AB0194 THE IMPACT OF TNFα INHIBITORS ON GLUCOCORTICOID USE AMONG PATIENTS WITH ARTHRITIDES - A NATIONWIDE STUDY

Background: Glucocorticoids (GCs) remain a common co-treatment in arthritic diseases. Tumour necrosis factor alpha inhibitors (TNFi) effectively reduce disease activity and should allow for GC tapering. There have been few studies regarding the effect of TNFi treatment on GC use, mainly focusing on rheumatoid arthritis (RA). Objectives: To estimate whether and to what extent TNFi treatment reduces the use of GC in inflammatory joint diseases in a nationwide cohort. Methods: The study was a nationwide registry study and included all adult biologic-naïve patients with RA, psoriatic arthritis (PsA), and axial spondyloarthritis (axSpA) registered in ICEBIO who received their first TNFi treatment (2005-2018). Each patient was matched on age, sex and calendar time to five individuals from the general population. Data on all filled prescriptions of oral GC for two years before and two years after initiation of TNFi were retrieved from the Icelandic Prescriptions Medicine Register for patients and matched comparators. Baseline patient demographics and disease activity scores were extracted from ICEBIO at the initiation of TNFi therapy (baseline) and at 18 months after initiation. Wilcoxon rank-sum test and paired t-test were used for comparisons before and after TNFi and between patients and comparators. GC use predictions were estimated using univariable and multivariable binomial logistic and quasi-Poisson regression models. Models were constructed for baseline and month-18 values. Results: A total of 1290 patients diagnosed with RA (41%), PsA (31%) and axSpA (28%) started treatment with a TNFi during the study period. They filled 4349 prescriptions of GC while 5413 comparators filled 1005 (p< 0.001). The proportion of patients with arthritis using GC increased steadily over the 24 months leading up to initiating TNFi; 20.8% were prescribed GC at 24 to 18 months, and 37.1% were prescribed GCs in the last six months before TNFi treatment. The largest increase in GC prescriptions leading up to TNFi initiation was seen among RA patients, or 21.0 percentage points (Figure 1). Following the initiation of TNFi the proportion got steadily lower with 19.6% of patients being prescribed GC 18 to 24 months after starting TNFi therapy. The decrease was most prominent among RA patients, or 29.1 percentage points (Figure 1). There was a significant difference in number of prescriptions (NP) for patients before and after initiation of TNFi (p<0.001). The patients in the top 25% of GC users before initiation of TNFi had lower DDDs following initiation of TNFi, and there was a statistically significant decrease in DDDs for all diseases (p<0.001 for all diseases; Figure 2). In multivariable analysis predicting the risk of NP of GC following initiation of TNFi, female sex, history of smoking, NPs and DDDs of GC before TNFi initiation and diagnosis of RA were positively associated with an increased risk of GC prescriptions following TNFi initiation in the baseline model. In the month 18 model, female sex, age, NPs before TNFi, smoking and HAQ scores were significantly associated with an increased risk of more NPs following initiation of TNFi. Only HAQ scores showed a significantly increased risk of NPs following initiation for RA patients; no disease-specific activity scores showed a significant correlation with NP. In multivariable analysis predicting the risk of being on GCs 18 to 24 months after initiation of TNFi, only female sex and NPs of GC before starting TNFi were statistically significant in the baseline and month 18 models. No disease activity scores were associated with an increased risk of being on GCs 1.5 to 2 years after initiation of TNFi. Conclusion: TNFi initiation decreases the proportion of arthritis patients on GC therapy and decreases DDD for top GC users. Disease activity does not seem to have an association in predicting the risk of filling prescriptions following initiation or still being on GC therapy two years after starting TNFi therapy. REFERENCES: NIL. Acknowledgements: ICEBIO study group. Disclosure of Interests: None declared.