The best method of quantifying cerebral perfusion pressure (CPP) has not been definitively established. Traditionally, CPP has been defined as the difference between arterial blood pressure (ABP) and intracranial pressure (ICP). This model is based on the assumption that cerebral veins collapse if their intraluminal pressure decreases below ICP and, therefore, the effective downstream pressure of the cerebral circulation is determined by a Starling resistor at the level of the veins. Recently, there has been increased interest in the arteriolar critical closing pressure (CrCP) model originally proposed by Burton.1 In this model, the CrCP of a Starling resistor located at the level of the arterioles determines the effective downstream pressure of the circulation and CPP is then defined as the difference between ABP and CrCP. According to this model, if ABP is constant, then changes in cerebral blood flow (CBF) are either due to a change in CrCP, a change in cerebrovascular resistance (CVR), or changes in both. Similarly, the mechanism of cerebral autoregulation, whereby CBF remains nearly constant over a range of ABP, must act via changes in CrCP, CVR or both. In practice, it may be difficult to determine if observed changes in cerebral hemodynamics are primarily mediated by changes in CrCP or CVR, because both are thought to be strongly influenced by arteriolar smooth muscle tone and therefore these variables can be expected to change simultaneously.
The most common approach to estimating CrCP in humans compares waveforms from simultaneous recordings of middle cerebral artery blood velocity (Vmca) and ABP. Within individual cardiac cycles, there is usually an approximately linear relationship between instantaneous ABP and Vmca. The inverse of the slope of the Vmca-ABP relationship is termed the resistance area product (RAP) and can be taken to be an index of CVR.2,3 When the line describing this relationship is extrapolated, it typically intersects the pressure axis at a positive value, implying that flow would cease if ABP decreased to the value indicated by the pressure-intercept. This apparent zero flow pressure (aZFP) has been taken to be an estimate of the CrCP of the cerebral circulation.4 This analysis assumes that instantaneous CBF within an individual cardiac cycle is determined by two variables: 1) an apparent cerebral perfusion pressure (aCPP) equal to the difference between ABP and aZFP, and 2) a CVR which is constant throughout the cardiac cycle. A variety of methods for determining aZFP have been described and these methods differ both in the technique for measuring ABP and the method of analyzing the Vmca-ABP relationship. In this study, ABP was measured continuously via a radial artery catheter and we compared two methods for defining the Vmca-ABP relationship: a linear regression method and a Fourier analysis method.
While multiple studies have established that aZFP varies inversely with changes in the arterial partial pressure of carbon dioxide (Paco2),4–8 the effects of drugs such as anesthetics and vasoactive drugs on aZFP and RAP have not been adequately investigated. In this study, we investigated the effect of reducing Paco2 on aZFP and RAP during isoflurane anesthesia. We also investigated the effects of an increase in ABP induced by the α-adrenergic agonist, phenylephrine, and we compared those effects in two conditions: when cerebral autoregulation was impaired by isoflurane and when autoregulation had been restored by hyperventilation.
The ABP and Vmca waveforms used in the present study were recorded during a previous investigation of the effects of isoflurane and Paco2 on cerebral autoregulation.9 Subjects were 11 healthy patients scheduled for elective surgery requiring general anesthesia with mechanical ventilation. The experimental procedures were approved by the local ethics review committee and informed consent was obtained from each patient.
The anesthetic technique, monitoring techniques, and techniques for manipulation of ABP and arterial carbon dioxide have been previously described.9 Briefly, anesthesia was induced with propofol and maintained with isoflurane in 100% oxygen. The end-tidal concentration of isoflurane was increased in each patient until periods of suppression were noted on the frontal electroencephalogram, then the end-tidal isoflurane concentration was reduced by 0.1% to 0.2% (to avoid burst-suppression on the electroencephalogram). The resultant end-tidal isoflurane concentration, 1.6% ± 0.2% (mean ± sd), was kept constant for the duration of the study. ABP was monitored via a radial artery catheter with the transducer referenced to the height of the external auditory meatus and arterial blood samples were obtained to determine Paco2. Subjects' lungs were mechanically ventilated and Paco2 was adjusted by altering respiratory rate and/or tidal volume and/or bypassing the CO2 absorber. Transcranial Doppler probes were positioned over the temporal windows to record Vmca and were held in constant position by a head frame. If required, a remifentanil infusion was commenced to achieve a mean ABP (MAP) below 80 mm Hg. Once hemodynamics had stabilized (as judged by a steady heart rate, Vmca and MAP) phenylephrine was infused to increase the ABP to above 100 mm Hg over approximately 2 min. This procedure was performed twice in each subject in random order: once at normocapnia (Paco2 38–43 mm Hg) and once at hypocapnia (Paco2 27–34 mm Hg). The ABP waveform, the spectral outline of the Vmca signal, and the capnograph were recorded at a sampling rate of 100 Hz on a personal computer via an analog-to-digital converter (ADInstruments, Sydney, Australia).
