Abstract
Background: Effective methods of treating an unstable distal radial
fracture are described in the literature, but there is no reliable method of
identifying an unstable fracture in time to initiate appropriate treatment.
The purposes of this study were to identify the predictors of fracture
instability and to construct a method of prospectively predicting the
radiographic outcome.
Methods: Data on approximately 4000 distal radial fractures were
prospectively recorded over a 5.5-year period. The database was validated by
reexamining a sample of it. Demographic data on the patients and mode of
injury, as well as the fracture classification and measurements, were recorded
at the time of presentation. Outcome measures consisted of radiographic
measurements made at one week and six weeks and assessment of carpal alignment
at six weeks. Univariate and multiple logistic regression analyses were
performed to identify the significance of the data obtained at presentation in
the prediction of early and late instability as well as the risk of malunion
and carpal malalignment.
Results: The predictors of early and late instability and malunion
differed according to the displacement of the fracture at presentation.
Patient age, metaphyseal comminution of the fracture, and ulnar variance were
the most consistent predictors of radiographic outcome. Dorsal angulation was
not found to be significant in the prediction of radiographic outcome for
displaced fractures. The degree to which the patient was independent was
predictive of malunion in minimally displaced and displaced fractures.
Formulas that are predictive of each of the seven radiographic outcome
measurements were constructed.
Conclusions: The study succeeded in identifying the factors that are
prognostic of the radiographic outcome for distal radial fractures. Formulas
to predict the radiographic outcome were constructed as the independent
prognostic significance of these factors was quantified. These formulas can be
used to inform the surgeon's decision about the nature of primary treatment of
fractures of the distal aspect of the radius. However, they must be validated
by further studies before they are used to impact the management of distal
radial fractures.
Level of Evidence: Prognostic Level I. See Instructions
to Authors for a complete description of levels of evidence.
Fractures of the distal aspect of the radius are
common1, and they
constitute a substantial proportion of the workload in orthopaedic trauma
practice2. Stable
fractures can be managed conservatively, with good anatomical and functional
results3,4.
However, the management of the unstable fracture of the distal part of the
radius continues to stimulate debate, particularly when such a fracture occurs
in an elderly patient. There is currently general agreement that there is a
close relationship between anatomy and
function3-9.
This implies that treatment should strive to regain as near an anatomical
position as possible, to optimize the functional outcome. A number of
treatment methods, including internal and external fixation techniques, have
been shown to restore and maintain the radiographic position until union of
the
fracture9-14.
Some of these techniques have been shown to be effective in the treatment of
this fracture in elderly
patients9,14.
The diagnosis of an unstable fracture of the distal aspect of the radius is
usually made by observation of the behavior of the fracture after initial
treatment in a cast. Our standard initial management of a displaced fracture
is closed manipulation followed by application of a cast. Instability is
diagnosed by radiographic examination between one and two weeks later. If
appropriate, definitive surgical treatment is instituted at that stage for
fractures displaying early instability. However, fractures that exhibit
instability after two weeks are not detected by this management protocol.
A reliable method of predicting instability at the time of presentation
would enable timely definitive surgical treatment to be undertaken with
evidence-based decision-making. The patient would not have to wait for the
diagnosis of instability to be made. Unnecessary manipulation of the displaced
unstable fracture could be avoided, and fractures with late instability would
be less likely to proceed to malunion.
Various studies have defined factors that are predictive of
instability15-17.
Computer algorithms have been used to predict the radiographic outcome for
individual fractures, with reasonable
results18.
Unfortunately, this predictive method has not been validated. In an attempt to
quantitatively predict anatomical outcome, we prospectively gathered data for
approximately 4000 distal radial fractures. The aims of this study were (1) to
identify the factors relating to the patient, mode of injury, and the fracture
that were of significant prognostic value with respect to radiographic
outcomes of the fracture, and (2) to use this data to construct mathematical
formulas that are predictive of radiographic outcomes of the fracture.
Definitions
The premorbid normal level of function of the patient was categorized as
independent, which was defined as the ability to go shopping without
assistance, or as dependent, which was defined as the need for assistance in
order to go shopping or an inability to go shopping.
Fracture displacement (Fig.
