Intravitreal triamcinolone acetonide for diabetic diffuse macular edema: preliminary results of a prospective controlled trial

Intravitreal Triamcinolone Acetonide for Diabetic Diffuse Macular Edema Preliminary Results of a Prospective Controlled Trial Pascale Massin, MD, PhD,1 Franc¸ois Audren, MD,1 Belkacem Haouchine, MD,1 Ali Erginay, MD,1 Jean-Franc¸ois Bergmann, MD,2 Rym Benosman, MD,1 Charles Caulin, MD,2 Alain Gaudric, MD1 Objective: To evaluate prospectively the efficacy and safety of 1 intravitreal injection of 4 mg of triamcinolone acetonide for refractory diffuse diabetic macular edema. Design: Interventional case series. Participants: Fifteen patients with bilateral diabetic macular edema unresponsive to laser photocoagulation. In all patients, one eye received the injection, and the other served as a control. Intervention: Intravitreal injection of 4 mg of triamcinolone acetonide under subconjunctival anesthesia. Main Outcome Measures: The main outcome measure was central macular thickness (CMT) at 1, 3, and 6 months, measured by optical coherence tomography. Secondary outcomes were Early Treatment Diabetic Retinopathy Study (ETDRS) scores, intraocular pressure, and cataract progression. Results: In this preliminary report, we give the results for 12 patients who had a follow-up of at least 3 months. Seven of them were followed up for 6 months. Before injection, CMT was 509.6⫾143.5 ␮m (mean ⫾ standard deviation [SD]) in injected eyes, versus 474.4⫾82.6 ␮m in control eyes. Four weeks after injection, it was 207.3⫾44.2 ␮m in injected eyes and 506.7⫾122.4 ␮m in control eyes (P⬍0.001, bilateral Wilcoxon test for paired samples), and after 12 weeks, 207⫾96.7 ␮m and 469.3⫾117.6 ␮m, respectively (P ⫽ 0.005). The difference between the CMTs of injected and control eyes was no longer significant at 24 weeks because of the recurrence of macular edema in 5 of 12 injected eyes. Before triamcinolone injection, the ETDRS score was 47.8⫾13 (mean ⫾ SD; range, 28 – 66) in injected eyes, versus 51.9⫾14.6 (range, 31–71) in control eyes. Twelve weeks thereafter, the corresponding values were 52.7⫾10.8 (range, 34 –70) and 50.8⫾14.3 (range, 29 – 69), respectively, and at 24 weeks, 54.7⫾7.6 (range, 47– 68) and 50.6⫾18.4 (range, 28 –71). At no time was the difference between the ETDRS scores for injected and control eyes significant. In 6 of the 12 injected eyes, intraocular pressure exceeded 25 mmHg, and was controlled by topical medication. Conclusion: Intravitreal injection of triamcinolone effectively reduces macular thickening due to diffuse diabetic macular edema, at least in the short term. Further studies are required to demonstrate that it provides visual benefit. Ophthalmology 2004;111:218 –225 © 2004 by the American Academy of Ophthalmology.

Macular edema (ME) is the main cause of visual impairment in diabetic patients.1 Its treatment is based mainly on laser photocoagulation. The Early Treatment Diabetic Retinopathy Study (ETDRS) trial showed that focal laser photocoagulation is beneficial for eyes with clinically signifiOriginally received: October 20, 2002. Accepted: May 23, 2003.

Manuscript no. 220670.

1

Department of Ophthalmology, Hoˆpital Lariboisie`re, Assistance Publique-Hoˆpitaux de Paris, Universite´ Paris 7, Paris, France.

2 Unite´ de Recherches The´rapeutiques, Hoˆpital Lariboisie`re, Assistance Publique-Hoˆpitaux de Paris, Universite´ Paris 7, Paris, France. Presented at: American Academy of Ophthalmology Annual Meeting, November 2002; Orlando. The authors have no proprietary interest in this study. Dr Audren, co–first author, contributed equally to the work. Correspondence to Pascale Massin, MD, PhD, Department of Ophthalmology, Hoˆpital Lariboisie`re, 2 Rue Ambroise Pare´, 75475 Paris CEDEX 10, France. E-mail: [email protected].

