Prevalence of Pulmonary Embolism Among Patients With COPD Hospitalized With Acutely Worsening Respiratory Symptoms

Key Points

Question  How common is pulmonary embolism among patients with chronic obstructive pulmonary disease who are admitted to the hospital with acutely worsening respiratory symptoms?

Findings  In this cross-sectional study with prospective follow-up that used a predefined pulmonary embolism diagnostic algorithm and included 740 consecutive patients with chronic obstructive pulmonary disease, pulmonary embolism was detected in 5.9% of patients.

Meaning  Among patients with chronic obstructive pulmonary disease admitted to the hospital with an acute worsening of respiratory symptoms, pulmonary embolism was detected in 5.9% patients using a predefined diagnostic algorithm.

Importance  The prevalence of pulmonary embolism in patients with chronic obstructive pulmonary disease (COPD) and acutely worsening respiratory symptoms remains uncertain.

Objective  To determine the prevalence of pulmonary embolism in patients with COPD admitted to the hospital for acutely worsening respiratory symptoms.

Design, Setting, and Participants  Multicenter cross-sectional study with prospective follow-up conducted in 7 French hospitals. A predefined pulmonary embolism diagnostic algorithm based on Geneva score, D-dimer levels, and spiral computed tomographic pulmonary angiography plus leg compression ultrasound was applied within 48 hours of admission; all patients had 3-month follow-up. Patients were recruited from January 2014 to May 2017 and the final date of follow-up was August 22, 2017.

Exposures  Acutely worsening respiratory symptoms in patients with COPD.

Main Outcomes and Measures  The primary outcome was pulmonary embolism diagnosed within 48 hours of admission. Key secondary outcome was pulmonary embolism during a 3-month follow-up among patients deemed not to have venous thromboembolism at admission and who did not receive anticoagulant treatment. Other outcomes were venous thromboembolism (pulmonary embolism and/or deep vein thrombosis) at admission and during follow-up, and 3-month mortality, whether venous thromboembolism was clinically suspected or not.

Results  Among 740 included patients (mean age, 68.2 years [SD, 10.9 years]; 274 women [37.0%]), pulmonary embolism was confirmed within 48 hours of admission in 44 patients (5.9%; 95% CI, 4.5%-7.9%). Among the 670 patients deemed not to have venous thromboembolism at admission and who did not receive anticoagulation, pulmonary embolism occurred in 5 patients (0.7%; 95% CI, 0.3%-1.7%) during follow-up, including 3 deaths related to pulmonary embolism. The overall 3-month mortality rate was 6.8% (50 of 740; 95% CI, 5.2%-8.8%). The proportion of patients who died during follow-up was higher among those with venous thromboembolism at admission than the proportion of those without it at admission (14 [25.9%] of 54 patients vs 36 [5.2%] of 686; risk difference, 20.7%, 95% CI, 10.7%-33.8%; P < .001). The prevalence of venous thromboembolism was 11.7% (95% CI, 8.6%-15.9%) among patients in whom pulmonary embolism was suspected (n = 299) and was 4.3% (95% CI, 2.8%-6.6%) among those in whom pulmonary embolism was not suspected (n = 441).

Conclusions and Relevance  Among patients with chronic obstructive pulmonary disease admitted to the hospital with an acute worsening of respiratory symptoms, pulmonary embolism was detected in 5.9% of patients using a predefined diagnostic algorithm. Further research is needed to understand the possible role of systematic screening for pulmonary embolism in this patient population.

Chronic obstructive pulmonary disease (COPD) is a frequent disease with high morbidity and mortality worldwide.¹ From 1969 to 2017, it was the third leading cause of death, and mortality is projected to increase in the next 20 years.^2,3 Acute exacerbations of COPD, defined by an acute worsening of respiratory symptoms leading to a change in medication, contribute to this mortality.¹

Quiz Ref IDBronchial infections or air pollution are major causes of acute exacerbation. However, several studies reported a high frequency of pulmonary embolism in these patients.^4-7 In a meta-analysis, patients admitted to the hospital for an unexplained acute COPD exacerbation had a 16% prevalence of pulmonary embolism, albeit with a wide 95% CI (8.3% to 25.8%).⁸ In a retrospective postmortem analysis, pulmonary embolism was the main cause of death in 21% of patients admitted for acute COPD exacerbation, independently of the premortem suspected cause of exacerbation.⁹ In addition, patients with COPD more often develop pulmonary embolism than deep vein thrombosis.¹⁰

