SAN DIEGO – The way Dr. Geoffrey T. Manley sees it, the classification of traumatic brain injury needs an extreme makeover.
For the past 35 years, clinicians have relied on symptomatology from the Glasgow Coma Scale (GCS) to classify traumatic brain injuries (TBIs) as mild, moderate, or severe, but such emphasis on symptoms “misses the point,” Dr. Manley, chief of neurotrauma and vice chairman of the department of neurosurgery at the University of California, San Francisco, said at the annual meeting of the American Association for the Advancement of Science.
“The brain is not like the heart, where if you lose a certain percentage of your heart muscle then you'll have an unexpected reduction in cardiac function. The brain is a unique organ in that it's an organ of functional connectivity. You can have very small lesions in discrete pathways, which can have a phenomenal impact on outcome. Many of these lesions can only be seen with MRI, which is not routinely used for TBI.”
He went on to note that the GCS was developed “before the advent of CT scans, so this is a very old system that we're using.”
In 2007, Dr. Manley and a working group of TBI experts–including Prof. Sir Graham Teasdale, who developed the GCS–convened to explore the potential for improving TBI classification (J. Neurotrauma 2008;25:719-38). It became clear to the group, Dr. Manley said, “that if we were going to try to change the field, we were going to have to start defining a common set of data elements and technical standards so that we could be able to collect the same information on patients from site to site and to make sure that assessment tools are applied in the same way.”
Common data elements are needed in TBI research “because accurate collection of structured data is essential, especially if you want to do meta-analyses and if you want to share data,” he added. “It reduces time, cost, and effort of initiating clinical trials and provides opportunities for lessons learned and best practices, even if a trial isn't considered successful.”
The group's recommendations call for the following:
▸ Broaden TBI trials. They should include less severely injured patients.
▸ Improve CT imaging classification. “The systems that we use now are different from hospital to hospital and radiologist to radiologist,” Dr. Manley said. “There is no standardization.”
▸ Increase use of early MRI. “Many of us have seen a lot of value in using MRIs,” he said. “We will get an MRI on a stroke patient in a moment, but we almost never get an MRI in a TBI patient. This is a cultural change that needs to happen in this field.”
▸ Examine phase II trials and surrogate end points more closely. TBI patients “have such a long recovery: an injury, an acute hospitalization, rehabilitation, and then you look at an outcome at 6 months or a year,” he said. “Lots of things happen during that time period.”
▸ Develop more complex statistical and bioinformatics tools. TBI studies “aren't like cancer studies,” he said. “You can't phenotype these patients as well as you can in studies of other diseases. We need some novel statistical methods to deal with the realities of studying these patients with life-threatening diseases.”
In March 2009, Dr. Manley and about 160 other representatives from 49 agencies and institutes, including the Department of Defense, the Department of Education, and the National Institute of Neurological Disorders and Stroke convened in Washington to begin an unprecedented effort to develop standards for TBI data collection and to better define and classify TBI.
The experts were divided into numerous work groups charged with assembling white papers on specific areas of TBI research, including demographics and acute clinical assessment, biospecimens and biomarkers, neuroimaging, posttraumatic stress disorder, and outcome measures. White papers from the various work groups will be published later in 2010, and the TBI Common Data Elements will be available online at www.nindscommondataelements.org
The next step in this multidisciplinary effort is to establish a prospective, multivariate TBI database to validate common data elements, including a contemporary snapshot of TBI and treatment and a cross-sectional overview of patients. “So rather than saying we're looking at patients with severe, moderate, or mild injury, we're going to be agnostic to [the label of] mild, moderate, and severe, and we'll look across the entire spectrum of injury,” Dr. Manley explained.
The database “will also allow the researchers to validate prognostic models, establish process indicators, and develop improved TBI classification,” he said.