Background To evaluate visual acuity gains, percentage of improvement from baseline and anatomical outcomes of a standardised decision-making algorithm integrating modern vitrectomy techniques with individualised intravitreal antibiotic therapy for postoperative endophthalmitis. Methods Patients with suspected postoperative endophthalmitis were stratified by presenting visual acuity and inflammatory biomarkers. Those with visual acuity of >20/100 and mild inflammation received intravitreal moxifloxacin and dexamethasone every 48 hours (P1). Patients with vision <20/150 and/or severe inflammation underwent immediate vitrectomy plus intravitreal antibiotics (P3). Surgeries followed a semistandardised protocol, with a 9-month follow-up. Primary outcomes included best-corrected visual acuity, mean Early Treatment Diabetes Retinopathy Study (ETDRS) letter gain, proportion achieving >15 letter improvement, proportion reaching 20/40 or better and overall percentage improvement from baseline. Results Of 367 patients completing follow-up, patients on P1 improved from 0.56 to 0.33 logMAR (p<0.01); 65% of patients gained ≥15 letters, and 76% achieved 20/40 or better. In P3, patients improved from 1.69 to 0.92 (p<0.05); 73% improved by ≥15 letters and 32% attained 20/40 vision or better. The overall proportion of improvement (i%) from baseline was 32% in P1 versus 59% in P3 (p<0.01). Conclusions This treatment algorithm improves visual outcome, accelerates recovery and may reduce disease burden. At 9 months, clinically meaningful gains in logMAR and ETDRS letters, including percentage of improvement from baseline, were observed. The clinical utility of modern vitrectomy remains under-recognised and warrants broader consideration. Intravitreal moxifloxacin and dexamethasone demonstrate promising efficacy and safety.
No takes yet. Share an insight, caveat, or question.
Vélez-Montoya et al. (2026) studied this question.
Synapse has enriched 5 closely related papers on similar clinical questions. Consider them for comparative context: