Article, Traumatology

Incidence and risk factors for venous thromboembolism in patients with nonsurgical isolated lower limb injuries

Original Contribution

Incidence and risk factors for venous thromboembolism in patients with nonsurgical isolated lower limb injuriesB,BB

Bruno Riou MD, PhDa,*, Christophe Rothmann MDb, Nathalie Lecoules MDc, Eric Bouvat MDd, Jean-Luc Bosson MDe, Philipe Ravaud MD, PhDf,

Charles Marc Samama MD, PhDg, Moussa Hamadouche MDh

aService d’Accueil des Urgences, Centre Hospitalo-Universitaire (CHU) Pitie’-Salpe^trie`re, Universite’ Pierre et Marie Curie (Paris 6), Assistance-Publique Ho^pitaux de Paris (AP-HP), Paris, France

bService d’Accueil des Urgences, Centre hospitalier re’gional (CHR) Notre Dame de Bon Secours, Metz, France

cService des Urgences, CHU Purpan, Toulouse, France

dUnite’ Fonctionnelle de Traumatologie du Sport, Ho^pital Sud, CHU de Grenoble, France

eService d’Informatique Me’dicale, CHU Grenoble, Unite’ mixte de recherche CNRS 5525,

Universite’ Joseph Fourier, Grenoble, France

fDe’partement d’e’pide’miologie, biostatistiques et recherche clinique, INSERM E03 57, CHU Bichat-Claude Bernard, AP-HP,

Universite’ Denis Diderot (Paris 7), Paris, France

gDe’partement d’anesthe’sie-Re’animation, CHU Avicenne, AP-HP, Universite’ Le’onard de Vinci (Paris 13), Bobigny, France hService de Chirurgie Orthope’dique et Traumatologique, CHU Cochin-Port Royal, AP-HP, Universite’ Rene’ Descartes (Paris 5), Paris, France

Received 9 August 2006; accepted 25 September 2006


Purpose: The aim of our study was to determine the incidence of venous thromboembolism in patients with nonsurgical isolated lower limb injury and to determine the risk factors associated with the development of the condition.

Methods: This observational study was conducted in French hospital emergency departments (EDs). Patients older than 18 years presenting with nonsurgical isolated lower limb injury below the knee in the ED were included. Deep VTE was diagnosed with Compression ultrasound. The final diagnosis of VTE was confirmed by an expert panel.

Results: Three thousand six hundred ninety-eight patients were included, and compression ultrasound examination was obtained in 2761 (75%) of them who were retained in the analysis. Deep venous thrombosis occurred in 177 patients and nonfatal pulmonary embolism in 1 patient. The incidence of VTE, mainly distal and asymptomatic, was 6.4% (95% confidence interval, 5.5%-7.4%). In a

B The trial was funded by Sanofi-Aventis. Sanofi-Aventis participated in the study design, collection of the data, and statistical analysis, with support from an independent statistical company (Sylia-Stat, Bourg la Reine, France). The statistical analysis plan, writing of the report, and decision to submit the paper were independent of Sanofi-Aventis, Paris, France.

BB All authors received consultancy fees, and B Riou and M Hamadouche received lecture sponsorships from Sanofi-Aventis, Paris, France.

* Corresponding author. Tel.: +33 1 42 17 72 49; fax: +33 1 42 17 72 64.

E-mail address: [email protected] (B. Riou).

0735-6757/$ – see front matter D 2007 doi:10.1016/j.ajem.2006.09.012

multivariate analysis, predictive variables of VTE were age of at least 50 years (odds ratio, 3.14; P b .0001), rigid immobilization (odds ratio, 2.70; P b .0001), no weight bearing (odds ratio, 4.11; P = .0015), and severe injury (odds ratio, 1.88; P = .0002). The discriminant analysis showed that age was the only variable independent of an antithrombotic prophylaxis associated with VTE. Conclusion: The incidence of VTE was 6.4% in patients with nonsurgical lower limb injury. Rigid immobilization, recommendation not to bear weight, severe injury, and age of at least 50 years should be considered as risk factors for VTE. Emergency physicians should also take age into account when prescribing antithrombotic prophylaxis.

