A Clinical Predictor of Lung Molecular Endotype Identifies Heterogeneity in Corticosteroid Response in Severe COVID-19: an Emulated Target Trial
et al., medRxiv, doi:10.64898/2026.06.08.26355201, Jun 2026
Target trial emulation of 429 mechanically ventilated patients with severe COVID-19 in the USA, showing that corticosteroids were associated with lower 28-day mortality in patients with a predicted Hyper-Inflammatory endotype (OR 0.62, 95% CI 0.39-0.99), but showed higher risk, without statistical significance, in those with a predicted Metabolic Dysregulation endotype (OR 1.15, 95% CI 0.82-1.61).
The results suggest that a clinically-deployable machine-learning predictor of lung molecular endotype - using routine bedside data like respiratory rate, albumin, mean corpuscular hemoglobin, and anion gap - could identify mechanically ventilated COVID-19 patients most likely to benefit from corticosteroids.
Corticosteroids in COVID-19 and ARDS carry harms - broad immunosuppression that impairs viral clearance and host defenses (driving secondary bacterial and fungal infection), hyperglycemia, impaired healing, and myopathy, offset by an anti-inflammatory benefit that depends on how much steroid-suppressible inflammation is driving lung injury.
Because benefit appears to rise with severity faster than harm, there is a crossover point where treatment goes from harmful to beneficial. Guidelines typically tie the treatment decision to the level of oxygen support, however this is a crude proxy for the underlying inflammatory biology: this study shows that even within a ventilated population, an inflammatory endotype can classify patients into groups with predicted benefit or no benefit.
A natural refinement is to model each patient's treatment effect as a function of oxygen requirements and a continuous inflammatory score, where each patient has their own benefit-harm crossover point. For example, a strongly hyperinflammatory patient might warrant treatment at high-flow oxygen, while a metabolically-dysregulated patient may cross into net benefit only if ventilated, or not at all.
Standard of Care (SOC) for COVID-19 in the study country,
the USA, is very poor with very low average efficacy for approved treatments1.
Only expensive, high-profit treatments were approved for early treatment. Low-cost treatments were excluded, reducing the probability of early treatment due to access and cost barriers, and eliminating complementary and synergistic benefits seen with many low-cost treatments.
|
risk of death, 8.0% lower, OR 0.92, p = 0.56, treatment 284, control 145, propensity score weighting, day 28, RR approximated with OR.
|
|
risk of death, 11.0% lower, HR 0.89, p = 0.49, treatment 284, control 145, propensity score weighting, day 28.
|
|
risk of death, 38.0% lower, OR 0.62, p = 0.04, predicted Hyper-Inflammatory endotype, propensity score weighting, day 28, RR approximated with OR.
|
|
risk of death, 33.0% lower, HR 0.67, p = 0.13, predicted Hyper-Inflammatory endotype, propensity score weighting, day 28.
|
|
risk of death, 15.0% higher, OR 1.15, p = 0.42, predicted Metabolic Dysregulation endotype, propensity score weighting, day 28, RR approximated with OR.
|
|
risk of death, 4.0% higher, HR 1.04, p = 0.87, predicted Metabolic Dysregulation endotype, propensity score weighting, day 28.
|
| Effect extraction follows pre-specified rules prioritizing more serious outcomes. Submit updates |
Sines et al., 10 Jun 2026, retrospective, USA, preprint, 12 authors, study period 1 January, 2020 - 31 December, 2022.
Contact: benjamin_sines@med.unc.edu.
