1. Standards. There are insufficient data to support a treatment standard for this topic.

2. Guidelines. There are insufficient data to support a treatment guideline for this topic.

3. Options. Mild or prophylactic hyperventilation (PaCO₂<35mmHg) in children should be avoided.

   Mild hyperventilation (PaCO₂ 30-35 mm Hg) may be considered for longer periods for intracranial hypertension refractory to sedation and analgesia, neuromuscular blockade, cerebrospinal fluid drainage, and hyperosmolar therapy.

   Aggressive hyperventilation (PaCO₂ <30 mm Hg) may be considered as a second tier option in the setting of refractory hypertension. Cerebral blood flow (CBF), jugular venous oxygen saturation, or brain tissue oxygen monitoring is suggested to help identify cerebral ischemia in this setting.

   Aggressive hyperventilation therapy titrated to clinical effect may be necessary for brief periods in cases of cerebral herniation or acute neurologic deterioration.

4. Indications from Adult Guidelines. The adult guidelines recommended1 at the level of a treatment standard that in the absence of increased intracranial pressure (ICP), chronic prolonged hyperventilation therapy (PaCO₂ of ≤25 mm Hg) should be avoided after severe TBI. At the level of a treatment guideline, it was recommended that prophylactic hyperventilation (PaCO₂ ≤35 mm Hg) therapy during the first 24 hrs after severe TBI should be avoided because it can compromise cerebral perfusion during a time when CBF is reduced.

   It was recommended as a treatment option that hyperventilation therapy may be necessary for brief periods when there is acute neurologic deterioration or for longer periods if there is intracranial hypertension refractory to sedation, paralysis, cerebrospinal fluid drainage, and osmotic diuretics. Jugular venous oxygen saturation, arterial jugular venous oxygen content differences, brain tissue oxygen monitoring, and CBF monitoring may help to identify cerebral ischemia if hyperventilation, resulting in PaCO₂ values <30 mm Hg, is necessary.

Aggressive hyperventilation therapy has been used in the management of severe pediatric TBI for the rapid reduction of ICP since the 1970s. In an uncontrolled study, Bruce et al.2 used a protocol that included aggressive hyperventilation and reported very good outcomes. This approach was based on the assumption that hyperemia was common after pediatric TBI. Hyperventilation therapy also was thought to benefit the injured brain through a variety of mechanisms including reduction of brain acidosis3, improvement of cerebral metabolism4, restoration of blood pressure autoregulation of cerebral blood flow5, and increasing perfusion to ischemic brain regions (local inverse steal)6.

More recent pediatric studies have shown that hyperemia is uncommon and also have raised concerns about the safety of hyperventilation therapy. Study of the effect of hyperventilation in children has focused on assessment of cerebral physiologic variables. The effect of hyperventilation therapy on outcome in infants and children with severe TBI has not been directly compared with other therapies such as hyperosmolar agents, barbiturates, hypothermia, or early decompressive craniectomy.

Hyperventilation reduces ICP by inducing hypocapnia. This leads to cerebral vasoconstriction and a reduction in CBF. This is accompanied by a reduction in cerebral blood volume, resulting in a decrease in ICP. However, hyperventilation is associated with a risk of iatrogenic ischemia. In an experimental model, Muizelaar et al.7 reported that the vasoconstrictor effect of hyperventilation was sustained for a period of <24 hrs. Chronic hyperventilation depletes brain tissue interstitial bicarbonate buffering and causes cerebral circulation to become hyper-responsive to subsequent increases in PaCO₂. In addition, the respiratory alkalosis that accompanies hyperventilation causes a left shift of the hemoglobin-oxygen dissociation curve, which may impair delivery of oxygen to tissue.

The assumption of benefit from hyperventilation recently has been challenged. Recent clinical studies in mixed adult and pediatric populations also have demonstrated that hyperventilation may decrease cerebral oxygenation and may induce brain ischemia8-11. After TBI, the CBF response to changes in PaCO₂ can be unpredictable and should be specifically monitored.

We searched Medline and Healthstar from 1966 to 2001 by using the search strategy for this question (see Appendix A) and supplemented the results with literature recommended by peers or identified from reference lists. Of 20 potentially relevant studies, two were used as evidence for this question (Table 1).

Diffuse cerebral swelling is a common finding in pediatric patients with severe TBI12, 13. Increased cerebral blood volume and CBF had been considered to be the unique cause of this diffuse swelling, and raised ICP in children and aggressive hyperventilation was advocated14. In the classic study by Bruce et al.2, 36 of 76 children with severe TBI were found to have diffuse cerebral swelling on CT scan. Six patients, ages 14-21 yrs, were found to have CBF that was normal or above normal. In three patients, CBF decreased back to control levels after diffuse swelling had resolved. Consequently, aggressive hyperventilation (PaCO₂ 23-25 mm Hg) was advocated and mannitol was discouraged.

There are now data to suggest that hyperemia is not as common as previously thought15. In a series of 80 normal, unanesthetized children, CBF ranged from 40 mL•100g-1•min-1 in the first 6 months of life to a peak of 108 mL•100g-1•min-1 at age 3-4 yrs, declining to 71 mL•100g-1•min-1 after age 9 yrs. Similarly, Chiron et al.16 demonstrated that CBF ranged from about 50 mL•100g-1•min-1 in normal neonates to a peak of 71 mL•100g-1•min-1 at 5 yrs. After age 19, CBF gradually decreased to adult levels. Thus, posttraumatic CBF may not be greater than normal in children. However, caution should be exercised in interpreting these studies because techniques used to measure CBF differed between reports.

