Advertisement

Anakinra or high-dose corticosteroids in COVID-19 pneumonia patients who deteriorate on low-dose dexamethasone: an observational study of comparative effectiveness

Open AccessPublished:November 17, 2022DOI:https://doi.org/10.1016/j.ijid.2022.11.017

      Highlights

      • Dexamethasone 6 mg is the standard of care for patients with COVID-19 requiring oxygen.
      • Neverthless, some patients with COVID-19 pneumonia worsen or fail to improve on this treatment.
      • Whether intensifying immunosuppression in such patients is helpful is unknown.
      • Neither high-dose corticosteroids nor anakinra reduced mortality in these patients.
      • Continuing dexamethasone 6mg in these patients was safer.

      Abstract

      Objectives

      To assess whether escalating to high-dose corticosteroids or anakinra compared with continuing low-dose corticosteroids reduced mortality in patients with severe COVID-19 whose respiratory function deteriorated while receiving dexamethasone 6 mg daily.

      Methods

      We conducted a retrospective cohort study between March 1 to December 31, 2020, of hospitalized patients with confirmed COVID-19 pneumonia. In-hospital death was analyzed using logistic regression with inverse probability of treatment weighting of receiving anakinra, high-dose corticosteroid (dexamethasone >10 mg daily), or remaining on low-dose corticosteroids on the day of first respiratory deterioration.

      Results

      We analyzed 6671 patients whose respiratory status deteriorated while receiving dexamethasone 6 mg daily for COVID-19 pneumonia, of whom 6265 stayed on low-dose corticosteroids, 232 were escalated to high-dose corticosteroids, and 174 to anakinra in addition to corticosteroids. The propensity score-adjusted odds of death were higher in the anakinra (odds ratio [OR] 1.76; 95% CI 1.13-2.72) and high-dose corticosteroid groups (OR 1.53; 95% CI 1.14-2.07) compared with those who continued low-dose corticosteroids on the day of respiratory deterioration. The odds of hospital-acquired infections were also higher in the anakinra (OR 2.00; 95% CI 1.28-3.11) and high-dose corticosteroid groups (OR 1.43; 95% CI 1.00-2.04) compared with low-dose corticosteroid group.

      Conclusion

      Our findings do not support escalating patients with COVID-19 pneumonia who deteriorate on low-dose corticosteroids to high-dose corticosteroids or anakinra.

      Keywords

      Introduction

      Some patients with COVID-19 pneumonia display clinical and laboratory signs of a hyperinflammatory response, including markedly elevated serum C-reactive protein (CRP) and ferritin levels, followed by rapid respiratory status decompensation and progression to multisystem organ failure. These observations led to the use of varying doses of corticosteroids and targeted anti-cytokine agents, including anakinra, an anti-interleukin (IL)-1 treatment, early in the pandemic.
      When a large, open-label randomized controlled trial (RCT) showed a significantly lower 28-day mortality with 6 mg dexamethasone daily (DEXA6) daily for up to 10 days compared with usual care (23.3% vs 26.2%) (
      • Recovery Collaborative Group
      • Horby P
      • Lim WS
      • Emberson JR
      • Mafham M
      • Bell JL
      • Landray MJ
      Dexamethasone in hospitalized patients with Covid-19.
      ) in patients with COVID-19 requiring supplemental oxygen, DEXA6 rapidly became standard of care. However, a sizable proportion of patients’ respiratory status deteriorates while treated with DEXA6. It is unknown whether intensifying immunosuppression with anakinra or high-dose corticosteroids improves survival compared with completing the DEXA6 protocol, and practice remains varied. To our knowledge, these questions have not been addressed in RCTs or observational studies.
      Evidence supporting the use of higher dose corticosteroids or anakinra compared with DEXA6 alone in newly hospitalized patients with COVID-19 pneumonia is inconclusive. A recently published RCT comparing dexamethasone 12 mg to DEXA6 in addition to oxygen support strategies found no difference in mortality between treatment groups regardless of severity of respiratory status at randomization (
      • Bouadma L
      • Mekontso-Dessap A
      • Burdet C
      • Merdji H
      • Poissy J
      • Dupuis C
      COVIDICUS Study Group
      High-dose dexamethasone and oxygen support strategies in intensive care unit patients with severe COVID-19 acute hypoxemic respiratory failure: the COVIDICUS randomized clinical trial.
      ). However, a pre-planned Bayesian analyses of a traditional RCT suggested that dexamethasone 12 mg may be beneficial in patients with more severe respiratory impairment compared with DEXA6, although the difference in mortality did not reach statistical significance in the primary analyses (
      • COVID STEROID 2 Trial Group
      • Munch MW
      • Myatra SN
      • Vijayaraghavan BKT
      • Saseedharan S
      • Benfield T
      • Perner A.
      Effect of 12 mg vs 6 mg of dexamethasone on the number of days alive without life support in adults with COVID-19 and severe hypoxemia: the COVID STEROID 2 randomized trial.
      ).
      In a randomized, open-label adaptive platform trial (

      REMAP-CAP. REMAP-CAP results, April 10, 2021. https://www.remapcap.org/covid19publications, 2021 (accessed 29 June 2022).

      ), anakinra plus corticosteroids had no effect on survival (

      REMAP-CAP. REMAP-CAP results, April 10, 2021. https://www.remapcap.org/covid19publications, 2021 (accessed 29 June 2022).

      ) compared with patients predominantly treated with anti-IL6 inhibitors plus corticosteroids. In this trial, a high proportion of patients were mechanically ventilated at treatment start (37% in anakinra-treated patients and 32% in the comparison treatment arms) (

      National Institutes of Health. COVID-19 treatment Guidlines: interleukin-1 inhibitors, October 19, 2021. https://www.covid19treatmentguidelines.nih.gov/therapies/immunomodulators/interleukin-1-inhibitors/, 2021 (accessed 29 June 2022).

