Intravitreal Triamcinolone for the Treatment of Macular Edema AssociatedWith Central Retinal Vein Occlusion

Objective  To investigate the safety and efficacy of intravitreal triamcinoloneacetonide as treatment for macular edema associated with central retinal veinocclusion (CRVO).

Methods  We reviewed the medical records of 13 consecutive patients (13 eyes)with macular edema associated with CRVO who were treated with an injectionof intravitreal triamcinolone acetonide (4 mg) at the University of Wisconsinand the Bascom Palmer Eye Institute. Each intravitreal injection was deliveredthrough the pars plana using a 27- or 30-gauge needle.

Main Outcome Measures  Change in Snellen visual acuity, clinical appearance of macular edema,measurement of foveal thickening with optical coherence tomography (OCT),and frequency of complications.

Results  The median age of the 13 patients was 67 years (interquartile range,57-77 years), and the median duration of symptoms before injection was 8 months(interquartile range, 4-9 months). Mean baseline visual acuity was 20/500in the affected eye. Mean visual acuity at the 6-month follow-up examinationwas 20/180 in the affected eye. All 13 patients completed the 6-month examination.Eyes with nonischemic CRVO (n = 5) demonstrated a significant improvementin visual acuity, whereas eyes with ischemic CRVO (n = 8) demonstrated a nonsignificantvisual acuity improvement. No patient had a decrease in visual acuity. Meanbaseline foveal thickness as measured by OCT was 590 µm (retinal thickening= 416 µm). Mean foveal thickness as measured by OCT at the 1-month follow-upexamination in 12 patients was 212 µm (retinal thickening = 38 µm).At the 3-month follow-up examination, mean foveal thickness as measured byOCT for 13 patients was 193 µm (retinal thickening = 19 µm). Betweenthe 3- and 6-month follow-up examinations, 4 patients developed a recurrenceof macular edema. Three of the 4 patients were retreated with a second injectionof triamcinolone. Two of these 3 patients experienced an improvement in visualacuity following retreatment. At the 6-month follow-up examination, mean fovealthickness as measured by OCT for 13 patients was 281 µm (retinal thickening= 107 µm). No adverse effects such as retinal detachment or endophthalmitisoccurred. One patient experienced an increase in intraocular pressure thatwas controlled with 2 aqueous suppressants.

Conclusions  Intravitreal injection of triamcinolone appears to be a possibly effectivetreatment in some patients with macular edema associated with CRVO. Patientswith nonischemic CRVO may respond more favorably than patients with ischemicCRVO, and retreatment may be necessary in some patients. In this case series,severe complications were not noted.

Central retinal vein occlusion (CRVO) is a common retinal vascular disorder.It has a characteristic clinical appearance with intraretinal hemorrhage,tortuous and dilated retinal veins, and in some cases optic disc edema. Macularedema is a frequent cause of visual acuity loss in patients with CRVO.^1-4

Macular edema from venous occlusive disease is caused by the initialoccurrence of thrombus formation at the lamina cribrosa or an arteriovenouscrossing. Green et al,⁵ in a histopathologicstudy of 29 eyes with CRVO, documented a new or recanalized thrombus of thecentral retinal vein in the area of the lamina cribrosa as a common pathologicfinding. Experimental studies in animals have demonstrated that a hypoxicenvironment in the retina is produced after venous occlusion.⁶ Thisis followed by functional and later structural changes in the retinal capillaries.These changes cause an immediate increase in retinal capillary permeabilityand accompanying retinal edema.

