1559

C OPYRIGHT 2012 BY T HE J OURNAL OF B ONE AND J OINT S URGERY, I NCORPORATED

Two to Twenty-Year Survivorship of the Hip in 810

Patients with Operatively Treated Acetabular Fractures
Moritz Tannast, MD, Soheil Najibi, MD, PhD, and Joel M. Matta, MD

Investigation performed at the Hip and Pelvis Institute, Santa Monica, California

Background: The aims of the study were (1) to determine the cumulative two to twenty-year survivorship of the hip after
open reduction and internal fixation of displaced acetabular fractures, (2) to identify factors predicting conversion to total
hip arthroplasty or hip arthrodesis, and (3) to create a predictive model that calculates an individuals probability of early
need for total hip arthroplasty or hip arthrodesis.
Methods: Eight hundred and sixteen acetabular fractures treated with open reduction and internal fixation by one
surgeon over a twenty-six-year period were analyzed. Cumulative two to twenty-year Kaplan-Meier survivorship analyses of
the hip, including best and worst-case scenarios, were performed with total hip arthroplasty or hip arthrodesis as the end
point. Univariate and multivariate Cox regression analyses were performed to identify negative predictors, which were then
used to construct a nomogram for predicting an individuals probability of needing an early total hip arthroplasty.
Results: The cumulative twenty-year survivorship of the 816 hips available for follow-up was 79% at twenty years. The
best and worst-case scenarios corresponded to cumulative twenty-year survivorship of 86% and 52%, respectively.
Significant independent negative predictors were nonanatomical fracture reduction, an age of more than forty years,
anterior hip dislocation, postoperative incongruence of the acetabular roof, involvement of the posterior acetabular wall,
acetabular impaction, a femoral head cartilage lesion, initial displacement of the articular surface of 20 mm, and
utilization of the extended iliofemoral approach.
Conclusions: Open reduction and internal fixation of displaced acetabular fractures was able to successfully prevent the
need for subsequent total hip arthroplasty within twenty years in 79% of the patients. The results represent benchmark
comparative data for any future and past studies on the outcome of surgical fixation of acetabular fractures.
Level of Evidence: Prognostic Level II. See Instructions for Authors for a complete description of levels of evidence.

A
cetabular fractures are life-altering injuries that com-
monly occur in young, active, and productive members
of society, although the number of elderly patients sus-
taining acetabular fractures has increased. The general aim of
surgical treatment of these challenging fractures is the preser-
vation of the native hip joint so that it can continue to function
for the remainder of the patients life1. Except for the contribu-
tion of Letournel and Judet2, this outcome has not been dem-
onstrated in a comprehensive study involving a large number of
patients with long-term follow-up spanning two decades.
The aims of this study were (1) to determine the cu-
mulative survivorship of the hip two to twenty years after
operative treatment of displaced acetabular fractures, (2) to
identify predictors indicating the future need for total hip ar-
throplasty or hip arthrodesis, and (3) to create a model for
prediction of the hip survivorship probability of an individual
patient with a specific acetabular fracture.
Materials and Methods
Study Group
T his study was approved by our local institutional review board. The study
was based on the data for 1319 consecutive acetabular fractures (1307
patients) treated by the senior author (J.M.M.) over a twenty-six-year period
(July 1980 to December 2006). Acetabular fractures that were treated

Disclosure: One or more of the authors received payments or services, either directly or indirectly (i.e., via his or her institution), from a third party in
support of an aspect of this work. In addition, one or more of the authors, or his or her institution, has had a financial relationship, in the thirty-six months
prior to submission of this work, with an entity in the biomedical arena that could be perceived to influence or have the potential to influence what is written
in this work. No author has had any other relationships, or has engaged in any other activities, that could be perceived to influence or have the potential to
influence what is written in this work. The complete Disclosures of Potential Conflicts of Interest submitted by authors are always provided with the
online version of the article.

J Bone Joint Surg Am. 2012;94:1559-67 d http://dx.doi.org/10.2106/JBJS.K.00444

 1560
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
T WO T O T W E N T Y -Y E A R S U R V I V O R S H I P O F T H E H I P A F T E R
V O LU M E 94-A N U M B E R 17 S E P T E M B E R 5, 2 012
d d
O P E R AT I V E LY T R E AT E D A C E TA B U L A R F R A C T U R E S

TABLE I Operative Approach According to Acetabular Fracture Type

Approach

Combined
Kocher- Extended Kocher-Langenbeck
All Langenbeck Ilioinguinal Iliofemoral and Ilioinguinal

