Prevalence of venous thromboembolism diagnosed in emergency department visits by cancer patients and associated healthcare resource utilization in the United States

To the Editor: Venous thromboembolism (VTE) is one of the leading causes of morbidity and mortality in cancer patients. Most cancer patients are followed and treated as outpatients and symptoms from VTE can prompt emergency department (ED) visits. Data is scarce regarding VTE diagnosed in the ED setting and associated outcomes. We investigated the prevalence of VTE diagnosed among cancer patients in the EDs and subsequent hospital admission rate, mortality, and healthcare resource utilization. The study cohort was obtained from the Nationwide Emergency Department Sample (NEDS) database, which is the largest nationally representative ED database in the United States. ED visits by patients with at least one cancer diagnosis from 2006 to 2012 were included in our analysis. VTE diagnosis was identified with International Classification of Disease, Ninth Revision, Clinical Modification (ICD-9 CM) codes. The utilization of imaging studies which could lead to diagnosis of VTE was identified using current procedural terminology code. Imaging studies of interests were duplex ultrasound of the extremity veins (DUV), ventilation-perfusion scan (V/Q scan), computed tomography pulmonary angiography (CTPA), and computed tomography of chest with intravenous contrast (CTC). CTC was included because it could detect pulmonary embolisms (PEs) incidentally albeit the study was ordered for other purposes. The prevalence of VTE among cancer patients during ED visits was estimated. Multivariable logistic regression analysis was used to estimate the effect of VTE status on subsequent hospital admission, mortality risk, and total hospital costs. Higher total hospital costs was defined as greater than the 75th percentile of the study population (>$43 964). A two-sided P value < .05 was considered statistically significant. There were 6 330 322 ED visits by cancer patients included in our study. The median age was 70 (interquartile range: 58–80) and 3 369 367 (53%) visits were female. The most common cancer types were breast (n = 1 011 229; 16%), prostate (n = 730 954; 11.6%), lung (n = 704 013; 11.1%), and colorectal cancer (n = 626 439; 9.9%) (Supporting Information Table). The highest prevalence of VTE prevalence in the ED was observed in pancreatic cancer (6.1%), brain and nervous system (3.7%), lung cancer (3.5%), and stomach cancer (3.1%). The lowest VTE rates were in patients with thyroid cancer (1.3%), head and neck cancer (1.3%), and leukemia (1.5%). Imaging studies which could lead to diagnosis of VTE were performed in 1.65% ED visits by cancer patients. From 2006 to 2012, the utilization of the four imaging studies all increased (Supporting Information Figure, CTPA: 0.21% to 0.97%; DUV: 0.36% to 0.92%; CTC: 0.37% to 0.57%; V/Q scan: 0.04% to 0.14%; all P values for trend < .0001). The trend for utilization of any of the above four imaging studies increased from 0.95% to 2.46% (P value for trend < .0001). From 2006 to 2012 (Table 1), DVT prevalence decreased from 1.46 cases (95% CI: 1.43-1.49) to 1.18 cases (95% CI: 1.16-1.20) per 100 ED visits (P value for trend < .0001); PE prevalence increased from 1.17 cases (95% CI: 1.15-1.20) to 1.28 cases (95% CI: 1.26-1.30) (P value for trend < .0001). VTE prevalence were 2.34 cases (95% CI: 2.30-2.37) in 2006 and 2.23 cases (95%: 2.20-2.26) per 100 ED visits (P value for trend < .0001). Of note, the changes over time were all within decimal ranges. In crude analysis, VTE diagnosis was associated with higher hospital admission rate (92% vs. 60.6%), higher mortality rate (6.4% vs. 3.66%), and higher total hospital costs (median: $31 000 vs. $22 800) (all P values < .0001). In multivariable logistic regression analysis after adjusting for age, sex, and cancer type, VTE diagnosis continued to be significantly associated with higher odds of hospital admission [adjusted odds ratio (AOR): 7.43, 95% CI: 7.29-7.57], mortality (AOR: 1.52, 95% CI: 1.49-1.55), and higher hospital costs [>75th percentile ($43 964), AOR: 2.69, 95% CI: 2.66-2.73] (all P values < .0001). To our knowledge, this study is the first and the largest population-level study describing the prevalence and associated outcomes of VTE diagnosed in the ED among cancer patients in the United States. This present analysis revealed relatively stable acute VTE prevalence (2.35 cases per 100 ED visits) but with variation by cancer types. Although the VTE prevalence remained relatively stable, there was a 2.6-fold increase (0.95% to 2.46%) of imaging studies use in cancer patients presented to EDs. While this may suggest increasing awareness of VTE in cancer patients by ED healthcare providers, this also reflects that efforts are needed to properly stratify VTE risk in the EDs. A more fundamental way to reduce VTE burden in cancer patients is probably to identify those with high VTE risk and consider primary outpatient prophylaxis. The PROTECHT1 and SAVE-ONCO2 trials using low-molecular weight or ultra-low-molecular weight heparins demonstrated reduced VTE risk in patients with locally advanced or metastatic solid tumors receiving outpatient chemotherapy. Although their results were statistically significant, the low event rates in this non-risk-adapted population have prevented the results from being adopted into guidelines or clinical practice. An individualized VTE risk assessment approach may better select those who will be benefited more from outpatient VTE prophylaxis and better justified the bleeding risk. To identify those high VTE risk patients, a well-validated Khorana score3 has been recommended by society guidelines to discriminate between high versus low-risk VTE patients. The benefit of low-molecular weight heparin (LMWH) thromboprophylaxis in patients with a score >=3 has recently been demonstrated.4 Two ongoing trials, CASSINI with rivaroxaban5 and AVERT with apixaban6 are seeking to expand upon this with a cutoff score of >=2, using more patient-friendly direct oral anticoagulants. There are inherent limitations of the current study. First, the NEDS database is administrative data. The initial clinical symptoms leading to ED visits were unavailable which prevented more granular analysis (for instance, ED visits for VTE related symptoms vs. incidental VTE diagnosis during ED visits). Second, VTE diagnosis was only based on ICD-9 codes and coding errors could exist. Third, detailed cancer stages and cancer treatment data were unavailable, which are important predictors for VTE. In summary, VTE diagnosis in cancer patients were associated with higher hospital admission rate, mortality rate, and higher hospital costs and is a significant burden to the healthcare system. Outpatient thromboprophylaxis has the potential to reduce such events, thereby reducing ED visits and associated healthcare resource utilization. Conception and design: Y.W. Chen, A.A. Khorana; Collection and assembly of data: Y.W. Chen, M.J. Wang; Data analysis and interpretation: Y.W. Chen, A.A. Khorana; Manuscript writing: Y.W. Chen, M.J. Wang, A.A. Khorana; Final approval of manuscript: Y.W. Chen, M.J. Wang, A.A. Khorana. Dr. Alok Khorana reports personal fees from Janssen, personal fees from Bayer, personal fees from AngioDynamics, personal fees from Pharmacyte, personal fees from Pfizer, outside the submitted work; Yu-Wei Chen and Min-Jung Wang have no conflicts. Additional Supporting Information may be found online in the supporting information tab for this article. Supporting information 1 Supporting information 2 Please note: The publisher is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing content) should be directed to the corresponding author for the article.