Fractures of the radial head or neck usually occur after a fall on the
outstretched arm1.
Undisplaced fractures usually result in minimal residual deficits independent
of
treatment2,3,
whereas displaced fractures have been described as having the risk of an
unfavorable
outcome4,5.
Even if short-term follow-up does not reveal functional deficits, residual
posttraumatic joint incongruity may lead to elbow osteoarthritis with
increasing problems over time.
The purpose of this study was to determine the incidence and the long-term
outcome of isolated, closed Mason type-II and III fractures of the radial head
or neck in adults. We hypothesized that these fractures have a low incidence,
minimal long-term functional deficits, and a high risk for the development of
radiographic evidence of degenerative changes.
The study was performed in a city of 264,937 inhabitants, of whom 208,765
were 16.5 years of age or older in November 1970. The incidence of a defined
fracture within this population could be determined retrospectively as
virtually all patients with a fracture are seen at the trauma unit at the only
emergency hospital in the city. City residents who sustain a fracture
elsewhere are later referred to the orthopaedic department for follow-up, and
at this visit the fracture is classified and entered into the hospital
registry. With this relatively comprehensive case documentation system, it is
possible to retrospectively identify individuals with a specific type of
fracture sustained over a specified period of time. Additionally, all
radiographs and reports have been saved in the archives of the hospital for
the past half-century; therefore, the fracture type can be reevaluated and
reclassified retrospectively with use of the primary radiographs. Fewer than
3% of all city residents who sustain a fracture are cared for outside of this
hospital system and thus are not included in the hospital registry.
Seven hundred and fifty-six individuals who sustained a radial head or neck
fracture between 1969 and 1979 were identified in the hospital registry. All
patients with a Mason type-II or III fracture who were still living in the
city were invited to return for reexamination, at a mean of nineteen years
(range, fourteen to twenty-five years) after the injury. No patient had
another major fracture or soft-tissue injury associated with the proximal
radial fracture. Of the 131 individuals still living in Malmo, seven refused
to participate in the study (an attendance rate of 95%) and twenty-four had
been children (younger than 16.5 years of age) at the time of injury and thus
were excluded. One hundred individuals who had been 16.5 years of age or older
at the time of the fracture were included in the study. The subjective results
for all 100 individuals were evaluated with a questionnaire, and seventy-five
also returned for a clinical and radiographic evaluation.
Seventy patients were women, and thirty were men. The mean age was
forty-seven years (range, eighteen to seventy-three years) at the time of the
injury. Seventy-seven injuries were the result of low-energy trauma, defined
as a fall or direct impact, and twenty-two were caused by high-energy trauma,
defined as a fall from >2 m or a motor-vehicle accident. Information
regarding the injury mechanism was missing for one patient. The left elbow was
affected in forty-eight patients and the right elbow, in fifty-two. There were
seventy-six Mason type-II fractures, twenty-three of which were radial neck
fractures, and twenty-four Mason type-III fractures, five of which were radial
neck fractures.
Primary treatment included an elastic bandage or a collar and cuff sling
with mobilization as soon as the pain allowed for forty-four patients
(forty-one Mason type-II fractures and three Mason type-III fractures; thirty
radial head fractures and fourteen radial neck fractures), cast immobilization
for a median of two weeks (range, one to three weeks) for thirty-four patients
(twenty-eight Mason type-II fractures and six Mason type-III fractures;
twenty-eight radial head fractures and six radial neck fractures), acute
radial head excision in nineteen patients (four Mason type-II fractures and
fifteen Mason type-III fractures; twelve radial head fractures and seven
radial neck fractures), open reduction of the radial head fracture prior to
plaster cast immobilization in two patients (two Mason type-II, radial head
fractures), and a repair of the medial collateral ligament in one patient (a
Mason type-II, radial neck fracture). The method of treatment was chosen
according to the preference of the surgeon on call when the patient presented
with the fracture. No complications were recorded during or after the surgery.
Nine patients (six Mason type-II fractures and three Mason type-III fractures;
eight radial head fractures and one radial neck fracture) had a secondary
radial head resection, after a mean of thirteen months (range, two to
forty-one months), because of residual pain, and one patient (a Mason type-II,
radial neck fracture) underwent neurolysis of the ulnar nerve three years
after injury.
The subjective outcome was assessed with a questionnaire that evaluated
activities of daily living; elbow pain on loading and at rest; and tenderness,
loss of strength, and range of motion of the affected elbow. Strength, range
of motion, and pain in the wrist and hand of the injured extremity were also
evaluated. The uninjured arm served as the control.
Seventy-five former patients returned for clinical examination, which was
performed by two of the authors who had not been involved in the treatment of
the patients. Flexion and extension of the elbow and the wrist, pronation and
supination of the forearm, and the angle of the extended elbow were measured
with a goniometer. Grip strength of the hand was evaluated with a Martin
vigorimeter (Werkstätten für Medizinmechanik; Heinrich C. Ulrich,
Ulm-Donau, Germany), and the circumference of the arm 10 cm proximal to the
tip of the olecranon and the circumference of the forearm 10 cm distal to the
tip of the olecranon were measured with a tape measure. The uninjured arm
served as the control. The difference in the strength of elbow flexion and
extension between the two sides was estimated by subjective comparison of the
elbows. The Tinel test in the cubital tunnel was performed in both elbows. The
outcome was also graded according to the classification system of Steinberg et
al.6, with the
understanding that that classification system was developed for children.
