Impact of trauma centre capacity and volume on the mortality risk of incoming new admissions.

INTRODUCTION: Trauma centre capacity and surge volume may affect decisions on where to transport a critically injured patient and whether to bypass the closest facility. Our hypothesis was that overcrowding and high patient acuity would contribute to increase the mortality risk for incoming admissions. METHODS: For a 6-year period, we merged and cross-correlated our institutional trauma registry with a database on Trauma Resuscitation Unit (TRU) patient admissions, movement and discharges, with average capacity of 12 trauma bays. The outcomes of overall hospital and 24 hours mortality for new trauma admissions (NEW) were assessed by multivariate logistic regression. RESULTS: There were 42 003 (mean=7000/year) admissions having complete data sets, with 36 354 (87%) patients who were primary trauma admissions, age ≥18 and survival ≥15 min. In the logistic regression model for the entire cohort, NEW admission hospital mortality was only associated with NEW admission age and prehospital Glasgow Coma Scale (GCS) and Shock Index (SI) (all p<0.05). When TRU occupancy reached ≥16 patients, the factors associated with increased NEW admission hospital mortality were existing patients (TRU >1 hour) with SI ≥0.9, recent admissions (TRU ≤1 hour) with age ≥65, NEW admission age and prehospital GCS and SI (all p<0.05). CONCLUSION: The mortality of incoming patients is not impacted by routine trauma centre overcapacity. In conditions of severe overcrowding, the number of admitted patients with shock physiology and a recent surge of elderly/debilitated patients may influence the mortality risk of a new trauma admission.

Pleural decompression and drainage during trauma reception and resuscitation.

This review examines pleural decompression and drainage during initial hospital adult trauma reception and resuscitation, when it is indicated for haemodynamically unstable patients with signs of pneumothorax or haemothorax. The relevant historical background, techniques, complications and current controversies are highlighted. Key findings of this review are that: 1. Needle thoracocentesis is an unreliable means of decompressing the chest of an unstable patient and should only be used as a technique of last resort. 2. Blunt dissection and digital decompression through the pleura is the essential first step for pleural decompression, as decompression of the pleural space is a primary goal during reception of the haemodynamically unstable patient with a haemothorax or pneumothorax. Drainage and insertion of a chest tube is a secondary priority. 3. Techniques to prevent tube thoracostomy (TT) complications include aseptic technique, avoidance of trocars, digital exploration of the insertion site and guidance of the tube posteriorly and superiorly during insertion. 4. Whenever possible, blunt thoracic trauma patients should undergo definitive CT imaging after TT to check for appropriate tube position.

Video techniques and data compared with observation in emergency trauma care.

Video recording is underused in improving patient safety and understanding performance shaping factors in patient care. We report our experience of using video recording techniques in a trauma centre, including how to gain cooperation of clinicians for video recording of their workplace performance, identify strengths of video compared with observation, and suggest processes for consent and maintenance of confidentiality of video records. Video records are a rich source of data for documenting clinician performance which reveal safety and systems issues not identified by observation. Emergency procedures and video records of critical events identified patient safety, clinical, quality assurance, systems failures, and ergonomic issues. Video recording is a powerful feedback and training tool and provides a reusable record of events that can be repeatedly reviewed and used as research data. It allows expanded analyses of time critical events, trauma resuscitation, anaesthesia, and surgical tasks. To overcome some of the key obstacles in deploying video recording techniques, researchers should (1) develop trust with video recorded subjects, (2) obtain clinician participation for introduction of a new protocol or line of investigation, (3) report aggregated video recorded data and use clinician reviews for feedback on covert processes and cognitive analyses, and (4) involve multidisciplinary experts in medicine and nursing.

Chest wall and lung mechanics during acute hemorrhage in anesthetized dogs.

