- Email: [email protected]
Re-evaluation of the hypoxia theory as the mechanism of hyperventilation-induced EEG slowing
Re-Evaluation of the Hypoxia Theory as the Mechanism of Hyperventilation-Induced EEG Slowing Yoko Hoshi, MD, PhD*‡, Hiroyuki Okuhara, BS†, Shinji Nakane, BS†, Koji Hayakawa, BS†, Norio Kobayashi, MD‡, and Naofumi Kajii, MD‡ To determine whether the well-accepted hypoxia theory accounts for hyperventilation-induced electroencephalogram (EEG) slowing, the authors monitored changes in cerebral oxygenation and end-tidal concentrations of carbon dioxide in 67 patients with epilepsy (age range ⴝ 5-12 years) during the hyperventilation activation test in a routine EEG examination. Relative concentration changes in cerebral oxygenated, deoxygenated, total hemoglobin, and oxidized cytochrome oxidase were measured by near-infrared spectroscopy in the frontal region. In all patients, except one who demonstrated EEG slowing, total and oxygenated hemoglobin decreased, and cytochrome oxidase was not reduced. EEG slowing occurred intermittently in 22 patients and was not synchronous with changes in either the cerebral oxygenation or end-tidal concentration of carbon dioxide. The degree of EEG slowing was diminished or the slow waves disappeared abruptly within 1 second after the cessation of hyperventilation in 22 patients when both the cerebral oxygenation and end-tidal concentration of carbon dioxide were still at low levels. The findings during the recovery periods do not confirm the hypoxia theory. It is thus supposed that more subtle mechanisms are the cause of EEG slowing. © 1999 by Elsevier Science Inc. All rights reserved.
It is well known that hyperventilation (HV) produces electroencephalogram (EEG) slowing. Beginning in the
1940s, the mechanisms of HV-induced EEG slowing were disputed [1-3]. However, in 1960, Meyer and Gotoh [4] concluded that EEG slowing in the monkey during HV resulted from cerebral ischemic hypoxia associated with vasoconstriction caused by a lowering of Paco2 (the hypoxia theory [1]). Since then the hypoxia theory has been supported by much clinical and experimental evidence obtained under conditions of moderate or severe hypocapnia, such as Paco2 less than 25 mm Hg [5,6]. However, there is a wide range of Paco2 levels over which EEG slowing occurs [2]. For example, in young children, EEG slowing can be produced easily under mildly hypocapnic conditions, which is unusual in adults. This difference is partly the result of age-related cerebral vascular carbon-dioxide responsiveness [7], although a striking difference was also demonstrated between normal subjects and patients with epilepsy [2]. HV is commonly used as a useful activation test in a routine EEG examination of children with epilepsy. Previously the authors examined the relationship between cerebral oxygenation and EEG in rats and found that high-voltage slow-wave activity appeared on the EEG when cerebral oxygenation was reduced to an extremely low level, at which the intracellular oxygen pressure was less than 0.03 mm Hg [8]. This observation means that the HV activation test could be a harmful procedure to children with epilepsy. The degree of HV-induced EEG slowing has been reported to be independent of the extent of the decrease in cerebral blood flow (CBF) [9]. These differences are believed to result from individual brain sensitivity to hypoxia [2,9], although there is no direct evidence to support that hypothesis. There is no doubt that marked hypocapnia causes cerebral ischemic hypoxia because of vasoconstriction, resulting in EEG slowing. However, the existence of individual differences calls into
From the *Biophysics Group; Research Institute for Electronic Science; Hokkaido University; †Laboratory of Electrophysiology; Hokkaido University Hospital; and ‡Department of Pediatrics; Hokkaido University School of Medicine; Sapporo, Japan.
Communications should be addressed to: Dr. Hoshi; Biophysics Group; Research Institute for Electronic Science; Hokkaido University; Sapporo 060-0812, Japan. Received February 26, 1999; accepted May 26, 1999.
Hoshi Y, Okuhara H, Nakane S, Hayakawa K, Kobayashi N, Kajii N. Re-evaluation of the hypoxia theory as the mechanism of hyperventilation-induced EEG slowing. Pediatr Neurol 1999;21:638-643.
Introduction
638
PEDIATRIC NEUROLOGY
Vol. 21 No. 3
© 1999 by Elsevier Science Inc. All rights reserved. PII S0887-8994(99)00063-6 ● 0887-8994/99/$20.00
question the presumption that the hypoxia theory can be applied universally. It is essential for verification of the hypoxia theory to measure cerebral oxygenation. Near-infrared spectroscopy (NIRS) allows real-time and continuous measurements of changes in the hemoglobin oxygenation state, blood volume, and the redox state of cytochrome oxidase (cyt ox) in the brain. The measurement of the redox state of cyt ox provides direct information about the intracellular oxygenation state. Using NIRS the authors monitored changes in the cerebral oxygenation and end-tidal concentration of carbon dioxide (ETco2) in children with epilepsy during an HV activation test. The authors also examined cerebral vascular carbon-dioxide responsiveness in normal healthy children as a control study. The aim of this study was to examine whether the hypoxia theory actually accounts for HV-induced EEG slowing. Subjects and Methods Subjects. The subjects were 67 well-controlled outpatients with epilepsy (25 females, 42 males) aged 5-12 years who underwent routine EEG examinations. Exclusion criteria included gross cerebral structural abnormality as defined by magnetic resonance imaging or associated cerebrovascular anomalies. Patients with absence seizures were also excluded. Before the study, informed consent was obtained from the subjects’ parents and from children older than 10 years. Because simultaneous measurement of EEG, ETco2, and cerebral oxygenation is stressful, several children consented to participate in the study on the condition that only ETco2 or cerebral oxygenation was measured. Thus ETco2 was measured separately from the NIRS measurement. ETco2 was measured in 15 patients (Group 1), and cerebral oxygenation was measured in 52 patients (Group 2). The clinical characteristics of these two groups are presented in Table 1. To examine cerebral vascular carbon-dioxide responsiveness, simultaneous measurement of cerebral oxygenation and ETco2 was performed in nine healthy children (five females, four males) aged 5-12 years. Near-infrared Spectroscopy. The basic principle of the NIRS apparatus used in this study has been previously published in detail [8]. A portable apparatus was built whereby near-infrared light from a halogen lamp passed through a lens system with a rotating disc containing four interference filters (700-, 730-, 750-, and 805-nm wavelengths). The concentration changes in oxygenated hemoglobin (oxy-Hb), deoxygenated hemoglobin (deoxy-Hb), and oxidized cyt ox were calculated by the following numeric formulas every 1 second: K⌬关oxy-Hb兴 ⫽ ⫺ 0.868⌬A 700-750 ⫺ 1.74⌬A 730-750 K⌬关deoxy-Hb兴 ⫽ 0.868⌬A 700-750 ⫺ 2.26⌬A 730-750 Ka 3⬙⌬关cyt ox兴 ⫽ 0.674⌬A 700-750 ⫺ 0.752⌬A 730-750 ⫹ 0.5⌬A 805-750 K is the apparent difference absorption coefficient of either oxy-Hb or deoxy-Hb at an arbitrary wavelength pair, and a3⬙ is a proportionality factor for oxidized cyt ox [8]. Because the value of either K or a3⬙ cannot be determined experimentally, the results were expressed in relative amounts rather than in absolute values of concentration. Summation of changes in oxy-Hb and deoxy-Hb gives changes in the total hemoglobin concentration (t-Hb). Under conditions with a constant hematocrit, the changes in t-Hb can be used as an indicator of changes in cerebral blood volume (CBV) within the optical field. The optodes were placed 3 cm from each other on the middle of the forehead. To shield from ambient light the forehead was wrapped in a band. Procedures. ETco2 was continuously monitored by a capnometer (Eliza, Gambo, Tokyo). Data were plotted every 10 or 20 seconds. The EEG was recorded with monopolar electrodes (reference electrodes,
Table 1.
Clinical characteristics of the two study groups
Sex (n) Female Male Age (yr, mean ⫾ SD) Seizure type (n) GTS GTCS GTCS ⫹ CPS CPS SPS Other Antiepileptic (n) VPA CBZ CZP VPA ⫹ CBZ, CZP, and/or PH PB PB ⫹ CBZ Other
Group 1 (n ⴝ 15)
Group 2 (n ⴝ 52)
6 9 9 ⫾ 1.86
19 33 8 ⫾ 1.72
5 5 2 2 1 0
15 8 13 8 5 3
4 6 2 1 0 1 1
17 19 4 6 1 2 3
Abbreviations: CBZ ⫽ Carbamazepine CPS ⫽ Complex partial seizure CZP ⫽ Clonazepam GTCS ⫽ Generalized tonic-clonic seizure GTS ⫽ Generalized tonic seizure PB ⫽ Phenobarbital SPS ⫽ Simple partial seizure VPA ⫽ Valproic acid
auricles). Silver-silver chloride electrodes were attached to the skin according to the International 10-20 system. All measurements were performed within 2 hours after a meal to avoid hypoglycemia. The subjects were in the supine position. After a 2-minute baseline recording the subject began HV with his eyes closed. HV was performed for 4 minutes in children with epilepsy and for 3 minutes in healthy children. Subjects hyperventilated voluntarily, but when the frequency of ventilation and the depth of breathing decreased below minimal values, a supervisor instructed them to adjust the rate. Measurements of cerebral oxygenation and ETco2 were discontinued 2 minutes after the cessation of HV. EEG slowing was defined when HV-induced diffuse high-voltage slow-wave activity with disappearance of background activities continued for more than 30 seconds. EEG slowing is generally divided into four degrees according to the severity of the slowing [9]. The definition used here corresponds to the most severe slowing. Comparisons were made by Student t test.
Results Relationship Between Cerebral Oxygenation and ETCO2 During HV in Healthy Children In seven of the nine healthy children, t-Hb and oxy-Hb decreased in parallel with the decrease in ETco2; cyt ox was not reduced (Fig 1). The behavior of the change in deoxy-Hb varied with each subject: an increase, no change, and a decrease in deoxy-Hb were observed. Two children did not demonstrate decreases in either t-Hb or oxy-Hb, yet ETco2 was considerably decreased. Figure 2 presents the NIRS traces obtained from one of these two
Hoshi et al: Mechanisms of HV-Induced EEG Slowing 639
ished and did not meet the definition of EEG slowing, or the slow waves disappeared abruptly within 1 second after the cessation of HV in four patients and the ETco2 was still lower than the level at which EEG slowing had begun. Relationship Between EEG Slowing and Cerebral Oxygenation in Group 2
Figure 1. Changes in cerebral oxygenation and end-tidal carbon dioxide (ETCO2) caused by hyperventilation (HV) in a healthy child. Baselines were selected from the starting point of the measurement; these values were taken as zero for each near-infrared spectroscopy signal. Changes from baseline were represented as relative amounts. Upward and downward trends demonstrate increases and decreases in values, respectively. HV was performed between two arrows marked HV. ⌬[oxy-Hb]: change in oxygenated hemoglobin; ⌬[deoxy-Hb]: change in deoxygenated hemoglobin; ⌬[t-Hb]: change in total hemoglobin; ⌬[oxidized cyt. ox.]: change in oxidized cytochrome oxidase.
