Intravitreal Reinjection of Triamcinolone for Exudative Age-RelatedMacular Degeneration

Objective  To evaluate the outcome of repeated intravitreal injections of triamcinoloneacetonide for the treatment of exudative age-related macular degeneration.

Methods  This prospective, comparative nonrandomized clinical interventionalstudy included 13 patients with progressive exudative age-related maculardegeneration with occult, or predominantly occult, subfoveal neovascularization.All patients had shown an increase or stabilization of visual acuity aftera first intravitreal injection of 25 mg of triamcinolone acetonide. They receiveda second intravitreal injection of 25 mg of triamcinolone acetonide 3.1 to18 months after the first injection. Mean ± SD follow-up time afterthe second injection was 5.2 ± 3.6 months (median, 5.3 months). A controlgroup included 24 patients with exudative age-related macular degenerationwho did not receive treatment for their maculopathy. The main outcome measureswere visual acuity and intraocular pressure.

Results  In the study group, mean ± SD visual acuity increased significantly(P = .005 and P = .003,respectively) from 0.17 ± 0.11 to 0.32 ± 0.26 and from 0.15± 0.14 to 0.23 ± 0.19, respectively, after the first and secondinjections. An increase in visual acuity was found for 10 patients (77%) afterthe first and second injections. In the control group, visual acuity did notvary significantly during follow-up (P = .81). Thedifference in change in visual acuity between the study group and controlgroup was significant (P = .01 [Snellen lines] and P = .05 [logMAR units]). The peak in visual acuity and,in a chronologically parallel manner, the peak in intraocular pressure elevationoccurred 2 to 5 months after each injection.

Conclusions  Repeated intravitreal injection of 25 mg of triamcinolone acetonidemay lead to an increase in visual acuity in patients with exudative age-relatedmacular degeneration, with the peak in visual acuity and intraocular pressureelevation occurring about 2 to 5 months after each injection.

Age-related macular degeneration is the most common irreversible causeof severe loss of vision in the elderly population in Western countries.^1,2 Ocular photodynamic therapy with verteporfinhas been demonstrated to reduce vision loss in patients with the classic orpredominantly classic type of exudative age-related macular degeneration.³ Despite its importance for quality of life of theindividual patient and its marked socioeconomic effect on society, therapyfor age-related macular degeneration has remained unsatisfactory for manypatients with occult subfoveal neovascularization.

The purpose of this study was to investigate whether the antiangiogenic,antiproliferative, and antiedematous effects of triamcinolone acetonide wouldimprove visual acuity in patients with exudative age-related macular degeneration.^4-6 Because previous studieshad not shown a therapeutically positive effect of topical, oral, or intravenousapplication of cortisone, we chose the intraocular injection of cortisoneto achieve high levels of steroids at the site of required action withoutprovoking major systemic adverse effects.

In a previous pilot study of patients with exudative age-related maculardegeneration with occult or predominantly occult subfoveal neovascularization,intravitreal triamcinolone was associated with a significant increase in meanvisual acuity 2 to 3 months after the injection.⁷ Meanvisual acuity declined toward the baseline values 3 to 5 months after theinjection. Patients who had responded favorably to the first injection receiveda reinjection when the visual acuity decreased after an initial increase.The purpose of this study was to report the outcomes of the latter patientswho received a reinjection of intravitreal triamcinolone for the treatmentof exudative age-related macular degeneration and to compare the patientsin this study group with those in a control group who received no treatmentfor exudative age-related macular degeneration.

