Internal and External Carotid Contributions to Near-Infrared Spectroscopy During Carotid Endarterectomy
Background and Purpose The internal carotid (ICA) and external carotid (ECA) contributions to changing concentrations of oxyhemoglobin (Hbo2) and deoxyhemoglobin (Hb) during carotid endarterectomy were assessed with the use of near-infrared spectroscopy (NIRS).
Methods NIRS optodes were placed on the forehead with an interoptode distance of 6 cm, and laser-Doppler flowmetry (LDF) was used to monitor the change in skin blood flow between the optodes. Hb, Hbo2, LDF, arterial blood pressure, and middle cerebral artery flow velocity were recorded continuously. The ECA was clamped 2 minutes before the ICA was clamped. Suitable multimodal recordings were achieved in 44 patients.
Results When the ECA was clamped, 76% of patients showed a fall in Hbo2 and 65% an increase in Hb. When corrected for changes in arterial blood pressure, an accompanying fall in cutaneous LDF predicted the fall in Hbo2 with high sensitivity (100%) and specificity (100%). Among those with no NIRS changes during ECA clamping, 56% had severe ECA stenosis or occlusion; none of these showed an accompanying fall in LDF. In contrast, when the ICA was clamped, substantial additional changes in NIRS occurred in 55% of cases, all of which were associated with a fall in flow velocity, but none with a change in LDF. Patients with a constant flow velocity after ICA clamping also showed no change in NIRS.
Conclusions Both the ECA and ICA vascular territories contribute to NIRS changes during carotid endarterectomy. The external carotid contribution to NIRS can be monitored with cutaneous LDF.
Near-infrared spectroscopy has been shown to be useful in monitoring changes in the chromophores Hbo2 and Hb in various tissues, including brain.1 2 3 4 5 In the adult head, where only reflectance spectroscopy (scattered light sampled by an ipsilateral receiving probe) is possible, an estimation of the path length factor has been calculated, allowing quantification.6 7 Thus, NIRS can demonstrate physiological changes in cerebral Hbo2 and Hb content during different components of the respiratory cycle,1 8 temporary carotid compression,9 cardiopulmonary bypass10 and CEA,11 and in response to external photic stimuli,12 cognitive challenges,2 and blood pressure changes.13
The contribution from extracerebral tissues in adult NIRS remains unresolved.14 15 Near-infrared light passes through a number of layers of tissue before reaching the brain and may scatter within the extracerebral layers (including the cerebrospinal fluid layer) without sampling brain tissue. These tissues are predominantly supplied by the ECA. Of these layers, skin contains the highest concentration of chromophores and is probably the most significant source of extracerebral contamination in adult NIRS.14 Various methods attempt to minimize extracerebral contamination by subtracting the cerebral and extracerebral contributions (spatial resolution).16 17 Alternatively, the different physiological properties of the cerebral and extracerebral circulation may be exploited. For example, during a CO2 challenge, cerebral blood flow increases markedly with CO2, whereas there are only slight accompanying changes in skin blood flow.18 Under these controlled conditions, relative NIRS signal changes can be confidently attributed to the cerebral compartment.
In this study we attempted to estimate the extracerebral contribution to NIRS changes seen during CEA.14 15 By clamping the ECA before the ICA under highly controlled intraoperative conditions, we observed the changes in chromophore concentration occurring during each phase of surgery. In addition, we monitored the relative changes in cutaneous blood flow using skin LDF and relative changes in intracranial flow using TCD. The respective changes in blood flow in the different anatomic compartments were compared with the NIRS changes seen during each part of the surgical procedure. In this manner, we may have identified a method for discriminating, with high specificity and sensitivity, the extracerebral contribution to the NIRS changes.
Subjects and Methods
CEA was performed between September 1995 and April 1996 on 50 patients (mean age, 69 years; range, 54 to 86 years) with symptomatic carotid stenosis. Twenty-five presented with transient ischemic attacks, 13 with minor strokes, and 12 with amaurosis fugax. All had ipsilateral carotid stenosis of 50% or greater. Twenty patients also had significant contralateral disease with moderate to severe carotid stenosis (n=15) or occlusion (n=5). All patients were independent before operation. Preoperative assessment included bilateral carotid and cerebral digital subtraction angiography and CO2 reactivity testing.19 20 21
CEA was performed under general anesthesia. Monitors used during surgery were as follows: (1) invasive ABP monitor (Servomed, Hellige); (2) bilateral cerebral function monitor (Multitrace 2, Lecromed) to assess cerebral cortical electric activities; (3) NIRS system (NIRO-500, Hamamatsu Photonics KK) placed on the ipsilateral frontal region1 ; (4) LDF (MBF3D monitor and modified P3 probe, Moor Instrument Ltd) to monitor cutaneous blood flow between the NIRS optodes; (5) TCD (PCDop 842, Scimed; and CDS, Neuroguard Inc) to monitor ipsilateral middle cerebral artery FV; and (6) end-tidal CO2 monitor (Multinex, Datascope Corp).
