Abstract
Background:
Many studies on hospital-associated venous thromboembolism (VTE) do not specifically assess the risk factors for events after discharge.
Objectives:
This study aimed to identify risk factors for postdischarge venous thromboembolism (PD-VTE) after a medical hospitalization.
Methods:
Patients discharged from a medical hospitalization at the University of Vermont Medical Center between January 2010 and September 2019 were followed up for inpatient and outpatient VTE events for up to 90 days. Age-adjusted, sex-adjusted, race-adjusted, and length of stay–adjusted Cox models estimated the hazard ratios (HR) and 95% CIs for potential risk factors for PD-VTE.
Results:
Among 22 599 admissions, there were 180 PD-VTE events (90-day cumulative incidence of 0.8%). The median time from discharge to PD-VTE was 29 days. Of previously identified in-hospital VTE risk factors, only history of VTE (HR, 3.34; 95% CI, 2.26-4.93) and active cancer (HR, 3.13; 95% CI, 2.31-4.23) were associated with increased risk of PD-VTE. Compared to <2-day hospital stays, longer hospital stays (HR, 1.84; 95% CI, 1.23-2.75, for 6-10 days and HR, 1.64; 95% CI, 0.93, 2.90, for 11+ days) were associated with increased risk of PD-VTE. Both mortality (HR, 1.22; 95% CI, 1.06-1.41 per SD higher) and readmission (HR, 1.30; 95% CI, 1.13-1.50, per SD higher) Elixhauser comorbidity indices were associated with increased risk of PD-VTE.
Conclusions:
Risk factors for VTE after discharge from medical hospitalization differ from risk factors for in-hospital VTE. These data support that characteristics of the hospitalization influence PD-VTE risk, and if indicated, PD-VTE risk should be assessed at discharge.
Keywords: comorbidity, length of stay, outpatients, patient discharge, venous thromboembolism
1 ∣. INTRODUCTION
Venous thromboembolism (VTE), including deep venous thrombosis and pulmonary embolism, occurs in 1 to 2 of every 1000 US adults annually. Approximately a third of VTE events are associated with hospitalization, occurring either while admitted (hospital-acquired) or up to 90 days after discharge (postdischarge [PD]) [1]. There is little available information about the incidence and specific risk factors for PD-VTE [2].
Up to 70% of hospital-associated VTE events (events occurring during hospitalization or within 90 days of discharge) occur after discharge [3]. Recognizing an opportunity to prevent VTE, there have been several trials that specifically investigated PD-VTE pharmacologic prophylaxis [4-8]. These trials did not meet their primary outcome due to lower than anticipated incidence of PD-VTE as well as higher than anticipated incidence of PD bleeding [4-8]. Interestingly, the 2 studies that randomized participants at discharge rather than admission demonstrated a borderline benefit in VTE prophylaxis outweighing bleeding risks [6,8].
The equivocal results of randomized trials of PD-VTE prophylaxis demonstrate the need for a better understanding of the risk factors for PD-VTE in medical patients. The objective of this study was to describe risk factors for PD-VTE and compare them with in-hospital VTE risk factors. We hypothesized that risk factors occurring during hospitalization and characteristics of the hospitalization would impact PD-VTE risk.
2 ∣. METHODS
2.1 ∣. Population
The cohort consisted of all patients aged 18 years or older who accessed a University of Vermont (UVM) health network primary care clinician and were discharged alive from medical hospitalization at the UVM Medical Center between 2010 and 2019. Individuals who were admitted for VTE or who had a VTE while hospitalized were excluded. UVM Medical Center is a 540-bed tertiary acute care hospital in Burlington, Vermont, which serves a referral population of 1 000 000 individuals and is the only hospital and imaging center in Chittenden County, Vermont (population of 160 000). A medical admission, congruent with Centers for Medicare and Medicaid Services definitions, was defined as spending at least 1 midnight in the hospital on the general medicine, family medicine, cardiology, hematology, oncology, or medical intensive care unit services under observation or inpatient status. The patients were followed up from the day of discharge for 90 days or until the first validated VTE event occurred: death, hospital readmission, loss to follow-up, or change to a non-UVM Medical Center primary care provider, or August 31, 2019, whichever occurred first. If an individual was readmitted to the hospital during 90-day follow-up period, the timeline was reset, and the individual was followed up anew for up to 90 days after discharge from the latter hospitalization. UVM Medical Center uses an Epic Systems Corporation electronic health record (EHR) for inpatient and outpatient care.
