Importance of subsegmental pulmonary embolism

Gary E. Raskob

In this issue of Blood, den Exter et al highlight the importance of subsegmental pulmonary emboli.1

Pulmonary embolism (PE) causes 100 000 or more deaths each year in the United States2 and is the primary diagnosis or a complicating condition in more than 300 000 patient hospitalizations.3

The diagnosis of PE has been revolutionized with the introduction of advanced computed tomographic pulmonary angiography (CTPA).4 The multi-row detector CTPA is highly sensitive and specific for PE, including the relatively smaller emboli confined to subsegmental pulmonary arteries.4 Consequently, the proportion of all emboli diagnosed in symptomatic patients that are confined to subsegmental arteries has increased from 4.7% with the single detector to 15% with multi-row detector CTPA.5 A recent systematic review6 has suggested that this improved detection of subsegmental PE represents overdiagnosis of relatively unimportant emboli for which the risks of anticoagulant therapy may not be warranted.

Now come den Exter and colleagues1 with important new data that strongly challenge this inference. These investigators used the combined data from 2 large prospective cohort studies of patients with clinically suspected PE to compare the thromboembolic risk profiles and clinical outcomes of patients with subsegmental PE by CTPA (116 patients, 15.5%) with those of patients with segmental or more proximal emboli (632 patients) and with those in whom PE was ruled out (2980 patients) on the basis of either an unlikely clinical probability combined with a negative d-dimer result or by negative CTPA. The methodology was strong, and follow-up for 3 months was complete in 99.8% of patients. The most important findings concern the outcomes on follow-up. Among the 116 patients with subsegmental PE, all of whom received anticoagulant therapy, symptomatic recurrent venous thromboembolism occurred in 4 patients (3.6%). In contrast, only 1.1% of the 2980 patients in whom PE was ruled out and who did not receive anticoagulant therapy had symptomatic venous thromboembolism on follow-up (hazard ratio for subsegmental PE, 4.3; 95% confidence interval 1.5-12.3). There was a gradient for mortality, which was highest among patients with segmental or more proximal PE (10.7%), followed by subsegmental PE (6.5%), and lower in patients in whom PE was excluded (5.4%, P = .01 compared with patients with subsegmental PE).1

How do we reconcile these results with the prior systematic review?6 Two facts are relevant. First, the studies that suggest that subsegmental PE does not require anticoagulant therapy all included imaging for deep-vein thrombosis using ultrasonography to detect and treat those patients with this source for recurrent embolism.5 The presence of deep-vein thrombosis is a key prognostic marker for recurrent venous thromboembolism7 and an independent predictor of mortality among patients with PE.8 Imaging for deep-vein thrombosis is therefore “… an important component of the management of subsegmental pulmonary embolism left untreated.”5 Second, den Exter and colleagues1 show that patients with subsegmental PE have persistent risk factors for recurrent venous thromboembolism, and these risk factors occur less commonly among the patients in whom PE was ruled out. The significant prevalence of active malignancy (18%) helps to explain the 3.6% incidence of recurrence despite anticoagulant therapy.1

What are the implications for current clinical practice? A symptomatic patient with confirmed subsegmental PE who does not have an absolute contraindication or risk factor(s) conferring a high risk of bleeding should be treated with anticoagulant therapy unless strong patient preference dictates otherwise. The safety and simplicity of anticoagulant therapy have been improved with the new oral anticoagulants.9,10 This development, together with the data of den Exter et al,1 tips the balance in favor of anticoagulant therapy for most patients. In patients with absolute contraindications or a high risk of bleeding, serial imaging for deep-vein thrombosis of the legs using compression ultrasonography is likely a safe alternative, providing several conditions are met: (1) the patient has adequate cardiopulmonary reserve to tolerate the existing embolus left untreated and/or a small recurrent embolus; (2) that ultrasonography can be done 3 or 4 times over a 10- to 14-day period to detect a new proximal deep-vein thrombosis before it leads to important recurrent PE; (3) that persistent risk factors for recurrent venous thromboembolism are absent; and (4) that alternate sources for recurrent PE are absent (eg, central venous catheterization).7 The ongoing prospective cohort study (NCT 01455818) of withholding anticoagulant therapy in patients with subsegmental PE who have negative results by serial ultrasonography for deep-vein thrombosis is awaited with interest.5 For now, PE, including subsegmental PE, should be considered a serious and unforgiving condition for which the benefits of contemporary anticoagulant therapy outweigh the risks for most patients.


  • Conflict-of-interest disclosure: The author receives, or has received, consultant fees for clinical research or advisory service from Bayer Healthcare, BMS, Daiichi-Sankyo, Janssen Pharmaceuticals, Johnson and Johnson, Pfizer, Quintiles, Sanofi-Aventis, and Takeda Research and Development.