Incidence and predictors of venous thromboembolism in post-acute care patients

Transcription

1 734 Schattauer 2010 Blood Coagulation, Fibrinolysis and Cellular Haemostasis Incidence and predictors of venous thromboembolism in post-acute care patients A prospective cohort study Gianluigi Scannapieco 1 ; Walter Ageno 2 ; Andrea Airoldi 3 ; Erminio Bonizzoni 4 ; Mauro Campanini 3 ; Gualberto Gussoni 5 ; Mauro Silingardi 6 ; Antonella Valerio 5 ; Chiara Zilli 1 ; Ido Iori 6 ; for the TERSICORE Study Group* 1 Internal Medicine, Cà Foncello Hospital, Treviso, Italy; 2 Department of Clinical Medicine, Insubria University, Varese, Italy; 3 Internal Medicine, Maggiore Hospital, Novara, Italy; 4 Institute of Medical Statistics and Biometry, University of Milan, Milan, Italy; 5 FADOI Foundation Research Centre, Milan, Italy; 6 Internal Medicine I, Arcispedale S. Maria Nuova, Reggio Emilia, Italy Summary Few studies have addressed the topic of venous thromboembolism (VTE) in patients hospitalised in rehabilitation facilities. This patient population is rapidly growing, and data aimed to better define VTE risk in this setting are needed. Primary aim of this prospective observational study was to evaluate the frequency of symptomatic, objectively confirmed VTE in a cohort of unselected consecutive patients admitted to rehabilitation facilities, after medical diseases or surgery. Further objectives were to assess overall mortality, to identify risk factors for VTE and mortality, and to assess the attitude of physicians towards thromboprophylaxis. A total of 3,039 patients were included in the study, and the median duration of hospitalisation was 26 days. Seventy-two patients (2.4%) had symptomatic VTE. The median time to VTE from admission to the long-term care unit was 13 days. According to multivariable analysis, previous VTE (hazard ratio 5.67, 95% confidence interval ) and cancer (hazard ratio 2.26, 95% confidence interval ) were significantly associated to the occurrence of VTE. Overall in-hospital mortality was 15.1%. Age over 75 years, male gender, disability, cancer, and the absence of thromboprophylaxis were significantly associated to an increased risk of death (multivariable analysis). In-hospital antithrombotic prophylaxis was administered to 75.1% of patients, and low-molecular-weight heparin was the most widely used agent. According to our study, patients admitted to rehabilitation facilities remain at substantially increased risk for VTE. Because this applies to the majority of these patients, there is a great need for clinical trials assessing optimal prophylactic strategies. Keywords Post-acute care facilities, thromboprophylaxis, venous thromboembolism Correspondence to Gianluigi Scannapieco, MD Cà Foncello Hospital Piazza Ospedale, Treviso, Italy Tel.: , Fax: gscannapieco@ulss.tv.it * A list of the participants in the TERSICORE Study Group is given at the end of the article. Financial support: The study was partially supported by an unrestricted grant from AstraZeneca Italy, without involvement in study design, management, analysis and reporting. Received: March 10, 2010 Accepted after major revision: April 23, 2010 Prepublished online: July 20, 2010 doi: /th Thromb Haemost 2010; 104: Introduction Venous thromboembolism (VTE) is a common clinical problem among hospitalised patients, and is associated with significant morbidity and mortality. Many studies have investigated risk factors for VTE and contributed to identifying high-risk groups among surgical and acutely ill medical patients who can benefit from routine antithrombotic prophylaxis (1). In contrast, few studies have focused on patients hospitalised in post-acute care facilities, who represent a rapidly growing population. These patients often have multiple risk factors for VTE in addition to age, such as prolonged immobilisation, recent surgery, and concomitant medical diseases (2, 3). Available data on the epidemiology of deep-vein thrombosis (DVT) and/or pulmonary embolism (PE) in these patients are limited and controversial. The reported incidence of symptomatic VTE in post-acute care patients varied from 0% to 5.2% (3 6). On the other hand, the incidence of DVT documented by routine ultrasonography was reported from 12.8% to 