A Multi-Center Retrospective Review of COVID-19 Outcomes in Patients with Lymphoid Malignancy

Background: Early reports suggested that cancer patients have a 1.7-fold increased risk of contracting SARS-CoV-2 and higher rates of severe events and mortality compared with general population. Patients with hematologic malignancies may have worse COVID-19 outcomes, due to an impaired humoral immune response from their underlying malignancy and concurrent anticancer therapy. In this multi-center, retrospective, observational study, we evaluate the associations of COVID-19 outcomes with patient and lymphoma disease characteristics.

Methods: EMRs at 10 study centers across the USA identified 519 patients with a diagnosis of lymphoma, CLL, or other lymphoid malignancies, who had a documented positive result of SARS-CoV-2 PCR or nucleocapsid antibody testing. Descriptive statistics were used to summarize the demographic and clinical characteristics. Logistic regression was used to evaluate the associations of individual characteristics with COVID-19 outcomes, adjusted for center (NYU vs. other). The interactions between each of the variables was also included in these models; since the interactions were generally small and non-significant, only the main effect of center was included. Two-sided p-values ≤0.05 were considered significant; there were no adjustments for multiple variables or endpoints. Each analysis was based on complete data for that analysis.

Results: Tables 1 and 2 provide demographic and clinical characteristics, respectively, of the 519 patients. The mean age was 61.9 years, with 296 (57%) male and 374 (72%) white patients. NYU had the largest cohort (318 patients), with the remaining centers contributing a range of 3 to 69 patients (median 14).

Logistic regression models for the association of each COVID-19 outcome with individual clinical and patient characteristics included adjustments for the center (NYU/other). While center effects were statistically significant, center by covariate interaction effects were not and are not included in the final models. The odds ratio (OR) estimates and p-values for each patient and CLL/lymphoma clinical variable are shown in Tables 3 and 4, respectively. The risks of experiencing a severe event, death, and hospital admission increased with age; for each 10 years of age increase, the ORs were 1.58 for experiencing severe events, 1.78 for death, and 1.65 for hospital admission. The risks of poor outcome were higher in males than in females (OR 1.93 for severe events, 1.85 for death, and 1.47 for hospital admission). Patients with Charlson Comorbidity Index (CCI) >5 had a higher risk of severe events (OR 2.46), mortality (3.30) and hospital admission (2.73) compared to patients with CCI ≤5.

Compared to patients with HL, patients with aggressive NHL had a higher risk of severe events (OR 4.05), mortality (4.68) and hospital admission (4.65). Patients with CLL similarly had a higher risk of severe events (OR 4.64), mortality (4.65) and hospital admission (5.93) compared to HL patients. Patients with indolent NHL had a higher risk of hospital admission (OR 3.95) but not a higher risk of mortality compared to HL. Patients in remission at the time of COVID-19 diagnosis had a lower risk of severe events (OR 0.42), mortality (0.36) and hospital admission (0.40) relative to those who were not in remission. Patients who received cytotoxic chemotherapy within 28 days of their COVID-19 diagnosis had a higher risk of severe events (OR 2.54), mortality (2.79), and hospital admission (2.31). Patients who received an anti-CD20 monoclonal antibody within 6 months of COVID-19 diagnosis had a higher risk of severe events (OR 2.60), mortality (2.17) and hospital admission (3.28).

Conclusions: In addition to demographic and comorbidity risk factors identified in previous studies, our study shows that patients with aggressive NHL and CLL, or patients who have received recent cytotoxic chemotherapy or anti-CD20 mAB, may be at risk for poor COVID-19 outcome. The difference in risk between NHL and HL patients is likely associated with young age of HL patients but may also be related to differences in underlying innate and adaptive immune defects. Patients at high risk for poor outcome should be a priority for studies of monoclonal antibody prophylaxis. If defects in humoral immunity are at the root of poor outcome, this may be compounded by poor response to vaccination. Multivariate analysis of this data will be completed in advance of the meeting.

Figure 1

View largeDownload slide

Figure 1

View largeDownload slide

Close modal