http://www.diva-portal.org

This is the published version of a paper published in Ophthalmology (Rochester, Minn.).

Citation for the original published paper (version of record):

Lindén, C., Qvarlander, S., Jóhannesson, G., Johansson, E., Östlund, F. et al. (2018)Normal-Tension Glaucoma Has Normal Intracranial Pressure: A Prospective Study ofIntracranial Pressure and Intraocular Pressure in Different Body PositionsOphthalmology (Rochester, Minn.), 125(3): 361-368https://doi.org/10.1016/j.ophtha.2017.09.022

Access to the published version may require subscription.

N.B. When citing this work, cite the original published paper.

Permanent link to this version:http://urn.kb.se/resolve?urn=urn:nbn:se:umu:diva-142423

Normal-Tension Glaucoma Has NormalIntracranial Pressure

A Prospective Study of Intracranial Pressure and IntraocularPressure in Different Body Positions

Christina Lindén, MD, PhD,1 Sara Qvarlander, PhD,2 Gauti Jóhannesson, MD, PhD,1 Elias Johansson, MD, PhD,3

Fanny Östlund, MD,1 Jan Malm, MD, PhD,3 Anders Eklund, PhD2,4

Purpose: To test the hypothesis that normal-tension glaucoma (NTG) is caused by an increased pressuredifference across the lamina cribrosa (LC) related to a low intracranial pressure (ICP).

Design: Prospective case-control study.Participants: Thirteen NTG patients (9 women; median 71 [range: 56e83] years) were recruited for investi-

gation with the same protocol as 11 healthy volunteers (8 women; 47 [30e59] years). A larger control group(n ¼ 51; 30 women; 68 [30e81] years) was used only for ICP comparison in supine position.

Methods: ICP and intraocular pressure (IOP) were simultaneously measured in supine, sitting, and 9� head-down tilt (HDT) positions. Transelamina cribrosa pressure difference (TLCPD) was calculated using ICP and IOPtogether with geometric distances estimated from magnetic resonance imaging to adjust for hydrostatic effects.

Main Outcome Measures: ICP, IOP, and TLCPD in different body positions.Results: Between NTG patients and healthy volunteers, there were no differences in ICP, IOP, or TLCPD in

supine, sitting, or HDT (P � 0.11), except for IOP in HDT (P ¼ 0.04). There was no correlation between visual fielddefect and TLCPD, IOP, or ICP and in any body position (P � 0.39). Mean ICP in supine was 10.3 mmHg(SD ¼ 2.7) in the NTG group (n ¼ 13) and 11.3 (2.2) mmHg in the larger control group (n ¼ 51) (P ¼ 0.24).

Conclusions: There was no evidence of reduced ICP in NTG patients as compared with healthy controls,either in supine or in upright position. Consequently, the hypothesis that NTG is caused by an elevated TLCPDfrom low ICP was not supported. Ophthalmology 2018;125:361-368 ª 2017 by the American Academy ofOphthalmology. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/).

Glaucoma is a multifactorial disease that may be defined as aprogressive optic neuropathy with characteristic changes ofthe optic nerve head and the retinal nerve fiber layer, andcorresponding visual field defects. The pathogenesis is stillnot fully understood. Elevated intraocular pressure (IOP) wasonce thought to explain the disease. However, not all peoplewith elevated IOP develop glaucoma and patients withnormal-tension glaucoma (NTG) do not have elevated IOP.Thus, other causes have to be sought. An attractive hypothesissuggests that reduced intracranial pressure (ICP) is involvedin NTG. Although the published research data are limited,1e3

this hypothesis is currently actively discussed.4e10

The optic nerve fiber bundles leave the eye through thelamina cribrosa (LC), i.e., the perforated part of the sclera atthe posterior pole of the eye. The LC separates the intra-ocular compartment from the retrobulbar optic nerve, whichis surrounded by the cerebrospinal fluid (CSF) in a thin layerbetween the optic nerve and the optic nerve sheath (ONS).The CSF pressure/ICP is thus transferred to the retrobulbarregion and has an influence on the tissue pressure behind theLC. Analyzing the LC biomechanics with a simplified

ª 2017 by the American Academy of OphthalmologyThis is an open access article under the CC BY-NC-ND license(http://creativecommons.org/licenses/by-nc-nd/4.0/). Published by Elsevier Inc.

lumped system comparing ICP and IOP excludes importantlocal biomechanical aspects such as hoop stresses in theperipapillary sclera, changes in LC morphology, and tissueproperties causing variability in structural support andremodeling, all important in understanding how pressure-related stresses affect the optic nerve head and the path-ways for axons to leave the eye.11e13 Additionally, theretrolaminar tissue compartment is formed by the pia materaround the optic nerve, which is capable of bearing stress,14

and occlusion of the ONS around the optic nerve can cause adisrupted CSF communication.15 Pressure-related stressesmust, however, originate from pressure differences betweencompartments. The difference between IOP and ICP, i.e.,the transelamina cribrosa pressure difference (TLCPD ¼difference between IOP and ICP at the level of the LC), hasthe potential to be one of the contributors to abnormalstresses and strains in the optic nerve head.12,16

Elevated IOP is a risk factor for primary open-angleglaucoma; a high TLCPD would then be generated by ahigh IOP. But a high TLCPD could also be caused by a lowICP, and thus cause damage even when IOP is normal.

