Change in level of consciousness is an important clinical sign in patients with brain injury. When consciousness is lost, further cerebral deterioration may be difficult to detect. Similarly, during anesthesia, assessment of mental status is not possible, shifting reliance toward some form of instrumental monitoring. The crucial factor for cerebral well-being is the maintenance of tissue oxygenation. Monitoring of cerebral oxygenation is not routine because existing techniques either are invasive , require a prolonged period of equilibration, or involve the use of ionizing radiation [2,3]. This is compounded by the lack of sensitivity and high false-positive rates of many of the available monitors .
Near infrared spectroscopy (NIRS) is a continuous, noninvasive bedside technique that may be used to monitor cerebral oxygenation. Its potential as a noninvasive cerebral oxygenation monitor was first outlined by Jobsis  in 1977, and it has since found widespread application in neonatal and adult medicine. NIRS exploits the relative transparency of biological tissues to near infrared light and the differential absorption properties of oxy- and deoxyhemoglobin. Because of the scattering properties of the underlying tissue, the path taken by the photons is more complex than the simple straight line distance between the transmitting and receiving optodes. It has been estimated that only one in a million of the emitted photons is actually detected , but this is sufficient to enable measurements to be made in the clinical setting.
When cerebral blood flow (CBF) is altered by extreme maneuvers, such as ventricular fibrillation induced during automatic defibrillator testing, NIRS can detect sudden changes in cerebral oxygenation . During less severe changes in CBF induced by interruption of carotid artery blood flow during carotid endarterectomy, NIRS can detect cerebral ischemia [8,9]. Kirkpatrick et al.  simultaneously measured NIRS, transcranial Doppler velocimetry, and cerebral function monitoring and demonstrated that ischemic events were detected with similar frequency by all three monitors. In addition, NIRS has been used to investigate small changes in cerebral blood volume in adult volunteers  and anesthetized patients  with normal brains. NIRS could provide important measurements of rapid and small changes in cerebral oxygenation.
IV anesthetics depress cerebral metabolism [13,14], with consequent reductions in oxygen consumption (CMRO2), CBF, and intracranial pressure . As CMRO2 decreases, CBF is reduced proportionately; this relationship is termed flow-metabolism coupling. After the administration of IV propofol and thiopental, flow-metabolism coupling usually remains intact , and cerebral saturation would be expected to remain unaltered or to be improved. Not all IV anesthetics have the same effect. Milde et al.  showed, in dogs, that etomidate can produce a rapid reduction in CBF accompanied by a slower reduction in CMRO2. This mismatching of flow-metabolism coupling, with a greater reduction in flow than demand, may induce cerebral desaturation. Edelman et al.  confirmed that, in anesthetized humans undergoing craniotomy with isoflurane anesthesia, the administration of incremental bolus doses of etomidate to produce burst suppression on the electroencephalogram was associated with a significant reduction in cerebral cortical PO2.
This difference between etomidate and other IV induction anesthetics is therefore a potential model to investigate whether NIRS can detect small changes that occur at the onset of anesthesia. The aim of this study was to compare the effect of different IV induction anesthetics on cerebral oxygenation using NIRS.
After institutional ethics committee approval, and with informed written consent, 36 unpremedicated ASA physical status I or II adult patients undergoing elective day-case surgery were enrolled into this study. The patients were randomized by a sealed envelope method to undergo induction of anesthesia with either etomidate, propofol, or thiopental by an anesthesiologist independent of the study. Each data set was coded by subject number and induction anesthetic and was analyzed by an investigator blinded to the code.
