Breakthrough COVID-19 in vaccinated patients with hematologic malignancies: results from the EPICOVIDEHA survey

Coronavirus disease 2019 (COVID-19) is a life-threatening infection in patients with hematologic malignancies (HM), associated with severe clinical presentation and high risk of death.^1-3 In April 2020, the European Hematology Association’s Specialized Working Group, Infections in Hematology, opened the EPICOVIDEHA (Epidemiology of COVID-19 Infection in Patients with Hematological Malignancies: European Haematology Association) registry to collect data on all adult patients with HM who developed COVID-19. It aimed to describe the epidemiology and risk factors, and reported a mortality rate of 31.2% among 3801 patients.⁴ In December 2020, nearly 1 year after the first described COVID-19 case, vaccines against severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) were approved and became available first to patients at high risk, including patients with HM.^5-8 The recently published recommendations from the European Conference of Infections in Leukemia (ECIL-9) identify the critical role of messenger RNA (mRNA)-based vaccines in the fight against COVID-19 and recommend their use in HM, although they may have more limited efficacy among severely immunocompromised patients.⁹ 

We collected data on adults with HM who developed breakthrough COVID-19, to assess the vaccine efficacy and the potential role of new emergent treatments against SARS-CoV-2. Our preliminary data, regarding the first 113 patients included, showed a significant decrease in the overall mortality rate in the postvaccination era (12.4%), which was, however, still remarkably higher than the rate observed in the overall population.¹⁰ To date, few reports have been published on the severity and outcomes of breakthrough COVID-19 in patients with cancer in general^11,12 and HM specifically,¹³ all showing high rates of severe clinical presentation, hospitalization, and death among these patients. This suggests that HM require close monitoring and increased medical attention when COVID-19 is diagnosed, regardless of previous anti-SARS-CoV-2 vaccine administration.

In this study, we analyzed the epidemiology and outcome of breakthrough COVID-19 in a large cohort of patients with HMs and evaluated anti-SARS-CoV-2 treatment received by these patients.

From 1 January 2021 to 10 March 2022, participating institutions documented episodes of COVID-19 in patients with HM that received anti-SARS-CoV-2 vaccination. Our analysis comprised data from the EPICOVIDEHA registry. EPICOVIDEHA (www.clinicaltrials.gov; National Clinical Trials identifier NCT04733729) is an international open web-based registry for patients with HMs infected with SARS-CoV-2.¹⁴ The EPICOVIDEHA registry was approved by the local ethics committee of the Fondazione Policlinico Universitario Agostino Gemelli, RCCS, Università Cattolica del Sacro Cuore of Rome, Italy (Study ID: 3226). When applicable, the respective local ethics committee of each participating institution had approved the project. EPICOVIDEHA methods have been described elsewhere.^4,14 The electronic case report form is accessible online at www.clinicalsurveys.net (EFS Summer 2021, Tivian, Cologne, Germany). Each patient that had been documented was reviewed and validated by infectious diseases and hematology experts from the coordination team. Inclusion criteria were: (1) active HM within the last 5 years before COVID-19 diagnosis, (2) aged ≥18 years, (3) laboratory-based diagnosis of SARS-CoV-2 infection, and (4) last vaccine dose ≥15 days before PCR confirmed SARS-CoV-2 infection. Data on baseline conditions pre-COVID-19 (ie, age, sex, status of HM at COVID-19 diagnosis, COVID-19–predisposing factors), HM clinical management (ie, last HM treatment strategy), vaccine type, spike protein concentration at diagnosis of COVID-19, COVID-19 diagnosis and management (ie, reason for diagnostic test, symptoms at onset, hospital stay during infection, treatments received for infection), and outcome (ie, mortality, attributable mortality [assessed by the medical team in charge of the patient], and last day of follow-up) were collected. Status of HM at COVID-19 onset and last follow-up was defined as active (onset and refractory/resistant), stable disease, or controlled (complete and partial response) based on the reports from the respective participating institution.

