J Neurosurg 103:597607, 2005

Cerebral acidbase homeostasis after severe traumatic brain injury

TOBIAS CLAUSEN, M.D., AHMAD KHALDI, M.D., ALOIS ZAUNER, M.D., MICHAEL REINERT, M.D., EGON DOPPENBERG, M.D., MATTHIAS MENZEL, M.D., PH.D., JENS SOUKUP, M.D., OSCAR LUIS ALVES, M.D., AND M. ROSS BULLOCK, M.D., PH.D.
Department of Anesthesiology and Intensive Care Medicine, Martin-Luther-University HalleWittenberg, Halle/Saale, Germany; Division of Neurosurgery, Virginia Commonwealth University Health System, Richmond, Virginia; Department of Neurosurgery, University of Miami Lois Pope Life Center, Miami, Florida; Department of Neurosurgery, Inselspital Bern, Switzerland; and Servico de Neurocirurgia, Centro Hospitalar de Gaia, Faculdade de Medicina da Universidade do Porto, Portugal
Object. Brain tissue acidosis is known to mediate neuronal death. Therefore the authors measured the main parameters of cerebral acidbase homeostasis, as well as their interrelations, shortly after severe traumatic brain injury (TBI) in humans. Methods. Brain tissue pH, PCO2, PO2, and/or lactate were measured in 151 patients with severe head injuries, by using a Neurotrend sensor and/or a microdialysis probe. Monitoring was started as soon as possible after the injury and continued for up to 4 days. During the 1st day following the trauma, the brain tissue pH was significantly lower, compared with later time points, in patients who died or remained in a persistent vegetative state. Six hours after the injury, brain tissue PCO2 was significantly higher in patients with a poor outcome compared with patients with a good outcome. Furthermore, significant elevations in cerebral concentrations of lactate were found during the 1st day after the injury, compared with later time points. These increases in lactate were typically more pronounced in patients with a poor outcome. Similar biochemical changes were observed during later hypoxic events. Conclusions. Severe human TBI profoundly disturbs cerebral acidbase homeostasis. The observed pH changes persist for the first 24 hours after the trauma. Brain tissue acidosis is associated with increased tissue PCO2 and lactate concentration; these pathobiochemical changes are more severe in patients who remain in a persistent vegetative state or die. Furthermore, increased brain tissue PCO2 ( 60 mm Hg) appears to be a useful clinical indicator of critical cerebral ischemia, especially when accompanied by increased lactate concentrations.

KEY WORDS traumatic brain injury cerebral acidbase homeostasis carbon dioxide lactate/ischemia multimodal monitoring

head injury is one of the principal causes of death and disability in Western society, especially among younger people. Annually, approximately 1.5 million US citizens sustain a mild, moderate, or severe TBI; 50,000 people die, 230,000 people are hospitalized and survive, and an estimated 80,000 to 90,000 people experience the onset of long-term disability as a result of these injuries.43 Despite tremendous progress made in the area of brain trauma research, our knowledge about the exact mechanisms and biochemical cascades that lead from trauma to the development of secondary brain injury is still poormany pieces of the puzzle remain unsolved. Relatively well-established factors initiating secondary brain damage include posttraumatic reduced CBF, hypoxia, and excessive release of excitatory amino acids.7,8,17,30 All

EVERE

these mechanisms may result in the development of cerebral acidosis, mainly due to the failure of oxidative energy metabolism.22,28,40,41 Tissue acidosis in turn has been associated with the formation of intracellular edema, increased [Ca ]i concentrations, enzyme inactivation, mitochondrial impairment, and free radical formation.41,42 Nevertheless, although brain tissue acidosis has been well examined and documented after stroke and global brain ischemia, comparatively little information is available on this condition following human TBI.15,22,29,49 The goal of the present study, one of the first continuous assessments of these parameters in a large cohort of patients with severe TBI, was therefore to relate the time course of cerebral acidbase homeostasis (brain tissue pH, PCO2, and lactate values) to good and bad outcomes, early after a severe head injury. Clinical Material and Methods The investigational protocol of these studies was approved by the Committee for Conduct of Human Research of VCUHS and Virginia Commonwealth University. Writ597

Abbreviations used in this paper: ATP = adenosine 3 ,5 -triphosphate; CBF = cerebral blood flow; CPP = cerebral perfusion pressure; ECF = extracellular fluid; GOS = Glasgow Outcome Scale; ICP = intracranial pressure; ICU = intensive care unit; SEM = standard error of the mean; TBI = traumatic brain injury VCUHS = Virginia Commonwealth University Health System.

J. Neurosurg. / Volume 103 / October, 2005

T. Clausen, et al.
TABLE 1 Characteristics in 151 patients with severe TBI*
No. of Patients GOS Score Total Male Female Mean Age (yrs) SD Range

5 (good recovery) 4 (moderate disability) 3 (severe disability) 2 (persistent vegetative state) 1 (death) total

