Table 1.
Study characteristics
| First author, year (country) | Population (sample size) | Time frame (y) | Mean age, y (SD)* | Study type (number of centers and study design) | Prophylaxis (%) | Outcome (number of events) | Diagnostic methods | Follow-up time | Variables in multivariate logistic regression, when applicable |
|---|---|---|---|---|---|---|---|---|---|
| Venous thromboembolism | |||||||||
| Spyropoulos et al7 (US) | Acutely ill medical patients (N = 15 156) | 2002-2006 | 68 (52-79)* | Prognostic model development (multicenter prospective cohort) | VTE prophylaxis (adjusted-dose warfarin, elastic stockings, LMWH, unfractionated heparin, intermittent pneumatic compression, and aspirin): 44% | Any symptomatic VTE (lower extremity DVT, and PE) (n = 184) | Clinically observed VTE | 92 d | IMPROVE RAM-related factors: |
| The model was adjusted for VTE prophylaxis. | DVT verified by positive venogram or compression ultrasonography. | Age >60 y; prior cancer, prior VTE, ICU/CCU stay, lower limb paralysis, immobility | |||||||
| PE verified by positive lung scan, pulmonary angiogram, or spiral CT scan. | Known thrombophilia | ||||||||
| Fatal PE was defined as PE diagnosed at autopsy or, in the absence of autopsy, when PE was considered the most likely cause of death. | |||||||||
| Mahan et al25 (Canada) | Acute medical patients (N = 417. VTE cases: n = 139; non-VTE controls: n = 278) | 2005-2011 | Cases: 68; controls: 65 | External validation (multicenter retrospective case-control) | VTE prophylaxis, 0% | Any symptomatic VTE (lower extremity DVT, and PE; n = 139) | PE, verified by a positive pulmonary angiogram, spiral computed tomography, or high-probability ventilation/perfusion scan or at autopsy. | 92 d | IMPROVE RAM-related factors: |
| Lower extremity DVT verified by positive compression ultrasonography, computed tomography, or magnetic resonance imaging, or at autopsy. | Previous VTE; known thrombophilia; lower limb paralysis; current cancer; immobilization ≥7; ICU/CCU stay; age >60 y | ||||||||
| Rosenberg et al26 (US) | Medical patients (N = 539: VTE cases: n = 135 and non-VTE controls: n = 404) | 2009-2013 | 67 | External validation (multicenter retrospective case-control) | Any prophylaxis in VTE cases: 49% (of those 44% pharmacological VTE and 5% mechanical) | Any symptomatic VTE (n = 135) | VTE events identified using ICD-9 codes | Within 90 d following the index admission | IMPROVE RAM-related factors: same as those in Spyropoulos et al7 |
| Any prophylaxis in non-VTE controls: 45% (of those 40% pharmacological VTE and 5% mechanical) | |||||||||
| The results were essentially unchanged when the cases and controls were stratified into groups that received VTE prophylaxis, including pharmacological prophylaxis during hospitalization and those that did not. | |||||||||
| Zakai et al27 (US) | Patients admitted to medical services (N = 900: VTE cases: 299; and non-VTE controls: 601) | 2002-2009 | cases: 6317; controls: 6615 | Prognostic model development (single-center retrospective case-control) | Pharmacological prophylaxis in VTE cases: 64.6% and in non-VTE controls: 62.2%; | Any symptomatic VTE (upper and lower extremity DVT and PE; n = 299) | VTE events identified using ICD-9 VTE discharge codes. | Discharge or transfer from medical service | Venous thrombosis prophylaxis (mechanical; pharmacologic); demographics (age, sex, BMI); medical history (myocardial infarction, COPD, diabetes, chronic kidney disease); conditions active on admission (fever, COPD, pneumonia, any infection). |
| mechanical prophylaxis in VTE cases: 31.8% and in non-VTE controls: 27.6%; | Codes confirmed by clinician review. | ||||||||
| the model was adjusted for both mechanical and pharmacological prophylaxis. | Records were reviewed by a research nurse and all hospital-acquired VTE cases and 20% of noncases were reviewed by a physician. | ||||||||
| Zakai et al28 (US) | Medical patients (N = 188: VTE cases: 65 and non-VTE controls: 123) | 2000-2002 | 68 | External validation (single-center retrospective case-control) | VTE prophylaxis (including warfarin, unfractionated heparin, low molecular weight heparin or intermittent pneumatic compression devices): in VTE cases: 59%; and non-VTE controls 47%. | Any symptomatic VTE (upper and lower extremity DVT and PE; n = 65) | VTE events identified using ICD-9 VTE discharge codes. | LOS case: 16 (10-28); controls: 6 (4-10) | Trauma last 3 mo; active cancer past year; admission fever; leg edema on admission; immobility >72 h; bacterial infection (cellulitis, pneumonia, sepsis, other); platelet count >350 × 109/L; use of VTE prophylaxis |
