Article Text
PDF
Abstract
Purpose To analyse the effects of intravitreal dexamethasone implant in patients suffering from diabetic macular oedema (DME) on the basis of their visual and functional response to antivascular endothelial growth factor (VEGF) loading dose, in order to early shift to corticosteroids in poorly responding patients.
Design Retrospective monocentric study.
Methods Data of patients with diabetes shifted to 0.7 mg dexamethasone implant after three injections of ranibizumab (RNB) and followed-up to 12 months were reviewed. Main outcome was the evaluation of short-term changes after dexamethasone implant injection, stratifying patients on the basis of best-corrected visual acuity (BCVA) and central macular thickness (CMT) after RNB loading dose. Secondary outcome was to investigate clinical gain maintenance at long-term follow-up.
Results Overall, 45 eyes of 45 patients (23 males, 51.1%), mean age 69.7±9 years, were included in the analysis. After 3 injections of RNB, 30 eyes (66.7%) had a poor visual response (−4.3±10.7 letters), while 15 eyes (33.3%) disclosed good visual outcome (+13.9±9.2 letters). Patients with poor visual response were associated with limited morphological improvement (p=0.04). After 1 month from dexamethasone, only poor responders showed relevant increase in BCVA (p=0.006) and reduction in CMT (p=0.002), in comparison to good visual response patients, featuring only minor clinical effects (p=0.3). The same trend was maintained up to 12 months, after a mean of 1.9±1.1 dexamethasone administrations.
Conclusion Visual and anatomical responses after RNB loading dose are significant predictors of both early term and long-term visual acuity improvement after switching to corticosteroids in patients with DME unresponsive to anti-VEGF.
- diabetic macular edema
- anti-VEGF
- dexamethasone
- ranibizumab
Statistics from Altmetric.com
Introduction
Diabetic macular oedema (DME) still represents the main cause of visual impairment in patients with diabetes mellitus (DM), and approximately 50% of patients with DME lose two or more lines of visual acuity within 2 years loss if left untreated.1 2 Laser photocoagulation has been considered the gold standard for treatment of DME for the last 30 years, with reduction of the risk of moderate visual loss only by 50% and limited visual improvement (≥3 lines visual gain equal to 3%).3 Intravitreal antivascular endothelial growth factor (anti-VEGF) agents, as monotherapy or in association with laser photocoagulation, have demonstrated to provide greater improvements in both visual acuity and macular thickness than laser treatment alone and currently constitute first-line therapy for most patients with DME.4 5 Nevertheless, an appreciable proportion of patients with DME respond suboptimally to anti-VEGF therapy.
Recently, in a subanalysis of Protocol I, it has been demonstrated that best-corrected visual acuity (BCVA) response at 12 weeks was significantly associated with ∆BCVA at 3 years follow-up, and that poorly responders to ranibizumab (RNB) loading dose (ie, eyes gaining <5 letters after three intravitreal RNB injections administered monthly) have limited/no chances to further improve with continuous RNB injections.6
Neither the Protocol I study nor the current analysis addresses the question of whether patients with suboptimal response to ranibizumab with prompt or deferred laser treatment are likely to achieve better visual acuity outcomes with alternative therapies, or whether any specific treatment pathway offers a potential efficacy advantage.
The use of steroids administered intravitreally has been introduced as a valid alternative in those patients non-responder to anti-VEGF treatment. Intravitreal injection of sustained-release dexamethasone implant (DEX implant 0.7 mg; Ozurdex; Allergan, Irvine, California, USA) has demonstrated effectiveness in recalcitrant DME, along with limited side affects.7 8 Since long-standing macular oedema can permanently damage retinal pigment epithelium (RPE) and photoreceptors, leading to retinal fibrosis and macular atrophy, prompt identification of poorly responding patients to anti-VEGF therapy is crucial for early shift to alternative anti-DME treatments.9
The primary aim of the present study is to analyse the functional and morphological effects of DEX in patients suffering from DME on the basis of their visual and functional response to RNB loading dose, in order to justify early shift to corticosteroids in poorly responders.