subjects. Although ETco2 decreased to about 2.5%, t-Hb and oxy-Hb increased and deoxy-Hb decreased. t-Hb and oxy-Hb decreased about 20 seconds before the cessation of HV. Although ETco2 was returning to the pre-HV level after the cessation of HV, decreases in t-Hb and oxy-Hb and an increase in deoxy-Hb were observed. The other subject demonstrated no changes in NIRS parameters during HV. After the cessation of HV, this subject also demonstrated decreases in t-Hb and oxy-Hb with an increase in deoxy-Hb; the redox state of cyt ox was not changed. Relationship Between EEG Slowing and ETCO2 in Group 1 EEG slowing was induced in seven of 15 patients. There was no difference in ETco2 in the resting period between patients with and without EEG slowing (5.1 ⫾ 0.49% vs 5.1 ⫾ 0.35%). The value of ETco2 when EEG slowing occurred varied with each patient. ETco2 at the end of HV was significantly lower in patients with EEG slowing than in those without EEG slowing (2.98 ⫾ 0.41% vs 3.89 ⫾ 0.64%, P ⬍ 0.001). EEG slowing occurred intermittently in five patients and was not synchronous with changes in ETco2 (Fig 3). The degree of EEG slowing was dimin-
640
PEDIATRIC NEUROLOGY
Vol. 21 No. 3
EEG slowing was induced in 28 of 52 patients. In all of these patients but one, decreases in t-Hb and oxy-Hb were observed. The behavior of the change in deoxy-Hb varied with each patient as it did in healthy children, and a reduction of cyt ox was not observed. Figure 4 illustrates a typical example of NIRS traces obtained from a subject with EEG slowing. In three patients, EEG slowing occurred before t-Hb and oxy-Hb decreased. Only one patient demonstrated no change in cerebral oxygenation throughout the EEG slowing (Fig 5). In this patient, t-Hb and oxy-Hb decreased and deoxy-Hb increased after cessation of HV. EEG slowing occurred intermittently in 17 patients, in whom disappearance of slow waves was not accompanied by reoxygenation of the brain (Fig 6). Attenuation of a continuous EEG slowing with brief statements, such as “I’m in pain,” was observed in three patients. This attenuation also occurred without accompanying reoxygenation of the brain. The degree of EEG slowing was diminished or the slow waves disappeared abruptly within 1 second after the cessation of HV in 18 patients, and cerebral oxygenation had not yet recovered. In seven of 24 patients without EEG slowing, decreases in t-Hb and oxy-Hb were observed during HV. Discussion Relationship Between Cerebral Oxygenation and ETCO2 in Healthy Children Cerebrovascular carbon-dioxide responsiveness varies with age [7,10], and a positive correlation between CBF (or CBV) and Paco2 has been observed over a wide age range [7,10-12]. Several NIRS studies in both humans and animals have demonstrated decreases in CBV (t-Hb) during HV [13,14]. The authors also observed that progressive decreases in ETco2 accompanied decreases in t-Hb and oxy-Hb in seven of nine healthy children. In contrast, two healthy children demonstrated no changes or, conversely, increases in t-Hb and oxy-Hb, with ETco2 considerably decreased. Performing HV is a difficult task and is accompanied by emotional stress. Previously the authors observed that emotional stress caused increases in t-Hb and oxy-Hb in the frontal region that reflect an increase in regional CBF (rCBF) responding to neuronal activation [15,16]. In addition, HV often increases alertness in a drowsy child, as a result of the increase in CBF. Thus the authors speculate that the increase in rCBF caused by emotional stress or arousal response exceeded or canceled the decrease in rCBF caused by vasoconstric-
Figure 2. Changes in cerebral oxygenation and ETCO2 caused by hyperventilation (HV) in a healthy child. In contrast to the subject in Figure 1, this subject demonstrated increases in oxy-Hb and t-Hb during HV. Abbreviations as in Figure 1.
tion in the frontal region, which resulted in no changes or increases in oxy-Hb and t-Hb, with CBF decreases in other brain regions. The decreases in t-Hb and oxy-Hb with an increase in deoxy-Hb after the cessation of HV observed in these subjects probably resulted from the combined effects of cerebral ischemia that became manifest because of the relief from the emotional stress and hypoxemia caused by suppression of breath.
Mechanisms of EEG Slowing The most prominent changes observed in the patients with EEG slowing were hypocapnia and decreases in t-Hb and oxy-Hb that seemed to be consistent with the hypoxia theory. However, one patient did not demonstrate decreases in either t-Hb or oxy-Hb, and EEG slowing was observed almost throughout the HV. In three patients,
Figure 3. Relationship between EEG slowing and ETCO2 in an epileptic patient. The bottom horizontal bars denote the occurrence of EEG slowing. EEG slowing occurred intermittently. Abbreviations as in Figure 1.
Hoshi et al: Mechanisms of HV-Induced EEG Slowing 641
Figure 4. Changes in cerebral oxygenation caused by hyperventilation (HV) in an epileptic patient with EEG slowing. The bottom horizontal bar denotes the occurrence of EEG slowing. Abbreviations as in Figure 1.
EEG slowing preceded decreases in t-Hb and oxy-Hb. Furthermore, unlike high-voltage slow-wave activity observed in rats in hypoxic hypoxia [8], HV-induced EEG slowing was not accompanied by the reduction of cyt ox. These data suggest that ischemic hypoxia in the frontal region is not essential for EEG slowing. As is presented in Figures 3 and 5, EEG slowing occurred intermittently in some subjects in whom the disappearance was not accompanied by increases in either ETco2 or cerebral oxygen-
Figure 5. Changes in cerebral oxygenation caused by hyperventilation (HV) in an epileptic patient with EEG slowing. Unlike the subject in Figure 4, this subject demonstrated intermittent EEG slowing. Abbreviations as in Figure 1.
