Abstract TP254: Hydration Status and Regional Brain Volumetrics in a Population of Hospitalized Stroke Patients

Background: Measuring hydration status after stroke is especially important due to alterations in cerebral perfusion, though to date is reliant on crude bedside observations of the patient or indirect laboratory markers (e.g. blood urea nitrogen (BUN)/creatinine ratios). We hypothesized that regional brain volumetrics could differ based on hydration status at the time pf presentation with stroke. In this study, we sought to use methods of computational anatomy to quantify whether or not there are differences in brain region shape in underhydrated versus at the time of acute stroke. Methods: Clinical data was obtained from sequential stroke patients admitted to the Johns Hopkins Comprehensive Stroke Center. We used blood urea nitrogen: creatinine ratio (BUN/creatinine ratio) >15 to indicate a state of relative dehydration at the time of hospital arrival. Once the list of brain volumes was connected to each patient, each patient’s data was then normalized for intracranial volume (ICV). The dehydrated and hydrated volumes per each brain area underwent the Kolmogorov–Smirnov test, which checked the likelihood that two sets were drawn from the same distribution. To correct for multiple hypothesis testing, the Benjamini–Hochberg procedure was applied to the resulting p-values to control the false discovery rate. For any p-values less than 0.05, the sets were considered statistically likely to come from different distributions. The analysis was done on volumes partitioned at level 3 and level 5 granularity. Results: At level 5 granularity, significant volumetric differences were observed in 14.92% of brain regions with 99% confidence. At level 3 granularity, differences were identified in 11.11% of regions with 95% confidence. In both analyses, the body of the corpus callosum and the lateral ventricles exhibited significant shrinkage in dehydrated patients. Additional shrinkage was observed in the frontal lobe, parietal lobe, and central sulci. Conversely, the medulla demonstrated an increase in volume in the dehydrated group at both granularities. Conclusions: The corpus callosum and lateral ventricles may demonstrate vulnerability to hydration status in this cohort of patients. These regions may serve as potential biomarkers for hydration status in acutely ill stroke patients. Future work will be needed to determine how these volume differences relate to functional outcomes and recovery.