Secondary Logo

Journal Logo

Neurosurgical Anesthesia

The Effect of Hyperventilation and Hyperoxia on Cerebral Venous Oxygen Saturation in Patients with Traumatic Brain Injury

Thiagarajan, Arimavalavan MD; Goverdhan, Puri D. MD, PhD; Chari, Promila MD; Somasunderam, Kanan MD

Author Information
doi: 10.1213/00000539-199810000-00019
  • Free


Patients with severe head injury may develop increased intracranial pressure (ICP), which can adversely affect cerebral perfusion pressure [1]. Hyperventilation is one treatment for reducing ICP in patients admitted to the intensive care unit. Acute hyperventilation induces cerebral vasoconstriction. A reduction in cerebral blood volume and the resultant decrease in ICP may improve cerebral perfusion pressure [2]. Cerebral vasoconstriction secondary to hyperventilation has also been known to cause cerebral ischemia.1 Brain function is exquisitely dependent on adequate oxygen delivery. If oxygen delivery to the brain can be increased, the adverse effect of hyperventilation may be minimized. However, hyperoxia has also been shown to cause cerebral vasoconstriction, which may lead to cerebral ischemia [4].

(1) Reivich MP, Cohen PJ, Greenbaum L. Alterations in the electro-encephalogram of awake man produced by hyperventilation: effects of 100% oxygen at 3 atmospheres (absolute) pressure [abstract]. Neurology 1966;16:304.

The measurements of jugular venous bulb oxygen saturation (SjvO2) and arteriovenous oxygen content difference (AVDO2) can provide indirect evidence of adequacy of cerebral blood flow (CBF) and global oxygenation [5,6]. The effect of hyperoxia during hyperventilation on SjvO2 has been studied in patients undergoing anesthesia for intracranial aneurysm and tumor surgery [7]. Although changes in SjvO2 in head-injured patients have been assessed as a function of PaCO2, the effect of hyperoxia on SjvO2 during hyper-ventilation has not been studied. This study was designed to evaluate the interaction of hyperoxia with hypocarbia on SjvO2 and AVDO2 in head-injured patients.


After obtaining ethics committee approval and informed consent from the patients' next of kin, 18 consecutive head-injured patients (aged 15-60 yr) admitted for elective hyperventilation in the intensive care unit were included in the study. We excluded patients with Glasgow Coma Scale (GCS) scores >or=to9, preexisting lung pathology, hemodynamic instability, hemoglobin (Hb) levels <10 g/dL, fracture of the base of the skull, skin infection in the neck, associated intracranial hematomas (confirmed by computerized tomography), and body temperature exceeding a range of 36.5-37.5[degree sign]C during the period of study.

The study was conducted 24 h after admission to the intensive care unit and at least 4 h after patients received their last dose of mannitol or furosemide. Morphine 0.025 mg [center dot] kg-1 [center dot] h-1 IV and diazepam 0.025 mg [center dot] kg-1 [center dot] h-1 IV were used for sedation. Pancuronium 1 mg/h IV was used for neuromuscular paralysis. Monitoring included electrocardiogram, pulse oximetry, arterial blood pressure with a radial artery catheter, central venous pressure, nasopharyngeal temperature, PETCO2, and airway pressure. The lungs were ventilated by using a Puritan Bennett 7200 AE respirator (Mervue, Galway, Ireland) to a PaCO2 of 30 mm Hg. The fraction of inspired oxygen was adjusted to PaO (2) 100-150 mm Hg. Using an aseptic technique, the jugular bulb was cannulated on the right side in all the patients using a 16-gauge, 20-cm catheter (Secalon[registered sign] Seldy; Ohmeda, Swindon, UK) without any complications. The tip of the catheter was positioned just medial to the mastoid process and confirmed radiographically. Catheter patency was maintained by a slow infusion of 0.9% NaCl.

Arterial and jugular venous bulb blood gases were measured using the Ciba-Corning 288 blood gas analyzing system (Ciba-Corning Diagnostics Corp., Medfield, MA). The jugular venous blood sample was drawn at a rate of <2 mL/min. The first samples were drawn at a PaO2 of 100-150 mm Hg and PaCO2 of 30 mm Hg. Minute ventilation was then increased by altering the respiratory rate to achieve a PaCO (2) of 25 mm Hg without altering the peak airway pressure. After another 30-min period of stabilization, arterial and jugular venous blood gases were measured. Minute ventilation was then decreased to restore PaCO2 to 30 mm Hg. After a 30-min period of stabilization, the arterial and jugular venous blood gases were again measured. The same procedure was repeated after increasing the PaO2 to 200-250 mm Hg.

