Infection has emerged as the chief cause of non-relapse mortality (NRM) post CD19-targeting chimeric antigen receptor T-cell therapy (CAR-T) therapy. Even though up to 50% of patients may remain infection-free, many suffer multiple severe, life-threatening, or fatal infectious events. The primary aim of this study was to explore severe and life-threatening infections post licensed CAR-T therapy in large B-cell lymphoma, with a focus on the role of disease burden and disease sites in assessing individual risk. We sought to understand the cohort of patients who experience ≥2 infections and those at the highest risk of infectious NRM. Our analysis identifies a higher disease burden after bridging therapy as associated with infection events. Those developing ≥2 infections emerged as a uniquely high-risk cohort, particularly if the second (or beyond) infection occurred during an episode of immune effector cell-associated neurotoxicity syndrome (ICANS) or while on steroids and/or anakinra for ICANS. Herein, we also describe the first reported cases of “CAR-T cold sepsis,” a phenomenon characterized by the lack of an appreciable systemic inflammatory response at the time of detection of infection. We propose a risk-based strategy to encourage heightened clinician awareness of cold sepsis, with a view to reducing NRM.

INTRODUCTION

CD19-targeting chimeric antigen receptor T-cell therapy (CAR-T) has transformed the treatment of relapsed and/or refractory (r/r) large B-cell lymphoma (LBCL), leading to Food and Drug Administration and European Medicines Agency approval of axicabtagene ciloleucel (axi-cel), tisagenlecleucel (tisa-cel), and lisocabtagene maraleucel (liso-cel).^1-3

Published reports indicate that infection, which occurs in 23%–42% of patients within the first month,^4-6 is the chief cause of non-relapse mortality (NRM).^7-10 The greatest morbidity and mortality from infection is observed in the first-month post-CAR-T, but ongoing infection-driven NRM is reported up to 12 months post infusion.^{7, 10} Early infections (<30 days post-CAR-T) are predominantly bacterial, and later infections are primarily viral,^{5, 11} while fungal infections are relatively less commonly observed.¹²

Published data suggest that risk factors for infection post-CAR-T include high-grade cytokine release syndrome (CRS),^{5, 6} corticosteroids,^{13, 14} number of prior lines of therapy,⁴ recent (preadmission) infection,⁴ prior hematopoietic stem cell transplantation, Immunoglobulin G (IgG) < 400 mg/dL,¹⁵ and bone marrow burden.¹⁶ More recently, Rejeski et al. reported that a high HEMATOTOX (HT) score was associated with infection, suggesting that impaired neutrophil, hemoglobin, and/or platelet counts pre-CAR-T combined with raised inflammatory blood markers (ferritin, C-reactive protein [CRP]) could potentially predict patients at higher risk of post-CAR-T cytopenia and infectious complications.^{8, 17}

While the management of CAR-T immunotoxicity, such as CRS and immune effector cell-associated neurotoxicity syndrome (ICANS) continues to evolve, with earlier immunomodulatory interventions leading to better control and lower morbidity, infections have therefore emerged as the main clinical barrier to the safe delivery of CAR-T in the clinic. Establishing which patients are at the highest risk of severe and life-threatening infection is key to improving safety, particularly in older patients and those with comorbidities.

Here we report the infectious complications arising in a cohort of 178 adults with r/r LBCL following axi-cel or tisa-cel. The objectives of the study were to identify pre-CAR-T factors associated with the highest risk of severe infection and infectious NRM, to raise clinician awareness, and to support the development of targeted monitoring and prevention strategies that could benefit patients at the highest risk.

METHODS

Patients

Data were extracted retrospectively from electronic medical records for consecutively treated patients at two CAR-T centers (University College London Hospital [UCLH] and King's College Hospital [KCH]) between January 2019 and March 2021 as part of a hospital service evaluation. CAR-T product selection (axi-cel; Kite Pharma, or tisa-cel; Novartis) was at the discretion of the treating center. Lymphodepletion (LD) with fludarabine and cyclophosphamide was administered as per the manufacturers' instructions. Uniform UK treatment eligibility criteria were applied, as previously described.¹⁸

Definitions and outcomes

The primary outcome was severe or life-threatening infection following CAR-T infusion. All-cause mortality was a secondary outcome. Infections during LD or on the day of CAR-T infusion were excluded from the analysis.

Infection was defined as positive microbiology, virology, histopathology, and/or radiological findings in conjunction with clinical symptoms. Invasive fungal infections were classified as proven, probable, and possible according to revised EORTC criteria.¹⁹ Culture-negative neutropenic fever was excluded due to the potential overlap with CRS.

Grading was mild (no treatment or oral antibiotics), severe (requiring intravenous therapy), or life-threatening (symptoms such as hemodynamic instability or requiring organ support) as previously established.^{6, 20} Fatal infection was defined as infection directly leading or contributing significantly to death, as determined by the responsible clinical team.

The date of onset of infection was recorded as the date of the diagnostic test. Early and late infections were defined as those occurring up to (≤90 days) and beyond 90 days of CAR-T infusion, respectively. Documented clearance of infection followed by subsequent recurrence were considered separate events. Blood cultures positive for more than one organism on the same day were classed as separate events. Routine monitoring for viral reactivation (cytomegalovirus [CMV], Epstein–Barr virus, adenovirus, and human herpesvirus 6 [HHV-6]) was not performed at either center.

Severe neutropenia was defined as a neutrophil count of <0.5 × 10⁹/L.²¹ Granulocyte-colony stimulating factor (GCSF) was not administered prior to Day +21 post infusion in the inpatient setting. The use of GCSF beyond the initial inpatient stay was not recorded for either center. The frequency of clinical follow-up post discharge was in line with European Society for Blood and Marrow Transplantation (EBMT) guidance.²²

Bridging therapy (BT) was defined as any lymphoma therapy delivered between T-cell apheresis and LD. CRS and ICANS were graded according to American Society for Transplantation and Cellular Therapy (ASTCT) consensus guidelines²³ and toxicity management strategies were at the discretion of the treating center. Steroid cumulative dosing for CAR-T toxicity was calculated as methylprednisolone equivalent (mg). Response to CAR-T therapy was determined locally and defined as per Lugano classification 2014.²⁴

In the absence of metabolic tumor volume to quantify disease burden, a “high burden” after bridging therapy (BT) was defined a priori as lactate dehydrogenase (LDH) pre-LD > ULN (249 units/L) and (stage 3/4 disease and/or ≥3 extra-nodal sites). “Low burden” post-BT was defined as LDH < 250 units/L and (stage 1/2 disease and/or <3 extra-nodal sites).

Fitness for an autologous stem cell transplantation (ASCT) was determined by the treating clinician based on physical suitability (e.g., age, frailty, comorbidities), irrespective of disease status.

Infection prophylaxis and intravenous immunoglobulin (IVIG)

All patients received prophylaxis against herpes simplex/varicella zoster virus (HSV/VZV) and Pneumocystis jirovecii pneumonia (PJP) with oral acyclovir and either inhaled pentamidine, oral co-trimoxazole or azithromycin, respectively. The duration of HSV/PJP prophylaxis and antifungal practice differed between centers (Supporting Information S6: Table S1). Prophylactic antibiotics (such as fluroquinolones) were not administered at any stage in either center. IVIG was commenced post infusion with regional IVIG panel approval in those with an IgG level <4 g/L and recurrent confirmed bacterial infections.

Statistical analyses

All analyses were performed using Stata 17 (StataCorp). Patient baseline characteristics, toxicity profiles, and outcomes were compared between the two participating centers using chi-squared tests (or Fisher's exact test if any cell frequency ≤5) for categorical variables and the Kruskal–Wallis test for continuous variables. Patient time-at-risk was calculated in days from the date of CAR-T infusion and follow-up was censored at disease progression, death, or end of follow-up (to June 14, 2021). Survival analysis was performed using Cox regression. A final multivariable Cox regression model (including a priori potential confounders age, sex, center, and CAR-T drug) was fitted by carrying forward variables identified as associated with the outcome (α ≤ 0.1) in univariate analysis and performing stepwise backward and forward adjustments to identify and select those variables that were independently associated with the outcome, supported by consideration of causal pathways and assessment of collinearity. Exposures occurring as episodes of variable length during time-at-risk were included as time-varying effects. The proportional hazards assumption was assessed by inspection of plots of log(−log(survival)) and scaled Schoenfeld residuals against the log of survival time (with a test of non-zero slope for the residuals). Graphs and figures were created using Stata and GraphPad Prism v9.5.1 (GraphPad Software).