361https://doi.org/10.1016/j.ophtha.2017.09.022ISSN 0161-6420/17

Table 1. Characteristics of Normal-Tension Glaucoma Patients

Value (N [ 13 Patients)

Age (years) 71 (56e83)Gender (F/M) 9/4NTG diagnosis (years) 10 (4e16)Height (m) 1.65 (1.56e1.93)Weight (kg) 77 (60e100)Body mass index 24.7 (21.1e35.7)Blood pressure, systolic 138 (110e162)Blood pressure, diastolic 85 (62e90)Central corneal thickness (mm) 534 (493e626)Best-corrected visual acuity (logMAR) 0.04 (�0.08 to 1.30)Visual field index (%) 61 (7e90)

NTG ¼ normal-tension glaucoma.All values except gender are presented as median (range).

Ophthalmology Volume 125, Number 3, March 2018

Investigations of this hypothesis in humans are at presentlimited to performing ICP measurements and assuming aCSF communication along the optic nerve to the retrobulbarregion. With a lumped comparison between IOP and ICP itis not possible to investigate the local mechanical strains inthe LC and optic nerve head. The lumped comparison will,however, give information on the maximum potentialpressure difference across the LCeretrobulbar region andthereby test the physiological plausibility of the theory of areduced ICP as a component in NTG.

In previous studies of TLCPD, posture has not beenaccounted for. Measurement of IOP is traditionally performedin a sitting position, whereas lumbar puncture is performed ina lateral decubitus position, which is also the situation in theabove-mentioned studies investigating TLPCD in glaucoma.We recently showed that the reduction in ICP when movingfrom supine to sitting is large compared to the small reductionin IOP, corresponding to a higher TLCPD in sitting than insupine position.17 These findings imply that (1) in scientificevaluation of the TLCPD importance in glaucoma,assessment of IOP and ICP must be performed with thepatient in the same posture; and (2) TLCPD in uprightposition is unknown in glaucoma and has potential to givenew understanding of the pathophysiology. Consequently,the aim of this study was to determine if patients withNTG have a reduced ICP and an increased TLCPDcompared with healthy subjects.

Methods

Our objective was to measure both ICP and IOP simultaneously ina single experimental procedure, which includes both upright andsupine posture, to investigate if ICP and/or TLCPD in patients withNTG are different, compared with healthy subjects. We alsocompared ICP in supine position to reference data from a previ-ously published cohort of healthy subjects.

Patients

Charts of all patients diagnosed with NTG at the 3 hospitals withinthe County Council of Västerbotten were reviewed. Patients whomet the inclusion criteria were invited to participate in this pro-spective study. All study procedures were undertaken at UmeåUniversity Hospital between February and May 2016. The studywas registered at ClinicalTrials.gov (ID: NCT02776449). Thestudy protocol followed the tenets of the Declaration of Helsinkiand was reviewed and approved by the Regional Ethical ReviewBoard at Umeå University. All patients and healthy volunteerssigned a written informed consent after oral and written explana-tion of the study.

Patients had unilateral or bilateral NTG, defined as untreatedand treated IOP readings in the patient history of maximum21 mmHg with an occasional measurement up to 24 mmHg, opticdisc damage, and corresponding visual field defects of glaucoma-tous origin. Patients with neurologic diseases, medications knownto affect ICP (e.g., carbonic anhydrase inhibitors), previous brainsurgery, or previous lumbar puncture were excluded, as werepatients with previous glaucoma surgery. All included eyes wereon antiglaucoma eye drops, but all IOP-lowering medication wasterminated 28 days before the IOP-ICP study day.

Seventeen patients with NTG entered the study. One patientwas excluded after the magnetic resonance imaging (MRI), owingto a cyst affecting the spinal cord, and 1 patient withdrew at the

362

time of lumbar puncture. Two subjects were excluded owing tomean IOP values >21 mmHg registered at the IOP-ICP study day.Thus, the NTG group consisted of 13 patients, 11 of whom hadbilateral disease. Characteristics of the NTG subjects are presentedin Table 1.

Healthy Controls

Eleven healthy volunteers (3 men/8 women) with a median age of47 (range, 30e59) years were included and investigated with thesame protocol as the NTG patients. Original data were collected toinvestigate TLCPD and its body position dependence and is pro-vided in the report of Eklund et al.17 The control subjects had nopast or present neurologic or ophthalmologic disease. To confirmtheir health they had undergone an ophthalmologic andneurologic examination.

Specifically for ICP comparison in supine position, previouslypublished reference material of 40 healthy subjects18 was added tothe group of 11 healthy volunteers. In this group of 40 subjects noIOP measurements were performed and the ICP was only measuredin supine position, but with the same lumbar puncture and dataacquisition method as the one used in the present study. The 2reference materials were merged to form the supine ICP controlpopulation, resulting in a group of 51 healthy subjects (21 men/30 women) with a median age of 68 (range, 30e81) years. Theage of the NTG group was not significantly different from theage of the supine ICP control group (n ¼ 51) (P ¼ 0.26).

