EXOGENOUS ENDOPHTHALMITIS INITIALLY TREATED WITHOUT SYSTEMIC ANTIBIOTICS

Purpose: In 1987, the authors reported the successful eradication of i nfection in 16 patients with culture-proven acute exogenous bacterial endophthalmitis using intravitreal but no systemic antibiotics. They r etrospectively reviewed additional consecutive cases since then to det ermine if the initial omission of systemic antibiotics remained reason able. Methods: Twenty patients had culture-proven endophthalmitis. Fou r patients initially received systemic antibiotics for orbital celluli tis (2 patients), prevention of a possible scleral buckle infection (1 patient), and ascending cholangitis (1 patient). The remaining 16 pat ients were treated initially with intravitreal antibiotics only. Findi ngs: Three of these additional 16 patients ultimately required systemi c antibiotics for orbital cellulitis (1 patient), infectious scleritis (1 patient), and prevention of central nervous system infection with Neisseria meningitidis (1 patient). Only in one patient who had a negl ected endophthalmitis and in whom an orbital cellulitis ultimately dev eloped were we unable to clear the intraocular infection. In the overa ll series of 32 patients, cultures yielded staphylococcal species in 1 6 eyes, gram-positive bacilli in 3, streptococcal infection in 5, gram -negative cocci in 1, and gram-negative bacilli in 7. Half of the 14 s pecimens (1 aqueous and 13 vitreal) collected at the time of 16 reinje ctions in 13 eyes yielded organisms. Half (16/32) of the eyes attained visual acuity of 20/40 or better; 87.5% (28/32) attained visual acuit y of 20/400 or better. Conclusions: Therapy with intravitreal antibiot ics without systemic antibiotics is reasonable, unless the infection h as extended (or is at risk to extend) beyond the globe. Such evidence includes an elevated temperature or leukocyte count, corneal ring absc ess, proptosis, loss of extraocular movements, scleral abscesses or in fectious scleritis, and, perhaps, the presence of a scleral buckle.