Neurocrit Care DOI 10.1007/s12028-015-0125-x
ORIGINAL ARTICLE
Prediction of Delayed Cerebral Ischemia After Subarachnoid Hemorrhage Using Cerebral Blood Flow Velocities and Cerebral Autoregulation Assessment Lionel Calviere • Nathalie Nasr • Catherine Arnaud Marek Czosnyka • Alain Viguier • Bernard Tissot • Jean-Christophe Sol • Vincent Larrue
•
Ó Springer Science+Business Media New York 2015
Abstract Background The risk of delayed cerebral ischemia (DCI) after subarachnoid hemorrhage (SAH) is associated with large cerebral artery vasospasm, but vasospasm is not a strong predictor for DCI. Assessment of cerebral autoregulation with transcranial Doppler (TCD) may improve the prediction of DCI. The aim of this prospective study was to assess the value of TCD-derived variables to be used alone or in combination for prediction of DCI. Methods We included consecutive patients with lowgrade aneurysmal SAH within 4 days of aneurysm rupture. Cerebral autoregulation was evaluated using the moving correlation coefficient Mx calculated from spontaneous fluctuations of cerebral blood flow velocities and arterial blood pressure. Transcranial color-coded sonography was performed to assess large artery vasospasm.
Results Thirty patients (19 women and 11 men; mean age ± SD 44.7 ± 12.1 years) were included. Twenty (66.7 %) patients had vasospasm. DCI occurred in six (20 %) patients after a median delay of 10 days (range 8–13 days). Cerebral autoregulation was impaired at baseline and at day 7 and then returned to normal at day 14. Neither cerebral autoregulation impairment nor large artery vasospasm alone was associated with DCI. In contrast, the combination of large artery vasospasm with worsening impairment of cerebral autoregulation from baseline to day 7 was significantly correlated to subsequent DCI (p = 0.007). Conclusions Early deterioration of cerebral autoregulation was strongly predictive of DCI in patients with large artery vasospasm after low-grade SAH. Our results suggest that consideration to both cerebral blood flow velocities and cerebral autoregulation may improve the prediction of DCI.
L. Calviere N. Nasr A. Viguier V. Larrue Department of Neurology, University Hospital of Toulouse, Toulouse, France
Keywords Cerebral autoregulation Transcranial Doppler Subarachnoid hemorrhage Vasospasm
L. Calviere (&) Neurovascular Unit, Department of Neurology, CHU Purpan, Hopital Pierre Paul Riquet, Place du Dr Baylac, 31059 Toulouse Cedex 9, France e-mail:
[email protected] Introduction
C. Arnaud Department of Biostatistics and Epidemiology, University Hospital of Toulouse, Toulouse, France M. Czosnyka Academic Neurosurgical Unit, Addenbrooke’s Hospital, University of Cambridge, Cambridge, UK B. Tissot J.-C. Sol Department of Neurosurgery, University Hospital of Toulouse, Toulouse, France
In spite of recent therapeutic progress, aneurysmal subarachnoid hemorrhage (SAH) remains a devastating disease. The case fatality rates are close to 40 % and among survivors only 25 % live without disability [1–3]. Once the aneurysm has been managed, delayed cerebral ischemia (DCI) is the leading cause of death and disability in patients who survive the initial SAH [4]. Identifying variables that might help in predicting DCI is clinically important to allow timely implementation of appropriate treatment. It is well established that DCI is more common
123
Neurocrit Care
in patients with vasospasm of large cerebral arteries [5–9]. However, the maximum rate of DCI among patients with large artery vasospasm is only 50 % [10]. Moreover, up to one-third of patients with DCI do not show evidence of large artery vasospasm [11, 12]. A few studies have shown that early impairment of cerebral autoregulation is associated with DCI after SAH [13–17]. Both large artery vasospasm and autoregulation impairment can be non-invasively investigated using transcranial Doppler (TCD) [18–20]. The purpose of this exploratory study was to assess the value of these TCD-derived variables used alone or in combination to predict DCI after low-grade SAH. More particularly, we assessed the value of the time-course changes of cerebral autoregulation and vasospasm for the prediction of DCI.
