SciELO - Scientific Electronic Library Online

vol.55 issue3Onset of 1% lidocaine for skin infiltrative anesthesiaHemodynamic and renal effects of high epidural clonidine doses in dogs author indexsubject indexarticles search
Home Pagealphabetic serial listing  

Services on Demand



Related links


Revista Brasileira de Anestesiologia

Print version ISSN 0034-7094

Rev. Bras. Anestesiol. vol.55 no.3 Campinas May/June 2005 



Effects of induced hypertension on brain compliance and perfusion pressure in experimental intracranial hypertension: comparison between cryogenic brain injury and subdural balloon*


Efectos de la hipertensión arterial inducida sobre la complacencia y presión de la perfusión encefálica en hipertensión intracraneana experimental: comparación entre lesión encefálica criogénica y balón subdural



Nelson Mizumoto, TSA, M.D.I; Humberto Katsuji Tango, TSA, M.D.II; Marcelo Lacava Pagnocca, TS, M.D.III

ISupervisor em Neuroanestesia do Instituto de Psiquiatria do HCFMUSP. Anestesiologista do Hospital Israelita Albert Einstein, Co-responsável do CET/HCFMUSP, Doutor em Ciências pela Disciplina de Anestesiologia da FMUSP
IIAssistente Colaborador do Instituto de Psiquiatria do HCFMUSP, Instrutor de Ensino do CET-HCFMUSP, Doutorando da Disciplina de Anestesiologia da FMUSP
IIISupervisor em Neuroanestesia da IMSCSP, Co-responsável do CET-IMSCSP. Doutor em Ciências pela Disciplina de Anestesiologia da FMUSP. Supervisor em Neuroanestesia da FMUSP





BACKGROUND AND OBJECTIVES: Traumatic brain injury (TBI) may increase intracranial pressure (ICP) and decrease brain compliance (BC). Different injuries are applied to TBI models studying the same variables. Since they are indistinctly used, the objective was to compare ICP and BC in two different TBI models.
METHODS: This study involved 18 male dogs anesthetized, ventilated and randomly distributed in two groups: SB - subdural balloon (n = 9) and CI - cryogenic injury (n = 9). ICP, BC and cerebral perfusion pressure (CPP) were evaluated in five moments: end of preparation (M0), normal brain (M1), beginning of injury (M2), end of injury (M3) and established injury (M4). BC is ICP variation during induced hypertension (IH) in 50 mmHg in M1 and M4. CPP = Mean Blood Pressure (MBP) - ICP. Paired Student's t test was used for the same group in different moments and Student's t test was used for two different samples in the same moment between groups.
RESULTS: MBP was similar for both groups in all studied moments (p = 0.31 in M0; p = 0.25 in M1; p = 0.31 in M2; p = 0.19 in M3; p = 0.05 in M4). ICP was similar between groups in M0 (p = 0.27) and M1 (p = 0.21), however different in M2 (p < 0.001). ICP was similar for both groups in M3 (p = 0.39) and M4 (p = 0.98), increased for SB in M1 (p = 0.04) and M2 (p = 0.01), but not in M3 (p = 0.36) and M4 (p = 0.12). For CI, ICP has increased in M1 (p < 0.01), M3 (p < 0.001) and M4 (p < 0.001), but not in M2 (p = 0.18). There has been CPP increase in M1 (p < 0.001) and M4 (p < 0.001), with no difference between groups (p = 0.16 in M1 and p = 0.21 in M4). There has been decreased CPP in M2 for both groups (p < 0.001), however more severe for CI (p < 0.001). In M3, there has been increased CPP for SB (p = 0.02) and decreased CPP for CI (p = 0.01), what has made CPP similar for both groups (p = 0.43). CPP has equally increased in M4 for both groups (p = 0.16).
CONCLUSIONS: Induced hypertension (IH) effect on CI model is comparable to what has been observed in the SB model. This type of injury should be better studied to establish precision in the ratio between injury extension and BC decrease, which seems to be a gradual and evolving process, with not totally understood limits.

