Journal of Craniovertebral Junction and Spine

: 2022  |  Volume : 13  |  Issue : 3  |  Page : 344--349

Unilateral atlanto-occipital injury: A case series and detailed radiographic description

Jacob Richard Lepard, Logan A Reed, Steven M Theiss, Sakthi Rajan Rajaram 
 Department of Orthopedic Surgery, The University of Alabama at Birmingham, Birmingham, Alabama, USA

Correspondence Address:
Jacob Richard Lepard
Department of Orthopaedic Surgery, University of Alabama at Birmingham, 510 20th St South, Faculty Office Tower, Birmingham, Alabama 35294


Context: Atlanto-occipital dissociation is a highly lethal ligamentous injury at the craniocervical junction (CCJ). Previous studies have described rare cases of milder forms of atlanto-occipital injury (AOI) which might be managed nonoperatively, but there is a paucity of literature on this subject. Aims: We retrospectively reviewed our institutional experience to characterize the injury patterns, treatments, and clinical courses of patients with unilateral AOI. Methods: We included patients with radiographic evidence of unilateral occipitocervical joint capsular disruption, distraction, or edema ± injury of the apical ligament, tectorial membrane, anterior atlanto-occipital membrane, posterior atlanto-occipital membrane, alar ligaments, or cruciate ligament. The long-term outcomes were gathered from medical records, and six patients were available for Neck Disability Index via phone call at the time of the study. Results: Eight patients were included in the study. The mean age was 45.1 years ± 26.5. Causes of trauma included motor vehicle collision for five patients (5/8, 62.5%), falls for two (2/8, 25), and assault for one (1/8, 12.5%). All patients had a widened condyle-C1 interval >2 mm. Three patients underwent occipitocervical fusion, one patient underwent atlantoaxial fusion, and another received subaxial fusions for other injuries. Three patients underwent no surgical intervention. All patients were seen at least once as an outpatient following hospital discharge. There were no delayed neurologic injuries or deaths. Conclusions: We propose that ligamentous injury at the CCJ functions more as a spectrum rather than dichotomous diagnosis, of which a subset can likely be safely managed nonoperatively.

How to cite this article:
Lepard JR, Reed LA, Theiss SM, Rajaram SR. Unilateral atlanto-occipital injury: A case series and detailed radiographic description.J Craniovert Jun Spine 2022;13:344-349

How to cite this URL:
Lepard JR, Reed LA, Theiss SM, Rajaram SR. Unilateral atlanto-occipital injury: A case series and detailed radiographic description. J Craniovert Jun Spine [serial online] 2022 [cited 2023 Jun 2 ];13:344-349
Available from:

Full Text


Atlanto-occipital dissociation (AOD) is a highly lethal ligamentous injury at the craniocervical junction (CCJ) with high risk for the neurologic deficit.[1] Before the recent decades, there was minimal literature on this injury pattern due to the low likelihood of patient survival to tertiary trauma care.[2] However, with improvements in cervical immobilization and prehospital care, more patients are surviving to diagnosis and treatment.[3] As the literature surrounding, AOD has grown, one evolving topic is the clinical and radiographic definition of the injury.[4],[5] Because of the, often, subtle nature of the injury and the extremely high risk of neurologic compromise with delayed diagnosis, a variety of radiographic criteria, largely based on plain film and computed tomography (CT), have been established.[4],[6] However, with the increasing use of magnetic resonance imaging (MRI) in these patients, it is now understood that what was once considered a single diagnosis of AOD is more likely to represent a spectrum of ligamentous injuries that may warrant more nuanced approaches to treatment.[7],[8]

In particular, incomplete or unilateral AOD is an entity that has been described in some published series but is often analyzed as part of the broader AOD cohort without separate consideration.[6] Others have also proposed the concept of a milder craniocervical injury within the spectrum of AOD that can be managed more conservatively than the more severe forms;[7],[8] however, clinical data regarding this entity are lacking. As a result, the purpose of this study was to utilize our multidisciplinary institutional level 1 trauma center experience to retrospectively characterize and describe the injury patterns and clinical courses of patients with incomplete unilateral occipitocervical injury, which we have termed “Atlanto-occipital injury (AOI).”


