key: cord-0681930-94ziaa82
authors: Hill, Joshua A.; Menon, Manoj P.; Dhanireddy, Shireesha; Wurfel, Mark M.; Green, Margaret; Jain, Rupali; Chan, Jeannie D.; Huang, Joanne; Bethune, Danika; Turtle, Cameron; Johnston, Christine; Xie, Hu; Leisenring, Wendy M.; Nina Kim, H.; Cheng, Guang‐Shing
title: Tocilizumab in hospitalized patients with COVID‐19: Clinical outcomes, inflammatory marker kinetics, and safety
date: 2020-11-22
journal: J Med Virol
DOI: 10.1002/jmv.26674
sha: ed8684d2852d0e57dc0f6b3d5cf5ce90c67fa129
doc_id: 681930
cord_uid: 94ziaa82

Coronavirus disease 2019 (COVID‐19) due to infection with severe acute respiratory syndrome coronavirus 2 causes substantial morbidity. Tocilizumab, an interleukin‐6 receptor antagonist, might improve outcomes by mitigating inflammation. We conducted a retrospective study of patients admitted to the University of Washington Hospital system with COVID‐19 and requiring supplemental oxygen. Outcomes included clinical improvement, defined as a two‐point reduction in severity on a six‐point ordinal scale or discharge, and mortality within 28 days. We used Cox proportional‐hazards models with propensity score inverse probability weighting to compare outcomes in patients who did and did not receive tocilizumab. We evaluated 43 patients who received tocilizumab and 45 who did not. Patients receiving tocilizumab were younger with fewer comorbidities but higher baseline oxygen requirements. Tocilizumab treatment was associated with reduced C‐reactive protein, fibrinogen, and temperature, but there were no meaningful differences in time to clinical improvement (adjusted hazard ratio [aHR], 0.92; 95% confidence interval [CI], 0.38–2.22) or mortality (aHR, 0.57; 95% CI, 0.21–1.52). A numerically higher proportion of tocilizumab‐treated patients had subsequent infections, transaminitis, and cytopenias. Tocilizumab did not improve outcomes in hospitalized patients with COVID‐19. However, this study was not powered to detect small differences, and there remains the possibility for a survival benefit.

Treatment with corticosteroids, which suppress immune responses nonspecifically, has not been shown to improve outcomes with other severe respiratory viral infections or acute respiratory distress syndrome in the general population. 10 However, recent clinical trial data demonstrate a mortality benefit of low dose dexamethasone for COVID-19 in hospitalized patients requiring supplemental oxygen or mechanical ventilation, 11 supporting the notion that hyperinflammation contributes to disease pathogenesis among patients with severe COVID-19 disease. The increasing use of cytokine-specific immunomodulatory agents in other conditions has led to intense interest in the use of these agents for COVID-19. The rationale for use of immunomodulatory therapies in patients with COVID-19 is extrapolated from the experience in CAR-T cell therapy recipients, in whom IL-6 receptor blockade with tocilizumab (Actemra) can rapidly attenuate symptoms of CRS. 4 In this context, tocilizumab does not appear to independently increase risk for infection, 12, 13 although infectious complications are reported with chronic use in rheumatoid arthritis, 14 particularly with concurrent immunosuppressants. Cytokine blockade for hyperinflammatory syndromes is supported by additional lines of evidence, including the finding that treatment with recombinant IL-1 receptor antagonist, anakinra (Kineret), may improve outcomes in patients with MAS-like features in the setting of sepsis. 15 Since the beginning of the pandemic, multiple observational reports have suggested that tocilizumab improves respiratory, radiological, and inflammatory parameters associated with COVID-19, although there are conflicting results with regard to clinical outcomes. [16] [17] [18] [19] [20] [21] [22] [23] [24] [25] [26] [27] [28] [29] Observational studies using either another anti-IL-6 receptor antagonist, sarilumab (Kevzara), or anakinra have shown similar findings. [30] [31] [32] [33] The safety of immunomodulatory therapy in the context of an active infection remains incompletely understood; IL-6 is an essential aspect of immune control for many viral infections, but overproduction of IL-6 may be detrimental to viral clearance and survival. 8, 34 Results from a randomized controlled trial indicate that tocilizumab did not result in reduced mortality or need for mechanical ventilation, 35 although there was evidence for shorter duration of hospitalization and intensive care unit stays in preliminary results from another trial, 36 and the incidence of adverse events were similar to the placebo groups. There remains a critical need to understand the efficacy and safety of tocilizumab in patients with COVID-19.

