A new study shows that watching how key markers move over days—not just a single snapshot—can reveal who is most at risk in the ICU.
Burn sepsis is a life-threatening infection that happens after a severe burn. It is one of the biggest challenges in burn ICUs. The body’s response is complex. It involves inflammation, immune suppression, and poor nutrition.
Doctors currently rely on single lab values to judge risk. But a single snapshot can miss the bigger picture. Two patients with the same lab result on day one can have very different outcomes.
This study looked at how 17 different markers move over 21 days. The goal was to find patterns that better predict who will survive.
The Old Way vs. The New Way
In the past, doctors checked a lab value and made a judgment. If albumin was low, they might worry. If inflammation markers were high, they might adjust treatment.
But this approach is static. It does not capture the body’s dynamic response to infection and healing.
This study flips that script. Instead of a single number, it tracks the trajectory—the path each marker takes over time. The researchers found that the shape of these paths tells a more accurate story.
Think of the body like a city during a crisis. Nutrition markers are like food supplies. Immune cells are the police force. Inflammation markers are the emergency sirens.
In a healthy response, supplies stabilize, the police regain control, and sirens quiet down. In a failing response, supplies run low, the police are overwhelmed, and sirens blare nonstop.
The study used a method called growth mixture modeling (GMM). This is like a smart camera that groups patients based on the shape of their city’s crisis response over time. It finds clusters of people who follow similar paths.
Researchers reviewed data from 712 adult burn sepsis patients admitted to a regional burn ICU between 2022 and 2025. They measured 17 biomarkers at five time points: days 1, 3, 7, 14, and 21. These markers covered nutrition (like albumin), immunity (like T-cells and immunoglobulins), and inflammation (like IL-6 and CRP).
The overall 21-day death rate was 17.9%. That means 81 patients died, and 631 survived.
First, the team compared the paths of survivors versus non-survivors. The paths were clearly different for many markers. This means the body’s response over time is tied to outcome.
Next, they used the GMM to find patient groups. They identified two main patterns:
1. High-risk phenotype: These patients had persistently low albumin and immunoglobulin G (IgG), and high IL-6 inflammation over the full 21 days. About 15.7% of this group died. 2. Low-risk phenotype: These patients had better trajectories. Their nutrition and immune markers improved, and inflammation fell. Only 8.8% of this group died.
The difference in death rates between the two groups was significant.
The study also checked if a single baseline marker could predict death. Some did, like albumin and IL-6. But when the team looked at all markers together, no single one stood out as independently powerful. This is likely because the markers are closely linked—they influence each other.
This doesn’t mean this treatment is available yet.
This research highlights a key shift in critical care: moving from static snapshots to dynamic profiles. The identified phenotypes reflect the real-world struggle of the body—fighting infection, trying to heal, and managing resources. While single markers have limits, the integrated pattern offers a clearer window into a patient’s true risk.
If you or a loved one is in the burn ICU with sepsis, this research is not a test you can ask for today. It is a step toward more personalized care. In the future, tracking these patterns could help doctors tailor treatments earlier. For now, it reinforces the importance of comprehensive care that addresses nutrition, infection control, and immune support.
This was a retrospective study, meaning it looked back at past data. It was done at one regional burn center, so the results may not apply to all hospitals. The sample size was decent but not huge. Also, the predictive performance of single markers was modest, showing this method is better used as a pattern tool, not a standalone test.
The next step is to validate these findings in new groups of patients, ideally in multiple hospitals. Researchers will also need to test if using these patterns to guide treatment actually improves survival. This could take several years of clinical trials before it becomes standard care.