The ABIS-study
All parents with children born between October 1st in 1997 and October 1st in 1999 in Southeast Sweden (N = 21.700) were asked to participate in the ABIS study (All Babies in Southeast Sweden), a prospective cohort study. Out of the 21.700 invited families 17.055 (78.6%) gave their informed consent to participate. Questionnaires were answered at birth and then at the age of 1, 3, 5, 8, 12-13 years of age, and different biological samples were collected at the follow-ups. During the first year of life a diary was used for daily registration of certain nutritional data, infections etc. In the questionnaires, parents were asked for duration of breastfeeding. Exclusive breastfeeding means no other food than breast-milk. Partial breastfeeding is defined as breastfeeding in addition to formula or other food. In this analyses we focus on total breastfeeding, defined as the duration of any breastfeeding (exclusive and partial breastfeeding). The registration by the mothers has been validated by comparison with breast-feeding data registered at the well-baby clinics showing very high agreement.
In earlier follow-up of questionnaire data (from the whole ABIS cohort) one year after birth showed that 78.4% of infants were exclusively breast-fed at 3 months, 10.1% at 6 months and 3.9% up to 9 months. Partially breast-fed children were 68.9% at 6 months, while 43.6% were partially breast-fed to at least 9 months of age (14). The median exclusive breastfeeding duration was four months and the median duration for total breastfeeding eight months. Maternal smoking, high maternal BMI, and being a single mother were associated with short-term exclusive breastfeeding. Both maternal and paternal age was positively associated with the duration of both exclusive and partial breastfeeding (Huus 2008).
In this study we analyzed a subpopulation of ABIS-children and parents since we were interested in studying different psychosocial factors impact on long-term HPA-axis activity i.e prolonged stress measured by cortisol in hair (8). For 126 ABIS children at 8 years of age we had sufficient hair samples to be analysed and also reliable information about breastfeeding, as well as other relevant data of the child, mother and family. Duration of breastfeeding was measured through a questionnaire to the mothers one year after the childbirth, when she answered about the duration in months of exclusive and partial breastfeeding. Possible confounding factors include child´s gender, child´s birth weight (low, normal and high), gestational age (early, normal or over time), type of delivery (vaginal, Cesarean section or other problems), mother parity (first born or earlier parity), age of mother at delivery (range 18-40 years), if mother smoked during pregnancy, and a composite index of early psychosocial vulnerability in the family including different indicators: single mothers, unemployment, low family income, low parental education, parents born abroad, mother experience of serious life events, not feeling safe or lack of social support during pregnancy (4).
Measures of cortisol in hair
Trained staff cut strands of hair from the posterior vertex area of the participants’ heads in accordance with guidelines published by the Society of Hair Testing (15). The hair was then enclosed in sealed plastic tubes marked with identification numbers and stored in room temperature until analysis. All analyses were performed at the laboratory of Clinical Chemistry at the University of Linköping. The first 3 cm of outgrowth were analyzed for cortisol concentrations using a competitive radioimmunoassay in methanol extract (Morelius 2004). At least 3 mg of hair was required for reliable measurements. The samples were dissolved in radioimmunoassay buffer and analyzed. Hair samples between 3-10 mg were required to maintain a total inter-assay coefficient of variation below 8% for hair extraction and measurement of cortisol by the radioimmunoassay. The method is fully described elsewhere (8,16)).
Statistical analysis
Statistical analyses was made by using the Statistical Package for the Social Sciences software, version 23 (IBM SPSS Statistics, IBM Corporation). The measured cortisol values were logarithmised before the statistical analysis due to possible skewness in the distribution. In the univariate tests analysis of variance (ANOVA) was used for comparisons of mean. For the association between continued variables, Pearson’s correlation coefficient was applied. Variables statistically significant in the univariate analysis were included in a multivariate linear regression model. A p-value <0.05 was considered statistically significant.