OBJECTIVES: In trauma and in surgical patients, respiratory mechanics may change because of many factors, including the hypotension induced by hemorrhage. The effects of acute hemorrhage on elastic and resistive characteristics of the respiratory system were studied. DESIGN: Prospective study. SETTING: Anesthesia research laboratory. INTERVENTIONS: Acute hemorrhagic shock was induced in 24 supine anesthetized/paralyzed, mechanically ventilated dogs by blood withdrawal over a 12-minute period to decrease systolic arterial pressure to 50 mmHg; additional blood was subsequently withdrawn to maintain this pressure for 2 hours. Total respiratory system dynamic compliance and resistance and lung and chest wall compliances and resistances were measured. MEASUREMENTS AND MAIN RESULTS: Total respiratory system dynamic compliance decreased from control (0.03 +/- 0.002 L/cmH2O) by the first 10 minutes of shock (p < 0.05) and was 9.8 +/- 2% lower than control 2 hours after the induction of shock because of decreases in both lung (9.6 +/- 3%) and chest wall (7.7 +/- 3%) compliances. Total respiratory resistance increased 12.8 +/- 3% from control (3.08 +/- 0.19 cmH2O/L/s) after 2 hours of shock (p < 0.05) because of an increase in chest wall resistance (21.6 +/- 8%, p < 0.05). Pulmonary resistance was not significantly increased (p > 0.05). In six control dogs, prepared similarly but not hemorrhaged, chest wall compliance and resistance did not change, but lung compliance gradually decreased by 17.8% during 150 minutes of anesthesia/paralysis. Lung resistance increased only after 100 minutes (p < 0.05). CONCLUSIONS: (1) Hemorrhagic shock caused slight changes in the chest wall, but effects on lung mechanics were a consequence of prolonged mechanical ventilation during anesthesia/paralysis, and (2) changes in respiratory mechanics caused by hemorrhagic shock are small and, unless other deleterious factors are present, would probably have little clinical significance.

Respiratory impedances and acinar gas transfer in a canine model for emphysema.

We examined how the changes in the acini caused by emphysema affected gas transfer out of the acinus (Taci) and lung and chest wall mechanical properties. Measurements were taken from five dogs before and 3 mo after induction of severe bilateral emphysema by exposure to papain aerosol (170-350 mg/dose) for 4 consecutive wk. With the dogs anesthetized, paralyzed, and mechanically ventilated at 0.2 Hz and 20 ml/kg, we measured Taci by the rate of washout of 133Xe from an area of the lung with occluded blood flow. Measurements were repeated at positive end-expiratory pressures (PEEP) of 10, 5, 15, 0, and 20 cmH2O. We also measured dynamic elastances and resistances of the lungs (EL and RL, respectively) and chest wall at the different PEEP and during sinusoidal forcing in the normal range of breathing frequency and tidal volume. After final measurements, tissue sections from five randomly selected areas of the lung each showed indications of emphysema. Taci during emphysema was similar to that in control dogs. EL decreased by approximately 50% during emphysema (P < 0.05) but did not change its dependence on frequency or tidal volume. RL did not change (P > 0.05) at the lowest frequency studied (0.2 Hz), but in some dogs it increased compared with control at the higher frequencies. Chest wall properties were not changed by emphysema (P > 0.05). We suggest that although large changes in acinar structure and EL occur during uncomplicated bilateral emphysema, secondary complications must be present to cause several of the characteristic dysfunctions seen in patients with emphysema.

Response time studies of a new, portable mass spectrometer.

OBJECTIVE: Mass spectrometers are frequently used by anesthesiologists perioperatively to monitor patients' respiratory function and levels of inhaled anesthetics. Due to size, complexity and expense, they are typically used in a time-sharing manner which degrades their performance. We assessed the accuracy of the Random Access Mass Spectrometer (RAMS), Marquette Electronics) which is small enough to be dedicated to a single patient. METHODS: We compared the 10-90% rise times for O2, CO2, N2O and isoflurane for the RAMS with different catheter configurations to those of a MedSpect mass spectrometer (Allegheny International Medical Technology) operating under ideal conditions. For CO2 the lag of the RAMS relative to the MedSpect was also measured. Next, perioperative conditions were stimulated by ventilating anesthetized dogs with a variety of inhalatory gases and ventilatory parameters, and the interchangeability of the two devices was assessed. RESULTS: When fitted with a catheter with minimal dead space the MedSpect had rise times of 0.11-0.12 sec while the RAMS had rise times of 0.07-0.12 sec and a delay of 0.19 sec compared to the MedSpect. The rise times and delay of the RAMS increased when using a larger catheter and water trap. Although there were statistically significant differences in some values for inhaled and end-tidal gases under simulated perioperative conditions, particularly at the higher frequencies, these differences were small and for most purposes not clinically significant. CONCLUSIONS: Our results demonstrate that the RAMS configured for clinical conditions performs nearly as well as the MedSpect under ideal conditions. The small differences between the two, confined almost entirely to their end-tidal CO2 values, could be due to differences in instrument calibration, by the larger sampling catheter commonly used in clinical settings, or by a combination of both factors. Therefore the RAMS is sufficiently accurate for clinical use and would alleviate problems associated with time-shared mass spectrometers.