Methods
This is a retrospective monocentric study. Data of consecutives patients with DME who switched to DEX implant after intravitreal injections of RNB at the Department of Ophthalmology of the University Scientific Institute San Raffaele in Milan between November 2014 and June 2016 were analysed. All patients signed a written general consent, which was approved by the ethics committee of San Raffaele Hospital, and the study was conducted in agreement with the Declaration of Helsinki for research involving human subjects.
Inclusion criteria were: (1) age ≥18 years; (2) diagnosis of DM, either type 1 or type 2; (3) presence of clinically significant DME according to EDTRS guidelines; (4) BCVA between 20/200 and 20/20; (5) central macular thickness (CMT) ≥300 µm as measured by spectral domain optical coherence tomography (SD-OCT); (6) follow-up of at least 12 months.10
Exclusion criteria were: (1) macular oedema secondary to other causes than diabetic retinopathy (eg, retinal vein occlusion, age-related macular degeneration, postsurgical macular oedema); (2) significant media opacities limiting OCT quality (eg, corneal opacity, cataract, vitreous haemorrhage); (3) history of ocular trauma or surgery ≤6 months before the first DEX injection; (4) any other intravitreal corticosteroid administration before DEX injection; (5) intravitreal anti-VEGF (bevacizumab, ranibizumab or pegaptanib) <1 month before the DEX injection; (6) uncontrolled glaucoma, defined as intraocular pressure (IOP) >25 mm Hg despite antiglaucoma medication in the study eye.
All the patients included in the study underwent three loading-dose intravitreal injections of RNB performed at fixed 4-week intervals for the first 12 weeks. Then, all patients, regardless of functional and anatomical characteristics, were shifted to 0.7 mg DEX implant injections continued at 4-month intervals until stable BCVA was reached (figure 1).
Study protocol design. RNB, ranibizumab; OCT, optical coherence tomography; BCVA, best-corrected visual acuity; DEX, dexamethasone.
Baseline demographic and clinical data of the study population, including age, sex, haemoglobin A1c (HbA1c) level, duration of DME, BCVA using standard ETDRS charts, IOP, lens status, CMT, previous and subsequent DME treatments were recorded. During the study period, all patients underwent complete ophthalmic evaluation, including BCVA, slit-lamp biomicroscopy, applanation tonometry, fundus biomicroscopy, SD-OCT (Spectralis; Heidelberg Engineering, Heidelberg, Germany) with CMT measurement after RNB loading dose, and at monthly basis after DEX administration. CMT was obtained through 19 horizontal lines (6×6 mm area) centred on the fovea, each with 9 averaged OCT B-scans-1024 A scans per-line at 240 µm intervals. Patients featuring either internal limiting membrane (ILM) thickening or frank epiretinal membrane were excluded from the analysis.
Based on the BCVA reading after three RNB injections, study eyes were categorised into two groups: (A) poor visual response (<5 letter improvement from baseline) and (B) good visual response (≥5 letter improvement), whereas based on the CMT at week 12±2, were divided into: (A) poor anatomical response (no oedema lowering or worsening from baseline) and (B) good anatomical response (any oedema lowering). To exclude that any association between baseline and subsequent BCVA responses have been due to truncation of treatment effect in eyes with well-preserved visual acuity (‘ceiling effect’), a further analysis was conducted after exclusion of eyes with baseline BCVA ≥95 ETDRS letters.
Statistical analysis including descriptive statistics for demographics and main clinical records, and comparative analysis (Student’s t-test analysis for independent and paired samples and one-way analysis of Variance with Bonferroni correction) were performed through GraphPad Prism 5.0 (GraphPad Software, San Diego, California, USA). The chosen level of statistical significance was two-sided p<0.05.
Main outcome of the study was to evaluate short-term visual and anatomical changes after DEX injection in patients with DME poorly or not responding to anti-VEGF loading dose, stratifying them on the basis of BCVA and OCT response at 12 weeks. Secondary outcome was to investigate clinical gain maintenance at long-term follow-up.
