AB0705 LONG TERM SAFETY OF CONVENTIONAL AND BIOLOGIC IL-6RI IMMUNOMODULATORS AS SECOND-LINE (2L) OR THIRD-LINE (3L) TREATMENT OF POLYMYALGIA RHEUMATICA

Abstract

BackgroundIn polymyalgia rheumatica (PMR), IL-6Ri therapy has been studied in limited number of patients in RCTs, with exposure up to 16-52 weeks,[1-3]while patients may require longer treatment in real world. Data are presented on safety of IL-6Ri and conventional immunomodulatory (CIM) therapy for up to 2 years (Y) in patients with PMR.ObjectivesEvaluate safety of IL-6Ri and CIM therapy in 2L or 3L treatment of PMR.MethodsThis was a retrospective study in US adults ≥50Y, with 1 inpatient/2 outpatient claims with PMR diagnosis ≥30 days apart, from national fee-for-service Medicare data (3/29/2016 - 6/30/2020). Patients were on ≤25 mg prednisone (PS) equivalent and started IL-6Ri (tocilizumab/sarilumab) or CIM (methotrexate/leflunomide/azathioprine) as 2L/3L therapy and had continuous enrollment ≥180 days prior to index. For 2L, index date was start of IL-6Ri/CIM with no prior use of IL-6Ri/CIM. For 3L, index date was start of IL-6Ri (with/without background CIM) or new CIM, after prior CIM. Follow-up was until the earliest of end of enrollment, death, treatment switch, or 2Y. Patients with seropositive RA, other arthritis or connective tissue disease, giant cell arteritis, organ transplant, or malignancy undergoing treatment were excluded. Stratifying by line of therapy, IL-6Ri patients were direct matched using 1:3 variable ratio matching on age, sex, index date and baseline PS equivalent dose category (<2.5, 2.5-<5, 5-<10, 10-<15, 15-<20, 20-25 mg), recency of prior CIM (3L only; 1-60, 61-180, 180+days), and then 1:1 propensity score matched using multivariate logistic regression to identify variables with SMD >0.1. We examined hospitalized infection (any and primary diagnosis), gastrointestinal perforation, major adverse cardiac events, malignancy, and drug induced liver injury. Except for infection, patients with prior events were excluded from each outcome-specific analysis. Outcomes were identified using validated or previously used claims-based algorithms. Incidence rates were reported per 100 patient-years through 2Y of follow-up, using exact Poisson for 95% CI, if events <5. Cox proportional hazards model was used to estimate adjusted hazard ratios (aHR) for infections (given paucity of other events). Sensitivity analyses examined events in Y1 and Y2 separately and stratified by 2L vs 3L.ResultsIn final cohort, covariate balance for 451 matched patients (2L: 183; 3L: 268) in each arm was good. Residual imbalances were small for disability status, steroid dose category, and chronic obstructive pulmonary disease. Incidence rates and aHRs for events are shown in Table 1; aHRs were not significant for IL-6Ri vs CIM. Sensitivity analyses were consistent with primary results and generally showed incidence was similar, or numerically lower for IL-6Ri vs CIM in Y2 and lower for CIM vs IL-6Ri in Y1, for most outcomes.Table 1.Pooled incidence rates and aHR stratified by 2L/3L of safety events for IL-6Ri and CIM-treated patients with up to 2Y exposureOutcomeExposure GroupEvents/100PYaHR (95% CI)IL-6Ri vs CIMHospitalized Infection, anyIL-6Ri14.4 (11.8 - 17.7)1.05 (0.79 - 1.41)CIM13.4 (10.8 - 16.6)Hospitalized infection, primary diagnosisIL-6Ri9.8 (7.7 - 12.5)1.16 (0.82 - 1.65)CIM8.4 (6.4 - 11.0)MalignancyIL-6Ri0.8 (0.3 - 1.9)NECIM2.0 (1.1 - 3.5)Major adverse cardiac eventsIL-6Ri2.1 (1.2 - 3.7)NECIM2.1 (1.2 - 3.7)Gastrointestinal perforationIL-6Ri0.1 (0.0 - 0.8)NECIM0.1 (0.0 - 0.8)Drug induced liver injuryIL-6Ri0.0 (0.0 - 0.5)NECIM0.0 (0.0 - 0.5)aHR, adjusted hazard ratios, controlling for covariates imbalanced with SMD >0.10; CI, confidence interval; CIM, conventional immunomodulatory therapy; NE, not estimated, inadequate sample size; PY, patient-years; IL-6Ri, interleukin-6 receptor inhibitorConclusionRates of serious adverse events for IL-6Ri vs CIM therapy with up to 2Y exposure were comparable in PMR.