Suboptimal performance of simple noninvasive tests for advanced fibrosis in Chinese patients with nonalcoholic fatty liver disease

Yun Hao Xun,Jian Gao Fan,Guo Qing Zang,Hong Liu,Yan Ming Jiang,Jing Xiang,Qian Huang,Jun Ping Shi
DOI: https://doi.org/10.1111/j.1751-2980.2012.00631.x
2012-01-01
Journal of Digestive Diseases
Abstract:OBJECTIVE: To evaluate the diagnostic accuracy of some noninvasive fibrosis models in Chinese patients with nonalcoholic fatty liver disease (NAFLD). METHODS: Consecutive biopsy-proven NAFLD patients were recruited from a single center from January 2005 to December 2010. Advanced fibrosis (stage 3 and 4) was defined using Kleiner criteria. The area under the receiver operating characteristic curve (AUROC) was used to compare the diagnostic accuracy of the NAFLD fibrosis score (NFS), FIB-4 index, aspartate transaminase (AST)/platelet ratio index (APRI), AST/alanine transaminase (ALT) ratio (AAR) and body mass index (BMI)-AAR-Diabetes (BARD) score. RESULTS: Of the patients with NAFLD, 79.6% were males with a mean age of 37.1 years, mean BMI of 26.1 kg/m(2) and 41.4% of them had nonalcoholic steatohepatitis, and 24 (15.8%) had advanced fibrosis. The AUROC of the FIB-4 index, APRI, AAR, NFS and BARD score for advanced fibrosis were 0.756, 0.742, 0.670, 0.653 and 0.642 (P < 0.05 for all), respectively. A concordant negative predictive value of approximately 90% was indicated whereas the positive predictive values were modest for all tests, and only the FIB-4 index yielded a higher positive likelihood ratio of 7.65. Using these cut-off values of tests for excluding advanced fibrosis could reduce the use of liver biopsy in 56.674.3% of the patients, with a minor false negative rate of 5.39.9%. CONCLUSIONS: Although slightly less accurate than liver biopsy, simple noninvasive tests can reliably exclude advanced fibrosis in Chinese NAFLD patients in our center. FIB-4 index performs better than the other tests under examination.
What problem does this paper attempt to address?