Optimal Oxygen Concentration during Induction of General Anesthesia

The study was approved by the Regional Ethics Committee (Uppsala, Sweden), and all patients gave their written informed consent to participate. Thirty-six healthy, nonsmoking females with American Society of Anesthesiologists physical status class I or II and with a body mass index (weight in kilograms divided by the square of the height in meters) of 31 or less, scheduled for elective hysterectomy due to benign disease, were enrolled in the study. Preoperative evaluation was carried out the day before surgery. All patients had normal electrocardiograms. Their arterial oxygen and carbon dioxide tensions were normal (ABL 505; Radiometer, Copenhagen, Denmark), and their spirometry and lung volumes (by helium dilution technique), measured in the supine position, were also normal. Patients’ oxygen uptake and carbon dioxide output at rest were measured, and major differences in metabolic rate between groups were excluded (integrative cardiopulmonary testing; Jaeger, Wurzburg, Germany). Patients were randomized to one of three oxygenation groups: 100, 80, or 60% oxygen (in nitrogen).

Premedication consisted of acetaminophen 1 g orally, and five patients were also given 10 mg diazepam orally. Anesthesia was induced and measurements were performed at the Radiology Department before transport to surgery. The patients were hydrated with 200 ml Ringer's acetate and 300 ml of the colloid dextran before induction to reduce the risk of postinduction hypotension. The electrocardiogram and peripheral oxygen saturation (Spo²) were continuously assessed during the procedure. End-tidal carbon dioxide (Fetco²) and end-tidal oxygen concentrations (Feto²) were continuously monitored. Blood pressure was measured intermittently. The monitor for all measurements was a CS 3 (Datex-Ohmeda, Helsinki, Finland).

Data collection started when the patients began to breathe the randomly chosen oxygen concentration through a tightly fitted mask. After breathing spontaneously for 1 min, the patients were given bolus doses of fentanyl, 0.2 mg, and alfentanil, 1 mg. A target-controlled propofol infusion (Diprifusor; AstraZeneca, Macclesfield, Cheshire, United Kingdom) was started 2 min after the beginning of preoxygenation, aiming for an initial target blood concentration of 8 μg/ml, which was reduced to 6 μg/ml after tracheal intubation. Rocuronium, 30 mg, was given approximately 30 s after starting the propofol infusion to facilitate tracheal intubation. All patients lost consciousness and stopped breathing 30–60 s after the start of the propofol infusion.

The experimental procedure during anesthesia in the different groups is outlined in figure 1.

The strict preoxygenation period with normal spontaneous ventilation was 1 min. This period was followed by an induction phase with decreased spontaneous ventilation. An oropharyngeal airway was used to facilitate positive-pressure mask ventilation when spontaneous breathing ceased. The oxygenation continued until the patient was fully asleep and a steady state in end-tidal oxygen concentration was achieved. The end-tidal values of oxygen and carbon dioxide at steady state were noted before ventilation was stopped. The mask was removed, and the patient's trachea was immediately intubated. Five and a half minutes elapsed from the start of preoxygenation to the start of the intubation maneuver. No fresh gas was connected to the endotracheal tube during the period of apnea. The apnea time was defined as the period after stopping the ventilation until the saturation, as measured by pulse oximetry, had fallen to 90%. 5The oxygen saturation was continuously registered. The fresh gas was connected to the patient, and manual ventilation with 40% oxygen in nitrogen was started in all three groups at the time when Spo²had decreased to 90%. The lowest oxygen saturation and the time when it occurred were registered, as was the time to return to a saturation of 96%. The patients breathed and were ventilated through the Mentell system (Anmedic, Vallentuna, Sweden), a hybrid low-flow anesthetic circuit. 6This circuit has a built-in positive end-expiratory pressure (PEEP) of 3 cm H²O. No other level of PEEP was used. A fresh gas flow of 6 l/min was used to prevent rebreathing during the period of preoxygenation and during the period of anesthesia induction. As spontaneous breathing ceased during induction, manual ventilation on mask began with the aim to keep end-tidal Pco²normal and fresh gas flow unaltered. This was achieved with tidal volumes of 7–10 ml/kg body weight and respiratory rates less than 10 breaths/min. After the period of apnea, the patient was connected to the respirator (Anmedic), and a standard ventilatory setting was used, consisting of a tidal volume of 10 ml/kg body weight and a respiratory rate of 12 breaths/min. A fresh gas flow of 6 l/min with 40% oxygen in nitrogen was used when ventilation was started after the apnea period. The fresh gas flow was lowered to approximately 1 l/min, still with 40% oxygen, to avoid hyperventilation, as soon as Spo²was raised to 96% and end-tidal carbon dioxide was normalized. Airway pressure did not exceed 25 cm H²O in any patient.

