Comment on “Intravitreal dexamethasone implants for diabetic macular edema”

Dear Editor,

We would like to address several issues with the study of Pareja-Rios et al^[1].

The study was retrospectively conducted and 3mo of treatment with laser or anti-vascular endothelial growth factor (VEGF) agents mean a period of time too short to label a patient as nonresponder to these therapies.

There was a selection bias attributable to inclusion in the study and pooled analysis of the eyes that were naive to treatment for diabetic macular edema (DME) and eyes that had received prior treatments (anti-VEGF therapy, laser or a combination thereof). Likewise, the results were globally analyzed as a unique group consisting of pseudophakic, naive, phakic, and vitrectomized patients. Taking together, these issues make interpretation of the outcomes challenging.

We hypothesized that a whole panoply of proinflammatory and proangiogenic cytokines, chemokines, and growth factors may be associated with the multifactorial pathophysiology of the DME. They are maximally expressed in the ischemic lesions of the long-standing DME and exacerbate the deterioration primarily caused by VEGF in the initially damaged macular ganglion cell complex^[2].

There were no details regarding the DME defined as retinal thickening or hard exudates at or within 1 disc diameter of the macula center and which is most commonly classified into either being clinically significant or not. Nothing was stated concerning the optical coherence tomography patterns of the DME (sponge-like swelling/cystoid macular edema/subfoveal neuroretinal detachment/mixed type) and the location of the cystoid type (ganglion cell layer/inner/outer nuclear layers) at enrollment and at the end of the study.

Initially, a comparison had to be carried out between the 3 groups of patients to establish whether or not they are comparable. Accordingly, this comparison should have been conducted only if there were no significant baseline differences between all variables of these 3 groups^[3]. Of note, there were obvious baseline differences between the pseudophakic, naive, and phakic groups concerning the following findings: the mean age of patients, the mean best-corrected visual acuity (BCVA), and the mean central macular thickness (CMT). Importantly, the baseline features of the patients in the treatment-naive group completely differ from those of the patients of the other two groups, namely, they have the lowest age, the best BCVA, and the greater CMT. In addition, they are the only one who have had recent cardiovascular events.

The following relevant data are missing from the study: the age of diabetes and the duration of the DME before entry into the study after diabetes onset; the qualitative status of the retinal inner layers, the outer nuclear layer, the external limiting membrane band, the ellipsoid zone, and the interdigitation zone at enrollment and at the end of the follow-up period; the qualitative status of the retinal pigment epithelial band-Bruch membrane complex and grading of the retinal pigment epithelium (RPE) changes (pigment migration within the neurosensory retina, RPE porosity, microrips or blowouts in the RPE, focal RPE atrophy, RPE thickening, presence of reticular pseudodrusen) at the enrollment and in month 12; the prevalence of the vitreoretinal interface abnormalities, the subfoveal thickness, and the quantification of the hyperreflective retinal foci at presentation and at the end of the follow-up period; and the proportion of the eyes with sustained retina dryness at the end of the study.

In the assessment of the 12-month results of this study we considered the current assertion according to which evaluation of outcomes should be guided by the anatomic measure data with visual changes as a secondary guide^[4]. Accordingly, the outcomes of this series are unsatisfactory. Specifically, despite a mean gain of 4.2 letters in BCVA, the CMT decreased to approximately 420 µm, a value much more than the cutoff of the upper level of the normal CMT. The persistence of this high value of the CMT after treatment highlighted unresolved macular edema owing to insufficient macular deturgescence and indicated that the disease process was still active and progressive requiring further treatment with anti-angiogenic agents^[5].

Nothing was stated regarding the diabetic choroidopathy, which consists in intrachoroidal vascular abnormalities and which may directly induce choroidal ischemia, leading to RPE dysfunction. The progressive thickening of the choroid layer caused by increasing the severity of the diabetic retinopathy (DR; from no DR to proliferative DR) and development of DME (being thickest in eyes with serous neuroretinal detachment type of DME) denotes progression of the diabetic choroidopathy^[6].

The currently available recommendations^[7] that the duration of ≥3-line improvement after a dexamethasone (DEX) implant is typically 2 to 3mo, and that reinjections generally will be performed after 4 to 5mo have not been taken into account by the authors. If these assertions had been considered, the design and outcomes of the present study would have been completely different.

We believe that the specific anti-VEGF drugs [e.g. bevacizumab (Avastin, Genentech, Inc., South San Francisco, CA, USA)/ranibizumab (Lucentis, Genentech, Inc.)/aflibercept (Eylea, Regeneron Pharmaceuticals, Inc. Tarrytown, NY, USA)] represent the front-line therapy for the treatment of DME but the VEGF inhibition alone may not be sufficient to decrease the inflammatory response. Therefore, the addition of a non-specific anti-VEGF substance (e.g. DEX implant), which inhibits the up-regulation of VEGF and suppresses the expression of the whole inflammatory factors, is mandatory^[5].

Altogether, regardless of the intravitreal pharmacotherapy chosen, namely, specific or nonspecific anti-VEGF agents, the efficacy of the treatment depends primarily on the promptness of the therapy after DME diagnosis^[2]. Both groups of anti-VEGF substances provide similar rates of vision improvement, but with superior anatomic outcomes and fewer injections in the corticosteroid implant-treated eyes. However, more patients receiving the corticosteroid implant lose vision mainly due to cataract^[2],^[5].