27 24 37 20 43 151

19 18 26 18 32 113

8 6 11 2 11 38

30.3 28.9 30.8 39.6 45.4 35.5

10.2 9.4 9.0 18.0 23.9 16.9

1753 1456 1654 1678 1090 1090

potension was treated aggressively by delivery of positive inotropic drugs and vasopressors; if arterial hypotension was caused by either impaired cardiac contractility or inadequate systemic vascular resistance or if hypotension was caused by a fluid deficit, an infusion of isotonic electrolyte solutions or plasma expanders was administered until normovolemia could be achieved. Furthermore, if hyperthermia was present, external cooling was applied until normothermia was obtained.
Brain Tissue Monitoring Microdialysis. A sterile microdialysis probe (CMA 20; CMA Microdialysis, Acton, MA) was introduced into cortical tissue that had no sign of an injury such as a hematoma or contusion on computerized tomography scans. The probe was usually placed in the right frontal lobe by using a special custom-made three-lumen polycarbonate bolt.49 We deliberately tried to refrain from inserting the probes into pathologically altered tissue, because the purpose of the monitoring setup was to detect secondary brain injury rather than damage directly caused by the primary injury. The skull bolt also served as an entrance for the ventriculostomy catheter and, in most cases, for a Neurotrend multiparameter sensor. Bolt placement and probe insertion as well as the introduction of the ventriculostomy catheter were performed as early as possible, either in the operating room after the initial evacuation of a mass lesion or immediately after arrival of the patient at the neuroscience ICU. The length of the active membrane of the microdialysis probe was 10 mm, with a molecular cut-off rate of 20 kD. The probe was perfused with sterile 0.9% NaCl solution at a flow rate of 2 l/minute. After a stabilization period of at least 1 hour, sampling was started to allow for an adequate equilibration between cerebral ECF and perfusate. Microdialysate samples were collected every 30 minutes into sealed glass tubes by using an automated refrigerated (4C) collector system (Honeycomb; Bioanalytical Systems, West Lafayette, IN). Cerebral ECF concentrations of lactate were analyzed by performing enzymic fluorometric assays on a CMA 600 Microdialysis Analyzer (CMA/Microdialysis, Solna, Sweden). The results were transferred to a relational database (Microsoft Access; Microsoft, Redmond, WA) for further analysis. Throughout this article cerebral lactate concentrations are presented as concentrations measured in the microdialysate without corrections for probe recovery rates. Neurotrend Sensor. The Neurotrend sensor, a minimally invasive multiparameter sensor for continuous measurement of brain tissue PO2, PCO2, pH, and temperature, was introduced via the polycarbonate skull bolt into unlesioned brain tissue. Before its insertion into the patient, the sensor was calibrated with special calibration gas mixtures in a tonometer under sterile conditions, according to the manufacturers specifications. The technical specifications of the Neurotrend sensor are shown in Table 2. All data that were obtained were digitally transferred from the Neurotrend patient data monitor to a Macintosh computer at intervals of 5 minutes, and from there into a relational database (Microsoft Access) for further analysis. No complications that could be ascribed to either the microdialysis catheter or the Neurotrend sensor were detected in these patients. J. Neurosurg. / Volume 103 / October, 2005

* SD = standard deviation.

ten informed consent was obtained from the next of kin for all patients.
Patient Population

A total of 151 patients with severe closed head injury (Glasgow Coma Scale Scores 38 at admission or within 12 hours after admission), who were admitted to the neuroscience ICU at VCUHS, were included in the study (Table 1). Medical exclusion criteria consisted of rapid neurological improvement after admission, the presence of brain death or near brain death, or any form of coagulopathy. The average age of the study participants was 35.5 16.9 years (mean standard deviation); 113 patients were male (75%) and 38 were female (25%). Patient outcome was graded according to the GOS at 3 months postinjury. At this time 27 patients (18%) demonstrated good recovery, 24 (16%) were moderately disabled, 37 (25%) suffered from severe disability, 20 (13%) remained in a persistent vegetative state, and 43 (28%) had died (Table 1). In 80 patients concurrent multimodal monitoring could be performed using microdialysis and the Neurotrend sensor system (Neurotrend; Codman, Johnson & Johnson Professional, Inc., Raynham, MA). Microdialysis measurements without Neurotrend monitoring were performed in an additional 59 patients and 12 patients were only monitored with the aid of the Neurotrend system because of delayed microdialysis probe failure. Thus, microdialysis monitoring was performed in 139 patients, whereas the Neurotrend sensor system was used in 92 patients. All patients were treated according to the VCUHS standard protocol for management of severe head injury, which emphasizes prevention of secondary brain damage and control of ICP. The mean arterial blood pressure, ICP, and CPP were monitored continuously. Intermittent arterial blood gas analysis was performed every 4 to 6 hours or when needed. Intubation and mechanical ventilation were used in all patients to maintain PaCO2 at approximately 35 mm Hg (4.7 kPa) and PaO2 at higher than 100 mm Hg (13.3 kPa). The goal of therapy was to achieve an ICP of less than 20 mm Hg by using a staircase algorithm that included sedation, analgesia, and neuromuscular paralysis; moderate head-up posture (1020); ventricular drainage; mannitol infusion; mild or moderate hypothermia; surgical decompression; and barbiturate-induced coma.13 Significant mass lesions were evacuated as soon as possible after their detection. Blood pressure management was directed to maintain the CPP higher than 70 mm Hg whenever possible. Arterial hy598

Cerebral acidbase homeostasis after severe TBI

TABLE 2 Technical specifications of the Neurotrend sensor system*
Factor Brain Tissue PO2 (mm Hg) Brain Tissue PCO2 (mm Hg) Brain Tissue pH Temp (C) Time to Respond (sec)

range of measurements accuracy (95% CI) drift per hr

10160 3.5 (1060) 10% (60110%) 0.3%

10160 3.5 (1080) 0.3%

6.87.8 0.035 (6.97.7) 0.005

1042 0.3

15

*According to the manufacturer. = not applicable.

Statistical Analysis

All clinical parameters were downloaded from the bedside monitor and logged into a VAX mainframe computer (Neurocore; VCUHS Databases) throughout the monitoring period. These data were then time-locked with the data on lactate concentrations, which were obtained from microdialysis as well as the Neurotrend data, and stored in a relational database (Microsoft Access) for further analysis. For statistical analysis the patients were grouped according to their GOS scores at 3 months after injury, into patients with good outcomes (GOS Scores 5 and 4), patients with severe disabilities (GOS Score 3), and patients with poor outcomes (GOS Scores 2 and 1). To elucidate the pathophysiological role of brain tissue pH, PCO2 and lactate, we compared the mean values of these parameters at different levels of CPP, as well as at different levels of brain tissue PO2. Because the ischemic threshold for O2 is still under dispute, we chose the following levels for brain tissue PO2 from the literature: values lower than 10 mm Hg were defined as an absolutely insufficient O2 supply; values between 10 and 20 mm Hg as a potentially insufficient O2 supply; and values higher than 20 mm Hg as a sufficient O2 supply. As for levels of CPP, we differentiated between values less than 70 mm Hg (currently considered as insufficient by the majority of authors,2,12,26,36,37 although there is some disagreement in that respect4,20), values between 70 and 90 mm Hg (the clinical standard in most neurosurgical ICUs), and values greater than 90 mm Hg. The analysis, including descriptive statistics, was performed using statistical software (StatView; Abacus Concepts, Berkeley, CA). An analysis of variance and the Fisher Protected Least Significant Difference test were performed to compare mean values within a group at several different time points. Means between different outcome groups at corresponding time points were compared using the unpaired t-test. Differences were considered statistically significant if the probability value was less than 0.05. To determine the degrees of relationship among brain tissue pH, PCO2, and lactate, both simple and multiple regression analyses were performed. Values are expressed as means SEMs unless otherwise indicated. Results
Lactate Concentrations

