e126(1)

C OPYRIGHT Ó 2017 BY T HE J OURNAL OF B ONE AND J OINT S URGERY, I NCORPORATED

Instability of the Hip Joint After Posterior Acetabular

Wall Fracture
Independent Risk Factors Remain Elusive
Jay H. Patel, MD, and Berton R. Moed, MD

Investigation performed at the Department of Orthopaedic Surgery, Saint Louis University School of Medicine, St. Louis, Missouri

Background: Exact determinants of hip instability have not been established for fractures of the posterior wall of the
acetabulum involving £50% of the wall. Therefore, examination of the hip under anesthesia (EUA) is routinely performed.
Recently, the superior exit point of the fracture has been reported to be an important identifiable risk factor. Pre-existing
adult hip dysplasia (developmental dysplasia of the hip [DDH]) is thought to have a similar role. The purpose of this study
was to determine if any known radiographic measurements and signs associated with DDH, or any fracture character-
istics, are independent risk factors for hip instability after fracture of the posterior wall of the acetabulum.
Methods: All patients with a posterior acetabular wall fracture (OTA 62.A1) treated at our institution between 2004 and
2015 were considered for the study. Inclusion criteria were an age of ‡18 years, an isolated posterior acetabular wall
fracture involving £50% of the acetabular wall, adequate imaging, and documented EUA results. Evaluated variables
included fracture fragment size, superior exit point of the fracture, center-edge angle, acetabular index, Tönnis angle,
lateralized head sign, crossover sign, posterior wall sign, ischial spine sign, and hip version. Data were examined using
univariate testing, followed by a multivariate logistic regression analysis.
Results: Sixty-eight patients met all of the inclusion criteria. Univariate analyses identified the posterior wall sign (p =
0.033), ischial spine sign (p = 0.030), and proximity of the superior exit point of the fracture to the acetabular dome (p =
0.044) as having a significant association with hip instability. However, multivariate logistic regression modeling revealed
that none of these factors were significant independent risk factors.
Conclusions: Consistent with previous studies, univariate analyses identified certain radiographic findings as significant
risk factors for hip instability in the setting of a fracture of the posterior wall of the acetabulum. However, subsequent
multivariate logistic regression modeling showed that no studied variable was an independent risk factor. Our results
indicate that important factors leading to hip instability are yet to be identified or the contributions of the measured
variables are relatively small. Therefore, EUA should remain the main clinical determinant of hip stability status.
Level of Evidence: Prognostic Level IV. See Instructions for Authors for a complete description of levels of evidence.

F
ractures of the posterior wall of the acetabulum are the
most common acetabular fracture type, accounting for
approximately one-quarter to one-third of all acetabular
fractures1-3. Computed tomography (CT) is routinely used for
the evaluation of these injuries1-6. Using CT, if £50% of the
wall is fractured, the stability status of the hip is uncertain4,5.
Therefore, patients are routinely taken to the operating room
for examination under anesthesia (EUA) with the use of dy-
namic stress views to determine further treatment4-6. A finding
of stability on EUA is predictive of hip joint congruity, excellent
radiographic outcome, and a good to excellent early clinical
outcome after nonoperative treatment of the fracture7. Frac-
tures with instability of the hip joint on EUA, or those with
subluxation of the femoral head at presentation, are treated
with open reduction and internal fixation (ORIF)7.
Adult developmental dysplasia of the hip (DDH) and
femoroacetabular impingement (FAI) have been found to have a
role in posterior hip dislocation because of disease-specific
morphologic features8,9. The association of posterior instability
of the hip joint with femoral shaft version, acetabular version,
and pre-existing pathologic conditions of the hip has been
studied, with findings indicating that reduced femoral shaft

Disclosure: No external funding was received for this study. The Disclosure of Potential Conflicts of Interest forms are provided with the online version of
the article (http://links.lww.com/JBJS/E466).