| The model was adjusted for prophylaxis. | |||||||||
| Zhou et al29 (China) | Medically ill patients (N = 1804: VTE cases: 902 and non-VTE controls: 902) | 2013-2016 | Cases: 6017; controls: 5717 | External validation (single-center retrospective Case-control) | Any prophylaxis: in VTE Cases: 4.1% and in non-VTE controls: 6.1% | Any symptomatic VTE (defined as DVT or PE; n = 902) | DVT verified by positive compression ultrasonography and/or contrast venography. | 6 mo after discharge | Caprini RAM factors* |
| VTE prophylaxis included any mechanical use (intermittent pneumatic compression devices or sole vein pump) or pharmacological use (unfractionated heparin, LMWH, warfarin, fondaparinux sodium, etc.) | PE verified by positive pulmonary angiogram, spiral computed tomography, or high probability ventilation/perfusion scanning or at autopsy. | Padua RAM factors† | |||||||
| The model was adjusted for VTE prophylaxis | |||||||||
| Barclay et al30 (US) | Chronic Liver disease (N = 1581) | 2008-2011 | 51(11) | Prognostic factor (single-center retrospective cohort) | Pharmacological VTE prophylaxis: 24.8% (Unfractionated heparin: 9.7%; low molecular weight heparin: 88.0%; or both: 2.3%) | Any symptomatic VTE (including DVT, PE or portal vein thrombosis-PVT) (n = 23) | VTE event identified in the medical record. | 4-7 d | VTE prophylaxis |
| The model was adjusted for pharmacological prophylaxis | VTE confirmed with radiologic testing. | Active malignancy | |||||||
| Trauma or surgery during hospitalization | |||||||||
| History of VTE | |||||||||
| Grant et al31 (US) | Hospitalized medical patients (N = 63 548) | 2011-2014 | 66 | External validation (multicenter retrospective cohort) | Pharmacological VTE prophylaxis: 60.9% | Any symptomatic VTE (defined as proximal upper- or proximal lower-extremity DVT and PE) (n = 670) | VTE was clinically suspected. | 90 d | Caprini RAM factors* |
| The model was adjusted for the pharmacological prophylaxis | VTE was Image confirmed. | ||||||||
| Majority of events were identified by medical record review, 44 (6.6%) of the events were confirmed via telephone follow-up. | |||||||||
| Rothberg et al34 (US) | Medical Patients (N = 46 503) | 2004-2005 | NR | Prognostic model development (multicenter retrospective cohort) | Pharmacological VTE prophylaxis: 30% | Any symptomatic VTE (n = 1 052) | VTE verified by lower extremity ultrasonography, venography, CT angiogram, ventilation/perfusion scan, or pulmonary angiogram) on hospital day 3 or later | 30 30 d | Any prophylaxis; female; length of stay ≥6 d; age (18-49; 50-64, and >65 y); primary diagnosis (pneumonia, COPD, stroke, congestive heart failure, urinary tract infection, respiratory failure, septicemia) |
| There was no difference in the model estimates for the factors when the model was adjusted for prophylaxis. | Secondary diagnosis of VTE provided using ICD-9 diagnoses. | Comorbidities (inflammatory bowel disease, obesity, inherited thrombophilia); cancer (18-49, 50-64, and >65 y). | |||||||
| Treatments (CVC, mechanical ventilation, urinary catheter, chemotherapy, steroids) | |||||||||
| Bembenek et al35 (Poland) | Early stroke patients (N = 299) | 2007-2009 | 75 (64-82)* | Prognostic factor (single-center, prospective cohort) | Oral anticoagulation: 7.1% | Any symptomatic or asymptomatic DVT (n = 9; 7 of which were distal) | The first ultrasonography was performed within the first 7 d and then 8-10 d after stroke onset by a trained physician blinded to patients’ baseline health status, to identify patients in whom DVT occurred early in the course of stroke. | Days 3 and 9 after stroke | Age for each additional 10 y; female; hypertension; congestive heart failure; atrial fibrillation; diabetes; smoking status (current and previous); prestroke disability (mRS 0-1 patient and mRS 0-2 patients); stroke severity (each additional 4 patients. NIHSS, NIHSS >7 patients, NIHSS >14 patients); decreased consciousness (≥1 patient in NIHSS, ≥2 patients in NIHSS); inflammatory markers (CRP >10 mg/L, fibrinogen >4 mg/dL) |
| The model was not adjusted for oral anticoagulation, but less than 10% of the included patients received prophylaxis. | |||||||||