Key words: ANIMALS: dogs; CENTRAL NERVOUS SYSTEM: brain compliance, intracranial hypertension, TRAUMA: brain


JUSTIFICATIVA Y OBJETIVOS: El trauma craneoencefálico (TCE) puede elevar la presión intracraneana (PIC) y reducir la complacencia encefálica (CE). Diferentes lesiones son aplicadas en modelos de TCE que estudian las mismas variables. Como son usadas indistintamente, el objetivo es comparar la PIC y la CE en dos modelos de TCE.
MÉTODO: Dieciocho perros machos, anestesiados, ventilados y distribuidos eventualmente en dos grupos: BS - balón subdural (n = 9) y LC - lesión criogénica (n = 9). Análisis de la PIC, CE y presión de perfusión encefálica (PPE) en cinco momentos: final de la preparación (M0), encéfalo normal (M1), inicio de la lesión (M2), término de la lesión (M3) y lesión establecida (M4). CE es la variación de la PIC durante la hipertensión arterial inducida (HAI) en 50 mmHg en M1 y M4. PPE = presión arterial media (PAM) - PIC. Prueba t de Student emparejado para el mismo grupo en diferentes momentos y t de Student para dos muestras diferentes para el mismo momento entre los grupos.
RESULTADOS: La PAM fue semejante en los grupos en los momentos estudiados (p = 0,31 en M0; p = 0,25 en M1; p = 0,31 en M2; p = 0,19 en M3; p = 0,05 en M4). La PIC fue semejante en los grupos en M0 (p = 0,27) y M1 (p = 0,21), pero diferente en M2 (p < 0,001). La PIC se volvió semejante en los grupos en M3 (p = 0,39) y M4 (p = 0,98), se elevó en BS en M1 (p = 0,04) y M2 (p = 0,01), pero no en M3 (p = 0,36) ni en M4 (p = 0,12). En el LC la PIC aumentó en M1 (p < 0,01), M3 (p < 0,001) y M4 (p < 0,001), pero no en M2 (p = 0,18). Hubo aumento de la PPE en M1 (p < 0,001) y M4 (p < 0,001), semejante en los grupos (p = 0,16 en M1 y p = 0,21 en M4). En M2 hubo reducción de la PPE en los grupos (p < 0,001), más intensa en LC (p < 0,001). En M3 hubo aumento de la PPE en BS (p = 0,02) y reducción en el LC (p = 0,01), lo que hizo la PPE semejante en los grupos (p = 0,43). En M4 hubo aumento de la PPE semejante en los grupos (p = 0,16).
CONCLUSIONES: El efecto de la HAI en el modelo de LC es comparable al observado en el modelo de BS. Ese tipo de lesión debe ser mejor estudiada para establecer precisión en la proporción entre la extensión de la lesión y la reducción de la CE, aparentemente, un proceso gradual y evolutivo cuyos límites aún no son totalmente conocidos.




Brain is an elastic organ which, in normal conditions, may mildly and rapidly change in volume by changing cerebrospinal fluid (CSF) and/or blood content, eventually changing intracranial pressure (ICP) depending on brain compliance (BC) 1,2. Brain compliance - also referred as elastance - may be understood as the measure of nervous tissue viscoelasticity, or stiffness. In mathematical terms, elastance represents volume variation observed in an elastic body in response to pressure changes, while compliance expresses the reverse ratio. In biological terms, the concept expresses the ability to compensate intracranial volume increase, and is especially useful in pathological situations 1,3. According to Monro-Kellie's doctrine, when the volume of one of the brain component increases, other(s) should proportionally decrease, otherwise ICP will increase. Sustained ICP increase above 15 mmHg is a pathological condition defined as intracranial hypertension (ICH) 2.

Traumatic brain injury (TBI), especially when severe, promotes ICH 4 - originated from water (edema), CSF (hydrocephalus), intravascular blood (brain swelling) or extravascular blood (extradural, subdural or intraparenchymatous hematomas) - in addition to impairing brain vascular auto-regulation (BVAR) 5 and may decrease perfusion pressure depending on mean blood pressure (MBP) 6. In addition, TBI patients suffer trauma-induced pain and also different manipulations during treatment, such as intubation and tracheal aspiration, skin punctures, vesical and nasogastric catheters, which are potentially nociceptive stimulations promoting MBP increase. This induced hypertension may increase blood volume in the intracranial space, especially if there is BVAR and BC decrease. For these reasons, TBI may change brain compliance and perfusion in different ways.

Type and intensity of those changes are dependent on the way, speed and intensity in which TBI promotes ICH, so, the experimental model is major determinant of observed results and inferred conclusions.

Known experimental models try to mimic the same phenomena of neuronal aggression after TBI. Two specific models are highlighted for having specific characteristics similar to those found in THI patients: decreased brain compliance and perfusion pressure.

