In the last 30 years antidepressant drugs have been used increasingly in the treatment of patients with chronic pain. This article reviews the results of some 40 placebo-controlled studies. It is difficult to make comparisons between the various studies because they often differ in terms of pain conditions, patient selection, antidepressant drug used, dosages, trial design, etc. However, in spite of this heterogeneity and other methodological problems it is clear that a wide range of pain conditions are responsive to antidepressant drug treatment, in particular: headache, migraine, facial pain, neurogenic pain, fibrositis, and probably arthritis and rheumatoid arthritis. More data need to be gathered in cancer pain, and in other conditions such as low back pain for which no, or very limited, effect has been shown. The beneficial effects of antidepressant drugs is in most cases of a mild to moderate degree, some time lag is necessary before it is completely manifest, and it tends to persist over time if drug treatment is continued in the long term. Strong evidence of efficacy is not evident for all the antidepressants, and there are probably significant differences in this respect between various drugs. The effect of a drug on pain does not seem necessarily to be related to its effect on mood. Further studies are needed to clarify this topic, and it will be necessary to examine specific pain conditions, compare different antidepressants, with reference to each other and to placebo, further investigate the role of drug plasma concentrations and control for the presence of concomitant psychiatric disturbances and for organic lesions responsible for the pain symptomatology.