J Bone Joint Surg Am. 2017;99:e126(1-7) d http://dx.doi.org/10.2106/JBJS.16.01427

 e126(2)
TH E JO U R NA L O F B O N E & JO I N T SU RG E RY J B J S . O RG
d
V O L U M E 99-A N U M B E R 23 D E C E M B E R 6, 2 017
d d
anteversion and intra-articular labral and chondral pathologies
play a role in posterior hip dislocation2,3,8,10. FAI morphologic
features such as a cam deformity or a retroverted acetabulum
also may play a role in traumatic posterior dislocations2.
Therefore, there may be a relationship between hip dysplasia and
instability after a fracture of the posterior wall of the acetabulum.
A variety of radiographic markers and signs are commonly used
to diagnose hip dysplasia. To our knowledge, no statistical cor-
relations or associations have been established between these
markers of dysplasia and hip stability status after a posterior
acetabular wall fracture. Recently, however, the proximity of the
superior exit point of the fracture to the acetabular dome was
reported to be an important identifiable risk factor11. Although
this is not a marker of dysplasia, it is a readily measurable
fracture variable and was of particular interest in our study.
The purpose of the current study was to determine if any
known radiographic measurements and signs associated with
pre-existing DDH or fracture characteristics are independent
risk factors for hip instability after fracture of the posterior wall
of the acetabulum. We hypothesized that independent risk
factors would not be identified.
Materials and Methods
A pproval from our institutional review board was obtained
for this study. All patients with a fracture of the posterior
wall of the acetabulum (OTA/AO 62.A1)12 treated at our in-
stitution between 2004 and 2015 were considered for inclusion.
This time frame was selected to coincide with the im-
plementation of our current picture archiving and communi-
cation system (PACS) for the storage of radiographs. Our
orthopaedic traumatologists routinely treat fractures involving
>50% of the posterior acetabular wall using ORIF without a
preoperative EUA. Therefore, our inclusion criteria were a
patient age of ‡18 years, a unilateral and isolated posterior
acetabular wall fracture involving £50% of the posterior wall of
the acetabulum, adequate radiographs and CT scans available
for review, and a documented EUA, an examination that has
been performed at our institution according to a standardized
method since 2003. Consistent with previously reported
studies, the exclusion criteria included skeletal immaturity;
other injuries involving the affected hip joint or proximal part
of the ipsilateral femur; previous injury to, or surgery on, the
affected hip before the index acetabular fracture; bilateral ac-
etabular fracture; an acetabular fracture that required open
reduction for other reasons (i.e., nonconcentric closed reduc-
tion of a hip dislocation or the presence of a large intra-
articular free fragment); concurrent pelvic ring fracture; and
the unavailability of CT scans or radiographs5,7. Current pro-
cedural terminology (CPT) codes were used to identify patients
who underwent EUA of the hip for a posterior acetabular wall
fracture. Chart review was performed to obtain patient de-
mographics, with recorded patient demographics including
patient sex and age, side of injury, and mechanism of injury. All
patients had anteroposterior radiographs of the pelvis obtained
at the time of initial presentation and before EUA, which
was used to determine hip status at presentation (located or
I N S TA B I L I T Y O F T H E H I P J O I N T A F T E R P O S T E R I O R A C E TA B U L A R
WA L L F R A C T U R E
dislocated). The best-quality anteroposterior radiograph ob-
tained prior to or following EUA was used in the evaluation of
the following data: center-edge angle, acetabular index, Tönnis
angle, lateralized head sign, crossover sign, posterior wall sign,
and ischial spine sign8,13. In addition, all patients had Judet
oblique pelvic radiographs obtained before EUA to provide
further evaluation of hip joint congruency as well as infor-
mation regarding the posterior wall fracture characteristics and
displacement. The contralateral hip was used for the evaluation
of the posterior wall sign when the size, displacement, or
location of the ipsilateral posterior wall fragment obscured
adequate evaluation. If a hip dislocation was present, only post-
reduction imaging was used for data gathering.
Each patient had a high-quality CT scan (£2-mm slice
thickness with bone windows) performed before the EUA but
after any associated hip dislocation had been reduced. Analysis of
these CT scans provided details of hip version, fracture fragment
size, and superior exit point of the fracture. CT scans were not
used in the evaluation of other dysplasia markers or signs8,13.
Acetabular version was determined by the method of Reynolds
et al., using a reference line drawn between the center of the pubic
symphysis and the center of the sacrum14. The angle formed by a
line parallel to this reference line and a line drawn between the
posterior and anterior acetabular rims at the level of maximum
femoral head diameter was recorded as acetabular version
(Fig. 1). Fracture fragment size was determined by the method of
