- Email: [email protected]
Effects of intravenous human albumin and furosemide on EEG recordings in patients with intracerebral hemorrhage
Clinical Neurophysiology 113 (2002) 454–458 www.elsevier.com/locate/clinph
Effects of intravenous human albumin and furosemide on EEG recordings in patients with intracerebral hemorrhage Zuchun Huang*, Weiwei Dong, Yong Yan, Qifeng Xiao, Yuan Man Department of Neurology, The First Affiliated Hospital, Chongqing University of Medical Sciences, Chongqing 400016, People’s Republic of China Accepted 20 December 2001
Abstract Objectives: To characterize alterations in continuous EEG monitoring that occurs during and after intravenous infusion of human albumin or furosemide in patients with intracerebral hemorrhage. Methods: Patients were rapidly administered 20% human albumin 50 ml or furosemide 40 mg intravenously with continuous EEG monitoring for 3 h before and after drug infusion in the neurological intensive care unit. Visual and spectral analyses of EEG recordings before and after mannitol administration were carried out. Results: The study consisted of 20 patients. Of 14 patients with human albumin treatment, a decrease in the slowing activity was visually noted in 9 cases after the drug infusion. The spectral analysis demonstrated that albumin-induced EEG changes increased in alpha power and decreased in delta power in the lesion hemispheres, especially in the central and middle temporal areas. The effects occurred after 30 min and were maximal 1 h after the end of the infusion, then remained significant for 2 h post-infusion. Of 6 patients with furosemide treatment, the EEG recordings before, during, and after the furosemide infusion were not statistically significantly different by visual and quantitative analyses. Conclusions: The results support the opinion that the available EEG monitoring techniques offer an inexpensive, non-invasive, and consistently reproducible technique for reflecting the therapeutic effects of therapeutics in lowering ICP and antiedema in stroke patients. q 2002 Elsevier Science Ireland Ltd. All rights reserved. Keywords: Brain edema; CT scan; Furosemide; Human albumin; Intensive care unit; Intracerebral hemorrhage; Monitoring EEG; Spectral analysis
1. Introduction EEG can provide a sensitive indication of cerebral metabolism, ischemia, hypoxia, and neural dysfunction (Jordan, 1993; Agarwal et al., 1998). Many neurointensivists share the opinion that EEG can be an integral part of monitoring in the intensive care unit (ICU) (Bricolo et al., 1978: Emmerson and Chiappa, 1988; Labar et al., 1991; Lesser et al., 1992; Jordan, 1993; Nuwer, 1997; Agarwal et al., 1998; Gotman et al., 1998). EEG changes induced by an intravenous dose of 20% mannitol 250 ml in patients with cerebrovascular disease (CVD) had been studied by Huang (1997). This study suggested that mannitol treatment could improve the slowing abnormal EEG as it exerts the action of dehydration of brain and reduces the increased intracranial pressure (ICP) in stroke patients with brain edema and raised ICP. * Corresponding author. Tel.: 186-23-6889-4268 (O) / 6889-5861 (H); fax: 186-23-6881-2985. E-mail address: [email protected] (Z. Huang).
This paper studied the correlation of computerized tomography (CT) scans, clinical evaluations, and the EEG changes before and after intravenous administration of human albumin and furosemide in the patients with intracerebral hemorrhage (ICH). The objectives of the study were to: (1) determine if the drug-induced EEG changes correlated clinically with the extent of brain edema and increased ICP; (2) evaluate EEG findings produced by the treatment in the patients. The hypothesis is that albumin- or furosemide-induced dehydration of brain and/or lowering of ICP is related to a decrease in delta and theta activity in the EEG recordings. 2. Materials and methods 2.1. Subjects The study was carried out in the Department of Neurology, First Affiliated Hospital, Chongqing University of Medical Sciences, China. Stroke patients admitted to the neurological ICU (NICU) between May 1995 and Septem-
1388-2457/02/$ - see front matter q 2002 Elsevier Science Ireland Ltd. All rights reserved. PII: S13 88- 2457(02)0001 5-9
CLINPH 2001685
Z. Huang et al. / Clinical Neurophysiology 113 (2002) 454–458
ber 2000 were assessed for selection in the study. Most patients were admitted through the emergency department, although others were direct transfers from other hospitals to the NICU. The attending neurologist and neurointensivist evaluated all patients upon admission to the NICU. The Glasgow Coma Scale (GCS) was performed on admission, and the functional outcome at discharge was assessed using the Barthel Index (BI). Non-contrasted cranial CT scans were performed in all cases. Each qualified patient was given screening examinations that included: two-lead electrocardiogram (ECG); complete blood count; blood chemistries (sodium, potassium, aminotransferase, asparate aminotransferase, gamma-glutamyltransferase, alkaline phosphatase, total bilirubin, and total protein); urine analysis; blood pressure (BP), and heart rate (HR) measurements. There were 3 criteria for selection of patients in the study: (1) ganglio-thalamic hemorrhage; (2) hospital admission within 2 weeks of stroke onset; and (3) no intravenous or ingested drugs 8 h prior to the participation in the study. Patients were excluded from the study if they: (1) had infections, cardiac, renal or liver disease, serum electrolyte disturbance, hypertension crisis; (2) had ICH related to trauma, coagulopathy, tumor, or arteriovenous malformation; (3) were less than 20 years of age or over 80 years old; (4) could not establish an accurate historical time of the stroke. The study was approved by the hospital Medical Ethics Committee. Informed consents were obtained from patients, nearest of kin, or both.