1) was characterized as minimally displaced when there was dorsal
angulation of
=10°5 and an
ulnar variance of <3
mm5,19,
as displaced when there was dorsal angulation of >10° and/or ulnar
variance of >3 mm, or as redisplaced when previous manipulative reduction
had been performed. An acceptable reduction was a fracture with dorsal
angulation of =0° and an ulnar variance of =3 mm following closed
reduction, and an unacceptable reduction was a fracture with dorsal angulation
of >0° and/or an ulnar variance of >3 mm following closed
reduction.
Carpal malalignment was deemed to be present when the long axes of the
radius and capitate failed to intersect within the carpus on the lateral
radiograph of the
wrist5. To permit an
assessment of the natural history of the fracture, this measurement was made
with use of the radiographs used for malunion (described below).
Fracture instability was divided into three types: early instability, late
instability, and malunion. Early instability was defined as a fracture that
was displaced (or redisplaced following closed reduction) radiographically
within two weeks after the injury. Late instability was defined as a fracture
that was displaced radiographically at the time of union (i.e., six weeks) but
had not previously demonstrated early instability. Malunion was defined as the
theoretical risk of a fracture being displaced at the time of union if the
fracture was untreated other than by closed reduction at presentation. This
allowed an assessment of the natural history of the fracture. For the purposes
of identifying factors predictive of malunion, any fracture that was displaced
(or redisplaced) at or prior to union was recorded as a malunion (as a
malunion would have occurred if no intervention other than closed manipulative
reduction had been undertaken).
Demographic Data
During the period from June 1988 to December 1993, 28,376 new fractures
were seen at the Orthopaedic Outpatient Department of the Royal Infirmary of
Edinburgh. The distal aspect of the radius was involved in 4024 (14.2%) of the
fractures, and they formed the study group. Seventy-nine percent (3173) of the
fractures were in female patients, and 21% (851) were in male patients. The
age range was fourteen to 100 years (mean, sixty-four years) for female
patients and fifteen to ninety-four years (mean, forty-two years) for male
patients. The mean age for all patients was fifty-nine years. Fifty-one
patients had a bilateral fracture.
Seven groups were considered for analysis, and the demographics of each
group are shown in Table I. The
number of fractures in each group varied for two reasons. With respect to
early instability and malunion, the difference in numbers was due to missing
data. With respect to late instability, the numbers are also reduced because
of the definition of late instability: only fractures that had not
demonstrated early instability could be used in this analysis.
Displaced fractures occurred in older patients and were more frequent in
females. These patients were less likely to be independent. The displaced
fracture was also more frequently due to a low-energy injury (i.e., a fall
from standing height).
Database Construction
Fracture management followed a standard protocol. The emergency-room staff
undertook the initial assessment and treatment. Fractures deemed to be in an
acceptable position were managed with a dorsal plaster-of-Paris slab. If the
fracture position was thought to be unacceptable, the emergency-room staff,
prior to application of a dorsal plaster-of-Paris slab, performed closed
reduction using intravenous regional anesthesia.
The exclusion criteria were (1) skeletal immaturity, (2) primary operative
treatment, (3) prior fracture malunion, and (4) missing data.
Primary operative treatment was selected for intraarticular fractures or
volarly displaced fractures (ninety-four patients), open fractures
(thirty-nine patients), fractures with severe nerve compression syndromes (ten
patients), and fractures complicated by compartment syndrome (three patients).
Patients with missing radiographic data from the evaluation at the time of
presentation were excluded from all statistical analysis, as fractures were
analyzed according to whether they were displaced or minimally displaced at
presentation. Patients with missing radiographic data from the one-week
evaluation were excluded from the analyses of early instability and late
instability. Patients with missing radiographic data from the six-week
evaluation (union) were excluded from analyses of late instability and
malunion.
The patients were evaluated clinically and radiographically at
approximately one and six weeks after the injury as per the protocol of the
orthopaedic trauma unit, which included radiographs of the normal, uninjured
wrist made at one week.
At approximately one week following the injury, the data on the patients
were reviewed by the senior author (M.M. McQ.) in a special research clinic.