The Vmca recording in one subject was not suitable for aZFP analysis due to frequent loss of Doppler signal during diastole; therefore, the present study is based on data from 11 of the 12 subjects in the previous study. Two methods were used to determine the aZFP: a linear regression technique (aZFPr) and a Fourier technique (aZFPF). These analyses were performed automatically using custom software written in MatLab (Mathworks, Natick, MA). To remove high frequency artifacts, the ABP and Vmca data were filtered with a low-pass filter with a 3 dB frequency of 4 Hz. For the linear regression technique, the least-squares method was used to determine the line of best fit between filtered ABP and filtered Vmca for each individual cardiac cycle. To account for time delays between the Vmca recording and the ABP recording, the ABP data were time-shifted to maximize the correlation coefficient between ABP and Vmca. For each cardiac cycle, aZFPr was determined from the MAP, mean Vmca (mca) and the slope of the regression line according to the formula: aZFP = MAP − mca/slope. For the Fourier technique, the start and end of each cardiac cycle were determined from the filtered data, and Fourier analysis was applied to the unfiltered data from each cardiac cycle individually. The amplitude of the fundamental (first harmonic) components of ABP and Vmca (ABP1 and Vmca1, respectively) were determined and aZFPF was calculated according to the formula: aZFPF = MAP − mca × (ABP1/Vmca1). This equation assumes that the relationship between ABP and Vmca is a straight line that passes through the point (MAP, mca) at a slope of Vmca1/ABP1. The following formula was used to calculate aCPP: aCPP = MAP − aZFP. The inverse of the Vmca/ABP slope was used to calculate RAP.
To account for the effect of respiration, for each experimental condition the results from individual cardiac cycles were averaged over either two or three complete respiratory cycles. Suitable Vmca recordings were available from both middle cerebral arteries in seven subjects, and in those cases the results from the left and right sides were averaged.
Cerebral autoregulation was assessed using the autoregulation index (ARI) as previously described.9 The ARI is a dimensionless number with 0 indicating Vmca increased passively in proportion to the increase in MAP, and 1.0 indicating “perfect” autoregulation with mca remaining unchanged, despite the change in MAP. Autoregulation is designated as significantly impaired when ARI ≤0.4.
We were interested in comparing the contributions of changes in aZFP and changes in RAP to the observed cerebrovascular responses. Panerai et al.10 described a method for assigning absolute values to the effects on mca of MAP, aZFP and RAP; however, we were unable to use their approach because one of their underlying assumptions (that the fractional change in RAP is very small) did not apply to our data. Instead, we adopted an approach of comparing hemodynamic variables as a ratio of their values before and after a change in conditions. By definition,
Therefore, the following equation can be used to describe relative changes in mca:
When MAP changes, the autoregulatory response usually limits the change in mca, requiring aZFP and/or RAP to change in the same direction as MAP. However, if neither aZFP nor RAP changed, mca would be determined by the equation:
where mca′ indicates the expected mca if aZFP and RAP remain constant and subscripts (1) and (2) indicate parameters before and after a change in MAP.
Following a change in conditions (such as a change in MAP or Paco2), aZFP and/or RAP may change and the associated mca may differ from mca′. The following equation describes the determinates of the ratio of the new mca (mca2) to mca′:
where subscripts (1) and (2) refer to parameters before and after the change in conditions. We used this equation to compare the contributions of changes in aZFP and RAP to the observed cerebrovascular responses. Expressed as a percentage, the effect of a change in aZFP is defined as
and the effect of a change in RAP is defined as
For example, if RAP increases causing the ratio RAP1/RAP2 to equal 0.4, this would be associated with a 60% reduction in the ratio mca/mca′. If, at the same time, aZFP increases causing the ratio (MAP2 − aZFP2)/(MAP2 − aZFP1) to equal 0.5, this would be associated with a further 50% reduction in mca/mca′. Both factors combined would then be associated with a mca 20% of the value that would have been expected had there been no change in either aZFP or RAP.
For comparison of paired data, the Wilcoxon's signed rank test was applied. For repeated measures, after confirming that the data did not fail a test for normality (using the D'Agostino-Pearson test), repeated measures one-way analysis of variance with Tukey's posttest for multiple comparisons was performed. Statistical analysis was performed with GraphPad Prism 5.0a for Macintosh (GraphPad Software, San Diego, CA). Data are summarized as mean ± sd unless otherwise stated.
Method of Calculating aZFP
The difference between the two methods of calculating aZFP (aZFPF − aZFPr) was 0.5 mm Hg ± 3.6 mm Hg (mean ± 2sd) (Fig. 1). Given the close agreement between the techniques, only the data based on aZFPF are presented in this report.
Effects of Increasing MAP With Phenylephrine
The effects of increasing MAP with a phenylephrine infusion are summarized in Table 1. As detailed in the previously published results,9 cerebral autoregulation was significantly impaired in 8 of 11 subjects during normocapnia but was not significantly impaired in any subject during hypocapnia. The ARI was 0.74 ± 0.12 and 0.36 ± 0.27 at hypocapnia and normocapnia, respectively (P = 0.007).
Significant impairment of autoregulation with normocapnia is indicated by the 19% increase in mca with increasing MAP (Fig. 2, lines A and D). With normocapnia, increasing MAP was associated with no significant change in aZFP, however, there was a significant increase in RAP (Table 1). The results shown in Table 1 are the combined results of the 11 subjects. In the subgroup of three subjects who were judged not to have significantly impaired autoregulation with normocapnia, increasing MAP was associated with a slight decrease in average aZFP (from 28 to 26 mm Hg), whereas the RAP increase (from mean of 0.69 to 0.95 mm Hg · cm−1 · s−1) was adequate to minimize the change in mca.
During hypocapnia, intact autoregulation is illustrated by the lack of significant change in mca (Table 1 and Fig. 2, lines B and C). The autoregulatory response during hypocapnia was associated with an increased aZFP, although this was insufficient to offset the increase in MAP. Consequently, aCPP increased by 40% ± 16%. There was also a significantly increased RAP, illustrated by a decrease in the slope of line C compared to line B in Figure 2. The relative contributions of aZFP and RAP to the autoregulatory response are summarized in Table 2. The increased RAP contributed a 23% reduction in mca compared to the mca that would have been expected had RAP and aZFP remained unchanged. In contrast, the increased aZFP contributed only a 7% reduction in mca.
Effects of Hypocapnia
Compared to the velocity at normocapnia, mca decreased with hypocapnia by 42% ± 8% and 48% ± 6% at the higher and lower MAP respectively (Table 1). The effect of hypocapnia at the lower MAP is illustrated by comparing lines A and B in Figure 2. Because comparative measurements at normocapnia and hypocapnia were made at closely matched ABPs, there would have been minimal change in mca if both aZFP and RAP had remained unchanged. Therefore, the equations described in the Methods section for determining the contributions of aZFP and RAP can be used to estimate their actual percentage effects on mca. For example, at the lower ABP range, aZFP and RAP both increased with hypocapnia, accounting for an 11% and 34% decrease in mca, respectively (Table 2), and these results are consistent with the overall 42% decrease in mca.*
Three interventions influencing cerebral hemodynamics were examined in this study: 1) reduction of Paco2 during isoflurane anesthesia, 2) phenylephrine-induced increase in ABP when cerebral autoregulation was impaired by isoflurane, and 3) phenylephrine-induced increase in ABP when autoregulation was restored by hypocapnia. Significant increases in both aZFP and RAP were seen when arteriolar tone was expected to increase, i.e., when Paco2 was reduced and when cerebral autoregulation responded to increasing ABP. In each situation examined, changes in RAP appeared to have a greater impact on the cerebrovascular response than changes in aZFP.
The theory of a precapillary CrCP was originally proposed by Burton who surmised that the CrCP of arterioles is determined by two forces tending to close the vessels: extramural pressure (ICP in the case of the brain) and arteriolar wall tension.1 Arteriolar wall tension arises from a combination of the stretched elastic components of the vessel wall and active contraction of vascular smooth muscle. Our subjects would not be expected to have increased ICP and any change in ICP during the study would be minimal, therefore changes in CrCP were likely to be mainly due to changes in arteriolar smooth muscle tone. However, we cannot exclude the possibility of some decrease in ICP with hyperventilation, which would tend to counteract the effect of increased vascular smooth muscle tone on CrCP. There is a paucity of data on normal values of aZFP in healthy adults, however, typical values are usually around 30 mm Hg.3 In the present study, the average aZFP during normocapnia was somewhat lower than this quoted norm, possibly reflecting the cerebral vasodilatatory effect of isoflurane.
