The management of severe traumatic brain injury

J Neurosurg 120:1451 1457, 2014 AANS, 2014 Pressure autoregulation monitoring and cerebral perfusion pressure target recommendation in patients with severe traumatic brain injury based on minute-by-minute monitoring data Clinical article *Bart Depreitere, M.D., Ph.D., 1 Fabian Güiza, Ph.D., 2 Greet Van den Berghe, M.D., Ph.D., 2 Martin U. Schuhmann, M.D., Ph.D., 3 Gottlieb Maier, M.D., 3 Ian Piper, Ph.D., 4 and Geert Meyfroidt, M.D., Ph.D. 2 1 Neurosurgery and 2 Intensive Care Medicine, University Hospitals Leuven, Belgium; 3 Klinik für Neurochirurgie, Universitätsklinikum Tübingen, Germany; and 4 Clinical Physics, Southern General Hospital, Glasgow, United Kingdom Object. In severe traumatic brain injury, a universal target for cerebral perfusion pressure (CPP) has been abandoned. Attempts to identify a dynamic CPP target based on the patient s cerebrovascular autoregulatory capacity have been promising so far. Bedside monitoring of pressure autoregulatory capacity has become possible by a number of methods, Czosnyka s pressure reactivity index (PRx) being the most frequently used. The PRx is calculated as the moving correlation coefficient between 40 consecutive 5-second averages of intracranial pressure (ICP) and mean arterial blood pressure (MABP) values. Plotting PRx against CPP produces a U-shaped curve in roughly two-thirds of monitoring time, with the bottom of this curve representing a CPP range corresponding with optimal autoregulatory capacity (CPPopt). In retrospective series, keeping CPP close to CPPopt corresponded with better outcomes. Monitoring of PRx requires high-frequency signal processing. The aim of the present study is to investigate how the processing of the information on cerebrovascular pressure reactivity that can be obtained from routine minute-by-minute ICP and MABP data can be enhanced to enable CPPopt recommendations that do not differ from those obtained by the PRx method, show the same associations with outcome, and can be generated in more than two-thirds of monitoring time. Methods. The low-frequency autoregulation index (LAx) was defined as the moving minute-by-minute ICP/MABP correlation coefficient calculated over time intervals varying from 3 to 120 minutes. The CPPopt calculation was based on LAx-CPP plots and done for time windows between 1 and 24 hours and for each LAx type. The resulting matrix of CPPopts were then averaged in a weighted manner, with the weight based on the goodness of fit of a U-shape and the lower value of the LAx corresponding to the U-bottom, to result in a final CPPopt recommendation. The association between actual CPP/CPPopt and outcome was assessed in the multicenter Brain Monitoring with Information Technology Research Group (BrainIT) database (n = 180). In the Leuven-Tübingen database (60-Hz waveform data, n = 21), LAx- and PRx-based CPPopts were compared. Results. In the BrainIT database, CPPopt recommendations were generated in 95% of monitoring time. Actual CPP being close to LAx-based CPPopt was associated with increased survival. In a multivariate model using the Corticosteroid Randomization After Significant Head Injury (CRASH) model as covariates, the average absolute difference between actual CPP and CPPopt was independently associated with increased mortality. In the high-frequency data set no significant difference was observed between PRx-based and LAx-based CPPopts. The new method issued a CPPopt recommendation in 97% of monitoring time, as opposed to 44% for PRx-based CPPopt. Conclusions. Minute-by-minute ICP/MABP data contain relevant information for autoregulation monitoring. In this study, the authors new method based on minute-by-minute data resolution allowed for CPPopt calculation in nearly the entire monitoring time. This will facilitate the use of pressure reactivity monitoring in all ICUs. (http://thejns.org/doi/abs/10.3171/2014.3.jns131500) Key Words traumatic brain injury cerebral perfusion pressure autoregulation intracranial pressure Abbreviations used in this paper: BrainIT = Brain Monitoring with Information Technology Research Group; CBF = cerebral blood flow; CPP = cerebral perfusion pressure; CPPopt = optimal CPP; CRASH = Corticosteroid Randomization After Significant Head Injury; DATACAR = Dynamic Adaptive Target of Active Cerebral Autoregulation; GOS = Glasgow Outcome Scale; ICP = intracranial pressure; IQR = interquartile range; LAx = lowresolution autoregulation index; L-PRx = low-frequency pressure reactivity index; MABP = mean arterial blood pressure; Mx = mean arterial Doppler flow velocity based autoregulation index; PRx = pressure reactivity index; RMSE = root mean squared error; TBI = traumatic brain injury. * Drs. Depreitere and Güiza contributed equally to this work. The management of severe traumatic brain injury (TBI) is primarily aimed at avoiding secondary brain damage, which mainly manifests itself as brain ischemia. Monitoring of intracranial pressure (ICP) has been a cornerstone in achieving this aim because of its imminent relation with intracranial volume in the Monro-Kellie doctrine 12,15 and its inherent relation with cerebral perfusion. It was widely agreed that secondary injury prevention should be targeted at keeping ICP below 20 25 mm Hg. 3 Recently, however, a large randomized controlled trial could not demonstrate a benefit of monitoring ICP to guide intracranial hypertension management. 4 The cerebral perfusion 1451

B. Depreitere et al. pressure (CPP) is the pressure gradient for cerebral blood flow to the brain, expressed as the difference between the mean arterial blood pressure (MABP) and the ICP. Based on experimental and clinical work by Rosner and Daughton, 20 the original recommendation was to target CPP at 70 mm Hg or higher. A randomized trial comparing CPP-targeted therapy with ICP-targeted therapy only 19 could not demonstrate a benefit for this strategy. Moreover, low CPP protocols, first introduced in Lund, Sweden, did not seem to negatively affect outcome. 7 As a consequence, the CPP target guideline of 70 mm Hg and later 60 mm Hg in TBI was abandoned in the 2007 revision of the Brain Trauma Foundation guidelines for the management of severe TBI. 3 In recent years, a more dynamic patient-tailored CPP target, based on the autoregulation capacity of the patient s cerebral vasculature, has been proposed. Cerebrovascular pressure autoregulation is the capacity of the cerebral vasculature to maintain a constant cerebral blood flow (CBF) through varying CPP by adapting the basal tone of the arteriolar smooth muscles. It is well known that autoregulation is often deficient in severe TBI, albeit that the degree and range of this dysfunction can vary among patients, and in time within the same patient. 24 Howells et al. were the first to successfully provide proof of concept for the idea of autoregulation-based management by demonstrating a relation between the ICP/MABP correlation coefficient and outcome in 2 distinct retrospective patient cohorts in 2 centers with different treatment protocols. 9 However, an adaptive CPP target based on the autoregulatory status of the patient has not been tested prospectively. To investigate this as a strategy, a continuous measure for autoregulation is needed. Czosnyka et al. have pioneered such a tool by developing the pressure reactivity index (PRx) as a continuous measure for cerebrovascular autoregulation, by calculating ICP changes in reaction to slow wave variability (30 200 seconds) of the blood pressure. 5 The PRx is calculated as the moving correlation coefficient between 40 consecutive 5-second averages of ICP and MABP based on signal capture at a frequency of at least 60 Hz. PRx was shown to correlate with outcome. 5,27 Moreover, PRx appeared to be affected by CPP in such a way that plotting PRx against CPP produced a U-shaped curve in about 60% of patients, enabling to define the CPP range correlating with optimal autoregulatory capacity as the CPP range corresponding to the bottom of the U-shaped curve (that is, CPPopt). 23,27 It was demonstrated in retrospective studies that outcome was better in patients with a mean actual CPP close to the mean CPPopt. 2,27 Because PRx calculation requires additional software solutions to extract continuous waveform data from the bedside monitor, thus far PRx monitoring has been limited to research-minded academic centers. Other measures of autoregulation, based on transcranial Doppler flow velocity measurement, 6 laser Doppler flowmetry, 13 or brain tissue oxygenation (PbtO 2 ) monitoring 10 require additional expertise and/or the placement of additional expensive probes. Moreover, none of the methods proposed so far is able to issue CPP recommendations during the entire critical phase after TBI. The present study investigates how the processing of the information on cerebrovascular pressure reactivity