Trauma
Determinants of functional outcome after
H. J. Kreder,
N. Rozen,
C. M. Borkhoff,
Y. G. Laflamme,
M. D. McKee,
E. H. Schemitsch,
D. J. G. Stephen
From the University
of Toronto,
Sunnybrook Health
Science Centre and St.
Michael’s Hospital,
Toronto, Canada

H. J. Kreder, MD, MPH,
FRCS(C), Associate Professor

D. J. G. Stephen, MD,
FRCS(C), Assistant Professor
Division of Orthopaedics
Sunnybrook Health Science
Centre, Suite MG365, 2075
Bayview Avenue, Toronto,
Ontario, Canada M4N 3M5.

N. Rozen, MD, Professor
Rambam Medical Centre, Haifa,
Israel.

C. M. Borkhoff, MSc,
Research Associate

M. D. McKee, MD, FRCS(C),
Associate Professor

E. H. Schemitsch, MD,
FRCS(C), Professor
Division of Orthopaedic
Surgery
University of Toronto, St
Michael’s Hospital, Toronto,
Canada.

Y. G. Laflamme, MD, FRCS(C),
Assistant Professor
Division of Orthopaedic Surgery
University of Montreal, Hôpital
du Sacré Coeur, Montreal,
Canada.
Correspondence should be sent
to Dr H. J. Kreder c/o Bev
Bulmer; e-mail:
bev.bulmer@sunnybrook.ca
©2006 British Editorial Society
of Bone and Joint Surgery
doi:10.1302/0301-620X.88B6.
17342 $2.00
J Bone Joint Surg [Br]
2006;88-B:776-82.
Received 24 October 2005;
Accepted after revision
19 January 2006
776
simple and complex acetabular fractures
involving the posterior wall
We have evaluated the functional, clinical and radiological outcome of patients with simple
and complex acetabular fractures involving the posterior wall, and identified factors
associated with an adverse outcome.
We reviewed 128 patients treated operatively for a fracture involving the posterior wall of
the acetabulum between 1982 and 1999. The Musculoskeletal Functional Assessment and
Short-Form 36 scores, the presence of radiological arthritis and complications were
assessed as a function of injury, treatment and clinical variables.
The patients had profound functional deficits compared with the normal population.
Anatomical reduction alone was not sufficient to restore function. The fracture pattern,
marginal impaction and residual displacement of > 2 mm were associated with the
development of arthritis, which related to poor function and the need for hip replacement.
It may be appropriate to consider immediate total hip replacement for patients aged > 50
years with marginal impaction and comminution of the wall, since 7 of 13 (54%) of these
required early hip replacement.
Approximately 35% of all acetabular fractures ior wall which were treated between 1982 and
have involvement of the posterior wall which is 1999 at a level-one tertiary referral trauma
the most common pattern,1 with approxi- centre. A total of 34 patients were excluded,
mately 76% of these injuries having additional six because they had stable fractures not
complex fractures.2 It has generally been per- requiring surgery according to Tornetta’s crite-
ceived that isolated fractures of the posterior ria,12,13 27 because they did not have clinical
wall have a good outcome, but recent reviews follow-up for at least one year and one because
have shown that 21% to 32% of patients have he underwent immediate hip replacement; this
poor results.2-9 Moed et al7 had the best clini- left 128 patients in the study, which was
cal results with a poor outcome in only 11% of approved by the institutional internal review
patients with a simple fracture of the posterior board.
wall.8,10,11 There have been only a few studies Anteroposterior plain radiographs of the
which have determined the functional outcome pelvis and iliac and obturator oblique views
using validated instruments, and considered were obtained in all patients at the initial pre-
complex and simple fractures of the posterior sentation. CT scans were carried out after the
wall separately. patient had been resuscitated and closed reduc-
We have studied the functional, clinical and tion of any associated posterior dislocation of
radiological outcome in patients with simple the hip carried out. The fractures were oper-
and complex fracture patterns involving the ated on using the Kocher-Langenbeck poster-
posterior acetabular wall with reference to the ior approach,1 with the addition of an ilio-
type of fracture, marginal impaction, commi- inguinal approach for selected fractures with
nution of the posterior wall, residual fracture associated T and transverse patterns that
displacement, injury to the femoral head, asso- required exposure of the anterior column. The
ciated injuries and increasing age. We also joint was irrigated and debrided by distracting,
examined the inter-relationship between the but without re-dislocating, the femoral head
radiological, clinical and functional results. from the acetabulum. The column fractures
were then reduced and secured by lag screws.
Patients and Methods Impacted fragments at the margins were ele-
We identified 162 skeletally mature patients vated and buttressed with autograft harvested
with acetabular fractures involving the poster- from around the greater trochanter. Finally, the
THE JOURNAL OF BONE AND JOINT SURGERY
 DETERMINANTS OF FUNCTIONAL OUTCOME AFTER ACETABULAR FRACTURES INVOLVING THE POSTERIOR WALL
fragments of the posterior wall were reduced and secured
by two or more lag screws, followed by buttress plating.
Those which were too small for fixation by lag screws were
buttressed by spring plates. The patients were given peri-
operative antibiotics for 24 to 48 hours after surgery. Cou-
madin or low-molecular-weight heparin was used as
thromboembolic prophylaxis until discharge from either
hospital or inpatient rehabilitation. Indomethacin was used
for six weeks after surgery, at the discretion of the treating
surgeon, as prophylaxis against heterotopic bone forma-
