Assessment of Cerebral Autoregulation Using Carotid Artery Compression
Background and Purpose A simple method of testing cerebral autoregulation by observing transcranial Doppler changes in middle cerebral artery flow velocity (FV) during a brief ipsilateral carotid artery compression (the transient hyperemic response test) was studied in 11 normal healthy volunteers. The aim of this study was to assess the reliability of the method and to compare derived autoregulatory indices with those of a standard noninvasive test of autoregulation, Aaslid's leg-cuff test.
Methods Volunteers were subjected to repeated carotid compressions and leg-cuff tests at different levels of CO2. Hypercapnia was induced using inhalation of a mixture of 5% CO2 in air. Hypocapnia was induced by moderate hyperventilation. To assess the influence of the duration of carotid compression, a series of carotid compressions lasting 3, 4, 5, 7, and 9 seconds were performed in random sequence. Monitored parameters included ipsilateral FV, end-tidal CO2, and arterial blood pressure. The transient hyperemic response ratio (THRR), calculated as the maximum increase of FV divided by baseline values after release of the carotid compression, was taken as the autoregulation index. This index was compared with the rate of autoregulation index derived from the leg-cuff test.
Results Both tests were significantly associated with end-tidal CO2 (ANOVA, P<.000001 for both carotid compression and cuff test). There was a linear correlation between THRR and autoregulation index (r=.86). However, the reproducibility of the THRR was more consistent than for the autoregulation index from single tests (13% versus 46%, respectively; P<.0001). Although the influence of the duration of carotid compression on THRR values was significant for carotid compressions lasting up to 5 seconds, there was no relation to the relative magnitude of FV drop during the compression.
Conclusions Brief (>5 seconds) carotid artery compression provides an index of cerebral autoregulation that is reproducible and is affected by CO2 tension in a fashion similar to autoregulatory indices derived from a standard leg-cuff test. The simplicity of the method provides a potentially useful addition to other noninvasive autoregulation tests for clinical assessments, particularly when repeated measurements are required.
Measurement of different cerebrovascular responses has gained clinical importance in the assessment of patients with a variety of cerebral pathology. Thus, impaired cerebrovascular reactivity to CO2 has been shown to be associated with worse outcome in patients after SAH1 and after severe head injury.2 Similarly, an impaired response to changes in ABP (autoregulation) is predictive of delayed ischemic deficits in SAH patients3 4 5 and a poor outcome after cerebral trauma.6 7 Methods that can assess these parameters reliably, repeatedly, and noninvasively are gaining clinical relevance.
TCD is a potential method for assessing cerebrovascular changes noninvasively.8 9 10 11 Testing an autoregulatory response may involve various ways of manipulating blood pressure, but drug-induced changes are cumbersome and not readily repeatable for sequential studies. The cuff test of Aaslid et al12 uses the sudden fall in blood pressure that occurs when a pair of leg cuffs is deflated. Recently, an alternative and much simpler method for testing autoregulation has been introduced, termed the “transient hyperemic response test.”13 The test was first applied in adult volunteers by Giller13 and theoretically analyzed by Czosnyka et al.14 The test assesses the response of MCA blood FV after a brief (<5 seconds) compression of the ipsilateral common carotid artery. Intact autoregulation is associated with vasodilatation during the period of carotid compression. Releasing the carotid compression results in a transient hyperemic overshoot when the perfusion pressure returns and acts on a dilated vascular bed (Fig 1⇓). If autoregulation is lost, the response is absent.
On the basis of previous studies,13 14 the carotid compression recordings from SAH patients allow an automatic computerized analysis of the hyperemic response.15 The test is simple, requires only TCD, but may be susceptible to examiner-dependent artifacts. The aim of this project was to compare the carotid compression method with the standard Aaslid leg-cuff test and to investigate the reliability and reproducibility of the test at different systemic CO2 levels.
Subjects and Methods
The study was approved by the Cambridge Local Research Ethics Committee.