that can be obtained from routine minute-by-minute ICP and MABP data can be enhanced to enable CPPopt recommendations in as much monitoring time as possible and to validate such method against the PRx method. The eventual goal is to provide neurosurgeons and neurointensivists with a practical software tool that should allow them to calculate pressure reactivity and optimal CPP at the bedside based on routine data capture, and, hence, enable autoregulation-steered CPP management without investment in additional equipment. Methods Data The Brain Monitoring with Information Technology Research Group (BrainIT) database 17,22 collected information from 264 consecutive TBI patients with ICP monitoring admitted to 22 neuro-icus in 11 European countries between March 2003 and July 2005. Among others, baseline risk factors, minute-by-minute ICP and MABP monitoring data, and Glasgow Outcome Scale (GOS) score at 6 months were registered. Complete data on ICP and MABP measurements for the first 48 hours of the ICU stay and GOS score at 6 months were available in 180 patients. The Multi-Centre Research Ethics Committee for Scotland approved the use of these data for scientific purposes on February 14, 2002. All BrainIT centers obtained local ethics committee approval, and the need for informed consent was waived. Artifacts in the data were removed during the validation phase of this database. Moreover, signals were independently and manually checked again by 2 clinicians in Leuven (first and last authors), and artifacts and questionable data were removed. The second data set comprised 21 patients who were admitted to the ICUs of the University Hospitals Leuven, Belgium, between September 2010 and March 2012 and Universitätsklinikum Tübingen, Germany, between February and December 2009. Both centers routinely monitor and collect data from their TBI patients, including 60-Hz waveform ICP and MABP data (for calculation of PRx and PRx-based CPPopt via ICM+ software, Cambridge Enterprises). Data collection also includes demographics and injury data, including baseline risk factors, and 6-month GOS scores. All data referring to the identity of the patients were removed, and ethics committee approval was obtained in Leuven as well as in Tübingen to use these data for retrospective data analysis. Artifacts were removed by 2 Leuven clinicians (first and last authors) who independently and manually checked all signals. Both data sets are described in detail in Table 1. LAx and DATACAR The low-resolution autoregulation index (LAx) for time interval x at a given time t(0) was calculated as the correlation coefficient between minute-by-minute measurements of ICP and MABP for the time interval [t(- x+1), t(0)]. For the present study, LAx values were calculated for the time intervals 3, 5, 10, 20, 30, 60, 90, and 120 minutes. 1452

CPP target recommendation in severe traumatic brain injury TABLE 1: Patient demographic information* Parameter BrainIT (n = 180) For continuous CPPopt calculation we applied Steiner s 23 and Aries 2 method of plotting CPP and PRx, and fitting a U-shaped curve, with the most negative values of the autoregulation index indicating a 5 mm Hg range of optimal CPP. However, instead of limiting this to the previous 4 hours of data as in the method of Aries, we applied this method on time windows of 1, 2, 4, 6, 8, 12, and 24 hours and did this for each LAx defined above. Hence, for each t(0), 45 plots were generated. These plots were given a weight based on 2 criteria: the better a U-shaped curve could be fitted and the lower the LAx-value corresponding to the plot-specific CPPopt, the higher the weight. The final resulting CPPopt of our method was computed as the weighted average of the plot-specific CPPopts (Fig. 1). This final CPPopt was then rounded and presented as the middle of a 5 mm Hg recommended CPP range. The rationale of this more elaborate method of CPPopt calculation was to optimize the capacity of detection of glimpses of active autoregulation in the registered monitoring data over the past period (with a maximum retrospective scope of 24 hours). We named this method the dynamic adaptive target of active cerebral autoregulation (DATACAR), in which the term dynamic refers to the use of time windows of varying lengths to determine the CPP zone where the most active autoregulation is present. Analyses Based on the BrainIT Database Based on the DATACAR method, a recommended CPPopt range