Timely course of SARS-CoV-2 infections and vaccinations in patients with hemato-oncological diseases: analysis of a real-life cohort

Background The severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) pandemic has particularly impacted patients with hemato-oncological malignancies, as they showed not only a higher propensity for severe courses but also weaker immune responses after vaccination. Still, data on the influence of pandemic waves and vaccinations on outcomes are rare. This study aimed to analyze the timely course of infections and vaccinations in a real-life cohort of patients with hemato-oncological diseases. Methods In this cohort study, 1817 patients with hemato-oncological diseases from 1 February 2020 to 15 December 2022 at the ‘Franz Tappeiner’ Hospital in Merano/Meran, Italy, were followed for SARS-CoV-2 infections and vaccinations. Results Of 1817 patients with hemato-oncological malignancies, 735 (40.5%) were infected at least once with SARS-CoV-2, and 1614 (88.8%) received one or more doses of the approved vaccinations. Patients receiving antineoplastic treatment had a lower SARS-CoV-2 infection rate [35.1% versus 41.0%; odds ratio (OR) 0.78, 95% confidence interval (CI) 0.64-0.95], but higher risk of hospitalization (13.4% versus 6.9%; OR 2.11, 95% CI 1.25-3.69) compared with untreated patients. Overall, the case fatality rate (CFR) was 3.4%. Unvaccinated patients were more prone to severe coronavirus disease 2019 (COVID-19) courses requiring hospitalization (OR 2.34, 95% CI 1.25-4.36) and had a higher CFR (7.3% versus 1.6%; OR 4.98, 95% CI 2.16-12.98) than their vaccinated counterparts. In the Delta wave, patients with two vaccinations had a lower infection risk (OR 0.18, 95% CI 0.10-0.35) and tendentially lower hospitalization rates (OR 0.25, 95% CI 0.05-1.29) than unvaccinated patients. In the Omicron wave, 345/1198 (28.8%) patients with three or more vaccinations had breakthrough infections, resulting in a similar risk for infection (OR 0.88, 95% CI 0.60-1.30) but numerically lower risk for hospitalization (24/345, 7.0%) than unvaccinated individuals (4/40, 10.0%). Scheduled visits were postponed in 128/335 (38.2%) patients due to COVID-19, and deferrals correlated with pandemic wave (P = 0.002) and vaccination status (P < 0.001). Conclusions SARS-CoV-2 infections and outcomes differ between distinct phases of the pandemic. Vaccination with variant-specific vaccines should be prioritized as general protective measures are increasingly lifted.


INTRODUCTION
Since the initial outbreak of severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) in 2019, the coronavirus disease 2019 (COVID- 19) pandemic has posed significant challenges on healthcare systems in the world. Until February 2023, w750 million SARS-CoV-2 infections were reported worldwide, including 6.8 million fatalities. 1 Soon it was recognized that individuals with comorbidities have a higher risk for severe clinical courses and death after SARS-CoV-2 infection. 2 Here, immunocompromised individuals were shown to be at particularly high odds for complicated and potentially fatal disease. This includes patients after organ transplant as well as patients with hematological or solid malignancies, especially those undergoing systemic antineoplastic treatment. [3][4][5] Based on these considerations, patients with cancer were among the first being prioritized for SARS-CoV-2 vaccinations according to recommendations of major hemato-oncological societies in the world. 6,7 However, hemato-oncological patients were initially excluded from the large clinical trials leading to approval of several SARS-CoV-2 vaccines, [8][9][10] and data on immune responses after vaccination are mainly derived from real-life cohorts. [11][12][13][14][15] Here, most studies measured SARS-CoV-2-specific antibody levels in serum as a surrogate parameter for humoral immune responses; still, real-life outcome data are rare. [16][17][18][19] In initial reports up to 33% of patients with cancer experienced severe courses, 3 but these numbers have considerably decreased subsequently, 20 probably due to improved clinical management of COVID-19 and the availability of SARS-CoV-2 vaccines, but also less pathogenic variants of concern (VOC). Still, it is challenging to define the real-life efficacy of vaccinations in specific populations, such as patients with cancer, as data derived from large, unselected cohorts are limited. Moreover, the relative contribution of less pathogenic VOC and vaccinations to milder clinical courses is still not fully understood in hemato-oncological patients. By contrast, mild courses of COVID-19 may also lead to deferral of antineoplastic treatment and potentially worse oncological outcomes, underscoring the importance of ongoing protective measures for particularly vulnerable individuals. 20 This cohort study aimed to describe the timely course of SARS-CoV-2 infections and vaccinations in a large real-life cohort of patients with hemato-oncological disease. Moreover, we sought to analyze the impact of VOC and vaccinations on infections, hospitalizations, and fatalities in distinct phases of the COVID-19 pandemic.