D 2007


venous thromboembolism is a frequent condition with a significant morbidity and mortality [1]. Thrombo- phylaxis has been recognized as highly effective at preventing Deep venous thrombosis , pulmonary embolism (PE), and fatal PE, and its cost-effectiveness has repeatedly been demonstrated [2]. The identification of risk factors is the cornerstone of an appropriate prevention of VTE. A considerable amount of literature has focused on the risk factors for surgical patients and hospitalized Medical patients. In contrast, very little data have been obtained from nonsurgical and ambulatory patients.

Lower limb injury is very common in persons of all ages and is associated with a considerable morbidity and cost [3,4]. Its incidence in young patients has increased because of the popularity of Recreational sports [5]. Nevertheless, the epidemiology and prevention of VTE in patients with nonsurgical lower limb injury has been poorly investigated. Most previously published studies have included patients requiring surgery and/or having knee injuries, which are obviously importantly associated with risk of VTE [6-11]. In addition, the population included in randomized studies [6-9] may differ from that observed in real life. The 7th American College of Chest Physicians Conference on Antithrombotic and Thrombolytic Therapy recently con- cluded that antithrombotic prophylaxis should not be recommended for patients with isolated lower limb injury. Although antithrombotic prophylaxis is considered as a standard of care in several European countries, risk factors have not been appropriately identified in this population [2]. The objective of this Epidemiological study was to determine the incidence of VTE in nonsurgical patients with isolated lower limb injury and to determine the risk

factors associated with the development of VTE.


This epidemiological study was conducted on a multi- center prospective basis and was performed from April to October 2004 at the emergency departments (EDs) of teaching (n = 37) and nonteaching hospitals (n = 413). Approval from the institutional review board was obtained, and all patients provided informed written consent.


The criteria for inclusion in the study were age of at least 18 years and isolated unilateral injury of the lower limb below the knee not requiring any surgery, with an expected immobilization ranging from 7 to 90 days. Only hospital- ization for less than 24 hours in the short-duration-stay ward of the ED was authorized. The exclusion criteria were any anticoagulant treatment before trauma, surgery during the 4 weeks before trauma, pathologic fracture, life-threatening or invalidating associated disease, and trauma occurring more than 48 hours before inclusion.

Investigators were randomly chosen from a list of 750 emergency physicians in 450 centers. They were asked to include 6 consecutive patients, with an expected ratio of fracture of 1 of 3. However, during the last 2 months of the study period, investigators were authorized to include more than 6 patients.

Study design

Patients were included in the study on their First visit to the ED. The emergency physician freely decided the treatment of trauma and the antithrombotic prophylaxis. The following data were recorded: age, sex, body mass index, personal and family history of VTE, ongoing pregnancy, cancer, cardiac (New York Heart Association [NYHA] class III or IV) or Respiratory insufficiency, diabetes, hormonal treatment, Tobacco use (N10 per day), venous insufficiency (presence of varices and/or cutaneous complication), and obesity (body mass index N30 kg d m–2). All patients underwent radiography to precisely assess the type of injury and their associated risk for VTE, and thus, the Ottawa Ankle Rules were not followed during the study [12]. The type of immobiliza- tion prescribed was classified as follows: (1) rigid (resin or plaster cast) or (2) nonrigid (all others, including orthesis, taping, and bandages). The recommendation provided for walking was classified as follows: (1) no weight bearing, (2) partial weight bearing, or (3) complete weight bearing. The diagnosis of the injury was recorded and its severity classified as follows: (1) severe (any injury with fracture or dislocation, complete tendon rupture, Achilles tendon rupture), (2) moderate (incomplete tendon ruptures), or (3) minor (sprain, contusions). The prescribed antithrombotic prophylaxis was classified as previously described [2]: (1) high-risk dose,

(2) moderate-risk dose, or (3) no prophylaxis.

Fig. 1 Study flowchart.

The second consultation was defined as the date of immobilization removal. As soon as possible after this consultation, the patient underwent compression ultrasound examination.