The mean dialysate lactate concentration of all patients during the entire monitoring period was 976 71 mol/L. During that time the mean lactate concentration was 785
J. Neurosurg. / Volume 103 / October, 2005

119 mol/L in patients with good outcomes (GOS Scores 4 and 5) and 1051 118 mol/L in patients with poor outcomes (GOS Scores 1 and 2). A continuous decline in the mean microdialysate lactate concentration of all patients from 1136 76 mol/L on Day 1 to 777 85 mol/L on Day 4 after injury was observed. The difference in mean lactate concentration between the 1st day after injury and each following day was statistically significant (p 0.05). The mean dialysate lactate concentration decreased from 951 138 mol/L on Day 1 to 672 109 mol/L on Day 2 after injury and then remained relatively stable for the rest of the monitoring period in patients with good outcomes. The difference between the 1st day after injury and the following days was not statistically significant in these patients. On the other hand, the lactate concentration decreased continuously from 1197 111 mol/L on Day 1 to 718 151 mol/L on Day 4 after injury in patients with poor outcomes. The differences in mean dialysate lactate concentration were statistically significant between the 1st day after injury and Days 3 and 4 (p 0.05). The mean lactate concentration was higher in patients who died or remained in a persistent vegetative state (GOS Scores 1 and 2, respectively) compared with patients with good recoveries and moderate disabilities (GOS Scores 5 and 4, respectively) throughout the first 4 days after the injury, even though the difference between these two groups was not statistically significant (Fig. 1). Increased lactate concentrations were observed during episodes of low CPP and/or poor brain tissue oxygenation. The highest lactate concentration was found at the lowest CPP level. When CPP was less than 70 mm Hg, the mean dialysate lactate concentration was 1027 58 mol/ L. When CPP ranged from 70 to 90 mm Hg, the lactate con27 mol/L mm Hg, and when the centration was 874 CPP was higher than 90 mm Hg, the mean lactate concentration was 910 42 mol/L. The difference in mean lactate concentration between the lowest CPP level and the medium CPP level was statistically significant (p 0.05). A similar pattern was observed when the mean lactate concentrations were compared at different levels of brain tissue PO2. The highest mean lactate concentration (910 114 mol/L) was found when the brain tissue PO2 was lower than 10 mm Hg. When the PO2 ranged from 10 to 20 mm Hg, the mean lactate concentration was 774 49 mol/L. The lowest mean lactate concentration (712 24 mol/L) was seen when the PO2 was higher than 20 mm Hg. The difference in dialysate lactate concentrations between the lowest ( 10 mm Hg) and highest ( 20 mm Hg) levels of brain tissue PO2 was statistically significant (p 0.05).
599

T. Clausen, et al.

FIG. 1. Bar graph depicting mean dialysate lactate concentrations during the first 4 days after TBI in patients with good outcomes and those with poor outcomes. The mean lactate concentrations are higher in patients who after 3 months postinjury remained in a persistent vegetative state or had died (GOS Score 2 or 1, respectively) compared with patients with good recoveries or moderate disabilities (GOS Score 5 or 4, respectively) throughout 4 days postinjury, even though the difference between the two groups is not statistically significant.

In addition, moderate elevations in dialysate lactate concentrations, which were not related to brain tissue PO2, CPP, or any other obvious hypoxic or clinical event, have been observed in some patients. These increases in lactate lasted for several hours and in a few cases for days, and were mostly accompanied by a moderate decrease in the concentration of brain tissue glucose.

The in vitro recovery rate for lactate, using this microdialysis system, was 42%.
Brain Tissue PCO2

The mean brain tissue PCO2 of all patients throughout the entire monitoring period was 48 1 mm Hg. Although pa-

FIG. 2. Bar graph demonstrating mean brain tissue PCO2 (ptiCO2) during the first 4 days after TBI. The mean brain tissue PCO2 in patients with good outcomes was 50 1 mm Hg on the 1st day after injury; this decreased to 45 1 mm Hg on Day 2 (p 0.05) and stabilized at approximately this level during the following days. The PCO2 in patients with poor outcomes was 53 3 mm Hg on Day 1 postinjury; this continuously declined to 45 1 mm Hg on Day 4.

600

J. Neurosurg. / Volume 103 / October, 2005

Cerebral acidbase homeostasis after severe TBI

FIG. 3. Graph showing the time course of brain tissue PCO2 during the first 4 days after TBI. The initially high mean brain tissue PCO2 of all patients was almost entirely due to patients with poor outcomes, whereas the mean PCO2 in patients with good outcomes was within the normal range throughout the monitoring period, despite slightly higher values during the first 24 hours after injury. Brain tissue PCO2 was significantly higher in patients with poor outcomes compared with patients with good outcomes at 6 hours after injury (71 10 and 47 3 mm Hg, respectively; p 0.05).

tients with good outcomes (GOS Scores 4 and 5) had a mean brain tissue PCO2 of 46 1 mm Hg, a mean brain tissue PCO2 of 52 3 mm Hg was seen in patients with poor outcomes (GOS Scores 1 and 2). The mean brain tissue PCO2 of all patients was significantly higher during the first 24 hours after the injury compared with all the remaining monitoring days. It decreased from 51 1 mm Hg on Day 1 postinjury to 46 1 mm Hg on Day 2, and remained at approximately that value for the rest of the monitoring period. The mean brain tissue PCO2 in patients with good outcomes was 50 1 mm Hg on Day 1 postinjury, decreased to 45 1 mm Hg on Day 2 (p 0.05), and stabilized at approximately this level over the following days. The brain tissue PCO2 in patients with poor outcomes was 53 3 mm Hg on Day 1 postinjury and continuously decreased to 45 1 mm Hg on Day 4 (Fig. 2). When we looked at the time course of brain tissue PCO2 in more detail, we found that the mean PCO2 of all patients reached its maximum value 6 hours after injury (the earliest time of monitoring), continuously declined for approximately 24 hours, and stabilized for the rest of the monitoring period. When we categorized patients by outcome, we found that this initial increase in PCO2 was almost entirely due to values obtained in patients who had died or remained in a persistent vegetative state (poor outcome; GOS Scores 1 and 2, respectively), whereas the mean PCO2 in patients with good recoveries or moderate disabilities (good outcome; GOS Scores 5 and 4, respectively) was within the normal range, as reported previously.10,16 Brain tissue PCO2 was higher in patients with poor outcomes than in patients with good outcomes almost throughout the entire monitoring period. At 6 hours postinjury the difference between these two outcome groups was significant (71 10 mm Hg compared with 47 3 mm Hg, p 0.05; Fig. 3).
J. Neurosurg. / Volume 103 / October, 2005

