Sensitivity and predictability of vitreous cytology, biopsy, and membrane filter culture in endophthalmitis.

PURPOSE

Forty-seven consecutive patients with endophthalmitis were prospectively studied to: 1) compare the undiluted vitreous biopsy culture and membrane filter culture of vitrectomy cassette fluid; 2) determine the sensitivity and specificity of initial smear examinations in relation to final culture; and 3) characterize vitreous cytology in culture positive and negative endophthalmitis.

METHODS

Examinations included smears (Gram stains, Diff Quik [Bacto Laboratories Pvt. Ltd., Liverpool, Australia] and unstained) and aerobic/anaerobic culture of undiluted vitreous biopsy specimens and diluted vitreous collected in the cassette. The cassette fluid was passed through 5-microns polycarbonate filter for cytology and a 0.22-microns polyvinylidene difluoride millipore filter for culture.

RESULTS

By either method culture results were positive in 27 (57.4%) patients. Culture positive of the vitreous biopsy alone was 44.6% and cassette fluid alone was 49% (P > 0.05). In the initial smear examination the sensitivity and specificity of the Gram stain were 66.6% and 84.2%, respectively. comparatively, Diff Quik provided lower sensitivity and specificity at 40.7% and 80.9% respectively. Examination of wet film of vitreous samples did not provide any useful information. Although polymorphs were seen in large numbers in all cultures that yielded positive results on examination, they also were seen in 80% of the sterile samples. There was no significant difference in the quantity of macrophages observed in infected and noninfected samples.

CONCLUSIONS

Initial smear examination and cytology have limited roles in the diagnosis of infectious endophthalmitis. Obtaining cultures of both an undiluted vitreous biopsy sample and the vitrectomy cassette fluid has a significant advantage compared with culture of only one sample.