Eleven healthy volunteers were examined, 6 men and 5 women ranging in age from 20 to 30 years. Carotid Doppler studies were carried out to exclude significant atheroma.
Volunteers were examined in a supine position while breathing through a face mask. The TCD probe (PC DOP 842, Scimed; 2-MHz probe) was positioned at the temporal window and fixed with a headband. The MCA was identified by recognition of the characteristic waveform, sound, typical FV (72±20), and depth of insonation.11 EtCO2 (901 Mk2, Morgan Instruments) was recorded continuously, as was the peripheral arterial oxygen saturation (pulse oximeter Multinex 4200, Datascope). The ABP was measured using an automated pressure cuff (Finapres, Ohmeda 2300).
Two large leg cuffs were wrapped around the volunteer's thighs. The cuffs were modified with larger tubing and inflated using a foot pump equipped with a pressure valve. The pressure within the cuffs was measured with a manometer. Changes in EtCO2 were induced by giving the volunteer a mixture of 5% CO2 in air and by voluntary moderate hyperventilation.
Signals of ABP, TCD, and EtCO2 were sampled at a frequency of 50 Hz and digitized with a 12-bit analogue-to-digital convertor (DT 2814, Data Translation). ABP, FV, and EtCO2 were calibrated in appropriate units. The signals were recorded and analyzed off-line using specific software (CVRTest by P. Smielewski, M. Czosnyka, and W. Zabolotny) with an IBM-compatible computer.
The examination consisted of two stages. (1) Five carotid compressions lasting 3, 4, 5, 7, or 9 seconds were performed. Sixty seconds was allowed between each compression to allow normalization of the cerebral blood flow to precompression levels. The order in which compressions were performed was randomized. (2) Two leg-cuff tests and two carotid compressions were performed at three different CO2 levels: normocapnia, hypercapnia, and hypocapnia. The leg cuffs were inflated to a pressure above systole for a period of 2 minutes and then deflated to induce a drop in systemic blood pressure. Before each of the tests, sufficient time (at least 2 minutes) was allowed for stabilization of baseline recordings.
Carotid compressions were accepted only when no further decrease in FV could be achieved and when stable conditions remained during the whole period of compression. If any confounding FV variations occurred, the compression was terminated and repeated 60 seconds later. Every attempt was made to achieve total occlusion each time the artery was compressed.
The methodology of the computerized analysis of the transient hyperemic response has been extensively described by Smielewski et al.15 Briefly, the THRR was calculated using the formulaTHRR|<|=|>|\frac|<|FVS_|<|hyperemia|>||>||<|FVS_|<|baseline|>||>|where FVS denotes systolic FV. The systolic value was used because it produced a stronger statistical association with clinical grades and showed smaller variability than the time-averaged value. Baseline systolic FV (FVSbaseline) was calculated using the average value of systolic FV from five heart cycles, ending with the one preceding the compression (Fig 1⇑). The hyperemic response (FVShyperemia) was calculated using the average systolic value of two heart cycles after the compression release with the exception of the very first cycle. Before the calculations, the FV signal was processed with the low-pass filter (cutoff frequency, 12.5 Hz) to eliminate spurious spikes.
The leg-cuff test index was calculated according to the formula of Aaslid et al:RoR|<|=|>|\frac|<|dCVR|>||<|dt|>||<|\cdot|>|\frac|<|1|>||<|CVR_|<|baseline|>||>||<|\cdot|>|\frac|<|ABP_|<|baseline|>||>||<||<|\Delta|>|ABP|>|where CVR is calculated as ABP divided by FV and ΔABP is the drop in blood pressure experienced after cuff deflation. The derivative of CVR (with respect to time) was calculated using linear regression of CVR against time for the first 3 to 5 seconds after cuff deflation.
