1. Clinical and Applied
Thrombosis/Hemostasis
Volume 16 Number 4
The ‘‘Critical Thrombosis Period’’ in July/August 2010 394-405
# 2010 The Author(s)
10.1177/1076029609355151
Major Orthopedic Surgery: When to http://cath.sagepub.com
Start and When to Stop Prophylaxis
David Warwick, MD, FRCS, FRCS (Orth), and
Nadia Rosencher, MD
Patients undergoing major orthopedic surgery are at literature reflects the need to balance the improved
high venous thromboembolism (VTE) risk, with morbid efficacy of initiating prophylaxis close to the surgery
and potentially fatal consequences. Anticoagulant VTE with increased risk of perioperative bleeding. Evidence
prophylaxis reduces rates of postoperative deep vein from pathology, epidemiology, and clinical studies sug-
thrombosis by up to 60% to 70% in these patients. gests the risk period for VTE begins at surgery and
Therefore, pharmacological prophylaxis with low- extends well beyond hospitalization—a crucial issue
molecular-weight heparins (LMWHs), vitamin K when considering how long to give prophylaxis—and,
antagonists, or fondaparinux is recommended by in the case of total hip arthroplasty, for at least 3
current guidelines. However, there remains an ongoing months after surgery. Literature supports the greater
debate regarding when to initiate and the optimal dura- use of ‘‘just-in-time’’ thromboprophylaxis initiation and
tion for prophylaxis. Here, we discuss the mechanisms after-discharge continuation of optimal prophylaxis in
underlying thrombus formation in patients undergoing orthopedic surgery patients. Providing optimal throm-
major orthopedic surgery, and we review the current boprophylaxis throughout the critical thrombosis
literature on the benefit-to-risk ratio associated with period where a patient is at VTE risk will ensure the
preoperative and postoperative initiation of thrombo- best reductions in VTE-related morbidity and
prophylaxis and also the benefit-to-risk ratio in cases mortality.
of neuraxial anesthesia. We also discuss the duration
of postoperative VTE risk following major orthopedic Keywords: deep vein thrombosis; hip arthroplasty;
surgery and assess the ‘‘critical thrombosis period’’ knee arthroplasty; prophylaxis; pulmonary embolism;
when prophylaxis should be provided. Current venous thromboembolism
Introduction (HFS), between 40% and 60% of patients are at risk
of subclinical deep vein thrombosis (DVT) and 4% to
Without prophylaxis, the frequency of venous throm- 10% develop pulmonary embolism (PE), of which
boembolism (VTE) after major orthopedic surgery is 5% might be fatal.1 High-risk orthopedic surgery can
high.1 Following total hip arthroplasty (THA), total also predispose an individual to the development of
knee arthroplasty (TKA), or hip fracture surgery chronic venous insufficiency, although the risk of
this sequela has not been quantified as comprehen-
sively as the risk of postoperative VTE.
From the Orthopaedic Surgery, University of Southampton,
Hampshire, United Kingdom (DW); and Department of Anaes-
Both international and US evidence–based
thesiology and Intensive Care, Groupe Hospitalier Cochin Port guidelines recommend the appropriate use of pro-
ˆ ´ ´
Royal, Assistance Publique – Hopitaux de Paris, Universite Rene phylaxis for the prevention of VTE following ortho-
Descartes, Paris, France (NR). pedic surgery.1,2 Specifically, current guidance
Address correspondence to: David Warwick, Orthopaedic from the American College of Chest Physicians
Surgery, University of Southampton, Southampton, Hampshire,
SO16 6UY, United Kingdom; e-mail: davidjwarwick@
(ACCP)1 and recently published International Con-
btinternet.com. sensus guidelines2 strongly recommend prophylaxis
394

2. The critical thrombosis period in major orthopedic surgery / Warwick and Rosencher 395
Table 1. Current Recommendations From the American College of Chest Physicians for Venous Thromboem-
bolism Prophylaxis in Orthopedic Surgery Patients
Thromboprophylaxis Regimen
Surgery Type Recommendation Type Dose/Initiation
Elective hip Grade 1A LMWH Usual high-risk dose started 12 hours before surgery or
replacement 12 to 24 hours after surgery or half the usual high-risk
dose started 4 to 6 hours after surgery, increasing to
the full usual high-risk dose the following day
Fondaparinux 2.5 mg started 6 to 24 hours after surgery
Adjusted-dose vitamin K INR target 2.5 (range 2.0 to 3.0) initiated preoperatively,
antagonists or the evening of the day of surgery
In patients with high Grade 1A Mechanical thromboprophylaxis When high bleeding risk decreases: pharmacologic
bleeding risk with VFP or IPC thromboprophylaxis be substituted for or added to
mechanical thromboprophylaxis (grade 1C)
Elective knee Grade 1A LMWH Usual high-risk dose
replacement
Fondaparinux
Adjusted-dose vitamin K INR target 2.5 (range 2.0-3.0)
antagonists
Grade 1B IPC (without anticoagulant)
In patients with high Grade 1A IPC When high bleeding risk decreases: pharmacologic
bleeding risk thromboprophylaxis be substituted for or added to
Grade 1B VFP mechanical thromboprophylaxis (grade 1C)
Hip fracture surgery Grade 1A Fondaparinux
Grade 1B LMWH
Grade 1B UFH
Grade 1B Adjusted-dose vitamin K INR target 2.5 (range 2.0-3.0)
antagonists
In patients with high Grade 1A Optimal mechanical prophylaxis When high bleeding risk decreases: pharmacologic
bleeding risk thromboprophylaxis be substituted for or added to
mechanical thromboprophylaxis (grade 1C)
When surgery likely Grade 1C LMWH
to be delayed
Grade 1C UFH
NOTES: LMWH ¼ low-molecular-weight heparins; INR ¼ international normalized ratio; VFP ¼ venous foot pump; IPC ¼ inter-
mittent pneumatic compression; UFH ¼ unfractionated heparin.
with low-molecular-weight heparins (LMWHs), ‘‘critical thrombosis period’’ when patients are at
vitamin K antagonists, or fondaparinux (Table 1). increased risk of developing VTE?