The study group in this prospective, comparative nonrandomized clinicalinterventional investigation included 13 consecutive patients (13 eyes; 9women; 7 right eyes) who had a progressive decrease in visual acuity due toexudative age-related macular degeneration with occult or predominantly occultsubfoveal neovascularization; who had received an intravitreal injection of25 mg of triamcinolone acetonide; who had shown an increase or stabilizationin visual acuity after the injection; and who experienced a rereduction invisual acuity several months after the injection. The study included all eyesfor which an intravitreal reinjection of triamcinolone was performed as treatmentfor exudative age-related macular degeneration. The reinjection took placeat a mean ± SD duration of 7.2 ± 4.1 months (median, 6 months;range, 3.1-18 months) after the first intravitreal injection. Mean ±SD follow-up time after the second injection was 5.2 ± 3.6 months (median,5.3 months; range, 1.3-13.3 months). Mean ± SD age of the patientswas 79.1 ± 8.3 years (range, 64.6-100.5 years; median, 79.0 years).Refractive error ranged between −1.0 diopter (D) and 3.5 D (mean ±SD, 1.33 ±1.59 D). Mean ± SD intraocular pressure prior to thefirst injection was 16.1 ± 3.6 mm Hg (median, 15 mm Hg). All patientswere fully informed about the experimental character of the therapy and signeda consent form. The ethics committee of the university approved the study,which followed the tenets of the Declaration of Helsinki.

The control group in the study included 24 patients (24 eyes; 16 women;13 right eyes) who had exudative age-related macular degeneration but didnot receive an intravitreal injection of triamcinolone or any other treatmentfor this disease. The reason for assigning these patients to the control groupwas that they did not want an intravitreal injection even though it was offeredto them. Mean ± SD refractive error was 1.56 ± 2.08 D (range,−2.50 D to 6.13 D), and mean ± SD age was 78.5 ± 8.7 years(range, 59.2-92.8 years; median, 78.0 years). Mean ± SD visual acuityat baseline measured 0.17 ± 0.10 (Snellen charts) (range, 0.03-0.30)and 0.85 ± 0.31 logMAR units.

Because of the distribution of patients in the study and control groups,these groups did not vary significantly in preoperative visual acuity (P = .29), preoperative intraocular pressure (P = .52), age (P = .99), refractive error(P = .62), sex (P = .59),right or left eye (P = .63), or length of follow-up(P = .86). All patients were fully informed aboutthe experimental character of the treatment.

Fluorescein angiography revealed subfoveal choroidal neovascularizationlocated mainly (>50% of the whole membrane) or completely underneath the retinalpigment epithelium in all eyes. None of the eyes included in the study underwentcataract surgery after the first intravitreal injection, neither as a singleprocedure nor as a combined procedure with the second injection.

All patients in the study group received an intravitreal injection of25 mg of crystalline triamcinolone acetonide in 0.2 mL of Ringer lactate solution.Prior to the intravitreal injection, topical 5% povidone-iodine (Alcon, FtWorth, Tex) was applied, and afterward the patients were completely draped.An eye speculum was inserted, and paracentesis was carried out to decreasethe volume of the eye. The injection of 25 mg (0.2 mL) of crystalline triamcinoloneacetonide was performed using a sharp 27-gauge needle through the temporalinferior pars plana 3 to 3.5 mm from the limbus. Then a combination antibioticointment (polymyxin B sulfate and neomycin sulfate) was applied. The triamcinolonehad been prepared by extracting 0.62 mL from the ampule (Volon A; Bristol-Myers-Squibb,Munich, Germany) containing 40 mg of triamcinolone acetonide in 1 mL. Theextracted volume was placed into a tuberculin syringe (1 mL) filled with Ringerlactate solution. A millipore filter (pore size, 5 µm; Sterifix Pury;Braun Melsungen AG, Melsungen, Germany) was placed on top of the syringe,and most of the contents of the syringe were pressed through the filter, withthe triamcinolone crystals remaining in the syringe. The syringe was thenrefilled with Ringer lactate solution, and the same procedure was repeated3 times. At the end, 0.2 mL of solution was left in the syringe and was injectedtransconjunctivally into the vitreous cavity.

Patients in the study group were reexamined the first day after theinjection, about 1 week after the injection, and in approximate monthly intervalsafter that. Patients in the control group were reexamined in approximate 2-monthintervals. Visual acuity (Snellen) was determined in a standardized fashionby an observer performing best-corrected refractometry.

