Developmental dysplasia of the hip

What is developmental dysplasia of the hip?

The term refers to a spectrum of pathology, ranging from mild acetabular dysplasia with a stable hip through more severe forms of dysplasia, often associated with neonatal hip instability, to established hip dysplasia with or without later subluxation or dislocation. The condition used to be known as congenital dislocation of the hip (CDH), but the term developmental dysplasia of the hip (DDH) better reflects both the variable presentation and the potentially progressive nature of the condition.^w5 Typically, cases present within the neonatal period, but some cases are recognised later (fig 1⇓).

Fig 1 Anteroposterior pelvic radiograph of a 2 year old who presented late with bilateral hip dislocations despite maternal concerns expressed from the first week of life. Hilgenriener’s line (H) is drawn through both tri-radiate cartilages; Perkins’ line (P) is perpendicular to this and drawn through the most superolateral point of the ossified acetabulum. The ossific nucleus of the femoral head (when present) should be in the inferomedial quadrant formed by the intersection of these two lines; in the dysplastic or dislocated hip the ossific nucleus lies in the superolateral quadrant

• Download figure
• Open in new tab
• Download powerpoint

This review concentrates on “idiopathic” developmental dysplasia of the hip. The management of dislocation secondary to neuromuscular conditions is outside the scope of this article.^w6-w8

How common is it?

Precise prevalence varies with age and definition of hip dysplasia. A worldwide systematic review of unscreened populations estimated the prevalence of clinically diagnosed, established hip dysplasia to be 1.3 per 1000.2 In populations screened clinically with Ortolani and Barlow tests, the prevalence is higher (1.6-28.5 per 10002), and it is higher still with ultrasound screening that uses a morphological definition of abnormality.^w9

What do we know about the causes?

The precise aetiology of developmental hip dysplasia is unknown, but genetic and environmental factors may act as internal or external influences, respectively (box 1). The left hip is most often affected; 20% of cases are bilateral and 80% are in girls.^w10 Findings from twin and family studies suggest high heritability consistent with a strong genetic susceptibility to onset of disease but not necessarily to progression or severity.6 Having a sibling with hip dysplasia increases risk by 5%.6

Vaginal delivery of babies with breech presentation is associated with a 17-fold increased risk of hip dysplasia, compared with a sevenfold increase for breech babies delivered by elective caesarean section.7

Why is early detection important?

The pathology of hip dysplasia changes with time, and so does management. In neonates, soft tissues are lax: a dislocated hip may be reduced by simple manipulation during clinical examination, and stabilisation occurs as the soft tissues tighten. An unstable or dislocatable hip may also stabilise spontaneously. When the femoral head is aligned with the centre of the acetabulum, the dysplastic acetabulum often normalises within the first months of life. If the hip remains dislocated, soft tissue contractures develop rapidly, and surgery is likely to be required to obtain and maintain joint reduction. The longer the hip is left in an abnormal position, the more the anatomy changes, developing abnormalities of both the proximal femur and the acetabulum such that, after the age of 18 months, one or both bones may need surgical correction to provide joint congruity and stability.

Treatment outcomes are difficult to interpret in the absence of a randomised control group, but most observational studies, despite their methodological weaknesses, have shown that early treatment of those most severely affected is important for a good outcome. There is no clear definition of a late diagnosis: we define diagnosis after 6-8 weeks as late since the success rate of simple, conservative treatment falls significantly after age 7 weeks.8 9

How is the diagnosis made?

Clinical examination

The 1986 guidelines of the Standing Medical Advisory Committee (SMAC) state that all babies should be screened clinically for congenital hip dislocation within 24 hours of birth; before hospital discharge; at 6 weeks; between 6-9 months; and at walking age.11 These guidelines were written before the change in nomenclature from congenital to developmental dysplasia and before ultrasound screening was available.

Early period

The Ortolani and Barlow tests, performed as part of the examination at birth and at 6 weeks, are used to detect hip instability. Signs can be difficult to elicit: the baby must be relaxed and lying without its nappy. The Ortolani test aims to relocate a dislocated femoral head; the Barlow test attempts to dislocate an articulated femoral head (fig 2⇓). Both tests will detect an unstable hip, but they will not detect a dislocated, irreducible hip (it is best detected by identifying limited abduction of the flexed hip) or a stable hip with abnormal anatomy, such as acetabular dysplasia (which is best detected by ultrasound examination). Benign “hip clicks” resulting from soft tissues snapping over bony prominences during hip movement should be distinguished from the “clunks” produced during the Ortolani manoeuvre as the dislocated femoral head is reduced and from the subluxation felt during the Barlow test. These tests are useful in neonates (table 1⇓), but they become difficult by 2-3 months of age owing to the development of muscle tone and soft tissue contractures.

