Effect of Laser Photocoagulation on the Retinal Vessel Diameter inBranch and Macular Vein Occlusion FREE

Retinal vein occlusion is the second most frequently occurring diseaseof retinal vessels. A histopathologically detectable thrombus in the occludedretinal vein^{1 - 2} or the pressureof an artery with artherosclerotic alterations at an arteriovenous crossing^{3 - 4} causes increased intravascular hydrostaticpressure and results in secondary macular edema that, in turn, is associatedwith severe vision loss. If the obstruction persists and macular edema cannotresolve spontaneously, chronic macular edema can develop.⁵ Insome eyes, severe ischemia may follow retinal vein occlusion that leads toneovascularization.

In previous studies, sectorial or grid laser photocoagulation in ischemicor edematous areas was reported to have beneficial effects in stabilizingvisual acuity (VA).^{6 - 8} Theeffect underlying the efficacy of laser therapy is not exactly known. Severaltheories for the mechanism underlying photocoagulation have been postulated:on the one hand, an effect on the retinal pigment epithelium with increasedproduction of the pigment epithelium–derived factor (PEDF)⁹ anda suppression of vascular endothelial growth factor (VEGF)–induced migrationgrowth has been proposed.¹⁰ On the other hand,a direct effect on retinal oxygen tension may occur. This may induce regulatoryvasoconstriction and indirectly reduce retinal blood flow, as shown for panretinalphotocoagulation applied in proliferative diabetic retinopathy.^{11 - 12} Accordingto Starling's law, reduced intravascular pressure in the capillary networkdue to reduced vascular caliber should then reduce retinal edema formation.The aim of this study was to investigate the response of retinal vessel diametersto laser photocoagulation treatment in patients with retinal vein occlusion.

The study was performed in adherence to the guidelines of the Declarationof Helsinki and the Good Clinical Practice guidelines, after written informedconsent was obtained. The study protocol was approved by the Ethics Committeeof the School of Medicine, University of Vienna, Vienna, Austria.

The patient group was prospectively recruited from our outpatient clinicthat specializes in retinal vascular and macular diseases. It consisted of7 men and 7 women who had ophthalmoscopic evidence of recent retinal veinocclusion, that is, a history of 3 to 9 months. The patients had to have 1of the following criteria on the study eye for inclusion: (1) a capillarynonperfusion area of 5 disc diameters, as demonstrated by fluorescein angiography,or (2) the presence of neovascularzation or evidence of macular edema andreduced VA due to vein occlusion.

Only patients with ocular media clear enough to permit safe laser photocoagulationwere included. Exclusion criteria were history of retinal photocoagulationtreatment, presence of other intraocular pathologic conditions, evidence ofmetabolic diseases such as diabetes mellitus, systemic hypertension higherthan 180/100 mm Hg, ametropia of 6 diopters (D) or more, and smoking morethan 20 cigarettes daily.

The ophthalmologic examination included the best-corrected VA (Snellen),intraocular pressure measured with applanation tonometry, slitlamp examination,biomicroscopy with a +90 D Volk lens, stereo fundus photography, and fluoresceinangiography of the involved area using a fundus camera set to a 40° angle(model CF60UV; Canon Co, Tokyo, Japan). Systolic, diastolic, and mean bloodpressures were measured on the arm by an automated oscillometric device (HP-CMSpatient monitor; Hewlett-Packard, Palo Alto, Calif). Pulse rate was automaticallyrecorded from a finger pulse-oxymetric device (HP-CMS patient monitor; Hewlett-Packard).Retinal vessel diameters were quantified using the retinal vessel analyzer(Carl Zeiss, Jena, Germany).^{13 - 14}

The retinal vessel analyzer is a commercially available system thatcomprises a fundus camera (model FF 450; Carl Zeiss), a video camera, a real-timemonitor, and a personal computer with an analyzing software for the accuratedetermination of retinal arterial and venous diameters.¹³ Everysecond a maximum of 25 readings of vessel diameter can be obtained. For thispurpose the fundus is imaged onto the charged-coupling device chip of thevideo camera. The consecutive fundus images are digitized using a frame grabber.In addition, the fundus image can be inspected on the real-time monitor and,if necessary, stored on a video recorder. Evaluation of the retinal vesseldiameters can either be done online or offline from the recorded videotapes.Owing to the absorbing properties of hemoglobin, each blood vessel has a specifictransmittance profile. Measurement of retinal vessel diameters is based onadaptive algorithms using these specific profiles. Whenever a vessel profileis recognized in the region of interest, the retinal vessel analyzer can followthis vessel as long as it appears within the measurement window.

Measurements were done at the main superior and inferior temporal ornasal vein and artery, 1 to 2 disc diameters (DD) from the optic disc. Thesame segment of the vessel was selected at all visits.

