This is a cross-sectional analysis of a longitudinal PsA cohort, in which all PsA patients referred to the Rheumatology Unit of the University of Molise from January 1, 2019 to September 1, 2019 were considered eligible.

Inclusion criteria were:

Age ≥ 18 years old,

PsA, satisfying ClASsification criteria for Psoriatic ARthritis (CASPAR) [9],

PsO confirmed by a dermatologist.

For each patient, age, sex, body mass index (BMI), smoking habits, PsO and PsA disease duration, predominant PsA subset at the time of the visit, PsO at the time of the visit (yes/no), body surface area (BSA) [10], uveitis, Crohn’s disease, ulcerative colitis, were collected.

To assess PsA disease activity, Patient Global Assessment, Physician’s global evaluation of disease activity [11], patient ‘pain on Visual Analogue Scale (VAS), tender/68 (TJC) and swollen/66 joint count (SwJC), Leeds Enthesitis Index [12], dactylitis (past, present, never) and C-reactive protein (CRP) were collected. Moreover, the following indices were calculated: Disease Activity for Psoriatic Arthritis and Minimal Disease Activity [13]. Finally, Health Assessment Questionnaire-Disability Index [14], Psoriatic Arthritis Impact of the Disease [15] and Patient Acceptable Symptoms State [16] were also performed.

For each patient, the current rheumatological therapy was collected. The Charlson Comorbidity Index (CCI), the Rheumatic Disease Comorbidity Index (RDCI) and the Functional Comorbidity Index (FCI) [17] (see Table S1 in the electronic supplementary material for details) were also calculated. Comorbidities were recorded based on previous diagnosis.

PsA patients were stratified in EOP (0–40 years) or LOP (over 40 years) and then compared to evaluate the potential differences in PsA disease characteristics, disease activity, impact of the disease, function, and comorbidity indices.

The study protocol followed the Declaration of Helsinki and was approved by the Institutional Review Board of the University of Molise (protocol n. 0002-09-2017).

Statistical analysis was performed using the R software (version 3.6.2). All demographic and clinical characteristics were summarized by using descriptive statistics. Normally distributed variables were reported by mean ± standard deviation (SD), and non-normally distributed variables by median and inter-quartile range (IQR). The distribution of normality was evaluated by Kolmogorov–Smirnov test. Categorical data are shown as number (n.) and percentage (%) of valid data.

To compare EOP and LOP PsA patients, the χ2-test for independence and Student’s t test or Mann–Whitney U test were used.

Univariate logistic regression models were performed to analyze the association between EOP/LOP (independent factors), with all the clinical characteristics and comorbidity indices that showed a statistically significant difference between the two groups.

Multivariate logistic regression models were performed to adjust the significant association by sex and age when appropriate. Odds ratios (OR) were used as a measure of association and a statistical significance was defined as a two-tailed p value ≤ 0.05.