Growth dynamics of Indian infants using latent trajectory models in pooled survey datasets.

Background: National survey data show that age- and sex-standardized weight and length measurements decline early in Indian children. In population-level longitudinal data, early detection of growth trajectories is important for the implementation of interventions. We aimed to identify and characterize distinct growth trajectories of Indian children from birth to 12 months of age residing in urban and rural areas.

Methods: Pooled data from four interventional and non-interventional longitudinal studies across India were used for the analysis. Latent class mixed modeling (LCMM) was employed to identify groups of children with similar trajectories over age. The trajectories named Classes of Children were created for length-for-age Z scores (LAZ) and weight-for-age Z scores (WAZ) based on place of birth, residential area, and maternal education.

Results: We identified two latent classes for LAZ in boys and three latent classes for LAZ in girls, and four classes for WAZ were identified in both boys and girls. The first class for LAZ, with the highest proportion of children (>80% of children), did not decline or increase with age; In boys, Class 1 was close to the WHO median, whereas in girls, Class 1 was lower than the WHO median from birth. The LAZ classes of remaining boys and girls declined with age (slope, = - 1.04; 95% CI: -1.09, -0.99 for boys and = - 0.69; 95% CI: -0.76, -0.63 for girls). The first trajectory of WAZ (approximately 50% of children) for boys ( =0.13; 95% CI: 0.11, 0.16) and the second trajectory of WAZ for girls ( =0.24; 95% CI: 0.18, 0.30) increased with age, while the remaining trajectories of WAZ declined with age.

Conclusion: There is heterogeneity in the growth of Indian children in the first year of life, which was identified by distinct types of growth trajectories. The predominant trajectories of both LAZ and WAZ did not decline with age, while most other trajectories demonstrated an initial decline.