Prediction of incident metabolic syndrome and dysglycaemia by adiposity and insulin resistance indices in Chinese women with polycystic ovary syndrome: a 4-year longitudinal study

Abstract

INTRODUCTION: Women with polycystic ovary syndrome (PCOS) are at increased risk of developing metabolic syndrome (MetS) compared to the general population. The visceral adiposity index for Chinese (CVAI), lipid accumulation product (LAP), triglyceride-glucose (TyG) index and triglyceride glucose-body mass index (TyG-BMI) were good clinical surrogate markers for visceral adiposity and insulin resistance (IR) as validated in previous studies. These are established composite indices derived from some components of the MetS. The present study aimed at investigating if CVAI, LAP, TyG and TyG-BMI were good predictors of incident MetS and dysglycaemia in women with PCOS. METHODS: This prospective observational study was based on metabolic data obtained from 459 women diagnosed with PCOS according to the Rotterdam criteria. These subjects were recruited in the Hong Kong Chinese population between 2010 and 2013. Subjects were recruited from a university hospital or community family planning clinics. Blood was taken after 8-hour fasting for metabolic assessment. Of them, 289 attended a follow-up visit around 4 years after recruitment with metabolic screening repeated. MetS was defined according to the joint interim statement (Alberti et al, 2009), and dysglycaemia was defined by the American Diabetes Association criteria. Prediction on incident MetS and dysglycaemia by CVAI, LAP, TyG and TyG-BMI were analysed by ROC curve. RESULTS: At recruitment, the prevalences of pre-existing MetS and dysglycaemia were 72 (15.7%) and 58 (12.6%) out of 459 PCOS subjects respectively. A total of 289 subjects attended follow-up at a median of 49 (25-75th percentile 48-51) months to repeat metabolic screening. At follow-up, a further 29 (7.5%) of the remaining subjects developed incident MetS, and 18 (6.2%) developed incident dysglycaemia. For the prediction of incident MetS, the areas under the ROC curve (AUROC) were highest for TyG-BMI (0.883, 95% CI 0.832 - 0.922), followed by CVAI (0.874, 95% CI 0.822 - 0.915) and LAP (0.843, 95% CI 0.788 - 0.889), which were significantly higher than TyG (0.736, 95% CI 0.672 - 0.793) (p<0.05). For predication of dysglycaemia, the AUROCs were significantly higher for TyG-BMI (0.765, 95% CI 0.703 - 0.819) and CVAI (0.763, 95% CI 0.702 - 0.818) compared to LAP (0.699, 95% CI 0.633 - 0.759) or TyG (0.619, 95% CI 0.551 - 0.683) (p<0.05). CONCLUSIONS: TyG-BMI and CVAI are the best clinical markers that predict incident MetS and dysglyaemia in women with PCOS. Disclosures and sources of support: The authors have no conflict of interest to declare. This study was supported by a grant from the Hong Kong Obstetrical and Gynaecological Trust Fund as well as internal research funding of the Department of Obstetrics and Gynaecology, The University of Hong Kong.