In England, 25,000 people die of venous thromboembolism each
year1. The
association between malignant disease and thromboembolism has long been
recognized, with venous thromboembolism being confirmed at the time of autopsy
in 50% of cancer
patients2.
Orthopaedic surgery is also known to be a risk factor for venous
thromboembolism3. It
could be hypothesized, therefore, that patients with bone or soft-tissue
tumors who undergo orthopaedic surgery might be at particularly high risk for
venous thromboembolism. However, there is little information in the
literature, and there is a need for benchmarking data about the risk of venous
thromboembolism in these patients. We identified one retrospective study of a
cohort of orthopaedic oncology patients, which demonstrated a 14% risk of
proximal deep venous thrombosis but a low rate of symptomatic pulmonary
embolism4.
Guidelines for thromboembolic prophylaxis in patients undergoing major
surgery are being developed in England and Wales by the National Institute for
Health and Clinical Excellence (NICE) but are not in place yet. Scottish
guidelines recommend the individual assessment of risk factors for venous
thromboembolism and promote mechanical and/or chemical prophylaxis for those
who are at increased
risk5, and the
seventh edition of the guidelines of the American College of Chest Physicians
(ACCP) has been
published6. Studies
have suggested that the extended use of low-molecular-weight heparin
significantly reduces the risk of proximal deep venous thrombosis in patients
undergoing abdominal surgery for the treatment of
cancer7 and that
there may be added benefits associated with the use of heparin in terms of its
antineoplastic
properties8. Despite
these findings, the role of thromboprophylaxis in orthopaedic patients with
cancer is not clear, and there is a need for guidance in the treatment of this
group of patients.
The primary purposes of the present retrospective study were to determine
the prevalence of venous thromboembolism in patients with trunk or extremity
bone or soft-tissue sarcoma in our unit and to identify risk factors for
venous thromboembolism in this group. The secondary purposes were to audit the
use of prophylaxis against thromboembolism in our unit and to develop local
guidelines for thromboprophylaxis.
The present investigation was a retrospective study that was based on a
review of clinical records. Patients who had presented to the North of England
Bone and Soft Tissue Tumour Service between 1998 and 2003 with a confirmed
diagnosis of primary trunk or extremity bone or soft-tissue sarcoma were
included. Patients were excluded if they were less than sixteen years of age,
had a benign diagnosis, had been managed elsewhere, or had been lost to
follow-up less than three months after treatment.
The data items that were retrieved from a review of the case notes included
the age and gender of the patient, the anatomical location and histological
type of the tumor, the results of biopsy, the type of surgery, the use of
adjuvant chemotherapy and radiation therapy, the presence of additional risk
factors for
thromboembolism5,
and the methods of thromboprophylaxis used. The anatomical location of the
tumor was defined as trunk, upper extremity, or lower extremity. Lower
extremity tumors were further classified, according to the location of the
greatest volume of the tumor, as hip/thigh (including the pelvis), knee, or
below-the-knee (including the foot and ankle). The signs and symptoms of
suspected thromboembolic events were recorded, as were the investigations that
were performed for the evaluation of suspected thromboembolic events and
whether or not the thromboembolic event was confirmed.
Statistical Analysis
Patients with and without venous thromboembolism were compared with regard
to age and gender, the anatomical location and histological type of the tumor,
surgical intervention, additional risk factors for thromboembolism, and the
methods of thromboprophylaxis; chi-square and two-sample t test analyses were
used as appropriate. The level of significance was maintained at p = 0.05
throughout.
Of the 314 patients who presented with bone or soft-tissue tumors between
1998 and 2003, 252 met the inclusion criteria. Sixty-two patients (including
twenty-five patients who were less than sixteen years of age, eleven patients
who were found not to have had sarcoma, and twenty-six patients who were
managed elsewhere) were excluded. The clinical records relating to nine
patients could not be located. Of the remaining 252 patients, ninety-four had
a diagnosis of primary bone sarcoma and 158 had a diagnosis of primary
soft-tissue sarcoma (Table I).
The mean age at the time of presentation was fifty-three years (range, sixteen
to ninety-four years), and 137 patients (54%) were male. The anatomical
distribution of tumors is shown in Table
II. Of the 252 patients, 249 underwent orthopaedic surgery;
specifically, 147 patients (59%) underwent excision of the tumor, forty-seven
(19%) underwent excision and reconstruction (with thirty-one patients
undergoing endoprosthetic reconstruction, twelve undergoing flap or vascular
reconstruction, and four undergoing excision and bone-grafting), thirty-three
(13%) had a primary amputation, thirteen (5%) had an open biopsy, and only
nine (4%) had other procedures. Three patients could not undergo surgery
because of medical comorbidities.
