Diabetic macular edema (DME) remains a leading cause of vision loss in working-aged adults, despite significantly improved outcomes after the introduction of new therapies such as vascular endothelial growth factor (VEGF) inhibitors.1–3 However, there is evidence that VEGF is not always the sole pathology-mediating cytokine in DME. A recent study that analyzed the biochemical composition of aqueous humor demonstrated increasing inflammatory cytokine levels, but no change in VEGF content, as the severity of diabetic retinopathy increased.4 Diabetic macular edema can manifest at any stage of the diabetic retinopathy severity spectrum and before the development of retinal ischemia5,6; therefore, its pathogenesis cannot be attributed solely to an increase in VEGF levels. In fact, Phase 3 clinical trial data have demonstrated significantly improved visual acuity associated with intravitreal triamcinolone acetonide and implants releasing dexamethasone and fluocinolone acetonide (FAc) compared with controls.7–9
The hypothesis that DME is multifactorial has been supported by clinical trial data. A prespecified subanalysis of the fluocinolone acetonide for macular edema (FAME) trials demonstrated that FAc-treated patients with chronic DME (cDME) achieved a greater visual benefit than patients with nonchronic DME (ncDME).10,11 Furthermore, recent emerging data suggest that even with VEGF inhibitors, efficacy may be lower when DME is long-standing.12 Unlike other landmark DME trials that included treatment-naive participants,8,13,14 the FAME trials had eligibility criteria that required prior focal/grid macular laser.7 As a consequence, there were no treatment-naive patients in the FAME trials; therefore, the population was enriched with patients with cDME. Thermal laser therapy was permitted as a rescue treatment in the FAME trials. Although nonprotocol treatments were discouraged, both triamcinolone acetonide and bevacizumab, which became the standard of care during this period, were used in all treatment arms at the various clinical sites of the FAME trials. Thus, the sham-control cohort in the FAME trials, which received intermittent therapy with a variety of agents, can be compared with the cohort that received continuous treatment with the FAc implant.
In this post hoc analysis, we compare the functional outcomes of patients with cDME and ncDME in the sham and 0.2 μg/day FAc-treatment arms based on three nonexclusive baseline visual acuity strata (≤20/100, ≤20/80, and ≤20/64). These strata evolved as a result of an analysis that was conducted as part of a health technology assessment to evaluate cost-effectiveness. This assessment included an examination of the efficacy of the 0.2 μg/day FAc implant across the three nonexclusive baseline visual acuity strata. Additionally, clinical trials typically enroll patients with a range of visual acuities rather than by specific vision strata; therefore, nonexclusive strata are more representative of how visual acuity groups would be defined in a clinical trial. We hypothesized that because patients within each treatment arm were treated similarly regardless of DME duration, differences in visual acuity response would reflect differences in the retinal microenvironment. We compared the degree of similarity in baseline characteristics between patients with cDME versus those with ncDME in the sham-control and 0.2 μg/day FAc arms and also the change in center point thickness (CPT) and visual acuity over the 36 months of the study by baseline vision strata.
FAME Study Design
The study population and design of the FAME trials have been described previously.7 Briefly, FAME comprised two multicenter, randomized, double-masked, sham-controlled, parallel-group studies performed under a single protocol (C-01-05-001, sponsored by Alimera Sciences, Inc, Alpharetta, GA) that compared FAc intravitreal implants (0.2 or 0.5 μg/day) with sham-control injection (± laser photocoagulation ± nonprotocol therapies [including VEGF inhibitors and triamcinolone acetonide]) in patients with DME. The former provided continuous therapy, whereas the latter was intermittent. Patients were eligible to participate in the FAME trials if they had a foveal CPT ≥250 μm despite ≥1 prior focal/grid macular photocoagulation treatment and best-corrected visual acuity (BCVA) between 19 and 68 Early Treatment Diabetic Retinopathy Study (ETDRS) letters (Snellen equivalent, 20/50–20/400). Key study assessments included BCVA (measured using ETDRS charts at 4 m or an electronic visual acuity tester at 3 m), foveal CPT (measured using the Fast Macular Scan protocol on a Stratus three optical coherence tomography instrument [Carl Zeiss Meditec, Dublin, CA]), and adverse events (AEs; in the case of glaucomatous change to the optic nerve, the University of Wisconsin Fundus Photograph Reading Center performed optic nerve head grading15).