RESULTS

Patient characteristics

Across the cohort of 178 patients, the median age at the time of CAR-T approval was 58 years (range = 19–80 years) with a male predominance (61.8%, 110/178) (Supporting Information S6: Table S2). Approximately two-thirds (63.5%, 113/178) of patients had de novo diffuse large B-cell lymphoma (DLBCL), 28.1% (50/178) had transformed indolent disease, and 6.2% (11/178) had primary mediastinal B-cell lymphoma (PMBCL). One-fifth (20.2%, 36/178) had a prior ASCT, and approximately one-third (36.0%, 64/178) had received three or more lines of chemotherapy. Most patients (89.9%, 160/178) received BT (Supporting Information S6: Table S3), 56.9% (91/160) and 43.1% (69/160) achieved stable disease (SD)/progressive disease (PD) and complete remission (CR)/partial remission (PR), respectively. Disease characteristics pre/post-BT are summarized in Supporting Information S6: Table S2. Patients treated at the two participating centers had similar characteristics post-BT except for an Eastern Cooperative Oncology Group Performance Status (ECOG PS) score of 2 in 13.3% (11/83) of KCH patients compared with only 1/95 UCLH patients (p = 0.004). Axi-cel was given to 139/178 (78.1%) of patients and tisa-cel to 21.9% (39/178) (Supporting Information S6: Table S4).

Where data were available (Supporting Information S6: Table S5), baseline median CRP (n = 142) and ferritin (n = 149) were 7.5 mg/L (interquartile range [IQR] = 2.7–34.4) and 595 µg/L (IQR = 264–1329), respectively; 6.9% (12/175) of patients started LD with neutrophils <1.2 × 10⁹/L (median = 3.2, IQR = 2.0–4.5), potentially indicating marrow toxicity from prior therapies in a minority. Post infusion 23% (41/178) received GCSF at ≥Day 21 in the inpatient setting. GCSF use post discharge was not recorded for either center.

Incidence of total and severe/life-threatening infection after CAR-T infusion

A total of 191 separate infectious episodes were recorded in 89/169 patients (52.7%) (9/178 were omitted as infection preceded or occurred on the day of cell infusion). 145/191 (75.9%) infectious events occurred within 90 days post-CAR-T infusion, and 46/191 (24.1%) events occurred later than 90 days (Supporting Information S6: Table S6). There was a clear predominance of early bacterial and later respiratory viral infections (Figures S1 and S2).

A total of 92 severe or life-threatening infections were reported in 53/178 patients (29.8%), of whom 20/53 (37.7%) experienced more than 1 severe/life-threatening event. The onset of the first infectious event was at a median of 14 days post-CAR-T infusion (IQR = 4–25 days, range = 1–504 days). Details of the causative organisms, subtype of infection, and bacterial site of recovery are illustrated in Supporting Information S6: Table S7, Figures 1 and 2, respectively, with a clear predominance of bloodstream infections. The most common organisms identified in blood were Staphylococcus species (n = 10), Escherichia coli (n = 7), Pseudomonas aeruginosa (n = 6), and Enterococcus spp. (n = 7); Escherichia coli (n = 2) or Klebsiella spp. (n = 2) in urine. 19/69 (27.5%) viral infections were deemed severe/life-threatening, the most common of which was SARS CoV-2 (n = 6). CMV reactivation was noted in the peripheral blood of seven patients, of whom two received systemic antiviral therapy due to concerns regarding high CMV titers. Evidence of end-organ disease was not confirmed in either case. Our group has previously published our experience of progressive multifocal leukoencephalopathy (JC polyomavirus) and HHV-6 encephalitis after CAR-T therapy.^25-27 The illustrative timeline of infections is demonstrated in Figure 3. The remaining 125 patients were censored at death (n = 2), disease progression (n = 66), or end of active follow-up (n = 57).

Details are in the caption following the image — **Figure 1**
Open in figure viewer PowerPoint

Early and late severe/life-threatening infections post CAR-T-cell infusion (n = 53 patients, 92 infections).

Of those who had ≥2 severe/life-threatening infections (n = 20), a median of 2 infections were reported per patient (range = 2–6) with a median onset of the second infection at 38 days post infusion (IQR = 19–48, range = 1–567). A total of 59 severe/life-threatening events occurred in those who had ≥2 severe/life-threatening infections, again with a predominance of early bloodstream and later viral infections. 6/29 (20.7%) who had one bloodstream infection isolated another organism in their blood on the same day (co-infection) and 4/29 (13.8%) had another episode of bacteremia a median of 14 days after the initial event (range = 6–18 days) with a different organism in 75% (n = 3/4). The infectious NRM rate in this group was 25% (5/20) (Table 1). Acknowledging the small numbers, we observed that 3/5 (60%) of deaths in this cohort occurred during the initial inpatient stay in the context of high-grade ICANS.

Table 1. Characteristics of LBCL CAR-T patients by number of severe/life-threatening infections.

	No severe/life-threatening infection (n = 125)	1 severe/life-threatening infection (n = 33)	2+ severe/life-threatening infections (n = 20)	p value
Age (years)	58 (49–65)	62 (53–66)	60 (51.5–70)	0.369
Sex (male)	79 (63.2%)	18 (54.6%)	13 (65.0%)	0.629
Subtype
DLBCL	82 (65.6%)	23 (69.7%)	8 (40.0%)	0.175
tFL/tMZL	34 (27.2%)	7 (21.2%)	9 (45.0%)
PMBCL	6 (4.8%)	2 (6.1%)	3 (15.0%)
Other/unknown	3 (2.4%)	1 (3.0%)	0 (0.0%)
CAR-T product
Tisa-cel	31 (24.8%)	5 (15.2%)	3 (15.0%)	0.446
Axi-cel	94 (75.2%)	28 (84.9%)	17 (85.0%)
Number of previous lines
1/2	83 (66.4%)	21 (63.4%)	10 (50.0%)	0.365
3+	42 (33.6%)	12 (36.4%)	10 (50.0%)
Pre-BT
Stage pre-BT
1/2	29 (23.2%)	2 (6.1%)	3 (15.0%)	0.049
3	17 (13.6%)	9 (27.3%)	1 (5.0%)
4	79 (63.2%)	22 (66.7%)	16 (80.0%)
LDH (units/L)	279 (219–387), n = 116	322 (223–639), n = 31	411 (252–560), n = 20	0.061
Extra-nodal sites (≥3 vs. <3)	5 (4.0%)	6 (18.2%)	3 (15.0%)	0.008
Post-BT
Stage post-BT
0	12 (9.6%)	4 (12.1%)	3 (15.0%)	0.230
1/2	23 (18.4%)	2 (6.1%)	2 (10.0%)
3	15 (12.0%)	6 (18.2%)	0 (0.0%)
4	75 (60.0%)	21 (63.6%)	15 (75.0%)
Extra-nodal sites post-BT (≥3 vs. <3)	15 (12.0%)	5 (15.2%)	3 (15.0%)	0.726
Response to BT (no BT = SD/PD)
SD/PD	75 (60.0%)	19 (57.6%)	14 (75.0%)	0.398
CR/PR	50 (40.0%)	14 (42.4%)	5 (25.0%)
Post-BT disease burden (N = 161)a
Low	63 (57.3%)	13 (40.6%)	8 (42.1%)	0.163
High	47 (42.7%)	19 (59.4%)	11 (57.9%)
Full blood count pre-LD
Hemoglobin ×10⁹/L	110 (98–121), n = 122	106 (95–118)	101 (90–117)	0.258
Neutrophils ×10⁹/L	3.1 (2.0–4.0), n = 122	2.8 (1.9–4.2)	4.8 (2.9–11.9)	0.023
Lymphocytes ×10⁹/L	0.62 (0.34–0.93)	0.82 (0.55–1.51)	0.26 (0.22–0.46)	0.001
Platelets ×10⁹/L	185 (133–248)	208 (124–259)	190 (152–266)	0.836
Marker of disease or inflammation pre-LD
LDH pre-LD (units/L)	255 (192–347), n = 110	272 (216–432), n = 32	325 (238–604), n = 19	0.033
CRP pre-LD or infusion (mg/L)	6.8 (2.6–30.0), n = 96	4.7 (2.2–18.0), n = 30	42.7 (5.8–138.8), n = 16	0.033
Ferritin pre-LD or infusion (µg/L)	530 (193–1087), n = 101	600 (309–1507), n = 32	1368 (844–2822), n = 16	0.002
Max CRS grade
0	27 (21.6%)	10 (30.3%)	2 (10.0%)	0.383
1	42 (33.6%)	9 (27.3%)	4 (20.0%)
2	49 (39.2%)	12 (36.4%)	12 (60.0%)
3/4	7 (5.6%)	2 (6.1%)	2 (10.0%)
Max ICANS grade
0	88 (70.4%)	24 (72.7%)	8 (40.0%)	0.030
1/2	22 (17.6%)	4 (12.1%)	4 (20.0%)
3/4	15 (12.0%)	5 (15.2%)	8 (40.0%)
Severe neutropenia	106 (84.8%)	23 (69.7%)	17 (85.0%)	0.124
Steroids (any)	42 (33.6%)	16 (48.5%)	15 (75.0%)	0.001
Steroid dose (if any)	627 (160–2016), n = 42	760 (333–1197), n = 16	2245 (987–5970), n = 15	0.004
Infectious NRM	0 (0.0%)	2 (6.1%)	5 (25.0%)	<0.001

