Dysautonomia and Systemic Interactions in Traumatic Brain Injury

Overview

Following brain injury, complex interactions between the nervous system and other organs are frequently encountered. Systemic effects may be induced by dysregulation of the hypothalamic-pituitary-adrenal axis and the autonomic nervous system. This observational study will investigate the link between clinical, physiological and biochemical expressions of dysautonomic reactions and physiological stress, and their relations to sympathetic activation in traumatic brain injury patients treated in the neurointensive care unit.

Following traumatic brain injury (TBI) complex interactions between the nervous system and other organs are frequently encountered. Systemic reactions may be induced by dys-regulation of the hypothalamic-pituitary adrenal (HPA) axis and the autonomic nervous system. Neuro-endocrine disturbances are common and up to 50 % of brain injured Neuro-intensive care (NICU) patients may exhibit a period of relative adreno-cortico insufficiency in the early phase of TBI, which in part may be centrally mediated. Catecholamine surge is thought responsible for cardio pulmonary reactions such as myocardial stunning, and may be an instrumental part of neurogenic pulmonary edema. An imbalance between the parasympathetic and sympathetic nervous system has been identified, and may even effect outcome, but is poorly understood. This is seen both in early and chronic stages of brain injury. Heart rate variability has been implicated as an indicator of dys-autonomic parasympathetic dysfunction, and has in small studies been related to TBI outcome. Recently a clinical definition of Paroxysmal Sympathetic Hyper-activation (PSH) has been suggested and identities as related to patient outcome. Thus, in TBI the picture of a triad of dysautonomic and hypothalamic-pituitary dysregulation and injury driven inflammation, with a potential of bi-directional cross-talk between central and peripheral immuno- modulators, is emerging. This study will aim to explore and integrate indicators of these three components as to define phenotypes. It will investigate the utility of medically approved (CE) Skin Conductance Algesimeter (Med-Storm ®) in relation to other parameters of physiological stress including, heart rate variability (HRV), intra-cranial pressure reactivity index (PrX), and products of the HPA axis such as, ACTH, adrenaline and nor-adrenaline including break down products, markers of brain trauma driven neuro and systemic inflammation. The investigators hypothesize that a limited number of composite patterns will emerge and may describe patient phenotypes with differing trajectories. CRF: Electronic case report form (eCRF) with pseudo anonymized data via a globally unique personal identifier (GUPI) to secure eCRF platform. Patient Key kept locally and GDPR compliant. Variables: Baseline variables including IMPACT calculator variables for co-variate adjustment and trauma time, including predictors from the from Karolinska Traumatic Brain injury Database. Additional parameters of injury severity assessment and outcome predictors: Intracranial injury severity scoring on computed tomography (CT) scan by abbreviated injury severity score, Marshall CT classification, Rotterdam CT score, Helsinki CT score and Stockholm CT score. Severity scoring on magnetic resonance imaging (MRI) (if available) focusing on the presence of diffuse axonal injury (DAI), burden and region of DAI. Concomitant drugs of interest, such as analgesics, sedation, alpha II agonists, betablockers, vasopressor support. Daily injury severity scores. Daily symptom assessment of paroxysmal sympathetic hyperactivity (PSH) in relation to guideline definition. Daily Therapy Intensity level (TIL) and individual components. Daily Pain assessment scores. Available relevant clinical lab data in hospital system such as S100B, ProBNP, Troponins TSH,T3,T4. High resolution physiological data, physiological monitoring during ICU stay. Physiological monitoring via ICM+ including intracranial pressure (ICP), Brain tissue oxygenation (PbtO2), ECG waveform and heart rate variability (HRV), Central Temperature (Temp) , Saturation (SaO2). Pulse reactivity Index (PrX). Med-Storm Skin Conductance Algesimeter measure of sympathetic activity. ICM + ® annotation tool: Time-stamped changes in sedation and potentially painful and stressful clinical interventions. Bio-sampling: Biomaterial bio-banked for analyses Collection of plasma (from 4 ml whole blood sample) Daily, day 1-7, sampled from arterial line when available clinically. Cerebral Microdialysis: hourly when clinically available, pooled. Cerebral spinal fluid(CSF): Daily (2 ml), when clinically available from EVD catheter, day 1-7. Urine: Daily mixed aliquot (20 ml) of a 6 hour measured collection, day 1-7. Samples centrifuged 20G-15min: Aliquots of 200 μl and frozen with sample to freezer time recorded. Outcome Follow-up: Extended Glasgow Outcome Scale (GOSE) questionnaire and/or interview at 3, 6 and 12 months. Extraction from Karolinska Traumatic Brain injury Database. Death Date. Planned analyses: Wet-lab Proteomics: protein profiling of brain enriched and inflammatory proteins from serum, CSF and cerebral microdialysis. Microparticles/ exosomes, microRNA of central origin in CSF, microdialysate and plasma. HPA axis parameters and breakdown products including cortisol, catecholamines, metanephrines. Specific parameters of inflammatory crosstalk periphery/brain including choline acetyl transferase, HMGB1 as well as peripheral and central modulators of levels of T and B cell subgroup and glial activation. Analytics: Signal decomposition of physiological variables: ECG derived metrics including decreased baroreflex sensitivity (BRS), low frequency (LF), high frequency (HF) and total power (TP). Entropy. HRV. Signal analysis including FFT and wavelet. ICP derived metrics including PrX. Dimensionality reduction techniques such as principle component analysis (PCA) and non-supervised clustering techniques. Time series analyses: Cross-correlations and Trajectory analyses, Deep learning