- Email: [email protected]
Objective measurement of hepatic encephalopathy in pigs by means of spectral analysis and evoked responses
298
Brain Research, 361)( 1985) 298-3()3 Elsevier
BRE 11277
Objective Measurement of Hepatic Encephalopathy in Pigs by Means of Spectral Analysis and Evoked Responses GERRIT H. DE GROOT, SOLKO W. SCHALM, MARINUS DE VLIEGER and CARIN C.D. VAN DER RIJT Departments of lnternal Medicine H and Electroneurology, Erasmus University, Rotterdam (The Netherlands) (Accepted April 2nd, 1985) Key words: acute hepatic failure - - hepatic encephalopathy - - electroencephalogram (EEG) - - spectral analysis - visual evoked potential - - brainstem auditory evoked potential
Objective measurement of hepatic encephalopathy by means of spectral analysis of the electroencephalogram (EEG), visualevoked potentials (VEP) and brainstem auditory-evoked potentials (BAEP) was studied in pigs with ischemic hepatic necrosis. The mean dominant frequency (MDF) and the relative power of the delta frequency band (%6 power) showed significant changes with increasing encephalographic grades of encephalopathy: MDF dropped from 7.0+ 0.5 (grade 0) to 2.7 + 0.3 Hz (grade 3 encephaiopathy) and %6 power increased from 52 + 7 (grade 0) to 83 _.+6% (grade 3). The patterns of the VEP in pigs corresponded to those of the human VEP. However, significant differences in either latency time of the peaks or the peak amplitude with increasing stages of he-' patic encephalopathy could not be found. The BAEP registered for pigs were reproducible but also were not as useful as spectral analysis for grading hepatic encephalopathy. INTRODUCTION Pigs are often used to study the pathogenesis or treatment of hepatic encephalopathy, because of their physiological similarity to the human being. Reliable methods for the objective measurement of encephalopathy in pigs are non-existent. Clinical neurological assessment of the grade of encephalopathy in pigs is subjective and unreliable, since it yields only a distinction between coma and non-coma. Conventional E E G readings have been proposed as a method for grading hepatic encephalopathy 1,5-7, but interpretation of the results remains subjective and therefore dependent on the experience and consistency of the electroneurologist. The introduction of quantitative analysis of the E E G offered the possibility of measuring hepatic encephalopathy objectively2,11. By this method the E E G can be expressed in terms of the mean dominant frequency (MDF) and the wave amplitude as a function of the frequency (power spectrum). Recently evoked potentials were introduced as another ob-
jective test for hepatic encephalopathy 4.s.9.J2. The present work was carried out to determine the reliability of quantitative E E G analysis a n d evoked potentials as a means of grading pigs with hepatic encephalopathy. MATERIALS AND METHODS Surgical preparation Fifteen large White pigs, weighing 28-33 kg, were used. Pigs underwent laparotomy for the introduction of a functional end-to-side portocaval shunt. All ligaments and peritoneal attachments to the liver were cut or coagulated. Subsequently a clamp was placed around the isolated hepatic artery. Five silver electrodes (diameter 1 mm) were inserted through the skull onto the dura for the recording of the E E G and evoked potentials: two over the frontal cortex, one m the vertex and two over the occipital cortex. The electrodes were anchored with sterile acrylic bone cement (Surgical Simplex, Howmedica, London. U.K.) and channeled percutaneously. Three
Correspondence: S.W. Schalm, Department of Internal Medicine II, University Hospital Dijkzigt, Dr. Molewaterplein 40, 3015 GD Rotterdam, The Netherlands. 0006-8993/85/$03.30© 1985 Elsevier Science Publishers B.V. (Biomedical Division)
299 days after construction of the portocaval shunt acute hepatic necrosis was induced by temporarily tightening the clamp to close off the hepatic artery for 4 or 6 h. Anesthesia was not used in this phase, since it may interfere with subsequent neurological assessments. Glucose (12 g/kg/24 h), penicillin G (9 megaU/24 h) and kanamycin (3 g/24 h) were infused continuously. The environmental temperature was maintained at 25 °C to ensure normothermia of the pigs with hepatic necrosis.
