Can Cerebrovascular Reactivity Be Measured With Near-Infrared Spectroscopy?
Background and Purpose We used near-infrared spectroscopy (NIRS) to monitor the cerebral oxygenation changes during CO2 reactivity tests.
Methods Fifty healthy volunteers were examined (age range, 19 to 68 years). The monitored parameters were as follows: transcranial Doppler (TCD) time-averaged middle cerebral artery flow velocity end-tidal CO2 (Etco2); change in concentration of cerebral oxyhemoglobin (Hbo2), deoxyhemoglobin (Hb), and total hemoglobin; mean arterial blood pressure; peripheral arterial oxygen saturation (Sao2); and extracranial tissue perfusion with the use of cutaneous laser-Doppler flowmetry. The examination protocol included both hypercapnia and hypocapnia. The cerebrovascular reactivity indexes were calculated as follows: TCD, relative change in flow velocity per 1 kPa increase in Etco2; NIRS, absolute change in Hbo2, Hb, and total hemoglobin concentration (micromoles per liter) per 1 kPa increase in Etco2.
Results Mean middle cerebral artery flow velocity was found to be 58 cm/s at a mean baseline Etco2 of 4.7 kPa. Mean cerebrovascular reactivities were as follows: TCD, 24%/kPa (SEM, 1.1); Hbo2, 2.06 μmol/L per kilopascal (SEM, 0.08); Hb, −0.63 μmol/L per kilopascal (SEM, 0.09); and total hemoglobin concentration, 1.44 μmol/L per kilopascal (SEM, 0.1). Statistical analysis revealed significant correlation between reactivities calculated with the use of NIRS and TCD (P<.001). Although some fluctuations were observed in Sao2 and laser-Doppler flux, they were not correlated with either Etco2 or NIRS.
Conclusions NIRS signal changes in Hbo2, Hb, and total hemoglobin concentration are very sensitive to alterations in Etco2, which are largely independent of extracranial tissue perfusion. NIRS may be developed as an alternative method for testing cerebrovascular reactivity and may be of particular clinical importance when the ultrasound window is poor.
The clinical importance of cerebrovascular reactivity testing is now recognized. Studies on hemodynamics in patients after head injury or subarachnoid hemorrhage and in those with carotid disease have shown that the state of cerebrovascular reactivity is associated with outcome.1 2 3 In some instances the reactivity has been used to direct the management of such patients.4
Several methods have been used to assess reactivity by measuring the response of the cerebral resistive vessels to physiological stimuli that affect tissue acidosis.5 Inhalation of 5% CO26 and intravenous administration of acetazolamide are commonly used.7 The quantification of the response of the blood vessels to the stimulus can be obtained by measuring cerebral blood flow,8 cerebral blood volume,9 or blood flow velocity.6 The ideal technique for testing should be simple, repeatable, and noninvasive. One technique that approaches these ideas is TCD.10 However, although the relative changes of cerebral blood flow velocity in the MCA provide an index of cerebrovascular reactivity,11 12 13 14 the accuracy and reliability of TCD have been questioned.15 16 17 The measurement of changes in blood flow may be biased, with some error introduced by small changes in diameter of the insonated vessels.16 18 19 In addition, blood flow velocity in the arteries of the circle of Willis may not detect the influence of extracranial-intracranial anastomoses. Finally, the insonation of the MCA is not possible in approximately 5% to 20% of patients.20 21
We have addressed these concerns by using the relatively new technique of NIRS, which is simple to apply and noninvasive.22 We have been encouraged by our own experience with NIRS in adults during carotid endarterectomy23 and in patients with severe head injury.24 It provides a real-time assessment of fluctuation in cerebral (Hbo2) and Hb and allows estimation of changes in cerebral blood volume.25 26 The principle of the technique has been described in detail.27 Briefly, the technique depends on the relative transparency of biological tissue to light in the infrared region, allowing the measurement of absorption by the key chromophores Hb and Hbo2. According to the Beer-Lambert law, changes in concentration of these chromophores are calculated. Summation of changes in Hbo2 and Hb produces values of changes in total hemoglobin concentration that are directly proportional to variations in regional cerebral blood volume.