      ). In contrast, a double-blind RCT in patients with moderate respiratory impairment, anakinra plus DEXA6 compared with DEXA6 alone suPAR-guided Anakinra treatment for Validation of the risk and Early Management Of seveRE respiratory failure by COVID-19 (SAVE-MORE) significantly reduced mortality (3.2% and 6.9%, respectively) and clinical worsening (
      • Kyriazopoulou E
      • Poulakou G
      • Milionis H
      • Metallidis S
      • Adamis G
      • Tsiakos K
      • Giamarellos-Bourboulis EJ.
      Early treatment of COVID-19 with anakinra guided by soluble urokinase plasminogen receptor plasma levels: a double-blind, randomized controlled phase 3 trial.
      ).
      The purpose of this study was to compare the effectiveness and safety of either intensifying immunosuppression or continuing the DEXA6 protocol in patients hospitalized for COVID-19 pneumonia who were not responding to DEXA6. To formulate this research question, we engaged with a panel of clinician stakeholders consisting of hospitalists, intensivists, infectious disease, and immunology specialists, who wanted to know whether escalating to high-dose corticosteroids or adding anakinra would be associated with lower in-hospital mortality or an increased risk of infections in patients whose respiratory status worsened or did not improve after starting DEXA6. Using granular electronic health record data from a large population-based sample of patients who received care at one of 36 hospitals in the Kaiser Permanente Southern California (KPSC) and Northern California (KPNC) regions, we performed a retrospective, observational study to generate real-world evidence to address these clinically important questions.

      Methods

      We performed this retrospective cohort study after receiving approval from the KPSC Institutional Review Board (#12396).

      Study population

      KPSC and KPNC provide care for >9 million members and operate 36 medical centers across California. To assemble the study cohort, we identified adults (aged ≥18 years) admitted between March 1, 2020, to December 31, 2020, with a positive SARS-CoV-2 polymerase chain reaction result in the 3 weeks before admission or during the hospitalization who met the following additional inclusion criteria: (i) received dexamethasone 6 mg or equivalent dose of another corticosteroid, (ii) required oxygen supplementation 12 hours before or after first dose of dexamethasone administration and (iii) experienced respiratory deterioration after the first 6mg dose of dexamethasone (see below for definition) or had moderate, severe, or very severe respiratory impairment that persisted for 24 hours or more after first 6 mg dose of dexamethasone (referred to as persistence hereafter). Exclusion criteria were as follows: (i) admission for labor and delivery; (ii) not receiving corticosteroid treatment; (iii) first corticosteroid dose was less than or greater than 6 mg of dexamethasone or equivalent; (iv) “do not intubate” orders, comfort care status before initiation of dexamethasone; (v) no supplemental oxygen required within 12 hours of starting 6 mg of dexamethasone; (vi) not meeting our definition for respiratory deterioration or persistence; (vii) receiving anakinra before starting dexamethasone 6 mg; (viii) escalation to anakinra or high-dose corticosteroids (>10 mg of dexamethasone or equivalent) before meeting our definition of respiratory status deterioration or persistence; or (ix) not meeting our definition of anakinra use or high- or low-dose corticosteroid use within 48 hours of respiratory status deterioration or persistence. To minimize immortal time bias, an additional exclusion criterion was receiving the first dose of anakinra or high-dose corticosteroids within 24 hours of the first dose of dexamethasone 6mg in those patients with persistently moderate, severe, or very severe respiratory impairment (Figure 1).
      Figure 1
      Figure 1Consort diagram of cohort derivation.
      Abbreviations: DEXA, dexamethasone; DEXA6, 6mg dexamethasone daily; DNI, do not intubate; KPSC, Kaiser Permanente Southern California; KPNC, Kaiser Permanente Northern California; O2, oxygen; Td, time of deterioration/persistence.

      Data collection

      We retrieved data from the KPSC Research Data Warehouse and the KPNC Virtual Data Warehouse on exposures, outcomes, and covariates of interest, the latter of which included patient demographic characteristics, comorbidity burden as reflected by the Elixhauser Comorbidity Score, corticosteroid use in the 12 months before admission, body mass index (BMI), inpatient and emergency department utilization in the year before admission, other inpatient treatments (including anakinra, tocilizumab, convalescent plasma, and hemodialysis) and selected inflammatory markers, including d-dimer and CRP values. We also calculated values for the Epic deterioration index (DI), a score based on inpatient vital signs and laboratory values that has been shown to predict hospital mortality in hospitalized patients with and without COVID-19 (
      • Singh K
      • Valley TS
      • Tang S
      • Li BY
      • Kamran F
      • Sjoding MW
      • Nallamothu BK.
      Evaluating a widely implemented proprietary deterioration index model among hospitalized patients with COVID-19.
      ).

      Respiratory status deterioration or persistence

      To define respiratory status deterioration or persistence, we categorized patients according to the intensity of respiratory support they required. The categories included six tiers: mild impairment A (>0–2 l/min via nasal cannula, mild impairment B (>2-4 l/min via nasal cannula), mild impairment C (>4-6 l via nasal cannula or <40% fraction of inspired oxygen [FiO2] via face mask), moderate (6-15 l/min via nasal cannula or 40-60% FiO2 via mask), severe (>15 l/min via nasal cannula or high flow system, >60% FiO2 via a mask or non-invasive ventilation), or very severe (invasive mechanical ventilation). We defined respiratory deterioration as the first progression from one level of respiratory support to a higher level of support for any duration of time and persistence as remaining in the moderate, severe, or very severe group for at least 24 hours after initiation of DEXA6.

      Corticosteroid groups after respiratory deterioration or persistence

      Low-dose corticosteroid use was defined as a total daily dose of 6-10 mg of dexamethasone or equivalent doses of other corticosteroids. High-dose corticosteroid use was defined as a total daily dose of >10 mg of dexamethasone or equivalent doses of other corticosteroids. We did not distinguish between oral or intravenous administration or whether doses were divided into two or more daily doses. Groups were defined based on the highest prescribed corticosteroid dose within 48 hours after respiratory deterioration or persistence.

      Anakinra use after respiratory deterioration or persistence

      Anakinra is a short-acting IL-1 inhibitor that is prescribed in combination with corticosteroids in the treatment of cytokine storm. It can be administered subcutaneously or intravenously, although in our system, the intravenous formulation was not available at the time of this study. Doses are adjusted in patients with renal failure. To account for variation in anakinra dosing based on clinicians’ preferences for high or low doses and to eliminate transient use that would not be expected to improve COVID-19-related cytokine storm, anakinra use after respiratory deterioration or persistence was defined as follows: (i) five or more consecutive days of treatment regardless of daily dose; (ii) at least three consecutive days of high-dose anakinra (≥400 mg per day or ≥200 mg per day for those with an estimated glomerular filtration rate (GFR) of <80 ml/min); or (iii) at least 1 day of anakinra at any dose and discontinuation due to death.

