Brain Magnetic Resonance Imaging Responses to Nonhypoxic Hypobaric Decompression
INTRODUCTION: The pathophysiological basis of neurological decompression sickness and the association between cerebral subcortical white matter (WM) change and nonhypoxic hypobaria remain poorly understood. Recent study of altitude decompression sickness risk evaluated acute WM responses to intensive hypobaric exposure using brain magnetic resonance imaging.
METHODS: Six healthy men (20 to 50 yr) completed 6 h of hyperoxic hypobaria during three same-day altitude chamber decompressions to pressure altitudes ≥ 22,000 ft (6706 m). Research magnetic resonance imaging sequences, conducted on the days preceding and following decompression, evaluated subcortical WM integrity, cerebral blood flow, neuronal integrity (fractional anisotropy), and neurometabolite concentrations.
RESULTS: No subcortical lesions were evident on diffusion weighted imaging and WM fractional anisotropy was unaffected. Mean WM blood flow was upregulated by 20% to over 25 mL · 100 g−1 · min−1. Gray matter flow was unchanged. There were no changes in gray matter or cerebellar neurometabolites. In parietal subcortical WM, levels of γ-aminobutyric acid (GABA) fell from (mean ± SD) 1.68 ± 0.2 to 1.35 ± 0.3 institutional units while glutathione (GSH) fell from 1.71 ± 0.4 to 1.25 ± 0.3 institutional units. Lactate increased postexposure in five subjects.
CONCLUSIONS: Postexposure decrements in GABA and GSH imply WM insult with loss of neuroprotection and oxidative stress. An association between decrements in GABA and GSH support a common origin, while GSH decrements also correlate with WM blood flow responses. WM lactate increments are prone to error but suggest dysregulation of subcortical microvascular flow. WM neurometabolite and blood flow indices did not normalize by 24 h postexposure.
Connolly D, Davagnanam I, Wylezinska-Arridge M, Mallon D, Wastling S, Lee VM. Brain magnetic resonance imaging responses to nonhypoxic hypobaric decompression. Aerosp Med Hum Perform. 2024; 95(10):733–740.
Exposure to hypobaric (low-pressure) decompression stress is associated with excess white matter hyperintensities (WMH) on magnetic resonance imaging (MRI) in military cohorts of altitude workers, specifically U-2 pilots and aerospace physiologists.1,2 Diffusion tensor imaging (DTI) MRI indicates that U-2 pilots also have decreased fractional anisotropy (FA) compared to controls, implying persistent diffuse axonal stress.3 Brain MRI conducted following aircrew trainee exposure to hypobaric hypoxia evaluated cerebral blood flow by arterial spin labeling (ASL) and neurometabolite concentrations on magnetic resonance spectroscopy (MRS).4 These indicated persistent elevation of white matter (WM) flow at 24 h, possibly related to neurometabolite responses, but could not discern the relative influences of hypoxia and hypobaria.
Few contemporary cohorts are exposed routinely to hypobaric decompression stress sufficient to incur significant risk of decompression sickness (DCS), but they include personnel involved in high altitude parachuting. In 2017, three cases of severe DCS, two neurological and one respiratory, arose in UK military parachutist despatchers working at 25,000 ft (7620 m), echoing the history of U-2 pilots experiencing DCS associated with increased activity and workload. A review of procedures resulted in more conservative altitude exposure limitations and denitrogenation requirements for high altitude parachuting.5 These procedures have been investigated, alongside the potential for safely conducting repeated, same-day ascents in a recent altitude chamber study that specifically evaluated the DCS risk to parachutist despatchers, reported in detail elsewhere.6 This study provided an opportunity to evaluate acute MRI responses to nonhypoxic hypobaria, as experienced by altitude workers who routinely breathe supplementary oxygen for prolonged periods during extended decompression.