Abbreviations: BT, bridging therapy; CR/PR, complete/partial remission; CRP, C-reactive protein; CRS, cytokine release syndrome; DLBCL, diffuse large B-cell lymphoma; ICANS, immune effector cell-associated neurotoxicity syndrome; LD, lymphodepletion; LDH, lactate dehydrogenase; PMBCL, primary mediastinal B-cell lymphoma; SD/PD, stable/progressive disease; tFL/tMZL, transformed follicular/marginal zone lymphoma.
^a High burden post-BT = LDH > ULN (249 units/L) AND (stage 3/4 and/or ≥3 extra-nodal sites); Low burden post-BT = LDH < 250 units/L AND (stage 1/2 disease and/or <3 extra-nodal sites).

Risk factors for severe or life-threatening infection after CAR-T infusion

In univariate analyses, pre-BT factors associated with an increased risk of severe or life-threatening infection after infusion included stage 3/4 (vs. 1/2) disease, ≥3 extra-nodal sites, and LDH. Significant post-BT factors included ferritin, LDH, ECOG PS, and “high burden” pre-LD (Supporting Information S6: Table S8; and Figure S3). Toxicity management strategies, such as the use of tocilizumab, anakinra, and steroids, and ≥grade 3 ICANS were also associated with increased risk in univariate analyses. In multivariable models, an independent signal was observed between a high burden of disease post-BT and an increased risk of severe/life-threatening infection (Supporting Information S6: Table S9A). The impact of EN sites on infection risk was weakened after bridging therapy and our “high burden” variable reflected mainly disease stage 3–4 and LDH pre-LD (Table 9B). Patients with high burden were almost twice as likely as low burden to experience severe/life-threatening infection (hazard ratio [HR] = 1.72; 95% confidence interval [CI] = 0.97–3.04; p = 0.064). Neither the use of BT nor the response to BT had a clear impact on infection risk. However, those achieving CR/PR versus SD/PD with BT were less likely to suffer ≥2 infection events (25% vs. 75%) (Table 1).

CAR-T immunotoxicity (CRS/ICANS) was evaluated and an association between CRS and infection was not observed. An overall effect of ICANS was evident (p = 0.056), but patients with ≥grade 3 ICANS had almost threefold higher odds of severe or life-threatening infection in multivariate analyses (HR = 3.67; 95% CI = 1.42–9.51; p = 0.007) compared with patients without ICANS. The use of tocilizumab, steroids, and anakinra was collinear with ICANS and did not retain significance in the multivariable model. In our data set, neither baseline neutrophil count pre-LD, neutrophils below the lower limit of normal (<1.2 × 10⁹/L) pre-LD nor severe neutropenia (<0.5 × 10⁹/L) were associated with an increased risk of infection (Supporting Information S6: Table S8). Using time-varying analysis, considering the effect of an episode of severe neutropenia (of variable duration) during the patient's time at risk, that is, after CAR-T infusion and before the onset of infection or censoring event, no overall time-varying effect of severe neutropenia on the risk of severe/life-threatening infection in our cohort was found. Patients without ICANS had severe neutropenia (<0.5 × 10⁹/L) for a median of 6 (IQR = 0–11) days compared with 9 (6–23.5) days for grades 1/2 and 7.5 (5–21) days for ≥grade 3 (p = 0.012). Among patients who developed ICANS (any grade), we found no effect of severe neutropenia on infection risk (HR = 0.88; 95% CI = 0.29–2.67; p = 0.82) and no effects if stratified by grade of ICANS.

Further exploration of severe or life-threatening infection by early (n = 48) versus late (n = 5) onset or by subtype of infection (bacterial n = 61, viral n = 19, fungal n = 13) was limited by small numbers.

Higher disease burden and infection risk

Patient demographics for those with a low and high burden of disease pre-LD were well matched in terms of age, number of prior lines of treatment, and CAR-T product choice (Supporting Information S6: Table S10). Those with a higher burden had higher baseline markers of inflammation (CRP and ferritin) and a lower baseline hemoglobin level. This cohort was also noted to have relatively higher baseline neutrophil counts and comparable lymphocyte counts pre-LD, suggesting that a prior chemotherapy-related immunological deficit was not contributing to infection risk. The risk of a severe/life-threatening event was higher for those with high versus low disease burden (39% vs. 25%), as was the risk of ≥2 events (14% vs. 10%); 100% (4/4) of early infectious NRM (<90 days) occurred in those with high-burden disease and 75% (3/4) who had ≥2 severe/life-threatening events (Table 1, and Supporting Information S6: Table S10). There were no clear differences in CAR-T immunotoxicity (CRS and ICANS), neutropenia, and steroid use between low and high-burden patients.

Looking specifically at those who developed ≥2 severe/life-threatening infectious events (n = 20), these patients were a particularly high-risk group with an infectious NRM rate of 25% (5/20) (Table 1). Pre-BT, patients in this high-risk cohort were more likely to have a higher stage of disease and higher LDH. Prior to LD, these patients also had a higher LDH and higher markers of inflammation (CRP and ferritin). Although these patients had comparable neutrophil counts, a lower lymphocyte count was noted, potentially suggestive of poor recovery post-BT. Post-infusion, this cohort was more likely to develop high-grade ICANS and require steroids and higher doses thereof. The second isolate was detected during an episode of ICANS or ongoing steroid use for ICANS in 37% of cases (7/20) with an associated mortality of 43% (3/7).

Extra-nodal sites and risk of infection

We observed a pattern of infection that appeared to correlate with commensals at the EN site or infection of the EN site itself. A total of 122 patients (68.5%) had lymphomatous involvement of ≥1 EN site at the time of CAR-T approval, of whom 39 (32%) acquired a severe/life-threatening infection. Of those with extra-nodal gastrointestinal (GI) tract involvement pre and/or post-BT (gastric, mesenteric, and small/large intestinal wall) and a severe/life-threatening event (n = 14), a GI commensal organism was isolated in blood in eight patients (57%) with ≥2 GI commensals detected in four patients (29%). Likewise, of those with extra-nodal lung parenchymal involvement pre and/or post-BT and a severe/life-threatening event (n = 7), 86% (6/7) developed a severe/life-threatening event involving the lung tissue during follow-up with 29% (2/7) suffering ≥2 lung-related events (Figures S4A and S4B).

Although GI organisms can commonly complicate the neutropenic posttreatment phase and lung tissue is a high-risk site for infection, it is possible that abnormal anatomy at EN sites, combined with T-cell-driven tumor lysis and necrosis may facilitate the breach of immunological defenses and entry of commensal organisms at these sites.

