Anesthesia & Analgesia

Skip Navigation LinksHome > November 2001 - Volume 93 - Issue 5 > Voluntary Hyperventilation Before a Rapid-Sequence Induction...
Anesthesia & Analgesia:
doi: 10.1097/00000539-200111000-00050
ANESTHETIC PHARMACOLOGY: (International Society for Anaesthetic Pharmacology): Research Report

Voluntary Hyperventilation Before a Rapid-Sequence Induction of Anesthesia Does Not Decrease Postintubation Paco2

Choinière, André MD; Girard, François MD, FRCPC; Boudreault, Daniel MD, FRCPC; Ruel, Monique RN; Girard, Dominique C. MD, FRCPC

Free Access
Article Outline
Collapse Box

Author Information

Department of Anesthesiology, CHUM, Hôpital Notre-Dame, Montreal, Canada

The i-STAT system was loaned by the diagnostic division of Abbott Laboratories for the duration of the study.

June 28, 2001.

Address correspondence and reprint requests to François Girard, MD, FRCPC, Department of Anesthesiology, CHUM, Hôpital Notre-Dame, 1560 Sherbrooke East, Montreal, Canada, H2L 4M1.

Collapse Box


To prevent hypercapnia, voluntary hyperventilation is recommended for patients with increased intracranial pressure before the induction of general anesthesia. We sought to determine whether this maneuver results in a lower Paco2 than breathing 3 min of oxygen 100% by face mask (preoxygenation) after intubation. Thirty patients requiring general anesthesia were randomly assigned to breathe either 3 min of oxygen 100% by face mask (Group P) or 1 min of oxygen 100% followed by 2 min of voluntary hyperventilation with oxygen 100% (Group H). All patients received a standard rapid-sequence induction of anesthesia followed by a 90-s period of apnea. Patients were then tracheally intubated and mechanically ventilated. Five arterial blood gas samples were taken: with room air, after preoxygen- ation or hyperventilation, after 60 and 90 s of apnea, and after tracheal intubation. Voluntary hyperventilation decreased Paco2 before rapid-sequence induction (hyperventilation, 30.0 ± 3.5 mm Hg versus preoxygenation, 37.9 ± 5.2 mm Hg;P < 0.0001), but after 60 s of apnea, both groups had similar Paco2 (hyperventilation, 36.1 ± 3.3 mm Hg versus preoxygenation, 35.6 ± 3.4 mm Hg;P = 0.673), and no benefit was found after intubation (hyperventilation, 40.5 ± 3.9 mm Hg versus preoxygenation, 41.4 ± 2.7 mm Hg;P = 0.603). We conclude that voluntary hyperventilation before rapid-sequence induction does not provide protection against potential hypercapnia during intubation.

Paco2 is an important factor in the regulation of cerebral circulation. Patients with increased intracranial pressure (ICP) frequently present for emergency surgery with a full stomach or with nausea and vomiting. Rapid-sequence induction of anesthesia is therefore often required, necessitating a brief period of apnea, which could result in a deleterious increase in Paco2. In an attempt to prevent hypercapnia, these patients are asked to voluntarily hyperventilate before general anesthesia induction (1–3).

Only one study, in anesthetized dogs, has suggested that hyperventilation could prevent increased ICP during apnea. In this study, Williams et al. (4) showed that after 2 min of apnea preceded by normocapnia, there was a 204% increase in ICP, whereas apnea after hypocapnia resulted in a 68% increase in ICP; this suggests that hyperventilation reduces the increase of ICP during a period of apnea. No study has demonstrated similar results in humans, nor has any study demonstrated the effectiveness of voluntary hyperventilation in preventing Paco2 increase after rapid-sequence induction.

The goals of this prospective, randomized study were to determine whether voluntary hyperventilation before rapid-sequence induction could significantly reduce Paco2 as compared with breathing 3 min of oxygen 100% by face mask (preoxygenation) and whether this difference in Paco2 would still be preserved after endotracheal intubation.

Back to Top | Article Outline


After IRB approval, written, informed consent was obtained from 30 ASA physical status I–III patients ranging in age from 30 to 70 yr. These patients were scheduled for elective surgery under general endotracheal anesthesia. A radial arterial catheter was used during surgery either because of the patient’s underlying medical condition or the nature of the surgical procedure. Patients with severe cardiovascular disease, chronic obstructive pulmonary disease, increased ICP, obesity (body mass index >30), gastroesophageal reflux, altered mental state, or difficult airway and those who smoked more than 25 cigarettes a day were excluded from the study.

