Diagnosis of pulmonary embolism: should we wait YEARS?

Venous thromboembolism (VTE), comprising deep vein thrombosis (DVT) and pulmonary embolism (PE), has a prevalence in the adult population of 1–2 cases per 1000. It is the third most common cause of cardiovascular disease and is an important cause of both morbidity and mortality resulting in an estimated 300 000 VTE-related deaths per year in Europe (Cohen et al, 2007). The diagnosis of PE is challenging because of the non-specific symptoms and signs in patients at presentation and the potentially dire consequences of a missed thromboembolic event. For adults presenting with features suggestive of PE, the current guidelines from the United Kingdom National Institute for Health and Care Excellence (NICE 2012), European Society of Cardiology (ESC; Konstantinides et al, 2014) and American College of Chest Physicians (Kearon et al, 2016) all recommend that a pre-test probability score is calculated (Wells et al, 2001; Le Gal et al, 2006). In patients with a “likely probability” of PE score, the guidelines recommend proceeding to imaging of the lungs utilising computed tomography pulmonary angiography (CTPA) whereas for those who have an “unlikely probability” of PE measurement of a biomarker in the blood, the D-dimer, is advised. The soluble blood coagulation protein fibrinogen is a dimeric molecule made up of three pairs of polypeptide chains Aα, Bβ and ϒ. Following thrombin generation, soluble fibrinogen is converted to insoluble fibrin (clot) by cleavage of fibrinopeptides A and B (FpA and FpB). This results in the formation of fibrin monomers which link via dimerization of the outer ‘D’ domains present on the terminal ends of the fibrin monomer ϒ chains. Under the action of activated factor XIII, there is crosslinking of fibrin. Plasmin fibrinolysis of cross-linked fibrin generates a series of fibrin degradation products (FDPs) of varying molecular weights, including D-dimers. These D-dimers are composed of two ‘D’ domains and one ‘E’ domain retaining the ϒ-ϒ cross-link and creating an unique signature that can be used as a biomarker (Fig 1). An issue with measuring D-dimer is the lack of standardisation in testing due to the number of commercial assays available, which utilise varying techniques and report differing cut-offs and units of measurement. Although the D-dimer is a degradation product of fibrin and arises in patients with VTE, there is a risk of false positive results observed in patients with malignancy, trauma, recent surgery, pregnancy and old age. Therefore, when deciding on a D-dimer assay as part of a VTE diagnostic strategy, it is important to choose one that has been validated and has a high degree of sensitivity to reduce the risk of false negative results. As the D-dimer rises with age, the number of patients with a level below a predetermined cut-off value falls significantly. In a study utilising an enzyme-linked immunosorbent assay D-dimer with a cut-off of 500 ng/ml, D-dimer excluded PE in 60% of patients less than 40 years of age but only 5% of patients over the age of 80 years (Righini et al, 2007). In an attempt to reduce the number of older patients undergoing CTPA an age-adjusted D-dimer for patients over 50 years of age (age × 10 ng/ml) has been prospectively validated. The Adjust-PE study excluded PE and avoided CTPA in 28% of patients with a low pre-test probability score and D-dimer less than 500 ng/ml (Righini et al, 2014). Addition of an age-adjusted D-dimer was able to exclude PE in an additional 12% of patients, with a false negative rate of <0·5% (Righini et al, 2014). More recently a new algorithm – YEARS (Fig 2) has been developed to be more easily applied and to further reduce the number of CTPA performed in patients of all ages. The YEARS algorithm considers only three items from the original Wells’ clinical decision rule, namely – clinical signs of DVT, haemoptysis and whether PE is the most likely diagnosis. In the absence of any of these three items, a D-dimer threshold of ≥1000 ng/ml is required to warrant a CTPA, however, a cut-off of ≥500 ng/ml is set if one or more of the three specified ‘YEARS items’ are present. CTPA was not indicated in 48% of patients with the YEARS algorithm, with a failure rate of 0·61%, compared with 34% of patients if a fixed D-dimer cut off of 500 ng/ml and the Wells’ criteria were utilised (van der Hulle et al, 2017). With the increasing sensitivity of CTPA the identification of filling defects in peripheral subsegmental pulmonary arteries is becoming more common and yet it is not clear if these represent acute thrombus or artefact. The clinical significance of these subsegmental PE (ssPE) and the need to treat with anticoagulation is also debatable (Carrier & Klok, 2017). In this month's journal the utility of the YEARS algorithm is further evaluated (van der Pol et al, 2018). In a post-hoc analysis of the original YEARS study, van der Pol et al (2018)compared the number of patients undergoing CTPA (48%) with those in the Christopher study (68%) an earlier prospective PE outcome study performed in the same Dutch hospitals as the YEARS study (van Belle et al, 2006; van der Hulle et al, 2017). In the YEARS study group PE was identified in 12% (n = 278) of those who underwent imaging, of whom only 28 patients (10% of all PEs diagnosed) were subsegmental. In the Christopher study group 16% of PEs were subsegmental – an absolute difference between the two studies of 6% (95% confidence interval [CI] 1·4–10) with an odds ratio of 0·58 (95% CI 0·24–0·55). These results appear to confirm the previous reports of an association between D-dimer level and location of PE, whereby proximal PE had higher median D-dimer levels when compared to ssPE (Klok et al, 2008). It is likely that age-adjusted D-dimers will be included in the next iteration of guidelines from NICE and the ESC, with the expected benefit being a further reduction in CTPA requests of at least 10%. It will be of interest to see whether or not any of the PE guidelines embrace the new YEARS algorithm, which also reduces the number of CTPA examinations by a similar amount compared to the current standard algorithm. In addition, the lower sensitivity of YEARS for ssPE appears to reduce the number of patients identified with isolated PE without compromising safety. The author declares no competing interests.