To estimate the association strength of dietary behaviour and sedentary habits in relation to childhood obesity in Spain.

A matched case–control study was carried out using data collected by sentinel network paediatricians in general practices.

Five Spanish autonomous communities.

Cases were 437 children (2–14 years old) with BMI >95th percentile according to Spanish reference tables. Controls were 751 children (2–14 years old; two paired per case) with BMI <84th percentile. Data were collected in two phases: individual (questionnaires filled in by sentinel paediatricians) and family (self-administered questionnaires filled in a family environment). Crude OR and adjusted OR (ORc and adj OR) for the given variables were calculated using a simple and multiple conditional logistic regression analysis.

The factors with the greatest effect on obesity were family history of obesity: both parents (adj OR = 11·2), mother but not father (adj OR = 9·1), father but not mother (adj OR = 6·1), siblings (adj OR = 2·7); and eating between meals (adj OR = 2·5) and consumption of sweets and soft drinks >2 times/week (adj OR = 2·0). The highest protection effect was found for five meals per day (adj OR = 0·5), the regular consumption of breakfast (adj OR = 0·5) and for eating fruit for dessert (adj OR = 0·6). Factors related to sedentary habits did not appear as noteworthy.

We have determined the association between certain dietary behaviour and family history with childhood obesity in several Spanish regions.