Abstract
Background
Few studies describe both inpatient and outpatient treatment and outcomes of patients with acute venous thromboembolism in the US.
Methods
A multi-institutional cohort of patients diagnosed with confirmed pulmonary embolism and/or deep venous thrombosis during years 2004 through 2010 was established from four large, US-based integrated healthcare delivery systems. Computerized databases were accessed and medical records reviewed to collect information on patient demographics, clinical risk factors, initial antithrombotic treatment, and vital status. Multivariable Cox regression models were used to estimate the risk of death at 90 days.
Results
The cohort comprised 5,497 adults with acute venous thromboembolism. Pulmonary embolism was predominantly managed in the hospital setting (95.0%) while 54.5% of patients with lower extremity thrombosis were treated as outpatients. Anticoagulant treatment differed according to thromboembolism type: 2688 (92.8%) of patients with pulmonary embolism and 1625 (86.9%) of patients with lower extremity thrombosis were discharged on anticoagulants, compared to 286 (80.1%) patients with upper extremity thrombosis and 69 (54.8%) patients with other thrombosis. While 4.5% of patients died during the index episode, 15.4% died within 90 days. Pulmonary embolism was associated with a higher 90-day death risk than lower extremity thrombosis (adjusted hazard ratio [aHR] 1.23 [1.04-1.47]) as was not being discharged on anticoagulants (aHR 5.56 [4.76-6.67]).
Conclusions
In this multicenter, community-based study of patients with acute venous thromboembolism, anticoagulant treatment and outcomes varied by thromboembolism type. Although case-fatality during the acute episode was relatively low, 15.4% of people with thromboembolism died within 90 days of the index diagnosis.
Keywords: venous thromboembolism, anticoagulation, deep venous thrombosis, pulmonary embolism, mortality
Introduction
Venous thromboembolism, primarily pulmonary embolism and deep venous thrombosis, affects an estimated 750,000 people in the United States (US) each year and causes more than 100,000 deaths annually.1-3 Yet there are relatively few large US-based studies that describe in detail the characteristics, treatment, and outcomes of venous thromboembolism. Many studies rely solely on diagnosis codes to identify thromboembolism, which may not accurately identify true thrombotic events4-8. Administrative databases also lack complete data on patient risk factors, acuity of illness, and treatment plan.9 Most studies focus on hospitalized patients,4-7, 10 despite acute management moving towards outpatient settings11, or only include patients discharged on anticoagulants.
To more completely describe the treatment and outcomes of patients with venous thromboembolism, we conducted a multicenter observational cohort study of patients with acute thrombosis within the Cardiovascular Research Network Venous Thromboembolism (CVRN VTE) consortium. Our goal was to describe treatment strategies, morbidity, and mortality outcomes of both inpatients and outpatients with acute thromboembolism.
Materials and Methods
Cohort Assembly
The CVRN VTE study is a retrospective cohort of patients with venous thromboembolism enrolled in one of four integrated healthcare delivery systems in the US. The four systems represented diverse geographic patient sources, with service provided during the study period to the following populations: Kaiser Permanente Northern California (>3.2 million members in Northern California during the study period), Kaiser Permanente Colorado (>460,000 members in the Denver, Colorado metropolitan area), Geisinger Health System, now Geisinger (~2.5 million members in central and northwest Pennsylvania), and Marshfield Clinic (>550,000 members in central and northwest Wisconsin). All clinical sites used electronic health record systems during the time period of the study. The study was approved by the institutional review board at each of the participating sites according to local policies and procedures. A waiver of informed consent was obtained due to the nature of the study.
The cohort was assembled by first searching computerized health plan databases for all individuals ⋥21 years of age with a clinical encounter (inpatient, emergency department [ED] or ambulatory) associated with either a primary or secondary International Classification of Diseases, Ninth Revision (ICD-9) diagnosis code of venous thromboembolism, between October 1, 2004 and December 31, 2010. Only the first encounter with a code for venous thromboembolism during the time period was reviewed. Because the focus of the study was on incident thrombosis, patients with a prior code for thrombosis while enrolled in the health plan, or prescribed anticoagulants within 4 years prior, were excluded. We included only subjects with continuous health plan and pharmacy benefits for at least 12 months prior to the index thrombotic encounter.
Events were categorized as pulmonary embolism (ICD-9 code 415.1x), lower extremity deep venous thrombosis (451.1x, 451.2, 451.81, 453.4x, 453.5x), upper extremity deep venous thrombosis (451.83, 451.84, 451.89, 453.72, 453.73, 453.74, 453.75, 453.76, 453.77, 453.82, 453.83, 453.84, 453.85, 453.86, 453.87), and other venous thrombosis (451, 451.9, 452, 453, 453.0, 453.1, 453.2, 453.3, 453.79, 453.8, 453.89, 453.9). We did not include pregnancy-related thromboembolism or superficial thrombophlebitis as part of this study.
Validation of Venous Thromboembolic Events
We identified 42,941 unique individuals with an incident clinical encounter assigned a diagnosis of thromboembolism meeting inclusion criteria. To attain a target sample size of ~5500 patients with valid, acute venous thrombosis, we initially selected 7,334 clinical encounters for review, comprising all available patients in two of the health systems and a random sample of patients from the other two systems. All available inpatient admission, transfer, discharge, ED, and outpatient encounter notes were redacted, as well as relevant radiology reports within 72 hours before and after the index venous thromboembolism date. Trained physician and pharmacist reviewers then reviewed medical records using a structured adjudication tool to determine whether the encounter represented a valid acute thromboembolism. Events were considered valid if there was radiographic, operative, or autopsy evidence of an acute thrombosis, or if a physician documented in the medical record that an acute venous thromboembolism had occurred during that episode of care. An encounter with a prior history of venous thromboembolism whose acute management was not contiguous with the current episode was not considered an acute event. Superficial venous thromboses/thrombophlebitis were not considered valid events.
