Alanine aminotransferase as an independent predictor of incident nonalcoholic fatty liver disease.

To the Editor: We read with interest the recent article by Chang et al. (1 ) reporting that higher serum alanine aminotransferase (ALT) concentrations, within the reference interval, independently predicted the incidence of nonalcoholic fatty liver disease (NAFLD) during a mean follow-up of 2.5 years in a large cohort of apparently healthy Korean men. Several prospective studies have previously shown that increased ALT concentrations, even within the reference interval, also predict the future development of type 2 diabetes (2 ) and cardiovascular events (3 ) independently of other known risk factors. In all of these studies, however, increased ALT concentrations have been used as a surrogate marker of NAFLD. Indeed, increased liver enzymes are usually thought to be a consequence (and not a cause) of liver injury in NAFLD and can possibly be used as predictors of NAFLD progression (2 ). We think the association of the 2 conditions, increased ALT concentrations and incident NAFLD, does not necessarily prove causation. The greater incidence of NAFLD—as diagnosed by ultrasound—among those with slightly increased ALT concentrations at baseline is most likely attributable to an underlying common mechanism, i.e., more severe insulin resistance in those with higher than in those with lower ALT concentrations. That the association between increasing serum ALT concentrations and incident NAFLD remained statistically significant even after adjustment for the homeostasis model assessment (HOMA)-estimated insulin resistance may be due simply to the fact that the HOMA score is not a good proxy measure of insulin resistance. Thus, we wonder how different the results would have been if the euglycemic clamp technique or methods that are more accurate had been used to measure insulin resistance. Interpretation of the results of the Chang et al. study are limited by another major caveat (1 ), also recognized by the authors, that the diagnosis or exclusion of NAFLD, both at baseline and follow-up, was based on liver enzymes and ultrasound imaging, but was not confirmed by liver biopsy. Indeed, it is known that liver enzymes may be within the reference interval in up to 70% of patients with diagnosed NAFLD and that the full histopathological spectrum of NAFLD may be present in patients with normal liver enzymes, which therefore cannot be reliably used to exclude the presence of NAFLD (2 ). Moreover, although liver ultrasonography is widely used for diagnosing NAFLD, this imaging method has good sensitivity and specificity only for detection of moderate and severe hepatic steatosis, but its sensitivity is reduced when hepatic fat infiltration on liver biopsy is 33% (4 ). Only liver biopsy can be used for diagnosing NAFLD and accurately determining the histological severity and prognosis of liver damage (2 ). Thus, although this limitation, if present, would probably tend to reduce (at least partly) the strength of the association between ALT concentrations and incident NAFLD shown by Chang et al. (1 ), we think some nondifferential misclassification of NAFLD on the basis of liver enzymes and ultrasonography is likely. That is, some of the study participants may not have NAFLD despite ultrasound detection of fatty liver and some may have underlying NAFLD with normal liver enzymes and negative ultrasound findings. The latter situation may have been partly confirmed during the short follow-up study ( 3 years) by the unexpectedly high incidence of NAFLD detected in the 5237 healthy participants with a mean age of 36 years (984 incident NAFLD cases, i.e., 75 new cases per 1000 person-years). In contrast, the incidence of high aminotransferase concentrations, as surrogate markers of NAFLD, was recently reported to be 30 new incident cases per 1000 person-years in a cohort of Japanese healthy individuals (age 35 years) who were free of NAFLD at baseline (5 ).