We read with interest the article by Dr. Tae Hoon Lee and colleagues,1 which demonstrates that aminotransferase activity can predict mortality in U.S. community residents. We think that the data collected by Dr. Lee and colleagues will be even more informative and clinically useful if supplemental analyses will be conducted.
The authors document that the risk of death from all causes increases as aminotransferase levels increase even within the normal range, but do not provide an estimate of the best thresholds to predict mortality. Survival analyses were performed after categorizing subjects in four subgroups, according to alanine aminotransferase (ALT) levels. Instead, we suggest the analysis of ALT as a continuous variable, using a receiver operating characteristic curve analysis. We believe this approach would be very helpful in clinical practice. As correctly underscored by the authors, we previously identified “healthy” or “desirable” ranges for ALT values, that is, the values expected in individuals at lowest risk for liver disease.2 Although these ranges were associated with lower mortality rates among Korean men,3 the optimal thresholds in Western populations remain to be assessed.
There is growing evidence that high aminotransferase values are associated with future mortality for both hepatic and nonhepatic causes. In this regard, ALT activity can be viewed not only as an indicator of liver disease, but also as a gauge of general health.4 Therefore, we fully agree with Dr. Lee and colleagues that aminotransferase measurement could have an important role as a preventive screening test, particularly in individuals at risk. It is time to provide clinicians with validated thresholds, defined according to different clinical outcomes.