ORIGINAL ARTICLE

Measurement of Intraocular andIntracranial Pressure: Is There

a Relationship?

Tyler Kirk, MD,1 Keith Jones, MD,2,3 Spencer Miller, MD,2,4 and James Corbett, MD1,2

Objective: To study whether noninvasive, intraocular pressure (IOP) measurements significantly correlate withstandard intracranial pressure (ICP) measurements.Methods: This prospective, blinded study enrolled 46 patients who were undergoing medically indicated lumbarpuncture (LP). IOP was measured by applanation tonometry immediately prior to measuring LP opening pressure.One patient was excluded due to unsuccessful ICP measurement.Results: In the 45 patients to successfully undergo IOP and ICP measurement, there was no significantrelationship between ICP and average IOP for both eyes (r ¼ �0.005). There was no significant relationshipbetween ICP and IOP in either eye, when studied individually(r ¼ 0.03 ocular dexter [OD], r ¼ �0.05 ocularsinister [OS]). There was no significant relationship between ICP and IOP when the eye best correlated to thepatient’s ICP was chosen (r ¼ �0.01).Interpretation: No significant relationship between ICP and IOP was observed. Noninvasive IOP measurements donot predict ICP.

ANN NEUROL 2011;70:323–326

The most frequently utilized method for measuring in-

tracranial pressure (ICP) is lumbar puncture (LP).

This is a well-studied invasive study, which carries

uncommon, but significant, risks, including cerebral or

cerebellar herniation, introduction of meningitis through

a contaminated needle, post-LP headache, spinal epidural

hematoma, and the potential development of dermoid at

injection site.1 It would be preferable to avoid these risks

if a satisfactory alternative existed to accurately measure

ICP noninvasively and repeatedly.

Several studies have attempted to identify a useful,

accurate, noninvasive method of measuring ICP. Two pri-

mate studies revealed an increase in intraocular pressure

(IOP) when ICP was increased, but the IOP elevation

was only modest and without a clear corresponding ele-

vation in IOP to match ICP elevations, no clinical utility

was suggested.2,3 Two human studies have found a corre-

lation between IOP and ICP, but without a consistency

to suggest IOP measurements as a reliable alternative to

standard ICP measurement techniques.4,5 Until 2006, all

human studies investigating a potentially viable noninva-

sive method for ICP measurement proved fruitless.2–8

Then Sajjadi and colleagues9 reported IOP measurement,

by Schiotz tonometry, to be highly predictive of ICP as

measured by LP. A subsequent retrospective case series

study, Han and colleagues,10 responded to this new evi-

dence by showing no correlation between ICP and IOP,

as measured by Goldmann applanation tonometry. While

the bulk of evidence suggesting IOP is not an adequate

substitute for invasive ICP measurement heavily out-

weighs the strong evidence from the single study by

Sajjadi and colleagues,9 there has yet to be a prospective,

blinded study to challenge the validity of the Sajjadi and

colleagues9 study.

Considering our knowledge of the anatomic and

physiologic relationship between the eye and cerebrospinal

fluid, it seemed unlikely that the IOP should show a direct

relationship to ICP. Still, the call put forth by the Sajjadi

and colleagues9 study for a viable noninvasive measurement

of ICP was strong, and demanded further study.

View this article online at wileyonlinelibrary.com. DOI: 10.1002/ana.22414

Received Aug 6, 2010, and in revised form Feb 21, 2011. Accepted for publication Feb 25, 2011.

Address correspondence to Dr Kirk, MD, Department of Ophthalmology, University of Mississippi Medical Center, 2500 North State Street, Jackson, MS

39216. E-mail: tkirk@umc.edu

From the Departments of 1Ophthalmology and 2Neurology, University of Mississippi Medical Center, Jackson, MS; 3Baptist Neurological Associates,

Jackson, MS; and 4Nellis Air Force Base, Las Vegas, NV.

VC 2011 American Neurological Association 323

Patients and Methods

Forty-six consecutive patients who were undergoing a medically

indicated LP by their treating neurologist for different diagnoses

(Table) were consented to take part in this study. This prospec-

tive, blinded study was approved by the University of Missis-

sippi Medical Center Institutional Review Board. Patients pre-

senting to the University of Mississippi Medical Center from

January 2008 until February 2009 were recruited for this study

after their neurologist’s recommendation for LP.

Patients less than 18 years old, prisoners, mentally, emo-

tionally or developmentally challenged patients, patients with a

history of glaucoma or funduscopic findings suspicious for glau-

coma, those using IOP lowering drops or oral acetazolamide,

and patients unable to consent or unable to sit up were excluded

(n ¼ 7). One patient who consented for study participation was

not included in data analysis because LP was unsuccessful despite

attempts by standard and interventional radiologic approaches.

