Memorandum by Mr David Warwick (VT 1)
A SYNTHESIS OF RECENT RESEARCH CONCERNING THE PREVENTION
OF VENOUS THROMBOEMBOLISM IN ORTHOPAEDIC SURGERY
Thromboprophylaxis is a controversial and changing
topic. Some new concepts are presented in this submission.
Some have questioned the very need for thromboprophylaxis.
Various reasons are given to include the relatively low frequency
of symptomatic thromboembolic events, the risk of bleeding, the
possibility of late infection, the reliance on surrogate endpoints
such as venography and finally "no evidence of effect"
(the beta error misinterpreted as "evidence of no effect").
However, in our current environment of risk management it would
be wise to remember that the weight of evidence supports the view
that thromboembolism is a potentially serious complication and
that on the balance of probability the risk can be diminished
(The THRIFT Group 1998, International Concensus Statement 2001
Geerts et al 2002). Anecdote aside, there is no evidence
that careful prophylaxis causes major wound bleeding, infection,
implant loosening or death.
The fatal PE rate without prophylaxis
after hip and knee replacement is probably in the region of 0.4%.
One might say that 0.4% is a rare and therefore unimportant event
rate. However, the individual who dies is 100% dead. With 1.25
million hip and knee replacements in Europe each year, 0.4% represents
5,000 fatalities annually or the capacity of an Airbus each month
. . . Even very low death rates are important.
Symptomatic venous thromboembolism (VTE)
is the most common complication after arthroplasty (around 4%).
This makes it more common than dislocation and infection combined.
Every orthopaedic surgeon accepts the need for antibiotic prophylaxis
to avoid infection and a meticulous surgical approach to avoid
dislocation. Why not accept the need for safe and effective thromboprophylaxis?
Fatal PE. John Charnley (Crawford et
al 1968) showed, by comparing phenindione with control in
a randomised trial of 900 patients, that extreme anticoagulation
merely exchanges a reduction in fatal PE for an increase in fatal
bleeding. However, more judicious anticoagulation can probably
reduce death. Fatal PE is very rare indeed in patients taking
prophylactic warfarin at adequate levels (for example, no fatalities
in 3,000 consecutive hip arthroplasty patients (Amstutz et
al 1989)). There is fairly good evidence that death rates
can be reduced by heparin. A meta-analysis (Collins et al
1988) of all the early randomised trials of heparin in orthopaedic
surgery and showed that the fatal PE rate was reduced by 66% in
those receiving heparin rather than placebo or nothing. The overall
death rate, as well as the fatal PE rate, was reduced by heparin.
However, there was an increase in bleeding of two-thirds in those
taking heparin (2% absolute increase) (Figure 1). Death rates
are now so low even without prophylaxis that a randomised study
is unlikely to be large enough (about 90,000 patients) to study
death as an end point (Warwick et al 1995). We will have
to rely upon surrogate endpoints.
Symptomatic VTE. It has been argued (Warwick
and Samama 2000) that the evidence for thromboprophylaxis is based
apon a surrogate outcomeusually venographyrather
than a clinical outcome. However, there is in fact evidence that
reduced DVT rates correspond with reduced symptomatic event rates.
The heparin meta-analysis paper (Collins et
al 1988) showed a similar risk reduction for both asymptomatic
DVT (scintigraphy or venography) of 67% (60.5% to 20.3%) and
fatal PE of 68% (1.9% to 0.6%). (Figure 1). However, this
paper was weakened because it was based apon a rather heterogenous
group of small studies of various orthopaedic patients diagnosed
primarily with iodinated fibrinogen.