642
PEDIATRIC NEUROLOGY
Vol. 21 No. 3
Figure 6. Changes in cerebral oxygenation caused by hyperventilation (HV) in an epileptic patient with electroencephalographic (EEG) slowing. In contrast to the subject in Figure 4, this subject demonstrated no significant change in cerebral oxygenation state during hyperventilation (HV). Abbreviations as in Figure 1.
ation. This intermittent EEG slowing might be a projected rhythm [17]. This rhythm, which can be observed in the earlier stage of hypoxia, is in general indicative of a disturbance arising from the midline structures and projecting to the surface [17]. Thus it is plausible that the central ischemic effects occur at the subcortical level and that the effects on the cortex are indirect. During the recovery periods, however, abrupt diminution of the degree of EEG slowing and disappearance of the slow waves occurred almost simultaneously with the cessation of HV in some subjects, with both cerebral oxygenation and ETco2 still at low levels. In addition the attenuation of EEG slowing with brief speaking was not accompanied by reoxygenation of the brain. These results suggest that more subtle mechanisms are significant for EEG slowing. Reviewing published reports, Patel and Maulsby [18] emphasized a similarity between the mechanism of HV-induced EEG slowing and the transition from drowsiness to slow-wave sleep, which had been originally proposed by Bonvallet and Dell [19] in the 1950s. That theory offers one possible explanation for the authors’ observations. That is, speaking and the instruction to stop HV are alerting stimuli. Such alerting stimuli might have attenuated EEG slowing without increases in cerebral oxygenation and ETco2. For a full explanation, however, further investigation is required. The results of the present study suggest that the effects of HV on the subcortical level are responsible for the EEG slowing. However, it is again unclear which factor primarily causes changes in neuronal activity at the subcortical level: hypoxia, hypocapnia, or other HV-associated changes. Direct measurements of the oxygenation state and carbon dioxide tension at the subcortical level are
necessary to answer this question. Recently the authors’ group developed optical computed tomography (CT) [20]. Unlike the conventional NIRS apparatus, optical CT provides quantitative information about the oxygenation state in the deep brain structures. Thus optical CT will contribute to clarifying this issue in the future. References [1] Davis H, Wallace WM. Factors affecting changes produced in the electroencephalogram by standardized hyperventilation. Arch Neurol Psychiatry 1942;47:606-25. [2] Gibbs FA, Gibbs EL, Lennox WG, Nims LF. The value of carbon dioxide in counteracting the effect of low oxygen. J Aviation Med 1943;14:250-61. [3] Kety SS, Scmidt CF. The effects of active and passive hyperventilation on cerebral oxygen consumption, cardiac output, and blood pressure of normal young men. J Clin Invest 1946;25:107-19. [4] Meyer JS, Gotoh F. Metabolic and electroencephalographic effects of hyperventilation. Experimental studies of brain oxygen and carbon dioxide tension, pH, EEG and blood flow during hyperventilation. Arch Neurol 1960;3:539-52. [5] Gotoh F, Meyer JS, Takagi Y. Cerebral effects of hyperventilation in man. Arch Neurol 1965;12:410-23. [6] Kennealy JA, Penovich PE, Moore-Nease SE. EEG and spectral analysis in acute hyperventilation. Electroencephalogr Clin Neurophysiol 1986;63:98-106. [7] Yamaguchi F, Meyer JS, Sakai F, Yamamoto M. Normal human aging cerebral vasoconstrictive response to hypocapnia. J Neurol Sci 1979;44:87-94. [8] Hoshi Y, Hazeki O, Kakihana Y, Tamura M. Redox behavior of cytochrome oxidase in the rat brain measured by near-infrared spectroscopy. J Appl Physiol 1997;83:1842-8. [9] Konishi T. The standardization of hyperventilation on EEG recording in childhood. I. The quantity of hyperventilation activation. Brain Dev 1987;9:16-20. [10] Vriens EM, Kraaier V, Musbach M, Wieneke GH, van Huffelen
AC. Transcranial pulsed Doppler measurements of blood velocity in the middle cerebral artery: Reference values at rest and during hyperventilation in healthy volunteers in relation to age and sex. Ultrasound Med Biol 1989;15:1-8. [11] Greenberg JH, Alavi A, Reivich M, Kuhl D, Uzzell B. Local cerebral blood volume response to carbon dioxide in man. Circ Res 1987;43:324-31. [12] Reuter JH, Disney TA. Regional cerebral blood flow and cerebral metabolic rate of oxygen during hyperventilation in the newborn dog. Pediatr Res 1986;20:1102-6. [13] Kamei A, Ozaki T, Takashima S. Monitoring of the intracranial hemodynamics and oxygenation during and after hyperventilation in newborn rabbits with near-infrared spectroscopy. Pediatr Res 1994;35: 334-8. [14] Pryds O, Greisen G, Skov LL, Friis-Hansen B. Carbon dioxiderelated changes in cerebral blood volume and cerebral blood flow in mechanically ventilated preterm neonates: Comparison of near infrared spectrophotometry and 133xenon clearance. Pediatr Res 1990;27:445-9. [15] Hoshi Y, Tamura M. Detection of dynamic changes in cerebral oxygenation coupled to neuronal function during mental work in man. Neurosci Lett 1993;150:5-8. [16] Hoshi Y, Tamura M. Near-infrared optical detection of sequential brain activation in the prefrontal cortex during mental tasks. Neuroimage 1997;5:292-7. [17] Saunders MG, Westmoreland BF. The EEG in evaluation of disorders affecting the brain diffusely. In: Klass DW, Daly DD, eds. Current practice of clinical electroencephalography. New York: Raven Press, 1979:343-79. [18] Patel VM, Maulsby RL. How hyperventilation alters the electroencephalogram: A review of controversial viewpoints emphasizing neurophysiological mechanisms. J Clin Neurophysiol 1987;4:101-20. [19] Bonvallet M, Dell P. Reflections on the mechanisms of the action of hyperventilation upon the EEG. Electroencephalogr Clin Neurophysiol 1956;8:170. [20] Oda I, Eda H, Tsunazawa Y, Takada M, Yamada Y, Nishimura G, Tamura M. Optical tomography by the temporally extrapolated absorbance method. Appl Opt 1996;35:169-75.