Abstract: It is made available under a CC-BY-NC-ND 4.0 International license . is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) medRxiv preprint doi: https://doi.org/10.64898/2026.06.08.26355201; this version posted June 10, 2026. The copyright holder for this preprint
- A Clinical Predictor of Lung Molecular Endotype Identifies Heterogeneity in Corticosteroid 1
- Response in Severe COVID-19: an Emulated Target Trial 2
3
- Benjamin J Sines, MD MSCR 1 *, Robert S Hagan MD PhD 1,4 , Xi Jiang PhD 2 , Ella Pavlechko
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
- PhD 2 , Scott McClain PhD 2 , Xin Hunt PhD 2 , Julia Florou-Moreno PhD 2 , Jake Acquadro 2 , Gabriel
- Risa 2 , Varunraj Valsaraj MS 2 , Jonathan C Schisler MS PhD 3† , Matthew C Wolfgang PhD 4†
- 1 Division of Pulmonary Critical Care Medicine, University of North Carolina at Chapel Hill,
- Chapel Hill, NC, United States
- 2 SAS Institute Inc, Cary, NC, United States
- 3 McAllister Heart Institute and the Department of Pharmacology, The University of North
- Carolina at Chapel Hill, Chapel Hill, NC, United States
- 4 Marsico Lung Institute and Department of Microbiology and Immunology, University of North
Carolina at Chapel Hill, Chapel Hill, NC, United States
†
Contributed equally as co-senior authors
Corresponding author:
Benjamin Sines, MD, MSCR Division of Pulmonary Diseases and Critical Care Medicine Department of Medicine University of North Carolina School of Medicine Bioinformatics 4103c 130 Mason Farm Road Chapel Hill, NC 27599 EM: benjamin\_sines@med.unc.edu
Conflict of Interest: The authors have declared that no conflict of interest exists. Word count: 3,524
It is made available under a CC-BY-NC-ND 4.0 International license . is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) medRxiv preprint doi: https://doi.org/10.64898/2026.06.08.26355201; this version posted June 10, 2026. The copyright holder for this preprint
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
ABSTRACT
Background: Corticosteroids reduce mortality in severe COVID-19 requiring oxygen or invasive mechanical ventilation, yet emerging data suggest that SARS-CoV-2-associated acute lung injury is biologically heterogeneous and that treatment response may vary across molecularly defined disease states. Lung-derived molecular endotypes of severe COVID-19-associated acute lung injury have been described, but direct molecular profiling is not routinely available at the bedside. We evaluated whether a clinical predictor of previously defined lung molecular endotype identifies heterogeneity in corticosteroid treatment effect among mechanically ventilated patients with COVID-19. Methods: We utilized a single-center cohort of 5,000 patients with COVID-19 treated at the University of North Carolina Hospital between January 1, 2020, and December 31, 2022, to emulate a target trial assessing the effect of corticosteroid receipt on mortality, length of stay, and incident organ support. Confounding was addressed through inverse probability of treatment weighting (IPTW). Outcomes for severely ill patients requiring mechanical ventilation were compared to the RECOVERY trial results, with subsequent moderation analysis and stratified analysis by..
DOI record:
{
"DOI": "10.64898/2026.06.08.26355201",
"URL": "http://dx.doi.org/10.64898/2026.06.08.26355201",
"abstract": "<jats:title>ABSTRACT</jats:title>\n <jats:sec>\n <jats:title>Background</jats:title>\n <jats:p>Corticosteroids reduce mortality in severe COVID-19 requiring oxygen or invasive mechanical ventilation, yet emerging data suggest that SARS-CoV-2-associated acute lung injury is biologically heterogeneous and that treatment response may vary across molecularly defined disease states. Lung-derived molecular endotypes of severe COVID-19-associated acute lung injury have been described, but direct molecular profiling is not routinely available at the bedside. We evaluated whether a clinical predictor of previously defined lung molecular endotype identifies heterogeneity in corticosteroid treatment effect among mechanically ventilated patients with COVID-19.