Abstract 2258: Predictors of hepatocellular carcinoma development in nonalcoholic fatty liver disease

Abstract

Abstract Background: The incidence of nonalcoholic fatty liver disease (NAFLD)-related hepatocellular carcinoma (HCC) is predicted to increase with increasing worldwide prevalence. We aimed to identify predictors of HCC risk in patients with NAFLD, an unmet clinical need. Methods: We identified all patients diagnosed with NAFLD in the electronic health records (EHR) of a large US-based health system from 2004 to 2018. Information on demographics, body mass index (BMI), smoking status, serum biochemistries, comorbidities, albumin-bilirubin (ALBI) score, AST/√ALT ratio, fibrosis-4 index (FIB-4), and cirrhosis status were extracted from the EHR. Over an average 5.5 year at risk from date of first recorded NAFLD to the last contact with the healthcare system, date of HCC diagnosis or death, whichever occurred first, we identified 371 patients diagnosed with HCC. Cox proportional regression method with a stepwise approach was employed to identify independent predictors of HCC incidence. A Poisson regression method was used to estimate the incidence rate of HCC in NAFLD patients with a specified set of identified risk factors in the stepwise Cox regression analysis. Results. Among the 39,424 NAFLD patients, 57.8% were women, 44.5% were ever smokers, 27.3% had diabetes, 39.9% had dyslipidemia, 4.7% had Fibrosis-4 index (FIB-4)>2.67 (advanced fibrosis), and 3.7% had cirrhosis. The mean (standard deviation) age and BMI at first recorded diagnosis of NAFLD were 57.0 (11.4) years and 33.7 (7.4) kg/m2, respectively. The corresponding figures in 371 incident HCC cases were 65.1 (10.6) years of age, 31.9 (6.9) kg/m2 of BMI, 47.7% women, 52.8% ever smokers, 38.3% diabetes, 39.6% dyslipidemia, 17.8% FIB-4 >2.67, and 11.6% cirrhosis. All of these were independent factors for HCC risk. The estimated incidence rate of HCC for a male patient with NAFLD who was 65 years old, ever smoked, had BMI 30 kg/m3 and a history of diabetes and dyslipidemia, but had neither advanced fibrosis nor cirrhosis, was 3.02 (95% confidence interval [CI], 2.37-3.84) per 1000 person-years. The corresponding figures were 8.20 (95% CI, 5.78-11.62) for a male with advanced fibrosis but no cirrhosis and the same other characteristics as described above, and 18.05 (95% CI, 12.19-26.72) for the same patient with cirrhosis. On the other hand, the incidence of HCC for a male cirrhotic patient without a history of smoking or diabetes was 9.53 per 1000 person-years, a 47% reduction in risk. Conclusion. In this large cohort of NAFLD patients, we identified age, sex, BMI, smoking, diabetes, dyslipidemia, advanced fibrosis and cirrhosis as independent risk predictors for HCC. Although the effect of smoking and diabetes was relatively small in low-risk patients, their impact on high-risk NAFLD, especially those with advanced fibrosis or cirrhosis is discernable. Smoking cessation and treatment of diabetes may be potential strategies for reducing HCC risk in high-risk NAFLD patients. Citation Format: Jian-Min Yuan, Renwei Wang, Hung N. Luu, Jaideep Behari. Predictors of hepatocellular carcinoma development in nonalcoholic fatty liver disease [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2022; 2022 Apr 8-13. Philadelphia (PA): AACR; Cancer Res 2022;82(12_Suppl):Abstract nr 2258.