Search by:

Source

Bouzrara S, Beyda R. Safety of treating low‐risk patients with acute pulmonary embolism at home. Academic Emergency Medicine. Published online May 25, 2025:acem.70044.

Study Population: 2694 low-risk patients with acute PE (selected by a validated risk stratification tool) discharged within 24 h of diagnosis for treatment at home

Efficacy Endpoints

All-cause mortality, recurrent VTE, and major bleeding events

Harm Endpoints

NA

Narrative

Discharging patients diagnosed with acute pulmonary embolism (PE) for treatment at home is a desirable disposition for hemodynamically stable patients who are deemed low risk for adverse outcomes. Outpatient PE treatment may potentially decrease resource utilization and cost of care by avoiding hospitalization. It could also increase patient satisfaction.1^, 2^, 3 Candidates for outpatient treatment are typically identified using clinical triage tools such as the PE Severity Index (PESI) score, simplified PESI (sPESI) score, and the Hestia rule. This approach has been associated with low adverse event rates.4 Applying the Hestia rule also allows the inclusion of hemodynamically stable underrepresented subgroups. Prior research included trials with small sample sizes that were conducted in single centers and underrepresented subgroups with serious comorbidities (e.g., cancer, cardiopulmonary disease, or higher risk PE). This systematic review and individual patient level meta-analysis discussed here was conducted to estimate all-cause mortality and adverse event rates in patients identified as low-risk acute PE by validated clinical risk stratification tools discharged for home treatment and summarizes randomized controlled trials and cohort studies aimed at answering this question.5 Outpatient treatment was defined as discharge within 24 h of PE diagnosis, randomization, or from ED arrival.

Eligible studies included in the analysis evaluated various algorithms for identifying acute PE patients, identified via computed tomography pulmonary angiography or high-probability ventilation/perfusion scans, appropriate for home treatment. Additionally, studies had to report at least one predefined outcome and include a minimum of 50 patients. The safety of home treatment was determined using risk stratification tools, including Hestia, PESI, or (s)PESI scores.5 Studies were excluded if they enrolled patients diagnosed with PE more than 48 h after hospitalization or if they exclusively included patients with cancer or unrelated medical conditions.

The primary outcomes assessed included 14-day and 30-day incidence of all-cause mortality and adverse events comprising a combined endpoint of all-cause mortality, recurrent venous thromboembolism (VTE), and/or major bleeding. Outcomes were also evaluated by calculating the relative risk (RR) of incidence of all-cause mortality and adverse events at 14 and 30 days in higher risk groups with preexisting cardiopulmonary comorbidities, abnormal troponin, abnormal (N-terminal pro-) B-type natriuretic peptide, right ventricular dysfunction, decreased kidney function, and/or cancer. Differences based on which applied clinical triage tool was utilized in combination with clinical judgment were also assessed.

For the patients who were discharged home, the 14-day incidence of all-cause mortality was 0.11% (95% confidence interval [CI] 0.0%–0.24%); the incidence of recurrent VTE was 0.34% (95% CI 0.12–0.56); and the incidence of major bleeding was 0.19% (95% CI 0.03–0.35). The composite outcome of adverse events at 14 days was 0.56% (95% CI 0.28–0.84).5 At 30 days, the incidence of all-cause mortality was 0.30% (95% CI 0.09%–0.51%), the incidence of recurrent VTE was 0.57% (95% CI 0.28–0.86), and the incidence of major bleeding was 0.45% (95% CI 0.19–1.71).5 The composite outcome of adverse events at 30 days was 1.2% (95% CI 0.79%–1.6%). The subgroup of patients with cancer, however, experienced an increased risk of 30-day mortality (RR 4.9, 95% CI 2.7–9.1).5

