PurposeTo study the morphological alterations in the individual retinal layers of patients with diabetic macular edema (DME) treated with dexamethasone intravitreal implant.MethodsRetrospective, observational study of patients with center‐involving DME treated with a single dexamethasone implant as primary or secondary treatment, and with at least two follow‐up visits. Central thickness of each retinal layer was obtained with optical coherence tomography automated segmentation (Heidelberg Spectralis Engineering, Germany), with manual correction as needed.ResultsTwenty one eyes of 18 type 2 diabetic patients were included (median HbA1c 7.1%). The median time of follow‐up after treatment was 8 weeks (visit 1) and 4 months (visit 2). At visit 1, a significant reduction in central macular thickness (CMT) (480 ± 24 to 325 ± 134), retinal nerve fibre layer (RNFL) (21 ± 1 to 18 ± 2), ganglion cell layer (GCL) (30 ± 2 to 22 ± 1), inner nuclear layer (94 ± 15 to 46 ± 6), outer plexiform layer (OPL) (35 ± 3 to 28 ± 1), outer nuclear layer (ONL) (176 ± 14 to 117 ± 18) and retinal pigment epithelium (RPE) (29 ± 7 to 15 ± 5) was noted (p < 0.05 for all layers). At visit 2 there was a significant increase in CMT, (424 ± 270), although still lower than the baseline; there was also a significant increase in RNFL (25 ± 3), GCL (30 ± 3), OPL (40 ± 3), ONL (146 ± 20) and RPE (24 ± 5) (p < 0.05). There was no difference in inner plexiform layer (IPL) or outer retinal layers thickness over time. Seven eyes had neurosensorial detachment (NSD) at baseline, versus 0 at visit 1 and 2 at visit 2.ConclusionsThe reduction in CMT noted in patients with diabetic macular edema treated with dexamethasone intravitreal implant is mostly due to resolution of NSD (which accounts to the automated RPE thickness) and the layers above external limiting membrane, except for IPL. This effect is highest at the first follow‐up visit.