Posttraumatic or secondary osteoarthritis may develop after
a fracture of the tibial plateau as a result of (1) the alteration
of the osseous anatomy leading to altered knee mechanics and loss
of cartilage and bone, (2) cartilage damage as part of the initial
injury, or (3) the presence of subchondral implants that may perforate
the articular cartilage. Posttraumatic osteoarthritis combined with
a compromised soft-tissue envelope can predispose the patient
to a less satisfactory functional outcome after any type
of surgical intervention. Surgical options for prevention of posttraumatic
arthritis range from arthroscopic débridement to arthrodesis.
One option for the treatment of end-stage posttraumatic
arthritis is total knee arthroplasty.
There is little information in the literature regarding the outcome
of total knee arthroplasty for posttraumatic arthritis after open
reduction and internal fixation of fractures of the tibial plateau1-3. The goal of this study was to
evaluate the results, after a minimum of five years of follow-up,
of total knee arthroplasty in patients who had had previous open
reduction and internal fixation of a fracture of the tibial plateau.
We identified fifteen consecutive patients who had undergone open
reduction and internal fixation of a fracture of the tibial plateau
and who eventually underwent a total knee replacement between 1986
and 1994, with a minimum five-year follow-up since the latter procedure.
The institutional review board approved the study protocol, and
all fifteen patients gave their informed consent. The average age
of the four men and eleven women was fifty-six years (range, thirty-seven
to sixty-eight years) at the time of the arthroplasty. The total knee
replacements (eleven left and four right) were performed at an average
of 38.6 months (range, eight months to eleven years) after the open
reduction and internal fixation.
Fixation of the fracture was accomplished with a medial plate only
in four patients, with a lateral plate only in seven, with both
a medial and a lateral plate in three, and with screw fixation only
in one. Between the time of the open reduction and internal fixation
of the fracture and the time of the total knee arthroplasty, six
patients had undergone nine procedures, including removal of implants
(two), high tibial osteotomy (one), arthroscopy (two), and scar
revision (one). All patients had pain as the primary symptom that
led to the decision to undergo arthroplasty.
Two fractures were not completely healed at the time of the total
knee replacement. Before the arthroplasty, eleven of the fifteen
patients had a flexion contracture, which ranged from 5° to 30°
(average, 11°). Five of the fifteen patients demonstrated medial
and/or lateral instability in the affected knee compared
with the contralateral knee. Before the arthroplasty, the tibiofemoral
alignment was measured on standing anteroposterior radiographs (14 ¥ 17
in [35.6 ¥ 43.2 cm]). Eleven of the fifteen
patients had a valgus deformity ranging from 10° to 25°, one patient
had a neutral alignment of 6°, and three patients had a varus alignment
ranging from 4° to —8°.
Surgical Considerations
All of the knees were aspirated preoperatively, and all cultures of
the aspirate were negative. Two patients had only a skin incision
done preoperatively in preparation for the total knee arthroplasty
to confirm the vascular integrity of the surgical flap. One incision
healed uneventfully, and the other healed slowly and required a
myocutaneous flap procedure, which was done simultaneously with
the total knee arthroplasty. That patient had had an infection after
the open reduction and internal fixation, which was treated with
chronic antibiotic suppression (Augmentin [amoxicillin
and clavulanate potassium]) for 107 months. Preoperative
cultures performed after discontinuation of the antibiotics were
negative, as were intraoperative gram stains and cultures.
A midline incision with a medial arthrotomy, incorporating a portion
of the previous scar whenever possible, was performed in all patients.
Eight patients who had had lateral incisions from the previous operations
required medial adjustment of the new midline incision to increase
the skin-bridge distance. Some incisions had a transverse component (a
hockey-stick-shaped incision), making incorporation difficult.
Lateral retinacular release of the patella was performed in five
patients, and a limited quadriceps release (rectus snip) was performed
in one patient. The internal fixation implant had been removed prior
to the arthroplasty in two patients and was removed at the time
of the arthroplasty in thirteen.
Five of the fifteen patients had an uncontained bone defect of the
lateral tibial plateau requiring autogenous bone-grafting. Metal
wedges were used in two of the fifteen patients, and one of these
patients required additional morselized bone graft. Two of the fifteen
patients had ununited fracture fragments that required open reduction
and internal fixation with screws. If there was ligamentous instability,
a constrained condylar knee prosthesis was used. If osseous deficiency
was such that metal augmentation was necessary, a stemmed component was
used.
