Diagnostic yield of vitrectomy in eyes with suspected posterior segment infection or malignancy

Purpose: To determine the yield of diagnostic pars plana vitrectomy in eyes with suspected posterior segment inflammation or malignancy when clinical examination and systemic laboratory testing did not yield a specific diagnosis. Design: Noncomparative interventional case series Participants: Eighty-seven consecutive patients (90 eyes) who underwent diagnostic pars plana vitrectomy from 1989 through 1999. Interventions: Vitreous samples were analyzed in a directed manner based on the preoperative clinical examination and systemic laboratory testing. Main Outcome Measures: Diagnosis from each test performed on the vitreous samples. Results: Diagnostic vitrectomy was performed alone in 6 eyes (7%) and as part of a therapeutic procedure in the remaining 84 eyes. The diagnostic tests performed most frequently included cytopathology (83%), microbiologic culture and sensitivity (43%), polymerase chain reaction (PCR) (36%), and intraocular antibody levels for T. canis (14%). Of these, intraocular antibody testing and PCR had the highest positive yield, 46% and 39%, respectively. Overall, directed vitreous analysis identified a specific cause in 35 eyes (39%). Of the 65 cases in which an underlying infection was suspected preoperatively, the procedure yielded a specific diagnosis in 27 (42%). When intraocular malignancy was considered preoperatively (71 eyes), a diagnosis of intraocular lymphoma was obtained in seven (10%). This difference between these diagnostic yields was significant (P = 0.02, Fisher's exact test). Conclusions: Diagnostic vitrectomy with directed vitreous fluid analysis yields a specific cause and guides subsequent therapy in a high percentage of cases. This procedure is a valuable adjunct in cases that cannot be diagnosed by less invasive methods.