Baseline Examinations

To assess baseline ocular status and to verify patient eligibility, allpatients underwent a comprehensive baseline ophthalmologic ex-amination including medical history, weight, height, slit-lamp ex-amination of the anterior segment, and biomicroscopy of thefundus. Visual acuity was assessed with Early Treatment DiabeticRetinopathy Study (ETDRS) charts and measurement of centralcorneal thickness was performed with IOLMaster 700 R (CarlZeiss Meditec AG, Jena, Germany). IOP was measured with bothGoldmann applanation tonometry (Haag-Streit, Bern, Switzerland)and the Applanation resonance tonometer (ART) (BioResonatorGood Eye, Umea, Sweden). Visual field was evaluated withHumphrey Field Analyzer 3, HT24-2 SITA Fast program (CarlZeiss Meditec AG, Jena, Germany). The visual field index (VFI) inthat program expresses the visual field as a percentage of a normalage-adjusted visual field and was used to assess disease severity. Aneurologic examination was obtained from all participants and an

Lindén et al � Intracranial and Intraocular Pressure in NTG

MRI investigation including standard anatomic sequences wasperformed; both had to be normal for inclusion.

Investigational Protocol for the IntraocularPressureeIntracranial Pressure Study Day

Measurements of ICP and IOP were performed simultaneously insupine, sitting, and head-down tilt (HDT) positions. To control thebody positions, ranging from supine to sitting, the tilt function ofthe backrest of the investigational bed was used (the legs were kepthorizontal in all positions). The body position was alteredaccording to the following procedure: (1) 15 minutes in supine (0�tilt angle); (2) 25 minutes when the patient was brought from thesupine toward the sitting position in 5 steps at tilt angles 8�, 16�,24�, 32�, and 40�, each of 5-minute duration; (3) 7 minutes sitting(69� tilt angle); (4) back to supine position for 7 minutes; (5) 7minutes in the HDT position (�9� tilt angle). ICP was continu-ously recorded at all tilt angles, whereas IOP was measured atphases 1, 3, 4, and 5 (Fig 1). The second supine position was addedbecause repeated IOP applanation measurements have been shownto reduce IOP19,20 and therefore a new supine baseline wasacquired before HDT measurements. A lower IOP is thus to beexpected in second supine position compared with first, whereasthe approximately 30 minutes between first supine IOP and sittingposition should make the sitting IOP unaffected by the first supineIOP assessment.

A lumbar puncture (18-gauge needle) was performed in sittingposition and the needle connected to a standardized pressure trans-ducer (Likvor CELDA system, Likvor AB, Umea, Sweden),whereafter the patient was placed in supine position and the studyprotocol started. Using this standardized high-precision method tomeasure lumbar CSF pressure, it has previously been shown thatlumbar CSF pressure agrees with the ICP.21 In all postures the zero-pressure reference level for ICP was placed at the auditory meatususing a built-in horizontal laser line. Further details on the ICPmeasurements set-up and protocol are reported in Eklund et al.17 Ineach body position the ICP was allowed to stabilize for at least 3minutes before continuous recording for 2 minutes. The mean ICPfor that period was used as the ICP estimate for that position.

IOP was measured with a handheld ART with a supporting armthat was placed on the forehead. The measurement technique of ARTis independent of gravity and could be used in all body positions. TheART technique has been described in detail elsewhere.22,23 Topicalanesthesia (Lidokain-Fluorescein, lidocaine hydrochloride 4% andfluorescein sodium 0.25%, Chauvin Pharmaceuticals, Kingston-upon-Thames, UK) was instilled in each eye. Three IOP measure-ments were obtained for each eye after the time slot for ICPaveraging. The mean was used for analysis. In the analysis only theIOP and VFI of the worse eye (i.e., with the lowest VFI) was used.

TranseLamina Cribrosa Pressure DifferenceCalculation

As the IOP measurements are taken at the cornea, this location canbe considered the reference point for the IOP. To calculate theTLCPD, the ICP and IOP values were adjusted for the hydrostaticpressure differences between their respective reference points andthe LC. These new values were defined as ICPLC and IOPLC,respectively. By definition, hydrostatic pressure differences can becalculated from density (r), gravity (g), and vertical height (h) ofinternal fluid columns (rgh). In any pressure measurement oneshould therefore adjust for the difference in the vertical level be-tween the reference (measurement) level and the level where wewant to know the pressure. In this case, the vertical distance is tothe LC from the cornea in the case of IOP, and to the LC from theauditory meatus in the case of ICP. For ICP we used the auditory

meatus as the reference point because the measured lumbar CSFpressure was already hydrostatically adjusted with respect to theauditory meatus to estimate the ICP. The vertical distancescontributing to hydrostatic pressure effects were estimated basedon distances determined from anatomic MRI sequences (Fig 2).Patients were examined by MRI of the brain with a T2-weightedsequence using a 3 Tesla GE scanner with a 32-channel headcoil (GE Discovery MR750, General Electric Healthcare, Wauke-sha, WI). For each subject the images were analyzed to determinethe distance between the relevant locations, as measured in thesuperior-inferior (ICP) and anterior-posterior (ICP and IOP) di-rections. The resulting vertical distances in supine, sitting, andHDT positions were then calculated using these distances and theangle of the backrest of the bed (0�, 69�, and �9�). The TLCPDwas defined as the difference between IOPLC and ICPLC.

Statistical Analyses

The study had a sample size resulting in a power of 0.8 to detect anICP difference between NTG and healthy subjects of 2 mmHg orhigher, assuming within-group standard deviations of 1.75 mmHg.Statistical analysis was performed using SPSS (PASW Statistics18, SPSS, Chicago, IL). The Student paired t test was used forstatistical comparisons within the NTG group. For comparisonswith the control groups independent t tests were used (noassumption of equal variance). Correlation between VFI andpressure variables was determined with Pearson linear correlationcoefficients; additionally, a mixed-model analysis was performedso as to be able to assess these relationships based on the IOPvalues from both eyes of all subjects with bi-ocular glaucomatousinjury and in all body positions. Data are presented as meanvalues � standard deviation (SD) when not stated otherwise. AP value < 0.05 was considered statistically significant.