We compared the effects of normocapnia and hypocapnia at carefully matched ABPs. In previous studies, significant changes in aCPP were reported when Paco2 was either increased or decreased in awake subjects, with little or no change in RAP.4,6,7 The implications of these previous results is that hypocapnia reduces CBF by decreasing CPP and that, contrary to common understanding, hypocapnia does not increase CVR. In contrast, we found RAP increased significantly with hypocapnia, and the increase in RAP appeared to have a greater effect on Vmca than the increase in aZFP. The discrepancy between our results and the results of previous studies in awake subjects is difficult to explain. It seems unlikely that Paco2 affects CBF via different mechanisms in awake compared to anesthetized subjects. It is worth noting that the earlier studies used continuous ABP waveforms obtained noninvasively from a servo-controlled cuff, either around the upper arm4 or a finger,6,7 whereas ABP waveforms in this study were obtained invasively from the radial artery.
In this study, we compared the effect of increasing ABP on aZFP and RAP in subjects with and without significant impairment of cerebral autoregulation. The results of this study suggest that the cerebral autoregulatory response is more dependent on changes in CVR than changes in CrCP because the calculated contribution of RAP was greater than the calculated contribution of aCPP, and some RAP response remained evident when autoregulation was designated significantly impaired. Panerai et al.10 have reported evidence suggesting that the autoregulatory response in awake subjects is primarily dependent on changes in RAP. Contradictory findings were reported by Moppett et al.,11 who investigated the effect of a norepinephrine infusion in awake volunteers. They found the autoregulatory response to be entirely dependent on changes in aZFP, with no change at all in RAP. The discrepancy between this result and the present study could be due to differences in experimental conditions: different vasopressor drugs were used and our subjects were anesthetized. Alternatively, the discrepancy could be due to differences in measurement techniques. Moppett et al.11 measured ABP by noninvasive oscillometry and aZFP was determined by extrapolation from only two points on the pressure-velocity relationship (mean and diastolic). We believe results based on oscillometry should be interpreted with caution because errors of ≥10 mm Hg in diastolic blood pressure are not uncommon.12 Also, if the relationships between pressures measured with a cuff around the arm and pressures in the middle cerebral artery were not constant, then interpretation of changes in RAP and aZFP would be invalid. Although this latter consideration applies to any method of measuring ABP in an extremity, it seems likely that analysis based on continuous recording over multiple cardiac cycles is more reliable than analysis based on oscillometry, which requires multiple cardiac cycles to provide a single ABP estimate.
In our study, aZFPr and aZFPF could be considered interchangeable with minimal bias and close limits of agreement between the two techniques. This result contrasts with the findings of Aaslid et al.13 who reported a significant bias with mean aZFPF 4.6 mm Hg more than mean aZFPr. However, Panerai et al.14 found a difference of 4.5 mm Hg in the opposite direction. Several authors have suggested that the Fourier method may be preferable to the regression method due to the first harmonic being less sensitive to artifacts and less sensitive to distortion of the ABP waveform within the arterial tree.13–15 The technique we used of averaging individual results from multiple cardiac cycles over an integral number of respiratory cycles may have reduced the effects of artifacts and reduced the variability of both analysis techniques.
Several assumptions are required when using RAP as an index of CVR and aZFP as an estimate of CrCP, and these assumptions are largely untested. First, it is difficult to know how closely radial artery pressure approximates the pressure waveform within the middle cerebral artery. Second, although there is some evidence that relative changes in Vmca over time are an accurate representation of relative changes in CBF, there is no direct evidence that instantaneous velocity is proportional to instantaneous flow within the cardiac cycle. For example, if there are significant pulsatile changes in middle cerebral artery diameter, then the velocity waveform would be expected to under-estimate the pulsatility of actual flow, which would lead to an overestimate of CrCP. Third, it is problematic to assume a constant resistance in a pulsatile system. If there is significant compliance in the system, then pressure and flow cannot be assumed to follow a linear relationship. Furthermore, compliance could be altered by changes in arteriolar tone, further complicating the interpretation of changes in aZFP and RAP. Indirect evidence, however, suggests the human cerebral arterial system is relatively noncompliant.16 Fourth, extrapolation of the pressure-velocity relationship to zero flow assumes these parameters would continue to follow the same linear relationship as they decreased below the physiological range; an assumption that has been questioned on theoretical grounds.3
Interpretation of our results would be confounded if phenylephrine had a direct action on cerebrovascular smooth muscle. However, sympathomimetic drugs are not thought to directly influence CBF,17 and phenylephrine increases CBF in situations where autoregulation is impaired,18 suggesting a lack of significant direct cerebral vasoconstriction.
