Clinical evidence of an interferon–glucocorticoid therapeutic synergy in COVID-19

The study addresses whether interferon (IFN) therapy can synergize with glucocorticoids (GC) in treating COVID-19 by compensating for GC-induced suppression of antiviral immunity. Severe COVID-19 is characterized by dysregulated immune responses and cytokine storm, for which GC like dexamethasone can reduce mortality in patients requiring respiratory support. However, GC may delay viral clearance and worsen outcomes in viral pneumonias, raising concerns for non-severe cases. Conversely, type I IFNs are central to antiviral defense and are notably blunted in severe SARS-CoV-2 infection. Prior reports suggest therapeutic IFN (including inhaled IFN-β) alone or in combination with antivirals shows promise, though timing is critical as late IFN may impair epithelial repair. Based on these data, the authors hypothesize IFN therapy may synergize with GC, improving outcomes by restoring antiviral responses while GC mitigates hyperinflammation. They test this in a retrospective cohort from a hospital in Hubei, China.

Background evidence includes: (1) GC reduce mortality in severe COVID-19 (e.g., RECOVERY trial showed dexamethasone benefits in patients on oxygen or mechanical ventilation), and meta-analyses support systemic GC in critically ill patients; (2) GC can impair innate antiviral immunity, delaying viral clearance and worsening outcomes in other viral pneumonias (SARS, MERS, influenza) and potentially in non-severe COVID-19; (3) SARS-CoV-2 induces low IFN responses, suggesting therapeutic IFN may counter viral evasion; (4) Clinical studies of IFN-β (alone or combined with antivirals) have shown promising outcomes in COVID-19; (5) Timing is crucial: late IFN may disrupt epithelial repair. These studies motivate assessing an IFN–GC therapeutic interaction and optimal timing/dosing in COVID-19.

Design: Non-interventional retrospective observational cohort study at Suizhou Zengdu Hospital, Hubei, China. Ethical approval obtained with waiver of informed consent; identifiers anonymized. Population: All 490 inpatients with PCR-confirmed COVID-19 (two consecutive positive qPCR tests from throat/nasal swabs, validated by local CDC) from Jan 15 to Mar 31, 2020 were screened. Records abstracted May 8–22, 2020 with quality control; additional import records abstracted Jun 20–Jul 30, 2020. Survivors followed for two weeks per local regulations; last follow-up May 22, 2020. Exclusions: (1) Incomplete treatment information (e.g., ≥5 days therapy elsewhere before admission or missing key labs/radiology); (2) Hospital stay <5 days; (3) Extended hospital stay for non-COVID conditions by >5 days; (4) Received no antiviral therapy for >2 days during hospitalization due to lack of symptoms, late diagnosis, adverse effects, or supportive care only. After exclusions, 387 records analyzed. Exposure definitions: GC exposure was conditional (more severe patients). Early IFN therapy defined empirically as initiation within 5 days of admission. IFN exposure was quasi-random relative to other antivirals in the early phase of the outbreak. Interventions: GC regimens recorded for cumulative doses ≥40 mg methylprednisolone-equivalent. GC mainly administered early during hospitalization at cumulative doses typically ≤150 mg dexamethasone-equivalent (≈80 mg methylprednisolone), with the majority ≤320 mg methylprednisolone-equivalent. IFN therapy: nebulized interferon (likely IFN-β; specified as IFN) 5 million units in 2 mL sterile water via nebulizer twice daily. Median initiation at 2 days post-admission among early IFN users; duration median 11 days in GC users vs 9 days in nonusers. Concomitant antivirals (e.g., lopinavir/ritonavir, umifenovir, oseltamivir) and human immunoglobulins were recorded if used consecutively ≥3 days. Outcomes: Primary—time to hospital discharge (LOS). Secondary—time from admission to symptom relief; prolonged viral shedding (PVS), defined as ≥1 positive SARS-CoV-2 PCR result ≥2 days after symptom relief. Discharge criteria: afebrile ≥3 days, improved respiratory symptoms/function, radiologic improvement, and two consecutive negative PCRs ≥24 h apart. Symptom relief: afebrile ≥3 days with improved respiratory symptoms/function; pre-relapse relief not counted in severe/critical relapses. Due to limited testing capacity early in the outbreak, PCR tests were not conducted between diagnosis confirmation and symptom relief. Covariates and confounders: Demographics, comorbidities (hypertension, diabetes), clinical severity at admission (moderate, severe, critical), oxygen saturation, respiratory rate, symptom counts, CT findings, lymphopenia/eosinopenia, timing/dose/duration of GC and IFN, and concurrent antivirals. Sensitivity analyses identified LVP/r, oseltamivir, oxygen saturation, gender, age, hypertension, and diabetes as confounders for LOS models. Sample size: Target n=250 (balanced across four groups) provided 90% power to detect a 15% change in LOS assuming mean LOS=20 days, α=0.05 (G*Power v3.1). Final analyzed n=387. Statistical analysis: Descriptive statistics with Fisher’s exact/Chi-square for categorical and Wilcoxon–Mann–Whitney for continuous variables. Time-to-event analyses via Kaplan–Meier and log-rank tests. Crude HRs via log-rank; adjusted HRs via Cox proportional hazards models adjusted for prespecified confounders (gender, hypertension, diabetes, oxygen saturation at admission, symptom count at admission, lopinavir/ritonavir, oseltamivir, and other indicated treatments). Proportional hazards assumptions assessed by Kaplan–Meier curves. PVS modeled with logistic regression adjusted for the same confounders among those who ceased viral shedding before discharge/death. Interaction between GC and early IFN evaluated; generalized log-gamma regression used for robustness checks of LOS and time to symptom relief. Two-sided α=0.05. Analyses conducted in R. Group distributions: Of 387, 118 (30.5%) received both GC and early IFN, 79 received GC alone, 87 early IFN alone, and 87 neither. Median time from symptom onset to admission: 6 days. Overall in-hospital mortality: 4.3%. Median LOS: 21 (IQR 15–26) days; median time to symptom relief: 12 (IQR 9–16) days. Overall PVS prevalence: 14.7%.

