Predicting infections in patients with haematological malignancies treated with chimeric antigen receptor T-cell therapies: A systematic scoping review and narrative synthesis.

Background: Chimeric antigen receptor T cells (CAR-T cells) are increasingly used to treat haematological malignancies. Strategies for preventing infections in CAR-T-treated patients rely on expert opinions and consensus guidelines.

Objectives: This scoping review aimed to identify risk factors for infections in CAR-T-treated patients with haematological malignancies.

Data Sources: A literature search utilized MEDLINE, EMBASE and Cochrane to identify relevant studies from conception until 30 September 2022.

Study Eligibility Criteria: Trials and observational studies were eligible.

Participants: Studies required ≥10 patients treated for haematological malignancy to report infection events (as defined by the study), and either (a) a descriptive, univariate or multivariate analysis of the relationship between infections event and a risk factors for infections, or (b) diagnostic performance of a biochemical/immunological marker in CAR-T-treated patients with infection.

Methods: A scoping review was conducted in accordance with PRISMA guidelines.

Data Sources: A literature search utilised MEDLINE, EMBASE and Cochrane to identify relevant studies from conception until September 30, 2022. Eligibility/Participants/Intervention: Trials and observational studies were eligible. Studies required ≥ 10 patients treated for haematological malignancy, to report infection events (as defined by the study), and either A) a descriptive, univariate or multivariate analysis of the relationship between infections event and a risk-factors for infections, or B) diagnostic performance of a biochemical/immunological marker in CAR-T treated patients with infection.

Assessment Of Risk Of Bias: Bias assessment was undertaken according to Joanna Brigg's Institute criteria for observational studies.

Methods Of Data Synthesis: Data were synthesized descriptively because of the heterogeneity of reporting.

Results: A total of 1522 patients across 15 studies were identified. All-cause infections across haematological malignancies were associated with lines of prior therapy, steroid administration, immune-effector cell-associated neurotoxicity and treatment-emergent neutropenia. Procalcitonin, C-reactive protein and cytokine profiles did not reliably predict infections. Predictors of viral, bacterial and fungal infections were poorly canvassed.

Discussion: Meta-analysis of the current literature is not possible because of significant heterogeneity in definitions of infections and risk factors, and small, underpowered cohort studies. Radical revision of how we approach reporting infections for novel therapies is required to promptly identify infection signals and associated risks in patients receiving novel therapies. Prior therapies, neutropenia, steroid administration and immune-effector cell-associated neurotoxicity remain the most associated with infections in CAR-T-treated patients.