Michael Weinrich, M.D., NICHD, NIH, DHHS, noted that this topic is very complex, and he listed several issues for participants to consider during their discussion:
- Incidence/prevalence. Mild traumatic brain injury (MTBI) is a major public health problem, each year affecting up to 1.5 million children in the United States.
- Case identification. Many of these injuries are inflicted, not accidental, and most such injuries go unreported. He asked how recollection bias could be minimized.
- Selection bias. One common event is the failure to recognize that any sequelae are associated with an event that wasn’t initially recognized as a significant injury.
- Definition of MTBI. Loss of consciousness (LoC) has been one of the primary requisites for classifying MTBI, but not all events that belong in the category involve LoC.
- Persistence of MTBI. Subsequent impacts may be more likely to occur, and the effects can be cumulative.
Dr. Weinrich added that MTBI poses some special issues for children:
- How does MTBI affect developmental trajectories?
- Are certain populations more susceptible biologically/genetically to experience MTBI and/or have more severe effects if they do?
- Frontal lobe injury is known to affect behavior, and thus, once incurred, may make a child more likely to experience subsequent MTBI.
Introductions and Background, Purpose, and Objectives for the Meeting
Dr. Langlois cited CDC epidemiological statistics, which showed that among children 0-14 years old, there are 2,600 deaths, 39,000 hospitalizations, and 400,000 visits to emergency departments (ED) each year. The 10:1 ratio of ED visits to hospitalizations does not apply to adult populations. Further, many cases of MTBI treated by primary care providers never get reported as such. Dr. Langlois emphasized that the workshop’s primary focus was to discuss the incidence, adverse outcomes, and measurements of MTBI.
Peter C. Scheidt, M.D., M.P.H., NICHD, NIH, DHHS, provided an overview of the National Children’s Study.
Topic 1: Identification of Mild TBI
Barry S. Willer, Ph.D., University of Buffalo, presented background information about MTBI, ascertaining cases, and reporting (prospective versus retrospective). He explained that there is no commonly accepted and operational definition of MTBI. Common medical practice, rehabilitation medicine, and the DSM-IV all use different standards. The Glasgow Coma Scale (GCS) brings some uniformity to the issue, but there is no consensus on what measures are best. Dr. Willer presented a number of considerations for the Study:
- The sample size of 100,000 is probably inadequate for deriving a reliable incidence rate of MTBI/concussion.
- Useful criteria must be developed and standardized to ascertain cases reliably and to avoid a multitude of minor events that are not appropriate for the Study.
- An achievable goal would be to determine factors and symptoms (for example, post-concussion syndrome) that predict long-term problems, such as interference with learning.
- Another feasible target would be to assess the mechanics and causes that lead to MTBI.
Robert M. Reece, M.D., Tufts University School of Medicine, described neurological injury inflicted on infants by others, commonly known as shaken baby syndrome (SBS) or shaken impact syndrome (SIS). Craniocerebral trauma is the predominant problem among infants and the most common cause of morbidity and mortality in this age group. In various studies, 50 percent or more of such injuries were inflicted. The younger the child, the more likely the injury is due to abuse. Dr. Reece described SBS in greater detail, including its pathophysiology and clinical results, associated injuries, and the continuum from mild to moderate to severe or fatal. A substantial numbers of victimized children go undiagnosed, for a number of reasons. Dr. Reece emphasized the need for recognition and reporting, each of which is an essential component in dealing with SBS, and incorporating them in the Study.
Dr. Langlois summarized the issues raised about Topic 1. She noted that the key challenge is going to be how to devise a system for incident reporting that effectively covers the full range of traumatic brain injuries:
- Ways to do that. Pilot studies.
- Scope. How to capture the full spectrum of injuries and obtain meaningful details about the mechanisms of the injury.
- Challenges. Getting parents of very young children to provide useful and timely information.
- Confounder issues. Inflicted TBI is an important and seriously underreported event, which can be a challenge to diagnose. Follow-up should last at least five years. Biomarkers could be useful, if primarily suggestive, for this subgroup. Perhaps instead of a definitive diagnosis, the Study could develop a catalog of probabilities, associations, and cofactors.
- May not fit the Study. Some issues (hyperthermia and maternal stress) may be outside the scope of the Study, but should be identified as important to the field.
Topic 2: Early Clinical Assessment and Diagnosis
The first presentation was by Jay Giedd, M.D., National Institute of Mental Health, NIH, DHHS, who presented an overview of his work using the latest neuroimaging techniques—primarily functional magnetic resonance imaging (fMRI)—to evaluate brain development in children. While relatively expensive, fMRI does not emit radiation and is safe to use repeatedly over time, even in infants and children of all ages and conditions. This would seem to be a good fit for a longitudinal study in the minimal risk category.
