Introduction

The COVID-19 pandemic, caused by the severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2), posed significant challenges to national healthcare systems between 2020 and 20221,2. It had a great impact on humanity, with 6.7 million confirmed deaths by December 20223with a particularly large impact on vulnerable populations4,5,6.This includes patients with hematological malignancies (HMs), who are a high-risk group for several reasons including underlying immune dysregulation due to the nature of their hematologic dysfunction, disease-specific complications, their lack of an appropriate response to vaccines, and intensive treatment regimens that cause deeper immune dysfunction7,8,9,10. Such aggressive management involves immunosuppressive chemotherapy, which, combined with disease-related immunodeficiency, predisposes these patients to severe viral, bacterial and fungal infections, including by SARS-CoV-2. Poorer clinical outcomes related to COVID-19 have been reported in hematologic patients relative to the general population6,11.

Among patients with HMs, those who have undergone bone marrow or hematopoietic stem-cell transplantation (HSCT) are a particularly vulnerable subset who require special attention. HSCT recipients have prolonged immune reconstitution periods, during which they remain highly susceptible to infections and complications related to immune dysfunction12,13,14. This increased vulnerability is due to a combination of several other factors, including preexisting comorbidities, graft-versus-host disease, and prolonged exposure to healthcare environments15,16. The combination of these factors with SARS-CoV-2 infection highlights the importance of understanding the clinical and demographic characteristics of populations with HSCT.

We analyzed the impact of the COVID-19 pandemic at a national level in hospitalized, hematologic patients with HSCT. The objective was to provide a comprehensive picture of patients with HMs who received HSCT, describing their demographic and clinical profiles during the pandemic. We also identified risk factors associated with mortality in HSCT recipients using a large dataset to have greater insights into their management.

Methods

Study design and population

We conducted a nationwide, retrospective study including patients with HMs who were hospitalized in Spain during the first 3 years of the COVID-19 pandemic (2020–2022). Data were extracted and collected from the Minimum Basic Data System at Hospitalization (MBDS-H) database provided by the Spanish National Health System. This is an administrative registry that includes demographic and clinical data extracted from hospitalization discharge reports. We collected data for approximately 35,000 patients with HMs, of whom 2,300 had undergone HSCT before or during the pandemic.

Data collection

MBDS-H uses discharge reports from nearly 95% of both public and private hospitals in Spain. It is estimated that 97% of all discharge reports are collected in this registry. The registry collects several demographic and clinical variables from patients: age, sex, date of admission/discharge, type of hospital, place of residence, ICU admission, time and occurrence of death (if applicable), and diagnosis/diagnoses at discharge. Diagnoses are encoded using the 10th Clinical Revision of the International Classification of Diseases (ICD-10-CM). We collected data using the coding for COVID-19 (U07.1) at any diagnostic position. We also identified codes on HMs including Hodgkin and non-Hodgkin lymphoma, multiple myeloma and plasma cell disorders, acute and chronic leukemia (lymphoid, myeloid, and monocytic), and myeloproliferative disorders (polycythemia vera, essential thrombocytosis, myelofibrosis, and chronic myelomonocytic leukemia). The patients were classified in relation to their HSCT status and were analyzed separately. We did not categorize groups by allogeneic or autologous HSCT due to the characteristics of the registry. We also identified certain comorbidities of interest, such as diabetes, hypertension, and other chronic conditions. No names or personal information were recorded, so the data were provided in a de-identified form to ensure patient privacy. No data on treatment or immunological status were provided. This study was approved by the research ethics committee of our institution.

A major limitation of the registry used is that a new dataset is generated in January of each year. Due to the large amount of data, the availability of the database has a delay of 1 year. Thus, Spanish Ministry of Health provided us with data through 31 December 2022.

Statistical analyses

We first performed descriptive and correlational analyses. Demographic and clinical characteristics were expressed using medians and interquartile ranges (IQRs) for continuous variables and frequencies and percentages for categorical ones. Correlational analyses were performed using the chi-square test for categorical variables and the Wilcoxon rank sum test for continuous variables.

We used multivariate logistic regression models to identify the risk factors associated with mortality in HSCT recipients. We included all variables, not only statistically significant (p < 0.05) ones, in univariate analyses. We used stepwise regression to fit the logistic regression model. This automatic procedure eventually chooses relevant, predictive variables. We adopted bidirectional elimination as the main approach for stepwise regression. The results from the fitted model are presented as univariate and adjusted odds ratios (ORs) with 95% confidence intervals (CIs).

