RHEUMATIC IMMUNE- AND NONIMMUNE-RELATED ADVERSE EVENTS IN PHASE 3 CLINICAL TRIALS ASSESSING PD-(L)1 CHECKPOINT INHIBITORS FOR LUNG CANCER:A SYSTEMATIC REVIEW AND META-ANALYSIS

Background Several adverse events (AEs) occurring during immune checkpoint inhibitors (ICIs) therapy are clearly related to their mechanisms of action, and in this case they are indicated as immune-related AEs (irAEs). Every organ may be affected, including the musculoskeletal system; myositis, polymyalgia rheumatica, arthritis or arthritis have been reported in several retrospective and prospective case series and cohorts, with an incidence between 1.5% and 22%. While arthritis, vasculitis, myositis, and polymyalgia rheumatica are usually defined as “irAE” in RCTs, other rheumatic musculoskeletal conditions such as arthralgia, myalgia, back pain and muscular pain are often reported under the umbrella of “general” AEs. Objectives We aimed to analyze rheumatic irAE and non-irAE due to immune-checkpoint inhibitors (ICIs) targeting programmed cell death-1 or its ligand PD-(L)1 in lung cancer patients from the available literature. Methods We performed a systematic review and meta-analysis of phase III randomized clinical trials (RCTs) assessing PD-(L)1 -ICIs in lung cancer patients, from inception until January 12 th , 2021. We extracted data of each trial to estimate odds ratio (OR) for rheumatic ir or non-irAE as classified in RCTs safety data. Sensitivity analyses (by ICI, treatment group and histology) were performed. Results Eighteen RCTs met the inclusion criteria (n=12172 subjects). The OR [95%IC] for rheumatic irAE in ICIs versus controls (either placebo or chemotherapy) was 2.20 [0.85,5.72]. Among rheumatic non-irAEs, both overall (any grade, Figure 1A) and severe (grade≥3, Figure 1B) back pain were significantly more frequent in ICIs versus controls (2.01 [1.09;3.73] and 2.90 [1.18;7.08], respectively). Figure 1. Forest plot showing pooled odds ratio (OR) for back pain (5 phase III trials) (A) and severe back pain (4 phase III trials) (B), respectively. The overall frequency of arthralgia and severe arthralgia was similar between ICIs and controls (1.13 [0.86, 1.47] and 1.69 [0.68, 4.20], respectively). By sensitivity analysis RCTs assessing ICIs in combination with chemotherapy versus chemotherapy alone showed a significant association with arthralgia (1.55 [1.15;2.10]). Similarly, the frequency of myalgia did not differ between ICIs and controls, but was significantly lower in RCTs assessing ICIs alone versus chemotherapy (OR 0.32 [0.24;0.42]). Muscular pain was not significantly increased with ICI. Conclusion Rheumatic irAEs are not increased in RCTs assessing PD-(L)1 inhibitors, not reflecting the real-life incidence, therefore likely underreported or misclassified. Back pain is significantly associated with PD-(L)1-ICIs regardless its severity, suggesting a possible implication of the PD-(L)1 axis in the development of inflammatory back pain in some patients. In addition, PD-(L)1-ICIs added on conventional chemotherapy are associated with a significantly higher frequency of arthralgia than ICI alone. This trend was seen in the other rheumatic AEs, suggesting that conventional chemotherapy might be a confounder in the interpretation of the occurrence of rheumatic AEs. Disclosure of Interests Antonello Veccia: None declared, Marie Kostine: None declared, Alice Tison: None declared, Mariachiara Dipasquale: None declared, Stefania Kingspergher: None declared, Guido Grandi: None declared, Orazio Caffo: None declared, Sandro Inchiostro: None declared, Giuseppe Paolazzi: None declared, Roberto Bortolotti: None declared, Divi Cornec: None declared, Alvise Berti Consultant of: GSK