Ricerca clinica sul trauma cranico: report di un Workshop Internazionale sull'efficacia comparativa
JOURNAL OF NEUROTRAUMA 29:32–46 (January 1, 2012)ª Mary Ann Liebert, Inc. Original ArticlesDOI: 10.1089/neu.2010.1599 Re-Orientation of Clinical Research in Traumatic Brain Injury: Report of an International Workshop on Comparative Effectiveness Research Andrew I.R. Maas, David K. Menon,2 Hester F. Lingsma,3 Jose A. Pineda,4 1 M. Elizabeth Sandel,5 and Geoffrey T. Manley 6AbstractDuring the National Neurotrauma Symposium 2010, the DG Research of the European Commission and theNational Institutes of Health/National Institute of Neurological Disorders and Stroke (NIH/NINDS) organizeda workshop on comparative effectiveness research (CER) in traumatic brain injury (TBI). This workshop re-viewed existing approaches to improve outcomes of TBI patients. It had two main outcomes: First, it initiated aprocess of re-orientation of clinical research in TBI. Second, it provided ideas for a potential collaborationbetween the European Commission and the NIH/NINDS to stimulate research in TBI. Advances in provision ofcare for TBI patients have resulted from observational studies, guideline development, and meta-analyses ofindividual patient data. In contrast, randomized controlled trials have not led to any identiﬁable major advances.Rigorous protocols and tightly selected populations constrain generalizability. The workshop addressed addi-tional research approaches, summarized the greatest unmet needs, and highlighted priorities for future research.The collection of high-quality clinical databases, associated with systems biology and CER, offers substantialopportunities. Systems biology aims to identify multiple factors contributing to a disease and addresses complexinteractions. Effectiveness research aims to measure beneﬁts and risks of systems of care and interventions inordinary settings and broader populations. These approaches have great potential for TBI research. Althoughnot new, they still need to be introduced to and accepted by TBI researchers as instruments for clinical research.As with therapeutic targets in individual patient management, so it is with research tools: one size does not ﬁt all.Key words: comparative effectiveness research; clinical research; clinical trials; methodology; systems biology;traumatic brain injuryIntroduction vehicles in low- and middle-income countries (Maas et al., 2008). Vulnerable road users (pedestrians, cyclists, and mo-Epidemiologic considerations torcyclists) are at high risk. The World Health OrganizationT he worldwide importance of traumatic brain injury (TBI) requires an effective, widely applicable response.The annual incidence of TBI is estimated at up to 500/100,000 (WHO) predicts that deaths from road trafﬁc incidents (pri- marily due to TBI) will double between 2000 and 2020, and that this increase will come exclusively in low- and middle-in the U.S. and Europe, and results in over 200 per 100,000 income countries (WHO/OMS, 2009).individuals being admitted to hospitals each year in Europe In developed, westernized societies the occurrence of TBI is(Langlois et al., 2006; Maas et al., 2008; Styrke et al., 2007; mainly increasing in people aged over 60 years (https:/ /Tagliaferri et al., 2006). Epidemiological surveillance suggests webgate.ec.europa.eu/idb/documents/2009-IDB-Report_that the nature of TBI is changing over time. Globally the screen.pdf; http:/ /www.cdc.gov/traumaticbraininjury/tbi_incidence is increasing, due mainly to greater use of motor ed.html). In the past, a TBI in this age group was thought to 1 Department of Neurosurgery, Antwerp University Hospital, Edegem, Belgium. 2 University of Cambridge, Cambridge, United Kingdom. 3 Department of Public Health, Center for Medical Decision Making, Erasmus MC, Rotterdam, The Netherlands. 4 Departments of Pediatrics and Neurology, Division of Critical Care Medicine, Washington University School of Medicine, St Louis,Missouri. 5 Kaiser Foundation Rehabilitation Center, Vallejo, California. 6 Department of Neurological Surgery, Brain and Spinal Injury Center, University of California–San Francisco, San Francisco, California. 32
TBI AND COMPARATIVE EFFECTIVENESS RESEARCH 33lead to a uniformly bad outcome; however, current experience cal outcomes in everyday clinical practice. To quote Goethe,is that a favorable outcome is possible and not uncommon. ‘‘Knowing is not enough; we must apply. Willing is not en-Nevertheless, the classic age range for inclusion in many ough; we must do.’’clinical trials excludes such people and makes trial data less These considerations provided the context for the DGrelevant to their treatment (Dhruva and Redberg, 2008). An- Research of the European Commission and NIH/NINDSother development is that the increasing burden of TBI due to to organize a workshop during the National Neurotraumamilitary conﬂicts, and the exposure of civilians in combat Symposium 2010 on comparative effectiveness researchzones has modiﬁed the epidemiology and clinical pattern of (CER).TBI (Risdall and Menon, 2011). Worldwide, TBI has devas- The aims of this workshop on CER were to: (1) perform atating effects on patients and their relatives and results in critical re-appraisal of approaches to clinical research in TBI;high socioeconomic costs to society. This necessitates a (2) identify the greatest unmet needs from a clinical perspec-widespread international effort to combat TBI and improve tive; (3) discuss the potential of CER in the ﬁeld of TBI; and (4)outcomes. explore the possibility and the added value of a joint EU-U.S. effort in this ﬁeld. This article summarizes the results of this workshop.Discordance between experimental and clinicalsuccess: The complexity of clinical research in TBI Approaches to clinical research in TBI: What has Research in basic science has disclosed that multiple (and has not) made a difference?mechanisms are involved in the pathophysiology of TBI(Marklund et al., 2006). This has led to the development of Clinical trials. Traditionally, the efﬁcacy of new treat-many neuroprotective agents with promising potential, but ments has been investigated in randomized controlled trialsnone have yielded beneﬁts in clinical testing. The gap between (RCTs). In TBI, these trials have been conducted on a range ofbench and bedside raises the question of why beneﬁts from neuroprotective agents and surgical and medical approachesseemingly effective interventions have not been seen in large (Maas et al., 2010a). Whereas various single-center studiesrandomized clinical trials. One explanation for this discrep- have reported beneﬁts of a range of interventions (for exampleancy may be that the clinical situation is far more complex and hyperbaric oxygen, mannitol, hypothermia, and decom-unpredictable than that seen with experimental models of pressive craniectomy), none of these ﬁndings were general-TBI. Experimentally, the type and degree of injury can be izable in multicenter RCTs. Furthermore, substantial selectionstandardized, while in the clinical situation wide variability bias may have existed in reporting beneﬁts in single-centerexists, both in type (e.g., diffuse and focal injuries) and in studies (Maas et al., 2010a). For example, of the 20 multicenterseverity of injury. In the clinical situation, pretreatment is studies on neuroprotective agents, only one reported a sig-impossible, and intervention within short therapeutic win- niﬁcant treatment beneﬁt: the HIT III study researchersdows can be challenging. Adverse effects of systemic insults, (Harders et al., 1996) found a beneﬁcial effect of the calciumsuch as hypoxia and hypotension, on outcome might easily channel modulator nimodipine, but targeted only patientsoverwhelm any beneﬁt from a new therapy. The identiﬁcation with traumatic subarachnoid hemorrhage. The generaliz-of relevant covariates and the development of robust prog- ability of this relatively small study (n = 123) is thereforenostic models by the CRASH (MRC CRASH Trial Colla- questionable, and taken in combination with negative resultsborators, 2008) and IMPACT (Steyerberg et al., 2008) from three other studies on nimodipine, these ﬁndings havecollaborations are very useful for dealing with the heteroge- not changed clinical practice.neity of the TBI population. However, adjustments for base- At the other extreme, a mega-trial of steroids in TBI wasline prognostic risk ignore the contribution of aspects of care, halted early because of increased mortality at 14 days in pa-such as the quality of intensive care and rehabilitation, on tients treated with corticosteroids (Roberts et al., 2004), butoutcome. For example, data from the NABISH-I study (Clif- did not show a signiﬁcant effect on quality of outcome atton et al., 2001) showed that inter-center variations in clinical 6 months (Edwards et al., 2005).care could produce substantial noise in results of clinical trials. It is not surprising that the past decade has seen a sharpRecent ﬁndings from the IMPACT studies have shown that decline in the number of clinical trials initiated in TBI, par-the risk of poor outcomes could differ between centers, and is ticularly in those on neuroprotective agents (Fig. 1). This mayup to three times higher than would be expected by chance be due to disappointment in the results of previous studies,after adjustment for baseline prognostic risk (Lingsma et al., along with the perception that TBI is a high-risk and costly2011). How much of the observed variability reﬂects vari- indication. Increasing overhead costs required by academicability in clinical management is unknown. research institutions in the EU and U.S. contribute signiﬁ- The impact that variability of care may have on clinical cantly to the high cost of clinical trials, in some casesoutcomes requires rigorous examination, because identiﬁca- amounting to or even exceeding 100% overhead, whichtion of underlying causes could result in real improvements in doubles the amount of funding required to conduct the trial.outcomes. Identiﬁcation of best clinical practices and efforts to The result is that academic institutions are pricing themselvesminimize recognized variances in care (Clifton et al., 2001), out of the market, and a clear shift in research has been seenthrough, for example, ‘‘Six Sigma’’ processes (Schweikhart towards the Far East and Latin America, where the potentialand Dembe, 2009), may yield beneﬁts. These may exceed for patient recruitment is higher and study costs are lowerthose provided by research addressing a single aspect in the (Maas et al., 2010a).continuum of care. It is therefore important to undertake more However, there is uncertainty whether ﬁndings obtainedimplementation-based research to ensure that current and in these settings may be extrapolated to higher-incomefuture effective therapies are translated into improving clini- countries, and about the importance of components of care
