Shelf acetabuloplasty for children with

Perthes’ disease and reducible subluxation of

the hip
PROGNOSTIC FACTORS RELATED TO HIP REMODELLING

W. J. Yoo, We evaluated radiological hip remodelling after shelf acetabuloplasty and sought to identify
I. H. Choi, prognostic factors in 25 patients with a mean age of 8.9 years (7.0 to 12.3) who had
T.-J. Cho, unilateral Perthes’ disease with reducible subluxation of the hip in the fragmentation stage.
C. Y. Chung, At a mean follow-up of 6.7 years (3.2 to 9.0), satisfactory remodelling was observed in
Y.-W. Shin, 18 hips (72%). The type of labrum in hip abduction, as determined by intra-operative
S. J. Shin dynamic arthrography, was found to be a statistically significant prognostic factor
(p = 0.012).
From Seoul National Shelf acetabuloplasty as containment surgery seems to be best indicated for hips in
University Children’s which there is not marked collapse of the epiphysis and in which the extruded epiphyseal
Hospital, Seoul, segment slips easily underneath the labrum on abduction, without imposing undue
Korea pressure on the lateral edge of the acetabulum.

„ W. J. Yoo, MD, Assistant
Professor
„ I. H. Choi, MD, Professor
„ T.-J. Cho, MD, Professor
„ C. Y. Chung, MD, Professor
Department of Orthopaedic
Surgery
Seoul National University
Hospital, 28 Yeongon-dong,
Jongno-gu, 110-744 Seoul,
Korea.
„ Y.-W. Shin, MD, Assistant
Professor
Department of Orthopaedic
Surgery
Inje University Sanggye Paik
Hospital, Sanggye-dong,
Nowon-gu, 139-707 Seoul,
Korea.
„ S. J. Shin, MD, Assistant
Professor
Department of Orthopaedic
Surgery
Jeju University Hospital, Ara
1-dong, 690-767 Jeju-do, Korea.
Correspondence should be sent
to Professor I. H. Choi; e-mail:
inhoc@snu.ac.kr
©2009 British Editorial Society
of Bone and Joint Surgery
doi:10.1302/0301-620X.91B10.
22149 $2.00
J Bone Joint Surg [Br]
2009;91-B:1383-7.
Received 3 December 2008;
Accepted after revision 26 May
2009
Shelf acetabuloplasty has been widely used to
treat Perthes’ disease for two different surgical
indications. When used as a salvage procedure,
it provides relief from pain and better cover of
the subluxed femoral head in patients with
coxa magna, coxa irregularis or hinge abduc-
tion during the late stage of the disease.1,2 It
has also been performed for containment of
subluxed hips in older children, usually over
seven years of age, during the fragmentation
stage of the disease.3-6 For this latter indica-
tion, it is believed to be beneficial in remodel-
ling the hip by preventing subluxation and
stimulating lateral acetabular growth4 in hips
with ‘reducible subluxation’, in which the fem-
oral head is centred within the acetabulum in
abduction without hinge abduction.3 However,
to our knowledge, the factors which predict
these beneficial effects of shelf acetabuloplasty
have not been investigated.
We have assessed hip remodelling after
shelf acetabuloplasty and sought to identify
prognostic factors associated with the mainte-
nance of reduction of the hip and the stimula-
tion of acetabular growth, in the expectation
that this would clarify the indications for this
procedure.
Patients and Methods
The study had institutional review board
approval. We reviewed the medical records and
radiographs of 25 children (22 boys and three
girls) with unilateral Perthes’ disease who had
undergone shelf acetabuloplasty between 1999
and 2005. The indications for surgery were age
seven years or older, ‘reducible subluxation’ of
the hip during the fragmentation stage and
abduction of the hip under anaesthesia of more
than 45° without hinge abduction. Intra-
operative dynamic arthrography was performed
in all hips to exclude those with hinge abduction
in which the centre of movement was located on
the lateral edge of acetabulum and the labrum
was deformed upwards in attempted abduction.
Before shelf acetabuloplasty was performed,
23 hips (92%) had been treated either by an
abduction cast and soft-tissue release with a
simultaneous adductor and psoas tenotomies in
12 and an adductor tenotomy performed in one
and a psoas tenotomy in one. In four hips an
abduction cast without soft-tissue release and in
five skin traction were used. The mean age at
surgery was 8.9 years (7.0 to 12.3) and at the
final follow-up it was 15.6 years (10.2 to 19.6),
giving a mean follow-up of 6.7 years (3.2 to
9.0). All the hips were in the fragmentation
stage at the time of operation. Five were in the
re-ossification stage and 20 in the residual stage
at the final follow-up.
There were ten group-III hips and 15 group-
IV hips according to the Catterall classifica-
tion,7 six group B, ten group B/C border and
nine group C according to the lateral pillar
classification of Herring, Kim and Browne8
and 25 group B according to the classification
of Salter and Thompson.9

