Are alanine transaminase (ALT) levels useful to screen for Non-alcoholic Fatty Liver Disease (NAFLD) in the community?

Abstract

BACKGROUND: The value of ALT in screening for NAFLD is controversial. Furthermore, the upper limit of normal (ULN) ALT is defined for western populations but not Asian populations. Objectives: To establish ULN ALT for an urban adult Sri Lankan population, and assess its value in screening for NAFLD in that community. METHODS: This was part of a community based study on non-communicable diseases - the Ragama Health Study. 35-64y adults were selected using stratified random sampling. They were screened by structured interview and liver ultrasound, and 10ml venous blood .was obtained. "Normal" adults were defined as those not using potentially hepatotoxic drugs, drinking alcohol within safe limits, HBsAg and anti-HCV negative, and no fatty liver on ultrasound. NAFLD was diagnosed on established ultrasound criteria, .safe alcohol consumption and being HBsAg and anti-HCV negative. The 95th percentile of ALT was taken as the ULN for this population. RESULTS: 3012 subjects participated in the study. Those with NAFLD (n=930) and "normals" (n=1716) were matched for age and sex. ULN ALT (U/l) was significantly higher in normal males than females (68 vs. 53; p<0.001, Student's t-test). ALT was significantly higher in NAFLD than normals (p<0.001). However, ALT>ULN had a sensitivity [positive predictive value (PPV)] of only 14.5% (58%) for males and 11.5% (60%) for females with NAFLD. When ULN ALT cut offs for western populations (males 30 IU/1, females 19 IU/1) were used, although sensitivity considerably increased, the PPV to detect NAFLD decreased.further [sensitivity (PPV): males 79.4% (42%); females 94.6% (41%)]. CONCLUSIONS: Although ALT was significantly higher in NAFLD than in normal adults, it does not seem a useful test to screen for NAFLD in the community.