Results

ICP and IOP in the various positions are presented in Table 2. Therewas no difference in IOP or ICP between NTG and healthy subjects,except in the HDT position, where IOP was significantly lower inNTG than in healthy subjects. For the larger cohort of healthyelderly subjects (n ¼ 51), ICP in supine position was11.3�2.2 mmHg. This was different neither from the (first) supinevalue for the NTG subjects of 10.3�2.7 mmHg (P ¼ 0.24) norfrom the second supine value of 11.2�2.4 mmHg (P ¼ 1.00). TheICP values for NTG and healthy subjects (n ¼ 11) in all bodypositions investigated in the present protocol (includingintermediate tilt angles between supine and sitting where no IOPmeasurements were performed; Fig 1) are shown in Figure 3.ICP did not differ between the 2 groups in any body position(P ¼ 0.39e0.97).

Values for Pressures at the Level of the LaminaCribrosa

IOPLC, ICPLC, and TLCPD in supine and sitting position arepresented in Table 3. There were no significant differences betweenNTG patients and healthy controls in these parameters in any bodyposition (including the second supine measurement and HDT)(P > 0.05). There was no correlation between visual field defect(i.e., VFI) and TLCPD, IOP, or ICP in any body position (rangeof R: �0.26 to þ0.08, P � 0.39). Additionally, the mixed-modelanalysis did not reveal any significant relationship between VFIand IOP, TLCPD, or ICP (P > 0.10), either using a full factorialmodel (i.e., including interaction effect between variables) or usinga model limited to main effects.

363

Figure 1. Illustration of the investigational protocol showing the changes in body position and the timing of the intracranial pressure (ICP) and intraocularpressure (IOP) measurements; the total measurement protocol duration was about 1 hour.

Ophthalmology Volume 125, Number 3, March 2018

ICP showed a larger change than IOP with respect to posture(Table 2). Between supine and sitting position, both ICP and IOPdecreased significantly (mean change in NTG patients,ICP: �10.8�4.1 mmHg; IOP: �3.8�2.8 mmHg, P < 0.01)(Table 2), whereas TLCPD increased (mean change: 7.0�4.6mmHg, P < 0.01) (Table 3).

Between the second supine measurement and the HDTposition, ICP increased significantly (mean change: 3.7�3.0mmHg,P < 0.01), whereas IOP showed no significant change (meanchange: 0.0�1.2 mmHg, P ¼ 0.98) and TLCPD decreased signifi-cantly (mean change: �4.0�1.9 mmHg, P < 0.01).

Discussion

The hypothesis that glaucoma is caused by an increasedTLCPD has led to the assumption that NTG may be caused

Figure 2. Illustration of the calculation of the intracranial pressure adjustedreference point) and the lamina cribrosa (the ICPLC) and the intraocular pressureference point) and the lamina cribrosa (the IOPLC). The total vertical distdescribed by heye: distance between the cornea and the lamina cribrosa, and hdistances were determined based on anatomic distances, assessed from magneticdifferent positions (supine: 0�, sitting: 69�, and head-down tilt: �9�).

364

by a low ICP,1,2,6,8,10,24 which would create a similarpressure relationship between the ICP and IOP as anelevated IOP. The present study is the first to prospectivelyevaluate ICP and IOP simultaneously in both supine andupright positions in patients with a well-established diag-nosis of NTG and healthy controls. No evidence of reducedICP in NTG patients was found and consequently the cur-rent hypothesis could not be supported by our findings.

The literature on the hypothesis of an increased TLCPDhas a common limitation in that it is based on assessments ofIOP in upright posture and ICP in horizontal posture.Because both pressures vary with posture, the pressuredifference over LC has not been evaluated per se. Thesestudies have primarily investigated if patients with glau-coma have a reduced ICP when they are lying down.1e3,25

We spend 16 hours per day upright, and because ICP is

for hydrostatic gradients between the auditory meatus (ICP measurementre adjusted for hydrostatic gradients between the cornea (IOP measurementances between the lamina cribrosa and the pressure reference points areear: distance between the auditory meatus and the lamina cribrosa. Theseresonance imaging in supine position, and the tilt angles of the body in the

Table 2. Intracranial Pressure and Intraocular Pressure in Different Body Positions

Posture

ICP IOP

NTG (N ¼ 13) Healthy (N ¼ 11) NTG (N ¼ 13) Healthy (N ¼ 11)

Supine (first) 10.3 (2.7) 10.5 (1.5) 18.9 (3.1) 17.2 (1.8)Sitting �0.5 (2.7) �0.8 (3.8) 15.1 (3.2) 14.5 (2.3)Supine (second) 11.3 (2.4) 11.5 (0.8)y 15.4 (2.5) 16.0 (1.9)HDT 15.0 (2.6) 15.8 (1.2)z 15.4 (2.7)* 17.5 (2.0)

HDT ¼ head-down tilt; ICP ¼ intracranial pressure; IOP ¼ intraocular pressure; NTG ¼ normal-tension glaucoma.All values presented as mean (SD) in mmHg.*Statistically significant difference between NTG and healthy subjects.yN ¼ 8 healthy controls; exclusions of values in these body positions were made because of suspected unreliable ICP estimation owing to reduced cere-brospinal fluid contact (as indicated by very low ICP pulsations).zN ¼ 7 healthy controls; exclusions of values in these body positions were made because of suspected unreliable ICP estimation owing to reduced cere-brospinal fluid contact (as indicated by very low ICP pulsations).