34 MAAS ET AL.FIG. 1. Number of clinical trials initiated over the past 30 years, differentiated by studies of neuroprotective agents versustherapeutic strategies. Adapted from Maas et al., 2010a.systems. As the search for effective therapies continues, there cols that are sometimes difﬁcult to replicate in the wideris a need for comparative studies in which differences, for clinical care spectrum. The discrepancy between clinical trialexample in access to acute and post-acute care, in trauma data and real world practice, and the need for other ap-organization, and in treatment, may be explored and the proaches to gathering evidence as a basis for clinical man-consequences for generalizability estimated. agement, was highlighted 15 years ago (Black, 1996). Guideline development. Some consistency in approach to Observational studies. Although relatively small incare has been achieved by management guidelines comparison to the number of clinical trials conducted in TBI,(www.tbiguidelines.org; www.nice.org.uk; Maas et al., 1997). observational studies have had a substantial impact uponMost of these guidelines are evidence-based, and follow a clinical management.rigorous and systematic analysis of the available literature. The Glasgow group initiated prospective data collection onDespite isolated reports that suggest otherwise (Cremer et al., severe head injury in 1968, and expanded this to an interna-2005), most reports show that adoption of such guidelines has tional level in 1972, including centers from the U.S. and theresulted in improvements in TBI mortality and functional Netherlands ( Jennett et al., 1976, 1977, 1980). These studiesoutcome (Bulger et al., 2002; Clayton et al., 2004; Elf et al., were facilitated by the introduction of the Glasgow Coma2002; Fakhry et al., 2004; Patel et al., 2002; Stein et al., 2010; Scale (GCS) and the Glasgow Outcome Scale (GOS), both ofSuarez et al., 2004; Varelas et al., 2006). Nevertheless, the level which are still universally accepted for classiﬁcation of initialof evidence underpinning the guidelines is, on average, only injury severity and outcome. Treatment approaches werelow. Of the 229 recommendations contained in seven guide- compared, and extensive prognostic studies revealed thatline publications, only four are based on class I evidence age, GCS score, and pupillary reactivity were the main pre-(Table 1). dictors of outcome in severe TBI, and prognostic rules were The limited number of recommendations that are sup- developed.ported by high-quality evidence is not unique to TBI, but has The Traumatic Coma Data Bank (TCDB; 1984–1987) pro-also been noted in other ﬁelds of medicine. A recent review of spectively collected data from four U.S. centers (Foulkes et al.,practice guidelines developed by the American College of 1991). This study highlighted the importance of systemicCardiology and the American Heart Association found that hypoxia and hypotension as determinants of outcomerelatively few recommendations were based on high-quality (Chesnut et al., 1993).evidence from overviews of RCTs, and many were based More recent reports suggest that physiological insultssolely on expert opinion, individual case studies, or standard continue to be an issue, both in the emergency room and in theof care (Tricoci et al., 2009). critical care unit context (Hlatky et al., 2004; Manley et al., The paucity of high-quality evidence reﬂects the many 2001). Numerous studies in animal models have conﬁrmeduncertainties about the beneﬁts and risks of multiple treat- the signiﬁcance of these clinical observations, while adding toment approaches, and the limited funding available for TBI our understanding of the pathological mechanisms involvedclinical research. We will never be able to mount adequately (DeWitt et al., 1995; Ishige et al., 1988; Statler et al., 2001). Apowered trials to address all of these uncertainties. Further- major advance employed in the TCDB was the Marshall CTmore, traditional clinical trials do not address the effect of classiﬁcation as a descriptive measure for the type of braindifferent clinical practices available in the real world, and damage and the risk of increased intracranial pressure (ICP).have limited external validity since the effect of treatment is Further studies have conﬁrmed its importance as a prognosticevaluated in selected populations, with management proto- parameter (Maas et al., 2007b).
TBI AND COMPARATIVE EFFECTIVENESS RESEARCH 35 Table 1. Evidence-Based Guidelines for Management of Traumatic Brain Injury (TBI): Strength of Recommendations Recommendations (n)Guideline Reference Topics (n) Class I Class II Class IIIPrehospital management Brain Trauma Foundation, 2000 7 0 5 12Penetrating brain injury Aarabi et al., 2001 7 0 0 12Pediatric guidelines Adelson et al., 2003 17 0 6 40U.K. guidelines for triage, U.K. National Institute for Health 27 3 16 107 assessment, investigation, and Clinical Excellence, 2003 and management of TBIField management of Brain Trauma Foundation, 2005 5 0 3 15 combat-related head traumaSurgical management of TBI Bullock et al., 2006 5 0 0 26Revised guidelines for management Brain Trauma Foundation, 2007 15 1 14 17 of severe TBITotal 83 4 44 229 aClassiﬁcation of evidence on therapeutic effectivenessClass I lack sufﬁcient patient numbers, or suffer from other methodological inadequacies that render them class II or III.Class II that were based on reliable data. Comparison of two or more groups must be clearly distinguished. Types of studies include observational, cohort, prevalence, and case-control. Class II evidence may also be derived from ﬂawed RCTs.Class III Types of studies include case series, databases or registries, case reports, and expert opinions. Class III evidence may also be derived from ﬂawed RCTs, cohort, or case-control studies. a From Guidelines for the Management of Severe Traumatic Brain Injury, 3rd ed., Brain Trauma Foundation, www.tbiguidelines.org. In the NICE guidelines (http://www.nice.org/), the grading scheme for level of recommendations was adapted from the Oxford Centre forEvidence Based Medicine levels of evidence as level A–D; for consistency, we considered grade A class I, grade B as class II, and grades C andD as class III. The core data survey conducted by the European Brain predictive value of many known predictors, has yielded newInjury Consortium (EBIC) concerned a 3-month data collec- predictors, and has resulted in the presentation of validatedtion period in moderate and severe TBI across 67 centers prognostic models for use in moderate and severe TBI(Murray et al., 1999). This study resulted in four publications (Steyerberg et al., 2008). These models have much wider ap-and demonstrated the evolution of CT lesions over time plications than only in the context of trial design (Lingsma(Servadei et al., 2000), as well as the importance of traumatic et al., 2010). Simulation studies demonstrated that broad in-subarachnoid hemorrhage (Servadei et al., 2002). Wide vari- clusion criteria, pre-speciﬁed covariate adjustment, and anations in intensive care management were reported across ordinal analysis all promote an efﬁcient trial, yielding gains inparticipating centers, without evidence of an association with statistical efﬁciency of over 40%. This means that smaller butimproved outcomes (Stocchetti et al., 2001). clinically relevant treatment effects can now be detected In an analysis of patients with head injury from the registry without increasing trial size or duration, and these ﬁndingsof the Trauma Audit and Research Network (TARN) in the have the potential to revolutionize the design and analysis ofU.K., Patel and associates (2005) found that improvements clinical trials in TBI (Maas et al., 2010b).over time in outcomes of patients with head injuries were less In summary, major advances in understanding and im-than those observed in other trauma patients. Patients with proving TBI care have come not from clinical trials, but fromsevere head injuries treated in a hospital without a neuro- observational studies, guideline development, and meta-surgeon had a 2.15-fold increase in the odds of death after analysis of individual patient data. Rigorously conductedadjustment for case mix, compared to those treated in a unit observational studies in large and diverse populations havewith neurosurgical facilities. the potential to identify better clinical practices and to reshape care for patients with TBI in the future. Meta-analysis of individual patient data. The IMPACT A critical re-appraisal of the experience within and acrossstudies (Maas et al., 2007a) brought together individual pa- clinical trials in TBI illustrates the inadequacies of approachestient data from eight RCTs and three observational studies to study design and analysis, clinical assessments, classiﬁca-(n = 9205; Marmarou et al., 2007). Relevant parameters from tion, and management.the individual studies were merged into a large dataset,forming a ‘‘culture medium’’ for exploring concepts for im- Bridging the gap between bench and besideproving the design of clinical trials in TBI (Maas et al., 2007a).The IMPACT studies have yielded important contributions to The possible causes for translational failures in clinicaladvance the ﬁeld of clinical research in TBI. Extensive prog- neuroprotection in TBI have been addressed in many previ-nostic analysis has deﬁned and quantiﬁed more precisely the ous publications (Green, 2002; Ikonomidou and Turski, 2002;