2.2 ∣. Outcome ascertainment
VTE was defined as deep vein thrombosis (either proximal or distal) of the upper or lower extremities or pulmonary embolism. PD-VTE consisted of outpatient diagnosed and treated VTE events and VTE events identified at the time of a readmission to the hospital up to 90 days after a medical discharge. Present on admission VTE, VTE occurring during hospitalization, and VTE diagnosed and treated as an outpatient were defined by previously validated algorithms using a combination of diagnosis codes, medications, and imaging studies [9,10]. Supplementary Table S1 has a detailed list outlining these factors and how they were abstracted from EHR.
2.3 ∣. Risk factor ascertainment
Risk factors known on the day of admission or before are referred to as history (such as a history of VTE). Baseline characteristics of the population were assessed using EHR data known at the time of admission. Race was assessed from that reported in EHR, with race/ethnicity recorded in a single field. Hospital length of stay (in days) was defined as discharge date minus admission date plus 1. Any portion of a day spent in the hospital counted as an entire day. Risk factors from a recently validated hospital-acquired VTE model were defined at admission as previously reported and included history of VTE, renal dysfunction, active cancer, anemia, elevated red cell distribution width (RDW), hyponatremia, and malnutrition [11]. The Elixhauser mortality and readmission indices were calculated based on discharge conditions using previously described and publicly available algorithms [12]. These indices are well validated and promoted by the Agency for Healthcare Research and Quality in the United States; the indices use diagnosis codes to predict the in-hospital mortality and readmission risk. Use of anticoagulant medications was defined at both admission and discharge: the former was defined based on the maximum administered dose during the day of or day after admission and was categorized as none, prophylactic, or full, as defined previously [11], while the latter was captured using the medication reconciliation process in EHR.
2.4 ∣. Statistical analyses
Baseline characteristics were summarized using frequencies and percentages or means and SDs, as appropriate. The incidence rate of PD-VTE was calculated as the number of PD-VTE events divided by the person-time at risk in the overall 90-day follow-up, as well as separately across different periods of follow-up (ie, 1-4, 5-14, 15-30, 31-60, and 61-90 days after discharge) in order to elucidate the time to event distribution of PD-VTE risk. Additionally, Cox proportional-hazards models were used to estimate hazard ratios (HRs) between risk factors of interest and PD-VTE while adjusting for age, sex, race, and length of stay. Each of the Elixhauser mortality and readmission indices were modeled per SD increment of the composite score [12]. Predicted hospital-acquired VTE risk from a previously developed and validated risk model [11]. Predicted risk was calculated from the risk assessment model and tertiles of risk calculated assuming a normal distribution of the predicted risk. To address data missingness for the hospital-acquired VTE risk model, variables with missing values were imputed using multiple imputation by chained equations, as has been described previously [11,13,14]. Statistical analyses were performed using SAS, version 9.4 (SAS Institute) and R, version 4.2.2 (R Foundation for Statistical Computing). The research reported conformed to the tenants of the Declaration of Helsinki and was determined exempt by the University of Vermont Institutional Review Board based on exemption category 4 under the 2018 Common rule [15].
3 ∣. RESULTS
Figure presents the selection of the study cohort. Among 193 974 admissions, 62 468 were to medical services, and 25 257 had a primary care provider within the UVM Health Network. We excluded 1980 admissions in which VTE was present on admission or acquired while hospitalized and 678 admissions for which the patient did not survive the admission. A total of 22 599 admissions were included in the analysis. Table 1 presents the characteristics of the study population including admitting service, length of stay, age, sex, and race.
FIGURE.
Population flow chart. VTE, venous thromboembolism.
TABLE 1.
Population characteristics.
