The Pattern of Recurrence in Diabetic Macular Edema Treated by Dexamethasone Implant: The PREDIAMEX Study
Abstract
Purpose:To assess the time to functional and anatomic recurrence of macular edema (ME) after a first intravitreal dexamethasone implant in eyes with diabetic macular edema (DME).Design: A 6-month observational, prospective, uncontrolled, multicenter, national case series.Participants: Thirty-seven patients included between January 2015 and June 2016.Methods: Patients were monitored at baseline and then monthly over 6 months after the first treatment.Main Outcome Measures: Different patterns of recurrence were defined: qualitative and quantitative anatomic recurrences and functional recurrence.Results: Median ME duration before the first dexamethasone implant was 2.04 months. All patients received a dexamethasone implant for the first time, but 73% of patients had not undergone any form of treatment pre- viously. The mean time from baseline to qualitative anatomic, quantitative anatomic, and functional recurrence was 4.22 months (95% confidence interval [CI], 3.80e4.65 months), 4.73 months (95% CI, 4.34e5.12 months), and 4.89 months (95% CI, 4.53e5.26 months), respectively. Almost all patients (7/8) who demonstrated a qualitative anatomic recurrence showed a subsequent quantitative anatomic and functional recurrence days later.Mean improvement in best-corrected visual acuity was 10.1 letters (95% CI, 6.7e13.4 letters) and 7.3 letters (95% CI, 4.1e10.6 letters) at months 2 and 6, respectively. The mean reduction in central subfield macular thickness was 206 mm (95% CI, 157e255 mm) and 146 mm (95% CI, 98e195 mm) at months 2 and 6, respectively.Conclusions: Dexamethasone implant is a functionally and anatomically effective treatment for DME in real- life practice. Qualitative anatomic recurrence seems to be an early sign of quantitative anatomic and functional recurrence.
INTRODUCTION
Diabetic retinopathy (DR) is the most common microvas- cular complication of diabetes.1 It is the leading cause of vision loss and blindness among adults in developed countries.2,3 The prevalence of diabetic macular edema (DME) increases from 0% to 3% in individuals with recent diagnoses of diabetes to 28% to 29% in those who have had diabetes for 20 years.4In 1985, the Early Treatment Diabetic Retinopathy Study (ETDRS) established macular laser as standard care treat- ment, but the limitations of laser treatment along with intense clinical research over the last 10 years have led to laser treatment being surpassed by intravitreal pharmaco- therapy as the first-line treatment for moderate to severe vision loss caused by DME.5Glucocorticoids were the first class of corticosteroids the treatment of macular edema (ME) secondary to retinal vein occlusion,8 for posterior inflammation such as noninfectious posterior uveitis,9 and for DME.10Patterns of recurrence have not been yet analyzed in DME after intravitreal injection of dexamethasone implant, as has been done already for ME after retinal vein occlusion (RVO).11 The objective of this 6-month study was to eval- uate the mean time to anatomic and functional recurrence of ME after the first dexamethasone implant (Ozurdex, Allergan, Irvine, CA) injection and also to estimate its ef- ficacy and safety. All parameters were reported prospec- tively and on a monthly basis throughout the duration of the study. shown by randomized clinical trials to be beneficial for DME.6,7 However, because of numerous adverse effects, they have been replaced by sustained-release steroid devices made specifically for intravitreal injection, in particular the dexamethasone implant. The implant has been approved for A 6-month observational, prospective, uncontrolled, multicenter national case series was conducted in France from January 2015 through June 2016. All patients received clear, detailed prior in- formation on the treatment and on the expected risks and benefits.The study was performed with informed consent in compliance with the Declaration of Helsinki and following all the guidelines for experimental investigations on human subjects.