The first of these models is obtained with gradual inflation of a balloon located in the intracranial space (extradural, subdural or intraventricular). Major advantage is to increase ICP in a controllable way 7, allowing for intensity to be maintained in desired times and values 3, thus helping the study of brain compliance and perfusion. However, although being possible to reach ICH values to decrease CPP 8, blood-brain barrier (BBB) injury is not so severe as compared to other models or to TBI in humans 9, which is a disadvantage of this model.

The second model creates a focal brain injury by intensive cooling (cryogenic injury). Histopathological brain changes 10 are observed in this model, such as BBB rupture, edema, intraparenchymatous hemorrhage and major impairment in BVAR mechanism 5, promoting increased intravascular volume as a consequence of stasis 8. However, the mass effect, as observed in the balloon model, is incipient.

In cryogenic injury models, BBB injury and resulting edema seem to be more important in decreasing BC as compared to balloon models in which mass effect seems to be the primary component to decrease BVAR, thus changing BC.

Both models have been often used to study TBI without taking those differences into consideration and were still not compared to check changes in BVAR and BC, especially in the presence of hypertension concomitant to ICH, so it is still not clear whether they are in fact comparable.

This study aimed at comparing the effect of induced hypertension on brain compliance and perfusion pressure in two experimental models in dogs: brain injury induced by subdural balloon inflation or by cryogenic cooling.



This study involved 18 mixed breed dogs of both genders, weighing 10 to 20 kg, in 10-hour fast, without preanesthetic medication and after evaluation and approval by the veterinary in charge of the experimental animals facility.

Dogs were placed in transparent acrylic boxes saturated and ventilated with 4% halothane in a 10 L.min-1 O2/N2O (1:1) gas flow. After unconsciousness and decreased respiratory rate to 12 rpm, animals were placed on Claude-Bernard device in the supine position and were intubated. Inspired halothane fraction was decreased to 2%, mechanically controlled ventilation was installed with tidal volume of 8, I/E ratio of 1:1 and frequency was adjusted to maintain expired CO2 in 33 ± 2 mmHg.

Pulse oximeter probe was placed on the tongue and ECG in DII lead on paws; tympanic thermometer was placed in right side, in addition to capnography and blood gases analyzer. Subcutaneous infiltration with 1 mL of 1% lidocaine was performed before every incision. Right femoral artery was dissected and catheterized for continuous mean blood pressure (MBP) monitoring and right femoral vein was dissected and catheterized for central venous pressure (CVP) monitoring. Left femoral artery was dissected and catheterized to collect the blood samples for biochemical dosages and ipsilateral femoral vein was dissected and catheterized for infusion of anesthetics and saline. Halothane inspired fraction was decreased to 1% and 2 µ fentanyl and 0.08 pancuronium were administered. Position was changed to "sphinx" position. Skullcap was surgically exposed. Circular trepanation with 1 cm diameter was performed 1.5 cm laterally to skull median line and intracerebral sensor was introduced to monitor ICP.

Animals were randomly distributed in two equal groups according to the scheme:

1 - Subdural balloon group (SB) - 3 mm x 5 mm trepanation and dural opening in the hemisphere contralateral to ICP catheter, introduction of 2 cm of Foley's catheter distal edge in the subarachnoid space. Tight sealing with acrylic resin. Slow and gradual balloon inflation with 0.9% NaCI solution for maintenance ICP around 40 mmHg during 20 minutes (Figures 1 and 2).

2 - Cryogenic injury group (CI) - Plastic funnel fixation with acrylic resin to the skullcap contralateral to ICP trepanation. The area delimited by funnel edge fixed to skullcap was a 25 mm diameter circumference. Then liquid nitrogen was poured through the funnel on the bone during 20 minutes (Figures 3 and 4).

Brain compliance test: is the core of the experimental model proposed in this study and depends on the functional status of brain vascular auto-regulatory mechanism. It consists on controlled MBP increase from a baseline value around 90 mmHg to the limit of 140 mmHg and in measuring resulting ICP and CPP.

Blood brain volume (BBV) is a function of brain blood flow (BBF). So, increased blood pressure within these pressure limits barely changes intracranial volume if BVAR mechanism is intact, and when it is changed, it is to a mildly lower value; so, this induced hypertension does not increase ICP 5. Conversely, if brain is aggressed and auto-regulation mechanism is functionally impaired 5, then blood pressure increase within physiological blood pressure interval, will passively dilate resistance arterioles and increase BBF, and as a consequence, increase BBV 8 and intracranial volume which ends up increasing ICP and decreasing brain compliance 2.

