A general anaesthetic technique that includes mechanical ventilation impairs pulmonary gas exchange . This is mainly because atelectasis develops in dependent lung regions, the area correlating with shunting of pulmonary blood flow, and because some regions develop a low ventilation/perfusion (V˙A/Q˙) ratio [2–4]. In anaesthetized patients with normal lungs, V˙A/Q˙ mismatch is usually less than 10% of the cardiac output and thus only slightly increased inspired oxygen is required to maintain oxygenation during anaesthesia and surgery. However, in clinical practice a higher inspired oxygen (close to 50%) is usually used to prevent possible hypoxia.
It is well known that high inspired oxygen can alter gas exchange . The usual explanation is development of absorption atelectasis and the inhibition of hypoxic vasoconstriction in low V˙A/Q˙ lung units [6,7]. This effect was shown for inspired oxygen above 60% , but studies have shown that even values less than 60% may also alter the distribution of V˙A/Q˙ [9,10]. These studies were carried out on awake subjects breathing spontaneously or during general anaesthesia and mechanical ventilation, but only in patients with diseased lungs or when mechanical ventilation was for 12 h.
Therefore, we chose to assess the possible early deleterious effect of inspired oxygen 50% administered intra-operatively to patients with healthy lungs undergoing laparotomy under general anaesthesia and mechanical ventilation of shorter duration (4–6 h) than in previous studies.
Patients and methods
Ethics Committee approval was obtained and all subjects gave informed consent. All patients were ASA Grade I and scheduled for elective abdominal surgery under general anaesthesia with mechanical ventilation. A computer generated code was used to allocate patients randomly to receive either 50% inspired oxygen or the lowest inspired oxygen (not allowed to exceed 30%) to maintain a pulse oximeter reading of at least 90%. Patients with a saturation of less than 97% immediately pre-operatively, while breathing air and before the injection of any intravenous (i.v.) sedation, were excluded. The operations were expected to last between 4 and 6 h and elective extubation at the end of surgery was planned. No haemorrhage was expected. Patients were withdrawn from the study if there were any intra-operative events likely to alter gas exchange or to necessitate altering the ventilator settings.
Anaesthesia and mechanical ventilation
All patients received oral lorazepam 1 mg the evening before surgery and oral hydroxyzine 2 mg kg−1 2 h before the start of surgery. Anaesthesia was induced with i.v. midazolam (0.03–0.1 mg kg−1), fentanyl (3–9 μg kg−1) and thiopentone (4–9 mg kg−1) and maintained with additional doses of fentanyl (1–1.5 μg kg−1) and isoflurane in oxygen/nitrous oxide as necessary. Muscle paralysis to allow tracheal intubation was obtained with pancuronium bromide (0.1 mg kg−1) and then maintained with intermittent doses of 2 mg i.v. according to the neuromuscular blockade monitor. Heat loss was limited by a heating pad. Ringer lactated solution was infused intra-operatively. If necessary, hydroxyethylstarch (Elohes, Biosedra Laboratory, France) was infused for volume expansion.
The patients' lungs were ventilated using a Dräger SA1 ventilator and circle absorber system. The same ventilator was used for all the patients. Tidal volume (VT) was at 10 mL kg−1 ideal weight  and the frequency was adjusted to maintain an end-tidal carbon dioxide pressure (PetCO2) of ≈4.3 KPa. FIO2, PetCO2, V˙T, and pulse oximetry were monitored continuously during surgery by a gas analyser and a spirometer (Capnomac, Datex Instrumentarium Corporation). Pulse oximetry was also monitored in the post-anaesthesia care unit. Post-operative analgesia was started only after the last blood sample for the study had been taken.
Procedure and statistics
Arterial blood samples were drawn by direct radial artery puncture for measurement of pH, PaO2 and PaCO2 (Corning 175 automatic pH/blood gas analyser system, Corning Medical Medfield MA). Measurements were made (temperature corrected) at steady state after induction of anaesthesia (T1) and during skin closure (T2). A third measurement (T3), 1 h after extubation, while breathing spontaneously was made only if saturation remained above 90% after a 10 min period breathing room air.