According to the system, a good result corresponds to full mobility and no
pain in the elbow; a fair result, to <20° of loss of motion in any
direction on objective examination or slight pain; and a poor result, to
=20° of loss of mobility in any direction or pain at rest.
On the basis of the primary radiographs, the fractures were classified as
involving the radial head or radial neck and according to the classification
described by Mason and later modified by Broberg and
Morrey7
(Fig. 1). This classification
was performed by a radiologist with no knowledge of the treatment or the
subjective or clinical outcome. Follow-up radiographs included anteroposterior
and lateral projections of the elbows. Subchondral cysts, subchondral
sclerosis, and/or osteophytes were defined as degenerative changes, and the
number of individuals with more than a 1-mm reduction in the joint space in
comparison with the joint space in the uninjured elbow was
recorded8. The
diameter of the radial head and miscellaneous pathological entities, such as
nonunion, avascular necrosis, proximal radioulnar synostosis, and
periarticular ossification, were also documented. The uninjured elbows served
as controls.
The study was approved by the ethical committee of the university hospital.
Comparisons of values for the two arms of the same individual were performed
with the Student t test between pairs and the chi-square test, with p <
0.05 indicating a significant difference. Data are presented as the mean and
the range or the mean and the standard deviation.
Between 1969 and 1979, 2965 patients sustained an elbow fracture. Seven
hundred and fifty-six (26%) of these patients sustained a fracture of the
radial head or neck, with 480 (64%) sustaining a Mason type-I fracture; 222
(29%), Mason type-II; thirty-six (5%), Mason type-III; and eighteen (2%),
Mason type-IV. Six hundred and seventy-two (89%) of the individuals with a
radial head or neck fracture were 16.5 years of age or older at the time of
the fracture. Thus, the annual incidence of fractures of the radial head or
neck was 2.9 per 10,000 individuals of 16.5 years of age or older at the time
of the fracture.
Of the 100 individuals included in the follow-up evaluation, seventy-seven
had no symptoms in the previously fractured elbow; twenty-one (twelve Mason
type-II fractures and nine Mason type-III fractures; seventeen radial head
fractures and four radial neck fractures) had occasional but not daily pain,
mainly when lifting heavy objects; and two (one Mason type-II fracture and one
Mason type-III fracture; both radial head fractures) had daily pain (one of
whom had pain at rest). Twelve individuals had a subjectively reduced range of
motion of the previously injured elbow, and five also noted weakness and
numbness in the hand of the previously injured extremity.
The previously injured elbows had less flexion (138° ± 8°
compared with 140° ± 7°; p < 0.001) and less extension
(—4° ± 8° compared with —1° ± 6°; p
< 0.001) than did the uninjured elbows, and there was less supination of
the forearm on the injured side (83° ± 11° compared with
86° ± 6°; p < 0.01)
(Table I). Nineteen previously
fractured elbows had a positive Tinel sign over the ulnar nerve compared with
eight of the unfractured elbows (p < 0.05). According to the Steinberg
classification, 84% of the outcomes were classified as good; 11%, as fair; and
5%, as poor.
The previously injured elbows had more degenerative changes seen on
radiographs than did the uninjured elbows, with cysts in forty-six injured
elbows compared with nine uninjured elbows, irregular subchondral bone in
fifty compared with eleven, and osteophytes in forty-two compared with nine.
The prevalence of degenerative changes was 76% in the injured elbows compared
with 16% in the uninjured elbows. All differences were significant (p <
0.001). Eight previously fractured elbows had a reduction in the joint space
of >1 mm compared with the joint space in the uninjured elbow, and six
uninjured elbows had a reduction in the joint space of >1 mm compared with
the joint space in the previously fractured elbow. On the average, the radial
head diameter in the extremities with a previously injured elbow was larger
than that in the uninjured extremities (24 ± 2 mm compared with 22
± 2 mm, p < 0.001). No cases of nonunion, avascular necrosis,
proximal radioulnar synostosis, or periarticular ossification were
identified.
Of the nine patients with a secondary radial head excision performed to
treat residual pain, only three subjectively rated the injured elbow as being
as good as the uninjured one at the time of follow-up; however, none
classified the previously fractured elbow as poor. When these results were
compared with the outcomes for the nineteen patients with primary radial head
excision, twelve of whom subjectively rated the previously injured elbow as
being as good as the uninjured one, the difference was not significant, with
the numbers available.
Twenty-five percent of all elbow fractures in this urban population were
isolated radial head or neck fractures. Eighty-nine percent of the radial head
and neck fractures were sustained by individuals 16.5 years of age or older,
so the annual incidence was 2.9 per 10,000 individuals in this age group. The
outcome of an isolated Mason type-II or III fracture of the radial head or
neck is predominantly favorable up to twenty-five years after the injury, with
no increased prevalence of reduced joint space in the previously injured
elbow.