Patients were randomly assigned to either the Preoxygenation group (Group P) or the Hyperventilation group (Group H). Patients were not premedicated. Standard monitoring was used. A 20-gauge arterial catheter was then inserted into the radial artery under local anesthesia. A first blood sample was drawn with a fraction of inspired oxygen of 0.21 (room air) for arterial blood gas (ABG) analysis and hemoglobin (Hb) and hematocrit determination. Patients in Group P (n = 15) were asked to breathe normally, for 3 min, 100% oxygen at 10 L/min through an anesthesia mask attached to a circle system with soda lime and the adjustable pressure-limiting valve fully open. Patients in Group H (n = 15) were asked to breathe normally for 1 min and were then coached into hyperventilating for 2 min, by using the same circuit as Group P. Hyperventilation consisted of vital capacity breathing at a rate of 15 breaths/min.

An ABG was obtained after the preoxygenation/hyperventilation period (post-O2). All patients then received a rapid-sequence induction regimen consisting of fentanyl 3 μg/kg, thiopental 5–6 mg/kg, and rocuronium 0.9 mg/kg IV. After the administration of rocuronium (defined as T0), patients were left apneic for 90 s with the face mask still applied with 100% oxygen at 10 L/min and the adjustable pressure-limiting valve fully open. Patients were then tracheally intubated and mechanically ventilated, ending the study period. ABGs were collected after 60 s and 90 s of apnea and after intubation. The time from rocuronium injection to intubation was recorded.

All ABG analysis was performed with the i-STAT system (i-STAT Corporation, Princeton, NJ) with the G3+ cartridge. By using this analyzer and cartridge, ABG analysis could be performed in approximately 2 min by using only three to four drops of whole blood. The i-STAT analyzer is very reliable and accurate (5,6), with results comparable to those performed by conventional laboratory blood gas analyzers (Paco2 mean difference ± sd, 0.97 ± 1.27 mm Hg) (6).

Statistical analysis was performed with the SAS statistical package, release 6.12 (1989–1996; SAS institute Inc., Cary, NC). Results are expressed as mean ± sd, except where otherwise noted. Differences in demographic and intraoperative data between the two groups were sought by using the χ2 test and Student’s t-test for nonparametric and parametric variables, respectively. Two-way analysis of variance for repeated measures was also used to characterize the evolution in time of Paco2 during the study period. A P < 0.05 was considered significant. Assuming a β of 0.2 and an α of 0.05, with an estimated sd of 3 mm Hg, 15 patients in each group were needed to detect a 3 mm Hg difference between the two groups.

Back to Top | Article Outline


There were no significant differences between the two groups when comparing age, sex, body mass index, ASA status, smoking habits, baseline ABG analysis, Spo2, and Hb value (Table 1). ABG analysis is presented in Table 2. There were no significant differences in Pao2 between the two groups at any time. When breathing room air, both groups had similar Paco2 (P = 0.115). Voluntary hyperventilation (Group H) was effective at decreasing Paco2 before rapid-sequence induction (P < 0.0001), but after 60 and 90 s of apnea and after intubation, both groups had similar Paco2 (P = 0.673, P = 0.437, and P = 0.603, respectively). After tracheal intubation, patients in Group H returned to a Paco2 similar to their baseline value (P = 0.547), whereas patients in Group P had a slightly higher Paco2 than their own baseline (3.33 mm Hg increase, P = 0.0002).

Table 1
Table 1
Image Tools
Table 2
Table 2
Image Tools

After preoxygenation/hyperventilation, all patients had an Spo2 of 100%. After 60 s of apnea, all patients were still 100% saturated. By 90 s of apnea, one patient in Group P and two in Group H had an Spo2 of <99%. At the time of intubation, one patient in Group P and two in Group H had an Spo2 of <98%. The time to tracheal intubation was similar in both groups (Group H, 115 ± 9 s versus P, 117 ± 11 s). No adverse event was noted during the study period.

Back to Top | Article Outline


This study is the first to investigate the effects of voluntary hyperventilation versus preoxygenation with normal ventilation on Paco2 during rapid-sequence induction. We assumed that Paco2 would start to increase as soon as respiration stopped and that a brief period of voluntary hyperventilation, caused by decreasing the initial Paco2, would result in a decreased Paco2 after intubation (1–3). This conception is based on studies that examined the rate of increase of Paco2 in humans under stable anesthesia (7,8). Our study clearly establishes that this is not the case during rapid-sequence induction of anesthesia.