In patients with VTE, a follow-up was performed at

90 days by telephone call. Complications related to Antithrombotic treatment were recorded. A severe hemor- rhagic complication was defined as being associated with death, hemorrhagic shock, blood transfusion, or requiring surgery for hemostasis or drainage.

Diagnosis of VTE

The final diagnosis was determined by an independent expert panel (see acknowledgment) that included a pulmo- nary physician, a cardiologist, and a CUS specialist, regarding the complete medical record including the CUS procedure. The final diagnoses recorded were distal DVT, proximal DVT, and PE. In cases of disagreement, a consensus was reached by the 3 experts.

Statistical analysis

Assuming an incidence of VTE between 1% and 14%, and a 12.5% dropout rate, we calculated that 4500 patients should be included to obtain a precision ranging between 0.3% and 1.1%.

Data are expressed as mean F SD, median, and 95% confidence interval for nongaussian variables (Kolmogorov- Smirnov test), or percentage and 95% confidence interval. Comparison between 2 groups was performed using the Student t test, the Mann-Whitney test, and Fisher exact method.

We performed a multiple forward logistic regression to assess variables associated with VTE. We included all significant preoperative variables in the univariate analysis ( P value of entry, .10) and variables that were thought to be prognostic or have been demonstrated to be prognostic in previous studies. The final decision concerning the list of

these variables was taken before the logistic regression was performed. The receiver operating characteristic (ROC) curve was used to determine the best threshold (minimizing the distance to the ideal point, with sensitivity = specificity = 1) for continuous variables to predict death. The Spearman coefficient matrix correlation was used to identify significant colinearity (q N 0.70) between variables. The odds ratio and its 95% confidence interval were calculated. The discrimination of the model was assessed using the area under the ROC curve [13], and the percentage of patients correctly classified. The calibration of the model was assessed using the Hosmer-Lemeshow statistics [14].

Because antithrombotic prophylaxis would be expected for patients with established risk factors, we expected that antithrombotic prophylaxis might not have been a signifi- cant variable in the logistic regression model. To further assess the possible role of antithrombotic prophylaxis, a propensity-score analysis was performed with regard to the use of anticoagulant [15]. End points were compared in each subgroup based on their propensity score using the Mantel- Haenszel test.

Lastly, we performed a discriminant analysis to explore variables explaining the separation of our population into 4 groups according to the presence or absence of VTE and antithrombotic prophylaxis.

Statistical tests were 2-sided, and P b .05 was considered significant. We used SAS software (SAS Institute Inc, Chicago, Ill).


A total of 3698 patients were studied, and CUS examination was obtained in 2761 (75%) patients who were retained for analysis (Fig. 1). There were no significant differences between patients lost during follow-up and those retained for analysis (data not shown).

There were 1391 (51%) men and 1358 (49%) women, with a mean age of 40 F 15 years (range, 18-89 years) and a mean body mass index of 25.1 F 4.4 kg d m–2. There were 222 (8%) obese patients; 136 (5%) patients had a personal history of VTE, and 464 (17%) reported a family history of VTE. There were only 2 patients with cancer and 1 with respiratory insufficiency. Injury was severe in 605 (22%) patients, moderate in 1215 (44%) patients, and minor in 941 (34%) patients. Recommendations for walking were no weight bearing in 1212 (44%) patients, partial weight bearing in 855 (31%), and total weight bearing in 690 (25%) patients. Antithrombotic prophylaxis was administered in 1680 (61%) patients, with a moderate-risk dose in 427 (25%) and a high- risk dose in 1253 (75%) patients. The remaining (39%) patients did not receive any prophylaxis. Antithrombotic prophylaxis was significantly more often prescribed in patients with a fracture (91% vs 41%, P b .001) and in patients with severe (95%) or moderate (75%) injuries compared with minor injuries (22%, P b .001).

