Essex-Lopresti described a traumatic radial head fracture with
associated disruption of the interosseous membrane and secondary proximal
migration of the
radius1. The
combined long-term effect of this injury on the radiocapitellar and distal
radioulnar joints is both pain and loss of
mobility2-5.
Reconstructive management of these lesions has been very difficult and has
yielded variable, mostly unfavorable
results2,6.
The accepted basis of management consists of reestablishing the longitudinal
relationship of the radius to the ulna. This allows anatomic realignment of
the proximal radioulnar joint, the distal radioulnar joint, the proximal and
distal ligaments, and the interosseous membrane. This can be accomplished by
reestablishing the length of the proximally migrated radius, by shortening the
ulna, or by a combination of the two.
In 1998 and 1999, four patients with established symptomatic proximal
radial translation secondary to prior trauma and radial head excision were
treated at our institution with a total of five frozen radial head allografts.
Treatment was tailored on the basis of each patient's specific problems.
All elbows were approached through a lateral incision extending through the
Kocher interval. After the scar around the radial head was released and the
proximal aspect of the radius was refreshed, a matching allograft radial head
was tailored to fit the radiocapitellar gap. Ulnar shortening osteotomy was
carried out in three of the four patients to treat ulnar carpal abutment
causing ulnar wrist pain associated with axial instability of the forearm. A
midline ulnar incision was used to apply a slotted plate (the Rayhack plate)
designed for ulnar shortening (Creative Medical Design, Tampa, Florida).
After proper length and alignment were achieved, the allograft was fixed
with one 2.7-mm plate in two patients and with two mini-plates in the other
two. Transverse end-to-end apposition of the graft-host site was used in the
two patients treated with a single plate, and a step-cut was used in the other
two patients. Autogenous cancellous bone graft was added to the osteotomy site
in two of the patients. Ulnar shortening was then performed in three patients
after the degree of ulnar variance had been measured.
Our institutional review board granted us approval to report our experience
with the use of allograft radial head replacement in the treatment of
Essex-Lopresti lesions in the four patients. The patients were informed that
data concerning their cases would be submitted for publication.
Case 1. A thirty-three-year-old man sustained a fracture
of the right radial head in an automobile accident in January 1998, and a
radial head excision was performed at that time. Elbow and wrist pain as well
as loss of elbow motion and forearm rotation subsequently developed. Six
months after the injury, the range of elbow motion was 35° to 138° and
the patient had 0° of forearm supination and 80° of pronation.
Radiographs showed 13 mm of positive ulnar variance. The patient underwent an
allograft radial head replacement with mini-plate fixation and an ulnar
shortening osteotomy. At three months postoperatively, he had no pain or
instability. The range of elbow motion was 45° to 105°, with 30°
of forearm pronation and 15° of supination. Although he had no pain,
radiographs showed incomplete union of the radial head allograft.
Seven years after the surgery, the patient felt that the elbow was
"normal" with no pain. The range of elbow motion was from 15°
to 135°, and pronation and supination were 75° and 70°,
respectively (Figs. 1-A through
1-D). Radiographs revealed that the graft had healed and had not
resorbed.
Case 2. A twenty-seven-year-old man sustained a right radial
head fracture in 1987. He was treated with radial head excision and insertion
of a silicone radial head implant. Over the next four years, the right elbow
became increasingly painful. By 1991, the silicone implant had failed and it
was excised. Because of radiographically evident proximal radial migration
with secondary ulnar carpal abutment, an ulnar shortening osteotomy was also
performed at that time. The elbow pain persisted, and the patient was referred
to our institution, where he was found to have a range of elbow motion of
5° to 140°, with 80° of pronation and 80° of supination. He
underwent an allograft radial head replacement with double-mini-plate fixation
in August 1995.
After the surgery, the pain remained moderate at the extremes of motion.
Radiographs made at seventeen months showed a frank nonunion of the allograft.
The range of elbow motion was now limited to 50° to 105°, with 70°
of pronation and no supination. Allogenic bone graft was applied to the
interface between the allograft and the radius. Motion at the interface was
noted intraoperatively. The nonunion persisted and the radial head collapsed,
and it was resected in January 1998. Insertion of a metallic radial head
component was considered, but a component of sufficient length to compensate
for the bone loss was not available. The elbow and wrist pain worsened, and
six months after the allograft resection the range of elbow motion was 20°
to 130°, with 80° of pronation and 80° supination. Wrist
radiographs showed a slight increase in the ulnar variance despite the
previous ulnar shortening procedure. The patient indicated that he had
obtained the most improvement after treatment with the radial head allograft,
even though it had not united. Thus, the patient underwent another allograft
radial head replacement in October 1998. Radiographs made eighteen months
later showed non-union of this allograft, and autogenous bone-grafting was
performed. Nonunion persisted, and the second radial head allograft was
resected nineteen months later. Following removal, the arc of elbow flexion
was 30° to 145°, with 80° of pronation and 20° of
supination.
Case 3. A twenty-six-year-old construction worker fell from a
height of 15 ft (4.6 m) in June 1998. He sustained a fracture-dislocation of
the right elbow and a fracture of the right radial head and neck. He was
treated immediately with radial head excision and a silicone radial head
replacement. Discomfort in the wrist and forearm developed six months
following the injury. Radiographs demonstrated proximal radial migration and
bone loss attributable to silicone synovitis. Eighteen months after the
injury, the silicone replacement was removed and a radial head allograft was
inserted and fixed with a 2.7-mm plate. The patient continued to have moderate
pain at twelve months, and radiographs showed a nonunion at the allograft
site. He underwent revision fixation and autologous and allograft
bone-grafting in November 2001. Examination eight months later revealed a
range of elbow flexion of 0° to 145°, with pronation and supination
limited to 70° and 55°, respectively. Radiographs made at that time
revealed recurrence of the graft resorption at the interface with the host
bone (Figs. 2-A through
2-D).
