Monitoring of Cerebral Autoregulation in Head-Injured Patients

Autoregulation of cerebral blood flow is important in preventing secondary insults to the injured brain.¹ Disturbed autoregulation may result in an abnormal balance between cerebral blood flow, blood volume, and the metabolic requirement of the cerebral tissues. Transient episodes of cerebral ischemia, hyperemia,² and/or uncontrollable increases in cerebral blood volume³ are thought to be detrimental.

Although abnormal cerebral autoregulation is seen after head injury, the underlying mechanisms effecting this disturbance are not fully understood.⁴⁵ Endothelial dysfunction,⁶ vasospasm,⁷ and release of free radicals⁸ may be involved. Accordingly, little is known regarding how to restore autoregulation. Some pursue an “autoregulation-oriented” therapy that attempts to maintain an adequate CPP⁹ by pressure support and mild hypocapnia.¹⁰ Some centers have suggested that disordered cerebral autoregulation correlates with a poor outcome.¹¹¹² Since the autoregulatory reserve may vary with time, reliable and repeatable clinical tests of autoregulation are essential to guide therapy.

Various methods of assessing the cerebrovascular hemodynamic reserve are available. These include tests of carbon dioxide reactivity,⁴ stress tests relying on mechanical¹³ or pharmacological¹⁴ alteration of arterial pressure, and the transient hyperemic response test after carotid artery compression.¹⁵ All may use TCD to assess the dynamic response of cerebral FV in the basal arteries to various stimuli. The averaged FV values encompass a complex array of factors, making interpretation of abnormalities difficult.¹⁶ However, the value of TCD in head-injured patients has been facilitated by methods of continuous analysis of FV waveform. Observation of the respiratory-related fluctuations in arterial pressure and FV to assess cerebral autoregulation has been recently reported.¹⁷ The FV pulse waveform is of particular importance, and in head-injured patients FV pulsatility correlates with the cerebral arteriovenous oxygen difference as CPP falls, which has clinical significance.¹⁸

In the present study we examined the possibility that continuous assessment of autoregulatory reserve can be achieved by comparing changes in FVm and FVs with spontaneous changes in CPP. We verified the hypothesis that a generous autoregulatory margin should be associated with a better outcome, and we attempted to study changes in autoregulatory reserve in time after head injury.

Subjects and Methods

Patients

Eighty-two patients (27 females, 55 males; age range, 6 to 75 years; mean age, 36 years) were admitted to Addenbrooke's Hospital suffering from head injuries with a mean GCS score of 6 (range, 3 to 13). Forty-five percent of the patients had hematomas seen on first (68%) or subsequent (32%) scans (45% extradural, 30% subdural, 25% intracerebral), of which 77% were evacuated surgically. No patients in whom bone flaps were removed were included in this study. The patients were paralyzed, sedated, and ventilated to achieve mild hypocapnia. Falls in ABP that reduced CPP to <60 mm Hg were managed with alternating colloid and normal saline infusions, with supplementary inotropic agents if necessary (constant infusion of dopamine 2 to 15 μg/kg per minute). If ICP rose to >25 mm Hg, boluses of mannitol (200 mL of 20% for ≥20 minutes) were administered.

Monitoring

We monitored ICP continuously using a fiber-optic transducer (Camino Direct Pressure Monitor, Camino Laboratories) inserted intraparenchymally into the frontal region. Arterial pressure was monitored directly in the radial or dorsalis pedis artery (System 8000, S&W Vickers Ltd). The MCA was insonated daily for 20 minutes to 2 hours, starting from the day of admission until discharge or day 8 after head injury, with the PCDop 842 Doppler Ultrasound Unit (Scimed). Signals were monitored during periods of stable respiratory parameters, free from physiotherapy, tracheal suction, and other disturbances.

Data Capture

Analog outputs from the pressure monitors and the TCD unit (maximal frequency envelope) were connected to the analog-to-digital convertor (DT 2814, Data Translation) fitted into an IBM AT laptop computer (Amstrad ALT 386 SX). Data were sampled, digitized, and stored on the hard disk with the software for the waveform recording (WREC, W. Zabolotny, Warsaw University of Technology). Digital signals were then processed with software developed in-house (ICMR, M.C.).

Time-averaged values of ICP, ABP, and CPP (CPP=ICP−ABP) were calculated with the use of waveform time integration for 5-second intervals. FVm, FVs, and FVd were calculated after careful spectral filtration to reduce the influence of noise and averaged within the same 5-second periods.