| Characteristic | Overall |
|---|---|
| No. of admissions (overall) | 22 599 |
| No. of admissions by service | |
| General medicine | 15 234 (67.4) |
| Hematology/oncology | 1238 (5.5) |
| Cardiology | 4880 (21.6) |
| Pulmonary/critical care | 1247 (5.5) |
| Any intensive unit stay | 2841 (12.6) |
| Demographics | |
| Age (y) | 66.9 ± 16.8 |
| Female | 12 239 (54.2) |
| Race/ethnicity | |
| White | 21 416 (94.8) |
| Black | 263 (1.2) |
| Hispanic | 7 (<0.1) |
| Asian | 166 (0.7) |
| Native American | 146 (0.6) |
| Native Hawaiian/other Pacific Islander | 14 (<0.1) |
| Other/unknown | 587 (2.6) |
| BMI (kg/m2) | |
| 0-18.5 | 966 (4.3) |
| 18.5-25 | 6874 (30.4) |
| 25-30 | 6418 (28.4) |
| 30-35 | 4029 (17.8) |
| 35+ | 4008 (17.7) |
| Missing | 304 (1.4) |
| Length of hospital stay (d) | |
| 1-2 | 10 872 (48.1) |
| 3-5 | 7355 (32.6) |
| 6-10 | 3042 (13.5) |
| ≥11 | 1330 (5.9) |
| Maximum anticoagulation on day of or day after admission | |
| None | 7315 (32.4) |
| Prophylactic | 10 370 (45.9) |
| Full | 4914 (21.7) |
| Anticoagulation on discharge (yes) | 1548 (6.9) |
| Elixhauser comorbidity indices | |
| Mortality | 6.4 ± 13.4 |
| Readmission | 8.2 ± 7.4 |
| Hospital-acquired VTE risk factors | |
| History of VTE | 1320 (5.8) |
| Renal dysfunctiona | 2017 (8.9) |
| Active cancer | 6418 (28.4) |
| Anemiaa | 10 277 (45.5) |
| Elevated RDWa | 8216 (36.4) |
| Hyponatremiaa | 5219 (23.1) |
| Malnutritiona | 876 (3.9) |
| Predicted hospital-acquired VTE risk tertiles | |
| Tertile 1 (<0.18%) | 6555 (29.0) |
| Tertile 2 (0.18%-0.33%) | 6902 (30.5) |
| Tertile 3 (>0.33%) | 6705 (29.7) |
| Missing values | 2437 (10.8) |
Values are n (%) or mean ± SD.
BMI, body mass index; RDW, red cell distribution width; VTE, venous thromboembolism.
Risk factors defined per Zakai et al [11]. Renal dysfunction was defined as a creatinine of 2.0 mg/L or higher, or on dialysis. Anemia defined as a hemoglobin <13.6 g/dL in male and <12.0 g/dL in female. An elevated RDW was defined as an RDW of ≥14.7%. Hyponatremia was defined as <136 mEq/L. Malnutrition was defined as a BMI <18.5 kg/m2.
Incidence rates of PD-VTE, overall and by timing of follow-up, are presented in Table 2. Overall, 180 PD-VTE events occurred in 90 days after discharge (cumulative incidence of 0.8%), corresponding to an incidence rate of 38.5 (95% CI, 32.9-44.2) events per 1000 person-years. The VTE rate was highest at 4 days after discharge (89.8; 95% CI, 52.3-127.3, per 1000 person-years) and gradually decreased to 27.1 (95% CI, 18.6-35.6) per 1000 person-years at 61 to 90 days after discharge. The median time from discharge to PD-VTE was 29 days.
TABLE 2.
Incidence of PD-VTE.
| Time in days | PD-VTE, n (%) | Person-years of follow-up |
PD-VTE rate per 1000 person-years (95% CI) |
|---|---|---|---|
| Overall (1-90 d) | 180 (100) | 4672 | 38.5 (32.9-44.2) |
| By timing of follow-up (d) | |||
| 1-4 | 22 (12.2) | 245 | 89.8 (52.3-127.3) |
| 5-14 | 37 (20.6) | 582 | 63.6 (43.1-84.1) |
| 15-30 | 35 (19.4) | 874 | 40.0 (26.8-53.3) |
| 31-60 | 47 (26.1) | 1531 | 30.7 (21.9-39.5) |
| 61-90 | 39 (21.7) | 1440 | 27.1 (18.6-35.6) |
PD, postdischarge; VTE, venous thromboembolism.
The prevalence of the investigated risk factors in people with and without PD-VTE and their associations with PD-VTE are shown in Table 3. Older age, but not sex or race, was associated with higher PD-VTE risk. Compared with individuals aged <45 years, those aged 45 to 64 years (HR, 1.89; 95% CI, 1.02-3.51), 65 to 74 years (HR, 1.99; 95% CI, 1.06-3.75), and ≥75 years (HR, 1.23; 95% CI, 0.66-2.29) had higher risk of PD-VTE. A hospital length of stay of ≥6 days was associated with increased risk of PD-VTE (HR, 1.84; 95% CI, 1.23-2.75 for 6-10 days, and HR, 1.64; 95% CI, 0.93-2.90, for ≥11 days) compared with a hospitalization of <2 days. The Elixhauser comorbidity indices were associated with PD-VTE risk: each SD increment in the mortality index was associated with 22% (95% CI, 6%-41%) increased hazard of PD-VTE, and each SD increment in readmission index was associated with 30% (95% CI, 13%−50%) increased hazard of PD-VTE. Anticoagulation, whether administered on the day of or day after admission or provided at discharge, was not associated with PD-VTE risk (Table 3).
