Developmental Dysplasia of the Hip (DDH)

Developmental dysplasia of the hip is a spectrum of instability rather than a single lesion: a shallow acetabulum that still holds the head, a head that has partly left the socket, and a head that has left it altogether. Treatment intensity follows that step, not the word "click". True dysplasia or dislocation needing treatment occurs in about 1-3 per 1000 live births.

View larger

Who. Girls outnumber boys about 4:1. The left hip is affected in 60% of cases, the right in 20% and both in 20%: in utero the left hip lies against the mother's spine, a position that limits abduction and promotes dysplasia.

Risk factors. Breech presentation, female sex, first-born status and a family history; breech is a strong risk factor in its own right.

Normal development. The femoral head and the acetabulum develop together, and movement stimulates that development; restricted movement causes dysplasia. The triradiate cartilage contributes to acetabular growth, the labrum deepens the socket, and the shape of the socket is influenced by where the head sits within it. Normal development therefore requires the hip to be reduced.

Why DDH develops. Ligamentous laxity from maternal hormones, mechanical factors such as breech position and oligohydramnios, and a shallow acetabulum that allows the head to dislocate. Once the head is out the acetabulum becomes shallower still, the labrum everts or inverts, the capsule elongates, and secondary changes accumulate if the hip is left untreated. Dysplasia proceeds to dislocation if untreated; early treatment prevents progressive deformity.

The acetabulum. Primary dysplasia is a shallow, saucer-shaped socket with poor anterior and lateral coverage. In dislocation the empty socket fills with fibrofatty tissue, the pulvinar, and in a high dislocation a false acetabulum forms on the ilium while the true acetabulum becomes shallow.

The femur. The ossific nucleus appears later on the affected side. A subluxed head loses its sphericity (coxa plana), lack of containment produces coxa valga with an increased neck-shaft angle, and increased femoral anteversion is common. This proximal femoral deformity complicates reduction.

Obstacles to reduction. In established DDH the following stand between the head and a concentric reduction, and in late cases they must be addressed surgically:

• Inverted labrum (limbus), folded into the joint, the Graf type IV hip
• Hypertrophied ligamentum teres, thickened and elongated
• Transverse acetabular ligament, hypertrophied and contracted across the inferior acetabulum
• Capsular constriction, the hourglass deformity of the capsule
• Iliopsoas tendon, tight and contracted where it crosses the capsule, with tight adductors alongside it
• Pulvinar, the fibrofatty tissue filling the socket

By severity. A dysplastic hip has an abnormal, shallow acetabulum with inadequate coverage but a located head. A subluxed hip has only partial contact between head and acetabulum. A dislocated hip has no contact at all. Severity guides treatment urgency and prognosis.

Teratologic DDH is a different condition: dislocation early in utero with severe soft-tissue contractures, associated with syndromes such as arthrogryposis, and more resistant to treatment.

Graf ultrasound types. Graf is drawn off the iliac baseline. The alpha angle, between the ilium and the bony acetabular roof, measures bony coverage; the beta angle, between the ilium and the labrum, measures cartilaginous coverage. Types IIa and IIb share the same angles at a different age: a young IIa is observed, a IIb is treated.

View larger

Tönnis radiographic grades. Used once the ossific nucleus is visible, after 4-6 months, and graded by how high the femoral head sits.

Every newborn should be screened. The two named tests answer different questions about the same hip.

Ortolani. Abduct the hip with anterior pressure on the greater trochanter. A clunk is the head reducing into the acetabulum, so a positive test means the hip was dislocated and is reducible.

Barlow. Adduct the hip with posterior pressure. A clunk is the head dislocating posteriorly, so a positive test means the hip is unstable and dislocatable.

View larger

Click versus clunk. Soft clicks are common and usually benign. A clunk is palpable movement of the head, and that is what counts: DDH is a clunk, not a click.

The rest of the newborn examination. Assess stability, look for a limb length discrepancy, and note asymmetric skin folds, which are less reliable.

The older infant and child. Once the neonatal period is past the signs change:

• Limited hip abduction, the most reliable sign in the older infant
• Galeazzi sign, a limb length difference
• Asymmetric gluteal folds
• Trendelenburg gait at walking age

Late presentation is more challenging to treat.

Ultrasound. The investigation of the unossified hip. Screen at 4-6 weeks when risk factors are present, and scan for diagnosis whenever there is clinical suspicion. The report is written in Graf's types, and the angles it quotes are the ones in the classification table.

View larger

View larger

View larger

View larger

Plain radiographs. Useful after 4-6 months, once the femoral head ossifies, and the investigation for older children and for follow-up. Know the lines and their normal values: they are routinely shown, and they underpin the decision to treat residual dysplasia.

View larger

Arthrography at examination under anaesthesia. The arthrogram shows whether a reduction is concentric and, if it is not, what is in the way. A medial dye pool of more than about 5-7 mm between the femoral head and the acetabular floor means a non-concentric reduction with soft tissue interposed. The classic obstructive signs are the rose-thorn appearance of an inverted labrum (limbus) indenting the contrast, an hourglass or figure-of-eight capsular constriction produced by the iliopsoas tendon crossing the capsule, and pooling from a hypertrophied pulvinar and ligamentum teres.