Statistical analysis was performed using commercially available software(SPSSWIN version 11.5; SPSS Inc, Chicago, Ill). Significance was set at P = .05 (2-tailed) for all statistical tests.

After the first injection, mean ± SD visual acuity had significantly(P = .005; Wilcoxon rank sum test) increased from0.17 ± 0.11 (median, 0.20; range, 0.03-0.30) to a maximum of 0.32 ±0.26 (median, 0.30; range, 0.04-0.90) (Figure1). Expressed in Snellen lines, 6 eyes (46%) increased in visualacuity by 2 or more. Converting visual acuity measurements to logMAR unitsshowed a change from 0.88 ± 0.70 to 0.67 ± 0.42 logMAR units.The increase in visual acuity was statistically significant for the measurementsobtained 2 months after the injection (P = .04).For the measurements performed 1 month after the injection, the differencebetween the baseline visual acuity and postinjection visual acuity was marginallysignificant (P = .09). Because of the selection ofpatients for the study, an increase in visual acuity was found in 10 (77%)of 13 patients. Three (23.1%) of 13 patients did not show a change in visualacuity. Comparing the last postoperative examination with the preoperativeexamination, an increase in visual acuity was found in 6 patients (46%), anda decrease in visual acuity was observed in 6 patients (46%). Mean visualacuity prior to the injection and at the end of follow-up prior to the secondinjection did not vary significantly (P = .81).

After the second injection, mean ± SD visual acuity increasedsignificantly (P = .003) from 0.15 ± 0.14(median, 0.08; range, 0.01-0.40) to a maximum of 0.23 ± 0.19 (median,0.25; range, 0.02-0.60) (Figure 2).Expressed in Snellen lines, 6 eyes (46%) increased in visual acuity by 2 ormore. Converting visual acuity measurements to logMAR units showed a changefrom 1.04 ± 0.50 to 0.88 ± 0.48 logMAR units. The increase invisual acuity was marginally statistically significant for the measurementsobtained 1 month after the injection (P = .11). Comparingthe last postoperative examination at the end of follow-up with the preoperativeexamination, an increase in visual acuity was found in 5 patients (38%), anda decrease in visual acuity was observed in 6 patients (46%). Two (15%) of13 patients did not show a change in visual acuity. Mean ± SD visualacuity prior to the injection and at the end of follow-up did not vary significantly(0.15 ± 0.14 vs 0.16 ± 0.16; P = .89).

In the control group, mean ± SD visual acuity at baseline (0.17± 0.10 [range, 0.03-0.30]; median, 0.18) and best visual acuity duringfollow-up (0.17 ± 0.12 [range, 0.03-0.50]; median, 0.14) did not varysignificantly (P = .81). Comparing the last examinationat the end of follow-up with the baseline examination, an increase in visualacuity was found in 5 patients (38%), and a decrease in visual acuity wasobserved in 6 (46%). Mean ± SD visual acuity was significantly worseat the end of follow-up than at baseline (0.12 ± 0.10 vs 0.17 ±0.10; P = .008). The differences in change in bestvisual acuity between the study and control groups, expressed in Snellen linesand logMAR units, were significant (P = .01 and P = .05, respectively). Correspondingly, the number ofpatients showing an increase in visual acuity by more than 1 Snellen lineduring follow-up was significantly higher in the study group than in the controlgroup (P = .005; χ² test).

After the first injection, mean ± SD intraocular pressure increasedsignificantly (P = .009; Wilcoxon rank sum test)from 14.7 ± 3.0 mm Hg (range, 10-20 mm Hg; median, 15 mm Hg) at baselineto a maximum of 20.2 ± 5.3 mm Hg (range, 15-30 mm Hg; median, 18 mmHg) during follow-up. The difference between the intraocular pressure measurementsprior to the injection and at the examination 4 months after the injectionwas marginally significant (P = .06). Four (31%)of 13 eyes developed maximal intraocular pressure measurements higher than21 mm Hg. In these patients, intraocular pressure could be normalized by topicalantiglaucoma treatment. All other patients had postoperative intraocular pressuremeasurements in the normal range.

