Impact of SARS-CoV-2 vaccination and monoclonal antibodies on outcome post–CD19-directed CAR T-cell therapy: an EPICOVIDEHA survey

The introduction of CD19-directed chimeric antigen receptor T-cell (CAR T-cell) therapy for patients with relapsed or refractory large B-cell lymphoma meant an incredible leap forward in the survival of these patients.^1,2 Indications for CAR T-cell therapy keep expanding in commercial and clinical trial settings. The major long-term adverse effect of CD19-directed CAR T cells is prolonged B-cell aplasia and subsequent (viral) infections.^3,4 

Early reports of patients with coronavirus disease 2019 (COVID-19) after CD19-directed CAR T-cell therapy showed a dismal outcome of just 50% overall survival (OS).^5,6 Updates on the outcome during the following years and subsequent severe acute respiratory syndrome coronavirus type 2 (SARS-CoV-2) variants are lacking. In addition, the impact of vaccination and treatment with monoclonal antibodies (MoAbs) or convalescent plasma on outcomes is not known.

Several studies assessed the impact of vaccination on the humoral immune response.^7-9 However, only a small subset of patients obtain a serological conversion after vaccination. This is in line with larger vaccination studies in hematological patients that underpin the need for B cells for mounting a humoral response to vaccination.^7,10 Although little data have been published about cellular response in this specific patient category,^11,12 in the general population, T-cell response is considered an important factor in successful SARS-CoV-2 vaccination.¹³ Nevertheless, T-cell function could be impaired because of the CAR T-cell conditioning therapy with fludarabine/cyclophosphamide or bendamustine, especially in the first few months.

We hypothesize that vaccination and treatment with MoAbs will yield protection against adverse outcome. Therefore, we conducted this study to assess the impact of vaccination and treatment with MoAbs on the outcome of patients with COVID-19 after CD19-directed CAR T-cell therapy.

In this retrospective observational multicenter study, data were collected of all adult patients who were diagnosed with COVID-19 and for whom the last line of treatment immediately before the infection for their hematological malignancy was CD19-directed CAR T-cell therapy. The data were collected within the EPICOVIDEHA survey,¹⁴ an initiative of the European Hematology Association infectious diseases working group. EPICOVIDEHA was approved by the local ethics committee of 276 the Fondazione Policlinico Universitario Agostino Gemelli - IRCCS, Università Cattolica del Sacro 277 Cuore of Rome, Italy (study ID: 3226). The study was conducted in accordance with the Declaration of Helsinki, and the trial was registered at www.clinicaltrials.gov as #NCT04733729. Deidentified data on demographics, comorbidities, outcome, underlying hematological malignancy, and treatment were collected at the survey on www.clinicalsurveys.net. The data cutoff date was 1 July 2022.

The aim of this study is to assess the impact of COVID-19 on survival. The hematological prognosis of patients with relapsed disease after CAR T-cell therapy is very poor. To avoid this competing risk, we chose not to present the data as OS, but as event-free survival (EFS), whereby an event was defined as death caused by COVID-19. Patients were censored at day of death because of a cause other than COVID-19. Deaths caused by COVID-19 were patients who either died solely because of COVID-19 or COVID-19 in combination with other diseases (eg, progression lymphoma). Death due to progression was limited to only those cases where it was clear that there was no relation to COVID-19. Data are presented up to day 90 after COVID-19 infection, when 75% of the patients had an event or were censored. Patients were considered vaccinated when they had at least the minimum number of vaccinations that were recommended by the manufacturer in the general population (eg, 2 vaccinations with Comirnaty, Spikevax and 1 if Jcovden) and the time between the last vaccination and COVID-19 must be at least 7 days.

For statistical analyses, SPSS v25.0 (IBM Corp, Chicago, IL) was used. Categorical variables are presented using frequency and percentage; continuous variables are shown by median and interquartile range (IQR). EFS was estimated with the Kaplan-Meier method. Risk factors were evaluated using a Cox proportional hazard model. Because of the limited number of events, we applied a limited multivariate analysis in the form of pairwise Cox regression. For half of the patients, the viral variant was not known, so the year of the COVID-19 was used in the multivariate analysis, assuming that in 2020 and 2021, most of the patients had an infection with a non-Omicron variant, and in 2022, all patients were infected with an Omicron variant.

