The congenital variety of craniovertebral junction (CVJ) instability has been associated with various anomalies requiring three-dimensional spatial orientation. This feature sets it apart from the other acquired types, thereby warranting specific attention to better understand existing types of abnormalities that pose additional challenges and develop strategies to adequately stabilize the C1–C2 joint despite these anatomical deviances. Natural history shows variable progression, and late presentation is common in developing countries due to limited accessibility, affordability, and awareness, aside from the condition being frequently asymptomatic [1]. Available literature has been inconclusive, leading to several controversies owing to the multiple variations in its management. These controversies involve whether or not supplemental cord decompression is required after stabilization, which surgical approach is ideal to achieve optimal results (e.g., anterior, posterior, or combined), the appropriate surgical maneuvers to achieve reduction, and the extent of fixation and implant for fixation that would achieve a delicate balance between C1–C2 joint stabilization and cervical spine mobility. Complications range from surgical site infections and vertebral artery (VA) injury to worsening neurological status and death [2].

This study provides a systematic approach to understanding the anomalies associated with congenital atlantoaxial instability (cAAI), their effects on management, and the postoperative clinical, radiological, and functional outcomes of treatment at a tertiary care center. We have attempted to summarize both common and rare anomalies we had encountered during our experience with handling cAAI cases and briefly discuss how they were managed.

We hypothesized that appropriate surgical intervention would lead to measurable improvements in angulometric parameters, correlating with favorable clinical and radiological outcomes.

Among the identified patients with AAI, 24.2% had cAAI. Among the patients with cAAI, 3.87% had syndromic association, with two patients having Down’s syndrome, one patient having Klippel-Feil syndrome, and one patient having Morquio’s syndrome. The remaining 75.8% of the patients had AAI of other etiologies, such as post-traumatic or infective AAI, which were excluded from this study. The average age at presentation was 25.04 years (range, 4–55 years), and 64% of the patients were male. The average duration of symptoms at presentation was 14.5 months. Symptoms included supra-axial neck pain (88%), subjective neck motion restriction (68%), gait instability (76%), and torticollis (52%).

Bony and vascular anomalies were observed in 92% and 36% of the patients, respectively. The VA was hypoplastic (16%) or had an anomalous V3 segment (16%). Two patients had a unique vascular anomaly in the form of an abnormal retropharyngeal course of the internal carotid artery in one patient and an abnormal extracranial course of the sigmoid sinus in the other. The most common bony anomaly encountered was occipitoatlantal assimilation (68%), with 44% of the patients having subaxial-level fusion, with the most common being a C2–C3 fusion. Further details regarding bony anomalies and their incidence are provided in Table 1. Basilar invagination was observed in 21 of the 25 patients. Among the included patients, 12% had syrinx formation and Arnold-Chiari malformation (Fig. 2A, B), while 84% had cord signal intensity changes.

Among the 25 patients, 7 (28%) had irreducible dislocations requiring anterior release to achieve reduction. OCF was performed in 17 patients (68%), C1–C2 TAS in three patients (12%), and posterior fusion with C1–C2 lateral mass fixation (Goel-Harms method) in five patients (20%). Anterior facet cages were inserted in two patients followed by OCF, whereas one patient required facet cage insertion posteriorly (Supplement). One patient required os-odontoideum excision and ventral decompression. Two patients (8%) required foramen magnum decompression (FMD). The mean C1C2A correction was 20.61° (standard deviation [SD]=12.24; 95% confidence interval [CI], 15.812–25.408), with the average postoperative C1–C2 lordosis corrected to 22.36° (SD=5.68; 95% CI, 20.133–24.587). The mean preoperative POCA and CCA were 135.3° (SD=9.19; 95% CI, 131.698–138.902) and 121.50° (SD=17.52; 95% CI, 114.632–128.368), respectively, which improved postoperatively to a mean POCA of 116.2° (SD=23.17; 95% CI, 107.118–125.282) and CCA of 143.9° (SD=8.99; 95% CI, 140.376–147.424). The mean preoperative ADI and SAC were 7.22 mm (SD=2.56; 95% CI, 6.216–8.224) and 8.06 mm (SD=2.86; 95% CI, 6.939–9.181), respectively, which improved to a mean postoperative ADI of 4.15 mm (SD=0.99; C95% CI, 3.762–4.538) and SAC of 15.3 mm (SD=3.14; 95% CI, 14.069–16.531). All craniovertebral parameters improved significantly (p<0.05, paired t-test).

