BACKGROUND: Although uncommon, severe post-phlebitic syndrome may be associated with persistent, intractable pain and swelling that interfere with work and leisure activities. This study was performed to determine whether intermittent compression therapy with an extremity pump benefits patients with this condition and, if so, whether the benefit is sustained. METHODS: The study was a randomized crossover trial. Over the period 1990 to 1996, all patients in the clinical thromboembolism program of an Ontario teaching hospital who had a history of deep vein thrombosis and intractable symptoms of post-phlebitic syndrome were recruited into the study. The study involved using an extremity pump twice daily for a total of 2 months (20 minutes per session). The patients were randomly assigned to use either a therapeutic pressure (50 mm Hg) or a placebo pressure (15 mm Hg) for the first month. For the second month, the patients used the other pressure. A questionnaire assessing symptoms and functional status served as the primary outcome measure and was administered at the end of each 1-month period. A symptom score was derived by summing the scores for individual questions. At the end of the 2-month study, patients were asked to indicate their treatment preference and to rate the importance of the difference between the 12 pressures. Treatment was considered successful if the patient preferred the therapeutic pressure and stated that he or she would continue using the extremity pump and that the difference between the therapeutic and placebo pressures was of at least slight importance. All other combinations of responses were considered to represent treatment failure. Patients whose treatment was classified as successful were offered the opportunity to keep the pump and to alter pressure, frequency and duration of pump use to optimize symptom management. In July 1996 the authors contacted all study participants whose treatment had been classified as successful to determine whether they were still using the pump and, if so, whether they were still deriving benefit. RESULTS: In total 15 consecutive patients (12 women and 3 men) were enrolled in the study. The symptom scores were significantly better with the therapeutic pressure (mean 16.5) than with the placebo pressure (mean 14.4) (paired t-test, p = 0.007). The treatment for 12 of the patients (80%, 95% confidence interval 52% to 96%) was considered successful. Of these, 9 patients continued to use the pump beyond the crossover study and to derive benefit. INTERPRETATION: The authors conclude that a trial of pump therapy is worthwhile for patients with severe post-phlebitic syndrome and that a sustained beneficial response can be expected in most such patients.