Loading...
The widespread use of computed tomography (CT) pulmonary angiography has facilitated the rapid diagnosis of pulmonary embolism (PE), but sometimes the only CT abnormality is the presence of one or several subsegmental emboli. Are these findings clinically important, and is it necessary to anticoagulate patients with subsegmental emboli? A recent meta-analysis forces us to think carefully about the management of these patients.
The meta-analysis included 14 observational studies involving 15,563 symptomatic patients (NEJM JW Emerg Med Apr 2018 and Acad Emerg Med 2018 Mar 2; [e-pub]). All patients underwent CT pulmonary angiography for suspected PE and had at least 30 days of follow-up to assess rates of recurrent PE or deep venous thrombosis (DVT) and bleeding complications. In the pooled population, the prevalence of subsegmental PE was 4.6%. Among those with subsegmental PE, there were no significant differences between anticoagulated patients and nonanticoagulated patients in incidence of recurrent PE/DVT (5.3% and 3.9%, respectively) or death (2.1% and 3.0%). However, administration or withholding of anticoagulation was not randomized, and the incidence of bleeding was only reported for anticoagulated patients (8.1%). Notably, most of the studies did not provide information regarding the presence of concurrent DVT at the time of PE diagnosis, so we don't know whether these subsegmental emboli were truly isolated or not. The one study that included only patients with isolated subsegmental PE (i.e., without detectable DVT) showed no difference in recurrent PE/DVT between patients who were and were not anticoagulated.
To put these findings in perspective, some background information is worth considering. Prior to 20 years ago, PE was usually diagnosed by ventilation-perfusion lung scanning. The landmark 1990 PIOPED study suggested that patients with subsegmental perfusion defects who were not anticoagulated had benign clinical outcomes (JAMA 1990; 263:2753). In the 1990s, CT pulmonary angiography became available, and a 2011 study found that — although the incidence of diagnosed PE doubled between 1993 and 2006 — mortality associated with PE remained the same, suggesting that we might be overdiagnosing PE and possibly overtreating clinically benign cases (NEJM JW Gen Med Jun 15 2011 and Arch Intern Med 2011; 171:831). In addition, a recent study found that roughly one quarter of CT pulmonary angiograms initially interpreted as positive might actually have been false-positive scans, owing in part to various scanning artifacts; in that study, a majority of false-positive readings involved subsegmental locations. Yet another often-overlooked question is whether some subsegmental “emboli” actually represent in situ thrombosis within the pulmonary artery — as has been postulated, for example, in patients with sickle cell disease (NEJM JW Gen Med Dec 15 2011 and Am J Respir Crit Care Med 2011; 184:1022).
What do guidelines say about anticoagulating patients with subsegmental PE? The 2016 guideline from the American College of Chest Physicians suggests “clinical surveillance over anticoagulation” for patients with subsegmental PE who have no more-proximal pulmonary artery involvement, no proximal leg DVT, and low risk for recurrent venous thromboembolism (NEJM JW Emerg Med Feb 2016 and Chest 2016; 149:315). In contrast, the guideline suggests “anticoagulation over clinical surveillance” for patients deemed at high risk for recurrent venous thromboembolism. The authors acknowledge that the quality of the evidence is low, and that bleeding risk and patient preferences should play a role. Guidelines from the European Society of Cardiology state that decisions to treat patients with isolated subsegmental PE and no proximal DVT “should be made on an individual basis, taking into account the clinical probability [of PE] and the bleeding risk” (Eur Heart J 2014; 35:3033). And finally, just three months ago, the American College of Emergency Physicians issued a “Clinical Policy” on management of PE and DVT that included the following statement: “Given the lack of evidence, anticoagulation treatment decisions for patients with subsegmental PE without associated DVT should be guided by individual patient risk profiles and preferences” (Ann Emerg Med 2018; 71:e59). The authors acknowledge that this is a “Level C” recommendation, based on consensus and not randomized trials.
Clinicians have a strong intuitive desire to anticoagulate all patients with PE; action is often favored over inaction. But treatment is not without risk, and severe intracranial and gastrointestinal bleeding complications are especially regrettable if the original indications for anticoagulation were marginal or uncertain. Potential bleeding complications should obviously be discussed with all PE patients who receive anticoagulation, regardless of PE size and location. However, for patients with isolated subsegmental PE — especially those we believe to be at high risk for bleeding — we should be willing to consider observation rather than treatment (after excluding concurrent DVT). In my view, not anticoagulating is not only as safe as, but perhaps even safer than, anticoagulation in selected patients with isolated subsegmental PE. In these cases, patient preferences should be an important element of the shared decision-making process.