Moed et al.15. Using this method, measurements were made at the
level of the largest posterior wall deficit of the fractured acetab-
ulum to determine the approximate medial-lateral dimension
(depth) of the fractured segment, and compared with a matched,
2-dimensional CT section from the normal, intact contralateral
acetabulum. Fragment size, expressed as a percentage, was
Fig. 1
The method of measuring acetabular version is shown. A reference line was
drawn between the center of the pubic symphysis and the center of the
sacrum (A). The angle (B) formed by a line parallel to this reference line (C)
and a line drawn between the posterior and anterior acetabular rims (D) at
the level of maximum femoral head diameter was recorded as acetabular
version.
 e126(3)
TH E JO U R NA L O F B O N E & JO I N T SU RG E RY J B J S . O RG
d
V O L U M E 99-A N U M B E R 23 D E C E M B E R 6, 2 017
d d
calculated from the ratio of the measured depth of the fractured
segment to the depth of the intact, matched contralateral ace-
tabulum (Fig. 2). The superior exit point was determined by the
method utilized by Firoozabadi et al.11. The superior exit point of
the fracture line was located where the fracture exited the ace-
tabulum’s articular surface. The superior exit point only refers to
the intra-articular fracture exit point, as the extra-articular
component can exit above or below the joint involvement. A
fracture exiting the acetabular dome was labeled as having a 0-
mm exit point11. The exit point in millimeters was identified
primarily using axial images. Sagittal and coronal slices of the CT
scan were used to cross-reference this value, when needed. All
evaluated CT scans were obtained after any associated hip dis-
location had been reduced.
During the time frame of the study, 79 patients were
identified as having undergone an EUA for a fracture of the
posterior wall of the acetabulum. Eleven were excluded ac-
cording to the selection criteria. Of these 11 patients, 7 had an
ipsilateral femoral fracture, 1 had undergone a contralateral total
hip arthroplasty, and 3 had missing radiographs or CT imaging,
leaving a total of 68 patients with an isolated posterior acetabular
wall fracture (OTA 62.A1) for study. Ipsilateral limb injuries
were excluded, as these injuries typically make EUA difficult,
thereby possibly compromising EUA results and quality.
The mean patient age was 34.0 years (range, 18 to 79
years; median, 29.5 years). Forty-six (67.6%) of the patients
were male and 22 (32.4%) were female. All patient injuries
were due to a high-energy mechanism; for the vast majority,
this was a motor-vehicle collision.
Fig. 2
The method of measuring posterior wall fragment size is shown. Fig. 2-A Fractured acetabulum. The approximate medial-lateral dimension (depth) of the
fractured segment was determined at the level of the largest posterior wall deficit. Fig. 2-B Contralateral, intact acetabulum. The percentage of fragment
size was calculated from the ratio of the measured depth of the fractured segment to the depth of the intact acetabulum measured to the medial extent of the
quadrilateral plate.
I N S TA B I L I T Y O F T H E H I P J O I N T A F T E R P O S T E R I O R A C E TA B U L A R
WA L L F R A C T U R E
For data analyses, all data were recorded as categorical
variables, with the exception of acetabular version, fracture
fragment size, and the superior exit point of the fracture, which
were recorded as continuous variables. The center-edge angle,
Tönnis angle, and acetabular index were classified categorically as
“dysplastic” or “normal” on the basis of the following definitions
of dysplasia: center-edge angle of <25°, Tönnis angle of >10°, and
acetabular index of >40°13,16. The femoral head was described as
“located” or “dislocated.” All classic radiographic dysplasia signs,
including the posterior wall, ischial spine, lateralized head, and
crossover signs, were described as “present” or “absent” as ob-
served with the use of traditional anteroposterior pelvic radio-
graphs8,13,15. The posterior wall sign is present when the rim of the
posterior wall is medial to the center of the femoral head15. The
ischial spine sign is present when the ischial spine is shown to
protrude into the true pelvis8. The lateralized head sign is present
when the medial aspect of the femoral head is >10 mm from the
ilioischial line. The crossover sign is present if the separate lines of
the posterior wall and anterior wall converge, making a figure of
8 on anteroposterior imaging13. For each patient, the findings for
these variables were compared with the recorded EUA result,
which identified the hip joint as being “stable” or “unstable.”
Statistical Analysis
The final data were analyzed by a PhD-level statistician using the
appropriate parametric and nonparametric tests, depending on
the variable type. Initially, all variables were evaluated with uni-
variate analyses. The Student t test was used for the continuous
variables of acetabular version, fracture fragment size, and
 e126(4)
TH E JO U R NA L O F B O N E & JO I N T SU RG E RY J B J S . O RG
d
I N S TA B I L I T Y O F T H E H I P J O I N T A F T E R P O S T E R I O R A C E TA B U L A R
V O L U M E 99-A N U M B E R 23 D E C E M B E R 6, 2 017
d d
WA L L F R A C T U R E

