Musculoskeletal infections remain a common problem. Because of better
staging systems, developed surgical techniques, antibiotics, and adjuvant
treatment modalities such as hyperbaric oxygen, the treatment strategy for
chronic osteomyelitis has changed to a great extent over the past twenty
years1.
Chronic osteomyelitis leads to necrosis of bone and soft tissues to a
variable extent. The dead bone forms a nidus for hosting pathogens. Moreover,
the host defense mechanisms are often not in an optimal condition to deal with
microorganisms, and antibiotic delivery to the infection site may be impaired
because of poor
circulation2.
Cierny et al. classified chronic osteomyelitis into four anatomical types,
a system that is known as the Cierny-Mader classification, and further staged
the pathology according to the extent of the local and systemic compromise in
the patient3. They
developed guidelines for management according to the system.
Appropriate radical débridement requires excision of all necrotic
bone and soft tissue, often resulting in limb instability. The unstable
situation requires some type of fixation and reconstruction of the resultant
bone and soft-tissue defects. Since the introduction of distraction
osteogenesis by Ilizarov, the technique has been employed successfully to
achieve union, correct deformity, reestablish limb-length equality, and
reconstruct segmental
defects4.
The time spent in an external fixator (the external fixation index) depends
on the length of distraction required and is not free of complications. When
the distraction phase is over, the consolidation phase, which is more than
double the distraction time, becomes difficult for the patient to tolerate.
Removal of the external fixator before satisfactory consolidation has occurred
is associated with fracture, deformity, and shortening occurring through the
distracted
callus5.
To decrease the time that the patient must remain in the frame, distraction
over an intramedullary nail has been investigated in both animal and clinical
studies6,7.
The main purpose of this combined technique is reduction of the external
fixation period to increase the comfort and activity level of the patient. In
this technique, as soon as distraction is finished, the nail is statically
locked and the external fixator is removed. The consolidation phase is then
allowed to proceed, with stabilization provided by the intramedullary nail.
Thus, complications associated with premature removal of the fixator can be
avoided.
This retrospective cohort study summarizes our experience with distraction
osteogenesis with use of an external fixator combined with an intramedullary
nail for the treatment of segmental defects and limb-shortening resulting from
radical débridement of a focus of chronic osteomyelitis.
The cases of a consecutive series of thirteen patients with Cierny-Mader
type-IVA and type-IVB osteomyelitis were reviewed retrospectively
(Table I). All laboratory
parameters to measure active infection such as C-reactive protein levels, the
erythrocyte sedimentation rate, and the white blood-cell count and
differential obtained at the time of nail insertion were normal. Demographic
data were collected after reviewing the medical record and the registry
maintained in our department of the patients undergoing distraction
osteogenesis.
The patients had had an average of 4.5 surgical procedures (range, one to
twenty procedures) before presenting to our clinic. All patients were assessed
for local skin condition, shortening, deformity, distal neurovascular status,
joint function, and nutritional index. Angiography was performed to identify
any vascular injury from previous interventions. Magnetic resonance imaging
was used to examine the entire long bone and identify any distant or skipped
infection and any foci of dead bone. The pathology was classified according to
the Cierny-Mader
system3.
There were seven tibiae and six femora. In the tibial group, two patients
had only a segmental defect, two patients had deformity associated with a
segmental defect, one patient had an infection around a total knee prosthesis
after tumor excision, and two patients had a combination of a segmental defect
and shortening. In the femoral group, only one patient had shortening and one
had a segmental defect but no shortening. The remaining four patients had a
segmental defect combined with deformity.
Operative Technique
Step I
Hardware removal and radical resection of dead bone with débridement
of the infected scarred soft tissue was performed, and representative tissue
cultures, including the sinus tract for all dead bone, were obtained. Cortical
bleeding, described as the so-called paprika sign, was accepted as an
indication of vital
tissue8. The dead
space was filled with custom-made antibiotic-impregnated
polymethylmethacrylate beads (a combination of 2.4 g of teicoplanin and 40 g
of polymethylmethacrylate powder). Patients who had an intramedullary implant
were managed with implant removal and insertion of an antibiotic-impregnated
polymethylmethacrylate cement rod in place of the nail and then with
immobilization of the limb in a custom-made
brace9. In the
remaining patients, stabilization was achieved with a temporary external
fixator. Small soft-tissue defects, resulting from débridement of
infected soft tissues and fistulae, ranging from 2 to 3 cm in size, were
closed during acute shortening in three patients (Cases 2, 10, and
11)10. Antibiotics
that were appropriate according to the findings on culture and sensitivity
were given for a minimum of six weeks or until the erythrocyte sedimentation
rate and C-reactive protein level returned to normal limits
(Table II).
