Computer navigation for total knee arthroplasty has become increasingly
popular because of its potential to improve the accuracy of placement of the
femoral and tibial
components1-4.
However, complications related to the navigation system have only rarely been
reported in the orthopaedic
literature3,5,6.
Ossendorf et al. reported on a femoral stress fracture that occurred after
computer-navigated total knee
arthroplasty7. This
nondisplaced stress fracture occurred at the pin-hole site for the navigation
trackers.
We present two cases of stress fracture that were associated with a
transcortical pin track that had been drilled to allow placement of navigation
trackers. These two cases provide information about the risk of the
development of a stress fracture after placement of a transcortical pin.
Hopefully the risk of a stress fracture can be reduced by making surgeons
aware of this complication. The patients were informed that data concerning
their cases would be submitted for publication, and they consented.
Case 1. A seventy-year-old woman presented to our institution
with painful osteoarthritis of both knees. The condition had begun six years
earlier. She subsequently underwent two computer-navigated total knee
arthroplasties: the right knee was replaced first, with use of a
low-contact-stress rotating-platform prosthesis (New Jersey LCS; DePuy,
Warsaw, Indiana); and the left knee was replaced one week later, with
insertion of a press-fit condylar rotating-platform prosthesis (PFC Sigma RPF;
DePuy). The operations were performed by one team (Y.-B.J. and H.-J.J.), the
members of which had together performed several hundred conventional total
knee arthroplasties and more than eighty computer-assisted procedures before
this operation. The Kolibri computed tomography-free knee navigation system
(BrainLAB, Munich, Germany) and Ci software (DePuy) were used.
With the patient under general anesthesia and with the use of tourniquet
control, a modified subvastus approach to the left knee was used. Navigation
trackers were fixed with non-threaded bicortical 2.8-mm Steinmann pins, two
each in the proximal part of the tibia and distal part of the femur. All of
the components were cemented, and a 12.5-mm-thick liner was used in both knees
(Figs. 1-A and 1-B). The
outcome during the first four weeks was excellent; the patient could walk
without restriction. A dull pain aggravated by weight-bearing activity
gradually developed in the distal portion of the left thigh during the fifth
postoperative week. Seven days after the onset of this pain, the left femur
fractured while the patient was standing during minimal weight-bearing.
Radiographs revealed a transverse femoral fracture at the junction of the
distal shaft and metaphysis (Fig.
1-C). The fracture line was scalloped through the drilled holes.
The fracture was treated with open reduction and internal fixation with a
locking compression plate (LCP; Synthes, Paoli, Pennsylvania) and cancellous
autograft. A bone biopsy, obtained to rule out a pathologic fracture, revealed
no worrisome findings. The patient had an uneventful recovery and achieved a
Knee Society score8
of 92 after six months.
Case 2. A sixty-four-year-old woman presented to our institution
with painful osteoarthritis in both knees. The condition had been present for
three years. She underwent a computer-navigated total knee arthroplasty with a
low-contact-stress rotating-platform prosthesis (New Jersey LCS; DePuy) in
both knees. The navigation system used was the same as that used for the
patient in Case 1. The right knee was replaced first, and the left knee was
replaced one week later. A modified subvastus approach was used, and all of
the components were cemented. A 12.5-mm-thick rotating platform tibial liner
was used in both knees (Fig.
2-A). For the first six weeks after surgery, the patient tolerated
unlimited walking but experienced vague pain and swelling at the site of the
left proximal tibial pin track after walking. The pain and swelling worsened.
Local heat or redness was not observed, and inflammatory serological markers
(erythrocyte sedimentation rate, c-reactive protein level, and white
blood-cell count) were within normal limits. Follow-up radiographs of the left
knee during the eighth postoperative week revealed only sclerosis at the
proximal tibial pin holes with no clear evidence of fracture
(Figs. 2-B and 2-C); however, a
technetium-99 bone scan showed an increased uptake at the site of the left
proximal tibial pin hole, suggesting a stress fracture
(Fig. 2-D). The patient was
instructed to avoid weight-bearing activity for four weeks. The pain
subsequently subsided and she achieved a Knee Society score of 95 at the sixth
postoperative month.