However, how and when patients admitted to the hospital for an acute COPD exacerbation should be screened for pulmonary embolism remains challenging. Well-established diagnostic management of suspected acute pulmonary embolism in the general population might be ineffective in the setting of COPD exacerbations, particularly when including a ventilation-prefusion lung scan.^11-14 Furthermore, clinical presentations of acute pulmonary embolism and COPD exacerbation are similar, making it difficult to determine whether pulmonary embolism should be suspected in this context.¹⁵

In the present multicenter cross-sectional study with prospective follow-up, the main objective was to determine the prevalence of pulmonary embolism in patients with COPD admitted to the hospital for acutely worsening respiratory symptoms.

The prevalence of symptomatic Pulmonary Embolism in Patients With an Acute Exacerbation of Chronic Obstructive Pulmonary Disease (PEP) study was conducted in accordance with the ethical principles stated in the Declaration of Helsinki, Good Clinical Practice, and relevant French regulations regarding ethics and data protection. The protocol and amendments were approved by a central independent ethics committee (Person Protection Committee Ouest-6, Brest, France), and written informed consent was obtained from all patients before inclusion. The protocol and statistical analysis plan can be found in Supplement 1.

The study was supervised by an academic steering committee. An independent data and safety monitoring committee periodically reviewed study outcomes (recurrent venous thromboembolism, bleeding events, and deaths).

Quiz Ref IDFrom January 8, 2014, to May 22, 2017, consecutive outpatients 18 years or older admitted for acutely worsening respiratory symptoms of COPD to 1 of the 7 French participating centers were eligible (ie, emergency department or respiratory care unit). Availability of previous pulmonary function tests showing a postbronchodilator forced expiratory volume in the first second of expiration:forced vital capacity ratio of less than 70% was required to validate the diagnosis of COPD or a prior diagnosis of COPD by a pulmonologist was required.¹⁶ Acutely worsening respiratory symptoms of COPD was defined by a sustained worsening of patients’ baseline symptoms (dyspnea, cough, sputum) beyond normal day-to-day variability and a required treatment modification.

The main exclusion criteria were contraindication to undergoing spiral computed tomographic pulmonary angiography (known allergy to iodine contrast agents); creatinine clearance lower than 30 mL/min/1.73 m² (to convert to mL/s/m², multiply by 0.0167); pregnancy; hospitalization for more than 48 hours before inclusion; anticoagulant treatment for a condition other than venous thromboembolism; pneumothorax at admission; severe COPD exacerbation making imaging tests unfeasible; a life expectancy of less than 3 months; and inability to give written informed consent (eMethods in Supplement 2).

Patients were included within 48 hours following hospital admission and immediately classified as having clinically suspected or not suspected pulmonary embolism and as having a diagnosis other than pulmonary embolism more or less likely. These assessments were based on physicians’ judgment; no clinical scores were applied during this step. Thereafter, a predefined standardized algorithm was used to initiate the diagnostic work-up within 48 hours.

Quiz Ref IDClinical probability of pulmonary embolism was assessed in all patients using the revised Geneva score.¹⁷ Patients with a high clinical probability (ie, revised Geneva score ≥11) directly proceeded to spiral computed tomographic pulmonary angiography and legs ultrasound. For patients with a low or intermediate clinical probability (revised Geneva score <11), a D-dimer test was performed, which was interpreted according to the usual cutoff of 500 ng/mL threshold as Fibrinogen Equivalent Units. In patients with a negative D-dimer test result (ie, D-dimer value <500 ng/mL, to convert to nmol/L, multiply by 5476), pulmonary embolism was excluded and no additional testing was performed; these patients were left without anticoagulant therapy.

In case of confirmed venous thromboembolism, anticoagulation was administered according to current guidelines.¹⁴ The diagnostic strategy is depicted in the eFigure in Supplement 2. The study protocol had no influence on prophylactic anticoagulation, which was started as recommended within 24 hours following admission in the absence of venous thromboembolism.¹⁸

All patients were followed up at 3 months. Patients were instructed to report to the study center immediately if any symptoms or signs suggestive of venous thromboembolism or bleeding occurred between the inclusion and the 3-month follow-up visit. At the end follow-up, all included patients were interviewed at a visit or by phone call by a physician using a structured questionnaire. All events occurring after hospital discharge were collected and charts were reviewed in case of hospital readmission for any cause.