As from this conclusion one may qualitatively infer the functional status of vascular auto-regulation 16 and quantify BC by MBP variation within the physiological interval of this phenomenon.

When test is negative, that is, when there is no significant ICP change, the assumption is that BVAR mechanism is intact (curve "a", Figure 5) and that parenchymal elasticity is within the linear region of the brain compliance curve (segment "a", Figure 6).

When the test is positive, that is, when there is significant ICP increase with increased MBP, the assumption is that there is functional BVAR impairment (curve "b", Figure 5) and that brain compliance is partially decreased during moderate ICH (segment "b") or significantly decreased in severe ICH (segment "c") of Langfitt's curve 8 (Figure 6).

Possible BBV change is a function of BVAR activity, but when present, it also depends on the agnitude of pressure variation. If mean blood pressure variation is always the same and all other conditions interfering with brain blood flow are maintained constant, one may consider observed ICP increase as dependent only on BVAR functional status and level of brain compliance.

So, considering that: 1) MBP increases are arbitrary and defined by investigators, thus similar in different study moments; 2) induced hypertensions (final less initial MBP) end up generating brain blood flow increases also similar and proportional to the intensity of MBP variation; and 3) ICP increase is just a consequence of these blood volume increases within the skull, then a ratio may be established between ICP variation as a function of MBP variation, that is, an absolute pressure ratio in pressure terms.

MBP increase was arbitrarily determined as 50 mmHg for being a sufficiently high value to increase ICP in case of BVAR incompetence and to be obtained through titrated and controlled infusion of a diluted norepinephrine solution. We started from a baseline value of approximately 90 mmHg toward a ceiling of approximately 140 mmHg, being careful not to go beyond the upper auto-regulatory phenomenon limit. We have also tried to rigorously start from the same MBP baseline value and to always repeat each MBP increase within the same time interval.

Analyzed parameters - MBP, ICP and CPP - were collected in the following moments:

(M0) - Moment zero - after stabilization, corresponding to 20 minutes after sensors insertion. In this period, sodium (osmolarity), pH, PaCO2, PaO2, O2 saturation (BVAR) and hematocrit (viscosity) of arterial blood were also evaluated to check whether groups were similar in these variables.

(M1) - Moment 1 - First brain compliance test: titrated norepinephrine infusion until target MBP and recording of resulting ICP and CPP - corresponding to normal brain response (BVAR and BC) to induced hypertension.

(M2) - Moment 2 - Brain injury was started 15 minutes after the end of norepinephrine infusion in M1, when hemodynamic variables had already returned to baseline values: subdural balloon inflation until ICP above 30 mmHg in Group SB or beginning of cryogenic injury by pouring liquid nitrogen through the funnel during 20 minutes in group CI.

(M3) - Moment 3 - 20 minutes after subdural balloon inflation in SB group with lowest ICP of 30 mmHg, and 20 minutes after end of cryogenic injury, that is, 40 minutes after cryogenic injury beginning in group CI, corresponding to end of injuries establishment in both groups.

(M4) - Moment 4 - Second compliance test: titrated norepinephrine infusion until target MBP and recording of resulting ICP and CPP - corresponding to injured brain response (BVAR and BC) to induced hypertension. Figure 7 shows the chronology of the experiment.

Animals were sacrificed at the end of the experiment by increasing anesthetic depth followed by bolus 10 mL of 19.1% potassium chloride injection. Statistical analysis used paired Student's t test to compare variations within the same group between moments M0 and M1, with intact brain; Student's t test for two different samples to compare both groups in moment M0 (normal pressure) and M1 (hypertension), to check whether animals belonged to the same sampling universe; Student's t test for two different samples was used to compare variables between groups in the following moments:

1. M2 when balloon inflation was started to reach ICP between 30 and 50 mmHg in the subdural balloon group and beginning of cooling in cryogenic injury group.

2. M3 20 minutes after balloon inflation in subdural balloon group and 40 minutes after cooling in cryogenic injury group.

