's Avatar

@sinha-lab.bsky.social

#science #sepsis #ALI #precision #datascience

8 Followers  |  23 Following  |  4 Posts  |  Joined: 16.11.2024  |  1.5056

Latest posts by sinha-lab.bsky.social on Bluesky

We identified two primary clusters — “Dysregulated” vs. “Undifferentiated” — with markedly different clinical outcomes and responses to antibiotics delay. Within the Dysregulated group, four distinct sub-clusters emerged, each with unique biological signatures and outcomes: · Cluster 3: Associated with renal dysfunction. · Cluster 4: Enriched for viral infections. · Cluster 5: Characterized by an immunosuppressed signature and adverse outcomes, often with bacterial culture positivity. · Cluster 6: Marked by gram-negative infections, high IL-6/IL-8, and paradoxically better outcomes despite greater vasopressor use. In one of the largest studies of its kind, these findings underscore the heterogeneity of the host response in acute infection early during hospitalisation, the role of the pathogen in driving host response, and demonstrate how data-driven phenotyping can reveal biologically and clinically meaningful subgroups. This framework can inform precision diagnostics, targeted antimicrobial strategies, and enrichment designs for clinical trials. A big thank you to my collaborators Matt Churpek, Alex Spicer, Siva Bhavani, and co-authors for making this possible, including Prenosis Inc. team, Bobby Reddy Jr, Greg Watson, and Carlos Lopez-Espina among others.

We identified two primary clusters — “Dysregulated” vs. “Undifferentiated” — with markedly different clinical outcomes and responses to antibiotics delay. Within the Dysregulated group, four distinct sub-clusters emerged, each with unique biological signatures and outcomes: · Cluster 3: Associated with renal dysfunction. · Cluster 4: Enriched for viral infections. · Cluster 5: Characterized by an immunosuppressed signature and adverse outcomes, often with bacterial culture positivity. · Cluster 6: Marked by gram-negative infections, high IL-6/IL-8, and paradoxically better outcomes despite greater vasopressor use. In one of the largest studies of its kind, these findings underscore the heterogeneity of the host response in acute infection early during hospitalisation, the role of the pathogen in driving host response, and demonstrate how data-driven phenotyping can reveal biologically and clinically meaningful subgroups. This framework can inform precision diagnostics, targeted antimicrobial strategies, and enrichment designs for clinical trials. A big thank you to my collaborators Matt Churpek, Alex Spicer, Siva Bhavani, and co-authors for making this possible, including Prenosis Inc. team, Bobby Reddy Jr, Greg Watson, and Carlos Lopez-Espina among others.

Our latest study is now out in Nature Communications! In a cohort of 3,802 emergency department patients with suspected infection, we applied unsupervised clustering of 29 plasma proteins to map the multivariate host-response landscape.

A big thank you to my collaborators for making this possible!

08.09.2025 20:10 — 👍 3    🔁 0    💬 1    📌 0
Dynamics of Subphenotypes in Critical Illness: When the Tick-Tock of the Clock Counts | American Journal of Respiratory and Critical Care Medicine

Dynamics of Subphenotypes in Critical Illness: When the Tick-Tock of the Clock Counts
@atscommunity.bsky.social #medsky

🔓 Open Access

🔗 tinyurl.com/2ra5jwbv

07.03.2025 19:16 — 👍 4    🔁 1    💬 0    📌 0
(A–C) Alluvial plots showing the proportion of patients in each phenotype and transition between phenotypes as classified using a parsimonious biomarker-based model on Days 0, 2, 3, and 4 and mortality status at Day 90. The plots are subdivided on the basis of cohort: (A) MARS, (B) ALVEOLI, and (C) CLOVERS. ALVEOLI = Assessment of Low Tidal Volume and Elevated End-Expiratory Pressure to Obviate Lung Injury; CLOVERS = Crystalloid Liberal or Vasopressors Early Resuscitation in Sepsis; MARS = Molecular Diagnosis and Risk Stratification of Sepsis.

(A–C) Alluvial plots showing the proportion of patients in each phenotype and transition between phenotypes as classified using a parsimonious biomarker-based model on Days 0, 2, 3, and 4 and mortality status at Day 90. The plots are subdivided on the basis of cohort: (A) MARS, (B) ALVEOLI, and (C) CLOVERS. ALVEOLI = Assessment of Low Tidal Volume and Elevated End-Expiratory Pressure to Obviate Lung Injury; CLOVERS = Crystalloid Liberal or Vasopressors Early Resuscitation in Sepsis; MARS = Molecular Diagnosis and Risk Stratification of Sepsis.

Recently published: www.atsjournals.org/doi/abs/10.1...

@atsblueeditor.bsky.social

Learnings:
-HyperI phenotype is dynamic
-Speed of transition matters prognostically
-Information from changing phenotypes is substantially additive to that extracted from resolution of organ failure.

07.03.2025 20:38 — 👍 0    🔁 0    💬 0    📌 0

Eerily good synopsis, I may be out of work soon.

04.03.2025 23:09 — 👍 0    🔁 0    💬 0    📌 0
Blue Briefing: Impact of Inflammatory Phenotypes on the Clinical Outcomes of Critically Ill Patients
YouTube video by American Thoracic Society Blue Briefing: Impact of Inflammatory Phenotypes on the Clinical Outcomes of Critically Ill Patients

This study finds that transitioning from a hyperinflammatory to hypoinflammatory phenotype can reduce the risk of mortality and improve the outcomes of critically ill patients.
@atscommunity.bsky.social #medsky

Read more here: tinyurl.com/2mw83zvy

Watch here: youtu.be/e_lhIPMaOZs

03.03.2025 21:24 — 👍 4    🔁 4    💬 1    📌 0

Sinha Lab is on Bluesky. Come find us #criticalcare #sepsis #ards #precisionmedicine #datascience #evolution

19.11.2024 17:33 — 👍 0    🔁 0    💬 0    📌 0

@sinha-lab is following 20 prominent accounts