Congenital Dislocation of the Knee

Congenital dislocation of the knee (CDK) is a knee that presents at birth in hyperextension or frank anterior dislocation: the tibia is displaced anterior to the femur, and the quadriceps mechanism and anterior joint capsule are contracted. It sits on a spectrum from simple genu recurvatum through subluxation to irreducible anterior dislocation, and in severe cases the knee cannot be flexed passively beyond a right angle.

Why it matters. Early recognition is critical: outcomes are excellent with prompt serial casting, but delayed or inadequate treatment leads to progressive deformity, growth disturbance and the need for complex surgery. CDK is rarely isolated. The high association with developmental dysplasia of the hip means every CDK baby must have hip screening, and the association with clubfoot, arthrogryposis and Larsen syndrome means the knee is often the presenting sign of something wider, so every newborn with CDK gets a full musculoskeletal examination.

Who. Approximately 1 per 100,000 live births, with a female predominance of roughly 2:1 to 3:1. It is bilateral in roughly 40-60 percent of cases. Most cases are sporadic; familial forms are reported.

What it costs untreated. Untreated CDK causes severe gait disturbance, and knee instability limits standing, walking and sitting. Chronic anterior displacement can distort the distal femoral physis. The cosmetic deformity and functional limitation carry their own psychosocial cost through childhood.

Two theories. The exact aetiology remains debated. The mechanical theory attributes CDK to intrauterine positioning, a breech presentation with the feet trapped under the chin or against the uterine wall, leading to sustained knee hyperextension. The muscular theory emphasises primary quadriceps fibrosis and contracture as the driving force, analogous to the contracture seen in arthrogryposis.

The case for each. The association with breech presentation and with oligohydramnios, which restricts movement and maintains hyperextension, supports mechanical compression; a first-born's tighter uterus leaves less room for movement. Against pure positioning, the absence of the cruciates suggests developmental arrest, and familial CDK and the syndromic forms (Larsen) support an inherent tissue abnormality. CDK may represent a continuum from positional deformity to true congenital absence of joint formation.

Pathoanatomy. In severe (Grade III) cases the quadriceps tendon, patellar ligament and anterior joint capsule are all contracted and shortened. The tibia is displaced anteriorly and laterally relative to the femur, and the hamstrings are attenuated and displaced anteriorly. The cruciate ligaments may be absent, attenuated or abnormally positioned; the table sets out how each structure changes across the grades.

View larger

The patella. The patella is frequently abnormal in CDK and is a source of later morbidity that is easy to overlook in the neonate. The neonatal patella is unossified, but the underlying cartilaginous patella is often small (hypoplastic), laterally positioned, or even absent, and the extensor mechanism may be laterally subluxed, features that are especially pronounced in Larsen syndrome. After the tibiofemoral dislocation is reduced, persistent patellar maltracking or recurrent lateral patellar instability may emerge as the child grows and the patella ossifies, sometimes requiring later extensor-mechanism realignment; see paediatric patellar instability for those options (lateral release, medial reefing, or distal realignment). Examine and document patellar position at presentation and follow it through growth, because a reduced tibiofemoral joint does not guarantee a stable patellofemoral joint.

History. Ask about the pregnancy and the delivery: breech presentation, oligohydramnios, any prenatal ultrasound findings, and whether the hyperextended knee was noted at delivery. Ask about consanguinity and familial joint disorders. Then ask about the other joints: a foot deformity, a hip click or limited abduction.

The knee. Inspect for hyperextension or anterior dislocation; the skin folds may be abnormal. Palpate the patella, which may be displaced or poorly palpable, and the tibiofemoral relationship. Assess passive flexion, because the arc is what grades the knee.

The rest of the baby. Every newborn with CDK must have:

1. Both knees examined, since the condition is so often bilateral
2. Both hips screened by clinical examination (Ortolani, Barlow, abduction) and by hip ultrasound
3. Both feet examined for clubfoot
4. Full musculoskeletal survey of elbows, shoulders and spine for other contractures or dislocations
5. Neurological examination to exclude myelodysplasia
6. Dysmorphology assessment for Larsen syndrome features: prominent forehead, hypertelorism, depressed nasal bridge

Failing to screen for DDH in a baby with CDK is a significant clinical error.

