Automatic Control of Veno‐Venous Extracorporeal Lung Assist

Veno‐venous extracorporeal lung assist (ECLA) can provide sufficient gas exchange even in most severe cases of acute respiratory distress syndrome. Commercially available systems are manually controlled, although an automatically controlled ECLA could allow individualized and continuous adaption to clinical requirements. Therefore, we developed a demonstrator with an integrated control algorithm to keep continuously measured peripheral oxygen saturation and partial pressure of carbon dioxide constant by automatically adjusting extracorporeal blood and gas flow. The “SmartECLA” system was tested in six animal experiments with increasing pulmonary hypoventilation and hypoxic inspiratory gas mixture to simulate progressive acute respiratory failure. During a cumulative evaluation time of 32 h for all experiments, automatic ECLA control resulted in a peripheral oxygen saturation ≥90% for 98% of the time with the lowest value of 82% for 15 s. Partial pressure of venous carbon dioxide was between 40 and 49 mm Hg for 97% of the time with no value <35 mm Hg or >49 mm Hg. With decreasing inspiratory oxygen concentration, extracorporeal oxygen uptake increased from 68 ± 25 to 154 ± 34 mL/min (P < 0.05), and reducing respiratory rate resulted in increasing extracorporeal carbon dioxide elimination from 71 ± 37 to 92 ± 37 mL/min (P < 0.05). The “SmartECLA” demonstrator allowed reliable automatic control of the extracorporeal circuit. Proof of concept could be demonstrated for this novel automatically controlled veno‐venous ECLA circuit.     

Continuous arterial and venous blood gas monitoring during cardiopulmonary bypass

A new monitoring technique, based on optical fluorescence chemistry, allows continuous monitoring of all blood gas variables during cardiopulmonary bypass. To evaluate the clinical performance of this monitor, we drew 220 arterial and 216 venous blood samples from 15 patients, and simultaneous blood gas values displayed by the monitor were compared with standard laboratory measurements. The continuous monitor predicted laboratory values with varying degrees of accuracy. (R2 values by linear regression: arterial oxygen tension 0.86, venous oxygen tension 0.36, arterial carbon dioxide tension 0.58, venous carbon dioxide tension 0.72, arterial pH 0.53, venous pH 0.58; pH 0.53, venous pH 0.58; p less than 0.0001). Monitor values of arterial oxygen tension overestimated laboratory values (bias = + 43.5 mm Hg), but the laboratory reference method likely underestimated true arterial oxygen tension in the high range achieved on bypass. Monitoring of venous oxygen tension was imprecise (precision = +/- 6.51 mmHg), regardless of whether stable conditions existed during the sampling period. Monitoring of carbon dioxide tension and pH showed small bias (carbon dioxide tension within 2 mm Hg, pH within 0.03) and good precision (carbon dioxide tension within 3 mm Hg, pH within 0.03). With the development of unstable conditions on bypass, monitor arterial oxygen tension values showed a changing relationship to corresponding laboratory values. In conclusion, arterial and venous carbon dioxide tension and pH monitoring provide acceptably accurate alternatives to laboratory measurement of these variables during cardiopulmonary bypass. Arterial oxygen tension monitoring accurately indicates changes in oxygen tension in the arterial oxygen tension range typically produced during extracorporeal circulation. Oxygen tension monitoring in the venous oxygen tension range is too imprecise for clinical decision-making purposes.     