Fig 2 Barlow test (left) and Ortolani test (right). In the Barlow test (baby’s right hip), the hip is adducted and flexed to 90°; the examiner holds the distal thigh and pushes posteriorly on the hip joint. The test is positive when the femoral head is felt to slide posteriorly as it dislocates. In the Ortolani test (baby’s left hip), the pelvis is stabilised by the examiner and each hip examined separately. In a baby with limited hip abduction in flexion, the hip is flexed to 90° and gently abducted while the examiner’s finger lifts the greater trochanter. In a positive test the femoral head is felt to relocate into the acetabulum. The dislocated femoral head (pictured on the opposite hip) has now been reduced back into joint12

• Download figure
• Open in new tab
• Download powerpoint

Table 1

 Developmental hip dysplasia in neonates

View this table:

• View popup
• View inline

Late diagnosis

Limited hip abduction when the hip is flexed to 90° is the single most important sign of a dislocated or dysplastic hip. Most relaxed infants can abduct the flexed hip fully (the examiner’s hand, holding the thigh, can touch the examination couch). Asymmetrical movement should alert the examiner to a potential problem, and limited abduction is a useful sign in cases of bilateral hip dysplasia, where asymmetry is absent.

The Galleazzi test identifies a “short” thigh (fig 3⇓). The child is supine with hips and knees flexed to 90° and the height of each knee is compared. Unilateral femoral shortening may signify hip dislocation or rarer abnormalities of the femur. False negative results may occur when bilateral hip dysplasia is present or when the pelvis is not level as when the nappy is left on, for example.

Fig 3 Galeazzi test: the child’s hips and knees are flexed to 90° and the relative heights of the knees noted. In this example, viewed from the patient’s right side, the right knee is “lower” and represents an apparent shortening of the right thigh because of a dislocated hip. (Child is under general anaesthesia before operative reduction of her hip)

• Download figure
• Open in new tab
• Download powerpoint

Additional signs, such as a discrepancy in leg length, a widened perineum on the affected side, buttock flattening, and asymmetrical thigh skin folds, may be present. None of these signs is particularly sensitive or specific. Asymmetric skin folds are found in 25% of normal babies: on its own this is not an important clinical finding.^w12

Making a clinical diagnosis of hip dysplasia in older children is often easier. All subluxed, dysplastic hips will show limited abduction when fully flexed (>90°) (fig 4⇓). The child may walk on its toes on the affected side or may present with a painless limp from weak hip abductor muscles secondary to the unstable hip and unequal leg lengths. Despite this, a delay in walking age is not common.13

Fig 4 Late presenting dysplasia of the right hip: asymmetry in abduction of the flexed hip (the right hip has reduced abduction in flexion); the right thigh seems short; thigh skin creases are asymmetrical. (The child was under general anaesthetic, so we could show limited abduction without hands obscuring the picture. Hands should be positioned as in fig 2 for this test)

• Download figure
• Open in new tab
• Download powerpoint

Imaging

Ultrasonography

Ultrasound scans of the hip are useful from birth until the age of 4-5 months, while the hips are cartilaginous. The Graf technique is based on a standardised coronal section through the mid-acetabulum and assesses hip morphology (static examination). It does not include assessment of stability (dynamic assessment), but a coexisting instability will influence the grading of pathology (fig 5⇓).14 Modifications allowing hip shape and hip stability to be assessed separately are now commonly used (table 2⇓, fig 6⇓).15

Fig 5 Graf’s technique for hip ultrasonography. The diagrams are usually presented in this orientation, but the ultrasound is commonly presented with a 90° change in orientation, as in fig 6. Line A is a continuation of the sonographic iliac wing and helps to define the percentage of the femoral head that is covered by (or contained within) the bony acetabulum, as described by Morin and modified by Terjesen^{w17 w18}

• Download figure
• Open in new tab
• Download powerpoint

Fig 6 Ultrasound scans performed using a modified (Rosendahl) Graf’s technique, based on the coronal standard section and measurements of the α-angle, with the femoral head centred in the acetabulum. The α-angle is a measure of acetabular depth: (a) normal, α≥60°; (b) immature, 50°≤α<60°; (c) mild dysplasia 43°≤α<50°; (d) significant dysplasia (with or without changes of the labrum), α<43°. This technique also includes a separate assessment of hip stability