Pearson's correlation coefficients¹⁹ between 36 consecutive samples of averaged CPP and FVm, FVs, and FVd were calculated for every 3-minute period. The correlation coefficients between CPP and FVd or FVm were not significantly different from one another with the exception of cases in which CPP was very low and no diastolic flow was recorded. Obviously, in such cases correlation of CPP with zero FVd values made little sense. Therefore, only the coefficients derived for FVm and FVs were included for final data analysis. Calculated correlation coefficients for FVm and FVs versus CPP, termed Mx and Sx, respectively, were then averaged over each recording.

Statistical Methods

The indices Mx and Sx and values of CPP, ICP, ABP, FV, and PI were averaged over all examinations made in different patients. Nonparametric statistical methods were generally used since the majority of variables did not have significantly normal distributions. The rank correlation coefficients were calculated between grade of initial coma (GCS score at admission), outcome (classified as favorable [good or moderate disability] or unfavorable [severe disability, persistent vegetative state, or dead] assessed 6 months after injury), patients' age, and the studied parameters. ANOVA (Kruskal-Wallis) was used to present differences in selected indices in separate outcome groups and according to days after injury. The Mann-Whitney U test was used to test for significance in difference from zero of Mx and Sx in consecutive days after head injury.

Results

Observation of Spontaneous Fluctuations of FV, ICP, and ABP

Recordings demonstrated typical patterns of spontaneous fluctuations of FV waveform during ICP plateau waves, intracranial hypertension, and incidental arterial hypotension.

During plateau waves of ICP (n=7 recordings), FVs remained constant, while FVd decreased (Fig 1). In contrast, during refractory intracranial hypertension (n=3 recordings) the rise of ICP produced a decrease in both FVs and FVd (Fig 1, bottom panel).

Incidents of arterial hypotension (Fig 2) were more frequent. During moderate decreases in ABP (mean CPP >70 mm Hg; n=15 recordings), FVs was stable, while FVd decreased. During more severe hypotension (mean CPP <60 mm Hg; n=8), both FVs and FVd decreased.

These observations suggest that FVs remains constant while FVd decreases during moderate falls in CPP, whereas both FVs and FVd decrease when CPP falls below 60 mm Hg.

Regression of FVs and FVd Versus CPP

Regression lines describing the baseline correlation between CPP and FV waveforms (FVs, FVm, and FVd) revealed three distinctive patterns (Fig 3), as follows. In group 1, all regression lines (for FVs, FVm, and FVd) had negative gradients (Sx <0, Mx <0; n=195 recordings in 49 patients). In group 2, the regression lines for FVs had a negative gradient, but those for FVm and FVd had positive gradients (Sx <0, Mx >0; n=95 recordings in 20 patients). In group 3, all regression lines had positive gradients (Sx >0, Mx >0; n=57 recordings in 13 patients).

Averaged values of GCS score on admission, outcome, and CPP in these three groups are presented in Table 1. In group 3, in which FVs and FVm were pressure-passive (Sx and Mx positive), both outcome and severity of injury were significantly (P<.05) worse than for groups 1 and 2.

Correlation of Mx and Sx With Severity of Injury, Outcome, Age, CPP, and ICP

Spearman's correlation coefficients of Sx and Mx, averaged over multiple investigations in each patient, with CPP were calculated versus severity of injury (GCS score on admission), outcome, age, mean CPP, and ICP. The detailed results are presented in Table 2 (correlation with clinical outcome and admission GCS score) and Table 3 (correlation with mean CPP, ABP, and ICP). Mx and Sx were significantly correlated with outcome and GCS score on admission.

Averaged values and 95% confidence limits (ANOVA) of Sx, Mx, and PI in patients with favorable and unfavorable outcomes are graphically presented in Fig 4. A greater F value indicates better differentiation of the y variable among different outcome groups. The highest F values are for Sx and Mx. For comparison, F values (ANOVA) of GCS score and PI versus outcome were lower.

