Secondary infections worsen the outcome of COVID‐19 in patients with hematological malignancies: A report from the ITA‐HEMA‐COV

Abstract The impact of secondary infections (SI) on COVID‐19 outcome in patients with hematological malignancies (HM) is scarcely documented. To evaluate incidence, clinical characteristics, and outcome of SI, we analyzed the microbiologically documented SI in a large multicenter cohort of adult HM patients with COVID‐19. Among 1741 HM patients with COVID‐19, 134 (7.7%) had 185 SI, with a 1‐month cumulative incidence of 5%. Median time between COVID‐19 diagnosis and SI was 16 days (IQR: 5–36). Acute myeloid leukemia (AML) and lymphoma/plasma cell neoplasms (PCN) were more frequent diagnoses in SI patients compared to patients without SI (AML: 14.9% vs. 7.1%; lymphoma/PCN 71.7% vs. 65.3%). Patients with SI were older (median age 70 vs. 66 years, p = 0.002), with more comorbidities (median Charlson Comorbidity Index 5 vs. 4, p < 0.001), higher frequency of critical COVID‐19 (19.5% vs. 11.5%, p = 0.046), and more frequently not in complete remission (75% vs. 64.7% p = 0.024). Blood and bronchoalveolar lavage were the main sites of isolation for SI. Etiology of infections was bacterial in 80% (n = 148) of cases, mycotic in 9.7% (n = 18) and viral in 10.3% (n = 19); polymicrobial infections were observed in 24 patients (18%). Escherichia coli represented most of Gram‐negative isolates (18.9%), while coagulase‐negative Staphylococci were the most frequent among Gram‐positive (14.2%). The 30‐day mortality of patients with SI was higher when compared to patients without SI (69% vs. 15%, p < 0.001). The occurrence of SI worsened COVID‐19 outcome in HM patients. Timely diagnosis and adequate management should be considered to improve their prognosis.

documented SI in a large multicenter cohort of adult HM patients with COVID-19.
Among 1741 HM patients with COVID-19, 134 (7.7%) had 185 SI, with a 1-month cumulative incidence of 5%. Median time between COVID-19 diagnosis and SI was 16 days (IQR: . Acute myeloid leukemia (AML) and lymphoma/plasma cell neoplasms (PCN) were more frequent diagnoses in SI patients compared to patients without SI (AML: 14.9% vs. 7.1%; lymphoma/PCN 71.7% vs. 65.3%). Patients with SI were older (median age 70 vs. 66 years, p = 0.002), with more comorbidities (median Charlson Comorbidity Index 5 vs. 4, p < 0.001), higher frequency of critical COVID-19 (19.5% vs. 11.5%, p = 0.046), and more frequently not in complete remission (75% vs. 64.7% p = 0.024). Blood and bronchoalveolar lavage were the main sites of isolation for SI. Etiology of infections was bacterial in 80% (n = 148) of cases, infected by SARS-COV 2 have been reported higher than in individuals without HM. 1 Several studies have been published describing the clinical characteristics of COVID-19 in HM patients and reporting an overall risk of death (ranging from 10% to 40%), [1][2][3][4][5] substantially higher than in patients with solid tumors. [5][6][7] Risk factors for mortality vary in different studies; however, COVID-19 severity, presence of comorbidities, age, phase of treatment and status of progressive disease are reported as the most relevant factors impacting on outcome. [7][8][9] It is known that infections of viral origin are often complicated by other pathogens able to impact the outcome. For instance, H1N1 influenza pandemic in 2009, but also other respiratory viruses, have been reported with variable percentage as associated to secondary pulmonary infections. 10,11 Additionally, in the current pandemic, secondary infections (SI) have been reported as a challenging issue influencing survival in critically ill patients. 2 HM patients are intrinsically immunocompromized due to the underlying disease and to the treatment; moreover, the high rate of hospitalization besides chemotherapy induced-neutropenia and impaired cellular-mediated or humoral response increases the risk of additional infections. 3,12 There are some convincing data concerning the contribution of SI to worsen the outcome of COVID-19 as reported in a comparison between survivors and non survivors of a large real-life study on HM patients. 7 However, little is known about the type and the evolution of SI during COVID-19 in HM patients. The underlying immune system impairment can largely vary among diseases also in relation to the degree of immune control and to the status of the HM. With regards to the type of HM, acute leukemia patients have a higher risk of bacterial and mycotic infections, particularly in profound neutropenia periods and patients with lymphoproliferative diseases receiving immunotherapy depleting B-or T-cell compartments can present viral reactivations. 5,6 In this ambispective study we present the available data on infectious events secondary to COVID-19 in the large population of HM patients enrolled in the ITAlian HEMatology Alliance on COVid-19 (ITA-HEMA-COV) project. The project already provided evidence on COVID-19 mortality 7 and seroconversion. 13 With this analysis, we aimed to assess the impact of SI on the outcome of patients with HM and COVID-19. This study is registered with ClinicalTrials.gov, NCT04352556, and the prospective part of the study is ongoing.