Moreover, brain tissue PCO2 was different at different levels of CPP. The mean brain tissue PCO2 was 50 1 mm Hg at CPP values lower than 70 mm Hg, 46 1 mm Hg at CPP values between 70 and 90 mm Hg, and 45 1 mm Hg at CPP values higher than 90 mm Hg. The difference between the lowest CPP level and the two higher levels was 0.0001). Furthermore, statistically highly significant (p considerable increases in brain tissue PCO2 were observed during periods of drastically decreased CPP. When the CPP decreased to lower than 40 mm Hg we observed a mean brain tissue PCO2 of 56 8 mm Hg. Episodes of increased PCO2 frequently coincided with episodes of poor brain tissue oxygenation. When brain tissue PO2 decreased to less than 10 mm Hg, the mean PCO2 was 49 1 mm Hg. In contrast, when the PO2 was between 10 and 20 mm Hg or higher than 20 mm Hg the mean PCO2 was 45 0.5 mm Hg or 45 0.2 mm Hg, respectively. The difference in mean brain tissue PCO2 between the lowest brain tissue PO2 level (10 mm Hg) and each of the two higher PO2 levels (1020 mm Hg and 20 mm Hg) was statistically highly significant (p 0.0005). Arterial PCO2 remained stable throughout the monitoring period. No relevant differences in mean PaCO2 between the various outcome groups were observed during the first 4 days after injury.
Brain Tissue pH

The mean brain tissue pH of all monitored patients was 7.13 0 for the entire monitoring period. In patients with good outcomes (GOS Scores 4 and 5) the mean pH was 7.13 0.01, whereas in patients with poor outcomes (GOS Scores 1 and 2) it was 7.12 0.01. The mean brain tissue pH of all patients was significantly lower on the 1st day after injury, compared with the following 3 days. It increased
601

T. Clausen, et al.

FIG. 4. Graph showing the time course of brain tissue pH during the first 4 days after TBI. The lowest mean brain tissue pH (ptiH) is found in patients with poor outcomes immediately postinjury. Approximately 21 hours postinjury the pH is similar in both outcome groups.

from 7.06 0.01 on Day 1 postinjury to 7.14 0.01 on Day 2 and further to 7.16 0.01 on both Days 3 and 4 (p 0.01). In patients with good outcomes the mean brain tissue pH increased from 7.08 0.02 on Day 1 postinjury to 7.13 0.01 on Day 2 and 7.16 0.01 on Day 3; it remained approximately at this level on Day 4 as well. Differences in the mean pH of patients with good outcomes between the 1st day after injury and the following days were statistically significant (p 0.01). In patients with poor outcomes the mean brain tissue pH was also significantly worse on the 1st day after injury compared with the following days. In these patients the pH increased from 7.02 0.02 on Day 1 postinjury to 7.15 0.01 on Day 2; it remained at that level for the rest of the monitoring period (p 0.01). Although brain tissue pH was considerably lower in patients with poor outcomes compared with patients with good outcomes on Day 1 postinjury, the difference between both groups was not statistically significant. When we looked at the time course of brain tissue pH in more detail, we found that the lowest mean pH occurred immediately after injury in patients with poor outcomes. Approximately 21 hours after injury the brain tissue pH was similar in both outcome groups (Fig. 4). In all patients the brain tissue pH was significantly inversely correlated with the brain tissue PCO2 (r = 0.53, p 0.0001) and, to a lesser degree, with the dialysate lactate concentration (r = 0.34, p 0.0001). The strength of the relationships between brain tissue pH and brain tissue CO2 and lactate, respectively, differed greatly between the different outcome groups. In patients with good recoveries and moderate disabilities (good outcome, GOS Scores 5 and 4, respectively) the correlation between brain tissue pH and PCO2 was relatively weak (r = 0.30, p 0.0001) and the correlation between brain tissue pH and lactate was even weaker (r = 0.13, p 0.01). In patients who died or remained in a persistent vegetative state (poor outcome, GOS Scores 1 and 2, respectively), however, brain tissue pH and PCO2 were correlated to a considerably stronger degree (r =
602

0.70, p 0.0001), whereas the correlation between brain tissue pH and lactate (r = 0.34, p 0.0001) was also higher than that found in patients with good outcomes. When the combined influence of brain tissue PCO2 and lactate concentration on brain tissue pH was examined, the same phenomenon was observed. Overall, in brain tissue the pH was inversely correlated with the PCO2 and lactate (r = 0.59, p 0.0001). This correlation was considerably stronger in patients with poor outcomes (r = 0.80, p 0.0001) compared with patients with good outcomes (r = 0.38, p 0.0001). Brain tissue pH varied substantially at the different levels of CPP, with significantly decreased pH values during episodes of low CPP. The lowest brain tissue pH (7.06 0.02) was found at CPP values lower than 70 mm Hg; at CPP values between 70 and 90 mm Hg, a mean pH of 7.13 0.01 was observed, whereas at values higher than 90 mm Hg, the mean pH was 7.17 0.01. The differences among all three levels were statistically highly significant (p 0.0001). A very similar pattern was seen when pH was analyzed for different levels of brain tissue oxygenation. The lowest pH was observed at the worst PO2 level and the highest pH was seen at the highest level of brain tissue oxygenation: when the PO2 was lower than 10 mm Hg, an average 0.02 was seen; when the PO2 brain tissue pH of 7.02 ranged from 10 to 20 mm Hg the pH was 7.09 0.01; and when the PO2 was higher than 20 mm Hg the mean pH was 7.16 0 (Fig. 5). Differences in the mean brain tissue pH between the lowest PO2 level ( 10 mm Hg) and each of the two higher levels (p 0.0001) as well as the difference between the medium PO2 level (1020 mm Hg) and the highest level ( 20 mm Hg) (p 0.001) were highly statistically significant. The mean arterial pH of all patients was 7.43 0, and this value remained stable throughout the monitoring period. In patients with good outcomes the arterial pH was 7.43 0, with no significant changes over time. The mean arterial pH in patients with poor outcomes was 7.42 0, and this demonstrated a slightly decreasing trend during
J. Neurosurg. / Volume 103 / October, 2005