The relative compression ratio describing magnitude of drop in FV during compression was defined asCompression Ratio|<|=|>|\frac|<|FVS_|<|baseline|>||<|-|>|FVS_|<|compression|>||>||<|FVS_|<|baseline|>||>||<|\cdot|>|100%where FVcompression was calculated as the average systolic value of the FV signal taken from the first two heart cycles of the compression (Fig 1⇑).
The data were first checked for normality using Shapiro-Wilks' W test. The data were then analyzed using either a parametric repeated measure of variance and Pearson correlation analysis or nonparametric Friedman ANOVA and Spearman's rank correlation analysis. Results from repeated tests during various levels of CO2 were averaged for further analysis, but the absolute relative difference between them (normalized by the average hyperemic increase above the baseline value) was taken as an index of reproducibility. This index was also used for analysis of influence of individual variations in compression ratio on THRR.
Effect of CO2 on THRR and Cuff Test Results
Close correlation was found between THRR and the level of EtCO2 (P<.00001, Figs 2⇓ and 3A). The average values at each CO2 level are given in the Table⇓. Aaslid's RoR also showed a close association with CO2 (P<.00001, Fig 3A⇓). Regression analysis of the THRR versus RoR gave a correlation coefficient of 0.86 (Fig 3B⇓). The mean RoRs calculated from different ABP drops after deflation of cuffs at different levels of CO2 are given in the Table⇓. The magnitude of the pressure drop was not related to the EtCO2.
Influence of the Duration of Carotid Compression and Compression Ratio
The relationship between the duration of carotid compression and THRR is shown in Fig 4⇓. The overall correlation was significant at a level of P=.015. However, for compressions lasting ≥5 seconds, THRR was independent of the compression time (there was no significant difference in THRR among groups with 5-, 7-, and 9-second compression).
The average compression ratio in normocapnic conditions varied between 36% and 57% among all 11 subjects, and the absolute difference in the compression ratio between two repeated tests ranged from 0.5% to 9.4%. No statistically significant correlation could be found between these inter-subject and within-subject variations of the compression ratio and the corresponding THRR (Fig 5⇓). The compression ratio showed a negative correlation to CO2 (P<.000002), with the smallest decrease in FV occurring at the lowest levels of EtCO2.
Reproducibility of the THRR
The total range of THRR values obtained was 1.105 to 1.29 (mean, 1.2; 95% confidence limits, 1.17 to 1.24). The relative difference between two repeated carotid compression tests for different levels of EtCO2 is summarized in the Table⇑. The overall variability of THRR was 13%, with 95% confidence limits of 8.9% to 17%. The reproducibility of THRR was not dependent on the levels of EtCO2. In contrast, the variability of the RoRs averaged 46% (95% confidence limits, 32.5% to 60%), with values at different CO2 levels summarized in the Table⇑.
Both the leg-cuff technique of Aaslid et al and the transient carotid compression test assess autoregulation by inducing a transient fall in cerebral perfusion pressure and monitoring the cerebral blood flow response using TCD flowmetry. The convenience of the latter presents an advantage, especially for clinical application in restricted areas (such as the intensive care unit) or when repeated measures are required. This study therefore addresses important issues concerning the effect of physiological variables on the transient hyperemic response, the effect of examiner-dependent variations, and the repeatability of the test result. Our data indicate that the carotid test performs well in comparison with the cuff test. Thus, despite theoretical concerns regarding manipulation of the carotid artery in the face of atheromatous disease or unruptured aneurysm, the THRT seems to provide a useful addition to the battery of noninvasive tests assessing cerebral autoregulation and warrants further evaluation in different clinical settings.