      Outcomes

      The primary outcome was in-hospital death. Secondary outcomes were hospital length of stay, total duration of mechanical ventilation, and infectious complications. We assessed the frequency of all hospital-acquired infections and specific infections (bloodstream, urinary tract, soft tissue, and pulmonary infections other than COVID-19) identified by using the International Classification of Diseases (ICD)-10 codes (Appendix Table 1) entered by professional medical coders on discharge.

      Statistical analysis

      We described continuous and categorical variables using means and SD and counts and percentages, respectively. To reduce the imbalance in demographic and clinical characteristics among the treatment groups, we fit propensity score (PS) models using multinomial regression to estimate the probability of receiving each treatment as a function of predictor variables that were plausibly associated with both exposure and outcome (
      • Imbens GW.
      The role of the propensity score in estimating dose-response functions.
      ). Covariates included age (continuous), sex (male/female), race and ethnicity (categorical: white, Black, Hispanic, Asian or other/unknown), smoking (categorical: never, former, current), BMI (categorical, normal/underweight, overweight, moderately obese or severely obese), episodes of inpatient, emergency department and urgent care utilization in the year before admission (continuous); Elixhauser comorbidity score (continuous), Epic DI (continuous), respiratory status (categorical: mild, moderate, severe or very severe), treatment with remdesivir, dialysis or convalescent plasma on or before day of deterioration/persistence (yes/no); steroid use 1 year before admission (yes/no); and month of admission (categorical). Patients with unknown BMI were not included in multivariable models (n = 112 who continued 6mg of dexamethasone and n = 2 in the high-dose corticosteroid escalation group).
      We analyzed in-hospital mortality by fitting logistic regression models with inverse probability of treatment weights based on PSs. We used robust sandwich estimators to obtain appropriate standard errors of treatment effects because the weighting induced a within-subject correlation in outcomes and an inflated sample size (
      • Hernán MA
      • Brumback B
      • Robins JM.
      Marginal structural models to estimate the causal effect of zidovudine on the survival of HIV-positive men.
      ;
      • van der Wal WM
      ipw: an R package for inverse probability weighting.
      ;
      • Xu S
      • Ross C
      • Raebel MA
      • Shetterly S
      • Blanchette C
      • Smith D.
      Use of stabilized inverse propensity scores as weights to directly estimate relative risk and its confidence intervals.
      ). PS-adjusted odds ratios (OR) were reported with continuing low-dose corticosteroids after first deterioration/persistence as the reference group. To examine the effect of CRP levels at the time of deterioration/persistence, we conducted sensitivity analyses restricted to those patients with CRP values available within 24 hours before deterioration/persistence, where the models described previously were additionally adjusted for the highest CRP value within 24 hours of the time of deterioration/persistence.
      All analyses were conducted using SAS 9.4 (Cary, North Carolina). Two-tailed tests were used with the threshold of 0.05 for statistical significance.

