To the Editor: We read with great interest the recent article by Hans Joergen Grabe, M.D., and colleagues (1) concerning alexithymic features as predictive factors of psychopathology in psychiatric patients. The article suggested the relevance of improving emotional awareness as a major issue of therapeutic interventions in patients with mental disorders. However, some methodological concerns should moderate the interpretation of these findings. Dr. Grabe and colleagues used a linear regression technique to calculate the relative magnitude of the prediction of psychopathology of several independent variables, such as Toronto Alexithymia Scale factors, Temperament and Character Inventory dimensions, age, and gender. However, they did not integrate in their analysis a dimensional assessment of depression. Now we know from several studies that alexithymia and depression, although not totally overlapping, are strongly associated, with depression acting as a strong mediator between alexithymic features and psychopathology (2). This is especially true for the difficulties identifying feelings subscale of the Toronto Alexithymia Scale, whose correlations with depression vary between 0.42 and 0.65 (3). It is not surprising that this factor was found to be the strongest predictor of psychopathology in the study by Dr. Grabe and colleagues. The authors should have included in the linear regression analysis a dimensional measure of depression, or they should have measured with a hierarchical regression analysis the additional part of the variance of psychopathology explained by alexithymic features beyond the variance accounted for by depression, as was done by Luminet and colleagues (4). These procedures would have strengthened their results and would have allowed a comparison with previous published studies that have shown that alexithymia is truly associated with higher levels of psychopathology and acts as a negative predictor of outcome beyond the influence of depression (5).