Time Profile of Disturbed Autoregulation After Head Injury

The time profile of the calculated indices shows a significant pattern of different autoregulatory status that fluctuates with time (P<.002, ANOVA) (Fig 5). The differences between patients with favorable and unfavorable outcomes were compared. On the first 2 days after injury, a positive Mx (P<.05) was seen in patients with unfavorable outcome. Autoregulation was also abnormal for FVs, resulting in an Sx that did not differ significantly from zero (P>.05). On average, both outcome groups had intact autoregulation during days 3 to 5 after injury (Mx <0, Sx <0; P<.05). Thereafter, patients with an unfavorable outcome showed a significant loss of autoregulation during days 6 to 8 (Mx >0; P<.05).

Discussion

Sx and Mx as Indices of Autoregulation

Do the indices Sx and Mx reflect cerebral autoregulation because they correlate with falling CPP? Mx becomes statistically positive for CPP <55 mm Hg, and Sx becomes statistically different from zero (nonnegative; P>.05) for CPP <40 mm Hg. Moreover, Sx and Mx correlate significantly with outcome and GCS score on admission. A poor outcome and low GCS score are associated with indices that indicate absent autoregulation. This finding supports the importance of autoregulation-oriented therapy. Although “optimal CPP”⁹ is necessary to maintain a safe autoregulation margin and sufficient perfusion, in some cases (n=6) autoregulation was disturbed even when CPP was >70 mm Hg and outcome was unfavorable. This may encourage the clinical use of Sx and Mx to continuously monitor the pressure reactivity reserve, even if CPP appears to be within a safe range.

Time Profile of Indices of Autoregulation

Time-dependent changes in autoregulation indices demonstrated that the first 2 days are critical from the point of view of maintenance of good autoregulatory reserve. Patients with unfavorable outcome had clearly disturbed autoregulation within this period, while patients with good/moderate outcome did not display disturbances in autoregulatory reserve. Therefore, if autoregulation-oriented management is to be accepted, disturbed autoregulation should be managed very intensively soon after admission. Continuous assessment of autoregulation should be used as part of a real-time multimodality monitoring system²⁰ if acute changes in autoregulatory status are to be realized.

Comparison With Experimental Studies

According to experimental studies in which methods to artificially raise ICP or decrease CPP by arterial hemorrhage²¹ were used, when CPP falls, FVs is initially stable, while FVm decreases. When the lower threshold for cerebral autoregulation is reached, as indicated by cortical blood flow measurement, FVs and FVm both fall.²² These experimental results encouraged the separate estimation of the correlation coefficient between CPP and FVs and FVm and their interpretation as indices of autoregulation. From these observations we suggest that when FVs is stable and FVm decreases with decreasing CPP, Sx should be approximately zero, while Mx becomes positive. Below the threshold for exhausted autoregulation, FVs and FVm become pressure passive, and both Mx and Sx become positive.

The negative values of Mx and Sx observed clinically cannot be explained from the aforementioned experiments. One possibility is that with a reactive cerebrovascular bed, vascular resistance overreacts to slow waves in CPP, producing a negative correlation between CPP, FVs, and FVm.

Importance of Analysis of CPP

Various methods of testing cerebral autoregulation have been described.^1112131517 Most are based on analysis of the response in TCD variables when ABP is altered during the test. In head injury, when the cerebral compensatory reserve is severely compromised,²³ any variation of ABP may provoke an acute response in ICP.²²⁴ When autoregulation is intact, changes in CPP may be greater than the change in ABP, since ABP and ICP are changing in opposite directions. When autoregulation is impaired, an increase or decrease in ABP provokes change in ICP in the same direction, with CPP remaining constant. This phenomenon, termed false autoregulation,⁴ may contaminate the results when arterial pressure either increases or decreases during the test. We emphasize the importance of analysis of CPP instead of ABP alone in testing cerebral autoregulation in head injury. Equivalent indices calculated with only ABP correlated with outcome and severity of injury less significantly (r=.23; P<.05) in our study.

Conclusions

The autoregulation indices Mx and Sx, calculated continuously with the use of TCD and CPP waveforms and presented as time trends on the computer screen, indicate the state of cerebral pressure reactivity. Positive values of Mx and Sx signify severely disturbed reactivity, which is predictive of a poor outcome. Cerebral autoregulation is significantly disturbed during the first 2 days after head injury in patients with unfavorable outcome.

Selected Abbreviations and Acronyms

• Download figure
• Download PowerPoint

• Download figure
• Download PowerPoint

• Download figure
• Download PowerPoint

• Download figure
• Download PowerPoint

• Download figure
• Download PowerPoint

Footnotes