17.4% (7, 8), and the prevalence of ultrasonographically detected DVT ranged from 14% to 16% (2, 9). However, some of the studies were not specifically focused on VTE (3, 4), were retrospective (5), were conducted on limited samples of patients (6) or on selected patients population (4, 5, 8). Available data suggest that the implementation of thrombophylactic strategies can reduce the incidence of VTE in post-acute care patients (7). However, no randomised clinical trials have addressed the efficacy and safety of thromboprophylaxis in this setting and, to our knowledge, there are no published international guidelines on this topic. Thrombosis and Haemostasis 104.4/2010

2 Scannapieco et al. VTE in post-acute care patients 735 Further, very few data are available on the rate of anticoagulant prophylaxis used in this clinical setting. In the study by Bosson et al. the percentage of patients receiving prophylaxis ranged from 20% to 87% across participating centres (2), so reflecting a high degree of uncertainty in the decision to use or not prophylaxis. In order to address the relative paucity of data on this growing patient population, we carried out a large multicentre prospective study to estimate the incidence of clinical VTE in an unselected population of patients admitted to rehabilitation facilities, to identify specific risk factors for VTE, and to assess the attitude of physicians towards antithrombotic prophylaxis. Materials and methods TERSICORE (ThromboEmbolic Risk Stratification In ChrOnically bedridden patients) is a prospective observational study in a cohort of unselected consecutive patients admitted to 24 rehabilitation facilities in Italy over a 12-month period. Data were collected under the scientific coordination of the Italian Federation of Internal Medicine (FADOI) Research Centre. The primary aim of the study was to evaluate the frequency of symptomatic VTE, during hospitalisation in the rehabilitation facility. VTE events had to be objectively documented by instrumental diagnostic work-up based on centre-specific procedures. Standard methods for diagnosis of DVT and PE were leg vein ultrasonography and CT scan, respectively. Secondary objectives were to assess the rate of overall mortality during hospitalisation, to identify specific risk factors for VTE and mortality, and to assess the attitude of physicians towards antithrombotic prophylaxis. At the time of hospital admission, information on the medical history (with particular attention to the presence of previous VTE or cancer), on the reason for hospitalisation (i.e. stroke, other neurologic diseases, heart failure, respiratory insufficiency, other medical diseases; orthopaedic or general surgery, neurosurgery, other types of surgery), and on the level of mobility based on the Rankin score were collected. The modified Rankin scale is a commonly used six-grade scale (0 = no symptoms at all / 6 = dead) for measuring the degree of disability or dependence in the daily activities, and it is the most widely used clinical outcome measure for stroke clinical trials (10, 11). During hospitalisation data were recorded concerning the occurrence of symptomatic VTE and of major bleeding events, the use of antithrombotic prophylaxis, and on patient outcome. No screening ultrasound was performed during the study. A bleeding event was defined as major if it was fatal or clinically overt and associated with a decrease in haemoglobin level of at least 2.0 g/dl over a 48-hour period or with transfusion of two or more units of whole blood or red cells, or occurred in a critical organ (brain, spine, pericardium, retroperitoneum, eye), or required an invasive intervention (12). Methods of prophylaxis were classified as unfractionated heparin (UFH, 5,000 IU bid or tid, or 12,500 IU bid), low-molecular-weight heparin (LMWH, 3,400 antixa IU/daily or >3,400 antixa IU/daily, or full anticoagulant dose), vitamin K antagonists, physical methods (elastic stockings, intermittent pneumatic compression), and combined prophylaxis (pharmacological plus physical methods). Aspirin or other antiplatelet agents were not considered as thromboprophylactic strategies for VTE. The study was approved by the Ethic Committees of all participating centres. Informed