More robust data are now available. A meta-analysis
(Hull et al 2001a) shows that extending the duration of
LMWH for about five weeks after hip replacement will reduce the
venographic DVT rate from 21% to 8.2% (risk reduction 61%). These
studies were large enough to show that the frequency of symptomatic
VTE was reduced by the same proportion from 4.5% to 1.7%, (risk
reduction 62%). (Figure 2). Similarly, when placebo was compared
with pentasaccharide in a double blind RCT for 4 weeks after hip
fracture surgery, the venographic DVT risk was reduced by 95.9%
(77/220 or 35% vs 3/208 or 1.4%) and the symptomatic event
rate by a similar 88.8% (9/330 or 2.7% vs 1/326 or 0.3%
(Erikkson et al 2003). (Figure 3).
Thus we can now be confident that venographic
surrogates do reflect clinical reality.
Chronic venous insufficiency. Whether
joint replacement predisposes to chronic venous insufficiency,
and whether this risk can be reduced by prophylaxis, is not yet
Until recently, most clinical trials studied
the use of prophylaxis in arthroplasty for only seven to 10 dayswhilst
the patient was in hospital. With this strategy, LMWH would reduce
the venographic DVT rate by 60% (2). However, there is consistent
evidence from several sources that half of symptomatic VTE after
knee replacement and two-thirds after hip replacement occur beyond
the second weekusually when the patient has been discharged
from hospital (White et al 1998, Colwell et al 1999,
Dahl et al 2000). The Total Hip Replacement Outcome Study
(Gregg et al 2000) shows that venous thromboembolism is
the commonest cause of readmission after hip replacement. As described
above, several recent randomised trials have consistently shown
that the risk of thrombosis (both venographic and symptomatic)
after hospital discharge in hip surgery can be reduced
by two-thirds if low molecular weight heparin or pentasaccharide
is continued for at least four weeks after surgery (Hull et
al 2001a, Cohen et al 2001, Eikelboom 2001, Erikkson
2003). (Figure 3,4). The advantage for extended prophylaxis in
knee replacement is not so clear.
These studies show that the number-needed-to-treat
(NNT) to prevent one symptomatic DVT or PE after hip replacement
is 37; from this figure, one can calculate cost effectiveness.
Because the cost of LMWH is relatively low, and the cost of investigation
or treatment of thromboembolism relatively high, this is likely
to be a cost effective approach (Friedman and Dunsworth 2000,
Sarasin and Bounameaux 2000).
Even if these statisics are not thought compelling
enough to extend prophylaxis for five weeks, it should be remembered
hospital stays after arthroplasty are falling. Discharge at four
days after surgery is not uncommon and minimally invasive day
case hip surgery is on the horizon. Even the most sceptical may
realise that prophylaxis for only one day is probably too short.
Thus the practical issue of administering prophylaxis after discharge
is realwhen do you need to stop it, who gives it, who will
pay for it . . .
Aspirin is superficially attractive. It is a
cheap, readily available, familiar tablet. Surely if aspirin is
given, the surgeon won't have to worry, the patient will be fine
and the lawyers will be out of work. Initial meta-analysis suggested
that aspirin might reduce the frequency of DVT and PE (ATC 1994).
However, the recent PEP study (PEP 2000) showed aspirin is not
as helpful as might have been hoped. (Figure 4) Over 13,000 hip
fracture patients were randomised to have either aspirin or placebo.
The death rate was identical in each group. The risk reduction
for symptomatic VTE from 2.5% to 1.6% was only about 30%half
what one would expect from LMWH and one-third from pentasaccharide)
(Figure 5). This reduced risk of VTE was matched by an increased
risk of transfusion, gastro-intestinal bleeding and wound bleeding
(Fig 5). In the supplementary group of 4,000 hip and knee replacement
patients, there was an insignificant difference in symptomatic
VTE (Cohen and Quinlan 2000). In other words, aspirin has a relatively
weak thromboprophylactic effect, carries an alternative complication
rate and its use might deprive patients of safer or more effective
prophylaxis. It is not recommended by the two largest evidence-based
Consensus groups (Geerts et al 2001, ICC 2001). Furthermore,
it is not licensed for thromboprophylaxis in the United Kingdom.
