Comparing immune-checkpoint inhibitor-mediated inflammatory arthritis to polymyalgia rheumatica: a prospective cohort study

Background Immune checkpoint inhibitor (ICI)-related inflammatory arthritis occurs in approximately 6% of ICI-treated patients, and is often long-lasting. The clinical presentation is heterogeneous, and often loosely resembles either rheumatoid arthritis (ICI-IA) with peripheral presentation, or polymyalgia rheumatica (ICI-PMR) with predominant bilateral shoulder and hip pain and stiffness. Our objective was to compare clinical features, immunosuppressive treatment, and cancer outcomes between ICI-IA and ICI-PMR. Methods Data were extracted from the Rheumatology Adverse Events Due to Immunotherapy Observational Study multicenter prospective registry, regarding patients with rheumatologist-diagnosed ICI-IA or ICI-PMR and no preexisting primary IA/PMR. Clinical features were collected at the first visit in the registry. Multivariable logistic regression was used to compare second-line immunosuppressive drug utilization between groups. Kaplan-Meier curves were constructed to compare time of prednisone tapering to a dose of (1) 10 mg and (2) 5 mg from the baseline registry visit. Cox proportional hazard models were used to compare progression-free survival (PFS). Date of first cancer progression from arthritis onset was extracted based on documentation of imaging and/or pathology in the medical record. Patients with ICI-PMR were also compared with an independent cohort of patients with primary PMR. Results We analyzed 490 patients: 418 with ICI-IA and 72 with ICI-PMR. Compared with patients with ICI-IA, patients with ICI-PMR were older (70 vs 63 years old, p<0.001), more likely to be male (p=0.034), had shorter onset after ICI initiation (87 vs 125 days, p=0.011) and had less peripheral synovitis (15% vs 72%, p<0.001). Time to a prednisone dose of 10 and 5 mg/day was comparable between groups. Patients with ICI-IA were more likely to receive a second-line immunosuppressive drug (OR 3.51 (95% CI 1.50 to 8.21, p=0.004)). PFS was comparable between groups. Compared with primary PMR (n=189), the sex ratio was inversed in ICI-PMR, but there was a similar proportion with peripheral arthritis (15%). Conclusions Patients with ICI-IA were more likely to require a second-line immunosuppressive drug than patients with ICI-PMR, but PFS was the same. Future therapeutic clinical trials in ICI-IA/PMR should stratify patients according to phenotype.