Crit Rev Neurosurg (1997) 7: 45–51 © Springer-Verlag 1997
Celso Agner Manuel Dujovny Mukesh Misra James I. Ausman
45
Transcranial cerebral oximetry
Introduction C. Agner · M. Dujovny (½) · M. Misra J. I. Ausman Department of Neurosurgery (M/C 799), Neuropsychiatric Institute, The University of Illinois at Chicago, College of Medicine, 912 South Wood Street, Chicago, IL 60612-7329, USA Tel.: (312) 996-4842 Fax: (312) 996-9018 e-mail:
[email protected]
The measurement of cerebral blood oxygen saturation using nearinfrared spectroscopy (NIRS) is pivotal in many clinical and surgical situations such as carotid stenosis, cranio-encephalic trauma, and arteriovenous malformations and during endovascular procedures for intracranial aneurysms, carotid-cavernous fistulae and other similar situations. In the articles reviewed here, the functioning of transcranial cerebral oximetry and all the advantages and disadvantages of its use in current practice are examined. Current cerebral function monitors are either too invasive or provide only an indirect picture of what is really happening inside the brain in some situations,
[1] Retrograde cerebral perfusion during hypothermic circulatory arrest reduces neurologic morbidity J Thorac Cardiovasc Surg (1995) 109: 259–268
Information. The authors describe 35 patients who underwent thoracic aortic arch replacement or resection of intracardiac tumor under hypothermic circulatory arrest. There were 27 men and 8 women with ages ranging from 21 to 83 years (mean 60 years). The mean retrograde perfusion time was 63 min, with values vary-
such as distal carotid disease, or when jugular venous oxygen saturation (SJO2) is monitored. Transcranial cerebral Doppler ultrasonography can not insonate the middle cerebral artery (MCA) in 10–15% of patients, precluding use of this procedure for monitoring. Additionally, an individual sufficiently experienced to correctly manipulate the ultrasound apparatus sometimes is not available. Therefore, the search for a noninvasive, accurate and continuous cerebral monitoring device continues. The papers reviewed here illustrate several areas where transcranial cerebral oximetry (TCCO) promises to provide these advantages in clinical practice.
ing from 35 to 128 min. Some 86% (30) of the patients had more than 45 min of hypothermic circulatory arrest, 34% (12) more than 60 min and 11% (4) more than 90 min. Cerebral oxygen saturation was measured using the Somanetics Invos 3100 device with NIRS using two wavelength sensors attached to the frontal scalp region. Retrograde cerebral perfusion (RCP) was regulated to maintain cerebral oxygen saturation within 10% of baseline. In the patient who was monitored with RCP via the right internal jugular vein, the left cerebral oxygen saturation was consistently 10–15% higher than on the right side. There was no alteration in any of the
46
patients with RCP via the superior vena cava. No motor deficits or sensory loss were observed in 94% (33) of the patients. One patient (RCP 92 min) had a documented left hemisphere infarct, and one perioperative death occurred due to an acute dissection and a perioperative myocardial infarction with cardiogenic shock (RCP 61 min). One late death occurred due to postoperative mediastinitis and another due to multiple pulmonary emboli. Making use of a simple and accurate technique, this study permitted the extension of circulatory arrest time far beyond the previously established value of 45 min, with good results and minimal immediate neurological deficits. Analysis. This article starts with an excellent historical background on the use of hypothermic circulatory arrest for the performance of cardiovascular surgeries. Svensson and co-workers showed in 1993 that cerebral ischemic periods longer than 45 min were associated with a high risk of stroke and those longer than 65 min with higher mortality. The major objective of the authors was to demonstrate that longer periods of hypothermia could be employed with a much lower incidence of complications in the patients where cerebral oxygenation was monitored and maintained at baseline levels. TCCO proved valuable during RCP as a means of verifying maintenance of adequate blood flow during the arrest period. The ability to adjust pump flow rates and head pressures accurately resulted in improved outcomes in patients experiencing arrest times of 45 min to 2 h. No other cerebral function monitoring devices were employed in the study due to the difficulties encountered in their use during circulatory arrest, hypothermia and reversed flow. While this precludes comparisons with these devices, it shows that they were not necessary to optimize the RCP process. Neuropsychological evaluation was also not performed, precluding any analysis of neurological deficits after operation.