consent was obtained from all patients. Statistical analysis A sample size of 3,000 patients was considered appropriate to document a reliable estimate of the incidence of symptomatic VTE events (primary endpoint), by hypothesising an event rate of 2.00% with 95% confidence intervals (CI) of the estimate equal to the event rate ± 1.00%. This hypothesis was based on the results of a preliminary, single-centre survey (13). Data were summarised using proportions if categorical, and mean ± standard deviation (SD) or median if continuous, and they Table 1: Characteristics of patients at baseline. Characteristics Gender Age Male / Female (%) 41.1 / 58.9 Mean ± standard deviation (SD) 40 years (%) years (%) years (%) years (%) 54.3 > 90 years (%) 8.9 Rankin score < 4 (%) (%) 85.0 Main reason for hospitalisation (%) Ischaemic stroke 18.0 Haemorrhagic stroke 3.1 Other neurologic disease 7.4 Heart failure 16.5 Respiratory insufficiency 9.5 Other medical diseases 28.7 Orthopaedic surgery 24.0 General surgery 4.0 Neurosurgery reasons 14.9 Other characteristics (%) Previous venous thromboembolism 5.4 Cancer ± Schattauer 2010 Thrombosis and Haemostasis 104.4/2010

3 736 Scannapieco et al. VTE in post-acute care patients were compared by means of the chi-square test or the Student s t-test as appropriate. The association between the occurrence of VTE or death and the presence of prognostic factors, was evaluated by means of two separate multivariable Cox regression analyses. Covariates for these analyses included age ( 75 vs. 75 years), gender, previous VTE, Rankin score ( 4 or <4), medical disease or recent surgery inducing hospitalisation, cancer, time free from antithrombotic prophylaxis (time-dependent covariate). Hazard ratios (HR) and 95%CIs were reported with two-tailed probability values. A p-value 0.05 was considered statistically significant. Statistical analysis was carried-out using SAS software, version Results Study population A total of 3,039 patients were included in the study, and details on their characteristics are specified in Table 1. Patients aged 75 Table 2: Venous thromboembolism, major bleeding and death in the study population. *According to evaluation by the investigator, no autopsy was performed. VTE, venous thromboembolism; DVT, deep-vein thrombosis; Hb, haemoglobin. Major events N. patients (%) VTE total 72 (2.4) Isolated DVT 47 (1.5) Pulmonary embolism 14 (0.5) DVT + pulmonary embolism 11 (0.4) Major bleeding 25 (0,8) Bleeding in critical organ 5 (0.2) Decrease of Hb >2 g/dl or need for transfusion 20 (0.6) Death 460 (15.1) Fatal VTE * 12 (0.4) Fatal bleeding 2 (<0.1) years or more were 63.2% of the overall study population, and the great majority of patients (85%) had moderately severe / severe disability (Rankin score 4 5). The median duration of hospitalisation in the rehabilitation facilities was 26 days (range 1 255). VTE and other major clinical events Seventy-two patients (2.4%) had symptomatic, objectively confirmed VTE ( Table 2). All DVT occurred at the lower limbs. The median time to VTE from admission to the long-term care unit was 13 days. At the time of VTE 49 of 72 patients (68.1%) were receiving antithrombotic prophylaxis. According to multivariable analysis ( Table 3), a previous episode of VTE (HR 5.67, 95% CI , p <0.001) and cancer (HR 2.26, 95% CI , p <0.01) were significantly and independently associated to the occurrence of VTE during hospitalisation. An episode of major bleeding was recorded in 25 patients (0.8%), and bleeding was fatal in two cases. Pharmacological prophylaxis was ongoing at the time of bleeding in 18 patients (72.0%). Overall in-hospital mortality was 15.1%. The median time from admission to the long-term unit to death was 18 days. In 12 of 460 deaths, the attending physicians attributed the event to PE. In all these patients, death was anteceded by recent symptomatic objectively confirmed PE. Age over 75 years, male gender, moderately severe / severe disability, cancer, and the absence of antithrombotic prophylaxis were significantly and independently associated to an increased risk of death, according to the results of multivariable analysis ( Table 4). Antithrombotic prophylaxis During hospitalisation antithrombotic prophylaxis