We obtained authorization from a national Protection to Persons and Property Committee (institutional review board number, IRB00009118). We certify that this research complies with all the applicable institu- tional and governmental regulations concerning the ethical use of human volunteers.All patients in the study were at least 18 years of age and had decreased visual acuity resulting from central DME involvement defined as retinal thickening involving the 1-mm (according to OCT) central subfield macular thickness (CSMT) with subretinal or intraretinal fluid, or both, were eligible. There was no limit regarding the time since the first diagnosis of DME. Patients were not necessarily naïve to treatment for DME, but patients having received previous intravitreal injections of dexamethasone implants in either eye were not accepted. All the previous treatments administered to each patient were identified clearly. For the patients with a history of antievascular endothelial growth factor treatment, a mandatory delay of a minimum of 3 months between the last antievascular endothelial growth factor injection and the first dexamethasone implant was imposed. We selected 1 eye per patient as the study eye. Eligible patients had to have best-corrected visual acuity (BCVA) of more than 34 ETDRS letters in the study eye. Central subfield macular thickness had to be 275 mm or more in the study eye (thickness of a circular area of 1 mm, concentric to the foveal center). Patients with glaucoma requiring more than 1 topical drug treatment to control intraocular pressure (IOP) were not included in the study.The main exclusion criteria included the presence of active oracute intraocular inflammation or infection, clinically significant epiretinal membrane, or vitreomacular traction.
Patients also were excluded if they had hypertensive retinopathy in either eye or any uncontrolled systemic disease. Patients treated with oral cortico- steroids, patients with any of the contraindications for dexameth- asone implant set out in the June 2013 marketing approval, patients with uncontrolled diabetes with glycosylated hemoglobin of more than 10%, and patients with a history of steroid-induced IOP increase of 5 mmHg or more also were excluded.Each patient received a single intravitreal injection of dexa- methasone implant at day 0. Only topical anesthetic eye drops (oxybuprocaine hydrochloride 1.6 mg/0.4 ml) were used in the study eye. The intravitreal injections of the dexamethasone implant were performed according to standard clinical practices published by the French Health Authority in January 2011. Given that the main objective of this study was to assess the pattern of recurrence, none of the patients included were injected with a second dexa-methasone implant before month 6, unless the rescue criteria were met, to prevent any macular alteration (vision loss >10 letters, increase of CSMT >100 mm, or both).Each patient underwent a standardized examination at the initialvisit and at each monthly follow-up visit, with measurement of BCVA in ETDRS letters, air-puff or applanation tonometer to measure IOP, lens status determination, fundus ophthalmoscopy, and spectral-domain (SD) OCT (Cirrus HD-OCT model 500 Zeiss; Carl Zeiss Meditec, Inc., Dublin, CA) to measure CSMT. Patients who received more than 1 injection received the same follow-up, with a monthly examination and recording of the same clinical and OCT data.The primary efficacy outcome was evaluation of the mean time to anatomic recurrence (increase in macular edema) or functional recurrence (decrease in BCVA) after treatment.
We used the same definitions as previously published,11 and therefore considered qualitative or strict anatomic recurrence when SD-OCT imaging showed new intraretinal cysts, little subretinal fluid, or both. We also defined quantitative or obvious anatomic recurrence as an increase in CSMT of 50 mm or more identified using SD-OCT imaging. Functional recurrence was defined as a loss of BCVA of 1 line or more in the study eye after treatment. We also assessed the mean change in VA from baseline BCVA at each visit and at 6 months, the mean change from baseline CSMT measured using SD-OCT, and the proportion of eyes with a minimum 3-line improvement from baseline BCVA.We also defined responder and nonresponder status. Responders were classified as follows: functional responder, BCVA improvement of a minimum of 1 ETDRS lines during follow-up; anatomic responder, central foveal thickness improve- ment of 20%; and complete responder, both anatomic and func- tional responder criteria apply.12e16Categorical variables were described using absolute and relative frequencies, and quantitative variables were described using me- dian, minimum, maximum, mean, and standard deviation (SD). Linear mixed-effects models were used to study BCVA and CSMT over time. This method allowed us to take into account the within- subject correlation of the repeated observations over time and the inclusion of patients with a varying number of measurements. Best-corrected visual acuity and CSMT were expressed using the absolute measured value or as change from the baseline value. The models gave estimates of the mean BCVA, mean changes in BCVA, mean CSMT, and mean changes in CSMT for each time point, with a 95% confidence interval (CI). The Kaplan-Meier product limit method was used to study the occurrence and delay of anatomic and functional recurrences. Patients contributed to the risk set within the 6 months after the first injection. Patients who did not demonstrate recurrence were considered right censored at the date of their last visit within the period. The R software pro- gram (R Development Core Team, Vienna, Austria) was used to perform all analyses, and for each test, the 0.05 significance level was used.
Results
This prospective study was conducted in 3 centers located in mainland France. Thirty-seven eyes of 37 patients were included between January 2015 and June 2016, with a minimum follow-up period of 6 months for all patients.The population characteristics are shown in Table 1. The mean age was 64.4 years (range, 29.3e84.7 years). The population was made up of more men than women (68% men). There was a small difference in terms of laterality (62% of left eyes). In terms of type of diabetes, 11% of patients had a type 1 diabetes and 90% had a type 2 diabetes. Concerning diabetes treatment, 54% of patients received a combination of oral antidiabetic drugs and insulin, whereas 11% and 35% received only insulin or oral antidiabetic drugs, respectively. In terms of the severity of DR, 14% of patients initially demonstrated mild nonproliferative DR, 46% had moderate nonproliferative DR, 30% had severe nonproliferative DR, and 5% had proliferative DR. Only 2 patients did not have any type of DR. We categorized patients with nonadvanced DR (absence of DR, mild DR, and moderate DR) and patients with advanced RD (severe and proliferative Table 2.
The mean time from baseline to qualitative anatomic recurrence, quantitative anatomic recurrence, and functional recurrence was 4.22 months (95% CI, 3.80e4.65), 4.73 months(95% CI, 4.34e5.12), and 4.89 months (95% CI, 4.53e5.26;Fig 1), respectively. The first pattern of recurrence for 8 patients was the qualitative anatomic one. Concerning these 8 patients, 1 patient did not demonstrate any quantitative anatomic or functional recurrence after 8 months of follow-up. The cysts for this patient remained perfectly stable during the follow-up period, with no impairment of visual acuity. Concerning the remaining 7 patients, secondary quantitative anatomic and functional recurrence developed. Qualitative anatomic recurrence thus occurred 10.6 days (SD, 20.5 days) before quantitative anatomic recurrence on OCT imaging and 17.5 days (SD, 23.1 days) before a loss of vision of at least 1 line (functional recurrence). The mean delay between functional recurrence and quantitative anatomic recurrence was 5.4 days (SD, 12.2 days), with the quantitative anatomic recurrence appearing first. Concerning retreatment, 69% of patients wereinjected before month 6 (n ¼ 18) because of rescue criteria.Functional Efficacy DR). Concerning lens status, 81% of patients were pseudophakic. The mean glycated hemoglobin the day of the first dexamethasone implant injection was 7.6% (median, 7.4%).Concerning patient history of panretinal photocoagulation (PRP), 39% of patients had complete PRP, 23% had incomplete PRP, and 38% had no history of PRP. Present patients had no history of focal or grid laser. None of the patients underwent PRP during follow-up, because this was already complete before the start of follow-up.
Regarding DME treatment, 73% of patients were naïve to any treatment and 27% had received at least 1 antievascular endothelial growth factor intravitreal injection before the first dexamethasone implant injection.The median time between diagnosis and first dexamethasone intravitreal implant injection was 2.04 months (mean, 29.16 months; range, 0e405.9 months). Each patient had a minimum follow-up of 6 months. During this period, the mean number of dexamethasone implant injections was 1.5 and ranged from 1 to 2.Anatomic and Functional RecurrenceThe proportion of eyes with qualitative anatomic, quantitative anatomic, or functional recurrence at each study visit is shown in achieving at least a 3-line improvement from baseline BCVA was 27% during follow-up, with a time lapse of 48 days (range, 7e157 days) after the first injection. During the 6-month follow- up, 84% of patients were functional responders with a mean time lapse of 50.7 days (range, 7.0e188.0 days) after intravitreal in- jection. Sixty-eight percent of patients reached a BCVA of at least 70 ETDRS letters or more with a mean time lapse of 44.3 days (range, 7.0e176.0 days) during follow-up. Six patients (16% of patients) were functional nonresponders, but of these, 2 patients had initial visual acuity of 70 ETDRS letters or more and final visual acuity of 74 ETDRS letters (possible ceiling effect).Anatomic EfficacyInitial mean SD CSMT was 479 145 mm. At month 2, mean CSMT was 281 mm (95% CI, 239e322 mm), with a mean reduc- tion of 206 mm (95% CI, 157e255 mm) compared with baseline.