Entire series 816 (100%) 352 (43%) 323 (40%) 129 (16%) 12 (1%)
Simple fracture type 241 (30%)
Anterior wall 12 (1%) 11 (92%) 1 (8%)
Anterior column 80 (10%) 77 (96%) 3 (4%)
Posterior wall 107 (13%) 104 (97%) 3 (3%)
Posterior column 14 (2%) 14 (100%)
Transverse 28 (3%) 20 (71%) 4 (14%) 4 (14%)
Associated fracture type 575 (70%)
Posterior column, 26 (3%) 26 (100%)
posterior wall
Transverse, posterior wall 143 (18%) 122 (85%) 1 (1%) 19 (13%) 1 (1%)
T-shaped 96 (12%) 60 (63%) 8 (8%) 21 (22%) 7 (7%)
Anterior column, 76 (9%) 3 (4%) 70 (92%) 3 (4%)
posterior hemitransverse
Both columns 234 (29%) 3 (1%) 152 (65%) 75 (32%) 4 (2%)
Duration of operation* (hr) 2.5 1.1 (1-9.3) 3 1.2 (1-12) 4.6 1.5 (1.5-10) 6 1.5 (3.5-8)
Blood loss* (mL) 800 650 (100-8000) 1000 700 (200-6000) 1700 1200 (300-7500) 1500 700 (700-2700)

*Values are given as the mean and the standard deviation, with the range in parentheses.

nonoperatively (eighty-seven hips [eighty-five patients]), those that required
acute primary total hip arthroplasty (twelve hips [twelve patients]), and those
in which the acetabular fracture was periprosthetic (twelve hips [twelve pa-
tients]) were excluded. Of the remaining 1208 acetabular fractures (1198 pa-
tients), 259 fractures (249 patients) had less than two years of follow-up and 133
hips (133 patients) were lost to follow-up, and these were excluded. Hips that
were converted to a total hip arthroplasty or hip arthrodesis at any time point
were included. This left 816 surgically treated displaced acetabular fractures
(810 patients) with complete data for analysis.
Radiographic Evaluation
All patients were evaluated with the use of five standard radiographic views in-
cluding an anteroposterior radiograph, two Judet views, and cephalad and caudad
3
views of the pelvis . All patients who were managed after 1983 were further
evaluated with computed tomography (CT) of the pelvis (737 acetabular fractures,
90%). The maximum displacement of the fracture on any of the five radiographs
was recorded without adjustment for magnification. The mean displacement (and
standard deviation) was 20 10 mm (range, 2 to 70 mm). Two hundred and
twenty-six acetabular fractures (28%) had a displacement of 20 mm.

Data Collection Classification of Fractures

Data were acquired and saved in a digital format according to a standardized
1 2
documentation protocol established in 1980 . The created database included
information on demographics, mechanism of injury, associated injuries in-
cluding neurologic and vascular injuries, any previous surgical treatment,
timing of surgery, classification of the fracture, preoperative and postoperative
radiographic evaluation, intraoperative and perioperative data, and complica-
tions as well as follow-up (see Appendix).
Clinical Evaluation
Eighty-six patients (11%) had a nerve palsy preoperatively. The sciatic nerve
was involved in seventy-eight (10% of the total cohort) of these patients and
the femoral or superior gluteal nerve in eight (1%). A fracture-dislocation
occurred in 203 hips (25%); the dislocation was anterior in six hips (1% of the
total cohort) and posterior in 197 (24%). One hundred and fifty-two dislo-
cations (six anterior and 146 posterior, 75%) were treated with closed re-
duction before the fracture surgery, and the remaining fifty posterior
dislocations were reduced at the time of the fracture reduction. Preoperative
traction was generally not used except in certain fracture patterns that put the
hip at risk for posterior dislocation or raised concern of femoral head wear
against a fracture edge (e.g., transtectal-T and transverse patterns; sixty-two
fractures, 8%).
Each fracture was classified on the basis of an evaluation of the five radiographs
as well as the CT scan according to the method of Letournel and Judet . Of the
816 acetabular fractures, 241 (30%) had a simple fracture pattern and 575
(70%) had an associated fracture pattern (Table I).
Surgical Technique
Three standard approaches were used: Kocher-Langenbeck, ilioinguinal, and
extended iliofemoral. The approach was chosen according to a previously es-
1 2
tablished surgical protocol and the recommendations of Letournel and Judet ,
which were based on the classification and configuration of the fracture. The
AMSCO and Kirschner orthopaedic tables (no longer manufactured) were used
from the initiation of the study in 1980 until 1985. The Judet Tasserit table (no
longer manufactured) was used from 1985 until 2006. The PROfx table
(Mizuho OSI, Union City, California) has been in use after 2003. The prone
position was used for the Kocher-Langenbeck surgical approach, the supine
position was used for the ilioinguinal approach, and the lateral position was
used for the extended iliofemoral approach. The goal of the surgery was fracture
reduction and fixation through a single approach when possible. When the
fracture reduction was deemed achievable through a single approach, an ilio-
inguinal or Kocher-Langenbeck approach alone was preferred rather than
the extended iliofemoral approach or a combined approach to minimize the
 1561
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
T WO T O T W E N T Y -Y E A R S U R V I V O R S H I P O F T H E H I P A F T E R
V O LU M E 94-A N U M B E R 17 S E P T E M B E R 5, 2 012
d d
O P E R AT I V E LY T R E AT E D A C E TA B U L A R F R A C T U R E S

Fig. 1-A

Fig. 1-B
Time dependence of the utilization of the different approaches (Fig. 1-A) and of the accuracy of reduction (Fig. 1-B).