The clinical, demographic, and radiographic data were recorded and entered
into a database either by the senior author or a research nurse. In addition
to standard demographic details, the patient's normal level of function (as
described earlier) and mode of injury were recorded. Radiographs of the
uninjured, normal wrist were made, and radiographs of the injured wrist were
repeated. The patients with a fracture that had maintained a good position
(i.e., no early instability was demonstrated) had the dorsal slab completed to
a below-the-elbow forearm cast with the wrist in slight flexion and ulnar
deviation. Patients with a fracture that had displaced (or redisplaced) were
admitted to the orthopaedic trauma unit for further intervention, unless the
patient had low functional demands and operative intervention was deemed
inappropriate.
The patients were subsequently evaluated at approximately six weeks.
Radiographs were repeated for the assessment of displacement and carpal
alignment. If surgical intervention had been undertaken prior to six weeks for
displaced fractures, then all radiographic measurements subsequent to surgery
were excluded from statistical analysis and the fracture was recorded as
having gone on to malunion.
Radiographic Measurement Techniques
All radiographs (those made at presentation, at the time of reduction, and
at the one-week and six-week evaluations) were measured manually with use of a
protractor and a ruler to provide values for the dorsal angle, radial
shift19, and ulnar
variance20. These
measurements are illustrated in Figure
1. The ulnar variance and radial shift are expressed as the
difference between the injured side and the normal, uninjured side. In the
cases of the patients for whom normal values were unavailable, the mean values
for the normal side were
used17. The
fractures were classified with use of both the Frykman and AO/OTA
classifications21,22.
The type of metaphyseal comminution was recorded, according to the location,
as absent or as involving the dorsal metaphysis, volar metaphysis, or both the
dorsal and volar metaphysis. Thus, comminution was a purely qualitative
measurement. Carpal alignment (as defined earlier) was assessed on the
radiographs made at the time of the final review. The senior author alone was
responsible for fracture classification and the assessment of comminution and
carpal alignment.
Database Validation
No widely accepted method of database validation is available. Validation
was performed by reexamining selected data for a sample of patients. Nineteen
data fields were reexamined for 116 patients (every thirtieth patient in the
database). Data fields included demographic information, mode of injury, and
all radiographic measurements. Any difference in categorical data was
considered an error. For quantitative data, a difference of >5° in
angular measurement and 2 mm in length was considered an error. Data for 331
fields were missing, leaving 1808 pairs of data for comparison.
The error rate was expressed simply as a percentage calculated according to
the following formula:
(pairs of data in
disagreement×100)total number of pairs of complete
data
Errors in categorical fields were rare. No difference in the error rates
was recorded for angular measurements compared with length measurements. The
overall rate of error was 3.5%.
Statistical Methods
The following variables were analyzed: (1) age, (2) gender, (3) mechanism
of injury, (4) independence, (5) type of comminution, (6) AO/OTA type, (7)
AO/OTA group, (8) AO/OTA subgroup, (9) Frykman classification, (10) dorsal
angle at presentation, (11) radial shift at presentation, (12) ulnar variance
at presentation, (13) dorsal angle at one week, (14) radial shift at one week,
and (15) ulnar variance at one week.
Univariate association between outcomes and potential predictors was tested
by chi-square tests for nominal data and Mann-Whitney tests for ordinal or
quantitative data. All variables achieving a p value of =0.05 in univariate
analysis were included in the regression analysis. Stepwise logistic
regression was then used to determine independently significant variables,
thus accounting for relationships or confounding between variables. The
parameter estimates produced by the regression analysis were then used to
produce a predictive formula for radiographic outcome. The instability score
calculated from the formula could then be converted to a percentage with the
use of an exponential equation. The goodness of fit of each logistic
regression analysis was assessed with the Cox and Snell R2
statistic.
The results of statistical analysis for minimally displaced and displaced
fractures and carpal malalignment are summarized in tables in the Appendix. It
is noteworthy that marked differences are seen between univariate and
regression analyses in all cases. For example, patient gender, mode of injury,
and original radial shift are almost invariably significant following
univariate analysis, and they are almost invariably of no prognostic value
following multiple logistic regression. The latter statistical method allows
demonstration of dependence between variables; thus, gender loses significance
as most females are older and males are younger.
Fractures That Were Minimally Displaced at Presentation
Prediction of Early Instability
Early instability occurred in 149 (10%) of 1486 patients with minimally
displaced fractures at presentation. Only four factors remained significant
following regression analysis. The most important factor was age (p <
0.001). Early instability occurred ten times more frequently in patients who
were more than eighty years old compared with patients who were less than
thirty years old. Comminution was also highly significant (p < 0.001).
