Immunological Cytopenias Induced By Anti-Programmed Cell Death (ligand) 1 Antibodies

Background. Anti-programmed death (ligand) 1 (PD-(L)1) antibodies are novel immunotherapies for cancer. Anti-PD(L)1 can induce immune-related adverse events (irAEs), which most frequently affect the skin, endocrine glands, lung, liver and digestive tract, however all organs including the hematopoietic system can potentially be involved. Hematologic irAEs (hem-irAEs) have not yet been extensively characterized. This study based on a pharmacovigilance academic registry aims to provide a comprehensive report of hem-irAEs.

Patients and methods. All grade 2 or higher hem-irAEs recorded in pharmacovigilance databases were recorded in this study over the period 2013-2018. Pharmacovilance database included REISAMIC (Registre des Effets Indésirables Sévères des Anticorps Monoclonaux Immunomodulateurs en Cancérologie) and ImmunoTOX committee of Gustave Roussy, and the French nationwide CeReCAI registry (Centre de Référence des Cytopénies Auto-Immunes de l'adulte). Prevalence of hem-irAE was calculated in the prospective REISAMIC registry. The hem-irAE's severity was assessed according to the Common Terminology Criteria for Adverse Events (version 4.03).

Results. Thirty-five patients (21 men and 14 women) with hem-irAEs related to anti-PD(L)1s were recorded. The prevalence of grade ≥2 hem-irAEs among patients treated with anti-PD(L)1s was estimated to 0.54%. The patients' median age was 65 years (range: 30-90), tumor types were melanoma (n=15), non-small-cell lung carcinoma (n=12), lymphoma (n=4), and other types (n=4). A concomitant medical history of chronic lymphocytic leukemia was found in 3 (9%) patients. The hem-irAEs were classified as neutropenia (n=9; 26%), auto-immune hemolytic anemia (n=9; 26%), immune thrombocytopenia (IT) (n=9; 26%), pancytopenia or aplastic anemia (n=5; 14%), bicytopenia (n=2; 6%) and pure-red cell aplasia (n=1; 3%). The maximum grade of severity reported was grade 2 in 3 (9%) patients, grade 3 in 8 (23%) patients and grade 4 in 24 (69%) patients. Two patients (6%) had a poor outcome and died in the course of the hem-irAE. Treatments given for the hem-irAE were steroids in 31 (89%) patients, granulocyte colony-stimulating factor in 13 (37%) patients, intravenous immunoglobulins in 6 (17%) patients and rituximab in 2 (6%) patients. The median time to onset from anti-PD(L)1 initiation was 10.1 weeks (range 0.9 - 198.0). With a median follow-up of 19 (range 2-180) weeks, the rate of resolution of hem-irAE was 67% (78% for IT and 40% for aplastic anemia patients, p=0.524). After resolution, 6/35 (17%) patients were rechallenged and 3 of them (50%) recurred the same hem-irAE.

Conclusion. The clinical spectrum of hematologic irAE is wide and dominated by neutropenia, hemolytic anemia and immune thrombocytopenia. Aplastic anemia is rarer and present as a life-threatening condition. The prevalence of grade ≥2 hematologic irAEs following anti-PD(L)1 was 0.54%. The recurrence rate after rechallenge was 50%. Further prospective investigations are warranted to better detect and manage hematologic immune-related adverse events.