For survival analysis, we used Cox proportional hazards regression models to identify predictors of mortality, and hazard ratios (HRs) were calculated. Mortality was considered the dependent variable, and length of stay was used as the time-to-event variable. Later, the proportional hazards assumptions were verified using Schoenfeld residuals.

Statistical analyses were conducted using R language version 4.3.2 (Vienna, Austria) on a Debian 12 GNU/Linux workstation. For all tests, the level of statistical significance was set at p < 0.05.

Ethics approval and consent to participate

This study was approved by the Ethical Board of Universidad Rey Juan Carlos (ID number 2610202334423). No identifying information was included in the manuscript. Because the authors used historical data, the Ethical Board considered informed consent was not necessary. So, the need for informed consent to participate was waived by our Ethical Board. All procedures involving human participants were conducted in accordance with the ethical standards of the responsible institutional and/or national research committee and with the 1964 Helsinki Declaration and its later amendments or comparable ethical standards.

Table 1 Characteristics of the studied cohort of patients with hematologic malignancies between 2020 and 2022 in Spain.
Full size table
Table 2 Correlation analyses of recipients of bone marrow transplantation.
Full size table
Table 3 Univariate and multivariate analyses of the cohort of HSCT recipients (binary logistic regression using mortality as the response variable).
Full size table
Table 4 Summary results for multivariate analysis using the Cox proportional hazards model predicting mortality.
Full size table

Results

Descriptive analyses

We collected data from 35,648 patients with HMs who were hospitalized with a COVID-19 diagnosis in 2020–2022. In this group, we identified 2,324 patients who had undergone an HSCT before or during the hospitalization. Table 1; Fig. 1 summarize the descriptive results of the cohort. During the 3-year period studied, we observed an increased number of hospitalizations among hematologic patients in 2022, with 17,342 hospital admissions (49% of all admissions) compared to 2020 and 2021. This increase was observed in patients both with and without HSCT (48% and 52%, respectively). There were more males than females in both groups, with the difference being greater among HSCT recipients (62% male). HSCT recipients were younger than non-recipients (median ages of 75 vs. 58 years old) and had a higher rate of ICU admissions. Mortality was significantly higher in those who did not undergo HSCT (20% vs. 13%, p < 0.001).

Fig. 1
figure 1

Distribution of admissions across the whole cohort (panel A) and in the subset of recipients of bone marrow transplantation (panel B). Panels C and D show the distribution of age between the two groups, displayed as distribution and boxplots.

Full size image

Regarding underlying HMs, non-Hodgkin lymphoma, acute leukemias, and chronic myeloproliferative disorders were more frequent in non-recipients, while multiple myeloma and malignant plasma cell neoplasms were more frequent in recipients. There were no differences regarding Hodgkin lymphoma between the groups.

Apart from chronic liver disease, chronic conditions were more prevalent in the non-recipient group.

Next, we analyzed the data for a subset of recipients that underwent HSCT, as shown in Table 2.

Among the 2,324 hospitalizations, there were 302 deaths (13%). Mortality was higher in 2022 than in the previous 2 years (118 individuals, 39%). The patients who died were older than those who survived (61 vs. 58 years old, p = 0.004). Men were more common in the two groups, regardless of survival status. ICU admissions, ICU length of stay, and hospitalization length of stay were also higher in patients who died, probably as surrogates of severity. Heart failure, dementia, and the presence of a solid neoplasm were highly correlated with mortality as comorbidities. Acute leukemias and chronic myeloproliferative disorders were the HMs that were more associated with mortality in HSCT recipients.

Figure 2 shows the evolution of mortality as both a time series plot (absolute count) and as a bar plot chart (percentage).

Fig. 2
figure 2

Evolution of in-hospital mortality of recipients of bone marrow transplantation between 2020 and 2022. Panel A shows a time series with deaths and survivors, while panel B shows the percentage of deaths among all hospitalizations over time.

