a comparison of ocular blood flow in untreated primary open-angle glaucoma and ocular hypertension

A Comparison of Ocular Blood Flowin Untreated Primary Open-angle

Glaucoma and Ocular Hypertension

JAN KERR, FRCOPHTH, PATRICIA NELSON, MSC,AND COLM O’BRIEN, MD, FRCS, FRCOPHTH

● PURPOSE: To compare ocular blood flow inuntreated primary open-angle glaucoma and ocularhypertension using scanning laser Doppler flow-metry and pulsatile ocular blood flow.● METHOD: Fourteen ocular hypertensive subjectsand 10 patients with primary open-angle glaucomawere matched for intraocular pressure, mean arte-rial blood pressure, and age. They had scanninglaser Doppler flowmetry images taken centered onthe optic disk. Pulsatile ocular blood flow readingswere performed in sitting, standing, and supinepositions. No subjects were receiving topical anti-glaucoma treatment, systemic b-blockers, calciumantagonists, or nitrates at the time of measure-ment.● RESULTS: Laser Doppler flowmetry results showeda significant reduction in blood velocity, volume, andflow at the lamina cribrosa and the temporal neuro-retinal rim in glaucoma compared to ocular hyper-tension (P < .05). No difference was found betweenthe groups at the nasal neuroretinal rim or the nasaljuxtapapillary retina. There was a significant increasein minimum velocity (P 5 .03) at the temporaljuxtapapillary retina in glaucoma compared to ocularhypertension. The ocular pulse amplitude, pulsevolume, and pulsatile ocular blood flow were signif-

icantly lower (P < .05) in the glaucoma groupcompared to ocular hypertension in sitting and stand-ing positions.● CONCLUSION: Having controlled for factorsknown to affect perfusion pressure, we foundevidence of reduced ocular blood flow in primaryopen-angle glaucoma compared with ocular hyper-tension. Our findings indicate a reduction in cho-roidal and short posterior ciliary artery circulationin primary open-angle glaucoma. Whether thesechanges in blood flow are a cause or a consequenceof glaucomatous optic atrophy is still unknown.(Am J Ophthalmol 1998;126:42–51. © 1998 byElsevier Science Inc. All rights reserved.)

T HE PRESENCE OF VASCULAR ABNORMALITIES IN

primary open-angle glaucoma is now widelyaccepted. Abnormalities of blood flow in

primary open-angle glaucoma have been shownusing colour Doppler imaging,1,2 fluorescein angiog-raphy,3 laser Doppler flowmetry,4 and pulsatile ocu-lar blood flow measurements.5

Rheology studies have demonstrated differencesin glaucoma patients with increased red cell aggre-gability6 and increased plasma viscosity in primaryopen-angle glaucoma compared with normals.7 Wehave recently shown increased activation of theclotting system in untreated primary open-angleglaucoma.8 Pillunat and associates 9 demonstratedchanges in optic nerve head autoregulation inglaucoma patients, and Grunwald and associates10

found altered retinal autoregulation in the peri-macular area in primary open-angle glaucoma using

Accepted for publication Dec 5, 1997.From the Princess Alexandra Eye Pavilion, Edinburgh, Scotland,

United Kingdom. This research was supported by grants from theInternational Glaucoma Association, London; the Guide Dogs for theBlind Association, Reading; and the Ross Foundation, Edinburgh, UK.

Reprint requests to Colm O’Brien, MD, Department of Ophthalmol-ogy, The Mater Hospital, Eccles St, Dublin 2, Ireland; fax: 353-1-8305693.

© 1998 BY ELSEVIER SCIENCE INC. ALL RIGHTS RESERVED.42 0002-9394/98/$19.00PII S0002-9394(98)00074-9

the blue field entopic phenomenon. These findingsall point toward a disturbance in the circulation ofblood to the eye in primary open-angle glaucoma.Whether this is a primary cause of the disease orsimply one of a complex collection of secondarychanges that occur in glaucoma is not yet clear.

Approximately 5% of the population have anintraocular pressure greater than 22 mm Hg and areclassified as being “ocular hypertensive.” Although aminority of this group go on to develop glaucoma(study results suggest from 1.1%11 to 34%12 progressto glaucoma), there is a definite association betweenelevated intraocular pressure and glaucoma, withapproximately two-thirds of glaucoma patients hav-ing an intraocular pressure greater than 22 mmHg.13,14 Therefore, ocular hypertensives are theoret-ically a heterogeneous group composed of preglau-coma patients and healthy subjects with intraocularpressures in the upper 5% of the normal range.