The characteristics of the patients who were diagnosed with venous
thromboembolism are summarized in Table
III. Thirty-seven of the 252 patients had clinically suspected
deep venous thrombosis. Of the thirty-seven patients with suspected deep
venous thrombosis, thirty-three underwent venous ultrasonography, one
underwent magnetic resonance imaging, one had a D-dimer level alone, one had
no documented investigation, and one received the diagnosis at another center
(investigation unknown). Nine patients were confirmed to have deep venous
thrombosis on the basis of venous ultrasonography and one patient received the
diagnosis at another center (investigation unknown), resulting in a rate of
clinically evident deep venous thrombosis of 4%. Two of these ten patients had
a deep venous thrombosis after biopsy and went on to have an additional deep
venous thrombosis after definitive surgery despite the use of chemical
prophylaxis (both patients received low-molecular-weight heparin following
definitive surgery, and one also received warfarin and an inferior vena cava
filter) (Table III).
Nine patients were clinically suspected of having pulmonary embolism
because of shortness of breath, chest pain, or tachycardia. Of these nine
patients, five were investigated with a radioisotope ventilation-perfusion
scan or computed tomographic angiogram, two had chest radiographs only, one
was confirmed to have had a pulmonary embolism at the time of autopsy, and one
received the diagnosis at another center (investigation unknown). One patient
was confirmed to have had a pulmonary embolism on the basis of a
ventilation-perfusion scan, one patient died of a pulmonary embolism, and in
one case the diagnosis was made at another center (investigation unknown),
resulting in an overall rate of pulmonary embolism of 1.2% and a rate of fatal
pulmonary embolism of 0.4%. The patient for whom the diagnosis was confirmed
at the time of an autopsy died suddenly following an open biopsy for the
evaluation of a large thigh tumor, having already undergone an inconclusive
needle biopsy. This was the only fatal pulmonary embolism in the series.
Therefore, the rate of clinically evident pulmonary embolism was 1.2% (three
of 252) and the rate of fatal pulmonary embolism was 0.4% (one of 252).
The demographic and clinical data for patients with and without venous
thromboembolism are compared in Table
IV. With the numbers available, there were no significant
differences between the patients with and without venous thromboembolism in
terms of age, gender, the type of tumor (bone or soft-tissue sarcoma), or the
use of low-molecularweight heparin. However, all patients with thromboembolic
events had tumors of the hip or thigh (p = 0.0006), with a mean maximum tumor
diameter of 16.6 cm (range, 11 to 23 cm). The anatomical site was the
acetabulum for one patient (Case 2) and the distal part of the femur for one
patient (Case 4); the rest of the tumors were all soft-tissue tumors of the
thigh (Table III). The rate of
clinically evident venous thromboembolism in patients with hip or thigh tumors
was 9.6% (thirteen of 136), compared with 0% for patients with tumors in other
anatomical sites. In the group of patients with tumors in the hip or thigh,
there were no significant differences between patients with and without venous
thromboembolism in terms of age, gender, the type of tumor (bone or soft
tissue), or the use of low-molecular-weight heparin, with the numbers
available. However, patients with venous thromboembolism had significantly
more additional risk factors than did those without (3.2 compared with 2.5; p
= 0.012).
The majority of thromboembolic events occurred before definitive surgery.
Two patients presented with deep venous thrombosis on referral to the unit;
neither had undergone biopsy or treatment before referral, although one was
noted to have had reduced mobility for two weeks prior to outpatient
consultation. Five of ten deep venous thromboses and two of three pulmonary
emboli occurred at a mean of twenty-seven days (range, four to sixty days)
after open biopsy but before definitive surgery. Three patients had deep
venous thrombosis at a mean of twenty-one days (range, eleven to thirty days)
after definitive surgery; one of these patients had been discharged from the
hospital at the time of diagnosis. One patient had a pulmonary embolism
following chemotherapy but before definitive surgery.
Patients with venous thromboembolism had significantly more additional risk
factors than did those without venous thromboembolism (mean, 3.2 compared with
2.4; p = 0.005) (Tables IV and
V). There was evidence of
appropriate assessment of risk factors for venous thromboembolism at the time
of admission and during the hospital stay as patient circumstances changed.
Two of the thirteen patients with venous thromboembolism had metastatic
disease at the time of presentation (Cases 6 and 12)
(Table III). Two of the
thirteen patients had a thromboembolic event while receiving chemotherapy
(Cases 6 and 8) (Table III),
and one patient had a pulmonary embolism following chemotherapy. None of the
patients with thromboembolic events had had radiation therapy at the time the
venous thromboembolism occurred.
Overall, 172 (69%) of 249 patients in the cohort received some form of
chemical or mechanical thromboprophylaxis after definitive surgery;
thirty-three patients received both. Of these 249 patients, 141 (57%) received
low-molecular-weight heparin. Mechanical prophylaxis was in the form of
thromboembolic deterrent stockings for twenty-seven patients and pneumatic
compression boots (A-V Impulse System; Novamedix, Andover, England) for
twenty-four patients. In five cases, the patient did not receive
thromboprophylaxis as prescribed in the operative note. Low-molecular-weight
heparin was given for a mean of thirty-nine days (range, six to 183 days)
after definitive surgery for patients with venous thromboembolism, compared
with thirteen days (range, one to ninety-two days) for those without venous
thromboembolism. However, in the majority of cases, low-molecular-weight
heparin was given as treatment of a venous thromboembolism rather than as
prophylaxis. Of the seven patients who had venous thromboembolism following
open biopsy, only one had received low-molecular-weight heparin as
thromboprophylaxis following this procedure. Twenty-two patients were
receiving aspirin for the treatment of coexisting conditions rather than as
thromboprophylaxis.
