Abstract
Background: The association of deep venous thrombosis and deep
musculoskeletal infection in children has been reported infrequently. The
purpose of the present study was to evaluate the characteristics of children
with osteomyelitis in whom deep venous thrombosis developed and to compare
them with those of children with osteomyelitis in whom deep venous thrombosis
did not develop.
Methods: A retrospective review of the records of children who were
managed at our institution because of a deep musculoskeletal infection between
January 2002 and December 2004 identified 212 children with osteomyelitis
involving the spine, pelvis, or extremities. Children in whom deep venous
thrombosis developed were compared with those in whom it did not develop with
respect to age, diagnosis, causative organism, duration of symptoms prior to
admission, laboratory values at the time of admission, surgical procedures,
and required length of hospitalization.
Results: Eleven children with osteomyelitis and deep venous
thrombosis were identified. The mean C-reactive protein level was 16.9 mg/dL
for the group of eleven patients with osteomyelitis in whom deep venous
thrombosis developed, compared with only 6.8 mg/dL for the group of 201
patients with osteomyelitis in whom deep venous thrombosis did not develop (p
= 0.0044). Staphylococcus aureus was the causative organism of
infection in all eleven children with deep venous thrombosis and in
ninety-three (46%) of the 201 children without deep venous thrombosis.
Methicillin-resistant strains of Staphylococcus aureus were
identified in eight of the eleven children with deep venous thrombosis and in
only forty-nine of the 201 children without deep venous thrombosis. The
children with osteomyelitis and deep venous thrombosis were older, had a
longer duration of hospitalization, had more admissions to the intensive care
unit, and required more surgical procedures than those with osteomyelitis but
without deep venous thrombosis.
Conclusions: Deep venous thrombosis in association with
musculoskeletal infection is more common in children over the age of eight
years who have osteomyelitis caused by methicillin-resistant
Staphylococcus aureus and who present with a C-reactive protein level
of >6 mg/dL. Diagnostic venous imaging studies should be performed to
assess for the presence of deep venous thrombosis in children with
osteomyelitis, especially those who have these risk factors.
Level of Evidence: Therapeutic Level III. See
Instructions to Authors for a complete description of levels of evidence.
Deep venous thrombosis and pulmonary embolism are uncommon in children,
with an estimated prevalence of <0.01%. The use of intravenous catheters,
extremity trauma, postoperative complications, and inherited coagulation
disorders account for most of these
events1.
Recent studies have demonstrated an association between deep
musculoskeletal infection in children and the development of deep venous
thrombosis and septic pulmonary embolism as part of a life-threatening
clinical syndrome of disseminated staphylococcal
disease2,3.
The emergence of community-acquired methicillin-resistant Staphylococcus
aureus as a leading infectious organism in children has occurred
concurrently with an increase in the frequency of deep venous thrombosis and
septic pulmonary embolism in children with musculoskeletal infection. As many
as 6% of children with osteomyelitis caused by methicillin-resistant
Staphylococcus aureus have been reported to have development of deep
venous thrombosis2.
Still, there have been few published cases of combined deep musculoskeletal
infection and thrombosis. A recent report from the Texas Children's Hospital
in Houston cited eight cases of osteomyelitis and deep venous thrombosis that
occurred at that institution between August 2001 and December 2004,
representing the largest series to
date2.
There is value in distinguishing the few children who have development of
deep venous thrombosis from the larger number of children who have deep
musculoskeletal infection without this complication. Unfortunately,
hematologic values are often normal in affected children, indicating that a
prothrombic tendency is not essential for the development of thrombosis in the
setting of musculoskeletal
infection4.
Martinez-Aguilar et al. recently reported that the Panton-Valentine leukocidin
(PVL) gene was encoded in the strains of Staphylococcus aureus
isolated from all five children who had development of deep venous thrombosis
in a series of twenty-eight children with musculoskeletal
infection4. While
such preliminary research might suggest that the presence of genetically
encoded virulence factors may be associated with a higher prevalence of deep
venous thrombosis associated with Staphylococcus aureus infections,
no clinically useful tests are currently available to assess for the presence
of these virulence
factors2,4.