</jats:p>\n </jats:sec>\n <jats:sec>\n <jats:title>Methods</jats:title>\n <jats:p>We utilized a single-center cohort of 5,000 patients with COVID-19 treated at the University of North Carolina Hospital between January 1, 2020, and December 31, 2022, to emulate a target trial assessing the effect of corticosteroid receipt on mortality, length of stay, and incident organ support. Confounding was addressed through inverse probability of treatment weighting (IPTW). Outcomes for severely ill patients requiring mechanical ventilation were compared to the RECOVERY trial results, with subsequent moderation analysis and stratified analysis by clinically predicted lung molecular endotype and vaccination status. The primary outcome was 28-day mortality. Secondary Outcomes were time to discharge alive and progression to additional organ support.</jats:p>\n </jats:sec>\n <jats:sec>\n <jats:title>Results</jats:title>\n <jats:p>This emulated target trial showed a directionally favorable but non-statistically significant association between corticosteroid treatment and reduced 28-day mortality in patients requiring mechanical ventilation for SARS-CoV-2 infection. A clinical predictor of lung molecular endotype moderated the effect of corticosteroids on 28-day mortality (p-value for interaction 0.038) and identified distinct predicted endotype-specific treatment effect. Corticosteroid treatment was associated with lower 28-day mortality in the predicted Hyper-Inflammatory endotype (OR 0.62, 95% CI 0.39, 0.99) but not in the predicted Metabolic Dysregulation endotype (OR 1.15, 95% CI 0.82, 1.61). We did not detect significant effect modification by vaccination status (p-value for interaction 0.65), although inference was limited by the small, vaccinated subgroup (28-mortality OR 0.78, 95% CI 0.37, 1.65 in vaccinated vs 0.94, 95% CI 0.70, 1.26 in unvaccinated).</jats:p>\n </jats:sec>\n <jats:sec>\n <jats:title>Conclusions</jats:title>\n <jats:p>In this target trial emulation of mechanically ventilated patients with severe COVID-19, corticosteroid treatment showed a directionally favorable but non-statistically significant association with reduced 28-day mortality in the overall cohort. However, a clinical predictor of lung molecular endotype identified significant heterogeneity in treatment effect, with benefit concentrated in the predicted Hyper-Inflammatory endotype and no apparent benefit in the predicted Metabolic Dysregulation endotype. These findings support prospective validation of clinically deployable endotype-guided corticosteroid treatment strategies in acute lung injury and ARDS.</jats:p>\n </jats:sec>",
"accepted": {
"date-parts": [
[
2026,
6,
10
]
]
},
"author": [
{
"ORCID": "https://orcid.org/0000-0001-5606-0238",
"affiliation": [
{
"name": "Division of Pulmonary Critical Care Medicine, University of North Carolina at Chapel Hill, Chapel Hill, NC, United States"
}
],
"authenticated-orcid": false,
"family": "Sines",
"given": "Benjamin J",
"role": [
{
"role": "author",
"vocabulary": "crossref"
}
],
"sequence": "first"
},
{
"affiliation": [
{
"name": "Division of Pulmonary Critical Care Medicine, University of North Carolina at Chapel Hill, Chapel Hill, NC, United States"
},
{
"name": "Marsico Lung Institute and Department of Microbiology and Immunology, University of North Carolina at Chapel Hill, Chapel Hill, NC, United States"
}
],
"family": "Hagan",
"given": "Robert S",
"role": [
{
"role": "author",
"vocabulary": "crossref"
}
],
"sequence": "additional"
},
{
"affiliation": [
{
"name": "SAS Institute Inc, Cary, NC, United States"
}
],
"family": "Jiang",
"given": "Xi",
"role": [
{
"role": "author",
"vocabulary": "crossref"
}
],
"sequence": "additional"
},
{
"affiliation": [
{
"name": "SAS Institute Inc, Cary, NC, United States"
}
],
"family": "Pavlechko",
"given": "Ella",
"role": [
{
"role": "author",
"vocabulary": "crossref"
}
],
"sequence": "additional"
},
{
"affiliation": [
{
"name": "SAS Institute Inc, Cary, NC, United States"
}
],
"family": "McClain",
"given": "Scott",
"role": [
{
"role": "author",
"vocabulary": "crossref"
}
],
"sequence": "additional"
},
{
"affiliation": [
{
"name": "SAS Institute Inc, Cary, NC, United States"
}
],
"family": "Hunt",
"given": "Xin",
"role": [
{
"role": "author",
"vocabulary": "crossref"
}
],
"sequence": "additional"
},
{
"affiliation": [
{
"name": "SAS Institute Inc, Cary, NC, United States"
}
],
"family": "Florou-Moreno",
"given": "Julia",
"role": [
{
"role": "author",
"vocabulary": "crossref"
}
],