Warfarin has been widely used in North America
and the United Kingdom for prophylaxis. Used carefully, death
is exceedingly rare and it is as effective as LMWH in reducing
venographic DVT. Its use is supported by the main Consensus Groups.
It can be delivered beyond hospital discharge to protect against
the risk of late onset VTE. However it has many drawbacks and
for this reason is regarded as more or less obsolete by many in
Western Europe and Scandinavia. It requires regular monitoring,
which is expensive and time consuming. If started too close to
surgery or at too high a dose, there will be a risk of bleeding.
If started judiciouslylater and at a lower dosethere
will be an interval of several days during which the patient will
be unprotectedat their most thrombogenic time. Warfarin
interacts with many drugs and alcohol. On objective comparison,
it is difficult to see an advantage for Warfarin over LMWH or
pentasaccharide, except that (arguably) it is more convenient
to continue beyond discharge.
Fondaparinux (Arixtra, Sanofi-Synthelabo, Guildford
UK) is a pentasaccharide which offers a new, effective and relatively
safe pharmacological approach. It precisely inhibits factor Xa,
which is a key component of coagulation. It has been meticulously
compared with LMWH in over 7,300 hip replacement, knee replacement
and hip fracture patients (Turpie et al 2002). The overall
VTE rate at 11 days after surgery (venographic DVT plus symptomatic
DVT or PE) was reduced from 13.7% with Enoxaparin to 6.8% with
Arixtra (odds reduction 55.2%; 95% confidence interval 45.8 to
63.1%, p<0.001) (Figure 6). Some of this advantage in VTE (and
disadvantage in bleeding) may be explained by a different timing
schedule than LMWHrather closer to surgery. With respect
to bleeding, it is as well to remember that an omelette cannot
be made without cracking an egg. Fondaparinux appeared to have
some increased minor bleeding (so called bleeding index) but no
major bleeding side effects in comparison with LMWH. It must be
given at least six hours after surgery and after removal
of the spinal/epidural catheter to avoid the risk of surgical
or neuraxial (ie spinal) bleeding. The case for pentasaccharide
has been critically analysed (Lowe et al 2003) with some
discussion about cost, the choice of endpoint, the relevance of
the bleeding and the means of reversal. Nevertheless, the risk
reductions presented for DVT are enticing.
Neuraxial (ie spinal or epidural) anaesthesia
conveys many benefits to orthopaedic patients (Rodgers et al
2000). The mortality after surgery is reduced by 30%, post-operative
analgesia is enhanced and it is even weakly thromboprophylactic
(Prins and Hirsh 1990). Initial European experience suggested
that neuraxial anaesthesia could be safely used in the presence
of prophylactic anticoagulants (Bergqvist 1992). However, more
recently the American Food and Drug Administration has raised
concerns that on occasions a spinal haematoma may develop. It
is therefore prudent to avoid giving neuraxial anaesthesia and
chemical prophylaxis within at least six hours of each other (Horlocker
The ideal chemical agent would be taken orally.
This would overcome the difficulty of reconciling the clear need
for extended prophylaxis with the pragmatic issue of who will
administer it. Aspirin (anti-platelet rather than anti-thrombotic)
may fulfil this, but the efficacy is weak and there is no good
evidence for its extended use. Warfarin has many disadvantages
described above. Melagatran (AstraZeneca, UK) is a recently-developed
direct oral thrombin inhibitor which has a number of important
advantages over warfarin. There is a wide therapeutic and safety
window; no monitoring is needed; it is not known to interact with
other medications. Recent trials show equivalence with LMWH in
prophylaxis after arthroplasty (Erikkson 2001, Heit 2001, Erikkson
2002). Future trials may show its efficacy for extended prophylaxis
in which case a pragmatic solution to this important problem would
There is a dilemma with chemical prophylaxis:
the closer to surgery that it is administered, the better the
thromboprophylaxis but the greater chance of bleeding complications.