Adelson et al.17 studied 30 children with severe TBI, all <8 yrs of age. Seventy-seven percent had CBF <20 mL•100g-1•min-1 on admission. Children were treated with a protocol including mild (PCO2 32-35 mm Hg) hyperventilation and barbiturate coma (60%). CBF was highest at 24-48 hrs (59.6 ± 4.5 mL•100g-1•min-1) and decreased (<50 mL•100g-1•min-1) after 3 days. Any child with global CBF of 20 mL•100g-1•min-1 or less at any time had a poor outcome. CBF of >55 mL•100g-1•min-1 was associated with a higher proportion of children with a good outcome.

Muizelaar et al.18 studied 32 children with severe TBI (age 3-18 yrs). The average CBF was only 44 ± 22 mL•100g-1•min-1, which is considerably lower than the average of 68 ± 4 mL•100g-1•min-1 found in four normal unanesthetized children. No correlation was found between CBF and ICP.

Although the effect of hyperventilation on long-term outcome has not been directly addressed in pediatric TBI, several reports have described the effects of hyperventilation on CBF and brain physiology. Stringer et al.19 studied local CBF and vascular reactivity before and after hyperventilation in 12 patients including three children with severe TBI. Hyperventilation-induced blood flow reductions affected both injured and apparently intact areas of the brain and were not reflected by ICP measurement.

Sharples et al.20 investigated CBF, arterial jugular venous oxygen difference, and cerebral metabolic rate in 21 children with TBI. No fundamental difference between adults and children in the pathophysiologic response of CBF to severe TBI was found. Absolute cerebral hyperemia was uncommon. Raised ICP was associated with low, rather than increased, CBF. Cerebral metabolic rate was initially normal in 81% of children with TBI. These data do not support the hypothesis that ICP increases as a result of excessive CBF in children with TBI. Based on this study, and on a subsequent study of cerebral vascular reactivity21, the authors recommended maintaining a normal PaCO₂.

Skippen et al.22 found that hyperemia was uncommon in children with severe TBI. However, CBF rates remained above the metabolic requirements of most children studied. A modest decrease in CBF and a much larger decrease in cerebral oxygen consumption were found at baseline. As PaCO₂ was reduced with hyperventilation, CBF was decreased in almost all patients despite decreased ICP and increased cerebral perfusion pressure. A clear relationship between hypocarbia and frequency of cerebral ischemia was seen. The frequency of regional ischemia (CBF <18 mL•100g-1•min-1) was 28.9% during normocapnia and increased to 73.1% for PaCO₂ <25 mm Hg.

The effect of hyperventilation therapy on outcome of infants and children with severe TBI has not been directly compared with other therapies such as hyperosmolar agents, barbiturates, hypothermia, or early decompressive craniectomy. Surprisingly, outcome data reported by Bruce et al.2 in the late 1970s, when aggressive hyperventilation (PaCO₂ 20-25 mm Hg) represented the cornerstone of therapy for pediatric TBI5, has not been surpassed by contemporary protocols and only rarely equaled23. Hyperventilation may have unique advantages as a therapy in severe pediatric TBI; however, it can only be supported as a second tier therapy based on the current evidence.

Key Elements from the Adult Guidelines Relevant to Pediatric TBI

The adult guidelines1 conclude that prophylactic hyperventilation (PaCO₂ <35 mm Hg) therapy during the first 24 hrs after severe TBI should be avoided because it can compromise cerebral perfusion during a time when CBF is already reduced. The guidelines1 strongly contend that chronic prolonged hyperventilation therapy (PaCO₂ <25 mm Hg) should be avoided after severe TBI in the absence of increased ICP. It was emphasized that the preponderance of the physiologic literature concludes that hyperventilation during the first few days following severe TBI, whatever the threshold, is potentially deleterious in that it can promote cerebral ischemia.

Specifically, CBF during the first day after injury is less than half that of normal individuals18, 24-31. During the first 24 hrs after injury, there is a direct correlation between CBF and Glasgow Coma Scale score or outcome24, 29. Hyperventilation reduces CBF22, 32-34 even further but does not consistently reduce ICP35, 36 and may cause loss of autoregulation37. Aggressive hyperventilation may cause arterial jugular venous oxygen and CBF to approach ischemic levels.

In a prospective randomized clinical trial by Muizelaar et al.8, 77 severe TBI patients were randomized to a group treated with chronic prophylactic hyperventilation (PaCO₂ of 25 ± 2 mm Hg) or to a group kept relatively normocapneic (PaCO₂ of 35 ± 2 mm Hg). At 3 and 6 months after injury, patients with initial Glasgow Coma Scale motor scores of 4-5 in the hyperventilation group had a significantly worse outcome than did patients in the normocapneic group. Statistically significant differences between the two groups were not found at 1 yr after injury; this was attributed to a type II statistical error since substantially fewer patients were available for 1-yr follow up.

Hyperemia may not be as common in severe pediatric TBI as previously reported. Hyperventilation can reduce CBF to potentially ischemic levels. Additionally, the cerebrovascular response to hyperventilation can be extremely variable following TBI. Studies in children with severe TBI raise the concern that the toxicity of hyperventilation may be similar to the toxicity that has been demonstrated in adults and related to adverse outcome. Unfortunately, the precise relationship between hyperventilation and outcome has not been studied in children with severe TBI.

• Studies to identify subgroups of patients who might benefit from hyperventilation are needed.
• Studies are needed to address the timing and duration of the optimal use of hyperventilation.
• Studies to determine the optimal monitoring technique of patients undergoing hyperventilation are lacking.
• Studies are needed to address the influence of age on the response to hyperventilation
• The effects on long-term outcome should be addressed in all aspects of research on hyperventilation. Hyperemia may not be as common in severe pediatric traumatic brain injury as previously reported.