http://wrap.warwick.ac.uk

Original citation: Sagoo, R. S., Hutchinson, Charles, Wright, A., Handford, C., Parsons, Helen, Sherwood, V., Wayte, S., Nagarajaan, S., NgAndwe, E, Wilson, M. H. and Imray, C.. (2016) Magnetic resonance investigation into the mechanisms involved in the development of high-altitude cerebral edema. Journal of Cerebral Blood Flow & Metabolism . ISSN 0271-678X Permanent WRAP url: http://wrap.warwick.ac.uk/75613 Copyright and reuse: The Warwick Research Archive Portal (WRAP) makes this work of researchers of the University of Warwick available open access under the following conditions. This article is made available under the Creative Commons Attribution-NonCommercial 3.0 (CC BY-NC 3.0) license and may be reused according to the conditions of the license. For more details see: http://creativecommons.org/licenses/by-nc/3.0/ A note on versions: The version presented in WRAP is the published version, or, version of record, and may be cited as it appears here. For more information, please contact the WRAP Team at: publications@warwick.ac.uk

http://wrap.warwick.ac.uk/http://wrap.warwick.ac.uk/75613http://creativecommons.org/licenses/by-nc/3.0/mailto:publications@warwick.ac.uk

Original Article

Magnetic Resonance investigation into themechanisms involved in the developmentof high-altitude cerebral edema

Ravjit S Sagoo1, Charles E Hutchinson1,2, Alex Wright3,Charles Handford4, Helen Parsons5, Victoria Sherwood6,Sarah Wayte6, Sanjoy Nagaraja1, Eddie Ng’Andwe1,Mark H Wilson7 and Christopher HE Imray2,8,9; for theBirmingham Medical Research and Expedition Society

Abstract

Rapid ascent to high altitude commonly results in acute mountain sickness, and on occasion potentially fatal high-altitude

cerebral edema. The exact pathophysiological mechanisms behind these syndromes remain to be determined. We report a

study in which 12 subjects were exposed to a FiO2¼ 0.12 for 22h and underwent serial magnetic resonance imaging sequencesto enable measurement of middle cerebral artery velocity, flow and diameter, and brain parenchymal, cerebrospinal fluid and

cerebral venous volumes. Ten subjects completed 22 h and most developed symptoms of acute mountain sickness (mean Lake

Louise Score 5.4; p< 0.001 vs. baseline). Cerebral oxygen delivery was maintained by an increase in middle cerebral arteryvelocity and diameter (first 6 h). There appeared to be venocompression at the level of the small, deep cerebral veins (116 cm3

at 2 h to 97 cm3 at 22 h; p< 0.05). Brain white matter volume increased over the 22-h period (574 ml to 587 ml; p< 0.001) andcorrelated with cumulative Lake Louise scores at 22h (p< 0.05). We conclude that cerebral oxygen delivery was maintained byincreased arterial inflow and this preceded the development of cerebral edema. Venous outflow restriction appeared to play a

contributory role in the formation of cerebral edema, a novel feature that has not been observed previously.

Keywords

Brain imaging, cerebral blood flow, high altitude, magnetic resonance imaging, physiology

Received 10 August 2015; Revised 8 November 2015; Accepted 27 November 2015

Introduction

On ascent to altitude, there is a reduction in barometricpressure and a consequent fall in the partial pressure ofinspired oxygen.1–3 Acute mountain sickness (AMS) ofvarying severity can develop in individuals followingrapid ascent to high altitudes.1,4 The symptoms ofAMS encompass a wide range of signs and symptoms,the most prominent of which include headache,vomiting, lethargy, and cerebral dysfunction.1,2 At theextreme end of the spectrum, there is a risk of thepotentially fatal high-altitude cerebral edema (HACE)and high-altitude pulmonary edema (HAPE)developing.5,6

The brain is particularly sensitive to a lack of oxygenand, therefore, several adaptive mechanisms arebelieved to occur to maintain adequate oxygen delivery.

1Department of Imaging, University Hospitals Coventry and

Warwickshire NHS Trust, Coventry, West Midlands, UK2Warwick Medical School, University of Warwick, Coventry, West

Midlands, UK3College of Medical and Dental Sciences, University of Birmingham,

Birmingham, UK4University Hospitals Birmingham NHS Foundation Trust, Birmingham,

UK5Division of Health Sciences, Warwick Medical School, University of

Warwick, Coventry, West Midlands, UK6Department of Medical Physics, University Hospitals Coventry and

Warwickshire NHS Trust, Coventry, West Midlands, UK7Department of Neurosurgery, Imperial College Healthcare NHS Trust,

London, UK8Department of Surgery, University Hospitals Coventry and

Warwickshire NHS Trust, Coventry, West Midlands, UK9Coventry University, West Midlands, UK

Corresponding author:

Christopher HE Imray, Department of Surgery, University Hospitals

Coventry and Warwickshire NHS Trust, Coventry CV2 2DX, West

Midlands, UK.

Email: chrisimray@aim.com

Journal of Cerebral Blood Flow &

Metabolism

0(00) 1–13

! Author(s) 2016

Reprints and permissions:

sagepub.co.uk/journalsPermissions.nav

DOI: 10.1177/0271678X15625350

jcbfm.sagepub.com

Over a variable time course, hyperventilation, increasedheart rate, increased blood pressure, a rise in the hem-atocrit and hemoglobin are all known to facilitatecerebral oxygen delivery (COD).7 Specific to the brain,Wilson et al.8 suggested that both a rise in middle cere-bral artery (MCA) diameter and flow velocity areresponsible for a rise in cerebral blood flow (CBF) insustained hypoxia. Until recently, cerebral arterial diam-eter was widely assumed to stay constant, while cerebralflow velocity was thought to be the sole responsiblefactor in maintaining cerebral oxygenation.9

The exact pathophysiological mechanisms in thedevelopment of AMS and HACE have yet to be fullydetermined. One of themain theories to explain the devel-opment of AMS is an increase in intracranial pressure(ICP).3 Direct ICP measurement is difficult to achieve inhealthy volunteers, but the ICP can be inferred indirectlyvia volume alterations in the different intracranial tis-sues.10,11 Hackett et al.12 have alluded to the developmentof cerebral edema and its association with AMS in vol-unteers subjected to sustained hypoxia, particularlywithin the corpus callosum. Other studies have showndevelopment of cerebral edema, as well as increases inbrain parenchymal volume and compensatory decreasesin cerebrospinal fluid (CSF) volume in volunteers sub-jected to sustained hypoxia, but these changes were seenin patients both with and without AMS.13

Vasogenic (extracellular water accumulation due toincreased permeability of the blood–brain barrier)12 andcytotoxic (intracellular)14 edema have been postulated asthe mechanisms that underlie HACE. Whether AMS is amilder form of vasogenic and/or cytotoxic, edemaremains unclear. However, the rise in parenchymalvolume or decrease in CSF volume could also be due toa rise in the CBF and/or efferent restriction of cerebralvenous outflow, thus increasing the ICP.15 In a high-altitude study, Wilson et al.16 showed that the headacheburdenwas positively related to the retinal vein distensionand individuals with relatively narrow transverse sinuseswere more prone to hypoxia-induced headaches.

