The purposes of this study were to evaluate the efficiency of occlusion strain gauge plethysmography in the diagnosis of symptomatic deep venous thrombosis in upper extremities, to compare early (one month) and late clinical evolution (many years) to plethysmography data and, finally, to set up normal and pathological values. Four groups of patients were tested. Sixteen patients with confirmed acute thrombosis, 6 patients with unconfirmed thrombosis, 15 with an old thrombosis (mean follow-up: 7 years) as well as 15 healthy volunteers. The outflows measurements were significantly lowered (p less than 0.05) in the arms with acute thrombosis compared to the contralateral arms, to the affected arms of patients with thrombosis excluded and to both volunteers' arms. Pathological values for the different outflow parameters could be defined as follows: Maximum venous output less than 110 ml/100 ml/minute, Outflow after 3 seconds less than 1.2 ml/100 ml, Strandness index less than 0.16. On month after the acute episode we found a good correlation between the usually favourable clinical evolution and venous haemodynamics assessed by plethysmography. The mild late post-thrombotic syndrome of most patients with an old thrombosis was not well demonstrated by plethysmography.