Patient selection

After obtaining an approval from the Institutional Review Board at the University of Alabama at Birmingham, a case series was retrospectively constructed utilizing the Nuance Primordial Software (Product Information) to search all documented radiology reads from January 1, 2006 to August 1, 2021, for the following phrases and terms: ([edema OR disruption OR avulsion OR tear] AND [“Occipitocervical joint” OR “alar ligament” OR “apical ligament” OR “tectorial membrane” OR “anterior atlanto-occipital membrane” OR “posterior atlanto-occipital membrane” OR “cruciate ligament”]) NOT knee.

A total of 727 results were yielded by this initial search. The radiology interpretations from the queried scans were used to screen potential patients for inclusion by one of the primary investigators Jacob Lepard (JL). Following the initial screening process, the CT cervical spine and MRI cervical spine images were directly evaluated for study inclusion. Patients were included, if they were older than 18 years with a history of recent traumatic injury and radiographic evidence of unilateral occipitocervical joint capsular disruption, distraction, or edema ± injury of the following structures: apical ligament, tectorial membrane, anterior atlanto-occipital membrane, posterior atlanto-occipital membrane, either of the alar ligaments or transverse segment of the cruciate ligament. All patients with evidence of contralateral occipitocervical joint injury were excluded from the study. In addition, patients were excluded, if there was evidence of nontraumatic craniocervical pathology like infection or chronic degeneration. Demographic data and clinical outcomes were retrospectively obtained by chart review and summarized using descriptive statistics. Phone contact was attempted for all patients at the time of the study for updated assessment of their disability status due to their cervical injury. For those patients in whom contact was successful, Oswestry Disability Index, Neck Disability Index (NDI), and pain Visual Analog Scales (VAS) were obtained.[9] Injury severity scores (ISS) were obtained from a prospectively collected trauma database with a range of 1–75 and a score >15 indicating a major traumatic injury.[10]

Radiographic measurement

At our tertiary care level 1 trauma center, all patients undergo Advanced Trauma Life Support protocol after traumatic injury with a standardized imaging protocol. Using CT imaging of the cervical spine, the following parameters were measured using previously published methods: revised condyle-C1 intervals (CCI), Powers ratio, basion-axial intervals (BAI), and basion-dental intervals (BDI).[1] Measurements were considered to be abnormal, if the CCI was >2 mm, Powers ratio >1, BDI >10 mm, or BAI <−4 mm or >12 mm.[11] In addition, cranial CT was utilized to evaluate for cervicomedullary traumatic subarachnoid hemorrhage, and when available, CT angiogram of the neck was used to identify associated blunt cervical vascular injury. The occipitocervical joints, ligamentous structures, and the CCJ were evaluated using MRI cervical spine. Prevertebral soft-tissue swelling was measured on the midsagittal CT at the level of the C2 endplate with thickness >6 mm representing abnormal edema.[12] In one patient (Patient 7), CT imaging of the cervical spine was not available, and so all radiographic parameters were measured using MRI only. No patients underwent cervical traction testing.



A total of eight patients were included in the study. There were six (6/8, 75%) male and two (2/8, 25%) female. The mean age was 45.1 years ± 26.5 and with a range of 19–91. There were six Caucasians (6/8, 75%) and two African–Americans (2/2, 25%). Causes of trauma included motor vehicle collision for five patients (5/8, 62.5%), falls for two (2/8, 25), and assault for one (1/8, 12.5%). Five patients (5/8, 62.5%) had an American Society of Anesthesiologists score of 3, two patients (2/8, 25%) had a score of 4, and one patient's score was not documented. ISS were available for eight patients. The average ISS for the cohort was 18.9 ± 10.9 with a range of 10–41 [Table 1].{Table 1}

Radiographic findings

On CT imaging, five patients had a right-sided sagittal CCI, and three patients had a left CCI that was >2 mm. One patient had a BDI >10 mm. No patients were found to have an abnormal BAI or Powers ratio. Significant prevertebral edema was present in three patients (3/8, 37.5%). One patient had noncompressive retroclival subdural hemorrhage, and there were no patients with cervicomedullary subarachnoid hemorrhage. One patient had a unilateral intimal injury of their vertebral artery at the site of a coincident subaxial injury. In the remaining seven patients, there was no evidence of vertebral artery injuries. Three patients had concomitant occipital condyle fractures, one patient had a Jefferson fracture, two had evidence of ligamentous injury at C1-2, two patients had cervical 2 fractures, and two patients had subaxial cervical spine injuries [Table 2].{Table 2}