There is some published data relating to ocularblood flow in ocular hypertension. Trew andSmith15 found no significant difference betweennormals and ocular hypertensives in comparingvalues for pulsatile ocular blood flow. By contrast,Loebl and Schwartz found evidence of impairedcirculation in ocular hypertensives using fluoresceinangiography. They noted an increased incidence offilling defects at the optic nerve head in ocularhypertensives compared with normals.16

The response of ocular blood flow to artificiallyelevated intraocular pressure in the acute situationhas been studied in animal models17,18 and inman.19,20 Best and Toyuku19 used a suction cup toelevate intraocular pressure in normal volunteersand found that the velocity of blood flow, asmeasured with fluorescein angiography, was reducedat increased levels of intraocular pressure. Thereduction in choroidal circulation was more markedthan that in the retina.19

Our aim was to compare ocular blood flow inuntreated primary open-angle glaucoma patientsand ocular hypertensives who were matched for ageand intraocular pressure, thus removing two impor-tant confounding factors. It has been demonstratedthat intraocular pressure21 and age22–24 may alterocular blood flow in the absence of disease. Wehoped to identify any intraocular pressure-indepen-dent changes in blood flow present in primary

open-angle glaucoma. We also removed the possibleeffects of treatment by recruiting newly diagnosedpatients who were not taking topical antiglaucomamedications and excluding patients who were tak-ing systemic drugs that had possible vascular effects.We used scanning laser Doppler flowmetry andpulsatile ocular blood flow to provide informationabout the ocular micro circulation and macro cir-culation, respectively.

PATIENTS AND METHODS

PATIENTS WERE NEW REFERRALS WHO WERE ATTEND-

ing the glaucoma clinic and were recruited for thestudy before starting treatment. Ethical approvalwas given for the study (Lothian Research EthicsCommittee), and informed consent was obtainedfrom all subjects. Glaucoma patients had intraocularpressures of greater than 22 mm Hg, optic diskcupping, and characteristic visual field defects whentested with the Humphrey visual field analyzer usingthreshold program 24-2. Ocular hypertensives hadintraocular pressures of greater than 22 mm Hg withnormal threshold visual fields. The ocular hyperten-sives had a wide range of optic disk appearances,with cup-to-disk ratios varying from 0.2 to 0.7. Allhad visual acuity of 20/30 or greater in the study eyeand no previous history of eye disease. No patientwas receiving topical treatment at the time ofmeasurement, and patients who were takingb-blockers, nitrates, and calcium channel blockerswere excluded from the study.

The subjects were chosen from a cohort of 20primary open-angle glaucoma patients and 50 ocularhypertensives on the basis of having good scanninglaser Doppler flowmetry images. We selected 10primary open-angle glaucoma and 14 ocular hyper-tensive patients, who were matched for age andintraocular pressure (Table 1). There were threepatients with a history of systemic hypertension inthe ocular hypertensive group, one of whom wastreated with bendrofluazide; the other two had notrequired treatment and were normotensive at thetime of measurement.

The Heidelberg retinal flowmeter was used toimage the fundus. The machine uses scanning laserDoppler flowmetry to quantify the circulation to the

OCULAR BLOOD FLOWVOL. 126, NO. 1 43

superficial layers of the optic nerve head and reti-na.25,26 The method, which has been shown to givereliable and reproducible results,26,27 combines laserDoppler flowmetry with a scanning laser system(wavelength, 780 nm) to produce pictures of thearea under investigation. The image produced is atwo-dimensional map of the retina with 256 3 64points representing retinal perfusion. The bright-ness of each point or pixel is coded by the value ofthe Doppler shift of laser light reflected from thatpoint; the brightest points represent the areas ofhighest flow.25 The observer acquires numericalinformation about “flow” (the distance traveled byall moving red blood cells inside the sample volumeper unit of time), “volume” (the number of movingred blood cells), and “velocity” (mean red blood cellspeed) by placing a “region of interest” square onthe image to select a sample volume 2,560 3 640 3400 mm in size.

We used a 2.5-degree 3 10-degree frame centeredon the optic disk and focused on the retinal vessels.Images were taken in the late morning. Pupils wereundilated, and ambient lighting was standardized.

For purposes of analysis, we divided the flowmeterimage into five areas (Figure): nasal retina, nasalneuroretinal rim, optic cup, temporal neuroretinalrim, and temporal retina.

One problem with the use of scanning laserDoppler flowmetry is the potential bias produced bysubjective positioning of the sample square on theimage. In our experience, it is possible to obtainvery different results within a given anatomical areasimply by moving the sample square a small dis-tance; this is true even when visible blood vesselsare avoided and occurs presumably because thecapillary network is not uniform. We attempted toreduce potential bias by increasing the number ofsamples taken in each area and recording themaximum and minimum sets of results (volume,velocity, and flow) obtained. Measurements weremade using a 10-pixel 3 10-pixel square, and asmany as 10 samples were taken in each area. Areasof peripapillary atrophy and visible blood vesselswere avoided when placing the square.