There is clear evidence to show that the use of these There has been some debate in the literature
effective prophylactic agents can reduce the rate of regarding the optimal timing for the initiation of
postoperative DVT by 60% to 70%.1-3 Mechanical prophylaxis to maximize efficacy while minimizing
methods of prophylaxis, including graduated bleeding risk.4 For example, delaying the initiation
compression stockings (GCS) and intermittent of thromboprophylaxis with LMWHs (including
pneumatic compression, offer an alternative for enoxaparin, dalteparin, nadroparin, and tinza-
patients in whom anticoagulant prophylaxis is con- parin) has been shown to result in suboptimal
traindicated, or they can be used as adjunct to antithrombotic effectiveness4,5; however, com-
anticoagulant-based prophylaxis.1 mencing prophylaxis too soon after surgery has
There are 2 important questions to consider been linked to an increased bleeding risk.6
when planning VTE prophylaxis in major orthopedic Whatever the anticoagulant agent, the balance
surgery patients. First, when should anticoagulant between the benefits (prevention of VTE, a major
prophylaxis be initiated to balance efficacy with cause of perioperative morbidity and mortality)
safety? And second, when should prophylaxis be and risks (major bleeding and epidural hematoma)
stopped to ensure adequate risk reduction during the is the central consideration when deciding

3. 396 Clinical and Applied Thrombosis/Hemostasis / Vol. 16, No. 4, July/August 2010
whether or not to provide anticoagulant prophy- of prophylaxis to cover the critical thrombosis
laxis. This balance depends on the pharmacology period.
of each agent and on the dosage, timing of admin-
istration, the type of surgery, and patient
characteristics. Pathological Drivers of VTE Associated
Guidance on what constitutes an ‘‘optimal dura- With Orthopedic Surgery
tion’’ of thromboprophylaxis after orthopedic surgery
is less clearly stated and continues to be a topic of Thrombosis formation begins during surgery in elec-
debate.1,5,7-9 It is known that the risk of VTE contin- tive orthopedic surgery and perhaps earlier in the
ues long after patient discharge. Studies of large- case of hip fracture. For over a century, scientists
scale patient databases suggest that approximately and clinicians have recognized Virchow’s triad of
50% to 75% of cases of surgery-related VTE manifest venous stasis, endothelial injury or damage, and
after a patient has been discharged from hospital.10-14 alterations to the constitution of the blood, which
For example, a US study of postoperative throm- together predispose an individual to thrombi forma-
boembolic events in over 19 000 primary THA tion.21 Such conditions prevail during major ortho-
patients and over 24 000 primary TKA patients pedic surgery. There is often disruption to blood
found that the diagnosis of VTE was made after hos- flow in the femoral vein and debris might be present
pital discharge in 76% of cases following THA and in a patient’s blood.22-24 Slow, turbulent blood flow
47% of post-TKA cases. This corresponds to a med- in valve cusps creates local areas of stasis and
ian time to VTE diagnosis of 17 and 7 days postpro- encourages the development of thrombi.23 During
cedure, respectively.10 Another European-based surgery, there is a significant reduction in both
prospective study looking at the annual incidence venous outflow and venous capacitance.24 After
of DVT after almost 5000 joint operations noted that orthopedic surgery, DVT events are mostly concen-
symptoms of DVT appeared on average 27 days after trated in the operated leg, emphasizing the role of
THA, 36 days after surgery for hip fracture, and 17 stasis.25,26 Local tissue damage associated with sur-
days after TKA.11 gery is also an important factor in proximal thrombo-
Although 1 study of THA and TKA patients sug- sis,25 and it is recognized that surgical technique can
gested that prophylaxis for 9 days with the LMWH affect both operative venous stasis and the risk of
enoxaparin was associated with a clinically accepta- thrombi formation.22,25,27 Femoral vein kinking, calf
ble incidence of symptomatic VTE and major bleed- vein distension, endothelial hypoxia, and vibration
ing,7 other studies have demonstrated that extending during surgery, all contribute to endothelial
LMWH prophylaxis for up to 35 days after THA damage.28
offers significant reductions in the incidence of At the site of surgery, cell microaggregates are
VTE.8,9,15,16 The potential for improving clinical formed which, together with bone marrow debris,
outcomes by extending the period of thrombopro- embolize centrally. Some of this debris and microag-
phylaxis after major orthopedic surgery has also been gregate formation can appear in the arterial side of
described in patients receiving warfarin or the circulation as well. Symptoms can occur during
fondaparinux.17,18 Additionally, prolonged prophy- surgery itself—a phenomenon well known to
laxis of over 30 days with the novel anticoagulants anesthetists as a fall in PaO2, cardiovascular instabil-
dabigatran and rivaroxaban has been shown to be ity, and lung shunts. Indeed, intraoperative death
well tolerated.19,20 can occur. These microaggregates can obliterate
Here, we evaluate the physiological and hema- small vessels within the brain leading to cognitive
tological drivers of VTE, which lead to a prolonged changes and, within the heart, leading to myocardial
hypercoagulable state in patients undergoing infarction.