The lungs were studied with computed tomography (CT; Siemens Somatom HiQ; Siemens, Erlangen, Germany). The patients were placed in the supine position, with the arms parallel to the body, horizontally on the CT table. All CT scans were carried out in apnea at the end of expiration. Three scans were made: a basal scan 10 mm above the dome of the right diaphragm, a scan at the hilus, and one at the apex of the lungs. A control CT was taken in the conscious patient to exclude any atelectasis before anesthesia. During anesthesia, the three scans were performed as soon as ventilation was started after apnea. In the 100% group, this occurred approximately 14 min from the start of preoxygenation. As the three groups were apneic for different time periods until the saturation had fallen to 90%, depending on their initial inspiratory oxygen concentration, the 60 and 80% groups were split into two subgroups each. One subgroup (first six patients) had their three scans performed immediately after controlled ventilation was started, as was the situation in the 100% group. The other subgroup (last six patients) had only a basal scan immediately after apnea, and then three scans were undertaken approximately 14 min after start of preoxygenation. Thereby, scans were taken at the same time point in the 60 and 80% subgroups as in the 100% group. Performing the scanning procedure including positioning took 1.5 min.

The radiologist was unaware of the group affiliation of the individual patient when making the examination of the CT scans. Calculations were made using the region of interest presented in the Somatom computer program. For tracing the atelectasis, the window level and width were set at −500 and ± 1,500 Hounsfield units, respectively. Atelectasis was defined as areas with attenuation values between −100 and +100 Hounsfield units and was expressed in square centimeters and in percent of the total lung area. 7The dorsal border was traced manually between the atelectasis and the pleura. The ventral border was drawn a few centimeters above the dense area, and the atelectasis was calculated by the region of interest program. The total lung area was calculated as the inner thorax area, excluding the heart and great vessels. Each region of interest was redrawn two times, and the mean value was used.

The sample size for this study was based on the knowledge from earlier studies 8,9and the assumption that a reduction of atelectasis of 50% or more would be of clinical importance. We did not assume the main variables in the study, i.e. , the amount of atelectasis and the apnea time, to have a normal distribution. Therefore, and since three groups were involved, the Kruskal-Wallis nonparametric one-way analysis of variance was applied, and a P  value of 0.01 or less was considered significant.

The three study groups were similar in demographic and physiologic data collected the day before the investigation (table 1).

No patient presented with atelectasis in the awake state before anesthesia induction. During anesthesia, immediately after the period of apnea, the mean atelectasis area in the basal scan was close to 10 cm²(5.6%) in the 100% group compared to 1.3 cm²(0.6%) and 0.3 cm²(0.2%) in the 80 and 60% groups, respectively(P < 0.001;table 2and fig. 2). The atelectatic area in the hilus scans was much smaller, although the relation between the three groups was the same as at the basal cut. The apex scans indicated less than 0.1 cm²atelectasis in any group.

The time from start of preoxygenation until apnea commenced, corresponding to the time of oxygenation with either 100, 80, or 60% oxygen, was approximately 5.5 min, with no difference between the groups (table 2).

All patients had an Spo²of 99% at the beginning of the apnea period. The mean time for Spo²to decrease to 90% during the apnea was close to 7 min in the 100% group, 5 min in the 80% group, and 3.5 min in the 60% group (P < 0.01;table 2and fig. 3). The pattern in which Spo²decreased during the period of apnea is shown in figure 4. Immediately after apnea, ventilation was commenced, and then CT scanning was performed and completed in 1.5 min. Thus, the total time from the initiation of preoxygenation to the completion of CT scanning was 14 min in the 100% oxygen group and 12 and 10.5 min in the 80 and 60% oxygen groups, respectively.

There was no statistically significant change in the atelectasis formation in the 80 and 60% oxygen groups immediately after apnea compared with when CT scans were repeated at 14 min (corresponding to the time of CT in the 100% oxygen group;table 2).

The lowest saturation, the time at which Spo²started to increase, the time to recover to normal saturation, and the difference between end-tidal Pco²before and immediately after apnea are summarized in table 3.

A major finding in the current study was that breathing 100% oxygen during preoxygenation and anesthesia induction caused significantly more atelectasis formation than did breathing 80 or 60% oxygen. This atelectasis causes intrapulmonary shunting, 10and the clinical impression is that it contributes to postoperative pulmonary complications. 11 

The anesthetic procedure was uneventful in all but one patient. This patient showed signs of heavy mucus production during the operation, and suction of the endotracheal tube was carried out several times. It must be observed that this patient had more atelectasis (19.3 cm²) at the basal scan directly after apnea than did any other patient. However, even after the exclusion of this patient, the atelectatic area (now 9.0 ± 4.5 cm²) remained significantly larger than the areas in the 80 or 60% oxygen groups (P < 0.01). This patient also had the shortest apnea tolerance (239 s) in comparison with the group average of 411 s. Excluding this patient from the results in the 100% group increased the mean apnea tolerance to 428 s in the 100% group. We chose not to exclude her, as she most certainly demonstrates a major mechanism in the development of a larger than normal amount of atelectasis during a routine induction of anesthesia.

