The results of a retrospective study comparing two equally large groups of patients treated either surgically for restoration of venous patency and valvular function (24 patients) or medically with heparine, oral anticoagulants and compression stockings (25 patients) are presented. Follow-up time was 7.6 and 7.9 years respectively, operative mortality nil. Assessment of venous function was based on clinical observations as well as on measurements of haemodynamic parameters. Non-fatal pulmonary embolism after onset of treatment occurred in both cohorts with an equal frequency of 13%. Patients operated on for ilio-femoral deep venous thrombosis (DVT) were with few exceptions totally independent of any form of adjunctive hosiery which was in sharp contrast to the conservatively managed group. If onset of DVT had occurred more than three days earlier and extended from the ilio-femoral axis to the popliteo-crural level, surgery usually failed and patients were no better off than in the comparable medical subgroup. The same pattern of late outcome was found for all other clinical and haemodynamic parameters; i.e. clinical signs of venous hypertension, valvular competence as judged by sonography, patient's self-assessment and the expelled volume and refilling time measured by dynamic plethysmography after standardized leg-work. The mean expelled volume was 1.1 +/- 0.5 ml/100 g/min for the surgical subgroup treated early for ilio-femoral DVT and 0.7 +/- 0.5/100 g/min for the corresponding medical group (p = 0.05). Recovery or refilling time was 50 +/- 21 sec for the surgical group and 28 +/- 26 sec for the medical group (p = 0.03).(ABSTRACT TRUNCATED AT 250 WORDS)