Step II (Intramedullary Nail Insertion, Application of External
Fixator, and Osteotomy)
After a period of six weeks, or when the C-reactive protein level and
erythrocyte sedimentation rate had returned to normal values, the patient
underwent removal of the antibiotic beads or cement rod. A biopsy specimen
obtained from the bone gap was sent for Gram-staining and frozen-section
analysis. The absence of microorganisms on Gram-staining and <5
polymorphonuclear leucocytes per high-power field were taken as an indication
of the resolution of infection. Antegrade nailing was used only for patients
with a segmental defect but without a limb-length discrepancy. Retrograde
nailing was used for the treatment of shortening combined with a segmental
defect. With retrograde nailing, the nails were locked distally and the excess
length of the nail was left in the soft tissues proximally. With distraction,
the nail glided distally until the correct length was achieved and the nail
was locked at the completion of lengthening. For the patients undergoing
segmental transport to treat a bone defect without a length discrepancy,
antegrade nailing was performed. Additional holes were predrilled at the
planned site of locking of the segment at the completion of bone transport to
prevent recoil of the segment.
Treatment of Femoral Defects: The patient was placed supine on a
radiolucent table with the limbs in a scissors position with a bolster below
the pelvis on the involved side. By means of a standard approach (through the
piriformis fossa for antegrade nailing and through a parapatellar incision for
retrograde nailing), the medullary canal was reamed over a guidewire to a
diameter 1.5 mm larger than that of the intramedullary nail to be used. With
lengthening procedures, the goal was to provide sufficient nail length on both
sides of the regenerated bone at the end of distraction. This necessitated the
use of an intramedullary nail that was longer than the length of the femur,
and retrograde nailing allowed the excess nail length to protrude into the
buttock until distraction was completed, by which time the nail would have
glided gradually to its correct position. The proximal part of the femur was
overreamed because the proximal part of the nail was larger. An appropriately
placed corticotomy was then done percutaneously with an osteotome. Finally, an
intramedullary nail (TriGen; Smith and Nephew, Memphis, Tennessee) of
appropriate size was inserted and locked proximally, distally, or on both
sides according to the planned distraction.
Two to three Schanz screws were inserted proximal to the level of the
osteotomy site, and two to three Schanz screws were then inserted distal to it
without contacting an intramedullary
nail11. At least 1
mm of free space should exist between the Schanz screws and the intramedullary
nail to prevent medullary infection triggered by a pin-site
infection12. This
can be achieved with use of the cannulated drill-bit technique described by
Paley et al.13.
Treatment of Tibial Defects: After the medullary canal was reamed
1.5 mm larger than the planned size of nail, the nail was inserted and a
three-ring circular external fixator was applied. A corticotomy was done at
the appropriate level. For the patient with shortening and a segmental defect,
an intramedullary nail of the eventual desired length of the tibia was
inserted and left proud proximally so that it could slide distally during
treatment.
Postoperative Care
Distraction was started on the seventh postoperative
day4. The rate of
the distraction was 1 mm per day, divided into four equal increments. An
epidural catheter was placed for postoperative pain management, and
range-of-motion exercises of the hip and knee were initiated as soon as the
patient's comfort allowed. In patients with a longer tibial intramedullary
nail, knee exercises were postponed until the nail came to lie inside the bone
during lengthening. Full weight-bearing with two crutches was started as soon
as possible.
Step III (Removal of the External Fixator and Static Locking of the
Nail)
After the distraction was completed, the nails were statically locked and
the external fixators removed. Autogenous cancellous bone graft was added at
the docking site. In two
patients8,14,
a nonvascularized fibular graft was inserted into the posteromedial
distraction site to provide additional support and decrease the amount of
forces transmitted through the nail until total consolidation. In both
patients, the distraction site was in the proximal third of the femur.
A functional assessment was done with use of the criteria of Paley et
al.15. The
functional results were based on five criteria: substantial limp, equinus
rigidity of the ankle, soft-tissue dystrophy (skin hypersensitivity,
insensitivity of the sole, or decubitus), pain, and inactivity (unemployment
because of the leg injury or an inability to return to daily activities
because of the leg injury). The result was considered excellent when the
patient was active and had none of the other four criteria, good when the
patient was active and had one or two of the other four criteria, and fair
when the patient was active and had three or four of the other criteria or had
had an amputation.
The bone was assessed for union, infection, deformity, limb-length
discrepancy, and mechanical insufficiencies at the docking site. The result
was considered excellent when the following criteria were met: union, no
infection, a deformity of <7°, and a length discrepancy of <2.5 cm
in the tibia and femur. The result was considered good when there was union
and any two of the other criteria, fair when there was union and one of the
other criteria, and poor when there was non-union or refracture or none of the
other criteria.