negative effect of the more extensive dissection and soft-tissue injury in the
latter approaches. The surgical approach selected was generally consistent for
six of the fracture types (Table I). An ilioinguinal approach was typically utilized
for anterior wall, anterior column, and anterior column plus posterior hemi-
transverse fracture patterns. The Kocher-Langenbeck approach was typically
selected for posterior wall, posterior column, and posterior column plus
posterior wall fractures. The approach for treatment of transverse, T-shaped,
transverse plus posterior wall, and both-column fractures was individualized
(Table I). The choice of surgical approach during the study varied slightly over
time. There was an increase in utilization of the ilionguinal and Kocher-
Langenbeck approaches and a decrease in utilization of the extended ilio-
femoral approach over time (Fig. 1-A). The orthopaedic table served as an
important reduction aid and was a major factor in decreasing the number of
combined approaches and extended iliofemoral approaches.
 1562
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
T WO T O T W E N T Y -Y E A R S U R V I V O R S H I P O F T H E H I P A F T E R
V O LU M E 94-A N U M B E R 17 S E P T E M B E R 5, 2 012
d d
O P E R AT I V E LY T R E AT E D A C E TA B U L A R F R A C T U R E S

TABLE II Accuracy of Reduction According to Fracture Type and Other Characteristics

Anatomical, Imperfect, Surgical Secondary

0-1 mm 2-3 mm Poor Congruence

Entire series (n = 816) 616 (75%) 148 (18%) 36 (4%) 16 (2%)

Simple fracture type (n = 241) 201 (83%)* 34 (14%) 6 (2%)
Anterior wall (n = 12) 6 (50%) 5 (42%)* 1 (8%)
Anterior column (n = 80) 65 (81%) 12 (15%) 3 (4%)
Posterior wall (n = 107) 98 (92%)* 8 (7%) 1 (1%)
Posterior column (n = 14) 11 (79%) 3 (21%)
Transverse (n = 28) 21 (75%) 6 (21%) 1 (4%)
Associated fracture type (n = 575) 415 (72%) 114 (20%) 30 (5%) 16 (3%)
Posterior column, posterior wall (n = 26) 24 (92%)* 2 (8%)
Transverse, posterior wall (n = 143) 114 (80%) 25 (17%) 4 (3%)
T-shaped (n = 96) 67 (70%) 22 (23%) 7 (7%)
Anterior column, posterior hemitransverse (n = 76) 51 (67%) 20 (26%) 5 (7%)
Both columns (n = 234) 159 (68%) 45 (19%) 14 (6%) 16 (7%)
Initial displacement
20 mm (n = 226) 160 (71%) 47 (21%) 14 (6%) 5 (2%)
<20 mm (n = 590) 456 (77%) 101 (17%) 22 (4%) 11 (2%)
Treatment delay
<21 days (n = 730) 564 (77%)* 129 (18%) 26 (4%) 11 (2%)
21 days (n = 86) 52 (60%) 19 (22%) 10 (12%)* 5 (6%)
Previous surgery
Yes (n = 5) 2 (40%) 2 (40%) 1 (20%)
No (n = 811) 614 (76%) 146 (18%) 35 (4%) 16 (2%)
Age
<40 yr (n = 386) 316 (82%)* 50 (13%) 15 (4%) 5 (1%)
40-65 yr (n = 318) 234 (74%) 63 (20%) 13 (4%) 8 (3%)
>65 yr (n = 112) 66 (59%) 35 (31%)* 8 (7%) 3 (3%)
Approach
Ilioinguinal (n = 323) 226 (70%) 73 (23%)* 15 (5%) 9 (3%)
Kocher-Langenbeck (n = 352) 288 (82%)* 54 (15%) 10 (3%)
Extended iliofemoral (n = 129) 92 (71%) 20 (16%) 10 (8%)* 7 (5%)*
Ilioinguinal and Kocher-Langenbeck (n = 12) 10 (83%) 1 (8%) 1 (8%)

*Significantly higher compared with hips without the specific criterion. Significantly lower compared with hips without the specific criterion.

Reduction and Fixation
The fracture reduction was initiated with the help of intraoperative longitudinal
and/or lateral traction provided by means of the orthopaedic table. The re-
duction was fine-tuned by direct manipulation of the bone with bone forceps to
achieve an anatomical reduction of the acetabulum and the innominate bone.
A plate and screw construct was utilized for fracture fixation in 721 hips
(88%). Screw-only fixation was utilized in the remaining ninety-five hips (12%)
in which large fragments and good bone quality were present. The duration of
the operation and the estimated blood loss varied according to the operative
approach (Table I). The intraoperative findings included free intra-articular
fragments in 169 hips (21%), injury to the cartilage and/or bone of the femoral
head in 190 hips (23%), and acetabular articular impaction in 163 hips (20%).
Accuracy of Reduction
The accuracy of reduction was assessed on the basis of measurements of the
greatest residual displacement of any of the six acetabular reference lines of
2
Letournel and Judet on the anteroposterior and two Judet radiographs without
adjustment for magnification. The reduction was graded as anatomical (0 to
1 mm of residual displacement), imperfect (2 to 3 mm), or poor (>3 mm). In
both-column fractures in which the acetabulum was reduced anatomically but
the displacement of the innominate bone altered the position of the joint,
the reduction was categorized as surgical secondary congruence. In addition,
congruency of the acetabular roof with the femoral head was assessed
postoperatively.
Follow-up
Clinical and radiographic follow-up was generally performed at six weeks, three
months, and one and two years after fracture reduction. All patients were
advised to return for follow-up at regular two to three-year intervals thereafter.
In addition to the regular follow-up, an attempt to contact every patient for the
purposes of this study was made by phone, mail, and/or e-mail. This most
recent study-specific follow-up information was acquired between October
2008 and September 2009. All patients were asked whether or not they had
undergone conversion to a total hip arthroplasty or hip arthrodesis and the
 1563
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
T WO T O T W E N T Y -Y E A R S U R V I V O R S H I P O F T H E H I P A F T E R
V O LU M E 94-A N U M B E R 17 S E P T E M B E R 5, 2 012
d d
O P E R AT I V E LY T R E AT E D A C E TA B U L A R F R A C T U R E S