• Download figure
• Open in new tab
• Download powerpoint

Table 2

 Relation of hip morphology to stability as assessed on ultrasound²³

View this table:

• View popup
• View inline

Morphologically normal hips have a marginal risk of developing dysplasia in later childhood.1 16 ^w13 Immature hips, although dislocatable, develop normally without treatment in 97% of cases.^{w14 w15} Reports on the natural course of mild dysplasia are sparse, but two prospective studies suggest that such hips tend to normalise spontaneously.17 ^w16

A second measure of acetabular shape or depth is the percentage of the femoral head covered by the bony rim of the acetabulum (fig 5⇑).^{w17 w18} With a centred femoral head, coverage <47% in boys and <44% in girls is pathological, affecting 1-6% of newborns.1 16 18 When the hip is both unstable and dysplastic, the coverage varies and measurement can be misleading.

It is important to remember that the clinical picture and the ultrasound report should agree—if they don’t, presume one is wrong. Ultrasonography and clinical examinations are both operator and situation dependent.

Radiography

In babies older than 4.5 months, as bone develops, anteroposterior pelvic radiographs show hip dysplasia as a delay in acetabular ossification or as dysplasia with or without a subluxed or dislocated femoral head (fig 1⇑, fig 7⇓). As in ultrasonography, a standardised position for the radiograph is crucial for an accurate diagnosis. The acetabular index measures the “shallowness” of the acetabulum: hips with values >2SD^w19 are termed dysplastic and those between 1 and 2 SD indicate delay in acetabular ossification.

Fig 7 Anteroposterior pelvic radiographs in 6 month old children. (a) The acetabular index (AI) is a measure of the acetabular inclination and dysplasia; (b) normal hip with large ossified femoral head and normal AI; (c) hip with high AI with delayed acetabular ossification (note that the femoral head ossific nucleus is small); (d) hip with dysplasia with greatly abnormal AI and small femoral head ossific nucleus. In both (b) and (c) Shenton’s line is disrupted, implying the hip is subluxed

• Download figure
• Open in new tab
• Download powerpoint

Once detected, how is a dysplastic or unstable hip treated?

Abnormal results must be identified and acted on promptly (fig 8 ⇓), as the window of opportunity for instigating effective treatment of neonatal hip instability is small. In accordance with SMAC guidelines and the National Screening Committee’s 2004 report on hip dysplasia,11 19 and www.library.nhs.uk/specialistlibrarysearch] clinicians need to take responsibility for their assessments and audit the results of their local screening programmes.

Fig 8 Assessing developmental dysplasia of the hip

• Download figure
• Open in new tab
• Download powerpoint

Early diagnosis

The primary aim of treatment is to achieve a stable, concentric reduction of the hip to enable normal joint development. Most unstable hips stabilise spontaneously by 2-6 weeks of age. Hips that remain dislocatable or pathologically unstable after this time, most of which also show dysplasia on ultrasound, need prompt treatment. A dynamic flexion-abduction orthosis known as the Pavlik harness, or “splinting,” is used to obtain and maintain hip reduction (fig 9⇓).

Fig 9 Properly applied Pavlik harness. The hips are in 100° of flexion and considerable abduction. In this position the hip is held reduced, allowing the hip capsule to tighten, which stabilises the hip. The harness is safe, but extreme positions of abduction should be avoided to lessen the risk of femoral head avascular necrosis, which could disturb growth of the proximal femur

• Download figure
• Open in new tab
• Download powerpoint

The harness remains in place at all times, allowing loose capsular soft tissues of the hip to tighten and tight hip adductor muscles to stretch. The harness may be adjusted as the child grows and the hip stabilises. The child is followed up clinically and with ultrasound scans (radiographs after age 4.5 months), the frequency of which will vary with the pathology being treated, until the hip is clinically stable and ultrasonography shows a stable, centred, and normal or only slightly immature hip (Graf type 1 or 2a+). A longitudinal study showed that when harness treatment was started by 90 days of age, only 5.7% of 332 babies required further treatment.20

A Pavlik harness is contraindicated with teratological hip dislocation, in children older than 4.5-6 months, when the hip is irreducible (Ortolani negative), and in cases of parental non-compliance. The main risk of splinting is that pressure on the immature, dislocated femoral head may cause avascular necrosis (reported in up to 2% of those treated) or occasionally a temporary femoral nerve palsy.5 9

Two small randomised trials have shown that stable hips with mild dysplasia on ultrasonography can be observed safely for six weeks before a decision to treat is made.17 21

When is surgery indicated?