"sequence": "additional"
},
{
"affiliation": [
{
"name": "SAS Institute Inc, Cary, NC, United States"
}
],
"family": "Acquadro",
"given": "Jake",
"role": [
{
"role": "author",
"vocabulary": "crossref"
}
],
"sequence": "additional"
},
{
"affiliation": [
{
"name": "SAS Institute Inc, Cary, NC, United States"
}
],
"family": "Risa",
"given": "Gabriel",
"role": [
{
"role": "author",
"vocabulary": "crossref"
}
],
"sequence": "additional"
},
{
"affiliation": [
{
"name": "SAS Institute Inc, Cary, NC, United States"
}
],
"family": "Valsaraj",
"given": "Varunraj",
"role": [
{
"role": "author",
"vocabulary": "crossref"
}
],
"sequence": "additional"
},
{
"affiliation": [
{
"name": "McAllister Heart Institute and the Department of Pharmacology, The University of North Carolina at Chapel Hill, Chapel Hill, NC, United States"
}
],
"family": "Schisler",
"given": "Jonathan C",
"role": [
{
"role": "author",
"vocabulary": "crossref"
}
],
"sequence": "additional"
},
{
"affiliation": [
{
"name": "Marsico Lung Institute and Department of Microbiology and Immunology, University of North Carolina at Chapel Hill, Chapel Hill, NC, United States"
}
],
"family": "Wolfgang",
"given": "Matthew C",
"role": [
{
"role": "author",
"vocabulary": "crossref"
}
],
"sequence": "additional"
}
],
"container-title": [],
"content-domain": {
"crossmark-restriction": false,
"domain": []
},
"created": {
"date-parts": [
[
2026,
6,
10
]
],
"date-time": "2026-06-10T16:40:12Z",
"timestamp": 1781109612000
},
"deposited": {
"date-parts": [
[
2026,
6,
13
]
],
"date-time": "2026-06-13T08:00:22Z",
"timestamp": 1781337622000
},
"funder": [
{
"award": [
"5T32HL007106-45"
],
"award-info": [
{
"award-number": [
"5T32HL007106-45"
]
}
],
"id": [
{
"asserted-by": "publisher",
"id": "https://ror.org/001d55x84",
"id-type": "ROR"
}
]
}
],
"group-title": "Intensive Care and Critical Care Medicine",
"indexed": {
"date-parts": [
[
2026,
6,
13
]
],
"date-time": "2026-06-13T08:55:27Z",
"timestamp": 1781340927637,
"version": "3.54.1"
},
"institution": [
{
"name": "medRxiv"
}
],
"is-referenced-by-count": 0,
"issued": {
"date-parts": [
[
2026,
6,
10
]
]
},
"license": [
{
"URL": "http://creativecommons.org/licenses/by-nc-nd/4.0/",
"content-version": "vor",
"delay-in-days": 0,
"start": {
"date-parts": [
[
2026,
6,
10
]
],
"date-time": "2026-06-10T00:00:00Z",
"timestamp": 1781049600000
}
}
],
"link": [
{
"URL": "https://syndication.highwire.org/content/doi/10.64898/2026.06.08.26355201",
"content-type": "unspecified",
"content-version": "vor",
"intended-application": "similarity-checking"
}
],
"member": "54368",
"original-title": [],
"posted": {
"date-parts": [
[
2026,
6,
10
]
]
},
"prefix": "10.64898",
"published": {
"date-parts": [
[
2026,
6,
10
]
]
},
"publisher": "openRxiv",
"reference": [
{
"DOI": "10.1056/NEJMoa2021436",
"doi-asserted-by": "publisher",
"key": "2026061301001190000_2026.06.08.26355201v1.1"
},
{
"DOI": "10.1001/jama.2020.17022",
"doi-asserted-by": "publisher",
"key": "2026061301001190000_2026.06.08.26355201v1.2"
},
{
"DOI": "10.1126/science.abc8511",
"doi-asserted-by": "publisher",
"key": "2026061301001190000_2026.06.08.26355201v1.3"
},
{
"DOI": "10.1164/rccm.202105-1302oc",
"doi-asserted-by": "publisher",
"key": "2026061301001190000_2026.06.08.26355201v1.4"
},
{
"DOI": "10.1016/j.chest.2021.04.062",
"article-title": "Identification of Distinct Clinical Subphenotypes in Critically Ill Patients With COVID-19",
"doi-asserted-by": "crossref",
"first-page": "929",
"journal-title": "Chest",
"key": "2026061301001190000_2026.06.08.26355201v1.5",
"volume": "160",
"year": "2021"
},
{
"DOI": "10.1038/s41467-021-25040-5",
"doi-asserted-by": "publisher",
"key": "2026061301001190000_2026.06.08.26355201v1.6"
},
{
"DOI": "10.1016/S2213-2600(20)30366-0",
"article-title": "Prevalence of phenotypes of acute respiratory distress syndrome in critically ill patients with COVID-19: a prospective observational study",
"doi-asserted-by": "crossref",
"first-page": "1209",
"journal-title": "The Lancet Respiratory Medicine",
"key": "2026061301001190000_2026.06.08.26355201v1.7",
"volume": "8",
"year": "2020"
},
{
"key": "2026061301001190000_2026.06.08.26355201v1.8",
"unstructured": "Ozonoff, A. et al. Phenotypes of disease severity in a cohort of hospitalized COVID-19 patients: Results from the IMPACC study. eBioMedicine 83, 104208 (2022)."
