Minimizing secondary injury after brain trauma is the primary goal of cerebral resuscitation. Induced hypocapnia can lead to a lifesaving reduction of intracranial pressure through cerebral vasoconstriction in patients with severe intracranial hypertension after cerebral trauma. Hyperventilation to Paco2 of approximately 25 mm Hg has been one of the cornerstones in the management of traumatic brain injury for >45 yr. Evidence suggests that hyperventilation of 5 days' duration in critically ill patients may adversely affect outcome.1 Moreover, it has been demonstrated in acute studies using various biochemical and metabolic indices that hyperventilation might turn borderline ischemia that is present after severe head trauma into frank ischemia, which will lead to neuronal cell death.2,3 Despite the absence of any outcome studies with hyperventilation for <5 days, the Brain Trauma Foundation guidelines recommend that chronic prolonged hyperventilation therapy should be avoided after severe traumatic brain injury but that hyperventilation for brief periods may be necessary in the management of neurologic deterioration. Neither “prolonged” nor “brief periods” are defined.4 To address the absence of outcome data on the effect of short periods of hyperventilation, we investigated the effects of 4 h posttraumatic mild to moderate hyperventilation on neurocognitive and advanced motor function and coordination as well as lesion volume in rats up to 3 wk after focal traumatic brain injury. We hypothesized that this period of hyperventilation would have no demonstrable effect compared with normocapnia.
The study was approved by the institutional animal care committee and performed in accordance with the German animal protection law “Deutsches Tierschutzgesetz.” Twenty-one male Sprague-Dawley rats (50 g) were delivered from Charles River Laboratories (Kisslegg, Germany) 42 days before operation day and kept in the animal facilities under standard laboratory conditions (12 h light/12 h dark, lights on at 6:00 am, 22°C, 60% humidity, and free access to water and standard rat chow). Rats were acclimatized in groups for 3 wk. Another week was allowed for habituation to the home cage and attached test arena before starting the 2-wk training period.
The cognitive performance was evaluated by investigators blinded to treatment allocation using the modified hole-board test.5,6 This test was designed to investigate cognitive, exploratory, and motivational variables in rats without negative enforcement. The home cage and the test arena were separated from each other by a transparent partition, perforated with holes to allow visual and olfactory contact between group mates. The hole board (20 × 40 cm) was placed in the middle of the test arena. The board was made of opaque gray polyvinyl chloride with 15 holes staggered in 3 lines covered by lids, which could easily be opened. After opening, coil springs forced the lids immediately back to their original position. Three lids were marked with white adhesive tape (every day different lids) and contained a food reward (puffed rice). All holes were flavored with the aroma of black currants (black currant flavor dissolved in water; concentration 0.02%) to cover the odor and smell of the puffed rice.
Seven days before the cognitive training period, all animals were habituated to the testing environment. During this phase, the partition was removed and the animals were allowed to explore both the test arena and the home cage compartment. During habituation, the holes of the board were neither marked nor baited. After being returned to their home cage, the animals received several pieces of puffed rice to habituate to the food reward.
During the training phase, rats were tested for 4 trials a day for 14 successive days with marked and baited holes. The sequence of marked holes was randomly changed every day. With completion of the training period, baseline conditions were obtained for 3 days. The following day, animals underwent surgical preparation with induction of controlled cortical impact (see below). After surgery, the animals were returned to their home cages for daily testing for 20 days.
The main variable evaluated with the hole-board test was deficits within the declarative or explicit memory. Deficits were assumed if rats visited nonbaited holes or did not visit baited holes referred to as wrong choices. Other observations included anxiety-related behavior (prolonged latency to enter the open field of the board) and exploratory motivation, which was assumed if the rats hesitated to search for the rewards (latency hole visit) or were not interested in their environment (i.e., showed less grooming behavior).
To evaluate advanced motor function and coordination, a beam walk (number of wrong steps with the left hindpaw) and beam balance test (maximum 60 s) were performed. The beam for the walk test was 2.5 cm, and the beam for the balance test was 1.5 cm wide. Both were 130 cm long and placed 60 cm above the ground. During the 3 days before controlled cortical impact, rats were trained, and after controlled cortical impact, rats were tested for 20 days.
