendall, Henderson and Downie29–Reference Chapman and Perez-Escamilla31). Other cohort studies, some conducted in countries with high initiation rates such as Denmark or Russia, found no difference in breast-feeding initiation or duration according to maternal obesity(Reference Michaelsen, Larsen, Thomsen and Samuelson32–Reference Baker, Michaelsen, Rasmussen and Sorensen35). These inconclusive results could be attributed to the different study designs, definitions of breast-feeding and overweight, but also to other confounding factors for which the analyses were controlled.
The mechanisms lying beneath the association of maternal obesity with poor breast-feeding practice are still unclear. Overweight/obese women have been shown to be more likely to have late arrival of milk than normal-weight women(Reference Rasmussen, Hilson and Kjolhede36, Reference Dewey, Nommsen-Rivers, Heinig and Cohen37) while the prolactin response to suckling has been shown to be blunted in obese women, providing support for a possible biological association between maternal obesity and the duration of breast-feeding(Reference Rasmussen and Kjolhede38). Alternatively, women with large breasts may have practical/mechanical difficulties in breast-feeding their babies(Reference Chapman and Perez-Escamilla31, Reference Walker39). In addition, several psychosocial factors related to obesity, such as low self-esteem or higher rates of postpartum depression, may play a role in the failure to breast-feed(Reference Chapman and Perez-Escamilla31, Reference Lacoursiere, Baksh, Bloebaum and Varner40–Reference Hill and Williams42).
Still, we acknowledge that there is a potential limitation of the current study; among our study cohort only 13·4 % of women were overweight or obese before their pregnancy, a rather low prevalence in comparison to published reports for the corresponding population(Reference Manios17, Reference Lobstein, Rigby and Leach43). This could be attributed to deliberate under-reporting, over-reporting or recall bias for the self-reported pre-pregnancy body weight and height(Reference Villanueva44). This is a common limitation in similar studies(Reference Brawarsky, Stotland, Jackson, Fuentes-Afflick, Escobar, Rubashkin and Haas20).
However, several studies have shown high correlations between stated and actual pre-pregnancy weights(Reference Roberts45–Reference Troy, Hunter, Manson, Colditz, Stampfer and Willett47). Self-reports and measurements of body weight are highly correlated (r = 0·86–0·99), varying, on average, by only 1·1 kg to 2·4 kg(Reference Stevens-Simon, Roghmann and McAnarney48, Reference Stewart49). In addition, self-reported weights provide reasonable estimates of BMI category since the reporting error (1–2 kg) is a small percentage of total body weight, and age-related changes in height are not a factor for prenatal populations(Reference Harris and Ellison50). The good correlation between reported and measured preconception body weights is likely to minimize misclassification bias by BMI category(Reference Gunderson and Abrams51).
Conclusions
In conclusion, the present study shows that mothers with high pre-pregnancy BMI are less likely to initiate breast-feeding while high gestational weight gain has no effect on either initiation or duration of breast-feeding in Greece. Other socio-cultural factors, such as maternal education and smoking, were also shown to be important for the initiation of breast-feeding. Health professionals should take into account that obese women need extra support to initiate and sustain breast-feeding when developing breast-feeding promotion programmes and policies.
Acknowledgements
The authors would like to thank Evdokia Oikonomou, Vivian Detopoulou, Christine Kortsalioudaki, Margarita Bartsota, Thodoris Liarigkovinos and Christos Vassilopoulos for their contribution to the completion of the study. The study was supported by a research grant from Friesland Foods Hellas.
Contribution of authors: Y.M. was responsible for the study design and supervision of the field study. E.G. carried out the data collection and performed the statistical analyses. K.K., E.I., A.A. and M.B. carried out the data collection. All authors contributed to the writing of the manuscript and interpretation of the data, reviewed its content and approved the final version submitted for publication.
Conflict of interest: The GENESIS study was supported with a research grant from Friesland Foods Hellas. Y.M. works as a part-time scientific consultant for Friesland Foods Hellas. None of the other authors had any personal or financial conflict of interest. The study sponsor had no interference in the study design, data collection or writing of the manuscript.