We undertook a sea-level study to observe the effectsof normobaric hypoxia in healthy individuals on thebrain and cerebral vasculature over a prolongedperiod of 22 h (AMS and HACE normally take12–24 h to develop), making multiple assessments toassess the time course of aspects of cerebral perfusionthat might be implicated in the development of HACE.We assessed a unification hypothesis of two separateconcepts to explain the development of HACE: firstly,the raised arterial inflow (with cytotoxic and vasogenicedema) and, secondly, restricted venous outflow the-ories.15 This unification hypothesis was first proposedin schematic format in 2009.3 Specifically, we aimed tostudy changes in the cerebral arterial caliber and flow vel-ocity, changes in the small cerebral venous and dural

venous sinus caliber, observe the time course of the devel-opment of cerebral edema and volume alterations of thebrain parenchyma and intracranial CSF. A novel featureof this study was that these variables were measuredrepeatedly and simultaneously over a 22-h period of sus-tainedhypoxia,which has not beenperformed previously.

Materials and methods

Subjects and study design

The study design was a prospective observationalcohort study, in which subjects were exposed tonormobaric hypoxia in a chamber setting. Subjectsself-assessed clinical symptoms and various magneticresonance (MR) sequences were used to assess changesin cerebral parameters. Coventry and WarwickshireNHS Research Ethics Committee provided ethicalapproval for this study. Written informed consent wasobtained from each study participant.

Twelve subjects (10 males, mean age: 26.2 years,range: 21–47 years) were studied over a continuousnormobaric hypoxic period of 22 h. None of the sub-jects had ascended> 1000m in the preceding sixmonths. Subjects were students interested in wildernessmedicine and recruited by word of mouth.

Hypoxia

Subjects were exposed to 22 h of continuous normoba-ric hypoxia (FiO2¼ 0.12, approximately equivalent toan altitude of 4400m) inside a 2m� 2m� 1.6mhypoxic tent. Three hypoxic generators were used(Everest Summit Hypoxic Generator, HypoxicSystems, New York, NY, USA). Build-up of carbondioxide within the tent was controlled with an airpump (Bair HuggerTM, 3M, Berkshire, UK), whichpumped the tent air through a two soda lime scrubber(SpherasorbTM – medical grade soda lime fromIntersurgical Ltd UK in canisters by Draeger MedicalUK Ltd). The soda lime was replaced at 12 h. Baselinephysiological measurements, venous blood samples,and magnetic resonance imaging (MRI) scans wereobtained at normoxia (FiO2¼ 0.21). Inspired oxygenand carbon dioxide concentrations were checkedevery 15min for the first 2 h, every 30min from2–10 h, and then every 2 h from 12–22 h. ExtendedMRI compatible tubing and a tight fitting mask thatwere connected to one of the three hypoxic generatorsenabled the subjects to remain hypoxic during the MRIstudies. The subjects were scanned at rest and did notundertake exercise during the study period.

Subjects ate and drank ad libitum were kept wellhydrated and slept at various points between the 12and 22-h time points.

2 Journal of Cerebral Blood Flow & Metabolism

Physiological measurements

Physiological measurements were obtained at baselinefollowed by 2, 4, 6, 11, and 22 h of continuous normo-baric hypoxia. Blood pressure was measured using anautomated cuff, peripheral arterial oxygen saturation(SaO2) and heart rate by near infrared finger pulseoximetry probe, and end tidal carbon dioxide(ETCO2) using an infrared capnometer (Dash 3000,GE Medical Systems, Wisconsin, USA).

Whole venous blood samples were obtained at base-line, 11 and 22 h of continuous normobaric hypoxia.Laboratory analysis of hematological (hematocrit andconcentration of hemoglobin (Hb), red blood cells,white blood cells, and platelets), biochemical (concen-tration of sodium, potassium, urea, and creatinine) andinflammatory variables (erythrocyte sedimentation rate(ESR), and concentration of C-reactive protein (CRP))was carried out on each sample.

Blood oxygen content (BOC) and MCA flow (f) werecalculated using equations (1) and (2), respectively.

BOC ¼ 1:36 �Hb � SaO2 ð1Þ

f ¼ � d2

� �2v ð2Þ

where d is the MCA diameter and v the MCA flowvelocity.The COD was determined using equation (3).

COD ¼ BOC � f ð3Þ

Headache and Lake Louise scores

The Lake Louise questionnaire and scoring systemwere used to assess symptoms associated with AMSat baseline and 2, 4, 6, 11, and 22 h of continuous12% normobaric hypoxia.17 A total Lake Louisescore of >3 including headache at any time point wasdefined as AMS. Cumulative Lake Louise scores werecalculated by adding the Lake Louise scores at eachtime point to give a total score for each subject.

MRI

Each subject underwent 3-T MRI scanning (GE SignaHDxt, Milwaukee, USA) at baseline, and 2, 4, 6, 11,and 22 h of continuous normobaric 12% hypoxia. Thefollowing sequences were obtained:

1. 3D time of flight (TOF) magnetic resonance angiog-raphy (MRA) of the circle of Willis was performed toenable measurement of MCA cross-sectional area anddiameter (repetition time (TR)¼ 27ms; echo time

(TE)¼ 3.4ms; flip angle (FA)¼ 20�; matrix¼ 384�256, field of view (FOV)¼ 210� 210mm, slice thick-ness (ST)¼ 1.2mm)

2. A sagittal single-slice 2D electrocardiogram triggeredphase-contrast (PC) MRA was performed to enableestimation of the MCA flow velocity (TR¼ 33ms;TE¼ 5.3ms; FA¼ 30�; matrix¼ 256� 192, FOV¼160� 160mm, ST¼ 5mm, velocity encoding param-eter (VENC)¼ 150 cm/s). The sagittal slice was cen-tered 1 cm proximal to the bifurcation of the rightMCA. The acquisition window was sampled with 25phases.

3. 3D PC magnetic resonance venography (MRV) of thedural venous sinuses was performed to enable estima-tion of the volumes of the superior sagittal and trans-verse venous sinuses (TR¼ 12ms; TE¼ 4ms;FA¼ 8�; matrix¼ 352� 224, FOV¼ 240� 204mm;acquired ST¼ 1.6mm; reconstructed ST¼ 0.8mm).

