Liu, Chuan-Chuan, Hung, Chung-Lieh, Shih, Shou-Chuan, Ko, Hung-Ju, Lu, Yi-Ting, Wu, Yih-Jer, Cheng, Hui-Yun, Hu, Kuang-Chun, Yeh, Hung-I., and Chang, Ray-E.
Summary: Background: Nonalcoholic fatty liver disease (NAFLD), which represents metabolic abnormality and reflects extra visceral fat deposition, has been shown to be a threat to public health and associated with cardiovascular risks. However, data regarding the differences in clinical presentation and related metabolic abnormalities in the aging group population remained scarce in the Taiwanese population. Methods: We subsequently examined 8,658 subjects participating general health evaluation in Mackay Memorial hospital from 2003 to 2007. Baseline characteristics, anthropometrics, medical history, and biochemical markers were all collected. Abdominal ultrasonography was performed in all subjects. Two estimated cardiovascular risk scores were calculated according to criteria of the U.S. National Cholesterol Education Program Adult Treatment Panel III as metabolic score and Framingham risk score. Univariate logistic regression model was used to examine whether the baseline characteristics, anthropometrics, histories, and biochemical markers were independently associated with NAFLD from various age groups (young vs. aging groups) classified by 60 years of age. Respective receiver operating characteristic curves (ROC) with area under the curve were generated to test the capability of both cardiovascular risk scores in NAFLD discrimination from different age groups. Results: Totally 7,204 subjects (mean age: 44.5±11 years, 36% female) were finally enrolled in our study. Subjects with NAFLD were observed to have high body weight, body mass index, and circumferential waist and significantly abnormal biochemical markers accompanying worsening lipid profiles when compared with those without NAFLD in both the young and aging group populations (all p <0.001) although alkaline phosphatase, total cholesterol, and low-density lipoprotein did not show significant differences with and without NAFLD in the aging group population. Diabetes history remained a strong independent predictor of NAFLD in both young and aging groups (odds ratio: 3.3, p <0.001 vs. 2.51, p =0.014). The prediction model by using different cardiovascular risk scores yielded a meaningful ROC value of 0.82 (young group) and 0.71 (aging group) (both p <0.001) for metabolic score with ROC value of 0.67 (young group) and 0.52 (aging group) for Framingham risk score (p <0.001 vs. 0.408, respectively). Conclusion: The prevalence of NAFLD demonstrated a bimodal distribution with age in different genders. Although baseline characters and biochemical markers were demonstrated to be potential screening tools in detecting such clinical abnormality, they actually exerted diverse capabilities in the prediction of NAFLD in the different age groups. Traditional cardiovascular risk scores were less effective in predicting NAFLD in the aging group population. [Copyright &y& Elsevier]