TABLE I Significant Results from Univariate Analysis

Estimated
Variable Statistical Test OR (95% CI)* T Value (95% CI)* P Value Power

Superior exit point Student t test NA 2.05 (0.08-6.11) 0.044 60%

of fracture
Posterior wall sign Pearson chi-square test 9 (0.86-94.24) NA 0.033 87%
Ischial spine sign Pearson chi-square test 9 (0.85-94.00) NA 0.030 85%

*OR = odds ratio, CI = confidence interval, and NA = not applicable.

superior exit point of the fracture, while a Pearson chi-square test
was used for all other variables. Subsequently, a forward multi-
variate logistic regression model was built using only the variables
found to be significant in the univariate analyses and applied to
describe the association between the measured risk factors and
hip instability. The level of significance was defined as p < 0.05.
Results
ranged from 14% to 48% (mean, 34.8%; 95% CI, 29.1% to
40.5%). Among the evaluated fracture characteristics, only the
proximity of the superior exit point of the fracture to the ac-
etabular dome (overall range, 0 to 30 mm; mean, 4.55 mm) was
a significant indicator of hip instability in the initial univariate
analysis (p = 0.044, t test) (Table I). Among the stable hips, the
mean distance of the exit point from the dome was 5.57 mm
(range, 0 to 30 mm; 95% CI, 3.66 to 7.49 mm). Among the

O f the 68 patients evaluated, 33 had left-side injuries and 35

had right-side injuries (p = 0.671, chi-square test). After
EUA, 21 (30.9%) of the 68 patients were determined to have an
unstable hip requiring operative intervention, and 47 (69.1%)
of the patients were determined to have a stable hip and were
treated nonoperatively. Among the 21 patients with an unstable
hip, 8 (38.1%) had a dislocation at the time of presentation,
and among the 47 patients with a stable hip, 9 (19.1%) had a
dislocation at presentation. However, univariate analysis indi-
cated that hip dislocation was not a significant indicator of
instability (p = 0.096, chi-square test). In addition, fracture
fragment size (overall range, 12% to 48%; mean, 31.0%) had a
nonsignificant effect on hip instability (p = 0.486, t test).
Within the stable group, the fracture fragment size ranged from
12% to 47% (mean, 27.1%; 95% confidence interval [CI],
22.7% to 31.3%). Within the unstable group, the fragment size
unstable hips, the mean distance from the dome was 2.48 mm
(range, 0 to 12 mm; 95% CI, 0.90 to 4.05 mm). However,
logistic regression analysis revealed that this variable was not a
significant indicator of instability (p = 0.10).
On initial univariate analysis of all recorded radiographic
markers of hip dysplasia, the posterior wall sign (p = 0.033, chi-
square test) and the ischial spine sign (p = 0.030, chi-square test)
were shown to have a significant association with hip instability
following a posterior acetabular wall fracture (Table II). All other
hip dysplasia variables had nonsignificant p values, including the
center-edge angle, Tönnis angle, acetabular index, crossover
sign, and lateralized head sign (p = 0.647, 0.170, 0.391, 0.133,
and 0.236, respectively; chi-square test). Measured acetabular
version was also shown to be nonsignificant as an indicator of
hip instability (p = 0.092, t test). Acetabular version in the stable
hips ranged from 11° to 32° (mean, 19.2°; 95% CI, 17.8° to

