ULTRASOUND AND CONGENITAL DISLOCATION OF

THE HIP

THE IMPORTANCE OF DYNAMIC ASSESSMENT

L. B. ENGESAETER, D. J. WILSON, D. NAG, M. K. D. BENSON

From Nuffield Orthopaedic Centre, Oxford

One-hundred newborn children at high risk of hip instability were prospectively assessed clinically and
by ultrasound. The decision to treat was based only on the clinical examination. At the age of three months
all the children were evaluated clinically and with an anteroposterior radiograph of the pelvis. None of the
standard ultrasound measurements of acetabular depth and femoral head cover correlated with the outcome
at three months. Dynamic assessment of stability was the only ultrasound technique that had a significant
relation with outcome.

Early diagnosis and treatment of congenital dislocation the clinical assessment alone, following a well-established
and dysplasia of the hip (CDH) are important to ensure protocol (Bradley, Wetherill and Benson 1987). Subse-
a successful outcome. Early diagnosis remains difficult: quent analysis allowed us to judge whether early
neither clinical examination nor plain radiographs of the ultrasound evaluation should have modified our treat-
pelvis are reliable indicators of dysplasia of the hip ment.
(Zieger and Schulz 1987). Ultrasound as a diagnostic aid
was popularised by Graf(1980, 1981); many subsequent
PATIENTS AND METHODS
papers have claimed the usefulness of this technique for
early diagnosis (Clarke et al 1985 ; Clarke 1986 ; Langer One-hundred infants aged under three months were
1987). investigated. They were consecutive referrals to a special
Because there is now confidence in ultrasound children’s hip clinic. Each was referred with a clinical
assessment, protocols for treatment have been described, abnormality, a history of breech delivery, or a family
dependent upon geometric analyses of the ultrasound history of congenital hip disease. Patients with known
picture. We are concerned that the natural history of the chromosomal disorders or neuromuscular diseases were
maturing neonatal hip has not been established ultrason- excluded. Most patients were referred from the maternity
ically. Furthermore it is not clear whether accurate unit of the John Radcliffe Hospital, Oxford, the remain-
predictions ofoutcome can be made from early ultrasound der being referred by general practitioners or other
examination. Since treatment carries risks ofhip damage, specialists. The babies were born between June 1987 and
care must be taken to ensure that unnecessary therapy is April 1988. In this period 5 079 babies were born at the
avoided. John Radcliffe Hospital and a total of 125 were referred
We have used ultrasound to examine prospectively to the hip clinic. The first 100 of these babies with
the hips of 100 babies referred with possible CDH. The complete clinical data, ultrasonic and radiographic
decision to observe or to treat these babies was made on measurements were included in this study.
At the infant’s first attendance details of the
pregnancy, labour, family history, and the clinical
L. B. Engesaeter, MD, PhD, Honorary Registrar in Orthopaedics findings were recorded. On the basis of the orthopaedic
D. J. Wilson, MRCP, FRCR, Consultant Radiologist
D. Nag, FRCR, Senior Registrar in Radiology examination each hip was designated as normal, clicking,
M. K. D. Benson, FRCS, Consultant Orthopaedic Surgeon subluxatable, dislocatable, reducibly dislocated, or irre-
The Nuffield Orthopaedic Centre, Windmill Road, Headington, Oxford
OX3 7LD, England ducibly dislocated (Bradley et al 1987). Children with
Correspondence should be sent to Dr. D. J. Wilson irreducibly dislocated hips were not treated in early
infancy. Those with reducibly dislocated hips were
© 1989 British Editorial Society of Bone and Joint Surgery
030l-620X/90/2025 $2.00 splinted with either von Rosen or Pavlic appliances.
JBoneJointSurg[Br] 1990; 72-B: 197-201. Children with subluxatable or dislocatable hips were

VOL. 72-B, No. 2, MARCH 1990 197

198 L. B. ENGESAETER, D. J. wILsoN, D. NAG, M. K. D. BENSON

splinted only if they did not gain clinical stability by

three weeks of age.
Ultrasonography was performed at the first visit.
The technique used was based on the principles described
by Graf (1984, 1986). A Diasonics DRF-100 realtime
sector scanner was used with a 10 MHz transducer (Les
Ulis, France). The ultrasound examination was per-
formed in the coronal plane with the transducer posi-
tioned on the lateral aspect ofthe flexed, slightly abducted
hip (Fig. 1). This method produces a sonogram that
resembles an anteroposterior radiograph of the right hip
(Figs 2 to 5). Two images were recorded for each hip. For
stability testing the hip was slightly adducted, exerting
longitudinal pressure along the thigh (Barlow’s test).
Movement of the femoral head of more than 1 mm away
from the acetabulum indicated ultrasonic instability.
Graf’s classification (1986) of hip dysplasia is based
on morphological aspects and angle measurements (alpha
and beta angles) (Fig. 6). The original classification
consisted of four major types which have since been
subdivided (Table I) (Langer 1987). To test reproducibil-
ity of the measurement all angles were remeasured by the
same observer at a later date. In addition the amount of
the femoral head covered by the bony acetabulum was
Fig. 1
measured on the sonograms, by the method described by
Standard scanning position used for testing stability.
Gentle pressure is exerted along the femur while it is
beinglifted with the fingersofthe hand holding the thigh.
The hand holding the probe is also used to stabilise the
infant’s back.