After the second injection, mean ± SD intraocular pressure increasedsignificantly (P = .02; Wilcoxon rank sum test) from16.1 ± 3.6 mm Hg (range, 12-23 mm Hg; median, 15 mm Hg) prior to thesecond injection to a maximum of 19.0 ± 5.0 mm Hg (range, 13-28 mmHg; median, 18 mm Hg) during follow-up. The differences between the intraocularpressure measurements prior to the injection and at the single examinationsperformed during follow-up were not statistically significant (P>.10). The increase in intraocular pressure was most marked for theexaminations 3 and 5 months after the reinjection of triamcinolone. Threeeyes (23%) developed maximal intraocular pressure measurements higher than21 mm Hg. For 2 of the 3 eyes, intraocular pressure could be normalized bytopical antiglaucoma treatment. One eye had to undergo filtering surgery becauseprogressive glaucomatous damage of the optic nerve developed, with intraocularpressure measurements of 30 mm Hg that could not be normalized with topicalmedication.

Two (67%) of the 3 eyes in which intraocular pressure increased afterthe second intravitreal injection had also shown an elevation in intraocularpressure after the first intravitreal injection. The third eye in which themaximal intraocular pressure was higher than 21 mm Hg after the second intravitrealinjection had had normal intraocular pressure measurements after the firstinjection. For this eye, only 1 measurement after the second injection washigher than 21 mm Hg. Two (50%) of the 4 eyes that had developed secondaryocular hypertension after the first intravitreal injection had normal intraocularpressure measurements after the second injection.

The frequency of intraocular pressure elevation and the amount of increasein intraocular pressure were statistically independent (P>.50) of the increase in visual acuity. For all patients in the studygroup, the maximal intraocular pressure was lower, but not significantly (P = .55), after the second intravitreal injection thanthe first (mean ± SD, 19.0 ± 5.0 mm Hg vs 20.2 ± 5.3mm Hg).

Treatment of exudative age-related macular degeneration with occultor predominantly occult subfoveal neovascularization has been inconclusiveso far. In contrast to the classic type of subfoveal neovascularization, forwhich photodynamic therapy with verteporfin has been shown to stabilize orincrease visual acuity, photodynamic therapy has been less successful forthe treatment of occult subfoveal neovascularization.³ Asan alternative to laser treatment, Penfold et al^8,9 andChalla et al¹⁰ started to inject triamcinoloneintravitreally in an effort to treat exudative age-related macular degeneration.Similarly to Challa and colleagues, Danis et al¹¹ performeda comparative study and reported a beneficial effect of triamcinolone in thestudy group compared with the control group. Additionally, Ranson et al¹² applied intravitreal triamcinolone for the treatmentof recurring subfoveal choroidal neovascularization after laser treatment.A recent investigation including 71 eyes with exudative age-related maculardegeneration demonstrated a significant increase in visual acuity after anintravitreal injection of 25 mg of triamcinolone acetonide.⁷ Theimprovement in visual acuity was significant for the examinations performed1 month (P = .04) and 2 months (P = .04) after the injection. About 3 to 5 months after the injection,visual acuity declined so that measurements obtained at the end of the follow-upperiod for that study did not differ significantly (P =.17) from the baseline values. Overall, however, 48 eyes (66.2%) had gainsin visual acuity during the follow-up period.

In our study, patients were re-treated with an intravitreal injectionof 25 mg of triamcinolone acetonide if the first intravitreal injection wasassociated with stabilization or increase in visual acuity. The results suggestthat the reinjection led to a significant (P = .003)improvement in visual acuity, with 10 (77%) of 13 patients showing an increasein visual acuity during the follow-up period. That figure was comparable withthe frequency of improvement in visual acuity after the first injection inthe same patients (10/13 [77%]). One may infer that eyes that have shown abeneficial effect with 25 mg of intravitreal triamcinolone acetonide may receivean intravitreal reinjection if the visual acuity deteriorates again. Our findingsconfirm those of a previous case report describing a patient who received5 intravitreal injections of 25 mg of triamcinolone acetonide and who aftereach injection demonstrated a reincrease in visual acuity.¹³