Pulsatile variation in intraocular pressure is pro-duced by the bolus of blood entering the eye during

TABLE 1. Group Characteristics

OHT POAG P Value

N 14 10

Sex, no. male/female 9/5 7/3 .77

Age (yrs) 63.4 6 7.5* 69.4 6 10.2 .13

IOP (mm Hg) 27.2 6 2.8 29.8 6 4.7 .10

Axial length (mm) 22.8 6 1.0 23.3 6 0.8 .29

Refractive errors (SEDS) 10.53 6 1.6 11.43 6 2.0 .27

BP (mm Hg) 115 6 18.5 109 6 16.5 .38

Systolic BP (mm Hg) 155 6 27 149 6 20 .53

Diastolic BP (mm Hg) 87 6 13.5 83 6 11.2 .45

Pulse pressure (mm Hg) 66 6 14.35 69 6 18.8 .75

Pulse rate (beats/min) 68 6 9 69 6 7 .76

Mean deviation (dB) 0.1 6 1.4 215.2 6 8.1 .0001†

CPSD (dB) 1.4 6 0.7 8.7 6 4.0 .0001†

Smokers (no.) 3 2 .90

History of hypertension (no.) 3 0 .12

MI/angina (no.) 0 0 1.00

Raynaud

syndrome/migraine (no.)

3 1 .50

OHT 5 ocular hypertension; POAG 5 primary open-angle glaucoma; IOP 5 intraocular pressure; SEDS 5 spherical equivalent diopters

sphere; BP 5 blood pressure; CPSD 5 corrected pattern standard deviation; MI 5 myocardial infarction.

*Mean 6 SD.†Significant (P # .05).

AMERICAN JOURNAL OF OPHTHALMOLOGY44 JULY 1998

cardiac systole. This ocular pulse can be quantifiedusing a pneumotonometer linked to an ocular bloodflow system, and a value for pulsatile ocular bloodflow can be derived. The calculation of pulsatileocular blood flow is based on two important assump-tions: (1) Langham’s pressure/volume relationshipfor the human eye is assumed to be standard for alleyes,28 and (2) it is assumed that venous outflow isconstant and not pulsatile.29 Choroidal blood flowaccounts for as much as 90% of ocular blood flow,18

so it has been suggested that pulsatile ocular bloodflow may give an indirect indication of choroidalflow. This suggestion is supported by the finding ofreduced pulsatile ocular blood flow in conditionsassociated with reduced choroidal volume such asretinitis pigmentosa,30 myopia, and choroidal atro-phy.31

For this study we used a system (OBF Labs,Swindon, UK) that we have found to give repro-ducible results.32 The pulsatile ocular blood flowmeasurements were made with the patients in sit-ting, standing, and supine positions. Blood pressurewas measured with an automatic cuff (CritikonDynamap, Tampa, Florida) in all three positions justbefore pulsatile ocular blood flow measurement.Patients were given 5 minutes’ rest in each positionbefore measurement. Intraocular pressure was mea-sured three times with a Goldmann tonometerbetween 9:30 AM and 2 PM, and an average of thesereadings was used for analysis.

Group comparisons were made using the unpairedStudent t test after ensuring that our results followeda normal distribution pattern. The chi-square testwas used for categorical data (SPSS package for PC,Chicago, Illinois).

For analysis of the scanning laser Doppler flow-metry results, we compared maximum with maxi-mum and minimum with minimum between thegroups (Table 2). The laser Doppler flowmeter givesresults for flow parameters as a value plus or minusone standard deviation. We looked at the standarddeviation for each result as an indicator of “vari-ance” within the sample area and compared thevalues in our two groups (Table 3). We correlatedthe scanning laser Doppler flowmetry results withvisual field loss (mean deviation on field testing) forthe primary open-angle glaucoma group using linearregression analysis. Statistical significance was set atP , .05.

RESULTS

TABLE 1 SUMMARIZES THE CHARACTERISTICS OF OUR

two groups. There was no significant difference inmean age, mean intraocular pressure, axial length,refractive error, blood pressure, or pulse rate. Thegroups had a similar incidence of cardiovasculardisease; there were three smokers in the ocularhypertensive group and two in the primary open-angle glaucoma group. Visual field results (meandeviation and corrected pattern standard deviation)were significantly different between the groups.

Regarding the optic cup/lamina cribrosa, wefound a significant reduction in minimum velocityand minimum flow in the optic cup in primaryopen-angle glaucoma compared with ocular hyper-tension (Table 2). The standard deviation or vari-ance of these results was also significantly lower inprimary open-angle glaucoma than in ocular hyper-tension (Table 3). There was a nonsignificant re-duction in all other parameters in primary open-angle glaucoma in this area.