major orthopedic surgery, and the impact of these There is also substantial local and systemic acti-
factors on the optimal timing of initiation and vation of coagulation during the operation. It has
duration of thromboprophylaxis. We review the been noted that after a fibrinolytic shutdown imme-
current clinical and epidemiological evidence diately after surgery, activation of coagulation
relating to the risks and benefits of preoperative increases and leads to a prolonged hypercoagulable
versus postoperative initiation of pharmacological state.28,29
prophylaxis and prolonged-duration prophylaxis, Thus, the present day understanding is that sys-
and evaluate what constitutes an optimal duration temic activation of coagulation, as well as

4. The critical thrombosis period in major orthopedic surgery / Warwick and Rosencher 397
peripherally derived DVT, should be considered a undergoing elective THA found that preoperative
target for prophylactic medication use. This is a initiation was associated with a reduced DVT inci-
potent argument toward the use of chemical prophy- dence compared with postoperative initiation
laxis, rather than mechanical methods alone, which (10.0% vs 15.3%, respectively; P ¼ .02), with major
would not influence the more systemic effects. bleeding also reduced in patients receiving preopera-
tive prophylaxis (0.9% vs 3.5%, respectively; P ¼
.01).5 Additional reviews support the improved effi-
When Should Prophylaxis Begin? cacy of initiating LMWH prophylaxis in orthopedic
patients close to the time of surgery.4,30 A greater
The ‘‘ideal’’ time for the initiation of thromboprophy- reduction in VTE risk was observed when half the
laxis should balance the optimal efficacy of the antith- usual dose of a LMWH was given to patients under-
rombotic agent with the associated risk of bleeding. going elective total hip replacement close to the time
The pathological drivers of VTE provide some founda- of surgery versus the full dose given either 12 hours
tion for debating when is the best time to initiate preoperatively or 12 to 18 hours postoperatively.4 A
prophylaxis. Anticoagulant prophylaxis initiated preo- systematic review of studies of LMWH prophylaxis
peratively and close to the time of surgery offers in patients undergoing THA, TKA, or HFS established
potential for the formation of fewer thrombi7 and can the risk of postoperative DVT, as confirmed by veno-
increase the risk of intraoperative bleeding.30 graphy, according to the timing of prophylaxis.30 This
Preoperative initiation of anticoagulant prophy- analysis found that postoperative DVT occurred in
laxis can also increase the risk of compressive neur- 19.2% of patients given LMWH 12 hours preopera-
axial hematoma if regional anesthesia is used.31 tively, in 12.4% of patients given perioperative pro-
Indeed, the general incidence of compressive spinal phylaxis, and in 14.4% of patients given
hematoma after neuraxial anesthesia is very low postoperative prophylaxis (started at least 12 hours
(1 in 150 000 patients), but the incidence of hema- after surgery).30
tomata increased to 1 in 3600 when analysis was Another theoretical benefit of preoperative
restricted to epidural anesthesia for TKA in women administration over postoperative administration is
aged older than 70 years.32 Both US and European protection against the systemic activation of coagula-
guidelines allow a clearance of 12 hours after tion, which is initiated during surgery. Clinical trials
LMWH injection for a neuraxial puncture, thus and meta-analyses comparing preoperative against
preoperative injection can be done 12 hours before postoperative administration have used venographic
surgery.33,34 Alternatively, prophylaxis may be surrogate endpoints considering only peripheral vein
started after surgery in case of neuraxial puncture.35 manifestations of thrombosis. At present, therefore,
Indeed, puncture can be difficult with many there are no data on this important aspect of the
attempts or may result in a bloody puncture. There- pathophysiology of thrombus formation and whether
fore, the risk of bleeding with a compressive epidural prophylaxis can influence it.
hematoma could outweigh the still-debated potential
benefits of preoperative initiation.
Postoperative initiation of anticoagulant prophy- Timing of Prophylaxis Initiation to
laxis avoids increasing the risk of bleeding during Achieve Peak Efficacy
procedures but might not protect against the forma-
tion of perioperative thrombi. Current clinical guide- The peak efficacy of both LMWHs and other antith-
lines suggest either preoperative or postoperative rombotic agents depends on the timing of the first
anticoagulant prophylaxis,1 with practices varying injection.4,5,30 A meta-analysis of studies directly
by region. For example, LMWH prophylaxis is gener- comparing VTE prophylaxis with fondaparinux or
ally initiated up to 12 hours preoperatively (40 mg LMWH (enoxaparin) reported that fondaparinux
once daily) in Europe in accordance with regimens started 6 hours postoperatively was more effective
approved by the regulatory authorities, whereas in in reducing VTE risk.36 However, in 2 of the 4 stud-
North America it is mostly initiated 12 to 24 hours ies in this meta-analysis, the first dose of enoxaparin
(30 mg twice daily) postoperatively.4 was not administered until 12 hours after surgery,
A meta-analysis of 6 trials of LMWHs comparing whereas in the other 2 studies enoxaparin was initi-
the effectiveness of preoperative versus postoperative ated 12 hours before surgery. Thus, the pooled data
initiation of LMWH prophylaxis in patients in this meta-analysis combine the results of

5. 398 Clinical and Applied Thrombosis/Hemostasis / Vol. 16, No. 4, July/August 2010
2 different approaches to the timing and use of States between January 1998 and December 2000
LMWH, and in our opinion do not allow a true reported an incidence of major bleeding of 2.6%, and
or consistent comparison of these agents as throm- data from randomized clinical trials have demon-
boprophylactic agents. Furthermore, recently strated little or no increase in the rates of clinically
reported clinical trials comparing new oral agents important bleeding with prophylactic doses of
and enoxaparin have also used different timing guideline-recommended thromboprophylactic
approaches.19,20 Ideally, future studies comparing agents. 1,39
However, the definition of perioperative
anticoagulants in this setting should be made with bleeding is vague and there is no consistent assess-
therapies initiated at the same time points or at opti- ment in different trials. Significant bleeding would
mal time points for both regimens, according to be regarded as that leading to a degraded surgical
the time to reach maximum concentration for outcome (wound dehiscence, infection, and delayed
each drug.37 hospital discharge), anesthetic risk (in particular,
It is likely that if a single dose of prophylactic intraspinal bleeding associated with neuraxial block-
agent is provided too far in advance of surgery, the ade), or general effect (retroperitoneal, cerebral, or
anticoagulant effect may have declined by the time intestinal bleeding). After neuraxial anesthesia, this
of the surgical insult. The ideal time to begin risk of bleeding is rare but can lead to paraplegia.
administering an LMWH for peak efficacy is Even if the complication is less catastrophic, after
between 2 hours before and 8 hours after surgery.4 lumbar plexus block, some compressive hematoma
‘‘Just in time’’ LMWH prophylaxis, given 4 to 6 can occur also but with only femoral nerve compres-
hours after surgery, is as effective as preoperative sion.40-42
LMWH given 2 hours before surgery,4 with no rela- Current ACCP guidelines acknowledge that the
tive increase in the risk of perioperative bleeding. evidence supports both the preoperative and the
However, the risk of bleeding and neuraxial com- postoperative initiation of LMWH prophylaxis in
pressive hematoma in elderly patients increases if patients undergoing elective hip replacement (grade
prophylaxis is initiated too close to the start of 1A recommendations; Table 1).1 In clinical practice,
surgery (ie, between À2 and þ4 hours of sur- we believe that mechanical methods of prophylaxis,
gery).4,30 It would seem prudent, therefore, to delay as recommended in the ACCP guidelines, could have
restarting anticoagulant medications until a stable a role in covering the perioperative period until it is
clot has been established; hemostasis of small decided safe for an individual patient to receive
vessels should be at least 8 hours.31,38 However, antithrombotic prophylaxis (grade 1C in patients at
this does not mean waiting 8 hours in each patient bleeding risk, grade 2A as adjunct to chemoprophy-
because the onset time (tmax) of most anticoagu- laxis).1 However, the lack of protection provided by
lants is substantial. Instead, it is recommended mechanical prophylactic methods against the sys-
waiting for 8 hours minus tmax. The result of this temic activation of thrombin must be remembered.