A comparison of blood and Fluosol-DA for cardiopulmonary bypass

Fluosol-DA was compared to blood as a pump prime for total cardiopulmonary bypass in the pig animal model. Nineteen pigs weighing between 14 and 22 kg were studied, nine with blood and ten with Fluosol. Metabolic and hemodynamic measurements were determined before, during and after 60 minute bypass to establish the adequacy of Fluosol to sustain perfusion as compared to blood. The measurements and subsequent calculations included blood gases, arterial and mixed venous oxygen content, oxygen extraction and consumption, cardiac output, systemic and pulmonary vascular resistance and arterial, venous, pulmonary artery and left atrial pressures. The result showed a significant decrease in hematocrit during bypass in the Fluosol group as compared to blood perfusion (20 vs. 30%). While the arterial oxygen content fell from control levels with Fluosol during bypass, in the blood prime group, oxygen content remained at pre-control levels. Whole body oxygen consumption decreased during bypass, in both groups equally, but this decrease did not lead to acidosis and was stable during recovery. Oxygen and carbon dioxide transport were adequately maintained during bypass in both Fluosol and blood groups. Systemic pressures remained stable during bypass and were lower, but stable, during recovery. Pulmonary vascular resistance was elevated in both groups during recovery which probably explains a concomitantly decreased cardiac output. There was a 40% mortality in both experimental groups secondary to postpump pulmonary hypertension. It is concluded that Fluosol is a satisfactory oxygen carrying agent to be used instead of blood during cardiopulmonary bypass, and in the pig model both blood and Fluosol were associated with a high incidence of pulmonary hypertension.     

An Educational Training Simulator for Advanced Perfusion Techniques Using a High‐Fidelity Virtual Patient Model

The operation of cardiopulmonary bypass procedure requires an advanced skill in both physiological and mechanical knowledge. We developed a virtual patient simulator system using a numerical cardiovascular regulation model to manage perfusion crisis. This article evaluates the ability of the new simulator to prevent perfusion crisis. It combined short‐term baroreflex regulation of venous capacity, vascular resistance, heart rate, time‐varying elastance of the heart, and plasma‐refilling with a simple lumped parameter model of the cardiovascular system. The combination of parameters related to baroreflex regulation was calculated using clinical hemodynamic data. We examined the effect of differences in autonomous‐nerve control parameter settings on changes in blood volume and hemodynamic parameters and determined the influence of the model on operation of the control arterial line flow and blood volume during the initiation and weaning from cardiopulmonary bypass. Typical blood pressure (BP) changes (hypertension, stable, and hypotension) were reproducible using a combination of four control parameters that can be estimated from changes in patient physiology, BP, and blood volume. This simulation model is a useful educational tool to learn the recognition and management skills of extracorporeal circulation. Identification method for control parameter can be applied for diagnosis of heart failure.     

Drew-Anderson technique attenuates systemic inflammatory response syndrome and improves respiratory function after coronary artery bypass grafting

BACKGROUND: Cardiopulmonary bypass causes inflammatory reactions leading to organ dysfunction postoperatively. This study was undertaken to determine whether using patients' own lungs as oxygenator in a bilateral circuit (Drew-Anderson Technique) could reduce systemic inflammatory response to cardiopulmonary bypass, improving patients clinical outcome following coronary artery bypass grafting. METHODS: A prospective randomized controlled trial involving 30 patients, divided in two groups of 15 patients each, undergoing elective coronary artery bypass grafting, was undertaken. In the Drew-group bilateral extracorporeal circulation using patient's lung as oxygenator was performed. The other patients served as control group, where standard cardiopulmonary bypass procedure was used. RESULTS: Pro-inflammatory and anti-inflammatory mediators were measured. Peak concentrations of proinflammatory interleukin-6, interleukin-8, were significantly lower in 15 patients undergoing Drew-Anderson Technique compared with the concentrations measured in 15 patients treated with standard cardiopulmonary bypass technique. Differences in patient recovery were analyzed with respect to time of intubation, blood loss, intrapulmonary shunting, oxygenation, and respiratory index. In patients undergoing uncomplicated coronary artery bypass grafting procedures bilateral extracorporeal circulation using the patients' own lung as oxygenator provided significant biochemical and clinical benefit in comparison to the standard cardiopulmonary bypass procedure. CONCLUSIONS: This prospective randomized clinical study has demonstrated that exclusion of an artificial oxygenator from cardiopulmonary bypass circuit significantly decreases the activation of inflammatory reaction, and that interventions that attenuate this response may result in more favorable clinical outcome.     