There is currently no adequate explanation as to why children with serious
musculoskeletal infection may have increased susceptibility to deep venous
thrombosis. The purpose of the present study was to investigate the prevalence
of deep venous thrombosis among children managed for deep musculoskeletal
infection at a tertiary pediatric medical center in the southwest United
States. On the basis of this population, we sought to characterize the most
likely clinical and laboratory presentation of children who are prone to the
development of this problem.
The medical records of children who were evaluated and managed for an
infection involving the spine, pelvis, or upper or lower extremities in the
emergency room or inpatient hospital of our institution from January 1, 2002
to December 31, 2004 were retrospectively reviewed. Multiple data fields were
recorded, including demographic characteristics, age at the time of admission,
date at the time of the onset of symptoms, primary musculoskeletal infection
diagnosis, location of infection, body temperature, radiographic studies and
results, laboratory studies and results, culture results, dates of admission
and discharge, surgical procedures performed, and complications.
The records of patients who were diagnosed with both deep musculoskeletal
infection and deep venous thrombosis were identified and subjected to
additional analysis. For those patients, numerous additional data fields were
recorded, including the location and method of diagnosis of deep venous
thrombosis, the results of lung imaging and the presence of septic pulmonary
emboli, whether or not the patient had a genetic predisposition to thrombosis,
whether or not the patient had a family history of thrombotic disorders, the
use of anticoagulants, the use of an intravascular filter, the placement of a
central venous catheter, the duration of hospitalization before the diagnosis
of deep venous thrombosis, and the duration of treatment of deep venous
thrombosis.
Children who had osteomyelitis and deep venous thrombosis were compared
with a cohort of children who had osteomyelitis without deep venous thrombosis
with respect to age; gender; the duration of symptoms prior to admission; the
mean body temperature at the time of admission; the mean values for the
erythrocyte sedimentation rate, C-reactive protein level, and white blood-cell
count; the causative organism; the number of surgical procedures; and the
duration of hospitalization. Statistical analyses were performed to compare
patients with and without deep venous thrombosis; independent sample t tests
were used when comparing means, and the Fisher exact test was used when
comparing rates. The level of significance was set at p = 0.05.
Between January 1, 2002 and December 31, 2004, 555 children were evaluated
and treated for deep musculoskeletal infection at the institution;
specifically, 212 patients had osteomyelitis, 117 had septic arthritis,
twenty-one had pyomyositis, and 205 had a deep abscess. Thirteen of these
children were identified as having a deep venous thrombosis. The primary
musculoskeletal infection that was diagnosed in the children with deep venous
thrombosis was osteomyelitis in eleven, septic arthritis in one, and
pyomyositis in one.
The mean age of the eleven children with osteomyelitis and deep venous
thrombosis was 10.9 years (range, three to fourteen years) (see Appendix). In
comparison, the mean age of the children with osteomyelitis who did not have
deep venous thrombosis was 7.4 years (range, 0.06 to 17.9 years) (p = 0.0065).
Symptoms were present for an average of 5.6 days prior to admission in the
group of patients with deep venous thrombosis, compared with 14.4 days in the
group of patients without deep venous thrombosis (p = 0.0007).
In all eleven children with deep venous thrombosis and osteomyelitis, the
infectious organism that was identified was Staphylococcus aureus
(Table I). Eight (72.7%) of
these eleven infections were caused by methicillin-resistant
Staphylococcus aureus, and three (27.3%) were caused by
methicillin-sensitive Staphylococcus aureus. All infections were
considered to be community-acquired. In comparison, Staphylococcus
aureus was the causative organism in ninety-three (46.3%) of the 201
children who had osteomyelitis without development of deep venous thrombosis.
Forty-nine (52.7%) of these ninety-three infections were caused by
methicillin-resistant Staphylococcus aureus, and forty-four (47.3%)
were caused by methicillin-sensitive Staphylococcus aureus. Of the
forty-seven children with osteomyelitis and cultures that were positive for
methicillin-sensitive Staphylococcus aureus, only three (6.4%) had
development of deep venous thrombosis. In comparison, of the fifty-seven
patients with osteomyelitis and cultures that were positive for
methicillin-resistant Staphylococcus aureus, eight (14%) had
development of deep venous thrombosis.
The location of the infection was adjacent to the site of deep venous
thrombosis in nine (82%) of eleven children. The most frequent locations for
osteomyelitis associated with deep venous thrombosis were the distal part of
the femur and the proximal part of the tibia
(Fig. 1).