Cerebral acidbase homeostasis after severe TBI

FIG. 5. Bar graph depicting mean brain tissue pH at different levels of brain tissue oxygenation after TBI. Brain tissue pH varies substantially between different levels of brain tissue oxygenation. The lowest pH is observed at the worst brain tissue PO2 (ptiO2) level and the highest pH is seen at the best brain tissue oxygenation: when the PO2 is below 10 mm Hg the mean brain tissue pH is 7.02 0.02; when the PO2 ranges from 10 to 20 mm Hg the pH is 7.09 0.01; and when the PO2 is higher than 20 mm Hg the mean pH is 7.16 0. The differences in mean brain tissue pH between the lowest PO2 level ( 10 mm Hg) and each of the two higher levels (p 0.0001) as well as the difference between the medium (1020 mm Hg) and highest ( 20 mm Hg) level of brain tissue PO2 (p 0.001) is highly significant.

the first 4 days after injury (from 7.43 0.01 on Day 1 to 7.39 0.01 on Day 4). No relevant correlations between brain tissue pH and the arterial pH or PCO2 were observed.
Influence of Different CPP Levels on Brain Tissue Oxygen Tension

The mean brain tissue PO2 of all monitored patients 1 mm throughout the entire monitoring period was 29 Hg. The mean PO2 in patients with good outcomes (GOS Scores 4 and 5) was 30 2 mm Hg, whereas the mean PO2 in patients with poor outcomes (GOS Scores 1 and 2) was 27 2 mm Hg. As would be expected, brain tissue PO2 varied considerably between different levels of CPP. At CPP values lower than 70 mm Hg the mean PO2 was 26 2 mm Hg, at CPP values from 70 to 90 mm Hg it was 31 1 mm Hg, and at CPP values higher than 90 mm Hg it was 32 1 mm Hg. The difference between the lowest CPP level and each of the two higher levels was statistically highly significant (p 0.001). Discussion Brain tissue acidosis has been reported to mediate brain damage after traumatic and ischemic injuries.23,24,28,29,34,45 A low tissue pH is believed to contribute to edema formation, putatively due to the activation of an Na /H antiporter system coupled with a passive Cl /HCO3 antiporter, which would enable an increased inflow of Na , Cl , and osmotically obligated water into the cell during acidosis.11,42 More importantly, acidosis may interfere with mitochondrial ATP generation, directly by promoting the opening of the mitochondrial permeability transition pore and indirectly by
J. Neurosurg. / Volume 103 / October, 2005

dislodging Ca from its protein-binding sites, thus increasing free [Ca ]i levels.25,41 Increased [Ca ]i concentrations in turn can impair mitochondrial metabolism. Furthermore, the formation of free radicals, one of the principal pathomechanisms leading to delayed cell death after TBI, has been shown to be enhanced at lower pH values.33 Several factors contribute to the regulation of brain tissue pH: 1) brain tissue PCO2 (via the carbonic anhydrase reaction); 2) physiological buffer systems (in the case of cerebral interstitial fluid, primarily the bicarbonate system); and 3) the metabolic production of acidic substances (chiefly lactate and unbuffered H ). The interrelationship between brain tissue PCO2 and pH will be discussed in greater detail later in this section. Although the roles of PCO2 and the bicarbonate buffer system in the regulation of cerebral pH are closely related (H HCO3 CO2 H2O), our knowledge about the significance of other buffer systems in cerebral interstitial fluid is rather limited. By far the most important acid generated by cerebral metabolism is lactic acid. Increased lactate production causes a decline in brain tissue pH when the physiological buffer systems are exhausted. An elevated generation of lactate in turn has been reported after TBI as a product of anaerobic glycolysis during cerebral ischemia or hypoxia; as a result of a failure in mitochondrial metabolism, even during preserved CBF and oxygenation; and during compensatory glycolysis following a massive posttraumatic release of glutamate and the attempt to restore disturbed ionic homeostasis.14,22,23,27,32,39,44,49 All these pathomechanisms may contribute, alone or together, to lactate-mediated brain tissue acidosis after TBI. In the present study, the lactate increases observed in the cerebral ECF of patients with severe head injuries may be attributed to several mechanisms. The lactate peak observed during the first 24 hours postinjury is most likely due to a
603

T. Clausen, et al.

FIG. 6. Possible mechanisms responsible for increases in brain tissue lactate after TBI. The posttraumatic increases in lactate seen in the cerebral ECF in severely head injured patients in the present study may be attributed to a variety of mechanisms. Glycolytic activity increases due to the need to restore the posttraumatically disturbed ionic homeostasis by means of the ATP-dependent Na /K exchanger. Excitatory amino acids such as glutamate, which are massively released after TBI, are removed from the extracellular space by astrocytes via an Na -dependent cotransport; the ensuing rise in the intracellular Na concentration activates Na /K -ATPase, thus stimulating glycolysis and lactate production in astrocytes. (Glutamate is then converted to glutamine and shuttled back to the neurons.) Severe hypoxic events, such as cerebral ischemia or hypoxia, may result in a substantial accumulation of brain tissue lactate after TBI. Mitochondrial impairment after TBI may instigate an increase in glycolytic activity, even in the absence of concurrent ischemia or hypoxia. Acidosis itself may, in turn, promote the formation of mitochondrial permeability transition pores with a subsequent failure in mitochondrial metabolism, thus initiating a vicious cycle. Gln = glutamine; Glt = glutamate; Lac = lactate; MPT = mitochondrial permeability transition.

combination of different mechanisms. 1) There is increased compensatory astrocytic glycolysis to restore posttraumatically disturbed ionic homeostasis.22,46 2) Excitatory amino acids, chiefly glutamate, are removed from the extracellular space by astrocytes by means of an Na -dependent cotransport; the resulting increase in the intracellular concentration of Na cotransported with glutamate by the electrogenic uptake system activates Na /K ATPase, thus stimulating glycolysis and lactate production in astrocytes.27,32,35,39 3) Lactate and H are accumulated after early hypoxic events (that is, anaerobic glycolysis). The last mechanism seems to be far more significant in patients with poor outcomes, in whom not only higher lactate levels but also a more profound tissue acidosis combined with a worse brain tissue oxygenation were regularly displayed in the present study. This tight linkage between lactate and tissue pH also indicates that anaerobic glycolysis is most likely also responsible for the increases in brain tissue lactate during severe
604

hypoxic events, such as major increases observed in ICP and prior to herniation and brain death.44,49 In these circumstances, elevated levels of lactate are typically accompanied by tissue acidosis and poor brain tissue oxygenation. Moderate elevations in brain tissue lactate concentrations, however, which were not related to any obvious hypoxic or other clinical event and were not accompanied by severe tissue acidosis or impaired oxygenation, might also be attributed to posttraumatic hyperglycolysis, which has been shown to occur as a primary metabolic event in human TBI, even in the absence of concurrent ischemia or hypoxia (Fig. 6).1,5,
19,22,35