TABLE II Hip Dysplasia Categorical Variables Studied

No. (%) of
Subjects with No. (%) of No. (%) of
Dysplastic Hip Presence of Subjects with Subjects with an
Variable Dysplasia Sign* a Stable Hip† Unstable Hip† P Value‡

Posterior wall sign 4 (5.9) 1 (25) 3 (75) 0.033

Ischial spine sign 2 (2.9) 0 (0) 2 (100) 0.030
Crossover sign 3 (4.4) 1 (33.3) 2 (66.7) 0.133
Lateralized head sign 3 (4.4) 3 (100) 0 (0) 0.236
Center-edge angle 5 (7.4) 3 (60) 2 (40) 0.647
Acetabular index 18 (26.5) 11 (61.1) 7 (38.9) 0.391
Tönnis angle 3 (4.4) 1 (33.3) 2 (66.7) 0.170

*The percentage is of the total number of patients (n = 68). †The percentage is of the number of hips in which the dysplasia sign was present.
‡Derived from univariate analysis.

TH E JO U R NA L O F B O N E & JO I N T SU RG E RY J B J S . O RG
V O L U M E 99-A N U M B E R 23 D E C E M B E R 6, 2 017
d d
TABLE III Results from Logistic Regression Analysis
Variable OR (95% CI)*
Superior exit point 0.85 (0.71-1.03)
of fracture
Posterior wall sign 0.15 (0.01-3.41)
Ischial spine sign 0.29 (0.20-0.46)
*OR = odds ratio, and CI = confidence interval.
20.6°), and in the unstable hips, acetabular version ranged from
6° to 25° (mean, 17.0°; 95% CI, 14.6° to 19.4°). Following
multivariate logistic regression analysis, none of the previously
significant hip dysplasia variables were found to be independent
indicators of hip instability (Table III).
The multivariate logistic regression model had a
Nagelkerke R2 value of 0.24 and a Hosmer-Lemeshow value of
0.71. Only a single case on the case-wise list was >2.5, and no
consistent pattern for the 3 variables that were significant in the
univariate analyses was found among the unstable hips. These
tests, and the case-wise list diagnostic tests, used to assess the
regression, were within satisfactory limits. Therefore, the
multivariate logistic regression model, as applied to our sig-
nificant variables on univariate analyses (ischial spine sign,
posterior wall sign, and superior exit point of the posterior wall
fracture), showed that none of the studied variables were in-
dependent risk factors for hip instability after a fracture of the
posterior wall of the acetabulum (Table III).
Discussion
H ip instability demonstrated on EUA, hip incongruency due
to intra-articular fragments, and fractures involving >50%
of the posterior wall are considered indications for operative
intervention for fractures of the posterior wall of the acetabu-
lum2,15,17,18. However, instability of the hip joint after a posterior
wall fracture often is difficult to predict, and an EUA is generally
needed for treatment decision-making4-6,11,15. Once thought to be
a marker of an unstable hip after a posterior acetabular wall
fracture, a history of dislocation of the hip joint at the time of
injury has been shown to not be a reliable indicator5,6,17,18. Our
objective was to demonstrate if other described fracture char-
acteristics or established measurements of hip dysplasia in the
presence of a posterior acetabular wall fracture are independent
indicators of hip joint instability. If so, this information would be
helpful in treatment decision-making by directing appropriate
care without the need for an EUA.
Unfortunately, none of our studied variables proved to
be significant independent indicators of hip instability.
Firoozabadi et al. concluded that, in unstable hips, the exit point
of a posterior acetabular wall fracture is closer to the dome11. In
their study using univariate analyses, they determined a mean
cutoff value of 5 mm for instability; the superior exit point of
the fracture was a mean of 5 mm from the dome in the unstable
hips and 9.5 mm in the stable hips. Initially, and consistent with
e126(5)
d
I N S TA B I L I T Y O F T H E H I P J O I N T A F T E R P O S T E R I O R A C E TA B U L A R
WA L L F R A C T U R E
the findings of the previous study11, we found that the superior
exit point of the fracture was significantly associated with in-
P Value stability on univariate analysis (t test). Intuitively, it would
seem that a more superior fracture exit point would increase
0.10 the risk of hip instability. On further analysis, however, this
variable became nonsignificant as an independent risk factor,
0.23 demonstrating no significant effect on hip instability for our