2.2. Experimental design Each patient was continuously monitored for 3 h by EEG, ECG, BP, HR, and oxygen saturation (SaO2) pre- and postdrug administration. Each patient received a 20 min baseline EEG before insertion of an intravenous catheter in an upper extremity. The vein was maintained with infusion of normal saline (NS) at 100 ml per 20 min. After baseline studies were completed, 20% albumin 50 ml or furosemide 40 mg was infused over 20 min or 5 min, respectively. The osmotic diuresis was monitored with a urinometer. Serum osmolarity, electrolytes, and BUN were sampled 1 h after the termination of the infusion. The Timeline is shown in Fig. 1 of experimental chart design. The EEG recordings were assessed visually and quantitatively in the baseline, 1 min pre-drug infusion, 1 min postdrug infusion, and at post-infusion times of 10, 30, 60, and 120 min.
Fig. 1. The chart of experimental design for EEG recordings before, during, and after the drug. Note: ‘O’ represents EEG sampling time.
455
2.3. EEG acquisition Eight channels of EEG were recorded from scalp disc electrodes at International 10/20 System Placement with a VISH Digital Electroencephalogram (VISH Biomedical Instrument, China.) The EEG was recorded from Fp3 to 4, C3 to 4, T3 to 4, and O1 to 2 using linked earlobes as reference. Closed-eye EEG recordings in the NICU were continuously displayed oscilloscopically and stored on magnetic disc for later visual and spectral analysis. 2.4. Visual assessment The baseline abnormalities and the pre- and post-drug infusion EEG observations were evaluated independently by two experienced electroencephalographers. The baseline EEGs were considered abnormal if there were intermittent slow activity (including lateralized or generalized intermittent delta activity – IDA), and persistent slow activity with paroxysmal epileptogenic abnormalities. The abnormal EEG patterns were divided into generalized, lateralized, and generalized with one-sided dominance, if findings were prominent in one hemisphere. 2.5. Spectral analysis Three 4 s EEG epochs from 8 channels of 1 min sampling time were selected and digitized by a 12 bit analog to digital converter at the rate of 200 Hz. All epochs had minimal artifacts. The Fast Fourier Transform performed spectral analysis on each epoch. The spectrum was divided into 5 frequency bands: delta (0.5–3.5 Hz), theta (4–7.5 Hz), alpha1 (8–10.5 Hz), alpha2 (11–12.5 Hz), and beta (13– 30 Hz). The absolute power in each of the 5 frequency bands was obtained for each monopolar lead. The mean absolute power in each band of the patient’s recording was then determined for each sampling time at 8 sites. With respect to the baseline condition, the percentages of drug-induced EEG changes in absolute power of each frequency band of each sampling time for 8 sites were computed. 2.6. CT scans As soon as the patients were medically stable, CT imaging was performed by an SCT-3000 TX Scanner (Japan) in 5 mm slices oriented parallel to the orbitomeatal line. Volume measurements expressed in milliliters were taken from CT scans by planimetry (Compensating Polar Planimeters Model 10) of ICH, lesions (ICH plus perihematoma low-density areas), and intraventricular hemorrhage (IVH). Lesion areas were calculated on each slice separately by tracing the perimeter of the appropriate high or low attenuation zone on the CT console. These values were multiplied by the slice thickness to yield the total lesion volume. Perihematoma low-density volume was calculated
456
Z. Huang et al. / Clinical Neurophysiology 113 (2002) 454–458
Table 1 Baseline characteristics of 20 patients with intracerebral hemorrhage
Gender, n Age, years (range) Admission MABP mmHg (range) Admission mental status, n Alert Drowsy Stupor Coma Admission GCS (range) CT interval from onset, days EEG interval from onset, days Interval between CT and EEG examinations, h CT results ICH volume, ml Lesions volume, ml Perilesional hypodense region volume, ml Midline shift, mm IVH volume, ml Oxygen satiation, % Heart rate, bpm Barthel Index at discharge Length of stay, days
Albumin group
Furosemide group
8M, 6F 63.8 ^ 11.5 (45–68) 139 ^ 29 (73–225)
3M, 3F 60 ^ 9.6 (56–71) 141 ^ 31 (78–214)
6 3 2 3 11.3 (15–3) 4.8 (6 h–2 weeks) 5.4 (7 h–2 weeks) 4.2 (1–18)
3 1 1 1 13 (15–6) 4.3 (5 h–13 days) 6.2 (8 h–14 days) 4.7 (1–20)
16.2 (4.1–56.5) 24.9 (4.1–87.4) 8.6 (0–30.9) 1.3 (0.6–8.7) 10.2 (3.5–28.6) 98.7 ^ 1.8 76.8 ^ 15.9 71.4 (9–100) 34.6 (18–46)
15.6 (5.4–50.6) 22.4 (5.4–80.4) 8.2 (0–31.5) 1.1 (0.5–1.6) 6.9 (4.5–9.2) 99 ^ 1.2 77.3 ^ 16.2 72.6 (9–100) 30.4 (18–35.2)
by subtracting ICH volume from the lesion volume. The extent of midline shift was measured according to Ropper (1986).
group, the serum osmolality pre-furosemide ranged from 280 to 300 mOsm/l in 4 cases. Changes in serum sodium, serum potassium, and BUN were not statistically significant.