Table 1 Comparison of patients with and without VTE


Without VTE (n = 2582)

With VTE (n = 177)



39 F 15

51 F 15


Age z50 y

636/2582 (25)

101/177 (57)



1323/2572 (51)

68/177 (38)



1249/2572 (49)

109/177 (62)

Body mass index (kgd m–2)

25.0 F 4.4

26.1 F 4.5


Body mass index N30 kgd m–2

314/2569 (12)

28/177 (16)


Personal history of VTE

119/2581 (5)

17/177 (10)


Family history of VTE

434/2581 (17)

30/177 (17)


Disease increasing risk for VTEa

386/2584 (15)

31/177 (17)



54/2584 (2)

5/177 (3)


Hormonal treatment

386/2584 (15)

31/177 (17)


Venous insufficiency

377/2584 (15)

47/177 (27)


Tobacco use

438/2583 (17)

31/177 (18)


Data are expressed as mean F SD or number (percentage).

a Includes cancer and respiratory insufficiency.

Deep venous thrombosis was diagnosed by CUS in 177 patients (6.4%; 95% confidence interval, 5.5%-7.4%). There were 172 distal DVTs (6.2%; 95% confidence interval, 5.4%-7.2%) and 5 proximal DVTs (0.2%; 95% confidence

interval, 0.1%-0.4%), 27 being symptomatic (1.0%; 95%

confidence interval, 0.7%-1.4%). Only 1 (0.04%; 95% confidence interval, 0%-0.02%) PE was diagnosed in a patient who experienced proximal DVT. The comparison between patients with VTE and those without is shown in Table 1. The incidence of DVT was 14.5% (95% confidence interval, 12.0%-17.6%) in patients with severe injury; 13.7% (95% confidence interval, 11.4%-16.4%) in patients older than 50 years; 10.3% (95% confidence interval, 8.8%- 11.9%) in patients with rigid immobilization; and 9.9% (95% confidence interval, 8.3%-11.7%) in patients with a no weight-bearing recommendation.

Table 2 Comparison of injury and treatment in patients with and without VTE

Data are expressed as mean F SD or number (percentage).

The following variables were considered in the logistic model: age, sex, body mass index, personal and family history of VTE, diseases increasing risk for VTE (including cancer), hormonal treatment, venous insufficiency, use of tobacco, injury severity, recommendation for walking, type of immobilization, duration of immobilization, and anti- thrombotic prophylaxis. No significant colinearity was noted between these variables, except for antithrombotic prophylaxis and type of immobilization (q = 0.74, P b .05). Missing values were observed for age (n = 2), body mass index (n = 15), personal (n = 3) and family history of VTE (n = 3), tobacco use (n = 2), type of immobilization (n = 4), and recommendation for walking (n = 4). Thus, the final logistic model was used for 2755 (99.8%) patients (Table 2). In the multivariate analysis, 4 variables were significantly associated with VTE: age, severity of the injury, type of


Without VTE (n = 2582)

With VTE (n = 177)


Radiography performed

2521/2584 (94)

173/177 (98)


Severity of injury


517/2584 (20)

88/177 (50)


1142/2584 (44)

73/177 (41)



925/2584 (36)

16/177 (9)

Type of immobilization


1336/2580 (52)

153/177 (86)



Duration of immobilization (d)

1244/2580 (48)

22 F 11

24/177 (14)

29 F 10


Recommendation for walking

No weight bearing

1092/2580 (42)

120/177 (68)

Partial weight bearing

804/2580 (31)

51/177 (29)


Total weight bearing

684/2580 (27)

6/177 (3)

Antithrombotic prophylaxis


1061/2581 (41)

16/177 (9)

Moderate-risk dose

390/2581 (15)

37/177 (21)


High-risk dose

1130/2581 (44)

124/177 (70)

(95% confidence interval)

No weight bearing

4.11 (1.72-9.86)


Age N50 y

3.14 (2.27-4.33)


Rigid immobilization

2.70 (1.66-4.38)


Severe injury

1.88 (1.34-2.62)


See text for definitions.

immobilization, and recommendation for walking. The Hosmer-Lemeshow statistic was 2.74 ( P = .84), indicating appropriate calibration. The area under the ROC curve was 0.778. The accuracy of the logistic model was 72%.