Characteristics of a Nonocclusive Pressure‐Regulated Blood Roller Pump

For decades, extracorporeal life support (ECLS) systems have relied on pumps designed for short‐term cardiopulmonary bypass. In the past, occlusive roller pumps were the standard. They are being progressively replaced by centrifugal pumps and devices developed specifically for ECLS. However, the ideal pump for long‐term bypass is yet to be created. One interesting alternative is the Rhône‐Poulenc 06 pump that is a nonocclusive pressure‐regulated blood pump developed in France in the 1970s. This pump is composed of a double‐stage rotor with three rollers at each level. The raceway tubing is stretched on the roller and pump occlusivity depends on the tension of the chamber on the rotor. The pump is able to deliver physiological blood flow values without generating dangerous negative or positive pressures. The specific design of the chamber allows the pump to generate a pulsatile flow, inducing minimal blood trauma, and to act as a bubble trap, making it inherently safe. This pump has been used for cardiopulmonary bypass, extracorporeal lung support, and more specifically single‐lumen single‐cannula venovenous membrane oxygenation for neonates, left‐heart or right‐heart assist, and venovenous bypass during liver transplant. In conclusion, this old‐fashion pump is perfectly adapted for any kind of short‐ or long‐term bypass.     

Cerebral blood flow response to changes in arterial carbon dioxide tension during hypothermic cardiopulmonary bypass in children

We examined the relationship of changes in partial pressure of carbon dioxide on cerebral blood flow responsiveness in 20 pediatric patients undergoing hypothermic cardiopulmonary bypass. Cerebral blood flow was measured during steady-state hypothermic cardiopulmonary bypass with the use of xenon 133 clearance methodology at two different arterial carbon dioxide tensions. During these measurements there was no significant change in mean arterial pressure, nasopharyngeal temperature, pump flow rate, or hematocrit value. Cerebral blood flow was found to be significantly greater at higher arterial carbon dioxide tensions (p less than 0.01), so that for every millimeter of mercury rise in arterial carbon dioxide tension there was a 1.2 ml.100 gm-1.min-1 increase in cerebral blood flow. Two factors, deep hypothermia (18 degrees to 22 degrees C) and reduced age (less than 1 year), diminished the effect carbon dioxide had on cerebral blood flow responsiveness but did not eliminate it. We conclude that cerebral blood flow remains responsive to changes in arterial carbon dioxide tension during hypothermic cardiopulmonary bypass in infants and children; that is, increasing arterial carbon dioxide tension will independently increase cerebral blood flow.     

The Effects of Intrathoracic Pressure During Continuous Two‐Lung Ventilation for Thoracoscopy on the Cardiorespiratory Parameters in Sevoflurane Anaesthetized Dogs

The cardiopulmonary effects of different levels of carbon dioxide insufflation (3, 5 and 2 mmHg) under two‐lung ventilation were studied in six sevoflurane (1.5 minimum alveolar concentration; MAC) anaesthetized dogs during left‐sided thoracoscopy. An arterial catheter, Swan–Ganz catheter and multianaesthetic gas analyser were used to monitor the cardiopulmonary parameters during the experiment. Baseline data were obtained before intrathoracic pressure elevation and the measurements were repeated at intervals after left lung collapse induced by insufflation with carbon dioxide gas. The intrapleural pressure levels used were 3, 5 and 2 mmHg. Arterial blood pressures, cardiac index, stroke index, left and right ventricular stroke work index, arterial haemoglobin saturation, arterial oxygen tension and systemic vascular resistance decreased significantly during hemithorax insufflation, whereas heart rate, right atrial pressure, mean, systolic and diastolic pulmonary arterial pressure, pulmonary capillary wedge pressure, pulmonary vascular resistance and arterial carbon dioxide tension significantly increased during intrapleural pressure elevation. Although carbon dioxide insufflation into the left hemithorax with an intrapleural pressure of 2–5 mmHg compromises cardiac functioning in 1.5 MAC sevoflurane anaesthetized dogs, it can be an efficacious adjunct for thoracoscopic procedures. Intrathoracic view was satisfactory with an intrapleural pressure of 2 mmHg. Therefore, the intrathoracic pressure rise during thoracoscopy with two‐lung ventilation should be kept as low as possible. Additional insufflation periods should be avoided, since a more rapid and more severe cardiopulmonary depression can occur.     