Methods Patients and Controls Patients aged 18 years or more consecutively treated for low-grade aneurysmal SAH in a tertiary academic center were prospectively included. Diagnosis of SAH was confirmed by computed tomography, magnetic resonance imaging (MRI) or lumbar puncture as recommended [21]. Cerebral aneurysms were identified from cerebral angiography. Low-grade SAH was defined as grade 1–3 on the World Federation of Neurosurgical Societies (WFNS) grading scale. The appearance of SAH on brain imaging was classified using the Fisher’s grading scale. Patients had to be included within 4 days of SAH and after aneurysm coiling, the standard therapeutic procedure in our center. Patients with no readable acoustic temporal bone window or with previous disease that might impair cerebral autoregulation (e.g., carotid stenosis, history of stroke or head injury) were excluded. Nine healthy controls from a previous study [22] served as controls for cerebral autoregulation. Assessment of Vasospasm We used transcranial color-coded sonography (TCCS) for assessment of large artery vasospasm. Bilateral examination of the middle cerebral artery (MCA) and anterior cerebral artery (ACA) was performed at baseline (i.e., within 4 days of initial SAH) and then every other day up to 14 days after SAH or until any detected vasospasm had resolved. Vasospasm was defined as a Lindegaard ratio (i.e., the ratio of MCA mean flow velocity to upper cervical internal carotid mean flow velocity) >3.0 with mean flow velocity >120 cm/ s for the MCA, and mean flow velocity >130 cm/s for the ACA. Severe vasospasm was defined as a Lindegaard ratio C5
123
for the MCA, and mean flow velocity C200 cm/s for the ACA [23, 24]. Assessment of Cerebral Autoregulation Cerebral autoregulation was quantified using the Mx method [25]. Mx measurement was based on spontaneous fluctuations of mean flow velocity in the MCA and mean arterial blood pressure. A 2-MHz probe (Multidop, DWL, Germany) and a rigid head frame (Lam rack, DWL, Germany) were used for unilateral insonation of the MCA. Measurement was performed on the side of aneurysm or, for non-lateralized aneurysm (anterior communicating artery), on the side of maximum flow velocity. Continuous, non-invasive arterial blood pressure recording was achieved using a servo-controlled finger plethysmograph (Finapres, Ohmeda, Netherlands) with the patient left hand positioned at heart level. The reliability of Mx calculation using non-invasive arterial blood pressure monitoring has been previously studied. Mx calculation from invasive and non-invasive arterial blood pressure correlated well [26]. Mean arterial blood pressure and mean flow velocity recordings were acquired simultaneously and continuously over 20–30 min epochs. Analog outputs from the pressure monitor and the TCD unit (maximal frequency outline) were connected to an analogic-to-digital convertor fitted into a computer supporting the Biosan software (Biological Signals Analyser) version 2.2 developed by P. Smielewski and M. Czonyka for data collection and assessment of DCA [25, 27]. Mx is a moving correlation coefficient between mean flow velocity and mean arterial blood pressure. Mean arterial blood pressures and flow velocities were averaged over 5-s intervals to minimize the effect of pulse and respiratory waves. Thirty-six 5-s periods were used to calculate the Pearson’s correlation coefficient between mean arterial blood pressure and mean flow velocity over 3-min periods of the recording time series. Mx is the mean of the Pearson’s correlation coefficients calculated from these consecutive samples. Mx close to 0 or negative corresponds to preserved cerebral autoregulation. Mx close to 1 indicates that fluctuations in FV depend on variations of arterial blood pressure, meaning that cerebral autoregulation is impaired. Mx was measured at baseline, i.e., within 4 days of initial SAH (Mx 1), at day 7 ± 2 (Mx 2), and at day 14 ± 2 (Mx 3). Delayed Cerebral Ischemia DCI was defined as neurological deterioration not attributable to another cause than ischemia. Infection, hydrocephalus, epileptic seizures, and metabolic disorders were excluded using brain imaging, blood studies, CSF
Neurocrit Care