Compared with the continuously increasing number of reports about the importance of brain tissue PO2, relatively little is known about the pathophysiological role of brain tissue PCO2 after TBI. Charbel and colleagues9 described elevated brain tissue CO2 levels, with a concomitant drop in brain tissue pH, in patients suffering from severe vasospasm
J. Neurosurg. / Volume 103 / October, 2005

Cerebral acidbase homeostasis after severe TBI

after aneurysmal subarachnoid hemorrhage. Increases in brain tissue CO2 accompanied by a concomitant decrease in brain tissue pH and PO2 have also been reported after experimental middle cerebral artery occlusion in a feline model and after temporary clipping of major cerebral arteries during aneurysm surgery in patients.16,21,47 The same constellation, a significant rise in brain tissue PCO2 with a parallel decline in brain tissue pH and PO2, was observed immediately after brain trauma in a feline fluid-percussion-injury model by our group recently.48 Furthermore, brain tissue CO2 levels up to 150 mm Hg have been described during cerebral circulatory arrest and brain death after human TBI.49 In the present study a significant increase in PCO2, with a concomitant drop in brain tissue pH, was found in patients with poor outcomes within 6 hours postinjury. Moreover, increases in brain tissue PCO2 were seen during severe hypoxic episodes, especially when these were due to insufficient CPP. In general, four main reasons for an increase in brain tissue PCO2 are theoretically possible: 1) an increase in arterial CO2 with consecutive diffusion into the brain; 2) an increase in cerebral CO2 production; 3) a disturbance in the CO2 clearance from the tissue; and 4) profound tissue acidosis with an increase in CO2 generation via the carbonic anhydrase reaction (H HCO3 CO2 H2O). The first reason is not significant in this study. It is important to note that arterial PCO2 was considerably lower than brain tissue PCO2 in all patients, and remained constant throughout the monitoring period because of the use of vigorously controlled end-tidal CO2directed ventilation. An increase in cerebral CO2 production is theoretically possible if the turnover rate of oxidative glucose degradation rises considerably, as in the Krebs cycle. Nevertheless, the results of several studies in animals and humans have demonstrated a major decline in cerebral O2 consumption after TBI.3,18,31,38 Moreover, in the present study the increases observed in brain tissue PCO2 were regularly accompanied by a drop in brain tissue PO2. Given that decarboxylation reactions (for example, decarboxylation of pyruvate, isocitrate, and -ketoglutarate before and during the tricarboxylic acid cycle) as part of oxidative glucose metabolism and ATP generation are the main metabolic source of cerebral CO2 production, and because these reactions are inhibited in the absence of sufficient O2 by increased concentrations of the reduced form of nicotinamide adenine dinucleotide, a decrease in CO2 production, instead of a rise, would be expected. Therefore, a rise in normal physiological CO2 production is not likely to be the source of the increased brain tissue PCO2 observed in our patients. The principle way to eliminate CO2 from brain tissue is diffusion into cerebral capillaries and removal by CBF, as long as there is a PCO2 gradient between brain tissue and blood. If blood flow decreases, the gradient, and thus the driving force for CO2 elimination from the tissue, will also decrease as less CO2 is quantitatively removed by cerebral perfusion per time unit. If relative or complete ischemia occurs, CO2 accumulates in the brain tissue, thereby creating a new diffusion equilibrium between brain tissue and blood in the affected area. Because significant ischemia accompanies a critical reduction in the O2 supply, pathological CO2 production will decrease at some point. A complete cessation of oxidative CO2 production, however, should occur only during absolute anoxia. Concomitantly, the anaerobic
J. Neurosurg. / Volume 103 / October, 2005

lactate generation will increase as long as glucose is available as a substrate. The ensuing lactate acidosis may further contribute to an increase in brain tissue PCO2 levels because, under these conditions, more bicarbonate is converted into CO2 and H2O in an attempt to restore the physiological pH milieu (H HCO3 CO2 H2O). Given that cerebral perfusion has been demonstrated to be critically impaired in a large number of severely head injured patients with poor outcomes immediately posttrauma,6 cerebral ischemia may be a key factor responsible for the large early increase seen in brain tissue PCO2 in patients in this study who died or remained in a persistent vegetative state. Moreover, the fact that brain tissue PCO2 usually normalized within 12 hours postinjury indicates restoration of CBF, and accords with the observations of Bouma and colleagues,6 who found global and regional cerebral ischemia during the first 4 to 12 hours in approximately 34% of patients with severe TBI by using the stable xenonenhanced computerized tomography method. This, as well as the fact that the highest levels of brain tissue CO2 were seen during episodes of critically decreased CPP in the present study, lends further support to the thesis that the initial increase in PCO2 in these patients was caused mainly by insufficient CO2 clearance due to impaired CBF. Given that severe tissue acidosis was present at the same time in patients with poor outcomes, one might argue that the increase in brain tissue PCO2 may be a result of acidosis rather than ischemia. Nevertheless, the two main measurable factors influencing brain tissue pH in these patients were brain tissue PCO2 and extracellular lactate concentration. Of these two, PCO2 was by far the more important factor in patients with poor outcomes. Furthermore, lactate concentrations only started to rise at this time, whereas PCO2 reached its highest values at 6 hours after the trauma. Therefore, we speculate that the initially low brain tissue pH seen in patients with poor outcomes was the result of increased brain tissue PCO2 rather than its cause. In the present study, brain tissue CO2 recovered considerably faster from the initial derangement than brain tissue pH did. During normalization of PCO2, however, the cerebral concentration of lactate continued to rise, peaking approximately 15 to 18 hours after the injury. Therefore we assume that the initial drop in brain tissue pH was chiefly due to the early increase in PCO2, whereas the prolonged recovery of pH resulted, in large part, from the delayed increase in lactate concentrations. Apart from the initial increase in brain tissue PCO2 in patients with poor outcomes, elevated values of brain tissue PCO2 were also observed during severe hypoxic events, in which they were typically associated with decreased brain tissue PO2 and pH and increased extracellular lactate concentrations. Whether the increases in brain tissue PCO2 seen during such a hypoxic event were directly due to ischemia or a consequence of tissue lactate acidosis, is difficult to differentiate. The fact that these PCO2 increases were more often observed when the hypoxic event was due to insufficient CPP, at least indicates a direct influence of ischemia as a pathogenetic factor. It is probable that both tissue acidosis and ischemia may contribute to varying degrees to an increase in brain tissue PCO2 under such circumstances. Conclusions The results of the present study demonstrate that brain
605