2.7. Statistical analysis
3.1. Albumin treatment group
The analyses were carried out using the Statistical Analysis System. One-way analysis of variance (ANOVA) was used to detect differences between the mean percentages of drug-induced EEG changes and EEG changes during baseline recordings in each band at each site (F, T, C, or O) in the hemispheres of all patients. If the ANOVA yielded significant results, Dunnett tests were performed to compare the mean percentages of EEG changes during baseline recordings with the EEG changes by drug at different sampling times. The results are expressed as the means and SD. A two-tailed P , 0:05 was considered significant.
3.1.1. Visual assessment
3. Results Baseline characteristics of patients are shown in Table 1, where 20 patients with ICH were studied. Fourteen of of the 20 patients were treated with human albumin and 6 with furosemide. No evident change in ECG, BP, or SaO2 was observed after albumin treatment. Post-treatment diuresis averaged 364 and 785 ml, over a 2 h period, after infusion of albumin and furosemide, respectively. Increase in serum osmolality ranged from 4 to 12 mOsm/l (mean, 6.1 mOsm/l) 1 h post-albumin infusion in 8 cases (serum osmolality ranged from 280 to 298 mOsm/l pre-albumin). No change in serum osmolality was found in the furosemide treatment
3.1.1.1. EEG pre-albumin. The EEG findings pre-testing were abnormal in 12/14 patients, including lateralized abnormalities in 4 cases, generalized slowing activity with one-side dominance in 5, and diffuse slowing activity in 3. Placebo infusion produced no significant change in 0.5– 30 Hz EEG activity during the monitoring period. 3.1.1.2. EEG post-albumin. After the administration of albumin, unchanged EEG was noted in 5 cases. Two of them had normal EEGs pre-drug testing, 3 had an ICH volume less than 10 ml, and 3 were alert. In 9 cases, the slowing abnormalities of EEG had definitely decreased after the infusion. The changes were usually noted 10–30 min post-infusion of albumin, but the obvious change occurred 1 h post-infusion. Of these cases 6 had a GCS score less than 11, 7 had an ICH volume more than 10 ml, 6 had a lesion volume more than 30 ml, 5 had midline shift in CT scan, 4 had ventricular penetration of the hematoma in CT scan. In 3 patients with diffuse abnormal baseline EEG, clear improvement of the EEGs in the nonlesion hemispheres was also noted. 3.1.2. QEEG analysis In Fig. 2 the mean percentage values of albumin-induced
Z. Huang et al. / Clinical Neurophysiology 113 (2002) 454–458
457
Fig. 2. The mean percentage of albumin-induced EEG changes in absolute power at central area in the lesion hemispheres of 14 stroke patients. The percentage of EEG changes in alpha and theta power changed 30 min post-infusion and reached their maximal level 1 h post-infusion of albumin, then remained significant until the 2 h post-infusion. Each point is the mean ^ SD for all subjects at the corresponding time. * or 1 indicate significant differences (P , 0:01 or P , 0:05) from EEG changes by placebo.
absolute power changes at central area of lesion hemispheres with respect to the baseline conditions are shown. The graft demonstrates significant changes in EEGs which occurred 30 min post-albumin infusion, and were at a maximum 1 h post-infusion, and remained significant for 2 h post-infusion. The EEG changes in the non-lesion hemisphere were similar to those which occurred at the lesion hemisphere, but the effects were not significant. 3.2. Furosemide treatment group The EEG findings pre-furosemide testing were normal in two cases, and abnormal in 4 with lateralized abnormalities in two cases, generalized slowing activity with one-side dominance in one, and diffuse slowing activity in one. However, the EEG recordings before, during, and after the infusion were not statistically significantly different by visual and spectral analysis in all 6 cases.
4. Discussion 4.1. Albumin and furosemide treatment for ICH patients with increased ICP Osmotic and renal diuretics are used to reduce ICP in
patients with brain edema or raised ICP. Human albumin solution is commonly used in treating low serum albumin concentrations and hypovolemia. It effectively replaces volume and support colloid osmotic pressure. Currently, the licensed indications for use of albumin are emergency treatment of shock, acute management of burns, and clinical situations associated with hypoproteinemia (McClelland, 1990). Stroke patients in late stage have the pathophysiological basis of hypoalbuminemia, albumin given for correcting the problem will help to control the brain edema. Furosemide is a potent renal (loop) diuretic that has been used to reduce ICP in animals and in man (Cottrell et al., 1977; Tornheim et al., 1979; Schettini et al., 1982). It could also prolong and intensify the effect of mannitol, so as to reduce the dosage of mannitol used (Roberts et al., 1987). And if it were used with albumin, the treatment would be more effective and safe due to albumin maintaining the colloid osmotic pressure in blood, which causes normal blood volume dehydration. The effect of mannitol became less after it was intravenously administrated repeatedly. In this case, furosemide, albumin, and furosemide plus mannitol or albumin alternatively may be a good choice (Lehman, 1990; Pickard and Czosnyka, 1995). 4.2. EEG changes by human albumin and furosemide In the present study, the abnormal EEGs evidently
458
Z. Huang et al. / Clinical Neurophysiology 113 (2002) 454–458
improved by albumin in most coma cases with large hemorrhage and evident ‘mass effects’ in the CT scan. The result was similar to our previous studies on stroke patients with mannitol, which suggest that the effects of improving abnormal EEG by albumin in stroke patients were reflecting the work of drug lowering ICP and dehydrating of brain edema. According to the time the EEG changes occur, the effects of albumin began 30 min post-infusion of 10 g albumin (20% albumin 50 ml) and were maximal 1 h post-infusion, then remained significant for 2 h post-infusion. Compared with the EEG-pharmacodynamic characteristics of intravenous mannitol in stroke patients, the onset of albumin effect was little later, but its duration was longer. In view of our findings we conclude that albumin produced expected changes in monitoring EEG, whereas furosemide produced a large diuresis without a concomitant decrease in slowing abnormalities in EEG. This difference is consistent with the fact that, unlike mannitol, furosemide does not increase renal free water clearance and ‘dehydrate’ the brain (Buhrley and Reed, 1972; Pinegin et al., 1979; Rottenberg and Posner, 1980). Cascino et al. (1983) were also unable to demonstrate statistically significant CT density shifts in edematous white matter in 4 brain tumor patients with furosemide treatment. The small samples and the dose used for testing in our study also contributed to the present result mentioned above. Further studies will be carried out in the future. No previous study has undertaken to monitor EEG changes in stroke patients following the intravenous administration of albumin or furosemide and relate observed changes in the monitoring EEGs to the effects of antiedema or lowering ICP treatment. Our results indicate that available EEG monitoring technique is sensitive enough to monitor the effects of antiedema and lowering ICP therapy in vivo. Further studies on various therapeutic interventions in the related disease group can help this method be used reasonably and effectively in clinic practice.