The primary objective of this study was to assess the epidemiology and the outcome of HM affected by breakthrough COVID-19. Secondary objectives were: (1) to estimate the relative frequency of disease severity, graded according to international standards in our patient population;^15,16 (2) to evaluate the relative frequency of intensive care unit (ICU) admission among participating patients; (3) to evaluate the overall case-fatality rate; (4) to explore the effect of cancer treatment phase (induction, consolidation, maintenance, palliative, or reinduction) on patient outcomes; (5) to explore the effect of vaccine doses administered on patient outcomes; and (6) to explore the effect of COVID-19 treatment on patient outcomes. Moreover, data collected were compared with data reported in our previously published study performed in the prevaccine era by using the same registry.⁴ 

No a priori sample size calculation was performed for this analysis. Categorical variables are presented with frequencies and percentages, and continuous variables with median, interquartile range (IQR), and absolute range. A univariable Cox regression model was performed with variables suspected to play a role in the mortality of patients with HM with COVID-19. Variables with a P value ≤.1 were considered for multivariable analysis. A multivariable Cox regression model was calculated with the Wald backward method. Mortality was analyzed by using Kaplan–Meier survival plots. A log-rank test was used to compare the survival probability of the patients included in the different models. A P value ≤.05 was considered statistically significant. No a priori sample size calculation was done for this exploratory study. SPSS version 25.0 was employed for statistical analyses (SPSS, IBM Corp, Chicago, IL). Patients with missing data in essential fields (ie, HM, chemotherapeutic program, vaccination status, COVID-19 management, or survival status) were considered as “not valid” and excluded from the final analysis. Among the valid cases, if a value in a specific variable was missing or unknown, it is indicated as such in the descriptive analysis. Patients with missing data in a certain variable were excluded from regression analyses in case that variable was included into such analyses.

A total of 94 centers in 26 countries, mainly from Europe, participated and registered 1583 cases. A list of enrolled cases from each participating country is available in the supplemental material (supplemental Figures 1 and 2A, available on the Blood website). Out of these 1583 cases, 35 were excluded because COVID-19 was diagnosed within 14 days from the first vaccine dose. Clinical characteristics of 1548 evaluable cases are reported in Table 1. Lymphoid malignancies were the largest subgroup, accounting for 1181 cases (76.3%); the most frequently reported diagnosis was non-Hodgkin lymphoma (NHL, 549 cases). Among myeloid malignancies, the most frequent diagnosis was acute myeloid leukemia (AML, 140 cases). We found a significantly different distribution of lymphoid/myeloid malignancies from that reported in the prevaccination era (prevaccination lymphoid malignancy cases, 67.3%, vs postvaccination, 76.3%, P < .001). At the time of COVID-19 diagnosis, most patients had a controlled malignancy (n = 821, 53%), 322 (20.8%) a stable disease, and the remaining 365 (23.6%) an active disease, with 185 cases registered at HM onset. The most frequently reported last HM treatment was immuno-chemotherapy or immunotherapy alone (n = 708, 42%), followed by targeted therapies (n = 311, 20.1%) and conventional chemotherapy (n = 234, 15.1%); 92 patients (5.9%) had received hematopoietic stem cell transplantation within 6 months before COVID-19 (allogeneic 76; autologous, 16) and 8 had chimeric antigen receptor T-cell therapy. Most patients presented at least 1 comorbidity (60.7%) and 180 (11.6%) had a history of smoking; a complete list of comorbidities and associated clinical outcomes is available in the supplemental material (supplemental Table 1).