Comparison of self-reporting of deficiencies in airway management with video analyses of actual performance. LOTAS Group. Level One Trauma Anesthesia Simulation.

We compared the performance deficiencies of airway management captured by three types of self-reports with those identified through video analysis. The three types of self-reports were the anesthesia record (a patient record constructed during the course of treatment), the anesthesia quality assurance (AQA) report (a retrospective report as a part of the trauma center's quality assurance process), and a posttrauma treatment questionnaire (PTQ), which was completed immediately after the case for the purposes of this research. Video analysis of 48 patient encounters identified 28 performance deficiencies related to airway management in 11 cases (23%). The performance deficiencies took the form of task omissions or practices that lessened the margin of patient safety. In comparison, AQA reports identified none of these performance deficiencies, the anesthesia records identified 2 (of 28), and the PTQs suggested contributory factors and corrective measures for 5 deficiencies. Furthermore, video analysis provided information about the context of and factors contributing to the identified performance deficiencies, such as failures in adherence to standard operating procedures and in communications.

Task complexity in emergency medical care and its implications for team coordination. LOTAS Group. Level One Trauma Anesthesia Simulation.

To elicit components of task complexity in emergency medical care, a study was conducted to contrast one medical procedure with two levels of task urgency in trauma patient resuscitation. Videotapes of actual resuscitation were reviewed to extract task characteristics of the procedure. Two levels of urgency were compared in the following areas: patient status, technical difficulty of tasks, the amount of available patient monitoring information, and the pace of work. Four components of task complexity in emergency medical care were identified: multiple and concurrent task, uncertainty, changing plans, and compressed work procedures and high workload. These components of task complexity pose challenges to team functions and can lead to problems in team coordination, such as conflicts in goals, tasks, and access to the patient. Training to increase explicit communications and improvements in the design of work procedures are necessary in order to meet the challenges of task complexity.

Tetrodotoxin infusion: nonventilatory effects and role in toxicity models.

OBJECTIVES: To determine the cardiovascular, autonomic, and neuromuscular effects of an IV infusion of tetrodotoxin (TTX) when ventilation is supported. METHODS: TTX was infused in 18 anesthetized beagles during conventional mechanical ventilation. TTX infusion continued at a rate of 9.3 micrograms/kg/hr until apnea occurred with 1 minute of ventilator disconnection. Measurements included intravascular pressures, heart rate (HR), cardiac output, blood gases, displacements of the rib cage and abdomen, O2 delivery, and responses to train-of-four and tetanic peripheral nerve stimulation. Results are expressed as mean +/- SD. RESULTS: During TTX infusion, all the dogs had discoordinate movements of the rib cage, abdomen, and limbs. Vomiting, urination, defecation, and increased salivation occurred. Nicotinic and muscarinic effects, neuromuscular blockade, and cardiovascular depression were produced by TTX. Apnea occurred in 72.0 +/- 17.0 minutes when a total of 119.0 +/- 17.4 micrograms of TTX was infused. At apnea, decreases in arterial pressure, cardiac index, HR, O2 delivery, and systemic vascular resistance occurred, while pulmonary artery pressure and pulmonary vascular resistance increased. Loss of response to tetanic stimulation was closely correlated with the dose of TTX that produced apnea. CONCLUSION: The clinical symptoms and signs of TTX poisoning are similar to those of anticholinesterase poisons, and TTX dosing as described by this model may serve as a surrogate for organophosphorus poisoning. The model may be useful to determine optimum therapies for TTX poisoning and, since TTX prevents sodium influx into cells, to investigate enhanced survival in animals suffering from ischemia.

Efficacy of several modes of continuous-flow insufflation for resuscitation of a canine model of acute respiratory arrest.

