ROLE OF DIAGNOSTIC PARS PLANA VITRECTOMY IN DETERMINING THE ETIOLOGY OF UVEITIS INITIALLY UNKNOWN.

PURPOSE: To estimate the success and safety of diagnostic pars plana vitrectomy (PPV) in determining the etiology of uveitis initially unknown and analyze their characteristics.

METHODS: The PubMed, Embase, and Ovid were searched up to October 2017 to identify relevant studies. The PRISMA guidelines were followed. Statistical analyses were performed with R version 3.3.1. Result in proportion were transformed by the Freeman-Tukey variant of arcsine square to attain symmetry of confidence intervals (CIs). The statistical heterogeneity was assessed by the chi-square test and I statistics. Sensitivity analysis and subgroup analyses were performed to identify the source of heterogeneity. Publication bias was assessed by the Egger test.

RESULTS: Sixteen studies involving 1,195 patients were finally included. The pooling result showed the successful rate of diagnostic PPV was 44% (95% CI [39%∼50%]). Among patients whose diagnostic PPV yielded a definitive diagnosis, 69% were infectious uveitis (95% CI [61%∼77%]), 23% were lymphoma (95% CI [17%∼30%]), and 4% were metastatic carcinoma (95% CI [2%∼10%]). Among patients diagnosed with infectious uveitis, the most frequent pathogens identified were viruses, followed by bacteria, Toxocara canis, Toxoplasma gondii, tuberculosis, and fungus. The incidence of postoperative cataract and postoperative retinal detachment was 19% (95% CI [8%∼29%]) and 5% (95% CI [1%∼10%]), respectively, and the rate of secondary vitrectomy surgery was 10% (95% CI [2%∼22%]). Postoperative visual improvement rate was 46% (95% CI [39%∼52%]), and the postoperative treatment strategy was changed in light of the results of diagnostic PPV in 20% of cases (95% CI [10%∼29%]).

CONCLUSION: For uveitis of unknown cause, diagnostic PPV is an effective, reliable, and relatively safe procedure for establishing the definite diagnosis and guiding further treatment. Positive therapeutic effect could also be achieved.