Children who require surgery arise from two groups: those who fail early splint or harness treatment, and those who are diagnosed late and are not suitable for such treatment. The most common operation is closed reduction with adductor or psoas tenotomy, followed by 3-4 months in a plaster cast or abduction brace. The older the child, the more likely it is that an extensive procedure will be required: open reduction with soft tissue stabilisation of the joint, followed by a cast. Over the age of 18-24 months, an additional pelvic or femoral osteotomy (or both) is often required to create a more normal anatomical shape and orientation of the hip joint (fig 10⇓).

Fig 10 (top) Preoperative anteroposterior pelvic radiograph showing a late presenting (age 2 years) left dislocated hip with considerable acetabular abnormality; (bottom) after open reduction of the joint, the iliac wing of the pelvis was “broken” (a pelvic osteotomy) and the fragments moved into position with better bony coverage of the femoral head. The osteotomy was held with two wires. The improved anatomy helps stabilise a reduced hip joint; movement promotes normal growth

• Download figure
• Open in new tab
• Download powerpoint

What is the long term outlook?

Premature degenerative joint disease3 and low back pain are potential long term sequelae of hip dysplasia, depending on type and duration of the untreated instability, treatment, age at which it was instigated, and the presence of avascular necrosis. Untreated dislocated hips may fare better than treated dislocated or dysplastic hips. Cases must be assessed individually; a unilateral hip dislocation is probably more amenable to surgery than a bilateral dislocation. Most surgeons are reluctant to treat bilateral dislocations after the age of 6-8 years because of the limited remodelling potential as children get older.

Screening

A successful screening programme is defined by several criteria, and international debate continues as to whether hip screening fulfils them. The main aim of screening is to reduce the prevalence of late diagnosis as early detection allows early treatment, reducing the need for surgical treatment and the risk of residual dysplasia.

One controversy is whether screening should be by clinical examination alone or with selective or universal ultrasound assessment. In the United Kingdom, clinical screening with the Ortolani and Barlow tests has failed to reduce the prevalence of late diagnosed developmental hip dysplasia.22^w20 Both clinical tests have a high specificity, but low sensitivity, especially with non-trained examiners.23 24^w21-w23.

Only two randomised controlled trials have addressed the effect of neonatal ultrasound screening on presentation of hip dysplasia: both compared universal to selective (high risk) screening, and one included a clinically screened control group.16 25 Trained examiners performed both the clinical screening and the ultrasound examinations. Although there were more late cases in the selective screening groups than in the universally screened groups, this was not statistically significant (p=0.22).16 Had non-trained examiners performed the clinical screening, the results may have been different. The advantage offered by ultrasound screening becomes evident only in comparison to clinical screening by non-trained examiners.16 ^w24

In countries where universal ultrasound screening is practised, initial treatment rates (and costs) have been high initially but, with experience, both have fallen.26 ^{w25 w26} Late diagnosis of dysplasia has become less common, and few children require surgical reduction.27 One German study showed an ascertainment adjusted rate of first operative procedures of 0.26 per 1000 live births (95% CI 0.22 to −0.32) after universal ultrasound screening, compared with a UK equivalent estimate of 0.78 (0.72 to 0.84) after clinical screening.4 28

In the UK and parts of Europe, selective ultrasound screening has become common practice, despite the uncertain evidence base for screening for hip dysplasia, including effectiveness of treatment, as highlighted in two systematic reviews.5 29 The US Preventive Services Task Force did not recommend universal ultrasound screening in North America and acknowledged that even when newborn and infant hip examinations are normal, a late presentation occurs in 1 in 5000 infants.30 In Coventry, three years after the instigation of a universal ultrasound screening programme, there had been no late presentations of hip dysplasia.1 26

Where are we now?

As early harness or splint treatment avoids operative treatment in some cases, early diagnosis is important. Many studies emphasise that if the neonatal clinical examinations are performed by experienced practitioners, be they physicians, surgeons, or physiotherapists, then the late diagnosis rate is low, and in this context universal ultrasound screening may add little more in terms of clinical efficacy than selective ultrasound screening.23 24 31 ^{w22 w27} Late diagnosed hip dysplasia is still common in the UK.^w28 This may be related to variable quality or completeness of clinical screening, which is often performed by junior and inexperienced clinicians, or to the selective use of ultrasound screening. Although no system at national level has been shown to eliminate late presenting developmental dysplasia of the hip, improved outcomes have resulted from conscientiously applied screening programmes implemented in different regions.23 24 27 31 ^{w25 w27}