A sectorial scatter photocoagulation combined with grid laser treatmentin case of macular edema in branch vein occlusion (BrVO) or grid laser treatmentonly in macular vein occlusion was performed with blue-green argon laser accordingto the recommendation of the Branch Vein Occlusion Study Group^{6 ,8} andMiller.⁷

The presence of macular edema was established by biomicroscopy. Thefundus photographs and the angiograms were graded by an experienced ophthalmologist(N.M.). For the grading of the fundus photographs, a transparent overlay witha circle with a 1-DD radius was fixed over the image. By using a Donaldson×5 stereoscopic viewer, the degree of macular edema was determined usingthe modified Airlie House Classification.¹⁵ Asmorphologic variables the following lesions were graded as follows: degreeof macular edema (0, no evidence; 1, questionable involment; and 2, definitelypresent), size of macular retinal thickening less than 1 DD from the center(0, no evidence; 1, questionable involment; 2, size of thickening less thanone half of the disc area; 3, size of thickening <1 disc area; 4, sizeof thickening <2 disc areas; and 5, size of thickening ≥2 disc areas),and maximal retinal thickness at the center (0, no evidence; 1, questionableinvolment; 2, thickness <1× reference; 3, thickness ≥1×reference but <2× reference; 4, thickness ≥2× referencebut <½ DD; and 5, thickness ≥½ DD). After 3 months, macularedema was graded as reduced if the size of retinal thickening and/or the maximalretinal thickness decreased, and graded as worsened if these scores increased.

Measurement of retinal vessel diameter was repeated 1 and 4 weeks aswell as 3 months after laser photocoagulation. At the visit 3 months afterlaser treatment best-corrected VA (Snellen), slitlamp examination, biomicroscopywith a +90 D Volk lens, stereo fundus photography, and fluorescein angiographywere additionally repeated.

Statistical analyses were done using the Statistica software package(Release 4.5; StatSoft Inc, Tulsa, OK). Data are presented as mean (SD). Forsubgroup analysis patients were divided into 2 groups: patients with macularvein occlusion and patients with BrVO. In the second group the effect of laserphotocoagulation on the affected arteries and veins was also studied. Visualacuity befor laser therapy was compared with VA 3 months after photocoagulationby t test for dependent samples. The number of laserburns was correlated to the change in vessel diameter by using linear regressionanalysis. The change of the retinal vessel diameters at the follow-up visitswas analyzed using analysis of variance (ANOVA). Probability values smallerthan .05 were considered statistically significant.

The mean (SD) age of the 14 patients was 66.3 (8.6) years (age range,51-80 years), and the mean (SD) duration of symptoms (decrease in VA) whenthey visited in our outpatient department was 6.4 (2.1) months (range, 3-9months). Five and 6 eyes were initially seen with temporal inferior BrVO andtemporal superior BrVO, respectively, whereas in 3 eyes macular vein occlusionwas detected. At the initial ophthalmologic examination all eyes manifestedmacular edema. Three months after laser treatment the macular edema disappearedin 3 cases only but was reduced in all eyes. The number of burns used forphotocoagulation ranged between 4 and 821 (mean [SD], 245 [222]; range, 4-140in the macular vein occlusion group and 101-821 in the BrVO group), usuallywith a 0.1-mm spot size, 0.05- to 0.1-second exposure, and the minimal energyto create a light gray burn (range, 90-400 mW; mean [SD], 139 [96] mW).

In the macular vein occlusion group, VA was not significantly changedafter 3 months (P = .23, t test;mean [SD] VA before therapy, 0.57 [0.45], and 3 months later, 0.47 [0.35]).The VA in patients with BrVO increased significantly (P = .02, t test) at the 3-month follow-up visit(mean [SD] VA before laser therapy, 0.30 [0.22], and 3 months later, 0.45[0.21]).

The results of the measurements of retinal vessel diameters are summarizedin Table 1. In the macular veinocclusion group, no significant change of the vessel diameter in any vesselwas observed during the follow-up period. In the BrVO group, all vessels understudy tended to constrict after laser photocoagulation. This tendency wasmore pronounced in the veins than in the arteries. The effect of laser treatmenton retinal vessel diameters was significant for superior temporal (P = .045, ANOVA) and inferior temporal branch veins (P = .03, ANOVA). Vasoconstriction was more pronounced in the occludedbranch veins (P = .009, ANOVA) compared with thenonaffected veins (P = .12, ANOVA) (Figure 1). In the retinal arteries, none of the effects reachedthe level of significance although a tendency was seen in almost all vesselsunder study, particularly in the affected arteries. The change of VA after3 months correlated with the change of vessel diameter 3 months after lasertreatment for occluded venular branches (r = 0.78, P = .02, linear regression). There was no correlation betweenthe number of laser burns and the change of vessel diameters in the affectedveins in this period (r = 0.12, P = .75, linear regression).