View larger

View larger

Age at diagnosis decides the treatment; each age band below has its own method and its own way of failing.

0-6 Months6-18 MonthsWalking Age

The harness. The Pavlik harness is first-line for newborn DDH when the hip is reducible. It holds both hips, even when only one is affected, in 100-110 degrees of flexion and 50-70 degrees of abduction, not forced: the position is the treatment. Wear is full-time at first, 23 hours a day, with weekly ultrasound to confirm that the hip is reduced.

View larger

View larger

What predicts success. Age at initiation: about 90-95% when the harness is started under 6 weeks, and falling thereafter. Success decreases to 50% after 3 months, and the systematic review in the evidence base found success anywhere from 7% to 99% across series, so quote a range rather than a number. Graf type matters too, with better results for types II and III and poorer for type IV with its inverted labrum, and bilateral cases may need longer treatment.

Duration. Until the hip is stable on ultrasound, typically 6-12 weeks full-time, then part-time weaning.

Complications. The harness has its own ways of harming a hip:

• Femoral nerve palsy, from excessive flexion
• Avascular necrosis, from forced abduction
• Inferior (obturator) dislocation, when too much flexion drives the head out of the bottom of the socket; stop the harness and reduce properly rather than congratulating yourself for keeping the hip flexed
• Skin irritation and pressure sores

When to stop. If the hip has not reduced after 3-4 weeks, discontinue the harness. Persisting risks Pavlik disease: the inverted labrum prevents reduction and is made worse by continued harness use.

Examination under anaesthesia with arthrogram. Assess stability and reducibility, identify the obstacles (inverted labrum, ligamentum teres), and determine the safe zone, the arc from reduction to redislocation. Between 6 and 12 months closed reduction is often what is required; from 12 to 18 months open reduction is usually needed.

Closed reduction. Reduce gently, with an adductor tenotomy if the adductors are tight, and confirm on the arthrogram that the reduction is concentric. Cast in the human position, 100 degrees of flexion and 40-50 degrees of abduction, and avoid the frog-leg position because of the AVN risk. The spica stays on for 6-12 weeks, with CT, or MRI, in the cast to confirm the reduction; the film must show the head opposite the true acetabulum, not a false one.

Convert to open reduction when any of the following is present:

• The reduction is not concentric on the arthrogram (a persistent medial pool or a rose-thorn limbus)
• The safe zone is under 20-25 degrees
• The reduction is unstable
• The hip is Graf type IV, with an inverted labrum

Over 18 months the hip needs comprehensive reconstruction: open reduction, femoral shortening in the child over 2-3 years, and a pelvic osteotomy for the dysplastic acetabulum. The approaches, the obstacles and the osteotomies are set out under Surgical Techniques.

Late DDH Outcomes

Late presenting DDH has significantly worse outcomes than early treatment. Counsel families: higher AVN risk (10-20%), residual dysplasia common, may require future surgeries. Early diagnosis through screening is key.

The approach. The medial (Ludloff) approach is preferred under 12 months: less dissection, it preserves the blood supply, and it avoids the need for a capsulorrhaphy. The anterior (Smith-Petersen) approach gives better visualisation in older children and allows a capsulorrhaphy.

Clearing the obstacles. The structures described under pathophysiology are dealt with in turn:

1. Tight psoas tendon, released at the lesser trochanter
2. Transverse acetabular ligament, divided
3. Elongated ligamentum teres, excised
4. Inverted labrum (limbus), reduced, or excised if severely deformed
5. Pulvinar, cleared from the acetabulum

Femoral shortening. Essential in children over 2-3 years. It decompresses the femoral head, which reduces the AVN risk, and it lets the hip reduce without excessive tension; usually 1-2 cm is removed, depending on the degree of dislocation. Where reduction would need excessive force in an older child, shortening is preferred to brute reduction. A varus derotation osteotomy corrects coxa valga and anteversion and improves coverage, and femoral procedures are often combined with acetabular surgery.

Pelvic osteotomies. Choose by age and by the deformity, for the residual dysplasia in front of you rather than as a reflex add-on to every open reduction. The postoperative film has to show a seated head and a more horizontal roof, not just metal.

• Salter innominate osteotomy, 18 months to 6 years: redirectional. A complete iliac cut reorients the whole acetabulum anterolaterally, pivoting through the pubic symphysis, with a graft holding the rotation.
• Pemberton and Dega: reshaping. Incomplete cuts that hinge on the triradiate cartilage, for younger children.
• Triple osteotomy and periacetabular osteotomy (PAO): at skeletal maturity; the PAO is the most common choice for adolescent and adult dysplasia.