The risk of DCS and likelihood of experiencing heavy loads of venous gas emboli were greater than expected, attributed to subjects’ levels of activity during decompression. To evaluate acute neurophysiological responses to nonhypoxic decompression stress, a subset of six subjects underwent research MRI sequences the day before and day following a series of three same-day altitude chamber ascents. MRI sequences included DTI to evaluate FA, ASL for cerebral gray matter (GM) and WM blood flow, and MRS for neurometabolites. In addition, diffusion weighted imaging (DWI) is capable of detecting microinfarcts as small as 1 mm in diameter.7 Accordingly, DWI sequences were performed to exclude acute subcortical microinfarcts, ischemia, or edema.8,9
Besides enhancing our understanding of the neurophysiology of decompression stress experienced by parachutist despatchers at high altitude, our aim was to complement a previous study of hypoxic hypobaria by investigating MRI responses to nonhypoxic hypobaria.4
METHODS
Subjects
The study adhered to the principles of the Declaration of Helsinki. The research was funded by the UK Ministry of Defense and the experimental protocol was approved in advance by the Ministry of Defense Research Ethics Committee, an independent body constituted and operated in accordance with national and international guidelines (1010/MODREC/19, version 1.6, dated 29 July 2021, ‘Human response to repeated altitude decompression stress’).
A subset of six healthy, nonsmoking men participated in the research MRI element of the study protocol. Three were in their third decade of life (ages 20, 26, and 28 yr) and three were in their fifth decade (ages 41, 46, and 50 yr). Cohort mean ± SD height was 181.2 ± 5.7 cm (range 173–188); weight 83.2 ± 10.6 kg (71.0–98.0); and body mass index 25.4 ± 3.2 kg.m−2 (21.9–29.0). All were screened on study entry by contrast echocardiography at the Royal Brompton Hospital, London, prior to participation, to exclude any right-to-left intracardiac (patent foramen ovale) or pulmonary shunt.
Subjects were also screened for pre-existing WMH by high resolution 3D fluid attenuated inversion recovery sequences using a 3.0 Tesla scanner with 32-channel head coil (Achieva, Philips Healthcare, Farnborough, United Kingdom) at the Sir Peter Mansfield Imaging Centre, University of Nottingham, United Kingdom. An eligibility threshold of no more than five punctate subcortical WMH was consistent with previous UK studies.10 However, no subject participating in acute phase MRI had more than two subcortical WMH on entry. Scans were reported for total WMH number by an experienced consultant neuroradiologist from the National Hospital for Neurology and Neurosurgery, London, United Kingdom.
Equipment
Decompressions were conducted in the high performance hypobaric chamber of the Altitude Research Facility at Ministry of Defense Boscombe Down, Wiltshire, United Kingdom.6 The day before (baseline, PRE) and day after (postexposure, POST) their altitude chamber exposures, each subject participating in acute phase MRI attended the Queen Square Imaging Centre, London, to undergo research MRI sequences on a 3 Tesla scanner with 48-channel air head coil (Signa Premier, GE Medical Systems, Chalfont St. Giles, United Kingdom).
Procedure
Subjects avoided hypobaric or hyperbaric environments (flying, diving, parachuting, mountaineering) in the 72 h prior to decompression and for 24 h afterwards. They also avoided alcohol and strenuous physical exertion for 48 h prior to decompression and 24 h afterwards. Otherwise, they ate and drank normally and ensured a good night’s rest before and after hypobaric exposure.