During the trials, patients could be retreated with their assigned study treatment between months 12 and 33 if progression of edema was evident (BCVA loss of ≥5 ETDRS letters or an increased foveal thickness of ≥50 µm from best reading in previous 12 months) according to the assessing (masked) investigator. At the beginning of week 6, patients were permitted to receive rescue (focal/grid or panretinal photocoagulation) laser therapy for persistent or recurrent DME. In the FAME trials, patients who received nonprotocol intermittent therapy (triamcinolone acetonide and VEGF inhibitors) were not excluded from statistical analysis.
Subanalysis of the FAME Data
This was a post hoc subanalysis in which patients were grouped by nonexclusive categories based on baseline BCVA. Groups were defined by a difference of 5 ETDRS letters (≤20/100 [≤53 letters], ≤20/80 [≤58 letters], ≤20/64 [≤63 letters]; Figure 1) and were not mutually exclusive.
In the current analysis, DME chronicity (nonchronic vs. chronic) was defined based on the median duration of DME reported by patients at the baseline in the FAME trials (3 years).10
The intent-to-treat principle was used for all efficacy analyses. The method of last observation carried forward was used to impute values for all missing data. For analyses of baseline characteristics, the primary efficacy endpoint, and secondary efficacy endpoints based on binary variables, a comparison between treatment arms was made using a Cochran–Mantel–Haenszel Chi-square test. For analysis of baseline characteristics and secondary endpoints that were continuous variables, comparisons between treatment arms were made using an analysis of variance model with treatment arm as fixed effects.
Baseline characteristics were similar among patients regardless of the treatment arm or DME duration; however, there was a greater percentage of phakic patients among those with ncDME than cDME (Table 1; because patients were categorized by DME duration, the differences in the duration of disease are to be expected). In a subgroup analysis of patients with DME based on the disease duration, the effect of 0.2 μg/day FAc versus sham control was similar between all patients and those who were pseudophakic at the baseline,11 which suggests that outcomes based on DME duration and baseline vision should not be affected by the baseline lens status.
Among patients assigned to sham control, retreatment rates were higher in patients with cDME compared with those with ncDME, and this difference reached statistical significance (33.9% vs. 20.5%; P = 0.039; Table 2). A similar difference was not observed in the 0.2 μg/day FAc arm. Table 2 also shows that the proportion of patients with cDME who received off-protocol therapies was higher than those with ncDME in the sham-control arm, and this differential was not seen in the FAc arm.
The incidence of cataract was greater in the 0.2 μg/day FAc arm compared with that in the sham-control arm, with little difference based on DME duration in each arm. Cataract extraction rates were higher among patients with cDME than those with ncDME, regardless of treatment assignment, although the difference in surgery rates was more pronounced in the sham-control arm compared with that in the 0.2 μg/day FAc arm (Table 3).
Within each treatment arm, use of intraocular pressure-lowering medication was similar regardless of DME duration. Only one patient in the sham-control arm required intraocular pressure-lowering surgery, and this was a patient with ncDME. Among patients who received 0.2 μg/day FAc, a similar percentage of patients with cDME and ncDME required intraocular pressure-lowering surgery.
Foveal Center Point Thickness Outcomes and Visual Acuity in Patients who Received Sham Control or 0.2 μg/day Fluocinolone Acetonide: Results Stratified by Diabetic Macular Edema Chronicity
Patients in the sham-control and 0.2 μg/day FAc arms achieved comparable improvements in CPT, regardless of DME chronicity (Figure 2, A and B, respectively). However, among those who received sham contral, a significantly greater percentage of patients with ncDME achieved a ≥15-letter improvement in BCVA compared with patients with cDME (27.8% vs. 13.4%; P = 0.012; Figure 2C). This finding was not driven by the frequency of rescue laser treatment or nonprotocol therapies received by the 2 groups, as they were comparable (Table 2). Among those treated with 0.2 μg/day FAc, a significantly greater percentage of patients with cDME achieved a ≥15-letter improvement in BCVA than those with ncDME (34.0% vs. 22.3%; P = 0.029; Figure 2D). This finding was not influenced by the frequencies of rescue laser treatment or nonprotocol therapies that were delivered in addition to the FAc as these were similar in the groups with ncDME and cDME (Table 2).
Visual Acuity in Patients who Received Sham Control: Results Stratified by Diabetic Macular Edema Chronicity and Baseline Visual Acuity
In the sham-control arm, a significantly greater percentage of those with ncDME gained ≥15 letters of BCVA compared with those with cDME (Figure 3A). However, the percentages of patients with cDME who gained ≥15 letters of BCVA were similar across baseline visual acuity strata. Within each baseline vision stratum, a significantly greater percentage of patients with ncDME gained ≥15 letters of BCVA compared with those with cDME.
