Outcomes of Patients with Active Cancers and Pre-Existing Cardiovascular Diseases Infected with SARS-CoV-2

Objective To determine the impact of acute SARS-CoV-2 infection on patient with concomitant active cancer and CVD. Methods The researchers extracted and analyzed data from the National COVID Cohort Collaborative (N3C) database between January 1, 2020, and July 22, 2022. They included only patients with acute SARS-CoV-2 infection, defined as a positive test by PCR 21 days before and 5 days after the day of index hospitalization. Active cancers were defined as last cancer drug administered within 30 days of index admission. The “Cardioonc” group consisted of patients with CVD and active cancers. The cohort was divided into four groups: (1) CVD (−), (2) CVD (+), (3) Cardioonc (−), and (4) Cardioonc (+), where (−) or (+) denotes acute SARS-CoV-2 infection status. The primary outcome of the study was major adverse cardiovascular events (MACE), including acute stroke, acute heart failure, myocardial infarction, or all-cause mortality. The researchers analyzed the outcomes by different phases of the pandemic and performed competing-risk analysis for other MACE components and death as a competing event. Results The study analyzed 418,306 patients, of which 74%, 10%, 15.7%, and 0.3% had CVD (−), CVD (+), Cardioonc (−), and Cardioonc (+), respectively. The Cardioonc (+) group had the highest MACE events in all four phases of the pandemic. Compared to CVD (−), the Cardioonc (+) group had an odds ratio of 1.66 for MACE. However, during the Omicron era, there was a statistically significant increased risk for MACE in the Cardioonc (+) group compared to CVD (−). Competing risk analysis showed that all-cause mortality was significantly higher in the Cardioonc (+) group and limited other MACE events from occurring. When the researchers identified specific cancer types, patients with colon cancer had higher MACE. Conclusion In conclusion, the study found that patients with both CVD and active cancer suffered relatively worse outcomes when they had acute SARS-CoV-2 infection during early and alpha surges in the United States. These findings highlight the need for improved management strategies and further research to better understand the impact of the virus on vulnerable populations during the COVID-19 pandemic.


Introduction
Since the beginning of the SARS-CoV-2 pandemic in 2019, more than 625 million cases have been con rmed worldwide, and more than 6 million people have died from a SARS-CoV-2 infection and associated complications; more speci cally in the United States, over 95 million cases have been con rmed and more than 1 million deaths have been reported (https://covid19.who.int/[accessed 10.31.22]).Individuals with select comorbidities and from certain populations have been disproportionately affected by SARS-CoV-2 infection and associated complications. 1,2For example, patients with cancers are considered at greater risk to have poorer outcomes from SARS-CoV-2-related sequelae due to their immunocompromised state associated with chemotherapy, which suppresses the immune system and increases their risk for poor outcomes. 3imilarly, cardiovascular disease (CVD) is an important risk factor for mortality in patients infected with SARS-CoV-2. 4 Patients with lower cardiac reserve, increased susceptibility to arrhythmia, and the inability to augment cardiac output in response to increased stress during an acute illness event could result in poor outcomes.
The risk of severe SARS-CoV-2 infection could be greater in patients with coexisting cancer and CVD, leading, lead to poorer outcomes.In a retrospective study conducted at a single healthcare system, patients with cancer and CVD were at an increased mortality risk than those with either condition alone. 5A SARS-CoV-2 cardiovascular disease registry showed that the presence of CVD risk factors itself was not a predictor of mortality in patients with cancer, likely due to signi cant interaction with CVD and cancer risk factors. 6The major limitations of these studies were that they were done in the early phase of the SARS-CoV-2 pandemic when optimal management strategies were evolving.Therefore, we sought to investigate the SARS-CoV-2 infection outcomes inpatients with concomitant active cancer and CVD.We conducted a large, retrospective, registry-based study using data from the National COVID Cohort Collaborative (N3C) supported by National Center for Advancing Translational Sciences of the National Institute of Health (NIH).