膜式氧合器在胎羊体外循环中的运用

目的　为了改进胎羊体外循环技术 ,探讨膜式氧合器在胎羊体外循环中的应用。　方法　将健康怀孕山羊8只 ,采用 Dideco 90 1膜式氧合器和滚轴泵建立胎羊体外循环 ,常温 (37℃ )转流 6 0分钟 ,氧合器内充低氧混合气体 (8%O2 和 92 % N2 ) ,监测胎羊的血压、心率、血气、血清乳酸和胎盘血管阻力。　结果　胎羊体外循环中动脉氧分压 (PO2 )和二氧化碳分压 (PCO2 )维持在宫内生理水平 ,胎羊心搏有力 ,血压正常。但胎羊 p H值缓慢下降 (P<0 .0 5 ) ,血清乳酸值明显增高 (P<0 .0 1) ,胎盘血管阻力显著上升 (P<0 .0 1)。停体外循环后胎羊出现低氧、高碳酸血症和酸中毒。　结论　胎羊体外循环影响胎盘功能 ,膜式氧合器可以代替胎盘气体交换功能 ,体外循环中胎羊生理低水平 PO2 是否适合其需要值得探讨。     

Changes in jugular bulb oxygenation in patients undergoing warm coronary artery bypass surgery (34-37 degrees C)

BACKGROUND AND OBJECTIVE: Imbalance between cerebral oxygen supply and demand is thought to play an important role in the development of cerebral injury during cardiac surgery with cardiopulmonary bypass. METHODS: We studied jugular bulb oxygen saturation, jugular bulb oxygen tension, arterial-jugular bulb oxygen content difference and oxygen extraction ratio in 20 patients undergoing warm coronary artery bypass surgery (34-37 degrees C) with pH-stat blood gas management. RESULTS: Only two patients showed desaturation (jugular bulb oxygen saturation < 50%) at 5 min on bypass, and none from 20 min onwards. Multiple regression models were performed after using bypass temperature, mean arterial pressure, cerebral perfusion pressure, haemoglobin concentration and arterial carbon dioxide tension as independent variables, and arterial-jugular bulb oxygen content difference, jugular bulb oxygen saturation, oxygen extraction ratio and jugular bulb oxygen tension as individual dependent variables. CONCLUSIONS: We found that jugular bulb oxygen saturation, jugular bulb oxygen tension and oxygen extraction ratio are mainly dependent on arterial carbon dioxide tension, and arterial-jugular bulb oxygen content difference is dependent on arterial carbon dioxide tension and the bypass temperature. Our results suggest jugular bulb oxygenation is mainly dependent on arterial carbon dioxide tension during warm cardiopulmonary bypass.     

