Effect of Brimonidine on Optic NerveBlood Flow in Rabbits

ANUJA BHANDARI, MD, GEORGE A. CIOFFI, MD, E. MICHAEL VAN BUSKIRK, MD,SELIM ORGUL, MD, AND LIN WANG, PHD

● PURPOSE: Brimonidine is a highly selective a2-adreno-receptor agonist that lowers intraocular pressure. Theaim of the present study was to analyze in vivo thevasomotor effects and the influence of brimonidine onblood flow within the optic nerve, by means of intralu-minal microvascular corrosion casting technique andintravascular injection of colored microspheres.● METHODS: New Zealand white rabbits received eitherbrimonidine tartrate 0.2% or placebo (vehicle) topicaldrops in one eye for 4 weeks. Intraocular pressures weremeasured at baseline and 4 weeks. The anterior opticnerve microvasculature of four rabbits was examinedwith corrosion castings for regions of focal vasoconstric-tion. Optic nerve blood flow was determined in 16 rabbitsby means of nonradioactive colored microspheres.● RESULTS: The vasoconstriction values of the shortposterior ciliary arterial branches in the brimonidine eyeswere 16.7% 6 3.7%. In the fellow untreated eyes, themean vasoconstriction was 16.6% 6 2.4%. In theplacebo-treated eyes, the average constriction was 15.9%6 3.2%; the fellow eyes showed a mean constrictionvalue of 16.1% 6 5.3%. There was no statistical differ-ence between any of the groups (P 5 .2). The optic nerveblood flow in the brimonidine-treated rabbits was 0.18 60.06 ml/mg/min and 0.17 6 0.04 ml/mg/min in thetreated and the fellow eyes, respectively. The differencebetween the optic nerve blood flow in the brimonidine-treated eyes and the optic nerve blood flow in all of theuntreated eyes (0.19 6 0.06 ml/mg/min) also was notstatistically different (P 5 .82).● CONCLUSIONS: Long-term application of brimonidine0.2% does not affect the blood flow or vasomotor activityof the anterior optic nerve. (Am J Ophthalmol 1999;

128:601–605. © 1999 by Elsevier Science Inc. Allrights reserved.)

I N RECENT YEARS, IT HAS BEEN INCREASINGLY RECOG-

nized that although intraocular pressure is an importantrisk factor in the pathogenesis of glaucomatous optic

neuropathy, other factors, such as optic nerve circulation,may play a role.1–8 Therefore, in evaluating the safety andefficacy of an antiglaucoma medication, it is important toassess the intraocular pressure–lowering effect of the med-ication, as well as the potential vasoactive effects of thedrug on the optic nerve blood flow.

Comparative studies evaluating relatively selective andnonselective topical b-adrenergic antagonists have re-ported dissimilar influences in the sparing of visual fields inpatients with glaucoma.9–11 This effect appears to beindependent of the intraocular pressure–lowering effects.Differences in the ability of the various b-adrenergicantagonists to alter regional blood flow have been cited aspotential reasons for these differences in visual function.However, there is little evidence that these medicationsreach the optic nerve in pharmacologically active amountsor that they have vasoactive effects on the optic nervevasculature.

Recently, a2-adrenergic agonists have begun to be usedas antiglaucoma medications. Both brimonidine tartrateand apraclonidine hydrochloride have been shown toeffectively lower intraocular pressure.12–16 However, themechanisms of the ocular hypotensive action of apra-clonidine and brimonidine are different.12,15 Brimonidine,a highly selective a2-adrenoreceptor agonist that lowersintraocular pressure by reducing aqueous humor produc-tion and increasing uveoscleral outflow, has been shown tolower intraocular pressure over the long term.17 Apra-clonidine, a clonidine derivative that does not cross theblood–brain barrier, lowers intraocular pressure by decreas-ing the rate of aqueous humor production.18 Althoughtopical apraclonidine administration has been shown tocause anterior segment vasoconstriction in the ciliary bodyin the rabbit eye, it does not appear to have any vasomotoreffects on the microvasculature of the optic nerve aftereither short- or long-term administration.19,20 Other dif-

Accepted for publication June 3, 1999.From Discoveries in Sight, Devers Eye Institute, Legacy Health System,

Portland, Oregon.This study was supported in part by grant R01 EY05231 from the

National Institutes of Health, Bethesda, Maryland, and an unrestrictedgrant from Allergan, Inc, Irvine, California.