Lindén et al � Intracranial and Intraocular Pressure in NTG

decreased substantially in upright posture (compared withhorizontal), very little is known of the true LC gradient.Accounting for the effect of postural dependency had thepotential to give new insights into glaucoma. We addressedthis knowledge gap in this study by (1) measuring ICP andIOP in both supine and upright positions, (2) assessing theintraposture differences in these pressures at the level of theLC, and (3) comparing NTG patients with a group ofhealthy volunteers investigated with the same protocol. Wefound no significant differences between NTG and healthysubjects, for ICP or TLCPD, regardless of body position.These findings indicate both that the craniospinal CSF

Figure 3. Intracranial pressure (ICP) at all evaluated body positions for the norm(black diamonds, n ¼ 11). The x-axis shows the tilt angle of backrest of the inposition, and �9� to head-down tilt position. The bars represent �1 SD.

dynamics are not disturbed in NTG and that the mainpostural control mechanism for ICP, the collapse of theinternal jugular veins,26,27 is intact in NTG patients andprobably not a part of the pathophysiology of glaucoma.This was further supported by the finding of no correlationbetween visual field defect and TLCPD, IOP, or ICP in anybody position.

Regarding the mean ICP in the horizontal posture, ourresults are in contradiction to most previous reports andreviews on ICP in glaucoma patients. Much of the literatureon this subject is in the form of reviews4,6,7,10,28 or based onindirect measurements of ICP,9,29e32 which is still not

al-tension glaucoma subjects (white circles, n¼ 13) and the healthy controlsvestigational bed, where 0� corresponds to supine position, 70� to sitting

365

Table 3. Adjusted Intracranial Pressure, Adjusted Intraocular Pressure, and TranseLamina Cribrosa Pressure Difference in Supine andSitting Positions

Posture

ICPLC IOPLC TLCPD [ IOPLC L ICPLC

NTG Healthy NTG Healthy NTG Healthy

Supine 7.0 (2.9) 6.6 (1.4) 20.7 (3.2) 18.9 (1.8) 13.7 (3.8) 12.3 (2.2)Sitting �4.9 (2.7) �4.7 (3.8) 15.8 (3.2) 15.1 (2.3) 20.7 (3.6) 19.8 (4.6)

ICPLC ¼ intracranial pressure adjusted for hydrostatic gradients between the auditory meatus (ICP measurement reference point) and the lamina cribrosa;IOPLC ¼ intraocular pressure adjusted for hydrostatic gradients between the cornea (IOP measurement reference point) and the lamina cribrosa;NTG ¼ normal-tension glaucoma; TLCPD ¼ transelamina cribrosa pressure difference.All values presented as mean (SD) in mmHg.NTG: n ¼ 13; healthy: n ¼ 11.

Ophthalmology Volume 125, Number 3, March 2018

considered reliable enough for clinical use,33e36 or case-studies.37e39 We have only identified 3 studies in whichdirect measurements of ICP were performed (Table 4). In aretrospective analysis Berdahl et al1 found a statisticallysignificantly lower ICP in NTG patients as compared withpatients with high-tension glaucoma or age-matched non-glaucoma controls from a primarily white dataset. Thatresult was later supported in a prospective study of a Chi-nese cohort.2 The control group in the first study consistedof people with refractive errors or cataract who hadundergone lumbar puncture for neurologic indicationunrelated to ophthalmic findings. The control group in thesecond study contained patients with various neurologicdiseases. In a recently published retrospective analysis ofNTG patients undergoing lumbar puncture duringcomputer-assisted cisternography, Pircher et al3 were notable to confirm either a reduced ICP or an increasedTLCPD. Because no control group was investigated, theycompared the results with those of previous studies. In thepresent study the control materials consisted ofprospectively collected data from healthy volunteers withno medical indication to perform lumbar puncture.

The similar ICP in NTG and healthy subjects was sur-prising to us, but it is noteworthy that our findings haddissimilarities to the studies by Ren et al2 and Berdahl et al1

regarding ICP primarily for the control group (Table 4). Astatistical comparison between the ICP in the NTGsubjects in the studies of Ren et al2 and Berdahl et al1 andthe present study reveals no significant difference (analysisof variance, P ¼ 0.62), whereas if we compare the ICP of

Table 4. Summary of Results from Studies Comparing IntracranialPressure in Normal-Tension Glaucoma and Controls

Author (Year) Design

NTG Controls

PN ICP N ICP

Berdahl et al1 (2008) Retrospective 11 9.3 (3.2) 68 12.7 (3.9) <0.01Pircher et al3 (2016) Retrospective 38 11.6 (3.7) na na naRen et al2 (2010) Prospective 14 9.5 (2.2) 71 12.9 (1.9) <0.001Present study Prospective 13 10.3 (2.7) 51 11.3 (2.2) ns

ICP ¼ intracranial pressure; na ¼ not available; ns ¼ not significant;NTG ¼ normal-tension glaucoma.