The PaO2/FIO2 ratio was used to assess intra-operative venous admixture. We compared intra-operative and post-operative oxygenation on the basis of the PaO2/FIO2 ratio at the time of incision and during surgical wound closure (within and between groups) and post-operative PaO2 between groups. As in Register's study , we considered a 10% difference to be clinically significant. Setting the α risk at 5% and β risk at 10%, 20 patients were needed in each group.
A two-ways Anova for repeated measures followed by the Neuman Keuls test was used to compare PaO2/FIO2. For the PaO2/FIO2 ratio, standard error of the mean (SEM) was calculated. For patients characteristics and intra-operative data standard deviation (SD) was calculated.
Between January and June 1993, 43 consecutive patients gave informed consent to the study: 20 received oxygen 50% and 23 received oxygen less than 30%. Three patients receiving the lower concentration were withdrawn: one because the surgery lasted only 2 h 30 min and two because intra-operative acute events necessitated increasing the inspired oxygen (intra-operative haemorrhage and hypoventilation). The other 20 patients in the group were given 25% inspired oxygen.
Patients' characteristics were similar as were their ventilator settings (Table 1). The time between the first and the second arterial sample, reflecting the duration of the surgical procedure, was longer in patients receiving oxygen 50% because the duration of surgery was longer. PaO2/FIO2 ratio, measured at the beginning and at the end of the surgical procedure, did not vary within or between groups (Table 2).
Post-extubation PaO2 was not available from two patients from each group. It was not significantly different between the two groups (P = 0.62): FiO2 = 0.25, PaO2 = 11.3 ± 0.4 kPa; FiO2 = 0.50, PaO2 = 11.6 ± 0.5 kPa.
We have shown that increasing the inspired oxygen from 25% to 50% does not influence intra- and post-operative gas exchange during surgical procedures lasting 4–6 h in anaesthetized patients with healthy lungs.
General anaesthesia increases venous admixture [1–4] as a consequence of the rapid development of atelectatic areas in dependent lung regions with low V˙A/Q˙ ratio. Moreover, most studies on oxygen induced oxygenation impairment have been carried out at inspired oxygen of more than 50%. However, venous admixture associated with oxygen breathing occurs even if less than 100% oxygen is used [5–7,9,10,12]. With lesser inspired oxygen concentrations, either the oxygen was given for more than 16 h or there was appreciable lung disease. For example, Register et al.'s patients  were mechanically ventilated for 16–24 h after cardiac surgery. In their study the post-extubation PaO2 of patients who had received 50% oxygen was significantly lower than in those who had received less than 30% oxygen, but cardiac surgery itself causes respiratory dysfunction and in the investigation, oxygenation had been prolonged. The patients in the present study were scheduled for abdominal surgery under general anaesthesia which certainly leads to intra- and post-operative atelectasis. They had healthy lungs and intra-operative mechanical ventilation was expected to have a duration of 4–6 h. The only difference between the groups was the time of the first and the second arterial sampling, because interval between the duration of surgery was greater in the patients who received 50% oxygen. If anything this should have exaggerated the effect of the higher concentrations of oxygen on pulmonary oxygenation, and thus enhance the power of the study.
A criticism of the present study is that we were looking for a 10% difference in PaO2 between groups. Some might argue this difference is not clinically significant. On the contrary, we believe that any improvement in gas exchange is worthwhile. Abdominal surgery can be complicated by post-operative respiratory complications, even in patients with no preexisting lung disease .
Another criticism is that the present study was based on blood gas analysis when there are methods that are more sensitive indices of gas exchange: direct measurement of V˙A/Q˙ ratio , measurement of alveolar arterial oxygen tension gradient , and evaluation of atelectasis by computorized tomography scan . These techniques are difficult to use in the immediate post-operative period; this is why data was selected that could be collected more easily at the bedside. The PaO2/FIO2 ratio correlates well with pulmonary venous admixture when the inspired gas is oxygen 100%, but there is evidence that using oxygen 100% overestimates the shunt, thus allowing assessment at lower inspired oxygen . Moreover, the PaO2/FIO2 ratio cannot have been altered by variation in PaCO2, which was monitored and kept constant throughout the study.