AVDO2 (vol%) was calculated from the formula: Equation 1 where PjvO2 is the partial pressure of oxygen in jugular blood (mm Hg).

The SjvO2 and AVDO2 values at different PaCO2 and PaO (2) values were compared using repeated-measures analysis of variance. A P value of < 0.05 was considered significant. Values are given as mean +/- SD.


Demographic data for the patients are summarized in Table 1. SjvO2 decreased from 66% +/- 4% to 56% +/- 3% when PaCO2 was decreased from 30 to 25 mm Hg at a PaO2 of 100-150 mm Hg and recovered to baseline values (66% +/- 2%) when the PaCO2 was restored to 30 mm Hg (P < 0.001) (Figure 1). A similar decrease, from 77% +/- 4% to 64% +/- 3%, was seen in SjvO2 at a PaO2 of 200-250 mm Hg when the PaCO2 was decreased from 30 to 25 mm Hg (P < 0.001) (Figure 1). However, the absolute value of SjvO2 was greater at a PaO2 of 200-250 mm Hg for any given PaCO2 (P < 0.001) (Figure 1). The lowest individual value of SjvO2 was 51% at a PaO2 of 100-150 mm Hg and a PaCO2 of 25 mm Hg. Hyperoxygenation improved SjvO2 at both PaCO2 values, but the highest values were seen at a PaO2 of 200-250 mm Hg and a PaCO2 of 30 mm Hg.

Table 1
Table 1:
Patient Data
Figure 1
Figure 1:
The effect of hyperoxia on percent oxygen saturation of jugular venous blood (SjvO2) at two levels of PaCO2. *P < 0.001 for SjvO2 at PaCO (2) 25-30 mm Hg at each PaO2. [dagger]P < 0.001 for SjvO2 between PaO (2) at each PaCO2 level.

AVDO2 increased significantly by 1.7 +/- 0.4 mL/100 mL of blood when the PaCO2 was decreased from 30 to 25 mm Hg at a PaO2 of 100-150 mm Hg and recovered to baseline values when the PaCO2 was restored to 30 mm Hg (P < 0.001). A similar significant increase in AVDO2 was observed within a PaO (2) range of 200-250 mm Hg (P < 0.001). At each PaCO2 value, the AVDO2 value was less than that observed at a PaO2 of 100-150 mm Hg (P < 0.001) (Table 2).

Table 2
Table 2:
Effect of Hyperoxia on AVDO2


Using controlled hyperventilation to reduce ICP in head-injured patients has been associated with cerebral oligemia [8], which has been implicated in worsened neurological outcomes [9,10]. Studies in animals and humans have shown that cerebral oxygenation can probably be maintained by hyperoxygenation of the arterial blood when hyperventilation is used [11]. However, a number of studies have reported a reduction in CBF with hyperoxia during normocapnia that might offset the potential benefit of hyperoxia [12-14]. Matta et al. [7] assessed the effect of hyperoxygenation on SjvO2 intraoperatively and concluded that hyperoxia during hyperventilation improves oxygen delivery to the brain in anesthetized neurosurgical patients. We found similar results in the present study in patients with head injury.

The intracranial events in head-injured patients can be different from those that occur with intracranial tumors or aneurysms. Cerebral vasoreactivity may be altered in vessels supplying the tumor or those surrounding the aneurysm, especially after subarachnoid hemorrhage [15]. Matta et al. [7] analyzed cerebral venous blood of the nonoperative side, which may not accurately reflect events on the side of surgery. We chose patients with only diffuse brain edema (diagnosed by computerized tomography) because the changes are more global. Measurements on one side are more likely to reflect global events. Because SjvO2 is markedly altered when an intracranial hematoma is evacuated [16], we excluded patients with intracranial hematomas. Most of the patients in the present study were young; thus, the cerebral vasculature was not likely affected by age-related changes.

