ABSTRACT
Background: New York City has been at the epicenter of the SARS-CoV-2 (COVID-19) pandemic.Immunocompromised cancer patients may be more vulnerable to COVID-related morbidity and mortality. Theobjectives of this study were to determine if patients with cancer have worse outcomes compared to their noncancercounterparts and to identify potential demographic and clinical predictors of morbidity and mortality among cancerpatients. Methods: We used data from a retrospective observational cohort of adult patients who tested positive for COVID-19 at New York-Presbyterian hospitals between March 3 and April 25, 2020. Patients with active cancer werematched 1:4 to noncancer controls on age, gender, and diabetes status. Using Kaplan-Meier curves and the log-rank test, we compared morbidity (intensive care unit admission and intubation) and mortality outcomes betweencancer patients and controls. We identified demographic and clinical predictors of worse outcomes using CoxProportional Hazard models. Hazard ratios and 95% confidence intervals were calculated for all estimates. Results: We included 445 COVID-19 positive adult patients of whom 89 had active malignancy. Among cancerpatients, the median age was 72 years, 54% were male, and 52% were non-white. Presenting symptoms weresimilar between cancer and noncancer groups. Nearly half of cancer patients were on active treatment includingcytotoxic and immunosuppressive therapy, and 40.9% of patients received cytotoxic treatment within 90 days ofadmission. Both patients with and without cancer received hydroxychloroquine in similar proportions (64% vs.65.5%), and more cancer patients received remdesivir (7.9% vs. 3.7%). Overall, age (HR 1.14;95% CI 1.00-1.29;p=0.049), male sex (HR 1.43;95% CI 1.04-1.96, p=0.07), dyspnea on presentation (HR 1.81, 95% CI 1.3-2.58;p=0.0005), and bilateral lung infiltrates (HR 1.94;95% CI 1.30-2.89;p=0.001) were associated with worseoutcomes. Observed complications were similar for cancer and noncancer patients, including myocardial infarction(3.4% vs. 4.2%), vasopressor requirements (24.7% vs. 26.2%), bacteremia (9% vs. 10.4%), and venousthromboembolic events (7.9% vs. 7.3%), respectively. There were no statistically significant differences in morbidityor mortality between cancer and noncancer patients (p=0.287). Conclusion: We demonstrate that COVID-19 hospitalized patients with active malignancies have comparablemorbidity and mortality to patients without cancer. In contrast to previous findings, we observed no differences in riskof ICU admission, intubation, or death between cancer and noncancer patients. Our findings suggest that activemalignancy may not be a contributive risk factor in comparison to other significant comorbidities that may be moreresponsible for the unfavorable prognosis of COVID-19 in cancer patients. We should consider the consequences oflimiting care for cancer patients on cancer-specific outcomes and mortality in the context of COVID-19.