T. Clausen, et al.
tissue acidosis occurs during the first hours after severe head injury in humans and that disturbances in acidbase homeostasis are far more pronounced in patients who die or remain in a persistent vegetative state. The main causes for this acidosis were increases in brain tissue PCO2, most likely as a result of an insufficient CO2 clearance due to impaired cerebral micro- or macrocirculation, and elevated lactate concentrations, probably caused by increased glycolytic activity. Interestingly, the changes in cerebral pH depend to a considerably greater degree on changes in brain tissue CO2 and lactate in patients with poor outcomes compared with patients with good outcomes. Our findings also indicate that a significant rise in brain tissue PCO2 may be indicative of critical cerebral ischemia. Because high diffusibility is an important property of the CO2 molecule, however, it can disperse into neighboring capillaries with relative ease. Therefore, a disturbance in CBF needs to affect a larger area of brain tissue before an impairment in CO2 removal by the blood flow may be consequential and detectable, in comparison with brain tissue PO2. We postulate that an increase in brain tissue PCO2 ( 60 mm Hg), especially when it is associated with a decrease in brain tissue PO2 and/or an increase in cerebral lactate concentration, is an indicator of severe ongoing cerebral ischemia. Furthermore, monitoring of brain tissue PCO2, pH, and lactate, along with brain tissue PO2, constitutes a useful tool to assist in the individual optimization of CPP and O2 delivery to ensure a sufficient CBF and optimized metabolic recovery in patients with severe head injury.
References 1. Alessandri B, Doppenberg E, Zauner A, Woodward J, Choi S, Bullock R: Evidence for time-dependent glutamate-mediated glycolysis in head-injured patients: a microdialysis study. Acta Neurochir Suppl (Wien) 75:2528, 1999 2. Anonymous: The Brain Trauma Foundation. The American Association of Neurological Surgeons. The Joint Section on Neurotrauma and Critical Care. Guidelines for cerebral perfusion pressure. J Neurotrauma 17:507511, 2000 3. Artru F, Philippon B, Flachaire E, Peyrieux JC, Boissel JP, Ferry S, et al: A controlled study of Dextran 40: effect on cerebral blood flow and metabolic rates in acute head trauma. Intensive Care Med 15:499504, 1989 4. Asgeirsson B, Grande PO, Nordstrom CH: A new therapy of posttrauma brain oedema based on haemodynamic principles for brain volume regulation. Intensive Care Med 20:260267, 1994 5. Bergsneider M, Hovda DA, Shalmon E, Kelly DF, Vespa PM, Martin NA, et al: Cerebral hyperglycolysis following severe traumatic brain injury in humans: a positron emission tomography study. J Neurosurg 86:241251, 1997 6. Bouma GJ, Muizelaar JP, Stringer WA, Choi SC, Fatouros P, Young HF: Ultra-early evaluation of regional cerebral blood flow in severely head-injured patients using xenon-enhanced computerized tomography. J Neurosurg 77:360368, 1992 7. Bullock R, Zauner A, Myseros JS, Marmarou A, Woodward JJ, Young HF: Evidence for prolonged release of excitatory amino acids in severe human head trauma. Relationship to clinical events. Ann NY Acad Sci 765:290298, 1995 8. Bullock R, Zauner A, Woodward JJ, Myseros J, Choi SC, Ward JD, et al: Factors affecting excitatory amino acid release following severe human head injury. J Neurosurg 89:507518, 1998 9. Charbel FT, Du X, Hoffman WE, Ausman JI: Brain tissue PO(2), PCO(2), and pH during cerebral vasospasm. Surg Neurol 54: 432438, 2000 10. Charbel FT, Hoffman WE, Misra M, Hannigan K, Ausman JI: Cerebral interstitial tissue oxygen tension, pH, HCO3, CO2. Surg Neurol 48:414417, 1997 11. Chesler M: Regulation and modulation of pH in the brain. Physiol Rev 83:11831221, 2003 12. Chesnut RM: Secondary brain insults after head injury: clinical perspectives. New Horiz 3:366375, 1995 13. Clausen T, Bullock R: Medical treatment and neuroprotection in traumatic brain injury. Curr Pharm Des 7:15171532, 2001 14. Clausen T, Zauner A, Levasseur JE, Rice AC, Bullock R: Induced mitochondrial failure in the feline brain: implications for understanding acute post-traumatic metabolic events. Brain Res 908: 3548, 2001 15. DeSalles AA, Kontos HA, Ward JD, Marmarou A, Becker DP: Brain tissue pH in severely head-injured patients: a report of three cases. Neurosurgery 20:297301, 1987 16. Doppenberg EM, Watson JC, Broaddus WC, Holloway KL, Young HF, Bullock R: Intraoperative monitoring of substrate delivery during aneurysm and hematoma surgery: initial experience in 16 patients. J Neurosurg 87:809816, 1997 17. Gentleman D: Causes and effects of systemic complications among severely head injured patients transferred to a neurosurgical unit. Int Surg 77:297302, 1992 18. Glenn TC, Kelly DF, Boscardin WJ, McArthur DL, Vespa P, Oertel M, et al: Energy dysfunction as a predictor of outcome after moderate or severe head injury: indices of oxygen, glucose, and lactate metabolism. J Cereb Blood Flow Metab 23:12391250, 2003 19. Goodman JC, Valadka AB, Gopinath SP, Uzura M, Robertson CS: Extracellular lactate and glucose alterations in the brain after head injury measured by microdialysis. Crit Care Med 27: 19651973, 1999 20. Grande PO, Asgeirsson B, Nordstrom C: Aspects on the cerebral perfusion pressure during therapy of a traumatic head injury. Acta Anaesthesiol Scand Suppl 110:3640, 1997 21. Hoffman WE, Charbel FT, Gonzalez-Portillo G, Ausman JI: Measurement of ischemia by changes in tissue oxygen, carbon dioxide, and pH. Surg Neurol 51:654658, 1999 22. Hovda DA, Becker DP, Katayama Y: Secondary injury and acidosis. J Neurotrauma 9 (Suppl 1):S47S60, 1992 23. Kalimo H, Rehncrona S, Soderfeldt B, Olsson Y, Siesjo BK: Brain lactic acidosis and ischemic cell damage: 2. Histopathology. J Cereb Blood Flow Metab 1:313327, 1981 24. Kraig RP, Petito CK, Plum F, Pulsinelli WA: Hydrogen ions kill brain at concentrations reached in ischemia. J Cereb Blood Flow Metab 7:379386, 1987 25. Kristian T, Bernardi P, Siesjo BK: Acidosis promotes the permeability transition in energized mitochondria: implications for reperfusion injury. J Neurotrauma 18:10591074, 2001 26. Lang EW, Chesnut RM: Intracranial pressure and cerebral perfusion pressure in severe head injury. New Horiz 3:400409, 1995 27. Magistretti PJ, Pellerin L, Rothman DL, Shulman RG: Energy on demand. Science 283:496497, 1999 28. Marmarou A: Intracellular acidosis in human and experimental brain injury. J Neurotrauma 9 (Suppl 2):S551S562, 1992 29. Marmarou A, Holdaway R, Ward JD, Yoshida K, Choi SC, Muizelaar JP, et al: Traumatic brain tissue acidosis: experimental and clinical studies. Acta Neurochir Suppl (Wien) 57:160164, 1993 30. Miller JD: Head injury and brain ischaemiaimplications for therapy. Br J Anaesth 57:120130, 1985 31. Muizelaar JP, Marmarou A, DeSalles AA, Ward JD, Zimmerman RS, Li Z, et al: Cerebral blood flow and metabolism in severely head-injured children. Part 1: Relationship with GCS score, outcome, ICP, and PVI. J Neurosurg 71:6371, 1989 32. Pellerin L, Magistretti PJ: Glutamate uptake into astrocytes stimulates aerobic glycolysis: a mechanism coupling neuronal activity to glucose utilization. Proc Natl Acad Sci USA 91: 1062510629, 1994

