OBJECTIVES: Comorbid of non-alcoholic fatty liver disease (NAFLD) in chronic hepatitis B (CHB) patients is common but no simple noninvasive diagnostic methods are available for the identification. This study aims to develop a noninvasive nomogram for accurate detection of NAFLD in CHB patients. PATIENTS AND METHODS: This study included 535 liver biopsy-proven CHB patients with or without comorbid NAFLD. Independent risk factors of NAFLD were identified by multivariate logistic regression analysis. The risk factors identified were then incorporated into the nomogram. Performance of the nomogram was assessed by calibration, receiver operating characteristic (ROC) curve and decision curve analysis. RESULTS: Of the 535 patients, 100 patients (18.69%) were diagnosed as CHB/NAFLD and 435 patients (81.31%) as simple CHB. Body mass index, serum uric acid and low-density lipoprotein cholesterol levels and diabetes mellitus were independent risk factors of NAFLD. The nomogram incorporating these 4 factors had an area under ROC curve (AUC) of 0.864, achieved good concordance index of 0.864 (95% confidence interval: 0.832-0.892) for predicting NAFLD in the patients and had well-fitted calibration curves. The nomogram had a significantly higher AUC than some previously reported models. The decision curve analysis yielded larger net benefit. CONCLUSION: This study developed a simple, noninvasive, effective and convenient nomogram that achieved an optimal detection of NAFLD in CHB patients. Using this nomogram, the risk for an individual patient to have NAFLD could be discriminated, leading to a rational clinical management.
OBJECTIVES: Comorbid of non-alcoholic fatty liver disease (NAFLD) in chronic hepatitis B (CHB) patients is common but no simple noninvasive diagnostic methods are available for the identification. This study aims to develop a noninvasive nomogram for accurate detection of NAFLD in CHB patients. PATIENTS AND METHODS: This study included 535 liver biopsy-proven CHB patients with or without comorbid NAFLD. Independent risk factors of NAFLD were identified by multivariate logistic regression analysis. The risk factors identified were then incorporated into the nomogram. Performance of the nomogram was assessed by calibration, receiver operating characteristic (ROC) curve and decision curve analysis. RESULTS: Of the 535 patients, 100 patients (18.69%) were diagnosed as CHB/NAFLD and 435 patients (81.31%) as simple CHB. Body mass index, serum uric acid and low-density lipoprotein cholesterol levels and diabetes mellitus were independent risk factors of NAFLD. The nomogram incorporating these 4 factors had an area under ROC curve (AUC) of 0.864, achieved good concordance index of 0.864 (95% confidence interval: 0.832-0.892) for predicting NAFLD in the patients and had well-fitted calibration curves. The nomogram had a significantly higher AUC than some previously reported models. The decision curve analysis yielded larger net benefit. CONCLUSION: This study developed a simple, noninvasive, effective and convenient nomogram that achieved an optimal detection of NAFLD in CHB patients. Using this nomogram, the risk for an individual patient to have NAFLD could be discriminated, leading to a rational clinical management.