| pubmed-article:7818045 | pubmed:abstractText | Continuous monitoring of blood gases and pH could add substantially to patient safety. During the last decade, efforts have been made to develop continuous optochemical blood gas sensors. The initial evaluation of such fibreoptic-based systems showed major patient-interface problems [11]. We evaluated a new intra-arterial blood gas monitoring system (PB3300, Puritan-Bennett, Carlsbad CA) under routine clinical conditions. METHODS. After institutional review board approval and with written informed consent, 38 sensors were tested in 25 patients with acute respiratory failure (e.g., the acute respiratory distress syndrome, complications after lung transplantation). Two conventional bench-top blood gas analysers (ABL 520 and ABL 300, Radiometer, Copenhagen) served as criterion standards. The mean differences (bias) and standard deviations (SD) of the differences (precision) were calculated according to the method of Bland and Altman [2]. In addition, linear regression analysis and correlation coefficients were calculated. The quality of blood pressure tracings was assessed using a grading system. RESULTS. The median sensor lifetime was 81.3 h; 869 blood samples (median 14 per sensor) were analysed for the comparison of continuous and conventional blood gas analysis. The ranges for measured parameters were: pH: 6.92 to 7.55; PCO2: 20 to 83 torr; PO2: 31 to 518 torr. The mean differences (SD) were: pH: -0.03 (0.03) or -0.4 (0.4)%; PCO2: -2.6 (4.1) torr or -6.9 (10.9)%; PO2: -3.4 (10.5) torr or -2.9 (7.0)%. The results of linear regression analysis and the correlation coefficients are depicted in Table 2. The mean grade of blood pressure tracings was satisfactory for the clinical setting. CONCLUSIONS. The continuous blood gas monitor is sufficiently accurate and precise for clinical use. Bias and precision are better than those known from former studies evaluating fibreoptic blood gas monitors under experimental conditions [7]. Cost-effectiveness was not an issue of this study. | lld:pubmed |