COVID-19 was mild, severe, or critical in 39%, 32.9%, and 9.8% of cases, respectively. Of the 1548 patients, 283 (18.3%) were asymptomatic and in most of them the diagnosis was made in screening programs (Table 1). We found a significantly lower rate of severe or critical cases compared with what we reported in the prevaccination era (prevaccination, 2425/3801, 63.8%, vs postvaccination, 661/1545, 42.7%; P < .001). Overall, 823 (53.2%) patients required hospitalization and among them 152 (18.1%) required admission to ICUs. The hospitalization and ICU admission rate was significantly lower than reported in the prevaccination era (53.2% vs 73%; P < .001% and 9.8% vs 18.1%; P < .001, respectively). The percentage of asymptomatic cases was 18.3% (283/1548), similar to that reported in our previous publication with data from the prevaccine era (17.8%, 675/3801).⁴ Viral genomes were studied in 753 cases (48.6%), with the different Omicron variant the most frequently detected viral strain (517/753, 68.7%). Most patients received 2 or 3 anti-SARS-CoV-2 vaccine doses (91%), mostly with mRNA-based technology (89%); only a few patients (8.6%) received a vector-based vaccine and a minority of them an inactivated vaccine (Table 1; supplemental Figure 2B-D). Anti-SARS-CoV-2 spike protein immunoglobulin G (IgG) levels were analyzed in 244 (15.8%) fully vaccinated patients, 2 to 4 weeks after the last vaccine dose; among these patients, 109 (44.7%) presented an antibody response (optimal, 75, 30.7%; weak, 34, 13.9%), whereas the remaining 135 (55.3%) were nonresponders. Most patients who did not have a serological response to vaccines were affected by lymphoid malignancies, as expected (126/135, 93.3%; Figure 1).

Overall, 906 patients (58.5%) received a specific treatment for COVID-19, whereas 642 (41.5%) were not treated or received symptomatic therapies (nonsteroidal anti-inflammatories, painkillers, antipyretics). Among patients who received a specific treatment for COVID-19, 311 (34.3%) were treated with monoclonal antibodies only, 246 (27.1%) with corticosteroids only, 218 (24.1%) with antivirals only, 108 (11.9%) with antiviral plus monoclonal antibodies, and the remaining 23 with convalescent plasma. Details on COVID-19 treatments and outcomes are displayed in the supplemental material (supplemental Table 2). Overall day-30 mortality (ie, from COVID-19 diagnosis) was 9.2% (143/1548); if we only consider patients who were symptomatic, the day-30 mortality rate was 0.3% (130/1265). The primary cause of death was COVID-19 in 97 patients (67.8%), a combination of COVID-19 and progressive HM in 39 cases (27.2%), and HM alone or combined with other reasons in the remaining 7 patients (4.8%). The mortality rate was significantly lower than that reported in the prevaccine era (prevaccine 31.2% vs postvaccine 9.2%; P < .001). Looking at 2 of the largest patient cohorts (ie, chronic lymphocytic leukemia and NHL) we evaluated the potential effect of chemotherapeutic treatment type on mortality rate. In patients with chronic lymphocytic leukemia, we did not observe any significant difference in terms of 30-day mortality rate among patients who had received immuno-chemotherapy (13.4%), immunotherapy alone (12.5%), or new targeted therapies (16.1%). On the contrary, in NHL we did observe a slightly higher mortality rate for patients recently treated with chimeric antigen receptor T cells (20%), compared with those treated with immuno-chemotherapy (8%), immunotherapy alone (14.3%), or targeted therapies (9.5%). Patient outcomes according to clinical characteristics, vaccine received, and specific treatments against SARS-CoV-2 are detailed in Table 2. As shown in Figure 2, we did not find any significant difference in terms of 30-day mortality rate among the different HM (P = .693), in contrast to that observed in the prevaccination era in which we reported a higher number of fatalities in patients with AML/myelodysplastic syndrome. In univariable analysis, the factors associated with a worse mortality rate were older age (P < .001), active HM disease (P < .001), and presence of 2 to 3 comorbidities (P < .001) (Table 3). Regarding age, patients aged <60 years showed a more favorable outcome (30-day mortality rate, 2.6%), compared with patients aged 60 to 69 years (7%), 70 to 79 years (14.8%), and ≥80 years (19.6%) (P < .001). Conversely, we observed a better clinical outcome for patients who received monoclonal antibodies (with or without antivirals; Figure 3). Analyzing the severity of COVID-19 presentation, a better clinical outcome was observed in patients treated with monoclonal antibodies alone for asymptomatic, mild, or severe disease and with monoclonal antibodies combined with antivirals in critical cases (Figure 4). We did not find differences in terms of outcome according to the number of vaccine doses received; however, a slightly better clinical outcome was evident among patients who received 3 to 4 doses vs 1 to 2 doses (P = .040, Table 3). We did not observe differences in survival when sorting patients according to viral strain detected (P = .664; Figure 5), or postvaccine antispike IgG levels (Table 2).