      Results

      Among 28,035 patients with COVID-19 admitted between March 1, 2020, to December 31, 2020, 21,364 were excluded for a final study population of 6671, of whom 2667 were from KPSC and 4004 were from KPNC. The main reasons for exclusion were lack of corticosteroid use; receiving high rather than low corticosteroid dose at initiation; oxygen requirements that did not worsen or remain persistently poor while on low-dose corticosteroids; or no supplemental oxygen requirements within 48 hours of low-dose corticosteroid initiation (Figure 1).
      Most of the patients whose respiratory status deteriorated or remained persistently poor while being treated with low-dose corticosteroids remained on low-dose corticosteroids (n = 6265, 93.9%), whereas 232 (3.5%) patients were escalated to high-dose corticosteroids and 174 (2.6%) were started on anakinra within 48 hours of the time of deterioration (Table 1). All anakinra-treated patients continued to receive concomitant corticosteroids. After respiratory status deterioration, the median daily dose of dexamethasone or equivalent was 6 mg (interquartile range [IQR] 6-6 mg) in the low-dose group, 12 mg (IQR 7.5-20 mg) in the high-dose group, and 6 mg (IQR 6-15 mg) in the anakinra group. The median duration of corticosteroid treatment was 6 days (IQR 4-9) in the low-dose group, 9 days (IQR 6-14) in the high-dose group, and 9 days (IQR 6-14) in the anakinra group. The median duration of anakinra use was 8 days (IQR 5-10), and the median anakinra dose was 400 mg (IQR 400-400). Of the patients in the low-dose corticosteroid group, 10.3% (n = 645) were later switched to high-dose corticosteroids (n = 541) or anakinra (n = 104) after a median of 9 days.
      Table 1Patient characteristics on day of deterioration.
      >10 mg dexamethasone (N = 232)Anakinra (N = 174)Remained on DEXA6 (N = 6265)Total (N = 6671)
      Age, mean (SD), years61.8 (15.7)59.9 (15.3)61.4 (15.3)61.4 (15.3)
      Male gender, n (%)144 (62.1%)127 (73.0%)3825 (61.1%)4096 (61.4%)
      Race and ethnicity, n (%)
       Asian31 (13.4%)19 (10.9%)723 (11.5%)773 (11.6%)
       Black16 (6.9%)12 (6.9%)486 (7.8%)514 (7.7%)
       Hispanic128 (55.2%)109 (62.6%)3540 (56.5%)3777 (56.6%)
       White45 (19.4%)23 (13.2%)1177 (18.8%)1245 (18.7%)
       Other
      Other race/ethnicity includes multiple, native American Alaskan, Pacific Islander, other, unknown
      12 (5.2%)11 (6.3%)336 (5.4%)362 (5.4%)
      Body mass index categories, n (%)
       Underweight1 (0.4%)0 (0.0%)46 (0.7%)47 (0.7%)
       Normal22 (9.5%)23 (13.2%)701 (11.2%)746 (11.2%)
       Overweight66 (28.4%)62 (35.6%)1740 (27.8%)1868 (28.0%)
       Moderate obese101 (43.5%)68 (39.1%)2678 (42.7%)2847 (42.7%)
       Severe obese40 (17.2%)21 (12.1%)988 (15.8%)1049 (15.7%)
       Unknown2 (0.9%)0 (0.0%)112 (1.8%)114 (1.7%)
      Smoking, n (%)
       Never/passive125 (53.9%)92 (52.9%)3441 (54.9%)3658 (54.8%)
       Quit82 (35.3%)45 (25.9%)1806 (28.8%)1933 (29.0%)
       Active5 (2.2%)4 (2.3%)135 (2.2%)144 (2.2%)
       Unknown20 (8.6%)33 (19.0%)883 (14.1%)936 (14.0%)
      Oral corticosteroids in the past 12 months, n (%)48 (20.7%)26 (14.9%)973 (15.5%)1047 (15.7%)
      Elixhauser comorbidities, median (IQR)5.0 (3.0, 7.0)3.0 (2.0, 5.0)4.0 (2.0, 6.0)4.0 (2.0, 6.0)
      Healthcare utilizations in the past 12 months
      Inpatient visit count, n (%)
       0 visit205 (88.4%)156 (89.7%)5604 (89.4%)5965 (89.4%)
       1-2 visits23 (9.9%)16 (9.2%)557 (8.9%)596 (8.9%)
       3 or more visits4 (1.7%)2 (1.1%)104 (1.7%)110 (1.6%)
      Emergency department visit count, n (%)
       0 visit139 (59.9%)117 (67.2%)3933 (62.8%)4189 (62.8%)
       1-2 visits72 (31.0%)49 (28.2%)1922 (30.7%)2043 (30.6%)
       3 or more visits21 (9.1%)8 (4.6%)410 (6.5%)439 (6.6%)
      Urgent care visit count, n (%)
       0 visit180 (77.6%)108 (62.1%)4767 (76.1%)5055 (75.8%)
       1-2 visits44 (19.0%)50 (28.7%)1214 (19.4%)1308 (19.6%)
       3 or more visits8 (3.4%)16 (9.2%)284 (4.5%)308 (4.6%)
      COVID severity on day of deterioration
      Time from DEXA6 initiation to day of deterioriation, median (IQR), days1.0 (0.8, 1.7)1.0 (0.8, 1.3)1.0 (0.9, 1.4)1.0 (0.9, 1.4)
      Epic deterioration index
      Epic deterioration index score missing in five patients who remained on DEXA6
      , median (IQR)
      51.8 (15.2)49.4 (13.3)48.5 (15.0)48.6 (14.9)
      Respiratory impairment, n (%)
       Mild A (>0 and <2 l/min via N/C)0 (0%)0 (0%)0 (0%)0 (0%)
       Mild B (>2 and <4 l/min via N/C)28 (12.1%)28 (16.1%)1297 (20.7%)1353 (20.3%)
       Mild C (>4 and <6 l/min via N/C or 40% fraction of inspired oxygen via mask)29 (12.5%)31 (17.8%)902 (14.4%)962 (14.4%)
       Moderate (6-15 l/min via N/C or 40%-60% via mask)39 (16.8%)35 (20.1%)1089 (17.4%)1163 (17.4%)
       Severe (high flow, Non-invasive ventilation, >15 l/min via N/C or >60% via mask)94 (40.5%)69 (39.7%)2447 (39.1%)2610 (39.1%)
       Invasive mechanical ventilation42 (18.1%)11 (6.3%)530 (8.5%)583 (8.7%)
      Other treatments on or before day of deterioration
       Remdesivir, n (%)147 (63.4%)103 (59.2%)4311 (68.8%)4561 (68.4%)
       Convalescent Plasma, n (%)15 (6.5%)22 (12.6%)310 (4.9%)347 (5.2%)
       Tocilizumab, n (%)0 (0.0%)0 (0.0%)4 (0.1%)4 (0.1%)
       Dialysis, n (%)1 (0.4%)3 (1.7%)98 (1.6%)102 (1.5%)
      Selected laboratory values on day of deterioration
      In patients with multiple values within 24 hours of deterioration, the highest value was recorded Abbreviations: DEXA6, 6mg dexamethasone; IQR, interquartile range; N/C, nasal cannula.
       C-reactive protein, n18015745594896
       Median (IQR), mg/l123.4 (68.0, 193.4)150.4 (79.6, 205.6)118.0 (63.4, 184.0)119.0 (64.2, 185.7)
       D-Dimer, n16916242784609
       Median (IQR), microgram fibrinogen equivalent units/ml0.9 (0.6, 1.5)1.0 (0.7, 2.0)1.0 (0.6, 1.7)1.0 (0.6, 1.7)
      a Other race/ethnicity includes multiple, native American Alaskan, Pacific Islander, other, unknown
      b Epic deterioration index score missing in five patients who remained on DEXA6
      c In patients with multiple values within 24 hours of deterioration, the highest value was recordedAbbreviations: DEXA6, 6mg dexamethasone; IQR, interquartile range; N/C, nasal cannula.
      Patients escalated to high-dose corticosteroids had a greater comorbidity burden, higher Epic DI scores, and more severe respiratory impairment (>15 l/min of oxygen or mechanical ventilation) at the time of deterioration compared with patients who remained on low-dose corticosteroids (Table 1). Patients who escalated to anakinra treatment had slightly higher Epic DI scores and CRP levels but similar severity of respiratory impairment at the time of deterioration and were more likely to be male and to have received convalescent plasma before deterioration compared with patients who remained on low-dose corticosteroids. All three groups were similar in age, BMI, race and ethnicity, smoking behavior, health care utilization in the 12 months before admission, and d-dimer levels at the time of deterioration. Treatment with remdesivir before deterioration was common, whereas hemodialysis was rare, and neither differed significantly between groups.
      Crude in-hospital mortality was higher among patients who received high-dose corticosteroids (41.9%, 95% CI 35.5-48.3%) or anakinra (32.8%, 95% CI 25.8-39.7%) compared with the low-dose group (26.2%, 95% CI 25.1-27.3%). These differences remained significant in the PS-adjusted models for both high-dose corticosteroids (OR 1.53; 95% CI 1.14-2.07) and anakinra (OR 1.76; 95% CI 1.13-2.72; Table 2). The average length of hospital stay was significantly longer in the anakinra and high-dose corticosteroid groups compared with the low-dose group in PS-adjusted models (4.8 and 2.3 days, respectively; Table 2). There was no significant difference in ventilator days in either the high-dose or anakinra group compared with the low-dose group after PS adjustment (Table 2).