36 MAAS ET AL.Maas et al., 1999; Narayan et al., 2002; Tolias and Bullock, Classiﬁcation2004). One of the aspects highlighted during the panel dis- General approaches to classiﬁcation of TBI are by causecussions of the workshop was the uncertain relevance of re- (mechanistic) and severity. In clinical practice the conven-sults obtained from experimental studies in rodents to human tional categorization of the severity of brain injury is primarilysubjects, reﬂecting differences in brain development, brain based on the GCS score. While the GCS represented a sig-anatomy, and physiology. While studies in lissencephalic niﬁcant advance in the characterization of TBI, more recentspecies may continue to be useful in elucidating and antago- studies suggest that its accuracy and prognostic power innizing disease mechanisms, a strong plea was heard to im- more severely injured patients may be reduced due to effectsplement more intermediate studies in larger gyrencephalic of early sedation, neuromuscular blockade, and ventilationmammals, such as pigs and sheep, before clinical translation is (Czosnyka et al., 2005; Stocchetti et al., 2004). Further, itattempted. An alternative or adjunctive approach to such should be realized that the commonly employed differentia-large animal studies may be more preparatory clinical re- tion into mild (GCS score 13–15), moderate (GCS score 9–12),search, including the use of strategies to optimize candidate and severe (GCS score 8 or less) injury is artiﬁcial, and thattherapies before initiation of formal Phase II trials in patients clinical severity lies on a continuum. Pathoanatomical and( Janowitz and Menon, 2010). In TBI, perhaps uniquely among pathophysiological processes may occur across the spectrumacute CNS diseases, conventional protocols for clinical care of clinical severity, but may vary in frequency and severity.provide unparalleled access to key biological compartments Moreover, patients with similar clinical severity as assessedthrough techniques such as microdialysis, ventriculostomy, by the GCS may have widely differing types of injury (Fig. 2).and jugular bulb catheterization. This access, and the common Additional major issues are the physiological vulnerability ofuse of serial imaging (universally with computed tomography patients with TBI, particularly of those with moderate or se-[CT], but increasingly with magnetic resonance imaging vere injury and extracranial injuries. Thus the classiﬁcation of[MRI], and sometimes with positron emission tomography TBI is multidimensional and complex. A recent authoritative[PET]), provides opportunities to examine disease mecha- workshop has made a strong case for a new and more com-nisms, drug pharmacokinetics, and proof of principle of prehensive categorization of TBI in clinical trials (Saatmantherapy in small, well-designed studies. We suggest that those et al., 2008). The workshop also recommended the develop-monitoring and investigation tools that are commonly used in ment of a common data set of TBI data elements, and thepatients with TBI may provide a route to facilitate a more establishment of a large prospective patient database acrossrational approach to translational research. This suggestion is the spectrum of injury severity. Common data elements haveunderpinned by existing research data on disease biology, been proposed (Maas et al., 2011), but a large prospective datadrug delivery, and treatment response obtained with these set is still required.methods ( Janowitz and Menon, 2010). Mechanistic targetingGenetic make-up and response to injury Different pathophysiological processes may occur simul- A major gap in knowledge concerns different responses to taneously or sequentially, and to varying degrees. The con-similar injuries. Such patient-speciﬁc differences could in part cept of mechanistic targeting—the ideal for clinical trials—be genetically determined, and much research will be needed will require reliable identiﬁcation of occurrence and timein the areas of genomics and metabolomics to elucidate the course of pathophysiological mechanisms in individualwide variability in the response to brain injury. The relevance patients. There is the emerging hope that multimodalityof genetics may be seen by the observation that recovery is monitoring may allow us to differentiate between patho-poorer in patients with stroke or TBI who have the APOE e4 physiologies that appear superﬁcially similar, but requireallele than in those who do not have this allele (Alexander different treatment. For example, the combination of con-et al., 2007). Other genes for which evidence exists for asso- ventional monitoring with brain tissue oximetry and micro-ciations with outcomes are the TP53, COMT, DRD2, CAC- dialysis may allow us to differentiate between classicalNA1A, and BCL2 genotypes ( Jordan, 2007; Zangrilli Hoh ischemia (which may respond to cerebral perfusion pressureet al., 2010). [CPP] elevation), and diffusion hypoxia (Menon et al., 2004), or mitochondrial dysfunction (Vespa et al., 2005), which mayPremorbid factors and comorbidity respond to normobaric hyperoxia (Nortje et al., 2008). The emerging ﬁelds of advanced MRI imaging and of proteomic The importance of premorbid factors and comorbidity biomarkers offer opportunities for detection and tracking ofhas often been neglected in TBI, on the assumption that pri- pathophysiological processes.marily young and otherwise healthy subjects are afﬂicted.Awareness of the increase of TBI in older people and in Individualizing clinical managementthose with prior problems has changed this. A recent analysisrevealed that one in four U.S. citizens now lives with at least Protocol-driven approaches are currently the standard intwo chronic comorbidities (Parekh and Barton, 2010). These the treatment of TBI. These approaches are often poorlycomorbidities imply impaired physiological reserve and a focused, utilizing a stepwise approach, with an escalatingrange of pre-existing medications (including anticoagulant intensity of therapy, regardless of the underlying patho-therapy and platelet aggregation inhibitors). These factors physiology. They adhere to the concept of a ‘‘one pill for ev-will inﬂuence both disease course (e.g., via more hemorrhagic erybody’’ approach. The introduction of novel monitoringexpansion), and the pharmacokinetic interactions and phar- technology, and advances in neuroimaging techniques, nowmacodynamic safety of new therapies in TBI. offer opportunities for advancing care from a one-size-ﬁts-all
TBI AND COMPARATIVE EFFECTIVENESS RESEARCH 37FIG. 2. Types of brain injury may differ greatly in patients with similar initial clinical severity as assessed by the GlasgowComa Scale. Adapted from Saatman et al., 2008 (EDH, extradural hematoma; DAI, diffuse axonal injury; SDH, subduralhematoma; SAH, subarachnoid hemorrhage; IVH, intraventricular hemorrhage).approach, to a more focused approach, targeted to the needs more relevant than the actual values of ICP. Second, the use ofof an individual patient. The concept of individualized man- ICP as a surrogate outcome marker neglects the side effects ofagement is closely related to the possibilities for mechanistic therapies (Coles et al., 2002, 2007). The importance of com-targeting. There are clear beneﬁts to be gained from the em- plications following, for example decompressive cra-ployment of personalized approaches. For example, the curve niectomy, was recently demonstrated in the DECRA trialof the relationship between CPP and outcome shows an in- (Cooper et al. 2011).verted ‘‘U’’ shape, suggesting that there is an optimum level of A more rational option would be to use deﬁnitive markersCPP, and that higher or lower levels are disadvantageous of tissue fate in the brain. Biomarkers may have great poten-(Balestreri et al., 2006). Indeed, the use of management tial in TBI (Table 2). Advances in this emerging ﬁeld offerschemes that incorporate vigorous CPP augmentation can hope for the identiﬁcation of biochemical and other markersresult in signiﬁcant cardiorespiratory complications (Ro- that are clinically relevant for quantifying and tracking dis-bertson et al., 1999). Further, there is accumulating evidence ease processes. Further, if we could use patients as their ownthat the optimum level of CPP varies between patients, and controls in determining if an intervention altered the trajec-the optimum level in a given patient may be identiﬁable using tory of neuronal loss, we could at least determine whether aautoregulatory indices (Howells et al., 2005; Steiner et al., neuroprotective intervention was initially effective, in that it2002), or brain oximetry (Spiotta et al., 2010). reduced incremental tissue injury. Serial MRI with diffusion tensor imaging (DTI; Niogi et al., 2008), and MR spectroscopyEarly end-points and mechanistic targets (Ashwal et al., 2006; Govind et al., 2010), may provide clini- cally viable methods of assessing such incremental neuronal Major advances in the ﬁelds of cardiovascular medicine, losses, but this approach needs to be tested in largeoncology, and AIDS research have resulted from the use ofearly (mechanistic) end-points, for example troponin inmyocardial infarction and CD4 counts in HIV. In TBI, how-ever, early end-points that reliably predict quality of recovery Table 2. Potential Uses for Biomarkersare not yet available. In the past, ICP has been used, either in Traumatic Brain Injury (TBI)explicitly or implicitly, as a surrogate end-point, especially inearly-stage trials in clinical TBI. While this approach has some Establishing a diagnosis of TBI (relevant to mild cases)merit, it is important to recognize two important confounders. Assessing the severity and nature of TBI First, modern neuro-ICU practices have substantially Monitoring the evolution of injury and recovery in individ-blunted our ability to use ICP as a surrogate marker in this ual patients or groups of patientsway. It is possible to control ICP by intensifying ICP/CPP Deﬁning treatments neededtherapies until the system terminally decompensates and in- Monitoring of treatment effects Mechanistic target for clinical trialstracranial hypertension becomes refractory to therapy. In this Prediction of outcomecontext, the intensity of ICP/CPP-targeted therapy may be