The end-tidal CO2 was kept stable during operation. ABP was maintained stable during the operation. A selective shunt was used when there was a persistent fall in cerebral function or TCD showed evidence of ischemia (fall to <40% of baseline).21
The NIRS system (NIRO-500, Hamamatsu Photonics KK) used in this study employs four laser diodes for measuring light of four wavelengths (775, 825, 850, and 904 nm).22 The laser light is carried to the patient by a fiberoptic cable (optode). Scattered light is collected by a second optode and detected by a photomultiplier tube. The three main chromophores monitored are Hbo2, Hb, and CytO2, each having a characteristic absorption spectrum for the individual transmitted wavelength. Sampling frequency was set at 2 Hz. This system allows the data to be displayed graphically and numerically as changes in chromophore concentration (micromoles per liter) according to the algorithm of Wray et al.23 The interoptode distance was maintained at 6 cm. We avoided placing the optodes near the temporalis muscle and the sagittal sinus. This occasionally required shaving a small area of the scalp. The optodes were kept in position by means of a plastic optode holder and elastic bandage.
LDF Skin Blood Flow Monitoring
Many of the commercially available LDF monitors use a wavelength that coincides with that adopted by the NIRS system. This interference prevents correct assessment of cerebral oxygenation. Thus, the LDF monitor manufactured by Moor Instrument Ltd was fitted with a compatible laser diode that used light in the visible spectrum (wavelength of 635 nm). The LDF probe was placed between the two optodes of the NIRS system to maintain a distance of 3 cm from the optodes. LDF did not provide an absolute value of skin blood flow, and only arbitrary units were used. Moreover, the zero reading on LDF did not correspond to biological zero. Therefore, LDF data were categorized as increased, no change, or decreased for comparison with other parameters.
ABP, Hb, and Hbo2 as determined by NIRS, FV, LDF, and cerebral function data were captured as analog signals. The signals were converted to digital format with the aid of a 12-bit bipolar analog-to-digital converter (DT 2814, Data Translation), registered, and processed on a laptop computer with specific software.24 The signals were time averaged over 4 seconds and were presented graphically on the screen of the laptop computer. Markers were inserted at the time of application and release of the carotid clamps.
Interrupted time series analysis for specified epochs was used to determine relative changes in the parameters according to the preceding events. A 1-minute forecast (with 95% confidence intervals of the parameter) was constructed from the 5-minute data preceding the clamping of ECA. This forecast was compared with the observed changes occurring at the time of ECA clamping. Similarly, a 30-second forecast was constructed from the 2-minute data before the clamping of ICA (Fig 1⇓). The forecast was based on the best-fitting line of the third order with the use of a standard statistical program (SPSS for Windows, SPSS Inc). This model was chosen to account for the possible changes expected after an abrupt drop in blood flow in the tissue (which might include an exponential decline in the tissue oxygenation), secondary dilatation of the microcirculation (due to autoregulation or opening of collateral vessels), and correction for drifting blood pressure. The changes in ABP, NIRS, TCD, and LDF parameters at the time of clamping of vessels were then classified as increased, no change, or decreased. The quantitative data were tested for normality with the Kolmogorov-Smirnov test. Interval data with severely skewed distribution were dichotomized for comparison. The relationships between the presence and absence of changes were compared with Fisher’s exact test.
Six cases were excluded from the analyses because of computer errors or other technical problems. Additional case episodes were excluded from analysis of specific events (clamping or release of an artery) when there was a significant change in ABP during that episode.
Rate of Changes in Chromophores
When ICA was clamped, there were variations in the rates of changes in chromophore concentrations between patients; NIRS reached steady state within 0.5 minute in 89% (16/18) of cases, and 94% (17/18) reached steady state within 1 minute. The rate of changes in concentration of chromophores was slower when the ECA was clamped; 24% (7/29) of the patients reached the plateau within 0.5 minute, 55% (16/29) reached the plateau within 1.5 minutes, and 21% (6/29) did not reach the plateau before the ICA was clamped.
