Best Evidence Topics
  • Send this BET as an Email
  • Make a Comment on this BET

Do patients with an asymptomatic sub-segmental pulmonary embolism need anticoagulation therapy?

Three Part Question

[In an asymptomatic patient with small subsegmental pulmonary embolism] does [anticoagulation therapy or no treatment] lead to [reduced mortality and a lower rate of recurrent venous thromboembolism?]

Clinical Scenario

A 49 year old male, who was previously fit and well, attended the Emergency Department as a trauma patient after being involved in a road traffic collision. He underwent whole-body computed tomography (CT) scanning and was found to have an incidental, sub-segmental pulmonary embolism (PE). You wonder whether this finding warrants anti-coagulation treatment as is protocol for larger, symptomatic pulmonary emboli, or whether the patient would have a similar outcome if they were to be discharged without treatment.

Search Strategy

Embase (1974 to 2014 June 17): 877 papers

Ovid Medline (1946 to June Week 2 2014): 355 papers

Cochrane Central Register of Controlled Trials (May 2014) & Cochrane Database of Systematic Reviews (2005 to May 2014): 226 papers.
(pulmonary embol$.mp. OR lung embol$.mp. OR pulmonary infarct$.mp. OR lung infarct$.mp. OR OR AND (anticoagula$.mp. OR exp heparin/ OR OR low molecular weight OR exp low molecular weight heparin/ OR OR exp enoxaparin/ OR exp dalteparin/ OR exp tinzaparin/ OR exp nadroparin/ OR exp bemiparin/ OR exp certoparin/ OR exp reviparin/ OR exp parnaparin/ OR exp ardeparin/ OR OR OR OR OR OR OR OR OR OR OR OR exp warfarin/ OR exp antivitamin K/ OR vitamin k OR exp coumarin/ or OR novel oral OR novel OR OR OR exp rivaroxaban/ OR OR exp apixaban/ OR OR exp edoxaban/ OR exp dabigatran etexilate/ OR exp dabigatran/ OR AND ( OR OR OR OR

LIMIT to Human and English Language

LIMIT to “therapy (maximises sensitivity)” or “therapy (maximises specificity)” or “therapy (best balance of sensitivity and specificity)”

Search Outcome

4 papers were relevant to the research question. Prospective Cohort Studies were the best level of evidence available (A Cochrane Review showed that no relevant randomised controlled trials (RCTs) have been completed to date).

Relevant Paper(s)

