Delineating Between Pathophysiologic Phenotypes of Hypoxic Ischemic Brain Injury After Cardiac Arrest

Purpose: To develop methods to differentiate between hypoxic ischemic brain injury patients exhibiting diffusion versus perfusion dependence. Secondarily, to investigate the underlying mechanisms responsible for diffusion limitation of oxygen delivery.

Hypotheses:

1. Patients with PD will exhibit increases in O2EF when CBF is reduced during hypocapnic ventilation while patients with DL will exhibit a minimal change in O2EF, indicating the presence of impaired diffusive O2 transport.
2. Serum brain biomarkers will be greater with DL vs. PD, indicating BBB permeability, glial and axonal injury.
3. The lactate / pyruvate ratio will be increased in patients with DL vs. PD, indicating mitochondrial dysfunction.
4. The clinical neurologic outcome will be worse in patients with DL compared to PD.
5. There will be a decreased proportion of macrophages and associated circulating immune cells in the jugular venous bulb vs. the arterial circulation in DL patients.

Justification:

This research project aims to prospectively differentiate between pathophysiologic phenotypes of hypoxic ischemic brain injury (HIBI), delineate the underlying mechanisms and determine the associated clinical outcomes. In Canada, approximately 40,000 people per year suffer a cardiac arrest. HIBI, the resultant cerebral insult following cardiac arrest, is the predominant determinant of adverse outcome post-cardiac arrest with only 5-15% patients experiencing favorable neurological outcomes. The remainder of HIBI survivors may experience significant neuropsychiatric sequelae and with significant healthcare costs (~$160,000 per patient annually). Given these dismal results, significant opportunities exist to improve outcomes.

HIBI pathophysiology is characterized by a primary ischemic injury during circulatory arrest and a secondary injury following return of spontaneous circulation (ROSC). This secondary injury stems in large part from inadequate post-resuscitation cerebral delivery of oxygen (CDO2). CDO2 is dependent upon cerebral blood flow (CBF), which is proportional to cerebral perfusion pressure (CPP) [mean arterial pressure (MAP) - intracranial pressure (ICP)]. Given the vast majority of HIBI pathophysiology is characterized by low - normal ICP, the primary physiological determinant of overall CDO2 in the critical care setting is MAP.

Observational studies have linked hypotension (decreased MAP and CBF) to adverse neurological outcome. However, randomized control trials of uniform MAP augmentation have failed to demonstrate clinical or biologic efficacy. This discrepancy between the perceived importance of optimizing MAP in the post-resuscitation period and the lack of clinical efficacy suggests there may be missing gaps exist in the understanding of the pathophysiology of HIBI. The simplistic approach of uniform MAP augmentation assumes that once oxygen is delivered to the cerebral capillary bed, normal oxygen diffusion occurs across the blood brain barrier (BBB) and normal cellular oxygen utilization culminates in neuronal aerobic metabolism. Cerebral metabolic homeostasis is predicated on the integration of convective oxygen delivery, diffusion across the BBB and normal cellular utilization by mitochondria. In other words, while increasing CBF and CDO2 supplies more oxygen to the brain, this strategy only addresses one component of the oxygen cascade and ignores the vital roles of oxygen diffusion and utilization. Recently, our research group has demonstrated that despite invasive neuromonitoring guided goal-oriented management, HIBI patients continued to experience a significant burden of brain hypoxia. In a post-hoc analysis, we demonstrated a unique pathophysiologic phenotype of in a sub-set of HIBI patients exhibiting diffusion limitation (DL) of oxygen delivery. In this instance, despite optimizing convective CDO2 to the cerebral capillary bed, diffusion of oxygen into the parenchymal tissue was inhibited in patients. The precise pathophysiologic mechanisms underpinning this finding in HIBI are unknown. Conversely, this sentinel work also demonstrated that the other sub-set of HIBI patients exhibited intact oxygen diffusion into the parenchyma which was dependent upon systemic perfusion. These two distinct pathophysiologic phenotypes of HIBI (diffusion limitation [DL] and perfusion dependence [PD]) carry significant implications for post-resuscitation care but key research questions remain. Specifically, confirmation of these phenotypes prospectively, delineating the mechanisms underpinning DL and investigating the clinical outcomes between DL and PD are of great importance.

Inflammation and secondary injury in DL: The precise mechanisms of underlying secondary injury and its contributions to diffusion limitation are unclear. Recently, our research group has demonstrated a net release of interleukin-6 from the cerebral parenchyma / neurovascular unit in HIBI patients exhibiting brain hypoxia and diffusion limitation. Interleukin-6, a key pro-inflammatory cytokine, is responsible for the initiation of the inflammatory cascade, leukocyte diapedesis, macrophage activation and release of additional downstream inflammatory mediators. The culmination of this dysregulated immune response is associated with neuron dysfunction, apoptosis and cell death. Therefore, understanding the proportion of immune cell subtypes participating in this dysregulated inflammatory cascade in HIBI is essential to developing therapies aimed at modulating the inflammatory cascade for the preservation of vulnerable neuronal tissue.