      Table 2Anakinra or high-dose corticosteroids on day of deterioration while treated with dexamethasone 6 mg and primary and secondary outcomes.
      Treatment
      Remained on 6 mg daily of dexamethasoneAnakinra>10 mg dexamethasone or equivalent corticosteroid dosing
      N6107174229
      Mortality, n (%)1599 (26.2)57 (32.8)96 (41.9)
      Unadjusted, OR (95% CI)Reference1.37 (1.00, 1.90)2.04 (1.56, 2.66)
      PS adjustment (IPTW) with robust standard errors, OR (95% CI)
      The PS model for all outcomes included age, gender, race and ethnicity, smoking, body mass index, Elixhauser Comorbidity Index, Epic deterioration index score, severity of respiratory impairment and dialysis on day 1 of admission; remdesivir, convalescent plasma, or anakinra use on day 1 of admission; and inpatient, outpatient, and urgent care utilizations and corticosteroid use in the 12 months before admission. Abbreviations: CI, confidence interval; IPTW, inverse probability of treatment weights; OR, odds ratio; PS, propensity score.
      Reference1.76 (1.13, 2.72)1.53 (1.14, 2.07)
      Hospital length of stay, mean (SD)12.1 (12.4)16.4 (12.8)15.6 (12.6)
      Unadjusted mean difference (95% CI)Reference4.4 (2.5, 6.3)3.6 (1.9, 5.2)
      PS adjustment (IPTW) with robust standard errors, mean difference
      The PS model for all outcomes included age, gender, race and ethnicity, smoking, body mass index, Elixhauser Comorbidity Index, Epic deterioration index score, severity of respiratory impairment and dialysis on day 1 of admission; remdesivir, convalescent plasma, or anakinra use on day 1 of admission; and inpatient, outpatient, and urgent care utilizations and corticosteroid use in the 12 months before admission. Abbreviations: CI, confidence interval; IPTW, inverse probability of treatment weights; OR, odds ratio; PS, propensity score.
      (95% CI)
      Reference4.8 (2.5, 7.1)2.3 (0.6, 3.9)
      Ventilator days, mean (SD)4.3 (10.3)5.6 (9.8)7.5 (12.0)
      Unadjusted mean difference (95% CI)Reference1.3 (-0.2, 2.8)3.2 (1.6, 4.8)
      PS adjustment (IPTW) with robust standard errors, mean difference
      The PS model for all outcomes included age, gender, race and ethnicity, smoking, body mass index, Elixhauser Comorbidity Index, Epic deterioration index score, severity of respiratory impairment and dialysis on day 1 of admission; remdesivir, convalescent plasma, or anakinra use on day 1 of admission; and inpatient, outpatient, and urgent care utilizations and corticosteroid use in the 12 months before admission. Abbreviations: CI, confidence interval; IPTW, inverse probability of treatment weights; OR, odds ratio; PS, propensity score.
      Reference0.9 (-0.5, 2.4)1.4 (0.0, 2.8)
      Hospital-acquired infection (%)14.225.323.1
      Unadjusted, OR (95% CI)Reference2.05 (1.45, 2.91)1.83 (1.33, 2.50)
      PS adjustment (IPTW) with robust standard errors, OR
      The PS model for all outcomes included age, gender, race and ethnicity, smoking, body mass index, Elixhauser Comorbidity Index, Epic deterioration index score, severity of respiratory impairment and dialysis on day 1 of admission; remdesivir, convalescent plasma, or anakinra use on day 1 of admission; and inpatient, outpatient, and urgent care utilizations and corticosteroid use in the 12 months before admission. Abbreviations: CI, confidence interval; IPTW, inverse probability of treatment weights; OR, odds ratio; PS, propensity score.
      (95% CI)
      Reference2.00 (1.28, 3.11)1.43 (1.00, 2.04)
      Bloodstream infection (%)10.619.519.2
      Unadjusted, OR (95% CI)Reference2.06 (1.40, 3.02)2.01 (1.44, 2.83)
      PS adjustment (IPTW) with robust standard errors, OR
      The PS model for all outcomes included age, gender, race and ethnicity, smoking, body mass index, Elixhauser Comorbidity Index, Epic deterioration index score, severity of respiratory impairment and dialysis on day 1 of admission; remdesivir, convalescent plasma, or anakinra use on day 1 of admission; and inpatient, outpatient, and urgent care utilizations and corticosteroid use in the 12 months before admission. Abbreviations: CI, confidence interval; IPTW, inverse probability of treatment weights; OR, odds ratio; PS, propensity score.
      (95% CI)
      Reference1.99 (1.22, 3.27)1.60 (1.10 ,2.33)
      Pulmonary infection (%)5.210.98.3
      Unadjusted, OR (95% CI)Reference2.24 (1.37, 3.66)1.65 (1.02, 2.68)
      PS adjustment (IPTW) with robust standard errors, OR
      The PS model for all outcomes included age, gender, race and ethnicity, smoking, body mass index, Elixhauser Comorbidity Index, Epic deterioration index score, severity of respiratory impairment and dialysis on day 1 of admission; remdesivir, convalescent plasma, or anakinra use on day 1 of admission; and inpatient, outpatient, and urgent care utilizations and corticosteroid use in the 12 months before admission. Abbreviations: CI, confidence interval; IPTW, inverse probability of treatment weights; OR, odds ratio; PS, propensity score.
      (95% CI)
      Reference2.31 (1.29, 4.14)1.13 (0.65, 1.98)
      Soft tissue infection (%)0.10.00.0
      Urinary tract infection (%)2.35.23.1
      Unadjusted, OR (95% CI)Reference2.34 (1.17, 4.68)1.35 (0.63, 2.93)
      PS adjustment (IPTW) with robust standard errors, OR
      The PS model for all outcomes included age, gender, race and ethnicity, smoking, body mass index, Elixhauser Comorbidity Index, Epic deterioration index score, severity of respiratory impairment and dialysis on day 1 of admission; remdesivir, convalescent plasma, or anakinra use on day 1 of admission; and inpatient, outpatient, and urgent care utilizations and corticosteroid use in the 12 months before admission. Abbreviations: CI, confidence interval; IPTW, inverse probability of treatment weights; OR, odds ratio; PS, propensity score.
      (95% CI)
      Reference2.14 (0.95, 4.80)1.39 (0.58, 3.35)
      a The PS model for all outcomes included age, gender, race and ethnicity, smoking, body mass index, Elixhauser Comorbidity Index, Epic deterioration index score, severity of respiratory impairment and dialysis on day 1 of admission; remdesivir, convalescent plasma, or anakinra use on day 1 of admission; and inpatient, outpatient, and urgent care utilizations and corticosteroid use in the 12 months before admission.Abbreviations: CI, confidence interval; IPTW, inverse probability of treatment weights; OR, odds ratio; PS, propensity score.
      Treatment with anakinra (PS-adjusted OR 2.00; 95% CI 1.28-3.11) or high-dose corticosteroids (PS-adjusted OR 1.43; 95% CI 1.00-2.04) was associated with higher odds of hospital-acquired infections in crude and PS-adjusted models compared with continuing low-dose corticosteroids at the time of deterioration (Table 2). Bloodstream infections and non-COVID-19 pneumonia were the most common hospital-acquired infections across all three groups. Anakinra use was associated with a significantly increased odds of bloodstream infections and hospital-acquired pneumonia in PS-adjusted models. High-dose corticosteroid treatment was associated with a significantly increased odds of bloodstream infections (Table 2).
      The results of sensitivity analyses in which we adjusted for the highest CRP values within 24 hours of deterioration among the 4896 (73.4%) with such available CRP values yielded similar results. In these analyses, treatment with anakinra (adjusted OR 1.94; 95% CI 1.22-3.07) or high-dose corticosteroids (adjusted OR 1.71; 95% CI 1.20-2.44) was associated with higher odds of in-hospital mortality compared with continuing low-dose corticosteroids at the time of deterioration.

