There is a structural failure so catastrophic that it has its own name — craniocervical dissociation — and a classification system built specifically to grade its severity. It means the skull is separating from the spine. The ligaments that normally anchor the occiput to C1 and C2 have been stretched, torn, or rendered functionally incompetent. And three measurements — the ADI, the BDI, and the BAI — are the numbers that reveal it. In high-energy trauma, this injury is frequently fatal at the scene. In connective tissue disorders, where it develops slowly and insidiously over years, it is frequently invisible to imaging teams who do not know to measure for it.

Measurement 1 of 3

The Atlantodental Interval (ADI)

How the ADI is measured — step by step

The measurement is straightforward in principle. On a true lateral image — lateral plain radiograph, CT sagittal reconstruction, or MRI sagittal T1/T2 — two parallel vertical lines are drawn: one along the posterior surface of the anterior C1 arch, and one along the anterior surface of the odontoid. The horizontal distance between them, in millimetres, is the ADI.

The critical pitfall is imaging position. In neutral supine imaging, the ADI may be within normal limits. In flexion, the anterior C1 arch moves forward while the odontoid remains anchored — the gap widens. In patients with hEDS and transverse ligament laxity, the supine neutral ADI may be 2–3 mm, appearing normal, while the flexion ADI reaches 6–8 mm, indicating frank atlantoaxial instability. Unless flexion views are specifically requested, this instability is invisible.

The Steel Rule of Thirds — and why it matters

At the level of C1, the spinal canal can be conceptualised in thirds. In most adults, the cross-sectional space at C1 is approximately 3 cm in diameter. The rule, attributed to Harold Steel, holds that at this level the space divides roughly equally between the odontoid (one third), the cord (one third), and the available buffer space (one third). This buffer is the reason why early atlantoaxial instability can be neurologically silent — the cord has room to be displaced before it is compressed. But the buffer is finite. Once ADI widens beyond 5 mm in adults, the buffer has been substantially consumed and cord compression becomes imminent. Beyond 10–12 mm, the cord is being directly compressed by the displaced odontoid. The Steel Rule explains why patients with ADI in the 3–6 mm range can have significant myelopathic symptoms — the cord buffer has been partially consumed — even though the ADI number appears only mildly elevated.

The PADI — Space Available for Cord

The Posterior ADI (PADI) — Space Available for the Cord

The PADI is arguably the more clinically urgent of the two interval measurements. The ADI tells you about ligament integrity. The PADI tells you about cord safety — which is more immediately actionable. A patient with ADI of 6 mm has significant instability; whether they have cord compression depends on the PADI. If the spinal canal at C1 is capacious (say, 17 mm PADI), the cord may be uncomfortable but not immediately threatened. If the PADI has fallen to 10 mm, the cord is at acute risk with any forceful cervical movement — a car accident, a fall, even aggressive manual therapy.

PADI below 13 mm is not a borderline finding. It is a structural warning that the cord is one bad movement away from acute injury. It should change what that patient is and is not permitted to do — immediately.

Measurement 2 of 3

The Basion-Dens Interval (BDI)

The basion: a landmark that requires care

The basion is the most anterior and inferior point of the clivus at the foramen magnum. It sounds simple. In practice, identifying it precisely on imaging is one of the most common sources of measurement error in craniocervical studies. On CT, the basion is best seen on thin-slice sagittal reconstructions; on MRI, it is identifiable but may be less precisely defined due to cortical bone appearing as signal void. A generous beam of cartilage in a patient with an incomplete cranial base ossification (relevant in young adults) can push the apparent basion anteriorly, artifactually increasing the BDI. Conversely, degenerative changes at the craniovertebral junction can obscure it. Measurement precision at this landmark directly determines diagnostic accuracy.