Infectious NRM and the risk of “cold sepsis”

There were 66 deaths among the 178 patients, with a 12-month overall survival of 68% from the date of CAR-T infusion (Supporting Information S6: Table S11). Of 66 deaths, the vast majority were due to progressive disease (PD), but a total of seven were directly due to infection (Supporting Information S6: Table S12) and occurred at a median of 46 days post infusion (range = 15–257). Overall, 3.9% of 178 infused patients died post-CAR-T from infectious complications before progression or end of follow-up. In 2/7 cases, death was due to SARS CoV-2 complications in two tisa-cel recipients. In 5/7 cases, death from infection occurred in axi-cel recipients with active (n = 4) or resolved (n = 1) high-grade ICANS treated with corticosteroids (median total methylprednisolone dose of 1867 mg by time of death, range = 213–7133 mg). Of those with high-burden disease pre-LD who subsequently developed high-grade ICANS (n = 15, 19.5%), mortality from infection was 27% (n = 4). This cohort accounted for 100% of early NRM deaths (<90 days). All four patients had active high-grade ICANS and were still receiving steroids at the time of death, 3/4 were receiving anakinra and 3/4 did not mount a fever during the fatal infectious event. Two patients (29%) had no neutrophil recovery at the time of death. Figure 4 illustrates the clinical trajectory of our three cases of cold sepsis, demonstrating the issue of normothermia with no CRP response in this high-risk group. No early infectious deaths occurred in low-burden patients.

Consideration of antibacterial prophylaxis in a selected subgroup

None of our patients received a prophylactic antibacterial agent. In our data set, 22/28 (79%) Gram-negative organisms isolated from 20 patients with a severe/life-threatening infection during the CAR-T inpatient stay were deemed sensitive to a fluoroquinolone (ciprofloxacin), including in two cases of early infectious NRM. Hence, this agent may have conferred benefits in select cases (data not shown). In our data set, those with higher disease burden were at higher risk of severe/life-threatening infection. However, disease burden alone may not warrant anti-bacterial prophylaxis. The subset of high-burden patients who develop high-grade ICANS, requiring ongoing steroids/interleukin blockade with/without neutrophil and/or neurological recovery, should be considered for antibacterial prophylaxis. This subset of patients is at high risk of sepsis, including cold sepsis and early infectious NRM. Lymphomatous involvement of specific EN sites (e.g., GI tract) may confer an additional risk of infection with commensals. We propose a risk-based strategy to guide prophylactic antibiotic use and encourage heightened clinician awareness of cold sepsis, with a view to reducing early NRM (Figure 5).

Use of IVIG after CAR-T therapy

Of 172 patients alive beyond 30 days, 17/172 (9.9%) received IVIG, commenced at a median of 3 months postinfusion (range = 1–21), and continued until death or end of follow-up in 88.2% (n = 15/17).

Median age was 55 years (range = 19–70) and the majority (82.4%; 14/17) had received axi-cel. In 13/17 (76%) patients, IVIG was commenced for recurrent sinopulmonary infections (Supporting Information S6: Table S13) and 46% (6/13) of these patients had sinopulmonary EN involvement on a PET scan prior to LD. Of those with sinopulmonary involvement at the point of submission for CAR-T (n = 22), 23% (n = 5) ultimately required IVIG post-CAR-T.

Only one episode of infectious NRM occurred in this cohort (SARS CoV2). Infection was a cause or significant contributor to death in the context of progressive disease in two other cases.

Data on T-cell reconstitution post-CAR-T therapy was available in only a subset of patients but demonstrated incomplete T-cell recovery in a significant proportion at 12 months (36%). Small numbers limited further analysis of these patients (Supporting Information S6: Table S14).

Infection in the context of progressive disease post-CAR-T therapy

With a median follow-up of 291 days (IQR = 175–484 days, range = 4–886 days), median progression-free survival from infusion for the entire cohort (n = 178) and those alive beyond 30 days was 251 days (IQR = 97–432 days, range = 4–886 days) and 261 days (IQR = 133–460, range = 33–886), respectively. Of the 83 patients who progressed after CAR-T, 65% (n = 54) commenced further therapy. A total of 17 patients (20.5%) developed a severe or life-threatening infectious event after CAR-T relapse. In 9 of 17 patients (53%), infection contributed significantly or directly led to death (Supporting Information S6: Table S15).

DISCUSSION

A myriad of host and treatment-related factors contribute to infection risk post CAR-T infusion.^{7, 12, 28-30} These factors are likely to have a cumulative effect. The incidence and distribution of all grade (52.7%) and severe/life-threatening (29.8%) infections in our cohort is similar to other published series in this patient group.^{8, 14} Acknowledging an exclusively LBCL population, we did not find a signal for variables previously reported by others, such as high-grade CRS^{5, 6} or a number of prior lines of therapy.⁴ Akin to other groups, we have demonstrated that impaired ECOG PS, steroids, markers of inflammation such as ferritin, and high-grade ICANS are associated with infection.^{8, 14, 31} However, our multivariate analysis identifies disease burden as independently associated with severe or life-threatening events in our cohort. Rejeski et al. retrospectively stratified infection risk within 90 days by the HT score, potentially identifying a subset that may benefit from antibacterial prophylaxis.⁸ Our study builds on this work in a large cohort of LBCL patients receiving standard-of-care CAR-T products, with a focus on the role of lymphoma disease burden and site of disease in helping to identify those with the highest mortality risk from infection.

Active malignancy is known to predispose to infection via a number of mechanisms. Disease alone, even in untreated patients, is thought to blunt innate and adaptive immune responses.^{32, 33} Cancer-associated cachexia with associated inflammation and malnutrition is a catabolic state that can compromise host defenses, such as tissue healing, mucosal integrity, secretory IgA function, complement pathways, and cellular immunity.^34-36 Tumor infiltration and obstruction may disrupt the integrity of anatomical barriers, such as the skin and mucosal surfaces. Pulmonary tissue immunology may be particularly vulnerable to local tumor invasion leading to impaired muco-ciliary clearance and postobstructive atelectasis. Likewise with gut, oropharyngeal, and skin infiltration by cancerous tissue, local or systemic infection with tissue-specific resident flora is a recognized phenomenon.^37-39 Acknowledging small numbers, our data raise the possibility that EN involvement, particularly of the GI and sinopulmonary tract, may increase susceptibility to infection, often with resident commensal organisms. Awareness of this phenomenon may allow for a higher index of suspicion for potential offending pathogens and tailored antibiotic use. For instance, a longer duration of antibiotics may be considered for certain patients, particularly those with EN GI tract involvement in the context of neutropenia.

The potential correlation of EN site with site-specific infection is also supported by our data on IVIG use post infusion. Although both congenital^{40, 41} and treatment-related^42-44 hypogammaglobulinemia are known to manifest with sinopulmonary complications, our data suggest that patients with EN sinopulmonary involvement prior to CAR-T infusion are at higher risk of this clinical phenotype. Acknowledging the absolute numbers for immunoglobulin replacement were small (n = 17, 9.9%), 46% of those requiring IVIG for sinopulmonary infection post CAR-T (6/13) had radiological evidence of sinopulmonary EN disease prior to LD. This supports our theory that anatomical disruption at high-risk sites in susceptible individuals, such as the sinopulmonary tract, may contribute to infection risk. This encourages vigilance in select patients and supports early instigation of IVIG replacement to prevent later infections and further sinopulmonary damage in this cohort.

Higher disease burden and markers of inflammation are known to correlate with higher grades of CAR-T immunotoxicity in similar real-world cohorts^{45, 46} and other disease indications.⁴⁷ Therefore, it could easily be inferred that the infection risk in high-burden patients may be attributable to the immunotoxicity itself or the employment of steroids and interleukin blockade to mitigate high-grade toxicity. However, in our data set, the risk of a severe/life-threatening event was substantially higher for those with high versus low disease burden (39% vs. 25%) and there were no clear statistically significant differences in CRS or ICANS grade or steroid/tocilizumab use between those with low and high burden, especially when considering those who developed any severe or life-threatening infectious event. The use of tocilizumab, anakinra, and steroids was significant for infection on univariate analysis only. On multivariate analysis, their use was collinear with ICANS and did not retain significance. Our data suggests that immunotoxicity and agents to control immunotoxicity were not the main drivers of severe/life-threatening infection, particularly when considering those who experienced a single event. However, those developing ≥2 severe or life-threatening infection events emerged as a uniquely high-risk cohort, with significantly higher rates of high-grade ICANS, steroid use, and infectious NRM (25%). The occurrence of a second (or beyond) severe/life-threatening event during an episode of ICANS or while on steroids ± anakinra for ICANS led to 43% mortality.