Early instability occurred six times more frequently in fractures with any
form of comminution (dorsal, volar, or both dorsal and volar) compared with
fractures with no comminution. The original dorsal angle and ulnar variance
were also of significance (p < 0.01). Early instability occurred five times
more frequently in fractures with a dorsal angle between 5° and 10°
compared with fractures that maintained a degree of volar angulation. The
frequency of early instability in fractures with an ulnar variance of
>0° was twice that in fractures with an ulnar variance of
=0°.
Prediction of Late Instability
Late instability occurred in 244 (22%) of 1125 patients. This was a lower
rate of late instability than that reported by other
investigators16.
Only the age of the patient (p < 0.001), the presence of comminution (p
< 0.01), and the dorsal angle (p < 0.001) and ulnar variance (p <
0.001) measured at one week retained significance following regression
analysis. Again, age was the most important factor, with late instability
occurring four times more frequently in patients more than eighty years old
compared with patients less than thirty years old. The presence of fracture
comminution of any type increased the frequency of late instability by a
factor of three.
Prediction of Malunion
If all 1333 minimally displaced fractures had been treated only with closed
reduction at presentation, then 27% (354) would have gone on to malunion.
Factors that were predictive of malunion following regression analysis were
age, dorsal angle, and ulnar variance at presentation (p < 0.001);
comminution, AO/OTA group, and radial shift at presentation (p < 0.01); and
independence, AO/OTA subgroup, and Frykman classification (p < 0.05). An
age of more than eighty years increased the frequency of malunion by sixfold
compared with an age of less than thirty years. The frequency of malunion is
three times as great in fractures with a dorsal angle of between 4° and
10° compared with fractures maintaining some degree of volar angulation.
Malunion is three times more common in fractures with any type of comminution.
In fractures with no involvement of the distal radioulnar joint (a Frykman
score of 1 to 4), malunion occurred more frequently in fractures involving the
ulnar styloid (a Frykman score of 2 and 4). Dependent patients more frequently
had a malunion.
Fractures Displaced at Presentation
Prediction of Early Instability
Early instability occurred in 682 (43%) of 1595 displaced fractures.
Following regression analysis, the factors retaining significance were age and
ulnar variance at presentation (p < 0.001) and the type of comminution (p
< 0.01). Early instability occurred in 62% of the fractures in patients who
were more than eighty years old, nearly three times more frequently than in
patients who were less than thirty years old. This is also two and one-half
times more frequent in patients more than eighty years old with a minimally
displaced fracture. Dependent patients again had greater rates of early
instability. Interestingly, in displaced fractures, the dorsal angle had no
effect on the frequency of early instability.
Prediction of Late Instability
Late instability occurred in 391 (47%) of 829 patients. After regression
analysis, the following factors were significant: age (p < 0.001) and the
dorsal angle (p < 0.001) and ulnar variance (p < 0.001) measured at one
week. As in minimally displaced fractures, the radiographic measurements at
presentation are of no significance in predicting late instability.
Prediction of Malunion
Malunion in displaced fractures was very common. If all 1236 displaced
fractures had been treated by closed reduction alone at presentation, then 60%
(744) would have gone on to malunion.
Age, ulnar variance and patient independence at presentation (p <
0.001), and the type of comminution (p < 0.01) retained significance
following regression analysis. Age had a striking effect: 82% of the patients
more than eighty years old had a malunion compared with 30% of those who were
less than thirty years old. Again, malunion occurred more frequently in
dependent patients. An ulnar variance of >2 mm increased the frequency of
malunion by 10%. As was seen for fractures with early instability, dorsal
angulation had no bearing on the frequency of malunion in displaced
fractures.
Prediction of Carpal Alignment in All Fractures
Carpal malalignment was present in 1121 (31.5%) of 3559 patients. Age,
independence, type of comminution, AO/OTA subgroup, and original dorsal
angulation (p < 0.001 for each) all retained significance following
regression analysis. In contrast to the prediction of instability, for which
the presence or absence of comminution was important, it was specifically
dorsal comminution that was associated with an increase in the frequency of
carpal malalignment. Carpal malalignment occurred more often in fractures of
AO/OTA subgroup 2, and less often in AO/OTA type-B fractures.
