824 - Prediction of serious complications in patients with cancer and pulmonary embolism: validation of the EPIPHANY index in a prospective cohort of patients from the PERSEO study.

Background

The EPIPHANY decision tree is the first algorithm developed to predict serious complications in patients with cancer and pulmonary embolism (PE) (PMID: 28267709).

Methods

The objective is to evaluate the discriminatory ability of EPIPHANY in a prospective multicenter cohort. Patients with PE diagnosed by objective methods were recruited from October 2017 to April 2019. The association between the increase in prognostic category and in complications at 15 days was assessed using the linear by linear association test.

Results

The sample contains 463 patients with PE (39.7% suspected and 60.3% unsuspected). 68.3% (n = 316) showed clinical or haemodynamic instability, while 31.7% (n = 147) were normotensive PE with apparent clinical stability. The breakdown of initial risk criteria is: sudden or progressive dyspnea (58.7%), tachycardia >110 lpm (17.3%), hypotension <100 mmHg (9.9%), hypoxemia <90% (11%), elevated haemorrhagic risk (18.6%), tachypnea >30 rpm (3.2%), thrombopenia < 50000 (2.2%) and major bleeding (0.9%). Based on the entire decision tree, 21.8% were classified as low (n = 101), 22% as intermediate (n = 102), and 56.15% as high risk (n = 260). The serious complications rate at 15 days increased significantly throughout these prognostic categories: 3%, 5.9% and 26.5%, for low, intermediate and high-risk patients, respectively (p = 0.001). Serious complications occurred in 16.8% (n = 78) after a median of 5 days, being the most frequent: respiratory failure (56.4%), major bleeding (21.8%) and hypotension (20.5%). 28 patients died within the first 15 days post-PE, of which 2 belonged to the medium- and 26 to the high-risk group. The 3 main etiologies of exitus were: mixed origin (57.1%), complications of PE (32.1%) and progression of cancer (28.1%).

Conclusions

EPIPHANY is the only model available for the classification of patients with cancer and PE, based on their short-term risk of complications, with potential implications for decision making.

Clinical trial identification

Editorial acknowledgement

Legal entity responsible for the study

Asociación de Investigación de la Enfermedad Tromboembólica Venosa de la Región de Murcia.

Funding

Leo Academy.

Disclosure

M. Sanchez Canovas: Research grant / Funding (institution): Leo Pharma. All other authors have declared no conflicts of interest.