Concept of Trajectory Modelling using the study

Concept of Trajectory Modelling using the study "The associations between early childhood BMI trajectories and body composition and cardiometabolic markers at age 10 in the Ethiopian iABC birth cohort".

The study uses latent class trajectory modeling to identify distinct BMI growth patterns from birth to 5 years (to understand growth dynamics) and is discussed in the context of low- and middle-income countries where the double burden of malnutrition is a concern

1. Concept of Trajectory Modelling - Associations of early childhood body mass index trajectories with body composition and cardiometabolic markers at age 10 years: the Ethiopian infant anthropometry and body composition ( iABC ) birth cohort study University of Copenhagen Jimma University, Ethiopia

2. The Global Health Problem of Childhood Overweight Childhood overweight is a major global health problem (37 million Under 5) - key risk factor for cardiovascular disease and type 2 diabetes in adulthood. - In high-income countries, accelerated BMI growth in early childhood has been linked with overweight and higher cardiometabolic (CM) risk later in life. -In middle-income countries, rapid weight and BMI gain in infancy and childhood have been associated with greater lean mass ( ) rather than fat mass in childhood and adulthood.

3. Importance of Understanding BMI Growth Patterns Early BMI growth trajectories better predicted later body composition and risk of obesity in childhood than a single-point time BMI. How early BMI associated with body composition, adiposity, and cardiometabolic risk track in later childhood in LMICs - to identify those at risk and provide timely interventions.

4. Aim & hypotheses Previous Findings: Identified 4 distinct BMI growth patterns in Ethiopian iABC birth cohort from ages 0 to 5. High early BMI growth linked to larger body size, higher FFM and FM, and triglycerides. Current Study Aim: Examine associations of BMI growth patterns with age 10 body measurements, composition, abdominal fat, and disease markers. Hypotheses: - Rapid BMI growth may lead to higher FM and markers related to lipid metabolism at age 10. - Slow BMI growth may result in lower FFM, body fat levels, and markers related to lipid metabolism at age 10.

5. Covariates Child anthropometry – 1.5, 2.5, 3.5, 4.5 , 6 mo & 1, 1.5, 2,3,4,5, 10 yrs. Sociodemographic Characteristics: Information obtained within 48 hours after delivery, including maternal age, highest educational status, and family wealth status. Child Gestational Age: Calculated using the Ballard score. Family Economic Status: Assessed using Wealth Index. Maternal Anthropometric Measurements: Height, weight Breastfeeding Status: Assessed between 4 and 6 months (EBF/predominant/Partial/Not breastfed)

6. Study Design and Participant Selection - iABC birth cohort established in December 2008 in Jimma town, Ethiopia. - Mothers giving birth in the maternity ward of Jimma University Specialized Hospital, and their newborns recruited within 48 hours after birth. - Eligibility: living in Jimma town, healthy and term newborn (≥37 weeks of gestation) with a weight of ≥1,500 g , and without any medical complications and congenital malformations. - From 0-5 years of age, the children were invited for a total of 12 visits.

7. Follow-up and Data Collection Follow-up visit conducted from June 2019 to December 2020, when children were 7-12 years old. Data collection procedures included overnight fast.

8. Statistical Analyses Outcomes studied at the 10-year follow-up included anthropometric measurements, body composition, abdominal fat, and cardiometabolic markers. Associations between categorical exposure variables (latent BMI trajectories from 0-5 years) and the continuous outcomes at 10 years, analyzed using multiple linear regression.

9. Latent Class Trajectory Modeling (LCTM)

10. Latent Class Trajectory Modeling (LCTM)

11. One size fits all Growth Parameters: These are the intercept (initial status) and slopes (change per unit of time) that describe growth in a sample. Assumptions: The model assumes a certain distribution of variance around these growth parameters, typically a normal distribution. Adjusted Predictions with 95% CIs

12. Latent classes Intercept Slope Var 1 Var 2 Var 3 Var 4 Time 0 Time 1 Time 2 Time 3

13. - Regression analyses were performed as complete case analyses (children with complete data on all covariates). - Multiple imputations were performed to impute missing data for children who attended the 10-year follow-up.

14. Model Adjustments Model 1 Child’s sex, exact age at the 10-year follow-up visit Model 2 Model 1 + child’s birth order, gestational age at birth, maternal age at delivery, maternal height, maternal highest educational status, family socioeconomic status (wealth index) Model 3 Model 2 + birth weight Model 4 Model 3 + current body size measurements

15. Participant Flowchart

16. Key Findings slow and rapid growth to high BMI trajectory had greater waist circumference and FMI. slow growth to high BMI had greater abdominal subcutaneous fat. stable low BMI had lower FFMI at 10 years of age. slow growth to high BMI trajectory had higher insulin and HOMA-IR. rapid growth to high BMI had higher C-peptide concentration. rapid growth to high BMI had lower total cholesterol concentration at 10 years.

17. Limitations of the Study High loss to follow up unmeasured exposures that can affect growth and body composition (between 5 and 10 yrs ) Breastfeeding not included in the model (? Literature) No causal-effect associations. Those who did not attend 10 yr follow up are much healthier (bias?) Not representative of the general population - healthy and term children from a tertiary centre .