366

the corresponding control material there was a statisticallysignificant difference (analysis of variance, P < 0.01).Thus, the main difference, compared with previousstudies, is a lower ICP in the control group. We have useda standardized technique and specialized equipment forICP recording, used for 40 years for diagnostic purposesin idiopathic intracranial hypertension and inhydrocephalus.40,41 The protocol includes careful avoid-ance of leakage when placing the needle, setting the zerolevel very precisely to the auditory meatus with a horizontallaser beam, and then recording the CSF pressure for 15minutes with the patient in a restful supine position in asilent room. Average pressure during the last 2-5 minutes isused as ICP in supine position. Both control subjects andNTG patients were investigated with the same protocol. Weare therefore confident that the ICP assessment in the pre-sent study is reliable. The controls (n ¼ 68) from Berdahlet al,1 with a mean ICP of 12.7 mmHg and mean age of 68years, were extracted from a larger cohort (n ¼ 12 118) thatwas later published in a study of ICP vs. age.42 The laterstudy showed a mean value for the age span 60e80 yearsof 10.3 mmHg (n ¼ 4284), which is close to the ICP inthe NTG group of Berdahl et al1 and lower than the ICPin the control group (n ¼ 51) of the present study. Thecontrol population in the study by Ren et al2 revealed ahigh ICP compared with known normal data. In summary,this study does not support a difference in ICP betweenNTG and healthy subjects for the horizontal posture.

A pertinent question is what magnitude of pressure dif-ference would be clinically relevant. There is some evidencethat a 30% reduction of IOP significantly reduces the pro-gression of NTG.43,44 Assuming an average IOP in NTGpatients of 16 mmHg,44 a therapeutic effect may be detectedif the IOP is reduced by 5 mmHg. One could argue that acorresponding ICP reduction would be required in NTG.The present study was designed to detect a pressuredifference in ICP of 2 mmHg. Thus, based on thenegative findings in this study there is no indication of aclinically relevant pressure difference responsible forglaucomatous damage.

Our estimates of TLCPD assumes a free communicationbetween the intracranial CSF space and the LC in all bodypositions. It has been suggested and indicated in an animalstudy14 that in normal physiology the ONS occludes aroundthe optic nerve when ICP is reduced below the surrounding

Lindén et al � Intracranial and Intraocular Pressure in NTG

intraorbital pressure.45 ICPLC would then be constrained to aminimum level corresponding to the intraorbital pressure,producing a safety mechanism against high TLCPD inupright position. This occlusion effect would give aTLCPD in sitting position of about 12.5 mmHg,17 whichis close to the value we found in supine posture and muchlower than the value of about 20 mmHg that is predictedby a fully communicating CSF system (Table 3). Apathologic stiffening of the ONS could lead to incompleteocclusion and full fluid contact, and thereby maintainedhydrostatic coupling and pressure transfer between the LCand the intracranial CSF space in upright position. TheICPLC would then correspond to ICP rather than to theintraorbital pressure, resulting in a TLCPD of 20 mmHg(as determined in this study) rather than the 12.5 mmHgthat could be expected with an ONS that occludes fromintraorbital pressure. This indicates that a stiff ONS couldresult in abnormal TLCPD in upright position, giving apotential pathophysiological component in NTG, evenwith a normal upright ICP. However, it has beendemonstrated that the CSF turnover in the optic nervesubarachnoid space is reduced in papilledema and inNTG,15 and that the optic canal cross-sectional area issmaller.46 Both these findings indicate that there is occlusionand an inhibited fluid communication between theintracranial CSF space and optic nerve subarachnoid spacein NTG, rather than failure of the ONS occlusionmechanism.

In conclusion, we assessed ICP and IOP simultaneouslyin both supine and upright positions in NTG patients andhealthy controls and found no evidence of reduced ICP orincreased TLCPD in NTG patients, indicating that the ICPregulatory system or disturbed CSF dynamics are not themajor components of the NTG pathophysiology.

References

1. Berdahl JP, Fautsch MP, Stinnett SS, Allingham RR. Intra-cranial pressure in primary open angle glaucoma, normaltension glaucoma, and ocular hypertension: a case-controlstudy. Invest Ophthalmol Vis Sci. 2008;49(12):5412-5418.

2. Ren R, Jonas JB, Tian G, et al. Cerebrospinal fluid pressure inglaucoma: a prospective study. Ophthalmology. 2010;117(2):259-266.

3. Pircher A, Remonda L, Weinreb RN, Killer HE. Translaminarpressure in Caucasian normal tension glaucoma patients. ActaOphthalmol. 2017;95(7):e524-e531.

4. Berdahl JP, Allingham RR. Intracranial pressure and glau-coma. Curr Opin Ophthalmol. 2010;21(2):106-111.

5. Jonas JB, Ritch R, Panda-Jonas S. Cerebrospinal fluid pressurein the pathogenesis of glaucoma. Prog Brain Res. 2015;221:33-47.

6. Jonas JB, Wang N, Yang D, et al. Facts and myths of cere-brospinal fluid pressure for the physiology of the eye. ProgRetina Eye Res. 2015;46:67-83.

7. Marek B, Harris A, Kanakamedala P, et al. Cerebrospinal fluidpressure and glaucoma: regulation of trans-lamina cribrosapressure. Br J Ophthalmol. 2014;98(6):721-725.

8. Morgan WH, Balaratnasingam C, Lind CR, et al. Cerebrospinalfluid pressure and the eye. Br J Ophthalmol. 2016;100(1):71-77.