The increase in retinal capillary permeability and subsequent retinaledema may be the result of a breakdown of the blood-retinal barrier, possiblymediated in part by vascular endothelial growth factor (VEGF), a 45-kDa glycoprotein.⁷ Antonetti et al⁸ demonstratedthat VEGF may regulate vessel permeability by increasing phosphorylation oftight junction proteins such as occludin and zonula occluden 1. This modelprovides, at the molecular level, a potential mechanism for VEGF-mediatedvascular permeability in the eye. The normal human retina contains littleor no VEGF; however, hypoxia causes up-regulation of VEGF production.⁹ Disease states characterized by hypoxia-induced VEGFup-regulation include CRVO.^7,9,10

Attenuation of the effects of VEGF introduces a rationale for the treatmentof macular edema from CRVO. Corticosteroids, a class of substances with anti-inflammatoryproperties, have been demonstrated to inhibit the expression of the VEGF gene.¹¹ Additionally, corticosteroids have been demonstratedto abolish the induction of VEGF by the pro-inflammatory mediators platelet-derivedgrowth factor and platelet-activating factor in a time- and dose-dependentmanner.¹²

This retrospective review of intravitreal triamcinolone acetonide asa treatment for macular edema associated with CRVO was initiated because ofthe potential for corticosteroids to attenuate VEGF-mediated retinal capillarypermeability as well as early reports of efficacy with this treatment. Intravitrealtriamcinolone as a treatment for macular edema from CRVO has previously beenreported in small case series.^13-16 Inthis article, we describe a larger series of patients with longer follow-upthan that reported previously.¹⁶

Informed consent was obtained from each patient. All patients had clinicalevidence of CRVO with intraretinal hemorrhage and dilated, tortuous veinsin all 4 quadrants in addition to macular edema. At baseline, none of thepatients had retinal or anterior segment neovascularization. All eyes with10 or more optic disc areas of nonperfusion on fluorescein angiography werejudged to have ischemic CRVO. Eyes with fewer than 10 optic disc areas ofnonperfusion on fluorescein angiography or eyes without an afferent pupillarydefect were judged to have nonischemic CRVO. All patients were observed forat least 1 month prior to undergoing this investigative treatment.

Patients underwent clinical examination including a nonmasked Snellenvisual acuity test, intraocular pressure measurement, stereoscopic fundusphotography, and optical coherence tomography (OCT) prior to the injectionof triamcinolone. Patients underwent postoperative follow-up with repeatedclinical examination including nonmasked Snellen visual acuity testing, intraocularpressure measurement, stereoscopic fundus photography, and OCT. Patients wereassessed for adverse events including (but not limited to) retinal detachment,infectious endophthalmitis, noninfectious endophthalmitis, vitreous hemorrhage,cataract, and elevated intraocular pressure.

We performed OCT using the Optical Coherence Tomograph 2 or 3 (CarlZeiss Ophthalmic Systems Inc, Dublin, Calif). The same model was used foreach patient throughout the study. Readings for central retinal thicknesswere obtained either by measuring retinal thickness directly from the axialscan with the largest thickness measurement or from the mean retinal thicknessin the central subfield (500-µm radius). The measurement from the centralsubfield was used whenever possible; in patient 1, patient 4, and patient5, only the retinal thickness measurement from the axial scan was available.In these patients, the axial scan measurements were used in place of the centralsubfield measurement to derive retinal thickening values, determined at baselineand follow-up.

For the purpose of this study, retinal thickening was calculated asfollows: retinal thickening = actual retinal thickness – normal retinalthickness. The actual retinal thickness was the value measured by the OpticalCoherence Tomograph 2 or 3. In all patients (except patient 1, patient 4,and patient 5), the actual retinal thickness value was derived from the meanretinal thickness in the central subfield (500-µm radius). Normal retinalthickness was the retinal thickness expected in a normal eye without evidenceof diabetes. Normal retinal thickness was estimated according to a study byHee et al¹⁷ in which the mean ± SD thicknessin the central subfield (500-µm radius) was 174 ± 18 µm.A separate study by Muscat et al¹⁸ showed similarestimates of normal retinal thickness using a central subfield with an 800-µmradius.