strategy is that a total of 8 hours will have elapsed The mechanical method is an addition to, rather
before peak anticoagulation is reestablished. In than substitute for, chemical methods. Sole use of
practice, the longer the tmax the shorter the delay mechanical prophylaxis methods should only be con-
before anticoagulant reinitiation. For example, if sidered for patients with contraindications to
the tmax is 4 hours (as it is for LMWH), the safety anticoagulants.1
delay would be 4 hours; but if the tmax is only 1 hour In summary, the timing of initiation of pharma-
(as it is for unfractionated heparin), the safety delay cological prophylaxis should, according to the
should be at least 7 hours.31 This safety delay time guidelines,1 be based on risk-to-benefit considera-
is confirmed by the decrease of major bleeding in tions for each patient. This should consider both
fondaparinux studies when initiation of the drug is an individual patient’s relative risk of bleeding com-
done 6 to 8 hours after the end of surgery.36 plications versus their risk factors for VTE, both of
which should be established prior to surgery. In
general, it should be assumed that the more effica-
The Impact of Bleeding in Major cious a compound is against thrombosis, the greater
Orthopedic Surgery the risk of bleeding; that risk can be mitigated by
giving the compound further away from the
A large-scale database including 23 518 patients who moment of surgery (or reducing dose) but at the
underwent major orthopedic surgery in the United cost of decreased efficacy.

6. The critical thrombosis period in major orthopedic surgery / Warwick and Rosencher 399
or is elective in nature. For example, in patients with
femoral fractures, it has been shown using whole
blood thromboelastography that a hypercoagulable
state persists for at least 6 weeks after surgical
repair.29 Additional studies comparing venous blood
flow in fractured versus nonfractured legs found that
disturbed venous outflow persisted in fractured legs
for as long as 42 days after surgery (Figure 1).24
Epidemiological Studies
Other evidence for a protracted period of VTE risk
after orthopedic surgery comes from studies show-
ing, as described above, that the discontinuation of
prophylaxis a week after THA is associated with
rebound hypercoagulability persisting for 5 weeks
of follow-up.28
Epidemiological studies have identified that VTE
risk persists for prolonged periods after major ortho-
Figure 1. Venous output pattern following surgery (mean + pedic surgery.48,49 This is evident in the late emer-
standard error of mean) comparing fractured versus gence of symptomatic VTE and deaths after normal
nonfractured legs.24 Values at day 0 are preopera- venography or cessation of prophylaxis. For example,
tive recordings. Only patients with preoperative the Norwegian Arthroplasty Register, providing a
measurements were included in the postoperative
10-year follow-up data for 39 543 THA patients,
days.
noted a higher mortality rate among THA patients
in the first 60 postoperative days, than in the general
What Is the Optimal Duration of population.50 The continuing burden of VTE after
Prophylaxis? surgery has also been highlighted by a study of DVT
and PE incidence among 19 586 THA and 24 059
Historically, the timing of prophylaxis and assess- TKA patients in a US discharge database.51 The cumu-
ment of VTE risk following major orthopedic surgery lative incidence of DVT or PE within 3 months of
have closely mirrored the usual period of hospitaliza- surgery was 2.8% and 2.1% following THA and TKA,
tion, which was 10 to 14 days.2,9,43-46 However, the respectively. The diagnosis was made after discharge
physiological and hematological disturbances associ- in 76% of THA and 47% of TKA cases and at a median
ated with major orthopedic surgery persist for longer time of 17 and 7 days after surgery, respectively.
than this period, depending on the type of surgery The majority of these patients received in-hospital
and any patient-related factors determining individ- VTE prophylaxis, with only 32% receiving outpatient
ual VTE risk.1 prophylaxis. This suggests that more intense, earlier
prophylaxis might be needed following TKA and
more prolonged prophylaxis might be required follow-
ing THA to reduce the risk of late VTE.51
Physiological Studies According to the UK National Total Hip
Femoral blood flow (capacitance and outflow) is sig- Replacement Outcome study—documenting out-
nificantly reduced for at least 6 weeks after THA47 comes after primary THA in 13 343 patients in a
compared with a period of 6 days of altered hemody- health service setting over 1 year—VTE is the most
namics after TKA.26 Furthermore, in THA patients common complication following surgery.52 The hos-
there are 2 timings for the onset of DVT, with peaks pital readmission rate within 12 months of THA was
in incidence about 4 days after surgery and 13 days 17%, with almost a third of readmissions resulting
after surgery.27 It has also been shown that venous from complications of THA. The most common
function and the subsequent development of DVT complication was VTE, with rates higher than those
are affected by whether surgery follows a fracture for dislocation, urinary tract infections, wound

7. 400 Clinical and Applied Thrombosis/Hemostasis / Vol. 16, No. 4, July/August 2010
Table 2. Median Time After Surgery to Presenta-
tion of Objectively Confirmed Venous Throm-
boembolism Following Major Orthopedic
Proceduresa
Median Time, Days
Pulmonary Deep Vein
Embolism Thrombosis
Surgery Diagnosis Diagnosis
Hip fracture 17 24
Total hip arthroplasty 34 21
Total knee 12 20
arthroplasty
a
Data from a study of 5607 patients all of whom received
low-molecular-weight heparin prophylaxis while in hospital3
complications, and perioperative fracture; fatal PE
occurred at a rate of 0.1%.
The risk of PE in THA and TKA patients is high
and sustained, despite the use of prophylaxis for
approximately 10 days.53 In a prospective registry of
4000 patients from a single center, the overall inci-
dence of confirmed PE was 1.3% after in-hospital
LMWH prophylaxis lasting 10 days following joint
surgery. After major hip surgery, the incidence of
PE was raised for at least 2 to 3 months.