Figure 2 - A diagram of a normal sonogram. Figure 3 - A normal

sonogram. Figure 4 - Sonogram showing a shallow acetabulum (Graf
type lIlA). Figure 5 - Sonogram showing a subluxated hip (Graf type
IV).

NORMAL
Fig. 2

Fig. 3 Fig. 4 Fig. 5

THE JOURNAL OF BONE AND JOINT SURGERY

 UlTRASOUND AND CONGENITAL DISLOCATION OF THE HIP 199

reduction and was repeated at splint removal. All

children, even those found to be clinically normal, were
reviewed and radiographed at three months. Plain
anteroposterior radiographs were used to assess the
alignment, centreing of the femoral head, and the
maturity of the bony roof measured as an acetabular
index (Morin et al 1985).
The babies were then classified in four groups by
outcome:
Group 1. Normal hips on clinical examination (stable
with full movement); pelvic radiograph showing nor-
mally centred hips with acetabular angles of 25#{176}
or less.
Group 2. Normal hips on clinical examination ; radiologi-
cally immature with acetabular angles greater than 25#{176}.
Group 3. Hips which had been splinted for clinical
instability, but at three months were clinically normal
Fig. 6 and radiologically mature with acetabular angles of 25#{176}
Diagram of sonogram with the lines drawn to measure or less.
Graf’s angles. Group 4. Hips in which further treatment was necessary
because of clinical instability or radiological dysplasia.
Coefficients of correlation were calculated with the chi-
squared test or Fisher’s exact test for contingency tables
with smaller numbers.

RESULTS

There were 64 girls and 36 boys. Delivery had been

induced in 29 and a Caesarean section done in 29. The
presentation was by extended breech in 30 and by breech
in eight. No correlation could be found between these
variables and the final outcome. The length of labour,
the sequence ofthe child in the family, or a family history
of CDH were unrelated to outcome.
Most babies were referred by paediatricians. Table
II shows the diagnosis of the referring physician. One or
two weeks usually elapsed between referral and ortho-
paedic assessment. The high spontaneous resolution of
a b minor instability is reflected in the finding of more stable
Fig. 7 hips by the orthopaedic surgeon (Barlow 1962). Fifteen
ofthis high risk group of 100 babies were splinted, 14 in
Diagrams of sonograms with lines drawn to measure (a) the
ratio of acetabular depth to femoral head size (A I /A2) and (b)
von Rosen splints and one in a Pavlic harness. Thus the
the ratio representing the amount of femoral head covered by treatment incidence was about 3 per 1 000.
the labrum (Bl/B2).
The mean age for the first ultrasound examination
was 25 days (range 5 to 81). Tables III and IV show the
Morin, Harcke and MacEwen (1985) and Zieger, Hilpert results of the ultrasound measurements based on Graf’s
and Schulz (1986) (Fig. 7). The amount of femoral head criteria (1984).
covered by the labrum was expressed as a ratio calculated The average difference between the original meas-
as follows : two lines were drawn parallel to Graf’s base urement of the bony acetabular angle (alpha) made at the
line, one through the tip of the labrum and the other clinic examination and that at review was 2.3#{176}
(s.d. 10#{176}).
through the most lateral aspect of the femoral head. The difference in the cartilage roof measurements was
Distances were measured at right angles to the base line greater, with a mean of3.6#{176}
(s.d. 16#{176}).
and expressed as ratios. The pelvis was radiographed at a mean age of 98
Follow-up after the initial assessment depended on days (range 73 to 205). Table V shows the radiographic
the clinical problem. Children in splints were regularly measurements.
monitored. Those with mild clinical instability were re- As a result ofthe clinical and radiographic evaluation
assessed both clinically and ultrasonically. Splintage was 54 babies were discharged after hip normality was
checked weekly. Ultrasound examination confirmed hip confirmed at three months (group 1); 30 had clinically