Acknowledgements We wish to express our appreciation to J.D. Bell DO of the USA, Prof. Yanping Hu for their technical collaborations.
References Agarwal R, Gotman J, Flanagan D, Rosenblatt B. Automatic EEG analysis during long-term monitoring in the ICU. Electroenceph clin Neurophysiol 1998;107:44–58. Bricolo A, Turazzi S, Faccioli F, Sciaretta G, Erculiani P. Clinical application of compressed spectral array in long-term EEG monitoring of comatose patients. Electroenceph clin Neurophysiol 1978;45:211–225. Buhrley LE, Reed DJ. The effect of furosemide on sodium-22 uptake into cerebrospinal fluid and brain. Exp Brain Res 1972;14:503–510. Cascino T, Baglivo J, Conti J, Szewczykowski J, Posner JB, Rottenberg DA. Quantitative CT assessment of furosemide- and mannitol-induced changes in brain water contents. Neurology 1983;33:898–903. Cottrell JE, Robustelli A, Post K, Turndorf H. Furosemide- and mannitolinduced changes in intracranial pressure and serum osmolality and electrolytes. Anesthesiology 1977;47:28–30. Emmerson RG, Chiappa KH. Electrophysiologic monitoring. In: Ropper AH, Kennedy SK, Zervas NT, editors. Neurologic and neurosurgical intensive care, 2nd ed. Rockville, MD: Aspen, 1988. p. 13. Gotman YSJ, Pasupathy A, Flanagan D, Rosenblatt B, Gottesman R. An expert system for EEG monitoring in the pediatric intensive care unit. Electroenceph clin Neurophysiol 1998;106:488–500. Huang ZC. Effect of intravenous mannitol on EEG of the patients with acute stroke. Electroenceph clin Neurophysiol 1997;103(1):199. Jordan KG. Continuous EEG and evoked potential monitoring in the neuroscience intensive care unit. J Clin Neurophysiol 1993;10:445–475. Labar DR, Fisch BJ, Pedley TA, Fink ME, Solomon RA. Quantitative EEG monitoring for patients with subarachnoid hemorrhage. Electroenceph clin Neurophysiol 1991;78:325–332. Lehman LB. Intracranial pressure monitoring and treatment: a contemporary view. Ann Emerg Med 1990;19:295–303. Lesser RP, Webber WRS, Fisher RS. Design principles for computerized EEG monitoring. Electroenceph clin Neurophysiol 1992;82:239–247. McClelland DBL. Human albumin solutions. BMJ 1990;300(6):35–37. Nuwer M. Assessment of digital EEG, quantitative EEG, and EEG brain mapping: report of the American Academy of Neurology and the American Clinical Neurophysiology Society. Neurology 1997;49:277–292. Pickard LD, Czosnyka M. Management of raised intracranial pressure. J Neurol Neurosurg Psychiatry 1995;56:845–858. Pinegin LE, Tibekina LM, Shakhmatova EI, Natochin YV. The effects of altered osmolality and intracellular fluid volume on brain water and electrolytes. Fiziol Zh SSSR 1979;65:122–127. Roberts PA, Pollay M, Engles C, Pendleton B, Reyndds E, Stevens FA. Effect on intracranial pressure of furosemide combined with varying doses and administration rates of mannitol. J Neurosurg 1987;66:440–446. Ropper AH. Lateral displacement of the brain and level of consciousness in patients with an acute hemispherical mass. N Engl J Med 1986;314: 953–958. Rottenberg DA, Posner JB. Intracranial pressure control. In: Cottrell JE, Turndorf H, editors. Anesthesia and neurosurgery, St. Louis, MO: Mosby, 1980. pp. 89–118. Schettini A, Stahurski B, Young HF. Osmotic and osmotic-loop diuresis in brain surgery – effects on plasma and CSF electrolytes and ion excretion. J Neurosurg 1982;56:679–684. Tornheim PA, Mclaurin RL, Sawaya R. Effect of furosemide on experimental traumatic cerebral edema. Neurosurgery 1979;4:48–52.