4. T2*-based susceptibility-weighted imaging (SWI) ofthe whole brain (SWAN� sequence, (susceptibility-weighted angiography) provided by GE Healthcare,Buckinghamshire, UK) was performed to assesschanges in the deep cerebral venous anatomy(TR¼ 42ms; TE¼ 24.7ms; FA¼ 15�; matrix¼320� 224, FOV¼ 220� 200mm; acquired ST¼2.6mm; reconstructed ST¼ 1.3mm).

5. Diffusion-weighted images of the whole brain(TR¼ 10,000ms; TE¼ 95ms; matrix¼ 256� 256;FOV¼ 240� 240mm; ST¼ 4.0mm; b¼ 0, 800,1000) were acquired at each time point, fromwhich apparent diffusion coefficient (ADC) mapswere calculated to assess for edema.

6. T1-based 3D fast spoiled gradient echo imaging(FSPGR) of the whole brain was performed to enableestimation of brain and CSF volume (TR¼ 7.8ms;TE¼ 3ms; FA¼ 12�; matrix¼ 256� 256; FOV¼220� 165mm; acquired ST¼ 2.6mm; reconstructedST¼ 1.3mm).

Image analysis

The 3D TOF MRA data were imported into a GEAdvantage Workstation 4.6 (GE Healthcare,Buckinghamshire, UK) to estimate the MCA diameterand cross-sectional area. Three raters who were blindedto the subject and time point at which the scan had beenacquired, undertook the measurements independently.The M1 segment of the MCA was identified bilaterallyusing the orthogonal multiplanar reconstruction(MPR) module and each MCA M1 segment was manu-ally divided into five equally spaced data points. Thecross-sectional view, the maximal diameter and cross-sectional area were manually measured at each datapoint in bilateral MCA M1 segments. A mean value

Sagoo et al. 3

for the MCA diameter and cross-sectional area at eachtime point was calculated from the 10 data points.

The 2D PC-MRA data were also imported into theGE Advantage Workstation 4.6. MCA flow velocitieswere obtained using the flow analysis module. TheMCA margins were defined automatically and adjustedmanually for each phase if necessary. Mean and peaksystolic flow velocities were determined automaticallyby the flow analysis module.

Analyze 11.0 (AnalyzeDirect, KS, USA) was used tocalculate the total volumes of the superior sagittal andtransverse venous sinuses using the 3D PC-MRV dataacquired at each time point. The semi-automaticthreshold technique (using standardized minimum andmaximum threshold values) was used to define andextract the appropriate venous sinuses to enable auto-mated volume calculation.

The deep venous space was calculated in byquantifying the total volume of all voxels correspond-ing with deep veins within the brain tissue, which havelow signal intensity on SWI. SWI were segmented toremove the skull and CSF using SPM8 (WellcomeTrust Centre for Neuroimaging, University College,London, UK). Using a Matlab (Mathworks Inc.,Cambridge, UK) program written in-house, a linearintensity normalization of the voxel data was per-formed to correct for global intensity variationsbetween time points. Voxels falling within a userdefined intensity window (Figure 1) were summed andused to calculate the deep venous space based on thevoxel dimensions. The 2-h time point was used as abaseline for these measurements due to an expected

increase from baseline to 2 h caused by the magneticsusceptibility of deoxyhemoglobin (the BOLD effect).

ADC maps were calculated after importing thediffusion-weighted images into the GE AdvantageWorkstation 4.6. Circular regions of interest (ROI) of1 cm2 were placed bilaterally within the frontal whitematter, lentiform nuclei (gray matter), occipital whitematter, cerebellar white matter, and the genu and sple-nium of the corpus callosum. The mean ADC valuewithin each right and left circular ROI was then calcu-lated. The right and left ADC values for a given ana-tomical region were averaged to give a single meanADC value for the region.

The T1-weighted 3D FSPGR data were importedinto the FMRIB Software Library (FSL) 5.0 (FMRIBGroup, Oxford, UK) to quantify brain parenchymaland CSF volumes. Initially, the Brain Extraction Tool(BET) was used to delete all non-brain tissue. TheFMRIB’s Automated Segmentation Tool (FAST) wasthen used to segment the brain image into differenttissue types (gray matter, white matter, and CSF).This tool also automatically corrects for any spatialintensity variations. Tissue volume quantification wascarried out using the FSLSTATS command for each ofthe segmented tissue images at each time point.

Statistical analysis

Analysis of repeated observations was performed usingthe linear mixed models procedure in SPSS v22 (IBM,NY, USA) to examine the effects of time exposed tohypoxia on each dependent variable. The linear mixedmodels procedure does not require a balanced design sois able to account for all missing data. Model param-eters were estimated using the restricted maximum like-lihood method and covariance was modeled asunstructured. Sidak’s adjustment for multiple pairwisecomparisons and paired t-tests were employed asappropriate. Study size was in part determined bypower calculations but also based upon similar studies.To assess the relationship of selected variables, bivari-ate correlations for ordinal data (Spearman’s Rho) andcontinuous data (Pearson’s correlation) were used asappropriate. Statistical significance was set at p� 0.05.

Results

Two of the 12 subjects were unable to complete thestudy due to severe AMS and had to be withdrawnafter the 11-h time point. The data set was otherwisecomplete. Changes in the basic physiological variableswith time exposed to normobaric hypoxia are displayedin Table 1, and Table 2 presents changes in the mea-sured radiological variables at each time point. Bothtables summarize the estimated marginal means of the

Figure 1. A histogram demonstrating voxel intensities for the

whole brain for all subjects at the specific time points. The

number of voxels falling below the vertical dashed line (the set

threshold level) was calculated to enable venous space

quantification.

4 Journal of Cerebral Blood Flow & Metabolism

Tab

le1.

Est

imat

ed

mar

ginal

mean

s,si

gnifi

cance

ofch

ange

,and

95%

CIs

for

eac

hm

eas

ure

dphy

siolo

gica

lvar

iable

and

tota

lLak

eLouis

ean

dhead

ache

score

susi

ng

multile

velm

odelin

g.