Mark R. Lovell, Ph.D., discussed concussion, which he has studied in a variety of sports settings, especially football. In sports settings, brain trauma is usually recognized and responded to right away. Dr. Lovell uses ImPACT, a computer-based test, to measure reaction time, working memory, and other neuropsychological domains. In Dr. Lovell’s program, the test is administered before the onset of a sports season to capture baseline data. These data are then compared to how an athlete performs at various times post-concussion. Whenever a concussion is suspected, team trainers immediately begin to evaluate memory and function. Functional MRI scans are quickly obtained (1-2 days) for evaluation, and at 5-8 days for follow-up.
Gerard A. Gioia, Ph.D., Children’s National Medical Center, Washington, DC, discussed executive function (EF), which orchestrates basic cognitive processes such as inhibiting or deferring a response, strategically planning a sequence of actions, and using working memory. Dr. Gioia and his colleagues have developed an instrument called the Behavior Rating Inventory of Executive Function (BRIEF), which assesses nine discrete areas of EF in children aged 2 to 18. Dr. Gioia suggested that BRIEF could be used to correlate long-term outcomes of EF with MTBI.
David A. Hovda, Ph.D., UCLA School of Medicine, gave a presentation about the pathophysiology of the injured brain and the effects on metabolism and development. A number of principles apply to planning MTBI research in the Study:
- Mismatch occurs when the cerebral blood flow (CBF) fails to correspond to the body’s metabolism.
- Uncoupling occurs when stimulation fails to evoke a dynamic CBF response in an injured person.
- Adenosine triphosphate (ATP) is the brain’s universal energy. TBI causes hyperglycolosis, which induces a loss of ATP and may starve and eventually kill brain tissue. Even a mild TBI throws brain cells into an energy crisis.
- Energy failure can occur even without classic ischemia.
Rachel Berger, M.D., M.P.H., Children’s Hospital of Pittsburgh, defined biomarkers as biochemicals that are found in the brain after a wide variety of insults (drugs, oxygen deprivation, and others), including TBI. These chemicals are released from the brain into the cerebrospinal fluid (CSF) and move into the blood serum, where they can be detected and measured. Dr. Berger focused her discussion on closed head injury (CHI), and she suggested that biomarkers could provide a wealth of useful information for the Study.
- Neuron-specific enolase (NSE). Measurable in both the CSF and the serum, the ELISA test can be done with only 2½ drops of serum, and can be completed in less than one hour.
- S100B is a growth factor implicated in neuroplasticity. It is comparably convenient to NSE but is not affected by hemolysis. However, its half-life is much shorter, and it disappears six to eight hours after the injury.
Dr. Langlois summarized discussion points raised by participants about neuroimaging:
- CTs are not useful. Quantitative MRI would be an optimum baseline, but it may not be feasible within the entire Study population. Are there useful pre-injury markers of vulnerability (for example, athletes or conditions such as ADHD) that would help identify subpopulations?
- It is crucial to get an image as soon as possible after the injury (ideally within a day or two).
- It is necessary to determine how often to repeat the scan (at least every five years). Scanning children under age 8 is problematic because of the possible need for sedation and the need to keep the protocol in the minimal risk category.
- Because of rapidly changing technology, it will be difficult to plan for inevitable advances in imaging over the course of the Study.
Dr. Langlois’ summary of the presentations on acute diagnosis and follow-up included:
- How to get uniform and immediate assessment and evaluation of the injury is a challenge because it is often difficult to recognize or—in the case of inflicted injuries—report TBI. The diversity of treatment centers, EDs, and other caregivers make developing a uniform reporting system a challenge.
- It is necessary to have an initial assessment, ideally at the time of the event (for example, by a trained person at a sports event), but at least within 48 hours. There is also the need for longitudinal tracking.
- Methods are crucial to good retention. Going into the home has proven doable, as have certain computer-based instruments and procedures (the BRIEF, for example). In most models, a strong, centralized national or regional data collection center is essential.
- Separate methods, systems, and assessments are needed for the very young.
The material presented about metabolism and biomarkers was summarized by Dr. Langlois:
- Assessing the full range (impoverished to enriched) of environmental characteristics is essential.
- Measuring the rate of acceleration in performance on standardized test scores could indicate a change in trajectories.
- Obtaining baseline measures is essential. Beyond cord blood and early pediatric procedures (circumcision), it will be necessary to decide if the Study will collect a serum sample from all subjects or from a subsample. Other considerations include when to get samples and how much to store for all possible future uses.
- Collecting blood within six hours after an injury is important. Study designers will still need to deal with the overarching issue of case finding.