The aims of this observational study were to comprehensively describe the clinical course and outcomes of hospitalized patients with COVID-19 who were treated with tocilizumab and compare to patients from the same time period who were not treated with tocilizumab, describe adverse effects, and to evaluate the kinetics of inflammatory markers after tocilizumab therapy.

We used the University of Washington (UW) Enterprise Data Warehouse to retrospectively identify all individuals admitted to three UW Hospitals across the Seattle metropolitan region between March 19th (the first date of tocilizumab administration) and April 24th, 2020 with SARS-CoV-2 detection in a respiratory sample.

Patients were excluded if they did not require supplemental oxygen, died or were placed on comfort measures within 2 days of admission (in the no tocilizumab cohort only), or received an alternative anticytokine therapy as part of clinical care or a trial. This study was approved by the UW Institutional Review Board.

Hospital system-wide guidelines pertaining to the use of COVID-19specific therapies were formulated using the limited available data by an interdisciplinary committee. During the study period, this guidance recommended use of open label hydroxychloroquine or investigational remdesivir in patients requiring supplemental oxygen and not participating on a corresponding clinical trial. The guidelines recommended consideration of a single dose of tocilizumab 400 mg intravenously in patients who had persistent fever and either impending or current respiratory failure, hemodynamic instability requiring vasopressor or ionotropic support, or a serum IL-6 level greater than five times the upper limit of normal (range, 0-6 pg/ml). Tocilizumab was contraindicated in patients with sepsis due to another infection, transaminases > 5 times the upper limit of normal, severe neutropenia (neutrophil count < 500 cells/mm 3 ), or severe thrombocytopenia (platelet count <50,000 cells/mm 3 ). Corticosteroids and other antivirals were not recommended at that time. Patients received prophylactic heparin or enoxaparin for prevention of deep vein thrombosis at the discretion of the treating physician. Laboratory testing for inflammatory markers (IL-6, CRP, ferritin, fibrinogen, lactate dehydrogenase [LDH], and D-dimer) were recommended at the time of admission, every 3 days, with clinical deterioration, and before discharge.

The primary outcome was sustained (≥3 days) clinical improvement within 28 days after baseline. For patients treated with tocilizumab, baseline was considered to be at the time of tocilizumab administration.

For patients not treated with tocilizumab, baseline was considered to be 2 days after hospitalization to align with the median number of days between hospitalization and tocilizumab administration in the tocilizumab cohort (median, 1.8; interquartile range [IQR], 0.8-3.8).

Clinical improvement was defined as a two-point reduction in patients' baseline status on a six-point ordinal scale (modified from the WHO guidelines), or live discharge from the hospital, whichever came first.

The ordinal scale consisted of 1, discharged; 2, hospitalized but not requiring oxygen; 3, hospitalized and requiring any supplemental oxygen; 4, hospitalized and requiring high-flow oxygen devices or noninvasive ventilation; 5, hospitalized and requiring invasive ventilation; and 6, death. 37 We also analyzed mortality within 28 days.

We described additional clinical outcomes of sepsis (defined as requiring hemodynamic support with vasopressors, with resolution at the time of vasopressor discontinuation), thrombotic events, bacteremia, and microbiologically confirmed pneumonia with another pathogen prompting antimicrobial therapy. We also described the kinetics of inflammatory laboratory markers, temperature, and heart rate.

We abstracted data from medical records and electronic databases from the time of index admission for COVID-19 through 28 days after baseline as defined above. We described the proportion of patients with clinical improvement (as defined above) overall and stratified by baseline severity categories. We estimated the cumulative incidence of clinical improvement, overall and within strata of baseline oxygen requirements, treating death as a competing risk. To mitigate differences in baseline characteristics between those who did and did not receive tocilizumab, we modeled the probability of treatment using baseline characteristics and calculated weights for inverse probability of treatment weighting (IPTW) analyses to weight the distribution of covariates to that of the overall population.