218

© 2004 by the American Academy of Ophthalmology Published by Elsevier Inc.

cant diabetic ME.2 However, for diffuse diabetic ME, this treatment has had limited results. Lee and Olk3 showed that with grid-pattern laser photocoagulation, diffuse diabetic ME resolved in 68% to 94% of cases, and visual acuity (VA) stabilized in 61%. However, VA decreased by ⱖ3 lines in 24.6% of the eyes in their series, despite treatment. Alternative treatments to laser photocoagulation for diffuse diabetic ME are currently under investigation. Recently, Antcliff reported that intravitreal triamcinolone was beneficial for cystoid ME secondary to uveitis,4 and Martidis, in a prospective but uncontrolled study,5 reported that intravitreal triamcinolone was also beneficial for diffuse diabetic ME. The aim of our study was to investigate, in a prospective controlled study using an interventional case series design, the efficacy and safety of 1 intravitreal injection of triamcinolone acetonide (TA) for refractory diffuse diabetic ME, using optical coherence tomography (OCT) to evaluate the anatomic results. ISSN 0161-6420/04/$–see front matter doi:10.1016/j.ophtha.2003.05.037

Massin et al 䡠 Intravitreal Triamcinolone Acetonide for Diabetic Macular Edema

Patients and Methods The protocol was approved by the Institutional Review Board of the Hoˆ pital Saint-Louis, Paris, France. Patients were recruited and enrolled at the Diabetic Retinopathy Unit of the Lariboisie`re Hospital in Paris.

Patient Eligibility Patients were included if they had bilateral diffuse diabetic ME on biomicroscopy, unresponsive to laser photocoagulation and with no sign of vitreomacular traction on either biomicroscopy or OCT examination. Diffuse ME was defined by macular thickening on biomicroscopy, involving the center of the macula, with few or no macular exudates, and generalized breakdown of the inner blood– retina barrier, with diffuse fluorescein leakage involving most of the macular area on fluorescein angiography. All focal leaks had previously been treated by laser photocoagulation. Central macular thickness (CMT) had to be ⬎380 ␮m (normal ⬍ 206 ␮m) on OCT in both eyes. Glycated hemoglobin (Hb A1c) had to be less than 9.5%, and systolic and diastolic blood pressure less than 150 and 90 mmHg, respectively. Patients with a history of glaucoma or ocular hypertension were excluded from the study. To detect corticosteroid-induced ocular hypertension, 1 month of treatment with topical 0.1% dexamethasone 3 times daily in both eyes was performed before inclusion; patients whose rise in intraocular pressure (IOP) exceeded 15 mmHg after this month of treatment were excluded from the study. Also excluded were patients with ⱖ1 disc diameter of capillary closure on fluorescein angiography, patients who had undergone cataract or vitreous surgery during the previous 6 months, and patients with renal insufficiency. Each eligible patient received complete oral and written information concerning the study protocol; patients who agreed to participate signed a consent form.

Preoperative Examination Four weeks before TA injection, baseline data were recorded. They included the duration of ME, number of laser photocoagulation sessions, objective refraction, best-corrected VA (BCVA) measured on the ETDRS chart, lenticular status using the Lens Opacities Classification System III classification,6 the results of applanation tonometry, fundus examination with a contact lens, fundus photography, fluorescein angiography, and macular mapping using OCT, blood pressure measurements, and Hb A1c levels. The 1 month of treatment with topical 0.1% dexamethasone was then prescribed. If no rise in IOP exceeding 15 mmHg was observed after 1 month, and if the patient fulfilled all the inclusion criteria, he or she was included in the study. Best-corrected VA measurement and OCT mapping were repeated before TA injection.

Triamcinolone Acetonide Injection For patients with symmetric diffuse diabetic ME (i.e., with a left-to-right thickness ratio of 0.8 to 1.25), the eye to be injected was determined by randomization just before injection, and the uninjected eye served as a control. The randomization sequence was generated from a random number table and was block randomized into groups of 4. For patients with asymmetric diffuse diabetic ME (i.e., with a left-to-right thickness ratio exceeding 1.25), the eye with the thicker macular thickness was injected, and once again, the uninjected eye served as a control. Four milligrams of TA was injected into the vitreous (total volume, 0.1 ml) with full asepsis under subconjunctival anesthesia

(Kenacort, Bristol-Myers Squibb, Paris, France). As additives, it contained benzyl alcohol, sodium chloride, sodium carboxymethylcellulose, and polysorbate 80. The injection was performed 4 mm posterior to the limbus, through the inferior pars plana, with a 30-gauge needle. All patients were given 200 mg of ofloxacine intravenously as antibioprophylactic treatment. As the aim of our study was to investigate the effect of one intravitreal TA injection, no reinjection was performed.