In total, 64 patients with COVID-19 and CD19-directed CAR T-cell therapy as their most recent treatment were identified in 34 centers across 15 countries (supplemental Figure 1). The median follow-up was 50.5 days (IQR, 25-90), the median age at COVID-19 diagnosis was 56 years (IQR, 47-65), and 38 (54.3%) were male. At the time of COVID-19 diagnosis, 42 patients (60%) had complete remission of their hematological malignancy, 3 (4.3%) had stable disease, and 17 (24.3%) had active disease; for 2 patients, the disease status was missing. The median number of previous treatment lines was 3 (IQR, 3-4). Sixteen patients (25.0%) had been treated previously with autologous hematopoietic stem cell transplantation (auto HSCT). The median time since transplantation and COVID-19 diagnosis was 602 days (IQR, 276-1108). The most often used CAR T-cell constructs were axicabtagene ciloleucel (50.0%) and tisagenlecleucel (42.2%). The median time since CAR T-cell infusion until COVID-19 diagnosis was 202.5 days (IQR, 85-450). In 50.0% of all cases, the SARS-CoV-2 variant was not determined. Of the patients for whom variants were tested, most were infected with an Omicron variant (28.1%). Twenty-six patients (40.6%) had received at least 1 vaccination before COVID-19. Comorbidities (hypertension, cardiovascular disease, pulmonary dysfunction, diabetes, renal impairment, liver disease, or smoking history) were present in 39 patients. Twenty-two patients (34.4%) had 1 comorbidity, 10 (15.6%) had 2 comorbidities, and 7 (10.9%) had 3 or more comorbidities. Full demographic and clinical characteristics are summarized in Table 1.

In total, 50 patients (78.1%) were hospitalized because of COVID-19, and 18 patients (28.1%) were admitted to intensive care units (ICU), underscoring the severe impact of COVID-19 in this patient population. The median duration of hospitalization was 25.0 days. COVID-19 treatment consisted of corticosteroids in 30 patients (46.9%), viral replication inhibitors in 18 patients (28.1%), convalescent plasma in 16 patients (25.0%), MoAbs in 14 patients (21.9%), and tocilizumab in 6 patients (9.4%). At last follow-up, 38 (59.4%) patients were alive (Figure 1). In 26 deceased patients, death was caused by the underlying hematological malignancy in 6 (23.1%) and due to COVID-19 in 20 (76.9%). EFS at day 90 was 68.8% (Figure 2A).

The number of patients diagnosed with COVID-19 each year was equally distributed over the years. In 2020, there were 23 patients diagnosed with COVID-19, in 2021, 18 patients and in 2022, 23 patients (until 1 July 2022). For full patient details, refer to supplemental Table 1. There was no difference in hospital admission rate between the years, and the duration of hospitalization in 2022 was significantly shorter (18 vs 34 days; P = .039). In 2022, the frequency of ICU admittance was significantly lower compared with 2020: 8.7% vs 47.8% (P = .004). When comparing patients infected with the Omicron variant and those with other variants, there was no difference in hospitalization rate (P = .568) or duration (P = .648), but patients infected with Omicron were less frequently admitted to the ICU (P = .020). We found a decrease in COVID-19–caused mortality with each passing year; refer to Figure 2B. Among patients with virus variant determined (N = 32, 14 patients with wild-type/alpha/delta variants, 18 patients with the Omicron variant), there was a significantly better EFS with the Omicron variant as compared with the wild-type and alpha and delta variants (P = .012; supplemental Figure 3).

In total, 26 patients had received at least 1 vaccination before their COVID-19 diagnosis. All but 2 were solely vaccinated with a mRNA-based vaccine. One patient received a vector-based vaccine, and 1 patient received a mRNA-based vaccine followed by a vector-based vaccine. Ten patients received the first vaccination before CAR T-cell therapy, and 16 patients received the first vaccination after CAR T-cell therapy. The antibody response was tested by local standards in 6 of the 26 vaccinated patients. Three patients had antibodies above the cutoff value of the test that was used. There were no significant differences between the vaccinated and unvaccinated groups in age, sex, hematological malignancy, number of previous therapies, CAR T-cell constructs, or cell counts at the time of COVID-19 diagnosis. The vaccinated group had a significantly longer interval between vaccination and CAR T-cell infusion, median 159 vs 386 days (P ≤ .001). In the vaccinated group, the Omicron variant was more common (10.5% vs 53.8%; P ≤ .001). There were no significant differences in treatment of COVID-19 except that MoAbs were more often used in the vaccinated group (7.9% vs 42.3%; P < .001). The median follow-up of the unvaccinated group was longer than in the vaccinated group (50.5 vs 36 days; P = .025). For full vaccination and patient details, refer to Table 2.