The purpose of this study was to investigate the NIRS response to changing Etco2 concentration and compare it with TCD flow velocity reactivity to CO2.
Subjects and Methods
The study was approved by the local research ethics committee. Fifty healthy volunteers were examined (25 men and 25 women; age range, 19 to 68 years; median age, 32 years). The volunteers were recruited from the academic and hospital staff and from patients awaiting lumbar spinal surgery.
In a quiet environment, volunteers were asked to breathe through a face mask while a TCD (PC DOP 842; 2-Mhz probe) probe and NIRS (NIR 1000, Hamamatsu Photonics) optodes were attached to their heads (Fig 1⇓). The TCD probe was positioned at the temporal window and fixed with a head band. Two NIRS optodes were placed on the side of the TCD probe on the frontal region with the receiving probe 2 cm from the midline and 2 cm above the supraorbital ridge. The transmitting probe was placed 6 cm away toward the hairline. Probes were secured with adhesive tape, and a pressure bandage was applied beneath a lightproof cloth. The spectrometer was allowed to reach operating temperature for 1 hour and then to stabilize for approximately 10 minutes, after which the readings were normalized. Etco2 (901 Mk2, Morgan Instruments) was recorded continuously, as was Sao2 (pulse oximeter Multinex 4200, Datascope) and MABP (Finapres, Ohmeda 2300) with the use of an automated pressure cuff. Additionally, in 15 volunteers the superficial blood perfusion was monitored with the use of laser-Doppler flowmetry (MBF3D, Moor Instruments) with a cutaneous probe placed between two NIRS optodes and attached to the skin with adhesive plaster. The laser light wavelength required alteration by the company to avoid interference with NIRS (new laser diode wavelength, 635 nm).
Signals of ABP, flow velocity, Sao2, Etco2, and laser-Doppler flux were sampled at a frequency of 50 Hz and digitized with a 12-bit analog-to-digital converter (DT 2814, Data Translation). ABP, flow velocity, Sao2, and Etco2 were calibrated in appropriate units, while laser-Doppler flux was measured in arbitrary units. A path length factor of 5.9328 and an intraoptode distance of 6 cm were adopted, and NIRS signals were recorded in micromoles per liter by an RS232 interface. Signals were filtered and averaged over consecutive 4-second periods. The software for on-line signal analysis29 summed the change in Hbo2 and Hb signals to give an estimate of changing total hemoglobin concentration. Spectral PI was calculated on-line as the amplitude of the fundamental harmonic of the flow velocity pulse waveform divided by the time-averaged flow velocity.12 The data were stored on an IBM personal computer for further off-line analysis.
CO2 Studies Protocol
The examination protocol consisted of five stages after baseline was established: baseline 0, 5 minutes of baseline recording; stage 1, 3 minutes of hyperventilation; stage 2, 5 minutes of rest period; stage 3, 5 minutes of increased Etco2; stage 4, 3 minutes of normal air respiration; and stage 5, 3 minutes of hyperventilation. A typical example of the examination recording is presented in Fig 2⇓.
The cerebrovascular reactivity indexes were calculated as follows: for TCD, percent change in flow velocity per 1 kPa change in Etco2; for NIRS, absolute change in Hbo2, Hb, and total hemoglobin concentration (micromoles per liter) per 1 kPa change in Etco2.
To pool the data from all volunteers, changes in the measured parameters between hypocapnia and hypercapnia were calculated. The parameters flow velocity, laser-Doppler flux, and MABP were additionally normalized by the baseline values. The preprocessed data were then analyzed by linear regression with the Pearson correlation coefficient. Confidence intervals for reactivities were calculated with the use of t statistics. Normality of the data was verified with the use of normal probability plots and Lilliefors’ test.
Data Exclusion and Adverse Effects
Two volunteers were excluded from the study: one with a very poor ultrasound window and the other with a very low hairline, which rendered proper positioning of NIRS probes impossible without a partial shave. There were no adverse effects of the examination on any of the volunteers.
