Connection Between COVID-19 and Mortality Among Cancer Patients

by Kesavan K.E.T. on Feb 23 2022 11:18 PM

Connection Between COVID-19 and Mortality Among Cancer Patients
More than 185 million people worldwide have been affected by COVID-19. There have been 4 million COVID-19–related deaths with more than 129,000 deaths in the UK alone.
Preliminary reports suggest that patients with cancer are twice as likely to die from COVID-19-related deaths. As a result, cancer services have made significant changes in treatment and follow-up strategies to reduce the risk of COVID-19 infection.

On March 18, 2020, the UK Coronavirus Cancer Monitoring Project (UKCCMP) was launched and their data identified that COVID-19 mortality in patients with cancer was primarily associated with age, sex, and comorbidities rather than recent systemic anticancer therapies (SACTs), a finding shared by other large prospective studies, and that tumor subtypes had differing mortality risks.

The study aimed to determine whether SACTs were associated with tumor subtype, demographic characteristics (age and sex), and morbidity and mortality using a large database of patients with cancer and COVID-19 infections associated with comorbidities.

Patient data were collected and were 18 years and older with active cancer. They and a clinical diagnosis of COVID-19. Patients were registered to this study by cancer centers when patients presented at hospital for either cancer treatment, which led to COVID-19 testing, or COVID-19 infection requiring treatment. Most data were collected and managed.

All 69 participating centers entered anonymized patient information. The information collected included ethnicity data, which was only entered if present in the patient’s medical record. Ethnicity information was collected following guidance from the UK government. All data were securely transferred and were curated following a uniform annotation process.

Patients undergoing active treatment fewer than 4 weeks prior to their COVID-19 test were classified as under active treatment. The study period was March 18 to August 1, 2020, and as such all patients were unvaccinated.

The UKCCMP was classified by the National Health Service Health Research Authority (HRA) as a public health surveillance project that did not require further ethical review or approval by the HRA. Overall, 2,515 patients with valid outcome, age, sex, and comorbidity data were included in all analyses unless otherwise stated.

The primary outcome of all-cause mortality was assessed by death during the primary hospitalization. A secondary outcome was COVID-19–specific death, for which COVID-19 was listed as a significant contributing factor within the same time frame. Data processing, statistical analysis, and visualization were performed.

Potential explanatory variables were analyzed using multivariable logistic regression adjusting for age, sex, and the presence of at least one of the following comorbidities: chronic kidney disease, chronic obstructive pulmonary disease, cardiovascular disease, diabetes, hypertension, or vascular diseases.

Of the 2,515 patients included in the study, 1,464 (58%) were male, and the median age was 72 years. Overall, 1,463 (58%) presented with at least one of the key comorbidities, and 272 (11%) were identified as Black, Asian, or another minority ethnic group.

A mortality rate of 38% was observed (966 patients); median follow-up for all-cause death was 7 (3-13) days. A total of 1,108 patients (44%) presented with mild symptoms of COVID-19, 701 (28%) with severe, 539 (21%) with critical, while 119 (5%) presented with no symptoms of COVID-19.

Overall, 493 of those who died (51%) presented with critical symptoms, compared with only 46 of those who survived (3%). A total of 131 patients (5%) were admitted to intensive care.

Exploratory analyses to test the evidence to support associations of patient characteristics, tumor characteristics, and treatment types with outcomes in a large cohort of UK patients with cancer were performed and a confirmed diagnosis of COVID-19 is included.

It was found that there was no evidence that recent chemotherapy (among 587 participants) was associated with increased all-cause mortality compared with patients who had received other anticancer treatment or no anticancer treatment.

The results were consistent with those obtained with the total population and the robustness of the results was confirmed by comparing patients with recent chemotherapy with propensity score–matched patients.

In an analysis of patients with solid cancer, chemotherapy (406 patients) was associated with lower all-cause mortality (chemotherapy vs. no chemotherapy and chemotherapy vs. no anticancer treatment). There were previous reports on an association between higher mortality and chemotherapy in patients with hematological cancer.

In this larger data set of 604 patients and following reclassification of immunomodulatory treatments into a separate category (58 patients), 181 patients with hematological cancer had received recent chemotherapy. In this group, it was no longer observed that a statistically significant association between chemotherapy and mortality when adjusting for age, sex, and comorbidities was present. Immunomodulatory treatment was associated with higher mortality. The association remained in an analysis of only patients with myeloma.

The association between recent chemotherapy and patient outcome was tested and stratified by cancer subtypes in multivariate analyses adjusted for age, sex (for non–sex-specific cancer types) and comorbidities. In the analysis of solid cancers, recent chemotherapy was associated with improved survival in patients with noncolorectal digestive organ, female genital organ, and breast cancers.

There was no association between recent chemotherapy and death for any individual hematological cancer type. They also tested the association between recent chemotherapy and patient outcome, stratified by cancer stage in a multivariate analysis of patients with solid cancer. Receiving chemotherapy was associated with lower mortality for patients with primary cancers as well as for patients with metastatic cancers. Nonpalliative chemotherapy (neoadjuvant, adjuvant, or radical) was associated with lower mortality compared with palliative chemotherapy. However, there was no significant association between chemotherapy and all-cause mortality in an analysis of only patients receiving palliative care.

