Use of a diabetes self-assessment score to predict nonalcoholic fatty liver disease and nonalcoholic steatohepatitis

Abstract

Nonalcoholic fatty liver disease (NAFLD) and nonalcoholic steatohepatitis (NASH) are strongly associated with obesity, insulin resistance, and type 2 diabetes. We recently developed and validated a self-assessment score in the Korean population to identify people at high risk for diabetes. The objective of this study was to evaluate whether the self-assessment risk score for diabetes can also be used to screen for the presence of NAFLD or NASH. The study population included 15,676 subjects (8313 men and 7363 women) over 20 years old who visited the National Health Insurance Service Ilsan Hospital in Korea between 2008 and 2010. Anthropometric, clinical, and laboratory data were analyzed during regular health checkups. Fatty liver disease was diagnosed using ultrasound, discrimination capability was assessed based on the area under the receiver operating characteristic curve (AUC), and evaluation measures, including sensitivity and specificity, were calculated. Multiple logistic analyses were also performed. We calculated a self-assessed risk score for diabetes (range: 0-11), and a cutoff of ≥5 identified 60% (50%) of men (women) at high risk for NAFLD, reflecting a sensitivity of 79% (85%), a specificity of 60% (66%), a positive predictive value (PPV) of 68% (51%), and a negative predictive value (NPV) of 73% (91%), with an AUC of 0.75 (0.82) for men (women). A cutoff point of ≥6 identified 43% (31%) of men (women) at high risk for NASH, reflecting a sensitivity of 80% (86%), a specificity of 64% (75%), a PPV of 30% (28%), and a NPV of 94% (98%), with an AUC of 0.77 (0.86) for men (women). The odds ratios that a 1-point increase in the diabetes risk scores would be associated with an increased risk for NAFLD and NASH were 1.20 [95% confidence interval (CI): 1.16-1.25] and 1.57 (95% CI: 1.49-1.65), respectively, in men, and 1.28 (95% CI: 1.21-1.34) and 1.89 (95% CI: 1.73-2.07), respectively, in women. The present study indicates that our self-assessment risk score for diabetes could be an effective primary screening tool for the presence of NAFLD or NASH.