Outcome Measures The main outcome measure was the change in CMT at 4, 12, and 24 weeks. Central macular thickness was defined by the average thickness of a central macular region 1000 ␮m in diameter, centered on the patient’s foveola, and automatically measured by OCT. Optical coherence tomography mapping was performed using commercially available equipment (Zeiss, Dublin, CA) and done through a dilated pupil by an experienced examiner who was aware of the clinical findings for each patient. The OCT examination comprised 6 radial 6-mm-long scans of each operated eye, centered on the patient’s fixation point, at 0°, 30°, 60°, 90°, 120°, and 150°. Retinal thickness was computed automatically, using OCT retinal mapping A5 software (Humphrey Instruments, Inc., Dublin, CA). This mapping averaged the 6 scans to give the CMT in a central area 1000 ␮m in diameter (i.e., the average of 100 measurements). The normal values for mean retinal thickness in this area are 170⫾18 ␮m.7,8 Good reproducibility of these measurements using OCT mapping has been demonstrated.8 In addition, a change in macular thickness exceeding 12% is significant, and not due to the variability of the method.8 Secondary end points were changes in ETDRS scores, IOP, and cataract progression. Patients were seen before injection, 1 week and 1 month after injection, and thereafter monthly until 6 months. Early Treatment Diabetic Retinopathy Study scores were measured before injection and 4, 12, and 24 weeks thereafter. Intraocular pressure and macular thickness were measured at each visit. Fundus photography and lens status evaluation were performed before injection and 6 months thereafter.

Statistical Analysis For statistical analysis, paired series were compared, each treated eye being paired with the untreated eye of the same patient. The study was designed to have a type 1 error of 5%, with a power of 95% for detection of the expected change in CMT, given an estimated standard deviation (SD) of 175 ␮m (Massin, unpublished data). With these parameters, we needed 16 eyes in each group (i.e., 16 patients) to complete the study. Statistical analyses were performed using the bilateral Wilcoxon signed rank test for paired data. The null hypothesis was rejected for P values of ⬍0.05.

Results Between November 2001 and September 2002, 17 patients fulfilled the inclusion criteria and agreed to participate in the study. Two were excluded because of an IOP rise exceeding 15 mmHg after 1 month of treatment with topical 0.1% dexamethasone. Therefore, 15 patients (30 eyes) were included in the study. In this preliminary report, we give the results for the 12 patients (5 males and 7 females) who had a follow-up of at least 3 months. Seven of them were followed up for 6 months.

219

Ophthalmology Volume 111, Number 2, February 2004 Table 1. Baseline Clinical Characteristics of 12 Patients with Diffuse Diabetic Macular Edema, before Triamcinolone Acetonide Injection

Right eyes (%) Lens Phakic Pseudophakic Mean number of prior sessions of laser photocoagulation (range) Retinopathy NPDR PDR ⫹ PRP Median BCVA (range) Mean (⫾ SD) ETDRS score Mean (⫾ SD) central macular thickness (␮m)

Injected Eyes (n ⴝ 12)

Control Eyes (n ⴝ 12)

P between Groups

6 (50)

6 (50)

NS

10 2 2.4 (1–3)

10 2 2.4 (1–4)

NS NS NS

11 1 20/125 (20/320 to 20/50) 47.8⫾13 509.6⫾143.5

11 1 20/100 (20/250 to 20/32) 51.9⫾14.6 474.4⫾82.6

NS NS NS

ETDRS ⫽ Early Treatment Diabetic Retinopathy Study; NPDR ⫽ nonproliferative diabetic retinopathy; NS ⫽ nonsignificant; PDR ⫽ proliferative diabetic retinopathy; PRP ⫽ panretinal photocoagulation; SD ⫽ standard deviation.

The mean (⫾ SD) age of the patients was 59⫾9.2 years (range, 36 –70). Two patients had non–insulin-dependent diabetes and 10 had insulin-dependent or insulin-requiring diabetes. The mean (⫾ SD) duration of diabetes was 16.3⫾8.1 years (range, 4 –30). The clinical characteristics of the 24 enrolled eyes before TA injection are shown in Table 1. No significant difference between the 2 groups of eyes was observed for the extent of macular capillary closure. Two of these eyes had nonproliferative diabetic retinopathy, and 22 had proliferative diabetic retinopathy that had been treated by panretinal photocoagulation more than 6 months before TA injection. Before TA injection, mean Hb A1c was 7.1⫾1.2% (range, 5.5–9.5%), and mean systolic and diastolic blood pressures were, respectively, 136.8⫾12.9 mmHg (range, 118 –150) and 75.5⫾11 mmHg (range, 50 –90). The precise duration of diabetic ME was difficult to determine in 3 patients, but ranged from 23 months to 72 months. All eyes had undergone a mean of 2.4 sessions of laser photocoagulation for diabetic ME more than 6 months before TA injection, in accordance with ETDRS guidelines. Macular edema was symmetric in 10 patients, and asymmetric in 2.