CONNECTIONS—Web Site Update Provided by Steven M. Leber, MD, PhD and Kenneth J. Mack, MD, PhD The following neuro-oncology site came highly recommended by a parent: A Starting Point to Connect With Resources Related to Pediatric Neuro-Oncology, hosted and reviewed by the University of Miami, Department of Neurosurgery http://www.med.miami.edu/neurosurgery/start_htm The address for the International Child Neurology Association was incorrectly listed in the April 1999 issue. The correct address is http://www2.mf.uni-lj.si/⬃icna/index.html. Please contact us ([email protected]) for additions.
Hoshi et al: Mechanisms of HV-Induced EEG Slowing 643
It is well known that hyperventilation (HV) produces electroencephalogram (EEG) slowing. Beginning in the
1940s, the mechanisms of HV-induced EEG slowing were disputed [1-3]. However, in 1960, Meyer and Gotoh [4] concluded that EEG slowing in the monkey during HV resulted from cerebral ischemic hypoxia associated with vasoconstriction caused by a lowering of Paco2 (the hypoxia theory [1]). Since then the hypoxia theory has been supported by much clinical and experimental evidence obtained under conditions of moderate or severe hypocapnia, such as Paco2 less than 25 mm Hg [5,6]. However, there is a wide range of Paco2 levels over which EEG slowing occurs [2]. For example, in young children, EEG slowing can be produced easily under mildly hypocapnic conditions, which is unusual in adults. This difference is partly the result of age-related cerebral vascular carbon-dioxide responsiveness [7], although a striking difference was also demonstrated between normal subjects and patients with epilepsy [2]. HV is commonly used as a useful activation test in a routine EEG examination of children with epilepsy. Previously the authors examined the relationship between cerebral oxygenation and EEG in rats and found that high-voltage slow-wave activity appeared on the EEG when cerebral oxygenation was reduced to an extremely low level, at which the intracellular oxygen pressure was less than 0.03 mm Hg [8]. This observation means that the HV activation test could be a harmful procedure to children with epilepsy. The degree of HV-induced EEG slowing has been reported to be independent of the extent of the decrease in cerebral blood flow (CBF) [9]. These differences are believed to result from individual brain sensitivity to hypoxia [2,9], although there is no direct evidence to support that hypothesis. There is no doubt that marked hypocapnia causes cerebral ischemic hypoxia because of vasoconstriction, resulting in EEG slowing. However, the existence of individual differences calls into
From the *Biophysics Group; Research Institute for Electronic Science; Hokkaido University; †Laboratory of Electrophysiology; Hokkaido University Hospital; and ‡Department of Pediatrics; Hokkaido University School of Medicine; Sapporo, Japan.
Communications should be addressed to: Dr. Hoshi; Biophysics Group; Research Institute for Electronic Science; Hokkaido University; Sapporo 060-0812, Japan. Received February 26, 1999; accepted May 26, 1999.
Hoshi Y, Okuhara H, Nakane S, Hayakawa K, Kobayashi N, Kajii N. Re-evaluation of the hypoxia theory as the mechanism of hyperventilation-induced EEG slowing. Pediatr Neurol 1999;21:638-643.
Introduction
638
PEDIATRIC NEUROLOGY
Vol. 21 No. 3
© 1999 by Elsevier Science Inc. All rights reserved. PII S0887-8994(99)00063-6 ● 0887-8994/99/$20.00
question the presumption that the hypoxia theory can be applied universally. It is essential for verification of the hypoxia theory to measure cerebral oxygenation. Near-infrared spectroscopy (NIRS) allows real-time and continuous measurements of changes in the hemoglobin oxygenation state, blood volume, and the redox state of cytochrome oxidase (cyt ox) in the brain. The measurement of the redox state of cyt ox provides direct information about the intracellular oxygenation state. Using NIRS the authors monitored changes in the cerebral oxygenation and end-tidal concentration of carbon dioxide (ETco2) in children with epilepsy during an HV activation test. The authors also examined cerebral vascular carbon-dioxide responsiveness in normal healthy children as a control study. The aim of this study was to examine whether the hypoxia theory actually accounts for HV-induced EEG slowing. Subjects and Methods Subjects. The subjects were 67 well-controlled outpatients with epilepsy (25 females, 42 males) aged 5-12 years who underwent routine EEG examinations. Exclusion criteria included gross cerebral structural abnormality as defined by magnetic resonance imaging or associated cerebrovascular anomalies. Patients with absence seizures were also excluded. Before the study, informed consent was obtained from the subjects’ parents and from children older than 10 years. Because simultaneous measurement of EEG, ETco2, and cerebral oxygenation is stressful, several children consented to participate in the study on the condition that only ETco2 or cerebral oxygenation was measured. Thus ETco2 was measured separately from the NIRS measurement. ETco2 was measured in 15 patients (Group 1), and cerebral oxygenation was measured in 52 patients (Group 2). The clinical characteristics of these two groups are presented in Table 1. To examine cerebral vascular carbon-dioxide responsiveness, simultaneous measurement of cerebral oxygenation and ETco2 was performed in nine healthy children (five females, four males) aged 5-12 years. Near-infrared Spectroscopy. The basic principle of the NIRS apparatus used in this study has been previously published in detail [8]. A portable apparatus was built whereby near-infrared light from a halogen lamp passed through a lens system with a rotating disc containing four interference filters (700-, 730-, 750-, and 805-nm wavelengths). The concentration changes in oxygenated hemoglobin (oxy-Hb), deoxygenated hemoglobin (deoxy-Hb), and oxidized cyt ox were calculated by the following numeric formulas every 1 second: K⌬关oxy-Hb兴 ⫽ ⫺ 0.868⌬A 700-750 ⫺ 1.74⌬A 730-750 K⌬关deoxy-Hb兴 ⫽ 0.868⌬A 700-750 ⫺ 2.26⌬A 730-750 Ka 3⬙⌬关cyt ox兴 ⫽ 0.674⌬A 700-750 ⫺ 0.752⌬A 730-750 ⫹ 0.5⌬A 805-750 K is the apparent difference absorption coefficient of either oxy-Hb or deoxy-Hb at an arbitrary wavelength pair, and a3⬙ is a proportionality factor for oxidized cyt ox [8]. Because the value of either K or a3⬙ cannot be determined experimentally, the results were expressed in relative amounts rather than in absolute values of concentration. Summation of changes in oxy-Hb and deoxy-Hb gives changes in the total hemoglobin concentration (t-Hb). Under conditions with a constant hematocrit, the changes in t-Hb can be used as an indicator of changes in cerebral blood volume (CBV) within the optical field. The optodes were placed 3 cm from each other on the middle of the forehead. To shield from ambient light the forehead was wrapped in a band. Procedures. ETco2 was continuously monitored by a capnometer (Eliza, Gambo, Tokyo). Data were plotted every 10 or 20 seconds. The EEG was recorded with monopolar electrodes (reference electrodes,
Table 1.