In a series of five healthy adults scheduled for various surgeries, Eger and Severinghaus (7) observed that Paco2 increased by 13.4 mm Hg during the first minute of apnea and by 3.0 mm Hg/min thereafter. After one hour of hyperventilation, the rate of increase of Paco2 in the first minute of apnea was significantly smaller than before hyperventilation (9.6 vs 13.4 mm Hg, P < 0.05).

Gentz et al. (8) studied the rate of increase of Paco2 during apneic oxygenation after shorter periods of hypocapnia. Nineteen patients (32–75 yr old, ASA status I–III) scheduled for elective surgery under general endotracheal anesthesia were studied. Once anesthetized and mechanically ventilated, and before each apneic period, patients’ Paco2 was held constant for approximately 10 minutes between 35 and 40 mm Hg (normocapnic group) or 25 and 30 mm Hg (hypocapnic group). Contrary to the results of Eger and Severinghaus (7), patients in the hypocapnic group had a larger rate of increase of Paco2 than the Normocapnic group in the first minute of apnea (9.4 vs 7.6 mm Hg, P < 0.009). The authors hypothesized that this finding was caused in part by widening of the arteriovenous gradient of Pco2 with a subsequent shift of CO2 stores.

Williams et al. (4) showed that hyperventilation could blunt the increase of ICP induced by apnea in anesthetized dogs. The authors studied the effect of two minutes of apnea on ICP, compliance, and cerebral blood volume in 19 adult dogs during normocapnia (Paco2 from 35 to 40 mm Hg), hypocapnia (Paco2 from 24 to 28 mm Hg), and hypercapnia (Paco2 from 50 to 58 mm Hg). Their results suggest that hyperventilated dogs have a smaller ICP increase (68%) during apnea than normoventilated dogs (204%). However, this study had several methodologic limitations. Surprisingly, dogs in the hypocapnia group had a preapnea ICP significantly higher than the normocapnia group (P < 0.05). After two minutes of apnea, although ICP increased more in the normocapnia group than the hypocapnia group (204% vs 68%), both groups attained similar ICP. Finally, dogs were hyperventilated for an undetermined period of time (but for at least 30 minutes), making results difficult to compare with those of previous studies.

The measurement of the rate of increase of Paco2 was not a primary objective of this study. However, Table 2 clearly shows that we obtained an increase of approximately 10.5 mm Hg in Group H in the two-minute period between the post-O2 and postintubation ABG. The increase was three times slower in Group P, with only 3.5 mm Hg during the same two-minute period. This is at least in part in agreement with the results obtained by Gentz et al. (8), with short periods of hyperventilation.

Paco2 is determined by a balance between CO2 metabolic production, elimination, and physiologic stores within the body. CO2 elimination is dependent on alveolar ventilation, cardiac output, and Hb level. During apneic oxygenation, oxygen is drawn in by mass movement and replaces the oxygen that crosses the alveolar/capillary membrane, but no significant amount of CO2 is eliminated. Cardiac output varies during anesthesia induction. Thiopental, by decreasing peripheral resistances, decreases afterload, resulting in a brief increase in cardiac output. An acute change in Hb level, which is the primary buffer for respiratory gases in the body, could have an effect on subsequent CO2 loading. In our study, all patients received the same induction regimen; they were apneic and had similar Hb levels during the study period. Consequently, after the preoxygenation/hyperventilation period, CO2 elimination should have been minimal in both groups.

CO2 physiologic stores are very large, estimated at approximately 120 L, and they have been theoretically described in terms of a multicompartment model (9) : a rapid compartment represented by blood, brain, kidneys, and other well perfused tissues, a medium compartment representing skeletal muscles and other tissues with moderate blood flow, and a slow compartment including bone and fat. Each compartment has its own time constant, and the long time constants of the medium and slow compartment buffer change in the rapid compartment. By using this model, we can assume that long periods of hyperventilation would permit equilibration of the different compartments and result in a new Paco2 steady state. After such hyperventilation, the Paco2 would then return to normal at a slower pace during apnea, explaining the results obtained by Eger and Severinghaus (7). Shorter periods of hyperventilation, however, would decrease only the CO2 stores in the rapid compartment, resulting in a shift of CO2 from the medium and slow compartments during an apneic period. This would bring an additional source of CO2 to the central compartment and explain the greater increase of Paco2 observed in this study and that by Gentz et al. (8), when patients were hyperventilated for brief periods of time as compared with normoventilated patients.