Table 3 Multivariate analysis of risk factors for VTE (n = 2755 patients)

Variables Odds ratio P

Four variables were significantly associated with the administration of antithrombotic prophylaxis: rigid immo- bilization, no weight-bearing recommendation, severe inju- ry, and moderate injury. The analysis of VTE occurrence according to the propensity score is shown in Table 3. However, as expected, the power of this analysis was very low because few untreated patients had a high propensity score (n = 40) and few treated patients had a low propensity score (n = 45) (Table 4).

The discriminant analysis was performed for 2758 patients, 38 patients being excluded because of missing values. Two canonical-structure axes were significant, the first one separating patients treated and those not treated with antithrombotic agents, and the second one separating patients with and without VTE. Three variables (severe injury, no weight bearing, rigid immobilization) significantly deter- mined axis 1, and only one variable (age) significantly determined axis 2 (Fig. 2).


In patients experiencing nonsurgical isolated lower limb injury, we observed that (1) incidence of VTE was 6.4%, and that of symptomatic DVT was 1.0%; and (2) main risk factors for VTE were severe injury, rigid immobilization, a no weight-bearing recommendation, and age of at least 50 years.

Table 4 Propensity-score analysis in patients with or without antithrombotic prophylaxis

In a randomized study of low-molecular-weight heparin, which included patients with a plaster-cast immobilization of at least 3 weeks, Jorgensen et al [9] reported an incidence of 10% and 17% of VTE in the treated and placebo groups, respectively. In contrast, Kock et al [10] observed an incidence of 0% and 4.3%, and Kujath et al [11] observed an incidence of 4.8% and 16.5% of VTE, in the treated and placebo groups, respectively. Lassen et al [16] observed an incidence of 9% and 19% in the treated and placebo groups, respectively, but they included patients with leg fracture who underwent surgery. Some studies report a higher incidence of VTE, up to 28%, but these have included injuries of the femur or the knee, which are associated with a higher risk of VTE [7].

Although the incidence of VTE was 6.4%, the incidence of symptomatic DVT was 1% and that of proximal DVT (which carries the higher risk for PE [17]) only 0.2%. The proportion of proximal thrombosis among all DVT in our study (3%) was lower than that reported by Lassen et al [16] (25%) in a population with a higher risk of DVT. Therefore, our results suggest that the prevention of VTE was generally appropriate in the population studied. Nevertheless, this was not the case in every subgroup because the incidence of VTE reached 13.7% in elderly patients.

The risk factors for VTE in patients with lower limb injury remain debatable. Indeed, although the type and duration of immobilization, the presence of fractures, and the recommendation concerning walking have been widely recognized for some time as risk factors, few previous studies have attempted to determine them precisely and accurately. Importantly, most of the studies included small numbers of patients and only performed univariate analyses [10,11,16]. Our study suggests that the emergency physi- cians appropriately stratified these risk factors when considering the prescription of antithrombotic prophylaxis and the low incidence of VTE and symptomatic DVT. Our study provides strong evidence that a severe injury (mainly an associated fracture), a rigid immobilization, and lack of weight bearing represent independent risk factors. However, age of at least 50 years should be also considered as a risk factor, and this was not considered by the emergency physicians when prescribing antithrombotic therapy. Depending on the studied clinical conditions, age has

Propensity-score groups





No. of patients


723 (67%)

313 (29%)

40 (4%)


45 (3%)

98 (6%)

1535 (91%)

Propensity score Nontreated

0.07 F 0.00

0.21 F 0.05

0.85 F 0.12



0.07 F 0.00

0.24 F 0.10

0.98 F 0.05



8 (1.1%)

3 (1.0%)

5 (12.5%)



0 (0%)

5 (5.1%)

156 (10.2%)

Fig. 2 Schematic result of the discriminant analysis. Two canonical axes were significant, the first one separating patients treated (T) and not treated (NT) with antithrombotic agents, the second one separating patients with (VTE) and without VTE (NVTE). Only 3 variables (2, severe injury; 3, no weight bearing; and 4, rigid immobilization) significantly determined axis 1, and 1 variable (1, age) significantly determined axis 2.

sometimes been recognized as being an independent risk factor for VTE [2]. Indeed, it is likely that associatED diseases greatly influence the statistical weighting because they are highly linked to age [18], explaining why age per se may not always be a significant risk factor. In contrast, in our healthy population, it is not surprising that age was significant. It is likely that age of at least 50 years is an additional limiting factor for early healing, weight-bearing capacity, and return to complete autonomy.