A constrained condylar knee prosthesis was used in eight patients,
one of whom received a Total Condylar III implant (Johnson and Johnson,
New Brunswick, New Jersey) and seven of whom received a Constrained
Condylar Knee replacement (Zimmer, Warsaw, Indiana). Three patients
were treated with a posterior stabilized knee prosthesis (Insall-Burstein
II; Zimmer), and two received a posterior-cruciate-sparing knee replacement
(Genesis I; Richards Medical, Memphis, Tennessee). Two custom-designed
constrained condylar knee prostheses (Hospital for Special Surgery,
New York, NY) were necessary: to correct malunion in one patient and
to treat substantial bone loss in the other.
Clinical results were assessed on the basis of Hospital for Special
Surgery knee scores obtained before the arthroplasty and at an average
of 6.2 years (range, 4 to 11.1 years) after it as well as with use
of Short Form-36 (SF-36) scores obtained by telephone at an average
of 7.7 years (range, 5.4 to 13.2 years) after the arthroplasty4,5. Standing anteroposterior and lateral
radiographs on 14 ¥ 17-in (35.6 ¥ 43.2-cm) film
were assessed with use of the Knee Society radiographic criteria6 at an average of 4.7 years (range,
three to nine years) after the arthroplasty.
Four of the fifteen patients had poor wound-healing and prolonged
drainage for more than seven days and were treated with oral antibiotics
and daily sterile dressings until the wound closed completely. Of
these four patients, three had a fulminant infection. Two of them,
who had diabetes, required removal of the prosthesis; an Escherichia
coli infection had developed in one, and a Streptococcus
faecalis infection had developed in the other. Knee arthrodesis
was performed in both patients. The third patient was successfully treated
with a two-stage exchange arthroplasty after an appropriate
course of intravenous antibiotics. That patient had received chronic
suppression of infection with oral antibiotics preoperatively and
had had a pre-arthroplasty skin incision that did not heal, thus
necessitating a myocutaneous flap procedure simultaneously with
the arthroplasty procedure. Postoperatively, a pseudomonas infection
developed in this patient. The outcomes of these three patients
were considered failures.
Another patient required an exchange of the polyethylene thirteen
years after the knee replacement (Figs 1-A, 1-B, 1-C, 1-D, and 1-E).
Two of the fifteen patients had a patellar tendon rupture within
the first month after the arthroplasty. One rupture occurred intraoperatively
while the surgeon was attempting to obtain exposure, and the other
occurred thirty days postoperatively as a result of a fall. One
of these patients required reconstruction with an extensor mechanism
allograft7 and the other, with
a semitendinosus autograft8.
The average degree of flexion contracture improved from 15(range, 5 to 30) before the arthroplasty
to 4° (range, 0 to 10) after the arthroplasty. The average active
postoperative arc of motion (excluding the knees that had an arthrodesis) was
105 (range, 70 to 135) compared with 87 (range, 20 to 125) preoperatively,
an average increase of 18. Lucencies were noted on the radiographs
of all of the knees; however, they were incomplete and primarily
located between the tibial base-plate and the plateau.
Three of the fifteen patients required manipulation under anesthesia
because of stiffness, and one of them required a second manipulation.
As a result, all three patients had flexion of >90 (range,
95 to 120) and nearly full extension (one had a 5 flexion contracture).
One patient had persistent medial and lateral instability on examination
but had no complaints.
The average Hospital for Special Surgery knee score was 51 points
(range, 20 to 74 points) before the arthroplasty and 80 points (range,
44 to 91 points) at the time of follow-up. Four of the fifteen patients
had an excellent result; eight, a good result; one, a fair result;
and two, a poor result. The average postoperative SF-36
scores (with higher scores indicating better function) were 58.0 ± 10.0 points for general health, 72.4 ± 20.4 points for bodily pain, 72.1 ± 25.0 points for mental health, 58.3 ± 32.1 points for physical functioning, 84.6 ± 37.6 points for physical role functioning, 81.0 ± 23.5 points for social functioning, and 57.7 ± 26.0 points for vitality4,5,9,10.