Cancer inpatient with COVID-19: a report from the Brazilian National Cancer Institute

Brazil has been recording a frightening exponential curve of confirmed cases of SARS-CoV-2 infection. Cancer patients with COVID-19 are likely to have a greater risk of complications and death. A retrospective search in the electronic medical records of cancer inpatients admitted to the Brazilian National Cancer Institute from April 30, 2020 to May 26, 2020 granted identification of 181 patients with COVID-19 confirmed by RT-PCR method. The mean age was 55.3 years (SD 21.1). The most prevalent solid tumors were breast (40 [22.1%]), gastrointestinal (24 [13.3%]), and gynecological (22 [12.2%]). Among hematological malignancies, lymphoma (20 [11%]) and leukemia (10 [5.5%]) predominated. The most common complications were respiratory failure (70 [38.7%]), septic shock (40 [22.1%]) and acute kidney injury (33 [18.2%]). A total of 60 (33.1%) patients died due to COVID-19 complications. By multivariate analysis, cases with admission due to symptoms of COVID-19 (p = 0.027) and with two or more metastatic sites (p <0.001) showed a higher risk of COVID-19-specific death. This is the first study in a cohort of Brazilian cancer patients with COVID-19. The rates of complications and COVID-19-specific death were significantly high. Our data prompts urgent and effective public policies for this group of especially vulnerable patients.


Introduction
The novel coronavirus, named severe acute respiratory syndrome coronavirus-2 (SARS-CoV-2) that causes the coronavirus disease 2019 (COVID-19) 1 , was first detected in Wuhan, the provincial capital of Hubei, China, in December 2019. SARS-CoV-2 has rapidly spread to many other countries worldwide becoming an unprecedented astounding and devastating pandemic in a short period of time.
Following an exponential upward trend, the increasing number of cases and death toll remain to concern the scientific community around the globe. Currently, more than 8.99 million cases are confirmed worldwide, with more than 469,500 deaths 2 . The first case of COVID-19 in Brazil was detected on February 26, 2020.
Standing out worldwide for having one of the steepest epidemiological curves, the country has already reached the second place in incidence with almost 1.08 million cases and second place in mortality with more than 50,500 deaths so far 3  Patients with cancer are more likely to have severe complications and even death when affected by COVID-19 [6][7][8] , mainly due to the effects of the immunosuppressive anticancer treatments, frequent use of corticosteroids, advanced age, comorbidities and pulmonary involvement (primary tumors or secondary lung metastases). Particularly in low-and middle-income countries, COVID-19 has brought a heavy burdening to the public health systems and induced new planning and adjustments in the clinical approach to cancer patients 9 .
. CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

(which was not certified by peer review)
The copyright holder for this preprint this version posted June 29, 2020. . https://doi.org/10.1101/2020.06.27.20141499 doi: medRxiv preprint Based on a few series previously published around the world, data evaluating the impact of COVID-19 outbreak in management and survival of patients with cancer are still very scarce, incomplete, with heterogeneous outcomes and descriptions [10][11][12][13][14][15] . Brazilian data in this specific field are still unknown due to the lack of publications.
The aim of this report was to describe the demographics, clinical characteristics and laboratory abnormalities of cancer inpatients with  admitted to the hospital ward of the Brazilian National Cancer Institute (INCA), exploring factors associated with death.

Study design and participants
This retrospective cohort was performed through a search on electronic medical records and compiled data of cancer inpatients admitted to INCA with laboratory-confirmed SARS-CoV-2 infection between April 30, 2020 and May 26, 2020. The hospital admission occurred for various medical reasons, including COVID-19 symptoms or any other clinical condition (for those cases with onset of symptoms throughout hospitalization or cancer inpatients who had contact to other COVID-19 cases). Outpatients tested positive for SARS-CoV-2 infection and patients with only non-invasive cancer (or pre-malignant conditions) were not the object of this study.
COVID-19 was diagnosed on the basis of the WHO interim guidance 16 , in which confirmation was defined as a positive result on real-time reverse transcriptase polymerase chain reaction (RT-PCR) assay of nasal-and oropharyngeal swab specimens using the U.S. Centers for Disease Control and Prevention (CDC) reagents and protocol 17 . Specimens were collected right after the hospital admission from those patients with COVID-19 symptoms and immediately after clinical suspicion from those admitted to hospital for diverse reasons unrelated to COVID-19.
The study was approved by the National Commission of Ethics in Research (CONEP) and conducted in accordance with the Good Clinical Practice guidelines.
. CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted June 29, 2020. . https://doi.org/10.1101/2020.06.27.20141499 doi: medRxiv preprint Written informed consent was waived due to the retrospective design and the emergency feature of the research. Only anonymized data was analyzed.