VOL. 91-B, No. 10, OCTOBER 2009 1383

1384 W. J. YOO, I. H. CHOI, T.-J. CHO, C. Y. CHUNG, Y.-W. SHIN, S. J. SHIN
Fig. 1
Diagram of the various radiological indices including shelf graft lateral
extrusion (a), shelf gap (b) medial joint space (c), the acetabular depth
(d), the tear-drop thickness (e), the height ratio of the lateral pillar versus
the contralateral epiphysis (f) (f1/f2), the Sharp angle (g) and the centre-
edge angle of Wiberg (h). The acetabular depth growth ratio is the
growth of the true acetabular depth in the affected hip/the growth of
acetabular depth in the contralateral normal hip.
Operative technique. Our surgical technique was the ‘iliac
splitting-expanding shelf’ procedure and is a modification
of the method described by Saito, Takaoka and Ono.10 The
senior surgeon (IHC) performed this procedure in most of
the patients. Briefly, the oblique head of rectus femoris is
divided and the margin of the joint capsule exposed, while
the abductor muscles are left intact. A trough is made with
a small osteotome just above the capsular margin and the
ilium is split in the sagittal plane from the anterior margin
of the pelvis to the mid-point between the anterior gluteal
line and the sciatic notch. The iliac cortical flap hinged
superiorly is raised laterally and a quadrilateral cortical
bone plate harvested from the inner wall of the ilium is
inserted into the trough. Cancellous bone and calcium sul-
phate tablets are packed into the space between the
expanded iliac wall and the inferior bone plate. The oblique
head of rectus femoris is reattached to the direct head to
stabilise the inferior shelf graft. Before closure of the
wound, the hip is taken through a full range of flexion and
abduction to ensure that there is no impingement between
the femoral head and the shelf graft. Post-operatively, bilat-
eral short-leg casts with a connecting bar are applied for a
mean of five weeks (3 to 6).
Evaluation of hip remodelling. This was assessed using
radiographs obtained at the time of operation and at the
final follow-up. Acetabular remodelling was evaluated with
respect to the growth of the lateral acetabular wall with
changes in the lateral protrusion of the shelf graft, the shelf
gap (the vertical distance between the lateral acetabular
margin and the shelf graft), the acetabular depth growth
ratio (growth of the true acetabular depth in the affected
hip/growth of the acetabular depth in the contralateral nor-
mal hip) and the Sharp angle.11 Remodelling of the femoral
head was evaluated using the diameter ratio of the femoral
head (diameter of the affected hip/diameter of the contral-
ateral normal hip) and the Stulberg classification.12
Changes in subluxation were evaluated using the centre-
edge angle of Wiberg13 and the medial joint space ratios
(medial joint space in the affected hip/medial joint space in
the contralateral normal hip) (Fig. 1).
Satisfactory remodelling after acetabuloplasty was
defined as an acetabular depth growth ratio of greater than
100%, an increase in the centre-edge angle and a decrease
in the medial joint space ratio (Fig. 2). We investigated the
relationship between satisfactory hip remodelling and the
following pre-operative clinical and radiological variables,
namely, the Catterall group, the lateral pillar group, the
height ratio of the lateral pillar versus the contralateral epi-
physis (height of lateral pillar/maximal height of the epi-
physis on the contralateral normal side), the head sphericity
index (distance between the flattened head and the best-fit
circle centre of the femoral head/radius of the best-fit circle
of the femoral head (Fig. 2c), the age at the onset of symp-
toms, the age at the time of operation, the Hilgenreiner-
labral angle,14 the tear-drop thickness, the pre-operative
medial joint space ratio and the centre-edge angle. The
Hilgenreiner-labral angle was defined as the angle between
the line parallel to the Hilgenreiner line and the line con-
necting the lateral acetabular margin and the labral tip on
intra-operative dynamic arthrograms (Fig. 2b). It was mea-
sured with the hips in neutral and in abduction between 30°
and 45°.
Statistical analysis. The Wilcoxon signed-rank test was per-
formed to evaluate the significance of differences between
radiological measurements at the time of operation and
those at the final follow-up. The Mann-Whitney U, the chi-
squared and the linear logistic regression tests were used to
identify prognostic factors related to satisfactory hip remod-
elling. A p-value of < 0.05 was considered to be significant.
Results
Subluxation of the femoral head decreased significantly
after shelf acetabuloplasty. The mean medial joint space
ratio of the affected hips versus the contralateral normal
hips decreased from 203% (61% to 417%) pre-operatively
to 143% (39% to 213%) at the final follow-up
(p = 0.00011). The mean centre-edge angle increased from
6.1° (-21° to +20°) pre-operatively to 14.3° (-5° to +28°) at
the final follow-up (p = 0.00036). A significant increase in
the centre-edge angle occurred despite a mean increase
in the diameter ratio of the femoral head of affected hips
versus that of the contralateral hips from 111%
(96% to 132%) pre-operatively to 122% (104% to 141%)
at the final follow-up (p = 0.00014). According to the Stul-
berg classification at the final follow-up one hip was in group
1, eight in group 2, 13 in group 3 and three in group 4. The
THE JOURNAL OF BONE AND JOINT SURGERY
 SHELF ACETABULOPLASTY FOR CHILDREN WITH PERTHES’ DISEASE AND REDUCIBLE SUBLUXATION OF THE HIP 1385

