Applicability of rapid antibiotic susceptibility testing in the management of bacterial endophthalmitis

To compare the applicability of the Kirby-Bauer disk diffusion method (conventional method) and the rapid antibiotic susceptibility test (RAST) for susceptibility testing of bacteria in the management of bacterial endophthalmitis.

Vitreous samples from 114 consecutive patients with endophthalmitis who were seen between June 2002 and October 2002 were subjected to microbiological processing (smears and cultures for bacteria and fungi). Forty-seven of 114 samples could be tested by RAST. In addition, the bacterial isolates from 47 patients were tested for antibiotic susceptibility by the conventional method, and minimum inhibitory concentrations (MICs) of various antibiotics were determined. Clinical and microbiological data for these 47 patients were analyzed.

Bacteria were isolated by routine culture from 18 (38.3%) of 47 samples; all 18 were tested by the conventional method, and MICs were determined for antibiotic susceptibility. Only 10 (55.6%) of these 18 samples were analyzed by RAST. Low inoculum and the slow-growing nature of the organisms (Streptococcus, 3; Staphylococcus epidermidis, 5) accounted for no growth on RAST plates in eight cases. Of four patients who were treated with repeated intraocular antibiotics, only one received RAST result-guided intraocular injection. Considering MIC testing as the gold standard, the sensitivity and specificity of RAST were either equal to (gentamicin, cefazolin, vancomycin, and chloramphenicol) or lower than (ciprofloxacin) those of the conventional technique.

This comparative study failed to demonstrate a distinct advantage of RAST as a routine procedure in the management of bacterial endophthalmitis because of the need for trained staff to perform RAST, the nearly 50% negativity of culture, and the occasionally unreliable results.