0h

(norm

oxia

)2

hhy

poxia

4h

hypoxia

6h

hypoxia

11

hhy

poxia

22

hhy

poxia

Sig

Frac

tion

of

insp

ired

oxyg

en

0.2

10.1

20.1

20.1

20.1

20.1

2<

0.0

5

Frac

tion

of

insp

ired

carb

on

dio

xid

e

0.0

30.2

30.2

60.2

50.2

30.2

7<

0.0

5

Peri

phera

loxyg

en

satu

ration,%

99.3

3(9

8.1

9–99.7

5)

82.9

2(7

8.9

0–86.9

3)

82.9

2(8

0.9

5–84.8

8)

84.1

7(8

0.8

1–87.5

3)

83.8

3(7

9.9

9–87.6

7)

84.9

8(8

2.7

7–87.1

9)

<0.0

01

End

tidal

CO

2,%

5.3

6(5

.09–5.6

3)

1.7

5(1

.24–2.2

7)

2.2

1(1

.87–2.5

5)

1.9

2(1

.57–2.2

7)

2.1

3(1

.70–2.5

6)

1.8

8(1

.43–2.3

3)

<0.0

01

Hear

tra

te,beat

s/m

in64.9

2(6

0.4

9–69.3

5)

71.5

0(6

4.2

4–78.7

6)

80.5

0(7

4.4

9–86.5

1)

78.2

5(7

2.2

4–84.2

6)

76.9

2(6

7.9

5–85.8

9)

85.6

9(8

0.4

5–90.9

3)

<0.0

01

Syst

olic

blo

od

pre

ssure

,m

mH

g

123.3

3(1

18.0

2–128.6

5)

126.4

2(1

18.9

2–133.9

2)

121.8

3(1

15.2

5–28.4

2)

127.9

2(1

22.4

8–133.3

5)

133.6

7(1

26.8

9–140.4

4)

130.7

2(1

22.8

4–138.6

0)

0.0

85

Dia

stolic

blo

od

pre

ssure

,m

mH

g

73.5

8(6

5.4

5–81.7

2)

68.9

2(6

2.2

6–75.5

8)

71.5

0(6

6.8

5–76.1

5)

71.0

8(6

3.8

9–78.2

8)

76.8

3(7

0.3

4–83.3

2)

76.0

5(6

8.4

5–83.6

5)

0.0

68

Tota

lLak

eLouis

esc

ore

01.3

3(0

.77–1.9

0)

1.3

3(.71–1.9

6)

2.0

0(0

.67–3.3

3)

2.7

5(1

.70–3.8

0)

5.4

0(3

.19–7.6

2)

0.0

04

Head

ache

score

00

0.3

0(0

–0.5

0)

0.6

0(0

.10–1.1

0)

1.4

0(0

.80–2.0

0)

1.6

0(1

.00–2.2

0)

<0.0

01

Hem

ogl

obin

conce

ntr

atio

n,g/

dL

15.0

9(1

4.6

2–15.5

6)

15.2

6(1

4.6

8–15.8

5)

15.4

4(1

4.6

3–16.2

5)

0.4

45

Pla

tele

tco

unt,�

10

9/L

236.5

(208.2

8–264.7

1)

239.9

6(2

15.2

9–264.6

3)

234.5

6(2

16.9

1–252.2

1)

0.3

89

Red

blo

od

cell

count,�

10

12/L

5.0

3(4

.88–5.1

8)

5.0

9(4

.94–5.2

4)

5.1

2(4

.85–5.3

9)

0.4

04

Hem

atocr

it0.4

4(0

.43–0.4

6)

0.4

4(0

.43–0.4

6)

0.4

5(0

.43–0.4

7)

0.6

23

Seru

mso

diu

m

conce

ntr

atio

n,m

mol/L

143.4

9(1

42.4

5–144.5

2)

141.9

3(1

40.8

–143.0

6)

142.3

(141.3

3–143.2

7)

0.0

9

Seru

mpota

ssiu

m

conce

ntr

atio

n,m

mol/L

4.5

3(4

.32–4.7

3)

4.2

9(4

.05–4.5

4)

4.6

3(4

.38–4.8

9)

0.1

86

Seru

mure

a

conce

ntr

atio

n,m

mol/L

5.7

3(4

.86–6.6

1)

5.6

4(4

.71–6.5

7)

5.6

7(5

.02–6.3

3)

0.9

06

Seru

mcr

eat

inin

e

conce

ntr

atio

n,mm

ol/L

78.8

1(7

0.8

5–85.4

8)

81.4

(72.8

3–90.5

)83.6

(73.7

6–90.5

8)

0.0

7

CI:

confid

ence

inte

rval

s.

Sagoo et al. 5

Tab

le2.

Est

imat

ed

mar

ginal

mean

s,si

gnifi

cance

of

chan

ge,

and

95%

CIs

for

eac

hm

eas

ure

dra

dio

logi

calva

riab

leusi

ng

multile

velm

odelin

g.

0h

(norm

oxia

)2

hhy

poxia

4h

hypoxia

6h

hypoxia

11

hhy

poxia

22

hhy

poxia

Sig

MC

Adia

mete

r,m

m3.1

0(2

.89–3.3

0)

3.1

9(3

.00–3.3

8)

3.1

8(2

.97–3.4

0)

3.2

3(3

.00–3.4

4)

3.1

2(2

.91–3.3

3)

3.2

3(3

.02–3.4

3)

0.0

65

MC

Aar

ea,

mm

29.3

5(8

.18–10.5

2)

9.8

9(8

.693–11.0

8)

9.8

4(8

.58–11.1

0)

10.0

8(8

.69–11.4

7)

9.3

9(8

.19–10.5

9)

10.0

2(8

.81–11.2

3)

0.3

4

MC

Apeak

syst

olic

flow

velo

city

,cm

/s70.4

1(5

5.9

2–84.9

0)

82.2

1(6

3.3

4–101.0

8)

90.6

8(7

0.5

4–110.8

2)

84.9

3(6

3.1

9–106.6

6)

83.9

1(6

2.0

8–105.7

4)

78.7

7(5

9.5

4–98.0

1)

0.1

97

MC

Am

ean

flow

velo

city

,cm

/s47.5

3(3

9.6

9–55.3

6)

57.0

9(4

5.9

2–68.2

7)

53.8

3(4

2.8

1–64.8

6)

55.4

6(4

3.7

1–67.2

1)

55.8

8(4

2.8

1–68.9

5)

53.4

4(4

3.1

5–63.7

2)

0.0

5

MC

Apeak

syst

olic

flow

,m

l/s

6.5

8(5

.01–8.1

4)

7.9

8(6

.08–9.8

8)

8.6

7(6

.95–10.4

0)

8.3

6(6

.43–10.2

9)

7.7

67

(5.6

8–9.8

5)

8.0

0(5

.67–10.3

2)

0.1

01

MC

Am

ean

flow

,m

l/s

4.4

9(3

.49–5.4

9)

5.4

9(4

.49–6.4

9)

5.1

4(4

.27–6.0

2)

5.4

4(4

.47–6.4

2)

5.1

5(3

.95–6.3

6)

4.8

2(3

.88–5.7

5)

0.2

92

Blo

od

oxyg

en

conte

nt,

mlO

2/d

Lblo

od

20.2

0(1

9.5

0–20.9

0)

16.8

8(1

5.7

9–17.9

6)

16.8

6(1

6.1

8–17.5

5)

