0046 Diffusion Tensor Imaging Evidence of Hypothalamic Injury in Traumatic Brain Injury Warfighters with Sleep Dysfunction

Abstract

Abstract Introduction While sleep disorders occur in 40–70% of chronic traumatic brain injury (TBI) patients, the pathophysiology remains unknown. Increasingly, DTI has been used to evaluate gray matter structures, but no prior studies have evaluated hypothalamic regions in TBI. We hypothesized that TBI patients with poor sleep quality by questionnaire and/or polysomnography (PSG) may have structural injury to hypothalamic sleep circuitry and that this may be detectable by diffusion magnetic resonance imaging (dMRI). We examined diffusion tensor parameters in warfighters using dMRI within the hypothalamus of poor sleepers and compared them to good sleepers. Methods A retrospective review of 92 warfighters with blast TBI and loss of consciousness included demographics, structural MRI, dMRI, PSG and Pittsburgh Sleep Quality Index (PSQI) questionnaire. Acquisition of diffusion-weighted and structural data was performed with three Tesla MRI. Using the California Institute of Technology probabilistic high-resolution in vivo atlas as a prior, the hypothalamic nuclei were segmented by applying diffeomorphic registration of T1- and T2-weighted structural images and mapped to dMRI space. Results TBI patients within the lowest quartile of hypothalamic fractional anisotropy (FA) measures demonstrated decreased total sleep time (320 +/- 52 minutes vs. 382 +/- 19, p=0.006) on PSG and had more sleep complaints on PSQI (p=0.029) compared to those with the highest quartile of FA measures. There was no difference in BMI, age or AHI among the quartiles. Radial, mean and axial diffusivity quartiles did not carry significant differences in TST or PSQI. Linear models did not show significant correlation between any imaging parameter and sleep quality measures. Conclusion Our results reveal microstructural differences in the hypothalami of military TBI patients that may be related to clinical sleep dysfunction. Biomarkers of sleep circuitry damage may further our understanding of sleep dysfunction after TBI. Lack of correlations in linear models may be a reflection of the small sample size or a complex interaction, and removal of outliers did not change our results. Larger longitudinal studies may help clarify the association between hypothalamic and brainstem circuitry structure after TBI and sleep dysfunction. Support This work was supported by a grant 130132 from USAMRMC.