Purchase this article with an account.
Yanxian Chen, Billy Heung Wing Chang, Mingguang He; Patterns in longitudinal growth of refraction in Southern Chinese children: cluster and principal component analysis. Invest. Ophthalmol. Vis. Sci. 2015;56(7 ):2938.
Download citation file:
© ARVO (1962-2015); The Authors (2016-present)
Refraction progression was studied simply as progression rate and fitted into predefined statistics models in previous literatures. We tried to use hypothesis-free analysis to investigate the patterns in refraction growth process in Chinese children, and to explore the possible risk factors of different components in progression.
The first-born twins (n=637, 51.3% girls) in Guangzhou Twin Eye Study with 6-year annual visit data were included in the analysis. The mean age at baseline was 10.6 ± 2.3 years. Measurements included cycloplegic autorefraction, autorefraction without cycloplegia for parents, anthropometry, questionnaire addressing near work and outdoor activities. A clustering method was conducted to group similar growth patterns. A principal component analysis (PCA) was used to extract the main differences in how the refraction evolves over time. Main outcome measures were cluster groups identified by cluster analysis, components extracted by PCA and the associations between risk factors and PCA scores.
Clusters 1 to 3 were classified according to the cluster analysis, representing steady, slow and fast progressing patterns of refraction respectively. Baseline age, baseline spherical equivalence (SE), paternal SE, maternal SE and proportion of two myopic parents were significantly different across the three groups. Three major components were identified using PCA: ‘Average refraction’, ‘Acceleration’ and ‘Stabilization’. Younger children with more myopic SE tended to have greater ‘Acceleration’ (both p<0.001 ) when baseline SE was ≤0 diopter. After controlling for baseline age and SE, changes of height and weight (both p<0.001), near work time (p=0.020) and having two myopic parents (p=0.028) were positively related to ‘Acceleration’, while paternal SE (p=0.047) and maternal SE (p=0.012) were negatively associated with ‘Acceleration’ respectively. Positive relationships between ‘Stabilization’ and female, changes of height and weight (p<0.001 for all) were found.
We revealed the variation in refraction growth process of children, and found out refraction progression consists of three major components. The results indicated that both genetic and environmental risk factors may contribute to the acceleration but not stabilization of myopia in children.
This PDF is available to Subscribers Only