was calculated for each minute of the first Leuven-Tübingen (n = 21) median age in yrs (IQR) 33 (21 51) 49 (31 66) % male sex 8 61.9 mechanism of trauma: % road traffic accident pedestrian assault work sport fall unknown/missing 37.8 8.9 8.9 2.8 3.3 31.7 6.6 23.8 66.7 9.5 GCS (total/15) on admission median (IQR) 7 (5 10) 7 (3 11) % missing 6.7 9.5 pupil reactivity at the accident: % both reacting 71.1 85.7 1 reacting 7.8 4.8 neither reacting 12.2 4.8 untestable/missing 8.9 4.8 presence of extracranial injury: % 58.3 33.3 median GOS (IQR) 4 (3 5) 4 (3 5) % mortality 17.0 19.0 median LOS in days (IQR) 13 (7 23) 23 (12 27) * GCS = Glasgow Coma Scale; LOS = length of stay. 48 hours of monitoring. We concentrated on the first 48 hours after admission to the ICU for 2 reasons. First, the largest differences in outcome in relationship with the autoregulatory status of the patients can be observed in the first 2 days of monitoring. 5,24 Second, we wanted to avoid bias introduced by differences in monitoring time duration. We calculated the percentage of time during which a CPPopt recommendation could be given and studied the contribution of the different time windows in the CPPopt calculation. The accordance between CPPopt and actual CPP and its association with survival was calculated in 6 ways: 1) the percentage of time for which the actual CPP was within the recommended CPPopt range; 2) the average absolute difference between actual CPP and CPPopt; 3) the average absolute difference when actual CPP was outside the CPPopt range; 4) the average difference where actual CPP was below CPPopt; 5) the average difference where actual CPP was above CPPopt; and 6) the average absolute difference between actual CPP and CPPopt was used in a multivariate logistic regression model together with the CRASH basic model risk factors age, GCS, pupillary reaction, and presence of extracranial injury. 16 Analyses Based on the Leuven-Tübingen Database The multicenter Leuven-Tübingen data set was used to compare the DATACAR-based CPPopt recommendations with the high-resolution PRx-derived CPPopt recommendations. First, the absolute values of DATACAR- and PRxbased CPPopt recommendations were compared. Second, the proportion of time a CPPopt recommendation could be calculated was evaluated for both methods. Third, the association of the accordance between actual CPP and CPPopt and mortality was assessed (univariately). Analyses were performed for the first 48 hours of monitoring. Data Analysis and Statistics The building of the indices (LAx variants and PRx), calculation of CPPopt for both methods, and the univariate and multivariate statistics were done using Matlab (version B2011b, MathWorks). Student t-tests were used to compare recommended calculated CPPopt values calculated by both the DATACAR and the PRx method. Wilcoxon ranking tests were used to assess the associations between actual CPP/CPPopt relations and survival. As for assessing the multivariate relation between survival on the one hand and the actual CPP/CPPopt relations and CRASH variables on the other hand, multivariate logistic regression was used. Results BrainIT Database Analyses The DATACAR method was able to produce a CPPopt recommendation in 95% (90% 97%) of monitoring time (median and IQR) in the first 48 hours. The 1-hour time window contributed to 18.7% of CPPopt recommendations and the 2-, 4-, 6-, 8-, 12-, and 24-hour windows to 20.4%, 19.1%, 14.7%, 11.9%, 9.7%, and 5.5%, respectively. 1453

B. Depreitere et al. Fig. 1. Example demonstrating the DATACAR method. The CPPopt is computed for Minute 1670 of the ICU stay. U-curves could only be fitted for time windows of 2, 4, 6, 8, and 12 hours. The CPPopts receive a weighting factor based on the goodness of fit of their U-shaped curve and the lower value of the LAx at CPPopt. In this figure, the weighting factor is expressed by the thickness of the line connecting to each U-curve plot. The curves derived from the 4-hour time window and LAx-20, LAx-30, LAx-60, and LAx-90 had the best fit and reached lower U-bottom values and were therefore assigned the highest weights. The combined CPPopt calculation then results from averaging all CPPopts taking their weight into account. Values on the x axis are minutes and those on the y axis are MABP in mm Hg. The percentage of monitoring time during which actual CPP was within the CPPopt range was significantly higher in survivors than in nonsurvivors (25.6% vs 19.7%, p = 1), and actual CPP was significantly closer to CPPopt in survivors (average absolute difference was 5.2 mm Hg for survivors and 6.9 mm Hg for nonsurvivors, p = 1). The average absolute differences outside the CPPopt range were significantly larger for nonsurvivors (8.5 mm Hg vs 7 mm Hg, p = 1). Deviations in the positive sense were higher in nonsurvivors (8.0 vs 6.8 mm Hg, p = 5), and the same was true for deviations in the negative sense (-8.4 vs -6.7 mm Hg, p = 1). In a multivariate logistic regression model using the CRASH outcome prediction model variables age, Glasgow Coma Scale (GCS) score, pupillary reactivity, and presence of extracranial injury as covariates, the average absolute difference between actual CPP and CPPopt in the first 48 hours after ICU admission remained an independent and negative predictor of survival (Table 2). Leuven-Tübingen Database Analyses Evaluating the difference per patient between CPPopts obtained by both methods resulted in a root mean squared error (RMSE) of 7.35 ± 1.72 mm Hg (± SD), a mean absolute error of 5.60 ± 1.83 mm Hg, and a mean error of -0.25 ± 3.89 mm Hg. Furthermore, 71.4% of all the differences between CPPopts for each minute that both methods produced a CPPopt over the first 48 hours for all patients were within -5 mm Hg and 5 mm Hg (Fig. 2). There was no statistically significant difference between the per-patient medians of CPPopt for DATACAR and PRx (p = 0.802), which had medians of 72.5 mm Hg (IQR 62.5 77.5 mm Hg) and 72.5 mm Hg (IQR 67.5 77.5 mmhg), respectively. The DATACAR method was able to produce a CPPopt recommendation in a median 97% (IQR 94% 98%) of monitoring time, whereas the PRx-based method could issue a recommendation in a median 44% (IQR 31% 55%) and could only issue a recommendation for more than 5% of the duration for 1 nonsurviving patient. Differences between nonsurvivors and survivors were only statistically significant for the DATACAR method regarding deviations in the negative sense (-1 mm Hg vs -6.1 mm Hg, p = 4); likewise, average absolute values outside the CPPopt range were significantly larger for nonsurvivors (9.6 mm Hg vs 5.8 mm Hg, p = 1). Discussion Building on the method developed by Steiner 23 and Aries 2 for CPPopt determination, the novelty of the current method lies in both the use of minute-by-minute ICP 1454

CPP target recommendation in severe traumatic brain injury TABLE 2: Logistic regression model for survival: CRASH model variables, including the average absolute difference between actual CPP and CPPopt range in the first 48 hours after ICU admission (Brain-IT database, n = 180)* Parameter CRASH Basic CRASH Basic + CPP CPPopt (mm Hg) Coefficient p Value Coefficient p Value CRASH basic model age 36 03 28 31 total GCS score 89 0.184 92 0.179 pupil reactivity 1.046 <01 0.997 <01 extracranial injury 0.637 0.173 0.648 0.180 CPP CPPopt (mm Hg) 0.246 17 * Values in boldface are statistically significant. First 48 hours. On admission. and MABP data and in the combination of correlation coefficients (LAx values) over different time intervals with CPPopt calculation over different time windows. The multiple resulting CPPopts were then combined in a weighted manner into a final CPPopt recommendation. Hence, autoregulation was investigated in a dynamic way by scanning different time scales to maximally exploit and optimize the potential information on cerebrovascular pressure reactivity capacity within routinely obtained monitoring data. This method is compatible with fully automated processing and, hence, can easily be programmed in common software, as in common patient data management systems. In the retrospective study based on the BrainIT database, the actual CPP being closer to the CPPopt recommendation was independently associated with increased survival. Moreover, no significant difference in the recommended CPPopt could be found when both the PRx method and the DATACAR method were compared in the Leuven-Tübingen database. Importantly, DATACAR was able to issue a recommendation almost all the time in the first 48 hours of ICU stay, which means a significant improvement of the performance of the CPPopt concept. Fig. 2. Difference between the middle of the CPPopt ranges derived by DATACAR and derived from PRx for each minute where both methods provided a CPPopt recommendation in the Leuven-Tübingen data set of 21 patients. In the landmark paper by Howells et al., 9 pressure reactivity was determined as a single coefficient for the entire monitoring period of the patient. Such a long time interval allows for retrospective determination of pressure reactivity but does not have the temporal resolution required for clinical decision making. 