TABLE III Survivorship According to Fracture Type and Other Characteristics

Survivorship (95% Confidence Interval)* (%)

Median Time
Two Years Five Years Ten Years Twenty Years to Failure

Entire series (n = 816) 91 (90-92) 88 (87-90) 85 (84-87) 79 (76-81) 1.5

Simple fracture type (n = 241) 91 (89-93) 86 (84-89) 84 (81-87) 73 (68-79) 1.3
Anterior wall (n = 12) 91 (82-100) 68 (53-84) 68 (53-84) 34 (9-59) 2.3
Anterior column (n = 80) 95 (92-97) 92 (88-95) 87 (83-91) 77 (70-85) 3.0
Posterior wall (n = 107) 88 (84-91) 82 (78-86) 81 (77-85) 76 (71-82) 1.2
Posterior column (n = 14) 100 100 100 100
Transverse (n = 28) 89 (83-95) 89 (83-95) 89 (83-95) 89 (83-95) 0.3
Associated fracture type (n = 575) 92 (91-93) 89 (88-91) 86 (84-87) 80 (78-83) 1.6
Posterior column, posterior wall (n = 26) 85 (78-92) 85 (78-92) 85 (78-92) 85 (78-92) 0.5
Transverse, posterior wall (n = 143) 89 (86-91) 85 (82-88) 81 (78-85) 74 (68-80) 1.5
T-shaped (n = 96) 89 (85-92) 85 (81-89) 77 (72-81) 74 (68-79) 1.6
Anterior column, posterior hemitransverse (n = 76) 92 (89-95) 92 (89-95) 88 (84-92) 75 (65-84) 1.3
Both columns (n = 234) 96 (94-97) 93 (91-95) 91 (89-93) 87 (83-90) 2.2
Initial displacement
20 mm (n = 226) 86 (84-89) 84 (81-86) 78 (75-81) 68 (63-73) 1.3
<20 mm (n = 590) 93 (92-95) 90 (89-91) 88 (86-89) 83 (81-85) 1.9
Treatment delay
<21 days (n = 730) 93 (92-94) 89 (88-91) 86 (85-88) 79 (77-82) 2.0
21 days (n = 86) 82 (78-86) 80 (75-84) 74 (69-79) 74 (69-79) 0.9
Previous surgery
Yes (n = 5) 60 (38-82) 30 (6-54) 0.8
No (n = 811) 92 (91-93) 89 (87-90) 85 (84-87) 79 (77-81) 1.6
Age
<40 yr (n = 386) 96 (95-97) 95 (94-96) 92 (91-94) 87 (84-89) 2.3
40-65 yr (n = 318) 88 (86-90) 83 (81-86) 81 (79-83) 74 (71-77) 1.3
>65 yr (n = 112) 83 (79-87) 79 (75-83) 70 (65-76) 51 (38-64) 0.8
>75 yr (n = 42) 80 (73-87) 74 (66-83) 65 (54-76) 0.6
Approach
Ilioinguinal (n = 323) 94 (94-96) 92 (91-94) 90 (88-92) 84 (81-87) 2.0
Kocher-Langenbeck (n = 352) 90 (88-91) 86 (84-88) 84 (82-86) 80 (77-83) 1.2
Extended iliofemoral (n = 129) 90 (87-93) 86 (83-90) 79 (75-83) 67 (61-73) 2.2
Ilioinguinal and Kocher-Langenbeck (n = 12) 92 (84-100) 83 (73-94) 73 (59-86) 3.5

*Based on Kaplan-Meier cumulative survivorship analysis. Calculated for failures only. Significantly lower compared with hips without the
specific criterion. Significantly higher compared with hips without the specific criterion.

