Study population and setting
The primary objective of the study was to evaluate whether there were any differences in all-cause mortality, or hospital discharge, among COVID-19 patients with cancer who were on anticancer treatment (cytotoxic chemotherapy or other therapies) compared to cancer patients who were not on cancer treatment. The study includes the first 800 cancer patients with symptomatic, laboratory-confirmed SARS-CoV-2 infection (via RT-PCR), enrolled into the UK Coronavirus Cancer Monitoring Project (UKCCMP). The UKCCMP began prospectively following active cancer patients with COVID-19 at 55 cancer centers throughout the UK on March 18, 2020. The study population included individuals with metastatic cancer or those who received anticancer treatment or surgery with cytotoxic chemotherapy or radiotherapy in the previous 12 months. Patients were followed through April 26, 2020.
Summary of Main Findings
The cohort included 800 cancer patients with COVID-19 (median age 69 years; 44% female). The most common cancer types involved digestive organs (19%), hematological (14%), and breast (13%). Of these, 43% had metastatic cancer, while 19% had primary localized tumors. The vast majority (79%) of participants had other comorbidities apart from cancer including cardiovascular disease, hypertension, and diabetes. Just over half (52%) of the cohort had mild COVID-19, 23% had severe COVID-19, and 22% had critical COVID-19, according to WHO guidelines. During follow-up, 226 patients (28%) died, and 93% of those deaths were principally attributed to COVID-19. Patients who died were older, more likely to be male, and had higher rates of comorbidities (cardiovascular disease, COPD, diabetes, hypertension). In multivariable analysis (adjusting for age, sex, and comorbidities), receipt of cancer treatment in the 4 weeks prior to COVID-19 diagnosis was not associated with mortality.
The study was a multi-center nationwide prospective cohort of cancer patients with laboratory-confirmed COVID-19 in the UK. The data collection process was rigorous, in real time, and included important clinical and demographic variables that were adjusted for in the analyses.
Selection bias was possible, given that symptomatic individuals requiring medical attention may be overrepresented, which may in part, explain the high overall mortality. Confounding by indication was also a possibility: for example, cancer patients well enough to receive recent treatment may have been more likely to survive COVID-19. The study did not include COVID-19 cases with radiological or clinical diagnosis (sans a positive RT-PCT); because RT-PCR can produce false-negative results, the study may have incorrectly excluded some patients. Additionally, the study stratified cancer treatment as receipt within previous 4 weeks vs. not in the past 4 weeks; this time frame may be too wide to account for the impact of recent cancer therapy on COVID-19 incidence or severity (e.g., patients treated 3 weeks ago are categorized the same as those treated 1 week ago). The study does not compare COVID-19 mortality among cancer patients to that of the general population, and the true burden of SARS-COV2 among cancer patients can also not be estimated with these data. Small sample sizes within types of cancer and cancer treatments limited the power to detect associations with mortality in sub-group analyses.
Before this study, there was mixed evidence from small studies on the association of cancer treatment and COVID-19 morbidity and mortality. This was the largest prospective cohort study to date evaluating whether chemotherapy or other anticancer treatment are associated with higher mortality among cancer patients with COVID-19.
This review was posted on: 9 June 2020