11 In the PRx methodology, slow variations in MABP are used to calculate MABP/ICP correlations. As a consequence, the PRx method requires high-resolution data (at least 60 Hz). It was demonstrated that PRx becomes clearly positive in episodes of ICP plateau waves, refractory intracranial hypertension, and arterial hypotension and hypertension, that episodes of positive PRx correlate with a disturbed transcranial Doppler derived index of autoregulation, and that averaged PRx negatively correlates with outcome, 5 all suggesting a strong effective relationship between PRx and actual autoregulation capacity. More recently, in a small study of 18 patients with intracerebral hemorrhage, a moving correlation coefficient of minuteby-minute MABP and ICP over a 20-minute time window, called L-PRx, has been shown to correlate well with the original PRx and was able to generate CPPopt recommendations in the same range. 21 We have been able to demonstrate in a previous study on the BrainIT database that the ICP/MABP correlation, calculated on minute-byminute data, was a significant predictor of future critical elevations of ICP as well as of unfavorable outcome. 8 These data support the results of the present study, that is, that the minute-by-minute resolution is sufficiently accurate to monitor pressure autoregulation in TBI patients. The results of the present study also add to the body of evidence for the hypothesis that a universal CPP threshold or even a fixed CPP target valid for all patients cannot be recommended 3,7,19 and that a dynamic CPP target based on pressure autoregulatory capacity likely represents a more promising concept. 2,9,21,27 Interestingly, it was shown that not only negative deviations from CPPopt resulted in poorer outcomes, but also positive deviations did. In other words, too high CPP, at the right side of autoregulatory capacity in the CBF-CPP plot, also resulted in a detrimental effect. The current study has a number of limitations. First, LAx and DATACAR have been developed by retrospective data analysis and have not yet been validated in a pro- 1455

B. Depreitere et al. spective clinical setting. Second, the presented method is computational. Studying the effect of arteriolar vasoconstriction and vasodilation through its effect on ICP will inevitably include a dampening effect. This is also true for other proposed indices of cerebrovascular reactivity, such as the Ursino model, based on the mathematical modeling of different vascular pressures and ICP including interactions based on CO 2 reactivity and CSF flow, 14,26 or the Ragauskas noninvasive method based on the assessment of the phase shift between MABP variation secondary to the respiratory cycle and ultrasound-based capture of intracranial blood volume waves. 18 Other models describe the transfer function analysis between arterial blood pressure and flow velocity in the middle cerebral artery, measured by transcranial Doppler. 1,25 Mx describes variations in flow velocity, measured at the middle cerebral artery vessels secondary to variations in CPP, 6 and has been shown to correlate with PRx. 5 We do have to keep in mind that all methods described represent attempts to indirectly assess the physiological phenomenon of changes in cerebrovascular smooth muscle tone secondary to altered intravascular pressure. None of the methods was validated directly by juxtaposing it to the actual smooth muscle cell phenomenon in an in vivo animal experiment. Nevertheless, all models do describe direct consequences of pressure reactivity and intact or deficient pressure autoregulation and, hence, can be considered to be reflections of the actual phenomenon, as shadows in Plato s cave. The association of the indices and derived CPPopt recommendations and outcome, albeit retrospectively, provides additional evidence for the hypothesis that the indices do describe an actual physiological phenomenon. The results of the current study demonstrate that continuous monitoring of pressure autoregulatory capacity based on standard resolution physiological monitoring data are equivalent to pressure autoregulation monitoring based on capturing waveform quality physiological data. Moreover, the tools presented here enable to significantly increase the percentage of monitoring time for which a CPPopt recommendation could be given. Consequently, DATACAR-based CPPopt recommendation enables for pressure autoregulation steered CPP management without investment in waveform data capture equipment, making the technique widely applicable in most ICUs. A recent randomized controlled trial in TBI patients in low-income countries could not demonstrate a difference in outcome of an ICP-lowering strategy based on invasive ICP monitoring versus management based on imaging and clinical examination. 