timing of such a procedure. Of the 1198 patients (1208 fractures), 810 patients
(816 fractures, 68%) were successfully contacted. Postoperative nerve palsies
were not specifically investigated since that was beyond the scope of this
report. Eighty-seven (11%) of the patients with adequate follow-up had
died by the time of the study follow-up. Only one hip in these patients had
been converted to a total hip arthroplasty. Mortality and conversion to total
hip arthroplasty in these patients were ascertained on the basis of the
hospitals demographic database and the follow-up information provided
by relatives.
Worst and Best-Case Scenarios
Patients with insufficient follow-up (less than two years) or no follow-up were
also analyzed to identify any differences with regard to demographics and
clinical, radiographic, and intraoperative findings. A worst-case scenario was
postulated on the basis of the assumption that all fractures with insufficient or
no follow-up would fail. A best-case scenario was postulated on the basis of the
assumption that no fractures with insufficient or no follow-up would fail.
Statistical Analysis
The chi-square test was used to calculate differences in categorical parameters
between a pair of groups, and the unpaired t test was used for continuous
parameters. The cumulative survivorship was analyzed according to the Kaplan-
Meier method, with the end point defined as total hip arthroplasty or ar-
4
throdesis of the hip . Univariate and multivariate Cox proportional-hazard
modeling was used to identify factors predictive of a poor outcome and to
calculate the corresponding hazard ratios. When a predictive factor with a non-
nominal scale was identified, the difference in survivorship between the two
groups was calculated with use of the log-rank test. The predictors from the
 1564
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 LU M E 94-A N U M B E R 17 S E P T E M B E R 5, 2 012
d d
final multivariate Cox regression model were used to construct a nomogram to
predict the early need for total hip arthroplasty (by two years postoperatively).
A p value of 0.05 was considered significant.
Source of Funding
One of the authors (M.T.) received personal funding from the Swiss National
Science Foundation (SNF) and the Association for Orthopaedic Research
(AFOR) to conduct this study.
Results
Accuracy of Reduction (Table II)
T he mean postoperative displacement at the level of the
articular cartilage was 0.9 1.9 mm (range, 0 to 20 mm).
The postoperative reduction was graded as anatomical in 616
hips (75%), imperfect in 148 (18%), poor in thirty-six (4%),
and surgical secondary congruence in sixteen (2%). The rate
of anatomical reduction was significantly higher in hips with
simple compared with associated fracture types, in hips
without delayed treatment, and in patients younger than forty
years at the time of the fracture (p < 0.001 for all). Posterior
wall fractures and posterior column plus posterior wall frac-
tures had a significantly higher rate of anatomical reduction
(p < 0.001 and p = 0.04, respectively). Both-column fractures
had a significantly lower rate of anatomical reduction (p =
0.001). The overall rate of anatomical reduction increased
consistently over time from 40% in 1980 to 92% in 2006
(Fig. 1-B).
Survivorship Analysis (Table III)
One hundred and six hips were converted to total hip arthro-
plasty before the specified minimum two years of follow-up.
The mean duration of follow-up for the remaining 710 hips
was 10.3 6.9 years (range, two to 28.6 years). The follow-up
was two to five years in 188 hips (23% of the total cohort), five
to ten years in 223 (27%), ten to fifteen years in 133 (16%),
fifteen to twenty years in seventy-seven (9%), and more than
twenty years in eighty-nine (11%). A total of 124 hips (15%)
failed; 120 were converted to a total hip arthroplasty and four to
a hip arthrodesis. The mean time to failure was 4.5 6.2 years
(range, 0.2 to 26.2 years), and the median was 1.5 years.
The cumulative survivorship for the entire series was
88% (95% confidence interval [CI], 87% to 90%) at five years
of follow-up, 85% (95% CI, 84% to 87%) at ten years, and 79%
(95% CI, 76% to 81%) at twenty years (Table III, Fig. 2). The
survivorship for both-column fractures at twenty years of
follow-up was significantly greater (87% [95% CI, 83% to
90%], p = 0.002), and the survivorship for anterior wall frac-
tures was significantly lower (34% [95% CI, 9% to 59%], p =
0.002). The other types of fractures were not significantly dif-
ferent in terms of survivorship. Survivorship was significantly
lower in patients who had been older than forty years at the
time of the surgery (70% [95% CI, 67% to 74%] at twenty
years, p < 0.001), patients who had undergone previous un-
successful surgical attempts at reduction (30% [95% CI, 6% to
54%] at five years, p < 0.001), and patients in whom an ex-
tended iliofemoral approach had been utilized (73% [95% CI,
75% to 83%] at fifteen years, p = 0.04).
T WO T O T W E N T Y -Y E A R S U R V I V O R S H I P O F T H E H I P A F T E R
O P E R AT I V E LY T R E AT E D A C E TA B U L A R F R A C T U R E S
Fig. 2
Survivorship curve (including 95% confidence interval) for all acetabular
fractures (n = 816). Two curve sections can be identified. Section A covers
the first 1.5 years and represents an exponential decrease of the survi-
vorship. Section B covers the time period from 1.5 to twenty years post-
operatively and represents a linear decrease of the survivorship. The first
50% of hips that failed did so by 1.5 years postoperatively.
Predictors of Outcome (see Appendix) and Nomogram for
Predicting Early Failure
Univariate analysis of possible negative predictors identified
nineteen factors, and nine of these were identified as inde-
pendent significant predictors of outcome on the basis of the
multivariate analysis. Six of those nine parameters were al-
ready predetermined at the time of injury or initial evalua-
tion: an age of over forty years, anterior dislocation, femoral
head cartilage lesion, involvement of the posterior wall,
marginal impaction, and initial displacement of 20 mm.
The three remaining negative predictors were directly related
to the surgical intervention: nonanatomical reduction, post-
operative incongruence of the acetabular roof, and utilization
of the extended iliofemoral approach. The median time to
failure ranged from 1.0 years for hips with incongruence
of the acetabular roof to 10.9 years for hips with anterior
dislocation.
A nomogram was constructed to allow prediction of the
probability of the need for a total hip arthroplasty by two years
postoperatively (Fig. 3).
Worst and Best-Case Scenarios
The worst-case scenario revealed a cumulative two-year sur-
vivorship of 61% (95% CI, 59% to 62%), a ten-year survi-
vorship of 57% (95% CI, 55% to 58%), and a twenty-year
survivorship of 52% (95% CI, 50% to 54%). The best-case
scenario revealed a cumulative two-year survivorship of 94%
(95% CI, 93% to 95%), a ten-year survivorship of 90% (95%
CI, 9% to 91%), and a twenty-year survivorship of 86% (95%
CI, 85% to 88%).
 1565
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
T WO T O T W E N T Y -Y E A R S U R V I V O R S H I P O F T H E H I P A F T E R
V O LU M E 94-A N U M B E R 17 S E P T E M B E R 5, 2 012
d d
O P E R AT I V E LY T R E AT E D A C E TA B U L A R F R A C T U R E S