Central Macular Thickness Mean (⫾ SD) CMT before TA injection was 509.6⫾143.5 ␮m in injected eyes, versus 474.4⫾82.6 ␮m in control eyes, a difference that was not significant (P⬎0.1). Four weeks after injection, mean (⫾ SD) injected and control eye CMTs were 207.3⫾44.2 ␮m and 506.7⫾122.4 ␮m, respectively, and after 12 weeks, 207⫾96.7 ␮m and 469.3⫾117.6 ␮m (Table 2, Fig 1). The difference between CMTs of injected and control eyes was significant at 4 weeks (P⬍0.001) and 12 weeks (P ⫽ 0.005, bilateral Wilcoxon test for paired samples). Twelve weeks after injection, CMT had returned to normal (⬍206 ␮m) in 9 of 12 eyes. In the other 3, it was subnormal in 1 eye (patient 11), reduced in 1 (patient 7), but had again increased in 1 eye (patient 10). Twenty-four weeks after injection, mean (⫾ SD) CMT was 426⫾182 ␮m in 7 of the 12 treated eyes, versus 430.6⫾157.2 ␮m in the 7 corresponding control eyes. At that time, the difference between injected and control eye CMTs was no longer significant (P⬎0.1). The difference between the relative changes in macular thickness of injected and control eyes was significant at 4 weeks (P ⫽ 0.0022) and 12

Table 2. Central Macular Thickness (CMT) in the Treated Eye of 12 Diabetic Patients before Intravitreal Injection of 4 mg of Triamcinolone Acetonide and 4, 12, and 24 Weeks Thereafter, and in the Control Contralateral Eye CMT in Treated Eyes

CMT in Control Eyes

Patient No.

Inclusion (n ⫽ 12)

4 wks (n ⫽ 12)

12 wks (n ⫽ 12)

24 wks (n ⫽ 7)

Inclusion (n ⫽ 12)

4 wks (n ⫽ 12)

12 wks (n ⫽ 12)

24 wks (n ⫽ 7)

1 2 3 4 5 6 7 8 9 10 11 12 Mean (SD)

443 525 380 487 527 414 461 723 396 419 477 863 509.6 ⫾ 143.5

181 188 164 183 207 148 285 192 185 244 225 285 207.3 ⫾ 44.2

163 189 159 179 192 133 289 173 172 486 212 137 207.0 ⫾ 96.7

197 612 247 574 586

545 527 380 483 411 424 468 658 390 458 405 547 474.4 ⫾ 82.6

576 526 608 461 562 740 327 636 378 413 463 390 506.7 ⫾ 122.4

552 591 529 501 476 685 313 317 316 475 382 494 469.3 ⫾ 117.6

694 355 240 499 548

SD ⫽ standard deviation. Values are expressed in microns.

220

466 277

425.9 ⫾ 182.1

367 311

430.6 ⫾ 157.2

Massin et al 䡠 Intravitreal Triamcinolone Acetonide for Diabetic Macular Edema Table 3. Comparison of Mean (⫾ Standard Deviation) Early Treatment Diabetic Retinopathy Study (ETDRS) Scores for Eyes Injected with 4 mg of Triamcinolone Acetonide (TA) and Control Contralateral Eyes

ETDRS Score Before TA injection 4 wks after injection 12 wks after injection 24 wks after injection Figure 1. Evolution of central macular thickness after intravitreal injection of triamcinolone acetonide in control and injected eyes. NS ⫽ nonsignificant.

weeks (P ⫽ 0.0038, bilateral Wilcoxon test for paired samples) but not at 24 weeks (P ⫽ 0.866). By 24 weeks, diffuse diabetic ME had recurred in 5 eyes (patients 2, 4, 5, 7, and 10), with a final CMT equal to or greater than the initial value. Diffuse diabetic ME recurred by 12 weeks after TA injection in 1 of these eyes, by 16 weeks in 3, and by 20 weeks in 1.