analysis, or electroencephalography as appropriate. DCI was confirmed by CT or MR imaging. Treatment including the routine use of nimodipine, pravastatin 40 mg/day, and isotonic intravenous hydration with sodium chloride was applied in all patients. We did not use triple-H therapy. Endovascular treatment of severe large artery vasospasm using angioplasty or intra-arterial nimodipine infusion was performed at the discretion of treating physician for each patient in the case of ischemic manifestations. Statistical Analysis We used non-parametric unpaired Mann–Whitney U test and Fisher’s exact test to compare the groups defined by the presence of DCI. We used repeated measures ANOVA to determine how Mx was affected by time of measurement and subsequent occurrence of DCI. The performance of selected TCD variables for DCI prediction was evaluated using predictive values, sensitivity, specificity, and likelihood ratios. We used logistic regression analysis to adjust for the WFNS grade. A value of p < 0.05 was considered statistically significant. The Bonferroni’s correction was applied for multiple comparisons. As we tested six associations between TCD-derived variables and DCI, significance after Bonferroni’s correction was set at p < 0.008. Statistical analysis was performed with SPSS (IBM). Ethics
165.6 ± 30 cm/s).The median delay from SAH to vasospasm was 6 days (range 3–11 days). Vasospasm was severe in 5 patients. Cerebral Autoregulation The median time between the occurrence of SAH and the baseline measurement of Mx (Mx 1) was 3 days (range 2– 4 days). Mx 1 (0.43 ± 0.2) and Mx 2 (0.46 ± 0.2) were significantly higher than Mx in controls (0.23 ± 0.1) (p = 0.003 and p < 0.0001, respectively). Mx 3 (0.37 ± 0.2) was not different from Mx in controls. There was no significant difference between Mx 1 and Mx 2. In contrast, Mx 3 was significantly lower than Mx 2 (p = 0.008) (Fig. 1). Mx 1 and Mx 2 were unrelated to large artery vasospasm (p = 0.55 and p = 0.5, respectively). DCI DCI occurred in six patients (20 %) with a median delay of 10 days after SAH (range 8–13 days). Characteristics of patients with or without DCI are summarized in Table 1. DCI was associated with a higher WFNS grade (p = 0.007). All patients with DCI had previously documented large artery vasospasm. Fourteen of the twenty-four patients without DCI also had large artery vasospasm. The association of DCI with vasospasm was not significant (p = 0.07); Mx 1 and Mx 2 (both measured before
The study was approved by our Institutional Review Board and our Ethics Committee. All patients gave written informed consent.
Results Thirty patients (19 women and 11 men; mean age ± SD 44.7 ± 12.1 years) were included. The aneurysms were treated with coiling in all patients. Three patients were not included due to no readable acoustic temporal bone window or previous disease that might impair cerebral autoregulation. The median value (range) of the WFNS grade and Fisher’s grade were 1 (1–3) and 3 (1–4), respectively. The median delay (range) to inclusion was 3 days (2–4 days). The control group for cerebral autoregulation included 5 men and 4 women with a mean age ± SD of 49.3 ± 5.3 years. Large Artery Vasospasm Twenty patients (66.6 %) had large artery vasospasm (mean Lindegaard ratio, 4.0 ± 0.9; mean flow velocity,
Fig. 1 Comparison of Mx between patients and controls at different times of measurement. Mx is a moving correlation coefficient between mean flow velocity and mean arterial blood pressure. Mx close to 0 or negative corresponds to preserved cerebral autoregulation, whereas Mx close to 1 indicates cerebral autoregulation impairment. The definition and calculation modalities are detailed in the ‘‘Methods’’ section. The x-axis of the graph represents the different times of Mx measurement and the y-axis the Mx values
123
Neurocrit Care
DCI occurrence) were not significantly higher in patients with DCI. Further analysis of the variations of Mx over time showed a non-significant increase of Mx from baseline to day 7 in patients with DCI (p for interaction = 0.38; Fig. 2). The combination of large artery vasospasm with early worsening of autoregulation defined as Mx 2 – Mx 1 > 0.1 was strongly correlated with DCI (p = 0.007; Table 1). This correlation remained significant after Bonferroni’s correction and in a logistic model adjusting for WFNS grade (p = 0.025). The sensitivity, specificity, predictive values, and likelihood ratios of this combined variable are shown in Table 2. The positive likelihood ratio (95 % confidence interval) was of 8.00 (1.9–33.9) and the negative likelihood ratio of 0.36 (0.1–1.1).