View larger

View larger

View larger

Avascular necrosis is the most feared complication, and the rate depends on how the hip was reduced:

• Pavlik harness: 0-5% with correct technique in experienced hands, against a published range of 0-28%; quote the low figure only when you can also say why it is low
• Closed reduction: 5-15%
• Open reduction: 10-25%

The risk factors are excessive abduction, prolonged immobilisation and older age at treatment. AVN may not manifest for 6-12 months after treatment.

Why the Pavlik figures differ so much. The 0-28% range is not noise; it is driven by technique and case mix. Failure to achieve reduction, forced abduction into the so-called safe zone, and persisting with a harness that is not working all push a unit toward the top of the range. The 0-5% figure is achievable only when reduction is confirmed early and the harness is abandoned promptly if it fails, which is why the Mubarak pitfalls paper, not the incidence figure, is the one to internalise.

Preventing it. Keep the hip in the middle of Ramsey's safe zone and never force abduction: forced frog-leg abduction compresses the lateral epiphyseal vessels against the acetabular rim, and that is how a reduced hip still gets AVN. Consider femoral shortening in the older child to reduce the pressure of reduction, and watch closely during treatment for signs of vascular compromise.

View larger

Residual dysplasia affects 10-20% of treated hips and is managed with an osteotomy at the appropriate age. A flattened, fragmented ossific nucleus (coxa plana) and a still-steep roof mean the child may need a later pelvic osteotomy even though the hip stayed reduced.

Redislocation rates are variable, and adequate immobilisation is the prevention. Femoral nerve palsy in the harness is rare and is prevented by proper application.

Prognosis. Better outcomes follow early diagnosis (under 6 weeks), Pavlik treatment and a concentric reduction; worse ones follow late diagnosis, open reduction, AVN and residual dysplasia. Most treated hips function normally if reduction is achieved early.

By age at diagnosis.

• Under 6 weeks: 90-95% success with the Pavlik harness alone
• 3-6 months: 70-80% success with the Pavlik harness; closed reduction may be needed
• 6-12 months: often closed reduction and spica
• 12-18 months: usually open reduction and spica
• Over 18 months: open reduction with pelvic and femoral osteotomy

What to follow. The acetabular index and the centre-edge angle, until skeletal maturity: residual dysplasia may need a pelvic osteotomy at the appropriate age, a Salter between 18 months and 6 years or a PAO in adolescence. Osteoarthritis risk is higher with residual dysplasia or AVN, and such hips may need a total hip replacement in adulthood.

Global Epidemiology

• Say which number you mean. Neonatal hip instability detected on clinical examination is considerably commoner than established DDH, and most of it resolves spontaneously in the first weeks; the figure of roughly 1-3 per 1000 live births refers to true dysplasia or dislocation requiring treatment, not to the number of babies with an abnormal newborn examination. Quoting the treated-DDH figure as though it were the rate of neonatal instability is a common slip, and it makes spontaneous resolution look far rarer than it is.
• Sonographic dysplasia is far more common (immature/dysplastic hips in up to 15-20% of healthy neonates on early ultrasound), which is why universal early ultrasound generates substantial overtreatment.
• Marked geographic variation: very high prevalence in populations that swaddle with the hips extended/adducted (some Indigenous North American, Eastern European and Middle Eastern groups); very low rates where babies are carried abducted/straddled (parts of sub-Saharan Africa and Southeast Asia).
• Risk factors (BFFF): breech presentation, female sex (about 4:1), first-born, positive family history; also oligohydramnios and other packaging deformities (congenital muscular torticollis, metatarsus adductus, talipes).

Screening Strategy - Side by Side

Treatment Guidance - Side by Side

Registry and Surveillance Data

• DDH is tracked through national paediatric surveillance units (e.g. UK and Australian/New Zealand paediatric surveillance units) that monitor late-presenting DDH (diagnosed after the newborn period) as a key quality indicator of screening performance.
• Late-presentation rates of roughly 0.5-1.5 per 1000 persist even in well-resourced programmes, underscoring that no screening strategy eliminates missed cases.
• Unlike arthroplasty, there is no implant registry for paediatric DDH; long-term data come from osteotomy cohort follow-up rather than registries.

High- vs Limited-Resource Practice

Universal versus selective ultrasound. The central screening debate. Universal ultrasound detects more dysplasia but generates overtreatment without a proven reduction in late dislocation (Rosendahl RCT). German-speaking Europe favours universal ultrasound; most other programmes use selective ultrasound for risk factors. Reasonable practitioners still disagree.

The first-line device. The Pavlik harness is the default, but systematic review evidence suggests rigid splints (von Rosen) may achieve comparable or better success with less residual dysplasia. No adequately powered randomised trial settles the question.

The ossific nucleus. Whether to delay reduction until the femoral ossific nucleus appears, to lower the AVN risk, is contested. Meta-analysis (Niziol 2017) found no consistent protective effect overall, which challenges routine delay.

Pre-reduction traction. Historical practice held that traction before reduction lowers AVN and improves reducibility. Several modern cohorts dispute any benefit and many centres have abandoned it, but the evidence remains low quality.