In the multivariable model, older age, active disease, and 2 to 3 comorbidities were the factors significantly correlated with a higher mortality, whereas receiving anti-SARS-CoV-2 treatment with monoclonal antibodies alone or combined with antivirals was independently associated with a lower mortality (HR, 0.155; 95% CI, 0.077-0.313; P < .001 and HR, 0.407; 95% CI, 0.206-0.803; P = .010, respectively) (Table 3). Survival and severity according to vaccine doses administration and postvaccine antispike IgG levels are shown in Figures 6 and 1, respectively.

In the prevaccination era, several studies reported a high COVID-19 mortality in HM.^1-4 From December 2020, anti-SARS-CoV-2 vaccines have been administered in patients with cancer, including those with HM.^7,8 Most published studies on HM confirmed the efficacy and safety of vaccines, particularly those using mRNA, however, most showing less efficacy in patients with lymphoid malignancies treated with immunosuppressive drugs.^17-22 

This study was performed in a large cohort of vaccinated patients with HM to evaluate epidemiology, risk factors for adverse clinical outcome, and treatments of breakthrough COVID-19. We found a predominance of lymphoid malignancies, higher than observed in our previous survey during the prevaccine era; this difference might be explained by the lower efficacy of vaccines in this patient population, as further suggested by the high rate of serological nonresponders among patients with lymphoid malignancies when evaluating antispike IgG levels. These data are consistent with data in a recent report describing COVID-19 breakthrough infections in a large cohort of patients with HM, mostly consisting of patients with lymphoid malignancies.¹³ Advanced age, presence of comorbidities, and active HM were confirmed in this study as factors that negatively influenced clinical outcome and survival; these were the same risk factors that had previously been reported in the prevaccination era.^1-4 Interestingly, in this study, the underlying malignancy did not have a significant effect on survival, which was different from our previous experience in nonvaccinated patients, where AML and myelodysplastic syndrome were associated with higher mortality risk.⁴ A potential explanation for this difference might be the better efficacy of anti-SARS-CoV-2 vaccines in myeloid malignancies^23-25 than in lymphoid malignancies;^17-22 however, we may hypothesize new specific anti-SARS-CoV-2 drugs and better COVID-19 management to be particularly important for patients with AML at risk of increased mortality if urgent chemotherapy is delayed. Similarly, as reported by other studies,¹³ we did not find any significant difference in terms of mortality among different treatments received for HM. As expected, severe and critical COVID-19 had a worse clinical outcome than mild cases, showing a strong correlation with an increased mortality rate both in univariable and multivariable analysis. Given the vaccine protection, the occurrence of respiratory symptoms, hospitalization rate, and severe or critical clinical presentations were significantly lower than in the prevaccination era but still substantially higher compared with that of the overall population.^26-29 However, it is worth underlining that ∼20% of patients were asymptomatic and SARS-CoV-2 infection was detected in screening programs. Interestingly, this percentage is analogous to that reported in our previously published study referring to the prevaccination era.⁴ Unfortunately, it is not possible to estimate the true incidence of breakthrough infections nor the true number of patients who were asymptomatic with our data as only patients with COVID-19 were included in the registry; this is a potential selection bias, hypothetically hampering the reliability of our results. To the best of our knowledge, only few studies evaluated the incidence and cumulative COVID-19 risk among vaccinated patients with cancer, showing an increased risk in patients with HM compared with the overall population.^30-32 In particular, Lee and coworkers³² recently published a comprehensive population-based test-negative case-control study in the United Kingdom, evaluating COVID-19 breakthrough infections among a large number of vaccinated patients with cancer and healthy control participants. The authors showed that vaccine efficacy at 3 to 6 months after the second dose was lower in the cancer cohort than in the control population and among patients with cancer, was lower in patients with HM, especially those affected by leukemia and lymphoma. Very recently, an Italian study evaluated the immunogenicity and clinical efficacy of anti-SARS-CoV-2 vaccine in 365 patients with HM. The authors showed an overall incidence of breakthrough infections of 2.98 per 10 000 person-days, significantly lower in patients who were seropositive after vaccination, whereas a clear correlation between T-cell immunity response and risk of postvaccine infection has not been found.³³ 