was administered to 75.1% of patients. The median duration of prophylaxis was 17 days, and the median length of stay in the rehabilitation unit for patients receiving prophylaxis was 29 days. Details on the type of prophylaxis are described in Figure 1. LMWH was by far the most widely used agent (79.9% of patients receiving prophylaxis, Variable Effect HR 95% CI P value # Age > 75 vs Gender Male vs Female Previous VTE Yes vs No < Rankin score 4 vs < Medical disease Yes vs No Recent surgery Yes vs No Cancer Yes vs No < 0.01 Time free from prophylaxis* Yes vs No Table 3: Multivariable Cox regression analysis to correlate prognostic factors and venous thromboembolism (VTE). # Wald Chi-Square test; *time-dependent covariate. HR, hazard ratio; CI, confidence interval. Thrombosis and Haemostasis 104.4/2010 Schattauer 2010

4 Scannapieco et al. VTE in post-acute care patients 737 Table 4: Multivariable Cox regression analysis to correlate prognostic factors and death. # Wald Chi-sSquare test; *time-dependent covariate. VTE, venous thromboembolism; HR, hazard ratio; CI, confidence interval. Variable Effect HR 95% CI P value # Age > 75 vs < Gender Male vs Female < 0.01 Previous VTE Yes vs No Rankin score 4 vs < < Medical disease Yes vs No Recent surgery Yes vs No < Cancer Yes vs No < Time free from prophylaxis* Yes vs No < mostly administered at high antithrombotic doses). Vitamin K antagonists were used in 10.4% of patients receiving antithrombotic treatment, in all cases with indications other than prophylaxis of VTE. Figure 2 describes the rates of antithrombotic prophylaxis in specific subgroups of patients. The use of prophylaxis was not related to age (76.8%, 73.1%, and 75.6% in patients with age less than 60 years, age within 61 75, and more than 75 years, respectively), but to the severity of disability at hospital admission (78.7% for Rankin score 4 5 vs. 60.3% in patients with Rankin score < 4, respectively, p < ). Discussion To our knowledge, this is the largest prospective cohort study aimed to assess the incidence of symptomatic VTE in patients admitted to rehabilitation facilities following surgical procedures or acute medical diseases. Despite a high use of antithrombotic prophylaxis, symptomatic objectively confirmed VTE was documented in 2.4% of patients, during a median observation period of 26 days. Independent predictors for VTE were the presence of cancer or a previous VTE, whereas neither severe disability nor the reason for hospital admission were independently associated with the risk of VTE. Fifteen per cent of patients died; a major bleeding event occurred in less than 1% of patients and only two of these events were fatal. The short-term risk of venous thromboembolic complications in hospitalised patients, both medical and surgical, is well defined and thromboprophylaxis is recommended for many high-risk patient categories. Recommended duration of prophylaxis ranges between a minimum of 14 days in medical setting, including stroke patients, and a maximum of 28 to 35 days in high-risk surgical patients, such as those undergoing cancer surgery or major orthopaedic procedures (1). Reports from clinical practice patterns suggest that, in most cases, the use of prophylaxis is limited to the time spent in the hospital, regardless of the duration of hospitalisation, and extended prophylaxis remains poorly prescribed (14 17). However, the risk of VTE may remain substantial in some patients, particularly in case of severe residual disability in association with persistent risk factors for VTE after discharge. Unfortunately, there is currently little available information on the true incidence of VTE in medical or surgical patients admitted to rehabilitation facilities, and the risk-to-benefit ratio of prolonged prophylaxis in this setting remains uncertain. Despite this gap of knowledge and the consequent absence of evidence-based recommendations, we observed that as many as 75% of patients admitted to Italian rehabilitation facilities were actually receiving some form of thromboprophylaxis, more commonly LMWH administered at high prophylactic doses. On the one hand this finding is surprising, given the lack of specific guidelines, and the results of a recent Italian survey suggesting