17.1

4(1

6.0

8–18.2

)17.5

0(1

6.4

0–18.6

0)

18.1

5(1

7.3

1–18.9

8)

<0.0

01

Oxyg

en

deliv

ery

,m

l/s

9.0

7(7

.09–11.0

5)

9.2

5(7

.47–11.0

3)

8.7

0(7

.13–10.2

8)

9.3

7(7

.42–11.3

3)

8.9

1(6

.90–10.9

3)

11.1

2(7

.63–14.6

1)

0.2

94

Whole

bra

inA

DC

valu

es,

AD

Cva

lues,�

10-4

/mm

3

8.9

0(8

.64–9.1

5)

9.0

8(8

.75–9.4

1)

9.1

3(8

.88–9.3

8)

9.1

4(8

.98–9.3

1)

9.1

9(8

.94–9.4

4)

9.4

0(9

.20–9.6

0)

<0.0

01

AC

Ate

rritory

AD

C(F

ronta

l)va

lues,

�10-4

/mm

3(R

&L

Mean

)

5.3

8(4

.91–5.8

5)

5.7

0(5

.20–6.2

0)

5.7

5(5

.37–6.1

3)

5.6

2(5

.23–6.0

1)

5.6

4(5

.16–6.1

3)

5.6

9(5

.24–6.1

5)

0.1

2

MC

Ate

rritory

(LN

)A

DC

valu

es,

�10-4

/mm

3(R

&L

Mean

)

6.8

7(6

.50–7.2

3)

7.1

5(6

.86–7.4

3)

7.0

3(6

.69–7.3

7)

7.2

1(6

.77–7.6

5)

7.0

4(6

.70–7.3

7)

7.1

8(6

.79–7.5

8)

0.1

57

PC

Ate

rritory

(Occ

ipital

)A

DC

valu

es,

�10-4

/mm

3(R

&L

Mean

)

6.1

9(5

.58–6.7

9)

6.1

7(5

.38–6.9

7)

6.2

0(5

.46–6.9

4)

6.3

2(5

.72–6.9

2)

6.2

5(5

.56–6.9

3)

6.4

7(5

.88–7.0

5)

0.0

61

Cere

bella

rA

DC

valu

es,

�10-4

/mm

3(R

&L

Mean

)

7.4

4(7

.13–7.7

5)

7.4

6(7

.17–7.7

6)

7.4

5(7

.03–7.8

6)

7.4

3(7

.12–7.7

3)

7.4

2(7

.17–7.6

6)

7.6

0(7

.34–7.8

7)

0.7

61

Corp

us

callo

sum

(Genu)

AD

Cva

lues,

�10-4

/mm

3(R

&L

Mean

)

16.9

0(1

5.9

6–17.8

5)

17.2

2(1

6.3

2–18.1

3)

17.4

1(1

6.6

6–18.1

6)

17.5

4(1

6.7

0–18.3

9)

17.0

8(1

6.0

2–18.1

4)

18.2

5(1

7.4

0–19.1

0)

<0.0

01

Corp

us

callo

sum

(Sple

niu

m)

AD

CVal

ues,

�10-4

/mm

3(R

&L

Mean

)

20.3

1(1

9.3

0–21.3

1)

20.6

2(1

9.5

0–21.7

5)

21.0

5(1

9.8

7–22.2

3)

20.7

1(1

9.7

3–21.6

9)

22.1

7(2

1.1

1–23.2

4)

21.8

9(2

0.7

8–22.9

9)

<0.0

01

Intr

acra

nia

lC

SFvo

lum

e,

ml

283.7

2(2

63.3

1–304.1

2)

278.9

9(2

52.3

0–305.6

8)

282.5

8(2

53.7

8–311.2

8)

281.4

9(2

56.7

4–306.2

5)

276.9

0(2

51.8

4–301.9

6)

269.6

1(2

48.1

4–291.0

8)

0.0

05

Gra

ym

atte

rvo

lum

e,

ml

679.0

2(6

36.6

3–721.4

0)

694.4

6(6

48.9

4–739.9

8)

685.8

9(6

46.8

1–724.9

7)

691.8

8(6

52.8

9–738.8

6)

687.3

7(6

47.4

8–727.2

7)

695.3

2(6

51.1

1–739.5

2)

0.0

02

White

mat

ter

volu

me,

ml

573.9

0(5

34.1

7–613.6

2)

574.2

4(5

31.5

8–616.8

9)

577.4

5(5

34.1

8(6

20.7

1)

575.1

1(5

34.5

4–615.6

7)

579.3

5(5

35.0

7–623.6

3)

586.7

6(5

44.7

5–628.7

7)

<0.0

01

Tota

lbra

inpar

ench

ymal

volu

me,m

l1252.9

1(1

175.9

4–1329.8

8)

1268.6

9(1

184.3

7–1353.0

1)

1263.3

4(1

187.2

2–1339.4

6)

1270.9

9(1

190.3

8–1351.5

9)

1266.7

3(1

190.9

2–1342.5

4)

1282.0

7(1

201.4

9–1362.6

5)

0.0

02

Superi

or

sagi

ttal

and

tran

svers

e

sinus

volu

me,m

l

10.5

1(0

.07–11.9

4)

10.9

0(9

.11–12.6

8)

10.3

9(9

.00–11.7

8)

10.2

3(8

.87–11.5

9)

10.2

1(8

.93–11.4

8)

10.2

1(8

.67–11.7

5)

0.6

92

Deep

cere

bra

lve

nous

volu

me,

ml

115.9

1(1

06.7

7–125.0

5)

106.7

4(9

9.0

0–114.4

8)

110.1

6(9

8.7

1–121.6

2)

105.2

1(9

4.5

4–115.8

8)

97.3

4(8

8.8

4–108.4

3)

0.0

03

CSF

:ce

rebro

spin

alflu

id;A

DC

:ap

par

ent

diff

usi

on

coeffic

ient;

AC

A:an

teri

or

cere

bra

lar

tery

;M

CA

:m

iddle

cere

bra

lar

tery

;PC

A:post

eri

or

cere

bra

lar

tery

;C

I:co

nfid

ence

inte

rval

s.

6 Journal of Cerebral Blood Flow & Metabolism

measured variables, confidence intervals, and the sig-nificance of change at each time point.

Ten of the 12 subjects (including the two who left thestudy early) developed a total Lake Louise score of 3 ormore at any individual time point (i.e. had a diagnosisof AMS). The mean total Lake Louise and headachescores for all subjects were seen to rise steadily at eachtime point, with the largest increase at 22 h.

SpO2 and ETCO2 significantly decreased at all hyp-oxic time points compared to baseline, whereas theheart rate rose significantly at all-time points comparedto baseline (p< 0.001 for each of the three variables).The systolic and diastolic blood pressures appeared tosteadily rise, although this trend was not statisticallysignificant (p¼ 0.085 and 0.068, respectively). No sig-nificant changes in any of the venous blood compo-nents were seen.

