Endophthalmitis in cataract surgery: results of a German survey.

OBJECTIVE: To document perioperative prophylactic treatment and to evaluate the risk factors for endophthalmitis after cataract surgery. DESIGN: Cross-sectional study via anonymous survey. PARTICIPANTS: Four hundred sixty-nine centers in Germany were queried. RESULTS: A total of 311 (67%) questionnaires were received, with each center reporting an average of 900 cataract surgeries per year (total, 340,633 surgeries in 1996). Respondents reported a total of 267 cases of endophthalmitis, which resulted in a mean responder-specific endophthalmitis rate of 0.148% versus a median rate of 0%. Statistical analysis via Poisson regression suggested that sclerocorneal incisions were associated with a reduced incidence of endophthalmitis (odds ratio, 0.35; 95% confidence interval, 0.24-0.51). Antibiotics used intraocularly (odds ratio, 0.65; 95% confidence interval, 0.43-0.98) and the preoperative application of diluted povidone-iodine on the conjunctiva (odds ratio, 0.59; 95% confidence interval, 0.36-0.99) were associated with a reduced risk of postoperative infection. Immune deficiencies (66%), diabetes mellitus (62%), occlusion of the lacrimal system (40%), and skin diseases (33%) were regarded as risk factors for endophthalmitis by the respondents. When cataract surgery is performed solely under inpatient conditions, the use of systemic antibiotics as well as the periocular injection of antibiotics at the end of the operation were associated (although not significantly) with a trend toward reducing the incidence of postoperative infection. Conversely, flushing the lacrimal drainage system, using eye shields, and cutting the eyelashes had no demonstrable effect in preventing endophthalmitis. The use of preoperative topical antibiotics (odds ratio, 2.38; 95% confidence interval, 1.21-4.68) and the performance of more than 20% of the surgeries in an outpatient center (odds ratio, 2.0; 95% confidence interval, 1.24-3.21) were associated with a detrimental effect on the development of endophthalmitis. CONCLUSIONS: Although the appropriate antibiotic agent and dosage are not yet established, the administration of intracameral antibiotics and the application of povidone-iodine on the conjunctiva significantly reduced the relative risk of postoperative endophthalmitis in this survey. Because the study was not individual based but rather on aggregate questionnaire, the results have to be interpreted with care.