A multicentre observational study of the prevalence, management, and outcomes of subsegmental pulmonary embolism

This multicentre observational study investigated the prevalence, risk factors, additional radiological findings, and outcomes of patients with SSPE over six years across three UK hospitals. Of 14,300 CTPA reports screened, 14.4% had a diagnosis of PE and of these 3.8% had a diagnosis of SSPE. There was no significant difference found across all groups in mortality, VTE recurrence or major bleeding at 3 months. This is the first SSPE study to include propensity-matched patients with more proximal or no PE diagnoses. This is important for comparing outcomes across PE groups that can vary by other factors (i.e., age, sex).

The prevalence of SSPE in observational studies has been reported to be as high as 13.5%, whereas our current study prevalence of 3.8% is similar to the pooled estimate of 4.6% (95% confidence interval 1.8-8.5%) in a meta-analysis [9, 27]. The 3-month SSPE outcomes in the current study were 0% for VTE recurrence, 3.8% for major bleeding and 1% for mortality. The uniformly low outcome frequency across all groups limited comparisons between SSPE and more proximal or no PE patients. A recent study by Ceccato et al., also reported low 3-month outcomes with no cases of VTE recurrence in either anticoagulant treated or untreated SSPE patients [12]. This compares with 5.3% VTE recurrence, 8.1% bleeding, and 2.1% mortality for the anticoagulation treated SSPE patients in the Bariteau et al., meta-analysis [10]. However, the studies included in this meta-analysis were heterogenous with many lacking standardised bleeding definitions and utilised a combination of major bleeding and CRNMB that may explain the discordance in bleeding rates. Our study primarily used secondary care hospital records and therefore we were unable to report on the CRNMB episodes.

As the majority (96.2%) of SSPE patients were treated with anticoagulation there were insufficient untreated patients to make comparisons between the two groups. The American College of Chest Physicians (ACCP) guidelines on antithrombotic therapy for VTE disease recommend that SSPE can be treated with anticoagulation or no-anticoagulation (surveillance) depending on the patient’s VTE risk profile [31]. If a no-anticoagulation strategy is considered, then bilateral lower limb ultrasound (or alternative imaging) should be performed, as the presence of proximal DVT would necessitate anticoagulation. Previous studies have either not reported the number of patients investigated for DVT [32], or reported incomplete rates of investigation [7]. Fernández-Capitán et al. report that 55% of patients with SSPE did not have a concurrent DVT, however not all patients in this study were investigated for DVT [7]. Interestingly, they also found that the VTE-recurrence rates were similar in SSPE patients with and without DVT.

International surveys of thrombosis experts have highlighted the clinical equipoise in SSPE management and to date no RCTs have compared anticoagulation to surveillance [17, 18]. The use of anticoagulation in VTE is not without risk and treatment with DOACs is associated with major bleeding in 1.1% and CRNMB in 6.6% of patients during pooled follow-up of 3–6 months in a meta-analysis [33]. The risk of bleeding in anticoagulant treated SSPE has been reported as 8.1%, and in other instances of untreated VTE it is 0.6% per patient year [27, 30]. In a recent large prospective cohort study of patients with untreated SSPE by Le Gal et al., the 90 day risk of major bleeding was found to be 0.7%, and minor bleeding 1.4% [19]. Unsurprisingly, the risk of major bleeding is higher in our study (3.8%) where most patients with SSPE were treated with anticoagulation. The reported rate of VTE recurrence was higher in the Le Gal study compared to the present study, with 3.1% recurrence within 90 days compared to 0% respectively. This may reflect more standardised monitoring and recording of VTE events during the follow-up period compared to a retrospective cohort. Despite the difference in anticoagulation treatment the mortality rates for both studies were similar, at 1.4% in the Le Gal study and 1.3% in the current study. It remains unclear if anticoagulation in SSPE results in a net-benefit of improved VTE recurrence compared with the bleeding risks. There are ongoing multicentre SSPE RCTs in the UK (STOP-APE; stopping anticoagulation for isolated SSPE [NCT04727437]), and Canada / Switzerland and the Netherlands (NCT04263038; Surveillance vs. Anticoagulation For low-risk patiEnts with isolated SubSegmental Pulmonary Embolism) [34, 35] to address this important question.

The additional radiological findings were highest (87.3%) in the SSPE group, with the most common diagnoses being pleural effusion, consolidation, and cardiomegaly. Whilst consolidation and cardiomegaly could be a consequence of lung infarction and right heart strain associated with PE, this is less likely with smaller embolic loads. Therefore, the symptoms that precipitated investigation for PE may have been driven by alternative pathology and the SSPEs may have been contributory or incidental findings. The majority of CTPAs in the SSPE group were considered technically adequate by the reporting radiologists, which is important as SSPE can be challenging to diagnose with greater false positives than more proximal PEs [36]. This could result in overtreatment of non-existent SSPEs and the potential harm from anticoagulant therapy. The VTE guidelines from the ACCP proposed clinical and diagnostic criteria that may make a diagnosis of SSPE more likely however, this was based on expert consensus and is yet to be validated in clinical trials [31]. One of these criteria is the presence of PE related symptoms, however in our current study there is a high incidence of additional radiological findings in the SSPE, and the no PE groups, suggesting some symptoms may be driven by alternative diagnoses. Furthermore, the European Society of Cardiology guidelines on acute pulmonary embolism (2019) advocate discussing single SSPEs with a radiologist and if necessary, seeking a second opinion to confirm the diagnosis [37].

The strengths of the current study are that it (i) included patients across several hospitals over a six-year period, (ii) utilised propensity matching between SSPE and main, lobar, segmental, and no PE groups, (iii) identified cases from CTPA reports which is more robust than clinical coding and, (iv) included standardised definitions of 3-month outcomes from VTE recurrent and major bleeding. There are several limitations to the current study. Firstly, PE and SSPE cases were identified from CTPA reports but were not all reviewed by a thoracic radiologist. There is significant inter-observer variation in the diagnosis of SSPE between general and thoracic radiologists and the number of false positives is reduced by increasing specialisation [38]. However, given the ubiquity and frequency of CTPAs, thoracic radiology review of all scans is often not performed at all institutions and therefore the current study represents the current real-world practice. Secondly, only CTPA reports were utilised for case finding, which did not capture PEs identified by other modalities including ventilation/perfusion (VQ) scans. This is likely to have had a negligible effect on prevalence estimates as CTPA was the first line investigation for PE at the study sites and VQ was not widely performed. However, it is possible that a number of PEs were identified on other CTs such as staging CT scans and these have not been included in this study. Thirdly, propensity matching for other factors that may have influenced PE outcomes including active cancer was not possible as this information was only available following case note reviews. Fourthly, there may have been further factors that influenced PE outcomes in addition to embolus anatomical location including the embolic load, degree of right heart strain, presence of concurrent acute illness, and level of pre-existing physiological reserve. Finally, the use of electronic patient records alone for follow-up data may have underestimated some outcomes that are not immediately recorded in hospital records. One instance of this may be reflected in mortality which is lower than expected for more central PEs. Alternatively, this may reflect lower mortality for patients with PE in more modern cohorts.

In conclusion, this multicentre observational study screened 14,300 CTPA reports to identify 2,055 (15%) PEs, of which 79 (3.8%) were SSPEs. There were a substantial number of additional radiological findings in the SSPE group that may have accounted for their symptoms and contributes to the challenge of balancing the risks and benefits of anticoagulation. There remains clinical equipoise in the management of SSPE that will be addressed by ongoing RCTs.