tion. After operation, the patients were mobilised with
weight-bearing on the uninjured side. Full weight-bearing
on the injured hip was delayed for 12 weeks from fixation.
All patients were reviewed clinically by the treating sur-
geon for a minimum of one year after operation. Clinical
details and information on the nature of the injury, associ-
ated injuries and treatment, complications and the length of
hospital stay were gathered from the hospital records by an
independent observer (CMB). Beginning in 1997, all patients
were reviewed clinically at six weeks, six months and annu-
ally thereafter, with the functional outcomes being measured
by the Musculoskeletal Functional Assessment (MFA)14,15
and the 36-item Short-Form Health Survey (SF-36)16 ques-
tionnaires at each clinic visit. All patients treated before
1997 who had not been reviewed recently were sent the
questionnaires by a research assistant (CMB). Failure to
return the questionnaire within two weeks was followed by
at least two attempts to contact each patient by telephone.
If the patient had moved, an attempt was made to locate
them by contacting the last-known family physician and
any relatives listed in the hospital records.
The MFA is a validated 101-item self-administered
health-status instrument designed to detect differences in
function among patients after traumatic musculoskeletal
injuries.14,15 A lower score indicates better function, with a
score above 40 in any domain indicating profound func-
tional impairment.5 It has recently been demonstrated to
perform well in patients with fractures of the acetabulum.11
The SF-36 is a well-established and validated general
health-status measure.16
Two orthopaedic fellows (NR, YGL) who were not
involved with the management of the patient and were
blinded to their outcome reviewed the initial, immediate
post-operative and final radiographs. Their consensus opin-
ion was recorded for each parameter. There was no instance
when consensus could not be reached. Pre-operative assess-
ment of the fracture was made using the initial antero-
posterior radiographs and Judet views1 as well as the pre-
operative CT scans. The fractures were classified into the
ten types as described by Letournel.17 Radiographs and CT
scans taken after reduction of any associated hip disloca-
tion were reviewed for the presence of loose intra-articular
fragments, posterior-wall displacement, fractures of the
femoral head, marginal wall impaction and comminution
of wall fragments. A fracture was considered to be commi-
nuted only if it consisted of three or more fragments with
VOL. 88-B, No. 6, JUNE 2006
777
attached articular cartilage. Non-articular cortical frag-
ments were not included in this count. Radiographs taken
within two weeks of surgery, at one year and at the final fol-
low-up were reviewed for maximum residual displacement
of the fracture and 67 patients (52%) were supplemented
by an available CT scan. All articular displacement was
considered, including the posterior wall or the articular
component of an associated-column fracture. The final
radiograph was used to assess osteoarthritis, the grade of
heterotopic ossification and the presence of avascular
necrosis. The grade of heterotopic ossification was deter-
mined according to the method of Brooker et al18 as modi-
fied by Moed and Smith.19 Radiological arthritis was
evaluated according to the method of Kellgren and
Lawrence20 and Gunther and Sun.21 The raters considered
avascular necrosis to have occurred if the femoral head
showed segmental sclerosis, or if there was segmental col-
lapse of the head which predated radiological changes on
the acetabular side and loss of joint space, in order to min-
imise misinterpretation of early degenerative arthritis as
avascular necrosis.1
Statistical analysis. This was performed using SPSS version
10.0 (SPS Inc., Chicago, Illinois). Routine descriptive statis-
tics were used to describe the clinical characteristics and
functional and radiological outcome. Bivariate compari-
sons were made using either the Mann-Whitney or Pearson
chi-squared tests. Specific hypotheses were tested by mod-
elling the outcome of interest (function, heterotopic ossifi-
cation, avascular necrosis, total hip replacement, and
residual displacement) as a linear function of the predictor
variables (fracture pattern, marginal impaction, wall com-
minution, residual displacement, damage to the femoral
head and also arthritis, heterotopic ossification and avascu-
lar necrosis), while adjusting for patient age, gender and
length of follow-up as potential confounding variables. A p
value < 0.05 was considered to be statistically significant.
For clarity, statistically significant bivariate relationships
were reported, but only if these remained statistically signif-
icant after adjusting for the potential confounders noted
above.
Results
According to Letournel’s classification17 there were 44 sim-
ple and 84 associated posterior-wall fractures. In the latter
group the complex fracture was transverse in 51 patients,
involved the posterior column in 22 and was a T-shaped
acetabular fracture in 11.
A total of 61 (48%) of patients were men under 50 years
of age who had sustained multiple injuries in a motor-vehi-
cle accident. The mean age of the patients was 41.6 years
(18 to 75), and 100 (78%) were men. A motorised vehicle
was involved in 108 (84%) of the 128 cases. The mean
Injury Severity Score was 23.4 (9 to 50). A total of 34
patients (27%) presented with an associated head injury,
which was generally mild, with a mean Glasgow Coma
Score of 12.7 (6 to 14). An associated pelvic fracture was
778 H. J. KREDER, N. ROZEN, C. M. BORKHOFF, Y. G. LAFLAMME, M. D. MCKEE, E. H. SCHEMITSCH, D. J. G. STEPHEN

