Diagnosis of pulmonary embolism in patients with haemoptysis: the POPEIHE study.

Abstract

To determine the prevalence and characteristics of pulmonary embolism (PE) in patients presenting with haemoptysis. Additionally, we assessed the efficiency and failure rates of different clinical diagnostic algorithms for PE in this patient population. We enrolled consecutive adult patients who presented to nine Italian emergency departments with haemoptysis as the primary complaint. PE diagnosis was ruled out in patients with a low pre-test probability in combination with a negative age-adjusted D-dimer (referred to as the "age-adjusted" D-dimer strategy), a negative computed tomography pulmonary angiography or when a clear alternative source of bleeding was identified, along with negative findings for venous thromboembolism during a 30-day follow-up. A total of 546 patients were included in the study. The prevalence of PE, including the 30-day follow-up, was 4.2% (95% CI 2.7-6.3%). The majority of these cases (78%) exhibited distal (segmental or subsegmental) emboli and there were no PE-related fatalities. The "age-adjusted" D-dimer strategy initially excluded PE in 24% of patients (95% CI 21-28%), with a failure rate of 0.8% (95% CI 0.0-4.1%). Retrospectively applied, the "clinical probability-adjusted" D-dimer strategies, specifically the YEARS and Pulmonary Embolism Graduated d-Dimer (PEGeD) algorithms, excluded PE in a significantly higher proportion (30% and 32%, respectively) compared with the "age-adjusted" D-dimer strategy (p<0.05 for both), with similar failure rates. PE is infrequent among patients presenting with haemoptysis, showing segmental or subsegmental emboli distribution. The "clinical probability-adjusted" D-dimer strategies seem to have significantly higher efficiency compared with the "age-adjusted" strategy.