[2] Near-infrared spectroscopy use in patients with head injury J Neurosurg (1995) 83: 963–970
Information. The authors analyzed 14 ventilated patients with history of closed cranio-encephalic trauma and no intracranial mass requiring elevation of bone flaps, extracranial trauma or scalp lacerations over a
12-month period. Twelve male and 2 female patients with ages ranging from 16 to 63 years were included in the study. Their Glasgow Coma Scale (GCS) values varied from 3 to 12. Intracranial pressure (ICP) and cortical perfusion, cerebral perfusion pressure (CPP) jugular venous oximetry, peripheral oxygen saturation and NIRS measurements were analyzed in the group and comparisons were made with the intent to establish the possible role of NIRS as a routine measure in the intensive care unit (ICU). A total of 886 h of data (range 8–160 h) was recorded from the 14 patients. Of this, 376 h (42%) was considered reliable for the analysis. The major problems encountered were unusable SJO2 readings (48%) and automatic rejection of the NIRS signal due to stray light (32%). The ICP probe was proven to be 100% reliable. Relative decreases in cerebral blood flow (CBF) and cortical perfusion values were compared to measurements of cerebral blood oxygen saturation, with excellent correlation with the values obtained from the NIRS device. Thirty-eight cerebral events were observed during the monitoring interval. In 8 (21%) of them, a fall in peripheral saturation to below 90% was observed with concomitant alterations in oxyhemoglobin (HbO2), Hb, flow velocity and laser Doppler flowmetry values. Desaturation of SJO2 was observed in only four of these events. In 10 events, a rise in ICP and fall in CPP preceded all changes in HbO2 and Hb by a mean of 3.6 min. SJO2 values were detected in five of these events, perfectly synchronized with NIRS measurements. Eight events were characterized by intracranial hyperemia, in which a synchronous rise in ICP, HbO2, Hb, flow velocity and laser Doppler flowmetry values occurred. SJO2 increased in two of these events. The last 12 events involved a rise in ICP and a fall in CPP, with concomitant changes in HbO2, Hb, flow velocity and laser Doppler flowmetry values in all the events and SJO2 changes in 9 events. Analysis. The introduction of this paper describes the authors’ concern to understand all the pathophysiological alterations seen in head trauma patients. Of particular interest is the measurement of cerebral oxygenation and its relationship with other measurements in a demanding clinical environment. Patient selection includes a wide range of pathology which may have adversely influenced the volume of artifact-free data which could be analyzed. The authors also describe additional technological and usage problems with the NIRS device. Despite these problems, the authors were able to show correspondence between
47
changes in NIRS and other cerebral function monitors at a level much higher than with traditional cerebral oxygenation monitoring using invasive jugular catheters.
[3] Validation in volunteers of a near-infrared spectroscope for monitoring brain oxygenation in vivo Anesth Analg (1996) 82: 269–277
Information. The authors analyzed 22 conscious, healthy volunteers with ages ranging from 22 to 35 years, separated into two groups. The first group was composed of 12 volunteers with ages ranging from 23 to 33 years in whom NIRS data were analyzed with blood oxygen saturation samples to create a mathematical algorithm for measurement of cerebral oxygen saturation. In the remaining 10 volunteers (aged 22– 35 years), the new algorithm was validated. All of them underwent radial arterial and jugular venous bulb catheterization. They were monitored with a pulse oximeter, an end-tidal carbon dioxide monitor, an electrocardiogram and an automated blood pressure monitor and were studied in the supine position. Hypoxic gas mixtures were administered in random order (FIO2 ranging from 0.06 to 0.13) through a tight-fitting mask and a closed system. The mixtures were breathed until a steady oximeter reading was obtained, at which time simultaneous arterial and jugular blood samples were withdrawn and heart rate, blood pressure, end-tidal CO2 (ETCO2) and pulse oximeter readings were recorded. Neurologic function was assessed by verbal communication. Volunteers inspired the mixture for at least 5 min, with 5 min between blood samples and 100% oxygen being administered in the intervals, until restoration of baseline oxygen saturation. Multiple linear regression analysis was applied to the recorded values and correlations were performed. Sensitivity ranged from 78.6 to 85.1% for the training data and from 47.8 to 91.7% for the validation data. Specificity ranged from 78.6 to 95.5% for the training data and from 88 to 94.7% for the validation data, showing decreased sensitivity and enhanced specificity with progressive hypoxia. Thus, cerebral oximetry was demonstrated to be accurate ancillary instrument for the determination of hypoxia. Analysis. The authors introduce the oximeter and all its components in a comprehensive and well-written way.
The basic intent of this paper was to validate an empirical algorithm, based on a mathematical model, for calculation of cerebral oxygen saturation using NIRS data. It also examines the sensitivity and specificity of transcranial cerebral oximetry in progressive hypoxia in a group of healthy volunteers. Individual correlations between cerebral oximetry and traditional blood sample analysis showed a very high level of agreement, with r2 values ranging from 0.79 to 0.99 (mean 0.93). However, inter-individual differences reduced the overall accuracy of the absolute TCCO measurements to ±11% compared with global blood sample analysis. A weakness of the study was that it evaluated accuracy during systemic hypoxia only – no attempt was made to modify CBF or cerebral oxygenation independent of systemic oxygenation. This makes it impossible to tell whether the TCCO signals were originated solely from brain tissue.
[4] The influence of contralateral carotid stenosis and occlusion on cerebral oxygen saturation during carotid artery surgery Eur J Vasc Endovasc Surg (1995) 10: 198–206
Information. The authors studied 54 patients scheduled for carotid endarterectomy with >70% ipsilateral carotid artery stenosis demonstrated by color Doppler ultrasound. ETCO2 was kept at 36–40 mmHg. Heart rate and arterial blood pressure were monitored continuously by electrocardiography and radial artery cannulation. A pulse oximeter displayed peripheral oxygen saturation. Transcranial Doppler (TCD), SJO2 and cerebral oxygen saturation (CsO2) were monitored simultaneously and compared with each other using Spearman’s test of correlation. After endarterectomy, ipsilateral peak TCD increased from 31 cm/s to 69 cm/s with a percentage increase of 35%. After declamping, SJO2 increased from 75% to 79% and CsO2 from 65% to 68%, whilst contralateral CsO2 went from 67% to 68%. The correlation coefficient of TCD and SJO2 was r2=0.59 (P