STUDY OBJECTIVE: To test the efficacy of several modes of continuous-flow insufflation on the maintenance of physiologic parameters in a model of respiratory arrest, and the effect of these modes on neurologic outcome. METHODS: Anesthetized dogs were slowly infused with tetrodotoxin over 75 minutes to the point of respiratory arrest. We used two different modes of continuous-flow insufflation: endobronchial insufflation (EI) of air 3 cm distal to the carina (.25 or 1.0 L.kg-1.min-1); and tracheal insufflation of oxygen (TRIO) 1 cm proximal to the carina (.08 or .2 L.kg-1.min-1). RESULTS: EI at either flow rate provided ventilation sufficient to allow the dogs to recover effective spontaneous breathing and be removed from ventilation after 4 hours. By this time, almost all cardiovascular variables and blood gas values were normal. TRIO at .2 L.kg-1.min-1 also resulted in successful recovery, although Pa02, as well as systemic and pulmonary arterial pressures and vascular resistances, remained increased at the end of the 4-hour period. TRIO at the low flow rate, however, resulted in deterioration of blood gas values and systemic arterial pressure; dogs required conventional mechanical ventilation after 45 minutes of low-flow TRIO. CONCLUSION: EI can be used to maintain oxygenation in acute respiratory arrest when conventional techniques are not feasible; TRIO at .2 L.kg-1.min-1 is also effective.

Liver function and morphology after resuscitation from severe hemorrhagic shock with hemoglobin solutions or autologous blood.

OBJECTIVE: To test the effects of three hemoglobin solutions on liver function and hepatic morphology after resuscitation from severe hemorrhagic shock. DESIGN: Prospective study. SETTING: Laboratory. SUBJECTS: Thirty-three beagle dogs. INTERVENTION: Hemorrhagic shock was induced in anesthetized dogs by removal of blood at a rate of 2 mL/kg/min until systolic blood pressure (BP) reached 50 mm Hg. BP was maintained at this level 2 hrs by further withdrawing 5 to 10 mL aliquots whenever BP increased > 50 mm Hg. Resuscitation was then initiated with autologous whole blood (n = 7), 4% pyridoxalated-hemoglobin-polyoxyethylene conjugate (4% PHP [n = 6]), 8% pyridoxalated-hemoglobin-polyoxyethylene conjugate (8% PHP [n = 9], or 8% stroma-free hemoglobin (n = 7). Four dogs were managed identically but were not resuscitated. Gross necropsy and histologic examination of the liver were performed on all dogs after 7 days, or earlier if death occurred. MEASUREMENTS AND MAIN RESULTS: In vitro interferences of PHP and stroma-free hemoglobin with liver function tests were determined and recommendations for interpretation of results from blood samples containing PHP and stroma-free hemoglobin were made. Blood was collected before, during, and after resuscitation from hemorrhagic shock. The dogs were then awakened and survivors were monitored daily with blood sampling until they were killed and necropsy was performed. After 7 days, the survival rate following hemorrhagic shock was 100% for whole blood and 4% PHP, 86% for stroma-free hemoglobin, and 33% for 8% PHP. Of the resuscitated dogs not surviving 7 days, all but one died within 27 hrs from coagulopathy. All dogs not resuscitated died within 1.75 hrs after 2 hrs of shock. Bilirubin, alkaline phosphatase, and lactic dehydrogenase concentrations could not be measured due to interferences of stroma-free hemoglobin and PHP. Aspartate (AST) and alanine (ALT) aminotransferase concentrations could be measured after dilution to overcome the interferences. Significant increases in AST and ALT values in all groups 24 hrs after resuscitation were attributed to hypoxic hepatocellular damage associated with the severity of the shock model rather than to the resuscitation fluid. Liver histology showed no changes attributed to toxic damage of hepatocytes in dogs resuscitated with stroma-free hemoglobin or PHP. However, changes, were less severe in dogs resuscitated with 4% PHP than in other groups. CONCLUSION: Morphologic studies at necropsy and liver function tests in dogs receiving hemoglobin solutions, compared with autologous blood, support the conclusion that the PHP and stroma-free hemoglobin solutions tested did not produce hepatic toxicity when used as resuscitation fluids in this model of severe shock.

Effects of PEEP on acinar gas transfer in healthy and lung-injured dogs.