Table I. Clinical details and injury characteristics by fracture pattern

Fracture pattern*
Isolated PW PW + TV PW + PC PW + T Overall p value
Number 44 51 22 11 128 NA¶
Mean age (SD) (yrs) 41.8 (13.1) 40.5 (13.4) 42.8 (14.0) 43.8 (11.1) 41.6 (13.1) 0.829
Gender, male (%) 38 (86.4) 38 (74.5) 16 (72.7) 8 (72.7) 100 (78.1) 0.440
Side, right (%) 17 (38.6) 22 (43.1) 8 (36.4) 4 (36.4) 51 (39.8) 0.958
Mean ISS† (SD) 23.5 (10.4) 25.3 (10.3) 18.7 (7.3) 23.4 (9.2) 23.4 (10.0) 0.147
Mean GCS‡ (SD) 14.4 (1.6) 14.5 (1.6) 15.0 (0) 15.0 (0) 14.6 (1.4) 0.376
Mean time to OR§ (SD) 6.0 (10.6) 6.2 (5.8) 4.4 (7.6) 2.6 (2.8) 5.5 (7.9) 0.484

Associated injury
Spine (%) 3 (6.8) 5 (9.8) 3 (13.6) 0 (0) 11 (8.6) 0.569
Chest (%) 17 (38.6) 24 (47.1) 4 (18.2) 4 (36.4) 49 (38.2) 0.124
Abdomen (%) 5 (11.4) 9 (17.6) 1 (4.5) 3 (27.3) 18 (14.1) 0.249
Upper limb (%) 15 (34.1) 15 (29.4) 2 (9.1) 1 (9.1) 33 (25.8) 0.093
Lower limb extremity (%) 21 (47.7) 25 (49.0) 7 (31.8) 7 (63.6) 60 (46.9) 0.376

Initial nerve deficit (%) 16 (36.4) 12 (23.5) 8 (36.4) 4 (36.4) 40 (31.3) 0.296
Initial vascular deficit 1 (2.3) 2 (3.9) 2 (9.1) 0 (0) 5 (3.9) 0.509

Acetabular fracture
Comminuted wall (%) 18 (40.9) 28 (54.9) 11 (50.0) 6 (54.5) 63 (49.2) 0.598

Marginal impaction (%) 10 (22.7) 23 (45.1) 16 (72.7) 4 (36.4) 53 (41.4) 0.001**

Intra-articular fragments (%) 13 (29.5) 16 (31.4) 4 (18.2) 4 (36.4) 37 (28.9) 0.623
Femoral head lesion (%) 8 (18.2) 16 (31.4) 3 (13.6) 2 (18.2) 29 (22.7) 0.277
* PW, posterior wall; TV, transverse; PC, posterior column, T, associated T type
† ISS, injury severity score
‡ GCS, Glasgow coma scale
§ OR, operating room
** significant, p < 0.05
¶ NA, not available

Table II. Clinical outcomes according to fracture pattern

Fracture pattern*
Isolated PW PW + TV PW + PC PW + T Overall p value
†
Number complications 44 51 22 11 128
Late THR (%) 3 (6.8) 6 (11.8) 4 (18.2) 3 (27.3) 16 (12.5) 0.205
DVT (%) 1 (2.3) 2 (3.9) 1 (4.5) 2 (18.2) 6 (4.7) 0.136
Infection (%) 2 (4.5) 3 (5.9) 1 (4.5) 2 (18.2) 8 (6.3) 0.345
Iatrogenic nerve injury (%) 1 (2.3) 1 (2.0) 0 (0) 0 (0) 2 (1.6) 0.869

Residual displacement in mm (%)

0 to 2 42 (95.5) 42 (82.4) 15 (68.2) 9 (81.8) 108 (84.4) 0.015‡
3 to 5 2 (4.5) 9 (17.6) 7 (31.8) 1 (9.1) 19 (14.8) 0.034‡
6 to 10 0 (0) 0 (0) 0 (0) 1 (9.1) 1 (0.8) 0.034‡
Arthritis (%)
Any (grade 1 to 4) 9 (20.5) 22 (43.1) 14 (63.6) 4 (36.4) 49 (38.3) 0.01‡
Mild (grade 1 and 2) 4 (9.1) 14 (27.5) 6 (27.3) 3 (27.3) 27 (21.1) 0.018‡
Severe (grade 3 and 4) 5 (11.4) 8 (15.7) 8 (36.4) 1 (9.1) 22 (17.2) 0.037‡
Heterotopic bone (%)
Indocid prophylaxis 26 (59.1) 32 (62.7) 15 (68.2) 8 (72.7) 81 (63.3) 0.783
Any (grade 1 to 4) 20 (45.5) 20 (39.2) 11 (50.0) 4 (36.4) 55 (43.0) 0.929
Mild (grade 1 and 2) 17 (38.6) 13 (25.5) 7 (31.8) 4 (36.4) 41 (32.0) 0.283
Severe (grade 3 and 4) 3 (6.8) 7 (13.7) 4 (18.2) 0 (0) 14 (10.9) 0.362
Avascular necrosis (%) 3 (6.8) 5 (9.8) 5 (22.7) 2 (18.2) 15 (11.7) 0.364
* PW, posterior wall; TV, transverse; PC, posterior column, T, associated T type
† THR, total hip replacement; DVT, deep-vein thrombosis
‡ significant, p < 0.05