An ophthalmology resident physician obtained IOP by

applanation tonometry with either Goldmann or Perkins to-

nometer. These tonometers are regularly calibrated for accuracy

within 1mm Hg. The spinal fluid opening pressure was mea-

sured by manometer in the lateral recumbent position after spi-

nal puncture at the L3-4 or L4-5 vertebra under local anesthetic

by a neurology resident, with the exception of 3 cases that

required interventional radiologic (IR) assistance for successful

LP. The mean time between IOP and ICP measurement was

under an hour, with the longest time between measurements

being 3 hours in the cases requiring IR assistance. Recordings

for IOP were completed on separate forms from those used for

recording ICP and these forms were not viewed until the con-

clusion of the study by the primary investigator. Patients with

IOP greater or equal to 21mm Hg were offered follow-up for

ocular hypertension. The objective was to examine the relation-

ship between IOP and ICP.

Statistical significance was calculated by Pearson’s prod-

uct-moment correlation coefficient, using Microsoft Excel.

Results

Forty-five patients were included in the study. Thirty-five

were women. One of these subjects was excluded from

analysis due to unsuccessful ICP measurement. The

female majority was due to the predominant diagnosis of

idiopathic intracranial hypertension (IIH; Table).

The average body mass index (BMI) of our patients

was 33.8 6 10. Average LP opening pressure was 243 6108mm H2O. When converted to millimeters of mer-

cury, ICP ranged from 6.75 to 42mm Hg with a mean

of 18.25 6 8mm Hg. The IOP ranged from 6 to 25mm

Hg, with the combined average IOP for both eyes being

13.4 6 3.0mm Hg. The eye with the highest IOP read-

ing was used to calculate the average IOP which was

13.9 6 3.1mm Hg. The mean IOP for individual eyes

was 13.2 6 3.3 for the right eye and 13.4 6 2.8 for the

left. The largest pressure difference between eyes was

5mm Hg.

There was no significant correlation between the

average IOP for both eyes and ICP (r ¼ �0.005) Fur-

ther, there was no significant correlation between IOP

from either eye and ICP when studied individually (r ¼0.03 ocular dexter [OD], r ¼ �0.05 ocular sinister

[OS]). Finally, in selecting the IOP with the best

correlation to ICP for each patient, there still was no

statistically significant correlation between IOP and ICP

(r ¼ �0.01; Fig).

Discussion

Due to the serious risks involved with long term invasive

ICP monitors and repeated LP, there have been many

attempts to find a noninvasive surrogate means of moni-

toring ICP.2–8 Some of these studies have demonstrated

slight elevations in IOP related to elevated ICP, but not

to a comparable level that would allow us to use IOP

measurements an appropriate substitute for conventional

direct ICP measurement.4,5 Our study adds further evi-

dence to the generally accepted view that IOP cannot be

used as a substitute for direct ICP measurement.

However, in 2007, Sajjadi and colleagues9 demon-

strated a clinically useful IOP elevation as measured by

FIGURE: Scatter plot relating mean intraocular pressure andintracranial pressure (r 5 20.01, n 5 45).

TABLE: Diagnoses of Patients

Diagnosis n

Idiopathic intracranial hypertension 32

Anterior ischemic optic neuropathy 3

Headache 3

Optic neuritis 2

Bacterial meningitis 2

Systemic lupus erythematosus 1

Neurosyphilis 1

Unknown 1

ANNALS of Neurology

324 Volume 70, No. 2

Schiotz tonometry that was beyond the reference range

and was highly correlated with ICP elevation measured

by LP opening pressure. These findings run contrary to

all previous or subsequent reports studying the potential

relationship between ICP and IOP. Shortly after the Saj-

jadi and colleagues9 report, Han and colleagues10 rebut-

ted with a retrospective case series revealing no correla-

tion between ICP and IOP, measured by LP opening

pressure and Goldmann applanation tonometry respec-

tively. In 2009, Czarnik and colleagues11 also found no

relationship between ICP and IOP when measured

simultaneously in an intensive care setting among coma-

tose patients on continuous ICP measuring.

While our study design adds further weight to the

findings of Han and colleagues,10 our study did have

some limitations. In 3 of our patients more than 1 hour

elapsed between IOP and ICP measurements because

they required interventional assistance for successful LP.

This may have compromised the accuracy of IOP to

ICP relationships in these patients due to generally

accepted diurnal IOP variations. There is a low likeli-

hood that very large IOP changes occurred within 1

hour.12,13 Most IOP and ICP measurements were per-

formed by a single neurology and ophthalmology resi-

dent, but other residents in the ophthalmology and neu-

rology departments did perform measurements. While

in most patients IOP was measured by Goldmann

applanation tonometry, 9 patients could not be moved

to the eye clinic or could not sit upright underwent Per-

kins applanation tonometry. While there are small dif-

ferences in these 2 IOP measuring devices, studies have

shown the accuracy of their measurements to be nearly

equivalent.14

Our study and the study by Han and colleagues10

utilized Goldmann applanation tonometry for measure-

ment of IOP in primarily a sitting position. This method

of IOP measurement is accepted as the gold standard for

clinical IOP measurement, but several studies have docu-

mented that IOP rises when changing from the seated

upright position to the horizontal position. These pos-

tural changes are larger among glaucomatous eyes, which

were excluded from our study. All nonglaucomatous eyes

in studies have shown small elevations in IOP based on

body position changes, usually less than 2mm Hg in nor-

mal subjects.15 Sajjadi and colleagues9 measured IOP by

Schiotz tonometry in the horizontal position. No studies

have examined whether there are greater posture-induced

IOP changes in patients with elevated ICP.