Weights were standardized using the marginal probability of treatment and truncated at the 98th percentile (values >4) to reduce variability. 38 We carried out multivariable Cox proportional-hazards modeling to evaluate the association of receipt of tocilizumab with (1) clinical improvement and with (2) mortality using IPTW models.

Robust variances were utilized in IPTW models to account for inclusion of the weights. 38 We also computed the restricted mean survival time (RMST) difference for time to clinical improvement and adjusted these models using IPTW. This analysis provides an intuitively appealing estimate of the difference in the mean number of days to clinical improvement between treatment groups. 39 Variables with a p < .2 in univariable analyses were candidates for inclusion in the multivariable model; tocilizumab was included in all models, irrespective of univariate results. Variables were retained in the multivariable model if their inclusion modified the tocilizumab hazard ratio by more than 10% or if they were significantly associated with the outcome. In Cox models and RMST calculations for time to clinical improvement, patients who died before Day 28 without clinical improvement were assigned a censored follow-up time at Day 28, essentially giving them the worst possible "infinite" outcome.

Additionally, we described the kinetics of changes in vital signs and laboratory results, as well as the occurrence of other clinical events including thrombosis and infections. SAS version 9.4 (SAS Institute) and RStudio were used for analyses.

Between March 19th and April 24th, 2020, 44 patients were treated with tocilizumab within our hospital system. During this time period, an additional 120 patients were hospitalized with COVID-19 but did not receive tocilizumab. After excluding patients who did not require supplemental oxygen, who died or were placed on comfort measures within 2 days of admission, or who were enrolled in a placebo-controlled trial of IL-6 receptor blockade with sarilumab, there were 43 patients in the tocilizumab cohort and 45 patients in the non-tocilizumab cohort ( Figure S1 ). Demographic and clinical characteristics of each cohort are shown in Table 1 and stratified by baseline severity in Table S1 .

Compared to patients who did not receive tocilizumab, patients treated with tocilizumab were more likely to be younger, of white Hispanic race and ethnicity, without chronic lung disease, full code status, and receiving invasive ventilation (42% vs. 20%) or high-flow oxygen (47% vs. 22%). The invasive ventilation groups were similar overall. Most patients in both groups were treated with hydroxychloroquine, a minority (30%) were treated with remdesivir or placebo as part of a double-blind randomized clinical trial, 40 

All patients were followed for up to 28 days after the baseline date or the time of death, whichever occurred first. In the tocilizumab cohort, 26 (60%) patients were discharged from the hospital, 9 (21%) died, and high-flow subgroup, and 16 versus 13 days in the invasive ventilation subgroup (Table S3) .

To mitigate differences in baseline characteristics between tocilizumab treated and non-tocilizumab treated patients, we generated standardized IPTW weights from propensity scores ( Figure S3A Figure S4 .

An adjusted Cox model using the truncated IPTW weights demonstrated a lower observed risk of mortality within 28 days among patients treated with tocilizumab compared to those not treated with tocilizumab (aHR, 0.57; 95% CI, 0.21-1.52; Table 2 ). However, this estimate was highly variable and was not significant; as such, we cannot infer any difference in mortality between the treatment groups. A Kaplan-Meier survival curve using the IPT weighted population is shown in Figure S5 . We also explored the potential benefit of tocilizumab for reducing thrombotic events. A numerically higher proportion of patients developed thrombotic events in the tocilizumab cohort (n = 4, 9%) than in the non-tocilizumab cohort (n = 2, 4%; Table S5 ). In addition, we evaluated potential adverse events associated with tocilizumab treatment. We found that a numerically higher proportion of patients in the tocilizumab cohort had bacteremia, secondary pneumonia, transaminitis, neutropenia, or thrombocytopenia within 14 days after baseline in the tocilizumab cohort (Table 3 ).

Laboratory values were similar at baseline for the tocilizumab and no tocilizumab cohorts (Figure 3 ). Among most patients who received tocilizumab, there was a rapid and sustained decrease in temperature ( Figure S6) , CRP, and fibrinogen. There was an initial increase followed by a decrease in IL-6 levels after administration of tocilizumab as expected given the mechanism of action of blocking the IL-6 receptor.