Among the 25 patients, 18 (72%) had neurological deficits (ASIA D=13, ASIA C=4, ASIA A=1), of whom 10 (40%) achieved full ambulation (Nurick 0–3), 5 (20%) were ambulatory with support (Nurick 4–5), and one was non-ambulatory (Nurick 6) at final follow-up after 1 year. Postoperative ASIA grades were ASIA A=1, ASIA C=2, ASIA D=5, ASIA E=15 (p>0.05 for neurological recovery, chi-square test). An individual analysis was conducted to assess the significance of neuroradiological outcome. Preoperative neurological deficits showed no significant association with underlying bony abnormalities, regardless of their type or syndromic context. However, a positive correlation was observed between preoperative neurological impairment and VA malformations. Following surgery, improvement in neurological outcomes showed no significant relationship with the presence of bony abnormalities, associated syndromes, or VA malformations. Tables 2 and 3 summarize the baseline comparators versus outcomes along with the p-values for these calculations.

Significant improvement in postoperative radiological outcomes was found in each anomaly subgroup when compared to preoperative craniocervical parameters. Surgical technique was selected based on the specific anomaly and the degree of reduction required. However, no significant differences in the correction obtained were observed among individual subgroups, indicating comparable correction across all anomaly subgroups.

Complications occurred in eight patients. Specifically, dural leak and transient hypoglossal nerve palsy were noted in one patient (4%) each, whereas surgical site infection occurred in three patients (12%), among whom one had a deep infection requiring lavage. One patient (4%) developed postoperative delirium, whereas another one patient (4%) developed VA injury and succumbed to brainstem dysfunction.

The C1–C2 joint allows for multiple degrees of motion across multiple planes. The C1 lateral mass is normally cuboidal, but patients with cAAI have a trapezoidal C1 lateral mass, which causes the normally flat C1–C2 facet to become anteroinferiorly inclined in the sagittal view. This inclination causes C1 to slip over C2, resulting in C1 overhang. Asymmetrical facet alignment leads to subluxation, which presents as torticollis [6]. Similarly, verticalization of the C1–C2 facet occurs (normal, 165°) with respect to the foramen magnum, which causes telescoping of C2 into C1 (i.e., basilar invagination) [6]. In the presence of os-odontoideum, the C1–Os complex translates anteriorly over C2, leading to AAI. The natural history of cAAI can be traced back to an altered bony anomaly, which leads to the concentration of stresses at the C1–C2 joint. Over a period of time, this stress concentration causes gradual attenuation of anatomical restraints due to the load exceeding the yield point of the collagenous ligaments. Once the limit of these protective mechanisms is crossed, frank instability, and eventually neurologic compromise, occurs. A similar theory, called the “Sandwich-Fusion theory,” has been described by Tian et al. [7] (Fig. 2C, D). Moreover, Goel et al. [8] suggested that changes, such as syrinx and basilar invagination, are protective rather than detrimental for the spinal cord and resolve after C1–C2 stabilization.

In our experience, approximately 25% of patients had AAI of congenital etiology, which contradicts the findings of Mehrotra et al. [9], who showed that 73% of patients had AAI of congenital etiology. This discrepancy may be attributed to our center being an apex tertiary health care center in a developing country that primarily caters to patients presenting with an infective (tuberculosis) or traumatic AAI. The average age at which patients presented in our study was 25.04 years, which is similar to the findings reported by Prajapati et al. [10] (29.93 years). Moreover, 40% of our patients with cAAI presented after 30 years of age. Given that AAI is frequently asymptomatic, its diagnosis is often incidental [1]. The symptomatology of our patients was similar to that reported in another study [11]. The most common symptom was supra-axial neck pain, which may go un- noticed by patients, contributing to late presentations, with studies reporting restricted ROM, torticollis associated with muscle spasm, sleep disturbances/apnea, dysphonia, high cervical myelopathy, stroke, or sudden death following trivial trauma. AAI in patients with Down’s syndrome often goes undiagnosed given that 30% have radiographic instability but only 1% are symptomatic. This may cause sudden neurological worsening, respiratory depression, and even sudden deaths [12].