Changes in oxyhemoglobin concentration ([HbO2]) and deoxyhemoglobin concentration ([Hb]) were recorded using a NIRO 500 spectrophotometer (Hamamatsu Photonics, Hamamatsu City, Japan) . It uses four pulsed semiconductor laser diodes using wavelengths of 775, 826, 850, and 909 nm with a pulse frequency of 1.9 kHz and is capable of detecting intracranial events [18,19]. The optodes were located in the temporoparietal region with an interoptode spacing of 4 cm and shielded from ambient light. A sampling interval of 0.5 s was used, and the changes in [HbO (2)] and [Hb] were calculated using a previously established algorithm [20,21] incorporating a differential pathlength factor of 6.26 . Beat to beat arterial saturation was recorded using a modified Novametrix 520A pulse oximeter (Novametrix Medical Systems Inc., Wallingford, CT) attached to an earlobe. ETCO2 was monitored by using a mainstream Novametrix 7000A capnograph (Novametrix Medical Systems Inc.) directly attached to a tightly fitting face mask. Data were continuously recorded by the NIRO 500 and sampled synchronously with the NIRS data using a laptop computer running Onmain V1.32a (Hamamatsu Photonics). Blood pressure and heart rate were monitored noninvasively by using a Datex Cardiocap (Datex Instrumentarium Corp., Helsinki, Finland) with a sampling interval of 60 s.
Measurements were made with the patients in the supine position. All patients breathed room air for the first 4 min of the study, after which the induction anesthetic was slowly administered IV until loss of verbal contact was achieved. At this stage, ventilation was gently assisted with a Bain circuit, using a fresh gas flow of 10 L/min (25% oxygen and 75% nitrous oxide) to maintain ETCO2 constant at its preinduction value. The patients held a 1-kg weight in their dominant hand, and the time from the commencement of administration of the induction anesthetic until the weight was dropped was taken as the time to loss of consciousness. Data were recorded for a further 3 min before the study concluded, after which anesthesia was deepened and surgery allowed to proceed. No IV fluids were administered after the induction of anesthesia until after the study.
Changes in [HbO2] and [Hb] with time were plotted for each subject, and the area under the concentration-time curves was calculated to provide an index of changes in cerebral oxygenation with time. The total amount of hemoglobin ([HbT]) within the field of view is represented by the sum of [Hb] and [HbO2]. Similarly, the area under the [HbT]-time curve was also calculated to provide an index of temporal changes in blood volume within the field of view.
Statistical analysis was performed by using SPSS 4.0 (SPSS Inc., Chicago, IL). The areas under the curves were analyzed by using a repeated-measures analysis of variance model using Tukey's method to determine the significance of individual changes between drugs. A P value <0.05 was taken as significant. All results are expressed as mean +/- SD.
NIRS data were successfully recorded in all cases. The three groups were well matched (Table 1). After induction of anesthesia, neither arterial hemoglobin saturation nor ETCO2 changed. Mean arterial blood pressure decreased by 11% in the propofol group but was unchanged in the etomidate and thiopental groups.
Within groups, the shapes of the Delta [HbO2] and Delta [HbT] curves were similar for a given drug. Although there was intersubject variability in terms of the magnitude of the changes, these were in the same direction for all patients in each group. There were dear differences among the three different drugs (Figure 1). From a stable baseline, the onset of anesthesia corresponded with changes in the [HbO2] signal in all cases. The area under the Delta [HbO2]-time curve (an index of changes in cerebral oxygenation) showed marked changes with respect to both time (P < 0.001) and induction anesthetic (P < 0.001) (Figure 2).
The area under the Delta [Hb]-time curve did not significantly change with respect to either time or induction anesthetic. The area under the Delta [HbT]-time curve (an index of changes in cerebral blood volume) similarly showed marked changes with respect to time (P < 0.001) and induction anesthetic (P < 0.001) (Figure 3).
Subgroup analysis showed that the decrease in the area under the Delta [HbO2]-time curve seen in the etomidate group was significantly different from the increases seen with propofol (P < 0.001) and thiopental (P < 0.001). Similarly, the decrease in the area under the Delta [HbT]-time curve seen in the etomidate group was significantly different from the increases seen in both the propofol (P < 0.001) and thiopental (P < 0.001) groups. The differences in the areas under the Delta [HbO (2)] and Delta [HbT]-time curves between propofol and thiopental were not significant.
In this study, we demonstrated that NIRS was able to continuously monitor, in real time, the effect on cerebral oxygenation of administration of IV anesthetics. The changes in [HbO2] and [HbT] occurred simultaneously very shortly after loss of consciousness and showed striking differences among drugs. It was possible to distinguish among the induction drugs in every case, as changes within groups were in the same direction in all patients. This study used the usual clinical induction dose of etomidate, in contrast to the larger doses used in the studies by Milde et al.  and Edelman et al. . Despite the smaller dose in our study, we were still able to demonstrate a reduction in cerebral oxygenation with etomidate.
