Abstract B133: Disparities in risk factors for nonalcoholic fatty liver disease (NAFLD) among a racially and ethnically diverse cohort

• Published in: Cancer epidemiology, biomarkers & prevention biomarkers & prevention, 2023-12, Vol.32 (12_Supplement), p.B133-B133
• Main Authors: Chu, Janet N., DeRouen, Mindy C., Canchola, Alison J., Cortella, Aly, Inamdar, Pushkar P., Daida, Yihe G., Gomez, Scarlett L., El-Serag, Hashem, Shariff-Marco, Salma
• Format: Article
• Language: English
• ISSN: 1538-7755; 1538-7755
• DOI: 10.1158/1538-7755.DISP23-B133

Background NAFLD has become the leading cause of chronic liver disease and is an important risk factor for liver cancer. About 25% of the U.S. adult population has NAFLD, with prevalence up to 70% among those with type 2 diabetes. Racial and ethnic disparities in NAFLD show Asian and Hispanic Americans at greater risk. Yet, few studies have examined risk factors and the magnitude of influence of these risk factors for NAFLD in racially and ethnically diverse populations. Methods This cohort included patients ≥ 18 years old who had at least two in-person visits at least one year apart between 2000 and 2017 at Kaiser Permanente Hawai`i, a large, integrated healthcare system. We excluded those who had NAFLD diagnosis at baseline (i.e., first year after the first in-person visit from 2000-2017). From the electronic medical record, we obtained data on sociodemographic (age, sex, race, ethnicity) and behavioral (tobacco use, care utilization) characteristics. We also extracted Current Procedural Terminology codes, International Classification of Diseases (ICD)-9 and -10 codes, prescription orders, vital signs, and laboratory results. We used small-area neighborhood data from the Hawaii Neighborhoods Data System. We ran sequential Cox regression models with outcome of incident NAFLD with age as the time-scale, underlying stratification by baseline year and decile of number of outpatient and ED encounters, and clustering by census tract. We added in variables one-by-one based on the order of influence. This was calculated as the percent difference in effect separately for: 1) sex (men compared to women) and 2) for each race and ethnicity group (compared to non-Hispanic White) from the base model when adding in each covariate separately. Results Of 373,032 patients followed for a median of 5.6 years, 8,994 had a diagnosis of incident NAFLD (2.4%). Variables associated with NAFLD and disparities by sex or race/ethnicity included having hypertension, high triglycerides, lower HDL, smoking, diabetes, overweight or obese body mass index (BMI), limited English proficiency, history of smoking, chronic hepatitis C, or low neighborhood SES. High triglycerides (triglycerides ≥150 mg/dL) had a substantial impact on NAFLD risk by sex; when high triglycerides were accounted for, risk among men compared to women changed from higher (HR=1.28; 95% CI:1.16-1.40) to lower (HR=0.74; 0.60-0.91) risk of NAFLD. Risk factors with the greatest impact on NAFLD risk differed across racial/