Early dynamic changes in inflammatory indices and their association with mortality in intermediate–high and high-risk pulmonary embolism patients admitted to the ICU: a retrospective observational cohort study

Acute pulmonary embolism (PE) is associated with high mortality, particularly in patients requiring intensive care unit (ICU) management. This study aimed to evaluate early dynamic changes in systemic inflammatory indices and their association with in-hospital mortality in intermediate–high- and high-risk patients with PE.

In this retrospective observational ICU cohort study, 172 adult patients with confirmed intermediate–high- or high-risk PE were included. Patients were managed with either reperfusion therapy or low-molecular-weight heparin (LMWH). Systemic immune–inflammation index (SII), systemic inflammation response index (SIRI), and aggregate index of systemic inflammation (AISI) were calculated at baseline (day 0) and on days 1–3. Percentage changes from baseline were assessed. The primary endpoint was to compare early inflammatory trajectories between treatment groups. The secondary endpoint was in-hospital mortality.

In both treatment groups, percentage changes in inflammatory indices were calculated relative to baseline. A greater reduction in %Delta SII on day 3 was observed in patients receiving reperfusion therapy compared with those treated with LMWH. In the linear mixed-effects analysis, a significant group-by-time interaction was detected for SII (p = 0.007), whereas no significant interaction was found for SIRI or AISI. In-hospital mortality occurred in 22.1% of patients. Non-survivors had consistently higher baseline and follow-up inflammatory indices. In multivariable analysis, baseline SIRI independently predicted in-hospital mortality (OR 1.147, 95% CI 1.039–1.267, p = 0.007). Baseline SII, SIRI, and AISI demonstrated moderate discriminative ability for mortality prediction (AUC range: 0.717–0.743).

In intermediate–high and high-risk PE patients managed in the ICU, early inflammatory trajectories differed between treatment groups, with a significant group-by-time interaction observed only for SII in longitudinal analysis. Baseline inflammatory burden, particularly SIRI, was independently associated with in-hospital mortality. Early inflammatory patterns may provide supportive prognostic information; however, these findings should be considered exploratory and require validation in prospective studies.