Author, date and country Patient group Study type (level of evidence) Outcomes Key results Study Weaknesses
Schultz et al
90 Consecutive trauma patients with injury severity score (ISS)>9 without symptoms suggestive of PE or DVT underwent CT chest, pelvis and lower extremities 3-7 days after admission. Exclusion criteria - confirmed DVT, PE or symptoms suggestive of thromboembolic diseases, mechanical ventilation, renal failure, iodine allergy, pregnancy, age<18 2b- Prospective cohort study of trauma patients to determine the incidence of silent PE, and to assess the consequences of withholding AC in patients with mild pulmonary clot burden.Symptomatic PE/recurrent VTE5 out of 22 silent PEs therapeutically treated with AC (4 major clot burden, and 1 minor clot burden with newly diagnosed co-existent DVT) 17/22, all with minor clot burden not treated. 7 lost to follow up. 10 remaining patients remained symptom free for 3 months 40 out of 90 given DVT prophylaxis (heparin or LMWH). 12/40 (30%) developed PE Underpowered with small sample size. The reference diagnostic test (e.g. pulmonary angiography) not utilised although as technology advances, CT's diagnostic accuracy fast improving. DVT prophylaxis protocols not stated. Highly complex (trauma) cases hence not generalisable to other population groups. No mention of inter-observer variability in CT interpretation, hence open to bias. No blinding between reference standard results and clinical details. CT from 1 patient shown as saddle embolus but still classified as asymptomatic on initial recruitment. Could this be an example of selection bias as recruitment was dependent on physicians' clinical judgement? ‘Three quarters of the patients with minor clot burden had subsegmental emboli’. No true figure given and not known whether it was the majority of these patients who were lost to follow-up.
MortalityNo reported deaths
Engelke et al
Consecutive chest CT scans of 1966 patients (mean age 60, range 15-96 years old) taken in the tertiary referral cancer centre from Oct 2001 to Sep 2002. 117 MDCTs positive for PE, of which 96 had complete clinical data. Of these 5 received thrombolytic therapy, 44 therapeutic AC, 21 prophylactic dose AC and 26 no treatment 2b - retrospective analysis of CT scans to assess the outcome of patients in whom PE was missed on reporting and hence not treated with AC. CT scans were reviewed by two radiologists to assess PE diagnosis and PE severity score using a validated radiographic method, while being blinded to clinical details; a further radiologist was required to diagnose 8 of the 117 PEs. MortalityPatients with prophylactic dose AC or no AC therapy had a better 30 day outcome (1 death from cardiac and renal failure) than those receiving therapeutic AC or thrombolysis (7 deaths - 5 from PEs, 2 from Intracranial Haemorrhage) P= 0.37 (relative risk 0.14; 95% CI 0.02-0.87). Of 43 patients with false-negative CTs, no death within 30days (P=0.07) Positive predictors of early death (within 30days)- PE severity score>28, use of systemic thrombolytic therapy, major haemorrhage, new onset cardiac or renal failure (P=0.001-0.043) Predictors of late death (30days to 1 year)- old age, malignancy, new onset renal failure (P= 0.001-0.043) Retrospective study design with small sample size. Baseline clinical characteristics of cohort and control groups would have been helpful in estimating pretest probability of PE since patients that gave a higher clinical suspicion of PE were more likely to be investigated and consequently have a PE (PE was seen in 27/38 patients with clinical suspicion of PE vs 25/38 patients without clinical suspicion of PE; P=0.025); this likely guided clinicians' decision on treatment options. Likewise, good short term prognosis in no treatment group could be explained by the relatively low rate of severe underlying disorders. Contrast enhancement was only adequate down to sub-segmental branches in 55 patients leading to further possible false negative scans. Fails to mention how many were successfully followed-up. Focuses mainly on short term prognosis, although it works out late predictors of death from multivariate analyses. It states briefly that there were 25 deaths after 30 days but fails to mention how these data break down across different treatment groups.
Recurrent PEPE recurrence was more common in those with clinically suspected PE diagnoses (n=2, P=0.002)
Safety (major and minor bleeding)Bleeding complications more frequent with therapeutic AC/thrombolysis (two deaths within 30 days from ICH, five major nonfatal haemorrhages). No AC/ prophylactic AC (one minor haemorrhage only; P-0.037
Sun et al,
67 patients with suspected PE and 113 patients with unsuspected PE (180 total) from a lung cancer cohort of 8014 patients. Anticoagulation therapy was administered to 99% (66/67) of suspected PE and 45% (51/113) of unsuspected PE. AC therapy median duration was 3.2 months and was either warfarin (n=100), unfractionated heparin (n=12) or LMWH (n=5) in the majority. 2b - Retrospective Cohort Analysis into the significance of anticoagulation for unsuspected PE in patients with lung cancerSurvival comparison between those with suspected and unsuspected PE6 patients in the suspected PE group died of PE. Overall survival was 4.2 (CI=2.6-5.8) months vs 9.3 (CI=5.3-9.1) months in patients with suspected PE and unsuspected PE respectively (p=0.001). This remained statistically significant after adjustment for other prognostic factors (Hazard Ratio=1.7; CI=1.1-2.6); this was also despite significantly less anticoagulation in the unsuspected PE group. INR control not discussed – was Warfarin therapeutic for an acceptable amount of time and did they receive sufficient heparin/LMWH prior to stabilisation. Results are not generalisable to all patients due to cancer patients having hypercoaguable states. The term unsuspected PE, unlike in some other studies, does not correlate with size or distribution of the PE and therefore cannot be translated to patients with a sub-segmental PE for example. Unknown time-frame from when PE occurred to its detection on routine CT scan in these patients. Patients with DVT were not excluded – this patient cohort would already receive AC therapy and are therefore irrelevant to the analysis. All patients in this study had segmental, main or lobar artery PE rather than subsegmental PE. As such this study does not directly address the three-part question. This study has been tabulated given the absence of higher level evidence.
Survival comparison between patients with or without AC therapy with an unsuspected PEPatients who were not treated had a 4.1-fold increase in mortality compared with those receiving anticoagulation therapy (CI=2.3-7.6)
Yoo et al,
Adults diagnosed with symptomatic isolated or asymptomatic incidental sub-segmental PE via CTPA, MDCT or pulmonary angiography. SSPE defined as clots located beyond the 5th order pulmonary arteries. Patients with segmental PE were excluded due to need for treatment. Cochrane Review of RCTs looking at anticoagulation vs no AC therapy in patients with sub-segmental PE. Planned to assess study quality using the following: random sequence generation, allocation concealment, blinding, incomplete outcome data, selective reporting and other bias. 3 month thrombo-embolic recurrence Safety outcome - major bleeding defined as fatal or clinically overt Secondary outcomes were: 6-month VTE risk, minor bleeding, all-cause mortality and compliance with treatment 481 records identified either via database searches, or the World Health Organisation international clinical trials registry platform.