Secondary End Points For both control and injected eyes, mean (⫾ SD) ETDRS scores before TA injection, and 4, 12, and 24 weeks thereafter, are shown in Table 3. At no time was the difference between the ETDRS scores for injected and control eyes significant. Also, at 12 weeks the mean (⫾ SD) visual gain was 4.9⫾6.7 (range, ⫺7 to 13) in treated eyes, versus ⫺1⫾3.9 (range, ⫺7 to 5) in control eyes. By 24 weeks, the corresponding values were 4⫾6.5 (range, ⫺4 to 13) and ⫺3.1⫾4.1 (range, ⫺11 to 1). The difference between the relative changes in ETDRS scores of injected and control eyes was not significant at 12 weeks (0.05⬍P⬍0.1) or at 24 weeks (P ⫽ 0.1, bilateral Wilcoxon test for paired samples). However, at 12 weeks BCVA had improved by ⱖ2 lines in 5 of 12 injected eyes, versus none in the control eyes. The mean (⫾ SD) values for IOP in injected and control eyes are given in Table 4. In 6 of the 12 injected eyes, IOP exceeded 25 mmHg, and was controlled by topical medication with apraclonidine or dorzolamide. The maximal IOP observed after TA injection was 32 mmHg. Intraocular pressure increased after injection at various times, ranging from 2 days to 6 weeks. No cataract progression or other injection-related complications were observed in our study.

Discussion The preliminary results of this prospective controlled study demonstrate the efficacy of one intravitreal injection of TA in reducing macular thickening due to refractory diffuse diabetic ME, at least in the short term. In all injected eyes, we indeed observed, compared with control eyes, a dramatic anatomic improvement in diabetic ME that had been resistant to all previous treatments (Fig 2). This improvement occurred as soon as 1 month after injection. These results confirm those previously reported by Martidis et al, who, in an uncontrolled study, reported 55% and 58% reductions in macular thickness by 1 and 3 months, respectively, after intravitreal injection of 4 mg of TA for diffuse diabetic ME.5 Jonas et al also reported favorable results with a higher dose of TA (25 mg).9,10 Our protocol was designed primarily to detect differ-

Injected Eyes

Control Eyes

47.8 ⫾ 13 53.4 ⫾ 11.3 52.7 ⫾ 10.8 54.7 ⫾ 7.6 n⫽7

51.9 ⫾ 14.6 52.5 ⫾ 15.1 50.8 ⫾ 14.3 50.6 ⫾ 18.4 n⫽7

P of ETDRS Variation between Groups ⬎0.1 ⬎0.1 ⬎0.1 ⬎0.1

ences between injected and control eyes, as regards anatomic outcome. We therefore chose change in CMT as the main end point, because in ME the reduction of macular thickness is a prerequisite for visual improvement. In addition, OCT allows objective, accurate, and reproducible assessment of macular thickness.8 In diabetic patients, glycemic control, blood pressure, and nephropathy may affect macular thickness.11,12 We therefore excluded patients with renal insufficiency from the study, and to minimize the effects of high blood pressure and glycemia on ME, we included only diabetic patients with satisfactory glycemic and blood pressure control. In addition, we used the paired eyes of each patient to avoid possible bias by these 2 factors. Note that, in patients 3, 6, 8, and 12, we observed significant fluctuations in the macular thickness of control eyes (Table 2), which underlined the need for a control goup. However, despite these fluctations in the control group, the differences between relative changes in retinal thickness were significant at 4 and 12 weeks. We did not demonstrate any significant difference between injected and control eyes as regards the changes in the ETDRS scores. There was no significant difference between injected and control eyes as regards initial VA before injection, duration of ME, and extent of macular capillary closure, a finding that might explain the lack of significant visual improvement with triamcinolone. Nevertheless, a trend towards improvement in VA was observed at 3 months in injected eyes. A future trial including a larger series of patients may allow demonstration of a significant benefit for VA. In addition, in the present investigation, the long-standing diabetic ME that all of the eyes had might have caused permanent retinal damage, which prevented visual improvement, so that a shorter interval between the Table 4. Mean (⫾ Standard Deviation) Intraocular Pressure (IOP) in Injected and Contralateral Control Eyes before Triamcinolone Acetonide (TA) Injection and 4, 12, and 24 Weeks Thereafter

IOP

Injected Eyes (mmHg)

Control Eyes (mmHg)

P of IOP Variation between Groups

Before TA injection At 4 wks At 12 wks At 24 wks (n ⫽ 7)