0.38 tested range of fracture fragment size (£50%). This finding
could be due to the fact that, for the majority of the fractures
that we evaluated, fragment size ranged between 20% and 50%,
a range that is typically indeterminate in defining hip stability15.
In addition, previous investigators have shown that even frac-
tures involving <20% of the posterior wall can be accompanied
by hip instability4,11.
It is hypothesized that, in dysplastic hips, altered biome-
chanics and altered force vectors may result in hip instability
outside the historical range19. As in our study, Firoozabadi et al.
did not find any association between hip instability and the
center-edge angle, acetabulum version, or posterior acetabular
wall fracture fragment size using univariate analyses11. We not
only studied the same variables but also examined the acetabular
index, Tönnis angle, posterior wall sign, ischial spine sign, lat-
eralized head sign, and crossover sign. Our data initially sug-
gested that 2 radiographic measurements used to diagnose
DDH, both of which were not studied by Firoozabadi et al., were
associated with instability of the hip after a posterior acetabular
wall fracture: the presence of the posterior wall sign and ischial
spine sign. However, logistic regression analysis showed that
these variables were not significant independent risk factors for
hip instability after a fracture of the posterior wall of the
acetabulum.
Our study had limitations. One limitation was that all of
the measurements were made by a single individual, introducing
the possibility of detection bias. However, this investigator was
not aware of the EUA results at the time of performing the
measurements, thereby limiting this possibility. This study was
not designed to evaluate interobserver reliability, and, therefore,
this was an accepted limitation in study design and data gath-
ering. In addition, the studied measurements require adequate
anteroposterior imaging of the pelvis and can be difficult to
evaluate in trauma patients. At our center, multiple preoperative
radiographs are often obtained as well as additional studies just
prior to hospital discharge. Some pelvic radiographs had varying
amounts of flexion/extension of the pelvis; however, the vast
majority were centered on the symphysis using an acceptable
technique. Acceptable technique for an anteroposterior pelvic
radiograph is defined as positioning of both lower extremities in
15° of internal rotation, with the tip of the coccyx and the pubic
symphysis 1 to 3 cm apart to provide minimal tilt, and the
obturator foramina symmetric in appearance when evaluated en
face to provide no rotation13. Our institution treats a high vol-
ume of patients with acetabular fractures and has protocols in
place to obtain consistent anteroposterior radiographs of the
pelvis. Only adequate preoperative and, if applicable, post-
reduction imaging of the hip using the above definition was used
in data gathering. For the specific measurement of the posterior
 e126(6)
TH E JO U R NA L O F B O N E & JO I N T SU RG E RY J B J S . O RG
d
V O L U M E 99-A N U M B E R 23 D E C E M B E R 6, 2 017
d d
wall sign, approximately one-quarter of hips had posterior ac-
etabular wall fracture fragments that were sufficiently large or
displaced to warrant evaluation of the contralateral hip as a
substitute for the ipsilateral hip. Therefore, the assumption was
made that hip dysplasia, if present, was bilateral. This assump-
tion was a potential study limitation. However, a number of
studies have provided widely varying data regarding the rate of
bilateral hip dysplasia, ranging from 4% to 84%20,21. Because the
range is so large, we decided that this was an acceptable limita-
tion and justified the analysis of the posterior wall sign using the
contralateral hip, when necessary.
Another potential limitation was the relatively small
sample size for this study. However, the overlap in the data of
these 2 groups (stable and unstable hips) indicates that a very
large sample size is required to ultimately test if there is any
difference that can be reliably replicated across samples. When
models require very large sample sizes to detect these differences,
the effect size is generally small and the clinical application be-
comes irrelevant. Furthermore, the power estimates on the 3
