Assessment of Autoregulation by Means of Periodic Changes in Blood Pressure
Background and Purpose The aim of this study was to test the hypothesis that the phase difference that occurs between an induced oscillation in blood pressure and the resultant oscillation in middle cerebral artery (MCA) flow velocity could reflect the competence of cerebral autoregulation.
Methods Fourteen volunteers performed 19 cycles of 10 seconds of squatting followed by 10 seconds of standing. Peak MCA velocity was measured with transcranial Doppler ultrasound, and blood pressure was measured with a servo-controlled finger plethysmograph held level with the head. Waveforms from each cycle were added to obtain averaged waveforms of arterial blood pressure and MCA velocity. These results were processed by Fourier analysis to extract the phase difference between the fundamental components of velocity and pressure. Each volunteer performed the exercise three times: first breathing normally, secondly hyperventilating (hypocapnia), and finally while breathing air containing 5% carbon dioxide (hypercapnia). Under these conditions the volunteers were expected to have normal, enhanced, and impaired autoregulation, respectively.
Results The measurements made with normal breathing showed a phase lead of velocity ahead of pressure of 46±14° (mean±SD). We noted a highly significant reduction in phase lead with hypercapnia (P<.00015) (Wilcoxon signed rank test, two-tailed) and a highly significant increase in phase lead with hypocapnia (P<.002).
Conclusions The results support our hypothesis and may lead to a technique for assessing the competence of cerebral autoregulation.
Cerebral autoregulation is the process by which cerebral blood flow is held relatively constant despite changes in cerebral perfusion pressure.1 A rapid quantitative method of assessing cerebral arterial autoregulation would be of value in a wide range of clinical situations. These may include timing of surgery for clipping of intracranial aneurysm, management of head injury, and the treatment of hypertension. Early investigations were limited to the steady-state behavior of cerebral blood flow. More recently attention has been focused on dynamic behavior, which covers a wide range of time scales.
Transcranial Doppler ultrasound provides a continuous measurement of blood flow velocity in the basal cerebral arteries.2 This technique is now being used in many centers to investigate the dynamics of cerebral autoregulation.3 4 5 6
To investigate the dynamics of cerebral autoregulation it is necessary to observe the effect on flow of clearly defined changes in pressure. Three useful pressure changes, by virtue of their mathematical simplicity, are random fluctuations (containing all frequencies), step changes, and periodic oscillations.
Random input has been investigated by Giller,5 who studied the responses of intracranial blood flow velocity to spontaneous fluctuations in blood pressure and found that the autoregulatory mechanism acted as a high-pass filter, able to smooth out the lower frequencies but not the higher ones. Moreover, in clinical situations in which impairment of the autoregulation was likely, the mechanism became less able to respond to relatively slower frequencies. (In effect, the cutoff frequency for the filter was decreased.)
The response to a step change in blood pressure has been investigated by Aaslid et al,3 who produced a step drop in arterial blood pressure (ABP) by rapidly releasing bilateral thigh tourniquets. With this technique Aaslid et al were able to demonstrate that full recovery can be seen as early as 4.1 seconds after the step decrease in ABP. They also showed that the recovery is slower during hypercapnia and faster during hypocapnia. The step drop in ABP lasts “for only 5-7 seconds before reflexes start to restore ABP”; the technique is therefore best suited to assessment of faster rates of autoregulation.
The aim of this study was to investigate the use of periodic variations in ABP as a means of assessing cerebral autoregulation, with particular attention to the phase of induced oscillation in middle cerebral artery (MCA) flow velocity (MCAV). It has been shown by earlier investigators studying autoregulation that cerebral vascular resistance is extremely sensitive to carbon dioxide partial pressure.7 8 Aaslid et al3 have recently shown that hypercapnia and hypocapnia respectively impair and enhance the dynamics of cerebral autoregulation. We therefore used changes in arterial carbon dioxide as a means of modifying autoregulation.