- Early IFN therapy benefits were concentrated among GC users: - Hospital discharge: adjusted HR 1.68 (95% CI 1.19–2.37) among GC users; not significant among nonusers (adjusted HR 0.81, 95% CI 0.59–1.12). - Symptom relief: adjusted HR 1.48 (95% CI 1.06–2.08) among GC users; not significant among nonusers (adjusted HR 0.87, 95% CI 0.64–1.19). - Prolonged viral shedding (PVS): adjusted OR 0.24 (95% CI 0.10–0.57) among GC users; trend in opposite direction and not significant among nonusers (adjusted OR 2.05, 95% CI 0.88–4.74). - Strong interaction between GC and early IFN in models for LOS and time to symptom relief (generalized log-gamma regression, p<0.001 for both endpoints), indicating potential therapeutic synergy. - GC timing and dose dependence: - Early IFN therapy was associated with earlier recovery primarily when GC were initiated early and at certain cumulative dose ranges. Analyses stratified by GC cumulative dose (methylprednisolone-equivalent) showed shorter hospitalization with early IFN at 40–120 mg and across higher dose strata; synergy was significant at doses comparable to those used in RECOVERY. - GC alone associated with worse recovery metrics (model-adjusted effects, Table 3): - Discharge: HR 0.56 (95% CI 0.39–0.80), p=0.002. - Symptom relief: HR 0.65 (95% CI 0.46–0.93), p=0.018. - PVS: OR 3.74 (95% CI 1.46–8.53), p=0.002. - Mortality and descriptive outcomes: - Overall mortality 4.3%; higher in GC-alone group (11.6%) vs combined early GC+early IFN group (1.7%), consistent with a negative association between early IFN and mortality, though mortality analyses were limited due to confounding by indication (GC given to more severe patients). - Median LOS across cohort: 21 days (IQR 15–26); time to symptom relief: 12 days (IQR 9–16).

Findings support a therapeutic synergy between interferon and glucocorticoids in COVID-19: early IFN therapy, when combined with GC, was associated with faster discharge, quicker symptom relief, and reduced prolonged viral shedding, whereas IFN showed no benefit in patients not receiving GC. This aligns with the dual-edged nature of GC—beneficial anti-inflammatory effects in severe disease but detrimental suppression of antiviral immunity—suggesting IFN may restore antiviral responses blunted by GC. Consistency with RECOVERY trial is noted: GC benefits primarily in patients requiring respiratory support; in less severe cases GC may be harmful unless antiviral immunity is supported. The observed GC–IFN interaction was timing- and dose-dependent, underscoring the importance of early IFN initiation alongside GC. Mechanistically, inhaled IFN may achieve high local concentrations in the respiratory epithelium sufficient to overcome GC-mediated suppression of IFN signaling, improving viral control while GC mitigates hyperinflammation. Preclinical studies support that inhaled IFN can counter GC-impaired antiviral responses. Clinically, these results argue for consideration of early inhaled IFN in patients receiving systemic GC for COVID-19, with attention to GC dosing and timing. However, the observational design and confounding by indication limit causal inference; randomized controlled trials are needed to confirm efficacy, define optimal timing, dosing, and patient selection, and assess effects on hard endpoints like mortality.

In a retrospective cohort of hospitalized COVID-19 patients, early interferon therapy was associated with improved recovery and reduced prolonged viral shedding specifically among patients treated with glucocorticoids, indicating a potential therapeutic synergy. The interaction was dependent on GC timing and cumulative dose, suggesting that early, appropriately dosed GC combined with inhaled IFN may optimize outcomes by balancing antiviral and anti-inflammatory effects. Prospective randomized trials should validate these findings, determine optimal IFN formulation, delivery route, and timing relative to GC initiation, and evaluate impacts on mortality and long-term outcomes.

- Observational, non-randomized design with confounding by indication (GC preferentially given to more severe patients), limiting causal inference. - Single-center study during early pandemic phase; generalizability may be limited. - Limited PCR testing capacity prevented testing between diagnosis and symptom relief; PVS definition relied on tests after symptom relief and may under- or over-estimate shedding duration. - Potential residual confounding from concomitant therapies (e.g., antivirals, immunoglobulins) and clinical management variations despite adjustments. - Inability to robustly assess mortality effects of IFN–GC synergy due to selective GC use in critically ill patients and low overall mortality. - Some exposure misclassification possible (e.g., exact dosing/timing), and small subgroup sizes in dose-stratified analyses reduce precision.