Fig. 3
Nomogram predicting the early need for total hip arthroplasty (or hip arthrodesis) within two years postoperatively. To use the nomogram, locate the age axis
and draw a line straight upward to the Points scale at the top to determine how many points the patient receives on the basis of his or her age. Repeat this
process for each of the other predictor variables, then sum the points for the individual predictors. Locate this sum on the Total Points axis and draw a line
straight downward to identify the predicted probability of the need for total hip arthroplasty within two years postoperatively.

Patients without sufficient follow-up did not differ sig-
nificantly from the analyzed patients with regard to their de-
mographic, radiographic, and operative data except for a
higher proportion of male patients (p = 0.025), a greater age at
surgery (p < 0.001), fewer preoperative nerve palsies (p = 0.03),
fewer posterior wall fractures (p = 0.034), fewer hips with
<20 mm of initial displacement (p = 0.046), fewer hips with a
nonanatomical reduction (p = 0.002), and fewer hips with
incongruence of the acetabular roof postoperatively (p = 0.008)
(see Appendix).
Discussion
tabular fractures. We found that 79% of the native hips could
be preserved successfully at a follow-up of twenty years. When
secondary total hip arthroplasty was necessary, 50% of all cases
required joint replacement within 1.5 years after fixation of the
fracture. Independent noncontrollable factors associated with
the need for total hip arthroplasty were age, anterior disloca-
tion, posterior wall involvement, a femoral head cartilage
lesion, marginal impaction, and the amount of initial dis-
placement. Independent controllable negative predictors
of outcome were the accuracy of reduction, restoration of a
congruent acetabular roof, and utilization of the extended ilio-
femoral approach. This statistical information was incorpo-