a low propensity among physicians to prescribe extended prophylaxis following hospital discharge (15). On the other hand, this finding suggests that physicians working in rehabilitation facilities are aware of VTE as a major clinical problem in their patient population. A substantial proportion of patients enrolled in our observational study was indeed at increased risk for VTE because of age (two thirds of the patients were older than 75 years) and because of the severe residual disability that was documented in the large majority of them. A few studies have evaluated the burden of VTE in patients admitted to post-acute care facilities (2 9, 13), but our study is the first to prospectively provide adequate estimates of the incidence of symptomatic VTE in this setting. The observed 2.4% incidence of symptomatic VTE reflects a high-risk setting and compares with other rates reported from previous observational studies carried Figure 1: Type of antithrombotic prophylaxis in the study population. UFH, unfractionated heparin; LMWH LD, low-molecular-weight heparin low dose ( 3,400 antixa IU/daily); LMWH HD, low-molecular-weight heparin high dose (>3,400 antixa IU/daily); OA, oral anticoagulant. Schattauer 2010 Thrombosis and Haemostasis 104.4/2010

5 738 Scannapieco et al. VTE in post-acute care patients Figure 2: Percentages of use of prophylaxis in specific categories of patients. What is known about this topic? Venous thromboembolism (VTE) is a common clinical problem among hospitalised patients, and is associated with significant morbidity and mortality. Few studies have addressed the topic of VTE in patients hospitalised in rehabilitation facilities. What does this paper add? To our knowledge, this is the largest prospective cohort study aimed to evaluate the incidence of symptomatic VTE in patients admitted to rehabilitation facilities following surgical procedures or acute medical diseases. Further aims were to assess overall mortality during hospitalization, identify specific risk factors for VTE and mortality, and assess the attitude of physicians towards antithrombotic prophylaxis. Despite a substantially high use of antithrombotic prophylaxis strategies, symptomatic VTE was objectively documented in 2.4% of patients during a median of 26 days of hospitalisation. Independent predictors for VTE were the presence of cancer or a previous venous thromboembolic event. A major bleeding event occurred in less than 1% of patients. Death was considered related to pulmonary embolism in 12 cases, whereas fatal bleeding occurred in two patients. Patients admitted to rehabilitation facilities following hospitalisation for surgery or medical disease remain at substantially high risk for VTE. Since this risk applies to the majority of these patients, there is a great need for clinical trials assessing optimal prophylactic strategies in this setting. out in high-risk populations. In study, the incidence of symptomatic VTE occurring up to 30 ± 5 days after cancer surgery was 2.1% (14); in the FOTO study, the incidence of symptomatic VTE occurring at three months after major orthopaedic surgery was 1.8% (18). Of interest, in our study the reason for admission, either recent surgery or a recent medical disease, was not associated with the risk for VTE, as well as age or the degree of dependency. Likewise, the results of a recent French study have identified previous history of VTE and cancer as independent predictors of VTE (9). As opposed to our study, older age and the degree of dependency were also associated with VTE in the French study. Taken together, these data suggest that all patients admitted to a long-term care facility should be evaluated for their risk of VTE regardless of the reason for admission, but with particular attention to their mobility and to the concomitant presence of major risk factors for VTE. The results of our study further stress the need for clinical trials in this setting. On the one hand the level of risk for thromboembolic complications appears of clinical relevance, thus supporting the need for anticoagulant prophylaxis, on the other hand the very high mean age and the complexity of this population place these patients at a potentially increased risk for bleeding. Therefore, from a general point of view, patients hospitalised in rehabilitation facilities need a careful assessment of both the risk for VTE and for bleeding, in order to apply pharmacologic strategies for thromboprophylaxis to those with a low risk for bleeding and high risk for VTE. Of interest, in this study there was a trend toward a higher mortality rate for VTE than for bleeding (2.6% and 0.4% of cases of death, respectively), and the absence of thromboprophylaxis was independently associated with a Thrombosis and Haemostasis 104.4/2010 Schattauer 2010