In the sham-control arm, a numerically, but not statistically significant, greater percentage of patients with ncDME achieved ≥20/40 visual acuity versus those with cDME (Figure 3B). The percentages of patients with ncDME who achieved a visual outcome of ≥20/40 were similar in the ≤20/64, ≤20/80, and ≤20/100 strata. However, the percentage of patients with cDME who achieved a visual outcome of ≥20/40 was lower in each successively lower baseline visual acuity stratum, with only 6.3% of patients in the lowest baseline vision stratum achieving ≥20/40 visual acuity. Within each baseline vision stratum, there was no statistically significant difference in the percentage of patients who achieved a visual outcome of ≥20/40 between those with ncDME and cDME.
Visual Acuity After Treatment With 0.2 μg/day Fluocinolone Acetonide: Results Stratified by Diabetic Macular Edema Chronicity and Baseline Visual Acuity
In the 0.2 μg/day FAc arm, a significantly greater percentage of patients with cDME compared with those with ncDME gained ≥15 letters of BCVA (Figure 4A). There was an increased percentage of patients who gained ≥15 letters of BCVA with each successively lower baseline visual acuity stratum, regardless of DME chronicity; however, for each baseline vision stratum, a numerically greater percentage of patients with cDME gained ≥15 letters of BCVA versus those with ncDME.
Among all patients who received 0.2 μg/day FAc, a significantly greater percentage of patients with cDME compared with those with ncDME achieved ≥20/40 visual acuity (Figure 4B). Similar percentages of patients with ncDME in the ≤20/64, ≤20/80, and ≤20/100 strata achieved a visual acuity outcome of ≥20/40; however, with each successively lower baseline visual acuity stratum, fewer patients with cDME achieved a visual outcome of ≥20/40.
This post hoc subanalysis of the FAME trials reports the efficacy and safety of 0.2 µg/day FAc (continuous therapy) versus sham control (various intermittent therapies) in patients with ncDME or cDME as a function of the baseline vision status. In this subanalysis, patients in each treatment arm were categorized by disease chronicity and baseline vision. Categorizing patients by baseline vision provided a reference point for the visual acuity improvement patterns typically observed in clinical trials. Patients with poor baseline vision generally experience greater visual gains after intervention compared with their counterparts with better vision.16–18 Patients with better vision at the time of intervention typically experience a treatment “ceiling effect” as a consequence of the limited potential for improvement.18 In the present report, patients with ncDME in the sham-control arm who were treated with intermittent therapy followed this trend. However, this trend was not observed among patients with cDME in the sham-control arm who were treated with intermittent therapy. By contrast, patients with cDME and poor baseline vision who were treated with continuous 0.2 μg/day FAc did experience improvements in vision in line with those previously observed. Categorizing patients by disease chronicity within each treatment arm allowed for observation of response to therapies with distinct mechanisms of action to support the hypothesis of change within the retina microenvironment associated with disease chronicity. Our findings show that disease chronicity plays an important role in the heterogeneity of functional outcomes in DME and that continuous low-dose corticosteroid therapy is particularly beneficial in patients who would otherwise be refractory to intermittent therapy.
The underlying pathology of an individual patient's DME may be manifested by their response to therapeutic agents with differing mechanisms of action. In the RIDE/RISE trials of patients with DME, those who received ranibizumab 2 years after randomization gained ≈ 2 letters of BCVA over 12 months (compared with an improvement of ≈ 10 letters over 12 months among those who received ranibizumab at randomization).2 Interestingly, the improvement was small despite an appreciable reduction in central foveal thickness (≈100 μm)2 suggesting the presence of persistent retinal morphologic abnormalities unrelated to vascular leakage. However, the literature is inconsistent regarding correlations between improved visual acuity and reductions in foveal thickness. The Diabetic Retinopathy Clinical Research Network Protocol A suggested little correlation between improved visual acuity and reduction of CPT.19 However, in the FAME trials, at 2 years postrandomization, patients who received FAc implants experienced both significant improvements in visual acuity and reductions in foveal thickness.7 In a previous publication arising out of the FAME trials, a prespecified subanalysis based on DME duration has demonstrated that patients with DME for >1.73 years gained nearly 6 ETDRS letters of BCVA after receiving 0.2 μg/day FAc for 12 months (The aforementioned median DME duration of 3 years was based on the year of diagnosis.)11 The median DME duration of 1.73 years is based on the specific day, month, and year of diagnosis with DME. There was a significant concordance in DME chronicity between the 2 algorithms, and ≈ 93% of patients retained their original categorization.11 Our data suggest that a low dose of corticosteroid may act through pathways that are not fully understood to improve functional recovery in chronic disease. Reports of improved vision with corticosteroid use among patients who previously experienced suboptimal results associated with anti-VEGF agents have been published,20–24 and at least one prospective randomized clinical trial addressing this specific question is ongoing.25
The difference in efficacy based on the DME duration and treatment arm suggests a change in cytokine composition within the retina microenvironment. Over the course of the study, the percentage of patients who experienced an improvement in BCVA of ≥15 ETDRS letters among those with ncDME who were treated with continuous 0.2 μg/day FAc therapy increased; however, this effect was significantly greater among patients with cDME. Corticosteroids per se are known to reduce VEGF levels, albeit not to the extent of anti-VEGF agents.26 Thus, the difference in functional outcomes based on DME duration among patients in the sham-control arm who received laser and intermittent therapy supports the notion that the retinal microenvironment composition varies with the duration of DME, particularly given the similar frequency of laser and nonprotocol therapy use among these patients. The efficacy data presented herein may reflect a disease that is primarily VEGF driven in patients with ncDME; however, as the duration of DME increases, VEGF may no longer be the primary pathological mediator and the corticosteroid effect is greater.