Of the initial 7,334 patients, 7,063 had medical records available for review. However, only 2,135 (30.2%) were determined upon review to represent valid thromboembolism events, with ambulatory encounters and those with thrombosis codes in the secondary position less likely to represent true acute events (results of the validation process described previously).8 To enhance the likelihood of finding valid encounters, we then randomly selected an additional set of patients for review, restricting to hospital/ED encounters with a primary diagnosis of thrombosis in years 2006 or later, until we reached a total of ~5500 valid events (Supplemental Figure).
Presentation, Treatment, and Outcomes of Acute Venous Thromboembolism
Clinical and demographic information were obtained from computerized health plan databases. Relevant medical conditions within 4 years prior to the index event were identified using inpatient or outpatient ICD-9 codes. Ascertainment of diabetes mellitus and malignant cancer was supplemented using regional diabetes and cancer registries from some of the health systems.12, 13 ICD-9 codes were used to calculate a Charlson comorbidity index score for each patient.14 Receipt of medications to manage chronic cardiovascular diseases were identified by searching health plan pharmacy dispensing databases within 30 days prior to the index thromboembolism date.
Clinician reviewers collected information on thromboembolism type, smoking history, and documented immobility or injury (such as surgery within 30 days, trauma or prolonged immobility within 7 days, indwelling venous catheters). Thromboembolic events were defined as “unprovoked” if they occurred in the absence of identifiable risk factors such as diagnosed cancer or documented history of recent surgery, immobility, or injury.
Reviewers assessed whether antithrombotic medications were prescribed at the end of the encounter and categorized as oral anticoagulants (specifically warfarin, as direct oral anticoagulants were not yet approved for use in thromboembolism), parenteral anticoagulants (e.g., low-molecular-weight heparins, fondaparinux), and antiplatelet agents (aspirin, clopidogrel, ticlopidine, dipyridimole, and cilostazol). For encounters without antithrombotic treatment at discharge, the encounter documentation was reviewed for potential reasons why treatment was not issued.
The primary outcome for the study was death from any cause within 90 days after the index thromboembolic event. Secondary outcomes included death during the index event, death within 30 days, and new functional impairment at discharge. Death was determined from medical record review, health plan databases, and the Social Security Index death files. Functional impairment was ascertained by review of the encounter documentation. Reviewers used their clinical judgement to determine whether the patient suffered functional consequences as a result of the thrombotic event and categorized as “at least some functional impairments related to the event”, “no functional impairments related to the event”, or “unknown”.
Statistical Analysis
The study objective was to compare anticoagulant treatment patterns and short-term outcomes by thromboembolism type (any pulmonary embolism, lower extremity thrombosis, upper extremity thrombosis, and other venous thromboses). We first used descriptive statistics to report clinical characteristics, treatment, and outcomes by event type, using ANOVA for continuous variables and χ2 tests for categorical variables when comparing amongst thromboembolism types. We then modeled death within 90 days as a function of event type and anticoagulant treatment, developing a multivariable Cox regression model that incorporated variables that have been or could be plausibly be related to short-term mortality, including risk factors for thrombosis, coexisting medical conditions, Charlson comorbidity risk score, and baseline use of medications to manage cardiovascular conditions. A two-sided p-value less than 0.05 was considered statistically significant. All analyses were performed using SAS version 9.1 (SAS Institute, Inc., Cary, NC).
Results
The cohort comprised 5,497 adults with confirmed acute incident venous thromboembolism. The mean age of the cohort was 65.7 years and 51.4% were women (Table 1). The majority of subjects (n=3,056) had pulmonary embolism with or without concomitant deep venous thrombosis, followed by patients with isolated lower extremity thrombosis (n=1,928). Relatively few patients had isolated upper extremity thrombosis (n=383) or other forms of thrombosis (n=130).
Table 1.