Reprint requests to G. A. Cioffi, MD, Discoveries in Sight, Devers EyeInstitute, 1040 NW 22nd Ave, Ste 200, Portland, OR 97210; fax: (503)274-4839; e-mail: gacioffi@lhs.org

© 1999 BY ELSEVIER SCIENCE INC. ALL RIGHTS RESERVED.0002-9394/99/$20.00 601PII S0002-9394(99)00223-8

ferences between these two a2-adrenergic agonists includevariable rates of ocular allergic reactions.

Given the pharmacologic differences between apra-clonidine and brimonidine, there may also be differences intheir vasoactivity in the optic nerve. The aim of the presentstudy was to analyze in vivo the vasomotor effects and theinfluence on local blood flow of brimonidine in the anterioroptic nerve, by means of intraluminal microvascular corro-sion casting technique and intravascular injection of coloredmicrospheres.

METHODS

NEW ZEALAND WHITE RABBITS OF EITHER SEX, WEIGHING

between 2.5 and 3.5 kg, received either brimonidinetartrate 0.2% (Alphagan; Allergan, Irvine, California) orplacebo drops (vehicle) in a randomly assigned eye. Thedrug was applied topically once a day to the assigned eyeand was administered for 4 weeks. Intraocular pressureswere measured (TonoPen; Oculab, Glendale, California)by averaging three consecutive measurements, each with5% reliabilty, at baseline and approximately 4 to 6 hoursafter the final dose of medication on the day of theterminal procedure. Intraocular pressures were measuredbefore the administration of anesthesia. The investigatorswere masked to both the medication group and the side oftreatment. The key for the treatment was not broken untilthe last experiment had been concluded. All experimentsconformed to the Association for Research in Vision andOphthalmology resolution on the use of animals in re-search.

For intraluminal microvascular corrosion castings, therabbits (n 5 4) were anesthetized with intravenous pen-tobarbital sodium. The castings of the ocular vasculaturewere obtained under controlled physiologic conditions,with mechanical ventilation provided by a small-animalrespirator while normal blood gas levels were maintained.Systemic blood pressure and rectal body temperature werecontinuously monitored and kept stable. Arterial bloodpressure was monitored via an indwelling arterial catheter.The detailed procedure has been described elsewhere.21

Batson no. 17 methylmethacrylate injection media, mod-ified to reduce the viscosity to approximately 11 centi-poise, was injected into the superior circulation throughthe ascending branches of the aorta. The injection pressurewas maintained at approximately 90 to 100 mm Hg for 15minutes, until the plastic began to polymerize and arrestflow in the vascular system. Two hours after injection, theeyes were enucleated, stored overnight in warm 10%formalin to complete the polymerization, and corroded in6-M potassium hydroxide. The resulting plastic vascularcastings were rinsed in running water and air dried.Whole-globe vascular castings were hemisected at theequator, and the posterior segments were mounted onaluminum stubs, sputter-coated with gold palladium, and

examined with a scanning electron microscope (EtecAutoscan, Hayward, California).

The optic nerve microvasculature in rabbits is primarilysupplied by the short posterior ciliary arteries via anarterial loop encircling the optic nerve.22 Direct arteriolarbranches from the short posterior ciliary arteries orbranches from the arterial circle that penetrate and supplythe optic nerve normally display focal zones of relativeconstriction near their branching point from the proximalarterial supply, compared with downstream vessel caliber.22

The primary and secondary arteriolar branches from thearterial circle were measured at these constriction zonesnear the branching point and 50 m distal to the end of theconstriction zone. The relative amount of branch pointconstriction at the focal zones near the branching pointwas expressed as a percentage constriction relative to thedistal diameter (100 3 [distal diameter 2 constricteddiameter/distal diameter]).