9. Lee SH, Kwak SW, Kang EM, et al. Estimated trans-laminacribrosa pressure differences in low-teen and high-teen intra-ocular pressure normal tension glaucoma: The KoreanNational Health and Nutrition Examination Survey. PLoS One.2016;11(2):e0148412.

10. Guy AH, Wiggs JL, Turalba A, Pasquale LR. Translating thelow translaminar cribrosa pressure gradient hypothesis into theclinical care of glaucoma. Semin Ophthalmol. 2016;31(1-2):131-139.

11. Downs JC, Girkin CA. Lamina cribrosa in glaucoma. CurrOpin Ophthalmol. 2017;28(2):113-119.

12. Stowell C, Burgoyne CF, Tamm ER, et al. Biomechanicalaspects of axonal damage in glaucoma: a brief review. Exp EyeRes. 2017;157:13-19.

13. Tamm ER, Ethier CR; Lasker/IRRF Initiative on Astrocytesand Glaucomatous Neurodegeneration Participants. Biologicalaspects of axonal damage in glaucoma: a brief review. Exp EyeRes. 2017;157:5-12.

14. Morgan WH, Yu DY, Alder VA, et al. The correlationbetween cerebrospinal fluid pressure and retrolaminar tissuepressure. Invest Ophthalmol Vis Sci. 1998;39(8):1419-1428.

15. Killer HE, Miller NR, Flammer J, et al. Cerebrospinal fluidexchange in the optic nerve in normal-tension glaucoma. Br JOphthalmol. 2012;96(4):544-548.

16. Feola AJ, Coudrillier B, Mulvihill J, et al. Deformation of thelamina cribrosa and optic nerve due to changes in cerebro-spinal fluid pressure. Invest Ophthalmol Vis Sci. 2017;58(4):2070-2078.

17. Eklund A, Johannesson G, Johansson E, et al. The pressuredifference between eye and brain changes with posture. AnnNeurol. 2016;80(2):269-276.

18. Malm J, Jacobsson J, Birgander R, Eklund A. Referencevalues for CSF outflow resistance and intracranial pressure inhealthy elderly. Neurology. 2011;76(10):903-909.

19. Bechrakis E. [On spontaneous decrease of pressure in appla-nation tonometry]. Ophthalmologica. 1966;151(5):604-614.

20. Johannesson G, Hallberg P, Eklund A, et al. Effects of topicalanaesthetics and repeated tonometry on intraocular pressure.Acta Ophthalmol. 2014;92(2):111-115.

21. Lenfeldt N, Koskinen LO, Bergenheim AT, et al. CSF pressureassessed by lumbar puncture agrees with intracranial pressure.Neurology. 2007;68(2):155-158.

22. Eklund A, Hallberg P, Linden C, Lindahl OA. An applanationresonator sensor for measuring intraocular pressure usingcombined continuous force and area measurement. InvestOphthalmol Vis Sci. 2003;44(7):3017-3024.

23. Johannesson G, Hallberg P, Eklund A, Linden C. Introductionand clinical evaluation of servo-controlled applanation reso-nance tonometry. Acta Ophthalmol. 2012;90(7):677-682.

24. Morgan WH, Yu DY, Cooper RL, et al. The influence ofcerebrospinal fluid pressure on the lamina cribrosa tissuepressure gradient. Invest Ophthalmol Vis Sci. 1995;36(6):1163-1172.

25. Berdahl JP, Allingham RR, Johnson DH. Cerebrospinal fluidpressure is decreased in primary open-angle glaucoma.Ophthalmology. 2008;115(5):763-768.

26. Qvarlander S, Sundstrom N, Malm J, Eklund A. Posturaleffects on intracranial pressure: modeling and clinical evalu-ation. J Appl Physiol (1985). 2013;115(10):1474-1480.

27. Holmlund P, Johansson E, Qvarlander S, et al. Human jugularvein collapse in the upright posture: implications for posturalintracranial pressure regulation. Fluids Barriers CNS.2017;14(1):17.

28. Roy Chowdhury U, Fautsch MP. Intracranial pressure and itsrelationship to glaucoma: current understanding and future

367

Ophthalmology Volume 125, Number 3, March 2018

directions. Med Hypothesis Discov Innov Ophthalmol.2015;4(3):71-80.

29. Jonas JB, Nangia V, Wang N, et al. Trans-lamina cribrosapressure difference and open-angle glaucoma. The centralIndia eye and medical study. PLoS One. 2013;8(12):e82284.

30. Jonas JB, Wang NL, Wang YX, et al. Estimated trans-laminacribrosa pressure difference versus intraocular pressure asbiomarker for open-angle glaucoma. The Beijing Eye Study2011. Acta Ophthalmol. 2015;93(1):e7-e13.

31. Siaudvytyte L, Januleviciene I, Ragauskas A, et al. The dif-ference in translaminar pressure gradient and neuroretinal rimarea in glaucoma and healthy subjects. J Ophthalmol.2014;2014:937360.

32. Wang N, Xie X, Yang D, et al. Orbital cerebrospinal fluidspace in glaucoma: the Beijing intracranial and intraocularpressure (iCOP) study. Ophthalmology. 2012;119(10):2065-2073.e1.

33. Bruce BB. Noninvasive assessment of cerebrospinal fluidpressure. J Neuroophthalmol. 2014;34(3):288-294.

34. Fleischman D, Bicket AK, Stinnett SS, et al. Analysis ofcerebrospinal fluid pressure estimation using formulae derivedfrom clinical data. Invest Ophthalmol Vis Sci. 2016;57(13):5625-5630.