Characteristics of 5,497 Patients with Acute Venous Thromboembolism
| TOTAL | Pulmonary embolism (with or without deep venous thrombosis) N=3056 |
Lower extremity thrombosis N=1928 |
Isolated upper extremity thrombosis N=383 |
Isolated other thrombosis N=130 |
P- Value17 |
|
|---|---|---|---|---|---|---|
| Age, mean years (SD) | 65.7 (15.8) | 65.7 (15.6) | 66.5 (16.1) | 62.8 (15.8) | 61.1 (15.0) | <0.001 |
| Women, n (%) | 2824 (51.4) | 1565 (51.2) | 991 (51.4) | 209 (54.6) | 59 (45.4) | 0.33 |
| Race, n (%) | <0.001 | |||||
| White | 4375 (79.6) | 2437 (79.7) | 1546 (80.2) | 302 (78.9) | 90 (69.2) | |
| Black | 436 (7.9) | 260 (8.5) | 135 (7.0) | 28 (7.3) | 13 (10.0) | |
| Asian | 161 (2.9) | 85 (2.8) | 46 (2.4) | 19 (5.0) | 11 (8.5) | |
| American/Pacific Islander | ||||||
| Other/unknown | 525 (9.6) | 274 (9.0) | 201 (10.4) | 34 (8.9) | 16 (12.3) | |
| Hispanic ethnicity, n (%) | 391 (7.1) | 200 (6.5) | 138 (7.2) | 40 (10.4) | 13 (10.0) | 0.02 |
| Treatment setting, n (%) | <0.001 | |||||
| Hospital | 4155 (75.6) | 2902 (95.0) | 878 (45.5) | 267 (69.7) | 108 (83.1) | |
| Emergency department or ambulatory clinic only | 1342 (24.4) | 154 (5.0) | 1050 (54.5) | 116 (30.3) | 22 (16.9) | |
| Baseline medical history, n (%) | ||||||
| Hypertension | 3165 (57.6) | 1748 (57.2) | 1118 (58.0) | 225 (58.7) | 74 (56.9) | 0.91 |
| Diabetes mellitus | 1178 (21.4) | 624 (20.4) | 416 (21.6) | 108 (28.2) | 30 (23.1) | 0.006 |
| Cancer | 2002 (36.4) | 1115 (36.5) | 665 (34.5) | 172 (44.9) | 50 (38.5) | <0.001 |
| Prior ischemic stroke | 125 (2.3) | 70 (2.3) | 43 (2.2) | 9 (2.3) | 3 (2.3) | 1.0 |
| Ischemic heart disease18 | 558 (10.2) | 312 (10.2) | 208 (10.8) | 35 (9.1) | 3 (2.3) | 0.02 |
| Heart failure | 514 (9.4) | 280 (9.2) | 171 (8.9) | 55 (14.4) | 8 (6.2) | 0.004 |
| Atrial fibrillation or flutter | 504 (9.2) | 271 (8.9) | 175 (9.1) | 52 (13.6) | 6 (4.6) | 0.006 |
| Valvular heart disease | 320 (5.8) | 162 (5.3) | 120 (6.2) | 32 (8.4) | 6 (4.6) | 0.07 |
| Chronic pulmonary disease19 | 1065 (19.4) | 623 (20.4) | 325 (16.9) | 104 (27.2) | 13 (10.0) | <0.001 |
| Chronic liver disease | 181 (3.3) | 75 (2.5) | 63 (3.3) | 19 (5.0) | 24 (18.5) | <0.001 |
| Hospitalization for sepsis | 195 (3.5) | 90 (2.9) | 74 (3.8) | 29 (7.6) | 2 (1.5) | <0.001 |
| Inflammatory bowel disease | 98 (1.8) | 50 (1.6) | 35 (1.8) | 9 (2.3) | 4 (3.1) | 0.51 |
| Hypercoagulable hematologic conditions20 | 124 (2.3) | 55 (1.8) | 49 (2.5) | 15 (3.9) | 5 (3.8) | 0.01 |
| Thrombophilia21 | 22 (0.4) | 8 (0.3) | 9 (0.5) | 3 (0.8) | 2 (1.5) | 0.06 |
| Prior hospitalization for intracranial hemorrhage | 103 (1.9) | 46 (1.5) | 48 (2.5) | 8 (2.1) | 1 (0.8) | 0.07 |
| Prior hospitalization for gastrointestinal hemorrhage | 161 (2.9) | 82 (2.7) | 57 (3.0) | 16 (4.2) | 6 (4.6) | 0.26 |
| Prior hospitalization for other hemorrhage | 41 (0.7) | 19 (0.6) | 19 (1.0) | 2 (0.5) | 1 (0.8) | 0.50 |
| Prior hospitalization for mechanical fall | 32 (0.6) | 17 (0.6) | 14 (0.7) | 1 (0.3) | 0 (0.0) | 0.54 |
| Baseline medication use, n (%) | ||||||
| Angiotensin converting enzyme inhibitor | 1106 (20.1) | 626 (20.5) | 371 (19.2) | 86 (22.5) | 23 (17.7) | 0.40 |
| Angiotensin II receptor blocker | 324 (5.9) | 193 (6.3) | 99 (5.1) | 25 (6.5) | 7 (5.4) | 0.34 |
| Beta blocker | 1730 (31.5) | 978 (32.0) | 603 (31.3) | 118 (30.8) | 31 (23.8) | 0.26 |
| Calcium channel blocker | 807 (14.7) | 470 (15.4) | 262 (13.6) | 59 (15.4) | 16 (12.3) | 0.29 |
| Diuretic | 1615 (29.4) | 924 (30.2) | 540 (28.0) | 112 (29.2) | 39 (30.0) | 0.41 |
| Aldosterone receptor antagonist | 152 (2.8) | 81 (2.7) | 42 (2.2) | 18 (4.7) | 11 (8.5) | <0.001 |
