Spondyloarthritis
RESULTS Characteristics of patients
Erosions and new bone formation (NBF) were scored as 0/1. ‘Active synovitis’ was considered to be present if both a GS score ≥ 2 and a PD signal ≥ 1 were detected. 21 In addition, hand flexor tendons, extensor tendon compartments of the wrist as well as ankle plantar flexors and dorsiflexor tendons were evaluated bilaterally for the presence of tenosynovitis with both B-mode and Doppler in a dichotomous manner (0/1) according to the OMERACT guidelines. 22 Finally, 18 entheses were exam- ined ultrasonographically to assess for inflammatory and structural entheseal lesions. The details and results of the US entheseal evaluation have been reported in a previous publication. 23 Statistical analysis Statistical analysis was performed with R software, V.4.5.1. Categorical data were expressed as percentages and compared using Fisher’s exact test or the χ 2 test. Contin- uous data were presented as mean±SD or median with IQR, as appropriate. Continuous data were compared using the Student’s t-test or the Wilcoxon test in the case of two groups, while the analysis of variance or the Kruskal-Wallis test was chosen for the comparisons between three groups, as appropriate. P values <0.05 were considered statistically significant. Convergent construct validity was investigated through correlations (Spearman’s rank correlation test) of clinical parameters and disease activity measures, with different synovitis US scores and lesions. Additionally, further anal- yses using univariable and multivariable regression were performed to explore the association of US synovitis lesions with the clinical and inflammatory parameters of participants. To identify the optimal combination of joint regions that best discriminated between patients with SpA and HC, we assessed the best subset of sites using regression analyses and the Least Absolute Shrinkage and Selection Operator (LASSO) method, which uses L1 regularisa- tion to select out less predictive variables. Multivariable logistic regression was then performed on SpA diagnosis, including the selected joints, in addition to age and BMI, to account for confounding. We performed receiver operating characteristic curve analysis to evaluate the diagnostic performance of the different US synovitis scores. The optimal cut-off values for each US synovitis score were determined using the Youden index, defined as (sensitivity+specificity−1). For the sample size calculation, we expected a syno- vitis prevalence of 20% in patients with SpA and 5% in HC, based on previous studies. We considered an alloca- tion ratio of 2:1 for SpA cases to controls. With a desired power of β =0.80 and a significance level of α =0.05, and accounting for continuity correction, the sample for a two-sample test with a two-sided hypothesis was estimated to be 78 cases and 42 controls. We therefore concluded that with approximately 80 patients and 40 controls, our study would be sufficiently powered to address the outlined objectives.
Altogether, 121 participants were included: 80 with SpA (39 axSpA and 41 PsA) and 41 HC, with a total of 2662 joints assessed ultrasonographically. Overall, 57.0% of the participants were male, without significant gender differences across the three groups. Patients with axSpA and HC were younger than patients with PsA, while BMI was higher in patients with PsA compared with patients with axSpA and HC (p<0.001). Both disease duration and diagnostic delay were significantly higher in axSpA compared with PsA (p<0.001, p=0.019, respectively). While patients with PsA showed a higher frequency of dactylitis (58.3%) compared with axSpA (11.4%), p<0.001, only 28.6% of patients with PsA had axial disease. Patients with SpA demonstrated higher FiRST and SF-12 scores compared with HC, but without significant differ- ences between SpA subgroups. The main demographic and clinical characteristics of all groups are reported in online supplemental table 2. Clinical examination All TJC and SJC were significantly higher in patients with SpA compared with HC (table 1). Patients with PsA gener- ally presented higher TJC and SJC compared with patients with axSpA, except for the TJC-28, which was comparable between patients with PsA and axSpA (p=0.306). The prevalence of ≥ 1 tender joint in patients with SpA was significantly higher compared with HC (p<0.001), with no significant difference between patients with axSpA and PsA (p=0.86). The prevalence of ≥ 1 swollen joint in patients with SpA was also significantly higher than in HC (p<0.001), while there was a substantially higher preva- lence in PsA compared with axSpA (p=0.002). There were no significant differences between SpA subgroups regarding the GPAQ score, inflammatory markers, disease activity, quality-of-life measurements or comorbidities (online supplemental tables 3 and 4). A higher percentage of patients with axSpA (67.9%) were receiving a tumour necrosis factor inhibitor compared with patients with PsA (39.5%) (p=0.04), while patients with PsA were receiving more interleukin-17 (IL-17) inhibitors and IL-12/IL-23 inhibitors compared with axSpA (p=0.043) (online supplemental table 5). Prevalence and distribution of US lesions Regarding joint US lesions, SH was the most preva- lent, while erosions were the least frequent across all groups. In the SpA group, GS revealed the presence of SH in 540/1760 joints (31.1%), with grade 1 synovitis in 417 joints (24%), grade 2 in 104 joints (6.0%) and grade 3 in 19 joints (1.1%). While in HC, grade 1 synovitis was detected in 151 joints (16.7%), grade 2 in 22 joints (2.4%) and grade 3 in two joints (0.2%). Accordingly, GS synovitis was detected more frequently in patients with SpA compared with HC (p<0.001). PD signal abnor- malities were seen in 43/1760 joints (2.5%) in patients with SpA vs 0.9% in HC (p=0.035), with no significant
3
Elsehrawy GG, et al . RMD Open 2026; 12 :e006802. doi:10.1136/rmdopen-2026-006802
Powered by FlippingBook