The detailed methodology is described elsewhere.6 Briefly, subjects arrived at 08:00 to undertake three consecutive altitude chamber ascents breathing 100% oxygen throughout. The first was to 25,000 ft (7620 m) pressure altitude for 60 min after denitrogenation for 60 min at 15,000 ft (4572 m). Second and third ascents were both to 22,000 ft (6706 m) for 90 min, each following 30 min of denitrogenation at 15,000 ft. Approximately 1 h separated consecutive ascents, breathing air normally at ground level (400 ft/122 m above mean sea level). Subjects ate and drank freely before and between ascents and were encouraged to remain well hydrated. They simulated the duties of parachutist despatchers throughout the day, including short spells of load carriage at ground level prior to each ascent. Ambulatory activities at altitude were representative of despatcher duties and equated to a “heavy” workload with respect to risk of DCS.11
Besides structural imaging, MRI scans included DWI, DTI, ASL, and MRS sequences. Axial DWI sequences were obtained using multishot, segmented echo-planar imaging with multiplexed sensitivity encoding, giving 36 slices of 3.0-mm thickness with 1.0-mm gaps. Five sets of DTI data were collected using single-shot, echo-planar imaging, spin-echo sequences, with varied diffusion-weighting parameter sets, each providing 54 axial slices of 2.5 mm thickness with no gaps. Axial perfusion images were performed with 3D fast spiral ASL sequences (frequency = 512; postlabeling delay = 1525 ms; slice thickness = 4 mm). Proton MRS data were acquired using a point-resolved spectroscopy chemical shift imaging technique from voxels placed to sample parietal lobe WM (20 × 20 × 20 mm; volume of interest 8.0 cm3), GM in the putamen (10 × 20 × 20 mm; 4.0 cm3), and cerebellum (20 × 20 × 20 mm; 8.0 cm3). Spectral quantitation employed a linear combination model technique for automatic phase adjustment, frequency alignment, baseline subtraction, and eddy current correction (LCModel version 6.3-1P; http://s-provencher.com/lcmodel.shtml).12 Unsuppressed tissue water signal (water scaling) was used to estimate metabolite concentrations based on typical values for GM and WM based on voxel location.13,14 Metabolite levels were corrected for cerebrospinal fluid partial volume within the MRS voxel, derived from segmented 3D T1-weighted volumes acquired in the same scanning session. An overview of the MRI protocol, with repetition and echo times, is in Table I.
Statistical Analysis
DWI imagery was reviewed for presence of any lesions by a consultant clinical neuroradiologist. DTI data are presented as normalized histograms of whole brain and white matter FA, with smoothed curves. FA values can range from 0 to 1, where 0 represents fully isotropic diffusion (i.e., water is free to diffuse equally in all directions) and 1 is wholly anisotropic (single axis diffusion). To normalize the histograms, each FA histogram “bin” value is divided by the peak height of the histogram, so the vertical axis also ranges from 0 to 1. This format facilitates direct visual comparison of PRE and POST whole brain and WM FA data for each individual and for the cohort as a whole. ASL data are reported as single value estimates of overall cerebral GM and WM blood flow for each individual with PRE/POST comparisons using paired data. Proton MRS analyses focused on neurometabolites reported previously.4 Specifically, these were N-acetylaspartate, inositol, glutamate (Glu) and glutamine, total choline, creatine (Cr) and phosphocreatine (PCr), and reduced glutathione (GSH), plus γ-aminobutyric acid (GABA) and lactate. POST minus PRE data were examined, both as individual neurometabolite concentrations and as ratios to total creatine (Cr + PCr), and paired data analyzed if a consistent response pattern was observed. All pairwise inferential statistical analyses employed paired t-tests (α = 0.05), with the normality assumption checked using the Shapiro-Wilk test. Possible associations were explored using simple and multiple linear regression (α = 0.05).
RESULTS
There was no evidence of focal subcortical microinfarcts, ischemia, or edema on DWI imagery, either PRE or POST. DTI data for individual and cohort PRE and POST whole brain FA are shown as normalized histograms in Fig. 1A, and similarly for WM FA in Fig. 1B. POST curves closely reproduce the PRE data for individual subjects and for the cohort as a whole. The POST curve for WM FA in subject #06 appears marginally left-shifted compared to his PRE data. However, there is no overall influence on cohort mean FA.
Citation: Aerospace Medicine and Human Performance 95, 10; 10.3357/AMHP.6445.2024
Individual and group data for PRE and POST WM and GM cerebral blood flow are shown in Fig. 2. Cohort mean POST GM cerebral flow increased by over 13%, but this is attributable to the responses of just two subjects, remaining largely unaffected in the remaining four, such that the overall variance in the cohort data remained much as before. On the other hand, POST WM flow was generally upregulated, rising in five subjects, with mean flow increasing by ∼20%. The upper end of the range remained steady at around 25 mL · 100 g−1 · min−1, consistent with expectation. However, the minimum value almost doubled, such that the variance between subjects decreased considerably, from mean (± SD) 21.1 ± 6.4 to 25.2 ± 1.5 mL · 100 g−1 · min−1. The increment represents a medium effect size (Cohen’s d = 0.73), but does not quite achieve statistical significance on a right-tailed paired t-test [t(5) = 1.8, P = 0.066].