Subjects and Methods
After obtaining approval from our local Research Ethics Committee, we investigated 14 volunteers (7 men, 7 women; mean age, 25 years [range, 18 to 38 years]). All the volunteers had normal blood pressure with no history of cardiac or cerebral pathology, and all gave informed consent. Each was asked to perform 19 cycles consisting of 10 seconds of squatting followed by 10 seconds of standing. Pilot studies had demonstrated that this maneuver gives an approximately sinusoidal oscillation in ABP. A period of 20 seconds challenges the faster components of the autoregulation mechanism without being too strenuous for the volunteers. The procedure was repeated during hypocapnia induced by hyperventilation to the point of light-headedness and during hypercapnia induced by inhalation of 5% carbon dioxide in air from a Douglas bag via a nonreturn valve and divers’ mouthpiece. In each case, the procedure was delayed until end-tidal carbon dioxide had remained stable for 1 minute.
Peak MCAV was measured with 2-MHz pulsed transcranial Doppler ultrasound (EME TC2-64) through the temporal window. ABP was measured with a servo-controlled finger plethysmograph (Ohmeda Finapres 2300) on a finger maintained at the level of the head. The Finapres is best suited to applications such as this in which rapid changes in blood pressure are the focus of interest. Its weaknesses, namely baseline drift and occasionally unpredictable offsets, are not significant here because it is principally the timing of the changes in blood pressure that is used in our analysis.
Both MCAV and ABP were recorded throughout at sampling intervals of 20 milliseconds with the use of a 12-bit analog-to-digital converter installed in an IBM-compatible computer. Subjects squatted or stood as indicated by “traffic lights” under the control of the computer, with the times of these indications recorded as trigger signals for later off-line averaging. End-tidal carbon dioxide was measured at 3-minute intervals with the use of an in-line capnometer (Hewlett-Packard, 47210A).
Off-line, the first cycle was discarded. Subsequent cycles, in which the subjects’ response pattern was established, were used in the analysis. The 18 cycles were averaged to reduce the relative amplitude of the cardiac pulsations. The averaged waveforms of ABP and MCAV were then processed by means of Fourier analysis to extract the phase difference between the fundamental components of the ABP and MCAV. The differences between normocapnic, hypercapnic, and hypocapnic results were tested for significance with the two-tailed Wilcoxon signed rank test.
The derived phase shifts are interpreted in the light of a linear model of autoregulation.
All subjects were normotensive at rest, and all completed the protocol. Mean±SD end-tidal carbon dioxide partial pressure was 37.4±3.2 mm Hg in the normocapnic series, 25.9±3.5 mm Hg in the hypocapnic series, and 50.6±2.5 mm Hg in the hypercapnic series.
Figs 1⇓, 2⇓, and 3⇓ show the results from a typical subject. The phase lead is clearly visible in the hypocapnic and normocapnic studies and is greatly reduced during hypercapnia. The mean±SD phase lead of velocity ahead of pressure during normocapnia was 46±14°, a time interval of 2.5 seconds, which cannot easily be explained by differences in the timing of signal detection.
Hypercapnia was shown to reduce the phase lead of the MCAV when compared with normocapnia (P<.00015); in fact, all subjects demonstrated a reduced phase lead. The phase lead during hypocapnia was increased compared with normocapnia (P<.002), with only one subject in the group showing anomalous behavior. Fig 4⇓ illustrates the changes in phase lead of MCAV with respect to ABP as carbon dioxide levels were varied.
The overall results are shown in the Table⇓; as expected, there was a mild trend for mean blood pressures to be higher during hypercapnia (P<.05). There was also a trend for the amplitude of the ABP oscillation to decrease with increasing end-tidal carbon dioxide (P<.005).
These results show that for normal subjects, when ABP is forced to oscillate with a period of 20 seconds, the MCAV also has an induced oscillation, the phase of which leads that of pressure. Furthermore, this phase lead decreases as arterial carbon dioxide pressure is increased.
The observed phase shifts may be interpreted as evidence of functioning autoregulation. Such phase shifts might be expected in a variety of different models of the dynamics of autoregulatory behavior. Here we argue their consistency with a simple linear model.