Clinical characteristics of the two study groups
Sex (n) Female Male Age (yr, mean ⫾ SD) Seizure type (n) GTS GTCS GTCS ⫹ CPS CPS SPS Other Antiepileptic (n) VPA CBZ CZP VPA ⫹ CBZ, CZP, and/or PH PB PB ⫹ CBZ Other
Group 1 (n ⴝ 15)
Group 2 (n ⴝ 52)
6 9 9 ⫾ 1.86
19 33 8 ⫾ 1.72
5 5 2 2 1 0
15 8 13 8 5 3
4 6 2 1 0 1 1
17 19 4 6 1 2 3
Abbreviations: CBZ ⫽ Carbamazepine CPS ⫽ Complex partial seizure CZP ⫽ Clonazepam GTCS ⫽ Generalized tonic-clonic seizure GTS ⫽ Generalized tonic seizure PB ⫽ Phenobarbital SPS ⫽ Simple partial seizure VPA ⫽ Valproic acid
auricles). Silver-silver chloride electrodes were attached to the skin according to the International 10-20 system. All measurements were performed within 2 hours after a meal to avoid hypoglycemia. The subjects were in the supine position. After a 2-minute baseline recording the subject began HV with his eyes closed. HV was performed for 4 minutes in children with epilepsy and for 3 minutes in healthy children. Subjects hyperventilated voluntarily, but when the frequency of ventilation and the depth of breathing decreased below minimal values, a supervisor instructed them to adjust the rate. Measurements of cerebral oxygenation and ETco2 were discontinued 2 minutes after the cessation of HV. EEG slowing was defined when HV-induced diffuse high-voltage slow-wave activity with disappearance of background activities continued for more than 30 seconds. EEG slowing is generally divided into four degrees according to the severity of the slowing [9]. The definition used here corresponds to the most severe slowing. Comparisons were made by Student t test.
Results Relationship Between Cerebral Oxygenation and ETCO2 During HV in Healthy Children In seven of the nine healthy children, t-Hb and oxy-Hb decreased in parallel with the decrease in ETco2; cyt ox was not reduced (Fig 1). The behavior of the change in deoxy-Hb varied with each subject: an increase, no change, and a decrease in deoxy-Hb were observed. Two children did not demonstrate decreases in either t-Hb or oxy-Hb, yet ETco2 was considerably decreased. Figure 2 presents the NIRS traces obtained from one of these two
Hoshi et al: Mechanisms of HV-Induced EEG Slowing 639
ished and did not meet the definition of EEG slowing, or the slow waves disappeared abruptly within 1 second after the cessation of HV in four patients and the ETco2 was still lower than the level at which EEG slowing had begun. Relationship Between EEG Slowing and Cerebral Oxygenation in Group 2
Figure 1. Changes in cerebral oxygenation and end-tidal carbon dioxide (ETCO2) caused by hyperventilation (HV) in a healthy child. Baselines were selected from the starting point of the measurement; these values were taken as zero for each near-infrared spectroscopy signal. Changes from baseline were represented as relative amounts. Upward and downward trends demonstrate increases and decreases in values, respectively. HV was performed between two arrows marked HV. ⌬[oxy-Hb]: change in oxygenated hemoglobin; ⌬[deoxy-Hb]: change in deoxygenated hemoglobin; ⌬[t-Hb]: change in total hemoglobin; ⌬[oxidized cyt. ox.]: change in oxidized cytochrome oxidase.