If LMWH is given before surgery (as recommended in Europe), then
because of a relatively short half life, serum levels will be
too low for any prophylactic effect. If LMWH is delayed until
after surgery (as recommended in North America), then thrombi
may have already begun to form during the very thrombogenic operation.
Prophylaxis needs to be given close, but not too close, to surgeryso
called "just in time prophylaxis" (Hull et al
2001b). The optimum moment to administer LMWH or Fondaparinux
is probably around six to eight hours after surgery.
The Art, rather than Science, of clinical medicine
is to apply knowledge in a balanced way, tailored to the needs
of the individual patient.
Thromboprophylaxis has been often regarded as
a dichotomy- either chemical or mechanical. This
risks throwing the baby out with the bathwater. The advantages
of chemicals (ease of use, relative cheapness and efficacy) must
be weighed against the potential for bleeding both into the surgical
wound and into the spinal cord following neuraxial anaesthesia.
The advantages of mechanical prophylaxis such as the Foot Pump
(no bleeding side effects, no interactions, reasonable efficacy)
must be weighed against the disadvantages (compliance, refitting
when mobilising, impracticality of extended use). A sensible approach
would be to use the Foot Pump as soon as possible after injury
or surgery and then to switch to chemical prophylaxis once the
risk of bleeding has subsided and for as long as the risk of thromboembolism
pertains. For patients with a particularly risk of thrombosis,
the two can be combined in the hope of a synergistic effect (although
this has not yet been studied). There are no clinical trials which
have directly addressed this approach and perhaps there will not
Care pathways and guidelines are becoming endemic.
They should ensure the routine and automatic provision of important
care, yet allow flexibility when individual patient circumstances
require. This should give the patient the benefit of best practice
and give the hospital protection against risk.
It is wise for each orthopaedic Department to
combine common sense with evidence and then publish guidelines
for thromboprophylaxis. These guidelines can then be incorporated
into care pathways. The author (hopefully not tainted by wild
speculation or subconscious prejudice) suggests that the following
would not be unreasonable:
Risk assessment: all patients should
have a risk assessment to detect particularly high risk (previous
VTE, family history, malignancy, likely prolonged immobility or
Hip replacement: regional anaesthesia
and early mobilisation should be encouraged. For those surgeons
who are comfortable with chemical prophylaxis (reassured by the
literature and their experience) then they should start LMWH or
pentasaccharide no less than six to eight hours after surgery
and regional block. For those surgeons who are concerned about
peri-operative bleeding, a mechanical device such as a Foot Pump
should be started in the recovery room. It should be continued
for the entire hospital stay for those thought to have a particularly
high risk of VTE. Otherwise it could be stopped once the patient
begins to mobilise. Chemical prophylaxis should be started once
the surgeon feels that the risk of bleeding has subsided and at
least six hours after any indwelling epidural catheter has been
removed. In an ideal world it should be continued for five weeks
(notwithstanding the logistical issues of funding and administration).
Hip fracture: In some patients, medical
and social co-morbidity may occasionally suggest a more holistic
approach. Otherwise, a mechanical device should be applied as
soon after injury as possible (ie in the Emergency Department).
The device should be used, and chemical prophylaxis instigated
in the same way and for the same duration as hip replacement.
Knee replacement: The risk of soft tissue
side effects is higher in knee replacement, yet VTE is more resistant
to prophylaxis. Regional anaesthesia should be encouraged. A mechanical
device should be started in recovery and continued, if tolerated,
for as long as the patient is in hospital. Chemical prophylaxis
should be started as soon after surgery/regional block as the
surgeon feels is safe and continued for the entire hospital stay.