The oxygen transporting capability of neo red cells (NRC) evaluated under total cardiopulmonary bypass

The oxygen transporting capability of an artificial oxygen carrier NRC was evaluated by employing it as a perfusate for total cardiopulmonary bypass. NRC is a type of liposome encapsulated hemoglobin. It has a particle size of approximately 220 nm, with a hemoglobin concentration of 5.6 g/dl and its P₅₀ is controlled to 45 Torr. Male beagles were used in the experiment. Approximately 80% of the estimated circulatory volume was exchanged with NRC and total cardiopulmonary bypass was initiated. Arterial oxygen tension and carbon dioxide tension were controlled to 400 Torr and 40 Torr respectively. The perfused we heated to 37°C. The rate of flow was altered during the experiment. Oxygen consumption reached a plateau at 9.3 ml/kg/min where oxygen delivery was 14.9 ml/kg/min. At this point the oxygen consumed per gram of hemoglobin from NRC was equivalent to that from dog red blood cells. This indicated that almost an equal amount of oxygen was consumed from NRC in comparison to red blood cells. Regarding oxygen transporting capability, NRC could be considered a candidate for perfusate in cardiopulmonary bypass.     

Analysis of the hemodynamic and ventilatory effects of laparoscopic cholecystectomy.

Laparoscopic cholecystectomy uses carbon dioxide, a highly diffusable gas, for insufflation. With extended periods of insufflation, patient arterial carbon dioxide levels may be adversely altered. Patients were selected for laparoscopic cholecystectomy using the same criteria as for open cholecystectomy. Twenty patients (group 1) had normal preoperative cardiopulmonary status (American Society of Anesthesiologists class I), while 10 patients (group 2) had previously diagnosed cardiac or pulmonary disease (class II or III). Demographic, hemodynamic, arterial blood gas, and ventilatory data were collected before peritoneal insufflation and at intervals during surgery. Patients with preoperative cardiopulmonary disease demonstrated significant increases in arterial carbon dioxide levels and decreases in pH during carbon dioxide insufflation compared with patients without underlying disease. Results of concurrent noninvasive methods of assessing changes in partial arterial pressures of carbon dioxide (end-tidal carbon dioxide measured with mass spectrographic techniques) may be misleading and misinterpreted because changes in partial arterial pressures of carbon dioxide are typically much smaller than changes in arterial blood levels and, unlike arterial gas measurements, do not indicate the true level of arterial hypercarbia. During laparoscopic cholecystectomy, patients with chronic cardiopulmonary disease may require careful intraoperative arterial blood gas monitoring of absorbed carbon dioxide.     

Use of extracorporeal membrane lung assist device (Novalung) in H1N1 patients

Abstract We present three patients with severe respiratory failure secondary to H1N1 influenza type A pneumonitis, in whom hypercapnia and respiratory acidosis were not controlled by the conventional mechanical lung ventilation or high‐frequency oscillatory ventilation. Use of a pumpless arteriovenous extracorporeal carbon dioxide removal device (Novalung?, Inspiration Healthcare Ltd, Leicester, UK) resulted in reduced carbon dioxide levels, improved pH, and a reduction in inspiratory pressures, allowing for a less‐harmful ventilator strategy. These cases demonstrate that the Novalung is a safe and effective device to use in patients with H1N1 pneumonitis refractory to the conventional therapy and may be an alternative to extracorporeal membrane oxygenation (ECMO) in selected cases. (J Card Surg 2011;26:449‐452)     

The effect of ventilation on systemic blood gases in the presence of left ventricular ejection during cardiopulmonary bypass

The effect of pulmonary ventilation upon systemic arterial blood gases during cardiopulmonary bypass in the presence of left ventricular ejection was evaluated in 20 adult male patients undergoing coronary artery bypass grafting. Following rewarming, establishment of a sinus rhythm, and production of a pulse pressure of at least 20 mm Hg on the arterial pressure trace caused by left ventricular ejection, arterial blood gases were obtained from the arterial and venous extracorporeal circuits and the radial arterial cannula. Patients were then randomly assigned to a nonventilation (n = 10) or a ventilation (n = 10) group. The ventilation group was given 10 breaths/min with 100% oxygen at a tidal volume of 10 ml/kg. Whereas the nonventilation group received apneic oxygenation at zero end-expiratory pressure. After 5 minutes the arterial blood gas data were again obtained. Significant findings (p less than 0.05) included decreases in systemic carbon dioxide tension and increases in systemic pH in the ventilation group and decreases in systemic oxygen tension in the nonventilation group. Although the changes in the arterial blood gases were significant, these changes occurred well within the limits of clinical acceptability. It is concluded that left ventricular ejection for short periods during full cardiopulmonary bypass does not necessitate pulmonary ventilation.     

