INTRODUCTION
Deep vein thrombosis (DVT) is a significant cause of morbidity and mortality worldwide, with an estimated annual incidence of 5 to 10 per 10,000 persons per year in the general population [1–4]. Among emergency department (ED) patients, the rate might be higher, with older literature estimating that 0.1% to 0.2% of ED visits were due to DVT [5]. The condition accounts for substantial healthcare utilization, particularly in the ED setting, where timely diagnosis and management are crucial to prevent adverse outcomes, particularly pulmonary embolism (PE). PE is the third most common cause of cardiovascular death, following coronary artery disease and stroke, with mortality rates for untreated acute PE reaching as high as 30% [6]. However, when venous thromboembolism is diagnosed and treated in a timely manner, the mortality rates drop significantly [7].
Previously, patients with DVT would be admitted to the hospital while they bridged onto anticoagulation therapy. Even in the 2000s, more than half of DVT patients were routinely admitted from the ED [5]. The introduction of direct oral anticoagulants (DOACs) has revolutionized the treatment landscape, offering effective and convenient alternatives to traditional anticoagulation therapies such as warfarin and low molecular weight heparin. This has allowed outpatient treatment for the majority of uncomplicated DVTs, according to the American Society of Hematology’s 2020 Guidelines for Management of Venous Thromboembolism [8]. Despite those developments, no recent literature has examined their effects on clinical practice, and updated, robust data are needed to inform clinical practice and health policy.
This study aims to fill that gap using the Epic Cosmos national database to conduct a comprehensive analysis of DVT presentations in EDs across the United States from 2016 to 2023. The primary objective of this study is to elucidate practice trends in the ED-based diagnosis of DVT during the past 8 years, using a large national database to report incidence, admission rates, and management strategies. By providing large-scale, contemporary data, this study aims to inform evidence-based practices and guide future research and policies for managing DVT.
METHODS
Ethics statement
This study was deemed non-human subjects by the Institutional Review Board of Rush University Medical Center. The study followed best practices and adhered to the STROBE (Strengthening the Reporting of Observational Studies in Epidemiology) guidelines [9].
Study design
This cross-sectional study on ED presentations for DVT reports the incidence, admission rates, and outpatient prescriptions during the 8-year period from January 1, 2016, to December 31, 2023.
We obtained data from the Epic Cosmos research platform (Epic Systems Corp), which is an aggregated database of electronic health records submitted by health systems across the United States. Cosmos data provide a representative sample of patients seeking healthcare and demonstrate consistent alignment with the US Census data [10]. Patient records were centrally de-duplicated and anonymized by Epic. At the time of submission, the Cosmos data set included 259 million unique patients from more than 37,000 hospitals and clinics.
We queried Cosmos using the International Classification of Diseases, 10th Revision (ICD-10) codes for acute DVT. The inclusion criteria were all patients aged ≥18 years who had an ED presentation with an ICD-10 code corresponding to acute DVT (Supplementary Material 1). We excluded nonextremity DVT (i.e., internal jugular vein thrombosis, subclavian vein thrombosis, superior vena cava thrombosis, inferior vena cava thrombosis, and iliac vein thrombosis), chronic DVT, and patients with concomitant PE. Subanalyses were performed after dividing the data into upper and lower extremity DVTs based on ICD-10 codes. The following outcomes were collected: total ED visits, total ED visits for DVT, admission rates, and anticoagulant prescriptions. The anticoagulants prescribed were warfarin, enoxaparin, dabigatran, apixaban, rivaroxaban, edoxaban, and betrixaban. We compared those prescriptions against Epic’s Medication Grouper function, which categorizes medications by pharmaceutical class and subclass, to ensure that all anticoagulants for DVT were captured. Data are presented using descriptive statistics.
RESULTS
During the 8-year study period, a total of 190,144,463 ED encounters were reported, with isolated DVT accounting for 368,044 (0.2%). The overall rate of isolated DVT remained stable over time (Supplementary Table 1). Among those with isolated DVT, 119,986 (32.6%) were admitted, and that admission rate was also stable over time (Fig. 1 and Supplementary Table 1). The most common reasons for ED presentation are presented in Table 1. The demographics of the study population are reported in Table 2.
Among those prescribed anticoagulation agents, the most common was apixaban (40.3%), followed by rivaroxaban (28.3%), enoxaparin (7.9%), warfarin (3.6%), and dabigatran (0.3%). Over time, apixaban demonstrated a marked increase in use from 12.4% in 2016 to 56.2% in 2023 (Fig. 2 and Supplementary Table 2).
Lower extremity DVT accounted for the majority of DVTs (88.5%), with upper extremity DVT present in 11.7% of DVT patients. Among those with lower extremity DVTs, 32.1% were admitted, with minimal change over time (Supplementary Table 3). Patients with upper extremity DVTs were admitted 37.0% of the time (Supplementary Table 4). Prescription medication demonstrated similar trends for both upper and lower extremity DVT, with a rapid increase in apixaban use to approximately 40% in both groups (Supplementary Tables 5, 6).