At month 6, mean CSMT was 327 mm (95% CI, 286e368 mm), with a mean reduction of 146 mm (95% CI, 98e195 mm; Fig 3). During the 6-month follow-up, 87% of patients were anatomic responders with a mean time to occurrence of 41.6 days (range, 7.0e98.0 days) after intravitreal injection, with 70% of patients reaching a CSMT of at least 300 mm or less with a mean time to occurrence of 42.5 days.SafetyThe results shown in Figure 4 focus on the issues relating to IOP. The measurements obtained during follow-up reported IOP of 25 mmHg or more for 14% of patients (5 patients) and 35 mmHg or more for 3% of patients (2 patients), and 8% of patients showing an increase of more than 10 mmHg. All the patients demonstrating intraocular hypertension were managed with topical treatment only (25% of the patients needed monotherapy, 60% needed dual therapy, and 15% required triple therapy). None of the patients needed filtering surgery. In all cases, IOP returned to normal under ocular hypotensive treatment before the second injection was administered, if needed. Concerning other ocular adverse events, 2 patients experienced subconjunctival hemor- rhages. There were no cases of intravitreal hemorrhage, endoph- thalmitis, or anterior chamber migration of the implant. Cataract progression and changes in DR stage were not examined in this 6-month study.
Discussion
anatomic recurrence appears first, with qualitative anatomic recurrence occurring on average at 4.11 months (SD, 0.86 month), functional recurrence on average at 4.31 months (SD,1.33 months), and quantitative anatomic recurrence on average at 4.40 months (SD, 1.14 months). In this study, qualitative anatomic recurrence occurred 14.40 days (SD,26.22 days) before quantitative anatomic recurrence and 14.4 days (SD, 42.18 days) before functional recurrence. The mean time lapse between quantitative anatomic and functional recurrence was 4.09 days (SD, 30.58 days), with the func- tional recurrence appearing first.In our study, the criteria used for retreatment before the sixth month were an increase in CSMT of more than 100 mm as compared with the lowest value recorded or a visual loss The present prospective observational study is to our knowledge the first study in the literature to assess patterns of recurrence after an intravitreal injection of dexametha- sone implant for DME. This study also assessed the functional and anatomic efficacy of the dexamethasone implant in treating DME in routine clinical practice. Although this was not a retrospective study, this study can be labelled a real-world study because it was an open-treatment obser- vational study and patients were managed according to our routine practice.
In this study, fellow-eye treatment was not reported. This could be a bias because some cases series have reported that intravitreal treatment may affect the fellow eye.17e21We used a classification of 3 different patterns of recur- rence previously published by our team in RVO.11 Thus, we distinguished between functional recurrence (loss of vision of at least 1 line), qualitative anatomic recurrence (new intraretinal cysts, little subretinal fluid on SD-OCT, or both), and quantitative anatomic recurrence (increase in CSMT of 50 mm or more, identified using SD-OCT imag- ing). Qualitative anatomic recurrence appears first, followed by quantitative anatomic and then functional recurrences. Indeed, the mean times from baseline to qualitative anatomic recurrence, quantitative anatomic recurrence, andfunctional recurrence were 4.22 months (95% CI, 3.80e4.65 months), 4.73 months (95% CI, 4.34e5.12 months), and 4.89 months (95% CI, 4.53e5.26 months; Fig 1), respectively. Qualitative anatomic recurrence occurred 10.6 days (SD, 20.5 days) before quantitative anatomic recurrence and 17.5 days (SD, 23.1 days) before functional recurrence. The mean time lapse between functional recurrence and quantitative anatomic recurrence was 5.4 days (SD, 12.2 days), with the quantitative anatomic recurrence appearing first.No identical studies were found in the literature for DME. Nevertheless in RVO, we demonstrated that qualitative of more than 10 ETRDS letters.