An expected finding was that the safe time of apnea, defined as keeping peripheral oxygen saturation more than 90%, was significantly longer when breathing 100% oxygen than 80 or 60% oxygen. In a patient with signs of a difficult airway or with any other reason to have maximal safety during induction, 100% oxygen is always preferable in order to have an extra minute or two for securing the airway. A “cannot ventilate, cannot intubate” situation may unpredictably occur during induction of anesthesia and supports the use of 100% oxygen at all times during preoxygenation. However, if morbidity or mortality from atelectasis formation can be proven, this has to be considered against the probability of an unpredictable “cannot ventilate, cannot intubate” situation. The risk–benefit ratio for the 60% group versus  the 80% group seemed unfavorable. There was a small and probably clinically unimportant difference in atelectasis formation between the groups, while there was 1.5 min less time to decrease to 90% saturation.

Did the prolonged apnea per se  increase the size of the atelectasis? The design of the study excluded the possibility of addressing this question, as we judged the information regarding apnea tolerance in the 80 and 60% oxygen groups to be of paramount interest. This limitation warrants a study without prolonged apnea to be properly investigated. For the time being, we can only compare with earlier studies of atelectasis formation during anesthesia with a short apnea for the intubation of the airway, in which the amount of atelectasis has been as large as in the present patients on 100% oxygen. 4,12,13 

Another key question was whether the prolonged apnea increased the size of the atelectasis more so in the 100% oxygen group than in the others, since the apnea period was longer in this group. We tested the impact of this time difference by performing CT scans in a subgroup of patients breathing 80 or 60% oxygen at the same time point as we did in the 100% oxygen group. There was no difference in atelectasis in the 60 or 80% groups when comparing values directly after apnea (10.5 or 12 min after start of preoxygenation) with values at the time point when the 100% oxygen group was studied (14 min). It must be observed, however, that the patients in the 60 and 80% groups were ventilated with tidal volumes of 10 ml/kg some minutes longer before the new CT scans at 14 min, in contrast to the patients in the 100% group, who were ventilated only during the 1.5 min it took to complete the scans. It may be argued that this longer period of positive-pressure ventilation could have prevented or reversed atelectasis formation in these subgroups. However, it has been shown that reopening of collapsed lung tissue requires inflation of the lungs up to 40 cm H²O, corresponding to a vital capacity maneuver. 14It can also be proposed that complete airway closure with gas trapping and gas absorption is one likely mechanism of atelectasis during anesthesia, as hypothesized in clinical studies 15and shown in mathematical modeling. 16Whether the lungs are ventilated need not affect gas absorption, provided that airways remain closed over the respiratory cycle.

Application of PEEP can open up closed airways and may reduce the formation of atelectasis. 10The level of PEEP in this study was only 3 cm H²O. It might have reduced atelectasis formation, but equally so in the three groups since all groups had the same PEEP. PEEP is effective only as long as it is applied, and during the period of apnea, there was apparently no PEEP.

The effect on postoperative atelectasis of using 80% oxygen during anesthesia was recently investigated by Akça et al.  17In their study, 80% oxygen was compared with 30% given during colon resection and for 2 h after the procedure. No difference in incidence and severity of atelectasis was seen the first postoperative day. In the light of our results, the amount of atelectasis in the study by Akça et al.  17may have originated from the use of 100% oxygen during induction of anesthesia. The use of 80% oxygen might be of borderline importance to produce atelectasis. If so, no difference might be expected between the 30 and 80% groups.

Although there were no statistical differences in the current study among patients regarding physiologic data, some details deserve a comment. In the 80% group, the mean hemoglobin value was 8–11 g higher per liter than in the other two groups. This might have enhanced the apnea tolerance in the 80% group. On the other hand, patients in the 100% group were 4 yr younger than in the 80% group, and the body mass index was slightly higher in the 80% group, as was the preoperative oxygen consumption awake at rest. These three factors may decrease apnea tolerance. However, this study was too small to discriminate between the significance of factors other than oxygen level in influencing the formation of atelectasis during anesthesia induction. Clearly, those other factors vary in impact in different patients, as illustrated by the large variation in atelectasis formation in the 100% group.

We may conclude that preoxygenation with 80% oxygen seemed to be beneficial in reducing atelectasis formation during anesthesia induction, compared with 100% oxygen, but with a reduction in apnea tolerance. This reduction in apnea tolerance can make a difference in unpredictable, complicated situations during anesthesia induction. Until a large clinical trial can prove significant morbidity from atelectasis during or after anesthesia, the standard of using 100% oxygen for preoxygenation should continue.

The authors thank registered nurses Malin Engdahl, Majny Thorell, and Siv Johansson (Department of Anesthesiology and Intensive Care), Lotta Hjelm and Kerstin Thuveson (Department of Radiology), Lena Lundgren (Department of Obstetrics and Gynecology), and Monica Quarfordt and Annika Öhrn (Department of Clinical Physiology, all at Central Hospital, Västerås, Sweden) for skillful technical assistance. The authors also thank Hans U. Rothen, M.D., Ph.D. (Department of Anaesthesiology and Intensive Care, University Hospital, Bern, Switzerland), for methodological advice and Ivar Ringqvist, Ph.D. (Professor, Centre for Clinical Research, Central Hospital, Västerås, Sweden), for valuable advice.