In all patients, the intravitreal injection of triamcinolone was performedin the outpatient setting. Topical 0.5% proparacaine hydrochloride (Bausch& Lomb, Tampa, Fla) was applied to the ocular surface followed by preparationof the eyelids and conjunctiva with 5% povidone iodine. An eyelid speculumwas used to stabilize the eyelids, and a cotton-tipped applicator soaked inthe anesthetic was then applied over the inferotemporal injection site for1 minute. Triamcinolone acetonide was injected slowly through the inferiorpars plana at a dose of 4 mg (0.1 mL). No attempt was made to remove or dilutethe vehicle. A 27-gauge or 30-gauge needle was used for the injection. Theinferior pars plana was preferred to minimize postprocedural floaters becausethe injected triamcinolone rapidly localizes to dependent areas of the vitreouscavity following treatment. Indirect ophthalmoscopy was performed followingthe injection to confirm proper intravitreal localization of the suspensionand perfusion of the optic nerve head. An aqueous tap using a 30-gauge needlethrough a corneal paracentesis was performed, if necessary, to decrease theintraocular pressure to normal levels following the injection.

Visual acuity and retinal thickening at the baseline and follow-up visitswere summarized using mean ± SD. Changes from baseline to follow-upwere assessed using the paired t test. Analyses ofvisual acuity were performed by converting Snellen visual acuity measurementsto logMAR equivalents. Results are presented in both logMAR units, which wereused for analysis, and the equivalent Snellen visual acuity notation. Analysesare presented for the entire series as well as for ischemic and nonischemicsubgroups.

Patients were assessed at baseline and at 1, 3, and 6 months followinginitial intravitreal triamcinolone injection. The baseline and follow-up patientdata are included in Table 1.The median age of patients included in this study was 67 years (interquartilerange, 57-77 years). The median duration of symptoms (according to patienthistory) prior to treatment was 8 months with an interquartile range of 4to 9 months. Eight eyes had ischemic CRVO, and 5 eyes had nonischemic CRVO.

Visual acuity measurements are summarized in Table 2. The baseline mean visual acuity was 20/500 in the affectedeye. For the 13 eyes included in this study, there was a significant improvementin visual acuity at 1, 3, and 6 months of follow-up. The mean visual acuityvalues at these time points were 20/160 (P = .007),20/150 (P = .008), and 20/180 (P = .002), respectively, in the affected eye.

These grouped visual acuity results appear to be influenced primarilyby the nonischemic eyes. Nonischemic eyes benefited with a statistically significantvisual acuity improvement at each time point: the mean visual acuity at baselinewas 20/400 compared with 20/60 at 1 month (P = .004),20/50 at 3 months (P = .008), and 20/50 at 6 months(P = .003). The ischemic eyes also demonstrated improvementin visual acuity, but the improvement was statistically significant only atthe 6-month follow-up examination; the mean visual acuity at baseline was20/600 compared with 20/325 at 1 month (P = .21),20/300 at 3 months (P = .21), and 20/400 at 6 months(P = .01).

We assessed gain or loss of lines of Snellen visual acuity. The meangain in visual acuity was 2.2 lines (range, 0 to + 8). Four of the 13 patientsexperienced a visual acuity gain of 3 or more lines at the 6-month follow-upvisit. The CRVO in each of these 4 patients was nonischemic. Visual acuityin 9 of 13 patients was unchanged (<3 lines of improvement). Eight of these9 eyes had ischemic CRVO. No patient experienced a loss of visual acuity followingtreatment.

All eyes had a reduction in retinal thickening as demonstrated by OCT(Table 3). The mean baseline retinalthickening for all eyes was 386 µm. At 1 month there was an 89% reductionwith a mean retinal thickening of 38 µm (P<.001),at 3 months there was a 95% reduction with a mean retinal thickening of 18µm (P<.001), and at 6 months the retinalthickening remained reduced at 72% of baseline with a mean retinal thickeningof 108 µm (P<.001).

Both nonischemic and ischemic eyes with CRVO demonstrated a statisticallysignificant reduction in retinal thickening. Mean baseline retinal thickeningof the ischemic eyes measured 385 µm and was reduced to 69, 14, and121 µm at 1, 3, and 6 months, respectively. Mean baseline retinal thickeningof the nonischemic eyes was 386 µm and measured −4, 26, and 86µm at 1, 3, and 6 months, respectively. A comparison of the reductionin thickening between the ischemic and nonischemic eyes was nonsignificantat each time point.