Recording procedure During the study periods the animals were placed in a dim room on their left side and covered with a sheet to ensure adequate relaxation. E E G recordings were taken before induction of hepatic necrosis and 24, 30, 48 and 54 h afterwards. It should be noted that pigs as soon as they tend to fall asleep have an 'abnormally' slow EEG. Therefore in order to obtain a meaningful E E G they had to be aroused each time. The potentials in 4 bipolar leads (left and right fronto-occipital, frontofrontal and occipito-occipital) were registered by an E E G apparatus (Ahrend v. Gogh, Amsterdam, The Netherlands). During 3 'epochs' of 100 s the signals of the fronto-occipital leads were stored simultaneously on a magnetic tape (Analog 7, Philips, Eindhoven, The Netherlands) for (quantitative) spectral analysis. The E E G channels were filtered with a band pass filter between 0.5 and 70 Hz. The stored data were fed into a computer (PDP 11/34) at a sampling rate of 51.2 Hz and a sensitivity of 11 bits/5 VI Each epoch of 100 s was divided into 10 periods. The power spectrum (the sum of wave amplitudes as a function of each frequency or frequency band) was calculated for each period of 10 s using the Fast Fourier Transformation, and the mean power spectrum for each 100 s epoch was obtained. The frequency resolution was 0.1 Hz. To minimize the effect of artefacts, only the range of 1.0-25.6 Hz was used to calculate the mean spectrum which was then analyzed (Fig. 1). The parameters calculated were the mean dominant frec]uency and the relative power of the delta, theta and alpha frequency range 3. The mean dominant frequency (MDF) was defined as: n
MDF = ~: fi Si/• i=l
i=l
Si
where fi is frequency i, S i is power of frequency i and n is the number of frequencies. The delta activity was defined as the activity of the frequencies 1.0-3.5 Hz, the theta activity that of the frequencies 3.5-8.0 Hz, the alpha activity that of the frequencies 8.0-13.0 Hz, and the beta activity that of the frequencies 13.0-25.6 Hz. The 3 epochs of 100 s were averaged. Spectra that showed high muscle activity in the beta frequency band or asymmetry between the two fronto-occipital leads were not included in the calculations. The E E G recordings were analyzed independently by an experienced electroneurologist; the 5 grade classification system described by Opolon et al.5,6 was used.
Visual-evokedpotentials (VEP) A photo stimulator with a xenon flashlight provided stimuli with a flash intensity of 1 J and a stimulation frequency of 1 Hz. The experiments were performed in a darkened room with the flashlight at a constant distance of 70 cm from the right eye. The EEG, monitored on an oscilloscope and on the E E G recorder, was stored on the magnetic tape during a 300 s artefact-free period. For the two fronto-occipital leads VEPs were obtained by averaging the response of 100 stimuli. Only the responses during the first 400 ms were used 10.
Brainstem-evokedpotentials (BAEP) Responses from electrodes at the vertex and ear lobes were recorded. Ten clicks/s were generated; the intensity of each click was 70 dB and it lasted 0.2 ms. Clicks were presented to the animal via an earphone. The average response in the 10 ms interval after the stimulus (n = 2000) was calculated by computer (PDP 11/34). Each animal was subjected to two mono-aural tests in a 5 min period.
Data evaluation Analysis of the VEP in normal pigs showed 7 waves (Fig. 2A). Peaks I, III, V and VII were positive, whereas peaks II, IV and VI were negative. All peaks occurred within 150 ms after visual stimulation. Analysis of the BAEP in pigs showed 5-6 positive waves within the first 6 ms (Fig. 3a). All waves found for normal animals were com-
300
100.0 power uV2/Hz.
100.0-
grade 0
power
~ ~
grade 3
uV2/Hz.
10.0
10.0-
1.0
1.0
0.1 0
I 5
I 10 freq.
I 15 [Hz.]
I 20
b
0. I 25
0
i
I
!
I
5
I0
15
20
25
freq. [Hz.]
Fig. 1. A visual representation of the EEG power spectrum (sum of all amplitudes as a function of the frequency) for normal pigs (a) and for pigs with hepatic encephalopathy(b).
pared with the waves found for animals with hepatic encephalopathy. Moreover each animal was used as its own control during the development of encephalopathy. The latency time was defined as the time (ms) between onset of the stimulus and the maximal amplitude of each peak. Amplitudes between two consecutive positive and negative peaks were measured. RESULTS
Survival Fifty percent of the pigs undergoing 4 h of ischemia of the liver survived. Two animals died from complications (air embolism, peritonitis) and two of hepatic coma. Six h of ischemia was followed by hepatic encephalopathy in all animals. Four of the 7 animals died of hepatic coma; one animal in hepatic coma, grade 4, was sacrificed after 72 h and two other animals had concurrent gastric hemorrhage.
EEG and spectral analysis (Figs, 4 and 5) The EEG, semiquantified into grades 1-5, grad-
ually deteriorated after induction of ischemia. Animals subjected to 4 h of ischemia developed mild encephalopathic E E G changes whereas 6 h of ischemia produced severe encephalopathic E E G changes within 30 h. The mean dominant frequency dropped significantly with each grade of eneephatopathy up to grade 3. No significant frequency differences were observed between grades 3 and 4 encephalopathy. The relative power of the delta activity was 52 + 7% in grade 0, and this increased significantly with each grade of encephalopathy up to grade 3 (83 _+ 6%). Delta activities for grades 3 and 4 did not differ significantly. An example of the power spectra for grade 0 and grade 3 encephalopathy is given in Fig. 1.