Discussion Our results showed an early impairment of cerebral autoregulation after low-grade aneurysmal SAH. Impairment of autoregulation alone was not correlated with subsequent DCI. However, early worsening of autoregulation on repeat examination was predictive of DCI in patients with associated large artery vasospasm. Several previous studies using various methods for autoregulation assessment have demonstrated that cerebral autoregulation is impaired in the first day after SAH [13– 17, 28–30]. Our study confirmed this finding in patients with low-grade SAH. Impaired autoregulation may reflect the fact that the perfusion pressure downstream of large artery vasospasm is so low that small vessels are already maximally dilated. Alternatively, autoregulation may be impaired because the capacity of small vessels to dilate is reduced independently of perfusion pressure. Several potential mechanisms including microembolization and vasospasm in the microcirculation may explain such a reduction in patients with SAH [11, 31–33]. In our study, autoregulation impairment was already present at baseline and was unrelated to later large artery vasospasm, suggesting that dysfunction of microcirculation was the primary cause of impaired autoregulation. Early impairment of autoregulation alone was not associated with DCI in our study. This finding is at variance with those in two recent studies which reported a significant association of early failure of autoregulation with subsequent DCI [13, 34]. Several reasons can be considered to explain this discrepancy. First, the relatively small sample size in our study may have resulted in a lack of statistical power. Also we only included patients with lowgrade SAH. The association of impaired autoregulation with DCI may possibly be weaker in these patients than in patients with more severe bleeding and intracranial hypertension. Finally, comparisons between studies are
123
Fig. 2 Differential evolution of Mx (baseline: Mx 1, day 7: Mx 2, and day 14: Mx 3) according to delayed cerebral ischemia (DCI). Mx is a moving correlation coefficient between mean flow velocity and mean arterial blood pressure. Mx close to 0 or negative corresponds to preserved cerebral autoregulation, whereas Mx close to 1 indicates cerebral autoregulation impairment. The definition and calculation modalities are detailed in the ‘‘Methods’’ section. The x-axis of the graph represents the different times of Mx measurement and the yaxis the Mx values. The increase of Mx at day 7 in patients with DCI was not significant (p for interaction = 0.38)
Table 1 Characteristics of patients with or without DCI DCI (n = 6)
No DCI (n = 24)
p value
Age (SD), years
50.8 (13.6)
43.3 (11.5)
0.2
Female
2
17
0.16
WFNS grade, median (range) Fisher’s grade, median (range)
2 (1–3)
1 (1–3)
0.007
3 (1–4)
4 (1–4)
0.2
Large artery vasospasm
6
14
0.07
Severe large artery vasospasm
1
4
1
Mx 1 (SD)
0.37 (0.24)
0.44 (0.19)
0.55
Mx 2 (SD)
0.50 (0.09)
0.45 (0.15)
0.5
Mx 2 - Mx 1 > 0.1
4
6
0.1
Large artery vasospasm and Mx 2 - Mx 1 > 0.1
4
2
0.007
difficult because of different methodologies for autoregulation assessment. We used a moving linear correlation coefficient calculated from mean flow velocity, whereas methods in previous studies included a moving linear correlation coefficient calculated from systolic flow velocity or near-infrared spectroscopy, or transfer function analysis [13, 34]. It is possible that the strength of the association of
Neurocrit Care Table 2 Accuracy measures of combined variable for predicting DCI (large artery vasospasm plus Mx 2 - Mx1 > 0.1) Sensitivity %
66.7 (22.7–94.7)
Specificity %
91.7 (72.9–98.7)
Positive predictive value %
66.7 (22.7–94.7)
Negative predictive value %
91.7 (72.9–98.7)
Positive likelihood ratio
8 (1.9–33.9)
Negative likelihood ratio
0.36 (0.1–1.1)
Values in parentheses are 95 % confidence intervals
