Performance of the 2015 ACR-EULAR classification criteria for gout in a primary care population presenting with monoarthritis.

Abstract

To test the performance of the 2015 ACR-EULAR gout classification criteria against presence of SF MSU crystals in a primary healthcare population. The criteria were applied to an existing dataset of consecutive patients with monoarthritis presenting to Dutch family physicians; all patients underwent microscopic SF analysis by design. The data had been prospectively collected to develop a diagnostic decision rule for gout in 2010. Diagnostic performance was assessed by calculating area under the receiver operating characteristic curve, sensitivity, specificity, positive predictive value (PPV), negative predictive value (NPV) and constructing calibration plots for the full version of the criteria (including SF analysis results of all patients) and the clinical-only version (not including SF analysis results). Performance of both versions was compared with the 2010 diagnostic rule. Of 381 patients enrolled into the study, 216 (57%) were MSU crystal-positive. The full and clinical-only versions of the criteria had satisfactory area under the receiver operating characteristic curve (0.96 and 0.87, respectively), high specificity (0.98 and 0.84), high PPV (0.98 and 0.84), but lower sensitivity (0.68 and 0.68) and NPV (0.70 and 0.67). Specificity and PPV of both versions were higher compared with 0.71 and 0.89 of the 2010 diagnostic decision rule. The decison rule had the highest sensitivity and NPV (0.99 and 0.97). This study presents the first external validation of the 2015 ACR-EULAR gout classification criteria in a primary healthcare setting. The criteria perform well in this setting in patients presenting with monoarthritis for the purpose of enrolling into gout clinical trials.