| Alpha blocker | 416 (7.6) | 224 (7.3) | 152 (7.9) | 34 (8.9) | 6 (4.6) | 0.38 |
| Statin | 1586 (28.9) | 941 (30.8) | 515 (26.7) | 110 (28.7) | 20 (15.4) | <0.001 |
| Other lipid-lowering agent | 136 (2.5) | 72 (2.4) | 54 (2.8) | 9 (2.3) | 1 (0.8) | 0.45 |
| Non-aspirin antiplatelet agent | 290 (5.3) | 152 (5.0) | 117 (6.1) | 19 (5.0) | 2 (1.5) | 0.08 |
| Antidiabetic treatment | 736 (13.4) | 411 (13.4) | 240 (12.4) | 66 (17.2) | 19 (14.6) | 0.09 |
| Non-steroidal anti-inflammatory drugs | 588 (10.7) | 353 (11.6) | 192 (10.0) | 34 (8.9) | 9 (6.9) | 0.08 |
| Current smoker, n (%) | 438 (8.0) | 228 (7.5) | 147 (7.6) | 43 (11.2) | 20 (15.4) | <0.001 |
| Hospitalization within 30 days prior to index thrombotic event | 1045 (19.0) | 580 (19.0) | 349 (18.1) | 94 (24.5) | 22 (16.9) | 0.03 |
| Recent immobility/injury, n (%) | 2302 (41.9) | 1230 (40.2) | 794 (41.2) | 252 (65.8) | 26 (20.0) | <0.001 |
| Indwelling venous catheter | 301 (5.5) | 76 (2.5) | 32 (1.7) | 188 (49.1) | 5 (3.8) | |
| Prolonged immobility | 977 (17.8) | 557 (18.2) | 363 (18.8) | 47 (12.3) | 10 (7.7) | |
| Trauma/surgery within 30 days | 1172 (21.3) | 621 (20.3) | 446 (23.1) | 94 (24.5) | 11 (8.5) | |
| Charlson comorbidity index, n(%) | <0.001 | |||||
| <1 | 1550 (28.2) | 878 (28.7%) | 568 (29.5) | 64 (16.7) | 40 (30.8) | |
| 1-2 (mild) | 1664 (30.3) | 947 (31.0) | 586 (30.5) | 90 (23.5) | 41 (31.5) | |
| 3-4 (moderate) | 910 (16.6) | 505 (16.5) | 317 (16.4) | 72 (18.8) | 16 (12.3) | |
| ≥ 5 (severe) | 1373 (25.0) | 726 (23.8) | 457 (23.7) | 157 (41.0) | 33 (25.4) |
Comparison by thrombosis types
Defined as acute myocardial infarction, unstable angina, percutaneous coronary intervention, or coronary artery bypass grafting
Defined as chronic obstructive pulmonary disease, emphysema, chronic bronchitis, chronic obstructive asthma, bronchiectasis, interstitial lung disease, pulmonary hypertension
Defined as polycythemia vera, paroxysmal nocturnal hemoglobinuria, homocystinuria, congenital deficiency of other clotting factors, defibrination syndrome, essential thrombocythemia, myelofibrosis with myeloid metaplasia, monoclonal paraproteinemia.
Defined as ICD-9 codes 289.81 (primary hypercoagulable state) and 289.82 (secondary hypercoagulable state)
Cancer was a common comorbid condition, present in 2,022 (36.4%) of patients overall and with a particularly high prevalence among people with upper extremity thrombosis (44.9%). Recent immobility and/or injury was identified in 2,302 (41.9%) of the cohort. In 188 (49.1%) of the 383 patients with upper extremity thrombosis, an indwelling venous catheter was identified as a potential provoking event. For 1,254 (22.8%) of patients, the event was categorized as unprovoked (e.g., no history of immobility, injury, recent hospitalization, or known cancer).
Treatment and outcomes of acute venous thromboembolism
The majority of patients received anticoagulants (oral and/or parenteral) at discharge, although treatment differed by thrombosis type (Table 2). Among survivors of the acute event, 92.8% of those with pulmonary embolism and 86.9% of those with lower extremity thrombosis were discharged on anticoagulants. Anticoagulant treatment was less common among patients with thromboses in other locations: 80.1% of patients with upper extremity thrombosis and 54.8% of patients with other thrombosis. Few patients (2.9%) were discharged on antiplatelet agents only. Among the 430 patients who survived the acute episode and were not prescribed antithrombotic therapy at discharge, bleeding risk and/or poor prognosis were identified by reviewers as a rationale to avoid anticoagulants in 75.3% and 61.2% of patients, respectively.
Table 2.