Visual acuity improvement associated with pharmacotherapies as a function of baseline vision has been explored in clinical trials. A subanalysis of the BOLT study, in which patients were treated with bevacizumab, demonstrated a numerically greater change in visual acuity at the study end among those with baseline visual acuity <54 ETDRS letters versus those with baseline visual acuity ≥54 letters.17 Also, the recently published Diabetic Retinopathy Clinical Research Network Protocol T trial reported numerically greater improvements across all treatment arms in patients with poor baseline vision compared with those who had better baseline vision.16 Our findings are in agreement with the results of these prior studies, with the exception of results in patients with cDME in the sham-control arm; however, treatment in these patients with continuous 0.2 μg/day FAc restored the expected outcome.
This report provides clinical evidence suggesting that as DME duration increases, the multifactorial nature of the disease becomes more prominent. Among those who received sham control (laser and intermittent therapies), distinct responder types were observed between patients with ncDME and cDME. Treatment patterns were similar among patients who received continuous 0.2 μg/day FAc, although greater efficacy was observed in patients with cDME compared with those with ncDME. In patients with cDME, the likelihood of achieving vision ≥20/40 with treatment diminishes as the vision worsens. Therefore, our findings suggest a need for close monitoring of such patients to avoid significant vision loss and are in accord with the indication approved by the European Medicines Agency for 0.2 μg/day FAc.27 Thus, 0.2 μg/day FAc represents a treatment option for patients who are unresponsive to alternative or intermittent therapies.
1. Wenick AS, Bressler NM. Diabetic macular edema: current and emerging therapies. Middle East Afr J Ophthalmol 2012;19:4–12.
2. Brown DM, Nguyen QD, Marcus DM, et al. Long-term outcomes of ranibizumab therapy for diabetic macular edema: the 36-month results from two phase III trials: RISE and RIDE. Ophthalmology 2013;120:2013–2022.
3. Korobelnik JF, Do DV, Schmidt-Erfurth U, et al. Intravitreal aflibercept for diabetic macular edema. Ophthalmology 2014;121:2247–2254.
4. Dong N, Xu B, Wang B, Chu L. Study of 27 aqueous humor cytokines in patients with type 2 diabetes with or without retinopathy. Mol Vis 2013;19:1734–1746.
5. Klein R, Klein B. Vision disorders in diabetes. In: National Diabetes Data Group, ed. Diabetes in America. 2nd ed. Bethesda, MD: National Institutes of Health, National Institute of Diabetes and Digestive and Kidney Diseases; 1995:293–338.
6. Klein R, Klein BE, Moss SE, et al. The Wisconsin epidemiologic study of diabetic retinopathy. IV. diabetic macular edema. Ophthalmology 1984;91:1464–1474.
7. Campochiaro PA, Brown DM, Pearson A, et al. Long-term benefit of sustained-delivery fluocinolone acetonide vitreous inserts for diabetic macular edema. Ophthalmology 2011;118:626–635.e2.
8. Boyer DS, Yoon YH, Belfort R Jr, et al. Three-year, randomized, sham-controlled trial of dexamethasone intravitreal implant in patients with diabetic macular edema. Ophthalmology 2014;121:1904–1914.