TABLE 3.
Age-adjusted, sex-adjusted, and race-adjusted association of risk factors with PD-VTE.
| Time in days | No PD-VTE (n = 22 419), n (%) | PD-VTE (n = 180), n (%) | HR (95% CI) | P-value | |||
|---|---|---|---|---|---|---|---|
| Age (y) | |||||||
| <45 | 2365 (10.6) | 12 (6.7) | Ref | - | |||
| 45-64 | 6651 (29.7) | 63 (35.0) | 1.89 (1.02-3.51) | .044 | |||
| 65-74 | 4815 (21.5) | 49 (27.2) | 1.99 (1.06-3.75) | .033 | |||
| ≥75 | 8588 (38.3) | 56 (31.1) | 1.23 (0.66-2.29) | .524 | |||
| Sex | |||||||
| Female | 12 140 (54.2) | 99 (55.0) | Ref | - | |||
| Male | 10 279 (45.8) | 81 (45.0) | 0.94 (0.70-1.26) | .684 | |||
| Race | |||||||
| White | 21 241 (94.7) | 175 (97.2) | Ref | - | |||
| Black | 263 (1.2) | - | 0.51 (0.21-1.25) | .141 | |||
| Other/Hispanic/unknown | 915 (4.1) | 5 (2.8) | 0.45 (0.14-1.43) | .175 | |||
| Body mass index (kg/m2) | |||||||
| 0-18.5 | 963 (4.3) | 3 (1.7) | Ref | - | |||
| 18.5-25 | 6829 (30.5) | 45 (25.0) | 1.21 (0.81-1.81) | .345 | |||
| 25-30 | 6366 (28.4) | 52 (28.9) | 1.45 (0.94-2.22) | .09 | |||
| 30-35 | 3988 (17.8) | 41 (22.8) | 1.27 (0.82-1.98) | .288 | |||
| 35+ | 3970 (17.7) | 38 (21.1) | 0.49 (0.07-3.58) | .486 | |||
| Missing | 303 (1.4) | 1 (0.6) | - | - | |||
| Length of stay (d) | |||||||
| 1-2 | 10 796 (48.2) | 76 (42.2) | Ref | - | |||
| 3-5 | 7300 (32.6) | 55 (30.6) | 1.13 (0.80-1.60) | .494 | |||
| 6-10 | 3007 (13.4) | 35 (19.4) | 1.84 (1.23-2.75) | .003 | |||
| ≥11 | 1316 (5.9) | 14 (7.8) | 1.64 (0.93-2.90) | .09 | |||
| Any intensive unit stay | 2817 (12.6) | 24 (13.3) | 0.95 (0.61-1.47) | .812 | |||
| Elixhauser comorbidity indicesa | |||||||
| Mortality, mean ± SD | 6.4 ± 13.4 | 9.1 ± 14.5 | 1.22 (1.06-1.41) | .005 | |||
| Readmission, mean ± SD | 8.2 ± 7.4 | 10.2 ± 7.3 | 1.30 (1.13-1.50) | <.001 | |||
| Hospital-acquired VTE risk factors | |||||||
| History of VTE | 1289 (5.7) | 31 (17.2) | 3.34 (2.26-4.93) | <.001 | |||
| Renal dysfunction | 1998 (8.9) | 19 (10.6) | 1.29 (0.79-2.09) | .307 | |||
| Active cancer | 6323 (28.2) | 95 (52.8) | 3.13 (2.31-4.23) | <.001 | |||
| Anemia | 10 196 (45.5) | 81 (45.0) | 1.20 (0.86-1.66) | .281 | |||
| Elevated RDW | 8147 (36.3) | 69 (38.3) | 1.31 (0.94-1.82) | .108 | |||
| Hyponatremia | 5174 (23.1) | 45 (25.0) | 1.18 (0.83-1.67) | .362 | |||
| Malnutrition | 873 (3.9) | 3 (1.7) | 0.39 (0.12-1.22) | .105 | |||
| Maximum anticoagulation by hospital day 2 of admission | |||||||
| None | 7261 (32.4) | 54 (30.0) | Ref | - | |||
| Prophylactic | 10 280 (45.9) | 90 (50.0) | 1.14 (0.81-1.60) | .461 | |||
| Full | 4878 (21.8) | 36 (20.0) | 0.95 (0.62-1.45) | .809 | |||
| Anticoagulation at discharge (yes vs no) | 1537 (6.9) | 11 (6.1) | 0.85 (0.46-1.57) | .607 | |||
| Predicted hospital-acquired VTE risk at admission | |||||||
| Tertile 1 | 6519 (29.1) | 36 (20.0) | Ref | - | |||
| Tertile 2 | 6854 (30.6) | 48 (26.7) | 1.30 (0.84-2.01) | .243 | |||
| Tertile 3 | 6642 (29.6) | 63 (35.0) | 1.79 (1.17-2.74) | .007 | |||
| Missing values | 2404 (10.7) | 33 (18.3) | 2.59 (1.61-4.17) | <.001 | |||
| Predicted hospital-acquired VTE risk at admission with multiple imputation | |||||||
| Tertile 1 | 7335 (32.7) | 41 (22.8) | Ref | ||||
| Tertile 2 | 7627 (34.0) | 62 (34.4) | 1.51 (1.01-2.42) | .044 | |||
| Tertile 3 | 7457 (33.3) | 77 (42.8) | 1.98 (1.34-2.92) | .001 | |||
HR, hazards ratio; PD, postdischarge; RDW, red cell distribution width; VTE, venous thromboembolism.
