PREDICTORS OF FATTY LIVER DISEASE SUBTYPES: NHANES ANALYSIS (1999-2002).

Objectives: Identifying individuals with sub-clinical liver disease is an important goal within clinical practice and has major public health consequences. The majority of liver disease not attributable to viral etiologies is related to fatty liver disease. Fatty liver disease is traditionally divided into alcohol related liver disease (ALD) and non-alcohol related fatty liver disease (NAFLD). The objective of this analysis is to define the risk factors for fatty liver disease subtypes in the United States population.

Methods: The National Health and Nutritional Evaluation Survey (NHANES) conducted from 1999-2002 was analyzed to compare ALD to NAFLD. All non-pregnant, adult subjects with negative viral serologies for hepatitis b and c were included. An elevated serum alanine aminotransferase (ALT) level was used as the marker to define liver injury. Logistic regression models were fit to compare the likelihood of having NAFLD as compared to ALD.

Results: 18,358 subjects were in the dataset. 6,653 subjects who met inclusion criteria had normal ALT values. 293 subjects had elevated ALT values and endorsed excessive alcohol consumption (ALD cohort). 479 subjects had elevated ALT values but minimal alcohol consumption (NAFLD cohort). Males are more frequently represented in ALD cohort ( 79.8%) than in the NAFLD cohort (43.9%) (p=0.0009). The ALD cohort was younger (mean age 41.9 years) than the normal cohort (mean age 47.0 years) (p=0.04). Logistic regression analysis was performed modeling the odds of being in the NAFLD group in reference to the ALD group. Females were 5.2 times more likely to be in the NAFLD cohort (OR 5.21, 95% CI 2.4-11.6). Age and ethnicity were not significantly different between the two cohorts. The presence of excess abdominal adiposity was 6.9 times more frequent in the ALD cohort (OR 6.86, 95% CI 2.21-21.24). The diagnosis of diabetes was 5.6 times more frequent in the NAFLD cohort (OR 5.59, 95% CI 1.26- 25.00). The NAFLD cohort was 18.5 times more likely to have enzyme elevations of more than 3 times the upper limit of normal (OR 18.5, 95% CI 2.0-166.7). Serum GGT, transferrin saturation, and ferritin, and AST/ALT ratio were not significantly different between then cohorts. The diagnosis of metabolic syndrome was similar between the cohorts.

Conclusions: Metabolic syndrome is not a useful tool when attempting to differentiate ALD from NAFLD. Female gender, the diagnosis of diabetes, and ALT elevations greater than 3 times the upper limit of normal are all present significantly more often in subjects with NAFLD. Serum markers such as ferritin and AST/ALT ratio were not found to be useful in differentiating between ALD and NAFLD.