Low serum IgG at apheresis predicts delayed humoral immune reconstitution and increased risk of infections after CD19 CAR T-cell therapy for B-cell lymphoma

Abstract Background: Infections account for approximately half of non-relapse mortality following CAR T-cell therapy. Most infections occur within the first 28 days post-infusion, driven by cytopenia resulting from lymphodepleting chemotherapy. However, humoral immunodeficiency due to B-cell aplasia can persist for over a year, leaving patients at sustained risk of infection. Predicting delayed humoral immune reconstitution and infection risk from baseline parameters available prior to infusion may provide insights for infection prevention strategies. In CD19 CAR T-cell therapy for B-cell lymphoma (BCL), the impact of pre-infusion humoral immunity on post-infusion infection risk remains controversial. Methods: We retrospectively analyzed patients with BCL who underwent apheresis for lisocabtagene maraleucel at our center between September 2021 and December 2024. Patients were stratified into two groups based on the median serum IgG levels at the time of apheresis. Endpoints included post-infusion infection, severe infection (CTCAE grade ≥3), infection-related mortality, hematologic recovery, IgG recovery, and treatment outcome (OS, PFS, relapse or disease progression, treatment-related mortality). Fever of unknown origin, whether neutropenic or not, was excluded. Blood counts and IgG levels were collected at days –5, 28, 90, 180, and 365 post-infusion. Multivariate analyses were performed for infectious outcomes. Results: A total of 56 patients were included in the analysis: 33 with DLBCL, 6 with HGBCL, 11 with transformed indolent lymphomas, 3 with FL, and 3 with PMBCL. The median serum IgG level at the time of apheresis was 713.5 mg/dL. Based on this, patients were divided into two groups: IgG levels <700 mg/dL (Low group, n = 27) and ≥700 mg/dL (High group, n = 29). There were no significant differences in age, sex, or ECOG PS between the two groups. Patients who had received ≥3 prior lines of chemotherapy was 44.8% in High group and 66.7% in Low group (p = 0.17). A history of hematopoietic stem cell transplantation tended to be more frequent in Low group (44.4% vs 20.7%, p = 0.11). Median follow-up was 19.9 months. At 1 year post-infusion, the cumulative incidence of infections was significantly higher in Low group (53.8% [95% CI: 28.1–73.9%] vs. 25.7% [95% CI: 7.9–48.3%]; p < 0.01). The incidence of severe infections was also higher in Low group (40.5% [95% CI: 20.3–59.8%] vs. 6.6% [95% CI: 0.3–27.0%]; p < 0.01). At 1 year, no infection-related mortality was observed in High group, whereas 17.7% (95% CI: 5.0–36.8%, p = 0.05) in Low group. In multivariate analysis, serum IgG <700 mg/dL remained a significant factor for all infections (HR: 3.11, p = 0.02) and severe infections (HR: 6.91, p < 0.01), after adjustment for age, PS, number of prior therapies, and transplant history. Comparison of hematologic parameters at five time points showed no differences in neutrophil, lymphocyte, or hemoglobin counts between groups. Platelet counts were higher in High group at day-5, 90, 180, and 365. Serum IgG levels were significantly lower in Low group at day -5 (455.0 [IQR: 396.0–529.0] vs. 705.0 [IQR: 639.0–805.5], p < 0.01), and remained lower at day 28 (460.5 [IQR: 418.0–525.0] vs. 624.0 [IQR: 586.5–783.25], p < 0.01), day 90 (367.5 [IQR: 275.75–442.25] vs. 540.0 [IQR: 453.0–625.75], p < 0.01), and day 180 (330.0 [IQR: 265.0–383.0] vs. 484.0 [IQR: 448.5–549.5], p < 0.01). At day 365, serum IgG levels recovered in High group but not in Low group (614.0 [IQR: 556.0–628.0] vs. 334.0 [IQR: 267.75–361.5], p = 0.01). A total of 9 patients (16.1%) received IVIG; all were from Low group and had experienced infections. At 1 year, OS was 79.6% in High group and 46.1% in Low group (p = 0.06), indicating a trend toward better survival. PFS was similar between groups (39.4% vs. 38.1%). Relapse or progression was 57.0% in High group and 43.2% in Low group (p = 0.45). TRM was higher in Low group (18.6% vs. 3.6%, p = 0.21), partly due to infection-related deaths. Conclusion: In liso-cel–treated BCL patients, IgG <700 mg/dL at apheresis was significantly associated with higher risks of overall and severe infections, independent of neutrophil and lymphocyte recovery. Additionally, low baseline IgG predicted supressed and delayed humoral immune recovery post-infusion. These findings highlight the potential utility of serum IgG as a simple pre-infusion marker to inform infection risk and guide prophylactic strategies in CD19 CAR T-cell therapy for BCL.