Quiz Ref IDThe primary outcome was pulmonary embolism diagnosed within 48 hours of admission. The diagnosis of pulmonary embolism was based on the presence of an intraluminal filling defect in a segmental or more proximal pulmonary artery on spiral computed tomographic pulmonary angiography.¹⁴ In case of inconclusive computed tomographic angiography (eg, isolated subsegmental pulmonary embolism¹⁹), pulmonary embolism diagnosis was based on (1) a segmental or larger perfusion defect with normal ventilation on ventilation-perfusion lung scan; (2) a proximal deep vein thrombosis on lower-limb ultrasonography showing noncompressibility of a proximal vein; or (3) documented pulmonary embolism on autopsy.¹⁴

The key secondary outcome was pulmonary embolism that occurred during the 3-month follow-up in patients deemed not to have had venous thromboembolism within 48 hours of admission and who did not receive anticoagulant treatment. Other secondary outcomes were (1) venous thromboembolism at admission, including pulmonary embolism and/or isolated proximal or distal deep vein thrombosis; (2) venous thromboembolism during 3-month follow-up in patients deemed not to have had venous thromboembolism at admission and who had not received anticoagulant treatment; (3) 3-month all-cause mortality in all patients; and (4), alternative diagnoses to venous thromboembolism at admission. Deep vein thrombosis diagnosis was based on lower-limb ultrasonography showing noncompressibility of a distal or a proximal vein.¹⁴ Diagnostic tests are described in the eMethods in Supplement 2. All outcomes were adjudicated by an independent clinical events committee.

The sample size was estimated to obtain an estimate of the primary outcome. For a prevalence of 10% (based on the outpatient subpopulation in the meta-analysis of Aleva et al⁸) and a 95% CI comprised between −25% and 25% of the observed percentage of pulmonary embolism at admission, 690 patients were required. This estimate was based on binomial distribution for the exact calculation of the bilateral 95% CI. Considering potential nonevaluable patients (incomplete diagnostic workup, consent withdrawal, etc), the final sample size was 750 patients.

For outcomes analyses, missing data were not replaced. Outcomes and case-fatality rates were calculated (1) in the whole population; (2) whether pulmonary embolism was suspected within 48 hours of admission; and (3) whether venous thromboembolism was diagnosed at admission. The upper limit of the unilateral 95% CI for the 3-month thromboembolic risk should not exceed 3% for a clinically appropriate diagnostic strategy.¹⁹ In a post hoc analysis, D-dimer was interpreted according to age-adjusted D-dimer threshold and subsequent absolute reduction of computed tomographic angiography was calculated.²⁰

All tests were 2-sided, and a P value <.05 was considered statistically significant. Because of the potential for type I error due to multiple comparisons, findings for analyses of secondary end points should be interpreted as exploratory. Statistical analyses were performed using SAS version 9.4 software (SAS Institute Inc).

From January 2014 to May 2017, 2268 consecutive patients with COPD and acutely worsened respiratory symptoms were admitted in the 7 participating hospitals and were immediately screened to participate in this study; of those, 1518 presented with exclusion criteria, mainly related to long-term anticoagulation or inability to give informed consent (Figure). Of the remaining 750 patients, 10 were not analyzed for the following reasons: 5 had been previously included at the time of a prior exacerbation, 3 were under guardianship, and 2 withdrew consent after inclusion. Among the remaining 740 patients, inclusion process and initiation of diagnostic workup was performed within 48 hours of admission.

The baseline characteristics of the 740 (mean age, 68.2 years [SD, 10.9 years]; 274 women [37.0%]) analyzed study patients are summarized in Table 1. The following data were missing at admission: tobacco smoking status for 4, prior spirometry for 63, and dyspnea scale for 15 patients. Overall, diagnostic protocol violations occurred among 139 of the 740 included patients (18.8%), 16 (11.5%) were considered major (Figure; eTable 1 in Supplement 2).

Among the 740 included patients, 17 (2.3%) had a high Geneva pretest clinical probability: pulmonary embolism was confirmed in 5 patients on spiral computed tomographic pulmonary angiography; 2 were associated with a distal deep vein thrombosis (Figure). For the remaining 12 patients without pulmonary embolism following computed tomographic angiography, an isolated distal deep vein thrombosis was diagnosed in 2 patients, legs compression ultrasonography was negative for 2 patients and was not performed for 8 patients.