| Fan et al36 (China) | Acutely ill medical patients (N = 458) | 2006-2007 | 77 (7) | Prognostic factor (multicenter prospective cohort) | Pharmacological VTE prophylaxis: 0% | Any symptomatic or asymptomatic VTE (DVT or PE; n = 45: 30 symptomatic and 15 asymptomatic) | VTE verified by compression ultrasonography at enrollment and 3-wk follow-up. | 90 d follow-up for symptomatic and 3 wk for asymptomatic | Univariate model with results provided; a multivariate analysis was conducted but results of each factor were not reported. |
| Mechanical VTE prophylaxis (graduated compression stockings): 0.4% | Symptomatic cases were all screened by lower limb color duplex ultrasonography. | ||||||||
| The model was not adjusted for the mechanical thromboprophylaxis. | |||||||||
| Kelly et al38 (United Kingdom) | Acute ischemic stroke (N = 102) | Not reported | 70 (12) | Prognostic factor (single-center prospective cohort) | VTE prophylaxis: 0% | Any symptomatic or asymptomatic VTE (defined as proximal DVT or PE; n = 41). | Patients were assessed weekly for clinical evidence of VTE. New increases in calf circumference from initial assessment of ≥3 cm (based on the Wells scoring system), local pain or tenderness for DVT, and oxygen saturations ≤92% and/or respiratory rate >20/min in a patient otherwise asymptomatic for PE. | 21 d | Age >70; BI ≤9, total anterior circulation infarcts, malignancy, and atrial fibrillation |
| VTE was classified as “unrecognized clinical” if associated with the aforementioned signs or symptoms that went unrecognized by the attending team. | |||||||||
| Magnetic resonance direct thrombus imaging was performed. If DVT was identified, thoracic imaging was performed to detect PE. All scans were reviewed independently by 2 reviewers who reached a consensus. | |||||||||
| Clinical events diagnosed conventionally and data from postmortem examinations were included. | |||||||||
| Ota et al38 (Japan) | Congestive heart failure (N = 161) | 2003-2008 | 69.3 (10.8) | Prognostic factor (single-center, prospective cohort). | Anticoagulant therapy in DVT cases: 38.9%; non-DVT cases: 44.1% | Any symptomatic or asymptomatic DVT (no PE was detected; n = 18). | DVT verified by standardized sonography criterion of venous noncompressibility. | 11.8 ± 11.5 d | NYHA functional class; poor IVC collapsibility; no anticoagulation therapy |
| Antiplatelet therapy in DVT cases: 66.7% and in the non-DVT: 62.9% | PE verified by pulmonary angiography. | ||||||||
| The model was adjusted for anticoagulant therapy. | |||||||||
| Yi et al39 (China) | Acute stroke patients (N = 1380) | 2009-2010 | 69.8 (11.6) | Prognostic factor (multicenter, prospective cohort) | Pharmacological VTE prophylaxis with warfarin or LMWH: 15% | Any symptomatic or asymptomatic VTE (any PE and any DVT) (n = 62; 32 symptomatic DVT and 30 asymptomatic DVT) | DVT verified by VDU, venous angiography or venous CTA examination. | 12 mo | For PE as an outcome: |
| The model was adjusted for prophylaxis. | PE verified by chest CTA or pulmonary angiography. | Age ≥70 y, bedridden, incidence of DVT | |||||||
| For DVT as an outcome: | |||||||||
| Age ≥70 y; bedridden, Wells score ≥2, NIHSS score of lower limbs ≥3, BI score, rehabilitative therapy, anticoagulant therapy, concentration of D-dimer evaluated at admission. | |||||||||
| Bleeding | |||||||||
| Decousus et al8 (Canada) | Acutely ill medical patients (N = 15 156) | 2002-2006 | 68.2 (51.8-78.9)* | Prognostic model development (multicenter, prospective cohort) | Pharmacological VTE prophylaxis: 48% (LMWH, 38.4%; unfractionated heparin, 11.1%, aspirin, 0.7%) | Major or clinically relevant bleeding (n = 230; 83 major and 147 nonmajor, but clinically relevant, bleeding) | Major bleeding was defined as a bleeding event contributing to death, clinically overt bleeding associated with a decrease in hemoglobin level of ≥2 g/dL or leading to transfusion of at least 2 units of packed RBCs, or bleeding within a critical organ (including intracranial, retroperitoneal, intraocular, adrenal gland, spinal, or pericardial bleeding). | 14 d | Active gastroduodenal ulcer; bleeding in 3 mo before admission; platelet count <50 × 109/L; age ≥85 vs <40 y; hepatic failure; severe renal failure GFR <30 vs ≥60 mL/min/m2; ICU/CCU; CVC; rheumatic disease; current cancer; age 40-84 vs <40 y; male sex; moderate renal failure (GFR 30-59 vs ≥60 mL/min/m2). |
| Mechanical VTE prophylaxis: 9% (elastic stockings, 5.4%; intermittent pneumatic compression, 3.8%) | Nonmajor but clinically relevant bleeding was defined as overt gastrointestinal bleeding (except for insignificant hemorrhoidal bleeding), gross hematuria (macroscopic and lasting longer than 24 h), substantial epistaxis that required intervention and was recurrent and/or lasted at least 5 min, extensive hematoma or bruising (>5 cm in diameter), intra-articular bleeding (documented by aspiration), menorrhagia or metrorrhagia (increased quantity or duration), or other bleeding important enough to be recorded on the hospital chart. | ||||||||