THRT Versus the Cuff Test
The cuff test is an established noninvasive technique for testing cerebral autoregulation.16 17 18 19 The method is based on a principle similar to that of the THRT in that the autoregulatory response is assessed by observing the MCA FV response to an abrupt, short-lasting fall in ABP. However, there is an important difference. Assuming that the intracranial pressure does not change significantly, the cuff test produces a global decrease in the cerebral perfusion pressure (ΔCPP≈ΔABP)12 and is thus less dependent on redistribution of blood flow within the circle of Willis. Provided the intracerebral pressure is low (as in normal volunteers), the derived RoR still provides a standard reference for assessment of the THRT. To make a comparison, we performed both tests at three different levels of Paco2, a known potent modulator of cerebrovascular tone, and the autoregulatory response.20 21 22 A low level of Paco2 enhances the autoregulatory response, while high levels of Paco2 abolish the mechanism completely. The anticipated modulatory effect was realized for both tests with almost identical sensitivity. In addition, direct comparison of THRR and RoR showed strong linear correlation, suggesting that despite concerns regarding the theoretical comparability of the tests, the results are similar. A remarkable finding was the low standard error of the mean values of THRR at each level of CO2. Thus, despite the nonquantifiable stimulus causing the fall in cerebral perfusion pressure, the hyperemic response is consistent between different subjects. However, there is a potential hazard when comparing the carotid compression test with other tests that cause a global reduction in cerebral perfusion pressure. If the varying conditions under which the tests are compared (eg, different CO2 level) influence in any way the distribution of flow in the circle of Willis, the response to carotid compression will reflect both changes in resistive vessel reactivity and diameter of the basal arteries. Fortunately, the effect of CO2 on the basal arteries is minimal,23 and results of our study therefore can be confidently considered as reflecting the genuine response of the small vessels.
Examiner- and Subject-Dependent Variability in the Carotid Compression Test
The THRT is based on a simple principle. Brief occlusion of the common carotid artery provokes a reduction in the perfusion pressure at the ipsilateral circle of Willis and is a stimulus to autoregulatory mechanisms.13 14 Because the drop in the ABP in the MCA is not known, the test assesses the hemodynamic response to a nonquantified stimulus. Providing the compression on the common carotid artery establishes total occlusion, the magnitude of drop in perfusion pressure is mainly governed by the effectiveness of the collateral circulation at the level of the circle of Willis. The heterogeneity of the anatomy of the circle of Willis is high,24 hence the difference in magnitude of the pressure drop is likely to differ significantly. Indeed, our present data showed a range of the relative drop in FV during compression of 35% to 57% among different individuals. According to simplified mathematical modeling,14 such a difference in stimulation should produce a marked difference in hyperemic response. However, correlation of the THRR with the strength of compression was not found to be significant either within or between individuals. Although it may be argued that the lack of significant association between individual variations in compression ratio and THRR is merely a reflection of the effect of confounding spontaneous fluctuations of cerebral blood flow rather than subtotal occlusions (Fig 6⇓), the same explanation cannot be used for inter-subject variability. Thus, despite the highly complex relationship between the magnitude of FV drop during compression and the resulting hyperemic response, the lack of correlation between THRR and the compression ratio suggests that many of the theoretical concerns may not be important in clinical practice. This is further confirmed by the significant association of THRR with clinical grades and outcome in patients after SAH15 and severe head injury.7 The problem of the nonquantified stimulus may be reduced if the test is used to monitor changes in autoregulation in the same patient. However, in pathological situations the effect of compression may vary with time. For example, in patients after SAH, spasms affecting the circle of Willis may significantly alter the changes in carotid perfusion pressure experienced when the test is repeated at different times. Another way of increasing the reliability of the THRR interpretation is to introduce a threshold above which the result is “positive” (good autoregulation) and below which it is “negative” (impaired autoregulation). This approach allowed the achievement of high statistical significance of correlations with major clinical grades in a preliminary study of SAH patients in which a threshold of 1.09 was used.15