By using the Hestia rule, the systematic review collected patient level data to assess the impact of medical comorbidities as well as indicators of PE severity, e.g., laboratory parameters and/or imaging findings suggestive of right ventricular strain. Using these data, RRs for adverse events and event rates were calculated for subgroups based on the presence or absence of predefined subgroup characteristics. When compared to those without comorbidities, the 14-day incidences of the composite adverse event rate in the higher risk patient subgroups were higher for patients with preexisting cardiopulmonary comorbidity (RR 3.5, 95% CI 1.5–7.9), for those with abnormal troponin (RR 2.5, 95% CI 1.3–4.9), patients with abnormal NT-proBNP (RR 3.9, 95% CI 1.6–9.8), but lower in patients with decreased renal function (RR 0.47, 95% CI 0.22–1.0).5 At 30 days, the incidences of the composite adverse event rate were also higher in patients with abnormal troponin (RR 2.9, 95% CI 1.5–5.7), in patients with abnormal NT-proBNP (RR 3.3, 95% CI 1.6–7.1), and in those with cancer (RR 2.7, 95% CI 1.4–5.2) but lower in patients with decreased kidney function (RR 0.35, 95% CI 0.14–0.88).5 Patients with severe impairment of renal function (GFR less than 30 mL/min) were excluded, however, from most studies and there was a low sample size of this subset of patients. Furthermore, patients with mild to moderate renal impairment who meet low-risk criteria by the (s)PESI or Hestia scores may not necessarily be at higher risk solely based on renal function parameters.5 In patients with comorbidities, it is also important to consider the absolute risk of adverse events. The 30-day adverse events rate exceeded 2.5%, while the mortality risk ranged between 0.4% and 0.6%.5 In the subgroups, the combined endpoint event rate in patients with cancer was 2.4%, previous VTE 1.6%, decreased kidney function 0.49%, preexisting cardiopulmonary disease 1.8%, abnormal troponin 2.6%, abnormal (NT-pro) BNP 2.9%, and signs of right ventricular overload of 2.7%. The all-cause mortality event rate in patients with cancer was 1.9%, previous VTE 1.8%, decreased kidney function 0.18%, preexisting cardiopulmonary disease 0.34%, abnormal troponin and abnormal (NT-pro) BNP 0.4%, and signs of right ventricular overload 0.55%.

There was no significant difference in all-cause mortality between patients selected by Hestia or (s)PESI.5 However, in comparison to (s)PESI, patients selected by the Hestia rule had a higher incidence of recurrent VTE of 0.52% (95% CI 0.17%–0.87%) versus 0.11% (95% CI 0.0%–0.41%) in 14 days and 0.80% (95% CI 0.36%–1.20%) versus 0.43% (95% CI 0.0%–1.00%) in 30 days. Major bleeding was also higher with an incidence of 0.35% (95% CI 0.06%–0.64%) versus 0.0% (95% CI 0.0%–0.0%) at 14 days and 0.62% (95% CI 0.24%–1.00%) versus 0.43% (95% CI 0.0%–1.00%) at 30 days.

The risk of bleeding events was higher in patients using vitamin K antagonists (VKAs) compared to direct oral anticoagulants (DOACs).6 Overall, 40% of patients were treated with VKAs and the incidence of major bleeding at 14 days was 0.30% in the VKA group compared to 0.13% for those treated with a DOAC.5

Caveats

The systematic review and meta-analysis discussed here provided strong evidence in support of the safety of home treatment for patients with acute PE identified as low risk by a validated risk stratification tool and supported by clinical judgment. It further investigated subgroups that were underrepresented in the literature providing further insight concerning possible risks to consider when managing these patients and shared decision making. Outpatient treatment of low-risk acute PE patients has the potential to decrease the cost of care and hospital length of stay, improve patient satisfaction, and decrease overall medical resource utilization.5 Furthermore, these potential benefits are present with a low risk of increased mortality and adverse event rates. In patient subgroups with higher risk of all-cause mortality, e.g., patients with cancer, or with higher rates of adverse events, e.g., patients with chronic cardiopulmonary disease, and/or with elevation of cardiac biomarkers, it is unclear if hospitalization can further decrease this risk. Special consideration should be given in discharging these patients based on the adverse event rates discussed above. These clinical variables may serve to guide discussions during shared decision making with patients regarding their disposition.