Among the 723 patients with a low or moderate pretest clinical probability, a D-dimer assay was performed in 712 patients. D-dimer levels were lower than 500 μg/mL in 212 patients and pulmonary embolism was considered excluded without further work up. Among the 500 patients with a positive D-dimer test result, 478 patients underwent spiral computed tomographic pulmonary angiography, 36 of whom had confirmed pulmonary embolism. Thirteen were associated with deep vein thrombosis (proximal, 6; distal, 7). Among the 442 patients with a negative computed tomographic angiographic result, 8 had an isolated deep vein thrombosis (proximal, 4; distal, 4), 339 had a negative leg compression ultrasonographic result, and 95 did not undergo leg compression ultrasonography. Among the 22 patients with positive D-dimer results who did not undergo computed tomographic angiography, ventilation-perfusion lung scans were performed for 6 patients and did not reveal pulmonary embolism (normal or low probability); of the 16 patients who did not have ventilation-prefusion lung scan, leg compression ultrasonography was negative in 1 patient and was not performed in the remaining 15 patients. Among the 11 patients with a low to moderate pretest clinical probability who did not have D-dimer testing, computed tomographic angiography confirmed pulmonary embolism among 3 patients and was negative in the remaining 8 patients.

Thus, among the 740 included patients, pulmonary embolism was confirmed in 44 patients (5.9%; 95% CI, 4.5%-7.9%) (Figure).

Within 48 hours of admission, an additional 10 patients had an isolated deep vein thrombosis (1.4%) (4 proximal; 6 distal) yielding an overall prevalence of venous thromboembolism of 7.3% (95% CI, 5.6%-9.4%) at admission (Figure, Table 2). Details regarding pulmonary emboli extent and deep venous thrombosis location within 48 hours of admission are shown in eTables 2 and 3 in Supplement 2. Final causes of acutely worsening respiratory symptoms in COPD patients are listed in Table 3.

At the 3-month follow-up, among the 686 patients deemed not to have pulmonary embolism or deep vein thrombosis at admission, curative anticoagulation was started in 14 for reasons other than adjudicated venous thromboembolism and 2 were lost-to follow-up. Among the remaining 670 patients, pulmonary embolism was diagnosed during follow-up in 5 patients (0.7%; 95% CI, 0.3%-1.7%): 2 patients had objectively confirmed pulmonary embolism, and death was related to fatal pulmonary embolism in 3 patients (Figure, Table 2). No isolated cases of deep venous thrombosis were detected at 3 months.

The overall 3-month mortality rate was 6.8% (50 of 740; 95% CI, 5.2%-8.8%). The adjudicated causes of deaths are shown in Table 4. Three-month mortality rates among patients with venous thromboembolism at admission was 14 (25.9%) of 54 vs 36 (5.2%) of 686 among those without it (risk difference, 20.7%; 95% CI, 10.7%-33.8%; P < .001) (eTable 4 in Supplement 2).

Among the 299 patients suspected at admission of having pulmonary embolism, 30 (10%) were diagnosed vs 14 patients (3.2%) of 441 who were not suspected at admission were diagnosed with pulmonary embolism (risk difference, 6.9%; 95% CI, 3.3%-11.0%; P < .001). Twenty-two patients (12.7%) of 173 for whom pulmonary embolism was considered by the physician as the most likely diagnosis (more than any other diagnosis) had a pulmonary embolism vs 22 patients (3.9%) of 567 for whom pulmonary embolism was considered a less likely diagnosis had a pulmonary embolism (risk difference, 8.8%; 95% CI, 4.3%-14.8%; P < .001). Three-month mortality was similar among these subgroups (Table 2; eTable 4 in Supplement 2).

In post hoc analysis (eTable 5 in Supplement 2), applying the age-adjusted D-dimer threshold would have reduced the use of computed tomographic angiographies performed by 22%; however, among the 110 patients 50 years or older who had D-dimer concentrations higher than 500 ng/mL and less than the age-adjusted threshold, 4 patients had venous thromboembolism diagnosed at admission and 1 had fatal pulmonary embolism within 3 months (4.5%, 95% CI, 2.0%-10.2%).