38 MAAS ET AL.prospective studies that map the temporal trajectories of le- consistent with post-traumatic stress disorder (PTSD) are re-sion evolution with conventional treatment. ported in up to 40% of military personnel returning from deployment following mild blast TBI (Hoge et al., 2008), al-Post-acute assessments and care though the study is controversial because it may have un- derestimated the percentage of patients with mild TBI who Consistent standardized methods are largely lacking to experienced PTSD (Correspondence NEJM, 2009). Questionstrack service utilization and changes in functional status fol- persist concerning the best methods for the determination oflowing acute care discharge. Where they do exist, general use the neurological and/or psychological aspects and etiologiesis often inhibited by the copyright protection imposed by of these interrelated disorders (Bryant, 2008). In one studydevelopers. In comparison to acute care studies, post-acute using survey techniques for military personnel injured in Iraqstudies are often small in size and research is fragmented. In and Afghanistan, researchers found that 12% of respondentsparticular, there is little continuity in research between acute reported a history consistent with mild TBI, and 11% screenedand post-acute care studies. Nevertheless, disparities in access positive for PTSD (Schneiderman et al., 2008). Advancedto post-acute care may inﬂuence the recovery process and neuroimaging techniques such as MRI spectroscopy offerconfound interpretation of outcomes. A major challenge in the opportunities for the diagnosis of brain abnormalities in thispost-acute care phase is posed by the highly variable time population (Hoge et al., 2009). Civilian TBI studies have rarelyperiods at which data are recorded, confounding compara- included measures of PTSD, and consequently the incidencebility of studies and interpretation of their results. Thus a of this combined diagnosis is unknown in civilian popula-great need exists for more prospective longitudinal studies tions. There is clearly a great need for further research in thisbridging the gap between acute and post-acute research in area, including studies that allow us to differentiate moreTBI. accurately between PTSD and the sequelae of TBI in all populations.Outcome The GOS is currently accepted by investigators and regu- Pediatric considerationslatory authorities alike as the standard end-point for judgingefﬁcacy in a cohort of patients with TBI. Outcome after TBI, Despite past and current prevention efforts, TBI continueshowever, is by deﬁnition multidimensional, including neu- to kill and disable more children and adolescents than anyrophysical disabilities, disturbances in mental functioning other cause (Forsyth et al., 2008; Langlois et al., 2005). There is(e.g., cognitive and executive functioning), and consequent a paucity of widely-applied, effective therapies, and resultsproblems in social reintegration. Moreover, even the Ex- from adult studies may not be applicable to children. Whiletended GOS is relatively insensitive in its upper ranges, and conducting research in pediatric TBI, researchers need tomay therefore be less suited for patients with milder injuries. carefully balance the need for better therapies and delivery ofRecently, the Neurological Outcome Scale for TBI (NOS-TBI) care with the consideration that children with TBI are a vul-has been proposed as an effective and simple tool to quantify nerable population in whom to conduct research. There isneurological deﬁcits after TBI (Wilde et al., 2010). This ap- only fragmented understanding of disease biology, and theproach already takes some multidimensional aspects of out- frequency and geographical distribution of pediatric TBIcome assessment into consideration. create unique challenges for researchers (Natale et al., 2006). Relatively few TBI studies have utilized Health-Related Translation of knowledge and effective therapies from theQuality of Life (HRQoL) measures. There is a need for com- bench to the bedside are of particular signiﬁcance in pediatricprehensive disease-speciﬁc instruments for HRQoL assess- TBI. While a body of research has provided important pre-ment in persons after TBI, such as the novel disease-speciﬁc clinical data, TBI models in immature animals have uniquescale Quality of Life in Brain Injury (QOLIBRI; Truelle et al., limitations, including differences in brain development be-2010; von Steinbuechel et al., 2005, 2010a, 2010b). There is a tween humans and the species used in the laboratory, injuryneed to develop a multidimensional approach to outcome mechanics, and evaluation of recovery (Prins and Hovda,assessment and classiﬁcation, including the patient’s quality 2003). Pre-clinical studies have and will continue to makeof life. contributions to the ﬁeld, but more work needs to be done to improve our understanding of the biology of TBI in children. It is through this understanding that the suitability of a givenPsychological health and TBI mechanism as a target for therapy can best be deﬁned. On- The workshop on ‘‘An Integrated Approach to Research in going efforts to understand the pathobiology of pediatric TBIPsychological Health and TBI,’’ organized by NIH/NINDS, will build on previous single-center studies with limitedthe Department of Defense, the Department of Veteran Af- sample sizes, and the consequent need to use pooled samplesfairs, and the National Institute on Disability and Re- (Kochanek et al., 2000). Such an approach has value, buthabilitation Research in October 2008, revealed the previously results in fragmented descriptions of the time course ofunrecognized relationship between these two seemingly un- the pathological cascades under study, and limits externalrelated disciplines (Thurmond et al., 2010). First, substance validity.abuse is common in victims of TBI and may often be a con- Variability in aspects of clinical care can affect the appli-tributing factor in injury mechanism. Second, depression and cability of controlled studies to the general population. Foranxiety disorders are relatively frequent in survivors, and this example, children with severe TBI can be admitted to adultmay adversely affect social reintegration and confound out- hospitals, adult trauma centers (with or without pediatriccome assessments. Mental health and chronic pain are con- qualiﬁcations), pediatric hospitals, or pediatric trauma cen-sidered key areas, both pre- and post-injury. Third, symptoms ters. Furthermore, it is estimated that 17,000,000 children in
TBI AND COMPARATIVE EFFECTIVENESS RESEARCH 39the United States do not have timely access to high-level pe- Table 3. Summary of Unmet Needs and Prioritiesdiatric trauma care (Carr and Nance, 2010). Transport deci- for Clinical Researchsions are inﬂuenced by geographic factors and emergency Bridge the gap between bench and bedsidemedical service practices. Variability of clinical care within - Promote interaction between basic scientists and clinicaland between pediatric centers adds to these challenges. Many researchers in order to better scale experimental modelsof the challenges described in the adult TBI population are to reﬂect human pathology both physically and patho-magniﬁed when it comes to pediatric TBI. Speciﬁc consider- logicallyations concern injury classiﬁcation, the transition from acute - Experimental work-up should include testing in largerto sub-acute care, and the selection of optimal outcome animalsmeasures that include measurements of quality of life and - Better optimization of candidate drug molecules in clinicalevaluations of psychological health (Winthrop, 2010). As re- disease through experimental medicine approachessearchers address these challenges, the dynamic nature of - Develop mechanistic end-points in human traumatichuman development needs to be taken into account when brain injury (TBI), such as biomarkers and advanced neuro-imagingevaluating injury type, physiological values (such as CPP and Integrate acute and post-acute care researchcerebral blood ﬂow [CBF]), and cell injury mechanisms and Develop a novel approach to the classiﬁcation of TBI, alsorecovery (e.g., language, executive function, and measures of considering extracranial injuriesindependence; Kapapa et al., 2010; Kochanek, 2006; Walker Develop a multidimensional approach to outcome assess-et al., 2009). ment, including the patient perspective (quality of life) During the workshop, a review of the international net- Explore the inﬂuence of gender and genetic makeup onworks available to conduct research in pediatric TBI revealed disease course and outcomenumerous opportunities and highlighted challenges. Identi- Develop tools to better capture clinical variabilityﬁcation and understanding of best clinical practices through Use information on clinical variability to develop and testCER may help optimize and homogenize care delivery, fa- strategies for individualized management Prediction researchcilitating mechanistic targeting and the implementation of - Outcome: update/validate prognostic modelsfuture effective therapies. Whether via RCTs, rigorously - Prediction of treatment responseconducted observational studies, or other forms of CER, un- - Prediction of the expected trend using monitoredderstanding pediatric TBI requires strong collaboration parametersamong multiple centers and across countries and continents. - Prediction of the risk of hemorrhagic expansion Involve information technology personnel and other expertsAlternative Approaches and Priorities for Clinical from unrelated ﬁelds in explorations of novel approaches toResearch in Traumatic Brain Injury classiﬁcation (pattern recognition), and prediction research (machine learning techniques) The issues summarized above highlight the complexity of ‘‘Open source’’ research: data sharing and data standard-the challenges facing clinical TBI research. Priorities for clin- izationical research are summarized in Table 3. A strong collabora- A particular focus on pediatric and elderly subpopulationstive international effort will be required to address these. Collaborate with psychological health and pain experts In the absence of possibilities for mechanistic targeting, the Ensure that improvements in therapy are applicable totraditional approach to clinical trials in TBI has been to de- settings where they are needed most (developing economies)crease heterogeneity in patient populations by employing Explore whether ﬁndings obtained in a particular setting (e.g., developing economies) may be extrapolated to otherrestrictive enrollment criteria. The disadvantage of this ap- settingsproach is that it is statistically inefﬁcient and decreases ex-ternal validity. The question is if attempts to limitheterogeneity are appropriate, or alternatively, that the ex- organisms or cell types are responsible for the pathology,isting heterogeneity in patient populations, management ap- there has been only limited success using this approach forproaches, and outcomes, may be used to advantage by more heterogeneous complex diseases, such as inﬂammatoryexploring these differences and analyzing the underlying disease, diabetes, and cardiovascular disease. In these com-causes for a given outcome or individual patient response to a plex disorders there is likely no single factor that is respon-selected therapy or intervention. Classical clinical trials, with sible for the disease. This is particularly true for disorders oflarge numbers and substantial costs, may not be suitable ve- the central nervous system, such as traumatic brain injury, forhicles for providing answers to all of the questions that we which there is signiﬁcant heterogeneity in the etiology, pa-have. We need alternative approaches to address these thology, mechanisms, and outcome.questions. One direction to in which to proceed is the pro- In contrast to the reductionistic, hypothesis-driven ap-spective collection of large, multi-scalar (demographics, proach that seeks to target a single variable responsible for thephysiology, proteomics, genomics, and outcome), longitudi- disease, a systems biology approach aims to identify multiplenal, high-quality clinical databases, associated with systems factors that contribute to the disease. Systems biology alsobiology and CER methods. addresses the complex interactions of these multiple variables in a multivariate, multidimensional manner, over time. Sys-A Systems Biology Approach tems biology is a rapidly growing interdisciplinary ﬁeld that Traditional clinical trials and studies have relied upon a combines biology, mathematics, statistics, and computer sci-hypothesis-driven, model-based approach. While this reduc- ence, to better understand complex biological processes.tionistic approach has been very successful in developing In this regard systems biology may be better considered antreatments for infectious diseases and cancer, where single ‘‘informational science’’ approach, in which hypotheses are