Results
Overall, 45 eyes of 45 patients (23 males, 51.1%), mean age 69.7±9 years, were included in the analysis and followed for 11.9±7 months. Six patients were affected by type 1 DM (13.3%) and mean HbA1c was 7.8%±1.9%. Eighteen eyes (40%) were phakic, whereas 27 (60%) had underwent cataract extraction and intraocular lens implant. Thirteen eyes (11.1%) had received grid macular photocoagulation for DME prior to RNB.
Baseline clinical characteristics, stratified on the functional and anatomical outcome at week 12, did not vary among groups (table 1).
Baseline and demographic and clinical characteristics of study eyes, categorised by subsequent BCVA letter gain and central macular thickness at week 12
Ranibizumab loading-phase clinical outcomes
At 12 week, data subanalysis based on visual and anatomical response performed after a mean of 3.5±1.1 RNB injections revealed that 30 eyes (66.7%) had a poor visual response (−4.3±10.7 letters), while 15 eyes (33.3%) disclosed good visual response (+13.9±9.2 letters). Stratifying patients on the bases of OCT changes, 18 eyes (40%) showed no oedema lowering or oedema worsening, while 27 eyes (60%) showed any CMT improvement. Patients with poor visual response were significantly associated with more limited morphological improvement (Fisher’s exact test for categorical variables p=0.04).
In the pooled population, a little increase in BCVA (66.9±23.2 letters, Δ0-12=+2.7, Student’s t-test analysis for paired samples p=0.4) and a slight decrease in CMT (502.1±165.1 µm, Δ0-12=−29.8, Student’s t-test analysis for paired samples p=0.4) were observed after the loading dose.
Poorly responding patients disclosed relative stability in both BCVA (58.5±24.8 letters, Δ0-12W=−4.4, Student’s t-test analysis for paired samples p=0.07) and CMT (529±180.9 µm, Δ0-12W=−6, Student’s t-test analysis paired samples p=0.4). On the contrary, patients with good visual response disclosed a significant functional (77.3±16.4 letters, Δ0-12W=+14, Student’s t-test analysis for paired samples p<0.001) and anatomic improvement (468.9±141.5 µm, Δ0-12W=−59.2, Student’s t-test analysis for paired samples p=0.04) (table 2).
Visual and anatomical outcome of study eyes after each treatment, categorised by BCVA and CMT at week 12
Clinical outcome of DEX injection
All the patients, regardless of functional and anatomical characteristics, underwent DEX implant injection after a mean of 2.34±1.1 weeks from last RNB. At 12 months follow-up, patients underwent a mean of 1.9±1.1 steroid administrations.
Globally, the study population showed a significant improvement both in visual function (70.8±21.2 letters, Δ12W-1M=+3.9, Student’s t-test analysis for paired samples p=0.04) and anatomic profile on OCT (374.1±127.3 µm, Δ12W-1M=−128.1, Student’s t-test analysis paired samples p=0.001) after 1 month from the switch to DEX. Considering the groups according to BCVA, only poor responders disclosed a relevant increase in BCVA (63.8±23 letters, Δ12W-1M=+5.3, Student’s t-test analysis for paired samples p=0.006) and reduction in CMT (376±136.1 µm, Δ12W-1M=−153, Student’s t-test analysis paired samples p=0.002), while patients with good visual response revealed only minor effect on BCVA (79.5±15.2 letters, Δ12W-1M=+2.4 Student’s t-test analysis for paired samples p=0.3), despite a relevant morphological improvement (371.7±119.6 µm, Δ12W-1M=−97.2, Student’s t-test analysis paired samples p=0.03) (table 2) (figure 2). Similarly, considering the groups according to CMT changes, patients featuring no CMT lowering or CMT worsening at week 12 disclosed a greater functional increase (from 58.2±25.8 to 64.6±24.6 letters, Δ12W-1M=+6.4, Student’s t-test analysis for paired samples p=0.001) compared with those having any oedema reduction after RNB (from 74.7±17.8 to 76.4±16.3 letters, Δ12W-1M=+1.7, Student’s t-test analysis for paired samples p=0.4).