In the outpatient clinic, patients were screened for local signs and
symptoms of infection and sinus formation or drainage, and the erythrocyte
sedimentation rate and C-reactive protein levels were monitorized serially.
Conventional radiographs were made every two weeks during the distraction
phase and once every month during the consolidation phase.
The external fixation index was calculated as the duration of external
fixation in days divided by the total amount of bone transported and/or the
amount of lengthening in centimeters. The radiographic consolidation index was
calculated as the time to the appearance of consolidation of at least three
cortices on the anteroposterior and lateral radiographs in days divided by the
total amount of bone transported and/or the amount of lengthening in
centimeters.
Thirteen patients treated with the combined technique were followed for a
mean of 47.3 months (range, thirty-six to fifty-nine months). The laboratory
studies are reported in Table
II. The mean hospital stay was nine days (range, five to fourteen
days). The mean total duration of treatment was nine months (range, five to
sixteen months). The average external fixation index was 13.5 days/cm (range,
12.0 to 15.0 days/cm). The mean radiographic consolidation index was 31.7
days/cm (range, 20.0 to 53.0 days/cm).
There were two failures defined by recurrent drainage and elevated
erythrocyte sedimentation rates and C-reactive protein levels. Eleven patients
were apparently disease-free and able to walk without support at the time of
the final follow-up.
All thirteen patients were evaluated with respect to the functional and
bone results. With use of the criteria described above, the results in terms
of bone-healing and those for function were excellent for eleven patients and
good for two patients. Complete union was achieved in all patients. Refracture
or malalignment was not observed after removal of the frame. The demographic
data on our patients are shown in Table
I (Figs. 1-A and
1-B, 1-C and 1-D,
1-E and 1-F).
Complications
With use of the classification system of
Paley5, minor
complications were the problems that did not require additional surgery and
major complications were identified as either obstacles that resolved with
additional surgery or true complications or sequelae that remained unresolved
at the end of the treatment period (Table
III).
Pain was the most common complaint during the distraction period, observed
particularly in patients requiring lengthening in excess of 4 cm. It was
relieved consistently by oral analgesics. The most frequent complication in
this study was pin-track infection, which was seen in six patients. Six
grade-1 infections, according to the system of
Paley5, were treated
by local care with use of Betadine (povidone-iodine) solution and oral
antibiotics (750 mg of ciprofloxacin twice daily), with resolution at all pin
sites. We did not perform any additional surgery for delayed maturation at the
distraction site or to treat the two transient knee flexion contractures. The
contractures resolved with intensive physiotherapy.
Two patients had failures in the form of recurrent drainage, and both
involved the tibia. Both patients underwent repeat débridement, removal
of the intramedullary nail, and application of an Ilizarov fixator and
administration of antibiotics for six weeks, with parenteral administration
during the initial three weeks and oral administration during the last three
weeks. Soft-tissue defects resulting from repeat débridement were
reconstructed with local gastrocnemius muscle flaps and split-thickness skin
grafts. Subsequently, in both patients, the infection was eradicated and the
nonunion healed, resulting in good function and a good radiographic outcome.
These two patients were fully able to walk without support at the time of the
most recent follow-up.
Infected nonunion of a long bone remains a therapeutic challenge. Such
patients usually have had numerous previous surgical interventions, resulting
in bone defects and soft-tissue compromise. Beginning with the study by
Papineau et al., many treatment modalities have been
described14-24.
All such studies have described a common problem of delayed bone-healing.
Several factors have been blamed for this delay, with the most important being
unsuccessful eradication of
infection17,18,23,24,
and infection has been reported to be the main cause of delayed union or
nonunion15,17,18,20,23,25-28.
Thus, complete cure of the infection is the mainstay of treatment in infected
nonunions.
Today, as a result of changing concepts and advanced reconstruction
techniques, chronic osteomyelitis can be cured. The concept of burning an
infection in the fire of an Ilizarov device, as described by
Ilizarov29, has
changed to the current philosophy that the only cure for osteomyelitis is
radical debridement until live and bleeding bone is reached, as described by
Cierny et
al.1,3,30.
All patients in the current series were managed according to the principles of
Cierny et al.
The extent of débridement necessary to obtain live and uninfected
bone usually results in a bone defect, which requires complex reconstruction.
This challenge can often be addressed by distraction osteogenesis.