VEP (Table I) The latency time for all peaks was measured during several grades of encephalopathy. The overall results did not show significant differences in peak latency or changes in amplitude between grades 0 and 4 encephalopathy. However, in some animals there
301 2.0 u v
20.0 uv
grade : 0
VEP, g r a d e 0
"V VI I
2.0 UV 2 0 . O u v II
VEP, g r a d e 3 grade : 3
A
\ i"
I
\
\,,
B
, V, v,
.
~
'
'
Fig. 3. a: BAEP of a normal pig. b: B A E P of a pig with grade 3 hepatic encephalopathy. Iv Fig. 2. a: VEP of a normal pig. b: VEP of a pig with grade 3 hepatic encephalopathy.
during all grades of encephalopathy. The latency times and amplitudes of the 5 peaks did not change significantly with the grade of encephalopathy (Fig.
was an increase in latency time of peak VII and an decrease in amplitude between peak IV and V during encephalopathy (Fig. 2b).
3b). DISCUSSION
BAEP (Table II)
Since clinical grading of hepatic encephalopathy in pigs is subjective and difficult, Opolon et al. 5 intro-
As a rule the 5 positive peaks could be identified TABLE I
VEP peak latency times (ms, mean + S.D.) in relation to the grade of encephalopathy + + , no definite peak discernable. *, insufficient data to calculate S.D.
Encephalopathy
Peaks
grade
I
H
II1
IV
V
VI
VII
0 1 2 3 4
33+8 31 + 2 24 + * 24 + * ++
40+5 36 + 5 35 + 5 38 + * 36 + 4
54+6 52 + 9 49 + 7 54 + 6 48 + 9
82+11 82 + 19 76 + 14 89 + 17 ++
107+18 101 + 15 103 + 12 122 + 27 ++
133+25 116 + 16 125 + 12 131 + * ++
154+17 151 + 10 143 + 15 160 + * 153 + 34
302
5
4 hours of ischemia
5
4
4
~3 D~
3
~ 2
2
LU
I
I
0
O-
I
I
I
I
I
I
I
I
I
1
0
24
30
48
54
0
24
30
48
54
n o u r s a f t e r i n d u c t i o n of ischemia
h o u r s a f t e r i n d u c t i o n of ischemia
Fig. 4. The course of the EEG grades for pigs that have undergone 4 and 6 h of ischemia of the liver. The 4-h animals with an EEG at 54 h survived 72 h and were then sacrificed. Clinicallythe 4 animals were alert at 72 h. The 6-h animal with an EEG at 54 h survived till 72 h with grade 4-5 encephalopathy.
duced 5 grades of encephalopathy based on conventional E E G recordings. However, the interpretation of the E E G remains subjective and dependent on the expertise of the observer. Our study showed that objective measurement of hepatic encephalopathy in pigs is possible by means of spectral analysis. The EEG activity, represented by the mean dominant frequency, exhibited a fairly close correlation with the grade of encephalopathy. The relative power of the delta band expressed as a percentage of the total power of the alpha, beta, theta and delta bands, increased significantly with deterioration of the EEG, reflecting an increase in slow wave activity. Although the mean values of MDF and percent delta power were significantly different for each grade of encephalopathy, individual measurements showed some
overlapping. Therefore both parameters should be used together to classify the grade of encephalopathy, although our results indicate that a reliable distinction between grades 3 and 4 encephalopathy was not possible. The change in E E G as displayed with spectral analysis are probably not specific for hepatic encephalopathy, but may also be observed in other metabolic conditions1.13. Since evoked potentials are said to allow differentiations of etiologies 12. visualand auditory-evoked potentials were also analyzed. VEPs in pigs have the 7 main components present in man 10. However, the reproducibility of the latency times and the amplitudes of two consecutive positive and negative peaks for each grade of encephalopathy was rather poor. In some individual animals we observed a decrease in amplitudes of two consecutive
TABLE II B A E P peak latency times (ms, mean + S.D. ) in relation to the grade o f encephalopathy
*, insufficient data to calculate S.D. Encep~lopathy gra~
0 1 2 3 4
Pea~ A
B
C
D
E
F
1.4±0.1 1.4±0.1 1.3±0.1 1.5±0.2 1.4±0.1
2.2±0.2
2.2±0.2 2.0±0.2 2.2±0.1 2.0±0.2
2.8±0.2 2.6±* 2.6±0.1 2.6±0.2 2.5±0.1
3.5±0.2 3.5±0.1 3.4±0.3 3.4±0.1 3.4±0.2
4.4±0.2 4.2±0.3 4.2±0:3 4,4±0.2 4.1±0.3
5.6±0.3 5.5±0.3 5.6±0.3 5.9±0.5 5.5±0.4
303 7-
tive procedure to measure hepatic encephalopathy in
6
pigs could not be confirmed, in contrast to the opti-
MDF Hz
5
1
4 3 2 1 0
~6 power
1 O0
iI
90 80 70 60-
mistic results obtained with rats and rabbits8,12. The B A E P s were measured to obtain information about the conduction time in the brainstem during hepatic encephalopathy. In contrast to VEPs, B A E P s were highly reproducible. Five well-defined peaks could be obtained even in coma (grade 4). However, significant changes in the latency times of any of the peaks were not detected during progression of the hepatic encephalopathy. In conclusion, objective measurement of hepatic encephalopathy in pigs by means of spectral analysis of the E E G is possible. Evoked potentials, however, are not sufficiently sensitive to follow the course of hepatic encephalopathy in pigs.