Supplemental oxygen could have masked arterial hypoxaemia caused by impaired gas exchange at the time of the third arterial sample, but post-extubation samples were taken when the patients were breathing room air. Finally nitrous oxide may have masked the improvement in gas exchange at lower inspired oxygen, its high delivery volumes leading theoretically to absorption atelectasis. However, three recent studies failed to find any effect of changing nitrous oxide concentrations on pulmonary oxygenation [17–19].
In conclusion, mechanical ventilation with 50% inspired oxygen for 4–6 h did not by itself worsen oxygenation in the present study. However, intra-operative mechanical ventilation is sometimes just the start of a more prolonged period of post-operative mechanical ventilation, which can cause problems with gas exchange if inspired oxygen is high. Intra-operative mechanical ventilation with 25% inspired oxygen proved feasible and safe for patients with no lung disease. Modern pulse oximetry will detect early desaturation. We therefore suggest that there is no need for a routine inspired oxygen concentration greater than 25% in healthy patients, especially as the analgesic effect of intra-operative nitrous oxide is greater at a concentration of more than 50%.
The authors gratefully acknowledge the secretarial help of Nathalie Papillon and Agnès Vanoutryve, and the valuable advice of Frédéric J. Mercier.
1 Marshall BE, Wiche MO. Hypoxemia during and after anesthesia. Anesthesiology
2 Strandberg A, Brismar B, Hedenstierna G, Lindquist H, Tokics L. Atelectasis during anaesthesia and in the post-operative period. Acta Anaesthesiol Scand
3 Hedenstierna G, Strandberg A, Tokics L, Lindquist H, Brismar B. Correlation of gas exchange impairment to development of atelectasis during anaesthesia and muscle paralysis. Acta Anaesthesiol Scand
4 Gunnarsson L, Strandberg A, Brismar B, Tokics L, Lindquist L, Hedenstierna G. Atelectasis and gas exchange impairment during enflurane/nitrous oxide anaesthesia. Acta Anaesthesiol Scand
5 Wagner PD, Laravuso RB, Uhl RR et al.
Continuous distribution of ventilation perfusion ratios in normal subjects breathing air and 100% oxygen. J Clin Invest
6 Douglas ME, Downs JB, Dannemiller FJ et al.
Changes in pulmonary venous admixture
with varying inspired O2
. Anesth Analg
7 Eiser NM, Jones HA, Hughes JMB. Effect of 30% O2
on local matching of perfusion and ventilation in chronic airway obstruction. Clin Sci Mol Med
8 Burger EJ, Macklem PT. Airway closure: demonstration by breathing 100% O2
at low lung volumes and by N2
washout. J Appl Physiol
9 Dantzker DR, Wagner PD, West JB. Instability of lung units with low V˙a/Q˙ ratios during oxygen breathing. J Appl Physiol
10 Register SD, Downs JB, Stock C, Kirby RR. Is 50% harmful? Crit Care Med
11 Metropolitan Life Insurance Company. Statistical Bulletin
12 West JB. Blood flow to the lung and gas exchange. Anesthesiology
13 Pedersen T, Eliasen K, Henriksen E. A prospective study of risk factors and cardiopulmonary complications associated with anaesthesia and surgery: Risk indicators of cardiopulmonary morbidity. Acta Anaesthesiol Scand
14 West JB. New advances in pulmonary gas exchange
. Anesth Analg
15 West JB. Ventilation/Blood Flow and Gas Exchange,
3rd Edn. Oxford: Blackwell Scientific Publications, 1992: 83–85.
16 Quan SF, Kronberg GM, Schlobom RM, Feeley TW, Don HF, Lister G. Changes in venous admixture
with alteration of inspired oxygen concentration. Anesthesiology
17 Roberts CJ, Parke TJ, Sykes MK. Effect of intraoperative inspired gas mixtures on postoperative nocturnal oxygen saturation. Br J Anaesth
18 Einarsson S, Stenqvist O, Bengtsson A, Houltz E, Bengtsson JP. Nitrous oxide elimination and diffusion hypoxia during normo and hypoventilation. Br J Anaesth
19 Jensen AG, Kalman SH, Eintrei C, Fransson SG, Morales O. Atelectasis and oxygenation in major surgery with either propofol with or without nitrous oxide or isoflurane anaesthesia. Anaesthesia