Case 4. A thirty-seven-year-old male carpenter fell from a
scaffolding in September 1998 and sustained a fracture of the left radial
head. With nonoperative treatment, he attained an arc of elbow motion of
30° to 100° and was unable to rotate the forearm. Four months after
the injury, he was diagnosed as having radial head malunion and posterior
subluxation. In January 1999, he underwent a radial head excision. Despite
this treatment and a subsequent manipulation, the range of motion never
improved. Sequential radiographs of the wrist made between November 1998 and
June 1999 showed progressive prominence of the distal part of the ulna with
abutment against the ulnar aspect of the carpus.
The patient presented to our institution in June 1999 with left lateral
elbow and ulna-sided wrist pain. The range of flexion of the elbow at that
time was 20° to 105°, and he had no rotation from a 20° pronated
position. Flexion and extension of the left wrist were limited to 20° and
40°, respectively. In August 1999, the patient underwent an allograft
radial head replacement with fixation with a 2.7-mm plate and screws. A 5-mm
ulnar shortening procedure with plate fixation was also performed. Despite a
130° pronation-supination arc intraoperatively, the range of elbow motion
at four months postoperatively was 25° to 130°, with 15° of
pronation and 15° of supination. Hardware loosening was noted on
radiographs, and the patient was treated with grafting of a nonunion of the
allograft-host interface, tension-band wire fixation of the nonunion, and
release of the distal and proximal radioulnar joint capsules. Seven months
following this procedure, the patient noted no relief of the wrist or elbow
pain. The elbow range of motion at this time was 15° to 130°, with
35° of pronation and 10° of supination. Radiographs showed resorption
of the radial head allograft. The patient then underwent resection of the
allograft and insertion of a cobalt-chromium radial head prosthesis (Avanta,
San Diego, California). One year following this procedure, the patient had
less pain in the left elbow and wrist, but elbow motion was restricted to
30° to 125°, with 50° of pronation and 40° of supination. He
did not return to carpentry work and sought retraining.
Diagnosing an acute Essex-Lopresti injury is often difficult, and
treating the sequelae of proximal radial migration can be
troublesome2,7.
The central tenet of treatment is the establishment of a stable
radiocapitellar joint through fixation of the native radial head or by
replacing the radial head with a prosthetic implant. The inadequacy of
silicone implants is well
known8-11,
and this has prompted some to suggest the use of an allograft or a metallic
component capable of withstanding the longitudinal radial
loads12-14.
When clinically relevant shortening has taken place, some authors have
advocated conversion to a one-bone forearm or the creation of a radioulnar
synostosis15,16.
Ulnar shortening is also commonly performed when a patient has carpal
impingement12.
Recently, we have used the mobilized anconeus as a space-filling pedicle
graft, with encouraging
results7, but
long-term follow-up is necessary before the value of that procedure can be
fully determined.
In this report on four patients who underwent a total of five radial head
allograft procedures because of documented proximal radial migration after
radial head resection, three of the five allografts failed and had to be
excised. Only one patient demonstrated satisfactory elbow function with a
useful range of forearm rotation and radiographic evidence of allograft union
at the time of follow-up.
In 1997, Szabo et al. reported their experience with frozen allograft
radial head replacement for the treatment of established symptomatic proximal
translation of the
radius14. Ilizarov
external fixation was used in three of their five patients to reestablish the
length of the radius and to unload the allograft fixation until union
occurred. They concluded that allograft replacement should be done cautiously
despite favorable outcomes in their series. Ilizarov fixation was not used in
any of the patients in our series. However, all but one patient underwent an
ulnar shortening osteotomy to correct the relative malalignment of the distal
radioulnar joint prior to, or concurrent with, the allograft procedure. Thus,
efforts to reestablish the radioulnar relationship distally with an ulnar
shortening osteotomy may inadvertently affect allograft survival adversely.
Also, it is possible that our patients had more severe deformity than did
those in the study by Szabo et al.
The optimal method of fixation of these allograft replacements is also
uncertain. A variety of plates have been used. A broad, stable host-allograft
interface is of utmost importance for a solid union. However, larger plates
may cause more extensive scar formation, affect allograft revascularization,
or interfere with the mechanics of the proximal radioulnar joint. The use of
more axially directed screws, which we now employ for acute fractures, may be
considered in the future. The benefit of adding bone graft at the interface
site is not clear because of the risk of synostosis or heterotopic
ossification. Szabo et al. attributed some limitation of forearm rotation in
their series to the insertion of additional bone graft and recommended against
grafting14.
The results in our four patients were not optimal. The recently developed
modular prosthetic radial head implants, which were not available when these
patients were managed, may prove to be a better solution. Alternatively, the
most acceptable final salvage option might still be the creation of a one-bone
forearm. This procedure was not done in any of our patients as they all
preferred to live with the symptoms rather than undergo further treatment.
The literature clearly documents an 80% rate of failure of efforts at
reconstruction in patients in whom an acute Essex-Lopresti lesion was
initially
missed2,6.
In our study, still another effort to solve the problem, with an allograft
radial head replacement, also proved unreliable for the treatment of patients
with an Essex-Lopresti injury.
A table presenting clinical details on all four patients 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). ?