.
50-
1
ACKNOWLEDGEMENTS
40 " 0
0 EEG-grades
Fig. 5. Spectral analysis data for pigs with various EEG grades of encephalopathy, expressed as the mean dominant frequency (MDF) (mean + S.D.) and the relative percentage of the delta power (mean _+ S.D.). The differences between grades 0, 1, 2 and 3 were significant (P < 0.01). positive and negative peaks and an increase in the latency time of peak VII. That V E P constitutes a sensi-
The authors are grateful to Prof. D.L. Westbroek and his staff of the Laboratory of Experimental Surgery, Rotterdam, The Netherlands for their invaluable help in performing the experiments, to Prof. M. de Vlieger and his staff for technical assistance with the performances of the E E G and evoked potentials and to Mrs. I. Pijpers for typing the manuscript. Supported by the Foundation for Medical Research F U N G O , The Netherlands.
REFERENCES 1 Conn, H.O. and Lieberthal, M.M., TheHepatic ComaSyndromes and Lactulose, Williams and Wilkins, Baltimore, 1979, pp. 20-28. 2 Hawkes, C.H., Macpherson, A.J.S., Pryor, H. and Townsend, H.R.A., The value of EEG frequency analysis in hepatic encephalopathy, J. R. Coll. Surg. Edinb., 15 (1970) 151-157. 3 Holm, E., Thiele, H., and Wolpert, E.M., Neurologische und psychiatrische Symptome bei akuten und chronische Lebererkrankungen, Therapiewoche, 30 (1980) 4790-4806. 4 Jewett, D.L., Romans, M.N. and Williston, J.S., Human auditory evoked potentials, Science, 167 (1970) 1517-1518. 5 Opolon, P., Lavallard, MC., Huguet, C.L., Bidalier, M., Granger, A., Gallot, D. and Bloeh, P., Hemodialysis vs. cross-dialysis in experimental hepatic coma, Surg. Gynecol. Obstet., 142 (1976) 845-854. 60polon, P., Rapin, J.R., Huguet, C., Hepatic failure coma treated by polyacrilonitrile membrane hemodialysis, Trans Am. Soc. Artif. Intern. Organs, XXII (1976) 701-711. 7 Parsons-Smith, B.G., Summerskill, W.H.J., Dawson,
8
9 10
11
12
13
A.M., Sherlock, S., The electroencephalogram in liver disease, Lancet, 2 (1957) 867-871. Schafer, D.F., Brody, L.E. and Jones, E.A., Visual evoked potentials: an objective measurement of hepatic encephaiopathy in the rabbit, Gastroenterology, 77 (1981) A38. Starr, A. and Achor, L.J., Auditory brainstem responses in neurological disease, Arch. Neurol., 32 (1975) 761-768. de Vlieger, M., Sadikoglu, S. and Eijndhoven, J.H.M., Visual evoked potentials, auditory evoked potentials and EEG in shunted hydrocephalic children, Neuropediatrics, 12 (1981) 55-61. Yaar, J., Shapiro, M.B. and Pottala, E.W., Spectral analysis of the EEG in hepatic encephaiopathy treated with levodopa, Electroenceph. Clin. Neurophysiol., 52 (1981) 617-625. Zeneroli, M.C., Penne, A. and ParrineUo, G., Comparative evaluation of visual evoked potentials in experimental hepatic encephalopathy and in pharmacologically induced coma-like states in rat, Life Sci., 28 (1981) 1507-1515. Romano, J. and Engel, G., Delirium: I. Encephalographic data, Arch. Neurol. Psychiatr., 51 (1944) 356-377.