40 MAAS ET AL.data driven and seek to describe the behavior of the entire Table 5. Common Characteristics of Comparativesystem. In lay terms, it is the description of the forest instead Effectiveness Researchof the trees. There are now numerous examples of the appli- The objective of directly informing a speciﬁc clinicalcation of the systems approach to complex biological prob- decision from the patient perspective, or a health policylems, particularly in microorganisms (Bischofs et al., 2009; decision from the population perspectiveSpiro et al., 1997), and more recently complex human disease Comparison of at least two alternative interventions, each(Chen et al., 2009; Sears et al., 2009). These studies demon- with the potential to be a best practicestrate how a systematic, integrative analysis that includes Description of results at the population and subgroupgenes, proteins, and behavior over time, can solve complex levelsproblems that are insufﬁciently addressed by the conven- The use of outcomes—both beneﬁts and harms—that aretional, reductionistic approach. important to patients Employment of methods and data sources appropriate forComparative effectiveness research the decision of interest Interventions conducted in settings that are similar to RCTs—generally considered to be the gold standard— those in which the intervention will be used in practiceaddress efﬁcacy rather than effectiveness. Efﬁcacy reﬂects thedegree to which an intervention produces the expected resultunder carefully controlled conditions chosen to maximize the was used in 43% of the Rotterdam cases and in 14% of thelikelihood of observing an effect if it exists. The study popu- Groningen cases. The authors concluded that the results oflation and setting of efﬁcacy studies may differ in important their study did not support the concept that an aggressiveways from those settings in which the interventions are likely management regimen would improve outcome. However,to be used. By contrast, CER intends to measure differences in the efﬁcacy with which aggressive management was im-outcome and to relate these to the package of care and its plemented and the impact on target pathophysiological var-constituent components in ordinary settings and broader iables (such as cerebral perfusion pressure and carbon dioxidepopulations. It can therefore be more relevant to policy eval- control) was not analyzed.uation and the health care decisions of providers and patients This study can be considered a CER study, although per-(Table 4; IOM, 2009). formed long before this type of research was recognized as a Many different ofﬁcial deﬁnitions of CER exist. Common separate and important entity or even had a name. It usedcharacteristics are presented in Table 5. observational data from a setting that represents clinical The call for CER does not mean that all research must have practice. It measured a patient-relevant outcome, and thethese characteristics. Early studies of an intervention should study aimed to inform medical decision making.also compare it to a placebo, standard care, or no intervention. Similarly, the comparative analysis of treatment resultsDuring early development of a new intervention, it is critical between Charlottesville (U.S.) and New Delhi (India) reportedto determine safety and efﬁcacy under a deﬁned set of cir- by Colohan and colleagues in 1989 qualiﬁes as CER. An al-cumstances (IOM, 2009). most 8% higher mortality was found in patients with a lo- calizing motor response in New Delhi (12.5% versus 4.8%).Examples of comparative effectiveness research The relative absence of pre-hospital emergency care and thein TBI delay in admission after head injury were considered as two The concept of CER in TBI is not new. In 1983 Gelpke and possible causes for these differences (Colohan et al., 1989).associates analyzed differences in outcome between two A more recent example of CER comes from the IMPACTcenters from the Netherlands (Rotterdam and Groningen) studies. Here, differences in outcome between centers wereparticipating in the International Data Bank of severe head quantiﬁed across 10 RCTs and three observational studies,injury. The 1-year survival rate in Rotterdam was 45% versus containing data on 9578 patients with moderate and severe63% in Groningen. The research question was whether the TBI (Lingsma et al., 2011). The between-center differences indifference in survival rate was due to differences in the initial unfavorable outcome (dead, vegetative state, or severe dis-severity of the injury or to a difference in management efﬁ- ability as measured with the Glasgow Outcome Scale) at 6cacy. Of the 18% difference in survival rate, 10.5% was due to months were estimated with a random effects logistic re-differences in severity of injury on admission. The remaining gression model. An odds ratio (OR) was estimated for each7.5% difference in survival rate was not explained, but may center by comparing the number of patients with unfavorablehave been caused by unmeasured variations in the initial outcome to the average, set at an OR of 1. The authors founddetermination of severity of injury or by differences in man- that the 95% range of ORs among centers was 0.55–1.83,agement. Groningen had a more conservative management meaning that there are centers in which the odds of unfa-approach than Rotterdam. For example, artiﬁcial ventilation vorable outcome are almost half the average, and centers where the odds of unfavorable outcome approach twice the average. There is thus a more than threefold difference in the Table 4. The Different Intents of Randomized probability over and above chance effects to have an unfa- Controlled Trials (RCTs) and Comparative vorable outcome between the centers, which could not be Effectiveness Studies (CERs) explained by adjustment for the most important predictors of outcome in TBI: age, GCS motor score, and pupil reactivity.Efﬁcacy Can it work? RCTEffectiveness Does it work? CER Limitations resulting from the nature of the IMPACT da- tabase (inconsistent recording of relevant variables across Adapted from Drummond et al., 2008. multiple studies) unfortunately precluded a detailed
TBI AND COMPARATIVE EFFECTIVENESS RESEARCH 41comparative analysis aimed at exploring possible underlying versus home or outpatient care? Management issues in thecauses in depth. The observed center differences clearly acute care of more severely injured patients include whichdemonstrate the potential for CER in TBI and the importance treatment modalities for treating ICP should be used and inof rigorously conducted, comprehensive, consistent, pro- what sequence. Uncertainty exists about the optimal timingspective data collection across multiple centers. One key step for extracranial surgery, and for the indications and timing ofin such efforts will be the development of common data ele- surgery for contusions and for treatment of raised ICP.ments for TBI (CDEs), and the implementation of web-based, There are four unique features of TBI that make CER aefﬁcient data collection tools (Maas, 2009; Maas et al., 2010c, feasible approach to address these uncertainties.2011). First, there are large between-center differences and be- A further example of the application of CER to TBI—and tween-country differences in both outcome and management.now in the setting of post-acute care—is based upon stan- On one hand these differences might be considered worri-dardized data collection performed by the Kaiser Permanente some. On the other hand they provide a major opportunity tohealth system in the U.S. Kaiser Permanente (KP) serves the compare alternative interventions/management strategies/health care needs of 8.2 million members in nine states and the care organization that all are possible best practices, in ev-District of Columbia, including California, Colorado, Georgia, eryday clinical practice.Hawaii, Maryland, Ohio, Oregon, Virginia, and Washington. Second, robust covariates and validated prognostic modelsThe KP system offers a unique opportunity to compare vari- have been developed speciﬁcally for TBI by the CRASH andation of care and outcomes. Much of the data concerning an IMPACT collaborations (MRC CRASH Trial Collaborators,encounter or episode of care is standardized, and permits 2008; Steyerberg et al., 2008). These models provide the pos-comparison of populations based on demographic charac- sibility to adjust for patient characteristics that affect outcome.teristics, care settings and trajectories, service delivery, and Third, advanced statistical models, including random ef-also outcomes by mining and analyzing data. In the KP fects models, are available to analyze differences betweenNorthern California Region (KP-NCAL), where membership centers.currently stands at 3.2 million members, the annual incidence Fourth, recommendations have been developed for stan-of acute brain injury is approximately 2500, and a large-scale dardization of data collection and coding of variables.genetics database is being created that will be available for TBI However, what is currently still