      Discussion

      We found that escalating immunosuppression with high-dose corticosteroids or anakinra in patients with COVID-19 pneumonia whose respiratory function deteriorated or remained persistently poor, whereas being treated with DEXA6 was not associated with reduced mortality. Both anakinra and high-dose corticosteroid treatment initiation on the day of deterioration were associated with higher odds of in-hospital death and hospital-acquired infections compared with those who remained on DEXA6. These findings add to the growing literature supporting DEXA6 as the dose of choice for patients with COVID-19 pneumonia requiring supplemental oxygen (

      National Institutes of Health. COVID-19 treatment guidelines: corticosteroids, May 31, 2022. https://www.covid19treatmentguidelines.nih.gov/therapies/immunomodulators/corticosteroids/, 2022 (accessed 29 June 2022).

      ) and do not support intensifying immunosuppression in patients whose respiratory status deteriorates on DEXA6.
      DEXA6 presumably improves survival in patients with COVID-19 pneumonia by dampening lung inflammation and systemic inflammation in the subset of patients that exhibit signs and symptoms of cytokine storm (
      • Mehta P
      • McAuley DF
      • Brown M
      • Sanchez E
      • Tattersall RS
      • Manson JJ
      HLH Across Speciality Collaboration, UK. COVID-19: consider cytokine storm syndromes and immunosuppression.
      ). Since early in the pandemic, many clinicians, including those at our hospitals (
      • Langer-Gould A
      • Smith JB
      • Gonzales EG
      • Castillo RD
      • Figueroa JG
      • Ramanathan A
      • Gould MK.
      Early identification of COVID-19 cytokine storm and treatment with anakinra or tocilizumab.
      ), have treated patients with severe COVID pneumonia, particularly those with in-hospital respiratory status deterioration and rising inflammatory markers like CRP and ferritin, with high-dose corticosteroids, anakinra, or tocilizumab, albeit with mixed results. Once DEXA6 became the standard of care, a common clinical challenge became how best to treat those that deteriorate or do not improve on this low dose of corticosteroids. Without evidence to guide them, some clinicians opted to intensify immunosuppression. The findings from this study, however, indicate that intensifying immunosuppression with high-dose corticosteroids or anakinra in these patients does not lead to improved survival compared with continuing the DEXA6 protocol for a full 10 days. We speculate that the lack of improvement we observed reflects the multiple non-hyperinflammatory pathophysiological processes that can lead to respiratory decompensation, including direct viral damage, hypercoagulability, and/or secondary infections.
      Our findings do not exclude the possibility that initiation of anakinra upon hospitalization in patients with COVID-19 without severe respiratory impairment is beneficial, as indicated by previous studies (
      • Kyriazopoulou E
      • Huet T
      • Cavalli G
      • Gori A
      • Kyprianou M
      • Pickkers P
      International Collaborative Group for Anakinra in COVID-19. Effect of anakinra on mortality in patients with COVID-19: a systematic review and patient-level meta-analysis.
      ,
      • Kyriazopoulou E
      • Poulakou G
      • Milionis H
      • Metallidis S
      • Adamis G
      • Tsiakos K
      • Giamarellos-Bourboulis EJ.
      Early treatment of COVID-19 with anakinra guided by soluble urokinase plasminogen receptor plasma levels: a double-blind, randomized controlled phase 3 trial.
      ). A meta-analysis of primarily observational studies suggested that anakinra may be effective when administered to patients with COVID-19 pneumonia early in the disease course and in those with elevated CRP (>100), ferritin or soluble urokinase plasminogen activator receptor levels (
      • Kyriazopoulou E
      • Poulakou G
      • Milionis H
      • Metallidis S
      • Adamis G
      • Tsiakos K
      • Giamarellos-Bourboulis EJ.
      Early treatment of COVID-19 with anakinra guided by soluble urokinase plasminogen receptor plasma levels: a double-blind, randomized controlled phase 3 trial.
      ). These findings led to a double-blinded RCT of anakinra plus DEXA6 compared with DEXA6 alone (
      • Kyriazopoulou E
      • Huet T
      • Cavalli G
      • Gori A
      • Kyprianou M
      • Pickkers P
      International Collaborative Group for Anakinra in COVID-19. Effect of anakinra on mortality in patients with COVID-19: a systematic review and patient-level meta-analysis.
      ) initiated within 2-3 days of hospitalization in patients with COVID-19 pneumonia without very severe respiratory impairment and with elevated soluble urokinase plasminogen activator receptor (SuPAR) levels and median CRP levels of 50 (
      • Kyriazopoulou E
      • Poulakou G
      • Milionis H
      • Metallidis S
      • Adamis G
      • Tsiakos K
      • Giamarellos-Bourboulis EJ.
      Early treatment of COVID-19 with anakinra guided by soluble urokinase plasminogen receptor plasma levels: a double-blind, randomized controlled phase 3 trial.
      ). This RCT, the SAVE-MORE trial, showed overall low mortality rates but a significant reduction in mortality in the anakinra-treated group compared with DEXA6 alone (3.2% and 6.9%, respectively). The two open-label RCTs (
      CORIMUNO-19 Collaborative group
      Effect of anakinra versus usual care in adults in hospital with COVID-19 and mild-to-moderate pneumonia (CORIMUNO-ANA-1): a randomised controlled trial.
      ;

