Nomogram for accurate and quantitative prediction of the risk of psoriatic arthritis in Chinese adult patients with moderate and severe plaque psoriasis.

Background: Psoriatic arthritis (PsA) is an inflammatory form of arthritis that appears approximately 7-10 years after psoriasis and remains undiagnosed in most of patients. Currently, only a few quantitative and succinct PsA-risk prediction models are available.

Objective: The aim of this study was to establish and validate a prediction model for quantitatively assessing the risk of PsA in moderate and severe plaque psoriasis patients.

Methods: A non-interventional and cross-sectional study was conducted. Demographic, clinical, and laboratory records were collected and blindly reviewed. Logistic regression was used to develop this prediction model. With C-index and calibration curve, internal validation was performed. Five-fold cross validation, external validation and decision curve analysis (DCA) were also applied to assess this model.

Results: Among 405 patients, 111 patients had PsA. Arthralgia (OR = 39.346; 95% CI: 20.139-82.579), C-reactive protein (OR = 2.008; 95% CI: 1.051-3.838), lymphocyte level (OR = 0.341; 95% CI: 0.177-0.621), hypertension (OR = 0.235; 95% CI: 0.077-0.660) and disease duration (OR = 1.033; 95% CI: 0.998-1.071) were identified as potential predictors affecting the risk of transition from moderate and severe PsO to PsA. C-index for the prediction nomogram was 0.911 (95% CI: 0.879-0.943), and was confirmed to be 0.905 through 1000-time bootstrapping internal validation. Cross validation and external validation were preformed and proved the accuracy and generalizability of this prediction model.

Conclusions: This study establishes a quantitative predictive nomogram with good predictive power for assessing the risk of PsA in patients with moderate and severe PsO.