Latent class analysis to identify childhood predictors of abnormal glycemic status in young adults with cystic fibrosis

Abstract

AIMS: Data on the clinical course of patients with cystic fibrosis (CF) from childhood to CF-related diabetes (CFRD) diagnosis in adulthood are limited. We evaluate whether childhood trajectories of parameters of interest in CF are associated with the risk of abnormal glucose tolerance (AGT) in early adulthood. Methods: Pediatric and adult data from 108 subjects with CF followed annually were paired. Participants were grouped according to predominant childhood trajectories for weight, height, body mass index, lung function, glycated hemoglobin levels, fasting glycemia, and 2h post-oral glucose tolerance test glucose levels. Multivariable logistic regression was performed to identify parameters that predict glucose tolerance status in adulthood. Results: Univariate analyses reveal that the risk of developing an AGT in adulthood is greater in subjects who are homozygous vs. heterozygous for the ΔF508 mutation, have pancreatic insufficiency vs. sufficiency, or have higher fasting glycemia values at 10 years old rising rapidly vs. lower values that are gradually rising until 17 years old. Multivariable logistic regression retains only fasting glycemia as a significant predictor for the occurrence of AGT in adulthood. Conclusions: Fasting glycemia may be a clinical marker of interest to better target children with CF at risk of developing an AGT in early adulthood.

Keywords

• Abnormal glucose tolerance
• Adult cohort
• Cystic fibrosis
• Growth curves
• Pediatric cohort