In CENTER-TBI we applied advanced statistical and neuroinformatics approaches to better understand
the complex interrelationships of different measures for injury severity (clinical, structural,
biochemical, prognostics, genetic, etc.). Novel approaches (genetic risk stratification, advanced
imaging, and emerging biomarkers) offer substantial gains in
and therapy stratification, but clinical experience and validation in TBI is limited.
Computing platform and Neuroinformatics Resource
(International Neuroinformatics Coordinating Facility)
coordinated the establishment of an informatics platform for acquisition, storage and analysis of
CDE-based clinical data. The goal was to develop a next generation open standards-based platform to
support advanced large-scale analytics and model building.
Such a platform also provides a model for future clinical studies on brain diseases and disorders.
This common open international platform is a resource for future TBI studies. The resulting analytical
environment provides large-scale data access, enabling ‘big data’ science that supports this and other
work packages, and allowing large-scale data analysis and modelling to be leveraged for TBI questions
in ways that have not previously been possible. Comprehensive computational models for TBI were derived
from this approach.
Several studies report genetic associations with TBI outcome, but their relevance is limited by small
sample size and by unsophisticated analysis, which takes little account of predicted outcome. Further
conventional common disease-common variant (CDCV) approaches have problems and serve TBI poorly, since
the admission phenotype is heavily influenced by injury exposure, rather than host response (which is
the substrate for genetic modulation). Finally, such approaches take no account of pre-test
probabilities of significant gene-phenotype relationships based on known disease biology, clustering of
gene effects in biological pathways or rich serial measures of phenotype routinely recorded in
intensively monitored TBI subjects.
Conventional MRI is widely thought to be superior to CT for characterising TBI pathology. However, a
recent study only compared late (~2 week) MRI with CT performed upon initial presentation, and rigorous
comparisons of contemporaneously collected CT and MRI are lacking. Such studies are essential for
establishing the role of MRI in routine imaging, especially of mild TBI, in particular as there is good
evidence that MR visibility of lesions in TBI changes with time. Advanced MRI (Diffusion Tensor Imaging
[DTI], Resting state fMRI [rs fMRI], and susceptibility weighted imaging [SWI]) offer opportunities for
better detection, characterization and quantification of injury. SWI and DTI abnormalities consistent
with traumatic axonal injury occur in 30% of patients without abnormalities on CT and DTI imaging
provides exquisite characterisation of lesion evolution, and delivers important pathophysiological
insights. In the subacute phase MR imaging offers a great potential for understanding and tracking
disease processes, prognosticating cognitive recovery, and mapping covert cognition in disorders of
consciousness. In the chronic phase, the ability to relate neuronal loss and white matter injury to
cognitive measures, to vulnerability for depression and to functional outcome may be key to
understanding impairment following injury and to the development of strategies for rehabilitation which
may be tailored to patients’ needs. Past studies have poorly addressed the changes in structural and
functional connectivity that underpins these deficits, and the full potential of advanced MR imaging in
TBI has not been ascertained due to small sample sizes (the largest study involves less than 150
patients), selection bias and inconsistent use of advanced imaging.
Blood biomarkers have recently received much attention for a possible role in diagnosing mild TBI,
tracking disease progression, and predicting outcome in TBI, but studies have generally involved small
patient numbers, with no rigorous assessment of the added diagnostic or predictive power over and above
that of established techniques/models in appropriately large patient numbers with good quality acute
data collection and outcome documentation. Further, there is no evidence that these approaches can make
the key transition from technically demanding research tools to robust clinical assets that can be used
in everyday practice. There is a substantial need for well-established laboratory markers that can be
used for diagnosis, disease monitoring and prediction in TBI, replicating (for example) the use of
troponin in acute myocardial infarction. Candidate biomarkers which we will initially target include
GFAP, S100, NSE and UCHL-1.
Comparative effectiveness research (CER)
Comparative effectiveness research provides a promising framework to identify best practices and improve
outcome after TBI. CER is the generation and synthesis of evidence that compares the benefits and harms
of alternative methods to prevent, diagnose, treat, and monitor a clinical condition or to improve the
delivery of care. The purpose of CER is to assist consumers, clinicians, purchasers, and policy makers
to make informed decisions that will improve health care at both the individual and population levels.
Until recently randomized controlled trials (RCTs) were considered the gold standard for investigating
benefit of interventions. RCTs generally employ strict enrolment criteria in order to study the
investigational intervention in the "cleanest" setting. The downside of this approach is that results
are only valid in such selected subpopulations and that generalizability to the real world context is
limited. Moreover, trials are expensive and in the heterogeneous field of TBI results have been
disappointing. A basic concept of CER is to study differences in care and outcome in observational
studies, thus turning natural variability into an asset. In CENTER-TBI, we will exploit the existing
heterogeneity in structure, process and outcome to compare treatments and interventions that are
standard practice in some centres and countries but not in others. Natural links exist between CER and
individualized approaches, since CER aims to identify the best treatment for the individual patient,
with a specific type of injury, severity, co-morbidities and other aspects that determine optimal
treatment. We see a great potential for CER in TBI because of various unique features: First, there are
large between-centre and between-country differences in both outcome and management. Second, robust risk
adjustment models have been developed specifically for TBI, providing the possibility to adjust for
patient characteristics that affect outcome. Third, advanced statistical models, including random effect
models, are available to analyze differences between centres.
Outcome after TBI is complex and multidimensional, and includes functional status, generic and
disease-specific health related quality of life (HRQoL),
cognitive performance, and emotional and psychosocial adjustment. CENTER-TBI examined the time
course, correlated predictors and moderators of different aspects of outcome, explored interdependencies
and tailor outcome assessment after TBI. Specific topics addressed include the relative applicability to
different (severity) strata, the incidence and development of PTSD in the three strata, relationships
between outcome and TBI characterization (MRI, biomarkers, sociodemographic and psycho-social data,
etc.). We aimed to develop a multidimensional outcome classification.
Full assessments, including neuropsychological testing, were performed as a cross-sectional study at
the primary outcome time points (3 months for ER stratum, 6 months for admission and ICU strata).
Furthermore, all patients undergoing longitudinal MRI studies received repeated neuropsychological
testing. Longitudinal follow up across multiple time points is based upon the use of an abbreviated
subset of assessments, which could be administered by postal questionnaire or web-based completion.