The mean MCA flow velocity increased and peakedafter 2 h of continuous hypoxia exposure (47.53 to57.09 cm/s; p¼ 0.05) before falling towards baselinevalues for the remainder of the study. The MCA diam-eter displayed a similar trend of an early rise beforedecreasing, although this was not statistically signifi-cant (p¼ 0.065). The calculated mean flow throughthe MCA did not increase or decrease significantlyover the 22-h course of the study when analyzedusing the linear mixed models procedure. However,performing paired t-test analyses showed that themean MCA flow rose significantly from 4.5 to6.1ml/s (p< 0.05) between the baseline and 2 h beforefalling towards baseline values. The calculated BOCwas consistently low at all-time points compared tothe normoxic, baseline levels (p< 0.001). CalculatedCOD did not change significantly throughout the hyp-oxic period.

The superior sagittal and transverse venous sinusvolumes did not change significantly in response tothe prolonged hypoxic challenge. However, by 22 h,there appeared to be an element of venocompressionat the level of the small, deep cerebral veins (116 cm3

at 2 h to 97 cm3 at 22 h; p< 0.05).A late increase in total brain parenchymal volume

was seen at 22 h compared to baseline (1253ml at base-line to 1282ml at 22 h; p< 0.05). Increases in both graymatter (679ml at baseline to 695ml at 22 h; p< 0.05)and white matter (574ml at baseline to 585ml at 22 h;p< 0.001) volumes contributed to the overall brainparenchymal volume increase. A compensatory signifi-cant decrease in total intracranial CSF was seen at 22 hcompared to baseline (284ml to 270ml; p< 0.05). Inaddition, the increase in total brain parenchymalvolume correlated strongly with the decrease in CSFvolume between 0 and 22 h (Pearson¼�0.67,p< 0.05) (Figure 2). Strong correlation was seenbetween the increase in white matter volume at 22 h

and the cumulative Lake Louise score for each patientat 22 h (Spearman rho¼ 0.65, p< 0.05) (Figure 3). Nosignificant correlation was observed between any othervariable and the headache or Lake Louise score.

Taking into account all the ADC ROIs at each timepoint, a significant rise in the mean whole brain ADCvalue was observed after exposure to 11 and 22 h of hyp-oxia (8.9� 10�4 to 9.2� 10�4 and 9.4� 10�4mm2/s,respectively; p< 0.001). With each regional ROI ana-lyzed separately, there were significant rises comparedto baseline in the ADC value within the genu(16.9� 10�4 at baseline to 18.3� 10�4mm2/s at 22 h;p< 0.001) and the splenium (20.3� 10�4 at baseline to22.8� 10�4 at 11 and 21.9� 10�4mm2/s at 22 h;p< 0.001). The rest of the ROIs did not display anysignificant increases or decreases in ADC value overtime, although there was an upward trend within theoccipital lobes (6.19� 10�4 at baseline to6.47� 10�4mm2/s; p¼ 0.06).

Discussion

The novel feature of this study is that arterial inflow,large and small vessel venous outflow, cerebral edema(ADC), brain parenchymal volumes, and intracranialCSF volumes have been measured repeatedly and sim-ultaneously over a 22-h period of sustained hypoxia.As a consequence, this study gives insights into thetime course of both the adaptive processes as well assome of the potentially maladaptive processes on acuteexposure to a FiO2¼ 0.12 (which is equivalent toapproximately 4400m). This is one of the first studiesto use 3T MRI scanning for more than 10 h and this isimportant as many of the more severe high-altitudesyndromes develop over a longer period of time – clas-sically after a night at the new altitude.

This combination of time (including overnight) andsimulated altitude represents the sort of stimulus thatwould be expected to cause a significant level of acutehigh-altitude illness and this was found to be the casewith most of the subjects. Ten of the 12 subjects devel-oped a total Lake Louise score of more than 3 (inkeeping with a diagnosis of ‘AMS’). The mean totalLake Louise and headache scores for all subjects wereseen to rise steadily at each time point, with the largestincrease at 22 h. COD was maintained by increases inMCA velocity and there is emerging evidence that thereis also an associated increase in MCA diameter.8,18,19

This is the first time that COD has been shown to bemaintained to 22 h when assessed by MR.

Late rises in ADC values within the genu and sple-nium of the corpus callosum confirm the build-up ofregional water content and are an indirect measure ofcerebral edema. This supports the findings observed byHackett et al.12 We found the observed increase in

Sagoo et al. 7

brain parenchymal volumes at 22 h was associated withdecreased brain CSF volumes (confirming ‘‘internalvalidation’’ of our methodology of independently mea-suring the segmented CSF and brain parenchymal vol-umes) and also a reduction in small venous vesselvolumes (SWI sequence) at 22 h, further increasingthe Starling pressure driving edema formation. Adecreased ability to buffer these parenchymal brainvolume changes will be reflected by rises in ICP andeventually HACE.3 Interestingly, a correlation wasfound between the cumulative Lake Louise score andan increase in brain white matter volume at 22 h. Thisstudy supports certain aspects of the recent Lawleyet al.20 publication where they found an increase inbrain volume (predominantly in the gray matter) at10 h, with an associated fall in the cerebral componentof CSF. They also found a statistically significant rela-tionship between change in ICP (measured non-inva-sively by MRI) and AMS severity (R2¼ 0.71, B¼ 2.3,p< 0.01) after 10 h. We studied our subjects for a fur-ther 12 h and found, as would be predicted, higherlevels of AMS symptoms. In addition, we found alater increase in total brain parenchymal volume (grayand white matter) at 22 h compared to baseline.

This study has shown for the first time a possiblemechanistic link between hypoxia, increased arterialinflow, a hypothesized rise in Starling pressures, andthe subsequent development of parenchyma brainedema and the clinical development of AMS. We,therefore, propose a unification hypothesis, whichinvolves the activation of train of etiologic factors pro-gressing from cerebral arterial to venous to parenchy-mal compartments, and also a clear staged time course(Figure 4):

1. Firstly, hypoxia results in increased CBF (in order tomaintain COD).8

2. The increased arterial inflow is associated with anincrease in venous outflow. The increase in venousoutflow requires increased ‘‘driving pressures’’ tomaintain CBF (raised arterial pressures).21

3. The venous outflow restriction or resistance couldtake place at a number of levels (Figure 5):a. Intracerebral as appears to be the case in this

study, which has not been demonstratedpreviously.

b. Intracranial (in the great veins and dural venoussinuses).16

Figure 2. A graph demonstrating the correlation between the percentage change of the total intracranial CSF volume and the total

brain parenchymal volume over the 22-h study period.

