Congenital Kyphosis

Congenital kyphosis is a rigid sagittal-plane deformity caused by abnormal vertebral development: failure of formation or failure of segmentation of the anterior vertebral body. It is rare, much less common than congenital scoliosis, and affects males and females equally. The apex often lies at the thoracolumbar junction (T10-L2), which puts the cord at maximum risk.

Why it is treated differently from congenital scoliosis. Congenital kyphosis carries a high risk of spinal cord compression and paraplegia if untreated, so neurology takes priority over growth. That changes the answer on bracing and on the timing of fusion.

Natural history. Type I progresses relentlessly, at 5-10 degrees a year, and untreated it progresses to more than 100 degrees, with a 25-50% risk of paraplegia. Type II progresses more slowly and is often less severe; it may remain stable or progress, and its paraplegia risk is lower but still significant. Type III is unpredictable and behaves like whichever component dominates. Deterioration can be rapid during growth spurts, even in normal infants.

Failure of formation (Type I). The anterior part of one or more vertebral bodies fails to form, leaving a posteriorly based wedge vertebra or complete aplasia of the body. The spine is forced into a sharp angular kyphosis at that level. The spinal cord is stretched over the apex, and as the kyphosis progresses the cord is progressively compressed against the posterior body.

Failure of segmentation (Type II). An anterior unsegmented bar forms, like a stalactite of bone connecting adjacent vertebrae anteriorly, and tethers anterior growth across several levels. The posterior elements remain normally segmented and keep growing, so the result is a more gradual kyphosis that usually progresses more slowly than Type I.

The Winter classification (1973) is the standard classification, and its three types follow the two failures above. Type I has the worst prognosis, with high risk of both progression and neurological injury.

• Type I, failure of formation: partial aplasia of the anterior body (wedge vertebra) or complete aplasia of the anterior body (aplastic vertebra)
• Type II, failure of segmentation: an anterior unsegmented bar
• Type III, mixed: a combination of Type I and Type II defects

View larger

Type II is two prognoses. McMaster's natural-history series separates Type II by where the bar sits, and the distinction changes surveillance.

• Midline anterior bar, producing pure kyphosis (15 patients): usually progresses slowly
• Anterolateral unsegmented bar, producing kyphoscoliosis (9 patients): much more severe, and behaves closer to Type I

So the first clinic visit asks where the bar sits as well as which Winter type it is. A bar off the midline adds a coronal deformity and a worse trajectory.

McMaster classification. McMaster's system is primarily for congenital scoliosis but can overlap with kyphosis. It describes:

• Hemivertebra (fully segmented, semi-segmented or incarcerated)
• Block vertebra
• Wedge vertebra
• Unilateral bar, with or without a hemivertebra

History. Start with the birth history: was the anomaly diagnosed antenatally, and has VACTERL screening been done? Walking and continence tell you about cord function. Ask whether the parents have noticed any worsening.

Neurology comes first. A full upper and lower limb examination is critical. Look for spasticity and clonus, hyperreflexia, and an ataxic or scissoring gait. The red flags for cord compression are:

• Ankle clonus
• Opisthotonus, hyperextension in an infant
• Hyperreflexia
• Delayed walking

The spine and skin. The classic finding is a sharp angular kyphosis, a gibbus, which is usually rigid when flexibility is assessed. Check for scoliosis, which often coexists, and for cutaneous stigmata such as a hairy patch or dimple, which point to concurrent spinal dysraphism.

Radiographs and CT. PA and lateral whole-spine radiographs identify the anomaly and measure the kyphosis. CT with 3D reconstruction defines the bony anatomy and is essential for surgical planning.

MRI. MRI is mandatory before surgery. It shows whether the cord is compressed, with myelomalacia or signal change, and rules out intraspinal anomalies: diastematomyelia, tethered cord and syrinx.

Systemic screening. Screening for the VACTERL association means an echocardiogram, a renal ultrasound and a GI and anorectal examination.

A sharp, rigid, angular kyphosis (gibbus) in a child is congenital until proven otherwise. Several other conditions produce a kyphotic spine, and they must be distinguished because the management and the neurological risk differ sharply.

Bracing does not work. The deformity is rigid, so bracing is ineffective, and casting is not used as it is in scoliosis. Observation may be considered in mild Type II with no progression, but that is rare.

Fuse early. In early-onset scoliosis fusion is delayed; in congenital kyphosis early fusion is often recommended to prevent paraplegia, and it is often the safest option even at age 1-2. The lethal endpoint here is paraplegia rather than trunk-height loss, so arresting a progressive Type I curve takes priority and the crankshaft phenomenon is an accepted, managed risk rather than a reason to delay. Fusion is better than paralysis.

Choosing the operation. The operative options and their indications:

• Posterior fusion in situ: for mild cases or young infants; stops progression
• Posterior fusion with compression instrumentation: the modern standard; screws or hooks apply posterior shortening
• Combined anterior and posterior fusion: for severe or rigid cases, with access from both sides
• Vertebrectomy or vertebral column resection (VCR): for severe fixed deformity with neurological compromise

The principle is posterior shortening. Scoliosis correction lengthens the concavity; kyphosis correction requires shortening the convexity, the posterior spine.

Cantilever correction. Screws are anchored at the proximal and distal ends, and a rod pre-bent to the desired lordosis is cantilevered into them, progressively reducing the kyphosis. The risk is screw pull-out, especially proximally, so the construct needs strong anchors.

Compression correction. A rod is placed in situ and a compressor applied across the kyphotic apex. It is safer, but the correction is less powerful.