9. Diabetic Retinopathy Clinical Research Network. A randomized trial comparing intravitreal triamcinolone acetonide and focal/grid photocoagulation for diabetic macular edema. Ophthalmology 2008;115:1447–1449, 1449.e1–10.
10. Campochiaro PA, Brown DM, Pearson A, et al. Sustained delivery fluocinolone acetonide vitreous inserts provide benefit for at least 3 years in patients with diabetic macular edema. Ophthalmology 2012;119:2125–2132.
11. Cunha-Vaz J, Ashton P, Iezzi R, et al. Sustained delivery fluocinolone acetonide vitreous implants: long-term benefit in patients with chronic diabetic macular edema. Ophthalmology 2014;121:1892–1903.e3.
12. Pearce I, Banerjee S, Burton BJ, et al. Ranibizumab 0.5 mg for diabetic macular edema with bimonthly monitoring after a phase of initial treatment: 18-month, multicenter, phase IIIB RELIGHT study. Ophthalmology 2015;122:1811–1819.
13. Diabetic Retinopathy Clinical Research Network, Elman MJ, Aiello LP, Beck RW, et al. Randomized trial evaluating ranibizumab plus prompt or deferred laser or triamcinolone plus prompt laser for diabetic macular edema. Ophthalmology 2010;117:1064–1077.e35.
14. Nguyen QD, Brown DM, Marcus DM, et al. Ranibizumab for diabetic macular edema: results from 2 phase III randomized trials: RISE and RIDE. Ophthalmology 2012;119:789–801.
15. Parrish RK II, Traverso CE, Green K, et al.; for the FAME Study Group. Quantitative assessment of optic nerve changes in patients with diabetic macular edema treated with fluocinolone acetonide vitreous implants. Ophthalmic Surg Lasers Imaging Retina 2016;47:418–425.
16. Diabetic Retinopathy Clinical Research Network, Wells JA, Glassman AR, Ayala AR, et al. Aflibercept, bevacizumab, or ranibizumab for diabetic macular edema. N Engl J Med 2015;372:1193–1203.
17. Sivaprasad S, Crosby-Nwaobi R, Esposti SD, et al. Structural and functional measures of efficacy in response to bevacizumab monotherapy in diabetic macular oedema: exploratory analyses of the BOLT study (report 4). PLoS One 2013;8:e72755.
18. Bressler SB, Qin H, Beck RW, et al. Factors associated with changes in visual acuity and central subfield thickness at 1 year after treatment for diabetic macular edema with ranibizumab. Arch Ophthalmol 2012;130:1153–1161.
19. Diabetic Retinopathy Clinical Research Network, Browning DJ, Glassman AR, Aiello LP, et al. Relationship between optical coherence tomography-measured central retinal thickness and visual acuity in diabetic macular edema. Ophthalmology 2007;114:525–536.
20. Elaraoud I, Andreatta W, Kidess A, et al. Use of flucinolone acetonide for patients with diabetic macular oedema: patient selection criteria and early outcomes in real world setting. BMC Ophthalmol 2016; 16:3–015-0178-9.
21. Arikan Yorgun M, Toklu Y, Mutlu M, et al. Efficacy of single-dose dexamethasone implantation in patients with persistent diabetic macular edema. Int Ophthalmol 2015;36:531–539.
22. Jeon S, Lee WK. Effect of intravitreal triamcinolone in diabetic macular edema unresponsive to intravitreal bevacizumab. Retina 2014;34:1606–1611.
23. Quhill F. Real-world experience of fluocinolone acetonide (0.2 mcg/day) intravitreal implant in the treatment of diabetic macular oedema. Eur Ophthalmic Rev 2015;9:42–46.
24. Elaraoud I, Attawan A, Quhill F. Case series investigating the efficacy and safety of bilateral fluocinolone acetonide (ILUVIEN) in patients with diabetic macular edema. Ophthalmol Ther 2016;5:95–104.
25. ClinicalTrials.gov. Phase II combination steroid and anti-VEGF for persistent DME. Available at: https://clinicaltrials.gov/ct2/show/NCT01945866
. Accessed August 29, 2016.
26. Sohn HJ, Han DH, Kim IT, et al. Changes in aqueous concentrations of various cytokines after intravitreal triamcinolone versus bevacizumab for diabetic macular edema. Am J Ophthalmol 2011;152:686–694.
27. ILUVIEN summary of product characteristics. Aldershot, United Kingdom: Alimera Sciences Limited; 2013. Available at: https://www.medicines.org.uk/emc/medicine/27636
. Accessed August 29, 2016.