Body composition patterns and breast cancer risk in Chinese women with breast diseases: A latent class analysis.

ABSTRACT

AIMS: To identify unobserved body composition patterns among Chinese women with breast diseases using latent class analysis (LCA) and to examine the relationship between body composition patterns and breast cancer (BC) risk. DESIGN: A descriptive, cross-sectional study. METHODS: Female patients (N = 1816) with breast diseases were included in the study from April 2016 - March 2017. Body composition measures were done by the bioelectrical impedance analysis. The LCA models were estimated using Mplus 8.1. RESULTS: Four latent classes were identified based on water, protein, minerals and body fat mass Class 1 - Low Muscle Mass class; Class 2 - High Body Composition class; Class 3 - High Fat class; and Class 4 - Normal Body Composition Class. Classes 2 and 3 are higher risk classes for developing BC compared with the other two classes (p < 0.05). Overall, age is positively associated with the odds of BC development (p < 0.001). However, age effect depends on the body composition patterns. Age effect on the odds of BC was statistically significant only for women who had least body fat mass (Class 1, OR = 1.110, 95% C.I. 1.084-1.136) or had normal body composition (Class 4, OR = 1.090, 95% C.I. 1.036-1.147). The effect of age was not statistically significant if women had higher risk body composition (e.g., in Classes 2 or 3). CONCLUSION: Latent Class Analysis is a useful person-centred analytical approach for identification of unobserved patterns of body composition and it could be used to predict the risk of BC and develop personalized interventions for body composition studies.