In this study, antithrombotic prophylaxis was not a sig- nificant protective factor for VTE. This result was expected and is not surprising. Emergency physicians understandably prescribed antithrombotic prophylaxis in patients with a high risk and did not prescribe it in patients with a low risk, explaining why this variable was not significant in the logistic model (Table 3). We performed a propensity-score analysis with regard to the use of anticoagulant, but again, the power was very low because few untreated patients had a high propensity score and few treated patients had a low propensity score. This result should not be misinterpreted, and we can only say that the protective role of antithrom- botic prophylaxis cannot be assessed in our study. The only result that can be interpreted in favor of a protective role of antithrombotic prophylaxis is the relatively low incidence of VTE and the very low incidence of proximal DVT and PE. Previously, the efficacy of antithrombotic therapy was demonstrated by randomized studies [9,10,16].

The incidence of complications related to antithrombotic prophylaxis was very low and usually not severe, suggesting a good balance between risk and benefit in this population as it was treated. Extending the indication of prophylaxis to patients aged at least 50 years might slightly modify this

risk; thus, further studies are needed to confirm this finding. This is important because an increased incidence in ankle injuries has been reported in elderly patients [19].

Because of a very low incidence of certain variables (cancer, severe diseases, hormonal treatment), the power of our study was not sufficient to identify their roles as potential risk factors, although they have been found to be independent risk factors for VTE in other medical con- ditions [20-22]. Because the incidence of obesity was not high in our population, our results may not apply to morbidly obese patients [23]. Lastly, diagnosis of PE did not rely on a systematic search based on computed tomographic scan or scintigraphy, and its incidence might have been underestimated. However, these methods were considered to be inappropriate considering the Very low risk of PE in this population.

In conclusion, in a large sample of patients with nonsurgical lower limb injury, the incidence of VTE was 6.4% and the incidence of symptomatic DVT was 1%. Rigid immobilization, recommendation not to bear weight, severe injury, and age of at least 50 years should be considered as risk factors for VTE. Emergency physicians appropriately took all of these factors into consideration, with the exception of age, in prescribing antithrombotic prophylaxis.


The following individuals also contributed to the study: Statistical analysis: Sylvie Brun (Sanofi-Aventis, Paris, France), Isabelle Dubroca (Sanofi-Aventis) and Sylvie

Lancrenon (Sylia-Stat, Bourg la Reine, France).

Independent expert committee: Philippe Girard (Insti- tut Montsouris, Paris, France), Marie The’re`se Barellier (Laboratorie d’exploration Fonctionelle, CHU Cote de Nacre, Caen, France), Antoine Elias (Me’decine cardiovas- culaire, CHU de Rangueil, Toulouse, France).

The authors thank Dr D.J. Baker, DM, FRCA (Dept. of Anesthesia and Critical Care, CHU Necker-Enfants Malades, Paris, France) for reviewing the manuscript.


  1. Geerts WH, Heit JA, Clagett GP, et al. Prevention of venous thromboembolism. Chest 2001;119(Suppl):S132-75.
  2. Geerts WH, Pineo GF, Heit JA, et al. Prevention of venous thromboembolism. The Seventh ACCP Conference on Antithrom- botic and Thrombolytic Therapy. Chest 2004;126(Suppl):S338-S400.
  3. Bertini N, Bleichner G, Cannamela A. L’entorse de cheville au service d’accueil des urgences. Re’an Urg 1995;4:491 – 501.
  4. Bridgman SA, Clement D, Downing A, Walley G, Phair I, Maffuli

N. Population based epidemiology of Ankle sprains attending accident and emergency units in the West Widlands of England, and a survey of UK practice for severe ankle sprains. Emerg Med J 2003;20:508 – 10.