T o our knowledge, no comprehensive study with long-term

follow-up has previously analyzed negative predictors of
outcome in a large series of consecutive patients with all sub-
types of acetabular fractures. The aim of the present study was
to document the cumulative long-term survivorship of the hip
over a two-decade period after open reduction and internal
fixation (ORIF) in a consecutive series of 816 displaced ace-
rated in a nomogram for clinical use; the nomogram predicts
the need for early total hip arthroplasty after fixation of an
acetabular fracture.
This study has some limitations. We evaluated the long-
term results by using conversion to total hip arthroplasty or hip
fusion as an indirect indication of the development of post-
traumatic osteoarthritis. We did not evaluate the clinical or
 1566
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 LU M E 94-A N U M B E R 17 S E P T E M B E R 5, 2 012
d d
radiographic follow-up data of the patients. Advantages of this
approach include the clarity of the end point, the availability of
follow-up information for patients who were unable to return
for follow-up visits, the retrospective availability of follow-up
information for patients who died, and the reduction of
unnecessary radiation exposure in asymptomatic patients.
Disadvantages include the lack of radiographic and clinical
follow-up information for comparison with the literature.
Symptomatic and asymptomatic cases of secondary osteoar-
thritis that did not lead to arthroplasty or arthrodesis are not
reflected in the survivorship curve. In addition, indications for
arthroplasty can vary among surgeons, making this analysis less
generally applicable.
Another limitation is the relatively large number of pa-
tients with no or insufficient follow-up. Although the loss-to-
follow-up quotient5 for this study was 3.2, the 79% overall
cumulative survivorship of the hip after twenty years is note-
worthy. The majority of the demographic, radiographic, and
operative parameters did not differ between the groups with
sufficient and insufficient follow-up (see Appendix). Moreover,
a greater proportion of hips with negative predictive factors
were identified in the evaluated group of 816 hips with ade-
quate follow-up compared with those with inadequate follow-
up (see Appendix). This comparison indicates that the actual
survivorship after twenty years may be better than 79%.
Two sections of the survivorship curve could be identi-
fied. The first curve section showed an exponential decrease
and covered the first 1.5 years after surgery, during which the
first 50% of the failures occurred (see Appendix). Hips with
multiple negative predictors typically failed during this early
time period, as predicted by the nomogram. The second curve
section showed a more linearly decreasing survivorship and
covered the 18.5-year time period starting 1.5 years postoper-
atively. During this time, secondary osteoarthritis could de-
velop even in the absence of negative predictors. Preexisting
osteoarthritis 6,7, natural wear of the joint, and substantial
traumatic cartilage damage might lead to this phenomenon. A
similar long-term survivorship pattern was assumed but not
proven by other authors6,8,9. Studies of the outcome of acetab-
ular fractures should have a follow-up of at least two years.
When evaluating and comparing long-term outcomes, the
linear decrease in survivorship over time should be taken into
account.
The Cox regression analysis technique used in the pres-
ent study is a powerful tool that allows prediction of the time-
dependent probability that an individual patient will need a
total hip arthroplasty. The constructed nomogram (Fig. 3) is
a graphical representation of the numerical probability of the
need for a total hip arthroplasty by two years. As shown in
the two illustrative case examples in the Appendix, individual
prognoses corresponded well with the actual patient follow-up.
The knowledge of the individualized prognosis provides the
patient and the treating orthopaedic surgeon with realistic
expectations. These could influence the postoperative treat-
ment regime, the future level of activity that is adopted, or even
the choice of work. The two-year time point for the nomogram
T WO T O T W E N T Y -Y E A R S U R V I V O R S H I P O F T H E H I P A F T E R
O P E R AT I V E LY T R E AT E D A C E TA B U L A R F R A C T U R E S
was chosen since an unfavorable early prognosis could justify
acute primary total hip arthroplasty instead of ORIF6. The
nomogram can be used as an adjunct to select patients in whom
acute primary total hip arthroplasty could be more beneficial
than ORIF. Primary total hip arthroplasty might be considered
in such patients in whom early joint degeneration is highly
probable. However, accurate fracture reduction is essential even
if primary total hip arthroplasty is planned. The decision re-
garding the choice between ORIF and total hip arthroplasty
must also take into account other factors in addition to the
individual prognosis, such as the invasiveness of the procedure,
the associated morbidity, and the costs associated with a second
surgery.
Many of the identified negative predictors are consistent
with those in the literature, including nonanatomical reduc-
tion1,6,10-12, incongruence of the acetabular roof1, an age of more
than forty years1,6,9, a femoral head cartilage lesion6,7,9, marginal
impaction6,12, delayed reconstruction6,10, and relocation of an
associated dislocation after more than six hours13. We were also
able to identify additional univariate and multivariate predic-
tors of a poor outcome, including anterior or posterior fracture
dislocation, initial displacement of 20 mm, the presence
of free intra-articular fragments, involvement of the posterior
acetabular wall, and utilization of the extended iliofemoral
approach. The strongest predictors that were controllable by
the surgeon were the accuracy of reduction and the restoration
of a congruent acetabular roof.
There are some discrepancies between our findings and
the literature. In contrast to other reports9, we did not find
superior survivorship of hips with simple fracture patterns.
Consistent with one previous report6, anterior wall fractures
had the worst survivorship. This might be related to the
somewhat higher proportion of elderly patients, marginal
impaction, and nonanatomical reduction associated with
these fractures (see Appendix). Also, involvement of the
posterior acetabular wall (often associated with substantial
cartilage damage; see Appendix) was a negative predictor.
Both-column fractures were the only fractures with a signif-
icantly better outcome at twenty years, despite the signifi-
cantly higher proportion of nonanatomical reduction, large
initial displacement, and more frequent utilization of the