Preoperative evaluation includes static/dynamic radiography, CT with angiography, and MRI [2]. Our suggested workflow and treatment selection have been summarized in Fig. 3. Certain high-risk groups, such as those with Down’s syndrome or congenital scoliosis, should be screened even when asymptomatic.

Similar to other studies [9], bony abnormalities, most commonly in the form of O–C1 assimilation (68%), were observed, which contributed to the irreducibility and may warrant extended fixation with OCF. Various bony abnormalities had been noted (Fig. 4A–F), each having implications on management (Fig. 5).

Preoperative CT is crucial for planning and ensuring a clinically and radiologically satisfactory outcome. When a vertically aligned facet is present with an irreducible type 1 basilar invagination, the ability to manually distract the facet posteriorly is limited. In such situations, an anterior approach with release, which utilizes the vertical facet, is necessary to achieve reduction and restore normal craniovertebral parameters.

In the treatment of basilar invagination and atlantoaxial dislocation, anterior release has traditionally been employed for irreducible cases [13]. However, recent studies have shown that posterior-only approaches can be safe and effective alternatives. Ma et al. [14] found that intra-articular distraction with customized spacers and posterior fixation can achieve realignment without anterior release. Duan et al. [15] reported successful reduction using posterior cage placement in congenital cases, thereby avoiding transoral morbidity. Similarly, Meng et al. [16] observed favorable outcomes with posterior distraction and occipitocervical fixation. These findings suggest that posterior-only techniques can reliably manage complex deformities with lower surgical risk in select patient profiles.

Facet dysmorphism contributes to the rotational component of AAI. In such cases, C1 inferior facet overhang occurs, which blocks access to the C1–C2 facet joint. Rongeurs are used to remove this bony overhang to visualize the facet and facilitate distraction. Removal of all offending structures, like the odontoid, can be another approach for managing asymmetrical facet alignment/dysmorphism. However, this approach has been associated with increased instability due to extensive loss of column support along with a risk for dural tears and meningitis. Therefore, facet distraction can be considered a better approach.

According to Wang’s criteria [13], an irreducible AAI needs to be managed with additional anterior release. We observed that although this is true for adults, in pediatric populations with a partially reducible AAD under anesthesia and some opening of the C1–C2 facet, the elasticity of the facet capsule and myo-ligamentous structures allows for effective reduction using posterior instrumentation alone using a joint jamming technique, often avoiding the need for global fusion.

Vascular anomalies are a major contributor to fatalities during corrective surgery for AAI. Occipitalized C1 is often associated with an anomalous course of the V3 segment of the VA (lying on the inferior surface of the C1 arch=“inverted V3”) [17], which complicates C1 lateral mass cannulation. This can be addressed by (1) subperiosteal dissection and mobilization of the inverted V3 segment caudally to identify the base of the C1 lamina in order to insert a C1 lateral mass screw or (2) skipping C1 lateral mass and performing fixation proximally up to the occiput with an occipital keel-plate construct. All but one patient with an abnormal course of VA required the anterior approach to release the tethering structures. The anomalous course of the artery may also be adjoining the C1–C2 facet joint posteriorly (persistent first intersegmental artery), which may complicate facetal distraction and fusion. In such cases, the anterior approach is preferred. Special care is required when dissecting around the dominant VA in patients with a contralateral hypoplastic VA. Notably, 16% of our patients had a hypoplastic VA, in contrast to another study wherein only 8.9% had the same [18]. Excessive bleeding due to compromise of the extracranial course of the sigmoid sinus can be avoided through preoperative identification. Another anomaly reported in the literature is a fenestrated VA [19]. Algorithmic management (Fig. 5) of common vascular anomalies is shown in Fig. 4G–J.