HRs are scaled per SD increment in Elixhauser comorbidity index.
Among 7 risk factors from a previously validated hospital-acquired VTE risk assessment model (history of VTE, renal dysfunction, active cancer, anemia, elevated RDW, hyponatremia, and malnutrition) [11], we observed that only 2—history of VTE (HR, 3.34; 95% CI, 2.26-4.93) and active cancer (HR, 3.13; 95% CI, 2.31-4.23)—were associated with increased PD-VTE risk. The top tertile (HR, 1.98; 95% CI, 1.34-2.92) and the middle tertile (HR, 1.51; 95% CI, 1.01-2.52) of predicted VTE risk were associated with increased risk of PD-VTE compared with the lowest tertile of risk (Table 3).
4 ∣. DISCUSSION
In this study of 22 599 medical admissions from the UVM Health Network, there were 180 total PD-VTEs within 90 days of discharge from medical hospitalization, corresponding to an incidence rate of 38.5 events per 1000 person-years and a cumulative incidence of 0.8%. The incidence rate was highest within the first 4 days after discharge, and the majority of PD-VTE events occurred within the first month after discharge. Among the investigated risk factors, older age, a history of VTE, active cancer, hospitalization length of stay 6 days or greater, and the Elixhauser mortality and readmission comorbidity indexes were associated with increased risk of PD-VTE. There was no association of anticoagulation at admission or at discharge with PD-VTE risk.
This study’s findings extend previous research on hospital-associated VTE. In a large sample of medical admissions within a managed care consortium from Northern California and linked administrative data from New South Wales, Australia, more than 70% of hospital-associated VTE events occurred after discharge, similar to the percentage in the current population from a previous analysis [3,10,16]. Previous models exploring risk factors for PD-VTE have used a combined outcome of PD-VTE and hospital-acquired VTE, making it difficult to decipher the risk factors specific for PD-VTE [17-20]. Our study extends this previous work by looking specifically at risk factors for only PD-VTE. We identified previous VTE and active cancer as risk factors for PD-VTE, which have also been shown to be risk factors for hospital-associated VTE risk in previous studies [11,17,21]. However, while previous studies combined risk factors for both hospital-acquired VTE and PD-VTE, we found risk factors for PD-VTE differ from the broader category of hospital-acquired VTE with older age, hospital length of stay, and Elixhauser mortality and readmission indexes, exhibiting associations with higher risk of PD-VTE.
Other studies assessing PD-VTE risk factors have shown differing results to this study. The Intermountain Risk Score (IMRS) for PD-VTE found older age and selected blood cell and chemistry test as predictive of PD-VTE [22]. Several of the predictive factors in the IMRS model—namely RDW, lower red blood cell (anemia), hyponatremia, and creatinine [22]—were tested and were not associated with PD- VTE in this study. The IMRS analysis included nearly 300 PD-VTE events, so was larger than this study, which could explain why some risk factors in this study did not show an association. However, the IMRS score performed poorly when validated in an independent population, suggesting laboratory studies and age alone are not sufficient for assessing PD-VTE risk [23].