IOP depends on aqueous production, episcleral

venous pressure, and trabecular meshwork flow.16 In the

case of elevated ICP, Sajjadi and colleagues9 theorize that

cerebrospinal fluid (CSF) encasing the optic nerve head

may communicate this elevated pressure into the eye.

There is neuroimaging evidence of posterior globe com-

pression due to elevated ICP,17 but compression of the

globe by increased ICP has not been shown to elevate

IOP in any study reported to this time. No study except

that by Sajjadi and colleagues9 has shown a clinically use-

ful correlation between elevated ICP and IOP.2–8 Corre-

sponding elevations in IOP among patients with IIH

have not routinely been reported, and elevated IOP is

not considered a symptom or sign of ICP. Based on their

findings, Czarnik and colleagues11 contend that there is

no pathophysiological relationship between ICP and IOP.

All studies prior to and succeeding Sajjadi and col-

leagues9 appear to corroborate the statement by Czarnik

and colleagues,11 including our own.

In our study, of 45 patients undergoing concurrent

measurement of ICP and IOP, there was no correlation

in readings. IOP measurement by Goldmann or Perkins

applanation tonometry is not an effective surrogate for

direct CSF pressure measurements of ICP.

Potential Conflicts of Interest

Nothing to report.

References1. Jankovic J. Laboratory investigations in diagnosis and manage-

ment of neurological disease. In: Bradley WG, Daroff RB, FenichelG, Jankovic J, eds. Neurology in clinical practice, 5th ed. Philadel-phia: Butterworth-Heinemann (Elsevier), 2008:451.

2. Lehman RA, Krupin T, Podos SM. Experimental effect of intracra-nial hypertension upon intraocular pressure. J Neurosurg 1972;36:60–66.

3. Smith RB, Aass AA, Nemoto EM. IOP and ICP during respiratoryalkolosis and acidosis. Br J Anaesth 1981;53:967–972.

4. Lashutka MK, Chandra A, Murray HN, et al. The relationship ofIOP to ICP. Ann Emerg Med 2004;43:585–591.

5. Sheeran P, Bland JM, Hall GM. IOP changes and alterations inICP. Lancet 2000;355:899.

6. Czarnik T, Gawda R, Latka D, et al. Noninvasive measurement ofICP: Is it possible? J Trauma 2007;62:207–211.

7. Firsching R, Schutze M, Motschmann M, Behrens-Baumann W.Venous ophthalmodynamometry: a non-invasive method for assess-ment of intracranial pressure. J Neurosurg 2000;93:33–36.

8. Firsching R, Schutze M, Motschmann M, et al. Non-invasive mea-surement of intracranial pressure. Lancet 1998;351:523–524.

9. Sajjadi SA, Harirchian MH, Sheikhbahaei N, et al. The relationbetween intracranial and intraocular pressures: study of 50patients. Ann Neurol 2006;59:867–870.

10. Han Y, McCulley TJ, Horton JC. No correlation between intraocu-lar and intracranial pressure. Ann Neurol 2008;64:221–224.

11. Czarnik T, Gawda R, Kolodziej W, et al. Associations between in-tracranial pressure, intraocular pressure and mean arterial pressurein patients with traumatic and non-traumatic brain injuries. Injury2009;40:33–39.

Kirk et al: IOP and ICP Relationship?

August 2011 325

12. David R, Zangwill L, Briscoe D, et al. Diurnal IOP variations: ananalysis of 690 diurnal curves. Br J Ophthalmol 1992;76:280–283.

13. Sacca SC, Rolando M, Marletta A, et al. Fluctuations of intraocularpressure during the day in open-angle glaucoma, normal-tensionglaucoma and normal subjects. Ophthalmologica 1998;212:115–119.

14. Tonnu PA, Ho T, Sharma E, et al. A comparison of four methodsof tonometry: method agreement and interobserver variability. BrJ Ophthalmol 2005;89:847–850.

15. Prata TS, De Moraes CG, Kanadani FN, et al. Posture-inducedintraocular pressure changes: considerations regarding body posi-tion in glaucoma patients. Surv Ophthalmol 2010;55:445–453.

16. Shields MB. Textbook of glaucoma. 4th ed. Baltimore: Williamsand Wilkins, 1998.

17. Madill SA, Connor SE. Computed tomography demonstrates shortaxial globe lengths in cases with idiopathic intracranial hyperten-sion. J Neuroophthalmol 2005;25:180–184.

ANNALS of Neurology

326 Volume 70, No. 2