Therefore, we propose that disease burden and inflammation increase baseline infection risk. Additionally, we propose that the emergence of ICANS with the associated immunosuppressive treatments in high-burden patients selects out those at the highest risk of repeated infections and infectious NRM. All early infectious deaths (<90 days) (n = 4) occurred in those with high burden who subsequently developed high-grade ICANS. The incidence of infectious NRM in our data set is comparable with other groups.^{7, 8, 48} We also report the first cases of “cold sepsis” after CAR-T therapy, namely the lack of an appreciable systemic inflammatory response, such as fever and raised CRP, at the time of detection of infection. The phenomenon of “cold sepsis” has been described in other patient groups and is historically associated with a higher mortality.^49-53 While the pathophysiology is debated,^{54, 55} the risk of a delayed diagnosis of sepsis is undoubtedly higher when the conventional clinical and laboratory flags for infection, namely fever and raised CRP, are absent. In the context of CAR-T therapy, the use of steroids and/or interleukin blockade may suppress a hyperthermic or inflammatory response in patients, many of whom may have incomplete neutrophil and/or neurological recovery. We highlight three cases of cold sepsis, occurring with or without neutrophil recovery in the context of ongoing steroid and/or anakinra use for high-grade ICANS. The absence of clinical or biochemical signs of infection supports the need for heightened clinical awareness, enhanced surveillance, and education of ward-based staff. It is possible that a risk-based mitigation strategy for cold sepsis incorporating biomarkers may permit risk stratification, earlier detection, and even pre-emptive treatment with early antibiotics.^56-58 Such a tool would require robust, validated biomarkers for infection and prospective validation in a clinical study. Pending the above, we propose a clinical risk-based strategy for cold sepsis and antibiotic use for patients receiving CAR-T therapy (Figure 5).

The role of antibacterial prophylaxis post-CAR-T infusion is unclear. Rejeski et al. retrospectively stratified infection risk within 90 days by the HT score, potentially identifying a subset that may benefit from antibacterial prophylaxis.⁸ IDSA guidance supports fluoroquinolones for those at high risk of febrile or prolonged neutropenia.⁵⁹ Its routine use to cover the neutropenic period post-LD and an infusion is not recommended by EBMT/European Haematology Association but can be considered at a local level in the context of prolonged neutropenia.²² Updated best practice recommendations suggest prophylaxis in select cases considering immune effector cell-associated hematotoxicity grading.⁶⁰ Based on our experience with this patient cohort, we propose that patients with “high burden” post-BT who subsequently develop high-grade ICANS, requiring ongoing steroids/interleukin blockade with/without neutrophil recovery, particularly those with ≥1 prior severe/life-threatening event, should be considered for prophylactic antibiotics to mitigate the risk of cold sepsis and early NRM. Consideration can also be given to those with EN involvement at specific sites (e.g., GI tract). However, the use of ciprofloxacin prophylaxis has to be balanced against the risk (both for the individual patient and the wider healthcare setting) of promoting the emergence of quinolone-resistant Gram-negative bacterial infection. Additional risks to be considered are breakthrough infections with intrinsically less susceptible organisms such as enterococci, as well as antibiotic-related adverse events including Clostridium difficile infection.^{59, 61, 62}

There is no consensus on the early or prophylactic use of myeloid growth factors to facilitate count recovery post CAR-T therapy and prospective clinical studies are required to establish the safety and optimal use of these agents. Historically, caution has been advised given the theoretical risk of exacerbating CRS and ICANS.^{63, 64} However, some have demonstrated the safety of this approach with a reduced incidence of febrile neutropenia in the absence of increased immunotoxicity risk.⁶⁵ A retrospective study by Miller et al. (n = 197) demonstrated that prophylactic GCSF led to faster neutrophil recovery and comparable treatment outcomes without increasing the risk of high-grade ICANS. Higher rates of ≥grade 2 CRS were reported in this study.⁶⁶

SARS CoV-2 was the most common severe or life-threatening viral infection in our cohort with two late fatalities. The reported incidence of severe disease and mortality in CAR-T patients has fallen, likely due to a combination of less virulent variants, vaccination programs, and pre-emptive SARS-CoV-2-directed pharmacotherapies.^67-69 Vaccine-induced T-cell responses may provide some protection in those with ongoing B-cell aplasia and impaired antibody responses.⁷⁰ However, despite advances, morbidity and mortality risk in this vulnerable cohort persists. Annual vaccination of CAR-T recipients and close contacts against Influenza (inactivated) is also recommended to reduce the risk of severe disease and hospitalization.⁶⁴ Reactivation of latent viruses such as CMV has been reported, with the incidence ranging from 21% to 44%.^{71, 72} CMV reactivation was noted in seven of our patients. However, the probability of progression to end-organ disease in this patient group remains unknown. In the absence of systematic viral polymerase chain reaction monitoring in most centers, Marquez-Algaba et al. propose a risk-based surveillance strategy for CMV reactivation.⁷²

This retrospective real-world study has limitations. In the absence of the ready availability of metabolic tumor volume, low and high disease burdens after bridging therapy were categorized using LDH, stage, and number of extra-nodal sites. The authors accept that this is not a universally accepted definition and larger data sets will be required to validate our findings. It is also possible that some severe/life-threatening infections, particularly post relapse, may not have been communicated to our teams. Skin commensals were included as potential pathogens if the clinical team treated them as such; some of these isolates may represent contaminants. Our data set did not include information on recent infections (including during BT), neutrophil recovery patterns, or HT scores (due to missing data points). Such variables are recognized risk factors for infection post-CAR-T infusion^{4, 8} and the authors acknowledge this as a limitation of our work.

With rapidly expanding CAR-T use, a clear understanding of the core factors driving CAR-T toxicity and the projected burden on hospital resources has become increasingly important. Outpatient or ambulatory care delivery is an attractive option for the substantial proportion of patients who are at lower risk of high-grade complications.⁷³ Proposed CRS, ICANS, and cytopenia risk scores have the potential to identify higher-risk LBCL candidates in the real world^{17, 74, 75} with continued improvement in postinfusion care and burden of morbidity. However, infection, the chief cause of NRM, has emerged as the biggest clinical challenge, leading to prolonged hospital admissions, intensive care unit involvement, and significant morbidity. The risk of infection post infusion is highly variable, with up to 50% of patients remaining infection-free^{8, 28, 31} while others may suffer multiple severe, life-threatening, or fatal infectious events. Knowledge of predisposing factors could allow a more nuanced patient-specific CAR-T management pathway, with targeted interventions in acute and longer-term settings.

In summary, disease burden is a key risk factor for infectious complications post CAR-T therapy. The site of the disease may have implications for offending pathogens and targeted antibiotic use. The requirement for IVIG in our data set was low but concentrated in those with a history of sinopulmonary EN involvement. Better bridging strategies aimed at achieving disease control pre-infusion should be explored, with the potential added benefit of reducing the risk of high-grade toxicity. Cold sepsis represents a significant mortality risk to patients receiving steroids and/or interleukin blockade for immunotoxicity post CAR-T and additional measures are required to maintain safety in this cohort. Prospective validation of pre-emptive strategies, such as the use of antimicrobial prophylaxis, early GCSF, and the development of novel infection biomarker-based screening approaches should be encouraged in the real world to support their use.

AUTHOR CONTRIBUTIONS

Maeve A. O'Reilly, Lorna Neill, Claire Roddie, and Simon M. Collin led the concept and design of this document and prepared the manuscript. Simon M. Collin also analyzed the data. Deborah Springell, Jeremy Mensah, Kathleen P. L. Cheok, and Katarzyna Jalowiec contributed to data collection. All authors contributed to the preparation and approval of the final submitted version.

CONFLICT OF INTEREST STATEMENT

M. O. R.: Honoraria Kite Gilead, Novartis, and Janssen, Advisory boards Kite Gilead and Autolus, conference support Kite Gilead and Novartis. S. M. C.: Employee of AstraZeneca Ltd. N. S.: Speaker fees Kite Gilead, Pfizer and Shionogi, research grant Kite Gilead. K. P. L. C.: Honoraria Kite Gilead. A. K.: Kite Gilead conference support, honoraria, advisory board, Novartis: honoraria, research funding, Advisory board: Roche, Abbvie, BMS. C. R.: Advisory boards and speakers fees Novartis, Kite Gilead, BMS, Amgen, Autolus. R. S.: Kite Gilead speakers bureau, honoraria, conference travel, Novartis speakers bureau, honoraria, conference travel. The other authors declare no conflict of interest.