The mechanism. In a skeletally immature child a solid posterior fusion tethers the back of the spine, but the anterior vertebral bodies keep growing because the anterior growth plates are still open. Anchored posteriorly, that growth has nowhere to go but to bend and rotate around the posterior tether like a crankshaft, so the deformity recurs and rotates despite a radiographically solid fusion mass.

Who is at risk. The markers of remaining spinal growth identify the child:

• Open triradiate cartilage, the classic high-risk marker
• Risser 0, with no iliac apophysis ossification
• Young chronological age and Tanner stage, and the peak-height-velocity window

These are exactly the patients in whom congenital kyphosis pushes the surgeon to fuse early, which is why crankshaft is a real concern here and not a theoretical one. When early fusion is unavoidable, the strategies below mitigate it.

General growth-modulating and growing-rod strategies for non-congenital early-onset deformity are covered in the early-onset-scoliosis topic.

Why the stakes are high. The cord is draped over the kyphotic apex and supplied by a watershed anterior spinal artery, so correction can both kink it and devascularise it. Multimodal SSEP and MEP monitoring is the standard of care, and vertebral column resection requires it.

Why both modalities. SSEP monitors the dorsal columns of the posterior cord. It is good for global cord integrity, but it can stay normal while the motor tracts are injured. Transcranial motor-evoked potentials (MEP) monitor the corticospinal tracts in the anterolateral cord, the territory of the anterior spinal artery, which is exactly what anterior-apex kyphosis correction puts at risk. MEP is the more sensitive alarm for the ischaemic anterior cord, and SSEP alone is insufficient.

Responding to a true alert. A real, reproducible loss, not artefact, calls for action in parallel rather than watching and waiting. The goal is to act before a transient monitoring change becomes a fixed deficit: in large series most monitoring changes occur before or during the correction and, if addressed promptly, do not result in permanent injury.

Neurological Complications

Surgical Complications

Long-term problems. Besides crankshaft, the curve can add on above or below the fusion, which is monitored with serial radiographs. Chronic pain may develop at the fusion ends and is managed with physiotherapy and pain management. Functional limitations include a short trunk and reduced spinal mobility, and occupational therapy helps with adaptation.

Surgical results. With early intervention, neurological function is preserved in more than 90%. Correction of 50-70% is achievable, the fusion rate is greater than 90% with a combined approach, and patient satisfaction is high when cosmesis is improved.

Prognostic factors. The untreated natural history by type is set out in the Overview; the factors that shape the result of treatment are these.

Long-term function. Most patients are independent in activities of daily living, and quality of life is generally good with successful treatment. Low-impact sport is possible once the fusion has consolidated, and a wide range of careers is open, though heavy labour is best avoided.

Global epidemiology. Congenital kyphosis is rare and substantially less common than congenital scoliosis. It affects males and females roughly equally, the apex lies at the thoracolumbar junction (T10-L1) in about two-thirds of cases, and roughly 60 percent of patients have at least one associated anomaly within the VACTERL spectrum. There is no high-quality population registry for congenital kyphosis; the evidence base is built from single-centre series, so practice is consensus-driven rather than registry-driven.

Side-by-side guidance. No single society publishes a dedicated congenital-kyphosis pathway, so principles are drawn across bodies:

The genuine point of difference between regions is not whether to operate but which technique (in-situ vs instrumented posterior fusion vs anterior release vs VCR) and how early, driven by deformity rigidity, magnitude and neurological status rather than geography.

High- vs limited-resource practice variation.

• Well-resourced centres: routine multimodal neuromonitoring, 3D-CT planning, single-stage posterior VCR, cell salvage and paediatric intensive care — enabling aggressive correction of severe deformity at acceptable risk.
• Limited-resource settings: later presentation with established neurology and larger curves is common; monitoring may be unavailable, pushing surgeons toward safer in-situ posterior fusion, the wake-up test, and staged rather than single-stage correction. Early prophylactic posterior fusion of a small progressive curve is the highest-value, lowest-cost intervention and should be prioritised wherever follow-up is reliable.

The literature is entirely retrospective Level IV, so several management questions remain genuinely unsettled, and they are favourite viva discussion points.

Timing of fusion in the very young. Early posterior fusion arrests progression and prevents paraplegia, but it sacrifices growth and risks the crankshaft phenomenon. How small a curve justifies fusing a toddler's spine is unresolved: most accept fusing a documented progressive Type I, but the threshold is judgement, not evidence.

In situ or instrumented posterior fusion. Modern pedicle-screw constructs allow correction and shortening, but instrumentation in tiny pedicles carries pull-out and neurological risk. Some still advocate simple in situ posterior fusion to halt progression in the youngest patients.

Whether anterior surgery is still needed. Combined anterior and posterior fusion historically reduced pseudarthrosis and addressed the crankshaft. Posterior-only osteotomy or VCR now achieves correction without thoracotomy morbidity in many hands, and whether anterior support still adds value in severe rigid curves is debated.

VCR or a lesser osteotomy. VCR gives the most powerful correction but carries its neurological complication risk and high blood loss. The boundary at which a PSO or an anterior release is enough is not defined by trials.

Decompression for an established deficit. When the cord is compressed with T2 signal change, the relative roles of anterior decompression and posterior shortening, and the neurological recovery that can be achieved, are uncertain. The prognosis with established myelomalacia is guarded.

T2 Hyperintensity at the Apex

• Indicates cord edema, gliosis, or myelomalacia.
• Edema: Potentially reversible if decompressed urgently.
• Myelomalacia: Established damage. Irreversible.

Clinical Correlation

• Patients with MRI signal change and neurological deficit have a worse prognosis even after surgery.
• The goal is to intervene BEFORE signal change develops.