extended iliofemoral approach (see Appendix). An explana-
tion for this better outcome is the secondary surgical con-
gruence of the articular surface despite deficiencies in
reduction. In both-column fractures, all articular fragments
displace with the femoral head. The involvement of the
innominate bone in the fracture might allow it to act as a
crumple zone, resulting in less articular cartilage damage.
Another explanation could be the more direct force trans-
mission that occurs in acetabular wall fractures, in which the
femoral head is displacing in relation to a stable fragment.
This can create marginal impaction of the acetabulum, marked
cartilage damage, and impaction of the femoral head.
As shown in our study, the extended iliofemoral surgical
approach was directly associated with a higher risk of subse-
quent total hip arthroplasty. Although we did not evaluate the
 1567
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 LU M E 94-A N U M B E R 17 S E P T E M B E R 5, 2 012
d d
patient radiographs, one could suggest that the problem with
the extended iliofemoral approach is osteonecrosis of the in-
nominate bone with disconnection of the inner and outer
pelvic tables6. However, this was not observed by the senior
author when these patients returned for total hip arthroplasty.
Over time, the use of this approach has decreased, whereas the
use of the ilioinguinal and Kocher-Langenbeck approaches
alone and the rate of anatomical reduction have increased (Figs.
1-A and 1-B). The primary reason for these trends was the use
of the orthopaedic table, but a secondary reason was the de-
velopment of sophisticated surgical reduction instruments for
these approaches. More extensive dissection does not neces-
sarily lead to a higher rate of anatomical reduction and a better
prognosis. However, the extended iliofemoral surgical ap-
proach is still being used by the senior author in both-column
fractures with involvement of the sacroiliac joint. A combined
ilioinguinal and Kocher-Langenbeck approach is rarely neces-
sary with this fracture pattern.
We observed several secular trends in our patient pop-
ulation (see Appendix). Five of the negative predictors in the
multivariate analysis (a femoral head cartilage lesion, involve-
ment of the posterior wall, an age of more than forty years,
initial displacement of 20 mm, and marginal impaction)
showed an increasing prevalence over time. Some of those
factors might be the result of the aging population or the
maturation of a highly selected referral practice.
In summary, 21% of all patients with operatively treated
displaced acetabular fractures needed conversion to total hip
arthroplasty within twenty years. Half of the conversions oc-
curred within the first 1.5 years, with a slow and steady rate of
total hip arthroplasty at longer follow-up. Factors that were
predictive of the need for early conversion to total hip ar-
throplasty were greater age, anterior dislocation, posterior wall
involvement, a femoral head cartilage lesion, marginal im-
paction, large initial displacement, nonanatomical reduction,
failure to restore a congruent acetabular roof, and utilization of
References
1. 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 Nov;78(11):1632-45.
2. Letournel E, Judet R. Fractures of the acetabulum. New York, NY: Springer; 1993.
3. Judet R, Judet J, Letournel E. Fractures of the acetabulum: classification and
surgical approaches for open reduction. Preliminary report. J Bone Joint Surg Am.
1964 Dec;46:1615-46.
4. Kaplan E, Meier P. Nonparametric estimation from incomplete observations.
J Am Statist Assn. 1958;53:457-81.
5. Murray DW, Britton AR, Bulstrode CJ. Loss to follow-up matters. J Bone Joint Surg
Br. 1997 Mar;79(2):254-7.
6. Mears DC, Velyvis JH, Chang CP. Displaced acetabular fractures managed
operatively: indicators of outcome. Clin Orthop Relat Res. 2003 Feb;(407):
173-86.
7. Tannast M, Kruger A, Mack PW, Powell JN, Hosalkar HS, Siebenrock KA. Surgical
dislocation of the hip for the fixation of acetabular fractures. J Bone Joint Surg Br.
2010 Jun;92(6):842-52.
T WO T O T W E N T Y -Y E A R S U R V I V O R S H I P O F T H E H I P A F T E R
O P E R AT I V E LY T R E AT E D A C E TA B U L A R F R A C T U R E S
the extended iliofemoral surgical approach. We observed an
increasing prevalence of many of the predictive factors over the
twenty-six years encompassed by the study. A nomogram was
constructed that can predict the likelihood of early conversion.
This nomogram can be helpful in selecting patients who could
potentially benefit from acute primary total hip arthroplasty
instead of ORIF alone. To our knowledge, the number of
evaluated fractures and the duration of follow-up in the present
study are unique for acetabular fractures in particular and
unusual for any type of articular fractures in general. The re-
sults represent benchmark comparative data for any future and
past studies on the outcome after surgical fixation of acetabular
fractures.
Appendix
Tables showing demographic data for patients with and
without sufficient follow-up, regression results, the dis-
tribution of negative predictive factors according to fracture
type, and secular trends as well as figures showing illustrative
examples of patients with good and poor predicted outcomes
are available with the online version of this article as a data
supplement at jbjs.org. n
Moritz Tannast, MD
Department of Orthopaedic Surgery,
Inselspital, University of Bern,
Freiburgstrasse, 3010 Bern, Switzerland.
E-mail address: moritz.tannast@insel.ch
Soheil Najibi, MD, PhD
Joel M. Matta, MD
The Hip and Pelvis Institute, St. Johns Health Center,
2001 Santa Monica Boulevard #1090,
Santa Monica, CA 90404
8. McKinley TO, Borrelli J Jr, DLima DD, Furman BD, Giannoudis PV. Basic science
of intra-articular fractures and posttraumatic osteoarthritis. J Orthop Trauma. 2010
Sep;24(9):567-70.
9. Liebergall M, Mosheiff R, Low J, Goldvirt M, Matan Y, Segal D. Acetabular fractures.
Clinical outcome of surgical treatment. Clin Orthop Relat Res. 1999 Sep;(366):205-16.
10. Brueton RN. A review of 40 acetabular fractures: the importance of early sur-
gery. Injury. 1993 Mar;24(3):171-4.
11. Oh CW, Kim PT, Park BC, Kim SY, Kyung HS, Jeon IH, Cheon SH, Min WK.
Results after operative treatment of transverse acetabular fractures. J Orthop Sci.
2006 Oct;11(5):478-84.
12. Kreder HJ, Rozen N, Borkhoff CM, Laflamme YG, McKee MD, Schemitsch EH,
Stephen DJ. Determinants of functional outcome after simple and complex ace-
tabular fractures involving the posterior wall. J Bone Joint Surg Br. 2006 Jun;
88(6):776-82.
13. Hougaard K, Thomsen PB. Traumatic posterior fracture-dislocation of the hip
with fracture of the femoral head or neck, or both. J Bone Joint Surg Am. 1988
Feb;70(2):233-9.