There are some limitations to this systematic review. First, the small number of events in subgroups resulted in wide 95% CIs which limits the precision of risk assessments. Second, the broad range of missing data (1%–62%) in some of the studies forced the imputation of variables when a study was included in the review. This level of missing data could impact the reliability and validity of the results. Hence, a sensitivity analysis was performed on the nonimputed data to ensure its robustness but excluded an analysis on the missing data and how randomly it was choses. Therefore, careful considerations should be given when deciding on discharging patients with comorbidities. Third, there was a lack of subgroup standardization in some studies and how patients with specific comorbidities, biomarker levels, or cancer status were defined or categorized in the original studies. Fourth, the data only included adverse event rates and did not mention other important outcomes such as unscheduled visits, patients’ satisfaction, quality of life, or cost-effectiveness. Fifth, some of the studies used in the systematic review excluded subgroups that could have resulted in an underestimation of the prognostic impact. This is particularly relevant to the question of the prognostic value of right ventricular dysfunction/overload and cardiac biomarkers and the ability of these variables to predict mortality and/or adverse events. Most notably, the highest proportion of missing data was from echocardiographic-assessed right ventricular dysfunction limiting the applicability of safe discharge in subgroup patients with comorbidities. Lastly, the systematic review did not conduct a sensitivity analysis specifically separating randomized controlled trials from cohort studies. These considerations may be factored into the disposition decision. The incidences of adverse events and mortality were evaluated at 14 and 30 days. While these time frames are relevant for acute outcomes, they may not capture longer-term complications or recurrences of PE, which could influence the overall assessment of home treatment safety. In addition to the aforementioned limitations, one must consider the methodological heterogeneity introduced by including both randomized controlled trials and observational studies and the potential for selection bias. Finally, using a definition of discharge within 24 h allows for reasonable translation of these data into clinical practice variation ranging from discharge directly from the ED and up to brief periods of observation and/or short hospital stays less than 24 h based on available resources and shared decision making and in consideration of patient-specific psychosocial factors.

Current guidelines in the management of PE utilize clinical algorithms to risk stratify patients. (s)PESI estimates the 30-day mortality while Hestia estimates the safety of discharging patients home for continued outpatient treatment. In this systematic review, both tools performed very well at identifying low-risk patients for outpatient treatment, although patients stratified using the (s)PESI tool were at slightly lower risk of adverse events as defined by recurrent VTE and major bleeding.5 In addition, incorporating these risk stratification tools with clinical judgment and the absence of right ventricular dysfunction/overload can reconcile some of the limitations of the data presented in this analysis.

Given these results and the low adverse event rates in low-risk PE patients triaged using the validated clinical criteria defined by the (s)PESI or Hestia scores, combined with clinical judgment, and in the absence of right ventricular dysfunction/overload, discharging carefully selected low-risk patients with home treatment is a reasonable disposition consideration. This approach is supported by the European Society of Cardiology's guideline which recommends using (s)PESI for initial risk assessment alongside clinical and imaging parameters to decide on patients’ treatment.4 When determining treatment for higher risk groups, additional precautions are necessary to consider in light of their demonstrated increased incidence of adverse events. The study identified several higher risk groups that warrant special attention, including patients with active cancer, those with elevated cardiac biomarkers such as troponin or NT-proBNP, and individuals with preexisting cardiopulmonary disease. One must consider additional psychosocial factors that contribute to these patients being appropriate for outpatient treatments.

In conclusion, based on low incidence of adverse events at 14 and 30 days, we have assigned an NNT color recommendation of green (benefits > harms) for home treatment for appropriately selected low-risk patients with acute PE. Both risk stratification tools discussed here, Hestia or (s)PESI scores, are very user-friendly and easy to calculate in the clinical area using online resources.7^, 8

The original manuscript was published in Academic Emergency Medicine as part of the partnership between TheNNT.com and AEM.

Author

Sophia Bouzrara, MD; Raymond Beyda, MD
Supervising Editors: Shahriar Zehtabchi, MD

Published/Updated

August 26, 2025