Anticoagulation by ancrod for carbon dioxide removal by extracorporeal membrane lung in the dog

Ten dogs were subjected to defibrinogenation with an intravenous perfusion of ancrod (1 unit/kg) (Arvin, Knoll AG, Ludwigshafen, Federal Republic of Germany) over a 2 1/2 hour period. Six of them were subjected to extracorporeal elimination of carbon dioxide with a polypropylene membrane lung by means of veno-venous bypass. The remaining four dogs did not undergo extracorporeal circulation and served as control subjects. In both groups, ancrod administration itself resulted in a marked drop in alpha 2-antiplasmin (33% and 67%, respectively, of the baseline values) and in slight but significant decreases in factor II and plasminogen activities of 25% and 20%, respectively (p less than 0.05), in the group subjected to carbon dioxide removal. There were no significant changes in platelet number or factor V and antithrombin III activities. During the 6-hour bypass period, platelet count and antithrombin III and factor II and V levels decreased significantly. No bleeding was observed. Histologic examination of lung biopsy tissue showed no pathologic features. Analysis of the membrane of the artificial lungs revealed no fibrin deposits. In the control group, except for a drop in alpha 2-antiplasmin levels (54%), no significant changes in hemostatic parameters occurred during the corresponding 6 hours. We conclude that, despite the drop in coagulation factors and in alpha 2-antiplasmin activity during bypass, ancrod can be considered as a valuable alternative anticoagulant for extracorporeal carbon dioxide removal.     

Effects of a platelet-activating factor antagonist on lung injury and ventilation after cardiac operation

BACKGROUND: Platelet-activating factor is a mediator of lung injury during cardiac operation. Platelet-activating factor antagonists reduce lung injury in animal models of cardiopulmonary bypass but there is no confirmatory evidence in clinical practice. METHODS: The effect of a low or high dose of a platelet-activating factor antagonist (Lexipafant) was assessed in a single center, double-blind, placebo-controlled, parallel group study. One hundred fifty patients undergoing coronary artery bypass grafting were randomized by minimization into three groups to receive placebo infusion, 10 or 100 mg of lexipafant for over 24 hours. Serial arterial oxygen and carbon dioxide tension, alveolar arterial oxygen gradient, and percent saturation were measured before operation and at 1, 6, 24, 48 hours, and 5 days after operation. RESULTS: Patient groups were similar with respect to age, sex, body surface area, and urgency of operation. Likewise, the groups were similar with respect to duration of cardiopulmonary bypass and the number and type of grafts. Maximum lung injury occurred at 48 hours when the arterial oxygen tension and percent saturation reached a nadir (both p < 0.001) accompanied by the maximum increase in the alveolar arterial gradient (p < 0.001). All measurements demonstrated partial recovery by 5 days but remained significantly (p < 0.001) impaired in comparison to baseline values. Duration of ventilation was similar in all groups. Lexipafant, at low or high dose, did not moderate lung injury after cardiopulmonary bypass and did not influence the duration of postoperative ventilation. CONCLUSIONS: Despite experimental and clinical evidence implicating platelet-activating factor in the pathophysiology of lung injury after cardiopulmonary bypass, no beneficial effect of a platelet-activating factor antagonist on lung function or ventilation could be demonstrated in this clinical trial.     

Measuring the efficiency of an artificial lung: 1. Carbon dioxide transfer.

