Ethnic and Sex Differences in Fatty Liver on Cardiac Computed Tomography: The Multi-Ethnic Study of Atherosclerosis

Editor's Note: Commentary based on Tota-Maharaj R, Blaha MJ, Zeb I, et al. Ethnic and sex differences in fatty liver on cardiac computed tomography: the multi-ethnic study of atherosclerosis. Mayo Clin Proc 2014;89:493-503.


Obesity has increased worldwide, now affecting over one billion people. Non-alcoholic fatty liver disease (NAFLD)—associated with obesity and currently the leading cause of persistent liver disease—affects about 30% of U.S. adults and has been independently associated with cardiovascular disease (CVD) events and all-cause mortality.1,2,3

Tota-Maharaj, et al. add to the incomplete knowledge regarding NAFLD by describing the race/ethnicity- and sex-specific prevalence of fatty liver in the presence and absence of both obesity and metabolic syndrome, as well as delineating fatty liver associative factors by race/ethnicity in a secondary analysis of the Multi-Ethnic Study of Atherosclerosis (MESA) cohort.4

A novel aspect of this report is the use of non-contrast cardiac computed tomography (CT) to diagnose NAFLD. Such a technique enables the study of NAFLD in large cohorts, far bigger than the study populations previously available for epidemiologic NAFLD research.


MESA is a prospective community-based epidemiologic study of the determinants of CVD in an age- and gender-balanced, multi-ethnic cohort. Enrollment of 6,814 participants identifying as whites, African-Americans, Hispanics, or Chinese ages 45-84 and without clinical CVD resulted from recruitment in six U.S. communities. Two consecutive, non-enhanced cardiac CT scans allowed coronary artery calcium scoring. Scans of 4,384 participants were adequate for assessing liver (L) and spleen (S) CT attenuation, allowing calculation of the L/S ratio for NAFLD diagnosis, defined as L/S ratio <1. Exclusion of patients with heavy alcohol use history, cirrhosis, and oral corticosteroid or amiodarone use produced a final study population of 4,088.

Tota-Maharaj, et al. compared fatty liver prevalence and severity and fatty liver associative factors among ethnicities, stratifying by gender, obesity, and metabolic syndrome (MetS). Multivariable ordinal logistic regression determined the effect of cardiometabolic risk factors on fatty liver prevalence and severity.


There was no statistically significant difference in NAFLD prevalence by sex (P=0.54), though it varied significantly by ethnicity (African American, 11%; white, 15%; Asian, 20%; Hispanic, 27%; P<0.001). Multivariable adjustment for age, sex, body mass index (BMI), waist circumference, HDL, high-sensitivity C-reactive protein (hs-CRP), triglycerides, homeostasis model assessment for insulin resistance (HOMA-IR), diabetes history, diabetes treatment, diet, and physical activity didn't attenuate observed inter-ethnic differences. Obesity, clinically defined MetS, and IR predicted two-fold increased fatty liver prevalence in all ethnicities after multivariable adjustment.

Although African Americans had the highest obesity prevalence, a smaller percentage of obese African Americans had NAFLD than other ethnicities (African American, 17%; white, 31%; Asian, 37%; Hispanic 39%; P<0.001). Hispanics—obese, non-obese, with or without MetS—had the highest NAFLD prevalence.

Increased prevalence odds ratio of NAFLD correlated with increased triglyceride levels in all ethnicities except whites. Hs-CRP was not significantly associated with NAFLD prevalence in Asians and African Americans, but was in whites and Hispanics. With multivariable adjustment, diabetes lost statistical significance for an association with NAFLD prevalence in all ethnicities. Waist circumference remained predictive of increased prevalence odds ratio of NAFLD, except in African Americans.


African Americans have a lower prevalence and Hispanics have a higher prevalence of NAFLD than do whites and Asians. Distinct ethnic variations in fatty liver prevalence even in patients with MetS or obesity suggest other factors at play in the phenotypic expression of NAFLD.


The leading cause of persistent liver disease, NAFLD is also robustly associated with CVD. In fact, CVD contributes more than liver disease to the morbidity and mortality burden of patients with NAFLD. Given the NAFLD to CVD association and the NAFLD association to increasing obesity and IR prevalence, a greater understanding of NAFLD incidence with respect to gender and ethnicity is essential in building strategies to prevent CVD.

The findings that African Americans have a lower prevalence and Hispanics have a higher prevalence of NAFLD than whites and Asians are consistent with work from others.5,6 The analysis by Tota-Maharaj and colleagues of factors that may explain ethnicity differences is the most comprehensive to date and provides important insight regarding NAFLD. Differences in age, sex, BMI, waist circumference, high-density lipoprotein, hs-CRP, triglycerides, HOMA-IR, diabetes history, diabetes treatment, diet, and physical activity did not explain the observed inter-ethnic differences. Even with MetS, obese African Americans exhibited lower NAFLD prevalence and severity than other ethnicities.

The authors speculate this occurs due to differences in the numerous genes involved in hepatic fat metabolism, which have been implicated in fatty liver pathogenesis. Some of these genes may be linked to differences in hepatic lipid and lipoprotein metabolism. Additional work by this group suggests a possible independent pathophysiologic role between dyslipidemia and NAFLD.7

Gaining understanding of factors explaining differences in NAFLD prevalence among ethnicities may lead to new targets for preventing and treating this emerging CVD risk.

In summary, clinicians should be aware of the close association of NAFLD with IR, obesity, metabolic syndrome, dyslipidemia and CVD risk, but should also be aware of the marked race/ethnicity differences in phenotypic NAFLD expression.


  1. Szczepaniak LS, Nurenberg P, Leonard D, et al. Magnetic resonance spectroscopy to measure hepatic triglyceride content: prevalence of hepatic steatosis in the general population. Am J Physiol Endocrinol Metab 2005;288:E462-8.
  2. Targher G, Day CP, Bonora E. Risk of cardiovascular disease in patients with nonalcoholic fatty liver disease. N Engl J Med 2010;363:1341-50.
  3. Targher G, Bertolini L, Rodella S, et al. Nonalcoholic fatty liver disease is independently associated with an increased incidence of cardiovascular events in type 2 diabetic patients. Diabetes Care 2007;30:2119-21.
  4. Tota-Maharaj R, Blaha MJ, Zeb I, et al. Ethnic and sex differences in fatty liver on cardiac computed tomography: the multi-ethnic study of atherosclerosis.Mayo Clin Proc 2014;89:493-503.
  5. Azuma K, Kadowaki T, Cetinel C, et al. Higher liver fat content among Japanese in Japan compared with non-Hispanic whites in the United States. Metabolism2009;58:1200-7.
  6. Browning JD, Szczepaniak LS, Dobbins R, et al. Prevalence of hepatic steatosis in an urban population in the United States: impact of ethnicity. Hepatology2004;40:1387-95.
  7. DeFilippis AP, Blaha MJ, Martin SS, et al. Nonalcoholic fatty liver disease and serum lipoproteins: the Multi-Ethnic Study of Atherosclerosis. Atherosclerosis. 2013;227:429-36.

Keywords: Ethnic Groups, Sex Characteristics, Cardiovascular Diseases, Liver Diseases, Obesity

< Back to Listings