Patient cohort
All patients with hematological diseases or solid tumors who were managed at the Hemato-Oncological Day Hospital Unit of the 'Franz Tappeiner' hospital in Merano/Meran, Italy, between 1 February 2020 and 15 December 2022 were included, regardless of whether they received antineoplastic treatment or not. Active anticancer treatment was defined as patients undergoing systemic treatment a maximum of 6 months before the infection (in line with national vaccination prioritization regulations). SARS-CoV-2 infections were diagnosed by rapid antigen testing and/or reverse transcriptase PCR of (naso-)pharyngeal swabs, where infection dates refer to the first positive testing result. Of note, SARS-CoV-2 testing was mandatory prior to each scheduled patient visit. Severity of COVID-19 was defined by treatment setting [home versus admission to non-intensive care unit (ICU) ward versus ICU]. Hospitalizations were only considered as such if COVID-19 was the reason for admission. Deaths and causes of death were recorded as documented in the official death registry. Deferrals of scheduled visits are defined as the date difference between the actual date of visit and the previously scheduled appointment. Data were retrieved retrospectively by chart review from electronic medical records.
Pandemic phases with regard to viral variants were defined based on Nextclade/Global Initiative for Sharing All Influenza Data (GISAID) data which aggregate reported results of SARS-CoV-2 genome sequencing. 21 Patients at risk in each pandemic wave were defined as all patients with a diagnosis of hemato-oncological disease before or during the respective pandemic phase.
All included patients were vaccinated according to national regulations. Administered vaccines included BNT162b2 (BioNTech/Pfizer), mRNA-1273 (Moderna), AZD1222 (Astra-Zeneca), Ad26.COV2.S (Janssen), NVX-CoV2373 (Novavax), and bivalent, variant-specific boosters BNT162b2 BA.1 and BNT162b2 BA.4/5 (BioNTech/Pfizer). To detect phase-specific changes dependent on vaccination status, outcomes for the Delta and Omicron waves were compared between unvaccinated patients at risk and those who have received two and three vaccinations, respectively, as all patients had been offered two doses by the beginning of the Delta wave and three doses by the beginning of the Omicron wave. In line with approval, one dose of Ad26.COV2.S was defined as equivalent to two doses of other vaccines.
All study procedures were conducted according to the Declaration of Helsinki with all applicable amendments as well as in compliance with institutional and national guidelines. The study was approved by the Institutional Ethics Review Board of the Südtiroler Sanitätsbetrieb (South Tyrolean Health Care Service, approval numbers 35/2020, 139/2021, 53/2022, 119/2022).

Statistical analysis
Independence of categorical data was assessed using the chi-square and Fisher's exact tests as appropriate. Distributions in metric data between groups were compared by applying ManneWhitney U or KruskaleWallis tests. A twosided P value of <0.05 was interpreted as statistically significant. Because of the hypothesis-generating study design, no correction for multiple testing was applied. 23 Statistical analysis was carried out using GraphPad Prism 9.5.0 (La Jolla, CA, USA) and R version 4.2.1 (The R Foundation for Statistical Computing, Vienna, Austria) using the packages ggplot2, lubridate, cowplot, viridis, and tidyr.