autoregulation failure with DCI varies with the methodology used for autoregulation assessment [35–37]. Several previous studies using TCD have demonstrated an association of large artery vasospasm with DCI [5, 7–9]. However, the predictive value of TCD is only moderate both because of the reasons that some patients with DCI do not have large artery vasospasm and other patients with large artery vasospasm do not develop DCI [10–12]. In our study, all patients with DCI had previously documented large artery vasospasm. This association was not statistically significant because of a small number of patients with DCI, which reduced the statistical power. DCI was strongly correlated with early worsening of cerebral autoregulation in patients with large artery vasospasm. Constructing a variable that encompasses autoregulation failure and large artery vasospasm is pertinent for the prediction of DCI because such a variable measures hemodynamic compromise in large cerebral arteries and at the level of microcirculation. The positive and negative likelihood ratios were 8 and 0.36, respectively, indicating that this variable could be clinically useful to help discriminate patients at high/low risk of DCI. There are some limitations to our study. The lack of oneto-one association of DCI with large artery vasospasm and autoregulation is at variance with previous reports and may have resulted from insufficient power of our study. Post hoc calculation of power based on our findings and using a two-sided alpha of 0.05 showed a power of only 43.68 % for Mx 2 - Mx 1 > 0, and 15.05 % for large artery vasospasm. Accordingly, confirmation studies would need a sample size of 75 and 45 patients, respectively. Nevertheless, our findings suggest that the combination of large artery vasospasm with early deterioration of autoregulation is more strongly associated with DCI than either variable in isolation. As our study was exploratory, we tested a lot of associations, which might have resulted in spurious findings. However, the correlation of worsening autoregulation with DCI in patients with large artery vasospasm was strongly significant, remained significant after Bonferroni’s correction for multiple comparisons, and is biologically plausible. Another limitation is that we performed only two measures of cerebral autoregulation during the first 7 days
after SAH. Daily examination could improve the accuracy of our methods. Non-invasive arterial blood pressure for Mx calculation was used in this study, whereas cerebral autoregulation using Mx was initially described with invasive arterial blood pressure [25]. Agreement between invasive and noninvasive assessment of Mx using Finapres was tested by Lavinio et al. [26].The two methods showed good correlation. Also Finapres replicated well the slow waves of arterial blood pressure that are used for Mx calculation with no significant dampening or amplification of the signal. Our assessment of cerebral autoregulation was based on measurement of flow velocities in only one MCA. Normal findings could not exclude impairment of autoregulation in other cerebral arterial territories. Thus, in view of treatment adaptation, it is possible that some patients could have been misclassified as not at risk for DCI. The main strength of our study is its prospective design. The selective inclusion of low-grade SAH can also be viewed as strength because recognition of neurological deterioration may be difficult in patients who are unconscious. We also used TCCS which is more accurate than TCD to ascertain large artery vasospasm [24, 38, 39]. The diagnosis of DCI was confirmed on brain imaging in all patients. In conclusion, we found that early worsening of cerebral autoregulation in patients with large artery vasospasm is strongly predictive of DCI in low-grade aneurysmal SAH. Our results suggest that consideration to both cerebral blood flow velocities and cerebral autoregulation may improve the prediction of DCI. Acknowledgments This work was supported by a grant from ‘‘la Direction de la Recherche Clinique et de l’Innovation (DRCI)’’ at Toulouse University Hospital, Toulouse, France. Portions of this work were presented in oral communication form at the National Neurosurgery Conference, Toulouse, France, 2012 and at the International Stroke Conference, Hawaı¨, Honolulu, 2013. Conflict of interest The authors report no conflict of interest concerning the materials or methods used in this study or the findings specified in this paper.
References 1. Hop JW, Rinkel GJ, Algra A, van Gijn J. Case-fatality rates and functional outcome after subarachnoid hemorrhage: a systematic review. Stroke. 1997;28(3):660–4. 2. Koffijberg H, Buskens E, Granath F, et al. Subarachnoid haemorrhage in Sweden 1987–2002: regional incidence and case fatality rates. J Neurol Neurosurg Psychiatry. 2008;79(3):294–9. 3. Qureshi AI, Suri MF, Nasar A, et al. Trends in hospitalization and mortality for subarachnoid hemorrhage and unruptured aneurysms in the United States. Neurosurgery. 2005;57(1):1–8 discussion 1-8.