| Procedures
Data on type and site of SI, laboratory parameters and patient outcome were collected for all participants since the date of the diagnosis of SI. Data on patient characteristics and outcomes were extracted by study investigators from electronic medical records or clinical charts, including age, sex, Charlson Comorbidity Index (CCI), type and status of HM, time since diagnosis of COVID-19 to date of SI diagnosis and COVID-19 severity.
Diagnosis of HM was made based on WHO classification of hematopoietic tumors. 14 We defined a patient as having progressive disease in case of malignancy not responding to active therapy and remission as no evidence of disease; partial remission was defined as reduction of tumor burden during or after active treatment while stable disease if a stability during the time, without active treatment was present.
We categorized as "recently treated" patients who had received chemotherapy within the previous three months since COVID-19 diagnosis, or immunotherapy/immunochemotherapy/biologic treatment within six months and as "on treatment" if the last hematological therapy was still ongoing at the time of diagnosis of COVID-19. Polymicrobial infection is defined when more than one type of organism were isolated from one or more blood cultures within a 72 h period. Bacteremia by coagulase-negative staphylococci (CoNS) and by Corynebacteria was considered only when supported by at least two positive blood cultures. Gram-positive and Gramnegative isolates were considered multi-drug resistant (MDR) according to Magiorakos et al. 16 Proven or probable invasive fungal infections (IFI), classified according to EORCT criteria, 17 were registered. We considered probable IFI as microbiologically documented if culture isolate or serum/bronchoalveolar lavage (BAL) galactomannan (GM) positivity was available. Information about antibiotic resistance profile of isolated bacteria were collected, if available.

| Endpoints
The primary endpoint of this study was to define the incidence of SI in HM patients with a diagnosis of COVID-19. Secondary endpoints were: (1) assessment of the association between patients' clinical characteristics at COVID-19 diagnosis and the occurrence of SI; (2) description of all infectious events; (3) evaluation of the impact of SI on the 30 days-mortality and on the overall survival (OS).

| Statistical analysis
Qualitative variables were described as counts and percentage of each category. Quantitative variables were summarized as median and interquartile range (IQR). Association between two qualitative variables was tested via Fisher's exact test. Wilcoxon test was used to compare quantitative variables between two groups of patients.
The cumulative incidence (reported together with its 95% confidence interval, 95%CI) was calculated from COVID-19 diagnosis to secondary infection or last-follow-up and was estimated with a competing risks approach, analyzing death without infection as a competing event.
The association between baseline characteristics and the occurrence of SI was evaluated with univariable and multivariable Fine and Gray models. Variables with a p-value<0. 2 A population of 1741 patients with HM and symptomatic COVID-19 was enrolled in this study. Among them, 134 patients (7.7%) had a SI with a 30-day cumulative incidence of SI of 5% (95%CI: 4.0%-6.2%).

| Characteristics of the patients
The baseline characteristics of patients with or without SI are reported in Table 1. Overall, we found differences in the distribution of HM between patients with SI and those without SI (P < 0.001). When only patients "recently treated" or "on treatment" were considered, no difference in cumulative incidence of SI was found according to type of last HM therapy (p = 0.079), also after adjusting for line of therapy (p = 0.092).

| Impact of secondary infections on outcome
Considering the entire cohort, after a median follow up of 1.  times in patients with SI compared with those without SI (HR = 2.9, 95%CI: 2.2-3.9, p < 0.001) (Supplemental Table 2).  SI were more frequent in recently treated patients; however, we did not find a type of treatment favoring SI more than the others.

| DISCUSSION
As expected and in line with the literature, 8  The finding of high frequency of Gram-negative isolates is in line with the most recent epidemiologic reports observed in HM patients. 8,21 Fungal infections were all invasive and they represented 9.7% of all SI, more than reported by Garcia-Vidal et al. 2  Our study has some limitations. First, due to its partial patients. Moreover, even if SI mainly developed after hospital admission we cannot exclude that a number of patients developed SI without being hospitalized. Another possible limit, also observed by other authors, 19 could be the lack of available complete and consecutive information of all infectious complications, particularly during the first wave, immediately after the pandemic outbreak, due to the hard management of the health care system.
In conclusion, SI severely impact on survival of COVID-19 HM patients; therefore, in these patients, from the early phases of COVID-19, great attention should be paid to diagnosing a potential SI and to starting treatment promptly to minimize the risk of death.