It is interesting to note the time course of oxygenation changes for the different drugs. Before the induction of anesthesia, the NIRS signals were stable, began to change as unconsciousness ensued, and were clearly distinguishable from baseline at 3 min. It was not possible to continue this observational study beyond this time because of the need to avoid emergence from anesthesia. With propofol and thiopental, the change in cerebral oxygenation occurred as consciousness was lost (Figure 1). However, with etomidate, the NIRS signals changed before loss of consciousness, in a similar manner to that found by Milde et al.  for CBF in dogs. We hypothesize that the reduction in CBF produced by etomidate was responsible for the reduction in [HbO2] observed with NIRS.
It is possible to infer changes in cerebral blood volume from changes in [HbT] if the hematocrit remains constant . It is unlikely that the hematocrit would be altered during the short period of this study; therefore, the change in the total hemoglobin signal is proportional to the change in blood volume within the field of view.
The relative effects of the observed changes in oxyhemoglobin changes detected by NIRS deserve comment. Such changes are conveniently expressed as changes in saturation, a ratio of [HbO2] to [HbT]. NIRS currently measures a change in [HbO2] and [HbT] from an arbitrary baseline, and it is therefore not possible to calculate saturation changes directly. However, if we substitute baseline values for [HbT]  and mean tissue saturation , our observed changes in [HbO (2)] 3 min after the induction of anesthesia equate to a reduction in cerebral saturation of 2% after the administration of etomidate and an increase of 8% after the administration of propofol and thiopental. The magnitude of the saturation changes seen in all three groups is sufficiently small to be clinically insignificant under most circumstances but may become so in patients in whom brain oxygen supply is critical.
NIRS samples from arterial, venous, and capillary compartments, but the relative contribution of each to the signal has not yet been defined. It is therefore not possible to determine whether the observed changes in saturation and blood volume occurred equally in all three compartments. Because arterial saturation was constant throughout this study, it is likely that the observed changes originate either from the capillary and venous compartments or because of a change in the flow-metabolism relationship.
Because CO2 tension was held constant at each patient's baseline, it is an unlikely contributor to the observed changes in NIRS signals. Although mean arterial pressure decreased in the propofol group, it remained within the range of autoregulation that is intact during propofol anesthesia. The change from breathing room air to 25% oxygen and 75% nitrous oxide is also unlikely to have accounted for the observed changes in the NIRS signals after the induction of anesthesia. We designed the study to measure the changes during routine IV induction of anesthesia; therefore, it involved this switch from room air to 25% oxygen. Although it might be argued that the small increase in the fraction of inspired oxygen (FIO2; 4%) may have been responsible for the increase in cerebral oxygenation in the propofol and thiopental groups, this cannot explain the decrease in the etomidate group, in which similar effects should also have been observed. Nitrous oxide would also be expected to have similar effects in all three groups. Furthermore, the clear temporal relationship between the NIRS signal changes and the administration of an IV induction anesthetic, as well as the reproducibility of the within-group changes, argues against these being an effect of the small change in FIO2 after the induction of anesthesia.
In conclusion, NIRS is able to detect small changes in cerebral oxygenation during the induction of general anesthesia. Anesthesia induction with propofol and thiopental was associated with a small increase in cerebral oxygenation, whereas etomidate was associated with a smaller reduction. NIRS was able to make measurements in real time and shows considerable promise as a bedside monitor. Its potential as a monitor of changes in cerebral oxygenation in the critically ill, or in those in whom consciousness is impaired, should be further investigated.
1. Dearden NM, Midgley S. Technical consideration in continuous jugular venous oxygen saturation measurement. Acta Neurochir (Wein) 1993;59:91-7.
2. Sakai F, Nakazawa K, Tazaki Y, et al. Regional cerebral blood volume and haematocrit measured in normal human volunteers by single-photon emission computed tomography. J Cereb Blood Flow Metab 1985;5:207-13.
3. Dahl A, Russell D, Nyberg-Hansen R, Rootwelt K. Effect of nitroglycerine on cerebral circulation measured by transcranial doppler and SPECT. Stroke 1989;20:1733-6.