Another theoretical factor that may influence the magnitude of the hyperemic response is the length of compression. Mathematical modeling suggests that compressions longer than the inherent autoregulation delay should generate a constant response.14 The time constant for autoregulation may be short. Some have indicated that the response is initiated within a few seconds and complete within less than 15 seconds.25 26 Others report an even faster response, with full restoration of blood flow to abrupt change in blood pressure occurring within 4.5 seconds.11 27 28 These observations of an ultrafast autoregulatory response are supported by our present data, which showed that the THRR was unaffected by the length of compression exceeding 5-second duration (up to 9 seconds). Although the average THRR seemed to reach the peak only for 7 seconds of compression, the difference in THRR among 5-, 7-, and 9-second compressions was not significant. Therefore, for clinical applications, compressions of 5- to 7-second duration should be used. It must be noted here, however, that there is also another slower mechanism, associated with the opening of collaterals, which causes partial restoration of the local cerebral perfusion pressure within a few minutes after carotid artery ligation.29
Finally, attention must be drawn to the moment of the compression release. Interpretation of the THRR is based on the assumption that increase in FV measured within a short (1 to 2 seconds) period after release of compression reflects the state of the arteriolar vasodilatation. However, in some cases, when the compression was released during end-diastole, there was a transient peak increase in FV (lasting <0.5 second) observed at the beginning of the hyperemic response (Fig 1⇑). This increase is probably associated with passive filling of the compliant arteries and should not be considered as evidence of active vasodilatation.19 Therefore, calculation of THRR should exclude data from the first heart cycle after release of compression.
Repeatability of the Results
The THRT results were found to be reproducible with relative variability being independent of CO2. Variability in the RoR parameter of the cuff test, however, was found to be much higher. This finding is inherent in the recommendations of Aaslid et al,12 suggesting that the test should be repeated four to six times to obtain an averaged response. There are two important factors that could be responsible for the difference in reproducibility seen between the THRT and cuff tests. First, the linear approximation for a pulsatile decrease in the estimated CVR after deflation of the leg cuffs may introduce errors unless several responses are averaged and a low-pass filter is applied. Second, the calculation of the RoR is derived from observations over 4 seconds after the cuff release. During this time, the ABP does not remain constant, modifying the pattern of changes in CVR and thus the linear approximation of the fall. In contrast, the THRT observes the initial 1- to 2-second period after compression release, which is comparable with the time delay in autoregulatory response.26 Because the disturbance in cerebral perfusion pressure is local (there was no change in ABP recorded during compression tests), restoration of the equilibrium in the circle of Willis is probably rapid. After initial passive filling of the arteries that immediately follows compression release (giving rise to an initial peak in FV, Fig 1⇑), the first two velocity waves used in the calculation of THRR represent flow through a maximally dilated cerebrovascular bed before a significant autoregulatory response.
It is recognized that carotid compression has to be avoided when the presence of atheromatous plaque or ulcer on the side of compression is expected.30 Another theoretical concern may arise when examining patients with aneurysm of any of the ipsilateral branches of the circle of Willis because of the theoretical risk of its rupture induced by the pressure pulse after the compression release. Although in the preliminary study on patients after SAH31 no complications have arisen from preoperative examinations, such a possibility cannot be entirely excluded. However, the Matas test32 (carotid compression and clinical observation) was traditionally used by many neurosurgeons as a crude determination of whether patients were fit for craniotomy for their aneurysm clipping.33
The carotid compression test provides a reproducible index of autoregulation that varies according to CO2 levels in a manner identical to that of the established noninvasive leg-cuff test. The result of the test is independent of the compression duration, providing the compression lasts for at least 5 seconds, and shows little sensitivity to the magnitude of drop in cerebral perfusion caused by the compression. The THRT proved to be simple in application and is suitable for regular examinations.
Selected Abbreviations and Acronyms
|ABP||=||arterial blood pressure|
|MCA||=||middle cerebral artery|
|TCD||=||transcranial Doppler ultrasonography|
|THRR||=||transient hyperemic response ratio|
|THRT||=||transient hyperemic response test|
This study was supported by the Raymond and Beverly Sackler Foundation and an overseas research scholarship award (P. Smielewski, PhD project). M. Czosnyka and P. Smielewski are currently on leave from the Institute of Electronics Fundamentals, Warsaw University of Technology (Poland).