Discussion
In the present retrospective study, we determined the rate of clinically
evident thromboembolic events in patients with trunk or extremity bone or
soft-tissue sarcomas in our center. Of the 252 patients, ten (4%) had a
confirmed deep venous thrombosis and three (1.2%) had a pulmonary embolism,
one of which was fatal (prevalence, 0.4%). The overall rate of clinically
detected thromboembolic events was similar to that seen in other groups of
patients undergoing orthopaedic procedures such as primary joint
replacement9 and
appears to be no higher than the 14% rate of proximal deep venous thrombosis
in orthopaedic cancer patients receiving mechanical thromboprophylaxis as
reported by Lin et
al.4. However, even
the relatively minor procedure of open biopsy may lead to thromboembolic
events in these patients. All of the venous thromboembolism events that we
identified occurred in patients with tumors that were located in the hip or
thigh region, and the majority occurred after open biopsy but before
definitive surgery.
Our study indicates that patients with large tumors in the hip or thigh
region are particularly at risk for venous thromboembolism. This has been
suggested in the literature
previously10, but
to our knowledge the present study is the first to demonstrate this finding in
a large cohort of patients. Large tumors in the thigh may impede venous return
from the lower extremity mechanically, or they may invade the veins. It seems
unlikely that large tumors have a systemic procoagulant effect as large tumors
in other anatomic locations were not associated with venous
thromboembolism.
The present study demonstrated that venous thromboembolism occurred after
open biopsy but before definitive surgery in seven of thirteen cases, raising
the question that open biopsy itself could be a risk factor. All three of the
patients who had development of deep venous thrombosis after definitive
surgery had undergone an open biopsy, and the patient who had development of a
pulmonary embolism following chemotherapy also had had a previous open biopsy.
Unfortunately, the limitations of the present study mean that we are unable to
comment on the contribution of open biopsy to the development of venous
thromboembolism. As we are not aware of any studies that have investigated
whether open biopsy is a risk factor, we believe that the type and timing of
biopsy should be included in prospective studies investigating this question.
Clearly, however, clinicians should be alert to the possibility that
thromboembolic events can occur after open biopsy, and consideration should be
given to excluding the presence of deep venous thrombosis before undertaking
definitive surgery.
Given that all thromboembolic events occurred in association with tumors in
the hip or thigh, patients with tumors in other anatomical sites are likely to
be at low risk of thromboembolism. Of the 116 patients with tumors in other
anatomical sites, only fifty were given low-molecular-weight heparin.
Nevertheless, there were no thromboembolic events in this group. The risks of
chemical thromboprophylaxis may therefore outweigh the benefits in selected
patients. Further defining this low-risk group would be a valuable aim of a
prospective study.
Overall, 60% of the patients received low-molecularweight heparin at the
time of definitive surgery and for the duration of their hospital stay.
Although we analyzed the use of low-molecular-weight heparin, the timing and
duration of administration were too diverse to allow us to draw conclusions
regarding its effectiveness in this cohort of patients. Low-molecular-weight
heparin is now widely accepted in Europe as a safe and effective
thromboprophylactic agent that is used for cancer patients during
hospitalization11
and has an extended role after discharge from the
hospital7.
Guidelines recommend the use of low-molecular-weight heparin for patients
undergoing major orthopaedic surgery such as joint
replacement12.
Anecdotal evidence suggests that orthopaedic oncology surgeons may be
reluctant to prescribe chemical thromboprophylaxis because of concerns about
bleeding complications. Findings from the RIETE (Registro Informatizado de
Pacientes con Enfermedad Tromboembólica) registry highlighted that
patients with cancer had a higher prevalence of major bleeding complications
than did patients without cancer and that patients with cancer were at a
significantly greater risk for recurrent bleeding complications and fatal
pulmonary
embolism13. Given
that sarcomas are highly implantable, it is certainly a theoretical concern
that wound complications such as bleeding or hematoma formation may lead to
wider contamination of a surgical field with tumor and an increase in the rate
of local recurrence. Balancing these considerations is difficult.
Our study had several limitations. First, the prevalence of venous
thromboembolism may have been underestimated because of the retrospective
design of the study, for two reasons: (1) thromboembolic events may have been
diagnosed and treated elsewhere given that patients were referred from a wide
geographical area, and (2) patients may have had subclinical thromboembolic
events that were not detected. Second, some of the documentation was found to
be poor, particularly that which related to the use of mechanical methods of
thromboprophylaxis and the recording of the body mass index for the patients.
In addition, because of the low numbers of thromboembolic events, we were
unable to identify significant risk factors for venous thromboembolism in this
group of patients.
We believe that a prospective study of thromboembolism in this group of
patients would be valuable for defining risk factors, determining the
contribution of biopsy to venous thromboembolism, and identifying low-risk
patients who do not need thromboprophylaxis. ?