lacking in TBI is the exis-studies. In 2002, The Division of Research in Oakland (KP- tence of a contemporary observational dataset with high-NCAL) created a large mild TBI registry that is also available quality, uniformly-collected highly granular, prospectivefor research studies. KP has internally funded a study of TBI to data. This is crucial to performing high-quality CER studies. Itdetermine the risk of development of Alzheimer’s disease that has been argued that data for CER research might be drawnwas to be completed at the end of 2010. from ongoing clinical trials. This approach would, however, With the availability of large databases and variability in violate a main principle of CER studies, namely that theyutilization of services, an opportunity exists to study the should be conducted in settings similar to those in which thevariation in outcomes of care delivery. For KP-NCAL neuro- intervention will be used in practice. Moreover, the experiencetrauma patients, care may be initially provided outside the from the IMPACT studies illustrates that in-depth CER anal-system, in a non-KP trauma center or an emergency room. For ysis of such data is not possible if speciﬁc research interests,example, KP-NCAL patients may receive acute trauma care at such as treatment-speciﬁc effects, have not been pre-San Francisco General Hospital, a Level 1 trauma center, or determined and data collection targeted accordingly. We ar-Kaiser South Sacramento, a Level 2 trauma center, or in a KP gue that contemporary prospective data collection is essentialemergency department. Patients with mild TBI are often seen and should be carried out through a coordinated effort in-in emergency departments, within or outside the KP inte- volving a large number of clinical centers in several countries.grated health care system, and released to home with a re- This initiative would require a signiﬁcant investment in termsferral to their primary care physicians. Others are referred of time and money. However, such an investment would beto specialists, for example at the regional rehabilitation certainly repaid by the results that would stem from the CERcenter or elsewhere, by local trauma centers. Physiatric co- analysis of the collected data. CER research has the potentialmanagement with a primary care provider, or evaluation by a to answer the many open questions in TBI, and further offersneuropsychologist, may also be an important model to study. opportunities for cost utility studies in the context of healthStudies of inpatient hospital versus skilled nursing facility technology assessment in TBI.versus home or outpatient care may be further developedusing this approach. Conclusions In many ways this workshop points to a paradigm shift inThe future potential of comparative clinical research in TBI. First, the joint organization by Euro-effectiveness research in TBI pean and U.S. funding agencies reﬂects the need for interna- There are many unanswered questions in TBI, that relate to tional collaboration. Second, we have come to realize thatthe process of care, trauma organization, and speciﬁc ap- approaches other than the reductionistic methods of strictlyproaches. Many uncertainties concerning best clinical practice controlled trials should be considered in clinical TBI research.exist in TBI. Improved clinical care of TBI patients will likely depend on How should acute trauma care and post-acute care be or- a range of research approaches, including systems biologyganized? Does time to referral to a specialist for post-acute and CER. These approaches are not yet widely used in clinicalcare matter? Can any preference in the post-acute care setting TBI research, ﬁrst because of unfamiliarity, second because ofbe found for inpatient hospital versus skilled nursing facility the lack of a rich and comprehensive human dataset that
42 MAAS ET AL.includes demographic, clinical, genomic, proteomic, imaging, sions. Their input was essential in making this workshop aand detailed outcome data across multiple time points. It is success. A particular word of thanks is due to Dr. Torilessential that CER studies comply with published guidelines Skandsen (Trondheim, Norway), for providing her notes offor such studies (Sox et al., 2010). The lower costs of this type the discussions, to Dr. Kenneth Curley for input in draftingof approach make it particularly attractive for studies that the manuscript, to Eno Lavrysen for administrative support inotherwise would be cost-prohibitive. Well conducted studies preparing the manuscript, and to Sir Graham Teasdale forwill translate into clinically meaningful results, which will advice in editing the ﬁnal version.directly inform decision making in clinical practice, and thusimprove care for TBI patients. An important barrier to sys- Author Disclosure Statementtems biology approaches is cultural, in which this new data-driven systems approach challenges the current reductionistic All of the authors conduct research in TBI and receiveapproach to clinical research, and requires a new way of funding for this research from various sources. None of thethinking about human biology and disease. It also requires authors received any honorarium or other compensation formultidisciplinary teams of investigators from disciplines that their participation at the workshop or for their contributionshave not previously worked together to apply and reﬁne in writing this manuscript.these promising mathematical and computational tools.Overcoming existing prejudice will require vision and sufﬁ- Referencescient and sustained funding. Aarabi, B., Alden, T.D., Chestnut, R.M., Downs, J.H., Ecklund, However relevant these approaches may be, it is important J.M., Eisenberg, H.M., Farace, E., Florin, R.E., Jane, J.A.,that we do not view them as the new panacea in TBI research. Krieger, M.D., Maas, A.I.R., Narayan, R.K., Potapav, A.A.,As with therapeutic targets in individual patient manage- Salazar, A.M., Shaffrey, M.E., and Walters, B.C. (2001). Man-ment, so it is with research tools in populations of patients: agement and prognosis of penetrating brain injury—guide-one size does not ﬁt all. lines. J. Trauma 51, S1–S86. The joint organization of the workshop by European and Adelson, P.D., Bratton, S.L., Carney, N.A., Chestnut, R.M., duU.S. funding agencies holds promise. It reﬂects the agencies’ Coudray, H.E., Goldstein, B., Kochanek, P.M., Miller, H.C.,interest in joining efforts to move the TBI ﬁeld forward, and Partingtan, M.D., Seldon, N.R., Warden, C.R., and Wright,this was conﬁrmed during the American Association for the D.W. (2003). Guidelines for the acute medical management ofAdvancement of Science meeting (Washington, D.C., Febru- severe traumatic brain injury in infants, children, and ado- lescents. Pediatr. Crit. Care Med. 4, 1–75.ary 2011), when the EU commission and the NIH jointly or- Alexander, S., Kerr, M.E., Kim, Y., Kamboh, M.I., Beers, S.R., andganized a session on ‘‘transatlantic synergies to promote Conley, Y.P. (2007). Apolipoprotein E4 allele presence andeffective traumatic brain injury research.’’ An international functional outcome after severe traumatic brain injury. J.collaborative initiative would be timely and of signiﬁcant Neurotrauma 24, 790–797.added value for the TBI ﬁeld. Constituting an observational Ashwal, S., Holshouser, B.A., and Tong, K.A. (2006). Use ofdataset with data of quality high enough to perform system advanced neuroimaging techniques in the evaluation of pe-biology approaches and CER studies requires a sizable eco- diatric traumatic brain injury. Dev. Neurosci. 28, 309–326.nomic effort. Given the observational nature of these ap- Balestreri, M., Czosnyka, M., Hutchinson, P., Steiner, L.A., Hiler,proaches, there is a limited probability that it would be M., Smielewski, P., and Pickard, J.D. (2006). Impact of intra-funded by industry. Governmental agencies are thus the most cranial pressure and cerebral perfusion pressure on severeplausible source of funding. Furthermore, the necessity of disability and mortality after head injury. Neurocrit. Care 4, 8–13.enrolling a large number of clinical centers across different Bischofs, I.B., Hug, J.A., Liu, A.W., Wolf, D.M., and Arkin, A.P.countries would rule out the possibility of the funding coming (2009). Complexity in bacterial cell-cell communication: quo-from a single country. The need for an international collabo- rum signal integration and subpopulation signalling in therative research effort is evident. We argue that this would Bacillus subtilis phosphorelay. Proc. Natl. Acad. Sci. USA 106,offer a much-needed opportunity to provide answers to the 6459–6464.many open questions in TBI research, and improve care for Black, N. (1996). Why we need observational studies to evaluateTBI patients. the effectiveness of health care. BMJ 312, 1215–1218. Brain Trauma Foundation, American Association of Neurologi- cal Surgeons (AANS), Congress of Neurological SurgeonsAcknowledgments (CNS), AANS/CNS Joint Section on Neurotrauma and Cri- tical Care. (2007). Guidelines for the Management of Severe The workshop on CER was supported by the European Traumatic Brain Injury, 3rd ed. J. Neurotrauma 24, 1–106.Commission DG Research and the National Institutes of Brain Trauma Foundation. http:/ /www.tbiguidelines.org. On-Health/National Institute of Neurological Diseases and line TBI guidelines. Accessed May 3, 2011.Stroke (NIH/NINDS). We are grateful to Dr. Ramona Hicks Bryant, R.A. (2008). Disentangling mild traumatic brain injury(NIH/NINDS) and Dr. Patrizia Tosetti (DG Research, Euro- and stress reactions. N Engl. J. Med. 358, 525–527.pean Commission) for their insightful planning and coordi- Bulger, E.M., Nathens, A.B., Rivara, F.P., Moore, M., MacKenzie,nation of the workshop. The authors of this manuscript were E.J., Jurkovich, G.J., and Brain Trauma Foundation. (2002).all speakers at the workshop. We wish to express our grati- Management of severe head injury: Institutional variations intude to Dr. Kenneth Curley, Dr. Alan Faden, and Prof. Nicole care and effect on outcome. Crit. Care Med. 30, 1870–1876.von Steinbuechel for contributing to the panel discussions. We Bullock, R.M., Chesnut, R., Ghajar, J., Gordon, D., Hartl, R.,are grateful to all attendees who participated in the workshop, Newell, D.W., Servadei, F., Walters, B.C., and Wilberger, J.E.many of whom traveled from afar speciﬁcally to attend the (2006). Guidelines for the surgical management of traumaticworkshop, and for their contributions to the general discus- brain injury. Neurosurgery 58, 1–58.