Regarding the neoretinal rim, at the temporal rimin the primary open-angle glaucoma group, therewas a significant reduction in maximum and mini-mum volume, maximum velocity, and maximum

FIGURE. Schematic representation of the optic nervehead and peripapillary retina. The area of the retinalflowmeter image is within the hatched area. The num-bers indicate the different areas of measurement, asdiscussed in the text: 1, nasal retina; 2, nasal neuro-retinal rim; 3, optic cup/lamina cribrosa; 4, temporalneuroretinal rim; and 5, temporal retina.


flow (Table 2). We found no significant differencein blood flow at the nasal neuroretinal rim.

Regarding the juxtapapillary retina, in the temporalretina, there was a significant increase in minimumvelocity in primary open-angle glaucoma comparedwith ocular hypertension (Table 2). The standarddeviation of the minimum velocity and the minimumflow were significantly increased in primary open-angleglaucoma compared with ocular hypertension (Table3). All other parameters in this area were increased toa nonsignificant extent in primary open-angle glau-

coma. No significant difference was found between thegroups in the nasal retina.

We correlated our scanning laser Doppler flow-metry results with the visual field indices in theprimary open-angle glaucoma group and found asignificant positive correlation between maximumvolume at the lamina cribrosa and mean deviation(r 5 0.76, P 5 .01), with lower maximum volumecorresponding to worse mean deviation.

We found a significant reduction in pulse ampli-tude, pulse volume, and pulsatile ocular blood flow

TABLE 2. Heidelberg Retinal Flow Meter Results (Mean 6 SD)

Variable POAG (N 5 10) OHT (N 5 14) P Value

Nasal retina

Maximum volume 26.3 6 6.8 26.2 6 5.1 .98

Minimum volume 14.1 6 3.7 15.9 6 0.8 .17

Maximum velocity 1.9 6 0.6 1.7 6 0.4 .25

Minimum velocity 1.0 6 0.3 1.0 6 0.3 .79

Maximum flow 542.3 6 166 502.8 6 120 .50

Minimum flow 304.9 6 111 285.5 6 93 .62

Nasal rim

Maximum volume 16.3 6 4.7 21.01 6 4.8 .32

Minimum volume 10.2 6 3.3 13.6 6 6.8 .15



Maximum flow 504.4 6 205.6 557.7 6 374.9 .70

Minimum flow 320.3 6 119.1 365.7 6 202.6 .56

Optic cup

Maximum volume 17.0 6 5.7 23.4 6 13.8 .18

Minimum volume 7.9 6 2.0 11.9 6 8.9 .19


Minimum velocity 0.6 6 0.2 0.9 6 0.5 .03*

Maximum flow 331.1 6 81.3 434.7 6 200.6 .09

Minimum flow 168.8 6 61.8 272.1 6 126.4 .03*

Temporal rim

Maximum volume 15.1 6 3.1 22.9 6 5.4 .0001*

Minimum volume 11.5 6 1.9 15.1 6 4.4 .03*

Maximum velocity 1.4 6 0.6 1.9 6 0.6 .05*


Maximum flow 401.9 6 169.7 562.7 6 182.1 .04*

Minimum flow 329.4 6 111.4 366 6 152.4 .42

Temporal retina

Maximum volume 28.3 6 9.5 25.0 6 3.7 .25

Minimum volume 16.1 6 7.5 15.8 6 5.6 .90


Minimum velocity 1.1 6 0.3 0.9 6 0.3 .03*

Maximum flow 622.3 6 265.9 478.5 6 75.6 .06

Minimum flow 326.2 6 81.2 248.1 6 101.9 .06

POAG 5 primary open-angle glaucoma; OHT 5 ocular hypertension.

*Significant (P # .05).


in primary open-angle glaucoma compared withocular hypertension. Pulse amplitude was signifi-cantly reduced in all positions in primary open-angle glaucoma compared with ocular hypertension.Pulse volume and pulsatile ocular blood flow weresignificantly reduced in primary open-angle glau-coma in sitting and standing positions (P , .05) butnot in the supine position. There was a non-significant reduction in pulsatile ocular blood flowand pulse volume in the supine position in theglaucoma group (Table 4).