MRI was performed on all included patients. Indications for MRI included evaluating injury at C2 or above in five patients, evaluating subaxial injuries in one patient, and cervical collar clearance in two patients that were intubated and sedated. Six patients had increased short tau inversion recovery (STIR) signal in an occipitocervical joint unilaterally that corresponded to increased CCI on CT imaging [Figure 1]. Two of the eight patients with unilaterally increased CCI did not have a corresponding STIR signal on MRI. Three patients had unilateral and three had bilateral alar ligamentous injury. Only one patient demonstrated injury to the atlanto-occipital membrane, while three had injury to the posterior atlanto-occipital membrane. One patient had complete and another had partial injury to the tectorial membrane. One patient had a complete rupture of the ascending portion of the cruciate ligament, and another four had injury to the transverse ligaments. Two patients were found to have a complete disruption; another had a partial tear to the apical ligament. Finally, seven patients had increased STIR signal change in the interspinous ligaments and soft tissues between the occiput and cervical 2 [Table 3].{Figure 1}{Table 3}


On presentation, six patients were Glasgow Coma Scale (GCS) 15 and neurologically intact and another two patients were GCS 10T without evidence of deficits on the neurological examination. Three patients underwent occipitocervical fusion, one patient underwent atlantoaxial fusion, and another received subaxial fusions for other injuries. Three patients underwent no surgical intervention and were treated with Aspen collar only. The average hospital stay was 11.3 ± 11.1 days with a range of 5–36 days. Four patients were discharged directly to the home, two patients were discharged to inpatient rehab, one required skilled nursing facility care, and one died of their traumatic injuries. There were no delayed complications or late-onset neurologic deficits in this cohort. All patients were seen at least once as an outpatient following hospital discharge. The average follow-up was 26.4 ± 45.1 months with a median of 10.7 months and a range of 0.9–126 months [Table 4].{Table 4}

Clinical outcomes

Three patients were available for a phone follow-up at the time of data collection. Of these, all three complained of ongoing neck pain. The mean NDI score was 33.5 ± 6.06 and the mean VAS was 4.3 ± 1.5 at a mean interval of 55.4 months from injury. One of the four patients underwent occipitocervical fixation, whose NDI was 28% and VAS was 4. The remaining two patients did not undergo occipitocervical fixation and had a mean NDI score of 36.3 ± 5.3. Based on the NDI disability scale, the patient who underwent occipital condyle fusion was rated as mild whereas of the three patients who did not require surgical fixation, two had a disability rating of moderate.


It is notable that before approximately 20 years ago, very few published series on AOD existed in the literature. As our understanding of this injury pattern and the biomechanics of the CCJ has grown, it is worth considering the possibility that injury to this mechanically complex region could represent a spectrum rather than binary classification of AOD versus not. In this case series, we have sought to emphasize this point by describing the radiographic findings and clinical course of eight patients with unilateral craniocervical injury that does not represent a complete AOD.

Horn et al. acknowledged the presence of these indeterminate cases in their institutional series of 33 survivors of AOD, in which five (5/33, 15.2%) were determined to be a Grade 1 injury, indicating normal CT findings based on the current diagnostic parameters (Powers ratio, BDI, and BAI), but with “moderately abnormal” MRI findings. All five of these patients were managed with external orthoses without late neurologic deterioration.[7] Similarly, Bellabarba et al. proposed a three-stage classification of AOD, in which Stage 1 patients were defined as “stable minimally or nondisplaced craniocervical injury, in which there is a sufficient preservation of ligamentous integrity to allow for nonoperative treatment.” Radiographically, these Stage 1 patients were defined as MRI evidence of ligamentous craniocervical injury with normal alignment on static radiographs and <2 mm distraction on provocative traction testing.[8] Because these Stage 1 patients were considered, by definition, to not be true AOD, Bellabarba et al. did not include them in their series of 17 survivors of AOD, and thus, their presentation and clinical course are not known. In our own series of eight patients with this injury pattern, the most presented after high-velocity motor vehicle crash or falls. On presentation, the vast majority were neurologically intact without lower cranial nerve deficit.