The invagination trap — BDI and basilar invagination

In basilar invagination — where the odontoid migrates superiorly into the foramen magnum — the BDI may appear reduced, not increased. The dens tip is closer to the basion because it has migrated upward toward it, not because the skull has migrated downward. This is the opposite pattern from dissociation. A reduced BDI in combination with an elevated Chamberlain's line measurement signals basilar invagination (covered in Article 4). An increased BDI signals distraction/dissociation. Always interpret BDI alongside the basilar invagination metrics — the directions of abnormality are diagnostically opposite.

Measurement 3 of 3

The Basion-Axial Interval (BAI)

BDI + BAI together — the Harris combined rule

In 1994, John Harris and colleagues proposed a combined rule: craniocervical dissociation is present when either the BDI exceeds 12 mm or the BAI falls outside the −4 to +12 mm range, or both. This combined application is now standard in trauma radiology — though, critically, it is rarely applied in the chronic CCI patient population, where mild-to-moderate dissociation develops insidiously over years rather than occurring acutely after impact.

BDI — what it catches

Vertical distraction of the skull away from the spine. Highly sensitive for atlanto-occipital ligamentous disruption causing cranial migration of the occipital bone away from the dens. Can miss purely horizontal (translational) dissociation. Reference: basion → dens tip direct distance.

BAI — what it catches

Horizontal translation of the skull on C2 — forward or backward. Catches anterior and posterior dissociation patterns that BDI may miss entirely. Together, BDI + BAI cover both axes of craniocervical dissociation. Reference: basion → posterior axial line horizontal distance.

The implication is direct: both measurements must be made. A patient with BAI of +15 mm (anterior dissociation) may have a BDI within normal limits if the skull is sliding horizontally without vertical distraction. Using BDI alone would miss the pathology entirely. The reverse is also possible. This is why the combined rule exists — not as a bureaucratic checklist but because the two axes of dissociation are genuinely independent and either alone is an incomplete screen.

The Traynelis Classification

The Traynelis Classification — Grading Craniocervical Dissociation

In 1986, Traynelis and colleagues published the classification system that remains in use for typing craniocervical dissociation (CCD) by the direction of skull displacement. Understanding the classification helps both clinicians and patients interpret reports — and understand why "Type I" does not mean "mild."

Traynelis Type II — vertical distraction — is the variant most frequently missed in chronic CCI presentations, because it does not produce the dramatic translational finding that radiologists are trained to look for. The BDI may be only mildly elevated (13–15 mm rather than dramatically widened), and in the absence of a known trauma mechanism, the finding can be overlooked or attributed to measurement variability. In patients with hypermobile connective tissue disorders and dysautonomia, a chronically elevated BDI in this range is not a measurement artefact. It is the structural basis of their symptoms.

Reference Table

All Four Measurements — Complete Reference

Measurement From → To Normal Abnormal Detects Modality Status
ADI Post. C1 arch → Ant. dens ≤ 3 mm (adult)
≤ 5 mm (child)		 > 3 mm adult;> 5 mm child Transverse lig. laxity; atlantoaxial subluxation Lat. X-ray, CT, MRI Flexion essential
PADI (SAC) Post. dens → Ant. post. C1 arch ≥ 13 mm < 13 mm Cord compression risk at C1; safety buffer Lat. X-ray, CT, MRI Cord safety critical
BDI Basion → Dens tip (direct) ≤ 12 mm > 12 mm Vertical distraction; skull migrating from spine CT preferred; MRI CCD — vertical
BAI Basion → Posterior axial line (horizontal) −4 to +12 mm > +12 or < −4 mm Horizontal translation; anterior or posterior CCD CT preferred; MRI; X-ray CCD — horizontal
PADI and ADI are complementary — as ADI increases, PADI decreases. Report both. BDI and BAI together constitute the Harris rule for craniocervical dissociation. Neither alone is sufficient. All four should be measured whenever CCI, atlantoaxial instability, or craniocervical dissociation is suspected. In children, ADI threshold is higher (≤ 5 mm) due to physiological ligamentous laxity; all other measurements use adult norms from approximately 8–10 years of age.
What It Feels Like

The Symptom Landscape — What These Structural Failures Do to People

Craniocervical dissociation in trauma announces itself with catastrophic neurological failure: tetraplegia, respiratory arrest, death. In connective tissue disorders, the same structural failure develops over a decade. The presenting symptoms are not acute paralysis — they are a progressively disabling constellation that gets misdiagnosed as anxiety, depression, functional neurological disorder, fibromyalgia, or hypochondria, because the structural basis is not being imaged correctly.

