Modern computed tomographic pulmonary angiography (CTPA) has excellent resolution and frequently detects distal, subsegmental pulmonary embolisms (SSPEs). These can be challenging to differentiate from artifact, and interobserver reliability is poor.1 Furthermore, the clinical significance of an isolated SSPE is questionable.2,3 For such cases, we determined how often clinicians opted for clinical surveillance over therapeutic anticoagulation in our center.
We retrospectively reviewed all CTPAs (n = 1408) at our tertiary care hospital from 2014 to 2016 as previously described.4 Studies with positive results (n = 223) were classified as proximal, lobar, segmental, or subsegmental based on the most proximal emboli. One scan contained only chronic pulmonary embolisms and was excluded. Doppler ultrasounds performed within 2 weeks of CTPA were adjudicated for deep vein thrombosis. We manually reviewed new prescriptions for anticoagulation and patient medical records for active neoplasm, emergency department visits or admission, changes in hemoglobin level, receipt of blood transfusion, and in-center mortality. Comparisons of proportions used χ2 test.
The McGill University Health Centre Research Ethics Board approved this study. Informed consent was waived due to the retrospective nature of the study.
Of 222 pulmonary embolisms adjudicated, 79 (36%) were subsegmental. Thirty-two of 79 (41%) had a Doppler ultrasound performed, and 8 (25%) were positive for deep vein thrombosis. Of the remaining 71 presumed isolated SSPEs, 39 (55%) were diagnosed in the emergency department. Sixty-two of 71 (87%) were systemically anticoagulated compared with 135 of 143 (94%) with more proximal embolisms (P = .07) (Table 1). Among the 9 patients with SSPE who did not undergo anticoagulation, the major determinants were bleeding at diagnosis or poor prognosis (Table 2).
Adverse events were common in both groups. During the 3 months following the initiation of anticoagulation for isolated SSPE, 26 patients (42%; 95% CI, 30%-55%) had emergency department visits or were readmitted for reasons unrelated to venous thromboembolism (VTE); 21 (34%; 95% CI, 22%-47%) had a decrease in hemoglobin level of 2 g/dL or greater (to convert to grams per liter, multiply by 10.0) and/or received a blood transfusion; and 10 (16%; 95% CI, 8%-28%) died. None of the deaths were related to VTE.
In our center, almost all isolated SSPEs were anticoagulated with a similar frequency to more proximal embolisms. Treatment of SSPE was associated with harm, and patients received anticoagulation for potentially insignificant emboli, often in the presence of terminal illness.
While our study was single center and retrospective, it illustrates that overtreatment of SSPE is common, and associated with harm. Owing to fear of complications, physicians are likely reticent to leave a pulmonary embolism untreated, even if it may represent imaging artifact or is discovered incidentally. The 2016 CHEST guidelines emphasize clinical surveillance over treatment in many cases of isolated SSPE and conclude that physicians are more likely to opt for clinical surveillance in patients with good cardiopulmonary reserve or a high risk of bleeding.5(p339) Practically, the likelihood of a true-positive study result increases with a higher pretest probability, adequate opacification of the arteries, high-quality images, and the presence of multiple filling defects. The decision to pursue clinical surveillance over anticoagulation should incorporate the following factors: the presence of other emboli, risks of therapy, a patient’s pulmonary reserve, hospitalization status, and their preference. Importantly, for isolated SSPE, clinical surveillance has not been associated with an increased recurrence of VTE over 3 months.3 For those with active cancer, an alternative approach that merits study is the use of prophylactic dose anticoagulation to balance this population’s higher risk of recurrence with the harms of anticoagulation.
Given that the majority of patients with SSPE at our center were treated, we identified an opportunity to intervene. Pending definitive studies, we are implementing a strategy to improve the pretest probability of patients undergoing CTPA4 in addition to facilitating a pathway of care that offers patients at low risk of VTE recurrence or with a limited life expectancy the option of clinical surveillance.
Accepted for Publication: May 4, 2018.
Corresponding Author: Emily G. McDonald, MD, MSc, Royal Victoria Hospital, 1001 Decarie Blvd, Room D5.5843, Montréal, QC H4A 3J1, Canada (emily.mcdonald@mcgill.ca).
Published Online: July 30, 2018. doi:10.1001/jamainternmed.2018.2971
Author Contributions: Drs Lee and McDonald had full access to all of the data in the study and take responsibility for the integrity of the data and the accuracy of the data analysis. Drs Lee and McDonald share equal credit for this work.
Study concept and design: Gallix, Lee, McDonald.
Acquisition, analysis, or interpretation of data: All authors.
Drafting of the manuscript: Raslan, McDonald.
Critical revision of the manuscript for important intellectual content: Chong, Gallix, Lee, McDonald.
Statistical analysis: Raslan.
Obtained funding: McDonald.
Administrative, technical, or material support: Chong, Gallix, Lee, McDonald.
Study supervision: Gallix, Lee, McDonald.
Conflict of Interest Disclosures: None reported.
1.Pena
E, Kimpton
M, Dennie
C, Peterson
R, LE Gal
G, Carrier
M. Difference in interpretation of computed tomography pulmonary angiography diagnosis of subsegmental thrombosis in patients with suspected pulmonary embolism.
J Thromb Haemost. 2012;10(3):496-498. doi:
10.1111/j.1538-7836.2011.04612.xPubMedGoogle ScholarCrossref