Continuous jugular bulb venous oximetry has been validated previously for periods of up to 12 h after calibration. We assessed a new Edslab venous oximetry catheter in 15 patients with brain injury in the intensive care unit. After insertion into the jugular bulb, the catheter was calibrated using a laboratory co-oximeter. Subsequent comparisons were made at varying intervals (range 20 min to 100 h); 78 paired samples were obtained during periods of stable recordings. Estimation of haemoglobin saturation by the methods correlated well (r2 = 0.97, P < 0.0001). Mean difference (d) was small (0.28% (SD 2.35%)) and limits of agreement (d +/- (SD x 1.96)) were acceptable (-4.88% to 4.32%). There was no appreciable drift and there was negligible bias. This catheter did not suffer from problems of light intensity described with other systems, and it provided acceptable accuracy for clinical use for periods of up to and exceeding 24 h after calibration.