Our study has strengths and weaknesses. We used longitudinal EHR data to identify a population cohort and validated algorithms to identify VTE events [9,10]. We also focused on PD-VTE events and assessed risk factors at the time of discharge, rather than solely at the time of admission, as most studies to date have done. UVM Medical Center has a diverse mix of community and tertiary care patients. While the hospital serves patients from both urban and rural areas, there is less racial/ethnic diversity than in the general US inpatient population. We did not specifically develop a risk assessment model for PD-VTE, because the observed number of PD-VTE events would have precluded a robust model development and validation. However, these data support using risk factors present before admission, at admission, and at discharge for future PD-VTE risk assessment model development strategies. Previous trials of PD-VTE prophylaxis did not meet their efficacy end points largely due to overestimating PD-VTE incidence and underestimating PD bleeding incidence [4-8]. The lack of information surrounding PD bleeding risk is a limitation of this study, but we performed an analysis on PD bleeding risk factors, which will aid in addressing the risk versus benefit interventions to reduce PD-VTE [24]. When combined with an understanding of PD bleeding risk, the estimates of absolute PD-VTE incidence from this study also serve to provide a foundation for determining whether a target population may benefit from PD pharmacologic VTE prophylaxis [25].
In conclusion, we identified demographic factors (older age), factors present on admission (active cancer, history of VTE, and Elixhauser readmission and mortality indices), and hospitalization- related factors available on discharge (increased hospital length of stay) as being associated with risk of PD-VTE. While there has not yet been a clearly identified population for whom the benefit of PD pharmacologic thromboprophylaxis outweighs the risks [26-28], these study findings are a first step toward future trials assessing benefit from targeted PD-VTE prophylaxis.
Supplementary Material
The online version contains supplementary material available at https://doi.org/10.1016/j.jtha.2025.04.005
ACKNOWLEDGMENTS
We appreciate and acknowledge the support and expertise of Mr. Michael Gianni, Senior Measurement Analyst, and the Data Management Office at the University of Vermont Health Network. We appreciate the support and guidance of investigators from the Study Design and Molecular Epidemiology Core of the Vermont Center for Cardiovascular and Brain Health. The sponsoring institution had no role in the analytic approach or review of data. The views presented in this study represent those of the authors and not necessarily of the National Institutes of Health. This study was funded by grant R01-HL141290 (N.A. Zakai) from the National Heart, Lung, and Blood Institute, National Institutes of Health, Bethesda, MD, USA. The sponsoring institution had no involvement in the analytic approach or data review. The views presented here represent those of the authors and not necessarily of the National Institutes of Health.
Funding information
National Heart, Lung, and Blood Institute, National Institutes of Health, Grant/Award Number: R01-HL141290 (to N.A.Z.) National Institute of General Medical Sciences, National Institutes of Health, Grant/Award Number: P20-GM135007.
Footnotes
DECLARATION OF COMPETING INTERESTS
There are no competing interests to disclose.
REFERENCES
- [1].Beckman MG, Hooper WC, Critchley SE, Ortel TL. Venous thromboembolism: a public health concern. Am J Prev Med. 2010;38(4 Suppl):S495–501. [DOI] [PubMed] [Google Scholar]