123
Neurocrit Care 4. van Gijn J, Kerr RS, Rinkel GJ. Subarachnoid haemorrhage. Lancet. 2007;369(9558):306–18. 5. Fisher CM, Roberson GH, Ojemann RG. Cerebral vasospasm with ruptured saccular aneurysm–the clinical manifestations. Neurosurgery. 1977;1(3):245–8. 6. Seiler RW, Grolimund P, Aaslid R, Huber P, Nornes H. Cerebral vasospasm evaluated by transcranial ultrasound correlated with clinical grade and CT-visualized subarachnoid hemorrhage. J Neurosurg. 1986;64(4):594–600. 7. Grosset DG, McDonald I, Cockburn M, Straiton J, Bullock RR. Prediction of delayed neurological deficit after subarachnoid haemorrhage: a CT blood load and Doppler velocity approach. Neuroradiology. 1994;36(6):418–21. 8. Grosset DG, Straiton J, du Trevou M, Bullock R. Prediction of symptomatic vasospasm after subarachnoid hemorrhage by rapidly increasing transcranial Doppler velocity and cerebral blood flow changes. Stroke. 1992;23(5):674–9. 9. de Rooij NK, Rinkel GJ, Dankbaar JW, Frijns CJ. Delayed cerebral ischemia after subarachnoid hemorrhage: a systematic review of clinical, laboratory, and radiological predictors. Stroke. 2013;44(1):43–54. 10. Kassell NF, Sasaki T, Colohan AR, Nazar G. Cerebral vasospasm following aneurysmal subarachnoid hemorrhage. Stroke. 1985;16(4):562–72. 11. Ohkuma H, Manabe H, Tanaka M, Suzuki S. Impact of cerebral microcirculatory changes on cerebral blood flow during cerebral vasospasm after aneurysmal subarachnoid hemorrhage. Stroke. 2000;31(7):1621–7. 12. Rabinstein AA, Friedman JA, Weigand SD, et al. Predictors of cerebral infarction in aneurysmal subarachnoid hemorrhage. Stroke. 2004;35(8):1862–6. 13. Budohoski KP, Czosnyka M, Smielewski P, et al. Impairment of cerebral autoregulation predicts delayed cerebral ischemia after subarachnoid hemorrhage: a prospective observational study. Stroke. 2012;43(12):3230–7. 14. Pickard JD, Matheson M, Patterson J, Wyper D. Prediction of late ischemic complications after cerebral aneurysm surgery by the intraoperative measurement of cerebral blood flow. J Neurosurg. 1980;53(3):305–8. 15. Lam JM, Smielewski P, Czosnyka M, Pickard JD, Kirkpatrick PJ. Predicting delayed ischemic deficits after aneurysmal subarachnoid hemorrhage using a transient hyperemic response test of cerebral autoregulation. Neurosurgery. 2000;47(4):819–25 discussions 25-6. 16. Ratsep T, Asser T. Cerebral hemodynamic impairment after aneurysmal subarachnoid hemorrhage as evaluated using transcranial doppler ultrasonography: relationship to delayed cerebral ischemia and clinical outcome. J Neurosurg. 2001;95(3):393–401. 17. Jaeger M, Schuhmann MU, Soehle M, Nagel C, Meixensberger J. Continuous monitoring of cerebrovascular autoregulation after subarachnoid hemorrhage by brain tissue oxygen pressure reactivity and its relation to delayed cerebral infarction. Stroke. 2007;38(3):981–6. 18. Aaslid R, Lindegaard KF, Sorteberg W, Nornes H. Cerebral autoregulation dynamics in humans. Stroke. 1989;20(1):45–52. 19. Lysakowski C, Walder B, Costanza MC, Tramer MR. Transcranial Doppler versus angiography in patients with vasospasm due to a ruptured cerebral aneurysm: a systematic review. Stroke. 2001;32(10):2292–8. 20. Soehle M, Czosnyka M, Pickard JD, Kirkpatrick PJ. Continuous assessment of cerebral autoregulation in subarachnoid hemorrhage. Anesth Analg. 2004;98(4):1133–9 table of contents. 21. Steiner T, Juvela S, Unterberg A, Jung C, Forsting M, Rinkel G. European Stroke Organization guidelines for the management of intracranial aneurysms and subarachnoid haemorrhage. Cerebrovasc Dis. 2013;35(2):93–112.