Effects of intravenous human albumin and furosemide on EEG recordings in patients with intracerebral hemorrhage Zuchun Huang*, Weiwei Dong, Yong Yan, Qifeng Xiao, Yuan Man Department of Neurology, The First Affiliated Hospital, Chongqing University of Medical Sciences, Chongqing 400016, People’s Republic of China Accepted 20 December 2001
Abstract Objectives: To characterize alterations in continuous EEG monitoring that occurs during and after intravenous infusion of human albumin or furosemide in patients with intracerebral hemorrhage. Methods: Patients were rapidly administered 20% human albumin 50 ml or furosemide 40 mg intravenously with continuous EEG monitoring for 3 h before and after drug infusion in the neurological intensive care unit. Visual and spectral analyses of EEG recordings before and after mannitol administration were carried out. Results: The study consisted of 20 patients. Of 14 patients with human albumin treatment, a decrease in the slowing activity was visually noted in 9 cases after the drug infusion. The spectral analysis demonstrated that albumin-induced EEG changes increased in alpha power and decreased in delta power in the lesion hemispheres, especially in the central and middle temporal areas. The effects occurred after 30 min and were maximal 1 h after the end of the infusion, then remained significant for 2 h post-infusion. Of 6 patients with furosemide treatment, the EEG recordings before, during, and after the furosemide infusion were not statistically significantly different by visual and quantitative analyses. Conclusions: The results support the opinion that the available EEG monitoring techniques offer an inexpensive, non-invasive, and consistently reproducible technique for reflecting the therapeutic effects of therapeutics in lowering ICP and antiedema in stroke patients. q 2002 Elsevier Science Ireland Ltd. All rights reserved. Keywords: Brain edema; CT scan; Furosemide; Human albumin; Intensive care unit; Intracerebral hemorrhage; Monitoring EEG; Spectral analysis
1. Introduction EEG can provide a sensitive indication of cerebral metabolism, ischemia, hypoxia, and neural dysfunction (Jordan, 1993; Agarwal et al., 1998). Many neurointensivists share the opinion that EEG can be an integral part of monitoring in the intensive care unit (ICU) (Bricolo et al., 1978: Emmerson and Chiappa, 1988; Labar et al., 1991; Lesser et al., 1992; Jordan, 1993; Nuwer, 1997; Agarwal et al., 1998; Gotman et al., 1998). EEG changes induced by an intravenous dose of 20% mannitol 250 ml in patients with cerebrovascular disease (CVD) had been studied by Huang (1997). This study suggested that mannitol treatment could improve the slowing abnormal EEG as it exerts the action of dehydration of brain and reduces the increased intracranial pressure (ICP) in stroke patients with brain edema and raised ICP. * Corresponding author. Tel.: 186-23-6889-4268 (O) / 6889-5861 (H); fax: 186-23-6881-2985. E-mail address: [email protected] (Z. Huang).
This paper studied the correlation of computerized tomography (CT) scans, clinical evaluations, and the EEG changes before and after intravenous administration of human albumin and furosemide in the patients with intracerebral hemorrhage (ICH). The objectives of the study were to: (1) determine if the drug-induced EEG changes correlated clinically with the extent of brain edema and increased ICP; (2) evaluate EEG findings produced by the treatment in the patients. The hypothesis is that albumin- or furosemide-induced dehydration of brain and/or lowering of ICP is related to a decrease in delta and theta activity in the EEG recordings. 2. Materials and methods 2.1. Subjects The study was carried out in the Department of Neurology, First Affiliated Hospital, Chongqing University of Medical Sciences, China. Stroke patients admitted to the neurological ICU (NICU) between May 1995 and Septem-
1388-2457/02/$ - see front matter q 2002 Elsevier Science Ireland Ltd. All rights reserved. PII: S13 88- 2457(02)0001 5-9
CLINPH 2001685
Z. Huang et al. / Clinical Neurophysiology 113 (2002) 454–458
ber 2000 were assessed for selection in the study. Most patients were admitted through the emergency department, although others were direct transfers from other hospitals to the NICU. The attending neurologist and neurointensivist evaluated all patients upon admission to the NICU. The Glasgow Coma Scale (GCS) was performed on admission, and the functional outcome at discharge was assessed using the Barthel Index (BI). Non-contrasted cranial CT scans were performed in all cases. Each qualified patient was given screening examinations that included: two-lead electrocardiogram (ECG); complete blood count; blood chemistries (sodium, potassium, aminotransferase, asparate aminotransferase, gamma-glutamyltransferase, alkaline phosphatase, total bilirubin, and total protein); urine analysis; blood pressure (BP), and heart rate (HR) measurements. There were 3 criteria for selection of patients in the study: (1) ganglio-thalamic hemorrhage; (2) hospital admission within 2 weeks of stroke onset; and (3) no intravenous or ingested drugs 8 h prior to the participation in the study. Patients were excluded from the study if they: (1) had infections, cardiac, renal or liver disease, serum electrolyte disturbance, hypertension crisis; (2) had ICH related to trauma, coagulopathy, tumor, or arteriovenous malformation; (3) were less than 20 years of age or over 80 years old; (4) could not establish an accurate historical time of the stroke. The study was approved by the hospital Medical Ethics Committee. Informed consents were obtained from patients, nearest of kin, or both.