8 Journal of Cerebral Blood Flow & Metabolism

c. Extracranially (elevated CVP, anatomic venousrestrictions).3,22

The required increase in ‘‘driving pressure’’ will inturn perturbate ‘‘Starling pressures’’ and is likely to bea contributing factor in the formation of parenchymaledema. As edema develops, there will be further restric-tion in venous outflow and this could result in a dan-gerous positive feedback loop. Reduction of thecerebral capacitance will eventually occur resulting inan increase in ICP and, ultimately, AMS/HACE. Theincreased CBF will increase the likelihood of vasogenicedema3,19 and localized cellular hypoxia makes cellssusceptible to cytotoxic edema.23,24 It should be notedthat in this study the ‘‘venous outflow restriction’’appears to be associated with smaller venous vesselsrather than the large vessels / anatomical variantsdescribed by Wilson et al.16 (who found that restrictionin transverse venous sinus outflow was associated withmore headaches and proximal venous distensionbetween 90 and 180min of exposure to 12% oxygen).The observed decreasing trend for transverse venoussinus volumes in the current study was modest without

reaching significance and, therefore, could be attributedto a small study sample.

The experimental design using a hypoxic tent adja-cent to the MRI room functioned well. Subjects werecomfortable, were kept well hydrated and in constantcommunication with the investigators throughout thetime of exposure to hypoxia. Care was taken to main-tain hypoxia while transferring from the tent to thescanner and during the period of the scan.

Study limitations

Limitations of this study include a small but stableaccumulation of carbon dioxide in the tent, despitesoda lime scrubbers, which may have contributed tothe increase in peak systolic flow velocity in theMCA. However, during scanning, re-breathing wasnot an issue and there were low ETCO2 readings.There were a relatively small number of subjects forinvestigation of a variable condition such as AMSwith symptoms usually occurring 6–12 h after arrivalat high altitude. However, peak symptoms would beexpected to occur somewhat later after 24 h exposure.Nevertheless, two subjects were withdrawn after 11 h of

Figure 3. A graph demonstrating the correlation between the percentage volume change of the brain white matter and the

cumulative Lake Louise score at 22 h.

Sagoo et al. 9

hypoxia and significant symptoms occurred in theremainder. As the subjects were all young, resultscannot necessarily be extrapolated to older people. Inmost reports, young subjects are more prone to developAMS.25 One theory that older people with more age-related atrophy might tolerate cerebral swelling betterhas been suggested.

Wider clinical relevance

The results of this study may have important transla-tional implications.26 ICP is the primary neurologicalparameter that guides therapy on neurointensive care.Understanding the cerebrovascular interactions thateffect ICP is vital in the management of both traumaand stroke patients. The optimization of cerebral per-fusion pressure (CPP) is determined by ICP(CPP¼Mean Arterial Pressure – ICP). Intracranialvenous outflow pressure closely correlates with ICP.27

The effects of carbon dioxide and hyperventilationon CBF are well known.28 The effects of hypoxia, andin particular the changes that occur over a period oftime, may aid the understanding of the interplay

between chest and intracranial physiology. Hypoxia,for example secondary to chest sepsis or adult respira-tory distress syndrome is common in the neurocriticalpatient.29 As such, the management of cerebral oxygen-ation, as well as carbon dioxide are integral to goodneurocritical care and is likely to be crucial to optimiz-ing the correct therapeutic regime.

In this study, hypoxia resulted in parenchymal brainswelling. Subsequent development of cerebral edemawas found, which would eventually result in a rise inICP. This was attributed to compression of the cerebralvenous system, restricting venous outflow, and possiblya Starling resistor mechanism.30 In an animal model,venous hypertension has been associated with increasesin brain volume and this was caused by vessel dilata-tion. Cerebral edema only occurred if there was alsoassociated cerebral tissue injury.31 Acute hypoxia cancause tissue injury through a number of different mech-anisms such as direct damage to basal membrane struc-tures,32 VEGF,24 or free radicals.25

Interestingly, there is corroborative clinical evidence.It has been shown in a study of sub-arachnoid hemor-rhage (SAH) patients that increased ICP values

Figure 4. A diagram illustrating the proposed sequence of events thought to contribute towards the development of AMS.

Source: Wilson MH and Imray CHE. The cerebral venous system and hypoxia. J Appl Physiol.Epub ahead of print 20 August 2015.DOI:

10.1152/japplphysiol.00327.2015.

10 Journal of Cerebral Blood Flow & Metabolism

correlated with increased venous flow velocities andwere associated with poor outcome.33 Flow velocityof the transversal sinus was thought to be a highly sen-sitive, reliable, and early predictor of outcome in SAH.An explanation is that raised ICP would reduce intra-cerebral venous diameters and in turn, if CBF were tobe maintained, a rise in venous velocities would beexpected and this would result in changes inthe Starling pressures at the capillary level.

Conclusions

In conclusion, this 22-h study has used a variety of MRimaging techniques to document the development ofHACE in normal individuals. There was an earlyincrease in CBF which maintained COD. Mild clinicalsymptoms appeared after a few hours and these grad-ually increased in severity over time. Cerebral edemadeveloped later and this occurred in conjunction withan increase in the brain parenchyma volumes and areduction in CSF volumes. Inability to buffer this

brain swelling would result in rises in ICP and this islikely to be associated with deterioration in the patient’scondition. A possible novel mechanism involving localedema causing a reduction in post capillary venousdiameters has been observed. In order to maintain thehigh CBF through smaller vessels, there must be alter-ations in the Starling pressures, and these could be oneof the causative factors, in parallel with reduced cere-bral buffering, in the development of life threateningraised intracerebral pressures seen in HACE. Thesame mechanism may be implicated in raised ICPfrom other causes.

Funding

The JABBS foundation kindly supported the costs of this

study.

Declaration of conflicting interests

The author(s) declared no potential conflicts of interest with

respect to the research, authorship, and/or publication of thisarticle.

Figure 5. A diagram illustrating the intracranial and extracranial sources thought to contribute to venous outflow restriction.

Source: Wilson MH and Imray CHE. The cerebral venous system and hypoxia. J Appl Physiol.Epub ahead of print 20 August 2015. DOI:

10.1152/japplphysiol.00327.2015.

Sagoo et al. 11

Authors’ contributions

CI, AW, CEH, SW, MW, and RS contributed to study design.

SW was involved in producing the parameters for the variousMR sequences. EN performed the MR scanning for all sub-jects. CI, AW, and CH participated in clinical/physiologicaldata collection and analysis. RS, CEH, VS, SW, and SN con-

tributed to radiological data collection and analysis. HP andRS performed the statistical analysis. RS, CI, and AW did theliterature review and wrote the manuscript. All authors critic-

ally reviewed the final draft of the manuscript and have givenapproval for the version submitted.