Brain Research, 361)( 1985) 298-3()3 Elsevier
BRE 11277
Objective Measurement of Hepatic Encephalopathy in Pigs by Means of Spectral Analysis and Evoked Responses GERRIT H. DE GROOT, SOLKO W. SCHALM, MARINUS DE VLIEGER and CARIN C.D. VAN DER RIJT Departments of lnternal Medicine H and Electroneurology, Erasmus University, Rotterdam (The Netherlands) (Accepted April 2nd, 1985) Key words: acute hepatic failure - - hepatic encephalopathy - - electroencephalogram (EEG) - - spectral analysis - visual evoked potential - - brainstem auditory evoked potential
Objective measurement of hepatic encephalopathy by means of spectral analysis of the electroencephalogram (EEG), visualevoked potentials (VEP) and brainstem auditory-evoked potentials (BAEP) was studied in pigs with ischemic hepatic necrosis. The mean dominant frequency (MDF) and the relative power of the delta frequency band (%6 power) showed significant changes with increasing encephalographic grades of encephalopathy: MDF dropped from 7.0+ 0.5 (grade 0) to 2.7 + 0.3 Hz (grade 3 encephaiopathy) and %6 power increased from 52 + 7 (grade 0) to 83 _.+6% (grade 3). The patterns of the VEP in pigs corresponded to those of the human VEP. However, significant differences in either latency time of the peaks or the peak amplitude with increasing stages of he-' patic encephalopathy could not be found. The BAEP registered for pigs were reproducible but also were not as useful as spectral analysis for grading hepatic encephalopathy. INTRODUCTION Pigs are often used to study the pathogenesis or treatment of hepatic encephalopathy, because of their physiological similarity to the human being. Reliable methods for the objective measurement of encephalopathy in pigs are non-existent. Clinical neurological assessment of the grade of encephalopathy in pigs is subjective and unreliable, since it yields only a distinction between coma and non-coma. Conventional E E G readings have been proposed as a method for grading hepatic encephalopathy 1,5-7, but interpretation of the results remains subjective and therefore dependent on the experience and consistency of the electroneurologist. The introduction of quantitative analysis of the E E G offered the possibility of measuring hepatic encephalopathy objectively2,11. By this method the E E G can be expressed in terms of the mean dominant frequency (MDF) and the wave amplitude as a function of the frequency (power spectrum). Recently evoked potentials were introduced as another ob-
jective test for hepatic encephalopathy 4.s.9.J2. The present work was carried out to determine the reliability of quantitative E E G analysis a n d evoked potentials as a means of grading pigs with hepatic encephalopathy. MATERIALS AND METHODS Surgical preparation Fifteen large White pigs, weighing 28-33 kg, were used. Pigs underwent laparotomy for the introduction of a functional end-to-side portocaval shunt. All ligaments and peritoneal attachments to the liver were cut or coagulated. Subsequently a clamp was placed around the isolated hepatic artery. Five silver electrodes (diameter 1 mm) were inserted through the skull onto the dura for the recording of the E E G and evoked potentials: two over the frontal cortex, one m the vertex and two over the occipital cortex. The electrodes were anchored with sterile acrylic bone cement (Surgical Simplex, Howmedica, London. U.K.) and channeled percutaneously. Three
Correspondence: S.W. Schalm, Department of Internal Medicine II, University Hospital Dijkzigt, Dr. Molewaterplein 40, 3015 GD Rotterdam, The Netherlands. 0006-8993/85/$03.30© 1985 Elsevier Science Publishers B.V. (Biomedical Division)
299 days after construction of the portocaval shunt acute hepatic necrosis was induced by temporarily tightening the clamp to close off the hepatic artery for 4 or 6 h. Anesthesia was not used in this phase, since it may interfere with subsequent neurological assessments. Glucose (12 g/kg/24 h), penicillin G (9 megaU/24 h) and kanamycin (3 g/24 h) were infused continuously. The environmental temperature was maintained at 25 °C to ensure normothermia of the pigs with hepatic necrosis.
