Background: A number of studies have shown that social support has a direct beneficial effect on well-being and also serves as a buffer to protect people from health problems due to excessive stress. Although preliminary studies report a positive relationship of alexithymia both with depression and reduced social support, there is no study examining whether the beneficial effect of social support on depression differs with the presence of alexithymia. Methods: A total of 120 workers aged 19-39 completed the 20-item Toronto Alexithymia Scale (TAS-20) to measure alexithymia, the Beck Depression Inventory-II (BDI-II) to evaluate depressive symptomatology, and the Job Content Questionnaire (JCQ) to assess job strain based on Karasek's demand-control-support model. The interrelationship among TAS-20, BDI-II and 3 subscales of JCQ (job demand, control, and support) were examined. Results: A significant association of depression with low support and high alexithymia was observed. Alexithymia was also associated with reduced support. Further, a statistically significant interaction between alexithymia and support in terms of their effect on depression was observed. Nonalexithymic individuals with low support showed a significantly higher depression score than those who received high support, while alexithymics did not differ in their depression score depending on the degree of support. Consistent results were obtained from the logistic regression analysis examining the odds ratio for depression by support by alexithymia; a significantly increased odds ratio for depression associated with low social support was observed only among nonalexithymics. Conclusions: Alexithymic individuals might be unable to benefit from social support because of their cognitive deficits of emotion. Copyright (C) 2003 S. Karger AG, Basel.