4 These findings suggest that the measurement of a certain clinical parameter per se is not likely to change outcome, but rather the reaction and therapeutic strategy upon changes in this clinical parameter. A clinical trial on autoregulation monitoring based CPP management in TBI patients is thus likely to have a complex design. The developed indices in the present study contribute to the feasibility of such trial, and to the rationale for autoregulation-driven CPP management. Meanwhile, prospective validation of the tool developed in the current study is adamant, and initiatives for such studies are being developed. Conclusions The present study presents a computational method based on which it is demonstrated that minute-by-minute ICP/MABP data contain relevant information for autoregulation monitoring in TBI patients. Moreover, the new method allowed for CPPopt calculation in nearly the entire monitoring time. The method, which can easily be programmed in an automated tool at the bedside, has the potential to make the technique of pressure autoregulation steered management widely applicable in most ICUs. Acknowledgments We wish to acknowledge the BrainIT steering members: Iain Chambers, Giuseppe Citerio, Bart Depreitere, Per Enblad, Barbara Gregson, Tim Howells, Karl Kiening, Julia Mattern, Pelle Nillson, Ian Piper, Arminas Ragauskas, and Juan Sahuquillo. Disclosure The present study was supported by the Foundation for Scientific Research Flanders (Fonds voor Wetenschappelijk Onderzoek Vlaanderen, FWO) (Research project G. 0904.11). Geert Meyfroidt receives funding from FWO as senior clinical investigator (1846113N). Greet Van den Berghe receives long-term structural research financing via the Methusalem program funded by the Flemish Government (METH/08/07). BrainIT data collection was supported by The European Framework Programma (FP5- QLRI-2000-00454, QLG-2002-00160, and FP7-IST-2007-217049). The authors report no conflict of interest concerning the materials or methods used in this study or the findings specified in this paper. Author contributions to the study and manuscript preparation include the following. Conception and design: Depreitere, Güiza, Meyfroidt. Acquisition of data: Depreitere, Schuhmann, Maier, Piper, Meyfroidt. Analysis and interpretation of data: Depreitere, Güiza. Drafting the article: Depreitere. Critically revising the article: all authors. Reviewed submitted version of manuscript: all authors. Approved the final version of the manuscript on behalf of all authors: Depreitere. Statistical analysis: Güiza. Administrative/technical/ material support: Van den Berghe, Maier, Piper. References 1. Aaslid R, Lindegaard KF, Sorteberg W, Nornes H: Cerebral autoregulation dynamics in humans. Stroke 20:45 52, 1989 2. Aries MJ, Czosnyka M, Budohoski KP, Steiner LA, Lavinio A, Kolias AG, et al: Continuous determination of optimal cerebral perfusion pressure in traumatic brain injury. Crit Care Med 40:2456 2463, 2012 3. Bullock MR, Povlishock JT: Guidelines for the management of severe traumatic brain injury, 3rd edition. J Neurotrauma 24 (Suppl 1):S1 S106, 2007 4. Chesnut RM, Temkin N, Carney N, Dikmen S, Rondina C, Videtta W, et al: A trial of intracranial-pressure monitoring in traumatic brain injury. N Engl J Med 367:2471 2481, 2012 5. Czosnyka M, Smielewski P, Kirkpatrick P, Laing RJ, Menon D, Pickard JD: Continuous assessment of the cerebral vasomotor reactivity in head injury. Neurosurgery 41:11 19, 1997 6. Czosnyka M, Smielewski P, Kirkpatrick P, Menon DK, Pickard JD: Monitoring of cerebral autoregulation in head-injured patients. Stroke 27:1829 1834, 1996 7. Eker C, Asgeirsson B, Grände PO, Schalén W, Nordström CH: Improved outcome after severe head injury with a new therapy based on principles for brain volume regulation and preserved microcirculation. Crit Care Med 26:1881 1886, 1998 8. Güiza F, Depreitere B, Piper I, Van den Berghe G, Meyfroidt G: Novel methods to predict increased intracranial pressure 1456