3. M4 after second norepinephrine dose infusion to promote hypertension in both groups.

Paired Student's t test was used to compare within group in the following moments:

1. Between M2 and M3 to check changes between these moments, because injury mechanisms were different for both groups.

2. Between M3 and M4 to check whether norepinephrine infusion had changed each specific type of injury.



Animals of both groups belonged to the same sampling universe when evaluated by hemodynamic parameters (mean blood pressure and heart rate), brain parameters (intracranial and perfusion pressure), metabolic parameters (sodium, potassium, partial CO2 arterial pressure, hematocrit and temperature) and morphometric parameters (weight) in M0 (Tables I, II and III).

Mean blood pressure (MBP): There has been MBP increase in M1 in groups SB (145.7 ± 6.2; p < 0.05) and CI (140.7 ± 11.5; p < 0.05). In addition, intensity of induced hypertension (IH) was similar for both groups in M1 (p = 0.25), as desired. After rest and stabilization period, corresponding to beginning of injury in both groups, MBP in M2 has returned to baseline values for both groups. Similarly, MBP at rest until beginning of second compliance test in M3 was also similar for both groups (SB = 97 ± 7.1 and CI = 91.8 ± 10.3; p = 0.31). After establishing the injury in M3, MBP was similar to M2 for SB (101.3 ± 8.1; p = 0.65) and CI (95.8 ± 9.4; p = 0.41). As expected, there has been significant MBP increase during IH both for SB (148.1 ± 10; p < 0.001) and CI (138.4 ± 8; p < 0.001). In this last test, MBP was slightly higher for SB as compared to CI (p = 0.04) (Table II and Figure 8).

Intracranial pressure (ICP): There has been significant ICP increase in M1 both for SB (9.5 ± 4.7; p = 0.04) and CI (11.3 ± 3.7; p = 0.01), however similar for both groups (p = 0.21). There has been significant ICP increase in M2 for SB (37 ± 7.4; p = 0.01) with balloon inflation, while ICP was maintained unchanged for CI (12 ± 3.7; p = 0.18). At this moment, ICP values became different (p < 0.001). In M3, ICP was maintained high in SB group, with no difference as compared to M2 (32 ± 12.9 and p = 0.39). In group CI, 20 minutes after cryogenic injury, or M3, ICP has significantly increased (26 ± 6.6; p < 0.001), again making ICPs similar (p = 0.39). In M4, during second compliance test, group SB increase was insignificant (38 ± 17.1; p = 0.12) however group CI increase was significant (35.7 ± 6.7; p < 0.001), although ICP values were similar for both groups (p > 0.98 (Table III and Figure 9).

Brain perfusion pressure (BPP): There has been significant CPP increase between M0 and M1, both for SB (136 ± 6.9; p < 0.001) and CI (126 ± 5; p < 0.001), however higher for group SB (p = 0.002). There has been CPP decrease in M2 for SB (60 ± 6.9; p < 0.001) and CI (79.8 ± 10.2; p < 0.001). CPP decrease for SB was higher as compared to CI (p < 0.001). In M3, since there has been CPP increase for SB (72.6 ± 13; p = 0.02) and decrease for CI (69.8 ± 9.9; p = 0.01) CPP has become similar for both groups (p = 0.43). With IH in M4, CPP has increased for SB (110.1 ± 17.1; p < 0.001) and CI (102.6 ± 8.3; p < 0.001). This increase was similar for both groups (p = 0.16) (Table IV and Figure 10).

Temperature: Tympanic temperature was significantly decreased in M2 for both groups: SB (35.4 ± 1.8; p = 0.05) and CI (33.5 ± 1.4; p = 0.006), being lower for CI (p = 0.02) and remaining as such until M4 (Table V).



Different controlled brain aggression models have been used to study changes during and after traumatic brain injury 9,20,29,31-35. Intuitively, the closest model to this type of aggression is skull contusion with solid objects.

However, this aggression often generates intracranial hematomas and bone fractures, which have unpredictable effects on brain compliance 17. So, animal models allowing gradual and controlled ICP increase, thus more predictable, are the most widely used.

Subdural or epidural balloon inflation, intraventricular fluid infusion or other intracranial volume expansion methods may be more predictable as to value and time one wishes to maintain intracranial pressure. However, they do not severely affect blood-brain barrier patency and, as a consequence, nervous tissue edema does not seem to be an important factor in these models. Some authors 18 suggest that after TBI, brain edema contribute to increase intracranial volume more than blood stasis caused by vasoparesis, that is, that ICP increase could be more a consequence of water volume increase, be it extravascular or intracellular, than of intravascular blood.