For blood flow measurements with colored microspheres,the rabbits (n 5 16; eight in each group) were anesthetizedwith intravenous pentobarbital sodium. This techniquehas been described in detail elsewhere.23 In brief, theanimals were placed on a thermostatically controlledheating pad to maintain the body temperature between38.0 C and 38.5 C. A tracheotomy was performed andmechanical ventilation provided by means of a small-animal respirator. Both brachial arteries were cannulatedwith a polyethylene catheter, one for continuous bloodpressure monitoring (connected to the mercury column ofa sphygmomanometer) and the other for collecting refer-ence blood samples and samples for blood gas analysis.Blood gases were controlled by adjusting ventilation.Colored polystyrene spheres 10.2 6 0.2 mm (mean 6 SD)in diameter (E-Z TRAC Ultraspheres; Interactive MedicalTechnologies Ltd, South Barrington, Illinois) were sus-pended in saline containing 0.05% Tween 80. After asternal thoracotomy, the left atrium was cannulated withpolyethylene tubing. One hundred million black micro-spheres were injected in a volume of 10 ml within 25seconds through the atrial catheter. This quantity ofmicrospheres has been shown to yield good reproducibilityfor optic nerve blood flow measurements in rabbits.23 Areference blood sample was collected during 1 minute froma catheter in the left brachial artery from the start of theinjection. No significant changes in blood pressure duringthe injection or collection period were observed. At theend of the collection period, the animal was euthanized bycross-clamping the ascending aorta. Both eyes were enu-cleated immediately.

The anterior optic nerves were carefully dissected fromthe eye under a dissecting microscope and weighed. Thetissue included the papilla, the region corresponding to thelamina cribrosa, and approximately 0.5 mm of a retrolami-nar segment. The tissues and the blood were chemicallydegraded and the microspheres were isolated by centrifu-gation. The microspheres in both optic nerves and in the

AMERICAN JOURNAL OF OPHTHALMOLOGY602 NOVEMBER 1999

reference blood sample were counted in a hemocytometer.The optic nerve blood flow was calculated by the followingformula: Ft 5 Fr 3 Ct/Cr, where Ft 5 regional blood flowin the tissue, Fr 5 reference sample flow rate (ml/min),Ct 5 microsphere count in tissue/weight of optic nervetissue, and Cr 5 microsphere count in reference bloodsample.

Statistical analysis was performed with a paired t test forthe interocular data and an unpaired t test between thegroups. Differences at a level of P , .05 were accepted asstatistically significant changes. The minimum sample sizeby means of this technique to detect the effect of bri-monidine on the blood flow has been calculated on thebasis of normal interocular variability from a previousstudy.23 Six pairs of eyes are required to detect a 20%difference between the treated and untreated eyes, whilesetting a significance level of 5%.

RESULTS

THE RABBITS WERE MAINTAINED WITHIN NORMAL PHYSIO-

logic measures, and arterial blood gases were constantlymonitored. Arterial partial pressure of oxygen was main-tained between 85 and 98 mm Hg, and partial pressure ofcarbon dioxide was maintained between 24 and 33 mm Hgfor all the animals being tested. For the colored micro-sphere blood flow experiments, the systolic blood pressurewas monitored through an indwelling arterial catheter andranged between 76 and 95 mm Hg. The baseline intraoc-ular pressure for all animals was 14.7 6 1.6 mm Hg(mean 6 SD). The average final intraocular pressure in thebrimonidine-treated eyes was 12.5 6 1.5 mm Hg, whichwas statistically lower than the pooled baseline intraocularpressures (P , .005) and statistically lower than thebaseline intraocular pressure in the treated eyes (P , .05).

In the study of intraluminal microvascular corrosioncasting, the average constriction values of the short pos-terior ciliary arterial branches in the brimonidine-treatedeyes was 16.7% 6 3.7% (mean 6 SD). In the fellowuntreated eyes, the mean vasoconstriction was 16.6% 62.4%. In the placebo-treated eyes, the average constrictionwas 15.9% 6 3.2%; the fellow eyes showed a meanconstriction value of 16.1% 6 5.3% (mean 6 SD).