Short-term Outcomes of 5,497 Adults with Acute Venous Thromboembolism
| TOTAL N=5497 |
Pulmonary embolism (with or without lower extremity thrombosis) N=3056 |
Lower extremity thrombosis N=1928 |
Isolated upper extremity thrombosis N=383 |
Isolated other thrombosis N=130 |
P- value |
|
|---|---|---|---|---|---|---|
| Disposition at discharge* | <0.05 | |||||
| In-hospital death | 247 (4.5%) | 158 (5.2%) | 59 (3.1%) | 26 (6.8%) | 4 (3.1%) | |
| Alive, no functional impairment | 3746(68.1%) | 1953 (63.9%) | 1422 (73.8%) | 276 (72.1%) | 95 (73.1%) | |
| Alive, with at least some functional impairment | 1185 (21.6%) | 797 (26.1%) | 301 (15.6%) | 61 (15.9%) | 26 (20.0%) | |
| Alive, unknown functional status | 319 (5.8%) | 148 (4.8%) | 146 (7.6%) | 20 (5.2%) | 5 (3.8%) | |
| Death within 30 days | 439 (8.0%) | 267 (8.7%) | 112 (5.8%) | 43 (11.2%) | 18 (13.9%) | <0.05 |
| Death within 90 days | 846 (15.4%) | 475 (15.5%) | 258 (13.4%) | 80 (20.9%) | 33 (25.4%) | <0.05 |
| Antithrombotic treatment among survivors of acute thromboembolism | N=5250 | N=2898 | N=1869 | N=357 | N=126 | <0.05 |
| Anticoagulant | 4668 (88.9%) | 2688 (92.8%) | 1625 (86.9%) | 286 (80.1%) | 69 (54.8%) | |
| Aspirin/other antiplatelet only | 152 (2.9%) | 32 (1.1%) | 94 (5.0%) | 20 (5.6%) | 6 (4.8%) |
Functional impairment ascertained by review of encounter documentation. Reviewers asked to determine whether the patient suffered functional consequences as a result of the thrombotic event
Death during the index thrombotic episode occurred in 5.2% of patients with pulmonary embolism and 3.1% of patients with lower extremity thrombosis (Table 2). Among patients who survived the index event, reviewers noted “at least some functional impairments related to the event” more commonly among patients with pulmonary embolism (26.1%) than patients with isolated lower extremity thrombosis (15.6%. p<0.01). The proportion of patients who died after the index event continued to rise after discharge: all-cause mortality was 8.0% at 30 days and 15.4% at 90 days (Table 2). Among all inpatients, 9.8% died within 30 days and 17.9% died within 90 days. In contrast, 2.5% and 7.7% of ED/ambulatory patients died within 30 and 90 days, respectively. Crude mortality at 90 days was highest in patients with other thrombosis (25.4%), followed by isolated upper extremity thrombosis (20.9%) and then pulmonary embolism (15.5%). However, after multivariable adjustment, no significant differences in 90-day death rates persisted between these three groups (Table 3). In a multivariable Cox regression model, patients with pulmonary emboli had a higher 90-day mortality risk compared to patients with lower extremity thrombosis, with an adjusted hazard ratio [aHR] of 1.2 (95% confidence interval [CI] 1.04-1.5). Not being discharged on anticoagulants was associated with significantly higher mortality risk, aHR 5.6 (95% CI: 4.8-6.7). Other clinical factors associated with a higher risk of death at 90 days included older age, current smoking, a diagnosis of cancer, and higher Charlson comorbidity scores (Table 3).
Table 3.
Significant predictors of death within 90 days in 5,497 adults with acute venous thromboembolism
| Variable | Adjusted Hazard Ratio* (95% CI) |
|---|---|
| Venous thromboembolism type | |
| Pulmonary embolism | ref |
| Lower extremity thrombosis | 0.81 (0.68-0.96) |
| Upper extremity thrombosis | 0.82 (0.64-1.05) |
| Other thrombosis | 0.85 (0.57-1.26) |
| No anticoagulant treatment at discharge | 5.6 [4.8-6.7] |
| Age (years) | |
| <60 | ref |
| 60-69 | 1.49 (1.18-1.89) |
| 70-79 | 1.56 (1.24-1.97) |
| ≥80 | 2.39 (1.89-3.02) |
| Current smoker | 1.34 (1.03-1.75) |
| Unprovoked thrombosis | 0.70 (0.57-0.87) |
| Treatment setting | |
| Inpatient | ref |
| Emergency department or outpatient | 0.50 (0.40-0.62) |
| Malignant cancer | 1.38 (1.16-1.63) |
| Charlson comorbidity index category | |
| <1 normal | ref |
| 1-2 mild | 1.42 (1.07-1.87) |
| 3-4 moderate | 1.79 (1.32-2.41) |
| ≥ 5 severe | 3.78 (2.81-5.09) |
Multivariable models also adjusted for: gender, race, Hispanic ethnicity, hospitalization in the prior 30 days, specific medical conditions (unstable angina, acute myocardial infarction, stroke/transient ischemic attack, peripheral arterial disease, chronic lung disease, chronic liver disease, diabetes mellitus, heart failure, dementia, inflammatory bowel disease, thrombophilia, sepsis, hypercoagulable hematologic conditions, intracranial hemorrhage, extracranial bleeding) and baseline medication use (ACE inhibitor, angiotensin receptor blocker, beta blocker, calcium channel blocker, diuretic, aldosterone receptor antagonist, alpha blockers, statin, non-aspirin antiplatelet agent, diabetic therapy).
Discussion
Our study provides a unique examination of the acute presentation, treatment, and outcomes of inpatients and outpatients with venous thromboembolism across four geographically diverse US-based health systems. Although most patients with thromboembolism were treated with anticoagulants, 1 in 10 people were not, frequently due to concerns for bleeding risk or poor clinical prognosis. The short-term outcomes after a diagnosis of thromboembolism are consequential, with 15.4% of patients dying within 90 days after initial diagnosis.
Strengths of our study include its diverse source population, inclusion of events from both inpatient and outpatient settings, and validation of index thromboembolic events and collection of data through review of medical records. These methods were then facilitated with additional data from comprehensive electronic medical records at each study site. We were able to identify health care utilization even if patients presented to facilities outside of the healthcare system, as all patients were enrolled in integrated healthcare delivery systems. In contrast, most large, contemporary studies addressing acute venous thromboembolism in the US have relied on administrative data obtained from hospital settings, providing an incomplete perspective.4, 5, 15, 16 Our prior work found that only 30.9% of ambulatory encounters with a thrombosis diagnosis actually represented an acute thrombotic event .8 With the advent of direct oral anticoagulants, it is likely that the outpatient treatment of thromboembolism will become even more prevalent.