We chose two levels of PaCO2, 30 mm Hg and 25 mm Hg, to assess whether hyperoxygenation improves SjvO2 during moderate and marked hyperventilation and allowed a stabilization period of 30 min for a given level of PaCO2 and PaO2. The lowest SjvO2 values in our study were 60% at a PaCO2 of 30 mm Hg and 51% at a PaCO2 of 25 mm Hg when the PaO2 was within 100-150 mm Hg and 67% and 59% with PaO2 at 200-250 mm Hg, respectively. These values are considerably higher than those obtained by Matta et al. [7], who noted relatively lower saturation values even at higher PaO2. Higher SjvO2 values may also be the result of a hyperemic response in head-injured patients, especially seen in the first 48 h after injury [17]. The lower SjvO2 value at lower PaCO2 is consistent with the decrease in CBF, assuming that cerebral metabolic rate remains constant. The higher AVDO2 value observed during marked hyperventilation probably reflects a decreased margin of safety or limited oxygen supply with impending tissue hypoxia. The decrease in the AVDO2 observed at a higher PaO2 suggests benefit from this therapy.

In conclusion, hyperoxia during acute hyperventilation in patients with head injury improves oxygen delivery to the brain, as indicated by improved SjvO (2) and AVDO2 values. Moderate hyperventilation to a PaCO2 of 30 mm Hg in combination with higher PaO2 may be beneficial in patients with head injury.


1. Mori K, Nakos S. Brain death. In: Miller RD, ed. Anesthesia. 4th ed. New York: Churchill Livingstone, 1994:2565.
2. Albert JW, John LB. The cerebral vascular response to reduction in arterial carbon dioxide tension. J Clin Invest 1961;40:1297-309.
3. Deleted in proof.
    4. Shinij N, John SM, Amano T, et al. Cerebral vasomotor responsiveness during 100% oxygen inhalation in cerebral ischemia. Arch Neurol 1983;40:271-6.
    5. Matta BF, Lam AM, Mayberg TS, et al. A critique of the intra-operative use of jugular venous bulb catheters during neurosurgical procedures. Anesth Analg 1994;79:745-50.
    6. Robertson CS, Narayan EK, Gokaslan ZL, et al. Cerebral arteriovenous oxygen difference as an estimate of cerebral blood flow in comatose patients. J Neurosurg 1989;70:222-30.
    7. Matta BF, Lam AM, Mayberg TS. The influence of arterial oxygenation on cerebral oxygen saturation during hyperventilation. Can J Anaesth 1994;41:1041-6.
    8. Cold GE. Does acute hyperventilation provoke cerebral oligemia in comatose patient with acute head injury? Acta Neurochir (Wein) 1989;96:100-6.
    9. Muizelaar JP, Marmarou A, Ward JD, et al. Adverse effects of prolonged hyperventilation in patients with severe head injury: a randomised clinical trial. J Neurosurg 1991;75:731-9.
    10. Elias-Jones AC, Punt JAG, Turnbul AE, Jaspan T. Management and outcome in the Trent region 1985-90. Arch Dis Child 1992;67:1430-5.
    11. Kennealy JA, McLennan JE, Loudon RG, McLaurin RL. Hyperventilation-induced cerebral hypoxia. Am Rev Respir Dis 1980;122:407-11.
    12. Busija DW, Orr JA, Rankin JHG, et al. Cerebral blood flow during normocapnic hyperoxia in the unanesthetized pony. J Appl Physiol 1980;48:10-5.
    13. Leahy FAN, Cates D, MacCullum M, Rigatto H. Effects of CO2 and 100% O2 on cerebral blood flow in preterm infants. J Appl Physiol 1980;48:468-72.
    14. Sheinberg M, Kanter MJ, Robertson CS, et al. Continuous monitoring of jugular venous oxygen saturation in head-injured patients. J Neurosurg 1992;76:212-7.
    15. Voldby B, Enevoldsen EM, Jenses FT. Cerebrovascular reactivity in patients with ruptured intracranial aneurysm. J Neurosurg 1989;62:59-65.
    16. Gopinath SP, Cormio M, Ziegler J, et al. Intraoperative jugular desaturation during surgery for traumatic intracranial hematomas. Anesth Analg 1996;83:1014-21.
    17. Obrist WD, Langfitt TW, Jaggi JL, et al. Cerebral blood flow and metabolism in comatose patients with acute head injury: relationship to intracranial hypertension. J Neurosurg 1984;61:241-53.
    © 1998 International Anesthesia Research Society