Fracture of the tibia after total knee arthroplasty is uncommon, with a
reported rate of less than
1%9. Stress
fractures of the tibia after total knee arthroplasty have been reported in
association with axial malalignment, improper component position, tibial
tuberosity osteotomy, component loosening, and resection-gu ide pinholes with
local osteoporosis and
osteolysis9-13.
Stress fractures of the femur after total knee arthroplasty have also been
reported. Supracondylar fractures of the femur have been attributed to
so-called notching of the anterior cortex of the femur in association with
rheumatoid arthritis, osteoporosis, and preexisting neurologic
disorders14-16.
In both of our patients, all of the components were cemented and there was
no evidence of malalignment or component malposition on radiographs made
immediately after surgery. A bone biopsy performed in the patient in Case 1
revealed no pathologic findings. Fractures were documented at more than six
weeks after total knee arthroplasty in both patients after a good initial
recovery. We therefore believe that these were cases of fatigue fracture,
probably related to the navigation instrumentation.
Complications caused by a navigation system, such as pin or drill breakage,
have rarely been
reported3,5,6.
Recently, Ossendorf et al. suggested that repeated drilling was the cause of
fracture in a patient who had a stress fracture following a computer-navigated
total knee arthroplasty, and the authors recommended unicortical pin fixation
to avoid
fracture7.
Fractures through drill holes have been documented, as the drilled holes
substantially decrease the breaking strength of the surrounding
bone17.
Furthermore, fresh screw holes in bone have been reported to act as stress
risers and to substantially weaken the bone to the effects of torsional
stresses and bending loads in a canine
model18,19.
Hipp et al. and McBroom et al. suggested that the shape and length of the
defect strongly correlate with the reduction in torsional strength of a long
bone20-22.
Brooks et al. studied torsional failures in canine femora in association
with stress concentrations. They found that the energy-absorbing capacity of
the bone was reduced by 55% if there was a drill hole, regardless of the hole
size18.
Close scrutiny of the immediate postoperative radiographs of our two
patients revealed that the medial cortex of the distal part of the femur (Case
1, Figs. 1-A and 1-B) and
proximal part of the tibia (Case 2, Fig.
2-A) was violated by one pin. A pin drilled into tubular bone can
be unicortical, bicortical, or transcortical
(Fig. 3). On plain radiographs,
a transcortical pin track has a different appearance from a bicortical pin
track; while the former looks like a tunnel, the later appears as two dots.
This finding was also reported in the case described by Ossendorf et
al.7. In both of our
patients, stress fractures occurred through a transcortical pin track. We
believe that the fractures resulted from a misplaced fixation pin that did not
pass through two cortices but passed close to the cortex or through the
cortex, creating a stress riser. Although pin holes may heal without problems,
delayed healing may be associated with transcortical holes leading to a
fatigue fracture through this region.
Transcortical drilling could cause not only a longer cortical bone defect
but also more thermal necrosis of bone. Necrotic bone remodels slowly, which
lengthens the time that a major stress concentration is present. Therefore, we
believe that transcortical drilling is an important risk factor for stress
fractures after computer-navigated total knee arthroplasty.
Ossendorf et al. suggested the use of unicortical pin fixation to avoid a
stress fracture7.
However, Kuo et al. postulated that the stress concentration factor for a
single-cortex defect is similar to that of a double-cortex defect of identical
dimensions23 and
that unicortical fixation could result in cutting errors caused by navigation
tracker motion in elderly patients with poor bone
quality3.
Currently, we use bicortical rather than unicortical fixation in patients
with poor bone quality, and we try to avoid transcortical pin placement.
Surgeons should be aware of the potential for stress fracture if navigation
pins are placed transcortically, and patients should be monitored accordingly.
Any reported pain in the thigh or tibia after the initial healing period of
three to four weeks following computer-assisted total knee replacement should
be considered a potential stress fracture. We believe that the risk of a
stress fracture due to a transcortical pin track can be reduced by the
surgeon's awareness of this complication.?