We thank Drs. Prati and Colli for their encouraging comments. With regard to their comment about using serum aminotransferase data as a continuous variable and determining thresholds to predict mortality, we would like to direct the reader's attention to figure 3 of the article. The figure describes the risk of mortality associated with serum aspartate aminotransferase (AST) and alanine aminotransferase (ALT) activity levels, with the reference being the local general population. In both men and women, increases in these enzyme activities were associated with increasing risk of death. The curves indicate that there is no clear-cut threshold: the lower the levels of AST and ALT, the lower the risk of mortality.
Our data are less well suited for a receiver operating characteristic curve analysis, because they entail survival data (including different lengths of follow-up and censoring), as opposed to simple, dichotomous (yes or no) outcome. Furthermore, given all the caveats about generalizability of the data as discussed in the article, we believe it would be prudent not to suggest AST/ALT threshold levels that would be “protective” of future mortality.