Additional support for prolonged and differing
VTE risk periods following THA and TKA is pro-
vided by a recent observational study reporting
data collected over 13 years.3 In over 5000
Figure 2. The time course of prophylaxis use (any type) versus
patients, all of whom had received in-hospital the cumulative incidence of venous thromboembo-
LMWH prophylaxis, the cumulative incidence of lism (VTE) following (A) total hip arthroplasty (n ¼
symptomatic VTE was 2.7% within 6 months fol- 6639) or (B) total knee arthroplasty (n ¼ 8236).14
lowing surgery (1.1% PE, 1.5% DVT, and 0.6% *Compared with prophylaxis use on the day when
both). In 70% of cases, VTE was diagnosed after most patients received it.
discharge. After surgery for hip fracture, DVT and
PE occurred at a median of 24 and 17 days, was discontinued, with equivalent rates of VTE
respectively. Deep vein thrombosis and PE between both groups 3 months after hospital dis-
occurred at a median of 21 and 34 days, respec- charge.54 A prospective study11 of 4840 patients
tively, following THA, and 20 and 12 days, respec- given routine prophylaxis for 10 days noted that the
tively, following TKA. The cumulative risk of VTE annual incidence of DVT following major orthopedic
lasted for up to 3 months after hip surgery and for procedures was 2.1%. More recently, data from the
1 month after knee surgery (Table 2).3 GLobal Orthopaedic Registry (GLORY) concerning
A number of studies have identified that the ben- VTE incidence in 6639 THA and 8326 TKA patients
efits of prophylaxis could be lost when medication is showed that 75% of VTE after THA and 57% after
discontinued too soon after orthopedic surgery. In- TKA occurred following discharge.14 In this registry,
patient prophylaxis with enoxaparin for a mean of over a quarter of all patients (26% of THA and 27%
7.3 days afforded significantly greater protection of TKA patients) were not receiving prophylaxis 7
from VTE than adjusted-dose warfarin during hospi- days after surgery, highlighting the discrepancy
talization in a group of over 3000 THA patients (P ¼ between the risk of VTE and the administration of
.0083). However, the benefits were lost once therapy prophylaxis (Figure 2).14

8. The critical thrombosis period in major orthopedic surgery / Warwick and Rosencher 401
and almost 70% received LMWH for 4 weeks after
surgery. The rate of symptomatic VTE (confirmed
by ultrasound or venography) was 1.34% after 3
months. This is in the lower range reported in the lit-
erature for patients with hip fracture (1% to 9%).61-63
Data from randomized clinical trials show a
trend toward fewer symptomatic VTE events with
extended-duration prophylaxis in patients who
have undergone elective total knee replacement58;
although this difference did not reach statistical
significance. A meta-analysis of prophylaxis studies
also supports the hypothesis that extended-duration
Figure 3. Rates of symptomatic venous thromboembolism prophylaxis is of benefit compared with standard
(VTE) following total hip arthroplasty from studies duration prophylaxis in TKA patients, with VTE
comparing in-hospital prophylaxis with prolonged incidences of 0% versus 1.4%, respectively (odds
(5 week) prophylaxis.6,8,9,15,57-59 ratio, 0.74; 95% confidence interval, 0.26-2.15).64
In summary, there is evidence to suggest that
altered blood flow and hypercoagulability persist for
Radiological Studies
weeks after major orthopedic surgery and that the
In our opinion, additional radiological evidence for cumulative risk of VTE continues for about 3 months
the extended risk of VTE is convincing. Orthopedic after surgery. There is also a growing tendency to
patients are still at risk of late VTE, even when dis- continue prophylaxis beyond the period of hospitali-
charged due to negative ultrasound or venography zation. Indeed, French guidelines recommend 42
results.7,44,45 For example, an overview of published days of prophylaxis following hip replacement.35
studies, in which more than 4000 orthopedic This commentary has a limitation as some arti-
patients were discharged after pharmacological pro- cles only report incidences of symptomatic VTE at
phylaxis, showed that in 68.1% of patients who were long-term follow-up, but not asymptomatic VTE
venographically negative, there were 30 reported events that may also result in long-term morbidity
episodes of symptomatic VTE after hospital dis- or even a fatal PE. This can lead to an underestima-
charge.45 Imaging studies over recent decades have tion of postoperative VTE risk, and especially post-
also suggested that VTE risk extends beyond the discharge risk.
period during which orthopedic patients typically
receive prophylaxis.27,43,55,56 Indeed, most of these
studies suggest that short periods of prophylaxis, Conclusions
lasting 7 to 10 days, often serve simply to delay the
interval between surgical operation and onset of The ‘‘critical thrombosis period’’ associated with
DVT, rather than preventing VTE entirely. The liter- major orthopedic surgery begins at the time of sur-
ature suggests that 20% to 30% of THA patients gery and persists beyond the assumed limits of hospi-
develop venographically confirmed DVT 4 to 5 weeks tal stay and the standard assessments made to
after surgery despite having had a normal day-10 or determine a patient’s risk of VTE. Thrombosis for-
day-7 venogram at the end of their prophylactic mation begins during (or even before in hip fracture)
therapy and hospital stay.8,9,15,57 Data from several surgery due to prevailing conditions that include
studies show that prolonging prophylaxis for up to endothelial injury, disruption in blood flow, venous
5 weeks after hospital stay for THA can reduce the stasis, and alterations in coagulability. It follows that
incidence of postdischarge VTE, both symptomatic efforts to prevent the formation of thrombi should
and venographic (Figure 3).6,8,9,15,17,18,57-59 begin as early as possible. Current guidelines advise
Real-world data from the ESCORTE study of that the timing of initiation of pharmacological pro-
approximately 7000 prospectively enrolled hip frac- phylaxis is a clinical decision that should consider
ture patients are also supportive of a lower incidence the trade-off between the risk of VTE and the bleed-
of VTE associated with prolonged prophylaxis.60 ing risks associated with antithrombotic therapy.1
Approximately 98% of hip fracture patients in this For agents such as the LMWHs, the current clinical
observational study received perioperative LMWH guidelines note that prophylaxis can be initiated

9. 402 Clinical and Applied Thrombosis/Hemostasis / Vol. 16, No. 4, July/August 2010
either before or immediately after surgery; whereas evidence and case for VTE prevention and strongly
for newer agents the current advice is begin prophy- advocate use of thromboprophylaxis, yet do not
laxis 6 to 8 hours after surgery. In practice, many provide strong recommendations on when to initi-
surgeons seek to adopt a ‘‘just in time’’ approach ate prophylaxis or the optimal duration for effec-
to prophylaxis, where initiation of prophylaxis tive prophylaxis. Both pathophysiological and
occurs several hours after surgery at a time when clinical evidence suggests that the risks begin ear-
the patient’s bleeding risk is deemed to be suffi- lier and extend for longer than the period during
ciently low (the precise time when pharmacological which most patients are provided with VTE pro-
prophylaxis is administered will depend on the phylaxis. The issue for an individual patient is to
patient’s individual risk of bleeding), while ensuring ensure that prophylaxis is given and continues for
the earliest possible antithrombotic activity against the known duration of risk of venous thrombosis.