If these criteria were not met, the retreatment criteria at the sixth month were a visual acuity loss of more than 5 ETDRS letters, CSMT of 275 mm or more with manifest edema, or residual edema considered by the practitioner to be clinically significant. We found a need for retreatment at 4.7 months, as in the study by Querques et al,22 where retreatment also was judged necessary after a mean of 4.7 months from the first implant. In real-world studies, the mean time toFigure 1. Graph showing the mean time from baseline to qualitative anatomic recurrence, quantitative anatomic recurrence, and functional recurrence. Figure 2. Graph showing functional efficacy: changes in mean best- corrected visual acuity (BCVA) in Early Treatment Diabetic Retinop- athy Study letters after intravitreal injection of dexamethasone intravitreal implant during the 6 months of follow-up. Red arrows represent the mean day of reinjection. Figure 4. Graph showing tolerance: changes in mean intraocular pressure (IOP) after intravitreal injection of dexamethasone intravitreal implant during the 6 months of follow-up. Red arrows represent the mean day of reinjection. CI ¼ confidence interval. retreatment after the first dexamethasone implant injection was 5.7 months (range, 4.8e6.7 months) in the Reldex study23 and 5.4 0.2 months in the Mozart study.24 In pivotal studies such as the Mead,25 Callanan et al,26 and BEVORDEX,27 studies, recurrence was not analyzed.Figure 3. Graph showing anatomic efficacy: changes in mean central subfield macular thickness (CSMT; in micrometers) after intravitreal in- jection of dexamethasone intravitreal implant during the 6 months of follow-up. Red arrows represent the mean day of reinjection. CI confi- dence interval.
It is very important to note that almost all the patients (7/8) in our study with qualitative anatomic recurrence experienced functional or quantitative anatomic recurrence a few days or weeks later. These results emphasize the importance of SD-OCT imaging for diagnosing early recurrence after intravitreal treatment. The benefit of earlier retreatment in RVO already has been shown in a multicenter study conducted by Coscas et al,28 who emphasized the need to retreat patients as soon as possible after observing a recurrence of ME, as well as the clinical benefits of doing so. In addition, Pelosini et al29 demonstrated a linear relationship between tissue integrity and visual acuity. Earlier retreatment, before the quantitative anatomic recurrence, may prevent retinal damage and preserve visual acuity. Moreover, other studies have shown that a delay in the treatment of ME decreases the probability of an improvement in visual acuity in comparison with the prompt initiation of treatment.30,31The present results suggest the need to conduct a new study to ascertain whether earlier reinjection, as soon as cysts appear, could be both anatomically and functionally benefi- cial for patients. We therefore recommend retreatment with dexamethasone implants within the 2-week period after the qualitative anatomic recurrence of DME, because the mean time between qualitative anatomic recurrence and functional recurrence is 17.5 days.
One limitation of our study is the small number of eyes included. However, this is mitigated by the prospective follow-up at monthly intervals to provide more precise data on the duration of action and recurrence.Regarding safety and hypertension, a transient increase in IOP of 25 mmHg or more was observed in 13.5% of pa- tients, and IOP of 35 mmHg or more was observed in 2.7% of patients (2 patients); 8.1% of patients demonstrated an increase of more than 10 mmHg. All cases were well controlled with topical antiglaucoma medication. The safety profile was excellent in this real-world study. In comparison with the MEAD study, the proportion of eyes with hyper- tension (IOP 25 mmHg) was 32% and 0.6% of eyes needed filtering surgery.25The SAFety Of implant DEXamethasone implant (SAFODEX) study reported the safety of the dexametha- sone implant in a series of 1000 injections. An IOP of 25 mmHg or more was found in 20% of patients and an IOP of 35 mmHg or more was found in 6% of patients, and 27% of patients showed an increase of more than 10 mmHg during the 3-year follow-up.32 However, safety was significantly better than in the Fluocinolone Acetonide in diabetic Macula Edema (FAME) study of the fluocinolone acetonide implant, in which 18% of eyes demonstrated an IOP of 30 mmHg or more and 4.8% of patients required glaucoma surgery during the 3-year follow-up.
In conclusion, the dexamethasone implant is a functionally and anatomically effective treatment for DME in real-world practice. Moreover, it is well tolerated, producing generally transient, moderate, and easily managed increases in IOP. Qualitative anatomic recurrence seems to be an early sign of quantitative anatomic and functional recurrence. Therefore, further studies should assess if earlier retreatment at the stage of qualitative anatomic recurrence could protect visual func- tion better. Moreover, additional longer studies would be of interest to assess the time to recurrence after a second, third, or further treatment with dexamethasone implant.