P4. Abstract Title: Prevalence of Pulmonary Embolism in Patients Accessing the Emergency Department, Reported by Country of Study: A Systematic Review and Meta-Analysis

Abstract

<h3>Introduction</h3> Pulmonary embolism (PE) has an incidence of 0.75-2.69 cases per 1,000 individuals in Western Europe and North America (NA), with no significant differences between these two regions. In the United States, there is evidence of PE overdiagnosis, with increased incidence of PE over time and minimal changes in the mortality rate. It is possible that this tendency towards overdiagnosis (and over-testing) is different in NA and Europe. Greater tendency toward overtesting would imply a lower prevalence of PE among suspects and a lower diagnostic yield of imaging techniques. <h3>Aim</h3> The aim of the study was to determine whether the prevalence of PE among patients tested for PE in European and North American emergency department (EDs) differed. The secondary objectives were to determine if the rates of diagnostic imaging use and the diagnostic yield from imaging differed between the two continents. <h3>Methods</h3> The protocol was registered in the PROSPERO database. We searched MEDLINE and EMBASE from inception to September 3, 2017 for all studies which reported on consecutive ED patients tested for PE, without language restrictions. We screened titles and abstracts and then the full texts. Risk of bias was assessed using a modified version "Tool to Assess Risk of Bias in Longitudinal Symptom Research Studies Aimed at the General Population", form the CLARITY group, McMaster University. These steps and data extraction were performed independently by at least two authors. Conflicts were solved through discussion or with the intervention of a third author. For each outcome (PE prevalence, rate of diagnostic imaging, diagnostic yield of CTPA and ventilation-perfusion [VQ] scan), we conducted a meta-analysis of proportions specifying a random-effects model. We reported the pooled estimates as percentages with their 95% confidence intervals (CIs). A multiple meta-regression was conducted to evaluate if the effect of the region on PE prevalence and diagnostic yield of CTPA was independent of the year of execution of the studies. A sensitivity analysis was conducted only including studies at low risk of bias. <h3>Results</h3> The search identified 3109 records (2562 after duplicates removal). After screening the full text of 107 articles, 44 studies were included in the systematic review. Twenty-three of the 44 included studies were at low risk of bias. The overall prevalence of PE was 20% (95% CI 18%, 22%). The prevalence of PE in NA was 8% (95% CI 6%, 9%), while the prevalence of PE in Europe was 23% (95% CI 21%, 26%). After adjusting for the year of execution of the studies, the OR for PE prevalence in studies conducted in Europe versus NA was 1.16 (95%CI 1.11, 1.22). The overall frequency of use of CTPA was 53% (95% CI 45%, 62%). The frequency of CTPA use was 38% (95% CI 24%, 51%) in NA and 59% (95% CI 51%, 68%) in Europe, respectively. After adjusting for the year of study, the OR for the use of CTPA for studies conducted in Europe versus NA was 1.25 (95%CI 1.06, 1.48). The overall frequency of use of V/Q scan was 35% (95% CI 20%, 51%). The adjusted OR for the use of CTPA for studies conducted in Europe versus NA was 1.17 (95%CI 0.88, 1.52). The overall diagnostic yield was 25% (95% CI 20%, 29%). The diagnostic yield in NA was 13% (95% CI 9%, 17%), while the diagnostic yield in Europe was 29% (95% CI 26%, 32%). The adjusted OR for studies conducted in Europe versus NA was 1.16 (95%CI 1.09, 1.24). The overall diagnostic yield for VQ scan was 13% (95% CI 10%, 16%). The adjusted OR for studies conducted in Europe versus NA was 1.03 (95%CI 0.94, 1.14). The sensitivity analyses conducted including only studies at low risk of bias showed results consistent with the main analyses. We did not detect publication bias. <h3>Conclusion</h3> We found that studies reporting on ED patients tested for PE in NA had a lower prevalence of PE diagnosis compared to those conducted in Europe. There was a lower rate of CTPA use and a lower diagnostic yield from CTPA in NA compared to Europe. Our results support the hypothesis that those tested for PE in North American EDs have a lower risk of PE compared to Europe.