Background The relationship between nonradiographic axial Spondyloarthritis (nr-axSpA) and Ankylosing Spondylitis (AS) is currently debated. In Sweden, anti-TNF agents were often used to treat patients with nr-axSpA, even before this phenotype was formally characterised by the 2009 ASAS criteria.
Objectives To compare clinical and laboratory patterns between patients with nr-axSpA and AS during three years of anti-TNF treatment.
Methods Patients with nr-axSpA (n=99) or AS (n=238), starting anti-TNF therapy at a Rheumatology university department in southern Sweden, were monitored between 1999 and 2013. Diagnoses were based on data collected at anti-TNF initiation, applying the 2009 ASAS criteria for nr-axSpA and the modified New York criteria for AS. Visits were routinely scheduled at 0, 3, 6, 12, 24 and 36 months. Study dropout was defined as ceasing anti-TNF therapy, whithout starting another anti-TNF agent within three months. Baseline differences between the nr-axSpA and AS patients were studied by χ2 and Mann-Whitney U-test, while differences in the three year development of VAS global, VAS pain, EQ-5D, doctors'evaluation of disease activity (on a 5-grade Likert scale), CRP and ESR were assessed by repeated ANOVA (including all timepoints), adjusting for sex, age, disease duration and the presence of peripheral arthritis. Last observation carried forward was applied from study dropout and to impute missing data. For EQ-5D, due to a higher proportion of missing data (37%), lacking values at 0 and 3 months were first imputed by linear regression. Analyses restricted to observed data were also conducted.
Results More men were identified in the AS group (nr-axSpA/AS: 62/76%; p=0.009), and significant (p<0.05) baseline differences were also seen in age (nr-axSpA/AS mean [SD]: 38 /43  years), self reported disease duration (9.1 [8.9]/16  years), doctors' evaluation of disease activity (1.7 [0.7]/2.0 [0.7]), CRP (8.9 /23  mg/l) and ESR (22 /28  mm/h). Study dropout was similar in both groups (nr-axSpA/AS: 27/23%; Log rank test p=0.355). From 0 to 36 months, no significant between-group differences were seen in the development of VAS global, VAS pain, EQ-5D or doctors' evaluation of disease activity, regardless of analysing observed or imputed data. The pattern of CRP levels differed by both methods (observed/imputed p=0.034/0.001), while there was a discrepancy in the results regarding ESR (p=0.109/0.031; curves visually similar to those presented for CRP).
Conclusions Despite baseline variables indicating a more severe and longstanding disease in the AS group, patient reported outcomes and doctors' evaluation of disease activitiy developed similarly between the nr-axSpA and AS patients following the start of anti-TNF treatment and during three years of follow up. The overall CRP level was higher in the AS group, with the greatest difference seen at treatment initiation.
Disclosure of Interest None declared