      REMAP-CAP. REMAP-CAP results, April 10, 2021. https://www.remapcap.org/covid19publications, 2021 (accessed 29 June 2022).

      ) that reported no difference in survival in anakinra-treated patients compared with standard of care (corticosteroids and anti-IL-6 treatments in one trial and supportive care in the other) both included a large fraction of patients with severe respiratory impairment and populations with much higher mortality rates, as was the case in this observational study, in contrast to the SAVE-MORE trial (
      • Kyriazopoulou E
      • Huet T
      • Cavalli G
      • Gori A
      • Kyprianou M
      • Pickkers P
      International Collaborative Group for Anakinra in COVID-19. Effect of anakinra on mortality in patients with COVID-19: a systematic review and patient-level meta-analysis.
      ). We excluded 207 patients that were started on anakinra after DEXA6 initiation but before meeting our definition of respiratory status deterioration or persistence. This raises the possibility that these patients were escalated due to rising inflammatory markers, as we previously recommended (
      • Langer-Gould A
      • Smith JB
      • Gonzales EG
      • Castillo RD
      • Figueroa JG
      • Ramanathan A
      • Gould MK.
      Early identification of COVID-19 cytokine storm and treatment with anakinra or tocilizumab.
      ). Whether anakinra reduces mortality without increasing the risk of hospital-acquired infections in patients with early or mild COVID-19 pneumonia is not addressed in the current study. Thus, the results of these RCTs and the findings reported herein and in our previous observational study (
      • Langer-Gould A
      • Smith JB
      • Gonzales EG
      • Castillo RD
      • Figueroa JG
      • Ramanathan A
      • Gould MK.
      Early identification of COVID-19 cytokine storm and treatment with anakinra or tocilizumab.
      ) could be viewed as consistent if early administration of anakinra is indeed critical.
      We also cannot exclude the possibility that escalating immunosuppression with anakinra or high-dose corticosteroids in patients with COVID-19 pneumonia who do not respond to DEXA6 is harmful. That anakinra and high-dose corticosteroid treatments can increase the risk of infections, as we observed, is an anticipated adverse event. It is plausible that either through increasing the risk of secondary infections or augmenting viral damage by inhibiting immune-mediated viral clearance mechanisms, intensifying immunosuppression later in the disease course increases mortality. Alternatively, our findings of increased odds of in-hospital infections and mortality, particularly in anakinra-treated patients, may be explained by unmeasured confounders, the main limitation of this study. Such potential unmeasured confounders include rising inflammatory markers and laboratory or other indices of (multi) organ failure. Although we did not find any decreased odds of death with intensified immunosuppression, it is very possible that residual confounding by indication with those patients judged to be at highest risk of death treated most aggressively explains the increased odds of mortality among anakinra or high-dose corticosteroid-treated patients we observed.
      Other limitations of this study are the relatively small number of patients escalated to high-dose corticosteroids or anakinra after respiratory status deterioration/persistence, precluding subgroup analyses stratified by severity of respiratory impairment or CRP levels on the day of deterioration. Lastly, whether these results can be extrapolated to vaccinated populations where far fewer patients experience severe respiratory impairment or to the omicron variant, or any future strain, is uncertain.
      Strengths of this study include the importance of the question and restricting the analysis to patients who were escalated to high-dose corticosteroids or anakinra only after deteriorating or failing to improve after initiation of DEXA6. This approach minimizes confounding by indication and the influence of potentially unmeasured changes in COVID-19 pneumonia care around the same time DEXA6 became standard of care.

      Interpretation

      Our findings, taken together with those from previous studies (
      • Bouadma L
      • Mekontso-Dessap A
      • Burdet C
      • Merdji H
      • Poissy J
      • Dupuis C
      COVIDICUS Study Group
      High-dose dexamethasone and oxygen support strategies in intensive care unit patients with severe COVID-19 acute hypoxemic respiratory failure: the COVIDICUS randomized clinical trial.
      ;
      • COVID STEROID 2 Trial Group
      • Munch MW
      • Myatra SN
      • Vijayaraghavan BKT
      • Saseedharan S
      • Benfield T
      • Perner A.
      Effect of 12 mg vs 6 mg of dexamethasone on the number of days alive without life support in adults with COVID-19 and severe hypoxemia: the COVID STEROID 2 randomized trial.
      ;
      • Langer-Gould A
      • Xu S
      • Myers LC
      • Chen A
      • Greene JD
      • Creekmur B
      • Gould MK.
      High-dose corticosteroids in patients hospitalized for COVID-19 pneumonia: an observational study of comparative effectiveness.
      ), support DEXA6 for up to 10 days as the dose of choice for patients with COVID-19 pneumonia requiring supplemental oxygen (

      National Institutes of Health. COVID-19 treatment guidelines: corticosteroids, May 31, 2022. https://www.covid19treatmentguidelines.nih.gov/therapies/immunomodulators/corticosteroids/, 2022 (accessed 29 June 2022).

