Forty patients suspected clinically the first episode of deep vein thrombosis (DVT) were studied using a new computerized strain gauge plethysmographic (SGP) system, the Phlebotest. A capacitance mode was used to secure the obtaining of maximum venous filling before deflation in every determination. Following parameters were calculated automatically and instantly: the maximum volume change (V sec), the outflow rate during the first second after deflation (F 1.0), the expelled volume in four seconds (EV 4.0) and the surface area over the curve during the first four seconds after deflation (IND). The eighty lower limbs of the patients were divided into three groups according to the Duplex scanning results: proximal (n = 19), distal (n = 13) and no DVT (n = 48) groups. The diagnostic accuracy of the Phlebotest parameters was determined and compared with that of another widely used SGP system, the Perivein. In this case the maximum volume change and the maximum venous outflow (MVO) were calculated manually. The results of the two systems were compatible. For the detection of proximal DVT, the parameters of venous outflow were better than maximum volume change and the best single parameter was maximum venous outflow: F 1.0 for Phlebotest and MVO for Perivein. The two parameters had a similar sensitivity of 79% when 50 and 27 ml/100 ml/min were used as the lower limits of normal respectively. The best criterion for the diagnosis of proximal DVT was F 1.0 < 50 ml/100 ml/min plus the EV 4.0/V sec ratio < 60%, which had a sensitivity of 90%, specificity of 96% and overall accuracy of 94%.(ABSTRACT TRUNCATED AT 250 WORDS)