- [2].Virani SS, Alonso A, Aparicio HJ, Benjamin EJ, Bittencourt MS, Callaway CW, Carson AP, Chamberlain AM, Cheng S, Delling FN, Elkind MSV, Evenson KR, Ferguson JF, Gupta DK, Khan SS, Kissela BM, Knutson KL, Lee CD, Lewis TT, Liu J, et al. American Heart Assoc. Council on Epidemiology and Prevention Statistics Committee and Stroke Statistics Subcommittee. Heart disease and stroke statistics-2021 update: a report from the American Heart Association. Circulation. 2021;143:e254–743. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [3].Jordan Bruno X, Koh I, Lutsey PL, Walker RF, Roetker NS, Wilkinson K, Smith NL, Plante TB, Repp AB, Holmes CE, Cushman M, Zakai NA. Venous thrombosis risk during and after medical and surgical hospitalizations: the medical inpatient thrombosis and hemostasis (MITH) study. J Thromb Haemost. 2022;20:1645–52. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [4].Goldhaber SZ, Leizorovicz A, Kakkar AK, Haas SK, Merli G, Knabb RM, Weitz JI, ADOPT trial investigators. Apixaban versus enoxaparin for thromboprophylaxis in medically ill patients. N Engl J Med. 2011;365:2167–77. [DOI] [PubMed] [Google Scholar]
- [5].Cohen AT, Harrington RA, Goldhaber SZ, Hull RD, Wiens BL, Gold A, Hernandez AF, Gibson CM. Extended thromboprophylaxis with betrixaban in acutely ill medical patients. N Engl J Med. 2016;375: 534–44. [DOI] [PubMed] [Google Scholar]
- [6].Hull RD, Schellong SM, Tapson VF, Monreal M, Samama MM, Nicol P, Vicaut E, Turpie AGG, Yusen RD. EXCLAIM (Extended Prophylaxis for Venous ThromboEmbolism in Acutely Ill Medical Patients With Prolonged Immobilization) study. Extended-duration venous thromboembolism prophylaxis in acutely ill medical patients with recently reduced mobility: a randomized trial. Ann Intern Med. 2010;153:8–18. [DOI] [PubMed] [Google Scholar]
- [7].Cohen AT, Spiro TE, Büller HR, Haskell L, Hu D, Hull R, Mebazaa A, Merli G, Schellong S, Spyropoulos AC, Tapson V, MAGELLAN investigators. Rivaroxaban for thromboprophylaxis in acutely ill medical patients. N Engl J Med. 2013;368:513–23. [DOI] [PubMed] [Google Scholar]
- [8].Raskob GE, Spyropoulos AC, Zrubek J, Ageno W, Albers G, Elliott CG, Halperin J, Haskell L, Hiatt WR, Maynard GA, Peters G, Spiro T, Steg PG, Suh EY, Weitz JI. The MARINER trial of rivaroxaban after hospital discharge for medical patients at high risk of VTE. Design, rationale, and clinical implications. Thromb Haemost. 2016;115:1240–8. [DOI] [PubMed] [Google Scholar]
- [9].Thomas RM, Wilkinson K, Koh I, Li A, Warren JSA, Roetker NS, Smith NL, Holmes CE, Plante TB, Repp AB, Cushman M, Zakai NA. Development of a computable phenotype using electronic health records for venous thromboembolism in medical inpatients: the Medical Inpatient Thrombosis and Hemostasis study. Res Pract Thromb Haemost. 2023;7:100162. 10.1016/j.rpth.2023.100162 [DOI] [PMC free article] [PubMed] [Google Scholar]
- [10].Wilkinson KS, Sparks AD, Gergi M, Repp AB, Al-Samkari H, Thomas R, Roetker NS, Zakai NA. Validation of the International Medical Prevention Registry on Venous Thromboembolism (IMPROVE) risk scores for venous thromboembolism and bleeding in an independent population. Res Pract Thromb Haemost. 2024;8: 102441. 10.1016/j.rpth.2024.102441 [DOI] [PMC free article] [PubMed] [Google Scholar]
- [11].Zakai NA, Wilkinson K, Sparks AD, Packer RT, Koh I, Roetker NS, Repp AB, Thomas R, Holmes CE, Cushman M, Plante TB, Al- Samkari H, Pishko AM, Wood WA, Masias C, Gangaraju R, Li A, Garcia D, Wiggins KL, Schaefer JK, et al. Development and validation of a risk model for hospital-acquired venous thrombosis: the Medical Inpatients Thrombosis and Hemostasis study. J Thromb Haemost. 2024;22:503–15. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [12].Moore BJ, White S, Washington R, Coenen N, Elixhauser A. Identifying increased risk of readmission and in-hospital mortality using hospital administrative data: the AHRQ Elixhauser Comorbidity Index. Med Care. 2017;55:698–705. [DOI] [PubMed] [Google Scholar]
- [13].Rubin DB. Multiple imputation for nonresponse in surveys. Wiley Series in Probability and Statistics. John Wiley & Sons; 1987. [Google Scholar]
- [14].van Buuren S, Groothuis–Oudshoorn K. mice: multivariate imputation by chained equations in R. J Stat Softw. 2011;45:1–67. [Google Scholar]
- [15].83 FR 28497—Federal Policy for the protection of human subjects: six month delay of the general compliance date of revisions while allowing the use of three burden-reducing provisions during the delay period. https://www.hhs.gov/ohrp/regulations-and-policy/regulations/45-cfr-46/revised-common-rule-regulatory-text/index.html. [accessed May 23, 2023].
