Coagulopathy subphenotypes and mortality in critically ill patients with sepsis: a multicentre retrospective cohort study using MIMIC-IV and SICdb
This multicentre retrospective study utilizing consensus clustering on MIMIC-IV and SICdb data identifies three reproducible sepsis-induced coagulopathy subphenotypes, revealing that while pan-coagulopathy consistently predicts higher mortality, isolated aPTT prolongation serves as a high-risk marker specifically in surgical ICU and SIC-positive settings, thereby supporting phenotype-informed prognostication and future trial design.
Original paper licensed under CC BY 4.0 (https://creativecommons.org/licenses/by/4.0/). This is an AI-generated explanation of the paper below. It is not written or endorsed by the authors. For technical accuracy, refer to the original paper. Read full disclaimer
Sepsis is a life-threatening reaction to an infection, where the body's own defense system turns against itself, damaging organs and tissues. One of the most dangerous ways this happens is through a breakdown in the blood's ability to clot properly. Normally, clotting is a precise mechanism that stops bleeding when a vessel is cut. In sepsis, this system can go haywire, causing blood to clot too easily inside the body or, conversely, to fail to clot at all, leading to dangerous bleeding. This chaotic state, known as sepsis-induced coagulopathy, affects more than half of critically ill patients and is a major reason why so many people die from the condition. For decades, doctors have tried to treat this by giving blood-thinning medications, but large clinical trials have mostly failed. The prevailing theory is that these treatments failed because they were given to everyone with sepsis, treating a complex, messy problem as if it were a single, simple disease.
A team of researchers set out to solve this puzzle by looking at the blood of thousands of sepsis patients to see if there were distinct subgroups, or "types," of blood clotting failure that had been overlooked. They gathered data from two different intensive care units: one in Boston, USA, and another in Salzburg, Austria. They focused on four standard blood tests that are routinely measured in hospitals: the time it takes for blood to clot via two different pathways, the level of a specific clotting protein, and the number of platelets, which are tiny cells that help form clots. Instead of looking at these numbers in isolation, they used a powerful computer method called consensus clustering. This technique repeatedly shuffles and re-analyzes the data to find stable patterns that are not just random noise, much like how a cartographer might draw a map by having many different surveyors trace the same coastline to find the true shape of the land.
The analysis revealed that sepsis patients do not all have the same kind of blood clotting trouble. Instead, they fall into three distinct groups. The largest group, making up about 57 percent of patients, had blood that looked relatively normal despite their severe illness; their clotting times and protein levels were preserved. A second, smaller group, about 10 percent, had a very specific problem: only one of their clotting pathways was significantly slowed down, while the rest of their blood clotting system remained mostly intact. The third group, comprising about 33 percent of patients, suffered from a "pan-coagulopathy," meaning their entire clotting system was failing. They had high levels of clotting time, low platelet counts, and low levels of clotting protein all at once.
When the researchers tracked who survived and who did not, the differences between these groups became clear. The group with the total system failure, the pan-coagulopathy, faced a significantly higher risk of death within 28 days compared to the group with normal blood, even after accounting for how sick they were in other ways. This suggests that the total collapse of the clotting system is a specific, deadly feature of sepsis that goes beyond just being generally very ill. The group with the single, isolated problem in their clotting pathway told a more complicated story. In the American hospital data, this specific blood pattern did not seem to increase the risk of death. However, in the Austrian surgical hospital data, this same pattern was linked to a risk of death that was twice as high as the group with normal blood.
This difference highlights a crucial discovery: the meaning of a blood test result can depend entirely on the context of the patient. The researchers found that the isolated clotting problem was only dangerous in the surgical setting and specifically among patients who already met the criteria for a specific sepsis-related clotting score. In other words, a doctor looking at a patient's blood might see a warning sign that is only relevant if the patient is also showing other specific signs of organ failure. The study also showed that these three groups could be identified using only the four standard blood tests available in almost every hospital, without needing expensive or specialized equipment.
The findings suggest that the reason past treatments for sepsis-related clotting have failed might be that doctors have been treating all patients the same way, rather than targeting the specific type of clotting failure they have. The group with the total system failure appears to be a clear candidate for future trials of blood-thinning drugs, as they consistently face higher mortality. The group with the isolated clotting problem, however, requires a more careful approach, as their risk depends heavily on where they are being treated and what other symptoms they have. By using these simple blood tests to sort patients into these three categories, doctors could potentially move away from a one-size-fits-all approach and begin to tailor treatments to the specific biological reality of each patient's illness.
Drowning in papers in your field?
Get daily digests of the most novel papers matching your research keywords — with technical summaries, in your language.