The strength of this study is that we were able to obtain reliable
epidemiological data by using an existing fracture registry. The registry
enabled us to identify most of the fractures within the defined population,
even when a fracture had been sustained twenty-five years
earlier9. If
fracture incidence is determined only with use of a questionnaire, the number
of fractures can be underreported by 40% or overreported by
20%10. Also, the
follow-up period in our study was one of the longest published to date, our
sample size was larger than that in most previous studies, and we performed an
additional radiographic evaluation of most of the former patients (Figs.
2-A and
2-B). The weakness of the study
is that the treatment was not randomized. This is an important limitation
since, although the treatment was somewhat consistent, it certainly was not
standardized. There also is a possibility of detection bias in the way in
which the range-of-motion and other outcome measures were obtained during the
examination.
Our findings are in accordance with those in the published literature.
Arner et al.3 and
Poulsen and
Tophöj2
reported that 99%3
and 95%2 of
individuals with a radial head or neck fracture had a good or excellent
outcome. However, Mason type-I fractures were included in those reports, so a
comparison with the current study must be done with caution. Of 239
individuals with a Mason type-II or III fracture in the study by Arner et al.,
two had a poor outcome, as classified with the Steinberg
classification6. Of
thirty patients with such a fracture in the study by Poulsen and Tophöj,
two had a poor result, defined as reduced working ability and/or limitation of
motion exceeding 10° in one or more directions.
A reduced range of motion, usually described as a stiff elbow, is a common
symptom after elbow trauma, and Morrey et
al.11,12
and other
authors5,13-15
have suggested that patients often have complaints when an elbow extension
deficit exceeds 30° and/or elbow flexion is <130°. Other authors
have suggested that patients have subjective complaints when the extension
deficit exceeds
20°6, and even
an extension deficit exceeding 10° has been described as a poor
outcome2. In our
series, decades after the injury, the patients had a mean residual deficit of
2° in elbow flexion, 3° in elbow extension, and 3° in supination
compared with the ranges of the uninjured elbow. These deficits are smaller
than those in several other
studies5,16,
perhaps because only uncomplicated Mason type-II and III fractures were
included in our study or because of our longer follow-up period. Only four
patients in the current study had a loss of extension of =20°, a
deficit that Steinberg et al. defined as a poor
outcome6. However,
none of these four individuals had pain at rest, although two had occasional
pain. According to the Steinberg classification, 84% of the results in our
study should be classified as good; 11%, as fair; and 5%, as poor. The poor
results were due to an extension deficit of =20° in four patients and
to pain at rest in one.
The outcome of the primary treatment must be classified as poor in the ten
patients who needed a secondary operation because of persistent pain following
the fracture. There was a nonsignificant tendency for the patients with
secondary excision of the radial head to have an inferior outcome: 67% had
minor complaints compared with 37% of those who had had a primary excision.
These data raise the question of whether a secondary excision of the radial
head produces a less favorable outcome than does a primary excision. An
alternative explanation is that individuals who need a later excision of the
radial head had a different type of fracture than those treated with a primary
operation, but this hypothesis cannot be verified in this study. It seems that
a simple radial head excision, regardless of when it is done, results in a
clinically acceptable long-term outcome in the majority of patients. This
notion is supported by the great majority of published
studies2,13,16-19.
However, most of these studies included few individuals, had a short to
medium-term follow-up, and varied with regard to the classification system
that was used, so comparison between studies is difficult. In contrast, other
studies have demonstrated unfavorable outcomes after radial head excision.
Ikeda and Oka5
reported an average extension deficit of 25° in individuals who had
undergone primary radial head excision following a Mason type-II or III
fracture. However, all of the cited studies comparing the outcomes of primary
and late excisions of the radial head following a Mason type-II or III
fracture, including the current study, must be regarded with skepticism, and
all conclusions should be drawn with caution. The evaluated fracture cohorts
were small, resulting in a high chance of a type-II error, and the extent of
the associated soft-tissue injuries was not always described. Additional
long-term studies of a larger sample of patients should be performed to
compare the outcomes of primary and secondary radial head excisions.
The present study also supports previous findings that radiographically
evident deformity develops in most patients with a fracture of the radial head
or
neck18,20,21.
We found degenerative changes in 76% of the previously injured elbows.
However, there was apparently no correlation between these radiographic
degenerative changes and pain or loss of motion, which occurred in only a
minority of patients. The radiographically visible enlargement of the radial
head also seems to be of minor clinical relevance as does the radiographic
appearance of a reduced joint space.
In summary, isolated closed Mason type-II or III fractures of the radial
head or neck in adults usually have a favorable long-term outcome. This is
also true for patients treated with either a primary or a secondary radial
head excision. We did not identify a less favorable outcome in individuals who
had a secondary procedure, including a radial head excision due to persistent
pain, but more studies with a larger sample size are needed to answer this
question. We concluded that the long-term outcome of a Mason type-II or III
radial head or neck fracture in an adult is favorable when the treatment
strategies that we described are used.