Prognostic peripheral blood biomarkers at ICU admission predict COVID-19 clinical outcomes

Melina Messing, Mypinder S. Sekhon, Michael R. Hughes, Sophie Stukas, Ryan L. Hoiland, Jennifer Cooper, Nyra Ahmed, Mark Stephen Hamer, Yicong Li, Samuel B. Shin, Lin Tung, Cheryl L. Wellington, Don D. Sin, Kevin B. Leslie, Kelly M. McNagny

This study, published in Frontiers in Immunology, presents a high-precision prognostic tool for COVID-19. By combining serum cytokine levels with deep "immunophenotyping" of rare immune cell subsets, researchers developed a two-step screen that predicts whether an ICU patient will recover quickly or face a prolonged stay/death with 91% accuracy.

TL;DR

Researchers identified a specific "signature" at ICU admission—elevated IL-10 and a rare CD11c^low monocyte subset—that predicts severe COVID-19 outcomes with 91% sensitivity and specificity. This tiered biomarker approach outperforms standard clinical metrics and could revolutionize how hospitals triage intensive care resources.

What Problem Does This Paper Address?

While we have become better at treating COVID-19, hospital ICUs still struggle with "prognostic uncertainty." When a patient is admitted, clinicians often cannot tell who will recover in a few days versus who will require weeks of mechanical ventilation or succumb to the disease.

Standard laboratory markers like C-reactive protein (CRP) or D-dimer are useful for identifying general inflammation but lack the "resolution" to predict specific clinical trajectories. This paper seeks to find "high-resolution" biological signals that can be measured at the moment of admission to guide resource allocation and personalized treatment.

Methods

The researchers employed a multi-omic approach on 90 ICU patients:

  • Cytokine Profiling: They used Simoa (Single Molecule Array) technology to measure ultra-low levels of four key cytokines (IL-1β, IL-6, IL-10, and TNFα).
  • Deep Immunophenotyping (CyTOF): On a subset of 42 patients, they used Mass Cytometry (CyTOF) with a 38-antibody panel. Unlike standard flow cytometry, CyTOF uses metal-tagged antibodies and mass spectrometry to look at dozens of cellular markers simultaneously, allowing for the discovery of very specific cell subpopulations.
  • Outcome Binning: They defined a new clinical metric: "Short-Stay" (SS, <6 days in ICU and survived) vs. "Long-Stay/Died" (LS/D, ≥6 days or death).
  • Validation: Findings from an "Initial Cohort" were validated against a "Replication Cohort" to ensure the biomarkers weren't just statistical noise.

Datasets

  • Cytokine Cohort: 90 COVID-19 patients admitted to the ICU at St. Paul’s Hospital and Vancouver General Hospital.
  • Initial & Replication Cohorts: Sub-groups of 14 and 28 patients, respectively, used for the intensive CyTOF analysis.
  • Healthy Controls: 8 age-matched individuals used to establish baseline immune profiles.

Key Insights and Ideas

  • Broad Markers Fail: The study found that broad immune shifts—like general lymphopenia (low T-cell counts) or high total monocyte counts—occurred in all ICU patients. Because these shifts are universal in severe COVID, they cannot be used to distinguish between "bad" and "worse" outcomes.
  • The "Aha" Subset: The real predictive power lay in a very specific, rare population: CD11c^low classical monocytes. These cells were significantly enriched in patients who eventually died or had long ICU stays.
  • IL-10 as a Pre-Screen: While IL-6 is often the "famous" cytokine in COVID-19, the researchers found IL-10 was a superior predictor of stay length. However, IL-10 alone wasn't enough; it required the cellular data to reach high diagnostic certainty.
  • Tiered Triage: The authors propose a "stepwise" diagnostic: first, check if IL-10 is ≥15 pg/ml. If it is, then check the levels of CD11c^low monocytes. This combined "filter" provides a massive boost in predictive power (Positive Likelihood Ratio of 10.1).

Key Results

  • High Accuracy: The combined IL-10 and monocyte screen achieved 91% sensitivity and 91% specificity in predicting LS/D outcomes.
  • Thresholds: The critical "danger zone" thresholds were identified as IL-10 ≥15 pg/ml and CD11c^low classical monocytes ≥2.7 x 10⁷/L.
  • Independence from Standard Care: Interestingly, these biomarkers were predictive regardless of whether patients received standard treatments like corticosteroids or tocilizumab (though IL-6 was excluded from the final model specifically because tocilizumab interferes with its measurement).

Limitations and Future Work

  • Cohort Size: While the results are statistically strong, the CyTOF analysis was limited to 42 patients. Larger, multi-center studies are needed to confirm these thresholds across different populations and viral variants.
  • Sample Processing: CyTOF was performed on frozen/thawed PBMCs. The authors acknowledge that freezing can impact certain cell surface markers, and fresh blood analysis might yield even more precise results.
  • Specificity to COVID-19: The study lacked a "non-COVID ICU control" group. It remains to be seen if this CD11c^low monocyte signature is unique to COVID-19 or a general feature of viral sepsis.
  • Translation: Moving from CyTOF (a complex research tool) to a simplified flow cytometry panel that can be run in a standard hospital lab is the next logical step for clinical adoption.

Summary updated

Abstract

The COVID-19 pandemic continues to challenge the capacities of hospital ICUs which currently lack the ability to identify prospectively those patients who may require extended management. In this study of 90 ICU COVID-19 patients, we evaluated serum levels of four cytokines (IL-1β, IL-6, IL-10 and TNFα) as well as standard clinical and laboratory measurements. On 42 of these patients (binned into Initial and Replication Cohorts), we further performed CyTOF-based deep immunophenotyping of peripheral blood mononuclear cells with a panel of 38 antibodies. All measurements and patient samples were taken at time of ICU admission and retrospectively linked to patient clinical outcomes through statistical approaches. These analyses resulted in the definition of a new measure of patient clinical outcome: patients who will recover after short ICU stays (< 6 days) and those who will subsequently die or recover after long ICU stays (≥6 days). Based on these clinical outcome categories, we identified blood prognostic biomarkers that, at time of ICU admission, prospectively distinguish, with 91% sensitivity and 91% specificity (positive likelihood ratio 10.1), patients in the two clinical outcome groups. This is achieved through a tiered evaluation of serum IL-10 and targeted immunophenotyping of monocyte subsets, specifically, CD11c<sup>low</sup> classical monocytes. Both immune biomarkers were consistently elevated ( ≥15 pg/ml and ≥2.7 x10<sup>7</sup>/L for serum IL-10 and CD11c<sup>low</sup> classical monocytes, respectively) in those patients who will subsequently die or recover after long ICU stays. This highly sensitive and specific prognostic test could prove useful in guiding clinical resource allocation.
BibTeX
@inproceedings{messing2022prognostic,
  author = {Melina Messing and Mypinder S. Sekhon and Michael R. Hughes and Sophie Stukas and Ryan L. Hoiland and Jennifer Cooper and Nyra Ahmed and Mark Stephen Hamer and Yicong Li and Samuel B. Shin and Lin Tung and Cheryl L. Wellington and Don D. Sin and Kevin B. Leslie and Kelly M. McNagny},
  title = {{Prognostic peripheral blood biomarkers at ICU admission predict COVID-19 clinical outcomes}},
  booktitle = {Frontiers in Immunology},
  year = {2022},
  doi = {10.3389/fimmu.2022.1010216},
}
Figure