Recording procedure During the study periods the animals were placed in a dim room on their left side and covered with a sheet to ensure adequate relaxation. E E G recordings were taken before induction of hepatic necrosis and 24, 30, 48 and 54 h afterwards. It should be noted that pigs as soon as they tend to fall asleep have an 'abnormally' slow EEG. Therefore in order to obtain a meaningful E E G they had to be aroused each time. The potentials in 4 bipolar leads (left and right fronto-occipital, frontofrontal and occipito-occipital) were registered by an E E G apparatus (Ahrend v. Gogh, Amsterdam, The Netherlands). During 3 'epochs' of 100 s the signals of the fronto-occipital leads were stored simultaneously on a magnetic tape (Analog 7, Philips, Eindhoven, The Netherlands) for (quantitative) spectral analysis. The E E G channels were filtered with a band pass filter between 0.5 and 70 Hz. The stored data were fed into a computer (PDP 11/34) at a sampling rate of 51.2 Hz and a sensitivity of 11 bits/5 VI Each epoch of 100 s was divided into 10 periods. The power spectrum (the sum of wave amplitudes as a function of each frequency or frequency band) was calculated for each period of 10 s using the Fast Fourier Transformation, and the mean power spectrum for each 100 s epoch was obtained. The frequency resolution was 0.1 Hz. To minimize the effect of artefacts, only the range of 1.0-25.6 Hz was used to calculate the mean spectrum which was then analyzed (Fig. 1). The parameters calculated were the mean dominant frec]uency and the relative power of the delta, theta and alpha frequency range 3. The mean dominant frequency (MDF) was defined as: n
MDF = ~: fi Si/• i=l
i=l
Si
where fi is frequency i, S i is power of frequency i and n is the number of frequencies. The delta activity was defined as the activity of the frequencies 1.0-3.5 Hz, the theta activity that of the frequencies 3.5-8.0 Hz, the alpha activity that of the frequencies 8.0-13.0 Hz, and the beta activity that of the frequencies 13.0-25.6 Hz. The 3 epochs of 100 s were averaged. Spectra that showed high muscle activity in the beta frequency band or asymmetry between the two fronto-occipital leads were not included in the calculations. The E E G recordings were analyzed independently by an experienced electroneurologist; the 5 grade classification system described by Opolon et al.5,6 was used.
Visual-evokedpotentials (VEP) A photo stimulator with a xenon flashlight provided stimuli with a flash intensity of 1 J and a stimulation frequency of 1 Hz. The experiments were performed in a darkened room with the flashlight at a constant distance of 70 cm from the right eye. The EEG, monitored on an oscilloscope and on the E E G recorder, was stored on the magnetic tape during a 300 s artefact-free period. For the two fronto-occipital leads VEPs were obtained by averaging the response of 100 stimuli. Only the responses during the first 400 ms were used 10.
Brainstem-evokedpotentials (BAEP) Responses from electrodes at the vertex and ear lobes were recorded. Ten clicks/s were generated; the intensity of each click was 70 dB and it lasted 0.2 ms. Clicks were presented to the animal via an earphone. The average response in the 10 ms interval after the stimulus (n = 2000) was calculated by computer (PDP 11/34). Each animal was subjected to two mono-aural tests in a 5 min period.
Data evaluation Analysis of the VEP in normal pigs showed 7 waves (Fig. 2A). Peaks I, III, V and VII were positive, whereas peaks II, IV and VI were negative. All peaks occurred within 150 ms after visual stimulation. Analysis of the BAEP in pigs showed 5-6 positive waves within the first 6 ms (Fig. 3a). All waves found for normal animals were com-
300
100.0 power uV2/Hz.
100.0-
grade 0
power
~ ~
grade 3
uV2/Hz.
10.0
10.0-
1.0
1.0
0.1 0
I 5
I 10 freq.
I 15 [Hz.]
I 20
b
0. I 25
0
i
I
!
I
5
I0
15
20
25
freq. [Hz.]
Fig. 1. A visual representation of the EEG power spectrum (sum of all amplitudes as a function of the frequency) for normal pigs (a) and for pigs with hepatic encephalopathy(b).
pared with the waves found for animals with hepatic encephalopathy. Moreover each animal was used as its own control during the development of encephalopathy. The latency time was defined as the time (ms) between onset of the stimulus and the maximal amplitude of each peak. Amplitudes between two consecutive positive and negative peaks were measured. RESULTS
Survival Fifty percent of the pigs undergoing 4 h of ischemia of the liver survived. Two animals died from complications (air embolism, peritonitis) and two of hepatic coma. Six h of ischemia was followed by hepatic encephalopathy in all animals. Four of the 7 animals died of hepatic coma; one animal in hepatic coma, grade 4, was sacrificed after 72 h and two other animals had concurrent gastric hemorrhage.
EEG and spectral analysis (Figs, 4 and 5) The EEG, semiquantified into grades 1-5, grad-
ually deteriorated after induction of ischemia. Animals subjected to 4 h of ischemia developed mild encephalopathic E E G changes whereas 6 h of ischemia produced severe encephalopathic E E G changes within 30 h. The mean dominant frequency dropped significantly with each grade of eneephatopathy up to grade 3. No significant frequency differences were observed between grades 3 and 4 encephalopathy. The relative power of the delta activity was 52 + 7% in grade 0, and this increased significantly with each grade of encephalopathy up to grade 3 (83 _+ 6%). Delta activities for grades 3 and 4 did not differ significantly. An example of the power spectra for grade 0 and grade 3 encephalopathy is given in Fig. 1.