Visual and anatomical outcome of study eyes categorised by BCVA letter gain and CMT reduction at week 12. Pooled patients, <5 letter and ≥5 letter improvement, ≤0 and >0 µm CMT lowering categories’ data at different checkpoints. The difference in visual acuity prior and after dexamethasone is greater in <5 letters gainers (p=0.006) and ≤0 µm CMT lowering (p=0.001). BCVA, best-corrected visual acuity; CMT, central macular thickness; RNB: Ranibizumab injection; OZU, Ozurdex injection.
The same trend was maintained up to 12 months from DEX administration. At month 4, poor responders showed significant differences in BCVA (63.6±22.7 letters, Δ12W-4M=+5.1, Student’s t-test analysis for paired samples p=0.05) and in CMT (355.8±116 µm, Δ12W-4M=−173.3, Student’s t-test analysis paired samples p=0.001), while patients with good visual response disclosed little changes in BCVA (77.8±18.6 letters, Δ12W-4M=+0.7, Student’s t-test analysis for paired samples p=0.8) and in CMT (381.5±120.8 µm, Δ12W-4M=−87.4, Student’s t-test analysis for paired samples p=0.06) compared with week 12 (table 2) (figure 2). At month 12, poor responders conserved their significant functional (64.1±23 letters, Δ12W-12M=+5.6, Student’s t-test analysis for paired samples p=0.02) and anatomical gain (363.6±32.5 µm, Δ12W-12M=−165.5, Student’s t-test analysis for paired samples p=0.01), whereas patients with good outcome disclosed insignificant changes in BCVA (79.4±11.9 letters, Δ12W-12M=+2.3, Student’s t-test analysis for paired samples p=0.3) and in CMT (359±14.3 µm, Δ12W-12M=−109.29, Student’s t-test analysis for paired samples p=0.2).
All these differences were consistently found also after exclusion of eyes with baseline BCVA ≥95 ETDRS letters.
Safety analysis
No serious ocular and systemic adverse events were observed in the study eyes. When compared with baseline IOP (14.89±1.66 mm Hg), it increased significantly at months 1 and 4 (16.76±3.04 mm Hg, Δ0-1M=+1.87, Student’s t-test analysis for paired samples p=0.001 and 16.74±3.22 mm Hg, Δ0-1M=+1.84, Student’s t-test analysis for paired samples p=0.004). Seven eyes (18.4%) developed IOP ≥20 mm Hg after DEX injection, and these patients were successfully treated with topic antiglaucoma medications. Cataract progression was observed in nine among the phakic eyes (50%) during the study period, and two patients underwent cataract extraction after the DEX injection.
Discussion
Over the past decade, great strides have been made in the management of diabetic retinopathy and DME. Intravitreal administration of anti-VEGF agents and corticosteroid drugs has revolutionised the approach of clinicians to DME therapy. Anti-VEGF injections have become the first-line medical treatment for DME; several randomised clinical trials have demonstrated that VEGF inhibitors result in a significant gain of visual acuity, improvement in diabetic retinopathy (DR) severity as measured by the ETDRS retinopathy severity scale and reduction of the risk to develop proliferative DR. However, these trials have shown that no more than 65%–70% of patients achieve ≥10 letter improvement in BCVA after 1 or 2 years of treatment.11 12
Our analysis of visual acuity changes in RNB-treated eyes confirms an important heterogeneity in the clinical response to anti-VEGF. More than half of total patients (66.7%) disclosed poor visual response after RNB loading-dose; interestingly, lower letters gain was significantly associated with poorer CMT improvement at week 12.