Descriptions of short-term outcomes following acute pulmonary embolism vary across different studies. The Multicenter Emergency Medicine Pulmonary Embolism in the Real World Registry (EMPEROR) study reported a relatively low 30-day mortality rate (5.4%) in patients with pulmonary embolism presenting to US-based emergency departments between 2005-200817, comparable to the international Registro Informatizado de la Enfermedad TromboEmbólica (RIETE) registry (5.9% mortality).18 In contrast, a Canadian study found a substantially higher (16.9%) 30-day mortality rate among hospitalized and ambulatory patients19 as did a Danish nationwide registry (19.9% 30-day mortality after hospitalization).20 The Nationwide Inpatient Sample, a nationally representative sample of US-hospital admissions, reported a declining risk of inpatient death after pulmonary embolism (7.1% in 1993 and 3.2% in 2012).21 Our study found that 5.2% of patients with pulmonary embolism died during the index event and 8.7% at 30 days.
Variation in short-term outcomes observed across studies may have several possible explanations. One possibility may be differences in anticoagulation management and delivery; for example, the four health systems in our study were all integrated healthcare delivery models of care. Our study included patients treated in outpatient settings, who may have less severe disease and better functional status. Previous studies have shown that selective outpatient treatment of acute pulmonary embolism is feasible and effective.22-25 Despite this, outpatient management of pulmonary embolism is still limited in clinical practice 18, 26-28 Different follow-up periods could account for differences in case-fatality rates, given secular trends of lower mortality from pulmonary embolism within the last two decades.4, 18, 21 Finally, the introduction and increased use of multidetector computed tomography may have contributed to including less severe or even clinically unimportant pulmonary emboli.7, 29
In contrast to pulmonary embolism, we observed less discrepancy in 30-day mortality after deep venous thrombosis in our cohort compared to findings from the Worcester VTE study (5.4%),30 a Canadian study (7%),19 and a Danish registry (4.5%).20 Our results are consistent with other studies that show pulmonary embolism as an independent predictor of short-term death compared to deep venous thrombosis alone.30-32
Most patients in our study were prescribed anticoagulants and anticoagulant treatment was associated with a significantly lower risk for 90-day death. Causality between treatment and mortality cannot be inferred, as it is highly likely there was channeling bias, where patients with limited life expectancy or at higher risk for adverse outcomes from anticoagulants were less likely to be prescribed treatment. Indeed, poor prognosis was commonly cited as a reason for avoiding therapy. Anticoagulation treatment varied by event type, with upper extremity thrombosis and thrombosis in unusual sites less likely to be treated with anticoagulants.
Our results have several implications for practitioners, policy makers, and researchers. Our results underline the persistent public health importance of venous thromboembolism and highlight the substantial short-term mortality after incident thrombosis. Efforts to prevent thromboembolism and ensure timely diagnosis and treatment should be undertaken. Given that more than 19 billion dollars annually are spent for the treatment of thromboembolism in the US,33 with a substantial amount accounting from in-hospital care,34 better education of clinicians about the appropriate risk stratification and the possibility of home treatment of patients could help to reduce costs.26
There are also several limitations to this study. We oversampled for hospitalized patients, so preferentially captured more pulmonary embolism events and our findings might have been slightly biased towards patients with more severe disease. As a retrospective and observational study of clinical care, variation in documentation could have contributed to misclassification of certain variables, such as prior injury or immobility. Our study specifically addressed incident thrombosis and did not include recurrent thromboembolic events, which have different risk factors and outcomes. Finally, our study did not include the experience of patients with thrombosis who were treated with the direct oral anticoagulants, which were not yet in widespread use during the study period.35 Nevertheless, VKAs remain commonly prescribed in the US and worldwide.36, 37
Conclusions
This large, multicenter cohort of patients with acute venous thromboembolism provides unique insights into treatment and outcomes of thromboembolism and demonstrated that anticoagulant treatment patterns and short-term mortality varied by event-type. Although case-fatality during the acute episode was relatively low, 15.4% of affected patients with acute thromboembolism died within 90 days of the incident event, highlighting the need for effective strategies to improve outcomes in this population. As treatment options continue to expand beyond warfarin to include direct oral anticoagulants, it is important to document how treatment and outcomes of thromboembolism evolves over time.
Supplementary Material
Supplementary Figure. Flow diagram of how 5,497 patients with valid, incident venous thromboembolism were identified for the CVRN VTE cohort
Acknowledgments
Funding
This study was supported by the National Heart, Lung, and Blood Institute of the National Institutes of Health (Grants R01HL103820 and U19HL91179). The funding source had no role in the study design, the collection, analysis, and interpretation of data, writing the report, or the decision to submit the article for publication
Footnotes
Conflict of interests: none for all authors
Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
References
- 1.The Surgeon General’s Call to Action to Prevent Deep Vein Thrombosis and Pulmonary Embolism, Office of the Surgeon General (US), National Heart, Lung, and Blood Institute (US), Rockville, MD: Office of the Surgeon General (US) 2008;https://www.ncbi.nlm.nih.gov/books/NBK44181/, accessed March 14, 2019. [PubMed] [Google Scholar]
- 2.Beckman MG, Hooper WC, Critchley SE and Ortel TL. Venous thromboembolism: a public health concern. Am J Prev Med. 2010;38:S495–501. [DOI] [PubMed] [Google Scholar]
- 3.Centers for Disease Control and Prevention (CDC). Venous Thromboembolism (Blood Clots). Data& Statistics, https://www.cdc.gov/ncbddd/dvt/data.html, accessed March 14, 2019.