Objectives:

Aim 1 (primary outcome): To confirm the presence of DL vs. PD in HIBI patients by quantifying changes in the oxygen extraction fraction (O2EF) that occur secondary to a reduction in CBF with hypocapnic ventilation.

Aim 2 (secondary outcome): To compare differences in serum brain biomarkers in HIBI patients with DL vs. PD.

Aim 3 (secondary outcome): To compare the difference in the lactate / pyruvate ratio in HIBI patients with DL vs. PD.

Aim 4 (tertiary outcome): To compare the difference in the neurological clinical outcome in HIBI patients with DL vs. PD.

Aim 5 (tertiary outcome): To compare the differences in the proportion of circulating immune cells (macrophages) and rare subtypes between the arterial and jugular venous bulb locations in DL patients.

Research Design: This will be a prospective observational study. Study Overview: This study will be conducted at Vancouver General Hospital. Informed consent will be obtained from the substitute decision maker. On day 1 neuromonitoring will commence, arterial and jugular bulb blood will be collected for blood gas analysis and quantification of serum brain biomarkers, and microdialysis measures will be made every hour. Arterial sampling is done from the arterial line that is placed as part of usual care. Jugular bulb sampling is done from a jugular catheter that is placed specifically for the purpose of this study. Further we will quantify both the PvO2-PbtO2 gradients and conduct a hypocapnic ventilatory test to determine the relationship between perfusion and O2EF. On days 2 \& 3, neuromonitoring will continue and blood for serum biomarker and blood gas analysis will again be collected along with hourly microdialysis. After day 3, neuromonitoring will be removed upon request of the attending physician. CT and MRI imaging performed as part of usual care will be analyzed for ischemic burden. Clinical outcomes will be collected at 6-months in survivors.

Statistical considerations. Sample Size. A paired t-test was used to approximate the sample size calculation for a linear mixed model. To differentiate DL vs. PD, Menon et al. found a difference in the change of O2EF of approximately 10% in between patients with PD and DL (15% [SD 7] vs. 5% [SD 4], p =0.03) with exposure to hypocapnia. Assuming a difference in the O2EF of 10% with a standard deviation of the difference of 9, assuming an alpha 0.05 and desired power 0.80, the sample size is 28 patients (G*Power). We will recruit 40 patients to allow for early mortality prior to completion of three monitoring days. Given our previous distributions of approximately 50% of the entire cohort being defined as DL and 50% as PD, we predict that we will have at least 14 patients in each group. Given our previous recruitment rate of 12 patients in 12 months for invasive neuromonitoring in HIBI after cardiac arrest at our institution, this sample size will be readily achievable within three years. Analysis Plan: Discrete variables will be summarized by frequencies and percentages. Continuous variables will be summarized by mean (SD) or mean (P10,P90) if data are skewed. We will consider a P-value <0.05 to be statistically significant.

1. Aim 1: The differences in the change of the O2EF in both study groups during hypocapnia ventilation will be examined to confirm the presence of DL versus PD. This analysis will be conducted with an independent samples two-tailed t-test on the change in O2EF secondary to hypocapnia between DL and PD patients. The ability of non-invasively determining DL and PD with O2EF to that of the invasive PbtO2-PvO2 gradient method will be compared. This analysis will be conducted with receiver operator characteristics analysis.
2. Aim 2: The differences in the concentrations of serum brain biomarkers between DL and PD patients will be examined using independent samples two-tailed t-tests.
3. Aim 3: The differences in the lactate / pyruvate ratios from microdialysis between DL and PD patients will be examined using independent samples two-tailed t-tests to assess for mitochondrial dysfunction.
4. Aim 4: The differences in the clinical neurologic (BTACT \& CPC) and quality of life outcomes (EuroQol-5) at 6 months between DL and PD patients will be examined using independent samples two-tailed t-tests.
5. Aim 5: The differences in the proportion of circulating immune cells (macrophages) and rare subtypes between the arterial and jugular venous bulb locations in DL patients will be examined using independent samples two-tailed t-tests.
6. Exploratory outcome analyses: The differences in the ischemic burden on head CT (ASPECTS score) and on brain MRI (volume of white matter hyperintensity and percentage of voxels on apparent diffusion coefficient mapping of < 650 x 106mm/s on diffusion weighted imaging) in both groups will be examined using independent samples two-tailed t-tests.

Sex based analysis: As the understanding of HIBI pathophysiology remains limited, it is unclear if sex-related differences exist. As a result, assessing sex-related differences is not a primary outcome of the current study, but will encompass an exploratory analysis that aims to inform future research and hypothesis generation. In our ICU the proportion of patients admitted for post cardiac arrest HIBI has been approximately 60% male and 40% female over the last 3 years. There is likely to be a similar ratio of males to females in the current study. All data will be stratified by sex, across all patients and within each patient group. Within each patient group (DL vs. PD) differences between males and females will be assed with independent samples t-tests. A linear mixed effects model analysis (fixed factors: patient group \& sex; random factor: subjects) with an interaction term for patient group*sex will also be performed. Although likely underpowered to detect significant effects, these stratified data and their related statistical outputs will be reported and published in a data supplement alongside the manuscript. Any significant effects will, however, be included in the main manuscript document.