VEP (Table I) The latency time for all peaks was measured during several grades of encephalopathy. The overall results did not show significant differences in peak latency or changes in amplitude between grades 0 and 4 encephalopathy. However, in some animals there
301 2.0 u v
20.0 uv
grade : 0
VEP, g r a d e 0
"V VI I
2.0 UV 2 0 . O u v II
VEP, g r a d e 3 grade : 3
A
\ i"
I
\
\,,
B
, V, v,
.
~
'
'
Fig. 3. a: BAEP of a normal pig. b: B A E P of a pig with grade 3 hepatic encephalopathy. Iv Fig. 2. a: VEP of a normal pig. b: VEP of a pig with grade 3 hepatic encephalopathy.
during all grades of encephalopathy. The latency times and amplitudes of the 5 peaks did not change significantly with the grade of encephalopathy (Fig.
was an increase in latency time of peak VII and an decrease in amplitude between peak IV and V during encephalopathy (Fig. 2b).
3b). DISCUSSION
BAEP (Table II)
Since clinical grading of hepatic encephalopathy in pigs is subjective and difficult, Opolon et al. 5 intro-
As a rule the 5 positive peaks could be identified TABLE I
VEP peak latency times (ms, mean + S.D.) in relation to the grade of encephalopathy + + , no definite peak discernable. *, insufficient data to calculate S.D.
Encephalopathy
Peaks
grade
I
H
II1
IV
V
VI
VII
0 1 2 3 4
33+8 31 + 2 24 + * 24 + * ++
40+5 36 + 5 35 + 5 38 + * 36 + 4
54+6 52 + 9 49 + 7 54 + 6 48 + 9
82+11 82 + 19 76 + 14 89 + 17 ++
107+18 101 + 15 103 + 12 122 + 27 ++
133+25 116 + 16 125 + 12 131 + * ++
154+17 151 + 10 143 + 15 160 + * 153 + 34
302
5
4 hours of ischemia
5
4
4
~3 D~
3
~ 2
2
LU
I
I
0
O-
I
I
I
I
I
I
I
I
I
1
0
24
30
48
54
0
24
30
48
54
n o u r s a f t e r i n d u c t i o n of ischemia
h o u r s a f t e r i n d u c t i o n of ischemia
Fig. 4. The course of the EEG grades for pigs that have undergone 4 and 6 h of ischemia of the liver. The 4-h animals with an EEG at 54 h survived 72 h and were then sacrificed. Clinicallythe 4 animals were alert at 72 h. The 6-h animal with an EEG at 54 h survived till 72 h with grade 4-5 encephalopathy.
duced 5 grades of encephalopathy based on conventional E E G recordings. However, the interpretation of the E E G remains subjective and dependent on the expertise of the observer. Our study showed that objective measurement of hepatic encephalopathy in pigs is possible by means of spectral analysis. The EEG activity, represented by the mean dominant frequency, exhibited a fairly close correlation with the grade of encephalopathy. The relative power of the delta band expressed as a percentage of the total power of the alpha, beta, theta and delta bands, increased significantly with deterioration of the EEG, reflecting an increase in slow wave activity. Although the mean values of MDF and percent delta power were significantly different for each grade of encephalopathy, individual measurements showed some
overlapping. Therefore both parameters should be used together to classify the grade of encephalopathy, although our results indicate that a reliable distinction between grades 3 and 4 encephalopathy was not possible. The change in E E G as displayed with spectral analysis are probably not specific for hepatic encephalopathy, but may also be observed in other metabolic conditions1.13. Since evoked potentials are said to allow differentiations of etiologies 12. visualand auditory-evoked potentials were also analyzed. VEPs in pigs have the 7 main components present in man 10. However, the reproducibility of the latency times and the amplitudes of two consecutive positive and negative peaks for each grade of encephalopathy was rather poor. In some individual animals we observed a decrease in amplitudes of two consecutive
TABLE II B A E P peak latency times (ms, mean + S.D. ) in relation to the grade o f encephalopathy
*, insufficient data to calculate S.D. Encep~lopathy gra~
0 1 2 3 4
Pea~ A
B
C
D
E
F
1.4±0.1 1.4±0.1 1.3±0.1 1.5±0.2 1.4±0.1
2.2±0.2
2.2±0.2 2.0±0.2 2.2±0.1 2.0±0.2
2.8±0.2 2.6±* 2.6±0.1 2.6±0.2 2.5±0.1
3.5±0.2 3.5±0.1 3.4±0.3 3.4±0.1 3.4±0.2
4.4±0.2 4.2±0.3 4.2±0:3 4,4±0.2 4.1±0.3
5.6±0.3 5.5±0.3 5.6±0.3 5.9±0.5 5.5±0.4
303 7-
tive procedure to measure hepatic encephalopathy in
6
pigs could not be confirmed, in contrast to the opti-
MDF Hz
5
1
4 3 2 1 0
~6 power
1 O0
iI
90 80 70 60-
mistic results obtained with rats and rabbits8,12. The B A E P s were measured to obtain information about the conduction time in the brainstem during hepatic encephalopathy. In contrast to VEPs, B A E P s were highly reproducible. Five well-defined peaks could be obtained even in coma (grade 4). However, significant changes in the latency times of any of the peaks were not detected during progression of the hepatic encephalopathy. In conclusion, objective measurement of hepatic encephalopathy in pigs by means of spectral analysis of the E E G is possible. Evoked potentials, however, are not sufficiently sensitive to follow the course of hepatic encephalopathy in pigs.