For those going home in less than 10 days, or for those with other
risk factors, chemical prophylaxis should be considered beyond
Other orthopaedic operations: There are
so few data on both epidemiology and prophylaxis that one has
to resort to common sense. If the risk assessment suggests a minimal
risk of VTE then the cost and potential side-effects of peri-operative
chemical prophylaxis may not be justified. However, if there is
a potential risk, then a sensibly-timed combination of mechanical
and chemical prophylaxis should be devised. Major lower limb
trauma and spinal surgery certainly carry a risk of symptomatic
thromboembolism yet carry a greater threat of bleeding. A longer
period of mechanical prophylaxis, followed by extended chemical
prohylaxis when safe, seems sensible. Day case arthroscopy
causes very occasional VTE problems yet practical prophylaxis
is difficult (the patient will have gone home before it is safe
to give chemicals). The best one can do is to carry out a careful
risk assessment and provide anyone with a higher risk with an
injection prior to discharge and follow with home-prophylaxis
until fully mobile.
Updating: Scientific knowledge, clinical
experience and attitudes to risk change with time. It takes effort
to get new medications on to a hospital formulary. It takes persuasion
to change habits. A group of interested individuals in a department
or hospital can keep the local guidelines acceptable and current.
Opinions about and options for thromboprophylaxis
are continuously developing; perhaps this article helps surgeons
to think about this important aspect of surgical care.
Amstutz H, Friscia D A, Dorey F,
Carney BT (1989). Warfarin prophylaxis to prevent mortality from
pulmonary embolism after total hip replacement. J Bone Joint
Anti-platelet Triallists Collaboration,
(1994) Collaborative overview of randomised trials of antiplatelet
therapy; reduction in venous thrombosis and pulmonary embolism
by anti platelet prophylaxis amongst surgical and medical patients.
British Medical Journal: 308:235-246.
Bergqvist D, Lindblad B, Matsch T.
Low molecular weight heparin for thromboprophylaxis and epidural/regional
anaesthesiais there a risk? Acta Anaesthesiol Scand 1992;
Cohen A, Bailey C S, Alikhan R, Cooper
D J. (2001) Extended thromboprophylaxis with low molecular weight
heparin reduces symptomatic venous thromboembolism following lower
limb arthroplasty. Thrombosis and Haemostasis; 85:940-941.
Cohen A, Quinlan D. (2000) PEP trial
(letter). Lancet: 356:247.
Collins R, Scrimgeour A, Yusef S,
Peto R (1988). Reduction in fatal pulmonary embolism and venous
thrombosis by perioperative administration of subcutaneous heparin.
N Eng J Med; 318:1162-73.
Colwell C W, Collis D K, Paulson
R, McCutchen J W, Bigler GT, Lutz S, Hardwick M E (1998). Comparison
of enoxaparin and warfarin for the prevention of venous thromboembolic
disease after total hip arthroplasty. Evaluation during hospitalisation
and three months after discharge. JBJS (A) 1999; 81A:932-939.
Crawford W J, Hillman F, Charnley
J (1968). A clinical trial of prophylactic anticoagulant therapy
in elective hip surgery. Internal Publication 14, Centre
for Hip Surgery, Wrightington Hospital.
Dahl O E, Gudmunsen T E, Haukeland
L. Late occurring clinical deep vein thrombosis in joint-operated
patients. Acta Orthop Scand 2000; 71:47-50.
Eikelboom J W, Quinlan D J, Douketis
J D. (2001) Extended duration prophylaxis against venous thromboembolism
after total hip or knee replacement: a meta-analysis of the randomised
trials. Lancet: 358:9-15.
Erikkson B I, Arfwidsson A C, Frison
L et al. (2002). A dose ranging study of the oral direct
thrombin inhibitor ximelagatran and its subcutaneous form Melagatran
compared with dalteparin in the prophylaxis of thromboembolism
after hip or knee replacement :METHRO I. Thromb Haem; 87:231-237.
Erikkson B I, Lassen M R. Duration
of prophylaxis against venous thromboembolism with fondapainux
after hip fracture surgery: a multicentre, randomized, placebo-controlled,
double blind study. Arch Int Med 2003; 163:1337-1342.
