1Department of Clinical Neurophysiology and 2Anaesthetic Laboratory, St Bartholomews Hospital, London, UK. 3Critical Care Research Program, Department of Anaesthesiology and Intensive Care, Kuopio University Hospital, Kuopio, Finland
Accepted for publication: April 14, 2000
![]() |
Abstract |
---|
![]() ![]() ![]() ![]() ![]() |
---|
Br J Anaesth 2000; 85: 4713
Keywords: intensive care; complications, epilepsy; anaesthetics i.v., thiopentone; monitoring; monitoring, electrocardiography; monitoring, electroencephalography
![]() |
Case report |
---|
![]() ![]() ![]() ![]() ![]() |
---|
On admission to the intensive care unit (ICU), intermittent positive-pressure ventilation was continued, together with a thiopental infusion titrated up to 500 mg h1 under the guidance of a continuously running paper electroencephalogram (EEG), which confirmed the presence of burst suppression and the absence of seizure activity. Standard monitoring included continuous electrocardiography (ECG), invasive arterial pressure, capnography, respiratory flow/volume measurements, hourly urine output measurements and repeated arterial blood gas analysis. The cardiorespiratory, haematological, hepatic, renal, biochemical and blood gas profiles were within normal limits. In particular, no signs of poor tissue oxygenation or pH imbalance were observed. Cranial CT and MRI scans showed no abnormalities. Additional therapy included acyclovir 750 mg 8-hourly and a crystalloid infusion at 100250 ml h1. No sympathomimetic therapy was required. No subsequent seizures were detected clinically or electroencephalographically, and the thiopental infusion was stopped on day 2, by which time the patient had received 7975 mg over 20 h. Regular intravenous phenytoin therapy (250 mg 12-hourly) was continued.
On day 3, with the relatives assent and research ethics committee approval, specialized monitoring was set up as part of the EU Biomed 1 IMPROVE intensive care project1 in order to establish an annotated data library. This involved acquiring and storing continuous physiological variables (including arterial pressure, ECG and capnographs at sampling rates between 25 and 100 Hz) and intermittent annotations (exact timing of drug doses, insertion of lines and any other interventions). This period of specialized monitoring lasted 24 h, during which time continuous digitized EEG was also recorded.2 No notable clinical events occurred in this 24 h period, and biochemical and blood gas profiles remained within normal limits. As on admission, there were no signs of raised intracranial pressure.
Subsequently, the patient made an uncomplicated recovery; his level of consciousness improved and the tracheal tube was removed on day 6. At this point his Glasgow Coma Scale score was 15 and gross intellectual functions were normal. He left the ICU on day 7 and went home on day 14.
Retrospective trend analysis revealed marked cyclical oscillations of the multisystem physiological variables over 24 h (Fig. 1). Assessments were made of raw data (peripheral and core temperature, mean arterial pressure, heart rate) and frequently used processed EEG parameters. The latter consisted of root mean squared amplitude (each EEG amplitude digital value within a sampling period is squared, a mean is subsequently taken and the square root of this is taken) and median power frequency (frequency at which half the spectral power is above and half is below). Peaks of relative hypertension occurred synchronously with peaks in EEG amplitude. Oscillations in heart rate and EEG frequency both showed a close inverse relationship to arterial pressure and EEG amplitude. Peripheral temperature and urine output also showed oscillations, though these did not show exact synchrony with the other parameters.
|
![]() |
Discussion |
---|
![]() ![]() ![]() ![]() ![]() |
---|
More orderly variations exist in both cardiac and neurological signals. First-, second-, third- and, more recently, fourth-order blood-pressure waves have been described and summarized elegantly by Seiver and colleagues.4 First- and second-order waves are descriptions of blood-pressure variations with each systolic beat and respiratory cycle, respectively. Third-order waves, with a cycle of 10160 s, are a consequence of disordered autonomic feedback. Fourth-order waves (cycles of minutes to hours, as seen in this case) have been thought to be due to poor oxygen delivery to the tissues and are associated with a poor prognosis.4 5 Electroencephalographers have described cyclical changes in association with sleep patterns;6 the presence or absence of these features has, in turn, relevance to prognostication in coma.7 8 Repetitive waves with cycling of minutes to hours are also seen in the EEG monitoring of patients with brain insult as a consequence of raised intracranial pressure.9
In this case, we considered three possible causes of the multiparameter oscillations. Classical fourth-order changes seemed unlikely, as severe systemic dysfunction was not evident. Covertly raised intracranial pressure may have been an alternative cause, though there was never any clinical or radiological evidence of it, or of persisting brain injury. Lastly, we could not exclude high-dose thiopental infusion as an important cause. Plasma levels were not measured, as these findings became apparent only on trend analysis after the cessation of monitoring. Our hypothesis is that high-dose thiopental may cause autonomic dysfunction, leading to loss of randomicity; this would be similar to the hypothesis of fourth-order oscillations caused by poor tissue oxygen delivery in ill patients.
The interesting phenomenon of synchronized cyclical oscillations of physiological parameters should not automatically instil concern, and may prove to be a recognized finding when high-dose thiopental infusion is used or there is covert brain dysfunction. This may, in time, be clarified with the increasing sophistication of polygraphic trend analysis in intensive care, without which our findings would not have been apparent. To validate these findings, prospective studies of high-dose thiopental infusions (including the simultaneous measurement of plasma levels) and intracranial pressure monitoring should be encouraged.
![]() |
Acknowledgements |
---|
![]() |
Footnotes |
---|
![]() |
References |
---|
![]() ![]() ![]() ![]() ![]() |
---|
2 Thomsen CE, Gade J, Nieminen K, Langford RM, Ghosh IR, Jensen K, et al. Collecting EEG Signals in the IMPROVE Data Library. IEEE Eng Med Biol 1997; 16: 3340
3 Synek VM. Prognostically important EEG coma patterns in diffuse anoxic and traumatic encephalopathies in adults. J Clin Neurophysiol 1988; 5: 16174[ISI][Medline]
4 Seiver A, Daane S, Kim R. Regular low frequency cardiac output oscillations observed in critically ill surgical patients. Complexity 1997; 2: 515
5 Carroll G. Fourth-order blood pressure waves. J Am Med Assoc 1990; 236: 856
6 Loomis AL, Harvey EN, Hobart G. Potential rhythms of the cerebral cortex during sleep. Science 1938; 81: 58997
7 Chatrian GE, White LE, Daly D. Electroencephalographic patterns resembling those of sleep in certain comatose states after injuries to the head. Electroencephalogr Clin Neurophysiol 1963; 15: 2729[ISI]
8 Evans BM, Bartlett JR. Prediction of outcome in severe head injury based on recognition of sleep related activity in the polygraphic electroencephalogram. J Neurol Neurosurg Psychiatry 1995; 59: 1725[Abstract]
9 Munari C, Calbucci F. Correlations between intracranial pressure and EEG during coma and sleep. Electroencephalogr Clin Neurophysiol 1981; 51: 1706[ISI][Medline]