Usefulness of aqueous humor analysis for the diagnosis of posterior uveitis.

PURPOSE: To assess the clinical usefulness of aqueous fluid analysis for the diagnosis and treatment of patients suspected of having infectious posterior uveitis (PU). DESIGN: Case-control study. PARTICIPANTS: From 2002 through 2005, 152 eyes from 152 patients with active PU (16 of whom were immunosuppressed) underwent diagnostic aqueous testing. As controls, 20 patients with Fuchs' heterochromic uveitis and 20 patients with age-related cataract were included. METHODS: Aqueous samples were examined by real-time polymerase chain reaction (PCR) and by pathogen-specific analysis of intraocular antibody production (Goldmann-Witmer coefficient [GWC]) for herpes simplex virus (HSV), varicella zoster virus (VZV), cytomegalovirus (CMV), and the parasite Toxoplasma gondii. MAIN OUTCOME MEASURES: Results of aqueous analysis and any adverse effects of aqueous sampling. Correlations between the results of aqueous testing and clinical characteristics as well as the treatment of patients. RESULTS: Of 152 patients, 44 (29%) had positive results for at least one diagnostic assay (37/136 [28%] immunocompetent and 7/16 [44%] immunocompromised patients). None of the controls had positive results using PCR or GWC. A positive result was obtained predominantly in patients with focal chorioretinitis (37/87 [40%]) and in extensive retinitis (7/9 [78%]), whereas in multifocal chorioretinitis, neuroretinitis, and retinal vasculitis only a few samples demonstrated positive results (2/19, 1/29, and 0/10, respectively). Of 37 immunocompetent PU patients with positive results, 28 (76%) cases were caused by T. gondii, whereas viral infections were most common in immunocompromised patients (5/7 [71%]). In immunocompetent and toxoplasmosis PU patients, GWC was the most informative assay (34/37 [92%] and 28/30 [93%], respectively), in contrast to immunosuppressed patients (PCR positive in 5/7 and GWC positive in 4/7). Independent of the immune status of patients, positive PCR results were observed more frequently in viral infections than in toxoplasmosis (P<0.001). As a consequence of aqueous analysis, change of treatment was necessary in 36 patients (24%). None of the patients experienced complications during or after aqueous sampling. CONCLUSIONS: Despite the posterior location of inflammation, aqueous analyses with PCR and GWC for HSV, VZV, CMV, and T. gondii revealed an infectious cause in 29% of patients with PU.