THE JOURNAL OF BONE AND JOINT SURGERY

 DETERMINANTS OF FUNCTIONAL OUTCOME AFTER ACETABULAR FRACTURES INVOLVING THE POSTERIOR WALL 779

Table III. Comparison of the 128 patients with and without functional out- patients (3%) with injury to the L5 nerve-root, one recov-
come results
ered fully and the other three, partially.
Completed Did not complete Subsequently, 16 patients (12.5%) required a total hip
functional outcome functional outcome
Variable* questionnaires questionnaires replacement (THR) at a mean of 2.9 years (0.4 to 10.4)
Number 84 44
after initial surgery. In eight patients this was within two
Mean age in years (SD) 42.6 (12.4) 39.6 (14.4) years of the acetabular fracture, in six at between two and
Male (%) 66 (78.6) 34 (77.3) ten years and in two after ten years. The incidence of THR
Mean ISS (SD) 23.6 (9.4) 23.2 (11.3) varied with the type of fracture from just below 6.8% (3 of
Mean GCS (SD) 14.7 (1.2) 14.2 (1.8)
Mean time to operation 4.7 (4.6) 7.2 (12.0)
44) for isolated fractures of the posterior wall to over
in days (SD) 27.3% (3 of 11) for those with an associated T component
(Table II), but there was no statistical significance (p = 0.21;
Associated injuries (%)
Upper limb 19 (22.6) 13 (29.5)
Mann-Whitney U test) between the type of fracture and the
Lower limb 42 (50) 18 (40.9) incidence of THR. Total joint replacement was more likely
to be required in posterior-wall fractures associated with
Fracture type (%) marginal impaction (p = 0.01), wall comminution (p =
Associated acetabular 61 (72.6) 23 (52.3)†
fracture 0.005) and in patients older than 50 years (p = 0.01). In
patients aged over 50 years of age with marginal impaction
Acetabular fracture (%) and comminution of the posterior wall, the likelihood of
Comminuted wall 40 (47.6) 23 (52.3)
Marginal impaction 33 (39.3) 20 (45.5)
requiring THR was 46%, compared with 9% for younger
Intra-articular fragments 20 (23.8) 17 (38.6) patients without these fracture characteristics (p = 0.002).
Femoral head lesion 22 (26.2) 7 (15.9) THR was required more often in patients with residual dis-
* ISS, injury severity score; GCS, Glasgow coma scale placement of the fracture of more than 2 mm compared
† significant, p < 0.05 with 2 mm or less, but this did not reach statistical signifi-
cance given the sample size (p = 0.06). THR was not asso-
ciated with the presence of a lesion of the femoral head after
adjusting for other factors.
present in 13 patients (10%) and 60 (46.9%) had an injury Radiological outcome. After open reduction and internal
to the lower limb. The fracture of the posterior wall was fixation, 108 (84%) of the reductions were within 2 mm of
comminuted with more than three articular fragments, as residual displacement (step or gap), but this varied accord-
defined above in 63 (49%) of the patients and marginal ing to the type of fracture (Table II). Only two simple frac-
impaction was present in 53 (41%) (Table I). The latter was tures of the posterior wall (4.5%) had displacement greater
most common in associated fractures of the posterior col- than 2 mm (both 3 mm) compared with seven with associ-
umn, occurring in 16 of the 22 cases (73%). ated fractures of the posterior column (31.8%).
Closed reduction of the hip dislocation was undertaken Radiological evidence of osteoarthritis was identified in
as soon as possible after injury but the precise timing was 49 patients (38.3%) at a mean of 5.3 years (1 to 18) (Table
not consistently available. Definite fixation of the fracture II). It was significantly more common (p = 0.01; Mann-
was achieved at a mean of 4.7 days from the day of injury Whitney U test) in fractures of the posterior-wall with asso-
(1 to 40). ciated posterior-column fractures (63.6%; 14 of 22) than in
Clinical outcome. Six patients (4.7%) developed deep-vein simple fractures of the posterior wall (20.5%; 9 of 44) and
thrombosis, one of whom also had a non-fatal pulmonary was more common when marginal impaction was noted
embolism, while a further eight patients (6.3%) met the cri- (p = 0.02) and when the residual step or gap was greater
teria for infection, defined as wound drainage or dehiscence than 2 mm (p = 0.001). Scuffing or impaction of the femo-
treated with antibiotics or by debridement. The latter pro- ral head did not have a significant independent effect on the
cedure was required in two patients (1.6%). All the infec- development of arthritis or the need for early THR
tions resolved with treatment. (p > 0.05). Patients with radiological osteoarthritis had sig-
The peroneal branch of the sciatic nerve was injured nificantly worse functional scores (p < 0.05). Avascular
iatrogenically in two patients (1.6%), both of whom recov- necrosis of the femoral head was identified most often in
ered fully. A total of 40 patients (31%) presented with a posterior-wall with associated column fractures (23%)
nerve injury, with complete recovery in 26, partial recovery (Table II). Unlike arthritis, it was not associated with mar-
in five and no recovery in nine. There were 22 patients ginal impaction or residual step or gap deformity (p >
(17%) with complete palsy of the sciatic nerve of whom 0.05).
only one recovered fully, 16 showed partial recovery and Mild heterotopic bone was seen in 41 (32%) patients
five showed no recovery. One patient with isolated palsy of (Table II). There was no association between prophylaxis
the tibial nerve recovered fully. Of 13 patients (10%) with and the development of heterotopic ossification (p = 0.853)
palsy of the peroneal nerve, two recovered fully, seven given the sample size available. The 84 patients who com-
recovered partially and four had no recovery. Of the four pleted the functional questionnaires also answered ques-