Fig. 2a Fig. 2b Fig. 2c

Fig. 2d Fig. 2e

Plain radiographs of an eight-year-old boy with a satisfactory outcome showing a) a subluxed femoral head in the fragmentation stage and b) the
Hilgenreiner-labral angle (the angle between the line parallel to the Hilgenreiner line (a) and that connecting the lateral acetabular margin and the
labral tip (b) on the intra-operative dynamic arthrogram) which was 28° in the neutral position and 46° in abduction. The labrum was assessed as
‘uncomfortable’ in the neutral position, but as ‘comfortable’ in abduction. c) The head sphericity index (defined as the ratio of the distance between
the flattened femoral head and the best-fit circle centre (a) to the radius of the best-fit circle (b) on the intra-operative arthrogram which was 0.73. d)
An abduction cast with tenotomy of adductor and psoas was followed by shelf acetabuloplasty. e) Nine years after surgery there is a round femoral
head with Stulberg group-2 deformity, increased acetabular depth of the true acetabular rim (arrow) suggestive of stimulated lateral acetabular
growth (acetabular depth growth ratio, 124%), a medial joint space ratio reduction of 59% and an increase in the centre-edge angle of 8°.

acetabulum also showed marked remodelling after shelf
acetabuloplasty. The acetabular depth growth ratio of the
affected hips versus the contralateral normal hips was a
mean of 138% (58% to 237%). Growth of the lateral
acetabular wall resulted in a progressive decrease in the
lateral protrusion of the shelf graft and the shelf gap. The
mean incorporation rate of the shelf graft into the lateral
acetabular wall was 1.4 mm/year (0.2 to 3.9). Deepening of
the acetabulum was calculated by subtracting the shelf-
graft incorporation (lateral wall growth) value from the
increase in acetabular depth. The mean acetabular deepen-
ing rate was 0.8 mm/year (0 to 3.5). The mean Sharp angle
decreased from 49.1° (41° to 57°) pre-operatively to 45.2°
(40° to 57° at the final follow-up (p = 0.0001) and that of
the contralateral hips from 46.4° (40° to 52°) to 41.0°
(34° to 52°) (p = 0.000036). The ratio of the Sharp angle of
the affected hips versus the contralateral normal hips did
not change significantly after shelf acetabuloplasty
(p = 0.052; Table I).
When we divided the patients into two groups based on
satisfactory or unsatisfactory hip remodelling, satisfactory
remodelling was found to have been achieved in 18 hips
(72%). In this group, the acetabular depth growth ratio was
a mean of 164% (111% to 237%), the mean centre-edge
angle increased by 12° (5° to 34°) and the mean medial joint
space ratio decreased by 81% (48% to 237%). In the unsat-
isfactory remodelling group the acetabular depth growth
ratio was a mean of 73% (58% to 89%), the mean centre-
edge angle increase -0.3° (-8° to +8°), and the decrease in the
mean medial joint space ratio was 67% (17% to 98%).
Univariate analysis revealed that the height ratio of the
lateral pillar versus the contralateral epiphysis and the
Hilgenreiner-labral angle in hip abduction were statistically
significant prognostic factors. The mean height ratio of the
lateral pillar versus the contralateral epiphysis (SD) was
0.64 (SD 0.14) in hips with satisfactory remodelling, but
0.46 (SD 0.11) in hips with unsatisfactory remodelling
(p = 0.0061). The mean Hilgenreiner-labral angle (SD) of

