Chronic obstructive pulmonary disease exacerbation purulence status and its association with pulmonary embolism: protocol for a systematic review with meta-analysis

Abstract

Introduction

Acute exacerbation of chronic obstructive pulmonary disease (AECOPD) increases the risk of pulmonary embolism (PE). AECOPD and PE have similar symptoms which results in a high proportion of patients with AECOPD undergoing imaging to rule out PE. Finding predictors and explanatory factors of PE in AECOPD, such as purulence status, could help reduce the need for imaging. This systematic review with meta-analysis aims to evaluate if there is an association between purulence status in AECOPD and PE diagnosis.

Methods and analysis

MEDLINE, EMBASE and CENTRAL will be searched from database inception to April 2024. Randomised trials, cohort studies and cross-sectional studies on the prevalence of PE in patients with AECOPD will be included if the prevalence of PE based on the AECOPD purulence status is available. There will be no restriction on language. The primary outcome will be PE at the initial assessment and secondary outcomes will be all venous thromboembolism (deep venous thrombosis (DVT) and PE) and DVT, respectively, diagnosed at the initial assessment. Relative risks with their 95% CI will be calculated by using a Mantel-Haenszel random-effect model to compare the association between the risk of PE and the AECOPD purulence status (purulent vs non-purulent/unknown). Subgroup analyses will be performed based on the type of study, systematic search of PE versus no systematic search of PE and localisation of PE. Risk of bias will be evaluated by the ROBINS-E tool, publication bias will be evaluated with the funnel plot. The manuscript will be drafted based on the Preferred Reporting Items for Systematic Reviews and Meta-Analysis statement.

Ethics and dissemination

This study does not require ethics approval. This work will be submitted for presentation at an international conference and for publication in a peer-reviewed journal.

PROSPERO registration number

CRD42023459429.

STRENGTHS AND LIMITATIONS OF THIS STUDY.

• An experienced patient partner from the Canadian Venous Thromboembolism Research Network patient partner platform was involved in the protocol elaboration.

• The acute exacerbation of chronic obstructive pulmonary disease (AECOPD) purulence status may not be homogeneous across studies, which may make it more challenging to pool some data.

• Not all studies report on the prevalence of pulmonary embolism (PE) according to the AECOPD purulence status; consequently, the data included in this systematic review may represent a limited proportion of all the data available on the prevalence of PE in patients with AECOPD.

Introduction

Acute exacerbation of chronic obstructive pulmonary disease (AECOPD) increases the risk of pulmonary embolism (PE)¹ due to increased systemic inflammation as well as in the airways.² Moreover, PE is associated with a fivefold increased risk of mortality in patients with chronic obstructive pulmonary disease (COPD).³ Diagnosing PE in the context of AECOPD is challenging for several reasons. First, due to confounding symptoms of AECOPD and PE, it is unknown when PE should be suspected in patients with COPD. Second, even when PE is not suspected, or when another diagnosis is more likely, the prevalence of PE [i.e., 4.5% (PEP⁴ and SLICE⁵)] is not low enough to safely exclude PE on clinical grounds only. Clinical decision rules and D-dimers, when applied to patients with AECOPD and whether PE is suspected or not, have lower clinical utility in AECOPD since >65% of the patients would need imaging to rule out PE if standard diagnostic strategy was used.⁴ In addition, negative effects are seen with CT pulmonary angiogram (CTPA) such as cost, radiation exposure, contrast-induced nephropathy and incidental findings. Furthermore, as the severity of the COPD progresses, AECOPD occurs more frequently⁶ and it is expected that the need to rule out PE will become more frequent. Finding predictors and explanatory factors of PE in AECOPD, such as the purulence status, could help reduce the need for imaging. Clinically, it would make sense that if the AECOPD is explained by an infectious process, then the PE would be less likely, and conversely, if the AECOPD is unexplained, it would make sense that PE would be more likely to be the explanation for the AECOPD. As a matter of fact, some studies showed a lower risk of PE or venous thromboembolism (VTE) in patients with purulent AECOPD.^7–9

Thus, the main aim of this systematic review with meta-analysis is to evaluate whether purulence status in AECOPD is associated with PE. We hypothesise that the risk of PE will be lower in purulent AECOPD compared with non-purulent or unknown purulent status AECOPD since the aetiology of the exacerbation is unknown in up to 30% of the AECOPD¹⁰ and PE could thus be an explanation in those cases. As a secondary aim, we would like to evaluate the association between AECOPD purulence status and the risk of VTE (deep venous thrombosis (DVT) of the lower extremity and PE) and the risk of DVT, respectively. We hypothesise that the risk of VTE and DVT, respectively, will be lower in patients with purulent AECOPD compared with non-purulent or unknown purulent status AECOPD.