The Probability of Fracture Instability and Carpal Malalignment
The formulas for instability scores (X) were calculated from the weighting
of the significance of factors following the regression analysis. These
formulas are listed below along with the Cox and Snell R2
statistic.
Minimally Displaced Fractures
For the prediction of early instability (R2 = 0.12): X = (0.03
× age) + 1.39 (if any type of comminution present) + (0.05 ×
dorsal angle at presentation) + (0.21 × ulnar variance at presentation)
— 4.82.
For the prediction of late instability (R2 = 0.18): X = (0.03
× age) + 0.93 (if any type of comminution present) + 1.45 (if dorsal
angle is >4° at one week) + 0.33 (if ulnar variance >—1 mm at
one week) — 3.92.
For the prediction of malunion (R2 = 0.2): X = (0.03 ×
age) + 1.04 (if any type of comminution present) + (0.06 × dorsal angle
at presentation) + (0.2 × ulnar variance at presentation) —
3.41.
Displaced Fractures
For the prediction of early instability (R2 = 0.09): X = (0.03
× age) + 0.38 (if there is any type of comminution) + (0.21 ×
ulnar variance at presentation) — 3.12.
For the prediction of late instability (R2 = 0.11): X = (0.02
× age) + (0.07 × dorsal angle at one week) + (0.22 × ulnar
variance at one week) — 1.53.
For the prediction of malunion (R2 = 0.13): X = (0.04 ×
age) — 0.8 (if independent) + 0.53 (if comminution type = dorsal) +
(0.09 × ulnar variance at presentation) — 1.65.
All Fractures
For the prediction of carpal malalignment (R2 = 0.12): X = (0.03
× age) — 0.56 (if independent) — 0.97 (if comminution type =
none) — 0.46 (if comminution type = dorsal and volar) + 0.34 (if AO/OTA
subgroup = 2) + (0.0017 × dorsal angle at presentation) —
2.14.
The probability of instability or carpal malalignment can be expressed as a
percentage with use of the following conversion equation:
Probability(%)=([ex]×100)/(1+ex)
where X is the value calculated from the predictive formulas above.
In practice, the prediction of malunion is of the most clinical relevance,
especially in displaced fractures.
Most work examining the factors that are predictive of radiographic outcome
has not made the distinction between early and late collapse of the fracture.
In addition, fracture instability has not been examined with stratification of
the position of the fracture at presentation. The heterogeneity of study
populations may explain the disagreement among authors as to the most
important factors determining fracture stability. Hove et
al.15 analyzed data
for 645 conservatively managed distal radial fractures. Using multiple
regression analysis, the authors found that the initial dorsal angulation,
radial length, and patient age were predictors of malunion. It was not
possible to determine whether these factors may have been different for early
and late fracture collapse, as all fractures that required treatment for
redisplacement at one week were excluded from the analysis.
Jenkins16 found
that the position of the fracture at presentation was a good indicator of the
fracture position at union. He also found that the absence of dorsal
comminution was protective against malunion in dorsal angulation. However, he
did not calculate the independent significance of these factors. He also
analyzed only radiographic data. All of the patients in his study had a
displaced fracture at presentation, but displacement was not defined other
than by the need for reduction. Lafontaine et
al.23 reported
predictors of fracture instability: age (>60 years), dorsal angulation
(>20°), dorsal comminution, intra-articular fracture (radiocarpal joint
surface), and associated ulnar fracture. Again, all fractures in the study
were initially displaced. A scoring system for fracture instability was
produced on the basis of the aforementioned factors. This scoring system has
been further evaluated, finding the age of the patient most useful in the
assessment of
instability24.
Abbaszadegan et
al.17 calculated
the independent significance of a range of factors; ulnar variance, Lidstrom
class24, and
patient age proved to be the most important. The authors also quantified the
risk of instability but only with respect to ulnar variance. Adolphson et
al.18 used
computerized analysis to quantify the risk of fracture instability. However,
this method of prediction has not been prospectively validated.
A list of significant predictors of fracture instability is
available25. The
presence or absence of these predictors will facilitate management
decision-making. However, assessment with use of these predictors is
qualitative. Fracture instability cannot be quantified with use of this
information; thus, the assessment of instability is ultimately based on
clinical judgment. The present study demonstrated that certain factors
identifiable at presentation have a highly significant relationship with
radiographic outcomes of distal radial fracture. From the results of the
statistical analysis, it was possible to construct weighted mathematical
formulas that predict these outcomes. These formulas are based on data
available in the emergency room and are thus applicable at presentation. If
validated, these formulas will be an invaluable adjunct in the decision-making
process for the management of distal radial fractures.