Additionally, awareness, grooming, ability to walk and climb, force on a rotating screen, and prehensile traction were assessed daily as described previously7 and summarized as activities of daily living.
On the day of trauma induction, fasted rats (369 ± 15 g) were anesthetized in a bell jar saturated with halothane. Rats were then tracheally intubated and their lungs mechanically ventilated (Harvard Rodent Ventilator, Model 683, Harvard Apparatus, South Natick, MA, Paco2 = 38–42 mm Hg, Capnomac, Datex, Helsinki, Finland) with 1–1.5 vol% halothane in N2O/O2 (Fio2 = 0.33). Catheters were inserted in the tail artery for continuous arterial blood pressure measurement and blood sampling and in the femoral vein for administration of norepinephrine if needed. Temperature probes were placed in the right temporal muscle and the rectum. Pericranial temperature was maintained constant during the experiment at 37.5°C using a servo-controlled overhead heating lamp. Also, needle electrodes were inserted subcutaneously to record electrocardiogram and measure heart rate (Cadiocap II, Datex). Rats were then placed in a stereotactic “U”-frame (Model 962, Kopf Instruments, Tujunga, CA) using atraumatic ear bars. After a midline incision, the scalp was retracted, exposing the right parietal bone. A 6-mm craniotomy was performed between bregma and lambda, and the coronal ridge using a high-speed dental drill with a 0.9-mm tip cooled with lactated Ringer's solution. After drilling 3 sides without injuring the dura mater, the bone flap was opened over the sagittal suture.
The controlled cortical injury device (SHT-3 CCI Controller, custom made, Johannes Gutenberg-Universität, Mainz, Germany)8,9 consisted of a pneumatic cylinder rigidly mounted on a crossbar. On the lower end of the rod, the impact tip (5 mm in diameter) was attached (i.e., the shaft that comes into contact with the exposed dura). To induce controlled cortical impact, this tip was vertically driven at a predetermined velocity (4 m/s), depth (1.75 mm), and duration (200 ms) of brain deformation.
Ten minutes before adjustment of the impact tip, shaft halothane concentration was decreased to 0.8 vol%, and baseline measurements were taken. After controlled cortical impact induction, animals were immediately removed from the injury device, the bone flap was closed and sealed with histoacrylic glue (Histoacryl, B. Braun, Tuttlingen, Germany). The scalp incision was closed with interrupted sutures, and 0.2 mL of bupivacaine 0.5% (Curasan AG, Kleinostheim, Germany) was infiltrated. Nitrous oxide was now replaced by air (Fio2 = 0.33), and animals were then randomized to either normoventilation (n = 10; Paco2 = 38–42 mm Hg) or hyperventilation (n = 11; Paco2 = 28–32 mm Hg) and their lungs ventilated for 4 h, respectively. During this period in both groups, physiologic variables were continuously monitored, maintained stable, and documented shortly after controlled cortical impact (0.00), after 1:00, 2:30, and 4:00 h (respiratory rate, pH, and glucose). Mean arterial blood pressure was kept constant by intermittent low-dose IV injection of norepinephrine (Arterenol, Aventis Pharma, Frankfurt a. M., Germany, 6 μg · kg−1 · h−1). After 4 h of posttraumatic ventilation, catheters were removed and the incisions closed and infiltrated with 0.2 mL bupivacaine 0.5%. After weaning from ventilation and tracheal extubation, the completely awake rats received buprenorphine 0.03 mg subcutaneously as analgesia and were returned into the hole-board cage.
On Day 20, animals were killed by decapitation in deep anesthesia and brains were removed and frozen at −70°C. From each animal, sets of 11 consecutive coronal 10-μm cryostat brain sections with 1-mm intervals from a defined zero point (macroscopically visible morphologic formation at −1600 μm ante bregma, Plate 11) were cut and mounted.10 These sections were stained with cresyl violet and examined by 1 investigator blinded to the treatment conditions.