surgery-related thrombosis. The use of mechanical By adopting a ‘‘just-in-time’’ approach to the
methods of prophylaxis to cover the delayed initia- initiation of thromboprophylaxis with LMWH,
tion of anticoagulant prophylaxis may be that is, starting therapy at a safe time before or
considered. after surgery, and by continuing with effective
Investigations of the pathophysiology of thromboprophylaxis beyond hospital discharge for
surgery-related thrombosis and hematological up to 5 weeks after major orthopedic surgery, it
changes that occur during the postoperative period could be possible to further reduce the burden
highlight that VTE risk continues well beyond the of VTE disease associated with otherwise
first few days after surgical injury. Evidence from life-enhancing surgery.
epidemiological studies, analyses of patient data-
bases, radiological investigations, and clinical trials
attest to a period of risk of postoperative VTE
extending for a considerable time after major ortho- Acknowledgments
pedic procedures. The duration of risk is particu-
larly prolonged in THA and HFS, persisting for at We acknowledge the professional writing support of
least 5 weeks after surgery. Dr Tim Norris and Dr Rachel Spice from Elsevier
Increasingly, patients undergoing major ortho- and support from sanofi-aventis. The manuscript
pedic surgery are discharged earlier than in the was reviewed by sanofi-aventis prior to its sub-
past. The trend toward more frequent use of mini- mission but the content was not influenced. The
mally invasive procedures and financial pressures opinions expressed are those of the authors and no
for early patient discharge is resulting in hospital other party.
stays of 3 to 4 days becoming common. Thus, unless
steps are taken to ensure patients receive effective
prophylaxis for an adequate duration, they will
remain at risk of a potentially fatal VTE event or a Declaration of Conflicting Interests
symptomatic DVT or PE event following hospital
discharge. Improvements in the implementation David Warwick has received honoraria for lectures
and quality of VTE prophylaxis are advocated by or consulting from sanofi-aventis, Orthofix, Boehrin-
various government and professional organizations, ger-Ingelheim, Novamedix, Bayer. Nadia Rosencher
for example, the Surgical Care Improvement Proj- has received honoraria for lectures or consulting
ect in the United States and the Venous Throm- from Bayer, Boehringer Ingelheim, BMS, Glaxo
boembolism in Hospitalized Patients Expert Smith Kline.
Working Group in the United Kingdom.65,66 These
initiatives are important in reducing the risk of mor-
tality and morbidity in patients undergoing major
orthopedic surgery. Funding
In summary, there is a wealth of evidence that
the ‘‘critical thrombosis period’’ associated with Editorial support has seen funded by sanofi-aventis.
major orthopedic surgery begins at the time of The authors or any of their associated institutions
surgery and extends for at least 5 to 6 weeks after have received no financial or any other compensa-
surgery. Current clinical guidelines present the tion for this work.

10. The critical thrombosis period in major orthopedic surgery / Warwick and Rosencher 403
References 11. Dahl OE, Gudmundsen TE, Haukeland L. Late occur-
ring clinical deep vein thrombosis in joint-operated
1. Geerts WH, Bergqvist D, Pineo GF, et al. Prevention of patients. Acta Orthop Scand. 2000;71(1):47-50.
venous thromboembolism: American College of Chest 12. White RH, Zhou H, Romano PS. Incidence of sympto-
Physicians Evidence-Based Clinical Practice Guidelines matic venous thromboembolism after different elective
(8th Edition). Chest. 2008;133(suppl):381S-453S. or urgent surgical procedures. Thromb Haemost.
2. Cardiovascular Disease Educational and Research 2003;90(3):446-455.
Trust; Cyprus Cardiovascular Disease Educational and 13. Oster G, Ollendorf DA, Vera-Llonch M, Hagiwara M,
Research Trust; European Venous Forum; Interna- Berger A, Edelsberg J. Economic consequences of
tional Surgical Thrombosis Forum; International Union venous thromboembolism following major orthopaedic
´
of Angiology; Union Internationale de Phlebologie. Pre- surgery. Ann Pharmacother. 2004;38(3):377-382.
vention and treatment of venous thromboembolism. 14. Warwick D, Friedman RJ, Agnelli G, et al. Insufficient
International Consensus Statement (guidelines accord- duration of venous thromboembolism prophylaxis after
ing to scientific evidence). Int Angiol. 2006;25(2): total hip or knee replacement when compared with the
101-161. time course of thromboembolic events: findings from the
˚
3. Bjørnara BT, Gudmundsen TE, Dahl OE. Frequency and Global Orthopaedic Registry. J Bone Joint Surg Br.
timing of clinical venous thromboembolism after major 2007;89(10):799-807.
joint surgery. J Bone Joint Surg Br. 2006;88(3):386-391. 15. Dahl OE, Andreassen G, Aspelin T, et al. Prolonged
4. Hull RD, Pineo GF, Stein PD, et al. Timing of initial thromboprophylaxis following hip replacement sur-
administration of low-molecular-weight heparin prophy- gery–results of a double-blind, prospective, randomised,
laxis against deep vein thrombosis in patients following placebo-controlled study with dalteparin (Fragmin).
elective hip arthroplasty: a systematic review. Arch Thromb Haemost. 1997;77(1):26-31.
Intern Med. 2001;161(16):1952-1960. 16. Lassen MR, Borris LC, Anderson BS, et al. Efficacy and
5. Hull RD, Brant RF, Pineo GF, Stein PD, Raskob GE, safety of prolonged thromboprophylaxis with a low mole-
Valentine KA. Preoperative vs postoperative initiation cular weight heparin (dalteparin) after total hip arthro-
of low-molecular-weight heparin prophylaxis against plasty–the Danish Prolonged Prophylaxis (DaPP)
venous thromboembolism in patients undergoing elec- Study. Thromb Res. 1998;89(6):281-287.
tive hip replacement. Arch Intern Med. 1999;159(2): 17. Prandoni P, Bruchi O, Sabbion P, et al. Prolonged
137-141. thromboprophyalxis with oral anticoagulants after total
6. Hull RD, Pineo GF, Francis C, et al. Low-molecular- hip arthroplasty: a prospective controlled randomized
weight heparin prophylaxis using dalteparin extended study. Arch Intern Med. 2002;162(17):1966-1971.
out-of-hospital vs in-hospital warfarin/out-of-hospital 18. Eriksson BI, Lassen MR;. PENTasaccharide in HIp-
placebo in hip arthroplasty patients: a double-blind, ran- FRActure Surgery Plus Investigators. Duration of pro-
domized comparison. North American Fragmin Trial phylaxis against venous thromboembolism with fonda-
Investigators. Arch Intern Med. 2000;160(14): parinux after hip fracture surgery: a multicenter,
2208-2215. randomized, placebo-controlled, double-blind study.