123
22. Nasr N, Traon AP, Czosnyka M, Tiberge M, Schmidt E, Larrue V. Cerebral autoregulation in patients with obstructive sleep apnea syndrome during wakefulness. Eur J Neurol. 2009;16(3):386– 91. 23. Lindegaard KF. The role of transcranial Doppler in the management of patients with subarachnoid haemorrhage–a review. Acta Neurochir Suppl. 1999;72:59–71. 24. Krejza J, Kochanowicz J, Mariak Z, Lewko J, Melhem ER. Middle cerebral artery spasm after subarachnoid hemorrhage: detection with transcranial color-coded duplex US. Radiology. 2005;236(2):621–9. 25. Czosnyka M, Smielewski P, Kirkpatrick P, Menon DK, Pickard JD. Monitoring of cerebral autoregulation in head-injured patients. Stroke. 1996;27(10):1829–34. 26. Lavinio A, Schmidt EA, Haubrich C, Smielewski P, Pickard JD, Czosnyka M. Noninvasive evaluation of dynamic cerebrovascular autoregulation using Finapres plethysmograph and transcranial Doppler. Stroke. 2007;38(2):402–4. 27. Smielewski P, Czosnyka M, Steiner L, Belestri M, Piechnik S, Pickard JD. ICM+: software for on-line analysis of bedside monitoring data after severe head trauma. Acta Neurochir Suppl. 2005;95:43–9. 28. Jaeger M, Soehle M, Schuhmann MU, Meixensberger J. Clinical significance of impaired cerebrovascular autoregulation after severe aneurysmal subarachnoid hemorrhage. Stroke. 2011;43(8): 2097–101. 29. Dernbach PD, Little JR, Jones SC, Ebrahim ZY. Altered cerebral autoregulation and CO2 reactivity after aneurysmal subarachnoid hemorrhage. Neurosurgery. 1988;22(5):822–6. 30. Yundt KD, Grubb RL Jr, Diringer MN, Powers WJ. Autoregulatory vasodilation of parenchymal vessels is impaired during cerebral vasospasm. J Cereb Blood Flow Metab. 1998; 18(4):419–24. 31. Stein SC, Browne KD, Chen XH, Smith DH, Graham DI. Thromboembolism and delayed cerebral ischemia after subarachnoid hemorrhage: an autopsy study. Neurosurgery. 2006; 59(4):781–7 discussion 7-8. 32. Romano JG, Forteza AM, Concha M, et al. Detection of microemboli by transcranial Doppler ultrasonography in aneurysmal subarachnoid hemorrhage. Neurosurgery. 2002;50(5):1026–30 discussion 30-1. 33. Sun BL, Zheng CB, Yang MF, Yuan H, Zhang SM, Wang LX. Dynamic alterations of cerebral pial microcirculation during experimental subarachnoid hemorrhage. Cell Mol Neurobiol. 2009;29(2):235–41. 34. Otite F, Mink S, Tan CO, et al. Impaired cerebral autoregulation is associated with vasospasm and delayed cerebral ischemia in subarachnoid hemorrhage. Stroke. 2014;45(3):677–82. 35. Zweifel C, Castellani G, Czosnyka M, et al. Continuous assessment of cerebral autoregulation with near-infrared spectroscopy in adults after subarachnoid hemorrhage. Stroke. 2010;41(9): 1963–8. 36. Budohoski KP, Czosnyka M, Smielewski P, et al. Cerebral autoregulation after subarachnoid hemorrhage: comparison of three methods. J Cereb Blood Flow Metab. 2013;33(3):449–56. 37. Budohoski KP, Reinhard M, Aries MJ, et al. Monitoring cerebral autoregulation after head injury. Which component of transcranial Doppler flow velocity is optimal? Neurocrit Care. 2012; 17(2):211–8. 38. Proust F, Callonec F, Clavier E, et al. Usefulness of transcranial color-coded sonography in the diagnosis of cerebral vasospasm. Stroke. 1999;30(5):1091–8. 39. Mariak Z, Krejza J, Swiercz M, Kordecki K, Lewko J. Accuracy of transcranial color Doppler ultrasonography in the diagnosis of middle cerebral artery spasm determined by receiver operating characteristic analysis. J Neurosurg. 2002;96(2):323–30.