Data collection and outcomes
The demographic and clinical characteristics, including tumor site, histological subtype, staging, site of metastases, cancer treatment within the last 60 days, the presence of comorbidities, COVID-19-related clinical signs and symptoms, and laboratory tests at diagnosis and throughout hospitalization were obtained from the electronic medical records. COVID-19-specific clinical treatments were also collected. The variables analyzed in order to feature disease severity were admission to the intensive care unit (ICU), mechanical ventilation, renal failure, hemodialysis, septic shock, and death. Patients transferred out from INCA to another hospital were censored on the date of transfer. Patients who had not been discharged from hospital were censored in the date of the last follow-up on May 31, 2020.

Statistical analysis
The statistical software package SPSS, version 21.0 (São Paulo, Brazil) was used for the analyses. All continuous variables were evaluated by the Kolmogorov-Smirnov test of normality. Categorical variables were shown in percentages or absolute values. The study endpoint was COVID-19-related mortality. Time of follow-up was calculated from the date of swab collection to hospital discharge, death, or censorship of patients who were transferred or still hospitalized at the end of the study.
Risk factors for death were assessed using logistic regression. Crude and adjusted odds ratios (OR) were calculated. Variables with a p-value <0.20 at the univariate analysis were included in the multivariate model by stepwise forward selection with the entry order based on their level of significance. All p-values <0.05 were considered statistically significant.

Results
. CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted June 29, 2020. the need for mechanical ventilation. Ten patients (5.5%) were transferred out from . CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

(which was not certified by peer review)
The copyright holder for this preprint this version posted June 29, 2020. At the time of analysis, a total of 60 patients (33.1%) had died due to COVID-19. For solid tumors, the COVID-19-specific mortality rate was 37.7% (52/138) and for hematological malignancies (leukemia, lymphoma and multiple myeloma) was 23.5% (8/34). Four out of seven (57.1%) patients with lung cancer died from COVID-19, as well 52.5% (21/40) of breast cancer patients ( Figure 2B). Table 4, mortality related to COVID-19 was significantly associated to older age (p <0.001 for patients between 60 to 74 years and p = 0.002 for patients aged 75 years or older), metastatic cancer (p <0.001), two or more sites of metastases (p <0.001), the presence of lung (p <0.001) or bone metastases (p = 0.001), palliative or best supportive care intent (p <0.001), higher C-reactive protein levels (p = 0.002), admission due to COVID-19 (p = 0.009), and antibiotics use (p = 0.02). Isolated or combined comorbidities and elevated D-dimer levels did not demonstrate increased risk of dying from COVID-19.

As shown in
. CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

(which was not certified by peer review)
The copyright holder for this preprint this version posted June 29, 2020. . https://doi.org/10.1101/2020.06.27.20141499 doi: medRxiv preprint Different modalities of cancer therapy, including systemic agents (chemotherapy, hormone therapy, targeted therapy, immunotherapy), surgical procedures or radiotherapy within 60 days before COVID-19 were not associated with mortality. Also, specific therapies during the COVID-19 course, such as oseltamivir, therapeutic anticoagulation, corticosteroids, ivermectin and chloroquine did not influence the risk of death (data not shown).
According to the multivariate analysis patients admitted due to COVID-19 symptoms (p = 0.027) and with two or more metastatic sites (p <0.001) showed a significantly higher risk of COVID-19-specific death ( Table 5).

Discussion
The findings of this cohort highlight, in detail, several significant aspects of the COVID-19 course in patients already diagnosed with cancer. Although the emergency period for case selection was considerably short, due to the large number of cancer patients admitted to INCA with COVID-19, 181 patients were successfully included for analysis. Women had greater representation, more than half of the patients aged 60 years or older and almost a quarter of the patients had also reported smoking. Patients with two or more comorbidities accounted for more than a quarter of the study population as well, in which hypertension and diabetes prevailed.
Almost half of the patients (83 [45.9%]) were hospitalized due to conditions unrelated to the SARS-CoV-2 infection which can be explained by patients asymptomatic for COVID-19 having been admitted to hospital for other cancerrelated clinical complications. An intra-hospital transmission may also be considered, raising an important issue about the remarkable risk of infection to patients admitted for elective procedures. As for the symptoms present at diagnosis, similarly to data of other series 7,8,10,12,14 , in the current cohort, dyspnea, cough and fever were all highly frequent.
The odds of some COVID-related complications were quite similar to the findings reported by Kuderer et al. 14 in an international prospective series in which more than 928 patients were analyzed. The rate of ICU admission of 14% (versus . CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

