A prospective double-blind randomized trial was conducted on 184 cancer patients with moderate to severe chronic pain to evaluate the analgesic efficacy and tolerability of diclofenac alone (50 mg q.i.d.) or in combination with a weak opioid (codeine 40 mg q.i.d.), or with an anti-depressant (imipramine, 10 or 25 mg t.i.d.). All demographic and clinical characteristics including cancer type, presence of bone metastases, baseline pain severity, neuropathic and nociceptive pain, and depressive state, were well balanced between the three treatment groups. The main analysis of the study was on the VAS scores at visit 2 (day 4). The mean VAS values for both associations imipramine plus diclofenac and codeine plus diclofenac were similar to the association placebo plus diclofenac. Patients on imipramine plus diclofenac and on placebo plus diclofenac were withdrawn mainly for inadequate efficacy, while patients on codeine plus diclofenac discontinued equally for inadequate efficacy or adverse events. In conclusion, in a short-term evaluation the addition of a tricyclic anti-depressant or a weak opioid to diclofenac did not provide further analgesia with respect to diclofenac administration alone.