Macular Edema and Intraocular Pressure Outcomes Following Periocular versus Topical Corticosteroids After...

Postoperative inflammation after phacoemulsification is routinely managed with topical corticosteroids, while periocular administration may provide a sustained anti-inflammatory effect.

Pseudophakic cystoid macular edema (PCME) remains a leading cause of suboptimal visual recovery after cataract surgery. Periocular corticosteroid injections may offer a single-dose alternative to topical regimens, minimizing adherence-related limitations.

Following PRISMA 2020 guidelines (PROSPERO CRD420250632011), PubMed, Embase, and the Cochrane Central Registered of Controlled Trials (CENTRAL) were searched from inception to December 1, 2025, for studies comparing intraoperative periocular injections versus postoperative topical corticosteroids in adults undergoing uncomplicated phacoemulsification. Primary outcomes were PCME incidence and postoperative intraocular pressure (IOP). Best-corrected visual acuity (BCVA) was a secondary outcome. Data were pooled using a random-effects model.

Ten studies involving 57,564 eyes were analyzed. PCME incidence was comparable between groups (5 studies, 56,766 eyes; 1.10% periocular vs 1.40% topical; Risk Ratio (RR), 0.84; 95% CI, 0.55-1.28; P = .42; I2 = 65%). Eight studies (1,829 eyes) showed no clear difference in postoperative IOP (Mean Difference (MD), 0.28 mmHg; 95% CI, -0.20 to 0.77; P = .25; I2 = 65%). BCVA outcome was similar between groups (4 studies, 1,188 eyes; MD, -0.01 logMAR; 95% CI, -0.03-0.01; P = .31; I2 = 0%). Follow-up varied by outcome, extending from approximately 1 month to 120 days.

Current evidence found no clear differences in PCME risk, postoperative IOP, or BCVA between periocular and topical corticosteroids following uncomplicated phacoemulsification. Findings should be interpreted cautiously given the heterogeneity and low certainty evidence. Clinical use should consider drug, formulation, dose, injection route, and the need for longer-term monitoring.