subjects. Although ETco2 decreased to about 2.5%, t-Hb and oxy-Hb increased and deoxy-Hb decreased. t-Hb and oxy-Hb decreased about 20 seconds before the cessation of HV. Although ETco2 was returning to the pre-HV level after the cessation of HV, decreases in t-Hb and oxy-Hb and an increase in deoxy-Hb were observed. The other subject demonstrated no changes in NIRS parameters during HV. After the cessation of HV, this subject also demonstrated decreases in t-Hb and oxy-Hb with an increase in deoxy-Hb; the redox state of cyt ox was not changed. Relationship Between EEG Slowing and ETCO2 in Group 1 EEG slowing was induced in seven of 15 patients. There was no difference in ETco2 in the resting period between patients with and without EEG slowing (5.1 ⫾ 0.49% vs 5.1 ⫾ 0.35%). The value of ETco2 when EEG slowing occurred varied with each patient. ETco2 at the end of HV was significantly lower in patients with EEG slowing than in those without EEG slowing (2.98 ⫾ 0.41% vs 3.89 ⫾ 0.64%, P ⬍ 0.001). EEG slowing occurred intermittently in five patients and was not synchronous with changes in ETco2 (Fig 3). The degree of EEG slowing was dimin-
640
PEDIATRIC NEUROLOGY
Vol. 21 No. 3
EEG slowing was induced in 28 of 52 patients. In all of these patients but one, decreases in t-Hb and oxy-Hb were observed. The behavior of the change in deoxy-Hb varied with each patient as it did in healthy children, and a reduction of cyt ox was not observed. Figure 4 illustrates a typical example of NIRS traces obtained from a subject with EEG slowing. In three patients, EEG slowing occurred before t-Hb and oxy-Hb decreased. Only one patient demonstrated no change in cerebral oxygenation throughout the EEG slowing (Fig 5). In this patient, t-Hb and oxy-Hb decreased and deoxy-Hb increased after cessation of HV. EEG slowing occurred intermittently in 17 patients, in whom disappearance of slow waves was not accompanied by reoxygenation of the brain (Fig 6). Attenuation of a continuous EEG slowing with brief statements, such as “I’m in pain,” was observed in three patients. This attenuation also occurred without accompanying reoxygenation of the brain. The degree of EEG slowing was diminished or the slow waves disappeared abruptly within 1 second after the cessation of HV in 18 patients, and cerebral oxygenation had not yet recovered. In seven of 24 patients without EEG slowing, decreases in t-Hb and oxy-Hb were observed during HV. Discussion Relationship Between Cerebral Oxygenation and ETCO2 in Healthy Children Cerebrovascular carbon-dioxide responsiveness varies with age [7,10], and a positive correlation between CBF (or CBV) and Paco2 has been observed over a wide age range [7,10-12]. Several NIRS studies in both humans and animals have demonstrated decreases in CBV (t-Hb) during HV [13,14]. The authors also observed that progressive decreases in ETco2 accompanied decreases in t-Hb and oxy-Hb in seven of nine healthy children. In contrast, two healthy children demonstrated no changes or, conversely, increases in t-Hb and oxy-Hb, with ETco2 considerably decreased. Performing HV is a difficult task and is accompanied by emotional stress. Previously the authors observed that emotional stress caused increases in t-Hb and oxy-Hb in the frontal region that reflect an increase in regional CBF (rCBF) responding to neuronal activation [15,16]. In addition, HV often increases alertness in a drowsy child, as a result of the increase in CBF. Thus the authors speculate that the increase in rCBF caused by emotional stress or arousal response exceeded or canceled the decrease in rCBF caused by vasoconstric-
Figure 2. Changes in cerebral oxygenation and ETCO2 caused by hyperventilation (HV) in a healthy child. In contrast to the subject in Figure 1, this subject demonstrated increases in oxy-Hb and t-Hb during HV. Abbreviations as in Figure 1.
tion in the frontal region, which resulted in no changes or increases in oxy-Hb and t-Hb, with CBF decreases in other brain regions. The decreases in t-Hb and oxy-Hb with an increase in deoxy-Hb after the cessation of HV observed in these subjects probably resulted from the combined effects of cerebral ischemia that became manifest because of the relief from the emotional stress and hypoxemia caused by suppression of breath.
Mechanisms of EEG Slowing The most prominent changes observed in the patients with EEG slowing were hypocapnia and decreases in t-Hb and oxy-Hb that seemed to be consistent with the hypoxia theory. However, one patient did not demonstrate decreases in either t-Hb or oxy-Hb, and EEG slowing was observed almost throughout the HV. In three patients,
Figure 3. Relationship between EEG slowing and ETCO2 in an epileptic patient. The bottom horizontal bars denote the occurrence of EEG slowing. EEG slowing occurred intermittently. Abbreviations as in Figure 1.
Hoshi et al: Mechanisms of HV-Induced EEG Slowing 641
Figure 4. Changes in cerebral oxygenation caused by hyperventilation (HV) in an epileptic patient with EEG slowing. The bottom horizontal bar denotes the occurrence of EEG slowing. Abbreviations as in Figure 1.
EEG slowing preceded decreases in t-Hb and oxy-Hb. Furthermore, unlike high-voltage slow-wave activity observed in rats in hypoxic hypoxia [8], HV-induced EEG slowing was not accompanied by the reduction of cyt ox. These data suggest that ischemic hypoxia in the frontal region is not essential for EEG slowing. As is presented in Figures 3 and 5, EEG slowing occurred intermittently in some subjects in whom the disappearance was not accompanied by increases in either ETco2 or cerebral oxygen-
Figure 5. Changes in cerebral oxygenation caused by hyperventilation (HV) in an epileptic patient with EEG slowing. Unlike the subject in Figure 4, this subject demonstrated intermittent EEG slowing. Abbreviations as in Figure 1.
642
PEDIATRIC NEUROLOGY
Vol. 21 No. 3
Figure 6. Changes in cerebral oxygenation caused by hyperventilation (HV) in an epileptic patient with electroencephalographic (EEG) slowing. In contrast to the subject in Figure 4, this subject demonstrated no significant change in cerebral oxygenation state during hyperventilation (HV). Abbreviations as in Figure 1.