Levels of LDH declined over days, and there was minimal change in ferritin and D-dimer levels. Heart rates rapidly declined but changes were not sustained over time ( Figure S6 ). In contrast, levels of all biomarkers were relatively stable in the no tocilizumab cohort.

Morbidity and mortality remain unacceptably high in patients with COVID-19. The clinical utility of directed immunomodulatory agents to preempt or reverse the cytokine-driven inflammation responsible for much of this morbidity has biological plausibility, 41 [17] [18] [19] [20] [21] [24] [25] [26] [27] [28] [29] One of the largest and most rigorous studies in mechanically ventilated patients indicated a significant survival benefit, despite a higher incidence of infections in the tocilizumab group. 26 However, these studies are limited by residual indication bias among those receiving tocilizumab. Rapidly evolving improvements in supportive care strategies may also confound findings of better outcomes with tocilizumab over time. The first available data from randomized controlled trials of tocilizumab in moderately ill hospitalized patients with COVID-19 in North America did not meet their primary endpoints for a reduction in intubation or death among patients randomized to tocilizumab. 35 In a larger trial, there was evidence for a shorter duration of hospitalization and intensive care unit stays. 36 Patients receiving tocilizumab did not have an increased incidence of adverse events in either trial.

In our study, we did not demonstrate an apparent clinical benefit of treatment with tocilizumab, contrary to what has been published in many observational studies. Although patients receiving tocilizumab in our study were younger and sicker on presentation, no clear benefits were seen in comparisons stratified by baseline disease severity, IPTW models that established more balance between the tocilizumab and no tocilizumab groups, or adjusted models. Our study was not confounded by the use of corticosteroids, unlike most other observational studies. This is relevant given emerging data of the benefit of corticosteroids in patients with severe disease. 42 The possibility still remains that there is a clinical scenario in which tocilizumab is beneficial, but this will be best F I G U R E 2 Cumulative incidence of sustained (≥3 days) clinical improvement within 28 days among patients who did and did not receive tocilizumab using primary, unweighted data. Death was treated as a competing risk event T A B L E 2 Multivariable Cox models of time to clinical improvement or death within 28 days using IPTW weights (truncated to account for outliers) to weight the distribution of covariates to that of the overall population demonstrated through randomized clinical trials. Some studies report higher rates of infections in patients receiving tocilizumab, [25] [26] [27] although this was not observed in the randomized trials 35, 36 In our study, a numerically higher proportion of patients in the tocilizumab cohort had a microbiologically confirmed secondary infection, although this was likely related to more patients requiring invasive ventilation in the tocilizumab group, resulting in more diagnoses of ventilator-associated pneumonias.

We found numerically higher rates of transaminitis and cytopenias in patients receiving tocilizumab, but these comparisons were limited by a small number of events and may also reflect differences in the distribution of disease severity between the two groups. Similar to the preliminary data from the randomized trials, our findings do not suggest obvious safety concerns related to tocilizumab use for COVID-19.

We also tested the novel hypothesis that thrombotic events may be reduced in tocilizumab-treated patients. This could be an additional benefit of tocilizumab based on the high frequency of venous thrombosis in patients with COVID-19. 43 It is biologically plausible that tocilizumab could reduce risk for thrombosis given studies demonstrating reduced plasma levels of the procoagulant Factor XIII (fibrin-stabilizing factor), D-dimer levels, and other inflammatory markers in patients receiving tocilizumab. [44] [45] [46] [47] Although we demonstrate a clear anti-inflammatory effect with rapid and marked reductions in serum levels of CRP and fibrinogen, a higher proportion of patients in the tocilizumab group in our study had thrombotic events, which may be due to the higher baseline disease severity in the tocilizumab group. It is possible that thromboembolic events were underdiagnosed overall during the study period. 43, 48 The low number of events and differences in characteristics between the patient subgroups limit conclusions, but the role of tocilizumab for reducing thrombosis may warrant additional study in randomized trials.