  1. Leppilahti J, Orava S. Total Achilles tendon rupture: a review. Sports Med 1998;25:79 – 100.
  2. Hjelmstedt A, Bergvall U. Incidence of thrombosis in patients with tibial fractures: a phlebographic study. Acta Chir Scand 1968;134: 209 – 18.
  3. Abelseth G, Buckley RE, Pineo GE, et al. Incidence of deep-vein thrombosis in patients with fractures of the lower extremity distal to the hip. J Orthop Trauma 1996;10:230 – 5.
  4. Lassen MR, Borris LC, Nakov RL. Use of the low-molecular-weight heparin rexiparin to prevent deep-vein thrombosis after leg injury requiring immobilization. N Engl J Med 2002;347:726 – 30.
  5. Jorgensen PS, Warming T, Hansen K, et al. Low Molecular Weight Heparin (Innohep) as thrombophylaxis in outpatients with a plater cast: a venographic controlled study. Thromb Res 2002;105:477 – 80.
  6. Kock HJ, Schmit-Neuerburg KP, Hanke J, Rudofsky G, Hirche

H. Thrombophylaxis with low-molecular-weight heparin in out- patients with plaster-cast immobilization of the leg. Lancet 1995; 346:459 – 61.

  1. Kujath P, Spannagel U, Habscheid W. Incidence and prophylaxis of deep venous thrombosis in outpatients with injury of the lower limb. Hemostasis 1993;23(Suppl 1):20 – 6.
  2. Stiell IG, Greenberg GH, McKnight RD, et al. Decision rules for use of radiography in acute ankle injuries. Refinement and prospective validation. JAMA 1993;269:1127 – 32.
  3. Hanley JA, McNeil BJ. The meaning and use of the area under a receiving operating characteristic (ROC) curve. Radiology 1982; 143:29 – 36.
  4. Lemeshow S, Hosmer DW. A review of goodness of fit statistics for use in the development of logistic regression models. Am J Epidemiol 1982;115:92 – 106.
  5. Joffe MM, Rosenbaum PR. Invited commentary: propensity scores. Am J Epidemiol 1999;150:327 – 33.
  6. Lassen MR, Borris LC, Nakov RL. Use of low-molecular weight heparin reviparin to prevent deep-vein thrombosis after leg injury requiring immobilization. N Engl J Med 2002;347: 726 – 30.
  7. Seinturier C, Bosson JL, Colonna M, Imbert B, Carpentier H. Site and clinical outcome of deep vein thrombosis of the lower limbs: an epidemiological survey. J Thromb Haemost 2004;3:1362 – 7.
  8. Samama MM, Cohen AT, Darmon JY, et al. A comparison of enoxaparin with placebo for the prevention of venous thromboem- bolism in acutely ill medical patients. N Engl J Med 1999;341: 793 – 800.
  9. Kannus P, Palvanen M, Niemi S, Parkkari J, Jarvinen M. Increasing number and incidence of low-trauma ankle fractures in Elderly people: Finnish statistics during 1970-2000 and projections for the future. Bone 2002;31:430 – 3.
  10. Lee AY, Levine MN, Baker RI, et al. Randomized comparison of low- molecular weight heparin versus Oral anticoagulant therapy for the prevention of recurrent venous thromboembolism in patients with cancer. N Engl J Med 2003;349:146 – 53.
  11. Burkman RT. Venous thromboembolism and oral contraceptives: current status and clinical implications. Treat Endocrinol 2002;1: 143 – 7.
  12. Walker ID. Venous and arterial thrombosis during pregnancy: epidemiology. Semin Vasc Med 2003;3:25 – 32.
  13. Stein PD, Beemath A, Olson RE. Obesity as a risk factor in venous thromboembolism. Am J Med 2005;118:978 – 80.

Leave a Reply

Your email address will not be published. Required fields are marked *