VOL. 91-B, No. 10, OCTOBER 2009

1386 W. J. YOO, I. H. CHOI, T.-J. CHO, C. Y. CHUNG, Y.-W. SHIN, S. J. SHIN
Table I. Acetabular remodelling after shelf acetabuloplasty
At the time of surgery
Mean lateral protrusion of shelf graft in mm (range)
Mean shelf gap in mm (range)
Mean sharp angle ratio (%)†
* Wilcoxon signed-rank test
† ratio of affected hips to contralateral normal hips
abducted hips with satisfactory hip remodelling was
47° (SD 11), but only 27° (SD 7) in abducted hips with unsat-
isfactory remodelling (p = 0.0022). For multivariate analysis,
we divided the type of labrum into two groups based on the
Hilgeneriner-labral angle as follows: ‘a comfortable’ labrum in
which the Hilgenreiner-labral angle was > 35° and an ‘uncom-
fortable’ labrum in which it was ≤ 35°. Multivariate analysis
using the linear logistic regression test showed that only the
type of labrum in hip abduction correlated significantly with
satisfactory hip remodelling (p = 0.012). Specifically, a com-
fortable labrum (Hilgenreiner-labral angle > 35°) at a hip
abduction of between 30° and 45° was found to be correlated
with satisfactory hip remodelling with increased acetabular
depth growth and the prevention of hip subluxation (Fig. 2).
Discussion
Shelf acetabuloplasty is known to be effective for remodel-
ling subluxed but reducible hips during the early stage of
Perthes’ disease in older children, usually over seven years
of age. In this group, femoral varus osteotomy carries the
risk of limb shortening with coxa vara, and innominate pel-
vic osteotomy fails to reduce a subluxed femoral head.3 This
surgical indication was described as ‘reducible subluxation’
by Daly et al.3 They confirmed the reduction of the femoral
head by dynamic arthrography and stated that shelf acetab-
uloplasty was contraindicated in unstable hips with hinge
abduction. Domzalski et al4 also performed intra-operative
dynamic arthrography to exclude non-reducible hips in
which the epiphysis could not be contained in the acetabu-
lum and the labrum tip moved upwards on attempted
abduction. Bursal and Erkula15 used MRI and physical
examination to exclude hips with hinge abduction. We also
excluded hips with hinge abduction using intra-operative
dynamic arthrography, and our typical indication for shelf
acetabuloplasty was hips with ‘impending impingement’, in
which the subluxed femoral head could be reduced fol-
lowed by abduction casting with or without soft-tissue
release (18 hips) or traction (5 hips).
Our findings are consistent with those reported by Daly et al3
in terms of the surgical indications, the age at surgery and
favourable results. These authors left the question as to why
some patients did badly unanswered. Our study advances
their findings by identifying the variables associated with
lateral acetabular growth stimulation after shelf acetabulo-
plasty. These relate to the type of labrum in hip abduction. In
hips with reducible subluxation, a ‘comfortable’ labrum in
Final follow-up p-value*
17.3 (12 to 25) 9.8 (4 to 20) 0.000012
5.6 (3 to 9) 2.1 (0 to 5) 0.000025
106.1 (96.0 to 122.5) 111.3 (90.9 to 145.5) 0.052
hip abduction implies that the lateral part of cartilaginous
femoral head is relatively round and does not impose pres-
sure on the lateral edge of the acetabulum, as long as reduc-
tion of the femoral head is maintained by the surgery. On the
other hand, an ‘uncomfortable’ labrum in hip abduction may
be a manifestation of compression of the lateral acetabular
edge by a flattened and hypertrophied lateral part of the car-
tilaginous femoral head. In these hips, the femoral head may
be containable within the acetabulum only in extreme