(which was not certified by peer review)
The copyright holder for this preprint this version posted June 29, 2020. . https://doi.org/10.1101/2020.06.27.20141499 doi: medRxiv preprint 17.7% in the current study) and the mechanical ventilation requirement rate of 12% (versus 19.3% in the current study) were also alike in proportional terms. Zhang et al. 6 also showed paralleled data with respect to other variables such as the demand of supplemental oxygen in 78.6% of cases (versus 71.8% in the current study).
Early data from non-cancer patients suggested that 14-19% of cases progress with severe complications, such as septic shock, respiratory failure, acute kidney injury and multiple organ failure 5,15,18 . In the present study, these rates were much higher, ranging from 18.2 to 38.7%, highlighting the increased likelihood of severe complications in cancer patients. Conversely, cerebrovascular and cardiovascular events were less frequent.
In total, 69 (38.1%) of 181 patients died. Herein, COVID-19-related mortality was considered as the endpoint. Consequently, nine patients who clearly have recovered from COVID-19, and died due to other cancer-related reasons, were excluded from this mortality analysis. Therefore, the overall COVID-19-related mortality rate reached almost one third of the cases (60 [33.1%]), which was higher than that reported by other series with cancer patients 7,8,10,12,14 , and far exceeding the mortality reported for non-cancer patients 5 . It is important to point out that some patients had the definition of non-invasive support after or even before the diagnosis of COVID-19 due to the severity of their advanced malignancies, which may have overestimated the mortality rate. In addition to this, a noticeable number is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

(which was not certified by peer review)
The copyright holder for this preprint this version posted June 29, 2020. . https://doi.org/10.1101/2020.06.27.20141499 doi: medRxiv preprint greater number of metastases, pulmonary metastases, palliative or supportive treatment and symptomatic patients at hospital admission were significantly associated with a higher risk of death. In this same context, the type of anti-cancer treatment received by patients within the previous 60 days did not influence survival outcome.
Except for the association of C-reactive protein with mortality (OR 1.04; p = 0.002), none of the laboratory markers were likely to predict a higher risk of mortality. Other laboratory exams, including inflammatory markers such as lactate, ferritin, fibrinogen, and lactate dehydrogenase were not regularly collected in this cohort, preventing related analyses. A prospective study in order to evaluate the immune response markers in our cohort of cancer patients with COVID-19 is being currently conducted.
As in the study performed by Kuderer et al. 14 , none of the specific therapies prescribed such as antiviral oseltamivir (used for the initial suspicion of influenza infection), therapeutic anticoagulation, ivermectin or chloroquine influenced the risk of death in the current cohort. The strong association between the use of antibiotics and the outcome of death can be explained by the fact that these patients showed a more serious condition than COVID-19, including coinfections.
Some important limitations are also worth mentioning in this study. As a single-center cohort in a country of continental proportions, such as Brazil, a selection bias may well exist, hindering an external validity. The missing data rate for some variables was considerably high due to the retrospective design of the study. There was no paired sample with non-cancer patients with COVID-19 or cancer patients without COVID-19 to provide a better comparison between the outcomes of morbidity and mortality. Due to the in-hospital follow-up only, there was no report of long-term morbidity. Finally, the general population of the study was very heterogeneous with several types of neoplasia and anti-cancer treatment, making it difficult to design a more reliable portrait by tumor site.
Lastly, finishing on a positive note, some strengths of the current study can also be recognized. The Brazilian National Cancer Institute is the most important national reference center for the treatment of cancer patients through the Brazilian . CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

(which was not certified by peer review)
The copyright holder for this preprint this version posted June 29, 2020.

Data sharing
Our data are accessible to researchers upon request for data sharing to the corresponding author.

Acknowledgments
The authors thank all the INCA health workers for taking care of cancer patients with COVID-19. This work was supported by grants to LCST from CNPq (306798/2019-0), to MAS from CNPq (305765/2015-9) and FAPERJ (E-. CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

(which was not certified by peer review)
The copyright holder for this preprint this version posted June 29, 2020. . CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

(which was not certified by peer review)
The copyright holder for this preprint this version posted June 29, 2020. . CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.
. CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

(which was not certified by peer review)
The copyright holder for this preprint this version posted June 29, 2020. . https://doi.org/10.1101/2020.06.27.20141499 doi: medRxiv preprint . CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

(which was not certified by peer review)
The copyright holder for this preprint this version posted June 29, 2020. Missing or NA 11 (6.1) Total 181 *Patients may have more than one site of metastasis or receive more than one type of anticancer therapy. NA=not applicable.
. CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

(which was not certified by peer review)
The copyright holder for this preprint this version posted June 29, 2020. . https://doi.org/10.1101/2020.06.27.20141499 doi: medRxiv preprint . CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.
. CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

(which was not certified by peer review)
The copyright holder for this preprint this version posted June 29, 2020.

<0.001
. CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

(which was not certified by peer review)
The copyright holder for this preprint this version posted June 29, 2020. . CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

(which was not certified by peer review)
The copyright holder for this preprint this version posted June 29, 2020. . CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

(which was not certified by peer review)
The copyright holder for this preprint this version posted June 29, 2020. . CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted June 29, 2020. . https://doi.org/10.1101/2020.06.27.20141499 doi: medRxiv preprint CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted June 29, 2020. . https://doi.org/10.1101/2020.06.27.20141499 doi: medRxiv preprint  CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted June 29, 2020.  MulFple myeloma . CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted June 29, 2020. . https://doi.org/10.1101/2020.06.27.20141499 doi: medRxiv preprint