Standard non-invasive monitoring during anaesthesia usually includes pulse oximetry saturated pressure of oxygen (SPO2) and end-tidal carbon dioxide (end-tidal CO2). They frequently fail to monitor blood gas changes during cardiac surgery due to reduction in perfusion, ambient light, hypothermia and the use of vasoconstrictors . It is also impossible to monitor SPO2 during cardiopulmonary bypass (CPB) because of the lack of pulsatile flow. Intermittent blood gas monitoring, the rule in cardiac surgery, often fails to detect the rapid changes that are frequent in the beginning and end of bypass. Continuous blood gas monitoring with the Paratrend 7 (Diametrics, Manchester, UK) could therefore be an interesting option. The aim of this prospective observational study was to compare conventional intermittent blood gas monitoring with the continuous Paratrend system in patients undergoing cardiac surgery with bypass.
The study was approved by the hospital Ethics Committee. Ten consecutive patients (age 70 ± 5 yr) were included with written informed consent. Anaesthesia was induced with etomidate, midazolam and sufentanil with cisatracurium or rocuronium for muscle relaxation. Anaesthesia was maintained with isoflurane. The inspired oxygen fraction during anaesthesia was 50% and this increased if needed to keep SPO2 ≥97%. Monitoring included invasive blood pressure (BP), eight-lead ECG, central venous pressure (CVP), SPO2 and end-tidal CO2. Blood gases were assessed intermittently in the laboratory and continuously with the Paratrend 7 system.
Two anaesthetists took part in the study. One was responsible for the standard patient care according to the departmental guidelines. He had access to the conventional blood gas analysis, SPO2 and end-tidal CO2. However, he was blinded to the Paratrend data. A second anaesthetist monitored the continuous Paratrend blood gas data. Abnormal Paratrend data were classified as Level 1 (abnormal, but not life-threatening values) or Level 2 (dangerous). Level 1 and 2 limits are defined in Table 1. Level 2 abnormalities, with the exception of arterial pressure of oxygen (PaO2) ≥26 kPa, were immediately communicated to the anaesthetist in charge of patient care. The number of therapeutic actions performed on the basis of the results of conventional blood gas analysis, SPO2 and end-tidal CO2 were noted.
An average of 6.6 blood gas values per patient was outside the predefined limits. The mean duration of Level 1 and 2 abnormalities for arterial pressure of oxygen and carbon dioxide (PaO2 and PaCO2, respectively), and pH in absolute time and as a percentage of operating time are shown in Table 1.
The main abnormalities found with the Paratrend system were hyperoxaemia and hypocapnia. These deviations were also discovered by intermittent blood sampling but with a significant delay. According to the Paratrend data, the patients were hyperoxaemic >31% of the operating time. PaO2 often exceeded 26 kPa (average duration 57 min). Hyperoxaemia has been implicated in microcirculatory deterioration  and exacerbation of the inflammatory process post-bypass .
The risk of hypoxaemia is substantial during cardiac surgery with CPB. It can be due to the patient's condition or to mechanical failure. However, episodes of intraoperative hypoxaemia found with the Paratrend were much less frequent than hyperoxaemia, representing only 8% of operating time. These episodes were revealed with a delay by traditional intermittent blood gas sampling, requiring the blind to be lifted on five occasions (Table 1). It seems that patients are voluntarily maintained hyperoxaemic, as anaesthetists fear hypoxaemia more than hyperoxaemia.
The information provided by pulse oximetry played only a minor part in the management of the patients. In 10% of operating time (bypass excluded), the SPO2 signal was not reliable. Anaesthetists tend to act only in cases of low SPO2. In our patients SPO2 never fell below 95%. SPO2 cannot discriminate fast changes and normal values can be associated with critical levels of PaO2, especially after weaning from CPB. Thus, continuous blood gas monitoring revealed episodes of hypoxaemia during anaesthesia that were not detected by pulse oximetry or intermittent blood gas analysis.
The Paratrend revealed that the patients were hypocapnic during more than 50% of the operation time. In 24 instances it was severe enough (PaCO2 < 4 kPa) to qualify for blind lifting. Hypocapnia was also detected in 50% of the conventional blood gas samples but was followed by ventilatory modifications in only one-third of the cases. Low endtidal CO2 (<4 kPa) was recorded on average during 60% of operating time, but was not followed by any therapeutic action. This voluntary maintenance of hypocapnia may be explained by the fact that anaesthetists fear hypoventilation more than hyperventilation, which is considered less dangerous. The CO2 level plays an important part in the adjustment of cerebral microcirculation, and extreme values have been held responsible for neurological disorders post-bypass .
Continuous blood gas monitoring is a useful tool to detect metabolic acidosis, which was present during 15% of operating time in our patients. We encountered eight serious episodes that required lifting of the blind. In the majority of the episodes, the acidosis was related to the haemodynamic state of the patient. Sometimes acidosis is the first sign of peripheral hypoperfusion, which can be masked by normal systolic BP.
In conclusion, we have shown that there may be long delays before blood gas abnormalities during cardiac operations are discovered by conventional intermittent blood sampling, in combination with pulse oximetry and capnography, as compared to continuous Paratrend monitoring. In 10% of cases, by the time we had obtained the blood gas results, the acid-base status of the patient had changed outside the limits. Rapid changes in acid-base status occur at the beginning and at the end of bypass. It is important to be aware of the limitations of the traditional intermittent technique.
J. C. Marty
Département d'Anesthésie Réanimation; Hôpital du Bocage; Dijon, France
1. Reich DL, Timcenko A, Bodian CA, et al.
Predictors of pulse oxymetry data failure. Anesthesiology
2. Joachimsson PO, Sjöberg F, Forsman M, Johansson M, Ahn HC, Rutberg H. Adverse effects of hyperoxemia during cardiopulmonary bypass. J Thorac Cardiovasc Surg
3. Kotani N, Hashimoto H, Sessler DI, et al.
Supplemental intraoperative oxygen augments antimicrobial and proinflammatory response of alveolar macrophages. Anesthesiology
4. Nevin M, Adams S, Colchester ACF, Pepper JR. Evidence for involvement of hypocapnia and hypoperfusion in aetiology of neurological deficit after cardiopulmonary bypass. Lancet