The average (mean 6 SD) weight of the examinedtissue among the 32 eyes of the 16 rabbits used to assessoptic nerve blood flow was 2.25 6 0.46 mg. The opticnerve blood flow (mean 6 SD) in the brimonidine-treatedrabbits was 0.18 6 0.06 ml/mg/min (range, 0.08 to 0.27ml/mg/min) and 0.17 6 0.04 ml/mg/min (range, 0.11 to0.24 ml/mg/min) in the treated and fellow eyes, respec-tively. This difference was not statistically significant (P 5.34; paired t test). In the placebo-treated rabbits, the opticnerve blood flow (mean 6 SD) in the treated eyes was0.24 6 0.07 ml/mg/min (range, 0.14 to 0.35 ml/mg/min)and in the fellow eyes was 0.21 6 0.06 ml/mg/min (range,

0.10 to 0.28 ml/mg/min). This difference between theplacebo-treated eyes and the contralateral eyes of the sameanimals was not statistically significant (P 5 .26; paired ttest). The difference between the optic nerve blood flow inthe brimonidine-treated eyes (0.18 6 0.06 ml/mg/min)and the optic nerve blood flow in all of the untreated eyes(0.19 6 0.06 ml/mg/min) was also not statistically different(P 5 .82; unpaired t test).

DISCUSSION

USING NONRADIOACTIVE COLORED MICROSPHERES, SUP-

plemented with the erosion vascular casting techniques,we examined the effect of long-term topical application ofbrimonidine tartrate 0.2%. This study demonstrated thatbrimonidine does not alter the optic nerve blood flow inthe rabbit eye. These results are in keeping with those ofanother study in which there were no demonstrabledecreases in retinal and choroidal blood flow in rabbitswith topical brimonidine.24

The use of nonradioactive colored microspheres is anewly described technique for measuring optic nerve bloodflow in animals. This method has been shown to have goodreproducibility.23 Additionally, the ability of this tech-nique to demonstrate alteration in optic nerve blood flowsecondary to induced optic nerve ischemia has also beenshown in earlier experiments.25

The current therapeutic aim of most antiglaucoma medi-cations is to reduce the intraocular pressure and therebypreserve visual function. Topically applied a2-adrenergicagonists efficiently reduce intraocular pressure.12–16 However,the effectiveness of intraocular pressure reduction in theprevention of optic nerve damage in individuals with glau-coma remains a point of debate.26 Some patients withglaucomatous visual field defects and optic nerve damagenever exhibit increased intraocular pressure above the statis-tically derived upper normal limit. Additionally, progressionof glaucomatous damage seems to be weakly correlated to thelevel of intraocular pressure.27 Studies comparing the influ-ence of various b-blockers on visual function have suggestedthat some medications may have a more beneficial effect onvisual field survival than others.9–11 The reason for such adifference is not clear and remains controversial. Similarstudies have not been done with a-adrenergic agonists. Manyclinical observations lend credence to the potential role ofmicrocirculatory changes in glaucoma, either from directischemic damage or as a cofactor that increases the suscepti-bility to increased intraocular pressure.1–4 Besides reducingintraocular pressure, most adrenergic medications havemarked vasomotor effects. Topically applied b-adrenergicantagonists (timolol and betaxolol) or a-adrenergic agonists(phenylephrine and apraclonidine) have been shown to causevasoconstriction in the ciliary body of albino rabbits.28,29

Long-term topical administration of phenylephrine, an a-ad-renergic agonist with no ocular hypotensive effects, has been

BRIMONIDINE AND OPTIC NERVE BLOOD FLOWVOL. 128, NO. 5 603

shown to induce a vasoconstriction in the optic nervevasculature of albino rabbits.28 This was thought to be causedby intraocular drug penetration rather than systemic effects,because the contralateral, untreated eyes did not demonstratevasoconstriction. If topical drugs applied over a long periodpenetrate the eye as far as the vascular beds of the optic nerve,these medications may have potentially beneficial or detri-mental effects on the function of the optic nerve.