The location of thrombosis was most frequently the femoral vein (six
children), often with extension into the popliteal vein (see Appendix).
Thrombosis was found in the iliac veins of two children and in the superior
vena cava of two additional children. Four thromboses were identified with
ultrasound. Three thromboses each were identified with magnetic resonance
imaging and computed tomography. One additional thrombosis was identified with
use of an echocardiogram. The majority of deep venous thromboses (72.7%; eight
of eleven) were recognized when imaging was employed to evaluate the deep
musculoskeletal infection (Figs. 2-A,
2-B, 2-C,
2-D). The diagnosis of deep
venous thrombosis was made an average of 3.8 days (range, one to eight days)
after admission.
Pulmonary imaging revealed evidence of septic emboli in six (54.5%) of the
eleven children with osteomyelitis and deep venous thrombosis. One additional
child had bilateral pneumonia and nodular opacities that may have been related
to septic emboli, although they were not formally diagnosed as such.
With regard to risk factors, only two of the nine children who were tested
for genetic coagulation disorders were found to have an abnormality; the
abnormality was heterozygous in both cases. One of these had a Prothrombin
20210A mutation, and the other had a Factor V Leiden mutation. Neither child
had a family history of deep venous thrombosis or other hematologic problems.
One additional child had a second-degree relative who had died from an acute
myocardial infarction at the age of fifty-three years.
A central venous catheter was used for the treatment of all eleven children
with osteomyelitis who were found to have deep venous thrombosis. However, the
catheter was still present in only three of the children at the suspected time
of onset of the thrombosis. In one of these three children, the deep venous
thrombosis occurred adjacent to the area of infection and remote from the
catheter location. In another child, a deep venous thrombosis developed near
the site of a femoral intravenous line, but this child also had osteomyelitis
in proximity to the thrombosis. In the third child, the thrombosis occurred
near the site of the central venous catheter, which was remote to the location
of the musculoskeletal infection.
The values for body temperature, white blood-cell count, erythrocyte
sedimentation rate, and C-reactive protein level for the eleven children with
osteomyelitis and deep venous thrombosis were measured at the time of
admission and were compared with the values for the children with
osteomyelitis who did not have development of this complication
(Table I). At the time of
presentation, children who had development of deep venous thrombosis had a
higher mean body temperature (38.3°C) than did those who did not have
development of deep venous thrombosis (37.2°C) (p = 0.024). The mean white
blood-cell count was 11.1 × 109/L for the patients who had
development of deep venous thrombosis, compared with 11.5 ×
109/L for those who did not have development of deep venous
thrombosis (p = 0.867). The mean erythrocyte sedimentation rate was 62.5 mm/hr
for the patients who had development of deep venous thrombosis, compared with
54.9 mm/hr for those who did not have development of deep venous thrombosis (p
= 0.492). The differences between the patients who did and did not have
development of deep venous thrombosis with regard to the values for the white
blood-cell count and erythrocyte sedimentation rate were not significant.
Finally, the mean C-reactive protein level was 16.9 mg/dL for the patients who
had development of deep venous thrombosis and only 6.8 mg/dL for those who did
not have development of deep venous thrombosis (p = 0.0044).
Seven (63.6%) of the eleven children with deep venous thrombosis and
osteomyelitis were admitted to the intensive care unit, with one child
requiring intubation. Of the 201 children with osteomyelitis and without deep
venous thrombosis, only six (3%) were admitted to the intensive care unit,
with two children requiring intubation.
Nine (81.8%) of the eleven children were managed with low molecular-weight
heparin after identification of the thrombosis. Anticoagulant therapy for two
children was initiated with intravenous unfractionated heparin but ultimately
was converted to low molecular-weight heparin for one child and to warfarin
for the other. One child was not managed with anticoagulants. One child had a
Greenfield intravascular filter (Boston Scientific, Natick, Massachusetts)
placed in the inferior vena cava to diminish pulmonary showering by septic
emboli. Bacterial endocarditis developed in one child with disseminated
infection. This child subsequently underwent tricuspid valvuloplasty because
of damage due to the bacterial endocarditis. The antibiotic treatment for
musculoskeletal infection was not altered by the finding of deep venous
thrombosis in the subgroup of children in this study.