- 4.Minges KE, Bikdeli B, Wang Y, Kim N, Curtis JP, Desai MM and Krumholz HM. National Trends in Pulmonary Embolism Hospitalization Rates and Outcomes for Adults Aged >/=65 Years in the United States (1999 to 2010). Am J Cardiol. 2015;116:1436–42. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Agarwal S, Clark D 3rd, Sud K, Jaber WA, Cho L and Menon V. Gender Disparities in Outcomes and Resource Utilization for Acute Pulmonary Embolism Hospitalizations in the United States. Am J Cardiol. 2015;116:1270–6. [DOI] [PubMed] [Google Scholar]
- 6.Admon AJ, Seymour CW, Gershengon HB, Wunsch H and Cooke CR. Hospital-level variation in ICU admission and critical care procedures for patients hospitalized for pulmonary embolism. Chest. 2014;146:1452–1461. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7.Wiener RS, Schwartz LM and Woloshin S. Time trends in pulmonary embolism in the United States: evidence of overdiagnosis. Arch Intern Med. 2011;171:831–7. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8.Fang MC, Fan D, Sung SH, Witt DM, Schmelzer JR, Steinhubl SR, Yale SH and Go AS. Validity of Using Inpatient and Outpatient Administrative Codes to Identify Acute Venous Thromboembolism: The CVRN VTE Study. Med Care. 2017;55:el37–el43. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9.Coleman CI, Kohn CG, Crivera C, Schein JR and Peacock WF. Validation of the multivariable In-hospital Mortality for PulmonAry embolism using Claims daTa (IMPACT) prediction rule within an all-payer inpatient administrative claims database. BMJ Open. 2015;5:e009251. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10.Menzin J, Preblick R, Friedman M, Menzin J, Frean M and Jacqueline Kwong W. Treatment patterns and outcomes among hospitalized patients with venous thromboembolism in the United States: an analysis of electronic health records data. Hosp Pract (1995). 2014;42:59–74. [DOI] [PubMed] [Google Scholar]
- 11.Huang W, Goldberg RJ, Cohen AT, Anderson FA, Kiefe CI, Gore JM and Spencer FA. Declining Long-term Risk of Adverse Events after First-time Community-presenting Venous Thromboembolism: The Population-based Worcester VTE Study (1999 to 2009). Thromb Res. 2015;135:1100–6. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12.Nichols GA, Desai J, Elston Lafata J, Lawrence JM, O’Connor PJ, Pathak RD, Raebel MA, Reid RJ, Selby JV, Silverman BG, Steiner JF, Stewart WF, Vupputuri S, Waitzfelder B and Group S-DS. Construction of a multisite DataLink using electronic health records for the identification, surveillance, prevention, and management of diabetes mellitus: the SUPREME-DM project. Prev Chronic Dis. 2012;9:E110. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13.Ross TR, Ng D, Brown JS, Pardee R, Hombrook MC, Hart G and Steiner JF. The HMO Research Network Virtual Data Warehouse: A Public Data Model to Support Collaboration. EGEMS (Wash DC). 2014;2:1049. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14.Charlson ME, Pompei P, Ales KL and MacKenzie CR. A new method of classifying prognostic comorbidity in longitudinal studies: development and validation. J Chronic Dis. 1987;40:373–83. [DOI] [PubMed] [Google Scholar]
- 15.Stein PD, Matta F and Hughes MJ. In-Hospital Mortality with Deep Venous Thrombosis. Am J Med. 2017;130:596–600. [DOI] [PubMed] [Google Scholar]
- 16.Tsai J, Grosse SD, Grant AM, Hooper WC and Atrash HK. Trends in in-hospital deaths among hospitalizations with pulmonary embolism. Arch Intern Med. 2012;172:960–1. [DOI] [PubMed] [Google Scholar]
- 17.Pollack CV, Schreiber D, Goldhaber SZ, Slattery D, Fanikos J, O’Neil BJ, Thompson JR, Hiestand B, Briese BA, Pendleton RC, Miller CD and Kline JA. Clinical characteristics, management, and outcomes of patients diagnosed with acute pulmonary embolism in the emergency department: initial report of EMPEROR (Multicenter Emergency Medicine Pulmonary Embolism in the Real World Registry). J Am Coll Cardiol. 2011;57:700–6. [DOI] [PubMed] [Google Scholar]
- 18.Jimenez D, de Miguel-Diez J, Guijarro R, Trujillo-Santos J, Otero R, Barba R, Muriel A, Meyer G, Yusen RD and Monreal M. Trends in the Management and Outcomes of Acute Pulmonary Embolism: Analysis From the RIETE Registry. J Am Coll Cardiol. 2016;67:162–170. [DOI] [PubMed] [Google Scholar]
- 19.Tagalakis V, Patenaude V, Kahn SR and Suissa S. Incidence of and mortality from venous thromboembolism in a real-world population: the Q-VTE Study Cohort. Am J Med. 2013;126:832.e13–21. [DOI] [PubMed] [Google Scholar]
- 20.Larsen TB, Lip GY and Gorst-Rasmussen A. Anticoagulant therapy after venous thromboembolism and 10-year mortality. Int J Cardiol. 2016;208:72–8. [DOI] [PubMed] [Google Scholar]