Disease severity in distinct phases of the pandemic
Absolute numbers of first documented SARS-CoV-2 infections according to vaccination status (unvaccinated versus 1 vaccination dose) and pandemic phase by predominant viral variant (as defined in the 'Patients and methods' section) are given in Figure 2A, whereas disease severity (ambulant versus hospital admission versus ICU admission) in the overall cohort is shown in Figure 2B. During the observation period, 331 patients died with 25 deaths attributable to COVID-19. Of these, 24 deaths occurred after the first and one death after the second SARS-CoV-2 infection. Overall, this translates to a case fatality rate (CFR) of 3.4% (25/734 infections after diagnosis of hemato-oncological disease) and 3.5% (24/680) after first infection. Notably, the CFR after first infection was significantly lower during Omicron waves (7/453, 1.5%) as compared with previous phases of the pandemic (17/227, 7.5%; OR 5.16, 95% CI 2.24-13.45; P < 0.001).

Impact of SARS-CoV-2 vaccination on infection rates and disease severity
Across pandemic phases beginning with the Alpha wave and the availability of SARS-CoV-2 vaccinations, unvaccinated patients were more likely to experience a clinical course requiring hospital admission (15/93, 16.1%) than their    By the beginning of the Delta and the Omicron waves, all patients were offered two or three vaccinations, respectively. Therefore, to further analyze the impact of (at the time) full vaccination regimens on SARS-CoV-2 infections and disease severity, we compared unvaccinated patients with those who had received two or three vaccination doses in the Delta and Omicron waves, respectively ( Figure 3A Figure 3A). Numerically, hospitalizations were less frequent in patients who had received two vaccination doses (2/21, 9.5%) compared with their unvaccinated counterparts (6/20, 30.0%; OR 0.25, 95% CI 0.05-1.29; P ¼ 0.130). Moreover, infection rates were similar in 2020 before general availability of SARS-CoV-2 vaccines (139/ 1329, 10.5%) compared with unvaccinated patients during the Delta wave (20/210, 9.5%; OR 1.11, 95% CI 0.68-1.80; P ¼ 0.679). Furthermore, there was a numerical trend toward lower hospitalization rates in the prevaccination era (23/139, 16.6%) compared with unvaccinated patients during the Delta wave (6/20, 30.0%; OR 0.46, 95% CI 0.16-1.36; P ¼ 0.145), suggesting similar transmissibility but higher pathogenicity of the Delta VOC compared with earlier phases of the pandemic.

Deferral of scheduled visits and antineoplastic treatment
Of all 691 patients who were infected at least once with SARS-CoV-2 after diagnosis of hemato-oncologic disease, scheduled visits including administration of antineoplastic treatment were postponed in 335/691 (48.5%) cases, of which 128/335 (38.2%) deferrals were related to SARS-CoV-2 infections.

