Because the use of radioactive fibrinogen uptake test (FUT) has become questionable both for ethical (risk of virus transmission) and technical (lack of sensitivity) reasons, we investigated the potential value of two alternative methods for screening of asymptomatic deep venous thrombosis following elective digestive surgery: liquid crystal contact thermography (LCCT) and measurement of plasma concentration of D-dimer (DD), as compared with bilateral ascending phlebography. Out of 194 patients, 185 underwent phlebography on the 8th (0-19, median and range) postoperative day. Despite prophylaxis with low-molecular-weight heparin and elastic stockings, DVT was detected on phlebography in 58 legs of 45 patients. Sensitivity of LCCT with respect to the presence of DVT was 55% (n = 184 patients) or 28% (n = 368 legs) with a specificity of 67% and 82%, respectively. These poor performances were obtained despite a good interobserver agreement for the LCCT assessments (overall kappa coefficient of 0.66 between three experts). The most accurate cut-off of DD for discriminating patients with or without DVT was 3,000 micrograms/l, as determined by ROC curve analysis. Sensitivity of a DD level of more than 3,000 micrograms/l for the presence of phlebographically documented DVT on the 8th postoperative day was 89% for a specificity of 48%. Thus, LCCT cannot be used for screening of postoperative, mainly asymptomatic DVT following general surgery. On the other hand, measurement of plasma DD may be useful for initial screening, a negative result (level less than 3,000 micrograms/l) allowing to exclude DVT (negative predictive value of 93%) and a positive result (positive predictive value of 35%) requiring confirmation by phlebography.(ABSTRACT TRUNCATED AT 250 WORDS)