Improving Outcomes for Patients With Life-Threatening Neurologic Illness

The objective of the proposed minimal-risk, non-interventional retrospective (January 2012 onwards) as well as prospective, cross-sectional observational study (i.e., NCCU Research Repository) is to better understand the common pathophysiological pathways of life-threatening neurological illnesses, examine how systemic complications (e.g., fever, hyperglycemia, undernutrition, and infections) and our standard-of care management of them may exacerbate brain injury, and determine the impact of injury and intervention on functionally-relevant outcomes.

A systematically collected cohort may have an especially important role in the NCCU where RCTs are difficult to implement or may lack generalizability, and wherein there are pitfalls in research derived solely from administrative databases. This relational database will capture all patients admitted to University of Maryland's dedicated 22-bed adult NCCU. The investigators will not initially collect all potentially relevant fields unique to specific diagnoses. However, future disease-specific studies will allow the addition of data fields in this flexible database environment. This approach will include the tiered acquisition of patient characteristics (age, sex), clinical characteristics, standard-of-care medical treatments and surgical interventions, physiological data (such as vital signs, cerebral blood flow, intracranial pressure, cerebral oximetry etc), laboratory data, and diagnostic studies--particularly neurophysiology, neuropathology, and neuroimaging. Some data elements are inherently less reliable when abstracted from the electronic medical record (e.g., a more detailed social history, establishing premorbid functional status, and independent confirmation of past medical history), and will require an informed consent process to allow the research team to approach subjects or their surrogates.

Similar databases exist at other centers for neurocritical care and have been central to the identification of characteristics, both predictive and associative, of patient outcomes. Done properly, this repository can facilitate many NCCU priorities including hypothesis generating, determining feasibility of single-site trials or involvement in multi-center trials, multimodal monitoring, biomarker research, quality improvement, and resident education.

All data will be collected from the patient medical record, patient bedside monitors, radiology PACS, and laboratory systems. The collection of data will not affect nor delay patient treatment at any instance.

Potential subjects will include patients admitted to University of Maryland's 22-bed NCCU for treatment of a life-threatening neurological illness. All patients 18 years of age and older will be eligible for participation in the study.

At the time study personnel approach the patient or the patient's surrogate to discuss the study, they will be provided with three options: (a) Full Consent, (b) Partial Refusal, or (c) Complete Refusal.

Subject's medical treatment will conform to existing clinical management protocols including routine ICU care. Standard ICU protocols call for assessments of mental status several times a day. The investigators would like to prospectively record the results of these exams in a rigorous, standardized manner. The investigators will attempt to correlate these assessments with clinical events and outcome. In addition to standard clinical treatment, details of each subject's hospitalization, demographic profile, past medical history, family history will be obtained by questionnaire and chart review. At discharge or 2 weeks after admission, a brief examination and mental status test will be performed. The subject and a family member or caregiver may also be asked to return for follow up evaluations 3 and 12 months after initial brain or spinal cord injury, in order to evaluate their longterm neurological status, disability, emotional status, cognitive functioning, and overall quality of life. These interviews may be conducted in person and will take approximately one hour each. For those patients unable to return in person, a shorter telephone interview will be conducted. Outcome measures may include the Glasgow Outcome Scale (GOS), the extended Glasgow Outcome Scale (eGOS), Modified Rankin Scale, Lawton Activities of Daily Living Index, Barthel Index, Sickness Impact Profile (SIP), NIH Stroke Scale, psychosocial assessments, memory and cognitive neuropsychometric tests (e.g. the Montreal Cognitive Assessment-MoCA).

Variables recorded during hospitalization will be examined for statistical correlation with neuropsychometric test results. Candidate pathophysiologic variables with independent predictive value for cognitive dysfunction, (e.g., including severity of illness upon admission, hydrocephalus, ischemia from vasospasm, and laboratory test results) will be identified using Pearson correlation and forward stepwise logistic regression. Relationships between acute pathology (e.g., lesions on neuroimaging or malignant electrical patterns on electroencephalography) and dysfunction within various cognitive domains (memory, attention, visuospatial ability, language) will be examined using contingency tables, Pearson correlation, linear regression, and the Wilcoxon signed rank test. If necessary, multivariate models will be used to confirm that significant independent relationships exist.

Blood samples will be analyzed to identify the factors causing secondary brain injury (e.g., the narrowing of the blood vessels (vasospasm) after initial hemorrhage).

For patients requiring an External Ventricular Drain (EVD) as part of their medical treatment, the cerebrospinal fluid (CSF) is drained into a collection bag, the amount of drainage recorded and the CSF discarded. The investigators may collect a 1 cc sample of the CSF daily prior to discard until it is medically warranted to remove the EVD or 14 days whichever comes first. NO EVD will be placed for the purposes of clinical research.