The cresyl violet–stained coronal sections were digitized using a camera (Evolution MP Camera, Media Cybernetics, Silver Spring, MD). The lesion area was measured by determining the cross-sectional injury in each image and multiplied by the exact thickness of the tissue between the slices. This slab volume technique was implemented on the image processing program Image-Pro Express 4.5 and created the lesion volume (in mm3) (Media Cybernetics).
Hemodynamic, biochemical, and blood gas variables were subjected to a 2-way analysis of variance with the within-groups factors time2, the between-group factor ventilation, and their interaction term (time2 × ventilation). If this interaction term was significant (P < 0.05), post hoc analyses were performed in a hierarchical manner by Bonferroni-corrected t-tests or paired t-tests. Effects of time levels were analyzed quadratically (time2) for physiologic as well as neurologic and cognitive values, focusing on biphasic changes of these variables during the observation period.
The variables of the modified hole-board test and those of the neurologic tests were subjected to a 2-way repeated-measurement analysis of variance with the within-group factor, postoperative day, the between-groups factor ventilation, and all possible interaction terms. Post hoc analyses were performed in a hierarchical manner by Bonferroni-corrected t-tests.
Lesion volume and lesion area were analyzed using t-test with the 2 groups as independent variables. The sample size estimate was based on a critical difference of 2 × sd of at least 1 quality of brain function, i.e., cognitive, motor, behavioral, or sensory function. A power of 0.9 and an α <0.05 was defined. Alpha was divided by 4 to address the 4 possible reasons for a difference (repeated measurements). The number calculated by this technique was n = 8 but we increased to n = 10 to address the complexity of the project.
Continuous data were expressed as mean ± sd and presented in figures as mean ± sem. P > 0.05 was considered significant. Statistical analyses were performed using SPSS 16.0 for Windows (SPSS, Chicago, IL).
One animal of the hyperventilated group died on Day 1 after controlled cortical impact for unknown reasons. The frozen brain from 1 animal of the normoventilated group was damaged and therefore lost for histologic evaluation. Results of the behavioral tests of the normoventilated animal and physiologic variables of both animals were included in the analysis.
Differences in physiologic data before and after controlled cortical impact are shown in Table 1. There was a study-related marked increase in respiratory rate and pH in the hyperventilated group compared with the normoventilated group and baseline levels. Blood glucose levels decreased over time with values always within the physiologic range of rats. There were no differences in mean arterial blood pressure (90 ± 10 mm Hg), heart rate (325 ± 25 bpm), or pericranial temperature (37.5°C) and at no time were rats hypoxemic (Pao2: 125 ± 15 mm Hg).
Animals in the hyperventilated group showed a significant deficit in declarative (explicit) memory on Days 1 and 2 after controlled cortical impact compared with baseline and over time compared with the normoventilated group (time2 × group, P < 0.05) (Fig. 1). There were no differences in any anxiety-related or motivational variables between groups (Fig. 2). Both groups showed motor function and coordination deficits in the beam walk and beam balance test for the first days after controlled cortical impact compared with baseline (Fig. 3). Overall, hyperventilated animals performed significantly worse than animals in the normoventilated group (time2 × group, P < 0.05). In both groups, performance returned to baseline levels on Day 6. No differences between groups or comparisons with baseline were observed in awareness, grooming, ability to walk, climb, or apply force on a rotating screen, prehensile traction, anxiety-related behavior, and arousal.
Lesion volume (mm3) was significantly increased in hyperventilated animals (69.66 ± 13.01) compared with normoventilated animals (48.34 ± 15.6). Lesion areas (mm2) are shown in Figure 4.
The results of this study indicate that 4 h of hyperventilation (Paco2 = 28–32 mm Hg) after controlled cortical impact transiently impairs declarative (explicit) memory and advanced motor function and coordination to a greater degree than normoventilation. Longer-term lesion volume after traumatic brain injury was significantly larger in hyperventilated animals. Activities of daily living were not affected by traumatic brain injury or mode of ventilation in rats.