In this study, we reported an overall 30-day-mortality rate of 9.2%, mainly driven by COVID-19 infection as a direct or contributing factor, which is significantly lower than in the prevaccination era.^1-4 Moreover, the 30-day mortality rate in patients who were symptomatic was 10.3%. The success of vaccination strategies is likely a major factor in the reported improvement but not the only factor. Previous reports suggest that COVID-19 management (eg, steroids, etc) have also affected outcomes, and newer variants may be less severe. Data reported in our study are comparable with other recently published reports that showed a significant mortality rate of COVID-19 breakthrough infections among patients with cancer.^11,12 or more specifically, among those affected by HM.¹³ 

In our study, we collected data on viral genotyping in approximately half of patients, among which the most prevalent variant was Omicron, accounting for more than two-thirds of patients. These data are not surprising if we consider the large number of patients registered between late 2021 and early 2022, months in which the Omicron variant was rapidly spreading throughout Europe.³⁴ Interestingly, we did not find any significant difference in terms of severity of clinical presentation and mortality rate between Omicron and other variants, similar to other small, recently published reports on HM^35,36 but different to reports on immunocompetent patients in which Omicron presents with better outcome than other variants.^34,37 

Most of the patients enrolled in our study received 2 or 3 vaccine doses; comparing clinical presentation and outcomes, we did not find consistent data supporting a better clinical outcome for patients who had received a higher number of vaccine doses, although a slight difference in the proportion of deaths was observed comparing those who received 1 to 2 vs 3 to 4 doses. However, in multivariable analysis, the number of doses did not significantly effect on the overall 30-day mortality. Several studies highlighted the role of a third vaccine dose as capable of restoring the immune response in serologically less responsive patients with HM.^38,39 However, there are insufficient data to consider patients with low antispike antibody titers at high risk of worse outcomes. Indeed, in our study we did not find any differences in terms of outcomes stratifying patients according to serological response after 2 to 4 weeks from the last vaccine dose. By using the World Health Organization international standards (binding antibody units per milliliter), we did not find a significantly better survival for patients with optimal response, compared with those with weak or no response, although these data were only available in a small percentage of patients (16%). This lack of direct correlation between serological response and survival might be, at least in part, explained by the putative role of anti-SARS-CoV-2–induced cellular immunity, as suggested by several studies,^23,24,40 because the presence of memory T cells might control the infection and prevent severe COVID-19, even if high titers of long-lasting neutralizing antibodies are not elicited.⁴¹ However, because a recently published study did not find a clear correlation between postvaccine T-cell immunity and vaccine clinical efficacy,³³ further studies are warranted to better understand this aspect. Another possible explanation is related to the role of the specific anti-SARS-CoV-2 treatments (ie, monoclonal antibodies and antivirals) that could have partially balanced the lack of protection of serological nonresponders. Indeed, from our survey, monoclonal antibodies with or without antivirals showed a high clinical activity irrespective of COVID-19 severity, showing the best efficacy when administered as single agents in patients with asymptomatic, mild, or severe disease, and when administered in combination with antivirals in patients with critical disease. The role of monoclonal antibodies in mitigating the negative effect of weak vaccine responses is supported by a recent randomized trial evaluating their role in immunocompetent people without serological response.⁴² Moreover, our multivariable model confirmed the positive effect on 30-day mortality risk for patients who had received monoclonal antibodies alone or combined with antivirals. We are aware that this study has limitations owing to the retrospective observational design and the possible selection bias owing to the large number of participating institutions. Moreover, viral genotyping and serological data were not available for all enrolled patients, and we did not know whether COVID-19 was first diagnosed in hospital or in the community, potential key information for discriminating patient risk and infection natural history. Further prospective studies that better evaluate the role of vaccine response in patients with HM are needed.