In addition, ICP in severe traumatic brain injury is increased to a critical value enough to impair brain perfusion 6. This extremely high intracranial pressure has not always been controlled in ICP studies 17, which not always had brain perfusion pressure decreased to these critical levels to the point of impairing vascular reactivity and brain blood flow.

In this model, we have tried to maintain ICP above 30 mmHg for group SB, as seen during severe traumatic brain injuries, and this has significantly decreased brain perfusion pressure from mean baseline value of 87 mmHg in M0 to 60 mmHg in M2 (p < 0.001). With severe enough intracranial hypertension, in addition to decreasing perfusion, there is also muscle tone decrease of brain arterioles and major functional impairment of auto-regulatory mechanism, leading to passive resistance vascular bed dilatation and systemic hypertension 8, further worsening pre-existing intracranial hypertension.

A group 6 has shown than when brain perfusion pressure is below 60 mmHg, brain blood flow starts to decrease. On the other hand, other authors 16 have shown in anesthetized dogs that if blood pressure limits are maintained between 50 and 150 mmHg, brain compliance is directly proportional to perfusion pressure in normal conditions, that is with intact brain and preserved auto-regulation; and is indirectly proportional when brain is aggressed and its auto-regulatory mechanism is functionally impaired. With normal auto-regulation, mean blood pressure variation within physiological intervals is unable to significantly change BBF, and as a consequence to significantly change brain blood volume (BBV) and compliance. As opposed, with abolished auto-regulatory mechanism, BBV is increased due to increased BBF when MBP is increased, thus decreasing compliance.

Brain compliance evaluation by non-volumetric means, that is, without changing intracranial volume, has been used in different studies 19,29,30. These authors have studied ICP curve's B wave morphology in the presence of IH but have not found significant correlation between morphological changes in this wave and brain compliance evaluated by volumetric methods. They have also suggested that the best evaluation method would be volume/pressure ratio used by other authors 2,20.

During transportation, surgery or ICU stay, patients are frequently submitted to nociceptive stimulations increasing blood pressure to undesirable levels, what ends up decreasing neuronal perfusion exactly when brain vascular reactivity mechanism is functionally impaired.

Trying to reproduce this situation, our study has evaluated brain compliance exactly by means of induced hypertension, which also increases intracranial volume when brain is injured 6,8.

Group SB intracranial hypertension was limited to values below 50 mmHg to evaluate compliance in moments following the establishment of the injury.

The comparison of compliance between balloon and cryogenic injury models was important, since we did not know how this model would behave with a previously non-described injury extension.

In addition, excessive subdural balloon inflation could shift adjacent structures leading to menyngeal hemorrhages, hernias, venous system engorgement and hydrocephalus by CSF drainage obstruction 1.

Cryogenic injury model has been used to evaluate blood-brain barrier patency both by gravimetric density and water volume in pathological exams of the aggressed region 10,14. Cryogenic injury also changes brain vascular reactivity in the involved area 5. A group has shown that brain edema and swelling are significant already in the first 15 to 30 minutes after cryogenic injury 10. However, most cryogenic injury experiments have not produced tissue damage extensive enough to increase ICH to values impairing brain perfusion pressure or causing enough damage to BVAR mechanism, because either cooled surface was small or length of cryogenic agent contact with brain was too short 11-15.

Historically, cryogenic injury has not been used to study ICP 11,12,14 because it seems that the extension of the injury described by these authors had not enough volume to significantly impair ICP and BVAR mechanism. Since the impaired volume described is not high, it seems that brain compliance compensation mechanisms, CSF and blood shifts, are effective in maintaining ICP within normal values 8,12.

Our proposal was to evaluate brain compliance using two different models, however for both the level of proposed aggression was higher than what is normally studied and closer to those observed in severe traumatic brain injury patients.

MBP (p = 0.31), ICP (p = 0.27) and CPP (p = 0.19) were similar for both groups in M0, confirming that they belonged to the same sampling universe. Other variables which could have also influenced brain blood flow 36 and, as a consequence, results - central temperature (neuronal metabolism, p = 0.17), plasma sodium (edemas of osmotic origin, p = 0.81), hematocrit (blood viscosity, p = 0.65) and partial CO2 arterial pressure (changes in auto-regulatory mechanism, p = 0.37) - were also evaluated. Similarly, there were no significant differences between groups.

Special care was taken not to increase blood pressure above upper BVAR level (> 150 mmHg), to prevent hypertension of acting on BVAR per se 21, or increasing capillary patency of normal brain 21-23.

