OP0023 Effect of anterior uveitis, psoriasis and inflammatory bowel disease on drug-survival for tnf-inhibitors in ankylosing spondylitis

Abstract

Background: Tumour necrosis factor inhibition (TNFi) is the mainstay treatment for ankylosing spondylitis (AS) with a high disease activity. Whereas disease activity and sex have been shown to affect drug-survival for TNFi in AS, the impact of typical AS-comorbidities, such as anterior uveitis, psoriasis and inflammatory bowel disease (IBD), on the drug-survival of TNFi in AS is less well understood. Objective(s): To determine the impact of comorbidity with anterior uveitis, psoriasis and IBD on drug-survival in patients with AS starting treatment with a first TNFi. Method(s): Swedish biologics-naive patients with AS starting a 1 st TNFi July 1 2006-December 31 2015, were identified in the Rheumatology Quality Register, and followed from treatment start until treatment discontinuation. Censoring occurred at the first of: December 31 2015, death, emigration, or loss of follow-up. Comorbidities, and potential confounders, were identified through linkage to six other national registers. We calculated survival curves and hazard ratios (HRs) for each comorbidity and the risk of TNFi discontinuation. HRs were adjusted for sex, age, CRP, peripheral arthritis, type of TNFi and BASDAI at baseline. Additional models were also adjusted for other chronic morbidities (cardiovascular disease, affective disease, diabetes, malignancies, chronic lung disease and chronic kidney failure), and for socioeconomic status (length of education, household income, sick-leave, country of birth and civil status), respectively. Result(s): 2577 patients (71% men) were identified, 27% had a previous history of anterior uveitis, 7% IBD and 6% psoriasis. A history of anterior uveitis was associated with a lower risk of TNFi discontinuation (HR 0.72; 0.62-0.83), whereas presence of psoriasis was associated with an increased risk (HR 1.48; 1.18-1.86). No association was found between presence of IBD and risk of TNFi discontinuation. Models adjusting for disease activity, morbidities, and socioeconomic status resulted in an attenuated association for psoriasis (table 1). The impact of each comorbidity on drug-survival is visualised in figure 1. Conclusion(s): Comorbidities typically associated with AS affect TNFi drug survival, but in different directions. Possible explanations include a differential effect of TNFi on these morbidities themselves, or other inherent differences in the AS inflammatory phenotype. The association with psoriasis decreased after adjustment, suggesting the influence of other factors. The role of inflammatory comorbidities in determining response and persistence of TNFi should be further examined. (Table Presented).