Citation: Aerospace Medicine and Human Performance 95, 10; 10.3357/AMHP.6445.2024
When comparing POST data with PRE values, no GM or cerebellar neurometabolite responses were evident. Table II details the neurometabolite data for parietal WM, including relevant statistical outcomes. It is possible that the combination of decompression stress, hyperoxia, and exertion interferes with cerebral WM energy balance. Nonetheless, notwithstanding minor fluctuations, total creatine (tCr) levels (i.e., Cr + PCr) remained broadly consistent across the cohort before and after the experimental exposures. This is reflected by many consistent PRE and POST metabolite ratios relative to tCr, e.g., for Glu, total choline, inositol, and N-acetylaspartate. These appear unaffected by hypobaric decompression.
In contrast, relative to the PRE baseline, GABA decreased consistently in all subjects following decompression (Fig. 3A). This was statistically significant (P < 0.05) on paired t-tests both for absolute values and for ratio values of GABA/tCr (Table I). Across the cohort, GSH absolute values also decreased significantly postexposure (P < 0.05), falling in five of the six subjects (Fig. 3B), although the decrease in ratio values (GSH/tCr) was not statistically significant (P = 0.083) (Table II). Surprisingly, POST lactate levels increased in five subjects, four of them from a zero PRE baseline (Fig. 3C). While not statistically significant and also prone to measurement error, lactate peaks are not expected findings on white matter MRS. One subject exhibited lactate pre-exposure but not subsequently, which is unexplained.
Citation: Aerospace Medicine and Human Performance 95, 10; 10.3357/AMHP.6445.2024
Linear regression indicates a very strong direct association between the magnitudes of the postexposure decrements in GSH and GABA [R2 = 0.69, F(1,4) = 8.82, P = 0.041], suggesting that these are not chance findings and that they share a common origin (Fig. 4A). Furthermore, the change in WM blood flow predicts the change in WM GSH [R2 = 0.74, F(1,4) = 11.37, P = 0.028], with increasing flow mitigating the GSH decrement (Fig. 4B). No association is evident between changing WM blood flow and either GABA or lactate levels.
Citation: Aerospace Medicine and Human Performance 95, 10; 10.3357/AMHP.6445.2024
DISCUSSION
The key outcomes from the current study are the altered WM neurometabolite concentrations. The association between GABA and GSH decrements (Fig. 4A) implies a common cause that decreases the concentrations of both metabolites. The reduction in GABA indicates a loss of neuroprotection and excitatory bias, although the POST ratios of Glu/GABA, both for absolute values and relative to tCr, increased in only five of the six subjects, owing to a relatively greater POST decrease in Glu in one subject. Epilepsy is associated with decreased GABA levels, which also correlate well with seizure control.15 A relationship between oxygen toxicity and decreased GABA has long been suspected.16
GSH is the main antioxidant in mammalian cells that reduces reactive oxygen species (ROS). The POST decrease in WM GSH is most obviously attributable to hyperoxia, as all subjects spent a total of 6 h breathing 100% oxygen, albeit at reduced ambient pressure. Hyperoxia imposes oxidative stress with local generation of increased ROS, both by stimulating enzymatic production and by leakage from mitochondrial electron transport chains, the latter proportional to oxygen tension.17 The reduction in WM GSH is likely to reflect both glial and neuronal stress.18,19 The failure of the ratio of GSH to tCr to achieve statistical significance probably reflects the minor fluctuations in Cr and PCr levels postexposure (Table II).
