Burn Sepsis Biomarker Trajectories and Patient Phenotypes from a 712-Patient ICU Cohort

A retrospective cohort study of 712 adult patients with burn sepsis admitted to a regional burn ICU between 2022 and 2025. The dataset contains longitudinal measurements of 17 biomarkers across nutrition, immunity, and inflammation at five time points (days 1, 3, 7, 14, 21). It was authored by Pei Xu and shared under a CC-BY-4.0 license.

Use Cases

• Predicting 21-day mortality in burn sepsis patients based on baseline biomarker levels like albumin and interleukin-6.
• Identifying distinct patient phenotypes using growth mixture modeling on integrated trajectories of albumin, IgG, and IL-6.
• Assessing the dynamic prognostic value of multicategory biomarker trajectories compared to static measurements.
• Studying associations between nutritional, immune, and inflammatory markers and clinical outcomes in a burn ICU setting.

Strengths

• Includes 712 patient records, providing a substantial cohort for analysis.
• Longitudinal data with 17 biomarkers measured at five specific time points over 21 days.
• Clear clinical outcome defined with a 21-day mortality rate of 17.9% (81 deaths).
• Phenotyping results are provided, identifying two distinct groups with mortality rates of 15.7% and 8.8%.

Limitations

• Column-level documentation is absent; field semantics must be inferred after download.
• Row count for the underlying data table is unknown, which may limit suitability assessment.
• The dataset is very small (19.6 KB), suggesting it may contain summary or processed results rather than raw patient-level data.

Provenance

Source

figshare, author Pei Xu.

Collection Method

Retrospective cohort study from a regional burn ICU.

Time Range

2022–2025.

Freshness

Last updated 2026-04-16 05:14:51.

Geography

null