The disadvantage of this treatment modality, however, is the long-lasting
consolidation period, which takes about two times as long as the treatment
period, causing great patient discomfort. Our combined treatment with an
intramedullary nail and an external fixator shortens the time needed for
external fixation. Bone-segment transfer over an intramedullary nail with use
of an external fixator is a well-established
technique6,7;
the use of the method in the treatment of segmental defects resulting from
débridement of osteomyelitis is a new concept. The main purpose of the
combined technique is to reduce the external fixation period, thus increasing
the comfort and activity level of the patient.
This established technique is seldom used in the treatment of osteomyelitis
because of the risk of the infection extending along the length of the
involved bone. We believe that this danger is overcome with the use of the
radical débridement technique and removal of all dead and infected
tissue as described by Cierny et
al.1,3,30.
Placement of antibiotic-loaded, custom-made bone cement beads and/or rods
provides sustained local antibiotic delivery and fills the dead space.
Intraoperative tissue cultures identify the appropriate antibiotic therapy to
be continued until the second stage is undertaken. Currently, we mix
teicoplanin powder with the bone cement. Teicoplanin is effective against
methicillin-resistant Staphylococcus aureus both locally and
systemically31. It
can also be administered as a single-dose intramuscular injection after the
patient has been discharged.
Total resolution of the infection is demonstrated by a return of the
C-reactive protein level and erythrocyte sedimentation rate to values to
within normal limits, usually within six weeks. In the current series, all
patients but two demonstrated a normal C-reactive protein level and
erythrocyte sedimentation rate by six weeks after the radical
débridement. Eradication of infection was also demonstrated by frozen
sections and cultures of tissue obtained from the bone ends during the
reconstructive procedure. This same infection treatment protocol and similarly
successful results have been reported for the treatment of infections around
total knee
replacements25-28,32-34.
Beside addressing the chronic infection, the radical débridement
should disturb the bone circulation as little as possible so that successful
results can be obtained with the secondary reconstruction. Periosteal
circulation should be maximally protected during débridement. Reaming
and intramedullary nail placement compromise medullary blood flow and,
following the second stage, bone-healing initially relies on periosteal
new-bone
formation6,35-37.
A review of the medical literature yielded a small number of studies
describing the combined technique for the treatment of chronic osteomyelitis.
Raschke et al. reported on the results of the treatment in four patients with
open tibial
fractures6. They
used an intramedullary nail and the monorail external fixator and had a mean
external fixation index of 17.9 days/cm and a mean consolidation index of 41.2
days/cm. Hoffmann et al. reported on the results of treatment in thirty
patients with infected tibial
nonunions37.
Fifteen patients had reconstruction with the combination of an intramedullary
nail and a special wire system, and the remaining patients were managed with
an intramedullary nail and an external fixator. The mean external fixation
index was 12.2 days/cm for the former group and 13.7 days/cm for the latter
group. That study did not report any data regarding consolidation indices.
The current retrospective study consisting of thirteen patients yielded a
mean external fixation index of 13.5 days/cm and a mean consolidation index of
31.7 days/cm. The indices were similar in the femur and the tibia. These
results are comparable with those in the previous studies mentioned above. In
addition, the excellent results in terms of the bone and function scores for
eleven of our thirteen patients demonstrate the success of the combined
method.
The main purpose of the current study was to investigate a method to obtain
eradication of infection and to demonstrate that this method can shorten the
external fixation index and decrease the complication rate. Studies on the
classic segment transport with the Ilizarov device in the treatment of
infected nonunions have described a mean external fixation index of 54.9
days/cm, which is much longer than the mean external fixation index in the
present study of 13.5
days/cm17-20,24.
The same studies noted a mean of two complications per patient, which is twice
the complication rate in our study. We believe that the lower complication
rate in our study resulted from both the shorter external fixation index and
our meticulous application technique, in which careful attention was paid to
radical débridement of all dead tissue and to the insertion of Schanz
screws with a clear space left between the screws and the intramedullary
nail.
Another valuable reconstruction method for the treatment of an infected
nonunion in a long bone is the use of vascularized bone
grafts38,39.
One study has noted a mean time to bone-healing of approximately seven months
in the femur and six months in the
tibia38. The total
number of major complications, such as recurrence of infection and graft
failure, exceeded the number of complications reported in the present study.
Besides, our combined technique provides two very important, superior features
compared with vascularized bone grafts. First, there is no donor-site
morbidity, and the vascularized fibula can be kept as a last alternative.
Second, with the combined technique, bone segment transfer and lengthening can
be achieved with the same procedure, but the vascularized bone graft can only
bridge the defect and cannot overcome a length discrepancy.
The weaknesses of our study are the absence of a control group and our
small number of patients. Nevertheless, we describe a new and successful
alternative technique for the treatment of the challenging problem of chronic
osteomyelitis. ?