Quiz Ref IDIn this cross-sectional study with prospective follow-up including 740 patients with COPD admitted to the hospital for acutely worsening respiratory symptoms, systematic screening for pulmonary embolism within 48 hours of hospital admission showed a pulmonary embolism prevalence of 5.9%, and an additional 1.4% of patients had an isolated deep vein thrombosis. The prevalence of pulmonary embolism reached 10% when it was suspected, and remained at more than 3% among patients without a clinical suspicion.²¹ Among the 670 patients for whom the diagnosis of venous thromboembolism was ruled-out at inclusion and who did not receive anticoagulant therapy during the 3-month follow-up, 0.7% developed pulmonary embolism.

The overall prevalence of pulmonary embolism in patients with COPD admitted to the hospital for acutely worsening respiratory symptoms in this study (5.9%) is lower than the 16.1% reported in a meta-analysis involving 880 patients.⁸ However, the comparison with the 7 studies included in this meta-analysis is difficult due to low sample sizes (<200 patients in the largest study). In addition, the heterogeneity across studies in terms of pulmonary embolism diagnostic algorithms, COPD assessment and study populations, with some studies recruiting only patients with no evident cause of exacerbation other than suspected pulmonary embolism and others analyzing only patients with severe acutely worsening respiratory symptoms of COPD renders definitive conclusions difficult.^5-8 Therefore, to our knowledge, this study is the largest prospective study assessing the prevalence of pulmonary embolism—whether the diagnosis was suspected or not—using a predefined and standardized diagnostic algorithm for pulmonary embolism within 48 hours following admission among patients with documented COPD (ie, using validated criteria¹⁶) who were admitted to the hospital for acutely worsening respiratory symptoms.^14,16 Because all pulmonary embolism events were adjudicated and confirmed by an independent clinical events committee using predefined, standardized, and validated criteria,¹⁴ this study provides a valid and reliable prevalence estimate of pulmonary embolism in this setting.

Among patients in whom pulmonary embolism was suspected at admission, the prevalence of pulmonary embolism (10%) was close to that observed in recent diagnostic management studies performed in unselected outpatients with a suspicion of pulmonary embolism.^22,23 However, in patients without a clinical suspicion of pulmonary embolism, the prevalence of venous thromboembolism (3.2%, Table 2) cannot be considered as negligible, which raises the question of whether performing a systematic search for pulmonary embolism even in this subpopulation should occur. The observation of additional patients with isolated proximal deep vein thrombosis further supports such consideration.

During the 3-month follow-up, only 0.7% of patients (95% CI, 0.3%-1.7%) considered not to have venous thromboembolism within 48 hours of admission and not treated with an anticoagulant agent developed pulmonary embolism. Among patients with clinically suspected pulmonary embolism, the upper limit of the 95% CI was also lower than the safety threshold of 3% proposed by the subcommittee of the International Society of Thrombosis and Hemostasis.²¹ It remains to be determined in larger studies if, after a negative computed tomographic angiographic scan, a leg compression ultrasound should be systematically added when patients with COPD and acutely worsening respiratory symptoms are admitted.²⁴ It is an important question because this practice is no longer recommended for patients with suspected pulmonary embolism.¹⁴

Consistent with other studies, mortality was higher among patients with pulmonary embolism than it was for patients without it.^10,25-27 As shown in Table 4, 42% (6 of 14) of adjudicated deaths among those in the venous thromboembolism group were related to cancer. It is also known that the prognosis of patients with cancer is poorer when associated with associated venous thromboembolism.²⁸ In practice, clinicians should consider the probability of an underlying cancer when pulmonary embolism is diagnosed in patients with COPD exacerbation.

This study has several limitations. First, patients with mild acutely worsening respiratory symptoms, as well as those with severe respiratory failure, were likely underrepresented, so that these finding may not apply to such patients. However, in this study, these criteria corresponded with usual practice in the setting of patients with COPD admitted to the hospital for acutely worsened respiratory symptoms. Second, among the 686 patients considered not to have venous thromboembolism within 48 hours of admission, 121 (17.6%) did not complete the full initial assessment for pulmonary embolism, predominantly due to limited access to leg ultrasound (Figure; eTable 1 in Supplement 2). These (mostly minor) protocol violations may have slightly underestimated venous thromboembolism prevalence at inclusion, but all these patients were followed-up and none had a symptomatic venous thromboembolic event within 3 months. Third, it remains possible that having a clinical suspicion of pulmonary embolism by emergency department physicians contributed to the decision to admit some patients, which potentially biased the sample toward a higher pulmonary embolism frequency in this study. Fourth, among patients with low to moderate pretest clinical probability, the age-adjusted D-dimer threshold was not applied, the study having started before the prospective validation of this threshold in unselected patients.²⁰ Based on post hoc analysis, whether the age-adjusted D-dimer threshold algorithm excludes pulmonary embolism and is a cost-effective diagnostic test for patients with COPD remains to be confirmed in larger prospective study.

