PURPOSE: To determine the usefulness of real-time polymerase chain reaction (PCR) assays in the diagnosis of postoperative bacterial endophthalmitis in clinically diagnosed infectious cases and to test for bacterial DNA in control samples collected from noninfected eyes. SETTING: Federal University of São Paulo, Brazil. DESIGN: Evaluation of diagnostic test or technology. METHODS: This study comprised patients with clinically diagnosed infectious endophthalmitis after cataract surgery and vitreous samples (from noninflamed eyes obtained through vitrectomy) and aqueous samples (at end of phacoemulsification) from control patients at a single university setting. Universal and gram-specific real-time PCR, Gram staining, and culture were performed. Sensitivity and cycle thresholds were determined. Clinical and microbiologic data were also assessed. RESULTS: The study evaluated 11 patients with infectious endophthalmitis (9 vitreous and 7 aqueous samples), 12 control vitreous samples, and 50 control aqueous samples. Gram and culture identified 80% and 75%, respectively, of patients with infectious endophthalmitis. Real-time PCR assays were positive in 91% of patients with a clinical diagnosis of endophthalmitis using aqueous samples, vitreous samples, or both. None of the 12 vitreous controls were positive by PCR. Two aqueous control samples were positive by real-time PCR. The cycle threshold cutoff value was 36 for universal PCR (sensitivity 93.8%; specificity 100%) and 38 for gram-specific PCR (sensitivity 93.8%; specificity 100%). Gram-positive microorganisms prevailed, and visual acuity varied according to the causative bacteria. CONCLUSIONS: Real-time PCR provided fast and accurate diagnosis of bacterial endophthalmitis. As a quantitative technique, it may be useful in distinguishing between contamination and infection based on the cycle thresholds value.
PURPOSE: To determine the usefulness of real-time polymerase chain reaction (PCR) assays in the diagnosis of postoperative bacterial endophthalmitis in clinically diagnosed infectious cases and to test for bacterial DNA in control samples collected from noninfected eyes. SETTING: Federal University of São Paulo, Brazil. DESIGN: Evaluation of diagnostic test or technology. METHODS: This study comprised patients with clinically diagnosed infectious endophthalmitis after cataract surgery and vitreous samples (from noninflamed eyes obtained through vitrectomy) and aqueous samples (at end of phacoemulsification) from control patients at a single university setting. Universal and gram-specific real-time PCR, Gram staining, and culture were performed. Sensitivity and cycle thresholds were determined. Clinical and microbiologic data were also assessed. RESULTS: The study evaluated 11 patients with infectious endophthalmitis (9 vitreous and 7 aqueous samples), 12 control vitreous samples, and 50 control aqueous samples. Gram and culture identified 80% and 75%, respectively, of patients with infectious endophthalmitis. Real-time PCR assays were positive in 91% of patients with a clinical diagnosis of endophthalmitis using aqueous samples, vitreous samples, or both. None of the 12 vitreous controls were positive by PCR. Two aqueous control samples were positive by real-time PCR. The cycle threshold cutoff value was 36 for universal PCR (sensitivity 93.8%; specificity 100%) and 38 for gram-specific PCR (sensitivity 93.8%; specificity 100%). Gram-positive microorganisms prevailed, and visual acuity varied according to the causative bacteria. CONCLUSIONS: Real-time PCR provided fast and accurate diagnosis of bacterial endophthalmitis. As a quantitative technique, it may be useful in distinguishing between contamination and infection based on the cycle thresholds value.