In conclusion, our survey has shown that vaccination and novel COVID-19 treatments have brought significant improvements in terms of mortality in patients with HM. To further improve the prognosis of these patients, the role of additional booster vaccine doses and the role of prophylactic monoclonal antibodies in patients with an ineffective response to vaccination should be investigated.

The authors thank Janina Leckler and all contributors for their utmost contributions and support to the project during a pandemic situation.

EPICOVIDEHA has received funds from Optics COMMIT (COVID-19 Unmet Medical Needs and Associated Research Extension) COVID-19 RFP program by GILEAD Science, United States (Project 2020-8223).

Contribution: L.P. served as the principal investigator; J.S.-G. and F.M. served as project manager and research assistant, respectively. L.P., J.S.-G., and F.M. contributed to study design, study supervision, and data interpretation and wrote the paper; A.B., P.C., M.H., P.K., A.P., F.P., O.A.C., and L.P. conceived the registry idea. L.P., J.S.-G., and F.M. did the statistical plan, analysis, and interpreted the data; all the authors recruited participants and collected and interpreted data; all authors contributed to manuscript writing and review of the manuscript; and all authors 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.

A complete list of the EPICOVIDEHA Survey collaborators appears in “Appendix.”

Correspondence: Livio Pagano, Fondazione Policlinico Universitario Agostino Gemelli-IRCCS-Università Cattolica del Sacro Cuore, Largo Francesco Vito 1, 00168 Rome, Italy; e-mail: livio.pagano@unicatt.it.

The EPICOVIDEHA Survey collaborators that contributed to this work are Laura Serrano, José-María Ribera-Santa Susana, Joseph Meletiadis, Panagiotis Tsirigotis, Nicola Coppola, Malgorzata Mikulska, Nurettin Erben, Caroline Besson, Maria Merelli, Tomás-José González-López, Jorge Loureiro-Amigo, Carolina García-Vidal, Elizabeth de Kort, Annarosa Cuccaro, Sofia Zompi, Florian Reizine, Olimpia Finizio, Rémy Duléry, Maria Calbacho, Ghaith Abu-Zeinah, Sandra Malak, Przemyslaw Zdziarski, Gina Varrichio, Athanasios Tragiannidis, Gaëtan Plantefeve, Rafael Duarte, François Danion, Maria Chiara Tisi, Ioanna Sakellari, Meinholf Karthaus, Ana Groh, Monica Fung, Ziad Emarah, Omar-Francisco Coronel-Ayala, Louis Yi Ann Chai, Mathias Brehon, Valentina Bonuomo, Dominik Wolf, Jana Wittig, Maria Vehreschild, Mario Virgilio Papa, Julia Neuhann, María-Josefa Jiménez-Lorenzo, Jan Grothe, Eleni Gavriilaki, Ramón García-Sanz, Nicole García-Poutón, Shaimaa Saber El-Ashwah, Matthias Eggerer, Raul Cordoba, Gökçe Melis Çolak, and Elena Arellano.