- 21.Smith SB, Geske JB, Kathuria P, Cuttica M, Schimmel DR, Courtney DM, Waterer GW and Wunderink RG. Analysis of National Trends in Admissions for Pulmonary Embolism. Chest. 2016;150:35–45. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22.Zondag W, Mos IC, Creemers-Schild D, Hoogerbrugge AD, Dekkers OM, Dolsma J, Eijsvogel M, Faber LM, Hofstee HM, Hovens MM, Jonkers GJ, van Kralingen KW, Kruip MJ, Vlasveld T, de Vreede MJ and Huisman MV. Outpatient treatment in patients with acute pulmonary embolism: the Hestia Study. J Thromb Haemost. 2011;9:1500–7. [DOI] [PubMed] [Google Scholar]
- 23.Kovacs MJ, Hawel JD, Rekman JF and Lazo-Langner A. Ambulatory management of pulmonary embolism: a pragmatic evaluation. J Thromb Haemost. 2010;8:2406–11. [DOI] [PubMed] [Google Scholar]
- 24.Zondag W, Kooiman J, Klok FA, Dekkers OM and Huisman MV. Outpatient versus inpatient treatment in patients with pulmonary embolism: a meta-analysis. Eur Respir J. 2013;42:134–44. [DOI] [PubMed] [Google Scholar]
- 25.Aujesky D, Roy PM, Verschuren F, Righini M, Osterwalder J, Egloff M, Renaud B, Verhamme P, Stone RA, Legall C, Sanchez O, Pugh NA, N’Gako A, Cornuz J, Hugli O, Beer HJ, Perrier A, Fine MJ and Yealy DM. Outpatient versus inpatient treatment for patients with acute pulmonary embolism: an international, open-label, randomised, non-inferiority trial. Lancet. 2011;378:41–8. [DOI] [PubMed] [Google Scholar]
- 26.Jimenez D and Yusen RD. Outpatient therapy for acute symptomatic pulmonary embolism diagnosed in the emergency department: Time to improve the evidence base. Thromb Res. 2017. [DOI] [PubMed] [Google Scholar]
- 27.Klil-Drori AJ, Coulombe J, Suissa S, Hirsch A and Tagalakis V. Temporal trends in outpatient management of incident pulmonary embolism and associated mortality. Thromb Res. 2017. [DOI] [PubMed] [Google Scholar]
- 28.Stein PD, Matta F, Hughes PG, Hourmouzis ZN, Hourmouzis NP, White RM, Ghiardi MM, Schwartz MA, Moore HL, Bach JA, Schweiss RE, Kazan VM, Kakish EJ, Keyes DC and Hughes MJ. Home Treatment of Pulmonary Embolism in the Era of Novel Oral Anticoagulants. Am J Med. 2016;129:974–7. [DOI] [PubMed] [Google Scholar]
- 29.Wiener RS, Schwartz LM and Woloshin S. When a test is too good: how CT pulmonary angiograms find pulmonary emboli that do not need to be found. BMJ. 2013;347:f3368. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 30.Spencer FA, Gore JM, Lessard D, Douketis JD, Emery C and Goldberg RJ. Patient outcomes after deep vein thrombosis and pulmonary embolism: the Worcester Venous Thromboembolism Study. Arch Intern Med. 2008;168:425–30. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 31.Sogaard KK, Schmidt M, Pedersen L, Horvath-Puho E and Sorensen HT. 30-year mortality after venous thromboembolism: a population-based cohort study. Circulation. 2014;130:829–36. [DOI] [PubMed] [Google Scholar]
- 32.Heit JA, Silverstein MD, Mohr DN, Petterson TM, O'Fallon WM and Melton LJ 3rd. Predictors of survival after deep vein thrombosis and pulmonary embolism: a population-based, cohort study. Arch Intern Med. 1999;159:445–53. [DOI] [PubMed] [Google Scholar]
- 33.Mahan CE, Borrego ME, Woersching AL, Federici R, Downey R, Tiongson J, Bieniarz MC, Cavanaugh BJ and Spyropoulos AC. Venous thromboembolism: annualised United States models for total, hospital-acquired and preventable costs utilising long-term attack rates. Thromb Haemost. 2012;108:291–302. [DOI] [PubMed] [Google Scholar]
- 34.Fanikos J, Rao A, Seger AC, Carter D, Piazza G and Goldhaber SZ. Hospital costs of acute pulmonary embolism. Am J Med. 2013;126:127–32. [DOI] [PubMed] [Google Scholar]
- 35.Kearon C, Akl EA, Ornelas J, Blaivas A, Jimenez D, Bounameaux H, Huisman M, King CS, Morris TA, Sood N, Stevens SM, Vintch JR, Wells P, Woller SC and Moores L. Antithrombotic Therapy for VTE Disease: CHEST Guideline and Expert Panel Report. Chest. 2016;149:315–52. [DOI] [PubMed] [Google Scholar]
- 36.Medical Expenditure Panel Survey (MEPS) 2005-2015 Agency for Healthcare Research and Quality (AHRQ), Rockville, MD: ClinCalc DrugStats Database version 18.0. http://clincalc.com/DrugStats/Top300Drugs.aspx, accessed March 14, 2019. [Google Scholar]
- 37.Lippi G, Mattiuzzi C, Cervellin G and Favaloro EJ. Direct oral anticoagulants: analysis of worldwide use and popularity using Google Trends. Ann Transl Med. 2017;5:322. [DOI] [PMC free article] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
Supplementary Figure. Flow diagram of how 5,497 patients with valid, incident venous thromboembolism were identified for the CVRN VTE cohort