6 Scannapieco et al. VTE in post-acute care patients 739 higher risk of mortality. Again, these observations need to be confirmed by properly designed clinical studies. One surprising finding of this study was the very high rate of fatal PE events. Indeed, as many as 12 of the 25 patients with an objective diagnosis of PE died. After careful revision of all cases, we could confirm that in all patients who died because of VTE according to clinical evaluation by the investigator, death was anteceded by symptomatic objectively documented PE. We believe this surprising finding is probably explained by the low threshold for diagnosing haemodynamically stable PE in this setting, thus suggesting that a number of less severe, symptomatic PE have been probably missed. On the other hand, there was a low rate of mortality attributed to bleeding. A possible explanation is that the majority of bleeding events recorded in this study were qualified as major because they were associated with a decrease in haemoglobin levels, but did not occur in critical organs. This study has a number of limitations. First, information on the incidence of VTE is only based on the onset of signs and symptoms and on the application of centre-specific protocols to obtain objective diagnosis of the disease. Since it is well known that most cases of DVT remain asymptomatic and are detectable only by means of routine screening, our approach certainly underestimates the true incidence of the disease, in particular in patients with recent surgery of the leg or neurological diseases. However, by only focusing on hard endpoints such as mortality, major bleeding and symptomatic VTE rather than on surrogate endpoints, our study truly reflects the frequency of events that are normally faced in daily clinical practice and that are of greatest relevance to clinicians. Second, we do not have information on the time elapsed between the surgical procedure or the onset of the acute medical disease and the admission to the long-term care facility. Furthermore, we do not have information on prophylactic strategies used before the admission to the long-term care facility. Third, we did not collect information on all known risk factors for VTE (i.e. obesity, hormone replacement therapy) but we decided to only gather data on those that have the strongest association. On the one hand this Appendix Members of the TERSICORE Study Group A. Bonanome, M.Barbujani (Adria RO); N.Suzzi (Roma); P. Pavan, M. Rossi (Treviso); C. Di Donato, M. Garuti (Vignola MO); S. Cucca, G. De Marchi (Mestre / Venezia); G. Chesi, E. Scalabrini (Scandiano RE); G. Rigoli, G. Armani (Negrar VR); M. Cafiero (Napoli); D. Panuccio, G. Saccoccio (Loiano BO); G. Patrassi, L. Scarano (Cittadella PD); A. Belelli, D. Di Viesti (Reggio Emilia); M. Belogi, S. Diotallevi (Cagli PU); Q. Messina, C. Zilli (Motta di Livenza TV); G. Scanelli (Ferrara); V. Manicardi, P. Magnani, A. Zollino (Montecchio RE); G. Civardi, E. Bassi (Fiorenzuola d Arda PC); S. Zamboni; S. Cuppini (Rovigo); D. Primon (Camposampiero PD); G. Allosia (Valenza AL); C. Cisari, A. Baricich (Novara); C. Pedace (Arezzo); V. Orlando (Somma Vesuviana NA); S. Cabodi (Birago di Vische TO); G. Carbognin, M.C. Pasini, A. Chiaramonte (Negrar VR); G. Landini (Firenze). Abbreviations Bid, bis in die (twice a day); CI, confidence interval; CT, computerised tomography; DVT, deep-vein thrombosis; g, grams; Hb, haemoglobin; HD, high dose; HR, hazard ratio; IU, International Unit; LD, low dose; LMWH, low-molecular-weight heparin; OA, oral anticoagulant; PE, pulmonary embolism; SD, standard deviation; tid, ter in die (three times a day); UFH, unfractionated heparin; VTE, venous