Subsegmental pulmonary embolism: anticoagulation or observation?
As the use of chest CT-angiograms in emergency departments and medical wards has risen by more than tenfold, so has the discovery of small pulmonary emboli of unclear clinical significance. These PEs are often isolated to distal (subsegmental) branches of the pulmonary artery, without concurrent deep venous thrombosis (DVT). Small distal PEs may be incidentally found in an asymptomatic patient; more often, these PEs are found on CTs obtained for dyspnea, but the clots seem too small to cause significant symptoms.
Should anticoagulation be prescribed to patients with subsegmental pulmonary embolism? If so, for how long? The short answer is, no one knows. Here's why.
Overdiagnosis of Subsegmental Pulmonary Embolism
Radiologists often overdiagnose small, distal pulmonary emboli. That is to say, a significant proportion of the subsegmental pulmonary emboli diagnosed and treated with anticoagulation for months or years do not really exist.
Evidence for this lies in the low interobserver agreement between blinded radiologists diagnosing distal clots on scans for suspected PE. (Both intepretations, clot and no clot, cannot be correct.) Subsegmental PEs are often subtle and hard to distinguish from artifact. Add real-world medicolegal defensiveness to the mix, and regular overdiagnosis of non-existent PEs by radiologists must result.
Risks of Subsegmental Pulmonary Embolism
The risk of death or disability directly resulting from an isolated subsegmental pulmonary embolism (or a small number of them occurring together) appears anecdotally to be very low. These clots by themselves rarely cause significant hemodynamic compromise or hypoxemia, if ever. (It's conceivable that patients with significant pre-existing cardiopulmonary compromise could be seriously harmed by isolated subsegmental PE.)
As with the majority of larger PEs (most of which do not cause major cardiac stress), almost all the risk posed by a subsegmental PE derives not from the clot's current effects, but from a hypothetical larger future embolism that might threaten harm in an untreated patient.
Recurrence Risks After Subsegmental Pulmonary Embolism
Unfortunately, there is little evidence to guide the treating physician on the risk of future pulmonary embolism (recurrence) after a subsegmental PE. A 2014 Cochrane review found no randomized trial evidence at all, concluding "we can not draw any conclusions." There was a randomized trial planned ten years ago, but it looks defunct.
Larger, unprovoked, clinically significant PEs do frequently recur in the absence of anticoagulation, and tend to do so in roughly similar severity to the initial presentation. In contrast, the natural history of subsegmental pulmonary emboli is not known. If subsegmental PEs followed the same pattern of reproducible severity as do larger PEs, recurrent subsegmental PEs would tend to be lower risk, or asymptomatic. But this has not been studied.
Models and calculators to predict venous thromboembolism recurrence risk and guide anticoagulation discontinuation generally after PE have been proposed, but none has been validated in a large prospective trial.
Subsegmental PEs: Limited Evidence on Outcomes
In a meta-analysis, authors examined the ~750 patients with subsegemental PEs among ~15,000 patients enrolled in 14 randomized trials for PE. The large majority (81%) were then treated with anticoagulation. Over 90 days, there was no difference in recurrent PE among those treated (5%) and those not (4%); 90-day mortality was 3% vs 2%. Among treated patients, 8% had bleeding.
In an observational study, 82 subsegmental PEs made up 15% of all the PEs found on CT. About half the patients with small PEs were anticoagulated, and two then had life-threatening bleeding (a ~5% risk). No patient had an identified recurrent PE, whether or not they were anticoagulated. Follow-up was through chart review and data extraction only.
In another retrospective observational study, among 25 patients with small distal PEs and 108 with inconclusive scans, none of whom were anticoagulated, none had a recurrent PE on chart review after three months.
These numbers are too small and the retrospective study designs too weak to drive decision making.
Guidelines on Treatment of Subsegmental PE
A leading specialty society advises that patients with subsegmental PE without deep venous thrombosis receive anticoagulation if the risk of recurrence is high, and surveillance if recurrence risk is low. Both recommendations are Grade 2C, i.e., suggestions based on expert opinion rather than good evidence.
This is helpful for its endorsement of observation as an appropriate option for selected patients. However, the recurrence risk for most patients is indeterminate (i.e., middle-aged, without active cancer, no past history of PE, without multiple risk factors), pushing most decisions toward treatment over observation.
How Long to Treat Subsegmental PEs?
Current guidelines suggest that patients with pulmonary embolism not provoked by transient risk factors (surgery, immobility, etc.) should be considered for "indefinite" periods of anticoagulation, i.e., lifelong.
Should patients with unprovoked isolated subsegmental PEs be anticoagulated for decades, then? Probably not, but no expert or group has yet gone on record advising shorter courses of anticoagulation for subsegmental PEs.
Risks of Treating Subsegmental PEs
Patients receiving full dose anticoagulation with warfarin were historically considered to have a rate of major bleeding of about 3% per year, with catastrophic intracranial hemorrhage as frequently as 1% per year.
Bleeding risks with newer generation direct oral anticoagulants appear to be lower. However, the risk of significant GI bleeding is still likely at least 1% per patient, per year, and intracranial hemorrhage still regularly occurs.
Among 14,000 patients taking a leading DOAC for atrial fibrillation, 1 in 200 per year had intracranial hemorrhage. The rates of ICH were lower in clinical trials testing DOACs for DVT-PE compared to atrial fibrillation; fewer patients were enrolled in the DVT-PE trials and they tended to be younger. The risk of intracranial hemorrhage occurs disproportionately in the elderly: rates in young people are well below the pooled rate, while rates in the very old may be significantly higher than the observed average.
Combining a DOAC with aspirin or an antiplatelet drug doubles the risk of brain bleeds. ICH appear to occur less often, and are less severe, with DOACs compared to warfarin.
When DOACs or warfarin are taken over years, cumulative bleeding risks are substantial, both from the perspective of the individual patient and in terms of population health. Placing patients at these levels of risk for extended periods after clinically insignificant (or even non-existent) pulmonary emboli may someday prove to have been unwise.
Additional Testing for Small PEs
Obtaining more raw information -- by repeating CT scanning after a subsegmental PE, or performing d-dimer testing, or adding tests such as ventilation-perfusion scan or echocardiogram -- may tip the clinician toward treating or not treating with anticoagulation.
None of these approaches has been validated, and there is no reason to believe that the additional data obtained provides any truth about the future risk of significant pulmonary embolism. Likewise, hypercoagulable workup should not be expected to provide useful information, because of the well-described pitfalls of such testing and its interpretation.
Leading experts suggest that patients with a low risk of recurrence should not receive anticoagulation after a subsegmental PE without a DVT. Unfortunately, at this time there is no reliable method to predict which patients are at low recurrence risk and should forego anticoagulation.
Until additional guidance arrives from future randomized or prospective observational trials (none are listed on clinicaltrials.gov) or statements by experts, physicians and patients must make their own best guesses regarding anticoagulation or observation after small distal pulmonary embolism.
Most times, this results in anticoagulation, for months or years. This approach may save lives by preventing recurrent PE, although this has never been demonstrated. Almost certainly, it causes avoidable bleeding in some number of patients who are not at significant risk for recurrent PE.
We have always anticoagulated most patients with unprovoked PE either for too long, or not long enough for the individual patient; that's unavoidable given the current limits of our predictive abilities. Our decision-making for incidental, subtle subsegmental PE exposes and amplifies this weakness, occasionally leading us toward an absurd extreme: anticoagulation for life for a clot that never existed. One hopes severe bleeding resulting from such cases is a rare occurrence -- but how would we know?