The patient with an ADI of 6 mm, PADI of 10 mm, and BAI of +14 mm who has been told their imaging is "essentially normal" is not a mystery. Their imaging is abnormal. The measurements just were not made.

Pitfalls

Where These Measurements Go Wrong

1
The supine neutral trap — the most common error
All four measurements are position-sensitive, and all can be normal or near-normal in supine neutral imaging. ADI is at its minimum with the neck in slight extension — exactly the position in a standard supine MRI head rest. A patient with hEDS and a pathologically lax transverse ligament may show ADI of 2.5 mm in this position and 7 mm in full flexion upright. BDI and BAI also worsen under gravitational load. Standard supine imaging, by design, captures the anatomy at its most stable. For dynamic instability, this is the worst possible imaging protocol — and it is the protocol used for virtually every MRI in current clinical practice.
2
Imprecise basion identification — produces BDI and BAI errors
The basion is the single most consequential landmark in craniocervical measurement. It is the reference point for both BDI and BAI. A 3–4 mm error in basion placement — entirely possible on low-resolution MRI or when the craniocervical junction is imaged at non-standard angles — produces errors in both measurements that cross diagnostic thresholds. CT sagittal reconstruction at ≤1 mm slice thickness is the gold standard for basion identification. When using MRI, a 3D volumetric sequence should be preferred. Always note on which imaging sequence and which plane the basion was identified.
3
Confusing the posterior axial line for BAI with other reference lines
The BAI uses the posterior axial line — the same line used in Grabb-Oakes: the posterior cortex of the C2 body extended superiorly. It is not the posterior C1 arch, not the posterior dura, and not the posterior edge of the odontoid tip. Substituting any of these reference points moves the reference line and produces systematically false measurements. In institutions where CCI morphometrics are not routinely performed, there is no standardised protocol for which reference lines to use, and improvisation produces unreliable results.
4
Measuring ADI but not PADI — missing the clinically critical number
The ADI is what radiologists typically know to measure, because atlantoaxial subluxation has been taught in radiology training for decades. The PADI — the space available for the cord — is less commonly included in reports. But the PADI is the number that tells you whether the cord is safe. An ADI of 5 mm with a PADI of 16 mm is a very different clinical situation from an ADI of 5 mm with a PADI of 10 mm. Both will generate the same sentence in a standard radiology report: "ADI is 5 mm, at the upper limit of normal." The cord compression risk is entirely invisible.
5
Applying adult thresholds to children
The ADI threshold in children under 8 years is 5 mm, not 3 mm — because paediatric ligamentous laxity produces a physiologically wider predental space. Applying adult thresholds to a 6-year-old with suspected atlantoaxial instability will generate false-positive findings. Conversely, in older children with connective tissue disorders (hEDS, Marfan, Down syndrome — all of which produce early atlantoaxial instability), the adult threshold should be applied from earlier in adolescence, as ligament maturation may be delayed or absent. Down syndrome in particular requires vigilance: atlantoaxial instability occurs in approximately 15% of cases and can be severe.
6
Missing BDI reduction in basilar invagination
As noted in the callout above: in basilar invagination, the dens migrates superiorly toward the basion, reducing the BDI. An unwary clinician who knows BDI >12 mm = dissociation might look at a BDI of 7 mm in a patient with basilar invagination and conclude the craniocervical relationship is normal. It is not — the dens is invaginating into the foramen magnum. BDI must always be read alongside basilar invagination metrics (McGregor's line, Chamberlain's line — see Article 4), never in isolation.
7
Not making the measurements at all
The most common error, by a large margin. Standard brain and cervical spine MRI protocols do not include craniocervical morphometric measurements. Radiologists who do not specialise in this region do not routinely make them. The ADI may be measured if obvious atlantoaxial subluxation is apparent on visual review — but the BDI, BAI, and PADI almost never appear in standard reports. A report describing the craniocervical junction as "unremarkable" or "within normal limits" in the absence of specific measurements is not reassurance. It is a statement that the measurements were not performed.
Associated Conditions