DISCUSSION
Herein, we analyzed the time course of SARS-CoV-2 infections in a cohort of patients with hemato-oncological diseases, reflecting real-life outcomes of COVID-19 in an unselected patient population throughout the pandemic. Overall, w40% of patients were infected at least once over the observation period, and w90% in our cohort received at least one vaccination dose. By contrast, w55% of the overall regional population was infected, 24 and w78% of those eligible for vaccination received at least one dose until 23 December 2022. 25 These numbers underscore the high awareness of oncological patients and caregivers for COVID-19 and protective measures including vaccination. Indeed, our cohort also presented with a higher rate of hospitalization and fatalities than in the overall population, supporting previous reports of more complicated clinical courses in patients with cancer, especially those undergoing systemic antineoplastic treatment. 3,20,[26][27][28] In line with data on attenuated humoral immune responses after vaccination and previous clinical data, 26,[29][30][31] patients receiving the Bcell-targeting agent rituximab were particularly prone to SARS-CoV-2 infections and hospitalization in our cohort.
Because of the long observation period from February 2020 to December 2022, our study allowed to gain further insights into the impact of distinct pandemic phases on infections and their outcomes while local protective measures were kept continuously maintained. Indeed, infections, hospitalizations, and deaths due to COVID-19 differed considerably between pandemic waves in both the general population and in patients with cancer as previously reported. [32][33][34] However, increased infection rates and a lower risk for hospitalization for Omicron variants probably result from a combination of virus-and host-related factors including increased transmissibility and reduced pathogenicity, as well as enhanced immunity in the population by both vaccination and previous infections. Indeed, while vaccination remains a protective factor against severe COVID-19 in patients infected with the Omicron variant, previous data also suggest reduced hospital admissions in unvaccinated patients compared with the Delta VOC. 35,36 As still w12% of patients in our cohort were unvaccinated, comparisons according to vaccination status could be made in distinct phases of the pandemic. Indeed, we could find a tendentially higher risk for hospitalization in unvaccinated patients during the Delta wave compared with vaccinated individuals and the prevaccination period, suggesting both a real-life efficacy of vaccination and a higher pathogenicity of the Delta VOC compared with pre-Alpha variants. We could also observe high rates of breakthrough infections during the Omicron waves, emphasizing the immune-evading abilities of these VOCs. While lockdowns were increasingly lifted for vaccinated individuals, mobility restrictions remained in place for the unvaccinated population, 37 probably contributing to relatively high infection rates in vaccinated patients, although hospitalization rates and fatalities remained low.
In the first days of the pandemic and in light of shortages in healthcare resources, disruption in cancer care has been discussed and specific subgroups such as patients receiving adjuvant treatment have been prioritized for treatment continuation. 38 Later on, reasons for treatment deferrals increasingly shifted toward individual SARS-CoV-2 infections and postinfection sequelae, with significant impact on cancer-related survival. 20 In our cohort, more than one- third of treatment disruptions were due to COVID-19 with a median duration of almost 3 weeks, potentially affecting outcomes of oncological treatments. Whereas previous reports showed decreased severity for Omicron BA.1/BA.2 variants also in patients with cancer, 16,39,40 our data suggest even milder courses for Omicron BA.4/5, although increased vaccination coverage and the increasing application of bivalent, variant-specific vaccines may have contributed. However, infections still inevitably lead to delays in anticancer treatment also in the case of mild clinical courses, highlighting the urgent need for continuous vaccination efforts with VOC-specific boosters especially in vulnerable patient populations as currently discussed. 41 Clearly, this study has limitations which mainly lie in its retrospective design in an unselected patient cohort, leading to inherent heterogeneity of the population, small sample sizes in certain subgroups, and missing information, particularly regarding performance status as this may represent a further confounder impacting outcomes after SARS-CoV-2 infection. Moreover, we did not consider disease stage in hematological malignancies as these include very heterogeneous entities with no uniform staging system. Furthermore, the low number of patients, infections, and hospitalizations in certain subgroups precluded multivariate analysis to correct for confounding factors. Whereas data on vaccinations and infections have been systematically selected and patients underwent regular SARS-CoV-2 testing prior to scheduled visits, asymptomatic and undetected infections between visits cannot be excluded, especially in patients who had follow-up visits in larger intervals as compared with those undergoing active antineoplastic treatment. In addition, data on further SARS-CoV-2 medication such as antiviral agents including nirmatrelvir/ritonavir or molnupiravir have not been recorded. Lastly, the number of ICU admissions was relatively low; here, triage decisions in earlier phases of the pandemic with shortage of ICU resources may have contributed. This precluded further distinction to non-ICU wards when considering disease severity.
In conclusion, our study provides a bird's eye view on infection and vaccination dynamics throughout the pandemic, considering the impact of distinct viral variants and SARS-CoV-2 vaccinations, which both have contributed to vastly differing infection numbers and disease severity. Although the frequency of severe courses is decreasing, vaccination of patients with cancer with variant-specific boosters should be further supported as antineoplastic treatments are still postponed due to COVID-19, potentially leading to adverse oncological outcomes.