abduction without frank hinge abduction under anaesthesia.
In growing children, acetabular depth increases as a
result of growth at the periphery of the acetabular cartilage
and the periosteal formation of new bone at the lateral
pelvic wall, whereas the acetabular height and width
increase because of growth of the triradiate cartilage.16 Our
observation that the growth of acetabular depth of hips
with an ‘uncomfortable’ labrum is only 73% of that of the
contralateral normal side suggests that femoral heads in
these hips may sublux again from the true acetabulum after
surgery, despite intra-operative reduction, and that subse-
quent pressure on the acetabular periphery may lead to the
inhibition of growth of the acetabular cartilage and its
growth plate. Based on our observations, shelf acetabulo-
plasty appears to be inadequate in terms of stimulating lat-
eral acetabular growth in hips with an ‘uncomfortable’
labrum. In this situation, a redirectional osteotomy which
enables the lateral acetabular margin, including the labrum,
to rotate to achieve containment of the femoral head may
be more beneficial for hip remodelling and, furthermore, it
may also substantially reduce the risk of labral tear.
Our operative technique has some advantages compared
with those previously described.17-19 First, immediate sta-
bility of the shelf graft can be expected, because it is sup-
ported firmly by the expanded lateral acetabular wall under
which cancellous bone and calcium sulphate tablets are
packed tightly. In addition, the abductor muscles are not
damaged during the operation, because graft is harvested
from the inner wall of the ilium. Therefore, post-operative
immobilisation is limited to the use of bilateral short-leg
casts with a connecting bar for three to six weeks, allowing
patients to sit comfortably. In previous studies hip spica
casts have been used for longer periods of time.3-5,18,19
The limitations of our study require consideration. First,
it is a cohort study without a control group. However, in
subluxed Perthes’ hips with impending impingement in
older children, it is, in our experience difficult to maintain
THE JOURNAL OF BONE AND JOINT SURGERY
 SHELF ACETABULOPLASTY FOR CHILDREN WITH PERTHES’ DISEASE AND REDUCIBLE SUBLUXATION OF THE HIP
the reduced position after abduction casting, soft-tissue
release or skin traction without a subsequent containment
bony procedure. This is why we could not set up a control
group in our study. Secondly, the number of patients
recruited was not large enough to achieve strong statistical
power, and thus a larger study may reveal that prognostic
factors found to be insignificant in our study are in fact
relevant. For example, the height ratio of the lateral pillar
versus the contralateral epiphysis was found to be a statisti-
cally significant prognostic factor by univariate analysis, but
this significance was lost during the multivariate analysis. It
may be proved to be significant in a larger study. Thirdly, the
radiological indices of hip remodelling, including the
acetabular depth ratio, were measured using antero-
posterior plain radiographs. Although we assumed that the
increase in the acetabular depth may reflect lateral acetabu-
lar growth stimulation, two-dimensional assessment is
inadequate to determine the real configuration of the
remodelled acetabulum. In future, it may be possible to elu-
cidate the mechanism of lateral acetabular growth stim-
ulation after a labral support procedure by advanced three-
dimensional MRI techniques which can detect macro-
molecular changes during cartilage growth or degeneration.
In summary, shelf acetabuloplasty for hips with reducible