A note on "40-100 percent", because it is quoted throughout this page and it is not really a figure. A range whose upper bound is 100 percent and whose lower bound is 40 percent cannot be used to estimate anything for an individual baby; it is the spread across small series that differ in whether they studied isolated or syndromic CDK, and in whether "DDH" meant frank dislocation or any sonographic dysplasia. Do not quote it as a prevalence in a viva. The only inference that survives the spread is the one that changes behaviour: the association is high enough that every CDK baby gets hip imaging irrespective of how normal the hip examination feels. The hip examination is itself less reliable here, because a stiff hyperextended knee makes Ortolani and Barlow testing difficult to perform properly, which is why the ultrasound is mandatory rather than confirmatory.

Positional or true? A positional hyperextended knee is soft and passively correctable to full flexion without resistance, and the tibiofemoral joint is congruent on radiographs. True CDK has a firm endpoint to passive flexion, the tibia is anteriorly displaced, and radiographs show anterior subluxation or dislocation. If you can flex the knee to 90 degrees easily with no firm resistance, it is likely positional and will resolve without casting.

Radiographs. AP and lateral views of both knees at diagnosis. On the lateral, look for anterior displacement of the tibia relative to the femur, the position of the patella and the physeal status. Grade I may appear near-normal with hyperextension only; in Grade III the tibia is obviously displaced anterior to the femoral condyles. The diagnosis is primarily clinical: radiographs confirm the degree of tibial displacement and guide grading, and casting is not delayed while waiting for advanced imaging.

Hip ultrasound. At 6 weeks, by the Graf technique or dynamic ultrasound, in every CDK baby regardless of the clinical hip examination. This is the most important investigation, because DDH coexists so frequently and has its own treatment window.

MRI. For Grade III or syndromic cases before quadricepsplasty. It shows cruciate ligament status, meniscal position and cartilage integrity, and identifies the absent or abnormal structures that affect the surgical approach.

Skeletal survey and genetics. For bilateral CDK, multiple joint dislocations or dysmorphic features. Look for spinal anomalies, other joint dislocations and features of skeletal dysplasia; request FLNB gene testing for Larsen syndrome, and chromosome analysis if indicated.

The principle. Gradual reduction by serial casting, first line for every grade; in Grade I, passive stretching accompanies it and casting runs about 4-6 weeks. Casts are changed weekly with progressively more knee flexion, and the goal is a congruent reduction with at least 90 degrees of flexion. Grade I and II usually reduce within 6-8 weeks; failure of closed treatment by 3 months of age indicates surgery.

Knee before hip. When CDK and DDH coexist, as they frequently do, the knee is reduced first. The reason is mechanical: the Pavlik harness and the hip spica hold the hip in flexion and abduction, which requires a flexed knee. A knee fixed in hyperextension or anterior dislocation cannot be positioned in either device, so treating the hip first is futile and may even displace it. Serial casting, or surgery, is therefore directed at restoring knee flexion first; once the knee flexes past roughly 90 degrees a Pavlik harness can be applied, and one device may then treat a reducible knee and a dysplastic hip together. "Concurrent" describes that stage, not the neonatal one. In bilateral Grade III disease the same sequencing, knees then hips, dictates the order of any staged surgery. The hip side of the plan is set out in DDH treatment options; where a neurological cause is suspected, see myelomeningocele.

Prevention. The most important preventable complication is missed DDH, which is why every child with CDK has hip screening by clinical examination at birth and ultrasound at 6 weeks. The second is physeal injury from forced manipulation: cast changes progress gradually, never forcing the knee beyond what it will comfortably accept. Finally, peroneal nerve palsy is prevented by avoiding excessive flexion in a single cast and by monitoring neurovascular status after every cast change.

How to read the table. The gradient is right, but "good functional result" for an operated Grade III understates what the long-term data show and overstates what you should promise. In Honcharuk's series the operated knees had less flexion in swing, less overall motion, greater coronal instability and slower gait than controls, and the authors concluded that surgery for type-3 CDK will not likely restore normal knee function because the joint is teratologic (PMID 38666580): the knee was never normally formed, so reduction is achievable but normality is not.

The reassurance, from the same unit. In Oetgen's series these children walked without braces, averaged 112 degrees of arc and scored comparably to normal controls, with limitations only in high-demand activity, and 7 of 9 knees showed instability on examination that none of them felt (PMID 20357585). Clinically detectable instability is not the same as functional instability, and both facts belong in the same conversation with a parent. Both are true: reduction and a functional knee are realistic; a normal knee is not. The diagnosis matters more than the operation, since idiopathic CDK had the most control-like gait, then Larsen syndrome, then arthrogryposis.