The pharmacokinetics of brimonidine on aqueous hu-mor dynamics has been studied in detail.12,13,30,31 Bri-monidine has rapid intraocular absorption, resulting inhigh concentrations in the iris-ciliary body, which maycontribute to its intraocular pressure–lowering effect. Thisresponse is mediated via a2-adrenoceptors or imidazolinereceptors, and differences among various species have beendemonstrated.31 The effect of brimonidine on the tissues ofthe posterior segment is less well defined at present. Datathat have emerged so far, including this study, indicate alack of vasoactivity in the posterior segment.24

For a topical medication to modulate the blood supply tothe anterior optic nerve, the medication should not onlybe vasoactive but also have access to the optic nervevasculature. This access can be through local penetrationor systemic absorption. Local intraocular penetration oftopical medications is likely to be insufficient to createhigh enough levels in the vitreous cavity. Even if thevitreous concentration is high, vasoactive drugs may affectonly the superficial (retinal) layers of the optic nerve andfail to penetrate the deeper layers of the anterior opticnerve or the retrobulbar vascular supply. Because most ofthe anterior optic nerve of mammals is supplied bybranches of the short posterior ciliary arteries, topicallyapplied medications probably do not reach sufficient in-traorbital levels to affect these vessels.22

Caution is required when speculating on optic nerveblood flow in humans on the basis of results from animalstudies. The present findings indicate that long-term ap-plication of brimonidine 0.2% does not affect anterioroptic nerve blood flow in rabbits. However, the effect onoptic nerve blood flow during long-term treatment withtopical a2-agonists may theoretically be different in hu-mans or in patients with glaucoma. As an example, it hasbeen demonstrated that the intraocular pressure–loweringeffect of brimonidine is mediated by different receptortypes in rabbits and primates.31 Therefore, it is possiblethat the optic nerve blood supply may behave differently indifferent species. Furthermore, pharmacologic vasomotoreffects may be different under increased intraocular pres-sure compared with normotensive conditions.

REFERENCES

1. Armaly MF, Krueger DE, Maunder L, et al. Biostatisticalanalysis of the collaborative glaucoma study, I: summary

report of the risk factors for glaucomatous visual-field defects.Arch Ophthalmol 1980;98:2163–2171.

2. Levene RZ. Low tension glaucoma. Surv Ophthalmol 1980;24:621–663.

3. Phelps CD, Corbett JJ. Migraine and low-tension glaucoma:a case-control study. Invest Ophthalmol Vis Sci 1985;26:1105–1108.

4. Carter CJ, Brooks DE, Doyle DL, Drance SM. Investigationsinto a vascular etiology for low-tension glaucoma. Ophthal-mology 1990;97:49–55.

5. Gasser P, Flammer J. Blood-cell velocity in the nailfoldcapillaries of patients with normal-tension and high-tensionglaucoma. Am J Ophthalmol 1991;111:585–588.

6. Van Buskirk EM. Glaucomatous optic neuropathy. J Glau-coma 1994;3(suppl):S2–S4.

7. Hayreh SS, Zimmerman MB, Podhajsky P, Alward WLM.Nocturnal arterial hypotension and its role in optic nervehead and ocular ischemic disorders. Am J Ophthalmol1994;117:603–624.

8. Van Buskirk EM, Cioffi GA. Glaucomatous optic neuropa-thy. Am J Ophthalmol 1992;113:447–452.

9. Messmer C, Flammer J, Stuumpfig D. Influence of betaxololand timolol on the visual field of patients with glaucoma.Am J Ophthalmol 1991;112:678–681.

10. Collignon-Brach J. Long-term effect of ophthalmic beta-adrenoceptor antagonists on intraocular pressure and retinalsensitivity in primary open-angle glaucoma. Curr Eye Res1992;11:1–3.

11. Kaiser HJ, Flammer J, Stuumpfig D, Hendrickson P.Long-term visual field follow-up of glaucoma patientstreated with beta-blockers. Surv Ophthalmol 1994;38(suppl):S156 –S160.

12. Toris CB, Gleason ML, Camras CB, Yablonski ME. Effects ofbrimonidine on aqueous humor dynamics in human eyes.Arch Ophthalmol 1995;113:1514–1517.

13. Nordlund JR, Pasquale LR, Robin AL, et al. The cardiovas-cular, pulmonary and ocular hypotensive effects of 0.2%brimonidine. Arch Ophthalmol 1995;113:77–83.

14. David R, Spaeth GL, Clevenger CE, et al. Brimonidine inthe prevention of intraocular pressure elevation followingargon laser trabeculoplasty. Arch Ophthalmol 1993;111:1387–1390.

