Diabetic macular oedema (DME) is the main cause of visual impairment (or visual acuity) in patients with diabetic retinopathy, as it leads to progressive thickening of the retina, which in the long term leads to progressive death of the photoreceptor cells. It is therefore important to continue to treat macular oedema that has been progressing for several months or even years (resistant DME).

The management of DME necessarily involves controlling diabetes (improving glycated haemoglobin levels) and blood pressure, but this is often not enough.

Thus, when DME is significant and leads to a decrease in visual acuity, treatments are administered directly into the eye (intravitreal injections). For some years now, corticosteroids have been injected into the vitreous body (the gel that fills the eyeball) through the white of the eye for their anti-inflammatory properties. Indeed, these drugs improve the permeability of the retinal vessels and thus reduce oedema. These intravitreal implants are most often used in patients who have already undergone cataract surgery (pseudophakic) because corticosteroids also tend to aggravate a cataract. Currently, there are two implants containing corticosteroids that can be injected: the dexamethasone implant and the fluocinolone acetonide implant. These two implants have different properties, particularly with regard to their duration of action.

Today, the overall management at 3 years and the quality of life associated with the treatments deserve to be evaluated.

This study is the first multicenter controlled trial comparing the two reference corticosteroid treatments in terms of overall cost of treatment and follow-up and patient quality of life, while considering their efficacy and side effects. This evaluation will make it possible to precisely define the respective place of each implant in the management of resistant DME.

Inclusion criteria*:

• Patient who has given free, written and informed consent;
• Major patient ;
• Patient with treated DME greater than 300 microns of central foveolar thickness still present after at least 2 years of treatment and responsible for a decrease in visual activity;
• Best corrected Visual acuity (BCVA) ≤ 80 letters ETDRS
• Patient who has received at least one anatomically and functionally effective dexamethasone (DXM) injection more than 5 months ago
• Patient who has received one anti-VEGF injection more than 3 months ago
• Pseudophakic patient with surgery older than 6 months.
• Patient with uni or bilateral diabetic macular oedema (in the case of bilateral diabetic macular oedema, the most affected eye will be treated).

Exclusion criteria*:

• Patient not covered by national health insurance;
• Patient under a measure of legal protection;
• Pregnant, parturient or breast-feeding woman;
• Patient of full age who is unable to give consent;
• Patient who has already participated in the study
• Patient for whom the follow-up imposed by the protocol is not feasible (relocation)
• Patients with a known hypersensitivity to the active substance or to one of the excipients of Ozurdex®, Iluvien® ; Patients with uveitis or a severe form of asthma
• Patients with pre-existing uveitis or glaucoma or active or suspected ocular or periocular infection, including most viral diseases of the cornea and conjunctiva, including active epithelial herpes simplex keratitis (dendritic keratitis), vaccinia, varicella, mycobacterial infections, and mycoses
• Glycated hemoglobin > 12%.

In the study eye:

• Patient who received an injection of Iluvien (fluocinolone acetonide) less than 24 months ago-
• Patient with untreated severe proliferative or non-proliferative diabetic retinopathy;
• Patient with pan-retinal photocoagulation or focal treatment less than 3 months old;
• Patient with capillary macro aneurysms accessible to focal laser
• Patient with ocular hypertonia > 21 mmHg despite a treatment of more than 2 molecules;
• Aphakic patients or patients with a history of capsule rupture and iridal or transcleral fixation implants
• Phakic patient