subluxation was found to produce favourable hip remodel-
ling in older children with early Perthes’ disease. In addi-
tion, we were able to identify a prognostic factor associated
with increased acetabular wall growth and the prevention
of hip subluxation. Shelf acetabuloplasty as a containment
procedure seems to be best indicated for subluxed hips in
which epiphyseal collapse is not advanced and the extruded
epiphyseal segment can slip easily underneath the labrum
on attempted abduction without imposing undue pressure
on the lateral edge of the acetabulum.
No benefits in any form have been received or will be received from a commer-
cial party related directly or indirectly to the subject of this article.
VOL. 91-B, No. 10, OCTOBER 2009
1387
References
1. Kruse RW, Guille JT, Bowen JR. Shelf arthroplasty in patients who have Legg-
Calvé-Perthes disease: a study of long-term results. J Bone Joint Surg [Am]
1991;73-A:1338-47.
2. Reinker KA. Early diagnosis and treatment of hinge abduction in Legg-Perthes dis-
ease. J Pediatr Orthop 1996;16:3-9.
3. Daly K, Bruce C, Catterall A. Lateral shelf acetabuloplasty in Perthes’ disease: a
review of the end of growth. J Bone Joint Surg [Br] 1999;81-B:380-4.
4. Domzalski ME, Glutting J, Bowen JR, Littleton AG. Lateral acetabular growth
stimulation following a labral support procedure in Legg-Calve-Perthes disease. J
Bone Joint Surg [Am] 2006;88-A:1458-66.
5. Jacobs R, Moens P, Fabry G. Lateral shelf acetabuloplasty in the early stage of
Legg-Calvé-Perthes disease with special emphasis on the remaining growth of the
acetabulum: a preliminary report. J Pediatr Orthop B 2004;13:21-8.
6. Willett K, Hudson I, Catterall A. Lateral shelf acetabuloplasty: an operation for
older children with Perthes’ disease. J Pediatr Orthop 1992;12:563-8.
7. Catterall A. The natural history of Perthes’ disease. J Bone Joint Surg [Br]
1971;53-B:37-53.
8. Herring JA, Kim HT, Browne R. Legg-Calve-Perthes disease. Part I: classification
of radiographs with use of the modified lateral pillar and Stulberg classifications. J
Bone Joint Surg [Am] 2004;86-A:2103-20.
9. Salter RB, Thompson GH. Legg-Calvé-Perthes disease: the prognostic signifi-
cance of the subchondral fracture and a two-group classification of the femoral
head involvement. J Bone Joint Surg [Am] 1984;66-A:479-89.
10. Saito S, Takaoka K, Ono K. Tectoplasty for painful dislocation or subluxation of
the hip: long-term evaluation of a new acetabuloplasty. J Bone Joint Surg [Br]
1986;68-B:55-60.
11. Sharp IK. Acetabular dysplasia: the acetabular angle. J Bone Joint Surg [Br]
1961;43-B:268-72.
12. Stulberg SD, Cooperman DR, Wallensten R. The natural history of Legg-Calvé-
Perthes disease. J Bone Joint Surg [Am] 1981;63-A:1095-108.
13. Wiberg G. Studies on dysplastic acetabular and congenital subluxation of the hip
joint. Acta Chir Scand 1939;83 (suppl 58):1-135.
14. Hilgenreiner H. Zur Frühdiagnose und Frühbehandlung der angeborenen Hüftge-
lenksverrenkung. Med Klin 1925;21:1385-8, 1425-9.
15. Bursal A, Erkula G. Lateral shelf acetabuloplasty in the treatment of Legg-Calvé-
Perthes disease. J Pediatr Orthop B 2004;13:150-2.
16. Ponseti IV. Growth and development of the acetabulum in the normal child: ana-
tomical, histological, and roentgenographic studies. J Bone Joint Surg [Am]
1978;60-A:575-85.
17. Gill AB. Plastic construction of an acetabulum in congenital dislocation of the hip:
the shelf operation. J Bone Joint Surg 1935;17:48-59.
18. Staheli LT, Chew DE. Slotted acetabular augmentation in childhood and adoles-
cence. J Pediatr Orthop 1992;12:569-80.
19. Wilson JC Jr. Surgical treatment of the dysplastic acetabulum in adolescence.
Clin Orthop 1974;98:137-45.