CAR-T therapy can effectively improve the remission and survival rates of patients with R/R B-ALL. However, adverse events, including CRS, ICANS, infections, and hematological toxicity, are associated with CAR-T therapy. In recent years, CAR-T therapy-related infections have attracted increasing attention. In some prospective clinical trials and retrospective studies, the incidence of infection was approximately 18–60% [11,12,13,14,15,16]. Our research suggests that the cumulative infection rate within 90 days after CAR-T transfusion was 67.1%, the early infection density was 1.94, and the late infection density was 0.8, which were similar to recent reports [3, 17,18,19].

Neutropenia is an important risk factor for bacterial infection, and early bacterial infection with CAR-T therapy may be related to multiple neutropenic episodes during this period. Neutropenia is more common after CAR-T therapy, which can be caused by many factors (including CRS and LD chemotherapy) [20,21,22,23]. In a clinical study of CAR-T therapy in patients with relapsed/refractory lymphoma, the incidence of neutropenia was 71%, and most of these cases (98%) occurred in the early phase after CAR-T therapy [24]. Fried et al. [22] reported that 72% patients (n = 38) with R/R B-ALL had severe neutropenia (grade ≥ 3), and the median occurrence time of neutropenia was by day 17 after the initiation of CAR-T therapy. Our data suggest that 16 children developed infections following CAR-T therapy, with a total of 20 bacterial infections in the early phase. All 16 children developed neutropenia, and seven of these patients had persistent neutropenia (lasting > 20 days). Therefore, we should pay attention to identify bacterial infection, perform anti-infection actively, and avoid cross-infection for patients with neutropenia after CAR-T treatment, especially those with persistent agranulocytosis. In addition, some studies have showed that patients could receive granulocyte–macrophage colony-stimulating factor (GM-CSF) or granulocyte colony-stimulating factor (G-CSF) treatment to decrease the duration of neutropenia, but this approach remains controversial [25,26,27]. In our center, none of the children was treated with GM-CSF or G-CSF. We plan to perform some researches to evaluate the advantages and disadvantages of this treatment for children who receive the CAR-T therapy in the future.

Viral infections mostly occur in the late phase after CAR-T infusion, which may be related to B-cell aplasia and hypogammaglobulinemia [28]. The “off-target” effect of CAR-Ts (CAR-Ts not only kill malignant B cells, but also target normal B cells) leads to the failure of B-cell regeneration, inducing hypogammaglobulinemia. The incidence of hypogammaglobulinemia varies across treatment centers, with reports indicating an incidence of 20–90% due to differences in research objects, definitions of hypogammaglobulinemia, and methods of immunoglobulin determination [12, 29,30,31]. In our study, 53 children had complete humoral immunity data including 27 (50.9%) with late-phase hypogammaglobulinemia. Among the 27 children with hypogammaglobulinemia, 11 developed infections, and the pathogens were identified in four children, of which three were viral infections. The infection density in the late phase was lower than that in the early phase, which may be related to the routine monthly infusion of gamma-globulin for patients in our center after CAR-T therapy (until 6 months after treatment), which reduces the incidence of hypogammaglobulinemia. Moreover, lymphocyte and neutrophil counts recovered over time in most cases. In our study, lymphocyte and neutrophil counts recovered in 54 (54/60, 90%) and 48 (48/71, 67.6%) children, respectively, in the late phase. At present, most reports show that respiratory viruses (including influenza virus, parainfluenza virus, metapneumovirus, and respiratory syncytial virus [32]) are the most common pathogens in the late phase of reinfusion, and only a small number of herpes viruses are observed. The reported incidence of cytomegalovirus (CMV) infection is 1–2%, with viremia constituting most cases, whereas organ damage is rare. However, an increasing number of fatal cases of viral infections have been reported in recent years [33, 34]. Respiratory tract infection was the most common in our study; however, the etiological examination of respiratory tract infection has not been checked routinely. Meanwhile, we detected the EBV, CMV, and human parvovirus B19 in the peripheral blood of patients (All patients were screened for virus including AIDS, hepatitis B and C, CMV, EBV and VB19, etc., before the treatment, and no infection was found.) by PCR every month. It showed that CMV and VB19 were common in our data, and all patients were viremia and mainly receive symptomatic treatment.

Fungal infections after CAR-T therapy are rare, with an incidence between 1 and 5% [9, 19], which may be related to persistent neutropenia or the long-term use of glucocorticoids. Our center usually administers antifungal treatment to children with fungal infections during CAR-T treatment. Only one case of urinary tract fungal infection (T. asahii) was detected, and the infection was successfully treated with voriconazole. The child had long-term cytopenia, neutropenia lasting up to 30 days, a history of glucocorticoid use, and persistent application of broad-spectrum antibiotics. These factors increase the risk of fungal infection in children. Therefore, antifungal drugs should be used preventively in children with high-risk factors.