No difference was observed in frequency of hospitalization (P = .847) or ICU admission (P = .054), but the duration of hospitalization was significantly shorter in the vaccinated patients: 7 vs 27.5 days (P = .022). Vaccination showed a marked reduction in the risk of COVID-19–caused mortality (33.3% vs 14.2% [P = .379]), refer to Figure 2C. Furthermore, there is no significant difference in EFS between the patients vaccinated before CAR T-cell therapy and those vaccinated after (P = .519; supplemental Figure 4).

Fourteen patients were treated with MoAbs. Sex, age, status of hematological malignancy, and the CAR T-cell construct did not differ from those without MoAbs. In patients treated with MoAbs, the median time since CAR T-cell infusion was longer than in the non-MoAb group (162 vs 358 days; P = .028). They were more often infected with an Omicron variant (22.0% vs 50.0%; P = .046) and were significantly more often vaccinated (85.7% vs 32.0% [<.001]). See Table 3 for full details. There is a significant reduction in the risk of death due to COVID-19 between those treated with MoAbs and those without (P = .036; Figure 2D). There were no significant differences in hospitalization rate or duration, nor in ICU admission rates.

In univariate analysis, factors associated with death attributed to COVID-19 were age (P = .020; hazard ratio, 1.055; 95% confidence interval, 1.009-1.104) and infection with other variants than Omicron. The risk of death due to Omicron was lower than with other variants (P = .040; hazard ratio, 0.110; 95% confidence interval, 0.014-0.900). For all regression results, refer to supplemental Tables 2 and 3.

To elucidate a possible interaction between interventions and characteristics (age and year of COVID-19 infection), a pairwise Cox regression was conducted. In this limited number of events, only age was still a significant factor. In addition, there is a trend for reduction in the risk of dying due to COVID-19 when treated with MoAbs. Surprisingly, the year of COVID-19 (and thereby the SARS-CoV-2 variant) seemed to have less impact on the patient’s outcome.

We report on the impact of vaccination and treatment in patients with COVID-19 after previous CD19-directed CAR T-cell therapy. The main findings were a 40% overall mortality and a 31% mortality rate due to COVID-19, and thus a better prognosis than the previously reported 50% overall mortality.^5,6 Although still poor compared with the general population^15,16 and patients with other hematological malignancies, in which the mortality was 9% after vaccination.¹⁷ Our results indicate that a large proportion (23%) of mortality is because of the progression of lymphoma, and taking this into account, the OS of 60% after 90 days is comparable to earlier studies.

Although not significant, there is a trend for better survival over the years, as shown by a reduction in COVID-19–caused mortality from 43.5% to 13.0%. The chance to be hospitalized or to be admitted to the ICU also reduced over time, and the duration of hospitalization is significantly shorter. There are several interacting factors that could cause a better outcome. In the early days of COVID-19, only patients who were admitted to the hospital were tested and diagnosed. Later on, testing became available for the outpatient clinic, resulting in a selection bias. With an improved understanding of COVID-19, new treatment options and interventions became available.

After observing the clinical course of vaccinated patients, this seems to be milder, with a significantly shorter duration of hospital stay and a trend showing the chance to be admitted to the ICU. This is concordant with other studies in patients with hematological malignancies.¹⁸ There was no difference in EFS between the patients receiving the first vaccination before CAR T-cell therapy compared with those receiving the first vaccination after CAR T-cell therapy. One can presume that all patients were already B-cell depleted because of previous anti–B-cell therapy, as the median number of previous therapies was 3, possibly resulting in an impaired vaccination response. Information about serological response to vaccination is unfortunately limited in this survey. Presumably, serologic responses to vaccination were not different among patients in vaccination studies, with only one-third showing seroconversion.¹⁹ Although half of the vaccinated patients were infected with the Omicron variant, the trend is still present when we focus only on the vaccinated non-Omicron group (data not shown). In addition, we found that patients who received CAR T-cell therapy benefit from the milder clinical course of the Omicron variant with less frequent ICU admittance (P = .019), although the duration of hospitalization is not different compared with the non-Omicron–infected patients (22.8 vs 26.7 days [P = .648]). The mortality is significantly lower in the Omicron group, with 8% vs 68% in the non-Omicron group at day 90 (P = .012). This is in line with previous studies in patients with hematological malignancies infected with Omicron, where the mortality was 8.4%.²⁰ 

Treatment with steroids, viral replication inhibitors, IL-6 inhibition, and convalescent plasma showed no improvement in survival or chance to be hospitalized (data not shown). Treatment with MoAbs showed the largest impact on EFS, with none of the patients dying because of COVID-19. Taking into account that all but 2 patients were also vaccinated at the time of infection, we can conclude that the combination of vaccination and treatment with MoAbs seems to give significant protection against COVID-19 mortality.

