Page 55 - SPONDYLOARTHRITIS IN COLOMBIA -

analysis were entered in a (backward selection) multivariable analysis if data availability was at least 80%, with SI-MRI ordered or HLA-B27-ordered as dependent variables.
Based on the hypothesis that IBP will primarily lead to ordering SI-MRI in male patients and in those that have short disease duration, interactions of characteristics with gender (male/female) and with disease duration (short/long) respectively were tested. Potentially relevant interactions (p≤0.1) were found with age of symptoms onset and age of assessment. These interactions were weak, not considered clinically relevant and not reported.
In the logistic regression analysis, the following variables were excluded because of too much missing information (≥20% of the cohort): BASFI, BASDAI, CRP, ESR, ASDAS-CRP and ASDAS-ESR. In addition, IBD and age of symptom onset were excluded because of collinearity. Statistical analyses were performed with SPSS 20.0 and STATA 12.0.
RESULTS
Patients’ characteristics
In total 581 patients with a clinical diagnosis of SpA were analysed. Most of them (72%) were males with a mean (SD) age of 34.6 (12.1) years, age at symptom onset of 28 (10.3) years and disease duration of 7.3 (9.7) years. Sixty percent of the cohort had less than 5 years of disease duration. Disease activity was rather high in the patients with axial SpA: BASDAI 5.4 (2.4), ASDAS-ESR 3.0 (0.9), and ASDAS-CRP 2.6 (0.8), and also in patients with peripheral SpA as assessed by the ASDAS-CRP (2.3 (0.9)). Information on SI-MRI and HLA-B27 was available in a varying proportions of the patients, ordered at the instigation of the rheumatologist: of all patients, 137 patients (24%) had SI-MRI and 441 patients (77%) had HLA-B27 testing performed.
53