Topic 3: Assessment and Follow-Up of Longer Term Outcomes
Keith Owen Yeates, Ph.D., Ohio State University, said that he preferred the term mild closed-head injury (CHI). Despite the appearance of frequent and persistent post-concussive symptoms (PCS) in some children, it has been demonstrated that no long-term cognitive changes occur with this class of injury. Controversy also surrounds whether the underlying cause of such symptoms is psychological or physiological, when it could be some combination of both. The confusion begins with the DSM-IV proposed criteria for post-concussional disorder, which uses psychiatric nomenclature while assuming an underlying pathophysiology. In contrast, the ICD-10 criteria are more directly psychological and functional. He said that some of the issues for the Study are defining CHI, comparing groups, evaluating developmental issues, measuring outcomes, timing assessments, and predicting variability within groups.
Robert Asarnow, Ph.D., University of California, Los Angeles, discussed his background paper that provided a comprehensive summary of the 40 English-language studies on outcomes of MTBI in children. He explained that many of the studies were flawed, and they used operational definitions that varied. Only five studies found cognitive impairments, and those effects were small and transitory. He then pointed out that most studies failed to control for pre-injury level of functioning. The most common neurobehavioral sequelae of MBTI are attentional and memory impairments, behavior problems, and psychiatric disorders. He added that it would be useful to determine which of these are risk factors for accidents in general and TBI in particular, and then determine their prevalence in the general population.
Dr. Asarnow suggested that the Study should develop measures to assess:
- The effect of MTBI on cognitive abilities. Focus on brain structures in the frontal and temporal lobe that are most vulnerable to TBI and develop age-appropriate measures with good test/retest reliability over the entire range of the Study.
- Neuropsychological factors. Verbal learning and fluency, higher order language function (pragmatics), secondary memory for both verbal and non-verbal material, and executive function/problem solving.
- Academic achievement. Sample contemporary curricula by region and develop norms with good test/retest reliability.
Vicki Anderson, Ph.D., University of Melbourne, Australia, presented her submitted paper, Social and Emotional Function in Children Following Traumatic Brain Injury. She suggested that the Study recognize:
- It is crucial to embrace social domains, where profound and lasting effects after MTBI are known to occur.
- Behavioral problems can manifest as distractibility, anxiety, hyperactivity, depression, aggression, poor impulse control, and a lack of emotional expression and perception.
- Social issues are compromised or reduced social information processing, self-awareness, self-esteem, social contacts, and judgment.
- Important domains to capture in a Study assessment are pre-injury function; physical and cognitive disabilities; injury-related factors, such as PTSD or other responses to the accident; other injuries; other family members’ injuries; how the family is affected; post-concussive physical symptoms; and access to appropriate educational and other social resources.
Dr. Langois summarized Topic 3 discussions:
- Include a comparison group of injured subjects.
- Whenever possible, devise instruments and assessments that ask the child about symptoms; data are degraded by relying on the family exclusively.
- Determine in advance how data will be analyzed as a critical step in deciding what and how to collect those data. Utilize new methods of analysis that can separate group and individual effects.
- Evaluate new areas or research and ways to assess post-traumatic symptoms in general; family factors assume high importance with this group.
She added that the Study is not only a research and scientific endeavor; it also could potentially answer important public health policy questions. Population-based data must be representative (for example, how many children are affected and to what degree). TBI is acknowledged as a leading cause of disability in children. The Study provides an opportunity to actually specify the magnitude of the problem. Additionally, the Study can make public health history by developing solid information on the prevalence of adverse outcomes from TBI:
- Go beyond pathology to look at deviation from normal or from previous trajectories. Many of the problems and sequelae can be subtle at first, and then lead to more evident behavioral problems over time. Sorting out longer-term effects could revolutionize the framing of the social problem and its treatment.
- Assess the history of medical condition and include some measures that may not be picked up for other parts of the Study (such as previous seizure disorder).
- Try to assess post-concussive symptoms as soon as possible. Time subsequent assessments to match standard transitions in social and intellectual development, with long-term analysis.
Problems may be subtle, appear only under stress, and may require new measures to solidify the data for longitudinal epidemiological conclusions. Consider collecting data on utilization of education resources, medical rehabilitation services, and gaps in service that people need but don’t receive.