There was also no association between higher mortality and multimodal SACT or combination chemotherapy compared with single-agent chemotherapy. They compared all-cause mortality of patients receiving other anticancer therapies in 4 weeks before testing positive for COVID-19 with patients who were not receiving that anticancer therapy in the same time period.

Surgery (93 patients) was associated with lower mortality compared with no surgery, while immunotherapy (102 patients), hormonal therapy (250 patients), radiotherapy (175 patients), and targeted therapies (255 patients) were not associated with higher mortality compared with no treatment with the same type of anticancer therapy. When compared with no cancer therapy, hormonal therapy (250 patients) and immunotherapy (102 patients) were significantly associated with lower mortality.

It was also observed that an association between less severe COVID-19 symptoms and immunotherapy exist compared with no cancer treatment. The association between patient outcome and recent immunotherapy (<4 weeks) in patients with different cancer subtypes was observed. While based on small numbers, no significant associations were observed in any single cancer type.

In patients with lung cancer receiving check-point inhibitor immunotherapy, chemotherapy, combination chemotherapy, and immunotherapy, targeted treatment, or radiotherapy in the 4 weeks before a COVID-19 diagnosis were not associated with mortality compared with no treatment with these therapies. Blood indices were also measured.

C-reactive protein levels greater than the median value, neutrophilia, lymphopenia, and neutrophil-to-lymphocyte ratio were significantly associated with higher mortality. Neutropenia was not significantly associated with higher mortality.

This UKCCMP data set provides a unique opportunity to investigate a population within one health care system, the NHS. The data will facilitate better risk stratification of patients with cancer who may be exposed to COVID-19 and will permit clinicians to devise individualized care plans with patients that minimize disruption to cancer care.

Overall, 38% of patients in UKCCMP died, and despite 49% of patients presenting with severe or critical COVID-19 illness, only 5% of patients were treated on an intensive care unit. These data suggest patients with cancer, COVID-19, and known comorbidities represent a particularly vulnerable group. Cancer teams need to work closely with intensivists and primary and secondary care teams to ensure patients with cancer are offered the appropriate level of treatment.

The studies claimed that there is no association between SACT and mortality and there was no association between anticancer treatment and all-cause mortality in patients with solid or hematological malignant neoplasms. An analysis based on propensity score matching confirmed the results identified using the regression models. Consistent with the lack of excess mortality associated with recent chemotherapy, a significant association between neutropenia and mortality was not observed.

The findings regarding chemotherapy are at odds with a recent report, which identified an association between recent chemotherapy and higher mortality, a finding with potentially significant policy and health care implications. Furthermore, missing clinical variables were imputed for more than 10% of hospitalized patients, unlike the UKCCMP.

Those without an active cancer diagnosis were included in the associations between chemotherapy and patient outcome. The outcomes of patients with historical cancers and COVID-19 infection are likely to be very different from those with active cancer. When a sensitivity analysis removing 70 patients (2.8%) in UKCCMP categorized as being on surveillance or in remission was performed, the association of recent chemotherapy and lower mortality became significant.

An association between high mortality and immunomodulatory drugs like lenalidomide, thalidomide, and pomalidomide used only to treat patients with myeloma was observed that was contradictory, and is known to have increased mortality following COVID-19 infection. The association remained in an analysis of only patients with myeloma. These findings need urgent validation in other data sets and are important given suggestions that their use might temper the inflammatory storm that drives severe COVID-19. It is necessary to understand the association of these drugs with antiviral immunity and efficacy of vaccination is also needed.

The analyzed data confirms the association between higher mortality and hematological malignant neoplasms and was consistent with UK primary care data. The COVID-19 immunological signature and postviral clearance immune state of patients with solid cancer is similar to the signature of people with COVID-19 infection but without cancer.

In contrast, patients with hematological cancer and COVID-19 have much less immune activation, high levels of CD8+ T-cell exhaustion, severe B cell cytopenia. and inconsistent antibody responses. The only solid cancer in the data set associated with significantly worse outcome was lung cancer. COVID-19 infection primarily affects the lung, and lung cancer often occurs in the setting of chronic tobacco smoke–mediated damage and reduced respiratory reserve.

Overall, 87% of the patients with lung cancer in this study who had available smoking history had current or former smoking habits. Smoke exposure has been associated with increased angiotensin-converting enzyme 2 expression, a key component of SARS-CoV-2 cell binding and entry. Importantly, however, they have found that there was no significant association between death and cytotoxic chemotherapy or immunotherapy in patients with lung cancer, which was consistent with other studies.


This study has also limitations that none of the associations observed between treatment and mortality would remain significant after multiple testing correction, with the exception of the analysis of palliative vs nonpalliative radiotherapy. The median follow-up of patients in this cohort was short, and these results did not consider patients who may have been discharged and then subsequently died. Performance status was not collected for this cohort. This is a serious limitation, but the impact is mitigated as much as possible by having information on patient comorbidities and adjusting for these in all analyses presented.

Finally, these findings of this study of patients with active cancer suggest that recent SACT is not associated with the outcomes from COVID-19 infection. This has relevance for the care of patients with cancer requiring treatment, particularly in countries experiencing an increase in COVID-19 case numbers.

The differences in outcomes among patients with hematological and lung cancers were observed. Recent chemotherapy was not associated with all-cause mortality, and recent immunotherapy was associated with less severe COVID-19 symptoms and lower mortality. The findings of this study of patients with active cancer suggest that recent SACT is not associated with inferior outcomes from COVID-19 infection.


Recommended Readings