In our study, we directly compared the performance of FPG, CVAI, VAI, LAP, TyG, WC, and BMI in detecting AGT and observed FPG demonstrated the lowest AUC value (0.695) and the lowest sensitivity (49.2%) among all the indices for detecting AGT. Thus, although FPG is an inexpensive assay and does not require mathematical calculations, it may not reliably detect AGT in our PCOS population, a finding consistent with previous studies [3,5].
We next evaluate the abilities of six obesity- and lipid-related indices (CVAI, VAI, LAP, TyG, BMI, WC) as indicators of detecting AGT in PCOS patients. In line with 2 h-PG as criteria for AGT, our data demonstrated that CVAI outperformed other indices with a higher correlation coefficient and a larger AUC for AGT detection in Chinese women with PCOS.
CVAI, is known as Chinese VAI. VAI, was introduced to estimate visceral adipose function for the Caucasians and is an index established with the use of BMI, WC, TG, and HDL-c [10,24]. In Caucasians, VAI was used as a valuable indicator of visceral adiposity and adipose tissue dysfunction to predict the risk for cardiovascular diseases and insulin resistance [24]. However, a previous study showed that VAI was poorly associated with adipose tissue area and has poor diagnostic performance to predict prediabetes and diabetes in Chinese [16]. In conformity with previous findings, we showed that the AUC of VAI for AGT was 0.708, which was lower than the AUC of CVAI (0.758), LAP (0.736), TyG (0.724), and BMI (0.725). Moreover, data based on correlation analyses also revealed similar tendencies. The main reason of the discrepancy may be ethnic difference and subject characteristic. VAI was developed initially for Caucasians. However, compared to equivalent Caucasians, Asian subjects have a greater proportion of body fat for a given BMI level [25] and are more prone to accumulate fat around the abdomen [26].
The CVAI index based on the combination of age and all the parameters of VAI (BMI, WC, TG and HDL-C) was developed to estimate visceral fat area for Chinese by Xia and colleagues [16]. In their study, among 6495 middle-aged and elderly Chinese, CVAI provided higher AUCs for the diagnosis of T2DM and prediabetes than BMI, WC and VAI [16]. Also, two prospective studies indicated that CVAI proved a better predictor of prediabetes and diabetes compared to BMI, WC, and VAI in the general Chinese population [15,27]. However, no data are currently available establishing any association between CVAI and AGT. In our cohort of PCOS patients, CVAI provided the strongest correlation with 2 h-PG and the largest AUC for detecting AGT, thereby suggesting that CVAI might be a more useful indicator of AGT than LAP, TyG, BMI, VAI, and WC in Chinese women with PCOS. Although the calculation of CVAI requires more parameters and more complicated formulas, the parameters (age, BMI, WC, TG, and HDL-c) are easily available and low cost in clinical practice, and the complex algorithms can easily be performed by Tablets or Smartphone. In our study population, ROC curves analysis suggested a threshold value of 32.61 in CVAI for the detection of AGT (73.4% sensitivity, 69.9% specificity, and 0.758 AUC).
LAP, a novel index based on a combination of WC and TG, was first introduced for the U.S. National Health and Nutrition Examination Survey [23]. Furthermore, it is proposed as a valuable marker indicating T2DM and prediabetes [9,12]. However, few studies have explored the association of LAP with AGT. The results of previous studies indicated that LAP performed better than BMI in identifying abnormal glucose regulation in the young Korean women and provided higher AUC in estimating the risk for IGT than BMI and waist-to-hip ratio in Austrian PCOS women [12,19], which was accordance with our results. However, these studies did not compare the utility of LAP with other novel obesity- and lipid-related indices. Our analysis showed that LAP performed superior to TyG, VAI, BMI, and WC, but inferior to CVAI in detecting AGT in Chinese PCOS population. According to the maximized Youden index, the LAP 26.19 (with AUC 0.736, sensitivity 70.5%, specificity 68.9%) was found to be the cut-off point for AGT in our study population, which was much higher than that in Korean PCOS women (12.98) [20]. This discrepancy could be partially due to ethnic difference.
TyG index, a composite indicator composed of TG and FPG, is a key predictor of T2DM and prediabetes in severl epidemiological studies. The TyG performed better than homeostasis model assessment of insulin resistance (HOMA-IR) and TG/HDL-c in predicting T2DM in Korean adults without initial T2DM [28]. A study conducted among 3307 elderly Colombian individuals suggested better discriminative power of the TyG index than BMI, WC, and VAI to predict prediabetes [11]. However, OGTT was not performed, and prediabetes was defined as a FPG of 100 to 125 mg/dL (5.6 to 7.0 mmol/L) in this study. In a prospective cohort study conducted on 4543 Chinese individuals without initial prediabetes or diabetes, Wen et al. reported that TyG provided a larger AUC for predicting prediabetes and isolated IGT than FPG, WC, and BMI [8]. However, this study did not compare the ability of TyG with other novel obesity- and lipid-related indices. In the current study, TyG was found inferior to CVAI and LAP for detecting AGT based on AUC and correlation coefficient.
Due to the fact that a plasma insulin assay is not yet available in all laboratories, has poor reproducibility and is costly [29], we do not evaluate the performance of the indices those using insulin for detecting AGT. Indeed, they do not perform better than obesity- and lipid-related indices studied in our report (data not shown).
Despite these relevant findings, it is important to point out the limitation of our study. All the subjects were recruited from Xuzhou Central Hospital, and only those with complete medical records. This population may not represent all the Chinese women with PCOS, with a possibility of bias in the results.
In conclusion, CVAI performed better than LAP, TyG, VAI, BMI, WC, and FPG for detecting AGT in Chinese women with PCOS. However, due to the lower sensitivity and specificity of CVAI, it may be used as a screening indicator of AGT in Chinese women with PCOS when it becomes unfeasible to conduct an OGTT.