In a 12-week crossover study, a formulation of paracetamol (500 mg) and metoclopramide (5 mg) was compared with paracetamol (500 mg) alone for adjunct analgesia in 39 patients with arthritis. Most were on stabilized therapy with anti-inflammatory drugs and all had a history of a tendency to develop gastro-intestinal symptoms. Patients were allocated at random to receive 2 tablets 3-times daily of either treatment for 6 weeks and were then crossed over to the alternative treatment at the same dosage for a further 6 weeks. The results showed that the paracetamol/metoclopramide formulation appeared more effective as an analgesic than paracetamol alone and gave a highly significant reduction in the mean score for upper gastro-intestinal symptoms when compared with paracetamol. There were no side-effects associated with the metoclopramide component. It is concluded that the paracetamol/metoclopramide formulation is a suitable adjunct analgesic for osteoarthritis and rheumatoid arthritis in those patients prone to gastro-intestinal disturbance.