      ) and do not support intensifying immunosuppression with high-dose corticosteroids or adding anakinra in patients whose respiratory status deteriorates after initiating DEXA6.

      Declaration of competing interest

      KB has received research support unrelated to this study from Moderna/Pfizer. LCM has received research support unrelated to this study from Boehringer Ingelheim. GJE has received research support unrelated to this study from Astra Zeneca. The other authors have no conflict of interest.

      Funding

      This project was supported in whole or in part by Kaiser Permanente's Garfield Memorial Fund. The funders were not involved in the design and conduct of the study; collection, management, analysis, and interpretation of the data; preparation, review, or approval of the manuscript; and decision to submit the manuscript for publication.

      Ethical approval

      We performed this retrospective cohort study after receiving approval from the KPSC Institutional Review Board (#12396).

      Author contributions

      ALG drafted the manuscript. SX performed the analysis. AC and JG extracted the data. BC coordinated the IRB approval and data transfer. MKG, LCM, VXL, KB, and JLA advised on the methods and interpretation of the results. ALG, VXL and MKG conceived of the idea, obtained the funding and oversaw the project. All authors critically reviewed the manuscript.

      Appendix. Supplementary materials

      References

        • Bouadma L
        • Mekontso-Dessap A
        • Burdet C
        • Merdji H
        • Poissy J
        • Dupuis C
        • COVIDICUS Study Group
        High-dose dexamethasone and oxygen support strategies in intensive care unit patients with severe COVID-19 acute hypoxemic respiratory failure: the COVIDICUS randomized clinical trial.
        JAMA Intern Med. 2022; 182: 906-916
        • CORIMUNO-19 Collaborative group
        Effect of anakinra versus usual care in adults in hospital with COVID-19 and mild-to-moderate pneumonia (CORIMUNO-ANA-1): a randomised controlled trial.
        Lancet Respir Med. 2021; 9: 295-304
        • COVID STEROID 2 Trial Group
        • Munch MW
        • Myatra SN
        • Vijayaraghavan BKT
        • Saseedharan S
        • Benfield T
        • Perner A.
        Effect of 12 mg vs 6 mg of dexamethasone on the number of days alive without life support in adults with COVID-19 and severe hypoxemia: the COVID STEROID 2 randomized trial.
        JAMA. 2021; 326: 1807-1817
        • Hernán MA
        • Brumback B
        • Robins JM.
        Marginal structural models to estimate the causal effect of zidovudine on the survival of HIV-positive men.
        Epidemiology. 2000; 11: 561-570
        • Imbens GW.
        The role of the propensity score in estimating dose-response functions.
        Biometrika. 2000; 87: 706-710
        • Kyriazopoulou E
        • Huet T
        • Cavalli G
        • Gori A
        • Kyprianou M
        • Pickkers P
        International Collaborative Group for Anakinra in COVID-19. Effect of anakinra on mortality in patients with COVID-19: a systematic review and patient-level meta-analysis.
        Lancet Rheumatol. 2021; 3: e690-e697
        • Kyriazopoulou E
        • Poulakou G
        • Milionis H
        • Metallidis S
        • Adamis G
        • Tsiakos K
        • Giamarellos-Bourboulis EJ.
        Early treatment of COVID-19 with anakinra guided by soluble urokinase plasminogen receptor plasma levels: a double-blind, randomized controlled phase 3 trial.
        Nat Med. 2021; 27: 1752-1760
        • Langer-Gould A
        • Smith JB
        • Gonzales EG
        • Castillo RD
        • Figueroa JG
        • Ramanathan A
        • Gould MK.
        Early identification of COVID-19 cytokine storm and treatment with anakinra or tocilizumab.
        Int J Infect Dis. 2020; 99: 291-297
        • Langer-Gould A
        • Xu S
        • Myers LC
        • Chen A
        • Greene JD
        • Creekmur B
        • Gould MK.
        High-dose corticosteroids in patients hospitalized for COVID-19 pneumonia: an observational study of comparative effectiveness.
        Int J Infect Dis. 2022; 125: 184-191
        • Mehta P
        • McAuley DF
        • Brown M
        • Sanchez E
        • Tattersall RS
        • Manson JJ
        HLH Across Speciality Collaboration, UK. COVID-19: consider cytokine storm syndromes and immunosuppression.
        Lancet. 2020; 395: 1033-1034
      1. National Institutes of Health. COVID-19 treatment Guidlines: interleukin-1 inhibitors, October 19, 2021. https://www.covid19treatmentguidelines.nih.gov/therapies/immunomodulators/interleukin-1-inhibitors/, 2021 (accessed 29 June 2022).

      2. National Institutes of Health. COVID-19 treatment guidelines: corticosteroids, May 31, 2022. https://www.covid19treatmentguidelines.nih.gov/therapies/immunomodulators/corticosteroids/, 2022 (accessed 29 June 2022).

        • Recovery Collaborative Group
        • Horby P
        • Lim WS
        • Emberson JR
        • Mafham M
        • Bell JL
        • Landray MJ
        Dexamethasone in hospitalized patients with Covid-19.
        N Engl J Med. 2021; 384: 693-704
      3. REMAP-CAP. REMAP-CAP results, April 10, 2021. https://www.remapcap.org/covid19publications, 2021 (accessed 29 June 2022).

        • Singh K
        • Valley TS
        • Tang S
        • Li BY
        • Kamran F
        • Sjoding MW
        • Nallamothu BK.
        Evaluating a widely implemented proprietary deterioration index model among hospitalized patients with COVID-19.
        Ann Am Thorac Soc. 2021; 18: 1129-1137
        • van der Wal WM
        ipw: an R package for inverse probability weighting.
        J Stat Softw. 2011; 43: 1-23
        • Xu S
        • Ross C
        • Raebel MA
        • Shetterly S
        • Blanchette C
        • Smith D.
        Use of stabilized inverse propensity scores as weights to directly estimate relative risk and its confidence intervals.
        Value Health. 2010; 13: 273-277