INTRODUCTION
Pulmonary embolism (PE) is rare in pediatric age, with an estimated incidence inferior to 1:100.000 in general pediatric population1,2. In the last decade, incidence increased in both community and hospitalized children3,5-7, although it is underdiagnosed, as evidenced by autopsy-based studies3,4. PE is associated with a high rate of adverse outcomes, with mortality up to 26%3,6,8. In pediatric age, almost all the events are associated with at least one risk factor or underlying disorder9. Pediatric PE occurs mostly in infants and teenagers1,2 and in this last group occurs twice as much in females, due to estrogen related risk factors such as pregnancy or combined hormonal contraceptives (CHC)2.
There is a paucity of data on pediatric PE. Given its rarity, most pediatric studies include venous thrombosis and thromboembolism (VTE), whereas PE only accounts for approximately 15% of VTE episodes7. The lack of well-performing pediatric probability models to assist in PE diagnosis is concerning8,10. Adult validated diagnostic prediction tools, such as the Wells criteria11 and the Pulmonary Embolism Rule-out Criteria (PERC) tool12, are often used13. However, they seem to lack sensitivity and specificity in the pediatric population14,15. Pediatric models have been proposed15,16 but were based in a small number of PE patients from single centers. PE management in children and adolescents are also often extrapolated from adult studies9, a population with different pathophysiology, morbidity and mortality. Thus, pediatric studies are needed to improve diagnostic tools, risk stratification and treatment options.
We intended to characterize patients with PE admitted in a tertiary hospital regarding their clinical presentation, risk factors, severity classification, treatment and outcomes. Secondarily, we intent to investigate the sensitivity of PE diagnostic prediction tools in this population.