Figure 1 is an illustrativecase of a patient with nonischemic CRVO who showed a good anatomical and functionalresponse to treatment (patient 10). This patient demonstrated a decrease inretinal thickness from greater than 600 µm to 150 µm at 1 monthwith a corresponding improvement in visual acuity from 20/200 to 20/60 inthe affected eye. During the next several months, this patient's macular edemarecurred, necessitating retreatment (Table1).

Figure 2 is an illustrativecase of a patient with long-standing (>12 months) ischemic CRVO who did notshow a good functional response to treatment although anatomically there wasa reduction in retinal thickening (patient 12). This patient demonstrateda decrease in retinal thickness from 600 to 230 µm during a 6-monthperiod. However, visual acuity throughout that period remained stable between2/200 and 20/400 in the affected eye (Table1). Hence, intravitreal triamcinolone injection resulted in anatomicalimprovement in this patient without corresponding functional improvement.

The mean increase in intraocular pressure was 2.3 mm Hg at 1 month (P = .04), 4.1 mm Hg at 3 months (P =.04), and 2.5 mm Hg at 6 months (P = .03). Therewas no statistically significant difference in intraocular pressure elevationbetween ischemic and nonischemic eyes. Only patient 10 required aqueous suppressantsfor an increase in intraocular pressure judged to be clinically significant.This elevation peaked at 20 mm Hg higher than baseline at the 3-month visit,and treatment with 2 aqueous suppressants was then initiated. The intraocularpressure normalized after 1 month of treatment, and aqueous suppressant therapywas subsequently discontinued. No patient developed infectious or noninfectiousendophthalmitis or retinal detachment or required cataract extraction forprogression to visually significant cataract during the follow-up period.One patient (patient 10) developed iris neovascularization at the 4-monthfollow-up visit. This was successfully treated with panretinal photocoagulation.

Between the 3-month and 6-month follow-up examinations, 4 patients developeda recurrence of macular edema. Three of the 4 patients underwent intravitrealreinjection with 4 mg of triamcinolone acetonide. Two of these 3 patientsexperienced improvement in visual acuity following retreatment.

At present, there is no proven effective treatment for patients withmacular edema from CRVO. The safety and efficacy of grid laser photocoagulationfor macular edema associated with CRVO was evaluated as part of the CentralVein Occlusion Study.⁴ That study demonstratedthat although there was a definite decrease in macular edema on fluoresceinangiography in the grid laser photocoagulation treatment group compared withthe control group, this did not translate to a visual acuity improvement.Therefore, currently no therapy has proved effective for decreased visiondue to macular edema associated with CRVO.

Other avenues for treating this common cause of vision loss have beeninvestigated, including laser-induced chorioretinal venous anastamosis. Ina pilot series, successful chorioretinal venous anastamosis with reductionof macular edema and improvement in visual acuity was noted in some patients.¹⁹ Potential complications include intravitreal neovascularizationand subsequent vitreous hemorrhage as well as choroidal neovascularizationat the anastamosis site. More recently, surgical decompression of the scleralring around the optic nerve (radial optic neurotomy) has been investigatedin patients with CRVO.²⁰ However, this is asignificant surgical intervention with inherent risks, recovery time, andexpense. Several other surgical and pharmacologic treatment modalities havealso been studied.^21-24

The triamcinolone used in this study is a commercially available corticosteroid(Bristol-Myers-Squibb, Princeton, NJ). Intravitreal injection of pure triamcinolonewas shown to be nontoxic in animal studies,^25-27 aswas the vehicle used in the commercial preparation.²⁸ Intravitrealtriamcinolone acetonide at a dose of 4 mg or higher has been used clinicallyfor a variety of conditions including choroidal neovascularization from age-relatedmacular degeneration, diabetic macular edema, and proliferative vitreoretinopathy.^29-34 Asa result of the safety profile demonstrated in animal models, the prior useof intravitreal triamcinolone in the clinical setting, and the potential forcorticosteroids to attenuate VEGF-mediated vascular leakage, we investigatedthe use of intravitreal triamcinolone in patients with macular edema associatedwith CRVO.

