Background: Chronic hepatitis B (CHB) can progress to cirrhosis, but there are limited noninvasive tools available to estimate cirrhosis risk, including in patients receiving antiviral therapy. This study developed and validated a simple model to assess risk in CHB patients.
Methods: The derivation cohort included 3000 CHB patients from 6 centers in the United States, with 52.60% receiving antiviral therapy. External validation was performed for 4552 CHB individuals from similar cohorts in Taiwan, with 21.27% receiving therapy. Cox proportional hazards regression analyses were used to screen predictors and develop the risk score for cirrhosis. Areas under receiver operating characteristic curves (AUROCs) were calculated for predictive value.
Results: Sex, age, diabetes, antiviral treatment status/duration, hepatitis B e-antigen, and baseline alanine aminotransferase/aspartate aminotransferase levels were significantly associated with increased cirrhosis risk. A 13-point risk score was developed based on these predictors. The AUROCs for predicting cirrhosis risk were 0.82 at 3 years, 0.85 at 5 years, and 0.89 at 10 years in the derivation cohort, and 0.82, 0.79, and 0.77 in the validation cohort, respectively.
Conclusions: We developed and validated a simple cirrhosis prediction model with an independent external cohort that can be applied to both treatment-naive and treatment-experienced CHB patients in diverse settings and locations.
【저자키워드】 risk prediction, scoring system, risk model, area under the curve, variable selection, clinical tools,