No RCTs meeting the inclusion criteria were found (1 retrospective study, 1 cohort study and 1 case series was excluded).
Highlights the need for a blinded, randomised-controlled trial to be under taken to assess the effectiveness and safety of AC versus no AC therapy in patients with isolated subsegmental PE, whether incidentally found or not.


Small sub-segmental PE, once the nemesis of CT pulmonary angiography (CTPA), can now be detected with confidence. A systematic review on the diagnostic accuracy of CTPA for PE by Hogg et al showed a pooled sensitivity of 0.89 (95% CI 0.83-0.95) and a specificity of 0.95 (95% CI 0.91-0.98) in evaluating all vessels through to the sub-segmental level. The gold standard test, pulmonary angiography, is also difficult to interpret at the sub-segmental level, with many inter-observer variations reported across studies. Stein et al noted that approximately 5% of inpatients and 0.6% of outpatients had incidental emboli diagnosed on contrast-enhanced single-detector row helical CT performed for other reasons. This incidence is likely to be higher now due to recent advances in technology with the new generations of multi-detector CT (MDCT). All above studies discussed used either death and/or recurrence of venous thromboembolism as primary end points. One of the difficulties in interpreting data with this methodological approach derives from the fact that data for true PE incidence are inconsistent. The post-mortem examination remains the most reliable method of diagnosis and yet death as a direct outcome of PE is a relatively infrequent phenomenon. As patients with PE often have multi-systemic diseases, determining that PE is the sole cause of death will be contentious and misleading. In all of the above studies reviewed, the clinical detection of PE recurrence was dependent on the discretion of the attending clinicians, and no standardised definition of a validated diagnostic test was used. Likewise, none of the studies explicitly stated if the follow up monitoring was conducted in a blinded manner. This lack of blinding between reference standard results and clinical details during follow-up, combined with the diagnosis of PE recurrence heavily relying on the clinical judgement of clinicians, means that outcomes need to be interpreted with caution. Historically, patients have received anticoagulation once a diagnosis of PE is made, incidental, sub-segmental or otherwise. This review alone may be unlikely to alter this traditional clinical practice, as there is no level 1 evidence yet for the treatment of sub-segmental PE in asymptomatic patients. It is disappointing that most studies focussed solely on mortality rates and recurrence of venous thromboembolism (VTE) as primary end points while failing to address any potential side effects of anticoagulation treatment. Only one retrospective study measured bleeding complications in anti-coagulated patients but here patients who had received thrombolytic therapy were also included, potentially skewing results. Future studies need to address treatment complications as this is a valid clinical concern. It is unlikely that ethics committees will approve trials of anticoagulation treatment versus placebos in symptomatic patients, but the clinical justification of anti-coagulating asymptomatic patients with sub-segmental PE, with its potential side effects of haemorrhage needs to be addressed in a trial setting. The urgency of this task will become even more relevant as diagnostic technology continues to evolve. It can no longer afford to remain within the boundary of an academic interest alone.

Editor Comment

Abbreviations: LMWH, low molecular weight heparin; MDCT, multi-detector CT scan; AC, anticoagulation

Clinical Bottom Line

Some data suggest that patients with clinically unsuspected PE may have better prognostic outcomes than those with symptomatic presentations, especially if the PE is at the sub-segmental level. However, the only direct comparison of anti-coagulation vs no anti-coagulation in patients with an asymptomatic, unsuspected PE suggests a survival benefit from anti-coagulation. Two of the papers included here look at trauma and cancer patients; such results cannot be gerneralised to all patients. Consequently clinicians will have to continue to use their clinical judgment in this patient group until level 1 evidence can be provided and potentially modernise current practice


  1. Schultz DJ, Brasel KJ, Washington L, et al. Incidence of asymptomatic pulmonary embolism in moderately to severely injured trauma patients. J Trauma 2004; 56: 727-733,
  2. Engelke C, Rummeny EJ, Marten K. Pulmonary embolism at multi-detector row CT of chest: One year survival of treated and untreated patients. Radiology 2006; 239: 563- 575.
  3. Sun JM, Kim TS, Lee J et al. Unsuspected pulmonary emboli in lung cancer patients: The impact on survival and the significance of anticoagulation therapy. J lungcan 2010;69(3):330–6.
  4. Yoo HHB, Queluz THAT, El Dib R. Anticoagulant treatment for subsegmental pulmonary embolism. Cochrane Database of Systematic Reviews. 2014; Issue 4. Art. No.: CD010222. DOI: 10.1002/14651858.CD010222.pub2.