18.2 ⫾ 3.2 20.8 ⫾ 4.6 20.5 ⫾ 3.6 18.7 ⫾ 3.6

18.2 ⫾ 2.5 16.7 ⫾ 3.7 16.7 ⫾ 2.7 17.6 ⫾ 2.8

⬎0.1 ⬍0.01 0.01 ⬎0.1

221

Ophthalmology Volume 111, Number 2, February 2004

Figure 2. Evolution of diabetic macular edema (ME) after intravitreal injection of triamcinolone acetonide (TA) in the right eye of patient 2. Before injection, best-corrected visual acuity (VA) was 20/50. a, Color fundus photograph after 3 sessions of laser photocoagulation. b, Late-phase angiogram showing cystoid ME. c, Six-millimeter optical coherence tomography (OCT) scan before TA injection: the central macula has a thickened appearance, with large intraretinal hyporeflective cystoid spaces. d, Corresponding macular mapping. Macular thickness is displayed as a 2-dimensional false-color map (left), with brighter colors (red and white) indicating areas of increased retinal thickness. It is numerically reported as averages in each of the 9 Early Treatment Diabetic Retinopathy Study areas (right). Central macular thickness (CMT) is 525 ␮m. e, Four weeks after TA injection, VA is 20/40. On OCT, CMT has dramatically decreased. Intraretinal cysts have disappeared. f, Corresponding macular mapping. Central macular thickness is 188 ␮m. g, h, Twenty-four weeks after TA injection, ME has recurred; CMT has greatly increased, to 612 ␮m; and VA has dropped to 20/50. D ⫽ diameter.

222

Massin et al 䡠 Intravitreal Triamcinolone Acetonide for Diabetic Macular Edema initial diagnosis of diabetic ME and TA injection might have improved the visual outcome. Slightly better visual results were obtained by Martidis et al5 and Jonas et al9,10; however, in these studies, VA measurements were not standardized. In the present series, the trend towards improvement in VA had disappeared by 6 months after injection, because diffuse diabetic ME recurred in 5 of 12 eyes, from 12 to 24 weeks thereafter. Among the 7 injected eyes followed up for 24 weeks, ME recurred in 4. Recurrences of ME were expected, and their number will no doubt increase with extended follow-up. Triamcinolone acetonide provides therapeutic effects for as long as TA depots are visible in the vitreous cavity, and Beer et al calculated that measurable concentrations of triamcinolone could be expected to last for approximately 3 months in nonvitrectomized eyes.13 In our study, the interval between TA injection and ME recurrence was in agreement with these findings. A similar recurrence of ME was also observed in previous studies of the effect of intravitreal TA.4,5,9 The main side effect observed here was IOP elevation, which occurred in 6 of 12 injected eyes (50%), at intervals ranging from 2 days to 6 weeks after injection. In our study, IOP elevation was controlled by topical medication in all cases. This elevation is a known side effect of corticosteroids administered topically or systematically in about one third of the general population.14 In previous studies of intravitreal TA administration, it was reported in 20% to 80% of patients.4,5,9,15–17 To detect susceptibility to corticosteroid-induced increased IOP, we tested all eyes with topical dexamethasone before TA injection. Although of questionable value,18 this test may have helped to avoid severe intractable rises in IOP. In any case, we think that patients with glaucoma or a history of corticosteroid-induced increased IOP should not be given intravitreal TA. We did not observe any cataract progression in our study. However, the incidence of cataract progression may well increase with longer follow-up and repetitive intravitreal injections. No other injection-related complications were observed in our study, such as retinal detachment or endophthalmitis.19 The dose of TA that we injected (4 mg) was chosen empirically, and was the dose injected in most previous studies.5,15,16,20 –24 Antcliff et al, who injected a lower dose (2 mg) for uveitic cystoid ME, observed an effect on ME similar to the effect seen here, but with a shorter interval of recurrence4; despite this lower dose, they also observed one serious case of IOP elevation, which required filtering surgery. Jonas, who injected 25 mg of TA, did not observe a longer period of TA efficacy, and IOP elevation was not greater than in our study.9,17 These studies provide no evidence that TA has any dose-related effects or complications. Nevertheless, further studies are necessary to define the optimal dose of TA to be injected. In conclusion, intravitreal injection of 4-mg TA seems effective in improving diffuse diabetic ME, at least in the short term. Intravitreal administration is especially advisable for diabetic patients, because it allows a high intraocular concentration of steroids while avoiding systemic side effects, especially blood glucose perturbations. The safety

of intravitreal corticosteroids has been supported by the results of previous animal studies and human trials.16,20 –22 However, the mechanism of corticosteroid action on this complication of diabetic microangiopathy remains unclear. It is not specific to diabetic microangiopathy, because intravitreal TA is also effective for ME secondary to uveitis, venous occlusion, and cataract surgery.4,23–25 Two main hypotheses are considered acceptable: (1) that corticosteroids might reduce retinal capillary permeability by increasing the activity and/or density of the tight junctions in the retinal capillary endothelium26 and (2) that corticosteroids might inhibit the metabolic pathway of the vascular endothelial growth factor, but this has never been documented in the eye.27 Intravitreal TA injection seems a promising treatment for refractory diffuse diabetic ME, but further studies are required to demonstrate that it improves vision. The occurrence of a relapse may justify further injections, whose tolerance and frequency will also have to be evaluated. Intraocular pressure elevation may be the limiting factor of this treatment. If intravitreal TA proves to be safe and effective, its place in the treatment of diffuse diabetic ME will have to be clearly defined; however, in this chronic disease, the need to repeat intravitreal injections every 4 to 6 months will increase the risk of injection-related complications, and may not be tolerated by the patients. In such cases, sustained drug delivery devices containing steroids that produce constant intraocular drug levels for an extended period may be an interesting alternative.28

