Developing a risk model for early diagnosis of metabolic syndrome in Chinese adults aged 40 years and above based on BMI/HDL-C: a cross-sectional study

Table 4 The sensitivity, specificity, AUC and cutoff value of different models

Models	Sensitivity	Specificity	AUC	(95%CI)	Cut off value
BMI/HDL-C ^a
Model 1	0.700	0.681	0.731	(0.693,0.768)	0.468
Model 2	0.688	0.714	0.752	(0.716,0.788)	0.462
Model 3	0.810	0.640	0.780	(0.746,0.814)	0.623
TG/HDL-C ^b
Model 1	0.670	0.651	0.699	(0.661,0.738)	0.471
Model 2	0.783	0.575	0.718	(0.680,0.756)	0.425
Model 3	0.725	0.728	0.779	(0.745,0.813)	0.717
VAI ^c
Model 1	0.703	0.684	0.712	(0.673,0.750)	0.480
Model 2	0.703	0.684	0.716	(0.678,0.755)	0.458
Model 3	0.878	0.580	0.780	(0.745,0.813)	0.500
WHTR ^d
Model 1	0.670	0.649	0.672	(0.632,0.712)	0.452
Model 2	0.606	0.714	0.684	(0.644,0.724)	0.549
Model 3	0.875	0.550	0.772	(0.737,0.807)	0.516

a: prediction model based on quartile of BMI/HDL-C. b: prediction model based on the quartile of TG/HDL-C. c: prediction model based on the quartile of VAI. d: prediction model based on the quartile of WHTR. Model 1: no confounding factors were adjusted. Model 2: adjusting for confounding factors including gender, education level, and smoking status. Model 3: further adjustment of confounding factors including BMI, BFM, PBF, VFA, AMC, WC, SCR, LDH, TG, HDL-C, and insulin. BMI, Body Mass Index; HDL-C, high-density lipoprotein cholesterol; TG, triglyceride; WHTR: waist-to-height ratio. VAI: visceral adiposity index

ISSN: 1472-6823