Among patients with chronic obstructive pulmonary disease admitted to the hospital with an acute worsening of respiratory symptoms, pulmonary embolism was detected in 5.9% of patients using a predefined diagnostic algorithm. Further research is needed to understand the possible role of systematic screening for pulmonary embolism in this patient population.

Corresponding Author: Francis Couturaud, MD, PhD, Département de Médecine Interne et Pneumologie, Hôpital de la Cavale Blanche, CHRU de Brest, 29609 Brest cedex, France (francis.couturaud@chu-brest.fr).

Accepted for Publication: November 11, 2020.

Author Contributions: Dr Couturaud had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.

Concept and design: Couturaud, Bertoletti, Hoffmann, Mismetti.

Acquisition, analysis, or interpretation of data: Couturaud, Bertoletti, Pastre, Roy, Le Mao, Gagnadoux, Paleiron, Schmidt, Sanchez, De Magalhaes, Kamara, Bressollette, Nonent, Tromeur, Salaun, Barillot, Gatineau, Mismetti, Girard, Lacut, Lemarié, Meyer, Leroyer.

Drafting of the manuscript: Couturaud, Bertoletti, Sanchez, Gatineau, Mismetti, Girard, Meyer, Leroyer.

Critical revision of the manuscript for important intellectual content: Couturaud, Bertoletti, Pastre, Roy, Le Mao, Gagnadoux, Paleiron, Schmidt, Sanchez, De Magalhaes, Kamara, Hoffmann, Bressollette, Nonent, Tromeur, Salaun, Barillot, Mismetti, Girard, Lacut, Lemarié, Meyer.

Statistical analysis: Gatineau.

Obtained funding: Couturaud.

Administrative, technical, or material support: Couturaud, Bertoletti, Schmidt, Bressollette, Barillot, Girard, Lacut, Meyer, Leroyer.

Supervision: Couturaud, Bertoletti, Paleiron, Meyer.

Other: Nonent, Tromeur.

Conflict of Interest Disclosures: Dr Couturaud reported having received grant support from Bristol-Myers Squibb–Pfizer; fees for board memberships or symposia from Bayer, Bristol-Myers Squibb–Pfizer, and AstraZeneca; and travel support from Bayer, Bristol-Myers Squibb–Pfizer, Daiichi Sankyo, Leo Pharma, InterMune, and Actelion. Dr Bertoletti reported receiving grant support from Bayer and fees for board memberships or symposia from Actelion, Aspen, Bayer, Bristol-Myers Squibb–Pfizer, and Merck Sharp and Dohme; and travel support from Aspen, Bayer, Bristol-Myers Squibb–Pfizer, Daiichi Sankyo, Leo Pharma, Merck Sharp and Dohme, and Actelion. Dr Sanchez reported receiving grant support from Bayer, Daiichi Sankyo, and Portola Pharmaceuticals and fees or nonfinancial support for consultancy activities from Actelion, GlaxoSmithKline, Boehringer Ingelheim, and Chiesi. Dr Mismetti reported receiving grants from Bayer; fees for board memberships from Bayer, Bristol-Myers Squibb–Pfizer, and Daiichi Sankyo; lecture fees from Bayer, Boehringer Ingelheim, Bristol-Myers Squibb–Pfizer, Daiichi Sankyo, and Sanofi; and fees for development of educational presentations from Bayer and Bristol-Myers Squibb–Pfizer. Dr Girard reported receiving personal fees and nonfinancial support from Bayer and Leo Pharma. Dr Lacut reported receiving personal fees from Bayer Healthcare, Bristol Myers Squibb, and Boehringer Ingelheim. Dr Meyer reported receiving grant support from Bayer, Boehringer Ingelheim, Leo Pharma, and Sanofi, being an uncompensated board member and a consultant for Bayer, Boehringer-Ingelheim, Bristol Myers Squibb, Leo Pharma, and Pfizer, and receiving travel support from Bayer, Boehringer Ingelheim, Daiichi Sankyo, Leo Pharma, and Sanofi. Dr Leroyer reported receiving grant support from Pfizer, fees for board memberships or symposia from Bayer and AstraZeneca, travel support from Bayer, Daiichi Sankyo, Leo Pharma, InterMune, and Actelion. No other disclosures were reported.

