Clustering Growth Trajectories Leading to Early Childhood Obesity in Groups and Characterizing Their Co-Morbidities

• Activity Number: 315
• Type: Contributed
• Date/Time: Tuesday, August 11, 2015 : 8:30 AM to 10:20 AM
• Sponsor: Section on Statistics in Epidemiology
• Author(s): Md Jobayer Hossain* and H. Timothy Bunnell and Samuel S. Gidding and Thomas H. Shaffer
• Companies: Nemours/A.I. DuPont Children's Hospital and Nemours and Nemours/A.I. DuPont Children's Hospital and Nemours
• Keywords: Cluster ; Piecewise ; Mixed Effects ; Random Coefficients ; Pathways ; Trajectories

Abstract

Early childhood growth patterns are interactive with co-morbidities and correlates of obesity. Understanding of distinct growth trajectories leading to obesity and their interaction with co-morbidities and correlates may explore this association. Birth to 5 years clinic-visit data were collected from 488 children identified obese at age of 5. Standardized scores of weight-for-length (age< 2 years) and BMI (>=2 years) were calculated and termed both BMIz. Loess plot reveals a cubic polynomial of mean BMIz over 5 yrs. A piecewise linear mixed effects model with knots at ages 8 and 31 months yielded best fit. A cluster analysis of random coefficients identified five heterogeneous growth patterns: sigmoid trend with abrupt weight gain(G1), early rapid weight gain and then stable(G2), low birth weight and then steady weight gain(G3), chronicle marginal obese(G4), bathtub(G5). A substantially higher prevalence of asthma, bronchitis, gastroenteritis in G1; dermatitis, respiratory abnormalities in G2; allergic rhinitis, neonatal jaundice, wheezing in G3; dermatophytosis in G4; Phimosis in G5 was observed. In short, pathways to early life obesity have their own underlying co-morbidities.