Admission antithrombin III activity as a risk-enrichment marker for septic shock below the ISTH overt-DIC threshold

Background International Society on Thrombosis and Haemostasis (ISTH) overt disseminated intravascular coagulation (DIC) scoring is a standard bedside framework for infection-associated coagulopathy but does not capture anticoagulant-pathway markers. The study examined whether admission antithrombin III (AT-III) activity provided risk-enrichment information for septic shock among adults admitted with infection below the ISTH 2001 overt-DIC threshold. Methods The study retrospectively analyzed 1,286 adults admitted with infection. AT-III activity was evaluated in the ISTH-complete and below-threshold subsets and compared with ISTH 2001 and 2025 SSC overt-DIC scores. Retrospective temporal adjudication compared first valid AT-III report availability with operational septic-shock onset. In a clinically indicated lactate-measured subset, the study assessed whether AT-III added discrimination beyond age, sex, and lactate. Firth-penalized regression evaluated the conditional association of AT-III with septic shock beyond age, sex, and raw ISTH components. Results Among 1,027 ISTH-complete patients, 974 (94.8%) were below the ISTH 2001 overt-DIC threshold, including 44 of 60 septic-shock cases. In this subgroup, 43 of 44 shock cases had AT-III activity <80%; the area under the curve (AUC) was 0.844 and the positive predictive value was 9.37% (43/459). Of 82 septic-shock admissions, 74 were temporally adjudicable: 62 were incident and 12 were prevalent or concurrent at AT-III report availability. In incident cases, AT-III report availability preceded operational shock onset by a median of 7.6 h (interquartile range (IQR) 5.0–9.6); sampling therefore also preceded onset. In the lactate-measured subset ( n = 518), AT-III increased AUC beyond age, sex, and lactate from 0.740 to 0.821; this increment persisted in the incident-shock analysis (0.680–0.792). In the primary Firth model, lower AT-III activity remained conditionally associated with shock (OR 3.10 per 1-SD decrease); the apparent and bootstrap-corrected AUCs were 0.900 and 0.888. Conclusion Admission AT-III activity was associated with septic shock within the large ISTH below-threshold infection subgroup and provided high-sensitivity, low-positive-predictive-value enrichment information. Retrospectively reconstructed temporal ordering demonstrated a report-release-to-operational-onset interval in incident cases but does not establish prospective prediction, prospective workflow performance, or clinical benefit. AT-III should not be used as a standalone trigger for escalation decisions. Prospective, time-stamped validation is required.