DISCUSSION

WE ARE NOT AWARE OF ANY OTHER STUDY THAT HAS

controlled for factors known to affect ocular perfu-sion pressure (intraocular pressure, age, blood pres-sure, and antiglaucoma treatment) before measuringdifferent parameters of ocular blood flow. We havefound evidence of reduced blood flow at the laminacribrosa and temporal neuroretinal rim and of in-creased flow in the temporal retina in untreatedprimary open-angle glaucoma compared with ocular

TABLE 3. Variance of Heidelberg Retinal Flow Meter Results (Mean 6 SD)

Variable POAG OHT P Value

Nasal retina

SD of maximum volume 15.4 6 2.8 14.7 6 4.1 .62

SD of minimum volume 11.84 6 2.4 12.0 6 3.0 .87

SD of maximum velocity 1.49 6 0.38 1.37 6 0.37 .45

SD of minimum velocity 1.19 6 0.3 1.0 6 0.3 .25

SD of maximum flow 435.9 6 97.6 417.0 6 130.0 .70

SD of minimum flow 345.4 6 97.8 307.7 6 88.8 .33

Nasal rim







Optic cup




SD of minimum velocity 0.83 6 0.22 1.3 6 0.45 .003*


SD of minimum flow 235.8 6 65.0 380.3 6 136.5 .003*

Temporal rim







Temporal retina




SD of minimum velocity 1.18 6 0.2 0.89 6 0.28 .01*


SD of minimum flow 342.9 6 62.4 255.9 6 89.8 .01*

POAG 5 primary open-angle glaucoma; OHT 5 ocular hypertension.



hypertension. Our pulsatile ocular blood flow resultsshow significantly lower pulse amplitude in allpositions and lower pulse volume and pulsatileocular blood flow in sitting and standing positions inprimary open-angle glaucoma compared with ocularhypertension.

There is very little published data comparingocular hypertension and primary open-angle glau-coma. Our findings concur with those of Trew andSmith,5 who found a significantly reduced pulsatileocular blood flow and pulse amplitude in primaryopen-angle glaucoma compared with ocular hyper-tension. Nicastro and associates,33 using a differentpneumotonometer, also found reduced ocular pulseamplitude in primary open-angle glaucoma com-pared with both normal controls and ocular hyper-tensives (primary open-angle glaucoma , N ,ocular hypertension). The primary open-angle glau-coma patients in that study were receiving topicaltreatment. In 1981, Perkins,34 using a recordingapplanation tonometer, found a significant differ-ence between primary open-angle glaucoma andocular hypertension, with a smaller pulse amplitudein (N , primary open-angle glaucoma , ocularhypertension), which concurs with our findings.

Until now, published data from scanning laserDoppler flowmetry has referred to normal subjectsfor comparison with those who have ocular hyper-

tension and primary open-angle glaucoma. Michel-son and associates4 compared patients with primaryopen-angle glaucoma with normals and found re-duced blood flow at the lamina cribrosa and nasaland temporal retina in the glaucoma patients. If weassume that our ocular hypertensive patients havemore “normal” blood flow than those with primaryopen-angle glaucoma, relating to their better opticnerve function, then our results agree with those ofMichelson and associates. Nicolela and associates,35

using scanning laser Doppler flowmetry, also foundreduced blood flow at the lamina cribrosa in thosewith primary open-angle glaucoma compared withnormals, with a nonsignificant increase in bloodflow at the temporal neuroretinal rim in primaryopen-angle glaucoma. They found reduced flow atsites on the temporal retina in patients with primaryopen-angle glaucoma compared with normals. Thesubjects in the study were a heterogeneous groupthat included normal pressure glaucoma patients;two-thirds of those in the group were also takingtopical antiglaucoma medication.

In contrast to previous studies, our subjects werenot taking any form of topical or systemic vasoac-tive medication. While some authors have foundthat topical treatment does not make a significantdifference to ocular blood flow,4,36 others havedemonstrated changes in flow produced by topicalantiglaucoma medication.37–39 The possibility re-mains that treatment may produce changes in localdistribution of blood flow or autoregulation.

Reduced retinal blood flow may be a function ofincreased intraocular pressure, and indeed this hasbeen demonstrated experimentally.40 In many pre-vious studies, patients with primary open-angleglaucoma have been compared with normals, butthe groups have not been matched for intraocularpressure.4,6,41 Because our groups were matched forintraocular pressure, we hope to have reduced theconfounding effect of this variable, to give a moreinformative comparison.

We encountered several technical problems inour use of the scanning laser Doppler flowmeter. Itwas difficult to obtain well-focused images withoutmovement artifact, particularly in elderly patientswith poor vision in the contralateral eye thatinterfered with fixation. We were unable to obtaingood images from approximately half of our original

TABLE 4. Pulsatile Ocular Blood Flow Results(Mean 6 SD)

OHT (n 5 14) POAG (n 5 10) P Value

Sitting

PA (mm Hg) 3.5 6 1.0 2.6 6 0.7 .025*

PV (ml) 4.6 6 1.4 3.4 6 0.9 .03*

POBF (ml/min) 635 6 161 479 6 149 .025*

Standing

PA (mm Hg) 3.1 6 0.9 2.2 6 0.4 .009*

PV (ml) 4.2 6 1.2 3.1 6 0.5 .01*

POBF (ml/min) 621 6 135 481 6 107 .01*

Supine

PA (mm Hg) 3.4 6 1.0 2.5 6 0.6 .04*

PV (ml) 4.4 6 1.3 3.5 6 0.9 .10

POBF (ml/min) 560 6 130 479 6 129 .15

OHT 5 ocular hypertension; POAG 5 primary open-angle

glaucoma; PA 5 pulse amplitude; PV 5 pulse volume;

POBF 5 pulsatile ocular blood flow.



primary open-angle glaucoma group of 20 patients.Nicolela and associates 35 reported similar problemsin a recent study.