VOL. 88-B, No. 6, JUNE 2006

780 H. J. KREDER, N. ROZEN, C. M. BORKHOFF, Y. G. LAFLAMME, M. D. MCKEE, E. H. SCHEMITSCH, D. J. G. STEPHEN

Table IV. Functional outcome according to type of fracture. The normalised Z scores are reported for the eight domains and physical function scores of
the SF-36. The score values represent the number of SD above (positive values) or below (negative values) the age- and gender-matched normal pop-
ulation. The SF-36 and MFA scores for simple posterior-wall fractures and associated fractures are compared, separate from overall comparisons of
the associated fracture types with p values in the corresponding columns

Fracture pattern*
Isolated PW Associated p value PW + TV PW + PC PW + T Overall p value
Number 23 61 38 13 10 84
SF-36
Physical component summary -1.94 -2.26 0.379 -2.15 -2.81 -2.01 -2.17 0.434
(-4.72 to 0.22) (-6.59 to 0.70) (-5.04 to 0.37) (-6.59 to 0.70) (-4.17 to -0.27) (-6.59 to 0.70)
Mental component summary -0.82 -1.19 0.195 -1.19 -1.44 -0.88 -1.08 0.411
(-3.71 to 1.38) (-3.91 to 0.93) (-3.91 to 0.78) (-3.45 to 0.34) (-1.85 to 0.93) (-3.91 to 1.38)
Physical function -1.75 -2.54 0.036‡ -2.52 -3.12 -1.98 -2.32 0.101
(-5.21 to 0.33) (-6.62 to 0.31) (-6.62 to 0.04) (-6.27 to 0.31) (-4.80 to 0.07) (-6.62 to 0.33)
Role physical -1.18 -1.00 0.618 -0.81 -0.89 -1.79 -1.05 0.287
(-3.42 to 0.49) (-3.42 to 0.69) (-3.03 to 0.69) (-2.67 to 0.61) (-3.42 to 0.52) (-3.42 to 0.69)
Bodily pain -1.19 -1.36 0.551 -1.49 -1.27 -0.94 -1.31 0.499
(-3.01 to 0.10) (-3.65 to 1.02) (-3.65 to 1.02) (-3.63 to 1) (-2.50 to 0.10) (-3.65 to 1.02)
General health -0.51 -0.95 0.164 -0.87 -1.58 -0.57 -0.82 0.139
(-3.02 to 1.11) (-4.52 to 1.28) (-3.74 to 1.28) (-4.52 to 0.98) (-2.85 to 1.20) (-4.52 to 1.28)
Vitality -0.74 -1.32 0.041‡ -1.41 -1.57 -0.62 -1.15 0.063
(-2.57 to 1.59) (-4.03 to 1.63) (-4.03 to 1.63) (-3.97 to 0.86) (-3.31 to 1.09) (-4.03 to 1.63)
Social function -0.55 -1.19 0.141 -1.10 -1.99 -0.64 -1.00 0.131
(-3.77 to 2.92) (-4.70 to 3.18) (-4.04 to 3.18) (-4.70 to 0.65) (-4.39 to 0.77) (-4.70 to 3.18)
Role emotional -1.49 -1.42 0.866 -1.35 -1.00 -2.15 -1.44 0.371
(-4.01 to 0.49) (-4.01 to 0.55) (-4.01 to 0.55) (-3.02 to 0.55) (-4.01 to 0.50 (-4.01 to 0.55)
Mental health -0.33 -0.84 0.140 -0.85 -1.33 -0.29 -0.69 0.165
(-3.40 to 1.47) (-5.21 to 1.10) (-3.57 to 1.10) (-5.21 to 0.67) (-3.05 to 1.10) (-5.21 to 1.47)

MFA†
Overall 28.04 38.66 0.017‡ 38.55 41.00 36.00 35.75 0.185
(5 to 61) (4 to 89) (6 to 77) (4 to 89) (15 to 73) (4 to 89)
Move (lower limb) 32.61 50.66 0.003‡ 50.13 55.38 46.50 45.71 0.036‡
(0 to 70) (0 to 95) (0 to 90) (10 to 95) (10 to 85) (0 to 95)
Fine (upper limb) 5.59 8.67 0.557 8.64 15.38 0.00 7.82 0.349
(0 to 71.43) (0 to 100) (0 to 100) (0 to 100) (0 to 0) (0 to 100)
Home 30.92 46.08 0.025‡ 45.03 51.28 43.33 41.93 0.168
(0 to 88.89) (0 to 100) (0 to 100) (0 to 88.89) (0 to 88.89) (0 to 100)
Activities of daily living 6.28 19.03 0.001‡ 19.15 18.80 18.89 15.54 0.063
(0 to 38.89) (0 to 77.78) (0 to 77.78) (0 to 66.67) (0 to 61.11) (0 to 77.78)
Sleep 41.30 45.63 0.570 46.05 51.28 36.67 44.44 0.750
(0 to 100) (0 to 100) (0 to 100) (0 to 100) (0 to 100) (0 to 100)
Leisure 52.17 66.80 0.050‡ 69.74 67.31 55.00 62.80 0.130
(0 to 100) (0 to 100) (0 to 100) (0 to 100) (0 to 100) (0 to 100)
Relationships 20.43 30.82 0.162 31.58 28.46 31.00 27.98 0.566
(0 to 100) (0 to 100) (0 to 100) (0 to 90) (0 to 80) (0 to 100)
Cognition 18.48 31.56 0.127 33.55 32.69 22.50 27.98 0.456
(0 to 100) (0 to 100) (0 to 100) (0 to 100) (0 to 100) (0 to 100)
Emotional 43.72 45.54 0.756 45.03 48.29 43.89 45.04 0.954
(11.11 to 88.89) (0 to 94.44) (0 to 94.44) (5.56 to 94.44) (11.11 to 77.78) (0 to 94.44)
Job 50.00 59.84 0.364 56.58 55.77 77.50 57.14 0.436
(0 to 100) (0 to 100) (0 to 100) (0 to 100) (0 to 100) (0 to 100)
* PW, posterior wall; TV, transverse; PC, posterior column; T, associated T type
† MFA, musculoskeletal functional assessment
‡ significant, p < 0.05