Study objectives

Primary objective

The primary objective is to evaluate the risk of PE in patients with purulent AECOPD compared with non-purulent or unknown purulent status AECOPD.

Secondary objective

The secondary objective is to evaluate the risk of VTE (including DVT of the lower extremity and PE) and the risk of DVT, respectively, in patients with purulent AECOPD compared with non-purulent or unknown purulent status AECOPD.

Methods and analysis

Eligibility criteria

Randomised trials, cohort studies (retrospective or prospective) and cross-sectional studies on the prevalence of PE in patients with AECOPD will be included if the prevalence of PE according to the AECOPD purulence status is available. AECOPD purulence status will be categorised as definitive purulent AECOPD (purulent AECOPD or purulent sputum), possible purulent AECOPD (clinical and/or radiological evidence of tracheobronchial infection or pneumonia), non-purulent AECOPD or unknown purulence status AECOPD.

Information sources and search strategy

MEDLINE, EMBASE and CENTRAL will be searched from inception to April 2024. Conference abstracts from the American Thoracic Society, American College of Chest Physicians, European Respiratory Society, British Thoracic Society, American Society of Hematology, International Society on Thrombosis and Haemostasis will be hand searched from January 2000 to April 2024. There will be no restriction on language. The search strategy (online supplemental appendix 1) will be reviewed by a research librarian with expertise in knowledge synthesis and translation.

Study records

Two reviewers (VM and LG) will independently screen all the titles and abstracts for potentially eligible studies. Full texts of potentially eligible studies will be obtained and screened by two reviewers independently. Both levels of screening will be conducted using Covidence systematic review software, Veritas Health Innovation, Melbourne, Australia. Any disagreements will be resolved by further discussion or by consulting a third reviewer (GLG). If the same cohort was published in multiple papers, the paper with the largest cohort providing the required information needed will be selected.

Data items

Two independent reviewers (VM and LG) will extract the data from included papers by using a standardised collection form. Collected data will include study characteristics (study ID, reference, study design), patients’ characteristics (number of patients, age, sex, body mass index, mean forced expiratory volume in 1 s, Global Initiative for Chronic Obstructive Lung Disease stage, prior personal or familial venous thromboembolic event, current tobacco use, active cancer (defined as current diagnosis of cancer, receiving treatment for cancer or not receiving treatment for cancer and not in complete response as per the International Society on Thrombosis and Haemostasis Common Data Elements), the number of previous AECOPD in the last year, pretest clinical probability, mean D-dimers level, VTE (PE and/or DVT), AECOPD purulence status), proportion of patients who had imaging to rule out VTE, whether or not all patients systematically had diagnostic imaging searching for PE (or VTE), localisation of PE, clinical setting (inpatients vs outpatients) and the use of independent adjudication. Study authors will be contacted if important information is missing.

Outcome measures

The primary outcome will be PE at the initial assessment. PE will include symptomatic PE involving subsegmental branches or more proximal arteries on CTPA, high probability on a planar ventilation/perfusion (V/Q) scan, at least one segmental mismatch or two subsegmental mismatches on a V/Q SPECT (EANM criteria)¹¹ and incidental PE found fortuitously on imaging and fatal PE. If the localisation of the PE was not mentioned in the article, the study will still be included, and subgroup analyses will be performed. Secondary outcomes will include VTE (proximal DVT and/or PE), proximal DVT and distal DVT, respectively, at the initial assessment. DVT will include DVT of the lower extremity, either symptomatic or incidental. In case it was not mentioned if the DVT was proximal or distal, the study will still be included, and subgroup analyses will be performed. The initial assessment will be defined as the first 48 hours from hospital admission if the patient is admitted, as the first 48 hours from the initial medical evaluation if the patient is managed as an outpatient or as defined by individual studies.