Anesthesia was maintained in adequate depth, however not excessive, with halogenate agent expired fraction below 1% (below DE95) due to the association of intravenous opioids in equal doses for both groups. We have also prevented hypercarbia to prevent cardiac arrhythmias and brain vessel dilatation with consequent intracranial blood volume increase during the infusion of norepinephrine associated to halothane 24,25.

Our study has shown ICP threshold increase in significantly in M1 for both groups (p = 0.05 for SB and p = 0.01 for CI). This increase, however, was mild and fugacious, without corresponding biological meaning. Equally important is that there were no significant differences between groups in intensity of this change (p = 0,21). This brief ICP increase during induced hypertension was considered normal. Mild and temporary increase in ICP in response to hypertension in normal brain has also been observed by other investigators 26,31.

Perfusion pressure in M1 has increased more in group SB as compared to group CI (p = 0.001), probably because there has been higher blood pressure increase concomitant with lower intensity increase of ICP in group SB as compared to group CI.

There has been sudden ICP increase for group SB in M2 (p < 0.001) and brain perfusion pressure decrease (p < 0.001) as we had proposed, and these values were maintained for 20 minutes with additional subdural balloon inflation whenever ICP tended to decrease, to obtain a certain level of brain vascular reactivity impairment. After this aggression period and decreased perfusion pressure, there has been mild, although not significant blood pressure increase in M3 for both groups (p = 0.19).

Increased blood pressure at rest for both groups may be explained as central nervous system response to ischemia generated by intracranial hypertension 27,28.

There has been no ICP decrease for group SB in M2 and M3 (p = 0.36). Although not significant, mild ICP decrease was probably due to compliance compensation mechanisms, CSF and blood shift to outside the skull while subdural balloon was deflated.

Although not representing significant differences when separately considered, increased blood pressure associated to decreased ICP has resulted in major brain perfusion pressure increase in M3 for group SB (p = 0.03).

Significant mean blood pressure increase in M4 (p < 0.001 for SB and CI) was expected because the effect of induced hypertension on aggressed brain ICP was exactly the objective of our study.

At this moment, in SB, although ICP increase has been observed, this increase was not significant (p = 0.12) as compared to previous moment M3 when blood pressure was normal for this same group. The result was increased brain perfusion pressure (p < 0.001) with norepinephrine infusion.

Maybe a more severe brain aggression was needed in group SB for further BVAR impairment. Time during which balloon has remained inflated could have been longer than 20 minutes or level of intracranial hypertension could have been above 50 mmHg.

On the other hand, 20-minute ischemia seems to have been reasonable and balloon inflation maintaining ICP above 50 mmHg could shift adjacent brain structures and change blood and CSF dynamics.

On the other hand, in M2, even when liquid nitrogen injury had been already started, ICP was still unchanged for group CI (p = 0.18), probably because changes in blood-brain barrier and brain vascular reactivity were not yet installed. Two authors have shown that both edema and brain swelling are manifested after brain cryogenic injury as early as 15 to 20 minutes after establishment of the injury and continue to evolve up to 12 hours after beginning of injured tissue reperfusion 10.

Since at this moment MBP was similar to baseline value (p = 0.93) and was also not different from group SB MBP (p = 0.31), CPP for group CI was higher as compared to group SB (p < 0.001).

In moment M3, 40 minutes after beginning of cryogenic injury, corresponding to 20 minutes of injury plus 20 minutes of rest for overcooled brain thawing and reperfusion, ICP started to increase (p < 0.001). At this moment, cryogenic injury repercussions on ICP and on brain vascular reactive mechanism were already present, in agreement to these authors' findings 10.

With this ICP increase, CPP was decreased (p = 0.01) in this group (CI). In the same moment M3, CPP was similar for both groups (p = 0.43). In M4, norepinephrine infusion has increased MBP (p < 0.001), leading to CPP increase (p < 0.001) but also to ICP increase (p < 0.001) for group CI. This suggests that decreased compliance, which is a direct consequence of cryogenic injury, was more important than the impairment of this variable observed in group SB when subdural balloon was deflated.

The difference between both studied models suggests that in the subdural balloon model brain compliance is closer to point "b" of intracranial pressure/volume curve, as described for the first time in 1965 8 (Figure 6). In the extensive cryogenic injury model, however, compliance seems to be closer to region "c" of the same curve, especially more lately. This suggests that, in the conditions of our study, cryogenic injury model would be more prejudicial to brain and would further change compliance and vascular auto-regulation mechanism.