Finally, we evaluated the effect of tocilizumab on clinical and laboratory markers of inflammation. Similar to other studies of tocilizumab in patients with CRS and COVID-19, 16, 49 we demonstrated a rapid and sustained effect on lowering body temperature, CRP, and fibrinogen. This may play a role in the higher rates of resolution of sepsis physiology that we describe in our cohort, as reported in another study, 20 and as suggested by the shorter duration of intensive care unit stays in a randomized trial. 36 The anti-inflammatory effects also suggest that the approximately 4 mg/kg dose used in our patients was sufficiently bioactive. Studies using subcutaneous tocilizumab demonstrate similar anti-inflammatory effects, and this mode of administration could be further studied for its clinical utility. 50 Whether the kinetics of inflammatory biomarkers correlate with response to tocilizumab warrants further study.

The primary limitations of our study include those inherent to a retrospective, non-randomized design. Higher baseline severity status in the tocilizumab cohort, as well as other differences in patient characteristics, resulted in imbalances in the patient cohorts. To control for some of these differences, we compared outcomes within baseline severity categories and performed IPTW adjusted analyses. For most comparisons, we were limited to detecting relatively large differences. To avoid instability in the models, we did not include chronic lung disease, which only affected three individuals in the tocilizumab group, or hydroxychloroquine, which does not affect outcomes of hospitalized patients with COVID-19. Although the use of remdesivir may have affected our findings, it was used in a minority of participants, and blinded receipt of remdesivir versus placebo was balanced across the two cohorts. The observation that tocilizumab was used in sicker patients indicates opportunities for larger studies of earlier administration. Given that obesity is a risk factor for worse outcomes with COVID-19 and that more than half of our participants had a body mass index greater than 30, weightbased dosing could be considered to attain more standardized drug exposure across patients. Strengths of this study include the relatively large cohort of patients requiring noninvasive and invasive oxygenation, along with careful adjustment for differences in patient characteristics and other confounders. We performed an RMST analysis for time to clinical improvement that is novel to studies of tocilizumab for COVID-19 and provide a tangible measure of treatment effect. 39 We performed comprehensive chart review and data abstraction to describe and analyze a variety of clinical and laboratory outcomes to evaluate the potential benefits and side effects of tocilizumab therapy, including novel comparisons of thrombotic events and resolution of sepsis physiology. Our data were not confounded by steroids, and most patients did not receive remdesivir.

In conclusion, there was no clear evidence of improved clinical outcomes among hospitalized patients with COVID-19 who received one dose of 400 mg of tocilizumab without concurrent steroids or effective antivirals. Tocilizumab was associated with a robust reduction in F I G U R E 3 Kinetics of inflammatory laboratory markers. Units for other laboratory values are ng/ml for ferritin, mcg/ml for D-dimer, mg/dl for fibrinogen. The presented data indicate medians and the associated interquartile ranges. If multiple values were available for the same day, the maximum value was used. The number of patients contributing results to each panel are in Table S6 . In Panel A, the late secondary increase in IL-6 levels was driven by two individuals who had progressive cardiopulmonary disease with concurrent bloodstream infections and ventilator-associated pneumonias at the time of secondary IL-6 increase. CRP, C-reactive protein (mg/L); IL-6, interleukin 6 (pg/ml); LDH, lactate dehydrogenase (U/L) inflammatory biomarkers; further study in patients with sepsis physiology may be warranted. Randomized controlled trials using tocilizumab in the context of concurrent effective antiviral and steroid therapy will be particularly important to guide evolving treatment strategies.

World Health Organization. Coronavirus disease (COVID-19): situation report-133

Characteristics of and important lessons from the coronavirus disease 2019 (COVID-19) outbreak in China: summary of a report of 72314 cases from the Chinese Center for Disease Control and Prevention

Association of inflammatory markers with the severity of COVID-19: a meta-analysis

Chimeric antigen receptor therapy

Macrophage activation syndrome in the era of biologic therapy

Pathogenic human coronavirus infections: causes and consequences of cytokine storm and immunopathology

Mechanisms of severe acute respiratory syndrome coronavirus-induced acute lung injury

The role of interleukin 6 during viral infections

Critical adverse impact of IL-6 in acute pneumovirus infection

Clinical evidence does not support corticosteroid treatment for 2019-nCoV lung injury

Dexamethasone in hospitalized patients with COVID-19-preliminary report

Preemptive mitigation of CD19 CAR T-cell cytokine release syndrome without attenuation of antileukemic efficacy

Tocilizumab not associated with increased infection risk after CAR T-implications for COVID-19?

