Miniaturized sensors, based upon the principles of optical fluorescence, can measure in vivo the pH, pCO2 value and pO2 value of blood. In this report we studied continuous intra-arterial blood gas monitoring in 27 patients undergoing cardiac surgery (no. 16 coronary artery by-pass grafting, no. 2 valvular surgery) and major vascular surgery (no. 9 abdominal aortic aneurysms). Total duration of continuous intra-arterial blood gas monitoring was 677 hours, with a ratio of 25.0 +/- 14.8 hours/patient (range 4-96 hours). The in vitro values of pH, pCO2 and pO2 were compared to simultaneous records from the fiberoptic sensor for each of the 283 arterial blood gas samples obtained, by means of linear regression and Bland-Altman method, in order to test the correlation and the agreement between the two methods of measuring. For pH average bias was -0.023 and intersensor precision was 0.028, with a strong correlation (R = 0.92; p < 0.001) and agreement. For pCO2 the average bias was 0.91 and the inter-sensor precision was 2.65, with a slight decrease in correlation (R = 0.89; p < 0.001) and agreement. For pO2 average bias was -2.69 and the intersensor precision was 12.16, with a strong correlation (R = 0.97; p < 0.001) and agreement. In addition, we tested the reliability of the system for values of pO2 above 100 mmHg and we found a strong correlation (R = 0.96; p20.001) and agreement even for these clinical conditions, largely out of physiologic parameters. This study demonstrates the feasibility and reliability of continuous intra-arterial three-component PB 3300 (Puritan Bennett) blood gas monitoring.