Validation of the Fautrel classification criteria for adult-onset Still’s disease

Abstract Objectives To validate the Fautrel classification criteria for adult-onset Still's disease (AOSD) and to compare the discriminative performance to that of the Yamaguchi criteria. Methods We retrospectively reviewed the medical charts of 426 patients who had serum ferritin level and percentage glycosylated ferritin assayed at the biochemistry laboratory of Bichat Hospital. Medical data were extracted by use of a standardized form. All clinical, biological, and imaging features were collected, as well, evidence favoring an alternative diagnosis, specifically symptoms suggestive of other immune-mediated inflammatory diseases (IMID) or active infections. Patients were classified as AOSD patients or controls according to a predefined procedure, including consultation with a multidisciplinary expert group. Algorithms corresponding to the Fautrel and Yamaguchi classification criteria were applied for each patient. Results In all, 54 AOSD and 278 control patients were included. For the Fautrel criteria, the sensitivity was 87.0%, specificity 97.8%, and positive and negative predictive values 88.7% and 97.5%, respectively. For the standard Yamaguchi set—without strict application of exclusion criteria—the sensitivity was 96.3%, specificity 98.9%, and positive and negative predictive values 94.5% and 99.3%, respectively. If we applied a stricter definition of exclusion criteria, the sensitivity of the Yamaguchi set decreased to 31.5%. As wall, 37 AOSD diagnoses were missed. Conclusion This study validates the Fautrel classification criteria with a cohort independent of that used for the original publication. This criteria set demonstrates good sensitivity and specificity, overcomes exclusion criteria, and includes glycosylated ferritin level. It also confirms the high discriminative power of the Yamaguchi criteria, albeit substantially affected by how exclusion criteria are interpreted.