Data Source
The N3C is an open science community database focused on analyzing patient level Electronic Health Record (EHR) data harmonized from 72 sites across the US. 7Brie y, the N3C includes inpatient and outpatient cases after January 1, 2020, with the potential for data at the hospital site before the case up to January 1, 2018.
Sites contributing patient-level data from the EMR are transferred to N3C, where the data is received and undergoes several steps to harmonize from into OMOP concepts and perform quality checks.A data use agreement with the University of West Virginia and N3C allows for access to the de-identi ed data in the N3C data Enclave.The research was reviewed by the UWV IRB (Protocol# 2101218288) and granted exemption status.All authors who performed analyses and had access to N3C data in the Enclave obtained individual institutional review board approvals from their respective institutions for this project and were also approved to use a limited data set.N3C approved the research as limited data set (level 3) by the N3C Data Use Request Committee, allowing access to actual dates of patient care interventions.

Patients, data, and outcomes
Patients in the N3C database were identi ed by searching for patients captured in predetermined concept sets created using the International Classi cation of Diseases, tenth revision (ICD-10) codes, RxNorm, and LOINC codes on the N3C platform.Patients with a history of CVD and a diagnosis of an active oncologic condition, and whether they were con rmed to have a positive (or negative) SARS-CoV-2 infection between January 1, 2020 and July 21, 2022, de ned as RT-PCR positive (or negative) test for SARS-CoV-2 within 21 days before or 5 days after the index admission.Cancers that were excluded included non-melanoma skin cancers, benign tumors, or unknown types.We de ned patients with active cancer if they received chemotherapy within 30 days before or after the index admission.Four cohorts of patient groups were de ned: 1) Patients with CVD but SARS-CoV-2 negative (CVD (-)), (2) Patients with CVD and active cancer but SARS-CoV-2 negative (Cardioonc (-)), (3) CVD and SARS-CoV-2 positive (CVD (+)), and (4) CVD and active cancer and SARS-CoV-2 positive (Cardioonc (+)).The concept sets also collected patient demographics, characteristics, comorbidities, and antineoplastic agents.The primary outcome of this analysis was major adverse cardiac events (MACE), composed of acute heart failure, stroke, myocardial infarction, and death as all-cause mortality.The outcomes of interest were evaluated over time for each cohort.Additional sub-group analyses of the outcomes were conducted based on the cancer type, for lung, breast, prostate, colorectal, and hematological cancers separately.To account for differences in the outcome as treatments became available and care approaches evolved, we also sub-grouped the comparison of patient outcomes into categories of SARS-CoV-2 variants (early, alpha, delta, and omicron) by each period of time when different variants of SARS-CoV-2 were predominant in the US.We de ned Early SARS-CoV-2 from January 1, 2020 to February 13, 2021; Alpha from February 14, 2021 to July 6,2021; Delta from July 7, 2021 to January 08, 2022 and Omicron January 09,2022 to July 21, 2022 (last date of inclusion).We created the speci c date cutoffs based on the prevalence of waning strain no more than 10% (https://nextstrain.org/)

Statistical Analysis:
We summarized patient characteristics using descriptive statistics such median and quartile or means and standard deviations, proportions, and included summary tables, bar-plots, box-plots, as appropriate.Chisquare test was used in the data analysis for categorical variables while Wilcoxon rank-sum test or Student's t-test were used in the data analysis for continuous variables.Overall survival was studied using the Kaplan-Meier method and compared across groups using the log-rank test.In the multivariable-adjusted data analysis, the logistic regression model was used to assess odds ratios for categorical outcome variables such as MACE (yes/no), while the Cox proportional-hazards model was used to assess the overall survival and hazard ratios for comparison between SARS-CoV-2 negative and SARS-CoV-2 positive among CVD and Cardioonc respectively, adjusting for patient baseline characteristics.Because death was a terminal event for other outcomes such as stroke, acute heart failure or MI and because there were more deaths in the sicker patient subgroups, the cumulative incidences from each component in MACE including stroke, acute heart failure and MI were estimated and compared between different disease groups using competing risk analysis adjusting for the death -the terminal event.All p-values presented are 2-sided, and a p-value < 0.05 implies the statistical signi cance of this study.All analyses were conducted using Statistical software R. The competing risk analysis was carried out with the function in the cmprsk package.
We also estimated the survival probability for the major cancer types (lung, breast, prostate, colorectal, and hematological) in the Cardioonc (+) patients.Kaplan-Meier curves were used to visualize the corresponding survival probability, and the log-rank test to test statistical differences in survival probability between SARS-

Prevalence of MACE and Death
We examined the prevalence of death and MACE in each of the four cohorts, both overall and strati ed by periods of time when various SARS-CoV-2 variants were dominant (

MACE events over time
The Kaplan Meier curve of probability of MACE over time is presented in Figs.1A-1B, showing signi cant differences between the four cohorts.Cardioonc (+) patients had the highest risk of MACE compared to the other groups (Fig. 1a).When we compared across cancer types, risk of MACE events was highest among persons who had colorectal cancer and had SARS-CoV-2.