Treatment of macular edema from CRVO with intravitreal triamcinolonehas been reported previously. Greenberg et al¹³ studiedboth eyes of a patient with bilateral macular edema from CRVO. Similarly,Jonas et al¹⁵ evaluated a patient with bilateralmacular edema from CRVO. We previously reported shorter-term results in asubset of the 13 patients in our current study.¹⁶ Thesestudies show that although many patients have a rapid anatomical responseto treatment, some of these patients do not experience improvement in visualacuity. The results from our present study suggest that patients with nonischemicCRVO have a high likelihood of both an anatomical and functional responseto intravitreal triamcinolone injection. Patients with ischemic CRVO alsohave a high likelihood of anatomical response. However, these patients donot appear to respond as well functionally. In this study, all 5 patientswith nonischemic CRVO responded anatomically, and 4 of these 5 patients hadsignificant visual acuity improvement (P =.003). Even though all patients with ischemic CRVO (n = 8) responded anatomically,the magnitude of visual acuity change was not as great as for those with nonischemicCRVO. None of the 8 patients with ischemic CRVO lost visual acuity.

Retreatment due to recurrence of macular edema with a concomitant reductionin visual acuity occurred in 3 of 4 patients between the 3- and 6-month follow-upexaminations. One of these 3 patients (patient 7) had both a reduction infoveal thickness and improvement in visual acuity following retreatment. Thesecond patient (patient 13) did not demonstrate an anatomical or functionalresponse following retreatment. Interestingly, the third patient (patient10) had an improvement in visual acuity following retreatment but did notexperience a reduction in foveal thickness as measured with OCT.

The results of this series indicate that intravitreal injection of triamcinolonemay be an effective treatment for some patients with macular edema from CRVO.Four (31%) of 13 patients gained 3 or more lines of visual acuity at the 6-monthfollow-up visit. This compares favorably with data from the Central Vein OcclusionStudy,⁴ in which 6% of patients in both thetreated and untreated groups gained 3 or more lines of visual acuity at the12-month follow-up visit and 5% of patients in the treated group and 1% ofpatients in the untreated group gained 3 or more lines of visual acuity atthe 4-month follow-up visit.

Most patients with CRVO, both ischemic and nonischemic, may have a favorableanatomical response to this treatment. However, a favorable visual acuityresponse appears more likely in patients with nonischemic CRVO. This studyalso suggests that retreatment may be necessary in some patients owing tothe recurrence of macular edema. Four of 13 patients in this series experienceda recurrence of macular edema between the 3- and 6-month follow-up examinations;with a longer follow-up period, even more patients might have experiencedrecurrence. Retreatment may be effective in reducing retinal thickening and/orimproving visual acuity in some patients.

In this study, the mean intraocular pressure was higher than the meanat baseline at all follow-up time points. One patient had an increase in intraocularpressure that required transient treatment with topical aqueous suppressants.No other adverse events were noted. However, the potential adverse effectsof corticosteroids include glaucoma and cataract and may increase in frequencywith longer follow-up and repeated injections. Administration of corticosteroidsvia intravitreal injection would add other potential risks such as retinaldetachment, vitreous hemorrhage, and infectious endophthalmitis. Therefore,further investigation is warranted to balance the risks of this treatmentmodality against the potential benefits. Additional investigation may alsoanswer issues not addressed in this article, such as the duration of treatmenteffect and the need for repeated injection.

Correspondence: Michael S. Ip, MD, Department of Ophthalmology, Universityof Wisconsin, Park West One, 406 Science Dr, Suite 400, Madison, WI 53711-1068(msip@wisc.edu).

Submitted for publication June 23, 2003; final revision received January20, 2004; accepted January 20, 2004.