Application value of obesity measures in predicting health risk of metabolic-associated fatty liver disease(MAFLD)
Objective To evaluate the predictive value of obesity measures on the health risk of metabolic-associated fatty liver disease(MAFLD).Methods A cohort of 49 155 individuals who underwent physical examinations at the Health Management Center of the hospital from June 1 to December 31,2023 was se-lected.Subjects were categorized into MAFLD and non-MAFLD groups and assessed for body mass index(BMI),waist hip ratio(WHR),visceral fat index(VAI)and lipid accumulation product(LAP).BMI was stratified into normal,overweight,and obese categories;waist circumference(WC)into normal,early and established central obesity.Logistic regression differentiated gender-based obesity metrics and bio-chemical profiles.The ROC curve assessed predictive accuracy for MAFLD.Results A total of 23 765 cases(48.35％)of MAFLD were detected,including 18 030 males(75.87％)and 5 735 females(24.13％).In the same gender,the levels of other indicators in MAFLD group were higher than those in non MAFLD group(all P＜0.001),except the level of high density lipoprotein cholesterol(HDL-C)was lower than that in non MAFLD group.Adjusted for confounders,MAFLD risk correlated positively with BMI,WC,WHR,VAI,and LAP.ROC analysis showed that the area under the curve(AUC)of BMI,WHR,VAI and LAP in predicting MAFLD in different gender groups were greater than 0.7,the AUC value of LAP was the largest,and the AUC value of women[0.896(0.891～0.900)]was higher than that of men[0.831(0.826～0.836)],and the best cut-off values were 25.49 and 39.26(P＜0.05).The BMI×LAP model offered su-perior predictive accuracy across genders,with AUC values of 0.846(0.841～0.850)for males and 0.908(0.904～0.913)for females(all P＜0.05).Conclusion The risk of MAFLD increases with higher levels of BMI,WHR,VAI,and LAP,with LAP showing a higher predictive value than other indicators,particularly in women.BMI×LAP becomes a powerful tool for different gender MAFLD health risk assessment.
NETL
NSTL
万方数据