Full size image

Univariate and multivariate logistic regression

We performed logistic regression to assess the effects of demographic and clinical characteristics, comorbidities, and HMs on in-hospital mortality due to COVID-19 in the cohort of patients with HM. In univariate analyses, HSCT status did not show a significant effect on COVID-19-related mortality (OR = 1.01, 95% CI: 0.89–1.14, p = 0.88). After adjusting for variables such as age, comorbidities, and type of hematologic malignancy in the multivariate analysis, HSCT was also not associated with an independent reduction in mortality (adjusted OR = 1.08, 95% CI: 0.94–1.24, p = 0.27). Then we developed a logistic regression model in the cohort of HSTC recipients only. This model evaluated the effect of age, sex, comorbidities, and HMs. Table 3 shows the results of univariate and multivariate analyses. Age and dementia were risk factors for mortality. ICU admission significantly predicted in-hospital survival (OR = 0.13, CI 0.09–0.19, p < 0.001), highlighting the protective effect on mortality. The presence of solid neoplasms and acute leukemias were also major risk factors for mortality (OR = 3.63 and OR = 1.72, respectively).

Survival analysis

Next, we explored predictors of in-hospital mortality. To evaluate time-to-event data, we conducted a survival analysis with the Cox proportional hazards model, as shown in Table 4. The dependent variable was in-hospital mortality, and the time-to-event was defined as the length of hospital stay in days. As is done in logistic regression, we included all demographic variables, clinical factors, comorbidities, and HMs. As mentioned in the Methods section, we applied stepwise regression including bidirectional elimination to refine predictor selection. Thus, the model automatically included the relevant variables and excluded the nonsignificant ones. As with the logistic regression model, we identified the following predictors for in-hospital mortality: age, dementia, solid tumors, and acute leukemias. Of note, ICU admission was again associated with a lower HR, reflecting that intensive care was a likely protective factor in severe cases.

Discussion

This nationwide study investigated the clinical and demographic profiles of HSCT recipients who also had an underlying HM and who were hospitalized during the COVID-19 pandemic. We also identified predictors for in-hospital mortality in this population. Our results provide insight into the patient characteristics of this population, and their risk factors for mortality, highlighting their vulnerabilities and elucidating their clinical course and outcomes.

Among the 2,324 HSCT recipients, the overall in-hospital mortality rate was 13%, notably lower than the 20% observed in non-HSCT patients. This difference highlights the complexity of HSCT patients, as this finding can be counterintuitive. When analyzed the entire cohort of 35,648 patients with HMs, our univariate and multivariate analyses did not show an independent protective effect of HSCT on COVID-19-related mortality. This finding contrasts with the lower mortality observed in transplanted patients compared to those who did not undergo HSCT (13% vs. 20%, p< 0.001). The apparent survival advantage in transplanted patients may be explained by selection factors, such as the inclusion of younger patients and those with fewer comorbidities in the HSCT group, rather than a direct effect of transplantation on immunity against COVID-19. Additionally, the lack of detailed data on the type of transplant (allogeneic vs. autologous) and pre-infection immunological status limits our ability to assess the specific impact of HSCT on clinical outcomes.

According to a recent narrative review, reports of COVID-19-related mortality among HSCT recipients are mostly based on single-center studies; those from multicenter studies involved a low number of individuals17. The review also indicated significant heterogeneity in patients, clinical settings, inclusion criteria, follow-up periods, and outcomes. Furthermore, most of these studies were conducted during the early pandemic, before vaccines and other specific treatments were yet available. Hence, the mortality rate among HM patients may be overstated. Nonetheless, it is clear that mortality in hematologic patients, regardless of HSCT status, is increased relative to the general population. However, there is evidence that morbidity and mortality outcomes may vary significantly between hematologic patients who do and who do not undergo HSCT. For example, Shah et al.18, found that, while HSCT recipients experienced high mortality (up to 22%), they had favorable outcomes compared to non-HSCT recipients. They interpreted their findings as being due to the adaptive immune response after HSCT against COVID-19. Thus, HSCT patients are able to modulate immune reconstitution following transplantation19,20. Although SARS-CoV-2 infection may cause lymphopenia in severe cases, HSCT recipients are able to recover T cells, in spite of marked lymphopenia and circulating B cells. Indeed, such patients can produce SARS-CoV-2 antibodies18,21. A later meta-analysis of 19 studies including 2,031 HSCT recipients reported a median age of 57 years, similar to our findings, and a mortality rate of 19%22.

In our study, we found that mortality was higher in 2022, albeit specifically in the first months of the year. This may reflect the evolution of the pandemic in terms of SARS-CoV-2 variants. The omicron subvariants BA.4, BA.5, and BA.2.75 were predominant in Spain beginning in the first months of 2022. They were reported to be more immune-evasive, and although they only caused mild disease in healthy individuals, they could cause severe disease in vulnerable populations, such as the elderly or immunocompromised, during these months6,23.

