TELOS-plus identifies normotensive patients with pulmonary embolism at high risk (>10%) of 30-day mortality

Background/Introduction Identifying normotensive patients with acute pulmonary embolism (PE) at high risk for fatal outcomes remains challenging. Existing risk models fail to detect a sufficiently large cohort with a 30-day PE-related mortality threshold sufficiently high (>10%) to justify acute reperfusion treatments (systemic thrombolysis or catheter directed thrombectomy). Recent advancements in PE prognostication suggest that markers of tissue oxygenation (lactate) and respiratory failure (RF), alongside a refined definition of right ventricular dysfunction (RVD), may improve risk stratification. Purpose To investigate whether including an easy-to use respiratory failure marker and a refined definition of RVD improve the performance of the TELOS score comparing to other scoring systems (Table 1). Methods We analyzed retrospective data of normotensive patients (2008-2019) from our regional registry of PE. PE diagnosis was confirmed with CTPA. The TELOS-plus score was developed using logistic regression to identify predictors of 30-day PE-related mortality. The model included severe RF, defined as O₂saturation <85% at room air or the need of high flow nasal cannula (HFNC) or non-invasive ventilation (NIV), and revised RVD criteria requiring at least two of the following: RV-LV ratio >1, transtricuspidal gradient >30 mmHg, RV-free wall hypokinesis or paradoxical septal motion. Additional TELOS variables included elevated lactate (>2 mmol/L) and troponin I (Table 1). Model performance was assessed via ROC analysis, and patient risk stratification was compared across models. Results Among 1,091 patients (median age 74, 54.6% female), the 30-day PE-related mortality was 3.8%. Patients with severe RF (12.8%) had a 15.7% mortality rate (OR 8.7, 95% CI 4.6–16.4). The revised RVD criteria identified 17.8% of patients, with a mortality rate of 10.3% (95% CI 6.7–15.3), compared to 4.6% (95% CI 2.6–8%) in those meeting only one RVD criterion (OR 2.4, 95% CI 1.2-4.8, p<0.05). Logistic regression confirmed revised RVD and severe RF as independent predictors of PE-related mortality. The TELOS-plus score (AUC 0.86) outperformed other models except SHIeLD (Figure 1). A TELOS-plus score >2 identified 17.1% (95% CI 15-19.5%) of patients at intermediate-high risk with 16.1% (CI95% 11-22.1%) mortality rate (Table 1). Conclusions The TELOS-plus score, incorporating severe RF and a more stringent RVD definition, identifies about 17% of PE patients at intermediate-high risk of 30-day mortality. It demonstrated superior performance compared to traditional cardiac-focused models and offers a clinically feasible alternative to SHIeLD. TELOS-plus provides an effective tool for improved risk stratification in PE and warrants further validation for broader clinical use.