References 1. Klein R, Klein BE, Moss SE. Visual impairment in diabetes. Ophthalmology 1984;91:1–9. 2. Early Treatment Diabetic Retinopathy Study Research Group. Photocoagulation for diabetic macular edema. Early Treatment Diabetic Retinopathy Study report number 1. Arch Ophthalmol 1985;103:1796 – 806. 3. Lee CM, Olk RJ. Modified grid laser photocoagulation for diffuse diabetic macular edema: long-term visual results. Ophthalmology 1991;98:1594 – 602. 4. Antcliff RJ, Spalton DJ, Stanford MR, et al. Intravitreal triamcinolone for uveitic cystoid macular edema: an optical coherence tomography study. Ophthalmology 2001;108:765– 72. 5. Martidis A, Duker JS, Greenberg PB, et al. Intravitreal triamcinolone for refractory diabetic macular edema. Ophthalmology 2002;109:920 –7. 6. Chylack LT Jr, Wolfe JK, Singer DM, et al, Longitudinal Study of the Cataract Study Group. The Lens Opacities Classification System (LOCS) III. Arch Ophthalmol 1993;111: 831– 6. 7. Massin P, Erginay A, Haouchine B, et al. Retinal thickness measurement in healthy subjects using optical coherence tomography mapping software. Eur J Ophthalmol 2002; 12:102– 8. 8. Massin P, Vicaut E, Haouchine B, et al. Reproducibility of retinal mapping using optical coherence tomography. Arch Ophthalmol 2001;119:1135– 42. 9. Jonas JB, Degenring RF. Intravitreal injection of crystalline triamcinolone acetonide in the treatment of diffuse diabetic

223

Ophthalmology Volume 111, Number 2, February 2004

10. 11. 12.

13. 14. 15. 16. 17. 18. 19.

macular oedema [in German]. Klin Monatsbl Augenheilkd 2002;219:429 –32. Jonas JB, Kreissig I, So¨ fker A, Degenring RF. Intravitreal injection of triamcinolone for diffuse diabetic macular edema. Arch Ophthalmol 2003;121:57– 61. Bresnick GH. Diabetic macular edema: a review. Ophthalmology 1986;93:989 –97. Stratton IM, Kohner EM, Aldington SJ, et al, UKPDS Group. UKPDS 50: risk factors for incidence and progression of retinopathy in type II diabetes over 6 years from diagnosis. Diabetologia 2001;44:156 – 63. Beer PM, Bakri SJ, Singh RJ, et al. Intraocular concentration and pharmacokinetics of triamcinolone acetonide after a single intravitreal injection. Ophthalmology 2003;110:681– 6. Schwartz JT, Reuling FH, Feinleib M, et al. Twin study on ocular pressure after topical dexamethasone. 1. Frequency distribution of pressure response. Am J Ophthalmol 1973;76:126 –36. Wingate RJ, Beaumont PE. Intravitreal triamcinolone and elevated intraocular pressure. Aust N Z J Ophthalmol 1999; 27:431–2. Young S, Larkin G, Branley M, Lightman S. Safety and efficacy of intravitreal triamcinolone for cystoid macular oedema in uveitis. Clin Experiment Ophthalmol 2001;29:2– 6. Jonas JB, Kreissig I, Degenrig RF. Intraocular pressure after intravitreal injection of triamcinolone acetonide. Br J Ophthalmol 2003;87:24 –7. Palmberg PF, Mandell A, Wilensky JT, et al. The reproducibility of the intraocular pressure response to dexamethasone. Am J Ophthalmol 1975;80:844 –56. Benz MS, Murray TG, Dubovy SR, et al. Endophthalmitis caused by Mycobacterium chelonae abcessus after intravitreal injection of triamcinolone. Arch Ophthalmol 2003;121:271–3.