15. Toris CB, Tafoya ME, Camras CB, Yablonski ME. Effects ofapraclonidine on aqueous humor dynamics in human eyes.Ophthalmology 1995;102:456–461.

16. Nagasubramanian S, Hitchings RA, Demailly P, et al.Comparison of apraclonidine and timolol in chronic open-angle glaucoma. Ophthalmology 1993;100:1318–1323.

17. Greenfield DS, Liebmann JM, Ritch R. Brimonidine: a newalpha2-adrenoreceptor agonist for glaucoma treatment. JGlaucoma 1997;6:250–258.

18. Gharagozloo NZ, Relf SJ, Brubaker RF. Aqueous flow isreduced by the alpha-adrenergic agonist, apraclonidine hy-drochloride (ALO 2145). Ophthalmology 1988;95:1217–1220.

19. Orgul S, Bacon DR, Van Buskirk EM, Cioffi GA. Opticnerve vasomotor effects of topical apraclonidine hydrochlo-ride. Br J Ophthalmol 1996;80:82–84.

20. Cioffi GA, Orgul S, Bacon DR, Van Buskirk EM. Acutevasomotor effects in the anterior optic nerve of topicalapraclonidine hydrochloride. J Glaucoma 1995;4(suppl):S15–S18.

21. Fahrenbach WH, Bacon DR, Morrison JC, Van Buskirk EM.Controlled vascular corrosion casting of the rabbit eye. JElectron Microsc Technique 1988;10:15–26.

22. Sugiyama K, Bacon DR, Morrison JC, Van Buskirk EM.

AMERICAN JOURNAL OF OPHTHALMOLOGY604 NOVEMBER 1999

Optic nerve head microvasculature of the rabbit eye. InvestOphthalmol Vis Sci 1992;33:2251–2261.

23. Orgul S, Cioffi GA, Bacon DR, Bhandari A, Van BuskirkEM. Measurement of optic nerve blood flow with nonradio-active colored microspheres in rabbits. Microvasc Res 1996;51:175–186.

24. Zhan GL, Toris CB, Gaffney MM, et al. Effects of apra-clonidine and brimonidine on rabbit ocular blood flow.ARVO abstracts. Invest Ophthalmol Vis Sci 1997;38(4,suppl):S783.

25. Orgul S, Cioffi GA, Wilson DJ, Bacon DR, Van Buskirk EM.An endothelin-1 induced model of chronic optic nerveischemia in the rabbit. Invest Ophthalmol Vis Sci 1996;37:1860–1869.

26. Rossetti L, Marchetti I, Orzalesi N, Scorpiglione N, Torri V,Liberati A. Randomized clinical trials on medical treatmentof glaucoma: are they appropriate to guide clinical practice?Arch Ophthalmol 1993;111:96–103.

27. Crick RP, Vogel R, Newson RB, et al. The visual field inchronic simple glaucoma and ocular hypertension: its char-acter, progress, relationship to the level of intraocular pres-sure and response to treatment. Eye 1989;3:536.

28. Sugiyama K, Bacon DR, Cioffi GA, Fahrenbach WH, VanBuskirk EM. The effects of phenylephrine on the ciliary bodyand optic nerve head microvasculature in rabbits. J Glau-coma 1992;1:156–164.

29. Van Buskirk EM, Bacon DR, Fahrenbach WH. Ciliaryvasoconstriction after topical adrenergic drugs. Am J Oph-thalmol 1990;109:511–517.

30. Gabelt BT, Robinson JC, Hubbard WC, et al. Apraclonidineand brimonidine effects on anterior ocular and cardiovascu-lar physiology in normal and sympathectomized monkeys.Exp Eye Res 1994;59:633–644.

31. Burke JA, Potter DE. Ocular effects of a relatively selectivealpha2 agonist (UK-14,304-18) in cats, rabbits and monkeys.Curr Eye Res 1986;5:665–676.

Authors InteractivetWe encourage questions and comments regarding this article via the Interneton Authors Interactivet at http://www.ajo.com/ Questions, comments, andauthor responses are posted.

BRIMONIDINE AND OPTIC NERVE BLOOD FLOWVOL. 128, NO. 5 605