Although infections following CAR-T therapy are common, life-threatening infections are rare. Hill et al. performed a retrospective analysis of 133 patients who underwent CD19 CAR-T therapy, and found that 30 children had 43 infectious episodes, but only two led to death [9]. A report on CAR-T treatment in children and adolescents showed that two of 39 patients died. One patient died of rhinocerebral mucormycosis and E. faecalis disseminated infection, and the other patient died of polymicrobial bloodstream infection with E. faecalis and S. epidermidis [3]. In this study, only one child died of grade 4 CRS complicated by infection, which was caused by A. baumannii infection in the respiratory tract and Staphylococcus wallichii bacteremia. Following treatment with imipenem, amikacin, and voriconazole, the oxygen saturation and blood pressure could not be maintained at stable levels, and the patient died of multiple organ dysfunction syndrome and septic shock. Early identification of infection and active anti-infection therapy are one of the effective strategies to reduce the mortality of CAR-T therapy-related infection. Therefore, when children have symptoms of infection, we should promptly identify the infection site, conduct pathogen identification, actively and empirically use broad-spectrum antibiotics, regularly evaluate the severity of infection, and adjust antibiotics according to the pathogen test results.

We also analyzed the clinical factors related to infection and found that they were related to tumor load, lymphodepleting chemotherapy, neutrophil deficiency and lymphocyte reduction, CRS and ICANS, etc. High tumor load and intensive lymphocyte clearance usually lead to an increased incidence of severe CRS and ICANS [35, 36], and patients with severe CRS have a higher risk of infection [3, 9]. CRS is one of the most common adverse reactions after CAR-T therapy and typically occurs 1–14 days after CAR-T infusion for a duration of approximately 1–10 days, with an incidence of 30–100% while the incidence of CRS grade ≥ 3 being approximately 10–30% [11]. CRS is mainly characterized by fever, hypotension, decreased pulse oxygen, and toxicity of various organs [32]. These symptoms are often difficult to distinguish from sepsis caused by bacterial infection, so we may inevitably overestimate the incidence of infection in the early phase. Patients with severe CRS often require admission to the ICU. Indwelling catheters (central venous catheters, urinary catheters, and tracheal catheters) in the ICU increase the risk of infection, and patients often require glucocorticoid and/or IL-6 receptor antagonist treatment, which may inhibit the ability of the patient’s immune system to respond effectively to pathogens [28]. A single-center study on rheumatoid arthritis showed that the use of IL-6 receptor antagonists was associated with an increase in infection [37]. However, Frigault et al. noted that IL-6 receptor antagonist use in patients after CAR-T treatment was not associated with the occurrence of infection [38]. Regarding the association between glucocorticoid use and infection risk, current findings have yielded conflicting results with some studies showing no increased infection risk [17] and others demonstrating an increased infection risk [39]. These contradictory results may be due to the selection bias of different studies as different research centers have different standards for the selection of research participants, definition of infection, and use of antibacterial drugs. In summary, owing to the comprehensive effects of many factors, severe CRS increases the risk of infection in patients.

Interestingly, we also found that Tregs were related to infections after CAR-T therapy, and the risk of infection increased with an increase in the peak value of the Treg proportion. Tregs are a subset of T cells with strong negative immunoregulatory functions that actively inhibit the activation, amplification, and function of other immune cells, thus regulating the intensity and duration of the immune response and maintaining immune homeostasis in vivo [40,41,42]. Sustained high expression of Tregs inhibits the activation and expansion of tumor antigen-specific effector T cells, affecting the curative effect of CAR-Ts, further aggravating the immune deficiency of patients, and increasing the risk of infection. Therefore, inhibiting the activation of Tregs, when necessary, may promote the tumor-killing effect of CAR-Ts and reduce the risk of infection. We intend to explore this intriguing aspect in our future clinical work.

After the remission of CAR-T therapy in our center, most children underwent hematopoietic stem cell transplantation within approximately 90 days, and long-term infection after CAR-T transfusion could not be tracked. Further research is required to generate robust data on etiology and immunology to address this limitation.

The incidence, pathogens, and severity of infection after CAR-T therapy are affected by many factors. Therefore, integrating the experience of CAR-T therapy for various hematological diseases is important to better understand related infectious complications and formulate the optimal infection management strategy to improve the safety and effectiveness of CAR-T therapy for children with R/R B-ALL.