Diagnosing Pulmonary Embolism in the context of Common Alternative diagNoses in primary care; the PECAN study.

Background and rationale: General practitioners (GPs) frequently encounter patients with (sub) acute shortness of breath. Pulmonary embolism (PE) is a relatively rare cause in primary care, yet potentially lethal and easy to miss. GPs have the difficult task to distinguish PE from common alternative cardiopulmonary diagnoses. Hereto clinical decision rules (CDRs) and D-dimer testing are available. Previous research from our group demonstrated that a low score on a CDR combined with a negative D-dimer safely rules-out PE in primary care. However, D-dimer testing is often falsely elevated, thus in the absence of PE. To add to complexity, this most often occurs in patients with cardiopulmonary diseases with symptomatology similar to that of PE, e.g. pneumonia or heart failure. This makes D-dimer testing inefficient, leading to unneeded, costly and potential harmful (risk of contrast nephropathy) referrals for CT pulmonary angiography (CTPA). Hence, GPs constantly balance over- and under-testing for suspected PE. Both are harmful and further improvement is needed. An important possible solution that could tackle this problem and help GPs better distinguish PE from mimicking cardiopulmonary diseases is the introduction of variable D-dimer thresholds (instead of a fixed threshold of 500 mcg/L). Such a diagnostic approach was recently tested in secondary care: the YEARS-strategy. In this strategy, the physician first scores three clinical items: (i) haemoptysis, (ii) clinical signs suggestive of deep venous thrombosis, and (iii) PE considered most likely diagnosis by the physician. If none of these items is present, a D-dimer threshold of 1000 mcg/L is applied, while if one or more items are present, the ‘classical’ threshold of 500 mcg/L is used. If a suspected patient is below either D-dimer threshold, PE is safely ruled out (no referral for CTPA). In a study performed in secondary care, this algorithm lead to an absolute reduction of 14% in referrals for CTPA with a completely similar safety (only 0.4% missed PE cases), as compared to a fixed D-dimer threshold of 500 mcg/L. Although appealing, this strategy first needs confirmation in primary care given the lower prior chance of PE as well as differences in case-mix of suspected patients in this setting. A second solution would be additional (beyond D-dimer) testing with point-of-care (POC) biomarkers, i.e. C-reactive protein for better diagnosing pneumonia or brain natriuretic peptides for making acute heart failure more likely. GPs seldom assess one single diagnosis (e.g. only PE, pneumonia or heart failure). Rather, they assess a patient with a constellation of symptoms like shortness of breath where multiple underlying causes are possible, all with different required actions. Hence, a triple POC biomarker approach may better classify underlying causes of shortness of breath, yet this first needs to be developed. Research objectives: Our primary objective is to validate the YEARS strategy in primary care in patients with (sub) acute shortness of breath and a suspicion of PE. Secondary objective is to quantify the added diagnostic information obtained from two other POC biomarkers (CRP and NTproBNP) in these patients. Study design: PECAN is a prospective diagnostic management study in primary care. We aim to include 750 participants (sample size, see full proposal) managed according to the YEARS-strategy using a point-of-care D-dimer assay. After 3 months, final diagnoses are adjudicated using a composite reference (see full proposal) based upon all available tests and information from the GP. Next, we will calculate (i) the proportion of missed PE cases in those not referred (i.e. the safety outcome, plus 95% CI) and (ii) the total number of non-referred patients as a proportion of all suspected cases (i.e. the efficiency outcome, plus 95% CI). Both safety and efficiency will be compared with the proportions from our own previous studies using a fixed D-dimer cut-off. For our secondary objective, additional blood will be drawn for CRP and NTproBNP testing. Analysis of the blood samples will be performed at the end of study, on POC assays (thus GPs remain blinded for test results during the study). The added diagnostic information of these tests beyond information already obtained from the YEARS-items and D-dimer testing will be quantified by means of multivariable regression analysis. Feasibility: We believe this study is feasible as (i) we have a strong network of GPs willing to participate in studies like this, (ii) our study group has ample experience with diagnostic studies and includes various disciplines like GPs, vascular internists, epidemiologists, and a patient representative, and (iii) we have access to multiple datasets from previous PE studies (most our own), allowing to compare our results with previous findings. All this enhances feasibility and knowledge transfer of our findings into the literature and guidelines.