A subanalysis of Protocol I demonstrated that BCVA response at 12 weeks was significantly associated with ∆BCVA after 3 years, and that poorly responders to RNB loading dose (ie, eyes gaining <5 letters after three intravitreal RNB injections administered monthly) have limited/no chances of further improvement with continuous RNB injections. For eyes with limited early BCVA response after the first 3 monthly intravitreal RNB injections, the switch to adjunctive or alternative anti-DME treatments may be appropriate to maximise final clinical outcomes.6
The evidence that intravitreal administration of DEX (both DEX 0.7 and 0.35 mg implants) is able to improve and maintain vision acuity in chronic DME, showing a good safety profile and tolerable economical burden in poorly responding patients to anti-VEGF is well documented in the literature.13–17 In fact, corticosteroids have shown capacity to inhibit leukostasis, adhesion and transmigration of leukocytes and to downregulate expression of prostaglandins, cytokines and growth factors, especially VEGF in both in vitro and in vivo settings. In addition to their anti-inflammatory function, corticosteroids alter the composition of endothelial basal membrane by changing the local ratio of laminin isoforms, suppressing basement membrane dissolution and strengthening tight junctions to limit permeability and leakage. Finally, long-term steroid usage may have a neuroprotective effect on the retina.18–20
The possibility for clinicians to early predict functional and anatomical outcome after DEX injection would represent a crucial achievement in DME management. Our findings indicate that early visual and anatomical response (change from baseline to week 12 in BCVA and in CMT, respectively) after RNB loading dose are significant predictors of both early term and long-term visual acuity improvement after switching to corticosteroids in patients affected by DME. In detail, eyes with limited early BCVA response (<5 letter gain) showed a three fold letter improvement with respect to eyes with strong early BCVA response (≥5 letter gain) (ΔBCVA12W-1M=+9% vs ΔBCVA12W-1M=+2.9%, respectively); these results were maintained for all the 12-month follow-up. Similarly, eyes with limited early OCT response (≤0 µm improvement) showed a nearly five fold letter improvement with respect to eyes with better OCT response (>0 µm improvement) (ΔBCVA12W-1M=+11% vs ΔBCVA12W-1M=+2.3%, respectively). The sensitivity analysis findings (after exclusion of eyes with baseline BCVA ≥95 ETDRS letters) excluded the possibility that the association between early BCVA response and BCVA after DEX was due to the limited range for BCVA improvement in eyes with relatively good baseline visual acuity, a phenomenon known as ‘ceiling effect’.21 The analysis of the subset of eyes with baseline BCVA <95 ETDRS letters demonstrated similar significant increments in ΔBCVA0-1M across both the limited and the strong early BCVA response categories and the limited and the strong early OCT changes.
No baseline characteristic was related with either poorer or better outcome after the first three injections of RNB. Neither the type of diabetes nor visual acuity and macular thickness did vary meaningfully among the different patients groups.
Limitation of this study is its retrospective nature and lack of detailed information on other factors that may potentially affect anti-VEGF treatment outcome (eg, diabetic retinopathy severity, presence of macular ischaemia on fluorescein angiography and duration of DME prior to therapy).
In conclusion, our study demonstrates that early (week 12) visual and anatomical changes after RNB loading dose are major predictors of clinical outcome up to 12 months after shifting to corticosteroids in DME management. Moreover, the study shows that in cases not optimally responding to three RNB injections, DEX injection is a rescue therapy and can still improve visual acuity in the worst cases. If confirmed in future prospective studies, our evidences could justify the early introduction of adjunctive or alternative anti-DME treatment with intravitreal corticosteroids in order to reach the greatest clinical improvement.
References
Footnotes
Contributors All the authors contributed to the conception or design of the work, the acquisition, analysis and interpretation of data, drafting the work, revising it critically for important intellectual content and gave final approval of the version to be published.
Competing interests MVC, MC, LQ, AR: none; GQ consultant for Alimera Sciences (Alpharetta, Georgia, USA), Allergan (Irvine, California, USA), Heidelberg (Germany), Novartis (Basel, Switzerland), Bayer Shering-Pharma (Berlin, Germany), Zeiss (Dublin, USA). Francesco Bandello consultant for Alcon (Fort Worth, Texas, USA), Alimera Sciences, Allergan, Farmila-Thea (Clermont-Ferrand, France), Bausch & Lomb (Rochester, New York, USA), Genentech (San Francisco, California, USA), Hoffmann-La-Roche (Basel, Switzerland), Novagali Pharma (Évry, France), Novartis Bayer Shering-Pharma, Sanofi-Aventis (Paris, France), Thrombogenics (Heverlee, Belgium), Zeiss, Pfizer (New York, USA), Santen (Osaka, Japan), Sifi (Aci Sant’Antonio, Italy).
Ethics approval San Raffaele Hospital Ethics Committee.
Provenance and peer review Not commissioned; externally peer reviewed.