| There were no differences in the estimates of associations when the model was adjusted for pharmacologic prophylaxis use. | |||||||||
| Mahan et al32 (US) | Medical patients (N = 327 578) | 2005-2009 | 69 | Prognostic factor (multicenter retrospective cohort) | All antithrombotic agent use: 9.4% | Major or clinically relevant bleeding (n = 29 264; 5 951 major and 23 313 minor bleeding) | Bleeding events were identified through the ICD-9-CM diagnosis codes. | Within 30 d after hospitalization | Age (40-54, 55-64, 65-74, ≥75); male; preindex risk factors (insufficient renal function, cancer, rheumatoid arthritis, gastroduodenal ulcer, blood dyscrasias, thrombocytopenia, liver disease, CVC, thromboembolic stroke, estrogen use); postindex risk factors (postdischarge antithrombotic meds use, rehospitalization, length of stay (2, 3-5, and ≥6 d) |
| Anticoagulants: 3.9% (warfarin, 3.6%; enoxaparin, 0.4%; heparin, 0.1% and other, <0.0%) | |||||||||
| Antiplatelets: 5.7% (clopidogrel: 4.6%; aspirin-dipyridamole, 0.9%; other, 0.3%) | |||||||||
| Anticoagulants and antiplatelets: 0.2% | |||||||||
| The model was adjusted for antithrombotic use. | |||||||||
| Patell et al33 (US) | Cancer patients (N = 3358) | 2012-2014 | 62(19-98)* | Prognostic factor (single- center retrospective cohort) | Antiplatelets: 14% | Major or clinically relevant bleeding (n = 69; 51 major and 18 nonmajor but clinically relevant bleeding) | Bleeding was assessed using the International Society on Thrombosis Hemostasis definitions of major bleeding and clinically relevant nonmajor bleeding | Median length of stay was 5 d (range, 0-152) days. | Reason for admission (anemia); BMI ≥40; cancer site: GI; low hemoglobin (<13 g/dL for men and <11.5 g/dL for women); low platelets (<150 000/μL) |
| Anticoagulants: 67% | Bleeding events were identified from discharge summaries of admissions being studied. To obtain details of event, documentation including diagnostic tests (imaging and procedures) and clinical notes was used. All bleeding events were confirmed manually by 2 investigators (third-year internal medicine residents at the time of the study). When unclear, individual cases were cross-reviewed, discussed and included if both agreed. No separate training was performed, and no coding was used to extract bleeding information. | ||||||||
| Antiplatelet agents on day of admission were not found to be statistically significant in univariate analysis, so were not added to the multivariate regression analysis mode. | |||||||||
| Anticoagulation exposure on admission was noted to be associated with a decreased risk of bleeding (OR, 0.5; 95% CI, 0.3-0.8; P = 0.004) although this was not significant in multivariable analysis (but the model adjusted for it). | |||||||||
BI, Barthel Index; BMI, body mass index; COPD, chronic obstructive pulmonary disease; CTA, computed tomographic angiography; GI, gastrointestinal; ICD-9-CM, International Classification of Diseases, Ninth Revision, Clinical Modification; IVC, inferior vena cava; LWMH, low-molecular-weight heparin; mRS, modified Rankin Score; NYHA, New York Heart Association; RBC, red blood cell.
*
Caprini factors: stroke; acute spinal cord injury or paralysis (<1 mo); hip, pelvis, or leg fracture (<1 mo); multiple trauma (<1 mo); age ≥75); history of VTE; family history of VTE; history of thrombophilia; heparin-induced thrombocytopenia; age (41-60); age (61-74); positive history of cancer; immobilizing plaster cast; congestive heart failure; COPD or pulmonary function; inflammatory bowel disease; severe long disease (including pneumonia); acute myocardial infarction; sepsis (<1 mo); surgery (<1 mo); postpartum (<1 mo); history of unexpected stillborn infant, recurrent spontaneous abortion (≥ 3) or premature birth; varicose veins; BMI >25 kg/m2; swollen legs (current); CVC present on admission; immobile or not ambulating; and hormone replacement therapy or oral contraceptives.
†
Padua factors: active cancer; previous VTE; reduced mobility; known thrombophilia; recent trauma and /or surgery; elderly age; heart and/or respiratory failure; acute myocardial infarction or ischemic stroke; acute infection and/or rheumatologic disorder; obesity; ongoing hormone treatment, and VTE prophylaxis.