DISCUSSION
This comprehensive analysis of DVT presentations to US EDs during an 8-year period identified several key trends. The overall rate of isolated extremity DVT remained stable at 0.2% of all ED encounters during the 8-year study period, but that is an increase compared with earlier studies [5].
The increase in DVT diagnosis could be secondary to increased detection during the past decade due to greater use of point-of-care ultrasound in EDs [11]. As data have suggested similar accuracy between point-of-care ultrasound and radiology-performed ultrasound for DVT [12,13], the lower threshold for performing this test at the bedside might have identified cases that would have otherwise been missed. However, it might also reflect a higher rate of false positives because point-of-care ultrasound is a user-dependent skill [12].
Patient-specific factors might also influence these findings. According to the US Census Bureau, between 2010 and 2020, the US population older than 65 years increased by 38.6%, with a further 9.4% increase from 2020 to 2023 [14,15]. In addition, the US Centers for Disease Control and Prevention reported a 39% increase in the age-adjusted prevalence of obesity among US adults from 1999 to 2018 (from 30.5% to 42.4%) and a 5.7% increase in severe obesity (from 4.7% to 9.2%) [16]. Multiple sources have demonstrated that advanced age is a risk factor for venous thromboembolism (VTE) [6,17,18], and obesity has also been suggested as an independent risk factor for VTE [19]. As such, this rising incidence might reflect the advancing age of the US population and the comorbid obesity epidemic.
From a socioeconomic perspective, we analyzed data after adoption and implementation of the Affordable Care Act, which significantly decreased the number of uninsured people. Therefore, increased insurance coverage might have prompted patients to obtain medical care for conditions they previously would have tried to manage without care, which could increase the incidence of DVT diagnoses.
Notably, despite the higher overall incidence compared with prior studies, the rate during the most recent 8-year period has remained stable. This might reflect initiatives to decrease the incidence of VTE. In the United States, many organizations, including the Joint Commission, have prioritized measuring and reporting VTE outcomes with the goal of reducing the incidence of and preventable harm from VTE [20]. Similarly, as awareness of the risk factors and causes of VTE has increased, patient education about DVT prophylaxis has also increased, along with pharmacoprophylaxis after high-risk procedures (such as certain orthopedic operations) [21,22].
Another notable finding of our study is the stable admission rate for isolated extremity DVT, with approximately one-third of patients (32.6%) being admitted. This rate is significantly lower than in historical data: admission rates were as high as 52% in the early 2000s [5]. The reduction in admissions can likely be attributed to a few factors, including greater coordination of home health care and widespread adoption of DOACs, which facilitate the outpatient treatment of uncomplicated DVT. This is consistent with the American Society of Hematology’s 2020 guidelines that recommend DOACs over warfarin for the treatment of VTE [8].
Apixaban saw a significant increase in usage during the 8 years of the study, from 12.4% in 2016 to 56.2% in 2023. This aligns with studies that have demonstrated that apixaban has better safety and efficacy profiles than other DOACs. One study comparing apixaban with rivaroxaban in terms of major ischemic or hemorrhagic events in patients with atrial fibrillation found a statistically increased risk of complications with rivaroxaban compared with apixaban [23]. Other studies among patients with VTE have reported lower risks of recurrent VTE and bleeding with apixaban than with rivaroxaban [24,25].
When we analyzed lower extremity DVT and upper extremity DVT separately, we found a difference in admission rates (32.1% for lower extremity vs. 37.0% for upper extremity). The higher admission rate for upper extremity DVT might reflect the perceived complexity and potential complications associated with upper extremity thrombosis, such as catheter-related thrombosis. Catheter-associated upper extremity DVT is the most common etiology, causing 93% of all upper extremity DVTs in one retrospective analysis of 373 patients [26]. Another study noted that the presence of a central venous catheter increased the risk of upper extremity DVT by up to 14-fold [27]. Admission to potentially exchange or remove centrally inserted catheters could also explain the higher admission rate. Despite those differences, the prescription trends for anticoagulants were similar across the groups, reinforcing the growing preference for DOACs in the treatment of all forms of DVT.
Notably, our study has several limitations that need to be considered. The Epic Cosmos research platform contains data only from hospitals using Epic as their main electronic medical record system that choose to contribute data. Thus, these data might not represent overall clinical practice in the United States but only that in hospitals that opt into a national research database, potentially representing more research-driven institutions. In addition, we could not exclude other concomitant medical conditions that might have influenced admission decisions or management strategy. Moreover, reliance on ICD-10 coding might have led to some missed diagnoses due to incorrect coding. Last, given that we are unable to analyze patient outcomes based on the practice changes we identified, we cannot make any conclusions or recommendations about the effectiveness of current management strategies.
In summary, our study provides a summary of DVT presentations and management across many US EDs during an 8-year period. We identified increased incidence rates and a significant shift toward outpatient management (compared with prior work) that is facilitated by the use of DOACs, particularly apixaban. These findings emphasize the need for ongoing research and continuous evaluation of clinical practices to ensure optimal outcomes for patients with DVT. Future studies should focus on the long-term outcomes of patients managed with these new approaches.
PDF Links
PubReader
ePub Link
XML Download
Full text via DOI
Download Citation 
Supplement1
Print 
