賴藤 貴志

To investigate the effect of exclusive breastfeeding on the likelihood of Japanese preschool children being overweight, population-based cross-sectional survey data from M town in Japan were used. Using the population registry of this town, all 616 preschool children were identified, and a self-administered questionnaire was sent to their parents. The exposure variable of interest was exclusive breastfeeding from birth to 6 months, and the outcome variable of interest was the children being overweight at preschool age. Statistical analyses used included logistic regression and sensitivity analyses. In the final analyses, we included 448 preschool children. Although all point estimates indicated a protective effect, logistic regression analyses showed no significant reduction in being overweight due to exclusive breastfeeding in the unadjusted model (odds ratio (OR)0.70, 95% confidence intervals:0.30-1.64), the model adjusted for birth weight (OR0.70, 95% CI:0.30-1.63), the model adjusted for child lifestyle (OR0.71, 95% CI:0.30-1.67), or the model adjusted for parental factors (OR0.46, 95% CI:0.15-1.37). In sensitivity analyses, point estimates were not significant, but a protective effect was observed. In conclusion, our results suggest that breastfeeding might have a protective effect on Japanese preschool children against being overweight, although statistical significance was not observed due to the limitation of the statistical power of the findings.

Obesity in children is a serious public health problem, and TV viewing is considered a potential risk factor. Since, however, no relevant association studies have been conducted in Japan, we evaluated the association between TV viewing and obesity using a population-based study conducted in a Japanese town. All 616 preschool children in the town were enrolled in February 2008, and a self-administered questionnaire to collect children's and parents' characteristics was sent to the parents. We dichotomized the time spent TV viewing and evaluated associations by logistic regression using a "less than 2h" category as a reference. The questionnaire was collected from 476 participants (77.3%), of whom 449 were available for the final analyses. Among them, 26.9% of preschool children reported 2 or more hours of TV viewing per day and 8.2% were defined as obese. In logistic regression analyses, there was no positive association in unadjusted (odds ratio [OR]1.11, 95% confidence interval [95% CI]:0.50-2.49) or adjusted models for exclusively breastfed status, sleep duration, or maternal factors (OR1.11, 95% CI:0.50-2.51). We also found no positive association between TV viewing and overweight status, possibly owing to the influence of social environment, low statistical power, or misclassification.