Despite the, often, dramatic instability exhibited by patients with a complete AOD, a unified diagnostic criterion remains elusive.[5] Instead of this, multiple radiographic screening measures have been devised, positive findings in any of which result in concern for AOD and consideration of soft-tissue imaging.[11] In particular, the revised CCI, which was originally proposed by Pang et al.,[13],[14] has been found to be highly sensitive in the diagnosis of AOD compared to other measures such as the Powers ratio, BAI, and BDI. The findings of our study are consistent with this, as all patients had an abnormal CCI unilaterally, which nearly always correlated with increased STIR signal on MRI. In contrast, only one patient had a mildly elevated BDI, and there were no abnormal BAI's or Powers ratios. Additional radiographic clues for craniocervical injury such as prevertebral swelling, vertebral artery injury, and cervicomedullary subarachnoid hemorrhage were not reliably present in our cohort. Given the definition for abnormal CT used by Horn et al., which did not include CCI, seven of our eight patients would have been considered a Horn Grade 1, while the single patient with an elevated BDI would perhaps have been a Grade 2. This highlights the importance of the CCI as a clinically important screening tool that can diagnose the mildest versions of AOI, in which there is clinically important ligamentous injury in the absence of gross instability. It also indicates that the common injury pattern in our cohort is likely consistent with the five Grade 1 patients managed nonoperatively by Horn et al.

This consideration of partial atlanto-occipital injuries begs the important question as to what degree of ligamentous injury can be tolerated and at which point the CCJ becomes unstable.

Biomechanical studies have attempted to elucidate the most important structures for the maintenance of stability at the CCJ with conflicting results. Phuntsok et al. performed a finite element analysis of the upper cervical ligaments finding that the greatest decrease in stability occurred with injury to the occipitoatlantal capsular ligaments.[15] In a cadaveric study, Radcliff et al. found the transverse and alar ligaments to be the primary stabilizers at the CCJ.[4] In their review of the craniocervical ligaments, Tubbs et al. noted that there is an ongoing controversy regarding the importance of multiple structures including the apical ligament and tectorial membrane.[16] The most likely synthesis of these studies is that multiple of these ligaments play important roles in the biomechanical integrity of the CCJ to a greater and lesser degree depending on the type of craniocervical motion being considered. While our own series lacks the power to definitively answer the question of what degree of ligamentous injury can be safely managed nonoperatively; it is noteworthy that only three of the eight patients with AOI in our series underwent occipitocervical fixation, while the remainder were managed with external orthosis. By definition, all patients in our series had unilateral injury to the occipitoatlantal capsular ligaments with varying degrees of integrity of the midline ligamentous structures. Despite all having similar injury patterns, there were no late neurologic deteriorations or missed instability in the nonoperative group compared with the operative. When contacted by phone, with a mean follow-up of 55 months from injury, the mean NDI score among the two nonoperative patients was higher than the one patient who underwent occipitocervical fixation. While this is a trend worth investigating, the low numbers available for phone contact preclude a solid conclusion regarding the impact of operative fixation versus conservative management on long-term pain scores.

Our study is limited by the retrospective nature of the data collection. In addition, while loss to follow-up is well-documented in the trauma population,[17] it remains a limitation in the phone follow-up component of our study such that less than half of the included patients were successfully contacted. Furthermore, because all treatment decisions were made by the consulting surgeon at the time of injury, it cannot be definitively known whether the same criteria for nonoperative versus operative management were used for all patients. Further investigation would benefit from a prospective cohort methodology, in which consistent operative criteria were used. Lastly, in focusing our search only to patients with unilateral injury we have likely introduced selection bias to the radiographic characteristics of our cohort. There are likely other anatomic varieties of incomplete AOD or AOI, which are not strictly unilateral, that might also have been included in this study. However, in attempting to demonstrate the concept of incomplete AOD, we sought to create as homogenous a cohort as possible regarding injury patterns. Future studies might include a multi-institutional collaboration to seek out larger numbers of these patients with incomplete AOD, from which better definitions and radiographic criteria might be derived.


We report here a series of eight patients who sustained atlanto-occipital injury as evidenced by unilateral injury of their occipitoatlantal capsular ligaments. While three underwent occipitocervical fixation, the remaining five required only external orthosis for their craniocervical injury. This raises the important consideration as, to whether injury to the biomechanically complex CCJ can be adequately simplified to a binary classification such as the presence or absence of AOD. We propose instead that ligamentous injury in this region is more accurately considered a spectrum, of which a subset can likely be safely managed nonoperatively. Further investigation will be required to adequately define safe operative criteria.

Financial support and sponsorship


Conflicts of interest

There are no conflicts of interest.


1Garrett M, Consiglieri G, Kakarla UK, Chang SW, Dickman CA. Occipitoatlantal dislocation. Neurosurgery 2010;66:48-55.
2Traynelis VC, Marano GD, Dunker RO, Kaufman HH. Traumatic atlanto-occipital dislocation. Case report. J Neurosurg 1986;65:863-70.
3Theodore N, Hadley MN, Aarabi B, Dhall SS, Gelb DE, Hurlbert RJ, et al. Prehospital cervical spinal immobilization after trauma. Neurosurgery 2013;72 Suppl 2:22-34.
4Radcliff K, Kepler C, Reitman C, Harrop J, Vaccaro A. CT and MRI-based diagnosis of craniocervical dislocations: The role of the occipitoatlantal ligament. Clin Orthop Relat Res 2012;470:1602-13.
5Kalani MA, Ratliff JK. Considering the diagnosis of occipitocervical dissociation. Spine J 2013;13:520-2.
6Gire JD, Roberto RF, Bobinski M, Klineberg EO, Durbin-Johnson B. The utility and accuracy of computed tomography in the diagnosis of occipitocervical dissociation. Spine J 2013;13:510-9.
7Horn EM, Feiz-Erfan I, Lekovic GP, Dickman CA, Sonntag VK, Theodore N. Survivors of occipitoatlantal dislocation injuries: Imaging and clinical correlates. J Neurosurg Spine 2007;6:113-20.
8Bellabarba C, Mirza SK, West GA, Mann FA, Dailey AT, Newell DW, et al. Diagnosis and treatment of craniocervical dislocation in a series of 17 consecutive survivors during an 8-year period. J Neurosurg Spine 2006;4:429-40.
9Bernstein DN, Greenstein AS, D'Amore T, Mesfin A. Do PROMIS physical function, pain interference, and depression correlate to the oswestry disability index and neck disability index in spine trauma patients? Spine (Phila Pa 1976) 2020;45:764-9.
10Abdelwahed HS, Martinez FE. ICU length of stay and factors associated with longer stay of major trauma patients with multiple rib fractures: A retrospective observational study. Crit Care Res Pract 2022;2022:6547849.
11Dahdaleh NS, Khanna R, Menezes AH, Smith ZA, Viljoen SV, Koski TR, et al. The application of the revised condyle-C1 interval method to diagnose traumatic atlanto-occipital dissociation in adults. Global Spine J 2016;6:529-34.
12Hadley MN, Walters BC, Grabb PA, Oyesiku NM, Przybylski GJ, Resnick DK, et al. Diagnosis and management of traumatic atlanto-occipital dislocation injuries. Neurosurgery 2002;50:S105-13.
13Pang D, Nemzek WR, Zovickian J. Atlanto-occipital dislocation: Part 1-Normal occipital condyle-C1 interval in 89 children. Neurosurgery 2007;61:514-21.
14Pang D, Nemzek WR, Zovickian J. Atlanto-occipital dislocation-Part 2: The clinical use of (occipital) condyle-C1 interval, comparison with other diagnostic methods, and the manifestation, management, and outcome of atlanto-occipital dislocation in children. Neurosurgery 2007;61:995-1015.
15Phuntsok R, Ellis BJ, Herron MR, Provost CW, Dailey AT, Brockmeyer DL. The occipitoatlantal capsular ligaments are the primary stabilizers of the occipitoatlantal joint in the craniocervical junction: A finite element analysis. J Neurosurg Spine 2019;1-9. doi: 10.3171/2018.10.SPINE181102.
16Tubbs RS, Hallock JD, Radcliff V, Naftel RP, Mortazavi M, Shoja MM, et al. Ligaments of the craniocervical junction. J Neurosurg Spine 2011;14:697-709.
17Madden K, Scott T, McKay P, Petrisor BA, Jeray KJ, Tanner SL, et al. Predicting and preventing loss to follow-up of adult trauma patients in randomized controlled trials: An example from the FLOW trial. J Bone Joint Surg Am 2017;99:1086-92.