7. Leclerc JR, Gent M, Hirsch J, Geerts WH, Ginsberg JS. Arch Intern Med. 2003;163(11):1337-1342.
The incidence of symptomatic venous thromboembolism 19. Eriksson BI, Dahl OE, Rosencher N, et al. Dabigatran
during and after prophylaxis with enoxaparin: a multi- etexilate versus enoxaparin for prevention of venous
institutional cohort study of patients who underwent hip thromboembolism after total hip replacement: a rando-
or knee arthroplasty. Canadian Collaborative Group. mised, double-blind, non-inferiority trial. Lancet.
Arch Intern Med. 1998;158(8):873-878. 2007;370(15):949-956.
8. Planes A, Vochelle N, Darmon JY, Fagola M, 20. Kakkar AK, Brenner B, Dahl OE, et al. Extended dura-
Bellaud M, Huet Y. Risk of deep-venous thrombosis tion rivaroxaban versus short-term enoxaparin for the
after hospital discharge in patients having undergone prevention of venous thromboembolism after total hip
total hip replacement: double-blind randomised com- arthroplasty: a double-blind, randomised controlled
parison of enoxaparin versus placebo. Lancet. trial. Lancet. 2008;372(9632):31-39.
1996;348(9022):224-228. 21. Dickson BC. Venous thrombosis: on the history of Virch-
¨
9. Bergqvist D, Benoni G, Bjorgell O, et al. Low-molecular- ow’s triad. UTMJ. 2004;81:166-171.
weight heparin (enoxaparin) as prophylaxis against 22. Warwick D, Martin AG, Glew D, Bannister GC. Mea-
venous thromboembolism after total hip replacement. surement of femoral vein blood flow during total hip
N Engl J Med. 1996;335(10):696-700. replacement. Duplex ultrasound imaging with and with-
10. White RH, Romano PS, Zhou H, Rodrigo J, Bargar W. out the use of a foot pump. J Bone Joint Surg Br.
Incidence and time course of thromboembolic outcomes 1994;76(6):918-921.
following total hip or knee arthroplasty. Arch Intern 23. Dahl OE, Aspelin T, Lyberg T. The role of bone trauma-
Med. 1998;158(14):1525-1531. tization in the initiation of proximal deep vein

11. 404 Clinical and Applied Thrombosis/Hemostasis / Vol. 16, No. 4, July/August 2010
thrombosis during cemented hip replacement surgery in considerations related to timing of administration. Am J
pigs. Blood Coagul Fibrinolysis. 1995;6(8):709-717. Cardiovasc Drugs. 2007;7(1):1-15.
24. Wilson D, Cooke EA, McNally MA. Altered venous func- 38. Bouma BN, Mosnier LO. Thrombin activatable fibrino-
tion and deep venous thrombosis following proximal lysis inhibitor (TAFI) – how does thrombin regulate fibri-
femoral fracture. Injury. 2002;33(1):33-39. nolysis? Ann Med. 2006;38(6):378-388.
25. Planes A, Vochelle N, Fagola M. Total hip replacement 39. Vera-Llonch M, Hagiwara M, Oster G. Clinical and
and deep vein thrombosis. A venographic and necropsy economic consequences of bleeding following major
study. J Bone Joint Surg Br. 1990;72(1):9-13. orthopedic surgery. Thromb Res. 2006;117(5):569-577.
26. McNally MA, Bahadur R, Cooke EA, Mollan RA. 40. Aveline C, Bonnet F. Delayed retroperitoneal haema-
Venous haemodynamics in both legs after total knee toma after failed lumbar plexus block. Br J Anaesth.
replacement. J Bone Joint Surg Br. 1997;79(4):633-637. 2004;93(4):589-591.
27. Sikorski JM, Hampson WG, Staddon GE. The natural 41. Hsu DT. Delayed retroperitoneal haematoma after failed
history and aetiology of deep vein thrombosis after total lumbar plexus block. Br J Anaesth. 2005;94(3):395.
hip replacement. J Bone Joint Surg Br. 1981;63-B(2): 42. Weller RS, Gerancher JC, Crews JC, Wade KL. Exten-
171-177. sive retroperitoneal hematoma without neurologic defi-
28. Dahl OE, Aspelin T, Arnesan H, et al. Increased activa- cit in two patients who underwent lumbar plexus block
tion of coagulation and formation of late deep venous and were later anticoagulated. Anesthesiology. 2003;98:
thrombosis following discontinuation of thrombopro- 581-585.
phylaxis after hip replacement surgery. Thromb Res. 43. Hampson WG, Harris FC, Lucas HK, et al. Failure of
1995;80(4):299-306. low-dose heparin to prevent deep-vein thrombosis after
29. Wilson D, Cooke EA, McNally MA. Changes in coagul- hip-replacement arthroplasty. Lancet. 1974;2(7884):
ability as measured by thrombelastography following 795-797.
surgery for proximal femoral fracture. Injury. 44. Robinson KS, Anderson DR, Gross M, et al. Ultrasono-
2002;32(10):765-770. graphic screening before hospital discharge for deep
30. ¨
Strebel N, Prins M, Agnelli G, Buller HR. Preoperative venous thrombosis after arthroplasty: the post-
or postoperative start of prophylaxis for venous throm- arthroplasty screening study. A randomized, controlled
boembolism with low-molecular-weight heparin in elec- trial. Ann Intern Med. 1997;127(6):439-445.
tive hip surgery? Arch Intern Med. 2002;162(13): 45. Ricotta S, Iorio A, Parise P, Nenci GG, Agnelli G. Post
1451-1456. discharge clinically overt venous thromboembolism in
31. Rosencher N, Bonnet MP, Sessler DI. Selected new orthopaedic surgery patients with negative venogra-
antithrombotic agents and neuraxial anaesthesia for phy—an overview analysis. Thromb Haemost.
major orthopaedic surgery: management strategies. 1996;76(6):887-892.
Anaesthesia. 2007;62(11):1154-1160. 46. Warwick D, Samama MM. The contrast between veno-
32. Moen V, Dahlgren N, Irestedt L. Severe neurological graphic and clinical endpoints in trials of thrombopro-
complications after central neuraxial blockades in Swe- phylaxis in hip replacement. J Bone Joint Surg Br.
den 1990-1999. Anesthesiology. 2004;101(4):950-959. 2000;82(4):480-482.