significant variables showed adequate power (Table I), while the
model did not show significant predictive results. In addition,
the findings of the Nagelkerke R2 and the Hosmer-Lemeshow
test indicated that the model was sufficient and that the study
findings were not due to lack of power in the sample. None-
theless, while we believe that the results support the lack of
predictor variables that may be used clinically to identify stability
versus instability, there is no doubt that our sample size was
small for this type of analysis. In conducting the logistic re-
gression analysis, dividing a small number of outcomes into
different cells may have created the discrepancy with the initial
analysis. Therefore, it may be possible to find a significant in-
dependent predictor with a larger sample size of unstable hips;
however, the sample size required is unknown.
The retrospective method of this report constitutes an ad-
ditional study limitation. However, we believe that this issue was
offset by the quality and quantity of the collected data. At our
institution, data on patients treated using EUA were prospectively
recorded. The CPT codes were used to identify the patients for
this particular study. Because of the data storage methods at our
institution, we were able to study many more factors than in
previous studies. In this way, the application of logistic regression
analysis permitted the identification of any independent risk
factors of hip instability, if they existed. To the best of our
knowledge, no previous studies have used this method to
examine DDH and fracture variables and their relationship to
instability in the setting of fracture of the posterior wall of the
acetabulum.
Consistent with previous studies, univariate analyses
showed that certain radiographic findings in posterior
References
1. Letournel E, Judet R. Fractures of the acetabulum. 2nd ed. New York: Springer; 1993.
2. Baumgaertner MR. Fractures of the posterior wall of the acetabulum. J Am Acad
Orthop Surg. 1999 Jan;7(1):54-65.
3. Upadhyay SS, Moulton A, Burwell RG. Biological factors predisposing to traumatic
posterior dislocation of the hip. A selection process in the mechanism of injury. J
Bone Joint Surg Br. 1985 Mar;67(2):232-6.
I N S TA B I L I T Y O F T H E H I P J O I N T A F T E R P O S T E R I O R A C E TA B U L A R
WA L L F R A C T U R E
acetabular wall fractures are significant risk factors for hip
instability. However, the data did not fit the subsequent for-
ward multivariate logistic regression modeling, and none of
the studied variables reached significance. Logistic regression
modeling was chosen for this analysis because the dependent
variable of interest (stable or unstable) was binary and lo-
gistic regression is a powerful, nonparametric statistical test
that does not require many of the assumptions of typical
regression. Logistic regression was first performed with just
continuous variables; next, a mixed set of continuous and
nominal level variables was analyzed to predict stable or
unstable fractures. None of the models successfully predicted
hip joint instability. This finding leads to the conclusions that
the studied factors cannot be used to predict instability and
other factors not in the study may contribute to instability.
Therefore, these conclusions support the original hypothesis
that fracture fragment size alone or fracture fragment size in
combination with other measures, such as Tönnis angle,
acetabular index, anteversion, or proximity of the fracture
exit point to the dome, cannot dependably predict stability or
instability of the hip.
In summary, with the sample size available, we found all
of the studied variables to be nonsignificant as independent risk
factors for hip instability after a fracture of the posterior wall of
the acetabulum. The results of this analytic approach indicate
that important factors leading to hip instability have yet to be
identified, or the contributions of the measured ones are
relatively small. We cannot exclude the possibility that other,
yet undetermined, factors play an important independent role.
These undetermined factors could possibly be related to
proximal femoral dysplasia, labral pathology, or capsular
injury. Additional studies are warranted to examine these