Surgical intervention for the treatment of the underlying osteomyelitis was
necessary a total of twenty-nine times for the eleven children with deep
venous thrombosis (mean, 2.6 procedures per child), compared with 181 times
for the 201 children without deep venous thrombosis (mean, 0.9 procedure per
child) (p < 0.0001) (Table
I). The mean duration of hospitalization was 30.6 days for
children who had osteomyelitis with deep venous thrombosis, compared with 9.5
days for children who had osteomyelitis without deep venous thrombosis (p =
0.0004).
Documentation of the resolution of deep venous thrombosis was noted in the
outpatient medical records of eight (72.3%) of the eleven children with
osteomyelitis and deep venous thrombosis, at an average of eleven weeks
(range, three to eighteen weeks). The documentation was based on follow-up
imaging studies for only three children. The remaining documentation was based
on a clinical assessment of the adequacy of the duration of anticoagulation.
There was no documented follow-up for three patients. The resolution of septic
pulmonary emboli was documented on the basis of follow-up chest radiographs,
made during the hospitalization, that were interpreted as demonstrating
improvement in comparison with the original chest radiographs.
Deep venous thrombosis is rare in children and is most often related to the
use of intravenous catheters, extremity trauma, postoperative complications,
and inherited thrombotic risk
factors1. Several
reports have also associated thrombosis and septic pulmonary emboli with deep
musculoskeletal infection, including osteomyelitis and septic
arthritis2. Despite
these reports, the relationship between musculoskeletal infection and the
prevalence of deep venous thrombosis has been limited, making it difficult to
characterize the children in whom this complication is most likely to develop.
In the present study, thirteen cases of deep venous thrombosis were found
among 555 children with deep musculoskeletal infection who were managed at
Children's Medical Center Dallas between January 1, 2002 and December 31,
2004. Eleven of the thirteen cases were noted to occur in children with
osteomyelitis. The comparison of this subgroup of children with the larger
cohort of 201 children with osteomyelitis but without deep venous thrombosis
revealed several characteristics that may be helpful for directing closer
attention to those children who may be prone to this complication.
In general, children in whom deep venous thrombosis developed were older
than the other children who had a deep musculoskeletal infection (10.9
compared with 7.4 years). The locations of the infections in the patients with
deep venous thrombosis exclusively involved the spine, pelvis, and lower
extremities, and most of the infections occurred adjacent to the major
vascular structures of the lower extremities. Children with deep venous
thrombosis presented with higher mean C-reactive protein levels than did those
who had osteomyelitis without deep venous thrombosis (16.9 compared with 6.8
mg/dL). These findings, along with the higher mean temperature at the time of
admission (38.3°C compared with 37.2°C), the increased number of
surgical procedures per child (2.6 compared with 0.9), the increased
requirement for admission to the intensive care unit (seven of eleven patients
compared with six of 201 patients), and the longer duration of hospitalization
(30.6 compared with 9.5 days) support the conclusion that children with deep
venous thrombosis have a more advanced degree of infection and clinical
illness at the time of admission in comparison with children without deep
venous thrombosis.
Laboratory cultures were positive for Staphylococcus aureus in all
eleven children who had deep venous thrombosis associated with osteomyelitis.
To our knowledge, only twenty-three cases of Staphylococcus aureus
osteomyelitis and concurrent thrombosis have been reported to
date2,3.
The strain of infectious organism may be associated with the development of
deep venous thrombosis. A recent report from Texas Children's Hospital
documented that seven (6.0%) of 116 children who were managed for acute
hematogenous osteomyelitis caused by methicillin-resistant Staphylococcus
aureus had development of deep venous
thrombosis3. The
present study demonstrated an even higher rate of deep venous thrombosis
associated with methicillin-resistant Staphylococcus aureus
osteomyelitis (14%), as compared with a rate of 6% associated with
methicillin-sensitive Staphylococcus aureus osteomyelitis.
There is some evidence suggesting that the presence of the PVL gene encoded
in strains of methicillin-resistant Staphylococcus aureus and
methicillin-sensitive Staphylococcus aureus may explain the
occurrence of complications such as deep venous thrombosis associated with
deep musculoskeletal
infection2. Of the
eight patients with osteomyelitis and deep venous thrombosis at Texas
Children's Hospital as reported in a previous study, seven exhibited the PVL
gene in Staphylococcus aureus strain
isolates3. A
relationship between PVL-positive strains and other complications such as
chronic osteomyelitis and prolonged hospitalization has also been
noted2. In our
retrospective review, the genetic makeup of cultured organisms was not
determined. We hope to analyze this aspect of Staphylococcus aureus
infection prospectively at our institution in the future.