33. Horlocker T, Wedel DJ, Benzon HT, et al. Regional 47. McNally MA, Mollan RA. Total hip replacement, lower
anesthesia in the anticoagulated patient: defining the limb blood flow and venous thrombogenesis. J Bone Joint
risks (the second ASRA Consensus Conference on Neur- Surg Br. 1993;75(4):640-644.
axial Anesthesia and Anticoagulation). Reg Anesth Pain 48. Warwick D, Williams MH, Bannister GC. Death and
Med. 2003;28(3):172-197. thromboembolic disease after total hip replacement. A
34. Tryba M. European practice guidelines: thromboembo- series of 1162 cases with no routine chemical prophy-
lism prophylaxis and regional anesthesia. Reg Anesth laxis. J Bone Joint Surg Br. 1995;77(1):6-10.
Pain Med. 1998;23(6 Suppl 2):178-182. 49. Warwick DJ, Whitehouse S. Symptomatic venous
35. Samama CM, Albaladejo P, Benhamou D, et al. Venous thromboembolism after total knee replacement. J Bone
thromboembolism prevention in surgery and obstetrics: Joint Surg Br. 1997;79(5):780-786.
clinical practice guidelines. Eur J Anaesthesiol. 50. Lie SA, Engesaeter LB, Havelin LI, Gjessing HK,
2006;23(2):95-116. Vollset SE. Mortality after total hip replacement: 0-10-
36. Turpie AG, Bauer KA, Eriksson BI, Lassen MR. Fonda- year follow-up of 39,543 patients in the Norwegian
parinux vs enoxaparin for the prevention of venous Arthroplasty Register. Acta Orthop Scand. 2000;71(1):
thromboembolism in major orthopaedic surgery: a 19-27.
meta-analysis of 4 randomized double-blind studies. 51. White RH, Romano PS, Zhou H, Rodrigo J, Bargar W.
Arch Intern Med. 2002;162(16):1833-1840. Incidence and time course of thromboembolic outcomes
37. Tribout B, Colin-Mercier F. New versus established following total hip or knee arthroplasty. Arch Intern
drugs in venous thromboprophylaxis: efficacy and safety Med. 1998;158(14):1525-1531.

12. The critical thrombosis period in major orthopedic surgery / Warwick and Rosencher 405
52. A Joint Report from The Royal College of Surgeons of 59. Heit JA. Low-molecular-weight heparin: the optimal
England and the British Orthopaedic Association. duration of prophylaxis against postoperative venous
National Total Hip Replacement Outcome Study. Final thromboembolism after total hip or knee replacement.
Report to the Department of Health. Revised June Thromb Res. 2001;101(1):V163-V173.
2000. http://www.rcseng. ac.uk/publications/docs/ 60. Rosencher N, Vielpeau C, Emmerich J, Fagnani F,
hip_replacement.html/attachment_download/pdffile. Samama CM. the ESCORTE group. Venous throm-
Accessed April 17, 2009. boembolism and mortality after hip fracture surgery: the
˚
53. Dahl OE, Gudmundsen TE, Bjørnara BT, Solheim DM. ESCORTE study. J Thromb Haemost. 2005;3(9):
Risk of clinical pulmonary embolism after joint surgery 2006-2014.
in patients receiving low-molecular-weight heparin pro- 61. Perez JV, Warwick DJ, Case CP, Bannister GC. Death
phylaxis in hospital: a 10-year prospective register of after proximal femoral fracture–an autopsy study. Injury.
3,954 patients. Acta Orthop Scand. 2003;74(3): 1995;26(4):237-240.
299-304. 62. Todd CJ, Freeman CJ, Camilleri-Ferrante C, et al. Dif-
54. Colwell CW Jr, Collis DK, Paulson R, et al. Comparison ferences in mortality after fracture of hip: the east
of enoxaparin and warfarin for the prevention of venous Anglian audit. BMJ. 1995;310(6984):904-908.
thromboembolic disease after total hip arthroplasty. Eva- 63. Lawrence VA, Hilsenbeck SG, Noveck H, Poses RM,
luation during hospitalization and three months after Carson JL. Medical complications and outcomes after
discharge. J Bone Joint Surg Am. 1999;81(7):932-940. hip fracture repair. Arch Intern Med. 2002;162(18):
55. Mannucci PM, Citterio LE, Panajotopoulos N. Low- 2053-2057.
dose heparin and deep-vein thrombosis after total hip 64. Eikelboom JW, Quinlan DJ, Douketis JD. Extended-
replacement. Thromb Haemost. 1976;36(1):157-164. duration prophylaxis against venous thromboembolism
56. Trowbridge A, Boese CK, Woodruff B, Brindley HH Sr, after total hip or knee replacement: a meta-analysis of
Lowry WE, Spiro TE. Incidence of posthospitalization the randomised trials. Lancet. 2001;358(9275):9-15.
proximal deep venous thrombosis after total hip arthro- 65. The Surgical Care Improvement Project. MedQIC.
plasty. A pilot study. Clin Orthop Relat Res. 1994;299: Oklahoma City, OK. Scip Project Information.
203-208. www.medqic.org/dcs/ContentServer?cid¼11229049304
57. Lassen MR, Borris LC, Anderson BS, et al. Efficacy and 22&pagename¼Medqic%2FContent%2FParentShellTe
safety of prolonged thromboprophylaxis with a low mole- mplate&parentName¼Topic&c¼MQParents. Accessed
cular weight heparin (dalteparin) after total hip arthro- April 17, 2009.
plasty–the Danish Prolonged Prophylaxis (DaPP) 66. DH/CMO. Report of the independent expert working
Study. Thromb Res. 1998;89(6):281-287. group on the prevention of venous thromboembolism
58. Comp PC, Spiro TE, Friedman RJ, et al. Prolonged enox- in hospitalised patients, March 2007. http://www.dh.gov
aparin therapy to prevent venous thromboembolism .uk/en/Publicationsandstatistics/Publications/Publicatio
after primary hip or knee replacement. Enoxaparin Clin- nsPolicyAndGuidance/DH_073944? IdcService¼GET_
ical Trial Group. J Bone Joint Surg Am. 2001;83-A(3): FILE&dID¼138346&Rendition¼Web. Accessed April
336-345. 17, 2009.
For reprints and permissions queries, please visit SAGE’s Web site at http://www.sagepub.com/journalsPermissions.nav.