20. Penfold PL, Gyory JF, Hunyor AB, Billson FA. Exudative macular degeneration and intravitreal triamcinolone: a pilot study. Aust N Z J Ophthalmol 1995;23:293– 8. 21. McCuen BW II, Bessler M, Tano Y, et al. The lack of toxicity of intravitreally administered triamcinolone acetonide. Am J Ophthalmol 1981;91:785– 8. 22. Hida T, Chandler D, Arena JE, Machemer R. Experimental and clinical observations of the intraocular toxicity of commercial corticosteroid preparations. Am J Ophthalmol 1986; 101:190 –5. 23. Greenberg PB, Martidis A, Rogers AH, et al. Intravitreous triamcinolone acetonide for macular oedema due to central retinal vein occlusion [letter]. B J Ophthalmol 2002;86:247– 8. 24. Benhamou N, Massin P, Haouchine B, et al. Intravitreal triamcinolone for refractory pseudophakic macular edema. Am J Ophthalmol 2003;135:246 –9. 25. Jonas JB, Kreissig I, Degenring RF. Intravitreal triamcinolone acetonide as treatment of macular edema in central retinal vein occlusion. Graefes Arch Clin Exp Ophthalmol 2002; 240:782–3. 26. Antonetti DA, Wolpert EB, DeMaio L, et al. Hydrocortisone decreases retinal endothelial cell water and solute flux coincident with increased content and decreased phosphorylation of occludin. J Neurochem 2002;80:667–77. 27. Fisher S, Renz D, Schaper W, Karliczek GF. In vitro effects of dexamethasone on hypoxia-induced permeability and expression of vascular endothelial growth factor. Eur J Pharmacol 2001;411:231– 43. 28. Jaffe GJ, Ben-nun B, Guo H, et al. Fluocinolone acetonide sustained drug delivery device to treat severe uveitis. Ophthalmology 2000;107:2024 –33.

Discussion by Jay S. Duker, MD The fact that corticosteroids delivered to the posterior segment of the eye are beneficial in the treatment of macular edema related to inflammation is widely accepted. The idea that similar therapy might benefit macular edema from noninflammatory causes such as diabetic retinopathy is not intuitively obvious. Nevertheless, there is growing anecdotal, retrospective, and prospective evidence that intravitreal corticosteroids in the form of either triamcinolone or a sustained-release corticosteroid implant are effective in reducing macular thickness in diabetic macular edema.1 Massin and coworkers report the first prospective controlled trial of intravitreal triamcinolone versus observation in eyes with diffuse diabetic macular edema that failed previous conventional treatment via laser photocoagulation. Their selection criteria and protocol were strict, to minimize complications. Pretreatment with topical corticosteroids was performed before intravitreal injection to eliminate eyes that were steroid responders. Only patients with bilateral diffuse macular edema secondary to diabetic retinopathy and unresponsive to previous focal laser photocoagulation were enrolled. After 2 patients were eliminated due to elevated intraocular pressure (IOP) on topical corticosteroids, a total of 12 patients were enrolled in this trial and observed for a minimum of 3 months postinjection. Interestingly, the authors chose anatomic success From the New England Eye Center, New England Medical Center, Tufts University School of Medicine, Boston, Massachusetts. Address correspondence to Jay Duker, MD, New England Eye Center, Tufts University School of Medicine, 750 Washington Street, Box 450, Boston, MA 02111-1533.

224

based on optical coherence tomography as the primary end point, whereas visual success based on improvement in best-corrected Early Treatment Diabetic Retinopathy Study vision was a secondary end point. All patients received 4 mg of triamcinolone acetonide intravitreally under sterile conditions and subconjunctival anesthesia. A 30-gauge needle was used, 4 mm posterior to the limbus. Only one eye received an injection, with the fellow serving as an untreated control. The protocol did not allow for reinjection at any time during the follow-up period. The results of the trial showed that at follow-up visits 4 weeks and 12 weeks after injection, intravitreal trimacinolone had a markedly beneficial effect on retinal thickness relative to the fellow eyes that were observed. In fact, the mean retinal thickness in treated eyes at week 4 was 207 ␮m (i.e., essentially normal), compared with a preinjection thickness mean of 509 ␮m. The concurrent control group had a mean foveal thickness at week 4 of 469 ␮m. By 24 weeks postinjection, when the therapeutic effect of the single injection had been gone for several weeks, at least, the beneficial anatomic effect had lessened considerably, with 5 of 12 treated eyes showing recurrence of significant thickening.2 No statistically significant difference in visual acuity between the treated and control eyes was documented, based on the total Early Treatment Diabetic Retinopathy Study letters score. However, at the 12-week visit, 5 of 12 treated eyes improved ⱖ2 lines, whereas none of the control eyes showed any significant visual improvement at any point in the study.