The most important predictive factors were the age of the patient, the type
of comminution of the fracture, and the position of the fracture at
presentation. We showed that early and late instability and carpal
malalignment increase relentlessly with age. In the prediction of malunion,
the presence or absence of comminution was important, whereas in the
prediction of carpal malalignment, the location of comminution (dorsal) was
important. This is unsurprising as carpal malalignment following distal radial
fracture is due almost exclusively to dorsal instability of the
carpus26.
The dorsal angle at presentation was not important in predicting malunion
in displaced fractures. This result is explained by the importance of
comminution. Dorsal angle and comminution are intimately related in displaced
fractures. For a greater dorsal angle to be present, the metaphysis must be
deficient, except in fractures in which the distal fragment is dorsally
translocated. Thus, of the two factors, only comminution retained significance
following regression analysis. The position of the fracture at presentation
was predictive of the position at union. Malunion occurred more frequently in
displaced fractures. Initial ulnar variance was consistently significant.
Dorsal angle was of variable significance. It was important in the fractures
that were minimally displaced at presentation and in the prediction of carpal
malalignment. Patients with low functional requirements were unlikely to have
anything other than nonoperative management of the fractures. Although the
independence of the patient was likely to influence the decision as to whether
the fracture required manipulation, it was still an independently significant
predictor of radiographic outcome.
In minimally displaced fractures, those classified as AO/OTA group 3 had a
significantly higher frequency of early instability (p < 0.05). These were
almost exclusively A3.2 and C3.2 fractures, both of which have metaphyseal
comminution. Similar findings were noted in the prediction of malunion in
minimally displaced fractures. In the prediction of carpal malalignment, the
AO/OTA type and subgroup were significant (p < 0.01 and p < 0.001,
respectively). Carpal malalignment was more frequent in fractures with no
continuity between the diaphysis and the articular surface (types A and C),
and in subgroup 2 (again, mainly A3.2 and C3.2) because of the metaphyseal
component of the fracture.
The Frykman classification was also significant in the prediction of carpal
malalignment (p < 0.05) and malunion (p < 0.05) in minimally displaced
fractures. The frequency of malunion in fractures involving the ulna and those
involving the radiocarpal joint increased by 15% and 20%, respectively. The
frequency of carpal malalignment was reduced by approximately half if
fractures were in group 1 or 3. Other investigators have reported the
importance of ulnar styloid
involvement27. In
contrast, the Frykman classification was of no prognostic value in displaced
fractures. In addition, it must be noted that the prognostic value of these
classification systems is affected by their marked interobserver
variability28,29.
In view of the size of the database, reexamination of all data was thought
to be impractical. The measurement and recording of data were consistent in
the samples of patient data that were reexamined. It is recognized that the
missing data restrict the data validation and the statistical analysis of the
study.
In an independent fifty-five-year-old patient with a displaced distal
radial fracture with no comminution and no shortening (ulnar variance =
0°), the value of X is calculated as follows:
X=(0.04×55)-0.8+(0.09×0)-1.65=-0.25
With the conversion formula used to produce a percentage probability of
malunion:
Probability=(e-0.25×100)/(1+e-0.25)=44%
In an independent eighty-five-year-old patient with a dorsally comminuted
displaced fracture with an ulnar variance of 2 mm, the value of X is
calculated as follows:
X=(0.04×85)-0.8+0.53+(0.09×2)-1.65=1.66
In addition, the probability of malunion is calculated as follows:
Probability=(e1.66×100)/(1+e1.66)=82%
Thus, the risk of malunion is higher in the older patient with comminution.
However, the significance of the difference in probabilities is yet to be
ascertained. To help in decision-making, these probabilities need to be
converted into a binary outcome: will this fracture go on to malunion or not?
Thus, prospective studies are required to validate the predictive formulas.
The formula must be specific (i.e., able to detect stable fractures),
otherwise the number of stable fractures treated surgically will be
unacceptably high. It must also be sensitive (i.e., able to detect unstable
fractures), otherwise there will be little reduction in the number of unstable
fractures that go on to malunion.