Inherited prothrombic disorders may have been involved in the development
of deep venous thrombosis in two of the patients in our study. However, there
are not enough data to draw a useful conclusion regarding clotting propensity
given the small number of children in our series.
Given the significantly greater number of surgical procedures per child in
the deep venous thrombosis cohort, it is reasonable to consider whether the
thromboses may have been a consequence of surgery, immobility, and prolonged
recumbency. However, it should be noted that the deep venous thromboses were
recognized an average of 3.8 days after admission, relatively early in the
course of hospitalization and treatment. Furthermore, the length of time
between the onset of symptoms and the date of admission was significantly
shorter for children with deep venous thrombosis than for those without deep
venous thrombosis (5.6 compared with 14.4 days). Given the almost negligible
rate of deep venous thrombosis in our institution among children who have
undergone surgical procedures following injury, it appears that the rate of
deep venous thrombosis among children with infection is not likely related to
the surgical procedures or prolonged bedrest. Rather, the higher rate of
surgical procedures in these children most likely reflects the increased
severity of their illness at the time of treatment and the need for more
aggressive surgical débridement to adequately control the
infection.
The findings of the present study are consistent with previous reports of
the association between osteomyelitis and the development of deep venous
thrombosis in pediatric
patients2,3.
A high index of suspicion should be maintained whenever an infection caused by
Staphylococcus aureus involves the spine, pelvis, or lower
extremities, especially in an older child or adolescent who presents with a
moderately elevated C-reactive protein level. For example, among the 212
children with osteomyelitis who were reviewed in the present study, only ten
were more than eight years old at the time of admission, had a mean C-reactive
protein level of >6 mg/dL, and had positive cultures for
methicillin-resistant Staphylococcus aureus. Of these, four were
found to have deep venous thrombosis. Under these circumstances, we believe
that imaging studies such as Doppler venous ultrasound or magnetic resonance
venography may be considered to evaluate the possible presence of deep venous
thrombosis. The increased need for intensive care monitoring and support of
these children is also an important finding of our review.
Methicillin-resistant Staphylococcus aureus may have a unique
propensity to cause deep venous thrombosis in association with musculoskeletal
infection. Additional study is necessary in order to evaluate potential
bacterial and host genetic factors that may be responsible for the thrombotic
tendency in these children. Physicians caring for patients who have
osteomyelitis in areas where community-acquired methicillin-resistant
Staphylococcus aureus strains are common should be aware of this
complication. ?
A table showing the clinical results on all study subjects is available
with the electronic versions of this article, on our web site at
(go to
the article citation and click on "Supplementary Material") and on
our quarterly CD-ROM (call our subscription department, at 781-449-9780, to
order the CD-ROM).
Stein PD, Kayali F, Olson RE. Incidence
of venous thromboembolism in infants and children: data from the National
Hospital Discharge Survey. J Pediatr.2004;145:
563-5.145563
2004
[PubMed][CrossRef]
Gonzalez BE, Teruya J, Mahoney DH Jr,
Hulten KG, Edwards R, Lamberth LB, Hammerman WA, Mason EO Jr, Kaplan SL.
Venous thrombosis associated with staphylococcal osteomyelitis in children.
Pediatrics.2006;117:
1673-9.1171673
2006
[PubMed][CrossRef]
Yuksel H, Ozguven AA, Akil I, Ergruder
I, Yilmaz D, Cabuk M. Septic pulmonary emboli presenting with deep venous
thrombosis secondary to acute osteomyelitis. Pediatr Int.2004;46:
621-3.46621
2004
[PubMed][CrossRef]
Martinez-Aguilar G, Avalos-Mishaan A,
Hulten K, Hammerman W, Mason EO Jr, Kaplan SL. Community-acquired,
methicillin-resistant and methicillin-susceptible Staphylococcus aureus
musculoskeletal infections in children. Pediatr Infect Dis J.2004;23:
701-6.23701
2004
[PubMed][CrossRef]