2.2. Experimental design Each patient was continuously monitored for 3 h by EEG, ECG, BP, HR, and oxygen saturation (SaO2) pre- and postdrug administration. Each patient received a 20 min baseline EEG before insertion of an intravenous catheter in an upper extremity. The vein was maintained with infusion of normal saline (NS) at 100 ml per 20 min. After baseline studies were completed, 20% albumin 50 ml or furosemide 40 mg was infused over 20 min or 5 min, respectively. The osmotic diuresis was monitored with a urinometer. Serum osmolarity, electrolytes, and BUN were sampled 1 h after the termination of the infusion. The Timeline is shown in Fig. 1 of experimental chart design. The EEG recordings were assessed visually and quantitatively in the baseline, 1 min pre-drug infusion, 1 min postdrug infusion, and at post-infusion times of 10, 30, 60, and 120 min.
Fig. 1. The chart of experimental design for EEG recordings before, during, and after the drug. Note: ‘O’ represents EEG sampling time.
455
2.3. EEG acquisition Eight channels of EEG were recorded from scalp disc electrodes at International 10/20 System Placement with a VISH Digital Electroencephalogram (VISH Biomedical Instrument, China.) The EEG was recorded from Fp3 to 4, C3 to 4, T3 to 4, and O1 to 2 using linked earlobes as reference. Closed-eye EEG recordings in the NICU were continuously displayed oscilloscopically and stored on magnetic disc for later visual and spectral analysis. 2.4. Visual assessment The baseline abnormalities and the pre- and post-drug infusion EEG observations were evaluated independently by two experienced electroencephalographers. The baseline EEGs were considered abnormal if there were intermittent slow activity (including lateralized or generalized intermittent delta activity – IDA), and persistent slow activity with paroxysmal epileptogenic abnormalities. The abnormal EEG patterns were divided into generalized, lateralized, and generalized with one-sided dominance, if findings were prominent in one hemisphere. 2.5. Spectral analysis Three 4 s EEG epochs from 8 channels of 1 min sampling time were selected and digitized by a 12 bit analog to digital converter at the rate of 200 Hz. All epochs had minimal artifacts. The Fast Fourier Transform performed spectral analysis on each epoch. The spectrum was divided into 5 frequency bands: delta (0.5–3.5 Hz), theta (4–7.5 Hz), alpha1 (8–10.5 Hz), alpha2 (11–12.5 Hz), and beta (13– 30 Hz). The absolute power in each of the 5 frequency bands was obtained for each monopolar lead. The mean absolute power in each band of the patient’s recording was then determined for each sampling time at 8 sites. With respect to the baseline condition, the percentages of drug-induced EEG changes in absolute power of each frequency band of each sampling time for 8 sites were computed. 2.6. CT scans As soon as the patients were medically stable, CT imaging was performed by an SCT-3000 TX Scanner (Japan) in 5 mm slices oriented parallel to the orbitomeatal line. Volume measurements expressed in milliliters were taken from CT scans by planimetry (Compensating Polar Planimeters Model 10) of ICH, lesions (ICH plus perihematoma low-density areas), and intraventricular hemorrhage (IVH). Lesion areas were calculated on each slice separately by tracing the perimeter of the appropriate high or low attenuation zone on the CT console. These values were multiplied by the slice thickness to yield the total lesion volume. Perihematoma low-density volume was calculated
456
Z. Huang et al. / Clinical Neurophysiology 113 (2002) 454–458
Table 1 Baseline characteristics of 20 patients with intracerebral hemorrhage
Gender, n Age, years (range) Admission MABP mmHg (range) Admission mental status, n Alert Drowsy Stupor Coma Admission GCS (range) CT interval from onset, days EEG interval from onset, days Interval between CT and EEG examinations, h CT results ICH volume, ml Lesions volume, ml Perilesional hypodense region volume, ml Midline shift, mm IVH volume, ml Oxygen satiation, % Heart rate, bpm Barthel Index at discharge Length of stay, days
Albumin group
Furosemide group
8M, 6F 63.8 ^ 11.5 (45–68) 139 ^ 29 (73–225)
3M, 3F 60 ^ 9.6 (56–71) 141 ^ 31 (78–214)
6 3 2 3 11.3 (15–3) 4.8 (6 h–2 weeks) 5.4 (7 h–2 weeks) 4.2 (1–18)
3 1 1 1 13 (15–6) 4.3 (5 h–13 days) 6.2 (8 h–14 days) 4.7 (1–20)
16.2 (4.1–56.5) 24.9 (4.1–87.4) 8.6 (0–30.9) 1.3 (0.6–8.7) 10.2 (3.5–28.6) 98.7 ^ 1.8 76.8 ^ 15.9 71.4 (9–100) 34.6 (18–46)
15.6 (5.4–50.6) 22.4 (5.4–80.4) 8.2 (0–31.5) 1.1 (0.5–1.6) 6.9 (4.5–9.2) 99 ^ 1.2 77.3 ^ 16.2 72.6 (9–100) 30.4 (18–35.2)
by subtracting ICH volume from the lesion volume. The extent of midline shift was measured according to Ropper (1986).
group, the serum osmolality pre-furosemide ranged from 280 to 300 mOsm/l in 4 cases. Changes in serum sodium, serum potassium, and BUN were not statistically significant.
