In order to investigate the efficacy and safety of long-term treatment with flupirtine in patients with chronic pain, in particular arthrosis and arthritis, a study was planned which, when completed, will encompass the treatment of 200 patients over a 12-month period. The present paper is a preliminary report of this ongoing study. The report deals with 104 patients: 55 of whom completed the 12-month treatment period and a 2-week follow-up phase, during which flupirtine was replaced by placebo in order to be able to detect drug-withdrawal effects. Forty nine patients withdrew from the study. Most of the patients were suffering from degenerative rheumatic arthrosis or inflammatory rheumatic arthritis. The average daily dosage was 300 mg. The incidence of drop-outs was highest in the first months with hardly any patients withdrawing in the last six months. Fifteen patients dropped out because of side effects (dizziness, nausea, sleep disturbances, and headache). Ten patients dropped out because of ineffectiveness, seven because of side effects plus ineffectiveness, and three because of side effects and other reasons. The remaining 14 patients dropped out because of other or non-medical reasons. For the 55 patients who completed the study, the analgesic took effect within 45 minutes to 2 hours, the duration of effect was 4-6 hours. Three-quarters of the patients responded to the drug, one-quarter did not. The analgesic effect remained constant during the 12-month treatment, as did the average number of capsules taken per month. There was no evidence that tolerance developed. The most frequent side effects were drowsiness (9% of patients), dizziness (11%), dry mouth (5%) and pruritus (9%). The withdrawal symptom scale completed every month during treatment (to determine baseline values) and every day throughout the 2-week placebo post-treatment phase showed no changes in the median. The mean value increased during the withdrawal phase, however, indicating that the symptomatology was more pronounced in some subjects. After withdrawal, the non-specific symptoms increased to a greater extent than symptoms from the opiate scale. The symptoms were present throughout the withdrawal phase. If the withdrawal phenomena had corresponded to the flupirtine's terminal half-life, then the symptoms ought to have been present mainly in the first few days. There was a slight trend for lowering systolic blood pressure but no changes in diastolic blood pressure or heart rate, nor changes in the ECG or laboratory analysis that could be related to flupirtine. These preliminary data suggest that flupirtine is safe when given for a period of one year.