tions regarding smoking. The prevalence of smoking was
similar (mean 65%) across the types of fracture (p = 0.94).
Of 55 smokers, 27 developed heterotopic ossification
(49%) compared with only six of 29 non-smokers (21%).
Thus the risk of developing this complication was 3.7 times
(risk ratio calculation; 95% confidence interval (CI) 1.3 to
10.5) greater for smokers than for non-smokers.
Functional outcome. Of the 128 patients, 84 (66%) com-
pleted the SF-36 and MFA questionnaires at a mean of 4.2
years (1 to 17) after operation. Of the remaining 44
patients, one had died, six refused to participate and the
other 37 could not be located. The patients who completed
the functional questionnaires had a similar (p > 0.05) age,
gender and associated injury profile as those who did not,
with the exception of significantly more associated acetab-
ular fractures (p = 0.031; Mann-Whitney U test) in the
group which completed the questionnaires (Table III). Thus
scores should be considered in their respective fracture type
as they are reported rather than in the aggregate (Table IV).
The mean overall MFA score was 35.75 (4 to 89).
Patients had particularly poor scores for function of the
lower limb (45.71 (0 to 95), sleep 44.44 (0 to 100), leisure
activities 62.80 (0 to 100), activities around the home
41.93 (0 to 100) and emotional issues 45.04 (0 to 94.44),
but they were significantly better for simple fractures of the
posterior wall compared with those with associated frac-