},
{
"DOI": "10.3389/fimmu.2022.989556",
"doi-asserted-by": "publisher",
"key": "2026061301001190000_2026.06.08.26355201v1.9"
},
{
"DOI": "10.1001/jama.2023.3651",
"doi-asserted-by": "publisher",
"key": "2026061301001190000_2026.06.08.26355201v1.10"
},
{
"DOI": "10.1038/s41467-020-20139-7",
"doi-asserted-by": "publisher",
"key": "2026061301001190000_2026.06.08.26355201v1.11"
},
{
"DOI": "10.1371/journal.pone.0295418",
"article-title": "Heterogeneity of COVID-19 symptoms and associated factors: Longitudinal analysis of laboratory-confirmed COVID-19 cases in San Antonio",
"doi-asserted-by": "crossref",
"first-page": "e0295418",
"journal-title": "PLoS ONE",
"key": "2026061301001190000_2026.06.08.26355201v1.12",
"volume": "18",
"year": "2023"
},
{
"DOI": "10.1164/rccm.202308-1490OC",
"doi-asserted-by": "publisher",
"key": "2026061301001190000_2026.06.08.26355201v1.13"
},
{
"DOI": "10.1007/s00134-022-06869-w",
"article-title": "Coronavirus disease 2019 subphenotypes and differential treatment response to convalescent plasma in critically ill adults: secondary analyses of a randomized clinical trial",
"doi-asserted-by": "crossref",
"first-page": "1525",
"journal-title": "Intensive Care Med",
"key": "2026061301001190000_2026.06.08.26355201v1.14",
"volume": "48",
"year": "2022"
},
{
"DOI": "10.1164/rccm.202401-0232ED",
"article-title": "Understanding Heterogeneity in Acute Care Trials: Resource Availability Impacts Outcomes",
"doi-asserted-by": "crossref",
"first-page": "469",
"journal-title": "Am J Respir Crit Care Med",
"key": "2026061301001190000_2026.06.08.26355201v1.15",
"volume": "209",
"year": "2024"
},
{
"DOI": "10.1136/bmj.k4245",
"doi-asserted-by": "publisher",
"key": "2026061301001190000_2026.06.08.26355201v1.16"
},
{
"DOI": "10.1056/EVIDe2400054",
"doi-asserted-by": "crossref",
"key": "2026061301001190000_2026.06.08.26355201v1.17",
"unstructured": "Yarnell, C. J. & Fralick, M. Heterogeneity of Treatment Effect — An Evolution in Subgroup Analysis. NEJM Evidence 3, (2024)."
},
{
"DOI": "10.1164/rccm.202311-2011ST",
"doi-asserted-by": "publisher",
"key": "2026061301001190000_2026.06.08.26355201v1.18"
},
{
"DOI": "10.1016/S2213-2600(19)30417-5",
"article-title": "Dexamethasone treatment for the acute respiratory distress syndrome: a multicentre, randomised controlled trial",
"doi-asserted-by": "crossref",
"first-page": "267",
"journal-title": "The Lancet Respiratory Medicine",
"key": "2026061301001190000_2026.06.08.26355201v1.19",
"volume": "8",
"year": "2020"
},
{
"DOI": "10.1007/s00134-014-3287-7",
"doi-asserted-by": "publisher",
"key": "2026061301001190000_2026.06.08.26355201v1.20"
},
{
"DOI": "10.1016/j.amjmed.2020.03.012",
"article-title": "Hyperglycemia in Medically Critically Ill Patients: Risk Factors and Clinical Outcomes",
"doi-asserted-by": "crossref",
"first-page": "e568",
"journal-title": "The American Journal of Medicine",
"key": "2026061301001190000_2026.06.08.26355201v1.21",
"volume": "133",
"year": "2020"
},
{
"key": "2026061301001190000_2026.06.08.26355201v1.22",
"unstructured": "The Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) Statement: Guidelines for Reporting Observational Studies. 18, (2024)."
},
{
"DOI": "10.1503/cmaj.200077",
"doi-asserted-by": "publisher",
"key": "2026061301001190000_2026.06.08.26355201v1.23"
},
{
"key": "2026061301001190000_2026.06.08.26355201v1.24",
"unstructured": "Hagan, R. et al. Multiple omic analysis from the lungs explains heterogeneity of ARDS with molecular pathways that suggest differential response to pharmacotherapy. manuscript in preparation 40 pages."
},
{
"key": "2026061301001190000_2026.06.08.26355201v1.25",
"unstructured": "Sines, B. Clinician-Informed Feature Engineering Improves Machine Learning Assignment of Molecular Endotypes in the Intensive Care Unit."
}
],
"reference-count": 25,
"references-count": 25,
"relation": {},
"resource": {
"primary": {
"URL": "http://medrxiv.org/lookup/doi/10.64898/2026.06.08.26355201"
}
},
"score": 1,
"short-title": [],
"source": "Crossref",
"subject": [],
"subtitle": [],
"subtype": "preprint",
"title": "A Clinical Predictor of Lung Molecular Endotype Identifies Heterogeneity in Corticosteroid Response in Severe COVID-19: an Emulated Target Trial",
"type": "posted-content"
}
Late treatment
is less effective
is less effective