The study population represents the real-world population treated with CD19-directed CAR T-cell therapy by distribution of sex, age, previous lines of therapy (including auto HSCT), and CAR T-cell constructs used.^21-23 The outcome data may be applicable to the general CD19-directed CAR T-cell–treated population.

Like in the general population and in previous studies of patients with hematological malignancies, advanced age and non-Omicron variant SARS-COV-2 infection are risk factors for death due to COVID-19.^15,17,24 Although the general population showed an improvement in survival over the past 3 years,¹⁶ we found that the CAR T-cell population does have limited benefit from improved treatment and vaccination options. A possible explanation could be that most vaccinated patients included in this study were vaccinated twice. As shown in previous publications, patients with hematological malignancies need to be vaccinated 3 times for an optimal response.¹⁹ The importance of a fourth vaccination and its protective effect were also shown in a previous EPICOVIDEHA study.¹⁸ Here, 6 patients received a third or even fourth vaccination, and none of these patients died because of COVID-19 (data not shown).

Recent studies show promising T-cell responses after vaccination in the CAR T-cell–treated patient category.^11,12 Unfortunately, this is not reflected in our data on survival or hospitalization rates. Probably this highlights the important role of humoral immunity in response to SARS-COV-2. After CD19-directed CAR T-cell therapy, there is a deep and long-lasting B-cell depletion.^3,4 Presumably this is the main reason for the poor serological response to vaccination.^7,10 By treatment with MoAbs, the patients seem to overcome the lack of B cells, and the combination with improved T-cell responses owing to vaccination reduces the mortality risk. A major limitation of MoAb treatment is the constant need to adjust the antibody if there is a new viral mutation, limiting its use in daily practice. But showing the high efficacy in this patient category underpins the need for the continuous development of new antibodies and should encourage investors to invest in this life-saving type of therapy.

A limitation of this study is the small sample size, its retrospective nature, the varying treatment protocols over time and country, and the fact that in half the patients, the viral variant is unknown. In addition, data for in-depth analyses of B-cell and T-cell recovery and the timing of interventions are missing. We need to be cautious with definitive conclusions and larger studies are therefore needed. Moreover, because of the restricted therapeutic interventions at relapse after CAR T-cell therapy, the mortality of COVID-19 could be overestimated. In 14% of the cases, it was reported that the malignancy contributed to death. Possibly, it was decided to give patients with active disease only part of the full COVID-19 treatment, including intensive care. Another bias could be the underreporting of outpatients in the survey, resulting in a higher relative mortality attributed to COVID-19.

In conclusion, we find that the survival of CAR T-cell recipients with COVID-19 is better than previously reported, that vaccination significantly reduces the duration of hospitalization, and that the combination of vaccination and MoAbs significantly reduces the risk of death due to COVID-19 in these patients. However, because of the retrospective nature of this study and the ongoing improvement in both COVID-19 treatment and viral variants, a causal relationship cannot be established.

As such, we strongly advise patients who received CAR T-cell therapy to get vaccinated and for clinicians to consider therapy with MoAbs if available. Thereby also highlighting the need to for further development of new MoAbs against SARS-Cov-2 and other viral infections.

The authors thank all contributors for their contribution and support.

Contribution: J.A.v.D. and J.S.-G. performed statistical analysis, interpreted the data, and wrote the manuscript; J.S.-G., F.M., L.P., and O.A.C. are in charge of the management of EPICOVIDEHA registry; and all authors recruited participants, collected and interpreted data, contributed to manuscript writing and review of the manuscript, and agreed to be accountable for all aspects of the work in ensuring that questions related to the accuracy or integrity of any part of the work are appropriately investigated and resolved.

Conflict-of-interest disclosure: The authors declare no competing financial interests.

Correspondence: Jaap A. van Doesum, University Medical Center Groningen, Hanzeplein 1, 9713 GZ, Groningen, The Netherlands; e-mail: j.a.van.doesum@umcg.nl; and Jon Salmanton-García, Faculty of Medicine and University Hospital Cologne, Cologne Excellence Cluster on Cellular Stress Responses in Aging-Associated Diseases, University of Cologne, Cologne, Germany; Department I of Internal Medicine, Excellence Center for Medical Mycology, Herderstraße 52-54, 50931, Cologne, Germany; e-mail: jon.salmanton-garcia@uk-koeln.de.