- [16].Stubbs JM, Assareh H, Curnow J, Hitos K, Achat HM. Incidence of in-hospital and post-discharge diagnosed hospital-associated venous thromboembolism using linked administrative data. Intern Med J. 2018;48:157–65. [DOI] [PubMed] [Google Scholar]
- [17].Spyropoulos AC, Anderson FA, Fitz Gerald G, Decousus H, Pini M, Chong BH, Zotz RB, Bergmann JF, Tapson V, Froehlich JB, Monreal M, Merli GJ, Pavanello R, Turpie AGG, Nakamura M, Piovella F, Kakkar AK, Spencer FA, IMPROVE Investigators. Predictive and associative models to identify hospitalized medical patients at risk for VTE. Chest. 2011;140:706–14. [DOI] [PubMed] [Google Scholar]
- [18].Woller SC, Stevens SM, Jones JP, Lloyd JF, Evans RS, Aston VT, Elliott CG. Derivation and validation of a simple model to identify venous thromboembolism risk in medical patients. Am J Med. 2011;124:947–54.e2. [DOI] [PubMed] [Google Scholar]
- [19].Spyropoulos AC, Lipardi C, Xu J, Peluso C, Spiro TE, De Sanctis Y, Barnathan ES, Raskob GE. Modified IMPROVE VTE risk score and elevated D-dimer identify a high venous thromboembolism risk in acutely ill medical population for extended thromboprophylaxis. TH Open. 2020;4:e59–65. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [20].Rothberg MB, Hamilton AC, Greene MT, Fox J, Lisheba O, Milinovich A, Gautier TN, Kim P, Kaatz S, Hu B. Derivation and validation of a risk factor model to identify medical inpatients at risk for venous thromboembolism. Thromb Haemost. 2022;122:1231–8. [DOI] [PubMed] [Google Scholar]
- [21].Darzi AJ, Karam SG, Spencer FA, Spyropoulos AC, Mbuagbaw L, Woller SC, Zakai NA, Streiff MB, Gould MK, Cushman M, Charide R, Etxeandia–Ikobaltzeta I, Germini F, Rigoni M, Agarwal A, Morsi RZ, Akl EA, Iorio A, Schünemann HJ. Risk models for VTE and bleeding in medical inpatients: systematic identification and expert assessment. Blood Adv. 2020;4:2557–66. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [22].Woller SC, Stevens SM, Fazili M, Lloyd JF, Wilson EL, Snow GL, Bledsoe JR, Horne BD. Post-discharge thrombosis and bleeding in medical patients: a novel risk score derived from ubiquitous biomarkers. Res Pract Thromb Haemost. 2021;5:e12560. 10.1002/rth2.12560 [DOI] [PMC free article] [PubMed] [Google Scholar]
- [23].Hyder SN, Han HB, Ash S, Horne BD, Stevens SM, Woller SC, Barnes GD. Predicting post-discharge venous thromboembolism and bleeding among medical patients: external validation of a novel risk score utilizing ubiquitous biomarkers. Thromb Res. 2023;227: 45–50. [DOI] [PubMed] [Google Scholar]
- [24].Gergi M, Wilkinson K, Sparks A, Roetker NS, Al-Samkari H, Smith NL, Plante TB, Cushman M, Repp AB, Holmes CE, Martin K, Zakai NA. Anticoagulation and other risk factors for clinically relevant bleeding after medical hospitalization: the Medical Inpatient Thrombosis and Hemostasis Study. J Thromb Haemost. 2025;23:2122–32. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [25].Roberts LN, Arya R. Venous thromboembolism risk assessment models for acutely ill medical patients—back to the drawing board? JAMA Netw Open. 2024;7:e249952. 10.1001/jama-networkopen.2024.9952 [DOI] [PubMed] [Google Scholar]
- [26].Schunemann HJ, Cushman M, Burnett AE, Kahn SR, Beyer– Westendorf J, Spencer FA, Rezende SM, Zakai NA, Bauer KA, Dentali F, Lansing J, Balduzzi S, Darzi A, Morgano GP, Neumann I, Nieuwlaat R, Yepes–Nuñ ez JJ, Zhang Y, Wiercioch W. American Society of Hematology 2018 guidelines for management of venous thromboembolism: prophylaxis for hospitalized and nonhospitalized medical patients. Blood Adv. 2018;2:3198–225. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [27].Rezende SM, Bauer KA, Zakai NA. Thromboprophylaxis in hospitalized and nonhospitalized medical patients: what’s new? Blood Adv. 2023;7:5199–201. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [28].Nicolaides AN, Fareed J, Spyropoulos AC, Kakkar RHL, Antignani PL, Avgerinos E, Baekgaard N, Barber E, Bush RL, Caprini JA, Clarke– Pearson DL, Van Dreden P, Elalami I, Gerotziafas G, Gibbs H, Goldhaber S, Kakkos S, Lefkou E, Labropoulos N, Lopes RD, et al. Prevention and management of venous thromboembolism. International Consensus Statement. Guidelines according to scientific evidence. Int Angiol. 2024;43:1–222. [DOI] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.