References

1. Hackett PH andRoach RC. Current concepts: high altitude

illness – a practical review.NEngl JMed 2001; 345: 107–114.2. Hackett PH and Roach RC. High-altitude medicine. In:

Auerbach PA (ed.) Wilderness medicine 2000. St Louis:Mosby, pp.1–37.

3. Wilson MH, Newman S and Imray CH. The cerebral

effects of ascent to high altitudes. Lancet Neurol 2009;8: 175–191.

4. Imray C, Booth A, Wright A, et al. Acute altitude ill-nesses. BMJ 2011; 343: d4943.

5. Hultgren HN, Spickard WB, Hellriegel K, et al. High

altitude pulmonary edema. Medicine 1961; 40: 289–313.6. Hackett PH and Roach RC. High altitude cerebral

edema. High Alt Med Biol 2004; 5: 136–146.7. Ward M, Milledge JS and West JB. High altitude physi-

ology, 3rd ed. London: Arnold Publishers, 2000.8. Wilson MH, Edsell ME, Davagnanam I, et al. Cerebral

artery dilatationmaintains cerebral oxygenation at extreme

altitude and in acute hypoxia – an ultrasound and MRI

study. J Cereb Blood Flow Metab 2011; 31: 2019–2029.9. Giller CA. The emperor has no clothes: velocity, flow,

and the use of TCD. J Neuroimaging 2003; 13: 97–98.10. Wilson MH and Milledge J. Direct measurement of intra-

cranial pressure at high altitude and correlation of ven-

tricular size with acute mountain sickness. Neurosurgery2008; 63: 970–974.

11. Wright AD, Imray CHE, Morrissey MSC, et al.

Intracranial pressure at high altitude and acute mountainsickness. Clin Sci 1995; 89: 201–204.

12. Hackett PH, Yarnell PR, Hill R, et al. High-altitude cere-bral edema evaluated with magnetic resonance imaging:

clinical correlation and pathophysiology. JAMA 1998;

280: 1920–1925.13. Kallenberg K, Bailey DM, Christ S, et al. Magnetic res-

onance imaging evidence of cytotoxic cerebral edema in

acute mountain sickness. J Cereb Blood Flow Metab 2007;

27: 1064–1071.14. Houston CS and Dickinson J. Cerebral form of high-alti-

tude illness. Lancet 1975; 2: 758–761.15. Wilson MH, Imray CHE and Hargens AR. The headache

of high altitude and microgravity – similarities with clin-

ical syndromes of cerebral venous hypertension. High Alt

Med Biol 2011; 12: 379–386.16. Wilson MH, Davagnanam I, Holland G, et al. Cerebral

venous system and anatomical predisposition to high-alti-

tude headache. Ann Neurol 2013; 73: 381–389.

17. Roach RC, Bartsch P, Hackett PH, Oelz O. The Lake

Louise AMS Scoring Consensus Committee. The Lake

Louise acute mountain sickness scoring system. In:

Sutton JR, Coates G and Huston CS (eds) Hypoxia and

molecular medicine: proceedings of the 8th international

hypoxia symposium, Lake Louise, Alberta, Canada, 9–

13 February 1993, pp. 272–274. Burlington, VT: Queen

City Printers.18. Ogoh S, Sato K, Nakahara H, et al. Effect of acute hyp-

oxia on blood flow in vertebral and internal carotid

arteries. Exp Physiol 2013; 98: 692–698.

19. Imray C, Chan C, Stubbings A, et al. Time course vari-

ations in the mechanisms by which cerebral oxygen deliv-

ery is maintained on exposure to hypoxia/altitude. High

Alt Med Biol 2014; 15: 21–27.20. Lawley JS, Alperin N, Bagci AM, et al. Normobaric hyp-

oxia and symptoms of acute mountain sickness: elevated

brain volume and intracranial hypertension. Ann Neurol

2014; 75: 890–898.21. Imray CHE, Myers SD, Pattinson KTS, et al. Effects of

exercise on cerebral perfusion in humans at high altitude.

J Appl Physiol 2005; 99: 699–706.22. Willmann G, Fischer MD, Schommer K, et al. Missing

correlation of retinal vessel diameter with high-altitude

headache. Ann Clin Transl Neurol 2014; 1: 59–63.

23. Dorward DA, Thompson AA, Baillie JK, et al. Change in

plasma vascular endothelial growth factor during onset

and recovery from acute mountain sickness. Respir Med

2007; 101: 587–594.24. Bailey DM, Kleger GR, Holzgraefe M, et al.

Pathophysiological significance of peroxidative

stress, neuronaldamage, and membrane permeability

in acute mountain sickness. J Appl Physiol 2004; 96:

1459–1463.25. McDevitt M, McIntosh SE, Rodway G, et al. Risk deter-

minants of acute mountain sickness in trekkers in the

Nepali Himalaya: a 24-year follow-up. Wilderness

Environ Med 2014; 25: 152–159.

26. Imray CHE, Grocott MPW, Wilson MH, Hughes A,

Auerbach PS. Extreme, expedition and wilderness medi-

cine. The Lancet 2015; 386(10012): 2520–2525.27. Johnston IH and Rowan JO. Raised intracranial pressure

and cerebral blood flow. 3. Venous outflow tract pres-

sures and vascular resistances in experimental intracra-

nial hypertension. J Neurol Neurosurg Psychiatry 1974;

37: 392–402.28. Coles JP, Minhas PS, Fryer TD, et al. Effect of hyper-

ventilation on cerebral blood flow in traumatic head

injury: clinical relevance and monitoring correlates. Crit

Care Med 2002; 30: 1950–1959.

29. Lee K and Rincon F. Pulmonary complications in

patients with severe brain injury. Crit Care Res Pract

2012; 2012: 8.30. Simard JM, Kent TA, Chen M, et al. Brain oedema in

focal ischaemia: molecular pathophysiology and theoret-

ical implications. Lancet Neurol 2007; 6: 258–268.31. Cuypers J, Matakas F and Potolicchio SJ. Effect of cen-

tral venous pressure on brain tissue pressure and brain

volume. J Neurosurg 1976; 45: 89–94.

12 Journal of Cerebral Blood Flow & Metabolism

32. Miserocchi G, Passi A, Negrini D, et al. Pulmonary inter-stitial pressure and tissue matrix structure in acute hyp-oxia. Am J Physiol Lung Cell Mol Physiol 2001; 280:

L881–L887.

33. Niesen WD, Rosenkranz M, Schummer W, et al.

Cerebral venous flow velocity predicts poor outcome in

subarachnoid hemorrhage. Stroke 2004; 35: 1873–1878.

Sagoo et al. 13