.
50-
1
ACKNOWLEDGEMENTS
40 " 0
0 EEG-grades
Fig. 5. Spectral analysis data for pigs with various EEG grades of encephalopathy, expressed as the mean dominant frequency (MDF) (mean + S.D.) and the relative percentage of the delta power (mean _+ S.D.). The differences between grades 0, 1, 2 and 3 were significant (P < 0.01). positive and negative peaks and an increase in the latency time of peak VII. That V E P constitutes a sensi-
The authors are grateful to Prof. D.L. Westbroek and his staff of the Laboratory of Experimental Surgery, Rotterdam, The Netherlands for their invaluable help in performing the experiments, to Prof. M. de Vlieger and his staff for technical assistance with the performances of the E E G and evoked potentials and to Mrs. I. Pijpers for typing the manuscript. Supported by the Foundation for Medical Research F U N G O , The Netherlands.
REFERENCES 1 Conn, H.O. and Lieberthal, M.M., TheHepatic ComaSyndromes and Lactulose, Williams and Wilkins, Baltimore, 1979, pp. 20-28. 2 Hawkes, C.H., Macpherson, A.J.S., Pryor, H. and Townsend, H.R.A., The value of EEG frequency analysis in hepatic encephalopathy, J. R. Coll. Surg. Edinb., 15 (1970) 151-157. 3 Holm, E., Thiele, H., and Wolpert, E.M., Neurologische und psychiatrische Symptome bei akuten und chronische Lebererkrankungen, Therapiewoche, 30 (1980) 4790-4806. 4 Jewett, D.L., Romans, M.N. and Williston, J.S., Human auditory evoked potentials, Science, 167 (1970) 1517-1518. 5 Opolon, P., Lavallard, MC., Huguet, C.L., Bidalier, M., Granger, A., Gallot, D. and Bloeh, P., Hemodialysis vs. cross-dialysis in experimental hepatic coma, Surg. Gynecol. Obstet., 142 (1976) 845-854. 60polon, P., Rapin, J.R., Huguet, C., Hepatic failure coma treated by polyacrilonitrile membrane hemodialysis, Trans Am. Soc. Artif. Intern. Organs, XXII (1976) 701-711. 7 Parsons-Smith, B.G., Summerskill, W.H.J., Dawson,
8
9 10
11
12
13
A.M., Sherlock, S., The electroencephalogram in liver disease, Lancet, 2 (1957) 867-871. Schafer, D.F., Brody, L.E. and Jones, E.A., Visual evoked potentials: an objective measurement of hepatic encephaiopathy in the rabbit, Gastroenterology, 77 (1981) A38. Starr, A. and Achor, L.J., Auditory brainstem responses in neurological disease, Arch. Neurol., 32 (1975) 761-768. de Vlieger, M., Sadikoglu, S. and Eijndhoven, J.H.M., Visual evoked potentials, auditory evoked potentials and EEG in shunted hydrocephalic children, Neuropediatrics, 12 (1981) 55-61. Yaar, J., Shapiro, M.B. and Pottala, E.W., Spectral analysis of the EEG in hepatic encephaiopathy treated with levodopa, Electroenceph. Clin. Neurophysiol., 52 (1981) 617-625. Zeneroli, M.C., Penne, A. and ParrineUo, G., Comparative evaluation of visual evoked potentials in experimental hepatic encephalopathy and in pharmacologically induced coma-like states in rat, Life Sci., 28 (1981) 1507-1515. Romano, J. and Engel, G., Delirium: I. Encephalographic data, Arch. Neurol. Psychiatr., 51 (1944) 356-377.