Funding/Support: The study was supported by grants from the Programme Hospitalier de Recherche Clinique (French Department of Health), and the sponsor was the University Hospital of Brest.

Role of the Funder/Sponsor: Neither sponsor had a role in the design and conduct of the study; collection, management, analysis, and interpretation of the data; preparation, review, or approval of the manuscript; and decision to submit the manuscript for publication. An academic steering committee assumed overall responsibility for all these steps.

Members of the PEP Study Group (all in France): Steering Committee: Francis Couturaud, MD (chair), and Christophe Leroyer, MD, Centre Hospitalo-Universitaire de Brest, Brest; Olivier Sanchez, MD, and Guy Meyer, MD, Hôpital Européen Georges Pompidou, Paris; and Laurent Bertoletti, MD, and Patrick Mismetti, MD, Centre Hospitalo-Universitaire de Saint-Etienne, Saint-Etienne. Coordinating Committee: Francis Couturaud, MD (chair), Sophie Barillot, MS, Elise Poulhazan, MS, and Anne Le Brestec, MS, Centre Hospitalo-Universitaire de Brest, Brest. Independent Central Adjudication Committee (Critical Events): Philippe Girard, MD (chair), Stéphane Lenoir, MD, and Christian Lamer, MD, Institut Mutualiste Montsouris, Paris; Statistical Analysis: Florence Gatineau, MS, Centre Hospitalo-Universitaire de Brest, Brest; Operation Team: Dauphou Eddy, MS, and Florence Morvan, MS, Centre Hospitalo-Universitaire de Brest, Brest; Computerised Tomographic Scan Panel: Michel Nonent, MD, Pierre-Yves Le Floch, MD, and Fabien Podeur, MD, Centre Hospitalo-Universitaire de Brest, Brest; Lung Scintigraphy Panel: Pierre-Yves Salaun, MD, Philippe Robin, MD, and Pierre-Yves Le Roux, MD, Centre Hospitalo-Universitaire de Brest, Brest; and Ultrasound Panel: Luc Bressollette, MD, and Clément Hoffmann, Centre Hospitalo-Universitaire de Brest, Brest; Investigators (in order of the number of patients enrolled): Francis Couturaud, MD, Christophe Leroyer, MD, Cécile Tromeur, MD, Romain Didier, MD, Karine Lacut, MD, Emmanuelle Lemoigne, MD, Raphael Le Mao, MD, Charles Orione, MD, Alexandre Fauché, MD, Christophe Gut-Gobert, MD, Claire De Moreuil, Centre Hospitalo-Universitaire de Brest, Brest (365); Olivier Sanchez, MD, Guy Meyer, MD, Benjamin Planquette, MD, and Jean Pastre, MD, Hôpital Européen Georges Pompidou, Paris (99); Pierre-Marie Roy, MD, and Frédéric Gagnadoux, MD, Centre Hopitalo-Universitaire d’Angers, Angers (84); Nicolas Paleiron, MD, Hôpital d’instruction militaire Clermont Tonnerre and Clermont Tonnerre, Hôpital d’instruction militaire, Brest (79); Patrick Mismetti, MD, Sandrine Accassat, MD, Laurent Bertoletti, MD, Elodie De Magalhaes, MD, Alain Viallon, MD, Centre Hospitalo-Universitaire de Saint-Etienne, Saint-Etienne (76); Mariam Kamara, MD, Centre Hospitalier Général de Quimper, Quimper (32); and Jeannot Schmidt, MD, Centre Hospitalo-Universitaire de Clermont-Ferrand, Clermont-Ferrand (15). No member of the PEP Study Group received compensation for his/her role in the study.

Meeting Presentations: The results of this study were presented in the late-breaking abstract session at the 29th European Respiratory Society Annual Meeting; October 1, 2019; Madrid, Spain.

Additional Information: Dedicated to Guy Meyer.