Tocilizumab in rheumatoid arthritis: a meta-analysis of efficacy and selected clinical conundrums

Interleukin-1 receptor blockade is associated with reduced mortality in sepsis patients with features of macrophage activation syndrome: reanalysis of a prior phase III trial *

Effective treatment of severe COVID-19 patients with tocilizumab

Impact of low dose tocilizumab on mortality rate in patients with COVID-19 related pneumonia

Tocilizumab for treatment of severe COVID-19 patients: preliminary results from smatteo COVID-19 registry (SMACORE). Microorganisms

Tocilizumab treatment for cytokine release syndrome in hospitalized COVID-19 patients: survival and clinical outcomes

Tocilizumab for treatment of patients with severe COVID À 19: a retrospective cohort study

Interleukin-6 receptor blockade with subcutaneous tocilizumab in severe COVID-19 pneumonia and hyperinflammation: a case-control study

Tocilizumab for the treatment of severe COVID-19 pneumonia with hyperinflammatory syndrome and acute respiratory failure: a single center study of 100 patients in Brescia

Pilot prospective open, single-arm multicentre study on off-label use of tocilizumab in patients with severe COVID-19

Tocilizumab therapy reduced intensive care unit admissions and/or mortality in COVID-19 patients

Profiling COVID-19 pneumonia progressing into the cytokine storm syndrome: results from a single Italian Centre study on tocilizumab versus standard of care

Tocilizumab for treatment of mechanically ventilated patients with COVID-19

Tocilizumab in patients with severe COVID-19: a retrospective cohort study

Outcomes in patients with severe COVID-19 disease treated with tocilizumab: a casecontrolled study

Efficacy and safety of tocilizumab in severe COVID-19 patients: a single-centre retrospective cohort study

Sarilumab use in severe SARS-CoV-2 pneumonia

Sanofi and Regeneron provide update on U.S. Phase 2/3 adaptive-designed trial in hospitalized COVID-19 patients. Press Release

Targeting the inflammatory cascade with anakinra in moderate to severe COVID-19 pneumonia: case series

Safety and efficacy of early highdose IV anakinra in severe COVID-19 lung disease

IL-6 ameliorates acute lung injury in influenza virus infection

Efficacy of tocilizumab in patients hospitalized with COVID-19

Tocilizumab in hospitalized patients with COVID-19 pneumonia

WHO R&D Blueprint novel coronavirus COVID-19 therapeutic trial synopsis

Marginal structural models and causal inference in epidemiology

How to quantify and interpret treatment effects in comparative clinical studies of COVID-19

Remdesivir for the treatment of COVID-19-preliminary report

Level of IL-6 predicts respiratory failure in hospitalized symptomatic COVID-19 patients

Effect of dexamethasone in hospitalized patients with COVID-19: preliminary report

Incidence of venous thromboembolism in hospitalized patients with COVID-19

Non-autoimmune combined factor XIII A and B subunit deficiencies in rheumatoid arthritis patients treated with antiinterleukin-6 receptor monoclonal antibody (tocilizumab)

Tocilizumab induced acquired factor XIII deficiency in patients with rheumatoid arthritis

Effect of interleukin-6 receptor blockade on surrogates of vascular risk in rheumatoid arthritis: MEASURE, a randomised, placebo-controlled study

Prothrombotic biomarkers in patients with rheumatoid arthritis: the beneficial effect of IL-6 receptor blockade

Autopsy findings and venous thromboembolism in patients with COVID-19

Kinetics and biomarkers of severe cytokine release syndrome after CD19 chimeric antigen receptor-modified T cell therapy

Subcutaneous tocilizumab treatment in patients with severe COVID-19-related cytokine release syndrome: an observational cohort study

The authors declare that there are no conflict of interests. 

The data that support the findings of this study are available from the corresponding author upon reasonable request.

Joshua A. Hill https://orcid.org/0000-0002-7665-7100Christine Johnston http://orcid.org/0000-0002-3073-0843