FUNDING

This research received no funding.

Open Research

DATA AVAILABILITY STATEMENT

The data that support the findings of this study are available from the corresponding author upon reasonable request.

Supporting Information

REFERENCES

1Neelapu SS, Locke FL, Bartlett NL, et al. Axicabtagene ciloleucel CAR T-cell therapy in refractory large B-cell lymphoma. N Engl J Med. 2017; 377(26): 2531-2544.
10.1056/NEJMoa1707447
CAS PubMed Web of Science® Google Scholar
2Schuster SJ, Bishop MR, Tam CS, et al. Tisagenlecleucel in adult relapsed or refractory diffuse large B-cell lymphoma. N Engl J Med. 2019; 380(1): 45-56.
10.1056/NEJMoa1804980
CAS PubMed Web of Science® Google Scholar
3Abramson JS, Palomba ML, Gordon LI, et al. Lisocabtagene maraleucel for patients with relapsed or refractory large B-cell lymphomas (TRANSCEND NHL 001): a multicentre seamless design study. Lancet. 2020; 396(10254): 839-852.
10.1016/S0140-6736(20)31366-0
PubMed Web of Science® Google Scholar
4Mikkilineni L, Yates B, Steinberg SM, et al. Infectious complications of CAR T-cell therapy across novel antigen targets in the first 30 days. Blood Adv. 2021; 5(23): 5312-5322.
10.1182/bloodadvances.2021004896
CAS PubMed Web of Science® Google Scholar
5Park JH, Romero FA, Taur Y, et al. Cytokine release syndrome grade as a predictive marker for infections in patients with relapsed or refractory B-cell acute lymphoblastic leukemia treated with chimeric antigen receptor T cells. Clin Infect Dis. 2018; 67(4): 533-540.
10.1093/cid/ciy152
CAS PubMed Web of Science® Google Scholar
6Hill JA, Li D, Hay KA, et al. Infectious complications of CD19-targeted chimeric antigen receptor-modified T-cell immunotherapy. Blood. 2018; 131(1): 121-130.
10.1182/blood-2017-07-793760
CAS PubMed Web of Science® Google Scholar
7Bethge WA, Martus P, Schmitt M, et al. GLA/DRST real-world outcome analysis of CAR-T cell therapies for large B-cell lymphoma in Germany. Blood. 2022; 140(4): 349-358.
CAS PubMed Web of Science® Google Scholar
8Rejeski K, Perez A, Iacoboni G, et al. The CAR-HEMATOTOX risk-stratifies patients for severe infections and disease progression after CD19 CAR-T in R/R LBCL. J Immunother Cancer. 2022; 10(5):e004475. doi:10.1136/jitc-2021-004475
10.1136/jitc-2021-004475
PubMed Web of Science® Google Scholar
9Jacobson CA, Locke FL, Ma L, et al. Real-world evidence of axicabtagene ciloleucel for the treatment of large B-cell lymphoma in the United States. Transplant Cell Ther. 2022; 28: 581.e1-581.e8. doi:10.1016/j.jtct.2022.05.026
10.1016/j.jtct.2022.05.026
CAS PubMed Google Scholar
10Lemoine J, Bachy E, Cartron G, et al. Non-relapse mortality after CAR T-cell therapy for large B-cell lymphoma: a LYSA study from the DESCAR-T registry. Blood Adv. 2023; 7: 6589-6598. doi:10.1182/bloodadvances.2023010624
10.1182/bloodadvances.2023010624
CAS PubMed Web of Science® Google Scholar
11Cordeiro A, Bezerra ED, Hirayama AV, et al. Late events after treatment with CD19-targeted chimeric antigen receptor modified T cells. Biol Blood Marrow Transplant. 2020; 26(1): 26-33.
10.1016/j.bbmt.2019.08.003
CAS PubMed Web of Science® Google Scholar
12Little JS, Aleissa MM, Beluch K, et al. Low incidence of invasive fungal disease following CD19 chimeric antigen receptor T-cell therapy for non-Hodgkin lymphoma. Blood Adv. 2022; 6(16): 4821-4830.
10.1182/bloodadvances.2022007474
CAS PubMed Web of Science® Google Scholar
13Baird JH, Epstein DJ, Tamaresis JS, et al. Immune reconstitution and infectious complications following axicabtagene ciloleucel therapy for large B-cell lymphoma. Blood Adv. 2021; 5(1): 143-155.
10.1182/bloodadvances.2020002732
PubMed Web of Science® Google Scholar
14Wudhikarn K, Palomba ML, Pennisi M, et al. Infection during the first year in patients treated with CD19 CAR T cells for diffuse large B cell lymphoma. Blood Cancer J. 2020; 10(8): 79.
10.1038/s41408-020-00346-7
PubMed Web of Science® Google Scholar
15Vora SB, Waghmare A, Englund JA, et al. Infectious Complications Following Cd19 Chimeric Antigen Receptor T-cell Therapy for Children, Adolescents, and Young Adults. Vol 7. Oxford University Press; 2020:ofaa121.
Google Scholar
16Maron GM, Hijano DR, Epperly R, et al. Infectious complications in pediatric, adolescent and young adult patients undergoing CD19-CAR T cell therapy. Front Oncol. 2022; 12:845540.
10.3389/fonc.2022.845540
CAS PubMed Web of Science® Google Scholar
17Rejeski K, Perez A, Sesques P, et al. CAR-HEMATOTOX: a model for CAR T-cell–related hematologic toxicity in relapsed/refractory large B-cell lymphoma. Blood. 2021; 138(24): 2499-2513.
10.1182/blood.2020010543
CAS PubMed Web of Science® Google Scholar
18Kuhnl A, Roddie C, Kirkwood AA, et al. A national service for delivering CD19 CAR-T in large B-cell lymphoma—the UK real-world experience. Br J Haematol. 2022; 198(3): 492-502.
10.1111/bjh.18209
CAS PubMed Web of Science® Google Scholar
19De Pauw B, Walsh TJ, Donnelly JP, et al. Revised definitions of invasive fungal disease from the European organization for research and treatment of cancer/invasive fungal infections cooperative group and the national institute of allergy and infectious diseases mycoses study group (EORTC/MSG) consensus group. Clin Infect Dis. 2008; 46(12): 1813-1821.
10.1086/588660
PubMed Web of Science® Google Scholar
20Young JAH, Logan BR, Wu J, et al. Infections after transplantation of bone marrow or peripheral blood stem cells from unrelated donors. Biol Blood Marrow Transplant. 2016; 22(2): 359-370.
10.1016/j.bbmt.2015.09.013
PubMed Web of Science® Google Scholar
21Taplitz RA, Kennedy EB, Bow EJ, et al. Antimicrobial prophylaxis for adult patients with cancer-related immunosuppression: ASCO and IDSA clinical practice guideline update. J Clin Oncol. 2018; 36(30): 3043-3054.
10.1200/JCO.18.00374
PubMed Web of Science® Google Scholar
22Hayden PJ, Roddie C, Bader P, et al. Management of adults and children receiving CAR T-cell therapy: 2021 best practice recommendations of the European Society for Blood and Marrow Transplantation (EBMT) and the Joint Accreditation Committee of ISCT and EBMT (JACIE) and the European Haematology Association (EHA). Ann Oncol. 2022; 33(3): 259-275.
10.1016/j.annonc.2021.12.003
CAS PubMed Web of Science® Google Scholar
23Lee DW, Santomasso BD, Locke FL, et al. ASTCT consensus grading for cytokine release syndrome and neurologic toxicity associated with immune effector cells. Biol Blood Marrow Transplant. 2019; 25(4): 625-638.
10.1016/j.bbmt.2018.12.758
CAS PubMed Web of Science® Google Scholar
24Cheson BD, Fisher RI, Barrington SF, et al. Recommendations for initial evaluation, staging, and response assessment of Hodgkin and non-Hodgkin lymphoma: the Lugano classification. J Clin Oncol. 2014; 32(27): 3059-3067.
10.1200/JCO.2013.54.8800
PubMed Web of Science® Google Scholar
