Breast bud detection: a validation study in the Chilean Growth Obesity Cohort Study

Background: Early puberty onset has been related to future chronic disease; however breast bud assessment in large scale population studies is difficult because it requires trained personnel. Thus our aim is to assess the validity of self and maternal breast bud detection, considering girl's body mass index (BMI) and maternal education. Methods: In 2010, 481 girls (mean age = 7.8) from the Growth and Obesity Chilean Cohort Study were evaluated by a nutritionist trained in breast bud detection. In addition, the girl(n = 481) and her mother(n = 341) classified the girl's breast development after viewing photographs of Tanner stages. Concordance between diagnostics was estimated (kappa, Spearman correlation) considering girls' BMI and mother's educational level. Results: 14% of the girls presented breast buds and 43% had excess weight (BMI z-score > 1, World Health Organization 2007). Self-assessment showed low concordance with the evaluator (K < 0.1) and girls with excess weight over-diagnosed more than girls of normal weight (44% vs. 24%, p-value < 0.05). Instead, mothers showed good concordance with the evaluator (K = 0.7, 95% confidence interval (CI) = 0.6-0.9), even in overweight girls and/or in mothers with low education (K = 0.7, 95% CI = 0.6-0.8). Conclusions: Mothers were able to adequately evaluate the appearance of breast bud despite low educational level and girls' excess weight. Mother could be a useful resource for defining puberty onset in epidemiological studies, particularly developing countries.