The mean MCA flow velocity was found to be 58 cm/s (SEM, 1.5) at the mean baseline Etco2 of 4.7 kPa (SEM, 0.065). The mean PI was 0.49 (SEM, 0.01). Both flow velocity and PI were strongly correlated with Etco2 (Fig 3A⇓), and the correlation coefficients as well as the mean reactivities are presented in Table 1⇓.
Hbo2 and Hb Signals
All NIRS parameters were significantly correlated with CO2 (Fig 4⇓). The correlation coefficients as well as the mean reactivities calculated with Hbo2, Hb, and total hemoglobin are presented in Table 2⇓.
Correlation Between TCD and NIRS Parameters
Correlation analysis between reactivities calculated with NIRS and TCD showed significant (P<.001) convergence of these two methods of assessment (Fig 5⇓). Both flow velocity and PI were significantly correlated with the NIRS parameters. The actual values of correlation coefficients are given in Table 2⇑.
MABP showed some tendency to increase during periods of hypercapnia in individual volunteers. However, the overall correlation of MABP and Etco2 was not significant (P>.2).
Some minor changes were observed in peripheral saturation (SD, 1.6%); the overall correlation of Sao2 with Etco2 and NIRS parameters was not significant (Fig 6⇓).
Laser-Doppler Flux Signal
Although there have been some significant fluctuations noted in cutaneous laser-Doppler flux in some cases (particularly during the hyperventilation stage), these changes were not correlated with Etco2 (Fig 6⇑). Furthermore, the variation in laser-Doppler flux signals showed no relationship to NIRS changes during the reactivity study (P>.35).
There was no correlation between age and flow velocity reactivity. PI reactivity showed some tendency to decrease with age, but the correlation was statistically not significant (Table 3⇓). In contrast, Hbo2 and Hbdiff (defined as HbO2 minus Hb) demonstrated a significant decrease in CO2 reactivity at ages over 35 years (Table 3⇓, Fig 7⇓).
NIRS as a New Method for Cerebrovascular Reactivity Testing
This study adopts NIRS in the reflection mode (optodes placed on the same side of the head close to each other). Spatial and quantitative concerns surrounding NIRS in this mode are still not resolved. The technique measures changes in concentration from an unknown baseline. Despite this, absolute changes in the concentrations of chromophores (Hb and Hbo2) may still provide a useful clinical index. Our own experience in monitoring patients with severe head injuries shows that NIRS provides a more sensitive indicator of desaturation events in sedated patients with severe head injuries than the Sjo2 monitor.24 In the present study the main question we have addressed is whether absolute changes in hemoglobin concentration can be detected during CO2 stimulation and whether such changes are comparable among normal individuals.
The results of the study show that NIRS parameters react to changes of CO2 in a uniform fashion, with large increases in Hbo2 and smaller decreases in Hb concentrations. The very significant linear correlation between changes in Hb and HbO2 concentrations and TCD flow velocity indicates that interindividual variations in NIRS responses may be largely explained by differences in cerebrovascular reactivities. Thus, our own data support the notion that, even without knowledge of the reference baseline concentrations of hemoglobin, responses of Hbo2 and Hb to CO2 in healthy volunteers can be taken as indexes of cerebrovascular reactivity.
Assumptions of NIRS Reactivity Assessment
Our interpretation of the NIRS data has been based on certain assumptions. First, metabolic rate does not change during inhalation of CO2, and therefore changes in concentrations of Hb and Hbo2 are purely due to hemodynamic effect (changing diameter of the resistive vessels) providing that the cerebral perfusion pressure stays constant. Should metabolic rate change at all during excitation of CO2, the amount of consumed oxygen would change, and therefore the concentrations of Hbo2 and Hb would change as well, rendering the reactivity assessment unreliable. Fortunately, as Siesjö30 reported, there is no change in metabolic rate observed except very high levels of Paco2 (>11 kPa).