Irreducibility, which has been defined variably as non-reduction on dynamic radiography or under GA, warrants additional intervention to restore high cervical anatomy. For the purpose of this study, we defined irreducibility as non-reduction on traction under GA and complete muscle paralysis for 10 minutes (Wang’s criterion) [13]. Methods of reducing irreducible AAIs include (1) direct posterior reduction and fixation, with C1–C2 fixation (Goel-Harms technique), facetal distraction, reduction, and bone grafting [20], and (2) anterior release, which can be done through endoscopic/open methods via the transoral/transnasal [21]/extrapharygeal approaches. Transnasal endoscopic odontoid resection has been described in literature [21]. Regarding the selection of treatment options, one school of thought has been to perform an all-posterior approach, thereby avoiding the morbidity of the anterior approach. However, an all-posterior approach may lead to eccentric reduction. Another school of thought involves performing an anterior release to appropriately address the anterior myoligamentous tether and achieve biomechanically sound reduction, though this adds to surgical complexity, time, and blood loss. Nonetheless, techniques for posterior instrumentation have evolved over the years and have been described by various authors in the literature [11]. We herein used the retropharyngeal approach (lesser wound complications) as opposed to the transoral approach, where pharyngeal breach increases chances of meningitis in cases of dural tear. Sarat et al. [22] described the verticalization of the facet with a sagittal alignment of >160° and pseudoarthrosis between C2 pars and occiput (Fig. 1A–C), which had been managed using the distraction, compression, extension, and reduction technique. In our experience, the retropharyngeal anterior approach can be advantageous for the management of the vertical facet, given that the facetal orientation is parallel to the direction of soft tissue dissection. This approach can adequately release the myoligamentous structures and facetal capsule, after which facetal distraction and manipulation can be conducted. Two patients required anterior C1–C2 facetal spacer cage placement with OCF and stabilization. A similar procedure has been described for rheumatoid cervical spine [23].

If reduction is difficult, the SAC can be increased via FMD or anterior decompression through anterior C1 arch/os-odontoideum resection with or without duraplasty. The need for FMD in patients with syrinx or Chiari malformations remains controversial. In fact, Wang et al. [24] achieved 93% syrinx resolution with FMD, whereas Salunke et al. [25] achieved syrinx resolution with stabilization without bony decompression in nearly all of their patients. Two patients underwent FMD to obtain further improvement in SAC, with one patient each having anterior C1 arch resection and os-odontoideum resection.

Anterior decompression and plating can be done for difficult, untreated, or revision cases [26] across C1–C2 or as a clival plate [27]. The aim of reduction is the restoration of craniocervical angulometry to within normal biomechanical limits. It is measured using various angles, such as C1C2A, POCA, and CCA. The average C1C2A of the patients in our study was 22.36°. Choi et al. [28] found that patients with postoperative C1–C2 lordosis of <22° had lesser subaxial compensatory kyphosis and better long-term prognosis. Tang et al. [29] found that maintaining the POCA within a normal range of 108.2°±8.1° improved postoperative functional outcomes [30]. Fig. 2E and F demonstrates the angulometric correction, Fig. 6 details the various fixation methods, and Table 4 summarizes angulometric data.

Surgical management of AAI carries the possibility of life-threatening complications. Wound complication, which was the most common complication in the current study, occurred in 12% of our patients. Two such patients were superficial infections managed with antibiotics. One patient required surgical intervention and lavage, while another developed a VA injury during instrumentation. Moreover, one of our patients developed transient hypoglossal nerve palsy, which recovered spontaneously 1 month after surgery. This injury may occur due to overzealous retraction [31]. The management of CVJ abnormalities requires highly skilled manipulation and should always be performed with utmost care to ensure good outcomes.

Despite attempting to evaluate outcomes and summarize current protocols, some inherent limitations of our study cannot be overlooked, such as the single-center nature of the study, retrospective design, absence of power analysis, limited sample size, and the potential for measurement bias during angulometric measurements. The small sample size may also reduce the generalizability of our findings to broader populations with various forms of AAI, which warrants further multicenter studies with larger cohorts for validation.

Around a quarter of the patients included herein presented with AAI of congenital etiology. Although cAAI often presents in young to middle-aged adults, it is frequently asymptomatic. Preoperative assessment of cAAI requires a bird’s-eye view of all the components that may affect the outcomes. Managing each aspect of this anomaly requires an amalgamation of knowledge and skills pertaining to the CVJ. In the current study, 28% of our patients required anterior release; however, the decision for selecting the anterior approach depends on patient- and anatomy-specific factors. Restoration of cranio-cervical anatomy, including lordosis at C1–C2 to <22°, is essential for satisfactory outcomes. Prospective multicenter trials would be of help to further understand this enigmatic entity.