Development and validation of new diagnostic criteria for acute retinal necrosis

Purpose: The purposes of this study are to develop and validate new diagnostic criteria for acute retinal necrosis (ARN) based on the ocular findings, clinical course, and virologic testing of intraocular fluids. Subjects and methods: The Japanese ARN Study Group, comprising 8 uveitis specialists and 1 statistician, was formed to develop new diagnostic criteria for ARN. The criteria used a combination of clinical features consistent with ARN including 6 early-stage ocular findings ([1a] anterior chamber cells or mutton-fat keratic precipitates; [1b] yellow-white lesion(s) in the peripheral retina [granular or patchy in the early stage, then gradually merging]; [1c] retinal arteritis; [1d] hyperemia of the optic disc; [1e] inflammatory vitreous opacities; and [1f] elevated intraocular pressure), 5 clinical courses ([2a] rapid expansion of the retinal lesion(s) circumferentially, [2b] development of retinal breaks or retinal detachment, [2c] retinal vascular occlusion, [2d] optic atrophy, and [2e] response to antiviral agents), and the results of virologic testing of intraocular fluids by means of either polymerase chain reaction or the Goldmann-Witmer coefficient for herpes simplex virus or varicella zoster virus. Various combinations of findings were analyzed to maximize the sensitivity, specificity, positive predictive value (PPV), and negative predictive value (NPV). The criteria were then used to retrospectively analyze patients who had been diagnosed as having ARN or control uveitis. Patients were followed at 1 of 7 tertiary uveitis clinics between 2009 and 2011. Results: Analysis of the data allowed delineation of 2 levels of diagnosis: “virus-confirmed ARN” (defined as the presence of both early-stage ocular findings 1a and 1b, the presence of any 1 of the 5 clinical courses, and a positive virologic test result) and “virus-unconfirmed ARN” (defined as the presence of 4 of 6 early-stage ocular findings including 1a and 1b, presence of any 2 of the 5 clinical courses, and a negative virologic test result, or when virologic testing had not been performed). The new diagnostic criteria were applied to 45 patients with ARN and 409 patients with control uveitis, resulting in a sensitivity of 0.89, a specificity of 1.00, a PPV of 1.00, and an NPV of 0.99. Conclusions: New diagnostic criteria for ARN were developed and found to achieve high statistical values.