Second, the path length factor remains the same for all individuals. The quantification of the concentrations of chromophores requires a knowledge of the path length traveled by the light as it traverses the tissue. Because of the effect of scattering of light by the tissue, the actual distance covered by the light is greater than the geometric spacing between optodes by a (DPF).31 The average value of the DPF was measured in the adult head by the time-of-flight measurements of an ultrashort optical pulse through the tissue.28 This study showed that for interoptode distances higher that 2.5 cm, the DPF remains approximately constant, with changes in the geometric position of the probes, and that its interindividual variation is surprisingly very small (5.93±0.42 [SD]). A recent study by Duncan et al32 using phase-resolved spectroscopy produced slightly greater DPF values (6.26±0.88 at a wavelength of 807 nm); however, the difference was statistically not significant. Therefore, our assumption of a constant DPF is probably valid. The accuracy and reliability of the technique will increase when the DPF measurement facility is incorporated into the spectrometer. Phase-resolved spectroscopy seems to be a suitable technique for this purpose since the equipment needed to perform the measurements is potentially compact, and it takes only a few seconds to record path length from several wavelengths.32
Third, the behavior of NIRS parameters is primarily governed by hemodynamic changes in the brain tissue. One of the main criticisms of NIRS is possible contamination by the influence of the extracranial circulation. To what degree that problem is important is still unknown. There has been an attempt to negate this influence by subtracting readings from two receiving probes placed in line with the transmitting probe.33 Since path length of light through extracranial tissue should be almost identical for both receiving probes, the difference in absorption coefficients between the two probes should reflect changes taking place in cerebral tissue alone. However, as demonstrated by Germon et al34 and Harris and Bailey,35 this difference is still greatly influenced by the extracranial circulation. We have adopted the principle that when only two probes are used, the relative contribution from the extracranial tissue becomes less as the interoptode distance increases.36 During carotid endarterectomy the change in NIRS parameters when the clamp was removed from the internal carotid artery was far greater than when the clamp was removed from the external carotid artery.23 In the present study we have made an additional attempt to evaluate the relationship between extracranial tissue circulation (assessed by cutaneous laser-Doppler technique)37 and changes in concentration of hemoglobin. The poor correlation between these variables provides further support for the notion that the influence of the extracranial circulation is small compared with the intracranial tissue. However, we acknowledge that the laser-Doppler sampling volume is small and may not detect significant changes in blood flow in the deeper parts of the scalp.
Positioning of Probes: Difficulties in Standardization
We believe that the standard positioning of the probes (as presented) is crucial to the success of NIRS in providing stable indexes of cerebral oxygenation change. The most practical approach from the point of view of standardization would be to place the probes above and along the eyebrow. The distance between the optodes should be as high as 5 to 6 cm, and the probes should sample only from the MCA territory, avoiding sinuses and muscles. The sinuses are thought to channel light away from the brain,36 and muscles increase the influence of extracranial circulation on the NIRS signals. The preferred place to put the optodes is along the midline but slightly away from it. That is not always possible unless the head can be shaved, and therefore in practice the place of attachment of the probes varies from individual to individual. This may introduce errors in patients with focal brain pathology, but further experience is needed to define them. A further problem of variation in probe position is the possible spatial variability of DPF on the head.
Spatial Resolution of NIRS
The sampling volume of NIRS in adults is small. Obrig et al38 demonstrated in their study on functional brain activation that NIRS response quality critically depends on optode positioning. Maximum responses were seen with the optodes positioned around the motor area for the hand. This finding indicates that the NIRS signal reflects localized cerebral oxygenation changes. In normal brain or in the brain without focal lesions, it is probably safe to assume that the NIRS measurements from the small region represent global changes of oxygenation of the cerebral tissue to a global hemodynamic challenge. However, additional concerns arise when patients with focal cerebral disease are examined or when tests that selectively stimulate focal areas of the cortex are applied.
Lack of correlation between flow velocity reactivity to CO2 and age in our data cannot be treated as conclusive because of the very uneven age distribution, which was skewed heavily toward young people. It is inevitable that the majority of volunteers are in the younger group aged 20 to 40 years. In our study only eight people (16%) were in the group aged older than 50 years, and none were in the group aged older than 70 years. Such an uneven age distribution may miss the effect of age on reactivity. Therefore, it was of considerable interest that this effect was detected by NIRS. The reactivity to CO2 expressed with the use of the Hbdiff parameter showed a significant (P<.009) tendency to decrease with age. This is in agreement with the study of Hoshi and Tamura,39 who showed that in elderly people mental work is not associated with an increase in total hemoglobin concentration (impaired cerebrovascular reactivity) but rather with reciprocal change in Hbo2 and Hb.