We suppose that in response to slowly changing blood pressure the autoregulation mechanism is effective in maintaining a constant blood flow. When blood pressure is altered very rapidly, however, autoregulation is ineffective, and blood flow follows the changes in pressure passively. Based on these assumptions, a differential equation may be used to model the system’s response.
To keep the model simple, we use a linear differential equation. In a linear model a sinusoidally oscillating input (ABP) induces a sinusoidal oscillation of the same frequency in the output (MCAV). For a given frequency of input, the phase of the output is determined by the characteristics of the system. In addition, we chose a second-order differential equation as sufficiently general to encompass the expected behavior of the system, allowing the possibilities of overshoot rebound and resonance.
Our model predicts an increasing phase lead with increasing competence of regulation, as we have observed. The phase lead will increase from 0° when autoregulation is absent to between 90° and 180° (depending on the exact model chosen) as autoregulation improves.
This model is linear, yet there are likely to be nonlinearities in the true system characteristics. For instance, autoregulation can operate successfully only within a limited range of blood pressure, and nonlinearities will arise at the limits of this autoregulation range. If the response is nonlinear, then the magnitude of phase lead is likely to depend not only on the frequency of the oscillations and the competence of the autoregulation system but also on the shape (if not sinusoidal) and the amplitude of the induced blood pressure oscillations. It is likely that a less restricted model would also manifest a phase shift dependent on autoregulation competence. Our observations are not sufficient to determine the most appropriate model to be applied.
A common criticism of Doppler measurements is that they measure velocity and not flow.9 If the diameter of the MCA is changing in response to our ABP oscillations, this will introduce further nonlinearities in the response that could have an impact on the measured phase. Aaslid et al4 have shown that the MCA diameter does not alter in response to a step drop in blood pressure. Our results suggest that even if this is a significant mechanism, it does not challenge the potential of this technique to investigate autoregulation.
Another possible factor involved in changes in MCAV with squatting and standing might be the effects of the associated changes in mental activity. However, this happens twice in each cycle, manifesting as a second harmonic. This term was small relative to the fundamental, so even if the response to squatting and standing was different, the effect on the fundamental is likely to be negligible.
This study supports the work of Aaslid et al3 4 and Giller5 6 but also suggests a new approach to the quantitative measurement of autoregulatory dynamics. It is still difficult to validate these assessments of autoregulation against other techniques. The mechanisms that govern autoregulation are complex, and these studies have only investigated the rapid response components of the system. Little would be gained by comparison with methods such as 133Xe washout. These techniques can only examine the response on a much slower time scale than most transcranial Doppler methods, including the one outlined in this article. These two approaches in effect measure such different aspects of the cerebral circulation that they are likely to remain complementary. It remains to be shown which of the two approaches will be the more valuable in guiding treatment or predicting outcome.
The authors are grateful to A.J. Murrills and Dr D. Laycock for their support throughout the study, to F. Clewlow and Dr G. Petley for their technical advice and valuable discussions, and to O. Sparrow, D. Lang, and Professor T. Shelley for their comments on the manuscript.
- Received November 9, 1994.
- Revision received January 19, 1995.
- Accepted February 14, 1995.
- Copyright © 1995 by American Heart Association
Harper AM. Autoregulation of cerebral blood flow: influence of the arterial blood pressure on the blood flow through the cerebral cortex. J Neurol Neurosurg Psychiatry. 1966;29:398-403.
Aaslid R, Lindegaard KF, Sorteberg W, Nornes H. Cerebral autoregulation dynamics in humans. Stroke. 1989;20:45-52.
Aaslid R, Newell DW, Stoos 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.
Lennox WG, Gibbs EL. The blood flow in the brain and the leg of man and the changes induced by alteration of blood gases. J Clin Invest. 1932;11:1155-1177.
Harper AM, Glass HI. Effect of alterations in the arterial carbon dioxide tension on the blood flow through the cerebral cortex at normal and low arterial pressures. J Neurol Neurosurg Psychiatry. 1965;28:449-452.
Kontos HJ. Validity of cerebral artery blood flow calculations from velocity measurements. Stroke. 1989;20:1-3.