Severe traumatic brain injury is a major cause of morbidity and mortality, particularly among young men.11 Experimental and clinical studies have shown that brain damage does not cease with the primary injury but progresses over subsequent hours and days. Subsequent or secondary injury occurs when factors not initially present worsen outcome, and cerebral ischemia is a common contributor. Treatment of the primary injury is virtually impossible; therefore, therapeutic management of patients with traumatic brain injury focuses on minimizing the extent of the secondary injury.
The use of hyperventilation, defined as the induction and maintenance of levels of Paco2 less than the normal range (<35 mm Hg), in patients with head trauma has been advocated for >45 yr.12 Its main effect is a decrease in intracranial pressure by cerebral arterial vasoconstriction with resultant decreased cerebral flow and blood volume. There is concern and controversy whether the cerebral vasoconstriction may be excessive and lead to ischemia with worsening of secondary injury. The controversies have been extensively reviewed elsewhere.13 Briefly, biochemical surrogate end points such as glutamate and lactate have been shown to increase significantly.14 Many other surrogates have been used including jugular venous oxygen saturation15 and regional cerebral blood flow16 with many, but not all, studies showing a few minutes of hyperventilation worsening the measured variable. This type of study has prompted the Brain Trauma Foundation guidelines.
However, even with expert guidelines recommending against it,4 deliberate hyperventilation continues to be widely practiced. According to a survey conducted before the Brain Trauma Foundation guidelines were published, 83% of board-certified neurosurgeons in North America used prophylactic hyperventilation.17 Even though this proportion has significantly decreased after publication of the first Brain Trauma Foundation guidelines in 1995, 36% of neurosurgeons in North America still prophylactically hyperventilate patients with traumatic brain injury.17 Two years after this publication, a survey by Huizenga et al.18 showed that 47% of emergency clinicians would elect to use prophylactic hyperventilation despite guidelines recommending otherwise. In the same year, Thomas et al.19 were able to demonstrate that 60% of already intubated traumatic brain injury patients were hyperventilated, and 70% had an inappropriately high assisted ventilation rate during transport to the hospital. A more recent study found that only 30% of severe head trauma patients were transported with a Paco2 in the appropriate target range.20
Despite current recommendations, there is little clinical evidence for a worse outcome after relatively short periods of mild to moderate hyperventilation. The most recent Brain Trauma Foundation guidelines address this matter as a key issue for future investigations.4 Although our clinical interest was to crudely “mimic” transportation from the accident scene to the hospital and initial evaluation and treatment or from the intensive care unit to imaging and then the operating room, we also needed to use a duration of hyperventilation long enough to actually produce a quantifiable injury. A previous study that used less sensitive behavioral testing and 5 h of hyperventilation found no difference in behavior but a worse histologic lesion.21 We estimated that using our more sensitive tests for 4 h would result in quantifiable injury that would allow detection of differences between groups if there were any.
To be able to broadly evaluate clinical outcome, several determinants including neurocognition (declarative memory, anxiety-related behavior, and arousal), advanced motor function and coordination (beam walk and beam balance test), activities of daily living, and anatomical variables (brain lesion volume) were studied. Declarative memory is the aspect of memory that stores facts and is a term for information that is available to conscious recollection and in people to verbal retrieval (i.e., it can be “declared”). Neurophysiologically, declarative memory requires the medial temporal lobe, especially the hippocampus and related areas of the cerebral cortex of the brain.22 Therefore, the extent and severity of a lesion in this area induced by the controlled cortical impact with secondary injury will impair the function of this memory system and is detectable by the hole-board test. The hole-board test has advantages over other cognitive tests used in animals, e.g., the Morris-water maze or the radial-arm maze.23,24 The rats are habituated with a minimum of stress. No food deprivation is necessary because the rats perform the test voluntarily, and little effort is needed for the rats to move the lid and find the reward.5 These factors likely enhance the motivation of the animals to perform the test in comparison with other cognitive tests. Furthermore, our use of a relatively long prehead trauma acclimation and training period makes continuing learning after the injury highly unlikely. A parallel experiment included a sham group, i.e., anesthesia but no brain injury, and the sham inhaled anesthetic group displayed no cognitive changes over time other than a brief period of motor (grooming) hyperactivity.25
Using the modified hole-board test, animals that received controlled cortical impact treated with hyperventilation showed significant short-term (4 days) worsening of declarative memory compared with normocapnic animals. Because anxiety-related behavior, motivation, and activities of daily living were neither influenced by ventilatory treatment nor differed compared with baseline, our results suggest a hyperventilation-associated secondary functional lesion, i.e., cognitive dysfunction after controlled cortical impact.
