The strength of agreement between two noninvasive methods of assessing lower limb arterial disease and their relationship to patient symptoms following exercise have been investigated. Color-duplex ultrasound (CDU) and ankle/brachial pressure index (ABPI) (before and afer exercise) measurements were obtained from 200 consecutive patients referred to a vascular investigations laboratory. From these patients, 290 limbs were available for study, comprising limbs without previous vascular surgery, from patients without diabetes and who could attempt a walking exercise test. The overall level of agreement between CDU and resting ABPI measurements was 83% (Kappa 0.66). The ABPI technique identified the more serious disease; a resting ABPI of less than 0.6 gave 100% agreement with CDU. With higher resting ABPIs the level of agreement became poorer: 83% (0.6 < or = ABPI <0.9) and 76% (normal ABPI > or = 0.9). The addition of postexercise ABPI measurements in determining significant arterial disease increased the strength of relationship between the two techniques by only 2% (85%, Kappa 0.69). The exercise test was generally limited by the most symptomatic limb in each patient, and the agreement between CDU and postexercise ABPI measurements in these limbs was higher at 93% (Kappa 0.81). In comparison, agreement for the least symptomatic group of limbs was found to be poor (69%, Kappa 0.37). Compared with symptoms after exercise, overall agreements with CDU and ABPI were both 67% (Kappa 0.27). The agreement was better (91%) when the resting ABPI was less than 0.6. The ABPI is biased toward the detection of more severe disease and is more consistent with CDU when the most symptomatic limbs are compared. The relationship between either test and symptoms after exercise is strong only for limbs with major disease.