thromboembolism. simplified approach may have missed some additional independent factors, on the other hand focusing on those major risk factors easily taken into account by the practicing clinician makes the results of our study and the message conveyed more generalisable. In conclusion, patients admitted to rehabilitation facilities following hospitalisation for surgical procedures or medical diseases are at substantially increased risk for VTE. Because this high risk applies to the majority of these patients, there is a great need for clinical trials assessing optimal prophylactic strategies. Acknowledgments The authors thank Davide Ghilardi, from FADOI Foundation Research Centre, for editorial assistance. References 1. Geerts WH, Bergqvist D, Pineo GF, et al. Prevention of venous thromboembolism: American College of Chest Physicians Evidence-Based Clinical Practice Guidelines (8 th Edition). Chest 2008; 133: 381S-453S. 2. Bosson JL, Labarere J, Sevestre MA, et al. Deep vein thrombosis in elderly patients hospitalized in subacute care facilities: a multicenter cross-sectional study of risk factors, prophylaxis and prevalence. Arch Intern Med 2003; 163: Schianchi T, Meschi T, Briganti A, et al. Post-acute long stay and extensive rehabilitation: study of the first year of work at a long stay university hospital unit. Ann Ital Med Int 2001; 16: Roth HJ, Lovell L, Harvey RL, et al. Incidence of and risk factors for medical complications during stroke rehabilitation. Stroke 2001; 32: Harvey RL, Lovell LL, Belanger N, et al. The effectiveness of anticoagulant and antiplatelet agents in preventing venous thromboembolism during stroke rehabilitation: a historical cohort study. Arch Phys Med Rehabil 2004; 85: Valderrama A, Del Castillo J, Diaz Granados CA, et al. Deep vein thrombosis in chronically bedridden elderly individuals. J Am Geriatr Soc 2006; 54: Sellier E, Labarere J, Bosson JL, et al. Effectiveness of a guideline for venous thromboembolism prophylaxis in elderly post-acute care patients: a multicenter study with systematic ultrasonographic examination. Arch Intern Med 2006; 166: Ambrosetti M, Salerno M, Zambelli M, et al. Deep vein thrombosis among patients entering cardiac rehabilitation after coronary artery bypass surgery. Chest 2004; 125: Sellier E, Labarere J, Sevestre MA, et al. Risk factors for deep vein thrombosis in older patients: a multicenter study with systematic compression ultrasonography in postacute care facilities in France. J Am Geriatr Soc 2008; 56: Rankin J. Cerebral vascular accidents in patients over the age of 60. II. Prognosis. Scott Med J 1957; 2: Farrell B, Godwin J, Richards S, et al. The United Kingdom transient ischaemic attack (UK-TIA) aspirin trial: final results. J Neurol Neurosurg Psychiatry 1991; 54: Schulman S, Kearon C. Definition of major bleeding in clinical investigations of antihemostatic medicinal products in non-surgical patients. J Thromb Haemost 2005; 3: Schattauer 2010 Thrombosis and Haemostasis 104.4/2010

7 740 Scannapieco et al. VTE in post-acute care patients 13. Scannapieco G, Rossi M, Baldessin F, et al. Incidenza e fattori di rischio per tromboembolia venosa nei pazienti ricoverati in lungodegenza. Giorn Ital Med Int 2003; 2 (Suppl 2): Agnelli G, Bolis G, Capussotti L, et al. A clinical outcome-based prospective study on venous thromboembolism after cancer surgery: project. Ann Surg 2006; 243: Gussoni G, Campanini M, Silingardi M, et al. In-hospital symptomatic venous thromboembolism and antithrombotic prophylaxis in Internal Medicine. Findings from a multicenter, prospective study. Thromb Haemost 2009; 101: Goldhaber SZ. Venous thromboembolism prophylaxis: quality, location (hospital vs home), and duration. Thromb Haemost 2009; 102: Kalka C, Spirk D, Siebenrock K-A, et al. Lack of extended venous thromboembolism prophylaxis in high-risk patients undergoing major orthopaedic or major cancer surgery. Electronic assessment of VTE prophylaxis in high-risk surgical patients at discharge from Swiss hospitals (ESSENTIAL). Thromb Haemost 2009; 102: Samama CM, Ravaud P, Parent F, et al. Epidemiology of venous thromboembolism after lower limb arthroplasty: the FOTO study. J Thromb Haemost 2007; 5: Thrombosis and Haemostasis 104.4/2010 Schattauer 2010