606

J. Neurosurg. / Volume 103 / October, 2005

Cerebral acidbase homeostasis after severe TBI

33. Rehncrona S, Hauge HN, Siesjo BK: Enhancement of iron-catalyzed free radical formation by acidosis in brain homogenates: differences in effect by lactic acid and CO2. J Cereb Blood Flow Metab 9:6570, 1989 34. Rehncrona S, Rosen I, Siesjo BK: Brain lactic acidosis and ischemic cell damage: 1. Biochemistry and neurophysiology. J Cereb Blood Flow Metab 1:297311, 1981 35. Reinert M, Hoelper B, Doppenberg E, Zauner A, Bullock R: Substrate delivery and ionic balance disturbance after severe human head injury. Acta Neurochir Suppl 76:439444, 2000 36. Rosner MJ, Coley IB: Cerebral perfusion pressure, intracranial pressure, and head elevation. J Neurosurg 65:636641, 1986 37. Rosner MJ, Daughton S: Cerebral perfusion pressure management in head injury. J Trauma 30:933941, 1990 38. Schmoker JD, Zhuang J, Shackford SR: Hemorrhagic hypotension after brain injury causes an early and sustained reduction in cerebral oxygen delivery despite normalization of systemic oxygen delivery. J Trauma 32:714722, 1992 39. Schurr A, Miller JJ, Payne RS, Rigor BM: An increase in lactate output by brain tissue serves to meet the energy needs of glutamate-activated neurons. J Neurosci 19:3439, 1999 40. Siesjo BK: Basic mechanisms of traumatic brain damage. Ann Emerg Med 22:959969, 1993 41. Siesjo BK: Pathophysiology and treatment of focal cerebral ischemia. Part I: Pathophysiology. J Neurosurg 77:169184, 1992 42. Siesjo BK: Pathophysiology and treatment of focal cerebral ischemia. Part II: Mechanisms of damage and treatment. J Neurosurg 77:337354, 1992 43. Thurman DJ, Alverson C, Dunn KA, Guerrero J, Sniezek JE: Traumatic brain injury in the United States: a public health perspective. J Head Trauma Rehabil 14:602615, 1999 44. Valadka AB, Goodman JC, Gopinath SP, Uzura M, Robertson CS: Comparison of brain tissue oxygen tension to microdialysisbased measures of cerebral ischemia in fatally head-injured humans. J Neurotrauma 15:509519, 1998 Welsh FA, Ginsberg MD, Rieder W, Budd WW: Deleterious effect of glucose pretreatment on recovery from diffuse cerebral ischemia in the cat. II. Regional metabolite levels. Stroke 11: 355363, 1980 Yoshino A, Hovda DA, Kawamata T, Katayama Y, Becker DP: Dynamic changes in local cerebral glucose utilization following cerebral conclusion in rats: evidence of a hyper- and subsequent hypometabolic state. Brain Res 561:106119, 1991 Zauner A, Bullock R, Di X, Young HF: Brain oxygen, CO2, pH, and temperature monitoring: evaluation in the feline brain. Neurosurgery 37:11681177, 1995 Zauner A, Clausen T, Alves OL, Rice A, Levasseur J, Young HF, et al: Cerebral metabolism after fluid-percussion injury and hypoxia in a feline model. J Neurosurg 97:643649, 2002 Zauner A, Doppenberg EM, Woodward JJ, Choi SC, Young HF, Bullock R: Continuous monitoring of cerebral substrate delivery and clearance: initial experience in 24 patients with severe acute brain injuries. Neurosurgery 41:10821093, 1997

45.

46.

47. 48. 49.

Manuscript received October 12, 2004. Accepted in final form May 16, 2005. This study was funded by the Lind-Lawrence Foundation and by Grant No. NS12587 from the National Institute of Neurological Disorders and Stroke, National Institutes of Health. Address reprint requests to: M. Ross Bullock, M.D., Ph.D., Division of Neurosurgery, Virginia Commonwealth University Health System, P. O. Box 980631, Richmond, Virginia 23298-0631. email: robulloc@mail2.vcu.edu.

J. Neurosurg. / Volume 103 / October, 2005

607