The last determinant of Paco2 is CO2 metabolic production. Previous studies (7,8) have been performed on anesthetized patients with stable metabolic rates. In our study, however, the metabolic rate was certainly not constant. Patients in Group H were asked to voluntarily hyperventilate for two minutes before anesthesia induction. This maneuver requires a significant effort, possibly resulting in increased CO2 production. Although anesthesia induction with thiopental lowers metabolic rate, patients who hyperventilated still had a greater CO2 charge to eliminate, which, combined with a widening of the arteriovenous CO2 gradient, resulted in a more important Paco2 increase. In fact, after only 60 seconds of apnea, patients who hyperventilated already increased their Paco2 to the level of Group P. When compared with that of similar groups in previous studies (7,8), the increase of Paco2 in Group P (3.3 mm Hg over two minutes) was smaller and clinically nonsignificant. This may be attributed to the metabolic sparing effect of thiopental and the resulting decrease in CO2 production.

We chose to voluntarily hyperventilate patients for two minutes because it was previously determined that one minute of hyperventilation could significantly decrease Paco2(10), and more than two minutes would be very strenuous for patients. Two minutes of hyperventilation was in fact sufficient for patients in Group H to decrease their baseline Paco2 by an average of 10.1 mm Hg.

In summary, after only 60 seconds of apnea, Paco2 in hyperventilated patients had increased to levels similar to those after preoxygenation with normal ventilation. Because ICP variations induced by changes in Paco2 are rapid (11), occurring in a matter of seconds after changes in Paco2, the anticipated ICP changes would then parallel the Paco2 values. Therefore, no benefit could be expected from this maneuver. Furthermore, the increase in patient effort and the anxiety potentially associated with it are not justified by its effectiveness and may in fact result in an increased ICP (2). We conclude that voluntary hyperventilation in patients with increased ICP before rapid-sequence induction should no longer be advocated as a means of blunting an increase in Paco2 and its possible repercussion on ICP.

Back to Top | Article Outline


1. Spiekermann BF, Stone DJ, Bogdonoff DL, Yemen TA. Airway management in neuroanesthesia. Can J Anaesth 1996; 43: 820–34.

2. Black S, Cucchiara RF. Tumor surgery. In: Cucchiara RF, Black S, Michenfelder JD, eds. Clinical neuroanesthesia. 2nd ed. New York: Churchill Livingstone, 1998: 343–65.

3. Bendo AA, Kass IS, Hartung J, Cottrell JE. Anesthesia for neurosurgery. In: Barash PG, Cullen BF, Stoelting RK, eds. Clinical anesthesia. 3rd ed. Philadelphia: Lippincott-Raven, 1997: 699–745.

4. Williams G, Roberts PA, Smith S, et al. The effect of apnea on brain compliance and intracranial pressure. Neurosurgery 1991; 29: 242–6.

5. Bingham D, Kendall J, Clancy M. The portable laboratory: an evaluation of the accuracy and reproducibility of i-STAT. Ann Clin Biochem 1999; 36: 66–71.

6. Sediame S, Zerah-Lancner F, d’Ortho MP, et al. Accuracy of the i-STAT bedside blood gas analyser. Eur Respir J 1999; 14: 214–7.

7. EgerEL II, Severinghaus JW. The rate of rise of PaCO2 in the apneic anesthetized patient. Anesthesiology 1961; 22: 419–25.

8. Gentz BA, Shupak RC, Bhatt SB, Bay C. Carbon dioxide dynamics during apneic oxygenation: the effects of preceding hypocapnia. J Clin Anesth 1998; 10: 189–94.

9. Nunn JF. Applied respiratory physiology. 3rd ed. London: Butterworths, 1987.

10. Benumof JL. Preoxygenation: best method for both efficacy and efficiency? Anesthesiology 1999; 91: 603–5.

11. Sulek CA. Intracranial pressure. In: Cucchiara RF, Black S, Michenfelder JD, eds. Clinical neuroanesthesia. 2nd ed. New York: Churchill Livingstone, 1998: 73–123.

© 2001 International Anesthesia Research Society


Become a Society Member

Article Level Metrics