Purpose: To compare the results of early versus late switch to the dexamethasone intravitreal implant Ozurdex® in patients with diabetic macular edema who had a poor response to vascular endothelial growth factor inhibitors. Methods: Retrospective and single-center study conducted, in a real setting, on consecutive diabetic macular edema patients who were switch to an intravitreal dexamethasone implant after a poor response to anti-vascular endothelial growth factor treatment. Study sample was divided into two groups: (1) early-switch group, included those eyes who received three anti-vascular endothelial growth factor injections before switch and (2) late-switch group, included those eyes that received six or more anti-vascular endothelial growth factor injections before switch. The primary end-point was the difference in mean change in best-corrected visual acuity and in central subfoveal thickness. Results: A total of 69 (31 early-switch group and 38 late-switch group) eyes were included. In the early-switch group, median (25–75 quartile range) best-corrected visual acuity significantly increase from 0.2 (0.2–0.5) at baseline to 0.4 (0.3 –0.7) at month 24 (p = 0.0043). Whereas, in the late-switch group, best-corrected visual acuity did not increase (p = 0.8602). Central subfoveal thickness was significantly reduced in both early- and late-switch groups, p = 0.0002 and 0.0038, respectively. The proportion of eyes obtaining a central subfoveal thickness reduction ⩾ 10% was significantly greater in the early-switch group than in the late-switch group (71.0% vs 47.4%, respectively, p = 0.0498). Three (9.7%) and 10 (26.3%) eyes have developed ocular hypertension during the study in the early- and late-switch groups, respectively, p = 0.0816. Conclusion: Early switch to Ozurdex in patients who did not adequately respond to anti-vascular endothelial growth factor therapy provided better functional and anatomical outcomes.