Overall Patterns of NIRS Signal Changes
Twenty-nine patients had no significant changes in ABP when the ECA and ICA were clamped. Of these, no significant changes in NIRS signals occurred during clamping of the ECA and ICA in 2 patients. In 9 patients the NIRS changes occurred only when the ECA was clamped and in 6 others only when the ICA was clamped. In the remaining 12 patients, the NIRS changes were biphasic, with significant NIRS signal changes after both ECA and ICA clamping (Fig 2⇓). The magnitude of the Hbo2 signal from the preclamped baseline suggests that 8 of these patients had greater changes when the ICA was clamped. Thus, the ICA contribution to NIRS signal changes was predominant in 14 patients, whereas the ECA dominated in 13 patients.
NIRS Signal Changes After ECA Clamping
At the time of ECA cross-clamp application, 37 patients maintained a stable ABP. On clamping, 28 (76%) showed a significant reduction in Hbo2, 24 of whom also had a reciprocal increase in Hb. Of the remaining 9 patients, 5 had severe (>50%) stenosis of the ECA demonstrated angiographically (Table 1⇓).
The NIRS changes after ECA clamping demonstrated a close association with cutaneous LDF changes. A fall in LDF was 100% sensitive and 100% specific in predicting Hbo2 changes after ECA clamping. For Hb changes, LDF was 80% sensitive and 100% specific in predicting the rise in Hb (Table 2⇓).
TCD monitoring of FV was achieved in 33 patients. There was a small transient reduction in FV (54 to 44 cm/s) in only one of these patients after ECA clamping. This particular patient had reverse flow in the ipsilateral ophthalmic artery demonstrated on the preoperative angiogram, that is, flow from ECA to ICA. Two other patients demonstrated reverse flow in the ophthalmic artery: in one TCD monitoring was not possible because of a poor bone window, and in the other no changes in FV were seen during ECA clamping.
NIRS Signal Changes After ICA Clamping
Thirty-three patients maintained a stable ABP at the time of ICA cross-clamp application. Of these, 18 (55%) showed a significant reduction in Hbo2 and a reciprocal increase in Hb. Patients with more severe ipsilateral ICA stenosis were more likely to show a fall in Hbo2 after ICA cross-clamping than those with less severe stenosis. On the other hand, patients with more severe ipsilateral carotid stenosis were less likely to show a fall in Hbo2 (Table 3⇓). Furthermore, we considered the degree of stenosis of both ICAs together: if contralateral ICA stenosis was the same as or more severe than that of the ipsilateral ICA, 13 of 14 cases demonstrated a fall in Hbo2; if contralateral ICA stenosis was less severe, then 14 of 19 cases had no change in Hbo2.
No patients showed a change in LDF after ICA clamping provided that ABP was stable. There was a significant association between NIRS changes and a fall in FV. The FV measurements were 100% sensitive to a fall in Hbo2 after ICA clamping (Table 4⇓). However, FV was only 36% specific to NIRS changes during ICA clamping. Nonetheless, the Hbo2 changes after ICA clamping correlated with the percent drop in FV (r=.68, P<.001).
NIRS Signal Changes on ECA Release
Thirty patients maintained a stable ABP by the time the ECA was released. Twelve (40%) showed no change in Hbo2 and Hb during this phase of surgery. The remainder demonstrated a rise in Hbo2 and a fall in Hb that was significantly associated with LDF increases. Thus, cutaneous LDF was 82% sensitive and 100% specific in predicting Hbo2 changes during ECA release (Table 5⇓). Two patients had an increase in FV without significant changes in ABP. Both of them also had associated rises in LDF and Hbo2 and a reduction in Hb. The increase was small compared with the subsequent increase in FV on ICA release.
NIRS Signal Changes on ICA Release
Forty patients had stable ABP when ICA was released; of these, 18 had an increase in Hbo2. The increase in Hbo2 was not significantly associated with an increase in FV (Table 6⇓). Two patients had increases in the LDF signal accompanying increases in FV and Hbo2.
Our results have shown that with the use of a current NIRS monitoring technique and a wide interoptode distance, both ICA and ECA contributions to NIRS signals are significant. However, cutaneous LDF monitoring is highly sensitive and specific in predicting the changes in the ECA component of NIRS signals during carotid surgery. Furthermore, NIRS changes during ICA clamping were unrelated to ECA blood flow provided that a stable blood pressure was maintained.
Significance of ABP Changes
Our previous experience with NIRS in head-injured patients indicated that NIRS was highly sensitive to ABP changes.25 The present study indicates that cutaneous LDF monitoring is also sensitive to ABP variation. However, the effect of blood pressure on the cerebral and extracerebral circulation is highly variable between patients and probably depends on factors such as pressure autoregulation.13 26 27 Therefore, it is necessary to control ABP when attempting to isolate the separate components of the NIRS signals. We used interrupted time series analysis to assess the change in ABP, and patients were excluded if changes from baseline trends were significant. The criteria adopted were stringent, and therefore a relatively large number of patients were excluded on the basis of ABP changes. In the majority of these cases, the ABP changes were small and did not affect the direction of change in NIRS signals. However, the effects on the magnitude of NIRS measurements remained uncertain.
Time Course of NIRS Signal Changes After Vascular Clamping
The NIRS signal changes were rapid and were completed within a minute after ICA clamping in most cases. However, after ECA clamping the change occurred over a longer time period; only 55% reached steady state within 1.5 minutes. This observation probably relates to the difference in tissue blood flow and collateral circulation in the brain and extracerebral tissues. This longer period for achieving steady state after clamping of the ECA has two implications. First, if the duration between clamping of the ECA and ICA is too short, the contribution from the ECA may be underestimated. Second, changes in the chromophores in the extracerebral tissue may continue beyond the time when the ICA was clamped. The contribution from the intracranial tissue may then be overestimated. To address these problems, we allowed 2 minutes between ECA and ICA clamping. However, we did not extend the interval until steady state was reached in all cases, since we were concerned that other influences, such as a change in blood pressure or opening of cutaneous collateral blood supply, might intervene during a prolonged interval. Furthermore, we did not wish to extend the length of the operation excessively. Thus, the interruption of the trend of change in chromophores was adopted as evidence of an ICA contribution in patients in whom the steady state had not been reached, and the maximum deviation from the expected trend was used to quantify that contribution.
Intracranial Contribution to NIRS Signal Changes
The present study provides further evidence that when the extracerebral contribution was removed (by clamping the ECA), NIRS reflected changes in cerebral hemoglobin oxygenation. NIRS changes after ICA clamping were always associated with FV changes and not with skin blood flow variation. These observations were dependent on the severity of the ipsilateral and contralateral ICA disease. Thus, most of the patients with unilateral carotid stenosis showed no change in NIRS, despite the fact that many patients had a fall in middle cerebral artery FV. These observations support the hypothesis that frontal NIRS readings during CEA are dependent on the integrity of the cortical collateral circulation.11 The time resolution of the NIRS in our monitoring settings is 4 seconds so that very transitory changes in intracranial chromophore concentration might not be detected.
Extracerebral Contribution to NIRS Signal Changes
Many investigators using NIRS are still of the opinion that the majority of signal changes are derived from the scalp.28 29 In a previous study we found very few NIRS changes on clamping and releasing of the ECA. We now believe that this was due to the use of the older NIRS system (NIRO-1000, Hamamatsu Photonics Inc), which employed larger optodes that were more sensitive to extraneous light. A pressure bandage was necessary to use this machine, which probably rendered the underlying skin tissue ischemic. The current NIRS system (NIRO-500, Hamamatsu Photonics Inc) has smaller optodes that are held in a plastic mold. This allows the mold to be used for a prolonged period without discomfort to the patient or risk of scalp necrosis. In this study we were primarily addressing the extracerebral contribution to the NIRS changes during CEA and did not apply a pressure bandage. Our results showed that there was a significant contribution to NIRS during CEA in a patient subgroup. These changes were small in the presence of significant ECA stenosis.