In taking measurements from the images, theplacement of a 10-pixel 3 10-pixel square on thetemporal rim unavoidably includes an area of opticcup when the rim is thin, which would produce atendency for results at the temporal neuroretinalrim to be artificially lower in primary open-angleglaucoma than in ocular hypertension. Neverthe-less, we decided to use the 10-pixel square becauseprevious researchers have used this size of samplearea,4,24,35 and reliability and reproducibility studiesrelate to a 10-pixel 3 10-pixel square.26,27 A similarproblem was encountered when taking measure-ments from the optic cup in ocular hypertensivesubjects, that is, the sample area included neuroreti-nal rim if the cup was small and there was a risk ofartificially elevated blood flow readings at the opticcup in ocular hypertension. In an attempt to avoidthis problem, we did not include images of ocularhypertensive subjects who had an optic cup toosmall to allow placement of the sample square.Finally, when focusing on the retinal vessels, thefloor of the optic cup in primary open-angle glau-coma may be deeper and therefore farther from thefocal plane than in ocular hypertension. This wouldtend to produce lower results in primary open-angleglaucoma.

In comparing primary open-angle glaucoma withocular hypertension, it is important to rememberthat ocular hypertensives form a heterogeneousgroup that contains some preglaucoma patients.This study did not attempt to differentiate betweentypes of ocular hypertension but simply comparedtwo groups with similar intraocular pressure, ofwhich one, the patients with primary open-angleglaucoma, have had optic nerve damage producingdemonstrable visual field loss. It is likely that thechanges in blood flow identified in this study relateto the damaged state of the optic nerve in primaryopen-angle glaucoma. We can only speculate as towhether these changes contribute to the develop-ment of primary open-angle glaucoma or are simplya consequence of the disease. In criticism of thisstudy, our group sizes were small and the results mustbe viewed with this in mind. We did not include anormal control group because of the inability to

match by intraocular pressure. It is possible thatboth ocular hypertensive and primary open-angleglaucoma patients differ from normal.

Our results suggest that the choroidal and retinalcirculation respond in different ways in primaryopen-angle glaucoma, with a decrease in choroidalblood flow, as demonstrated by reduced scanninglaser Doppler flowmetry parameters at the laminaand reduced pulsatile ocular blood flow in theglaucoma group, and an apparent increase in retinalcirculation, as shown by the increased minimumvelocity at the temporal retina. The blood supply ofthe retina and choroid are anatomically different.The choroid derives its supply from the short pos-terior ciliary arteries which also supply the laminacribrosa, and the retinal supply comes from thecentral retinal artery. A difference also exists in thecontrol mechanisms of these two vascular beds. Thechoroid receives an autonomic nerve supply, andstudies have demonstrated increased choroidalblood flow in response to parasympathetic stimula-tion42 and reduced flow in response to stimulation ofthe cervical sympathetic chain.43 By contrast, theretinal vessels do not have an autonomic supply,and blood flow is autoregulated in response to localtissue demand and accumulation of metabolites.43

These anatomic and physiological differencesmay help to explain our findings. A reduction inparasympathetic activity in the choroid could pro-duce reduced choroidal blood flow leading to rela-tive retinal ischemia. In addition to causing damageto the photoreceptors and visual field loss, thisischemia may also produce a reactive autoregulatoryvasodilation in the retinal vessels. Autonomic dys-function has been previously suggested as a possibleunderlying abnormality in glaucoma.44 Jordan andassociates45 demonstrated parasympathetic denerva-tion hypersensitivity to dilute pilocarpine in glau-coma patients, and autonomic neuropathy has beensuggested as an explanation for the reported in-creased incidence of primary open-angle glaucomafound in diabetic patients.46,47

We suggest that our findings could be explainedby the hypothesis that primary open-angle glaucomais associated with autonomic dysfunction in whichrelative underaction of the parasympathetic systemproduces reduced choroidal blood flow and de-creased flow at the lamina cribrosa.


In the untreated state, retinal blood flow may beincreased in response to tissue hypoxia.