THE JOURNAL OF BONE AND JOINT SURGERY

 DETERMINANTS OF FUNCTIONAL OUTCOME AFTER ACETABULAR FRACTURES INVOLVING THE POSTERIOR WALL
tures in most domains (Table IV). Patients with associated
fractures had significantly worse MFA scores (mean 55.38;
10 to 95) for function of the lower limb than all other frac-
ture patterns (p = 0.036; Table IV).
Because age- and gender-specific population norms are
available for the SF-36, we were able to compute standard-
ised z-scores for each SF-36 domain and summary score.
The mean scores for all SF-36 domains were significantly
worse than the mean scores for the age- and gender-
matched normal population (p < 0.001; Students t-test).
After adjusting for age, gender and length of follow-up,
the functional outcome as measured by the MFA and SF-36
was consistently worse for patients with associated fracture
patterns compared with those with simple fractures of the
posterior wall and for those with radiological osteoarthritis
(p < 0.05). As noted above, the development of arthritis
was significantly associated with marginal impaction and
residual displacement (p < 0.05). There was also a trend for
a worse functional outcome in multiple MFA and SF-36
domains for patients with marginal impaction (0.05 > p <
0.08), and for those with residual displacement of more
than 2 mm (0.05 > p < 0.08). Age, gender, femoral-head
lesions, posterior-wall comminution, the presence of
heterotopic ossification and additional injuries involving
the limbs, head, spine or abdomen were not consistently
associated with functional outcome (p > 0.10; Students t-
test).
Functional outcome scores were available for 11 of the
16 patients (69%) who had undergone THR. Function as
measured using the MFA and SF-36 remained seriously
impaired in these patients, despite uncomplicated total
joint replacement. The overall MFA score of 32.55 (11 to
61) for patients who had required THR was similar to that
of those who had not undergone this procedure (p = 0.61).
There was no statistically significant difference for any of
the SF-36 or MFA domain scores between patients who had
undergone THR and those who had not (p > 0.05).
Of the 84 patients surveyed, 37 (44%) were still receiv-
ing disability payments, and 31 (37%) were involved in
legal action at the time of follow-up. These individuals were
more likely than other patients to have sustained a head
injury (p = 0.06) and to have an associated type of poster-
ior-wall fracture (p = 0.05). Those involved in legal action
or on disability had significantly worse MFA and SF-36
scores than patients not in this situation (p < 0.01), inde-
pendent of the presence of head injury and type of fracture.
Discussion
We found that patients with simple and associated fractures
of the posterior wall of the acetabulum had severe func-
tional deficits as measured by the SF-36 and MFA at a mean
of four years after injury. Functional outcome was signifi-
cantly worse in patients with radiological arthritis and in
those with associated posterior wall and posterior-column
fractures. The latter group had the worst function scores
for the lower limb of any type of fracture. They were asso-
VOL. 88-B, No. 6, JUNE 2006
781
ciated with marginal impaction in 16 cases (72.7%) and
with inability to achieve reduction within 2 mm in seven
cases (31.8%). There were 14 patients (63.6%) with asso-
ciated posterior-wall/posterior-column fractures who devel-
oped radiological arthritis during the period of study.
Marginal impaction was independently associated with the
development of arthritis, while comminution of the poste-
rior wall was associated with a higher likelihood of residual
displacement, which in turn was associated with arthritis.
Marginal impaction, along with older age and arthritis, was
directly and independently associated with the need for
THR during the period of study. These findings are consis-
tent with those of Wolinsky et al22 who found that the pres-
ence of marginal impaction, intra-articular fragments, a
fracture of the posterior wall and an older age correlated
highly with the need for THR after operative fixation of
fractures of the acetabulum.
Borrelli et al10 distinguished between patients with a
mean MFA score of 7 and those with a mean score of 32 in
a detailed review of 15 patients, including nine with frac-
tures involving the posterior wall. Those patients with
higher scores had weaker hip flexion and extension, but the
small sample size did not reveal any clear relationship
between the type of fracture, reduction and MFA score. In
a review of a heterogeneous group of 150 patients after
acetabular fracture, Moed et al11 noted a mean overall
MFA score of 24.9. Scores for fractures of the posterior
wall were not reported separately, and no association was
found between the MFA score and the type of fracture, gen-
der, age and the presence of associated systemic injuries. In
a previous review of 94 patients with isolated fractures of
the posterior wall, Moed et al7 found that a poor clinical
outcome was more common in patients over 55 years of age
and in those with intra-articular comminution, delay in
reduction of the hip of more than 12 hours and avascular
necrosis, but marginal impaction was of borderline inde-
pendent statistical significance with respect to the clinical
outcome. Saterbak et al8 noted deficits of comparable sever-
ity to our findings in a group of 42 patients using the MFA
score. They found that early failure was more common in
patients with comminution of the posterior wall (three or
more fragments) and extension of the fracture into the
dome of the acetabulum and/or the ischium. Murphy et al23
noted that associated fracture patterns and older age were
more likely to be associated with suboptimal reduction of
the fracture and worse Merle d’Aubigné scores. Our data
showed a clear and strong association between the type of
fracture and the functional outcome measured using vali-
dated instruments (MFA and SF-36). We could not demon-
strate an association between age and the quality of post-
operative reduction or between age and function. However,
our population was relatively young, with a mean age of
41.6 years (18 to 75). It is possible that bone quality, rather
than age, is the relevant factor but this has not been quan-
tified in our study or to our knowledge in any previous
work.
782 H. J. KREDER, N. ROZEN, C. M. BORKHOFF, Y. G. LAFLAMME, M. D. MCKEE, E. H. SCHEMITSCH, D. J. G. STEPHEN
Matta2 found that the presence of a lesion of the femoral
head resulted in a worse clinical outcome, but marginal
impaction, intra-articular fragments, initial displacement of
the fragment and associated injuries did not. We were also
unable to demonstrate a relationship between associated
injuries or intra-articular fragments with outcome but we
found that marginal impaction was associated with the
development of arthritis and also the need for THR.
Lesions of the femoral head were of no independent signif-
icance with respect to any of our end-points. We did not