Conditions Frequently Found Alongside ADI / BDI / BAI Abnormalities

In CCI patients, abnormal ADI, BDI, or BAI measurements are rarely isolated structural findings. They coexist with, and in many cases contribute causally to, a cluster of conditions that are often diagnosed and treated separately — even when they share a single structural root.

Associated structural and neurological conditions

Chiari Malformation Type I: Cerebellar tonsillar herniation ≥ 5 mm below the foramen magnum. In some patients, Chiari is primary — a congenital or structural underdevelopment of the posterior fossa. In others, it is secondary: the crowding and mechanical distortion at the craniocervical junction caused by CCI and atlantoaxial instability deforms the posterior fossa, physically pushing the cerebellar tonsils into the foramen magnum. Treating Chiari with foramen magnum decompression without addressing the underlying CCI leaves the primary structural driver in place — and can result in recurrence, or worse, postoperative worsening.

Tethered Cord Syndrome (Occult or Symptomatic): Abnormal fixation of the conus medullaris to surrounding structures, preventing normal upward migration during growth and producing traction on the cord. In CCI, craniocervical dissociation and spinal cord traction can create a dual-traction scenario — the cord being stretched from above (by the unstable skull/upper cervical complex) and from below (by a tethered conus). Tethered cord with CCI is a recognised co-occurring combination in hEDS, particularly in patients with bladder dysfunction, lower extremity sensory changes, and low back pain added to their craniocervical symptom burden.

Syringomyelia: CSF-filled cavity within the cord, often caused by disruption of normal CSF flow dynamics at the craniocervical junction. When atlantoaxial instability causes dynamic obstruction of the foramen magnum during movement, CSF pressure gradients can drive fluid into the central canal of the cord. A syrinx on MRI — especially a cervical syrinx in a patient with CCI symptoms — should always prompt craniocervical morphometric measurement, because its presence may indicate that the junction is not just unstable but is actively disrupting CSF flow on a cyclical basis.

Basilar Invagination: Covered in detail in Article 4, but relevant here because elevated BDI/BAI and basilar invagination frequently coexist. The structural crowding caused by an invaginating odontoid contributes to ventral brainstem compression (captured by Grabb-Oakes), angular brainstem deformity (captured by CXA), and horizontal displacement of the craniocervical relationship (captured by BAI). The measurements form an integrated picture.

Retro-odontoid Pannus: Fibrocartilaginous tissue accumulating behind the odontoid in response to chronic atlantoaxial instability. When the ADI is chronically elevated, the C1-C2 interface undergoes repetitive abnormal micro-motion. The body responds by laying down fibrovascular tissue that further narrows the PADI. This means a patient with a chronically elevated ADI who has a "safe" PADI on current MRI may have a progressively shrinking PADI over months to years, as pannus accumulates — even without any acute new injury.

The CCI + Chiari + Tethered Cord triad — a specific diagnostic pattern

Patients with hEDS who present with the combination of CCI (abnormal ADI/BAI/BDI), Chiari malformation, and tethered cord syndrome represent a specific and under-recognised subgroup. This triad has been described by several craniocervical neurosurgeons — most notably in the series reported by Henderson and colleagues — and is associated with severe symptom burden, complex treatment decisions, and a risk of suboptimal outcomes if any one element is addressed in isolation. In a patient with hEDS, myelopathy, dysautonomia, and Chiari on imaging: always measure for CCI and always ask about bladder, bowel, and lower extremity symptoms that might indicate tethering. These are not separate diagnoses — they are a system failing together.