4. Kirkpatrick PJ, Smielewski P, Czosnyka M, et al. Near-infrared spectroscopy use in patients with head injury. J Neurosurg 1995;83:963-70.
5. Jobsis FF, Noninvasive, infrared monitoring of cerebral and myocardial oxygen sufficiency and circulatory parameters. Science 1977;198:1264-7.
6. Cope M, Delpy DT. System for long-term measurement of cerebral blood and tissue oxygenation on newborn infants by near infra-red transillumination. Med Biol Eng Comput 1988;26:289-94.
7. Levy WJ, Levin S, Chance B. Near-infrared measurement of cerebral oxygenation. Anesthesiology 1995;83:738-46.
8. Kirkpatrick PJ, Smielewski P, Whitfield PC, et al. An observational study of near-infrared spectroscopy during carotid endarterectomy. J Neurosurg 1995;82:756-63.
9. Mason PF, Dyson EH, Sellars V, Beard JD. The assessment of cerebral oxygenation during carotid endarterectomy utilising near infrared spectroscopy. Eur J Vasc Surg 1994;8:590-4.
10. Kirkpatrick PJ, Lam J, Al-Rawi P, et al. Determination of near infrared spectroscopy (NIRS) thresholds for severe cerebral ischaemia during carotid surgery [abstract]. J Cereb Blood Flow Metab 1997;17(Suppl 1):S603.
11. Gupta AK, Menon DK, Czosnyka M, et al. Non-invasive measurement of cerebral blood volume in volunteers. Br J Anaesth 1997;78:39-43.
12. Owen-Reece H, Elwell CE, Goldstone J, et al. Investigation of the effects of hypocapnia upon cerebral haemodynamics in normal volunteers and anaesthetised subjects by near infrared spectroscopy (NIRS). Adv Exp Med Biol 1994;361:475-82.
13. Stullken EH, Milde JH, Michenfelder JD, Tinker JH. The nonlinear responses of cerebral metabolism to low concentrations of halothane, enflurane, isoflurane and thiopental. Anesthesiology 1977;46:28-34.
14. Alkire MT, Haier RJ, Barker SJ, et al. Cerebral metabolism during propofol anesthesia in humans studied with positron emission tomography. Anesthesiology 1995;82:393-403.
15. Michenfelder JD. Anesthesia and the brain. New York: Churchill Livingstone, 1988.
16. Milde LN, Milde JH, Michenfelder JD. Cerebral functional, metabolic, and hemodynamic effects of etomidate in dogs. Anesthesiology 1985;63:371-7.
17. Edelman GJ, Hoffman WE, Charbel FT. Cerebral hypoxia after etomidate administration and temporary cerebral artery occlusion. Anesth Analg 1997;85:821-5.
18. Villringer A, Planck J, Hock C, et al. Near infrared spectroscopy (NIRS): a new tool to study hemodynamic changes during activation of brain function in human adults. Neurosci Lett 1993;154:101-4.
19. Owen-Reece H, Elwell CE, Wyatt JS, Delpy DT. The effect of scalp ischaemia on measurement of cerebral blood volume by near-infrared spectroscopy. Physiol Meas 1996;17:279-86.
20. Wray S, Cope M, Delpy DT, et al. Characterization of the near infrared absorption spectra of cytochrome aa3
and haemoglobin for the non-invasive monitoring of cerebral oxygenation. Biochim Biophys Acta 1988;933:184-92.
21. Essenpreis M, Elwell CE, Cope M, et al. Spectral dependence of temporal point spread functions in human tissues. Appl Opt 1993;32:418-25.
22. Duncan A, Meek JH, Clemence M, et al. Optical pathlength measurements on the adult head, calf and forearm and the head of the newborn infant using phase resolved optical spectroscopy. Phys Med Biol 1995;40:295-304.
23. Edwards AD, McCormick DC, Roth SC, et al. Cerebral hemodynamic effects of treatment with modified natural surfactant investigated by near infrared spectroscopy. Pediatr Res 1992;32:532-6.
© 1999 International Anesthesia Research Society
24. Meixensberger J, Dings J, Kuhnigk H, Roosen K. Studies of tissue PO2
in normal and pathological human brain cortex. Acta Neurochir (Wein) Suppl 1993;59:58-63.