The COVID-19 pandemic has spread to most countries in the world and caused more than 900,000 deaths. It has altered our lives, and despite robust public health interventions, only a handful of countries have managed to temporarily eradicate the disease. From the outset, there has been an urgent need to identify those most susceptible to catching COVID-19 and those at greatest risk of severe infection or death. For hematologists involved in the care of patients with blood cancers, given the disease and treatment-associated immune-suppression and abnormal inflammatory states, there was a major concern that patients with hematologic malignancies might be at increased risk. Many countries adopted an approach that recommended universal shielding for patients with blood cancers. In certain cases, therapeutic interventions were attenuated. This advice came about from expert consensus but had a limited evidence base. As we face a prolonged period coexisting with the virus, there is an urgent need for more data to help define additional personalized approaches to shielding and to therapy choices in patients with blood cancers. These data are beginning to emerge.
The UK OpenSafely study1 examined primary care records of 17.3 million adults that were pseudonymously linked to 10,926 COVID-19–related deaths. Risk for COVID-19–associated death was linked with being male; older age; lower socioeconomic status; and having diabetes, severe asthma, and various other medical conditions. Strikingly, persons with a recent (i.e., in the past 5 years) history of hematologic malignancy had an at least 2.5-fold increased risk of COVID-19–related death. This important study provided robust evidence that patients with hematologic malignancies were at increased risk of COVID-19–associated death. As hematologic malignancies encompass a diverse range of different diseases, there is an urgent need to define disease-specific risks.
In an effort to add further clarity to this discussion across cancers, including hematologic malignancies, Dr. Lennard Lee and colleagues presented a large, prospective cohort study from the UK Coronavirus Cancer Monitoring Project (UKCCMP). One thousand forty-four patients were recruited from 61 cancer centers throughout the United Kingdom. All patients were adults with active cancer who were diagnosed with COVID-19 infection between March 18 and May 8, 2020, by SARS-CoV-2 polymerase-chain reaction. This was defined as those with either metastatic cancer or those receiving any anticancer treatment (including systemic chemotherapy/immunotherapy, radiotherapy, or surgery) within the past 12 months. Patients with skin cancer were not included. Patients were observed from diagnosis until either discharge or death at a median of six days (interquartile range [IQR], 2-11 days). The median age of the cohort was 70 years (IQR, 60-77 years), and 56.9 percent were men.
To estimate susceptibility to COVID-19 within different subtypes of cancer, they compared the UKCCMP cohort with the most recent English cancer registry. This registry collects data on all patients diagnosed with cancer in England during 2017. Any cancer included in the World Health Organization ICD-10 classification is included, excluding nonmelanoma skin cancer. They found that significantly more men were present in the UKCCMP cohort compared to the Office for National Statistics cancer registry, suggesting that men with cancer are more likely to have symptomatic COVID-19 compared to women (56.9% vs. 51.3%; overall response [OR], 1.26; 95% CI, 1.12-1.43; p=0.0002). There was no difference in age between the two cohorts. Patients with hematologic cancer were significantly over-represented in the UKCCMP cohort (21.5% vs. 9.5%) with ICD-10 codes for leukemia (OR, 2.8; 95% CI, 2.2-3.6; p<0.0001), myeloma (OR, 2.0; 95% CI, 1.4-2.8; p=0.0001), and lymphoma (OR, 1.6; 95% CI, 1.3-2.1; p<0.0001), all at increased risk. Surprisingly, lung cancer (OR, 0.75; 95% CI, 0.61-0.91; p=0.003) and prostate cancer (OR, 0.72; 95% CI, 0.59-0.88; p=0.0008) were significantly less represented in the UKCCMP cohort.
Disease severity was analyzed using multivariable logistic regression corrected for age and sex. Hematologic cancers were associated with severe or critical COVID-19 infection (OR, 1.57; 95% CI, 1.15-2.15; p=0.0043) and were more likely to require intensive care admission for ventilation (OR, 2.73; 95% CI, 1.43-5.11; p=0.0019). Risk of all-cause case fatality was increased in patients with leukemia (OR, 2.25; 95% CI, 1.13-4.57; p=0.023) but not other subtypes of hematologic malignancy. Of the patients with hematologic malignancies, 47.6 percent received chemotherapy within four weeks compared to 29.5 percent of patients in the rest of the UKCCMP cohort. Patients with hematologic cancer who received chemotherapy were at increased risk of death during hospital admission with COVID-19 (OR, 2.09; 95% CI, 1.09-4.08; p=0.028).
In summary, these findings are concerning for patients with hematologic cancer, especially those receiving chemotherapy with regards to susceptibility to, and risk of, severe outcome following COVID-19 infection. The question is what can be done to understand which patients are at particular risk, and how might we best protect them? Expansion of telemedicine clinics and alteration of chemotherapy regimens is widespread, with all these changes occurring on the presumption of increased risk and clinicians understandably acting quickly rather than waiting for the data before changing practice. Dr. Lee and colleagues provide further evidence to support these changes, but the picture is still incomplete, and much more disease-specific granularity is required. Despite the sizeable cohort, numbers are small for specific subtypes of hematologic malignancy with the ICD-10 codes grouping diverse spectra of disease together; for instance, the definition for leukemia includes acute, chronic, myeloid, and lymphoid leukemia. We also do not yet know how many patients have had mild or asymptomatic infections as well as the more severe infections reported here (i.e., there may be a major issue with ascertainment bias). Perhaps most crucially, we need large-scale serology surveys to determine an accurate denominator for the infection fatality rate in defined patient groups to help identify those at most risk. This will allow risk prediction models to be developed for patients with blood cancer. Prospective cohort studies will also be crucial to assess durability of immune response following COVID-19 (or vaccines) in patients with hematologic malignancies.
Dr. Salisbury and Dr. Mead indicated no relevant conflicts of interest.