Advanced motor function and coordination were impaired in both groups during the first days after traumatic brain injury with hyperventilated animals performing about 3 times worse than normoventilated animals. A similar study in rats subjected to more aggressive hyperventilation (Paco2 = 20.3 ± 0.7 mm Hg) did not show differences between groups in a beam balance test.21 The trauma depth in that study was larger (2.5 mm compared with 1.75 mm), and rats were trained for only 1 day compared with 3 days in this study. These differences between studies might account for the different results.
The lesion volume in this study was significantly larger in the hyperventilated group, which is consistent with a worsening of the secondary injury. We evaluated lesion volume when scar formation was most likely completed, and there was minimal edema. Additional time points to assess histopathologic damage, for example, when secondary damage is most prominent (24 h to 3 days), would have been helpful to determine a time course of events. However, emphasis in this study was placed on longer-term outcome with sequential behavioral and motor function tests rather than tracking sequential histologic events. The traumatic brain injury model used is well characterized and creates a highly reproducible lesion in a very controlled fashion. Relevant stages of the pathophysiology of human traumatic brain injury, including contusion and axonal injury, are reproduced in this model. The evaluation, procedure, and compromising factors of this method were described elsewhere.26
Animals in this study were fasted for 8 h with free access to water even though in hindsight there are suggestions that, when rats are handled stress-free and glucose levels range within normal limits, this would not have been necessary. Even hyperglycemic rats subjected to mild cortical impact injury had adverse effects only when a secondary ischemic insult was added after the impact injury.27 Therefore, fasting versus nonfasting does not seem to be an important factor in controlled cortical injury models.
In conclusion, 4 h of posttraumatic mild to moderate hyperventilation (Paco2 = 28–32 mm Hg) enhanced histologic damage but only transiently impaired neurocognitive performance, motor function, and coordination in rats subjected to traumatic brain injury. These experimental data, if clinically applicable, suggest that short-term hyperventilation as deliberately and inadvertently seen in traumatic brain injury patients immediately after the injury may cause transient neurocognitive deterioration but no long-term effect.
The authors thank Doris Droese and Anne Frye for their expert technical assistance and Kerstin Heimann, DVM, for assistance in evaluating the lesion volume.
1. Muizelaar JP, Marmarou A, Ward JD. Adverse effects of prolonged hyperventilation in patients with severe head injury: a randomized clinical trial. J Neurosurg 1991;75:731–9
2. Obrist WD, Langfitt TW, Jaggi JL, Cruz J, Gennarelli TA. Cerebral blood flow and metabolism in comatose patients with acute head injury. Relationship to intracranial hypertension. J Neurosurg 1984;61:241–53
3. Coles JP, Fryer TD, Coleman MR, Smielewski P, Gupta AK, Minhas PS, Aigbirhio F, Chatfield DA, Williams GB, Boniface S, Carpenter TA, Clark JC, Pickard JD, Menon DK. Hyperventilation follwing head injury: effect on ischemic burden and cerebral oxidative metabolism. Crit Care Med 2007;35:568–78
4. BrainTraumaFoundation. Hyperventilation—recommendations. J Neurotrauma 2007;24:S87–9
5. Ohl F, Holsboer F, Landgraf R. The modified hole board as a differential screen for behavior in rodents. Behav Res Methods Instrum Comput 2001;33:392–7
