Background/Purpose: Immune checkpoint therapy with programmed cell death protein 1 (PD-1) inhibitors has led to significant survival benefits in the treatment of multiple cancers. This strategy interrupts inhibitory signals within the immune system, but this may additionally cause immune-related adverse events (irAEs) that resemble a spectrum of well-recognised autoimmune conditions. Detailed study of these may also further the understanding of the development of spontaneous autoimmune disease. At present, characterisation of rheumatic irAEs is limited as reports from clinical trials have inadequate detail for this purpose and the current literature largely consists of case series. Few reports describe in detail all rheumatic irAEs in a whole cohort of patients receiving PD-1 inhibitors, which would enable determination of frequency and context to this emerging entity. To the best of our knowledge, we describe the largest such series to date.

Methods: A retrospective chart review was performed on all patients dispensed nivolumab or pembrolizumab in 2016 at our center. Patients diagnosed with a non-cutaneous irAE were identified as were patients with undiagnosed symptoms possibly consistent with a rheumatic irAE. Information recorded included details regarding the malignancy and its therapy, the nature of the irAE and its treatment and outcomes.

Results: Of the 149 patients, 54 (36.2%) had at least one non-cutaneous irAE. Thyroid irAEs were the most common, occurring in 22 patients (14.8%) followed by rheumatic irAEs, which were diagnosed in 18 patients (12.1%). Of these, 11 were de novo diagnoses and 7 were exacerbations of existing autoimmune disease. A further 4 had symptoms possibly consistent with a rheumatic irAE but without a diagnosis from the treating physician. Inflammatory arthritis was the most common rheumatic irAE, with a diagnosis in 15 patients. The median time to developing a rheumatic irAE was 5 months (range 1-28). Rheumatic irAEs were more common in patients with an oncological response to therapy (RR 10.0, p<0.01) and receiving combination therapy with ipilimumab (RR 3.5, p = 0.016). Most patients were treated with corticosteroids and only 4 patients required introduction of DMARDs. Confirmation by a rheumatologist only occurred in 10 patients but supportive findings on imaging were available in 13 patients, including 8 patients not reviewed by a rheumatologist.

Conclusion: Rheumatic irAEs are a relatively common consequence of PD-1 inhibitor therapy, and frequently require the use of corticosteroids. They would benefit from systematic investigation to explore, amongst other characteristics, the apparent association with oncological response to therapy. Our findings also suggest that many rheumatic irAEs are not routinely referred to rheumatological services. While imaging may be able to assist with retrospective identification, if insights are to be made regarding the pathogenesis of the spontaneous diseases from irAEs, consistent rheumatological assessment will be required. A reliance on conventional mechanisms may be insufficient for this purpose and electronic biosurveillance may be necessary.

« Back to ACR/ARHP Annual Meeting 2017

ACR Meeting Abstracts - https://acrabstracts.org/abstract/rheumatic-immune-related-adverse-events-due-to-programmed-cell-death-protein-1-pd-1-inhibition-for-cancer-comprehensive-analysis-of-a-whole-cancer-cohort/