In line with previous studies, older age at diagnosis of COVID-19 was associated with poor outcomes24. We likewise found that dementia and leukemia were significant predictors of mortality, in both logistic regression and survival analyses. Dementia is associated with age, probably reflecting a certain degree of multicollinearity between these two variables. These comorbidities may be key determinants of poorer outcomes, not only in HSCT recipients, but also more generally in patients with HMs18,25.

We found that acute leukemias were also significantly associated with mortality, likely due to their aggressive clinical course and the profound immunosuppression inherent in their management26. Our results align with those of Sharma et al.14, who included 318 patients diagnosed with COVID-19 in a survival analysis and studied mortality during the first months of the pandemic. They found that solid tumors, plasma cell disorders (i.e., multiple myeloma and malignant plasma cell neoplasms), and lymphoma were relevant predictors of mortality. Our results also support Shah et al.18. , who also found that active malignancy, particularly leukemia, was associated with mortality in HSCT recipients.

In our study, ICU admission was also predictive of mortality in both logistic regression (OR = 0.13, 95%CI 0.09–0.19, p < 0.001) and in survival analysis (HR = 0.42, 95%CI 0.31–0.56, p < 0.001). However, it may act as a confounding factor rather than a direct protective factor, and our finding requires careful interpretation. We believe that this is likely to have been influenced by survivorship bias. That is, patients who survive long enough to be admitted to the ICU are younger or have a less severe case of the disease and may inherently differ from those who did not match ICU admission criteria. The latter group of patients might experience rapid clinical deterioration or death before ICU admission. Furthermore, ICU admission could act as a surrogate marker for the availability and intensity of advanced medical care, not as a factor that directly mitigates mortality risk. Thus, this finding should be interpreted with caution, and clinicians should consider their clinical decision-making and resource availability when interpreting the role of ICU admission in morbidity and mortality risk among HSCT recipients. Indeed, future studies would benefit from incorporating additional variables, such as the timing of ICU admission, the onset of clinical deterioration, the management of complications, and the immunological status of patients prior to assessing the role of ICU admission as a potential confounder versus a direct protective factor.

Limitations and strengths

The potential limitations of our study include the inherent characteristics of the registry data itself. Detailed immunological and treatment data were not available. We did not collect information on graft type, i.e., allogeneic and autologous HSCT. In addition, we lacked data on the year of transplantation, as these data were not available in the administrative dataset. Patients with graft-versus-host disease were not identified. More importantly, no information on the immunologic profile before the SARS-CoV-2 infection (CD4+/CD8 + count) or on the status of the HM (relapse or progression of disease after transplantation) were available. We could not assess the impact of specific treatments such as corticosteroids, antivirals, or immunosuppressive therapies on mortality.

The lack of data on immunologic status, the diversity of graft sources, and the distinct immunosuppression therapies employed may have impacted the interpretation of our findings, although we focused on trends over time and the description of HSCT recipients with HMs using available data from the national registry, reporting the results for a large cohort to fully characterize risk factors for mortality. Therefore, the strengths of our study include its nationwide scope and its large sample size. We also used a robust methodological approach. By analyzing this centralized administrative database, we captured comprehensive information to depict an overall picture and enhance the generalizability of our findings.

Nonetheless, further research is needed to assess the role of immunological recovery, graft-versus-host disease, other transplant-specific factors, and the biomarkers of immune function that may be associated with mortality due to COVID-19 infection to provide further insight into risk stratification for HSCT recipients.

Conclusions

We characterize the clinical and demographic profiles of the HSCT population among patients with HMs. Our findings provide demographic, clinical, and disease-specific factors that influenced outcomes in HSCT recipients during the COVID-19 pandemic. Although mortality in this group is higher than in the general population, the rate is lower than that for non-HSCT recipients, likely due to superior immunological status. Our findings are consistent with previous studies. They demonstrate the increased risks associated with all patients with HMs during the COVID-19 pandemic, even those who underwent an HSCT. The relatively favorable outcomes in HSCT recipients form a contrast with some early reports, but this discrepancy may be due to differences in study design, included population, and pandemic phase. Overall, our findings highlight the need for tailored management strategies and continued vigilance in patients who either have an HM or have undergone HSCT.