measure initial displacement of fragments.
The clinical outcome after posterior-wall fractures can be
poor despite anatomical reduction.1-9 Matta2 reported 22
fractures of the posterior wall in a series of 262 acetabular
fractures. Despite anatomical reduction as determined on
plain radiographs in all 22 cases, a good or excellent clini-
cal result was obtained in only 68%. He suggested that
plain radiographs may not demonstrate imperfections in
the articular surface. Moed et al24 have shown that CT
reveals more incongruity and correlates better with out-
come than plain radiography. Unfortunately, routine post-
operative CT was not available for all of our cases, but we
did incorporate the CT findings in the 67 patients in whom
they were available. Our raters found it difficult to separate
gap from step deformity and the decision was taken to con-
sider the maximum displacement, whether this was a pure
gap or a gap with step deformity involving the fragments of
the posterior wall or an associated fracture of the articular
column. We felt that this was a conservative approach to
the reporting of residual displacement, but if gap is less
important than step, this approach may dilute the relation-
ship of displacement with outcome. Finally, we may have
overlooked misalignment or ‘hinging’ of the fracture which
does not result in step or gap deformity since this is difficult
to assess by plain radiography.
Individuals with fractures of the acetabulum involving
the posterior wall have profound functional deficits com-
pared with the normal population. Fractures of the poster-
ior wall with associated fractures of the posterior column
appear to have a particularly poor prognosis. Anatomical
reduction alone is not sufficient to restore function to nor-
mal, but residual displacement above 2 mm along with
marginal impaction is associated with the development of
early radiological arthritis, which in turn is related to poor
function and the need for THR. Although late THR did not
restore function to the age- and gender-matched control
groups, it may be appropriate to consider immediate THR
for older patients with severe marginal impaction and com-
minution of the posterior wall given the high likelihood of
their requiring THR soon after fixation of the fracture.25-27
The association of smoking with an apparent increased risk
of heterotopic ossification requires further investigation,
although the presence of this complication was not inde-
pendently associated with poor outcome.
Dr Kreder is the recipient of a career development award from the Orthopaedic
Research and Education Foundation.
This project was supported by a grant from the AO-ASIF Foundation.
No benefits in any form have been received or will be received from a com-
mercial party related directly or indirectly to the subject of this article.
References
1. Letournel E, Judet R. Fractures of the acetabulum. Second ed. Berlin: Springer-Ver-
lag, 1993.
2. Matta JM. Fractures of the acetabulum: accuracy of reduction and clinical results in
patients managed operatively within three weeks after the injury. J Bone Joint Surg
[Am] 1996;78-A:1632-45.
3. Brumback RJ, Holt ES, McBride MS, et al. Acetabular depression fracture
accompanying posterior fracture dislocation of the hip. J Orthop Trauma 1990;4:
42-8.
4. Matta JM, Merritt PO. Displaced acetabular fractures. Clin Orthop 1998:230:83-97.
5. Mayo KA. Open reduction and internal fixation of fractures of the acetabulum:
results in 163 fractures. Clin Orthop 1994;305:31-7.
6. Moed BR, Carr SE, Watson JT. Open reduction and internal fixation of posterior
wall fractures of the acetabulum. Clin Orthop 2000;377:56-67.
7. Moed BR, WillsonCarr SE, Watson JT. Results of operative treatment of fractures
of the posterior wall of the acetabulum. J Bone Joint Surg [Am] 2002;84-A:752-8.
8. Saterbak AM, Marsh JL, Nepola JV, Brandser EA, Turbett T. Clinical failure
after posterior wall acetabular fractures: the influence of initial fracture patterns.
J Orthop Trauma 2000;14:230-7.
9. Wright R, Barrett K, Christie MJ, Johnson KD. Acetabular fractures: long-term
follow-up of open reduction and internal fixation. J Orthop Trauma 1994;8:397-403.
10. Borrelli J Jr, Goldfarb C, Ricci W, Wagner JM, Engsberg JR. Functional out-
come after isolated acetabular fractures. J Orthop Trauma 2002;16:73-81.
11. Moed BR, Yu PH, Gruson KI. Functional outcomes of acetabular fractures. J Bone
Joint Surg [Am] 2003;85-A:1879-83.
12. Tornetta P 3rd. Non-operative management of acetabular fractures: the use of
dynamic stress views. J Bone Joint Surg [Br] 1999;81-B:67-70.
13. Tornetta P 3rd. Displaced acetabular fractures: indications for operative and non-
operative management. J Am Acad Orthop Surg 2001;9:18-28.
14. Engelberg R, Martin DP, Agel J, et al. Musculoskeletal function assessment
instrument: criterion and construct validity. J Orthop Res 1996;14:182-92.
15. Engelberg R, Martin DP, Agel J, Swiontkowski MF. Musculoskeletal function
assessment: reference values for patient and non-patient samples. J Orthop Res
1999;17:101-9.
16. Ware JE Jr, Sherbourne CD. The MOS 36-item short-form health survey (SF-36): I.
Conceptual framework and item selection. Med Care 1992;30:473-83.
17. Letournel E. Acetabulum fractures: classification and management. Clin Orthop
1980;151:81-106.
18. Brooker AF, Bowerman JW, Robinson RA, Riley LH Jr. Ectopic ossification fol-
lowing total hip replacement: incidence and a method of classification. J Bone Joint
Surg [Am] 1973;55-A:1629-32.
19. Moed BR, Smith ST. Three-view radiographic assessment of heterotopic ossifica-
tion after acetabular fracture surgery. J Orthop Trauma 1996;10:93-8.
20. Kellgren JH, Lawrence JS. Radiological assessment of osteo-arthritis. Ann Rheum
Dis 1957;16:494-502.
21. Gunther KP, Sun Y. Reliability of radiographic assessment in hip and knee osteo-
arthritis. Osteoarthritis Cartilage 1999;7:239-46.
22. Wolinsky PR, Davison BL, Shyr Y, Talwalker V, Johnson KD. Predictors of total
hip arthroplasty in patients following open reduction and internal fixation of acetab-
ular fractures. Procs 12th annual Orthopaedic Trauma Association Meeting 1996;9:29
(abstract).
23. Murphy D, Kaliszer M, Rice J, McElwain JP. Outcome after acetabular fracture:
prognostic factors and their inter-relationships. Injury 2003;34:512-17.
24. Moed BR, Carr SE, Gruson KI, Watson JT, Craig JG. Computed tomographic
assessment of fractures of the posterior wall of the acetabulum after operative treat-
ment. J Bone Joint Surg [Am] 2003;85-A:612-22.
25. McKee MD, Garay ME, Schemitsch EH, Kreder HJ, Stephen DJ. Irreducible
fracture-dislocation of the hip: a severe injury with a poor prognosis. J Orthop Trauma
1998;12:223-9.
26. Mears DC, Velyvis JH. Acute total hip arthroplasty for selected displaced acetabu-
lar fractures: two to twelve-year results. J Bone Joint Surg [Am] 2002;84-A:1-9.
27. Mears DC, Velyvis JH, Chang CP. Displaced acetabular fractures managed opera-
tively: indicators of outcome. Clin Orthop 2003;407:173-86.
THE JOURNAL OF BONE AND JOINT SURGERY