TBI AND COMPARATIVE EFFECTIVENESS RESEARCH 43Carr, B.G., and Nance, M.L. (2010). Access to pediatric trauma principles for the improved conduct of health technology as- care: alignment of providers and health systems. Curr. Opin. sessments for resource allocation decisions. Int. J. Technol. Pediatr. 22, 326–331. Assess. Health Care 24, 244–258; discussion 362–368.Chen, J., Arnow, B.J., and Jegga, A.G. (2009). Disease candidate Edwards, P., Arango, M., Balica, L., Cottingham, R., El-Sayed, gene identiﬁcation and prioritization using protein interaction H., Farrell, B., Fernandes, J., Gogichaisvili, T., Golden, N., networks. BMC Informatics 10, 73. Hartzenberg, B., Husain, M., Ulloa, M.I., Jerbi, Z., Khamis, H.,Chesnut, R.M., Marshall, L.F., Klauber, M.R., Blunt, B.A., Bald- ¨ Komolafe, E., Laloe, V., Lomas, G., Ludwig, S., Mazairac, G., win, N., Eisenberg, H.M., Jane, J.A., Marmarou, A., and ˜ ´ Munoz Sanchez Mde, L., Nasi, L., Olldashi, F., Plunkett, P., Foulkes, M.A. (1993) The role of secondary brain injury in Roberts, I., Sandercock, P., Shakur, H., Soler, C., Stocker, R., determining outcome from severe head injury. J. Trauma 34, Svoboda, P., Trenkler, S., Venkataramana, N.K., Wasserberg, 216–222. J., Yates, D., and Yutthakasemsunt, S. MRC CRASH trialClayton, T.J., Nelson, R.J., and Manara, A.R. (2004). Reduction in collaborators. (2005). Final results of MRC CRASH, a rando- mortality from severe head injury following introduction of a mised placebo-controlled trial of intravenous corticosteroid in protocol for intensive care management. Br. J. Anaesth. 93, adults with head injury—outcomes at 6 months. Lancet 365, 761–767 1957–1959.Clifton, G.L., Choi, S.C., Miller, E.R., Levin, H.S., Smith, K.R. Jr., Elf, K., Nilsson, P., and Enblad, P. (2002). Outcome after trau- Muizelaar, J.P., Wagner, F.C. Jr., Marion, D.W., and Luerssen, matic brain injury improved by an organized secondary insult T.G. (2001). Intercenter variance in clinical trials of head program and standardized neurointensive care. Crit. Care trauma—experience of the National Acute Brain Injury Study: Med. 30, 2129–2134. Hypothermia. J. Neurosurg. 95, 751–755. Fakhry, S.M., Trask, A.L., Waller, M.A., Watts, D.D., and IRTCColes, J.P., Fryer, T.D., Coleman, M.R., Smielewski, P., Gupta, Neurotrauma Task Force. (2004). Management of brain- A.K., Minhas, P.S., Aigbirhio, F., Chatﬁeld, D.A., Williams, injured patients by an evidence-based medicine protocol G.B., Boniface, S., Carpenter, T.A., Clark, J.C., Pickard, J.D., improves outcomes and decreases hospital charges. J. Trauma and Menon, D.K. (2007). Hyperventilation following head 56, 492–499. injury: effect on ischemic burden and cerebral oxidative me- Forsyth, R.J., Parslow, R.C., Tasker, R.C., Hawley, C.A., and tabolism. Crit. Care Med. 35, 568–578. Morris, K.P. (2008). Prediction of raised intracranial pressureColes, J.P., Minhas, P.S., Fryer, T.D., Smielewski, P., Aigbirihio, complicating severe traumatic brain injury in children: impli- F., Donovan, T., Downey, S.P., Williams, G., Chatﬁeld, D., cations for trial design. Pediatr. Crit. Care Med. 9, 8–14. Matthews, J.C., Gupta, A.K., Carpenter, T.A., Clark, J.C., Foulkes, M.A., Eisenberg, H.M., Jane, J.A., Marmarou, A., Mar- Pickard, J.D., and Menon, D.K. (2002). Effect of hyperventila- shall, L.F., and the TCDB research group. (1991). The trau- tion on cerebral blood ﬂow in traumatic head injury: clinical matic coma data bank: design, methods, and baseline relevance and monitoring correlates. Crit. Care Med. 30, 1950– characteristics. J. Neurosurg. 75, 8–13. 1959. Gelpke, G.J., Braakman, R., Habbema, J.D., and Hilden, J. (1983).Colohan, A.R., Alves, W.M., Gross, C.R., Torner, J.C., Mehta, Comparison of outcome in two series of patients with severe V.S., Tandon, P.N., and Jane, J.A. (1989). Head injury mortality head injuries. J. Neurosurg. 59, 745–750. in two centers with different emergency medical services and Govind, V., Gold, S., Kaliannan, K., Saigal, G., Falcone, S., Ar- intensive care. J. Neurosurg. 71, 202–207. heart, K.L., Harris, L., Jagid, J., and Maudsley, A.A. (2010).Cooper, D.J., Rosenfeld, J.V., Murray, L., Arabi, Y.M., Davies, Whole-brain proton MR spectroscopic imaging of mild- A.R., D’Urso, P., Kossmann, T., Ponsford, J., Seppelt, I., Reilly, to-moderate traumatic brain injury and correlation with P., Wolfe, R., the DECRA Trial Investigators and the Austra- neuropsychological deﬁcits. J. Neurotrauma 27, 483–496. lian and New Zealand Intensive Care Society Clinical Trials Green, A.R. (2002). Why do neuroprotective drugs that are so Group. (2011) Decompressive craniectomy in diffuse trau- promising in animals fail in the clinic? An industry perspec- matic brain injury. N. Engl. J. Med. 364, 1493–1502. tive. Clin. Exp. Pharmacol. Physiol. 29, 1030–1034.Correspondence NEJM (2009). Care of war veterans with mild Harders, A., Kakarieka, A., and Braakman, R. (1996). Traumatic traumatic brain injury. N. Engl. J. Med. 361, 536–538. subarachnoid hemorrhage and its treatment with nimodipine.Cremer, O.L., van Dijk, G.W., van Wensen, E., Brekelmans, G.J., German tSAH Study Group. J. Neurosurg. 85, 82–89. Moons, K.G., Leenen, L.P., and Kalkman, C.J. (2005). Effect of Hlatky, R., Contant, C.F., Diaz-Marchan, P., Valadka, A.B., and intracranial pressure monitoring and targeted intensive care Robertson, C.S. (2004). Signiﬁcance of a reduced cerebral on functional outcome after severe head injury. Crit. Care blood ﬂow during the ﬁrst 12 hours after traumatic brain in- Med. 33, 2207–2213. jury. Neurocrit. Care 1, 69–83.Czosnyka, M., Balestreri, M., Steiner, L., Smielewski, P., Hoge, C.W., Goldberg, H.M., and Castro, C.A. (2009). Care of Hutchinson, P.J., Matta, B., and Pickard, J.D. (2005). Age, in- war veterans with mild traumatic brain injury. N. Engl. J. tracranial pressure, autoregulation, and outcome after brain Med. 360, 1588–1591. trauma. J. Neurosurg. 102, 450–454. Hoge, C.W., McGurk, D., Thomas, J.L., Cox, A.L., Engel, C.C.,DeWitt, D.S., Jenkins, L.W., and Prough, D.S. (1995). Enhanced and Castro, C.A.V. (2008). Mild traumatic brain injury in U.S. vulnerability to secondary ischemic insults after experimental soldiers returning from Iraq. N. Engl. J. Med. 358, 453–463. traumatic brain injury. New Horiz. 3, 376–383. Howells, T., Elf, K., Jones, P.A., Ronne-Engstrom, E., Piper, I.,Dhruva, S.S., and Redberg, R.F. (2008). Variations between and Nilsson, P. (2005). Pressure reactivity as a guide in the clinical trial participants and Medicare beneﬁciaries in evi- treatment of cerebral perfusion pressure in patients with brain dence used for Medicare national coverage decisions. Arch. trauma. J. Neurosurg. 102, 311–317 Intern. Med. 168, 136–140. Erratum, Arch. Intern. Med. 168, http://www.cdc.gov/traumaticbraininjury/tbi_ed.html. Trau- 774. matic brain injury in the United States: emergency department ¨Drummond, M.F., Schwartz, J.S., Jonsson, B., Luce, B.R., visits, hospitalisations and visits 2002–2006. Accessed May 3, Neumann, P.J., Siebert, U., and Sullivan, S.D. (2008). Key 2011.