Subgroup Analysis
We examined risk of MACE in subgroups based on periods of dominant SARS-CoV-2 variants in a survival analysis, which showed differences between groups (Figs.2a-2b).During the early SARS-CoV-2 period, the risk of MACE was highest among Cardioonc (+) patients (Fig. 2a).Across all SARS-CoV-2 variant periods, there were differences in risk of MACE; in the early SARS-CoV-2 and Delta variant periods, risk of MACE was highest among Cardioonc (+) patients (Fig. 3-5).

Competing Risk Analysis
As shown in Table 4, When we accounted for death as a competing risk on MACE with CVD (-) as a reference, we found that the patients in the Cardioonc (-) cohort had reduced HRs for stroke (HR:0.79;p < 0.001), acute heart failure (HR:0.76;p < 0.001), and myocardial infarction (HR:0.73;p < 0.001).In the Cardioonc (+) cohort, HR was not signi cant after including death as a competing risk for all three MACE components.In the CVD (+) cohort, had a higher hazard ratio for myocardial infarction (HR: 1.06; p < 0.001) despite accounting for deaths in the cohort.

Discussion
The aim of our study was to report MACE outcomes in patients who had a cardiovascular diseases diagnosis with or without concomitant active cancers and, with or without documented SARS-CoV-2 infection.The N3C database, a large aggregate of patient data from 72 hospital EMRs across the US, provided a reliable and valuable dataset to address the knowledge gap.These results suggest that the SARS-CoV-2 infection, irrespective of cancer status, was associated with a higher rate of MACE; however, patients with active cancer with underlying cardiovascular conditions who were hospitalized with acute SARS-CoV-2 infection had poorer outcomes.The outcomes were signi cantly different amongst the groups earlier in pandemic, but there was no signi cant difference by cancer or CVD status in the later phases of pandemic (Central Illustration).When we looked at the most common cancer types, all cancers in the early course had similar probabilities for the events, but the patients with colon cancer had the worse intermediate term outcomes.
A study published by Ganantra and colleagues reported that the cardiooncology patients with SARS-CoV-2 had poorer outcomes than those with COVID-19 infection but did not have concomitant cancer and heart diseases. 5They observed that patients with active cancer and CVD are at elevated risk of MACE.However, the study included only four centers from Massachusetts, USA, and only included patients in the early phase of the pandemic.In our study, our data suggest similar results for patients admitted to the hospital in the early pandemic, while MACE outcomes were not signi cantly different for later phases of the pandemic (Table 3).
The lack of knowledge about SARS-CoV-2 and its management, scarcity of hospital resources, training and availability of staff, the severity of viral strains, and unavailability of vaccinations are possible explanations for worse outcomes in the early pandemic.
Patients with active cancers are immunosuppressed, making them susceptible to infections given their attenuated immune response.Several studies involving patients with cancer and SARS-CoV-2 infection have shown that these patients are vulnerable to infections and poor outcomes. 8Earlier reports have shown that patients with CVD or CV risk factors have a higher case-fatality rate than those without CVD from SARS-CoV-2 infection. 9,10When we compared Cardioonc patients with or without SARS-CoV-2 infection, the presence of SARS-CoV-2 was associated with worse outcomes.The Cardioonc patients are even more prone to blunted immune response and multi-organ failure are less likely to hemodynamically compensate in the setting of active infection. 11en we considered death as a competing risk for stroke, acute heart failure and acute myocardial infarctions, the patients in the Cardioonc (-) cohort had a reduced risk for all three events.These ndings suggest that death served as a competing risk for cardiovascular events from occurring, when compared patients who were in CVD (-) cohort.From our database, however, we cannot elucidate exact causes of death.
CVD (+) patients had higher HR of myocardial infarctions, but reduced HR for stroke and acute heart failure events.In other words, death precluded the occurrences of stroke and acute heart failure in the CVD (+), while deaths precluded all three events were in the Cardioonc (-) cohort.In contrary, HR did not change signi cantly for Cardioonc (+) even after accounting for deaths.The most likely explanation is that the majority of patients in Cardioonc (+) cohort suffered deaths, and therefore did not experience other events.In all four cohorts, the events were mainly driven by deaths, which is an important highlight of our study.
At the beginning of the pandemic, early 2020, when nationwide shelter in pace, social distancing and other preventive measured were initiated, necessary cardiac procedures decreased signi cantly from pre-pandemic levels, which further in uenced the outcomes in both cardio-oncology and CVD patients with active SARS-CoV-2 infections compared to their SARS-CoV-2 negative counterparts. 12As the pandemic progressed, easing restrictions for cardiac testing and procedures, availability of vaccination, and diminishing strength of SARS-CoV-2 viral strains may have eliminated the differences in outcomes by CVD or active cancer diagnosis in the later phases of the pandemic.The N3C database does not reliably capture the vaccine status of patients if the vaccines were administered outside of the healthcare systems.Therefore, we cannot draw any conclusion regarding the effectiveness of vaccinations in our study population.4][15] Thus, the increasing SARS-CoV-2 vaccination over time likely in uenced the SARS-CoV-2 outcomes later in the pandemic.A study reported by Tehrani and colleagues reported the ndings that differ from our study, where the authors did not see increased risk of MACE in patients with cancer and cardiac conditions as compared to patients with cancer alone. 6Possible explanations for the discrepancies are the inclusion of patients with a history of cancer vs. active cancer status in our study, de ned as patients with a cancer diagnosis and recent use of antineoplastic agent use; and the inclusion of all phases of pandemic in our study vs. Limited pandemic phase in the previous study.
We found that patients with active colorectal cancer at the time of infection had a higher cumulative probability of MACE than lung, breast, prostate, and hematologic malignancies.Interestingly, in the patients without acute SARS-CoV-2 infections, patients with lung cancers had a higher cumulative probability of MACE events.In a Veterans Affairs Healthcare System study, these authors reported poorer outcomes in patients with hematologic malignancies and infected with SARS-CoV-2 patients. 16In another study, liver and pancreatic cancers were associated with higher mortality rate. 17The exact mechanism behind the ndings is unclear and may need further study.Regardless of cancer subtypes, the patients with cancer and active SARS-CoV-2 infection had higher cumulative probabilities of MACE in the acute phase of infection, and sustained at 20 month follow-up.