Can NIRS Replace Other Modalities Used for Cerebral Reactivity Testing?
TCD provides an established index of CO2 reactivity in the form of a flow velocity ratio. Some investigators consider PI to be superior to flow velocity for the assessment of reactivity.40 This ratio of pulse amplitude to the mean value is reported to be a good index of cerebral vasodilatation during hypercapnic challenge.41 This observation was also confirmed in the present study. PI showed a significant decrease during vasodilatation induced by increase in Etco2.
Both TCD reactivity indexes were found to be significantly correlated with NIRS reactivity indexes. However, some discrepancy in the reaction of NIRS and TCD to CO2 occurred, which may have been associated with errors introduced by inaccurate measurement of the interoptode distance, variations in DPF caused by interindividual anatomic variations and differences in probe positioning, and extracranial contamination of the NIRS signal. Furthermore, reactivity assessed by TCD flow velocity may be biased, with some error caused by small changes in diameter of the insonated vessel.16 18 19
The two techniques measure different quantities at different sites. TCD assesses “big tube” blood flow within the MCA, while NIRS monitors relative changes in cerebral tissue oxygenation. This may partly explain the temporal differences in TCD and NIRS signal changes. The NIRS trends seem to lag behind the flow velocity trend in that the Hbo2 and Hb signals reached a plateau response to CO2 approximately 1 minute later than the flow velocity. This inertia cannot be caused by the algorithm because in our intraoperative monitoring of carotid endarterectomy, the responses of Hbo2 and Hb to the clamping of the internal carotid artery were immediate.23 Therefore, the blood volume and state of oxygenation of hemoglobin behave differently for direct manipulation of perfusion pressure and differently for manipulation of tissue acidosis with the use of inhalation of CO2. However, the reason for this phenomenon remains unclear to the authors.
We stress that the results obtained in this study represent reactivity in normal people. In cerebral pathologies the correlation between TCD and NIRS may differ because of additional factors such as extracranial-intracranial anastomosis. A study in patients with carotid stenosis, which addresses this problem, is currently under way.42 Another important issue is assessment of interindividual variability in reactivity indexes, which requires repeated assessments of reactivity in each individual and awaits address.
This study demonstrates that NIRS detects linear changes of Hb and Hbo2 in response to CO2 alterations and indicates that it may be further developed to provide a supplementary method for testing cerebrovascular reactivity. This may prove of particular clinical importance in patients in whom TCD is precluded because of a poor ultrasound window.
Selected Abbreviations and Acronyms
|ABP||=||arterial blood pressure|
|DPF||=||differential path length factor|
|Hbdiff||=||Hbo2 minus Hb|
|MABP||=||mean arterial blood pressure|
|MCA||=||middle cerebral artery|
|Sao2||=||peripheral arterial oxygen saturation|
|TCD||=||transcranial Doppler sonography|
We would like to thank all the participating volunteers for their help in conducting the study.
- Received May 29, 1995.
- Revision received August 31, 1995.
- Accepted September 2, 1995.
- Copyright © 1995 by American Heart Association
Hedera P, Traubner P, Bujdakova J. Short-term prognosis of stroke due to occlusion of internal carotid artery based on transcranial Doppler ultrasonography. Stroke. 1992;23:1069-1072.
Klingelhöfer J, Sander D. Doppler CO2 test as an indicator of cerebral vasoreactivity and prognosis in severe intracranial hemorrhages. Stroke. 1992;23:962-966.
Bishop CCR, Powell S, Rutt D, Browse NL. Transcranial Doppler measurement of middle cerebral artery blood flow velocity: a validation study. Stroke. 1986;17:913-915.
Piepgras A, Schmiedek P, Leinsinger G, Haberl RL, Kirsch CM, Einhöupl KM. A simple test to assess cerebrovascular reserve capacity using transcranial Doppler sonography and acetazolamide. Stroke. 1990;21:1306-1311.