25Shah M, Kuhnl A, Shields G, et al. Human herpesvirus 6 encephalitis following axicabtagene ciloleucel treatment for refractory diffuse large B cell lymphoma. HemaSphere. 2021; 5:e535. doi:10.1097/HS9.0000000000000535
10.1097/HS9.0000000000000535
PubMed Web of Science® Google Scholar
26Mackenzie S, Laurence A, O'Reilly M, Peggs KS, Roddie C. Progressive multifocal leukoencephalopathy in the era of chimeric antigen receptor T-cell therapy. Lancet Haematol. 2021; 8(12): e870-e873.
10.1016/S2352-3026(21)00316-1
CAS PubMed Web of Science® Google Scholar
27Mackenzie S, Shafat M, Roddy H, et al. Pembrolizumab for the treatment of progressive multifocal leukoencephalopathy following anti-CD19 CAR-T therapy: a case report. EJHaem. 2021; 2(4): 848-853.
10.1002/jha2.274
PubMed Google Scholar
28Jennifer M. Logue L, Elisa Zucchetti Z, Christina A. Bachmeier B, et al. Immune reconstitution and associated infections following axicabtagene ciloleucel in relapsed or refractory large B-cell lymphoma. Haematologica. 2021; 106(4): 978-986.
10.3324/haematol.2019.238634
PubMed Web of Science® Google Scholar
29Gudiol C, Lewis RE, Strati P, Kontoyiannis DP. Chimeric antigen receptor T-cell therapy for the treatment of lymphoid malignancies: is there an excess risk for infection? Lancet Haematol. 2021; 8(3): e216-e228.
10.1016/S2352-3026(20)30376-8
CAS PubMed Web of Science® Google Scholar
30Wudhikarn K, Perales MA. Infectious complications, immune reconstitution, and infection prophylaxis after CD19 chimeric antigen receptor T-cell therapy. Bone Marrow Transplant. 2022; 57: 1477-1488.
10.1038/s41409-022-01756-w
CAS PubMed Web of Science® Google Scholar
31Beyar-Katz O, Kikozashvili N, Bar On Y, et al. Characteristics and recognition of early infections in patients treated with commercial anti-CD19 CAR-T cells. Eur J Haematol. 2022; 108(1): 52-60.
10.1111/ejh.13712
CAS PubMed Web of Science® Google Scholar
32Kuderer NM, Choueiri TK, Shah DP, et al. Clinical impact of COVID-19 on patients with cancer (CCC19): a cohort study. Lancet. 2020; 395(10241): 1907-1918.
10.1016/S0140-6736(20)31187-9
CAS PubMed Web of Science® Google Scholar
33Hiam-Galvez KJ, Allen BM, Spitzer MH. Systemic immunity in cancer. Nat Rev Cancer. 2021; 21(6): 345-359.
10.1038/s41568-021-00347-z
CAS PubMed Web of Science® Google Scholar
34Kumar Chandra R. Nutrition, immunity, and infection: present knowledge and future directions. Lancet. 1983; 321: 688-691. doi:10.1016/s0140-6736(83)91980-3
10.1016/S0140-6736(83)91980-3
Google Scholar
35Donnelly JP, Blijlevens N, van der Velden WJ. Host impairments in patients with neoplastic diseases. Infect Complicat Cancer Patients. 2014; 161: 1-41.
10.1007/978-3-319-04220-6_1
Google Scholar
36Salimonu LS, Ojo-Amaize E, Williams AIO, et al. Depressed natural killer cell activity in children with protein-calorie malnutrition. Clin Immunol Immunopathol. 1982; 24(1): 1-7.
10.1016/0090-1229(82)90082-4
CAS PubMed Google Scholar
37Zembower TR. Epidemiology of infections in cancer patients. Infect Complicat Cancer Patients. 2014; 161: 43-89.
10.1007/978-3-319-04220-6_2
Google Scholar
38Axelrod PI. Infections complicating mycosis fungoides and Sézary syndrome. J Am Med Assoc. 1992; 267(10): 1354-1358.
10.1001/jama.1992.03480100060031
CAS Web of Science® Google Scholar
39Awada A, Van Der Auwera P, Meunier F, Daneau D, Klastersky J. Streptococcal and enterococcal bacteremia in patients with cancer. Clin Infect Dis. 1992; 15(1): 33-48.
10.1093/clinids/15.1.33
CAS PubMed Web of Science® Google Scholar
40Cardenas-Morales M, Hernandez-Trujillo VP. Agammaglobulinemia: from X-linked to autosomal forms of disease. Clin Rev Allergy Immunol. 2022; 63(1): 22-35.
10.1007/s12016-021-08870-5
PubMed Web of Science® Google Scholar
41Groth D, Wright H, Marsh R, et al. X-linked agammaglobulinemia: infection frequencies in 226 patients from the USIDNET Registry. J Allergy Clin Immunol. 2020; 145(2): AB80.
10.1016/j.jaci.2019.12.676
Web of Science® Google Scholar
42Fishman JA, Hogan JI, Maus MV. Inflammatory and infectious syndromes associated with cancer immunotherapies. Clin Infect Dis. 2019; 69(6): 909-920.
10.1093/cid/ciy1025
CAS PubMed Web of Science® Google Scholar
43Cabanillas F, Liboy I, Pavia O, Rivera E. High incidence of non-neutropenic infections induced by rituximab plus fludarabine and associated with hypogammaglobulinemia: a frequently unrecognized and easily treatable complication. Ann Oncol. 2006; 17(9): 1424-1427.
10.1093/annonc/mdl141
CAS PubMed Web of Science® Google Scholar
44Casulo C, Maragulia J, Zelenetz AD. Incidence of hypogammaglobulinemia in patients receiving rituximab and the use of intravenous immunoglobulin for recurrent infections. Clin Lymphoma Myeloma Leuk. 2013; 13(2): 106-111.
10.1016/j.clml.2012.11.011
CAS PubMed Web of Science® Google Scholar
45Jacobson CA, Hunter BD, Redd R, et al. Axicabtagene ciloleucel in the non-trial setting: outcomes and correlates of response, resistance, and toxicity. J Clin Oncol. 2020; 38(27): 3095-3106.
10.1200/JCO.19.02103
CAS PubMed Web of Science® Google Scholar
46Locke FL, Rossi JM, Neelapu SS, et al. Tumor burden, inflammation, and product attributes determine outcomes of axicabtagene ciloleucel in large B-cell lymphoma. Blood Adv. 2020; 4(19): 4898-4911.
10.1182/bloodadvances.2020002394
CAS PubMed Web of Science® Google Scholar
47Schultz LM, Baggott C, Prabhu S, et al. Disease burden affects outcomes in pediatric and young adult B-cell lymphoblastic leukemia after commercial tisagenlecleucel: a pediatric real-world chimeric antigen receptor consortium report. J Clin Oncol. 2022; 40(9): 945-955.
10.1200/JCO.20.03585
CAS PubMed Web of Science® Google Scholar
48Nastoupil LJ, Jain MD, Feng L, et al. Standard-of-care axicabtagene ciloleucel for relapsed or refractory large B-cell lymphoma: results from the US lymphoma CAR T consortium. J Clin Oncol. 2020; 38(27): 3119-3128.
10.1200/JCO.19.02104
PubMed Web of Science® Google Scholar
49Forstner C, Patchev V, Rohde G, et al. Rate and predictors of bacteremia in afebrile community-acquired pneumonia. Chest. 2020; 157(3): 529-539.
10.1016/j.chest.2019.10.006
PubMed Web of Science® Google Scholar
50Henning DJ, Carey JR, Oedorf K, et al. The absence of fever is associated with higher mortality and decreased antibiotic and IV fluid administration in emergency department patients with suspected septic shock. Crit Care Med. 2017; 45(6): e575-e582.
10.1097/CCM.0000000000002311
PubMed Web of Science® Google Scholar
51Wrobel CA, Drazner MH, Ayers CR, et al. Delayed febrile response with bloodstream infections in patients with continuous-flow left ventricular assist devices. J Investig Med. 2019; 67(3): 653-658.
10.1136/jim-2018-000893
PubMed Web of Science® Google Scholar
52Weatherall SL, Chambers AB, Mermel LA. Do Bacteremic patients with end-stage renal disease have a fever when presenting to the emergency department? A paired, retrospective cohort study. BMC Emerg Med. 2020; 20(1): 2.
