Early changes in immune cell subsets with corticosteroids in patients with solid tumors: implications for COVID-19 management

BACKGROUND: The risk–benefit calculation for corticosteroid administration in the management of COVID-19 is complex and urgently requires data to inform the decision. The neutrophil-to-lymphocyte ratio (NLR) is a marker of systemic inflammation associated with poor prognosis in both COVID-19 and cancer. Investigating NLR as an inflammatory marker and lymphocyte levels as a critical component of antiviral immunity may inform the dilemma of reducing toxic hyperinflammation while still maintaining effective antiviral responses. METHODS: We performed a retrospective analysis of NLR, absolute neutrophil counts (ANCs) and absolute lymphocyte counts (ALCs) in patients with cancer enrolled in immunotherapy trials who received moderate-dose to high-dose corticosteroids. We compared paired presteroid and available poststeroid initiation values daily during week 1 and again on day 14 using the Wilcoxon signed-rank test. Associated immune subsets by flow cytometry were included where available. RESULTS: Patients (n=48) with a variety of solid tumors received prednisone, methylprednisolone, or dexamethasone alone or in combination in doses ranging from 20 to 190 mg/24 hours (prednisone equivalent). The median NLR prior to steroid administration was elevated at 5.0 (range: 0.9–61.2). The corresponding median ANC was 5.1 K/µL (range: 2.03–22.31 K/µL) and ALC was 1.03 K/µL (0.15–2.57 K/µL). One day after steroid administration, there was a significant transient drop in median ALC to 0.54 K/µL (p=0.0243), driving an increase in NLR (median 10.8, p=0.0306). Relative lymphopenia persisted through day 14 but was no longer statistically significant. ANC increased steadily over time, becoming significant at day 4 (median: 7.31 K/µL, p=0.0171) and remaining significantly elevated through day 14. NLR was consistently elevated after steroid initiation, significantly at days 1, 7 (median: 8.2, p=0.0272), and 14 (median: 15.0, p=0.0018). Flow cytometry data from 11 patients showed significant decreases in activated CD4 cells and effector memory CD8 cells. CONCLUSIONS: The early drop in ALC with persistent lymphopenia as well as the prolonged ANC elevation seen in response to corticosteroid administration are similar to trends associated with increased mortality in several coronavirus studies to include the current SARS-CoV-2 pandemic. The affected subsets are essential for effective antiviral immunity. This may have implications for glucocorticoid therapy for COVID-19.