Interestingly, the peak of the increase in visual acuity occurred about2 to 5 months after the injections, with no marked difference in the timeof peaks between the first injection and reinjection (Figure 1 and Figure 2).This suggests that a reinjection of triamcinolone may be performed about 3to 5 months or more after an initial injection if the first is associatedwith an increase in visual acuity. In a chronologically parallel manner, thepeak of the elevation in intraocular pressure was about 2 to 5 months afterthe injection. This shows that after an intravitreal injection of triamcinolone,patients must undergo close follow-up for several months to detect a steroid-inducedincrease in intraocular pressure. Besides the chronological correlation betweenan increase in visual acuity and an elevation in intraocular pressure, thepostinjection increase in visual acuity was statistically independent of theelevation in intraocular pressure.

There are limitations to this study. The number of patients treatedwas relatively small; however, the postinjection visual acuity measurementswere significantly better than the baseline values when all patients withan intravitreal reinjection were included. Thus, the small number of patientsmay support our conclusions. Another important limitation of the study isits design as a nonrandomized comparative investigation. Because the firstintravitreal injection of triamcinolone was associated with an increase invisual acuity, it was difficult to convince patients to take part in a randomizedtrial in which some of them might not receive treatment. Furthermore, comparingthe patients in the study group with those in the control group showed significantdifferences in the change in visual acuity during follow-up. This suggeststhat the reincrease in visual acuity after the second injection of triamcinolonewas caused by the treatment and not the natural course of the disease.

Another limitation of the study is the method used to measure visualacuity. Instead of the charts used for the Early Treatment Diabetic RetinopathyStudy,¹⁴ visual acuity was determined usingSnellen charts in a standardized fashion by an observer performing best-correctedrefractometry. However, the same method for assessment of visual acuity wasapplied to the study group as well as the control group, so this flaw in thestudy methods might have been partially compensated for. A further limitationis that although intravitreal triamcinolone may increase cataract, cataractsurgery was not performed in combination with or after the intravitreal injection.The vision-reducing effect of progressive cataract may have hidden a vision-improvingeffect of triamcinolone, so again this limitation of the study may supportits conclusion.

An additional limitation of our study may be the relatively high doseof triamcinolone injected into the eye. In all preceding studies at othercenters applying intravitreal triamcinolone acetonide for the treatment ofexudative age-related macular degeneration or other macular diseases, a doseof 4 mg was used.^8-12,15-17 Thereason we used 25 mg of triamcinolone acetonide was that since the beginningof our ongoing investigations, now involving more than 400 patients with variousdiseases, we have used the same dose of 25 mg of triamcinolone acetonide andhave not yet seen adverse effects that may be attributed to that dose. Anadditional reason to use the high dose for this study was that in precedingstudies on the intravitreal use of 4 mg of triamcinolone acetonide for patientswith exudative age-related macular degeneration, the therapeutic effect hasnot been clear. This especially holds true for the prospective randomizedstudy performed by Gillies et al¹⁷ on the intravitrealuse of 4 mg of triamcinolone acetonide for the treatment of exudative age-relatedmacular degeneration.