In conclusion, this study identified at presentation the patients in whom
instability and carpal malalignment were more frequent. The position of the
fracture at presentation influences the radiographic outcome. The age of the
patient, the type of comminution, and the ulnar variance at presentation are
important predictive factors. We produced a predictive formula for carpal
malalignment. This phenomenon is related to the degree of malunion of the
fracture, but it may occur in fractures with relatively little displacement.
This may go some way to explaining the results of studies that have
demonstrated that functional outcome is not necessarily related to
radiographic outcome and that functional outcome may deteriorate with
time30-33.
As far as we are aware, this is the only study identifying independently
significant predictors of the radiographic outcome of distal radial fractures.
It is also, as far as we know, only the second study to produce a method of
prospectively quantifying the risk of fracture instability. It is hoped that a
user-friendly method of prediction can be produced from the mathematical
formulas derived in the study. If validated, these predictive formulas have
the potential to remove the delay in definitive fracture treatment and to
reduce the number of late corrective procedures that need to be performed.
Tables presenting the results of the statistical analysis for minimally
displaced and displaced fractures are 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). ?
Robertsson GO, Jonsson GT, Sigurjonsson
K. Epidemiology of distal radius fractures in Iceland in 1985. Acta
Orthop Scand. 1990;61:
457-9.61457Â
1990Â
[PubMed][CrossRef] Â
Abbaszadegan H, Conradi P, Jonsson U.
Fixation not needed for undisplaced Colles' fracture. Acta Orthop
Scand. 1989;60:
60-2.6060Â
1989Â
[CrossRef] Â
Cooney WP. Management of Colles'
fractures. J Hand Surg [Br].
1989;14:
137-9.14137Â
1989Â
[PubMed] Â
McQueen M, Caspers J. Colles fracture:
does the anatomical result affect the final function? J Bone Joint Surg
Br. 1988;70:
649-51.70649Â
1988Â
Â
McQueen MM, Hajducka C, Court-Brown CM.
Redisplaced unstable fractures of the distal radius: a prospective randomised
comparison of four methods of treatment. J Bone Joint Surg Br.
1996;78:
404-9.78404Â
1996Â
[PubMed] Â
McQueen MM. Epidemiology of
fractures of the radius and ulna. In: McQueen MM, Jupiter JB, editors.
Musculoskeletal trauma series: radius and ulna. Oxford: Butterworth-Heinemann;
1999.Â
1999Â
Â
Fernandez DL. Correction of
post-traumatic wrist deformity in adults by osteotomy, bone-grafting, and
internal fixation. J Bone Joint Surg Am.
1982;64:
1164-78.641164Â
1982Â
[PubMed] Â
McQueen MM, Michie M, Court-Brown CM.
Hand and wrist function after external fixation of unstable distal radial
fractures. Clin Orthop Relat Res.
1992;285:
200-4.285200Â
1992Â
[PubMed] Â
McQueen MM. Redisplaced unstable
fractures of the distal radius. A randomised, prospective study of bridging
versus non-bridging external fixation. J Bone Joint Surg Br.
1998;80:
665-9.80665Â
1998Â
[PubMed][CrossRef] Â
Krishnan J, Chipchase LS, Slavotinek J.
Intraarticular fractures of the distal radius treated with metaphyseal
external fixation. Early clinical results. J Hand Surg [Br].
1998;23:
396-9.23396Â
1998Â
[PubMed] Â
Melendez EM, Mehne DK, Posner MA.
Treatment of unstable Colles' fractures with a new radius mini-fixator.
J Hand Surg [Am]. 1989;14:
807-11.14807Â
1989Â
[PubMed][CrossRef] Â
Jenkins NH, Jones DG, Johnson SR,
Mintowt-Czyz WJ. External fixation of Colles' fractures. An anatomical study.
J Bone Joint Surg Br.
1987;69:
207-11.69207Â
1987Â
[PubMed] Â
Seitz WH Jr, Froimson AI, Leb R, Shapiro
JD. Augmented external fixation of unstable distal radius fractures. J
Hand Surg [Am]. 1991;16:
1010-6.161010Â
1991Â
[CrossRef] Â
Orbay JL, Fernandez DL. Volar
fixed-angle plate fixation for unstable distal radius fractures in the elderly
patient. J Hand Surg [Am].