Limitations
Despite the fact that N3C provides robust data to study SARS-CoV-2 infected patients, there are inherent limitations associated with administrative-style databases.The database does not provide severity of cardiovascular conditions.We cannot ascertain the details of metastatic and functional status from the information available in the database.We observed that there are signi cant discrepancies amongst the availability of variables, including laboratory and vital signs data that were missing for many organizations, and therefore, we chose to include or analyze incomplete and missing data.All authors agreed not to de ne a conditional based on laboratory or biomarker values (i.e., de ning acute myocardial infarction based on troponin elevation), and restricted the case de nition to the presence of speci c diagnoses codes.Follow-up data outside of the participating institutes was not available which introduced bias related to missing followup information.De nitions of active cancer varies across literature; however, we restricted our cohort only to cancer therapy within 14 days of index hospitalization to ensure that there is no inclusion of patients in cancer remission and to capture people with active cancer treatment.

Conclusion
Patients with acute SARS-CoV-2 infection in the early phase of pandemic suffered higher rates of MACE.The Cardioonc patients with active SARS-CoV-2 infection had the poorer outcomes.The effect of acute SARS-CoV-2 on MACE mitigated after early SARS-CoV-2 phase in all four cohorts.

Declarations
Authorship: Authorship was determined using ICMJE recommendations.
Funding The project described was supported by the National Institute of General Medical Sciences, 5U54GM104942-04.The content is solely the responsibility of the authors and does not necessarily represent the o cial views of the NIH.
N3C DUR ID: Figures  A and B: Major Adverse Cardiac Events (MACE) over time by study cohort (Figure 4A) and cancer type (Figure 4B) during the Delta period A and 5B: Major Adverse Cardiac Events (MACE) over time by study cohort (Figure 5A) and cancer type (Figure 5B) during the Omicron period

Figure
Figure 4

Table 3
Odds of Major Adverse Cardiac Event (MACE) from multivariable logistic models, adjusting for the baseline data including age, gender, BMI and other available patient characteristics.

Table 4
Hazard Ratios (HR) Accounting for Death as a Competing Risk from multivariable Cox models, adjusting for the baseline data including age, gender, BMI and other available patient characteristics.