VOL. 72-B, No. 2, MARCH 1990

200 L. B. ENGESAETER, D. J. WILSON, D. NAG, M. K. D. BENSON

normal hips, but bony immaturity on radiographs Table I. Classification of hip dysplasia

indicated the need for further follow-up (group 2); 12 had

Type Description
been splinted for six to 12 weeks during the first three
months, but then proved to have clinically and radio- IA/B Mature hip

graphically normal hips (group 3). Four had hips that hA Immature hip (normal finding up to age ofthree months); indicating
retarded development of the bony edge of the acetabulum
needed further treatment at three months (group 4); in
all four a further period in a Pavlic harness was judged IIB Retarded ossification of the cartilaginous edge in infants aged three
necessary. months and older

All hips classified as Graf type I had a successful IIC ‘Critical zone’. Delayed ossification of the cartilaginous edge, with
a deficient bony acetabular contour
outcome, as did many oftypes II and III. However, there
was no statistically significant correlation between any lID Decentring hip ; lateralisation of the femoral head
of the ultrasound measurements and the final outcome.
lilA Eccentric hip; the cartilaginous edge is displaced but normal in
Specifically, the Grafangles, Graftype, Morin acetabular echo pattern
depth measurements, and labrum cover ratios showed no
IIIB Eccentric hip ; the cartilaginous edge is compressed with an altered
correlation with the radiographic and clinical findings at echo pattern suggesting structural alteration
three months. This reflects the frequency of a satisfactory
IV Eccentric hip ; the cartilaginous edge is dislocated downwards
outcome despite a poor ultrasound classification. Dy- behind the femoral head
namic ultrasound testing correlated with the orthopaedic
surgeon’s initial clinical assessment for the left hip
(p < 0.01), but not with that ofthe right hip. The results
ofthe dynamic component ofthe ultrasound examination
Table II. Assessment of clinical stability by referring
for both hips showed a strong predictive value when
physician and by orthopaedic surgeon
compared with the outcome. Stable hips had a signifi-
cantly better outcome than unstable hips (p < 0.00 1). Referring physician Orthopaedic surgeon

Right Left Right Left

DISCUSSION Normal 61 48 72 53

There has been increasing recognition that clinical Click 18 24 15 24

examination alone, even in the hands of dedicated Subluxatable 6 10 3 10

experts, will not detect all cases of hip instability at birth.
Dislocatable 5 10 1 2
The diagnosis may be even more elusive in the days and
weeks after delivery. Ultrasound promised, simply and Dislocated 2 2 3

non-invasively, to increase the accuracy of diagnosis,

thereby reducing the number of babies treated and also
the number that slip through the screening net (Berman
and Klenerman 1986). Our study does not address the notable that there was a fairly wide scatter, with
question of population screening, rather it concerns the differences of 10% for the roof angle and 16% for the
value of ultrasound criteria in predicting the clinical cartilaginous labrum angle being within one standard
outcome. To this end we deliberately selected a high-risk deviation of the mean.
group and the overall incidence of abnormality should be We did not find any significant correlation between
taken in the context ofthe 5 079 children born during the the Graftypes and outcome. Graf(l986) claims that hips
study period. Several techniques for analysing the hip by oftype IIC or worse have such poor osseous development
ultrasound have been evaluated. In our hands, only (alpha > 43#{176})
that they require treatment irrespective of
ultrasound dynamic stability testing assisted the clinical age. This practice seems also to be that recommended by
decision whether to treat or just observe a newborn with Langer (1987). If we had followed this advice 80 infants
suspect hips. would have been treated. However, on the basis of the
Ifthe numbers in the various Graftypes in our study clinicalexamination, we found that only 15 babies needed
(Table IV) are compared with thatofa normal postpartum treatment.
distribution (Langer 1987), it is clear that we achieved Graf’s (1986)classification system, with eight groups
our aim of studying a group at high risk of hip instability. of sonographic appearance, is complicated. Therefore,
Langer reported 0.85% of hips with Graf-type IIC or other ways of assessing the hip by ultrasound have been
worse in an unselected population ; we found 64% of the developed. The ratio that expresses the proportion of the
hips to be IIC or worse. Allowing for the total number of femoral head covered by the bony acetabulum measured
deliveries from which our patients were selected, the in our study has been used by Morin et al (1985) and
figures are much the same. The average difference Terjesen, Bredland and Berg (1988). They claim that this
between the Graf angles measured on review and those ratio is equally accurate and easier to measure. The
from the original observations was small. However, it is unstable hips in our study did not have a significantly