In summary, during isoflurane anesthesia, aZFP and RAP both increased when Vmca was reduced by hypocapnia. The same variables also increased with the cerebral autoregulatory response to phenylephrine-induced increases in ABP. When autoregulation was impaired, aZFP did not change when ABP increased, although RAP retained some of its responsiveness to ABP. These results imply that, during isoflurane anesthesia, changes in arteriolar tone control CBF by altering both the resistance and the effective downstream pressure of the cerebral circulation. Quantitatively, changes in resistance appeared to have a greater influence than changes in effective downstream pressure.
1.Burton AC. On the physical equilibrium of small blood vessels. Am J Physiol 1951;164:319–29
2.Evans DH, Levene MI, Shortland DB, Archer LN. Resistance index, blood flow velocity, and resistance-area product in the cerebral arteries of very low birth weight infants during the first week of life. Ultrasound Med Biol 1988;14:103–10
3.Panerai RB. The critical closing pressure of the cerebral circulation. Med Eng Phys 2003;25:621–32
4.Aaslid R. Cerebral hemodynamics. In: Newell DW, Aaslid R, eds. Transcranial Doppler. New York: Raven Press, 1992:49–55
5.Weyland A, Buhre W, Grund S, Ludwig H, Kazmaier S, Weyland W, Sonntag H. Cerebrovascular tone rather than intracranial pressure determines the effective downstream pressure of the cerebral circulation in the absence of intracranial hypertension. J Neurosurg Anesthesiol 2000;12:210–16
6.Hancock SM, Mahajan RP, Athanassiou L. Noninvasive estimation of cerebral perfusion pressure and zero flow pressure in healthy volunteers: the effects of changes in end-tidal carbon dioxide. Anesth Analg 2003;96:847–51
7.Panerai RB, Deverson ST, Mahony P, Hayes P, Evans DH. Effect of CO2
on dynamic cerebral autoregulation measurement. Physiol Meas 1999;20:265–75
8.Garnham J, Panerai RB, Naylor AR, Evans DH. Cerebrovascular response to dynamic changes in pCO2
. Cerebrovasc Dis 1999;9: 146–51
9.McCulloch TJ, Boesel TW, Lam AM. The effect of hypocapnia on the autoregulation of cerebral blood flow during administration of isoflurane. Anesth Analg 2005;100:1463–7
10.Panerai RB, Moody M, Eames PJ, Potter JF. Cerebral blood flow velocity during mental activation: interpretation with different models of the passive pressure-velocity relationship. J Appl Physiol 2005;99:2352–62
11.Moppett IK, Sherman RW, Wild MJ, Latter JA, Mahajan RP. Effects of norepinephrine and glyceryl trinitrate on cerebral haemodynamics: transcranial Doppler study in healthy volunteers. Br J Anaesth 2008;100:240–4
12.de Greeff A, Reggiori F, Shennan AH. Clinical assessment of the DINAMAP ProCare monitor in an adult population according to the British Hypertension Society Protocol. Blood Press Monit 2007;12:51–5
13.Aaslid R, Lash SR, Bardy GH, Gild WH, Newell DW. Dynamic pressure-flow velocity relationships in the human cerebral circulation. Stroke 2003;34:1645–9
14.Panerai RB, Sammons EL, Smith SM, Rathbone WE, Bentley S, Potter JF, Evans DH, Samani NJ. Cerebral critical closing pressure estimation from Finapres and arterial blood pressure measurements in the aorta. Physiol Meas 2006;27:1387–402
15.Czosnyka M, Smielewski P, Piechnik S, Al-Rawi PG, Kirkpatrick PJ, Matta BF, Pickard JD. Critical closing pressure in cerebrovascular circulation. J Neurol Neurosurg Psychiatry 1999;66: 606–11
16.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;22:1148–54
17.Olesen J. The effect of intracarotid epinephrine, norepinephrine, and angiotensin on the regional cerebral blood flow in man. Neurology 1972;22:978–87
18.Strebel S, Lam AM, Matta B, Mayberg TS, Aaslid R, Newell DW. Dynamic and static cerebral autoregulation during isoflurane, desflurane, and propofol anesthesia. Anesthesiology 1995;83: 66–76
*(1−0.11) × (1−0.34) ≈ (1−0.42).