CPP target recommendation in severe traumatic brain injury during intensive care and long-term neurologic outcome after traumatic brain injury: development and validation in a multicenter dataset. Crit Care Med 41:554 564, 2013 9. Howells T, Elf K, Jones PA, Ronne-Engström E, Piper I, Nilsson P, et al: Pressure reactivity as a guide in the treatment of cerebral perfusion pressure in patients with brain trauma. J Neurosurg 102:311 317, 2005 10. Jaeger M, Schuhmann MU, Soehle M, Meixensberger J: Continuous assessment of cerebrovascular autoregulation after traumatic brain injury using brain tissue oxygen pressure reactivity. Crit Care Med 34:1783 1788, 2006 11. Johnson U, Nilsson P, Ronne-Engström E, Howells T, Enblad P: Favorable outcome in traumatic brain injury patients with impaired cerebral pressure autoregulation when treated at low cerebral perfusion pressure levels. Neurosurgery 68:714 722, 2011 12. Kellie G: Appearances observed in the dissection of two individuals; death from cold and congestion of the brain. Trans Med Chir Soc Edinb 1:84, 1824 13. Lam JM, Hsiang JN, Poon WS: Monitoring of autoregulation using laser Doppler flowmetry in patients with head injury. J Neurosurg 86:438 445, 1997 14. Lodi CA, Ter Minassian A, Beydon L, Ursino M: Modeling cerebral autoregulation and CO2 reactivity in patients with severe head injury. Am J Physiol 274:H1729 H1741, 1998 15. Monro A: Observations on Structure and Functions of the Nervous System. Edinburgh: Creech and Johnson, 1783 16. Perel P, Arango M, Clayton T, Edwards P, Komolafe E, Poccock S, et al: Predicting outcome after traumatic brain injury: practical prognostic models based on large cohort of international patients. BMJ 336:425 429, 2008 17. Piper I, Citerio G, Chambers I, Contant C, Enblad P, Fiddes H, et al: The BrainIT group: concept and core dataset definition. Acta Neurochir (Wien) 145:615 629, 2003 18. Ragauskas A, Daubaris G, Petkus V, Ragaisis V, Ursino M: Clinical study of continuous non-invasive cerebrovascular autoregulation monitoring in neurosurgical ICU. Acta Neurochir Suppl 95:367 370, 2005 19. Robertson CS, Valadka AB, Hannay HJ, Contant CF, Gopinath SP, Cormio M, et al: Prevention of secondary ischemic insults after severe head injury. Crit Care Med 27:2086 2095, 1999 20. Rosner MJ, Daughton S: Cerebral perfusion pressure management in head injury. J Trauma 30:933 941, 1990 21. Santos E, Diedler J, Sykora M, Orakcioglu B, Kentar M, Czosnyka M, et al: Low-frequency sampling for PRx calculation does not reduce prognostication and produces similar CPPopt in intracerebral haemorrhage patients. Acta Neurochir (Wien) 153:2189 2195, 2011 22. Shaw M, Piper I, Chambers I, Citerio G, Enblad P, Gregson B, et al: The brain monitoring with Information Technology (BrainIT) collaborative network: data validation results. Acta Neurochir Suppl 102:217 221, 2008 23. Steiner LA, Czosnyka M, Piechnik SK, Smielewski P, Chatfield D, Menon DK, et al: Continuous monitoring of cerebrovascular pressure reactivity allows determination of optimal cerebral perfusion pressure in patients with traumatic brain injury. Crit Care Med 30:733 738, 2002 24. Sviri GE, Aaslid R, Douville CM, Moore A, Newell DW: Time course for autoregulation recovery following severe traumatic brain injury. Clinical article. J Neurosurg 111:695 700, 2009 25. Tiecks FP, Lam AM, Aaslid R, Newell DW: Comparison of static and dynamic cerebral autoregulation measurements. Stroke 26:1014 1019, 1995 26. Ursino M, Giannessi M: A model of cerebrovascular reactivity including the circle of Willis and cortical anastomoses. Ann Biomed Eng 38:955 974, 2010 27. Zweifel C, Lavinio A, Steiner LA, Radolovich D, Smielewski P, Timofeev I, et al: Continuous monitoring of cerebrovascular pressure reactivity in patients with head injury. Neurosurg Focus 25(4):E2, 2008 Manuscript submitted July 11, 2013. Accepted March 18, 2014. A portion of this work was presented in abstract form at the BrainIT meeting in Leuven, Belgium, on December 8, 2012. Please include this information when citing this paper: published online April 18, 2014; DOI: 10.3171/2014.3.JNS131500. Address correspondence to: Bart Depreitere, M.D., Ph.D., Neurosurgery, University Hospitals Leuven, Herestraat 49, Leuven B-3000, Belgium. email: bart.depreitere@uzleuven.be. 1457