10.1186/s12873-019-0298-2
CAS PubMed Web of Science® Google Scholar
53Wang ME, Neuman MI, Nigrovic LE, et al. Characteristics of afebrile infants≤ 60 days of age with invasive bacterial infections. Hosp Pediatr. 2021; 11(1): 100-105.
10.1542/hpeds.2020-002204
CAS PubMed Google Scholar
54Fullerton JN. Increased mortality in “cold sepsis”: the result of a frozen immune response? Crit Care Med. 2015; 43(6): 1327-1329.
10.1097/CCM.0000000000000987
PubMed Web of Science® Google Scholar
55Wiewel MA, Harmon MB, van Vught LA, et al. Risk factors, host response and outcome of hypothermic sepsis. Crit Care. 2016; 20: 328.
10.1186/s13054-016-1510-3
PubMed Web of Science® Google Scholar
56Rejeski K, Blumenberg V, Iacoboni G, et al. Identifying early infections in the setting of CRS with routine and exploratory serum proteomics and the HT10 score following CD19 CAR-T for relapsed/refractory B-NHL. HemaSphere. 2023; 7(4): e858.
10.1097/HS9.0000000000000858
CAS PubMed Web of Science® Google Scholar
57Powell MZ, Mara KC, Bansal R, et al. Procalcitonin as a biomarker for predicting bacterial infection in chimeric antigen receptor T-cell therapy recipients. Cancer Med. 2023; 12(8): 9228-9235.
10.1002/cam4.5665
CAS PubMed Web of Science® Google Scholar
58Luo H, Wang N, Huang L, et al. Inflammatory signatures for quick diagnosis of life-threatening infection during the CAR T-cell therapy. J Immunother Cancer. 2019; 7: 271.
10.1186/s40425-019-0767-x
PubMed Web of Science® Google Scholar
59Taplitz RA, Kennedy EB, Bow EJ, et al. Antimicrobial prophylaxis for adult patients with cancer-related immunosuppression: ASCO and IDSA clinical practice guideline update. J Clin Oncol. 2018; 36(30): 3043-3054.
10.1200/JCO.18.00374
PubMed Web of Science® Google Scholar
60Rejeski K, Subklewe M, Aljurf M, et al. Immune effector cell-associated hematotoxicity: EHA/EBMT consensus grading and best practice recommendations. Blood. 2023; 142(10): 865-877. doi:10.1182/blood.2023020578
10.1182/blood.2023020578
CAS PubMed Web of Science® Google Scholar
61Trecarichi EM, Tumbarello M, Spanu T, et al. Incidence and clinical impact of extended-spectrum-β-lactamase (ESBL) production and fluoroquinolone resistance in bloodstream infections caused by Escherichia coli in patients with hematological malignancies. J Infect. 2009; 58(4): 299-307.
10.1016/j.jinf.2009.02.002
PubMed Web of Science® Google Scholar
62Bow EJ. Fluoroquinolones, antimicrobial resistance and neutropenic cancer patients. Curr Opin Infect Dis. 2011; 24(6): 545-553.
10.1097/QCO.0b013e32834cf054
CAS PubMed Web of Science® Google Scholar
63Abid MB. Granulocyte colony-stimulating factor usage in recipients of chimeric antigen receptor T-cell immunotherapy. J Clin Oncol. 2022; 40(13): 1508-1509.
10.1200/JCO.21.02694
CAS PubMed Web of Science® Google Scholar
64Hill JA, Seo SK. How I prevent infections in patients receiving CD19-targeted chimeric antigen receptor T cells for B-cell malignancies. Blood. 2020; 136(8): 925-935.
10.1182/blood.2019004000
PubMed Web of Science® Google Scholar
65Liévin R, Di Blasi R, Morin F, et al. Effect of early granulocyte-colony-stimulating factor administration in the prevention of febrile neutropenia and impact on toxicity and efficacy of anti-CD19 CAR-T in patients with relapsed/refractory B-cell lymphoma. Bone Marrow Transplant. 2022; 57(3): 431-439.
10.1038/s41409-021-01526-0
CAS PubMed Web of Science® Google Scholar
66Miller KC, Johnson PC, Abramson JS, et al. Effect of granulocyte colony-stimulating factor on toxicities after CAR T cell therapy for lymphoma and myeloma. Blood Cancer J. 2022; 12(10): 146. doi:10.1038/s41408-022-00741-2
10.1038/s41408-022-00741-2
PubMed Web of Science® Google Scholar
67Cheok K, Kirkwood AA, Creasey T, et al. Improved COVID-19 outcomes in CAR-T patients in the age of vaccination and preemptive pharmacotherapeutics. Leuk Lymphoma. 2023; 64(12): 1-5. doi:10.1080/10428194.2023.2248329
10.1080/10428194.2023.2248329
PubMed Google Scholar
68Cheok K, Kirkwood AA, Menne T, et al. Severe presentations and high mortality from SARS-CoV-2 in patients undergoing chimeric antigen receptor (CAR-T) therapy: a UK NCCP analysis. Leuk Lymphoma. 2022; 63(8): 1980-1984.
10.1080/10428194.2022.2057487
CAS PubMed Web of Science® Google Scholar
69van Doesum JA, Salmanton-García J, Marchesi F, et al. Impact of SARS-CoV-2 vaccination and monoclonal antibodies on outcome post–CD19-directed CAR T-cell therapy: an EPICOVIDEHA survey. Blood Adv. 2023; 7(11): 2645-2655.
10.1182/bloodadvances.2022009578
PubMed Web of Science® Google Scholar
70Atanackovic D, Luetkens T, Omili D, et al. Vaccine-induced T-cell responses against SARS-CoV-2 and its Omicron variant in patients with B cell–depleted lymphoma after CART therapy. Blood. 2022; 140(2): 152-156.
10.1182/blood.2022016175
CAS PubMed Web of Science® Google Scholar
71Chen G, Herr M, Nowak J, et al. Cytomegalovirus reactivation after CD19 CAR T-cell therapy is clinically significant. Haematologica. 2023; 108(2): 615-620.
10.3324/haematol.2022.281719
PubMed Web of Science® Google Scholar
72Márquez-Algaba E, Iacoboni G, Pernas B, et al. Impact of cytomegalovirus replication in patients with aggressive B cell lymphoma treated with chimeric antigen receptor T cell therapy. Transplant Cellular Therapy. 2022; 28(12): 851.e1-851.e8.
10.1016/j.jtct.2022.09.007
CAS PubMed Google Scholar
73Nasta SD, Hughes ME, Namoglu EC, et al. Outcomes of tisagenlecleucel in lymphoma patients with predominant management in an ambulatory setting. Clin Lymphoma Myeloma Leuk. 2022; 22(8): e730-e737.
10.1016/j.clml.2022.04.012
CAS PubMed Web of Science® Google Scholar
74Greenbaum U, Strati P, Saliba RM, et al. CRP and ferritin in addition to the EASIX score predict CAR-T–related toxicity. Blood Adv. 2021; 5(14): 2799-2806.
10.1182/bloodadvances.2021004575
CAS PubMed Web of Science® Google Scholar
75Pennisi M, Sanchez-Escamilla M, Flynn JR, et al. Modified EASIX predicts severe cytokine release syndrome and neurotoxicity after chimeric antigen receptor T cells. Blood Adv. 2021; 5(17): 3397-3406.
10.1182/bloodadvances.2020003885
CAS PubMed Web of Science® Google Scholar

Citing Literature

Volume8, Issue1

January 2024

e29

This article also appears in:

Infections in Hematology

Filename	Description
hem329-sup-0001-Figure_S1.JPG61.2 KB	Supporting information.
hem329-sup-0002-Figure_S2.JPG50.2 KB	Supporting information.
hem329-sup-0003-Figure_S3.JPG69.4 KB	Supporting information.
hem329-sup-0004-Figure_S4A.JPG71.2 KB	Supporting information.
hem329-sup-0005-Figure_S4B.JPG76.6 KB	Supporting information.
hem329-sup-0006-Supplementary_Tables_Revision_CLEAN.docx99.1 KB	Supporting information.

High pretreatment disease burden as a risk factor for infectious complications following CD19 chimeric antigen receptor T-cell therapy for large B-cell lymphoma

Graphical Abstract

Abstract

INTRODUCTION