The performance of a hollow fibre artificial lung ('Capiox E') was analysed by measurement of the 'parallel deadspace' of the device under varying conditions in 21 patients. The efficiency with which carbon dioxide was exchanged was determined by the time available for equilibration between the blood and gas phases. When this equilibration coefficient was less than 12 seconds per litre of blood flow per litre of gas flow, there was a marked reduction in the efficiency of gas exchange. Under certain conditions, the 'counter-current' design of the device apparently permitted the clearance of carbon dioxide at a partial pressure greater than that which was found in the mixed venous blood. This anomalous behaviour may represent in vivo confirmation of the Haldane effect.     

Non-invasive cardiac output measurement using a fast mixing box to measure carbon dioxide elimination

This study investigated the accuracy of a new technique for measuring cardiac output using the derivative Fick principle based on the ratio of change in the partial pressures of end-tidal and mixed expired carbon dioxide produced by short periods of partial rebreathing. A prospective clinical study involving 24 patients following cardiopulmonary bypass for coronary artery bypass grafting or valvular surgery was undertaken in the intensive care unit of a university-affiliated hospital. Haemodynamic measurements were performed after admission to the intensive care unit. Cardiac output was measured simultaneously by bolus pulmonary artery thermodilution and by a non-invasive carbon dioxide partial rebreathing technique. Cardiac output measurement using the new technique demonstrated a significant but consistent underestimate, with a bias of -0.60 +/- 0.87 l/min. This new adaptation of the partial rebreathing technique is reliable in measuring cardiac output in postoperative patients. Reasons for the consistent discrepancy between thermodilution and partial rebreathing techniques are discussed.     

Pulsatile Augmentation Device for Extracorporeal Circulation

Abstract: A device has been designed, constructed, and tested to provide pulsatile pressure/flow to a standard extracorporeal bypass circuit. The pulsatile augmentation device is pneumatically driven similar to an artificial heart ventricle except that there are no valves. It is constructed of polyurethane by vacuum forming and high frequency welding. Drivers used are a modified Arrow-Kontron in-traaortic balloon pump or the Utah artificial heart driver. In vitro testing with fresh bovine blood demonstrated acceptable blood compatibility and hemodynamic function. In vivo testing for 4 h in a right and left heart extracorporeal bypass circuit showed good pulse augmentation in pulmonary and systemic bypass circuits. The device shows promise for adding pulse to standard cardiopulmonary bypass and to extracorporeal right heart circulatory assist circuits.     

Oxygenator exhaust capnography as an index of arterial carbon dioxide tension during cardiopulmonary bypass using a membrane oxygenator

We have studied the relationship between the partial pressure of carbon dioxide in oxygenator exhaust gas (PECO2) and arterial carbon dioxide tension (PaCO2) during hypothermic cardiopulmonary bypass with non- pulsatile flow and a membrane oxygenator. A total of 172 paired measurements were made in 32 patients, 5 min after starting cardiopulmonary bypass and then at 15-min intervals. Additional measurements were made at 34 degrees C during rewarming. The degree of agreement between paired measurements (PaCO2 and PECO2) at each time was calculated. Mean difference (d) was 0.9 kPa (SD 0.99 kPa). Results were analysed further during stable hypothermia (n = 30, d = 1.88, SD = 0.69), rewarming at 34 degrees C (n = 22, d = 0, SD = 0.84), rewarming at normothermia (n = 48, d = 0.15, SD = 0.69) and with (n = 78, d = 0.62, SD = 0.99) or without (n = 91, d = 1.07, SD = 0.9) carbon dioxide being added to the oxygenator gas. The difference between the two measurements varied in relation to nasopharyngeal temperature if PaCO2 was not corrected for temperature (r2 = 0.343, P = < 0.001). However, if PaCO2 was corrected for temperature, the difference between PaCO2 and PECO2 was not related to temperature, and there was no relationship with either pump blood flow or oxygenator gas flow. We found that measurement of carbon dioxide partial pressure in exhaust gases from a membrane oxygenator during cardiopulmonary bypass was not a useful method for estimating PaCO2.