We measured cardiorespiratory variables and 133xenon washout from a nonperfused lung region (XeW) in six anesthetized/paralyzed dogs, mechanically ventilated with 60% O2 at different positive end-expiratory pressures (PEEP). XeW in this technique represents directly measured acinar gas transfer (3). Measurements were repeated after induction of lung injury by lavaging the lungs 11 to 13 times with 600 ml saline. In control dogs, lung compliance (CL), alveolar ventilation (Valv), and XeW all decreased with increasing PEEP from 0 to 25 cm H2O (p < 0.05), while lung resistance (RL) did not change. After lavage, CL, Valv, and XeW below 15 cm H2O PEEP were all less than control values (p < 0.05), while RL was higher than control values. As PEEP increased from 0 to 20 cm H2O, Valv and XeW increased, but CL did not change; RL decreased only from 0 to 5 cm H2O. At 20 cm H2O PEEP, Valv and CL were not different from control values (p > 0.05), and XeW was higher than control values (p < 0.05). At estimated alveolar volumes above 400 ml, values for XeW before and after lavage were similar. We conclude that, during severe lung injury: (1) increasing PEEP to moderate levels will increase acinar gas transfer but, after a certain lung volume is reached, further increases in PEEP will have effects similar to the healthy condition; (2) overall mechanical properties of the lung do not reflect the responses to PEEP of the lung periphery.

An audio-video system for automated data acquisition in the clinical environment. LOTAS Group.

OBJECTIVE: Our objective was to develop an audiovideo data acquisition system that facilitates studying the activities of anesthesia care providers in the clinical environment. METHOD: Ceiling-mounted miniature video cameras, vital sign monitors, and videocassette recorders (VCRs) were interfaced to digital computers in two patient admitting areas and two operating rooms of a trauma center. This video data acquisition system network (VASNET) is simple to operate. Insertion of a videotape activates the system and begins video overlay of updated vital signs onto the video image every 5 sec. Recorded data is passed via a local area network, allowing remote monitoring of the data acquisition process. To facilitate analysis of the video at a later time, the image, soundtrack, and vital signs data are stamped with the same time code. Each tape is initialized by recording the data file name and wall clock time for 30 sec at the start of taping. This initialization enables comparison of the video recordings with anesthesia, surgical, and nursing records. RESULTS: During 2 years of operation, VASNET was used to record over 100 cases of acute trauma management. Vital signs overlaid onto the video image identified when patient monitors were in use and providing data. Participants found videotape review useful in assessing their own performance. VASNET was nonintrusive and acquired data with minimum user interaction. In one operating room, separate from the trauma center, VASNET was installed to function as a remote monitor, with the option of videotaping. Although users were aware of when videotaping occurred, once patient management was underway, the activities of the anesthesia care providers did not appear to be influenced by the videocassette recording. Equipment maintenance was not excessive. The most frequent problems were changes to the VCR control settings and disconnection of the power supply or interface connections. CONCLUSIONS: Videotapes of the process of anesthetizing and resuscitating trauma patients provided a record of the activities of anesthesia care providers. Video vignettes may be useful training tools. Excerpts from real scenarios can be incorporated into anesthesia stimulators. The soundtrack and timing of real events from such video acquisition may be useful in the development of multimedia simulations of trauma patient resuscitation. The data collection may be useful for research into human performance, ergonomics, training techniques, quality assurance, and certification of anesthesia care providers in trauma patient management. Potential additional applications of VASNET include remote monitoring of patients in the operating room, in the intensive care unit, during transportation, in hazardous environments, and in the field. Such VASNET telemetry may facilitate the availability of expert opinions during medical and other consultations.

Oxygen transport and cardiovascular effects of resuscitation from severe hemorrhagic shock using hemoglobin solutions.

