To provide a historical perspective on outcomes of surgery for infants with stage 5 retinopathy of prematurity (ROP) as performed by a single surgeon over a 24-year period. Retrospective observational cohort study. Clinical practice of the Charles Retina Institute. The 601 infants with stage 5 ROP in at least one eye who were referred for surgery between 1977 and 2001. Charts were reviewed for demographic data and postoperative anatomical and visual acuity outcomes in eyes that underwent surgery. The postoperative anatomical status of the retina and visual function were assessed after surgery. Anatomical outcome was categorized as success (macula attached), partial success (macula detached), failure (total retinal detachment), or lost eye (opaque cornea, secluded pupil, or phthisis). Visual function was classified as>20/200, 20/200-5/200, hand movement, light perception, or no light perception. Multivariable logistic regression was used to explore relationships between outcomes and baseline characteristics. Only a minority of eyes had prior cryotherapy (15%) or laser photocoagulation (7%) therapy. The mean follow-up was 44 months for the 956 eyes treated surgically. The 608 eyes with available follow-up data were classified as follows: 28% success, 5% partial success, 55% failure, and 11% lost eye. Visual function of light perception or better was achieved in 74% of the 183 eyes with data on visual acuity. Controlling for other baseline factors, early postpartum age at the time of surgery was a statistically significant predictor of failure/lost eye (OR=2.08, 95% CI 1.09-3.97) and no light perception (OR=5.13, 95% CI 1.45-18.14). Surgery for stage 5 ROP on the fellow eye was also a predictor of failure/lost eye (OR=2.38, 95% CI 1.39-4.08). Surgery resulted in anatomical success for approximately one third of infant eyes with stage 5 ROP, and only a minority of eyes (8 of 183) achieved visual acuity better than 5/200. However, some initially successfully attached retinas redetached. Although this study is limited by follow-up and may represent a group of patients with a more vascularly active disease state due to the low proportion of patients with prior peripheral ablation, this cohort of infants provides results against which future interventions may be compared.