- Received June 21, 1996.
- Revision received August 20, 1996.
- Accepted September 19, 1996.
- Copyright © 1996 by American Heart Association
Klingelhofer J, Sander D. Doppler CO2 test as an indicator of cerebral vasoreactivity and prognosis in severe intracranial hemorrhages. Stroke. 1992;23:962-966.
Czosnyka M, Kirkpatrick P, Guazzo E, Whitehouse H, Smielewski P, Pickard JD. Assessment of the autoregulatory reserve using continuous CPP and TCD blood flow velocity measurement in head injury. In: Nagai H, Kamiya K, Ishii S, eds. Intracranial Pressure, IX. Tokyo, Japan: Springer Verlag; 1994:593-594.
Czosnyka M, Smielewski P, Kirkpatrick P, Pickard JD. Monitoring of cerebral autoregulation in head-injured patients. Stroke. 1996;27:1829-1834.
Ringelstein EB, Sievers C, Ecker S, Schneider PA, Otis SM. Noninvasive assessment of CO2-induced cerebral vasomotor response in normal individuals and patients with internal carotid artery occlusions. Stroke. 1988;19:963-969.
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.
Aaslid R, Lindegaard KF, Sorteberg W, Nornes H. Cerebral autoregulation dynamics in humans. Stroke. 1989;20:45-52.
Tiecks FP, Lam AM, Aaslid R, Newell DW. Comparison of static and dynamic cerebral autoregulation measurements. Stroke. 1995;26:1014-1019.
Aaslid R, Newell DW, Stooss R, Sorteberg W, Lindegaard KF. Assessment of cerebral autoregulation dynamics from simultaneous arterial and venous transcranial Doppler recordings in humans. Stroke. 1991;22:1148-1154.
Ekstrom-Jodal B, Haggendal E, Linder LE, Nilsson NJ. Cerebral blood flow autoregulation at high arterial pressures and different levels of carbon dioxide tension. Eur Neurol. 1972;6:6-10.
Haggendal E, Johansson B. Effects of arterial carbon dioxide tension and oxygen saturation on cerebral blood flow autoregulation in dogs. Acta Physiol Scand. 1965;66:27-53.
Raichle ME, Stone HL. Cerebral blood flow autoregulation and graded hypercapnia. Eur Neurol. 1972;6:1-5.
Kontos HA, Wei EP, Navari RM, Levasseur JE, Rosenblum WI, Patterson JL. Responses of cerebral arteries and arterioles to acute hypotension and hypertension. Am J Physiol. 1978;234:H371-H383.
Aaslid R. Visually evoked dynamic blood flow response of the human cerebral circulation. Stroke. 1987;18:771-775.
Symon L, Held K, Dorsch NWC. A study of regional autoregulation in the cerebral circulation to increased perfusion pressure in normocapnia and hypercapnia. Stroke. 1973;4:139-147.
Leech PJ, Miller JD, Fitch W, Barker J. Cerebral blood flow, internal carotid artery pressure, and the EEG as a guide to the safety of carotid ligation. J Neurol Neurosurg Psychiatry. 1974;37:854-862.
Jennett B, Miller JD, Harper AM. Effect of carotid artery surgery on cerebral blood flow. London: UK: Excerpta Medica: Elsevier/North Holland Biomedical Press; 1976.
Smielewski P, Czosnyka M, Iyer V, Piechnik S, Whitehouse H, Kirkpatrick P, Pickard JD. Computerised transient hyperaemic response test: a method for the assessment of cerebral autoregulation and its relationship to outcome. In: Nagai H, Kamiya K, Ishii S, eds. Intracranial Pressure, IX. Tokyo, Japan: Springer-Verlag; 1994:632-633.