2.7. Statistical analysis
3.1. Albumin treatment group
The analyses were carried out using the Statistical Analysis System. One-way analysis of variance (ANOVA) was used to detect differences between the mean percentages of drug-induced EEG changes and EEG changes during baseline recordings in each band at each site (F, T, C, or O) in the hemispheres of all patients. If the ANOVA yielded significant results, Dunnett tests were performed to compare the mean percentages of EEG changes during baseline recordings with the EEG changes by drug at different sampling times. The results are expressed as the means and SD. A two-tailed P , 0:05 was considered significant.
3.1.1. Visual assessment
3. Results Baseline characteristics of patients are shown in Table 1, where 20 patients with ICH were studied. Fourteen of of the 20 patients were treated with human albumin and 6 with furosemide. No evident change in ECG, BP, or SaO2 was observed after albumin treatment. Post-treatment diuresis averaged 364 and 785 ml, over a 2 h period, after infusion of albumin and furosemide, respectively. Increase in serum osmolality ranged from 4 to 12 mOsm/l (mean, 6.1 mOsm/l) 1 h post-albumin infusion in 8 cases (serum osmolality ranged from 280 to 298 mOsm/l pre-albumin). No change in serum osmolality was found in the furosemide treatment
3.1.1.1. EEG pre-albumin. The EEG findings pre-testing were abnormal in 12/14 patients, including lateralized abnormalities in 4 cases, generalized slowing activity with one-side dominance in 5, and diffuse slowing activity in 3. Placebo infusion produced no significant change in 0.5– 30 Hz EEG activity during the monitoring period. 3.1.1.2. EEG post-albumin. After the administration of albumin, unchanged EEG was noted in 5 cases. Two of them had normal EEGs pre-drug testing, 3 had an ICH volume less than 10 ml, and 3 were alert. In 9 cases, the slowing abnormalities of EEG had definitely decreased after the infusion. The changes were usually noted 10–30 min post-infusion of albumin, but the obvious change occurred 1 h post-infusion. Of these cases 6 had a GCS score less than 11, 7 had an ICH volume more than 10 ml, 6 had a lesion volume more than 30 ml, 5 had midline shift in CT scan, 4 had ventricular penetration of the hematoma in CT scan. In 3 patients with diffuse abnormal baseline EEG, clear improvement of the EEGs in the nonlesion hemispheres was also noted. 3.1.2. QEEG analysis In Fig. 2 the mean percentage values of albumin-induced
Z. Huang et al. / Clinical Neurophysiology 113 (2002) 454–458
457
Fig. 2. The mean percentage of albumin-induced EEG changes in absolute power at central area in the lesion hemispheres of 14 stroke patients. The percentage of EEG changes in alpha and theta power changed 30 min post-infusion and reached their maximal level 1 h post-infusion of albumin, then remained significant until the 2 h post-infusion. Each point is the mean ^ SD for all subjects at the corresponding time. * or 1 indicate significant differences (P , 0:01 or P , 0:05) from EEG changes by placebo.
absolute power changes at central area of lesion hemispheres with respect to the baseline conditions are shown. The graft demonstrates significant changes in EEGs which occurred 30 min post-albumin infusion, and were at a maximum 1 h post-infusion, and remained significant for 2 h post-infusion. The EEG changes in the non-lesion hemisphere were similar to those which occurred at the lesion hemisphere, but the effects were not significant. 3.2. Furosemide treatment group The EEG findings pre-furosemide testing were normal in two cases, and abnormal in 4 with lateralized abnormalities in two cases, generalized slowing activity with one-side dominance in one, and diffuse slowing activity in one. However, the EEG recordings before, during, and after the infusion were not statistically significantly different by visual and spectral analysis in all 6 cases.
4. Discussion 4.1. Albumin and furosemide treatment for ICH patients with increased ICP Osmotic and renal diuretics are used to reduce ICP in
patients with brain edema or raised ICP. Human albumin solution is commonly used in treating low serum albumin concentrations and hypovolemia. It effectively replaces volume and support colloid osmotic pressure. Currently, the licensed indications for use of albumin are emergency treatment of shock, acute management of burns, and clinical situations associated with hypoproteinemia (McClelland, 1990). Stroke patients in late stage have the pathophysiological basis of hypoalbuminemia, albumin given for correcting the problem will help to control the brain edema. Furosemide is a potent renal (loop) diuretic that has been used to reduce ICP in animals and in man (Cottrell et al., 1977; Tornheim et al., 1979; Schettini et al., 1982). It could also prolong and intensify the effect of mannitol, so as to reduce the dosage of mannitol used (Roberts et al., 1987). And if it were used with albumin, the treatment would be more effective and safe due to albumin maintaining the colloid osmotic pressure in blood, which causes normal blood volume dehydration. The effect of mannitol became less after it was intravenously administrated repeatedly. In this case, furosemide, albumin, and furosemide plus mannitol or albumin alternatively may be a good choice (Lehman, 1990; Pickard and Czosnyka, 1995). 4.2. EEG changes by human albumin and furosemide In the present study, the abnormal EEGs evidently
458
Z. Huang et al. / Clinical Neurophysiology 113 (2002) 454–458
improved by albumin in most coma cases with large hemorrhage and evident ‘mass effects’ in the CT scan. The result was similar to our previous studies on stroke patients with mannitol, which suggest that the effects of improving abnormal EEG by albumin in stroke patients were reflecting the work of drug lowering ICP and dehydrating of brain edema. According to the time the EEG changes occur, the effects of albumin began 30 min post-infusion of 10 g albumin (20% albumin 50 ml) and were maximal 1 h post-infusion, then remained significant for 2 h post-infusion. Compared with the EEG-pharmacodynamic characteristics of intravenous mannitol in stroke patients, the onset of albumin effect was little later, but its duration was longer. In view of our findings we conclude that albumin produced expected changes in monitoring EEG, whereas furosemide produced a large diuresis without a concomitant decrease in slowing abnormalities in EEG. This difference is consistent with the fact that, unlike mannitol, furosemide does not increase renal free water clearance and ‘dehydrate’ the brain (Buhrley and Reed, 1972; Pinegin et al., 1979; Rottenberg and Posner, 1980). Cascino et al. (1983) were also unable to demonstrate statistically significant CT density shifts in edematous white matter in 4 brain tumor patients with furosemide treatment. The small samples and the dose used for testing in our study also contributed to the present result mentioned above. Further studies will be carried out in the future. No previous study has undertaken to monitor EEG changes in stroke patients following the intravenous administration of albumin or furosemide and relate observed changes in the monitoring EEGs to the effects of antiedema or lowering ICP treatment. Our results indicate that available EEG monitoring technique is sensitive enough to monitor the effects of antiedema and lowering ICP therapy in vivo. Further studies on various therapeutic interventions in the related disease group can help this method be used reasonably and effectively in clinic practice.