In summary, we have found an altered pattern ofocular blood flow in untreated primary open-angleglaucoma patients compared with ocular hyperten-sives, as measured by scanning laser Doppler flow-metry and pulsatile ocular blood flow. Whetherthese changes are a cause or a consequence ofglaucomatous optic atrophy is unknown. We suggestfurther research into local and systemic autonomicfunction in primary open-angle glaucoma in addi-tion to studies of the effect of treatment on ocularblood flow.

REFERENCES

1. Rankin SJA, Walmer BE, Buckley AR, Drance SM. ColorDoppler imaging and spectral analysis of the optic nervevasculature in glaucoma. Am J Ophthalmol 1995;119:685–693.

2. Butt Z, O’Brien C, McKillop G, Aspinall P, Allan P. ColorDoppler imaging in untreated high- and normal-pressureglaucoma. Invest Ophthalmol Vis Sci 1997;38:690–696.

3. Schwartz B, Rieser JC, Fishbein SL. Fluorescein angio-graphic defects of the optic disk in glaucoma. Arch Oph-thalmol 1977;95:1961–1974.

4. Michelson G, Langhans MJ, Groh MJM. Perfusion of thejuxtapapillary retina and the neuroretinal rim area inprimary open-angle glaucoma. J Glaucoma 1996;5:91–98.

5. Trew DR, Smith SE. Postural studies in pulsatile ocularblood flow: II.: chronic open-angle glaucoma. Br J Oph-thalmol 1991;75:71–75.

6. Hamard P, Hamard P, Dufaux J, Quesnot S. Optic nervehead blood flow using a Doppler velocimeter and haemo-rheology in primary open-angle glaucoma and normalpressure glaucoma. Br J Ophthalmol 1994;78:449–453.

7. Klaver JHJ, Greve EL, Goslinga H, Geijssen HC, Heuvel-mans JHA. Blood and plasma viscosity measurements inpatients with glaucoma. Br J Ophthalmol 1985;69:765–770.

8. O’Brien C, Butt Z, Ludlam C, Detkova P. Activation of thecoagulation cascade in untreated primary open-angle glau-coma. Ophthalmology 1997;104:725–730.

9. Pillunat LE, Stodtmeister R, Wilmanns I, Christ T. Auto-regulation of ocular bloodflow during changes in intraocu-lar pressure. Graefes Arch Clin Ophthalmol 1985;223:219–223.

10. Grunwald JE, Riva CE, Stone RA, Keates EU, Petrig BL.Retinal autoregulation in open-angle glaucoma. Ophthal-mology 1984;91:1690–1694.

11. Armaly MF. Ocular pressure and visual fields: a ten-yearfollow-up study. Arch Ophthalmol 1969;81:25–40.

12. Lundberg L, Wettrell K, Linner E. Ocular hypertension: aprospective twenty-year follow-up study. Acta Ophthalmol1987;65:705–708.

13. Klein BEK, Klein R, Sponsel WE, et al. Prevalence of

glaucoma: the Beaver Dam eye study. Ophthalmology1992;99:1499–1504.

14. Coffey M, Reidy A, Wormald R, Xing Xian W, Wright L,Courtney P. Prevalence of glaucoma in the West of Ireland.Br J Ophthalmol 1993;77:17–21.

15. Trew DR, Smith SE. Postural studies in pulsatile ocularblood flow. I. Ocular hypertension and normotension. Br JOphthalmol 1991;75:66–70.

16. Loebl M, Schwartz B. Fluorescein angiographic defects ofthe optic disk in ocular hypertension. Arch Ophthalmol1977;5:1980–1984.

17. Sossi N, Anderson D. Effect of elevated intraocular pressureon blood flow. Arch Ophthalmol 1983;101:98–101.

18. Alm A, Bill A. Ocular and optic nerve blood flow atnormal and increased intraocular pressure in monkeys(macaca iris): a study with radioactive labelled micro-spheres including flow determinations in brain and someother tissues. Exp Eye Res 1973;15:15–19.

19. Best M, Toyofuku H. Ocular haemodynamics during in-duced ocular hypertension in man. Am J Ophthalmol1972;74:932–939.

20. Blumenthal M, Best M, Galin MA, Gitter KA. Ocularcirculation: analysis of the effect of induced ocular hyper-tension on retinal and choroidal blood flow in man. Am JOphthalmol 1971;71:819–824.

21. Schwartz B, Kern J. Age, increased ocular and bloodpressures, and retinal and disk fluorescein angiogram. ArchOphthalmol 1980;98:1980–1986.

22. Williamson TH, Lowe GDO, Baxter GM. Influence of age,systemic blood pressure, smoking and blood velocity onorbital blood velocities. Br J Ophthalmol 1995;79:17–22.

23. Rojanapongpun P, Drance S. Velocity of ophthalmic arte-rial flow recorded by Doppler ultrasound in normal subjects.Am J Ophthalmol 1993;115:174–180.