Pickard JD, Boisvert DPJ, Graham DI, Fitch W. Late effects of subarachnoid haemorrhage on the response of the primate cerebral circulation to drug induced changes in arterial blood pressure. J Neurol Neurosurg Psychiatry. 1979;42:899-903.
Powers WJ, Raichle ME. Positron emission tomography and its application to the study of cerebrovascular disease in man. Stroke. 1985;16:361-376.
Aaslid R, Lindegaard KF, Sorteberg W, Nornes H. Cerebral autoregulation dynamics in humans. Stroke. 1989;20:45-52.
Czosnyka M, Guazzo E, Iyer V, Kirkpatrick P, Pickard JD, Smielewski P, Whitehouse H. Testing of cerebral autoregulation in head injury by waveform analysis of blood flow velocity and cerebral perfusion pressure. Acta Neurochir Wien. 1994;60(suppl):468-471.
Dahl A, Russell D, Nyberg-Hansen R, Rootwelt K, Bakke SJ. Cerebral vasoreactivity in unilateral carotid artery disease. Stroke. 1994;25:621-625.
Kontos HA. Validity of cerebral arterial blood flow calculations from velocity measurements. Stroke. 1989;20:1-3.
Halsey JH. Effect of emited power on waveform intensity in transcranial Doppler. Stroke. 1990;21:1573-1578.
Gomez CR, Gomez SM, Hall IS. The elusive transtemporal window: a technical and demographic study. J Cardiovasc Technol. 1989;8:171-172.
Reynolds EOR, Wyatt JS, Azzopardi D, Delpy DT, Cady EB, Cope M, Wray S. New non-invasive methods for assessing brain oxygenation and haemodynamics. Br Med Bull. 1988;44:1052-1075.
Kirkpatrick PJ, Smielewski P, Czosnyka M, Menon DA, Pickard JD. Near infrared spectroscopy in head injury patients. J Neurosurg. In press.
Wyatt JS, Cope M, Delpy DT, Richardson CE, Edwards AD, Wray S, Reynolds EOR. Quantification of cerebral blood volume in human infants by near-infrared spectroscopy. J Appl Physiol. 1990;68:1086-1091.
Elwell CE, Cope M, Edwards AD, Wyatt JS, Delpy DT, Reynolds EOR. Quantification of adult cerebral hemodynamics by near-infrared spectroscopy. J Appl Physiol. 1994;77:2753-2760.
Van der Zee P, Cope M, Arridge SR, Essenpreis M, Potter LA, Edwards AD, Wyatt JS, McCormick DC, Roth SC, Reynolds EOR, Delpy DT. Experimentally measured optical pathlengths for the adult head, calf and forearm and the head of the newborn infant as a function of interoptode spacing. In: Goldstick TK, Cabe MML, Maguire DJ, eds. Oxygen Transport to Tissue, XIII. New York, NY: Plenum Press; 1992.
Siesjö BK. Brain Energy Metabolism. New York, NY: John Wiley & Sons; 1978.
Germon TJ, Kane NM, Manara AR, Nelson RJ. Near-infrared spectroscopy in adults: effects of extracranial ischaemia and intracranial hypoxia on estimation of cerebral oxygenation. Br J Anaesth. 1994;73:503-506.
Obrig H, Kleinschmidt A, Merboldt KD, Dinargl U, Frahm J, Villringer A. Monitoring of cerebral blood oxygenation during human brain activation by simultaneous high-resolution MRI and near-infrared spectroscopy. Proc Soc Magn Reson. 1994;1:67. Abstract.
Fürst H, Hartl WH, Janssen I. Patterns of carebrovascular reactivity in patients with unilateral asymptomatic carotid artery stenosis. Stroke. 1994;25:1193-1200.
Czosnyka M, Richards HK, Whitehouse HE, Pickard JD. Relationship between transcranial Doppler pulsatility and cerebrovascular resistance: an experimental study. J Neurosurg. In press.
Smielewski P, Czosnyka M, Kirkpatrick P, Whitehouse H, Pickard JD. Cerebrovascular reactivity in patients with carotid stenotic disease: study using TCD and near-infrared spectroscopy. In: Proceedings of the British Neurosurgery Research Group. J Neurol Neurosurg Psychiatry. In press.