2004;29:
96-102.2996Â
2004Â
[PubMed][CrossRef] Â
Hove LM, Solheim E, Skjeie R, Sorensen
FK. Prediction of secondary displacement in Colles' fracture. J Hand
Surg [Br]. 1994;19:
731-6.19731Â
1994Â
Â
Jenkins NH. The unstable Colles'
fracture. J Hand Surg[Br].
1989;14:
149-54.14149Â
1989Â
[PubMed] Â
Abbaszadegan H, Jonsson U, von Sivers K.
Prediction of instability of Colles' fractures. Acta Orthop
Scand. 1989;60:
646-50.60646Â
1989Â
[CrossRef] Â
Adolphson P, Abbaszadegan H, Jonsson U.
Computer-assisted prediction of the instability of Colles' fractures.
Int Orthop. 1993;17:
13-15.1713Â
1993Â
[PubMed] Â
van der Linden W, Ericson R. Colles'
fracture. How should its displacement be measured and how should it be
immobilized? J Bone Joint Surg Am.
1981;63:
1285-8.631285Â
1981Â
[PubMed] Â
Melone CP Jr. Articular fractures of the
distal radius. Orthop Clin North Am.
1984;15:
217-36.15217Â
1984Â
[PubMed] Â
Müller ME, Nazarian S, Koch P,
Schatzker J. The comprehensive classification of fractures of long
bones. Berlin: Springer;1990.Â
1990Â
Â
Frykman GK. Fracture of the distal
radius including sequelae—shoulder-handfinger syndrome, disturbance in
the distal radio-ulnar joint and impairment of nerve function. A clinical and
experimental study. Acta Orthop Scand.
1967;Suppl 108:
3.3Â
1967Â
Â
Lafontaine M, Hardy D, Delince P.
Stability assessment of distal radial fractures. Injury1989;20:
208-10.20208Â
1989Â
[PubMed][CrossRef] Â
Lidstrom A. Fractures of the distal end
of the radius. A clinical and statistical study of end results. Acta
Orthop Scand. 1959;Suppl 41:
1-118.1Â
1959Â
Â
Ruedi TP, Murphy WM, editors. AO
principles of fracture management. New York: Thieme;
2000. p 358.358Â
2000Â
Â
Bickerstaff DR, Bell MJ. Carpal
malalignment in Colles' fractures. J Hand Surg [Br].
1989;14:
155-60.14155Â
1989Â
[PubMed] Â
Oskarsson GV, Aaser P, Hjall A. Do we
underestimate the predictive value of the ulnar styloid affection in Colles'
fractures? Arch Orthop Trauma Surg.
1997;116:
341-4.116341Â
1997Â
[PubMed][CrossRef] Â
Illarramendi A, Gonzalez Della Valle A,
Segal E, De Carli P, Maignon G, Gallucci G. Evaluation of simplified Frykman
and AO classifications of fractures of the distal radius. Assessment of
interobserver and intraobserver agreement. Int Orthop.
1998;22:
111-5.22111Â
1998Â
[PubMed][CrossRef] Â
Flikkila T, Nikkola-Sihto A, Kaarela O,
Paakko E, Raatikainen T. Poor interobserver reliability of AO classification
of fractures of the distal radius. Additional computed tomography is of minor
value. J Bone Joint Surg Br.
1998;80:
670-2.80670Â
1998Â
[PubMed][CrossRef] Â
Roumen RM, Hesp WL, Bruggink ED.
Unstable Colles' fractures in elderly patients. A randomised trial of external
fixation for redisplacement. J Bone Joint Surg Br.
1991;73:
307-11.73307Â
1991Â
[PubMed] Â
Warwick D, Field J, Prothero D, Gibson
A, Bannister GC. Function ten years after Colles' fracture. Clin Orthop
Relat Res. 1993;295:
270-4.295270Â
1993Â
Â
Altissimi M, Antenucci R, Fiacca C,
Mancini GB. Long-term results of conservative treatment of fractures of the
distal radius. Clin Orthop Relat Res.
1986;206:
201-10.206201Â
1986Â
Â
Tsukazaki T, Takagi K, Iwasaki K. Poor
correlation between functional results and radiographic findings in Colles'
fracture. J Hand Surg [Br].
1993;18:
588-91.18588Â
1993Â
[PubMed] Â