Treatment Landscape

What Abnormal Measurements Lead To — Treatment Overview

These measurements are not academic exercises. They are the quantitative basis on which the most consequential treatment decisions in CCI are made. The treatment landscape ranges from conservative management and orthotics to major reconstructive neurosurgery — and the appropriate choice depends critically on which measurements are abnormal, by how much, in what direction, and in what clinical context.

Conservative & Non-Surgical

Cervical orthotic support: A rigid or semi-rigid cervical collar can reduce dynamic atlantoaxial subluxation and provide temporary relief of positional symptoms. It is a management tool, not a cure. Prolotherapy / PRP / dextrose injections: Proliferant injections targeting the weakened craniocervical ligaments, particularly the transverse atlantal, alar, and apical ligaments. Evidence is limited but accumulating in the hEDS population. Some patients achieve meaningful ADI reduction and symptom relief. Physical therapy: Targeted stabilisation of the deep cervical musculature — particularly the semispinalis capitis and deep suboccipital muscles — to provide muscular compensation for ligamentous incompetence. Must be carefully selected; aggressive cervical manipulation is contraindicated when PADI is reduced.

Surgical — Indications & Options

Surgical indications: Progressive myelopathy, PADI < 13 mm (especially with cord signal change on MRI), failed conservative management over ≥ 6 months, severe dysautonomia attributable to craniocervical compression, acute neurological deterioration. Posterior occiput-to-C2 fusion (craniocervical fusion): The definitive surgical treatment for craniocervical dissociation and severe atlantoaxial instability. Anchors the occiput to C1 and/or C2 with instrumentation and bone graft. Eliminates motion at the junction — which is its goal, but also its cost. C1-C2 fusion alone: Appropriate when instability is isolated to the atlantoaxial joint without occipital involvement. Preserves some residual craniocervical motion. Odontoidectomy: Transoral or endoscopic resection of the dens when irreducible odontoid compression is the primary pathology — combined with posterior fusion for stability.

The surgical decision in CCI is among the most nuanced in all of spine neurosurgery — because the patients are often young, the symptoms are often multisystem, the measurements are often "borderline" by standard trauma thresholds (developed for acute injury, not chronic instability), and the surgical risks are non-trivial. The craniocervical junction is the most vascular and neurologically dense region of the spine. The appropriate specialist is a neurosurgeon with specific experience in craniocervical instability in the hypermobility population — not a general spine surgeon, and not a surgeon whose CCI experience is limited to acute traumatic dissociation.

For Patients

What to Ask For — Practical Steps

If you have existing imaging: Ask specifically whether the ADI, PADI, BDI, and BAI have been measured and reported. In most cases they have not. You can ask your clinician to request a formal re-read by a radiologist experienced in craniocervical morphometrics, or to refer you to a specialist who will perform these measurements themselves. Neurosurgeons who specialise in CCI in connective tissue disorders routinely measure all four values on any imaging they review.

If you are requesting new imaging: The referral should specify craniocervical morphometric measurements including ADI, PADI, BDI, and BAI. It should also request flexion and extension sequences — because neutral supine imaging is the least sensitive protocol for dynamic instability. If upright MRI or open-bore MRI is accessible to you, this is strongly preferred for CCI evaluation, particularly when positional symptoms are a prominent feature.

When reviewing a report: The absence of these measurements from a radiology report does not mean the values are normal. It almost always means they were not checked. A report describing the craniocervical junction as "unremarkable" or "no significant abnormality" in the absence of specific measurements is a report that cannot speak to the presence or absence of craniocervical instability — because the measurement that would reveal it was not made.


The ADI, BDI, and BAI are not obscure academic measurements. They are the numbers that quantify, precisely and reproducibly, whether the skull is maintaining its structural relationship with the spine. When those relationships are lost — when the atlas is sliding forward, when the skull is separating from the axis, when the cord has only 9 mm of space left in a canal that should give it 15 — the body that contains these structures is in structural danger. The measurements exist. The thresholds are known. The only remaining question is whether anyone will make them.