ation. This intermittent EEG slowing might be a projected rhythm [17]. This rhythm, which can be observed in the earlier stage of hypoxia, is in general indicative of a disturbance arising from the midline structures and projecting to the surface [17]. Thus it is plausible that the central ischemic effects occur at the subcortical level and that the effects on the cortex are indirect. During the recovery periods, however, abrupt diminution of the degree of EEG slowing and disappearance of the slow waves occurred almost simultaneously with the cessation of HV in some subjects, with both cerebral oxygenation and ETco2 still at low levels. In addition the attenuation of EEG slowing with brief speaking was not accompanied by reoxygenation of the brain. These results suggest that more subtle mechanisms are significant for EEG slowing. Reviewing published reports, Patel and Maulsby [18] emphasized a similarity between the mechanism of HV-induced EEG slowing and the transition from drowsiness to slow-wave sleep, which had been originally proposed by Bonvallet and Dell [19] in the 1950s. That theory offers one possible explanation for the authors’ observations. That is, speaking and the instruction to stop HV are alerting stimuli. Such alerting stimuli might have attenuated EEG slowing without increases in cerebral oxygenation and ETco2. For a full explanation, however, further investigation is required. The results of the present study suggest that the effects of HV on the subcortical level are responsible for the EEG slowing. However, it is again unclear which factor primarily causes changes in neuronal activity at the subcortical level: hypoxia, hypocapnia, or other HV-associated changes. Direct measurements of the oxygenation state and carbon dioxide tension at the subcortical level are
necessary to answer this question. Recently the authors’ group developed optical computed tomography (CT) [20]. Unlike the conventional NIRS apparatus, optical CT provides quantitative information about the oxygenation state in the deep brain structures. Thus optical CT will contribute to clarifying this issue in the future. References [1] Davis H, Wallace WM. Factors affecting changes produced in the electroencephalogram by standardized hyperventilation. Arch Neurol Psychiatry 1942;47:606-25. [2] Gibbs FA, Gibbs EL, Lennox WG, Nims LF. The value of carbon dioxide in counteracting the effect of low oxygen. J Aviation Med 1943;14:250-61. [3] Kety SS, Scmidt CF. The effects of active and passive hyperventilation on cerebral oxygen consumption, cardiac output, and blood pressure of normal young men. J Clin Invest 1946;25:107-19. [4] Meyer JS, Gotoh F. Metabolic and electroencephalographic effects of hyperventilation. Experimental studies of brain oxygen and carbon dioxide tension, pH, EEG and blood flow during hyperventilation. Arch Neurol 1960;3:539-52. [5] Gotoh F, Meyer JS, Takagi Y. Cerebral effects of hyperventilation in man. Arch Neurol 1965;12:410-23. [6] Kennealy JA, Penovich PE, Moore-Nease SE. EEG and spectral analysis in acute hyperventilation. Electroencephalogr Clin Neurophysiol 1986;63:98-106. [7] Yamaguchi F, Meyer JS, Sakai F, Yamamoto M. Normal human aging cerebral vasoconstrictive response to hypocapnia. J Neurol Sci 1979;44:87-94. [8] Hoshi Y, Hazeki O, Kakihana Y, Tamura M. Redox behavior of cytochrome oxidase in the rat brain measured by near-infrared spectroscopy. J Appl Physiol 1997;83:1842-8. [9] Konishi T. The standardization of hyperventilation on EEG recording in childhood. I. The quantity of hyperventilation activation. Brain Dev 1987;9:16-20. [10] Vriens EM, Kraaier V, Musbach M, Wieneke GH, van Huffelen
AC. Transcranial pulsed Doppler measurements of blood velocity in the middle cerebral artery: Reference values at rest and during hyperventilation in healthy volunteers in relation to age and sex. Ultrasound Med Biol 1989;15:1-8. [11] Greenberg JH, Alavi A, Reivich M, Kuhl D, Uzzell B. Local cerebral blood volume response to carbon dioxide in man. Circ Res 1987;43:324-31. [12] Reuter JH, Disney TA. Regional cerebral blood flow and cerebral metabolic rate of oxygen during hyperventilation in the newborn dog. Pediatr Res 1986;20:1102-6. [13] Kamei A, Ozaki T, Takashima S. Monitoring of the intracranial hemodynamics and oxygenation during and after hyperventilation in newborn rabbits with near-infrared spectroscopy. Pediatr Res 1994;35: 334-8. [14] Pryds O, Greisen G, Skov LL, Friis-Hansen B. Carbon dioxiderelated changes in cerebral blood volume and cerebral blood flow in mechanically ventilated preterm neonates: Comparison of near infrared spectrophotometry and 133xenon clearance. Pediatr Res 1990;27:445-9. [15] Hoshi Y, Tamura M. Detection of dynamic changes in cerebral oxygenation coupled to neuronal function during mental work in man. Neurosci Lett 1993;150:5-8. [16] Hoshi Y, Tamura M. Near-infrared optical detection of sequential brain activation in the prefrontal cortex during mental tasks. Neuroimage 1997;5:292-7. [17] Saunders MG, Westmoreland BF. The EEG in evaluation of disorders affecting the brain diffusely. In: Klass DW, Daly DD, eds. Current practice of clinical electroencephalography. New York: Raven Press, 1979:343-79. [18] Patel VM, Maulsby RL. How hyperventilation alters the electroencephalogram: A review of controversial viewpoints emphasizing neurophysiological mechanisms. J Clin Neurophysiol 1987;4:101-20. [19] Bonvallet M, Dell P. Reflections on the mechanisms of the action of hyperventilation upon the EEG. Electroencephalogr Clin Neurophysiol 1956;8:170. [20] Oda I, Eda H, Tsunazawa Y, Takada M, Yamada Y, Nishimura G, Tamura M. Optical tomography by the temporally extrapolated absorbance method. Appl Opt 1996;35:169-75.
CONNECTIONS—Web Site Update Provided by Steven M. Leber, MD, PhD and Kenneth J. Mack, MD, PhD The following neuro-oncology site came highly recommended by a parent: A Starting Point to Connect With Resources Related to Pediatric Neuro-Oncology, hosted and reviewed by the University of Miami, Department of Neurosurgery http://www.med.miami.edu/neurosurgery/start_htm The address for the International Child Neurology Association was incorrectly listed in the April 1999 issue. The correct address is http://www2.mf.uni-lj.si/⬃icna/index.html. Please contact us ([email protected]) for additions.
Hoshi et al: Mechanisms of HV-Induced EEG Slowing 643