Weaker clot firmness on ROTEM tracked neonatal sepsis deaths, but barely moved the model

In a prospective single-center cohort of 168 term neonates meeting sepsis criteria, 23 died within 28 days. FIBTEM clot firmness was lower in non-survivors, but adding it to nSOFA plus INR raised model AUC only from 0.862 to 0.878; the 0.016 gain was not statistically significant (DeLong P=.455). ROTEM described a hypocoagulable severity phenotype, not a validated stand-alone mortality tool.

The hot take

A richer clot trace can describe physiology without earning a place as a mortality score.

In a prospective cohort of 168 term neonates meeting sepsis criteria, non-survivors had slower, weaker clot formation across conventional tests and rotational thromboelastometry (ROTEM). The physiological pattern was coherent. The prediction result was less persuasive: adding FIBTEM maximum clot firmness to nSOFA plus INR increased AUC from 0.862 to 0.878, an absolute gain of 0.016 that was not statistically significant (DeLong P=.455).

What the cohort actually was

The investigators enrolled term neonates of 37 to 42 completed weeks at Vietnam National Children's Hospital, a single national tertiary referral center in Hanoi, from June 2024 through December 2025. Eligibility required at least two clinical signs and two laboratory abnormalities under the 2010 European Medicines Agency neonatal sepsis criteria. Major congenital or chromosomal anomalies, severe perinatal asphyxia, recent plasma or platelet exposure, and death within 24 hours that prevented standardized coagulation assessment were among the exclusions.