Acknowledgements We wish to express our appreciation to J.D. Bell DO of the USA, Prof. Yanping Hu for their technical collaborations.
References Agarwal R, Gotman J, Flanagan D, Rosenblatt B. Automatic EEG analysis during long-term monitoring in the ICU. Electroenceph clin Neurophysiol 1998;107:44–58. Bricolo A, Turazzi S, Faccioli F, Sciaretta G, Erculiani P. Clinical application of compressed spectral array in long-term EEG monitoring of comatose patients. Electroenceph clin Neurophysiol 1978;45:211–225. Buhrley LE, Reed DJ. The effect of furosemide on sodium-22 uptake into cerebrospinal fluid and brain. Exp Brain Res 1972;14:503–510. Cascino T, Baglivo J, Conti J, Szewczykowski J, Posner JB, Rottenberg DA. Quantitative CT assessment of furosemide- and mannitol-induced changes in brain water contents. Neurology 1983;33:898–903. Cottrell JE, Robustelli A, Post K, Turndorf H. Furosemide- and mannitolinduced changes in intracranial pressure and serum osmolality and electrolytes. Anesthesiology 1977;47:28–30. Emmerson RG, Chiappa KH. Electrophysiologic monitoring. In: Ropper AH, Kennedy SK, Zervas NT, editors. Neurologic and neurosurgical intensive care, 2nd ed. Rockville, MD: Aspen, 1988. p. 13. Gotman YSJ, Pasupathy A, Flanagan D, Rosenblatt B, Gottesman R. An expert system for EEG monitoring in the pediatric intensive care unit. Electroenceph clin Neurophysiol 1998;106:488–500. Huang ZC. Effect of intravenous mannitol on EEG of the patients with acute stroke. Electroenceph clin Neurophysiol 1997;103(1):199. Jordan KG. Continuous EEG and evoked potential monitoring in the neuroscience intensive care unit. J Clin Neurophysiol 1993;10:445–475. Labar DR, Fisch BJ, Pedley TA, Fink ME, Solomon RA. Quantitative EEG monitoring for patients with subarachnoid hemorrhage. Electroenceph clin Neurophysiol 1991;78:325–332. Lehman LB. Intracranial pressure monitoring and treatment: a contemporary view. Ann Emerg Med 1990;19:295–303. Lesser RP, Webber WRS, Fisher RS. Design principles for computerized EEG monitoring. Electroenceph clin Neurophysiol 1992;82:239–247. McClelland DBL. Human albumin solutions. BMJ 1990;300(6):35–37. Nuwer M. Assessment of digital EEG, quantitative EEG, and EEG brain mapping: report of the American Academy of Neurology and the American Clinical Neurophysiology Society. Neurology 1997;49:277–292. Pickard LD, Czosnyka M. Management of raised intracranial pressure. J Neurol Neurosurg Psychiatry 1995;56:845–858. Pinegin LE, Tibekina LM, Shakhmatova EI, Natochin YV. The effects of altered osmolality and intracellular fluid volume on brain water and electrolytes. Fiziol Zh SSSR 1979;65:122–127. Roberts PA, Pollay M, Engles C, Pendleton B, Reyndds E, Stevens FA. Effect on intracranial pressure of furosemide combined with varying doses and administration rates of mannitol. J Neurosurg 1987;66:440–446. Ropper AH. Lateral displacement of the brain and level of consciousness in patients with an acute hemispherical mass. N Engl J Med 1986;314: 953–958. Rottenberg DA, Posner JB. Intracranial pressure control. In: Cottrell JE, Turndorf H, editors. Anesthesia and neurosurgery, St. Louis, MO: Mosby, 1980. pp. 89–118. Schettini A, Stahurski B, Young HF. Osmotic and osmotic-loop diuresis in brain surgery – effects on plasma and CSF electrolytes and ion excretion. J Neurosurg 1982;56:679–684. Tornheim PA, Mclaurin RL, Sawaya R. Effect of furosemide on experimental traumatic cerebral edema. Neurosurgery 1979;4:48–52.