The finding of elevated POST lactate levels in five subjects was unexpected. Lactate estimates are subject to considerable measurement error (SD > 30%), so must be interpreted with caution. While lactate is present in cerebrospinal fluid, care was taken to avoid sampling ventricular cerebrospinal fluid within the parietal WM voxels. The appearance of lactate in multiple subjects suggests recent WM anaerobic metabolism despite systemic hyperoxia.20 Arterialization of microbubbles does not explain WM microvascular ischemia, as no bubbles appeared in the left side of the heart at any time, and two younger subjects produced no venous gas emboli at all.6 Instead, hyperoxia-induced cerebral vasoconstriction reduces tissue blood flow, while microvascular flow heterogeneity increases shunting to the venous circulation and may lower local tissue oxygen availability.21 The suggestion of subsequent upregulation of WM blood flow, at least in some subjects, is consistent with persisting elevation of WM lactate. At around 25 mL · 100 g−1 · min−1 (Fig. 2), the upregulated POST WM flows measured in this study are at the upper end of MRI estimates of WM perfusion. The data support previous reports of persistently increased WM blood flow in response to decompression involving brief hypobaric hypoxia during what were otherwise hyperoxic exposures.4
While hyperoxia appears to have promoted acute reduction of both GSH and GABA, subjects had already been normoxic for 24 h by the time of postexposure MRI. However, ROS promote cell damage and activate endothelial cells, platelets, and neutrophils, inducing proinflammatory cytokines and chemokines. A companion study of blood biomarkers indicates that all six subjects are likely to have experienced a cytokine-mediated acute inflammatory response that does not resolve within 24 h, with persisting neutrophil activation, complement activation, and synthesis of acute phase proteins all likely to contribute to ongoing oxidative stress.22 Additionally, increased circulating monocytes and elevation of monocyte chemoattractant protein-1 and CD14 microparticles suggest possible monocyte activation and neuroinflammatory response.22,23 Notably, lactate elevation may result from macrophage accumulation in areas of acutely inflamed WM microvasculature stressed during decompression.24
The MRS pattern of decreased WM GSH with increased lactate is consistent with astrocytic injury or stress.25 Our companion study also demonstrated acutely elevated serum glutamate and high mobility group box protein 1, with delayed elevation of glial fibrillary acidic protein and brain-derived neutrotrophic factor at 24 h postexposure, independently supporting the contention that nonhypoxic hypobaric decompression generates WM stress.22,23
Altitude workers and divers both exhibit increased subcortical WMH associated with intensive exposure to hyperoxic decompression stress.1,2,26 Subcortical WMH reflect microvascular disease and develop in the region of a vascular watershed between deep penetrating and superficial cortical branches of cerebral vessels.27 Of note, WM GABA release has a neuroprotective effect following hypoxic/ischemic stress and chronic impairment of this mechanism may be associated with WM change.28,29 The current data provide evidence of acute WM stress responses to hyperoxic decompression, likely associated with microvascular dysregulation, that do not resolve fully within 24 h and may be associated with neuroinflammatory responses. In this context, the role of circulating microparticles to increase the permeability of the blood-brain barrier and to traverse it remains relevant.30,31 Microparticles remain implicated as possible mediators of the inflammatory response to decompression stress.32–34
The major limitation of the current study is the small cohort of male subjects undertaking research MRI, necessarily constrained by resources, which has compromised statistical power, such that this must be regarded as a pilot study. As a result, the increase in WM blood flow postexposure, while appearing relevant, has not achieved statistical significance. Likewise, the elevation of lactate in parietal WM warrants independent validation. Nonetheless, the strong associations between the magnitudes of GABA and GSH decrements, and between the upregulation of WM flow and GSH decrements, suggest that these are genuine rather than chance responses to hyperoxic decompression stress (Fig. 4).
In summary, these data provide evidence of acute neurophysiological responses indicating WM stress following nonhypoxic hypobaric decompression and that these responses do not resolve fully within 24 h. Repetitive localized microvascular disturbances and associated neuroinflammatory responses may underpin the association between intensive occupational altitude exposure and propensity to long-term white matter change.
A) Pre- (blue) and postexposure (red) whole brain fractional anisotropy (FA) showing six sets of individual data and group means. B) Corresponding data for white matter FA.
Pre- and postexposure global white matter (left) and gray matter (right) cerebral blood flow from arterial spin labeling MRI, represented as interquartile boxplots (upper graphs) showing median (horizontal bar) and mean (X), and individual subject responses (lower graphs).
Pre- and postexposure parietal white matter neurometabolite concentrations for: A) γ-aminobutyric acid (GABA); B) glutathione (GSH); and C) lactate. Data expressed in institutional units (iu) showing cohort means (± SE) and individual responses (dashed lines).
Linear regression analyses indicating: A) an association between the magnitudes of the postexposure decrements in GABA and GSH, supporting a common origin; and B) an association between altered postexposure WM blood flow and the GSH decrement.
Contributor Notes