In several preceding studies, a single intravitreal injection of triamcinolonehad already been used as treatment for exudative age-related macular degeneration.In 1998, Challa et al¹⁰ evaluated the safetyand efficacy of intravitreal triamcinolone after 18 months of follow-up inpatients with exudative age-related macular degeneration who were consideredunsuitable for laser photocoagulation. In that nonrandomized clinical pilotstudy, 30 eyes of 28 patients were treated with an intravitreal injectionof triamcinolone acetonide (4 mg). Of the 20 eyes with an initial visual acuityof 0.10 or better, vision was maintained in 11 eyes (55%), whereas 6 eyes(30%) had severe visual loss (≥6 Snellen lines). Visual acuity improvedin 3 of 10 eyes with an initial acuity of 20/400 or worse. The authors concludedthat a single intravitreal injection of 4 mg of triamcinolone acetonide isreasonably well tolerated and may be helpful in the treatment of exudativeage-related macular degeneration. In a randomized clinical trial, Danis etal¹¹ examined the effects of an intravitrealinjection of 4 mg of triamcinolone acetonide on the visual and clinical courseof exudative age-related macular degeneration in 27 patients who were comparedwith a nontreated control group. The authors found that visual acuity wassignificantly (P<.005) better in the treated groupcompared with control subjects at 3 and 6 months of follow-up. Intraocularpressure elevation was seen in 25% of treated patients but was controlledwith topical medications. Progression of cataract was more frequently detectedin the treated group. The authors concluded that intravitreal triamcinolonemay lead to improvement in visual acuity in exudative age-related maculardegeneration.

The results of both our study and the other investigations describedare partially in contrast to a recent study by Gillies et al,¹⁷ whofound no effect of 4 mg of intravitreal triamcinolone acetonide on the developmentof severe visual loss during a follow-up period of 1 year. Their treated grouphad more cataract, so a progressive opacification of the lens, reducing visualacuity, may have hidden the positive effect of triamcinolone. One reason forthe discrepancy between the investigation by Gillies and colleagues and ourstudy may be the difference in dose of triamcinolone. Another reason may bethat reinjections were not performed in the study by Gillies and colleagues.Their results fit with ours in that the peak in visual acuity occurred about2 to 5 months after the injection. At the end of their follow-up, the visualacuity was no longer better than prior to the intravitreal injection. Interestingly,Gillies and colleagues found a statistically significant and therapeuticallypositive effect of intravitreal triamcinolone on the size of subfoveal neovascularization3 months after the injection. Their study is in agreement with experimentalstudies regarding an angiostatic effect of intravitreal cortisone on subretinalneovascularization and other types of intraocular blood vessel proliferation.^5,6 An additional reason for the discrepancybetween the study by Gillies and colleagues and ours as well as the precedingstudies may be that their investigation included patients with the classictype of subfoveal neovascularization, which is associated with a worse prognosiscompared with occult subfoveal neovascularization.

The main adverse effect of intravitreal triamcinolone observed in thisstudy was elevation in intraocular pressure. However, maximal intraocularpressure measurements were lower, but not significantly (P = .55), after the second intravitreal injection than after the first.Only 1 eye that showed elevated intraocular pressure after the second injectionhad normal intraocular pressure measurements after the first injection. Onemay infer that if the intraocular pressure remains in the normal range aftera preceding intravitreal injection of 25 mg of triamcinolone acetonide, anelevation in intraocular pressure after a second injection is not probable.Future study may reveal whether a steroid-induced increase in intraocularpressure is less marked after a second injection than after the first. Withrespect to other adverse effects, one may infer that the repeated intravitrealinjection did not markedly damage the intraocular tissues; visual acuity didnot decrease after the reinjection but reincreased significantly. Furtherindication for the tolerability of the second intravitreal injection of 25mg of triamcinolone acetonide was that the eyes did not show other signs ofintraocular toxicity, such as intraocular inflammation or bullous keratopathydue to corneal endothelial dysfunction.

In conclusion, our data suggest that the repeated intravitreal injectionof 25 mg of triamcinolone acetonide for the treatment of exudative age-relatedmacular degeneration is associated with a reincrease in visual acuity in patientswho, as triamcinolone responders, had shown an improvement in visual acuityafter a preceding intravitreal injection of 25 mg of triamcinolone acetonide.The peak in visual acuity and the peak in intraocular pressure elevation occurabout 2 to 5 months after each injection.

Corresponding author: Jost B. Jonas, MD, Universitäts-Augenklinik,Theodor-Kutzer-Ufer 1-3, 68167 Mannheim, Germany (e-mail: Jost.Jonas@ma.augen.uni-heidelberg.de).

Submitted for publication February 13, 2003; final revision receivedJune 27, 2003; accepted September 10, 2003.