Perioperatively Inhaled Hydrogen Gas Diminishes Neurologic Injury Following Experimental Circulatory Arrest in Swine

Abstract

This study used a swine model of mildly hypothermic prolonged circulatory arrest and found that the addition of 2.4% inhaled hydrogen gas to inspiratory gases during and after the ischemic insult significantly decreased neurologic and renal injury compared with controls. With proper precautions, inhalational hydrogen may be administered safely through conventional ventilators and may represent a complementary therapy that can be easily incorporated into current workflows. In the future, inhaled hydrogen may diminish the sequelae of ischemia that occurs in congenital heart surgery, cardiac arrest, extracorporeal life-support events, acute myocardial infarction, stroke, and organ transplantation.

Keywords: CPB, cardiopulmonary bypass; GFAP, glial fibrillatory acidic protein; H2, hydrogen gas; PDI, Psychomotor Development Index; SNDS, Swine Neurodevelopment Score; circulatory arrest; hydrogen gas; ischemia-reperfusion injury; neuroprotection; •OH, hydroxyl radical.

Graphical abstract

Figure 1

Presumed Mechanism of H₂ Action in the Setting of Ischemia Ischemic insults create tissue hypoxia, stimulating a complex cascade (not shown) that results in the release of superoxide (O₂^−•). When O₂^−• is present in excess (i.e., when compensatory mechanisms become saturated), it directly causes the reduction of transition metal ions, including Fe³⁺ and Cu²⁺, which in turn, generates hydroxyl radicals (•OH) by the Fenton reaction. The •OH is the strongest of the oxidant species and is the direct effector of DNA injury and lipid membrane peroxidation, which releases HNE and MDA, causing direct cellular injury and stimulating apoptosis. Unlike O₂^−• and H₂O₂, there is no known detoxification system for •OH; therefore, scavenging •OH is a critical antioxidant process. Molecular hydrogen (H₂), which freely permeates the cell wall and diffuses into the cytosol and mitochondria, reduces the hydroxyl radical to water and thus mitigates •OH-mediated tissue injury. HNE = 4-hydroxyl-2-nonenal; MDA = malondialdehyde; SOD = superoxide dismutase.

Figure 2

Study Protocol Neonatal swine were acclimated with bottle feedings 5 time per day (blue marks) for 5 days prior to experimentation. On the day of experimentation, swine were anesthetized and instrumented for cardiopulmonary bypass. Ischemic injury included circulatory arrest for 75 min at 25°C. Swine were then rewarmed and decannulated and underwent mechanical ventilation using a standardized intensive care protocol. Swine were ventilated for a total of 24 h (including pre- and post-operative treatments) with or without 2.4% inhaled hydrogen therapy (n = 8 swine per group). Videotaped neurologic examinations (green marks) took place prior to and daily following the circulatory arrest period. Swine underwent sedated brain MRI followed by terminal cerebral perfusion for histopathologic examination on postoperative day 3. F = Friday; M = Monday; MRI = magnetic resonance imaging; R = Thursday; RN = nursing care; S = Saturday/Sunday; W = Wednesday.

Figure 3

Clinical Neurologic Outcomes (A) Hydrogen-treated swine exhibited significantly higher rates of neurologically intact survival, which was defined as an SNDS ≤120 at the time of death (log-rank test: p = 0.0035). The overall rates of survival were similar between groups (log-rank test: p = 0.1435). (B) Hydrogen-treated swine exhibited significantly lower SNDS at 1, 2, and 3 days post injury (2-way ANOVA: p < 0.0001). Data for day 1 represent 8 swine per group; data for days 2 and 3 represent 8 swine in the hydrogen group and 6 in the control group (2 swine allocated to the control group died from refractory seizures and could not be successfully extubated). ***p < 0.001; **p < 0.01 for daily differences according to Sidak’s multiple comparisons post test. ANOVA = analysis of variance; DHCA = deep hypothermic circulatory arrest; SNDS = Swine Neurodevelopment Score.

Figure 4

Radiographic Differences Between Groups Axial T2 images (A) were assessed for radiographically apparent injuries (B), which were outlined as moderate (green) or severe (red) on a voxel-per-voxel basis. These areas of injury were corroborated by review of apparent diffusion coefficient mapping (C), which were similarly outlined (D). (E) Areas of injury were rendered in 3 dimensions and overlaid onto an image of the brain to provide a visual image of the differences in the volume of cranial injury. Data were based on brain MRI of 8 swine in the hydrogen-treated group and 6 in the control group (2 swine allocated to the control group died from refractory seizures and could not survive to day 3). (F) H₂-treated animals exhibited a significantly lower volume of injury than did control animals (Student t-test: p = 0.0463). The line represents median, boxes are interquartile ranges, and error bars are minimum and maximum values.

Figure 5

Histopathologic Differences Between Groups Regions of radiographically apparent injury (A) (arrows) correlated well with histologically apparent injury (B), shown here by fluorescence microscopy (original magnification: ×1, using a Rhodamine filter). Neuronal injury was scored between 0 (normal) and 5 (severe neuronal injury, necrosis) for each region through identification of hypereosinophilic and/or apoptotic neurons by using both light (C) and fluorescence (D) microscopy (original magnification: ×60; bars = 50 μm). Open arrows = hypereosinophilic and apoptotic neurons. (E) As a group, hydrogen-treated swine exhibited significantly lower histologic injury scores than controls (2-way repeated measures ANOVA according to Sidak's test results: p = 0.0044). Data are based on histopathology for 8 swine in the hydrogen-treated group and 6 in the control group (2 swine allocated to the control group died from refractory seizures and did not survive to day 3). Data are means; error bars are SEM.