44 MAAS ET AL.https://webgate.ec.europa.eu/idb/documents/2009-IDB-Report_ vere head injury in adults. European Brain Injury Consortium. screen.pdf. Injuries in the European Union: Statistic summary Acta Neurochir. (Wien.) 139, 286–294. 2005–2007. Accessed May 3, 2011. Maas, A.I., Harrison-Felix, C.L., Menon, D., Adelson, P.D.,Ikonomidou, C., and Turski, L. (2002). Why did NMDA receptor Balkin, T., Bullock, R., Engel, D.C., Gordon, W., Langlois- antagonists fail clinical trials for stroke and traumatic brain Orman, J., Lew, H.L., Robertson, C., Temkin, N., Valadka, A., injury? Lancet Neurol. 1, 383–386. Verfaellie, M., Wainwright, M., Wright, D.W., and Schwab, K.IOM (Institute of Medicine). (2009). Initial National Priorities for (2011) Standardizing data collection in traumatic brain injury. Comparative Effectiveness Research. The National Academies J. Neurotrauma 28, 177–187. Press: Washington, D.C. Maas, A.I., Harrison-Felix, C.L., Menon, D.K., Adelson, D.,Ishige, N., Pitts, L.H., Berry, I., Nishimura, M.C., and James, T.L. Balkin, T., Bullock, R., Engel, D.C., Gordon, W., Langlois, J., (1988). The effects of hypovolemic hypotension on high- Lew, H.L., Robertson, C., Temkin, N., Valadka, A., Verfaellie, energy phosphate metabolism of traumatized brain in rats. M., Wainwright, M., Wright, D.W., and Schwab, K. (2010c) J. Neurosurg. 68, 129–136. Common data elements for traumatic brain injury: Re-Janowitz, T., and Menon, D.K. (2010). Exploring new routes for commendations from the Interagency Working Group on neuroprotective drug development in traumatic brain injury. Demographics and Clinical Assessment. Arch. Phys. Rehab. Sci. Transl. Med. 2, 27 rv1. Med. 91, 1641–1649.Jennett, B., Teasdale, G., Braakman, R., Minderhoud, J., and Maas, A.I., Marmarou, A., Murray, G.D., Teasdale, G.M., and Knill-Jones, R. (1976). Predicting outcome in individual pa- Steyerberg, E.W. (2007a). Prognosis and clinical trial design in tients after severe head injury. Lancet 1, 1031–1034. traumatic brain injury: the IMPACT Study. J. Neurotrauma 24,Jennett, B., Teasdale, G., Galbraith, S., Pickard, J., Grant, H., 232–238. Braakman, R., Avezaat, C., Maas, A., Minderhoud, J., Vecht, Maas, A.I., Roozenbeek, B., and Manley, G.T. (2010a). Clinical C.J., Heiden, J., Small, R., Caton, W., and Kurze, T. (1977). trials in traumatic brain injury: Past experience and current Severe head injuries in three countries. J. Neurol. Neurosurg. developments. Neurotherapeutics 7, 115–126. Psychiatry 40, 291–298. Maas, A.I. (2009). Standardisation of data collection in traumaticJennett, B., Teasdale, G., Fry, J., Braakman, R., Minderhoud, J., brain injury: key to the future? Crit. Care. 13, 1016. Heiden, J., and Kurze, T. (1980). Treatment for severe head Maas, A.I., Steyerberg, E.W., Butcher, I., Dammers, R., Lu, J., injury. J. Neurol. Neurosurg. Psychiatry 43, 289–295. Marmarou, A., Mushkudiani, N.A., McHugh, G.S., and Mur-Jordan, B.D. (2007). Genetic inﬂuences on outcome following ray, G.D. (2007b). Prognostic value of computerized tomog- traumatic brain injury. Neurochem. Res. 4–5, 905–915. raphy scan characteristics in traumatic brain injury: resultsKapapa, T., Pﬁster, U., Konig, K., Sasse, M., Woischneck, D., from the IMPACT study. J. Neurotrauma 24, 303–314. Heissler, H.E., and Rickels, E. (2010). Head trauma in children, Maas, A.I., Steyerberg, E.W., Marmarou, A., McHugh, G.S., part 3: clinical and psychosocial outcome after head trauma in Lingsma, H.F., Butcher, I., Lu, J., Weir, J., Roozenbeek, B., and children. J. Child Neurol. 25, 409–422. Murray, G.D. (2010b). IMPACT recommendations for im-Kochanek, P.M., Clark, R., Ruppel, R.A., Adelson, P.D., Bell, proving the design and analysis of clinical trials in moderate M.J., Whalen, M.J., Robertson, C.L., Satchell, M.A., Seidberg, to severe traumatic brain injury. Neurotherapeutics 7, 127–134. N.A., Marion, D.W., and Jenkins, L.W. (2000). Biochemical, Maas, A.I., Steyerberg, E.W., Murray, G.D., Bullock, R., Baeth- cellular, and molecular mechanisms in the evolution of sec- mann, A., Marshall, L.F., and Teasdale, G.M. (1999). Why ondary damage after severe traumatic brain injury in infants have recent trials of neuroprotective agents in head injury and children: Lessons learned from the bedside. Pediatr. Crit. failed to show convincing efﬁcacy? A pragmatic analysis and Care Med. 1, 4–19. theoretical considerations. Neurosurgery 44, 1286–1298.Kochanek, P.M. (2006). Pediatric traumatic brain injury: quo Maas, A.I., Stocchetti, N., and Bullock, R. (2008). Moderate vadis? Dev. Neurosci. 28, 244–255. and severe traumatic brain injury in adults. Lancet Neurol. 7,Langlois, J.A., Rutland-Brown, W., and Thomas, K.E. (2005). The 728–741. incidence of traumatic brain injury among children in the Manley, G., Knudson, M.M., Morabito, D., Damron, S., Erickson, United States: differences by race. J. Head Trauma Rehabil. 20, V., and Pitts, L. (2001). Hypotension, hypoxia, and head in- 229–238. jury: frequency, duration, and consequences. Arch. Surg. 136,Langlois, J.A., Rutland-Brown, W., and Thomas, K.E. (2006). 1118–1123. Traumatic brain injury in the United States: emergency de- Marklund, N., Bakshi, A., Castelbuono, D.J., Conte, V., and partment visits, hospitalizations, and deaths. Centers for McIntosh, T.K. (2006). Evaluation of pharmacological treat- Disease Control and Prevention, National Center for Injury ment strategies in traumatic brain injury. Curr. Pharm. Des. Prevention and Control: Atlanta. 12, 1645–1680.Lingsma, H.F., Roozenbeek, B., Li, B., Marmarou, A., Murray, Marmarou, A., Lu, J., Butcher, I., McHugh, G.S., Mushkudiani, G.D., Maas, A.I., and Steyerberg, E.W. (2011). Large between- N.A., Murray, G.D., Steyerberg, E.W., and Maas, A.I. (2007). center differences in outcome after moderate and severe IMPACT database of traumatic brain injury: design and de- traumatic brain injury in the IMPACT study. Neurosurgery scription. J. Neurotrauma 24, 239–250. 68, 601–607; discussion 607–608. Menon, D.K., Coles, J.P., Gupta, A.K., Fryer, T.D., Smielewski,Lingsma, H.F., Roozenbeek, B., Steyerberg, E.W., Murray, G.D., P., Chatﬁeld, D.A., Aigbirhio, F., Skepper, J.N., Minhas, P.S., and Maas, A.I. (2010). Early prognosis in traumatic brain in- Hutchinson, P.J., Carpenter, T.A., Clark, J.C., and Pickard, J.D. jury: from prophecies to predictions. Lancet Neurol. 9, 543– (2004). Diffusion limited oxygen delivery following head in- 554. jury. Crit. Care Med. 32, 1384–1390.Maas, A.I., Dearden, M., Teasdale, G.M., Braakman, R., MRC CRASH Trial Collaborators, Perel, P., Arango, M., Clayton, Cohadon, F., Iannotti, F., Karimi, A., Lapierre, F., Murray, G., T., Edwards, P., Komolafe, E., Poccock, S., Roberts, I., Shakur, Ohman, J., Persson, L., Servadei, F., Stocchetti, N., and Un- H., Steyerberg, E., and Yutthakasemsunt, S. (2008). Predicting terberg, A. (1997). EBIC—guidelines for management of se- outcome after traumatic brain injury: practical prognostic