Although there were no significant ICP differences between groups during second compliance test (M4) (p = 0.98), the evolution of both brain injuries has shown clearly different trends along time throughout the experiment.

In group CI, edema and loss of BVAR seemed to progress with time, what also suggests a more severe, prolonged and irreversible injury (apoptosis). Although induced hypertension was somewhat higher for group SB as compared to CI, repercussions on ICP, thus on brain compliance, where even higher with extensive cryogenic injury.

ICP changes with cold injury models were not always apparent in previous studies 11,12,14,15, probably because in these models, injury has been small and not enough to significantly impair increased ICP compensation mechanisms, resulting in a model of aggressed brain but with injury not large enough to increase ICP, which was maintained virtually unchanged in all those models.

According to some investigators, brain edema during traumatic head injury is the most important component of the pathophysiology of this type of aggression 4. So, we decided for an aggression that would largely change not only vascular reactivity, as in other models, but also blood-brain barrier patency.

Other investigators have studied hemodynamic variables and ICP during brain cryogenic injury and have determined that injury volume is closely related to brain blood flow changes 12.

Cryogenic injury extension of this study seems to clearly show studied changes, however literature is still scarce and inconclusive about this model, especially in quantifying brain cooling time, in determining direct contact point with liquid nitrogen, that is, whether bone, dura or brain, and in determining the extension of cooled surface. We decided to circumscribe a small skullcap surface and expose it to liquid N2 for a longer time as compared to other authors. Tympanic temperature for group CI was significantly decreased during cryogenic injury establishment in M2 (p = 0.006), being different from group SB temperature (p = 0.02) and remaining as such until the end of the experiment. This temperature decrease was especially due to proximity to thermal sensor and injury mechanism, wich is intensive cooling.

The conclusion was that experimental cryogenic injury model could be used to study brain compliance. It is comparable to subdural balloon model in terms of brain dynamic variables, at least during the time interval of our study.

We have also concluded, in agreement with other findings 12, that injury extension is critical to determine BVAR mechanism impairment and that it should be better standardized in further studies, with different cooling intensities and times in the cryogenic injury model to establish the ratio between injury and resulting compliance decrease.

Brain lesion by liquid nitrogen cooling was measured in two moments: fresh and after fixation with formalin for two weeks to establish correlation between cooling time and lesion extension. The whole injured brain hemisphere was stained with hematoxylin-eosin and analyzed under optic microscopy during the development of pilot of a different study in the same research line of our group, which is being prepared for publication.

For cryogenic injuries, especially if extensive, it has been advocated that, seemingly at least, increased ICP seems to be a gradual and evolving process and that its limits are still lacking better explanation.

Pathophysiological evolution of cryogenic lesion is clearly different from subdural balloon inflation injury, where maybe more inflation or longer intracranial hypertension would be needed. However, repercussions related to brain structures shift, which is not followed by proportionally severe blood-brain barrier injury, do not well represent changes observed in severe traumatic head injury patients, where parenchymatous contusion predominates, as in the cryogenic injury model, presenting more intensive and severe tissue injury of blood-brain barrier as compared to hematomas simulated by the balloon model.



01. Johnston IH, Rowan JO - Raised intracranial pressure and cerebral blood flow: 3. Venous outflow tract pressures and vascular resistances in experimental intracranial hypertension. J Neurol Neurosurg Psychiatry, 1974;37:392-402.        [ Links ]

02. Miller JD - Intracranial volume-pressure relationships in pathological conditions. J Neurosurg, 1976;20:203-213.        [ Links ]

03. Leech P, Miller JD - Intracranial volume-pressure relationship during experimental brain compression in primates. 1. Pressure responses to changes in ventricular volume. J Neurol Neurosurg Psychiatry, 1974;37:1093-1098.        [ Links ]



Correspondence to
Dr. Marcelo Lacava Pagnocca
Address: Rua Piracuama 316/11 Perdizes
ZIP: 05017-040 City: São Paulo, Brazil

Submitted for publication August 23, 2004
Accepted for publication January 17, 2005



* Received from Laboratório de Investigação Médica em Anestesiologia Experimental - LIM-08 Disciplina de Anestesiologia do Departamento de Cirurgia da Faculdade de Medicina da Universidade de São Paulo, SP. Apoio da FAPESP - Fundação de Amparo à Pesquisa do Estado de São Paulo (Processo nº 99/02663-8)