Assessment of risk of bias in included studies

The risk of bias of included studies will be evaluated by two independent reviewers (VM and LG) by using the ROBINS-E tool.¹² Publication bias will be assessed by conducting and evaluating the funnel plot for the primary outcome. A symmetrical funnel plot indicates the absence of publication bias.

Data synthesis

The prevalence of PE, VTE and DVT, respectively, at initial assessment will be calculated with its 95% CI by using the binomial exact method¹³ for each study. Data will be pooled using Review Manager V.5.3 (The Cochrane Collaboration, Oxford, England). Relative risks (RR) with their 95% CI will be calculated by using a Mantel-Haenszel random-effects model to compare the association between the risk of PE in patients with purulent AECOPD and the risk of PE in patients with non-purulent/unknown purulence status AECOPD. Events will be categorised in the definitive purulent AECOPD group if it was mentioned purulent AECOPD or the sputum was described as purulent. Events will be categorised in the possible purulent AECOPD group if there is clinical and/or radiological evidence of tracheobronchial infection or pneumonia. Similar analyses will be conducted to evaluate the association between the risk of VTE and the risk of DVT, respectively, and the AECOPD purulence status. Forest plots will be presented. If some studies cannot be pooled in the RR analysis evaluating the association between the risk of PE and the type of AECOPD, pooled proportions of PE of patients with purulent AECOPD and with non-purulent/unknown purulence status AECOPD, respectively, will be calculated using StatsDirect statistical software. I² will be calculated to evaluate heterogeneity and will be considered significant if I² is >50%. Subgroup analyses will be performed based on the type of study (randomised trials vs prospective cohort studies vs retrospective cohort studies vs cross-sectional studies), systematic search of PE (or VTE) vs no systematic search of PE (or VTE) and localisation of PE (or DVT). Sensitivity analyses will be performed by including only studies at low risk of bias. The manuscript will be drafted based on the Preferred Reporting Items for Systematic Reviews and Meta-Analysis statement.

Patient and public involvement

An experienced patient partner from the Canadian Venous Thromboembolism Research Network patient partner platform revised the protocol and approved the design and conduct of the study, as well as the outcome measures.

Ethics and dissemination

Since this is a systematic review with meta-analysis of published studies, ethics approval and patients’ consent will not be required. We aim to submit this work for presentation at an international conference and for publication in a peer-reviewed journal.

Discussion

This systematic review with meta-analysis aims at comparing the association between the risk of PE in patients with purulent AECOPD and the risk of PE in patients with non-purulent/unknown purulence status AECOPD. Finding predictors or explanatory factors for PE in patients with AECOPD, such as AECOPD purulence status, could help reduce the need for imaging. If the risk of PE is shown to be lower in patients with purulent AECOPD compared with non-purulent or unknown status AECOPD, this new information may help improve PE diagnostic algorithm in reducing the need for imaging in ruling out PE and thus, improve the care of patients with AECOPD. Moreover, if the prevalence of PE is shown to be very low in patients with purulent AECOPD and low enough to exclude PE without further investigations, this will certainly reduce the need for imaging in ruling out PE and subsequently, reduce the side effects of CTPA.

We acknowledge that this study may have some limitations and that we may face some challenges when conducting it. First, only a certain number of studies on the prevalence of PE in patients with AECOPD have reported the prevalence of PE based on the AECOPD purulence status. The data included in this systematic review may thus represent a limited proportion of all the data available on the prevalence of PE in patients with AECOPD. Second, the definition of the AECOPD purulence status may not be homogeneous across studies which could make it more challenging to pool the data. Finally, although we will analyse all patients with AECOPD, there might be some heterogeneity within this population (e.g., patients admitted vs treated as an outpatient).

Improving PE diagnostic algorithm for patients with AECOPD is of high importance to reduce the burden of imaging since PE and AECOPD share similar symptoms but also to minimise the proportion of missed PE. This systematic review with meta-analysis aims at evaluating if AECOPD purulence status could be a predictor of PE in order to improve the care of patients with COPD.

Ethics statements

Patient consent for publication

Not applicable.