Topic 4: Factors Influencing Outcomes
Ramon Diaz-Arrastia, M.D., Ph.D., University of Texas Southwestern Medical School, discussed genetic factors that affect outcome after MTBI. Animal studies have identified one particular polymorphism, apolipoprotein E (Apo εE-4), which is associated with poor outcome after TBI. Another, tumor necrosis factor-ą (TNF-ą), is elevated in the plasma and CSF of people with TBI, and, when counteracted by a binding protein, results in smaller lesions. A number of TNF antagonists are currently in clinical use, illustrating an example of how genetic information can translate into therapeutics fairly quickly. Because the promise of such research is so great, Dr. Diaz-Arrastia proposed that the Study establish a DNA bank. This would require researchers to collect clinical information, obtain sufficient DNA, develop genotyping assays, and devise hypotheses to probe phenotype-genotype links.
Summary
Keith Owen Yeates, Ph.D., Children’s Hospital, Ohio State University, summarized the issues and questions that emerged during the two days of the workshop. He said that the Study faces three primary issues in its TBI protocols:
- Definition and case finding. The major issue is that no good definition exists that can be adopted (or adapted) by the Study. The definitions that are currently in use derive from school-age studies and would not properly cover younger Study participants. Studying these groups has inherent problems (variability, subjectivity, and/or private motivations, for example) due to parental reporting. Inflicted injuries are much more likely in these groups, as well. It is very difficult to define the mild-to-moderate boundary, as well as the actual mechanism, especially with inflicted TBI.
- What happened (assessment of injury). Regardless of the strategies used to evaluate the injury (neuroimaging, biomarkers, brain metabolism, the biomechanical cause and experience, and signs and symptoms), the developmental stage/age of the child has an underlying effect that must drive which measure is used and how comparisons are made.
- What was the effect of what happened (assessment of outcomes). Each of four different domains has unique challenges. Cognitive functioning will have to be tailored to the age and stage of the participant at the time of the assessment. Post-concussive syndromes need to be recognized and woven into the assessment. Social function is a vital and underappreciated domain. As with cognitive functioning, it must be assessed against a background of normal developmental stages. Quality of life deserves a distinct effort, despite the difficulties in assessing younger populations.
Researchers will need to examine such factors as injury versus non-injury, the child versus the environment, and factors before and after the injury. None of these is likely to operate independently of the others, and so interactions must be accommodated. Finally, developmental variation over the age range must be factored into these assessments.
Dr. Langlois said that several take-home messages emerged during the workshop:
- The Study needs to address the full spectrum of TBI, from mild through moderate to severe. It is important to know the prevalence of adverse health outcomes in the full population of TBI in children.
- It is useless to pose valid questions about outcomes without first answering the question: Outcomes of what? Without some kind of STAT system, there is no chance to verify the clinical event with biomarkers, take useful images, and make cognitive assessments that are essential for later developmental comparisons. ED records for verification are much more likely to yield useful data if tapped within a week or two of the event. Recall bias will be a constant challenge. Recall bias is a threshold issue; who reports an injury and what they report about it will change over time. The incident report can trigger a number of confirmatory assessments, as well as post-injury baselines for structure and function.
- This component of the Study will succeed to the extent that parents and others become actively engaged in the issue. When long-term relationships are cultivated, subjects tend to self-report more reliably. Retention strategies must be ongoing and continually revised for success.
Participants
Vicki Anderson, Ph.D., Royal Children’s Hospital, University of Melbourne, Australia
Robert Asarnow, Ph.D., School of Medicine, University of California, Los Angeles
Adelaide Barnes, NICHD, NIH, DHHS
Rachel Berger, M.D., M.P.H., Children’s Hospital of Pittsburgh, University of Pittsburgh
Ruth A. Brenner, M.D., M.P.HNICHD, NIH, DHHS
Ramon Diaz-Arrastia, M.D., Ph.D., University of Texas Southwestern Medical School
Jay N. Giedd, M.D., National Institute of Mental Health, NIH, DHHS
Gerard A. Gioia, Ph.D., Children’s National Medical Center
David A. Hovda, Ph.D., University of California, Los Angeles
Jean Langlois, Sc.D., National Center for Injury Prevention and Control, CDC, DHHS
Harvey Levin, Ph.D., Baylor College of Medicine
Mark R. Lovell, Ph.D., University of Pittsburgh Medical Center
Mary Ellen Michel, Ph.D., National Institute of Neurological Disorders and Stroke, NIH, DHHS
Carol E. Nicholson, M.D., M.S., F.A.A.P., NICHD, NIH, DHHS
Robert M. Reece, M.D., Tufts University School of Medicine
Wesley Rutland-Brown, M.P.H., National Center for Injury Prevention and Control, CDC, DHHS
Peter C. Scheidt, M.D., M.P.H., NICHD, NIH, DHHS
Michael Weinrich, M.D., NICHD, NIH, DHHS
Barry S. Willer, Ph.D., University of Buffalo
Keith Yeates, Ph.D., Children’s Hospital, Ohio State University