In the present study sectorial laser photocoagulation in BrVO resultedin a significant decrease of the temporal branch vein diameter and tendedto decrease retinal arterial diameters. This is in keeping with the findingsof previous studies,^{16 - 17} whichdetected retinal vascular changes after grid photocoagulation.

Interestingly, the change of VA at 3 months after laser treatment wascorrelated to the decrease of the diameter of occluded veins, but not to thenumber of laser burns. Based on this finding one may hypothesize that earlyvasoconstriction as assessed with the retinal vessel analyzer may have predectivecharacter for the success of laser treatment. This needs, however, to be confirmedin large-scale studies.

Following retinal vein occlusion, the inner blood-retinal barrier isdamaged resulting in abnormal fluid homeostasis.¹⁸ Occlusionof a retinal branch vein leads to an increase of the intravascular pressureand a decrease of the oxygen concentration, resulting in abnormal permeabilityof the perifoveal retinal capillaries. If the volume exceeds a certain quantity,the retinal pigment epithelium as the outer blood-retinal barrier is not ableto sufficiently transport the accumulated fluid from the retina to the choriocapillarisand intraretinal edema develops.

The mechanism underlying the beneficial effect of photocoagulation inBrVO is still unclear. The following 3 hypotheses have been discussed:

1. After photocoagulation the oxygen flux increasesfrom the choroid to the inner retina leading to constriction of retinal vessels. This hypothesis is supported by the evidence that photocoagulationresults in an increase in the oxygen flux from the choroid to the inner retina.^{11 - 12} The oxygen tension plays a significantrole in retinal autoregulation because arterioles dilate when oxygen tensiondecreases and constrict when it increases resulting in modulation of retinalblood flow.¹⁹ In our study, the observed vasoconstrictionof the retinal branch vessels in response to photocoagulation may well resultfrom the increase of retinal oxygen tension in the ischemic area. This, inturn, decreases intravascular tension in capillaries and veins associatedwith passive vasoconstriction of the veins according to Laplace's law. Atthe same time, the decreased intravascular pressure leads to reduced edemaformation.

2. The increased concentration of VEGF during hypoxianormalizes after photocoagulation. Previous studies^{15 ,20 - 22} showan up-regulation of VEGF during vaso-obliterative and hypoxic phases in retinaldisease, which normalizes after photocoagulation.⁹ Thephotocoagulation-induced decrease in VEGF may well contribute to retinal branchvein constriction because VEGF is an important vasodilator in retinal vesselin healthy and diabetic rats.²² This occursas a direct effect of the laser photocoagulation on the cells of the retinaand the retinal pigment epithelium and/or develops as a consequence of improvedoxygenization of retinal tissue.

3. The increased concentration of PEDF and reducedexpression of VEGF after laser treatment lead to stabilization of the blood-retinalbarrier. The high concentration of VEGF and the reduced concentrationof PEDF can induce increased retinal vascular permeability leading to breakdownof the blood-retinal barrier.^{23 - 27} Retinallaser photocoagulation increases PEDF production.⁹ Thisfactor exerts effects opposite to VEGF by suppressing the VEGF-induced retinalendothelial growth and migration.^{10 ,28} Inaddition, neuroprotective properties have been described.^{29 - 31} Inthis way increased concentration of PEDF and decreased concentration of VEGFafter laser treatment can be indicative for inhibition of neovascularizationand resorption of macular edema through stabilization of the blood-retinalbarrier after laser treatment.

Which of these mechanisms underlying the therapeutic effect of laserphotocoagulation in retinal vein occlusion is predominant remains to be established.The fact that the increase of VA correlates with the change of diameter ofaffected branch veins in BrVO is compatible with the idea that the first 2mechanisms play an important role, but do not exclude the third possibility.

In macular vein occlusions we did not detect significant changes invessel diameter or changes in VA after grid photocoagulation. In eyes withminimal laser treatment, where the areas of photocoagulation are small, itmight well be that the effect of oxygen tension is localized thereby affectingcapillary but not branch vein diameters.

Our results show that sectorial retinal photocoagulation of BrVO hasa vasoconstrictive effect on occluded veins. The correlation between the changein VA and the change in vessel diameters suggests that branch vein constrictionafter photocoagulation may be an early indicator of the success of laser treatment.Whether this holds true remains, however, to be investigated in a large numberof subjects.

Correspondence: Leopold Schmetterer, PhD, Department of Medical Physics,Währinger Strasse 13, 1090 Vienna, Austria (Leopold.Schmetterer@univie.ac.at).

Submitted for publication March 10, 2003; final revision received October31, 2003; accepted December 12, 2003.