Some NIRS systems claim to minimize extracranial contributions by adopting two or more pairs of receiving optodes placed in line with the transmitting probe, providing the substrate for spatial resolution.30 However, this approach does not eliminate the extracranial influence,28 29 and other methods are necessary. Cutaneous LDF was used to assess the skin contribution to forearm NIRS,27 and the use of cutaneous LDF as an index of changes in the extracerebral blood flow in combination with NIRS was described by Smielewski et al.18 Since scalp, skull, and dura are supplied by the ECA, the absence of change in cutaneous LDF indicates the absence of change in all tissues supplied by the ECA. Cerebral NIRS changes were associated with relative changes in cerebral blood flow (FV), whereas extracerebral NIRS changes were paralleled by relative changes in cutaneous blood flow (LDF). Thus, our experience indicates that cutaneous LDF may help to define the extracerebral component of NIRS changes in adults, providing the substrate for physiological resolution of the signal changes rather than relying on more traditional methods of anatomic resolution.
Influence of EC-IC Anastomosis
An ECA contribution by means of an EC-IC anastomosis was observed in one patient in whom the FV fell after ECA clamping. The EC-IC anastomosis was shown on preoperative angiography as reversed flow in the ophthalmic artery. Contribution from EC-IC anastomoses during clamping of the ECA and ICA was not common in our series. These cases may be identified by preoperative angiography or TCD.
On the contrary, after a period of clamping of the carotid arteries, new EC-IC anastomoses were detected with the use of LDF and FV monitoring. In two of our subjects, when ECA clamping was released there was a small increase in FV without ABP changes. These were accompanied by both an increase in LDF and corresponding changes in NIRS signals. The increases were not artifacts, and the FV waveforms were normal. The EC-IC flow was not shown on preoperative angiography, and there was no change in FV when the ECA was clamped. We believe that after clamping of the ICA in the absence of an intraoperative shunt, the intracranial vessels dilate with reduced resistance to blood flow, thus encouraging movement of extracerebral blood into the cerebrum. Similarly, we observed a number of occasions when LDF increased when ICA was released. This may represent flow from the cerebrum into a dilated cutaneous vascular bed.
There was a significant contribution to NIRS signal changes from both the ICA and ECA in patients undergoing CEA, even when a wide interoptode distance of 6 cm was adopted. The ECA contribution was related to skin blood flow and can be monitored by specifically gated LDF. LDF provided a good correlation with changes in Hbo2 and Hb in the tissues supplied by the ECA. With improvement in LDF, comonitoring of relative cutaneous blood flow with NIRS may prove useful for estimating the extracerebral contribution to NIRS signal changes in the adult head.
Selected Abbreviations and Acronyms
|ABP||=||arterial blood pressure|
|ECA||=||external carotid artery|
|ICA||=||internal carotid artery|
|TCD||=||transcranial Doppler sonography|
The authors wish to thank Moor Instrument Ltd for supplying the modified LDF monitor and Neuroguards Inc for donation of the CDS TCD monitor for the duration of the study. The LDF and NIRS monitors were purchased with a University of Cambridge equipment research grant.
- Received July 12, 1996.
- Revision received August 30, 1996.
- Accepted August 31, 1996.
- Copyright © 1997 by American Heart Association
Jobsis FF. Noninvasive infrared monitoring of cerebral and myocardial oxygen sufficiency and circulatory parameters. Science. 1977;198:1264-1267.
Reynolds EO, 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.
Van der Zee P, Cope M, Arridge SR, Essenpreis M, Potter LA, Edwards AD, Wyatt JS, McCormick DC, Roth SC, Reynolds EO, 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. Adv Exp Med Biol. 1992;316:143-153.
Hampson NB, Camporesi EM, Stolp BW, Moon RE, Shook JE, Griebel JA, Piantadosi CA. Cerebral oxygen availability by NIR spectroscopy during transient hypoxia in humans. J Appl Physiol. 1990;69:907-913.
Germon TJ, Young AE, Manara AR, Nelson RJ. Extracerebral absorption of near infrared light influences the detection of increased cerebral oxygenation monitored by near infrared spectroscopy. J Neurol Neurosurg Psychiatry. 1995;58:477-479.
Germon TJ, Evans P, Barnett N, Wall P, Nelson RJ. Optode separation determines sensitivity of near infrared spectroscopy to intra- and extracranial oxygenation changes. J Cereb Blood Flow Metab. 1996;15:617. Abstract.
Smielewski P, Kirkpatrick P, Minhas P, Pickard JD, Czosnyka M. Can cerebrovascular reactivity be measured with near-infrared spectroscopy? Stroke. 1995;26:2285-2292.
Mancini DM, Bolinger L, Li H, Kendrick K, Chance B, Wilson JR. Validation of near-infrared spectroscopy in humans. J Appl Physiol. 1994;77:2740-2747.
Germon J, 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.