Erikkson B I, Ogren M, Agnelli G
et al (2001). The oral direct thrombin inhibitor ximelagatran
and its subcutaneous form Melagatran compared with enoxaparin
as thromboprophylaxis after total hip or knee replacement (METHRO
III). Thromb Haem 86S; OC46.
Friedman R J, Dunsworth G A. (2000)
Cost analysis of extended prophylaxis with enoxaparin after hip
arthroplasty. Clinical Orthopaedics and Related Research:
Geerts et al. (2001) Prevention
of venous thromboembolism. Chest; 119:132-175S.
Gregg P J, Devlin H B National Total
Hip Replacement Outcome Study 2000. www.rcseng.ac.uk
Heit J A, Colwell C W, Francis C
W et al (2001). Comparison of the oral direct thrombin
inhibitor, ximelagatran compared with dalteparin in the prophylaxis
of venous thromboembolism after knee replacement. Arch Int
Horlocker T T (2001). Low molecular
weight heparin and neuraxial anesthesia. Thrombosis Research
Hull R D, Pineo G F, Stein P D, Mah
A F, MacIsaac S M, Dahl OE, Butcher M, Brant R F, Ghali W A,
Bergqvist D, Raskob G E. (2001a), Extended out of hospital low-molecular
weight heparin prophylaxis against deep venous thrombosis in patients
after elective hip arthroplasty: a systematic review. Ann Intern
Med. 2001 Nov 20; 135(10):858-69.
Hull R D, Pineo G F, Stein P D, Mah
A F et al (2001b) Timing of low molecular weight heparin
prophylaxis against deep venous thrombosis in patients following
elective hip arthroplasty. Arch Int Med; 161:1952-1960.
International Consensus Statement.
International Angiology, 20:1-37 2001.
Lowe G D, Sandercock P A, Rosendaal
F R. Prevention of venous thromboembolism after major orthopaedic
surgery: is Fondaparinux an advance? Lancet; 362:504-5.
PEP. Prevention of pulmonary embolism
and deep vein thrombosis with low dose aspirin: Pulmonary Embolism
Prevention trial. Lancet 2000:355:1295-1302.
Pre-operative vs postoperative initiation
of low molecular weight heparin prophylaxis against venous thromboembolism
in patients undergoing elective hip replacement. Arch Int Med
Prins M, Hirsh J. A comparison of
general anesthesia and regional anesthesia as a risk factor for
deep vein thrombosis following hip surgery: a critical review.
Thromb Haemost 1990; 64:497-500 Prins and Hirsh.
Rodgers A, Walker N, Schug S, McKee
A, Kehlet H, van Zundeert A, Sage D, Futter M, Saville G, Clark
T, MacMahon S. (2000). Reduction of post-operative mortality and
morbidity with epidural or spinal anaesthesia: results from overview
of randomised trials. B Med J; 321:1493-7.
Sarasin F P, Bounameaux H. Out of
hospital anti-thrombotic prophylaxis after total hip replacement:
low molecular weight heparin, warfarin, aspirin or nothing. (2002)
Thrombosis and Haemostasis: 87:586-92.
THRiFT Consensus Group. (1998) Risk
of and prophylaxis for venous thromboembolism in hospital patients.
Turpie A G G, Bauer K A, Eriksson
B I, Lassen MR. (2002). Fondaparinux vs Enoxaparin for the prevention
of venous thromboembolism in major orthopaedic surgery. Arch
Int Med; 162:1833-1840.
Warwick D J, Samama M M. (2000) The
contrast between radiological and clinical outcomes in orthopaedic
thromboembolism. J Bone Joint Surg 2000.
White R H, Romano P S, Zhou H, Rodrigo
J, Bargar W (1998) Incidence and time course of thromboembolic
outcomes following total hip or knee arthroplasty Arch Intern
Med; 158:1525-1531 1998.