Search

The objective of this study was to estimate the frequency of breastfeeding and to identify variables associated with median breastfeeding length. It is a cross-sectional population-based study carried out in a representative sample of 261 children under 2 years old randomly selected in the urban area of Itupeva city, SP, Brazil. The breastfeeding patterns recommended by WHO were used. The frequency and the median length of breastfeeding were estimated by life tables technique. The Wilcoxon test was used to identify association between breastfeeding and the categorical variables (p0.05) in the bivariate analysis. For the multivariate analysis, the variables associated with the breastfeeding length (Wilcoxon test; p0.10) were inserted in a Cox Regression model. Almost 100% of the children have initiated breastfeeding and 41.0% were in breastfeeding at the moment of the interview. The median length of breastfeeding was 7.2 months and exclusive breastfeeding was only 28 days. It was observed association between the breastfeeding length and the following variables: father's schooling, mother's marital status, number of children, sequence of birth, bottle-feeding and pacifier use. The Cox regression model selected as significant variables: sequence of birth (HR=1.68; IC95%=1.03-2.73), bottle-feeding use (HR=15.20; IC95%=3.69-62.01) and pacifier use (HR=2.84; IC95%=1.95-4.19). The current pattern of breastfeeding in Itupeva is still far from the WHO recommendations; therefore, encouragement and support to breastfeeding need to be intensified. Variables influencing breastfeeding should be considered in the attempt to increase breastfeeding. Health workers should dedicate all their efforts to increase the exclusive breastfeeding period in children under 6 months of age and increase the total breastfeeding period.