variables. However, at the present time, EUA remains the only
reliable diagnostic method for determining hip stability status
after a fracture of the posterior wall of the acetabulum.
Therefore, EUA should remain as the main clinical determi-
nant of hip stability status. n
NOTE: The authors thank Heidi A. Israel, PhD, for her assistance with the statistical analysis.
Jay H. Patel, MD1
Berton R. Moed, MD1
1Department of Orthopaedic Surgery, Saint Louis University School of
Medicine, St. Louis, Missouri
E-mail address for B.R. Moed: moedbr@gmail.com
ORCID iD for B.R. Moed: 0000-0003-4908-8181
4. Reagan JM, Moed BR. Can computed tomography predict hip stability in posterior
wall acetabular fractures? Clin Orthop Relat Res. 2011 Jul;469(7):2035-41. Epub
2011 Jan 28.
5. Davis AT, Moed BR. Can experts in acetabular fracture care determine hip sta-
bility after posterior wall fractures using plain radiographs and computed tomogra-
phy? J Orthop Trauma. 2013 Oct;27(10):587-91.
 e126(7)
TH E JO U R NA L O F B O N E & JO I N T SU RG E RY J B J S . O RG
d
V O L U M E 99-A N U M B E R 23 D E C E M B E R 6, 2 017
d d
6. Tornetta P 3rd. Non-operative management of acetabular fractures. The use of
dynamic stress views. J Bone Joint Surg Br. 1999 Jan;81(1):67-70.
7. Grimshaw CS, Moed BR. Outcomes of posterior wall fractures of the acetabulum
treated nonoperatively after diagnostic screening with dynamic stress examination
under anesthesia. J Bone Joint Surg Am. 2010 Dec 1;92(17):2792-800.
8. Steppacher SD, Albers CE, Siebenrock KA, Tannast M, Ganz R. Femoroacetabular
impingement predisposes to traumatic posterior hip dislocation. Clin Orthop Relat
Res. 2013 Jun;471(6):1937-43. Epub 2013 Feb 20.
9. Krych AJ, Thompson M, Larson CM, Byrd JW, Kelly BT. Is posterior hip instability
associated with cam and pincer deformity? Clin Orthop Relat Res. 2012 Dec;470
(12):3390-7.
10. Philippon MJ, Kuppersmith DA, Wolff AB, Briggs KK. Arthroscopic findings fol-
lowing traumatic hip dislocation in 14 professional athletes. Arthroscopy. 2009
Feb;25(2):169-74. Epub 2008 Nov 1.
11. Firoozabadi R, Spitler C, Schlepp C, Hamilton B, Agel J, Routt MC, Tornetta P
3rd. Determining stability in posterior wall acetabular fractures. J Orthop Trauma.
2015 Oct;29(10):465-9.
12. Marsh JL, Slongo TF, Agel J, Broderick JS, Creevey W, DeCoster TA, Prokuski L,
Sirkin MS, Ziran B, Henley B, Audigé L. Fracture and dislocation classification
compendium - 2007: Orthopaedic Trauma Association classification, database and
outcomes committee. J Orthop Trauma. 2007 Nov-Dec;21(10)(Suppl):S1-133.
13. Clohisy JC, Carlisle JC, Beaulé PE, Kim YJ, Trousdale RT, Sierra RJ, Leunig M,
Schoenecker PL, Millis MB. A systematic approach to the plain radiographic
I N S TA B I L I T Y O F T H E H I P J O I N T A F T E R P O S T E R I O R A C E TA B U L A R
WA L L F R A C T U R E
evaluation of the young adult hip. J Bone Joint Surg Am. 2008 Nov;90(Suppl
4):47-66.
14. Reynolds D, Lucas J, Klaue K. Retroversion of the acetabulum. A cause of hip
pain. J Bone Joint Surg Br. 1999 Mar;81(2):281-8.
15. Moed BR, Ajibade DA, Israel H. Computed tomography as a predictor of hip
stability status in posterior wall fractures of the acetabulum. J Orthop Trauma. 2009
Jan;23(1):7-15.
16. Sharp IK. Acetabular dysplasia: the acetabular angle. J Bone Joint Surg Br.
1961 May;43(2):268-72.
17. McNamara AR, Boudreau JA, Moed BR. Nonoperative treatment of posterior wall
acetabular fractures after dynamic stress examination under anesthesia: revisited. J
Orthop Trauma. 2015 Aug;29(8):359-64.
18. Moed BR, Kregor PJ, Reilly MC, Stover MD, Vrahas MS. Current management of
posterior wall fractures of the acetabulum. Instr Course Lect. 2015;64:139-59.
19. Berkes MB, Cross MB, Shindle MK, Bedi A, Kelly BT. Traumatic posterior hip
instability and femoroacetabular impingement in athletes. Am J Orthop (Belle Mead
NJ). 2012 Apr;41(4):166-71.
20. Hattori T, Inaba Y, Ichinohe S, Kitano T, Kobayashi D, Saisu T, Ozaki T. The
epidemiology of developmental dysplasia of the hip in Japan: findings from a na-
tionwide multi-center survey. J Orthop Sci. 2017 Jan;22(1):121-6. Epub 2016 Sep 5.
21. Okano K, Takaki M, Okazaki N, Shindo H. Bilateral incidence and severity of
acetabular dysplasia of the hip. J Orthop Sci. 2008 Sep;13(5):401-4. Epub 2008
Oct 9.