35. Koskinen LD, Malm J, Zakelis R, et al. Can intracranialpressure be measured non-invasively bedside using a two-depth Doppler-technique? J Clin Monit Comput. 2017;31(2):459-467.

36. Zhang X, Medow J, Iskandar B, et al. Invasive and noninva-sive means of measuring intracranial pressure: a review.Physiol Meas. 2017;38(8):R143-R182.

37. Chen BH, Drucker MD, Louis KM, Richards DW. Progressionof normal-tension glaucoma after ventriculoperitoneal shunt to

368

decrease cerebrospinal fluid pressure. J Glaucoma.2016;25(1):e50-e52.

38. Yusuf IH, Ratnarajan G, Kerr RS, Salmon JF. Juvenile-onsetnormal tension glaucoma from chronic, recurrent low cere-brospinal fluid pressure. J Glaucoma. 2016;25(8):e738-e740.

39. Li Z, Yang DY, Lu Y, et al. Intracranial hypotension andco-existent normal-pressure glaucoma: the Beijing intracranialand intraocular pressure study. Chin Med J (Engl).2013;126(8):1588-1589.

40. Ekstedt J. CSF hydrodynamic studies in man. 2. Normal hy-drodynamic variables related to CSF pressure and flow.J Neurol Neurosurg Psychiatry. 1978;41(4):345-353.

41. Malm J, Sundstrom N, Cesarini KG, et al. Implementation of anew CSF dynamic device: a multicenter feasibility study in562 patients. Acta Neurol Scand. 2012;125(3):199-205.

42. Fleischman D, Berdahl JP, Zaydlarova J, et al. Cerebrospinalfluid pressure decreases with older age. PLoS One. 2012;7(12):e52664.

43. Collaborative Normal Tension Glaucoma Study Group. Theeffectiveness of intraocular pressure reduction in the treatmentof normal-tension glaucoma. Am J Ophthalmol. 1998;126(4):498-505.

44. Collaborative Normal Tension Glaucoma Study Group.Comparison of glaucomatous progression between untreatedpatients with normal-tension glaucoma and patients withtherapeutically reduced intraocular pressures. Am J Oph-thalmol. 1998;126(4):487-497.

45. Morgan WH, Yu DY, Balaratnasingam C. The role of cere-brospinal fluid pressure in glaucoma pathophysiology: the darkside of the optic disc. J Glaucoma. 2008;17(5):408-413.

46. Pircher A, Montali M, Berberat J, et al. The optic canal: abottleneck for cerebrospinal fluid dynamics in normal-tensionglaucoma? Front Neurol. 2017;8:47.

Footnotes and Financial Disclosures

Originally received: July 5, 2017.Final revision: September 14, 2017.Accepted: September 19, 2017.Available online: October 30, 2017. Manuscript no. 2017-1538.1 Department of Clinical Sciences, Ophthalmology, Umeå University,Umeå, Sweden.2 Department of Radiation Sciences, Biomedical Engineering, Umeå Uni-versity, Umeå, Sweden.3 Department of Pharmacology and Clinical Neuroscience, Neurology,Umeå University, Umeå, Sweden.4 Umeå Center for Functional Brain Imaging, Umeå University, Umeå,Sweden.

Presented at: Swedish Ophthalmological Society Annual Meeting, Umeå,Sweden, August 23e25, 2017.

Financial Disclosure(s):The author(s) made the following disclosure(s): J.M. and A.E.: Patents eLikvor, Umea, Sweden related to the CELDA device; Royalties e Likvor.

A.E.: Patent e ART IOP tonometry device assigned to BioResonatorGoodEye, Umea, Sweden.

Supported by the Swedish National Space Board, Swedish ResearchCouncil (grant no. 2015-05616), Swedish Society for Medical Research andVästerbottens Läns Landsting. The funding organizations had no role in thedesign or conduct of this research.

HUMAN SUBJECTS: Human subjects were part of this study protocol.The study was registered at ClinicalTrials.gov (ID: NCT02776449). Thestudy protocol followed the tenets of the Declaration of Helsinki and wasreviewed and approved by the Regional Ethical Review Board at UmeåUniversity. All patients and healthy volunteers signed a written informedconsent after oral and written explanation of the study.

Author Contributions:

Conception and design: Lindén, Malm, Eklund

Data collection: Lindén, Qvarlander, Jóhannesson, Johansson, Östlund,Malm, Eklund

Analysis and interpretation: Lindén, Qvarlander, Jóhannesson, Eklund

Obtained funding: Not applicable

Overall responsibility: Lindén, Qvarlander, Jóhannesson, Johansson,Östlund, Malm, Eklund

Abbreviations and Acronyms:ART ¼ applanation resonance tonometer; CSF ¼ cerebrospinal fluid;HDT ¼ head-down tilt; ICP ¼ intracranial pressure; IOP ¼ intraocularpressure; LC ¼ lamina cribrosa; MRI ¼ magnetic resonance imaging;NTG ¼ normal-tension glaucoma; ONS ¼ optic nerve sheath;TLCPD ¼ transelamina cribrosa pressure difference; VFI ¼ visual fieldindex.

Correspondence:Christina Lindén, MD, PhD, Department of Clinical Sciences, Ophthal-mology, Umeå University, SE-901 85 Umeå, Sweden. E-mail: christina.linden@umu.se.