THE JOURNAL OF BONE AND JOINT SURGERY

 ULTRASOUND AND CONGENITAL DISLOCATION OF THE HIP 201

Table III. Review of ultrasound evaluation TableIV. Numberof Table V. Evaluation of radiographs
hips in Graf groups taken at age three months
Right Left Right Left
side side Right Left
de de side side
Inverted labrum
Yes 2 3 IA 5 7 Hip position
No 98 97 Normal 94 94
lB 4 7 Doubtful 5 5
Ossific nucleus Subluxated 1 1
Yes 0 0 IIA 25 26
No 100 100 Acetabular index
IIC 12 9 Mean 24.7 24.3
Grafangle (mean) S.d. 5.9 5.7
Alpha 44 44 D 9 8
Beta 77 78 Ossific nucleus
lIlA 26 29 Yes 24 24
Dynamic testing No 76 76
Stable 86 89 IIIB 19 14
Unstable 14 11
Cover of femoral head IV 0 0
Acetabulum Al :A2 (mean) 0.36 0.36
Labrum Bl : B2 (mean) 0.63 0.64

lower ratio than the stable ones, which is contrary to the hand. Ultrasound highlights minimal displacement : 1 to
findings of Terjesen et al (1988). The ratio depicting 2 mm of movement are easily detectable.
cover of the femoral head by the cartilaginous labrum Our study suggests that in early infancy Graf’s
likewise bears no relation to clinical outcome. classification and Morin’s ratios do not reliably predict
We could not show any significant correlation the need for splintage. In our hands, the only ultrasound
between the alpha angle on the initial ultrasound and the assessment which related to the clinical outcome was
acetabular index at three months, which is in accordance dynamic testing ofthe hip. This concurs with the findings
with the findings of Morin et al (1985). of Saies, Foster and Lequesne (1988).
Although there was a significant correlation between This study was supported by a grant from the Oxford Health Authority.
clinical and ultrasonic examination of the left hip this The ultrasound equipment was in part purchased by the Arthritis and
Rheumatism Council for Research. We thank the research secretary
was not matched in the right hip. A right-handed Jean Glynn for her help in collecting the data and preparing the script.
No benefits in any form have been received or will be received
observer may be less accurate in detecting minor from a commercial party related directly or indirectly to the subject of
instability of the hip examined by the non-dominant this article.

REFERENCES

Barlow TG. Early diagnosis and treatment of congenital dislocation of Graf R. Sonographic diagnosis of hip dysplasia and hip dislocation.
the hip. JBoneJoint Surg[Br] 1962; 44-B :292-301. Revised from reprint
by Schuler P. Graft R. Sonographie in der
Orthop#{228}die in Braun - Gunther - Schwerk Ultrascha/idiagnostik
Berman L, Kienerman L. Ultrasound screening for hip abnormalities:
Ecomed Verlag: 1986; 4.Erg.Lfg. 7.
preliminary findings in 1001 neonates. Br MedJl986; 293 :719-22.
Langer R. Ultrasonic investigation of the hip in newborns in the
Bradley J, Wetberill M, Benson MKD. Splintage for congenital
diagnosis of congenital hip dislocation : classification and results
dislocation of the hip: is it safe and reliable? J Bone Joint Surg
ofa screening program. Skeletal Radiol 1987; 16:275-9.
[Br] 1987; 69-B :257-63.
Clarke NM. Sonographic clarification of the problems of neonatal hip Mona C, Harcke HT, MacEwen GD. The infant hip : real-time US
instability. J Pediatr Orthop 1986; 6:527-32. assessment of acetabular development. Radiology 1985 ; 157:
673-7.
aarke NMP, Harcke HT, McHugh P, et al. Real-time ultrasound in
the diagnosis of congenital dislocation and dysplasia of the hip. J Sales AD, Foster BK, Lequesne GW. The value of a new ultrasound
BoneJoint Surg[Br] 1985; 67-B :406-12. stress test in assessment and treatment of clinically detected hip
instability. J Pediatr Orthop 1988 ; 8:436-41.
Graf R. The diagnosis of congenital hip-joint dislocation by the
ultrasonic Combound treatment. Arch Orthop Trauma Surg 1980; Terjesen 1, Bredland 1, Berg V. A new measuring method for the
97:117-33. evaluation of newborns’ hip joints by ultrasound. Acta Orthop
Scand 1988; 59:112-3.
Graf R. The ultrasonic image of the acetabular rim in infants : an
experimental and clinical investigation. Arch Orthop Trauma Surg Zieger M, Hilpert S, Schulz RD. Ultrasound of the infant hip. Part I.
1981 ; 99:35-41. Basic principles. Pediatr Radio! 1986; 16:483-7.
Graf R. Classification of hip joint dysplasia by means of sonography. Zieger M, Schulz RD. Ultrasonography of the infant hip. Part III:
Arch Orthop Trauma Surg 1984; 102:248-55. Clinical application. Pediatr Radio! 1987; 17:226-32.

VOL. 72-B, No. 2, MARCH 1990