24. Groh MJM, Michelson G, Langhans MJ, Harazny J. Influ-ence of age on retinal and optic nerve head blood circula-tion. Ophthalmology 1996;103:529–534.

25. Michelson G, Schmauss B. Two dimensional mapping ofthe perfusion of the retina and optic nerve. Br J Ophthal-mol 1995;79:1126–1132.

26. Michelson G, Schmauss B, Langhans MJ, Harazny J, GrohMJM. Principle, validity, and reliability of scanning laserDoppler flowmetry. J Glaucoma 1996;5:99–105.

27. Chung ES, Kim DM. Variability of measurements in theoptic nerve head and peripapillary retinal blood flow byscanning laser Doppler flowmetry. Invest Ophth Vis Sci1996;37:S264.

28. Eisenlohr JE, Langham ME, Maumenee AE. Manometricstudies of the pressure/volume relationship in living andenucleated eyes of individual human subjects. Br J Oph-thalmol 1962;46:536–548.

29. Silver DM, Farrell RA, Langham ME, 0’Brien V, SchilderP. Estimation of pulsatile ocular blood flow from intraocularpressure. Acta Ophthalmol 1989;67:25–29.

30. Langham ME, Kramer T. Decreased choroidal blood flowassociated with retinitis pigmentosa. Eye 1990;4:374–381.

31. James CB, Trew DR, Clark K, Smith SE. Factors influenc-ing the ocular pulse-axial length. Graefes Archive Oph-thalmol 1991;229:341–344.

32. Butt Z, O’Brien C. Reproducibility of pulsatile ocular bloodflow measurements. J Glaucoma 1995;4:214–218.

33. Nicastro NJ, Requa RJ, Campo RV, Barnet RW. Decreasedocular pulse amplitudes in glaucoma. Glaucoma 1987;9:45–52.


34. Perkins ES. The ocular pulse. Curr Eye Res 1981;1:19–23.35. Nicolela MT, Hnik P, Drance SM. Scanning laser Doppler

flowmeter study of retinal and optic disk blood flow inglaucomatous patients. Am J Ophthalmol 1996;122:775–783.

36. Nicolela MT, Buckley AR, Walman BE, Drance SM. Acomparative study of the effects of timolol and latanoproston blood flow velocity of the retrobulbar vessels. Am JOphthalmol 1996;122:784–789.

37. Morsman CD, Bosem ME, Lusky M, Weinreb RN. Theeffect of topical beta-blocking agents on pulsatile ocularblood flow. Eye 1995;9:344–347.

38. Boles Carenini B, Brogliatti B, Boles Carenini A. Pulsatileocular blood flow and antiglaucomatous drugs. New Trendsin Ophthalmol 1992;7:195–200.

39. Harris A, Spaeth GL, Sergott RC, Katz LJ, Cantor LB,Martin BJ. Retrobulbar arterial haemodynamic effects ofbetaxolol and timolol in normal tension glaucoma. Am JOphthalmol 1995;120:168–175.

40. Harris A, Joos K, Kay M, et al. Acute IOP elevation withscleral suction: effects on retrobulbar haemodynamics. Br JOphthalmol 1996;80:1055–1059.

41. Wolf S, Arend O, Sponsel WE, Schulte K, Cantor LB,Reim M. Retinal haemodynamics using scanning laserophthalmoscopy and haemorheology in chronic open-angle glaucoma. Ophthalmology 1993;100:1561–1566.

42. Stjernschantz J, Alm A, Bill A. Effects of intra-cranialoculomotor nerve stimulation on ocular blood flow inrabbits. Modification by indomethacin. Exp Eye Res 1976;23:461–470.

43. Bill A, Sperber GO. Control of retinal and choroidal bloodflow. Eye 1990;4:319–325.

44. Mapstone R, Clark CV. Autonomic neuropathy in ocularhypertension. Lancet 1985;2:185–187.

45. Jordan K, Perry SR, McCormack PJ, Downes RN. Cholin-ergic supersensitivity of the iris in primary open-angleglaucoma: a clinical study. Eye 1988;2:233–237.

46. Sigsbee B, Torkelson R, Kadis G, Wright JW, Reeves AG.Parasympathetic denervation of the iris in diabetes melli-tus. J Neurol Neurosurg Psych 1974;37:1031–1035.

47. Mitchell P, Smith W, Chey T, Stat M, Healy PR. Open-angle glaucoma and diabetes: the Blue Mountains study,Australia. Ophthalmology 1997;104:712–718.

Authors InteractivetWe encourage questions and comments regarding this article via theInternet on Authors Interactivet at http://www.ajo.com/ Questions, com-ments, and author responses are posted.


a comparison of ocular blood flow in untreated primary open-angle glaucoma and ocular hypertension

Documents