OBJECTIVE: To test the short-term efficacy of three hemoglobin solutions in restoring cardiac output, intravascular pressures, oxygen transport (DO2), and oxygen consumption (VO2) after resuscitation from severe hemorrhagic shock. DESIGN: Prospective study. SETTING: Research laboratory. SUBJECTS: Beagle dogs. INTERVENTIONS: After anesthesia and instrumentation, hemorrhagic shock was induced for 2 hrs by blood withdrawal to maintain systolic blood pressure at 50 mm Hg. Resuscitation then occurred with one of four different resuscitation fluids. One group of dogs was not resuscitated. Survival rate was monitored for 8 days. MEASUREMENTS AND MAIN RESULTS: In 33 beagle dogs, cardiovascular variables (DO2 and VO2) were compared after resuscitation with 8% stroma-free hemoglobin, 4% or 8% pyridoxalated-hemoglobin-polyoxyethylene conjugate (PHP44 and PHP88, respectively), or autologous whole blood. The dogs were anesthetized, paralyzed, mechanically ventilated (FIO2 of 0.21), and instrumented with arterial and pulmonary artery catheters. An average of 63% of estimated blood volume was removed to maintain systolic blood pressure at 50 mm Hg for 2 hrs. The dogs then were either not resuscitated (n = 4) or resuscitated with 8% stroma-free hemoglobin (n = 7), PHP44 (n = 6), PHP88 (n = 8), or whole blood (n = 8), with a volume equivalent to the withdrawn blood. Cardiovascular variables, DO2, VO2, oxygen extraction ratios, and blood concentrations of lactic acid and catecholamines were determined before, and for < or = 6 hrs after, resuscitation from hemorrhagic shock. Blood smears were microscopically examined. In addition, the survival rate was monitored for 8 days after resuscitation. By 2 hrs of hemorrhagic shock, there was a large decrease in DO2 (p < .05) and an increase in oxygen extraction ratio from 0.27 to 0.70 (p < .05). There was a 3.5-fold increase in lactate concentrations and a 25-fold increase in catecholamine concentrations as compared with preshock values. All dogs not resuscitated died within 1.75 hrs after 2 hrs of shock. After resuscitation with whole blood, all cardiovascular and oxygen transport variables returned to approximately prehemorrhage values and remained so throughout the measurement period. After resuscitation with any hemoglobin solution, DO2 returned transiently to control values. However, recovery of DO2 was short-lived in all hemoglobin solution groups, and, by 4 hrs postresuscitation in all groups, DO2 was less than the DO2 of the dogs receiving whole blood (p < .05). These changes were associated with decreases in total hemoglobin concentrations compared with the values immediately before resuscitation (p < .05). In addition, with resuscitation using the PHP solutions, blood smears demonstrated aggregation of red blood cells and platelets. On day 8 after hemorrhagic shock, the survival rate was 100% for whole blood and PHP44, 86% for 8% stroma-free hemoglobin, and 33% for PHP88. CONCLUSIONS: Resuscitation from severe hemorrhagic shock with 8% stroma-free hemoglobin, PHP44, or PHP88 is equally effective in restoring cardiac index and vascular pressures as using whole blood. However, resuscitation with the three hemoglobin solutions only transiently restored DO2 after hemorrhagic shock. The subsequent reduction of DO2 compared with the DO2 value using whole blood was due mostly to hemodilution. With the two PHP solutions, formation of red blood cell aggregates probably resulted in sequestration of red cell mass and additional loss of oxygen carrying capacity.

In vitro oxyhaemoglobin saturation measurements in haemoglobin solutions using fibreoptic pulmonary artery catheters.

We compared in vitro oxyhaemoglobin saturations using two pulmonary artery catheters (catheter SO2), with oxyhaemoglobin saturations (SO2) measured by the IL282 co-oximeter and derived partial oxyhaemoglobin saturations (partial SO2) at different oxygen tensions (PO2) in six solutions: whole blood, 50:50 mixture of whole blood and Plasmalyte A (haemodiluted blood), 50:50 mixture of whole blood and 8% pyridoxylated haemoglobin-polyoxyethylene (PHP) conjugate (WB-PHP), 75:25 mixture of 8% PHP and Plasmalyte A solution (PHP66), 50:50 mixture of 8% PHP and Plasmalyte A solution (PHP44) and stroma-free haemoglobin solution (SFH). Calculated P50 values (PO2 vs SO2) were 3.79, 3.58, 3.49, 3.15, 3.04 and 2.07 kPa, respectively. However, if partial SO2 was used the curves were shifted to the left, reducing P50. Catheter SO2 correlated well with SO2 in whole blood (r2 > 0.99 for both catheters), haemodiluted blood (r2 > 0.98 for both catheters) and WB-PHP solution (r2 = 0.94 for both catheters). In PHP44 (r2 = 0.64 and r2 = 0.57), PHP66 (r2 = 0.40 for the Oximetrix and r2 = 0.25 for the Edwards catheter) and SFH solutions (r2 = 0.33 for the Oximetrix and r2 = 0.22 for the Edwards catheter) both catheters performed poorly. We conclude that mixed venous oxyhaemoglobin saturations measured by oximetric pulmonary artery catheters are inaccurate in the presence of haemoglobin solutions. For accuracy a multi-wavelength co-oximeter should be used if blood containing PHP or SFH is to be analysed.