6. Ohl F, Oitzl MS, Fuchs E. Assessing cognitive functions in tree shrews: visuo-spatial and spatial learning in the home cage. J Neurosci Methods 1998;81:35–40
7. Englelhard K, Werner C, Reeker W, Lu H, Möllenberg O, Mielke L, Kochs E. Desflurane and isoflurane improve neurological outcome after incomplete cerebral ischaemia in rats. Br J Anaesth 1999;83:415–21
8. Dixon CE, Clifton GL, Lighthall JW, Yaghmai AA, Hayes RL. A controlled cortical impact model of traumatic brain injury in the rat. J Neurosci Methods 1991;39:253–62
9. Lighthall JW. Controlled cortical impact: a new experimental brain injury model. J Neurotrauma 1988;5:1–15
10. Palkovits M, Brownstein MJ. Maps and guide to microdissection of the rat brain. 1 ed. New York: Elsevier Science Publishing Co., 1988
11. Jennett B. Epidemiology of head injury. Arch Dis Child 1998;78:403–6
12. Lundberg N, Kjallquist A, Bien C. Reduction of increased intracranial pressure by hyperventilation. Acta Psychiatr Scand 1959;34:5–64
13. Stocchetti N, Maas AIR, Chieregato A, van der Plas AA. Hyperventilation in head injury: a review. Chest 2005;127:1812–39
14. Marion DW, Puccio A, Wisniewski SR, Kochanek P, Dixon CE, Bullian L, Carlier P. Effect of hyperventliation on extracellular concentrations of glutamate, lactate, pyruvate, and local cerebral blood flow in patients with severe traumatic brain injury. Crit Care Med 2002;30:2619–25
15. Imberti R, Bellinzona G, Langer M. Cerebral tissue PO2
changes during moderate hyperventilation in patients with severe traumatic brain injury. J Neurosurg 2002;96:97–102
16. Settakis G, Lengyel A, Molnar C, Bereczki D, Csiba L, Fülesdi B. Transcranial doppler study of the cerebral hemodynamic changes during breath-holding and hyperventilation tests. J Neuroimaging 2002;12:252–8
17. Marion DW, Spiegel TP. Changes in the management of severe traumatic brain injury: 1991–1997. Crit Care Med 2000;28:16–8
18. Huizenga JE, Zink BJ, Maio RF, Hill EM. Guidelines for the management of severe head injury: are emergency physicians following them? Acad Emerg Med 2002;9:806–12
19. Thomas SH, Orf J, Wedel SK, Conn AK. Hyperventilation in traumatic brain injury patients: inconsistency between consensus guidelines and clinical practice. J Trauma 2002;52:47–53
20. Warner KJ, Cuschieri J, Copass MK, Jurkovich GJ, Bulger EM. The impact of prehospital ventilation on outcome after severe traumatic brain injury. J Trauma 2007;62:1330–8
21. Forbes M, Clark R, Dixon C, Graham S, Marion D, DeKosky ST, Schiding J, Kochanek PM. Augmented neuronal death in CA3 hippocampus following hyperventilation early after controlled cortical impact. J Neurosurg 1998;88:549–56
22. Tulving E, Schacter DL. Priming and human memory systems. Science 1990;247:301–6
23. Hodges H. Maze prozedures: the radial-arm and water maze compared. Brain Res Cogn Brain Res 1996;3:167–81
24. Morris R. Developments of a water-maze procedure for studying spatial learning in the rat. J Neurosci Methods 1984;11:47–60
25. Eberspächer E, Eckel B, Englelhard K, Müller K, Hoffman WE, Blobner M, Werner C. Effects of sevoflurane on cognitive deficit, motor function, and histopathology after cerebral ischemia in rats. Acta Anaesthesiol Scand 2009;53:774–82
26. Eberspächer E, Heimann K, Hollweck R, Werner C, Schneider G, Englelhard K. The effect of electroencephalogram-targeted high- and low-dose propofol infusion on histopathological damage after traumatic brain injury in the rat. Anesth Analg 2006;103:1527–33
27. Cherian L, Goodman JC, Robertson CS. Hyperglycemia increases brain injury caused by secondary ischemia after cortical impact injury in rats. Crit Care Med 1997;25:1378–83