Down’s syndrome (DS), or trisomy 21, is a disorder caused by a chromosomal abnormality (an extra chromosome 21). Children with DS often have obvious growth retardation and multiple deformities (1). One-third of DS patients involve instability of the craniocervical junction (CVJ) (2), which is the main cervical spine problem in children with DS. Although CVJ instability is mostly asymptomatic (3, 4), the initial presentation of cervical instability can be catastrophic, often accompanied by respiratory insufficiency and potential risk of death.

Patients with DS are more prone to CVJ instability than the general population due to low bone density, inherent ligamentous laxity, high joint flexion, odontoid hypoplasia, and os odontoideum (OsO) (5, 6). Some children may develop atlantoaxial dislocation after minor trauma, resulting in severe spinal cord injury. Currently, the treatment of CVJ instability in children with DS, especially atlantoaxial dislocation, was mainly based on braces and halo external fixators, with poor treatment effects and many complications (7). However, there are few reports of internal fixation system therapy. Therefore, this study retrospectively analyzed and summarized the clinical data of 5 cases of atlantoaxial dislocation diagnosed with DS admitted to our hospital from 2017 to 2020, aiming to explore the clinical characteristics of atlantoaxial instability (AAI) or atlantoaxial dislocation (AAD) caused by pediatric DS and the effect of surgical treatment with posterior atlantoaxial screw fixation.

The general information of 5 children are shown in Table 1. The average age was 4.2 years (3–5 years), and the average follow-up was 22 months (18–30 months). 3 cases (60%) had impaired neurological function (abnormal gait) and 2 cases (40%) had fixed/painful torticollis. All children have atlantoaxial dislocation. 5 cases underwent preoperative CT examination, and OsO was found in all 5 cases.

The preoperative average ADI was 10.6 mm (range 8–14 mm) in the flexion position, 8.0 mm (range 5–12 mm) in the neutral position, and 7 mm (range 3–12 mm) in the extension position. All children received preoperative MRI, 60% (3/5) showed changes in spinal cord signals, and 40% (2/5) showed dural sac compression. Every child underwent posterior atlantoaxial screw fixation (3.5-mm diameter), and the average fusion level was 1.8 (1–2). All 5 cases wore the head-neck-chest brace for 3–6 months after the operation.

Complications occurred in 2 cases. 1 case had dural tear between the posterior arch of the atlas and the lamina during the process of separation and exposure and recovered well after timely suturing. 1 case had internal fixation breakage of the prosthetic joint and underwent revision surgery. Posterior occipital fixation was performed again, with changed angle and direction of the screw, a new fixed pivot screw, and increased amount of bone graft. After enough brace wearing, fusion was finally obtained. No vertebral artery injury occurred in all cases. At the last follow-up, all cases were fused according to imaging examination results, and the neurological function were all ASIA grade E (Table 2). Two cases were shown in Figure 1 and Figure 2.

OsO is the most common type of odontoid deformity (10), which is mostly accompanied by AAI, leading to atlantoaxial dislocation (AAD), spinal cord compression or paralysis. The etiology of OsO is still unclear, but the incidence of OsO is high in the DS population, especially in children and teenagers (11–13). In this study, all 5 children were accompanied by OSO and AAD.

In recent years, people have tried to apply the posterior cervical screw-rod system implantation for the treatment of atlantoaxial dislocation in DS patients. Although the sample size is limited, the results are encouraging. Ito et al. (14) used screw fixation to treat 2 DS patients with AAI, with improved neurological symptoms and good fusion. Yang et al. (15) reported 12 cases of CVJ instability or dislocation in DS patients with posterior cervical screw rod system fixation and achieved good results. Although nonunion occurred in 4 cases, there were no catastrophic complications such as death or quadriplegia. All 5 children in our study underwent posterior atlantoaxial screw fixation or occipitocervical fusion. 1 case had dural tear between the posterior arch of the atlas and the lamina during the process of separation and exposure and recovered well after timely suturing. 1 case had internal fixation breakage of the prosthetic joint and underwent revision surgery. Posterior occipital fixation was preformed again, with changed screw angle and direction, a new fixed pivot screw, and increased amount of bone graft. After enough brace wearing, fusion was finally obtained. This child had poor atlas development and large span of bone grafting area, and the child did not wear external fixation bracing postoperatively as required, which may be the reasons for the occurrence of bone nonunion. For children with DS, free odontoid ossicles and the discontinuity of the posterior arch of the atlas are prone to occur due to their poor coordination and inherent ligamentous laxity (16). Younger age, poor nutrition and bone quality, internal fixation firmness and holding force are all factors to consider when choosing a long-term firm external fixator after surgery. Therefore, for children with DS, we generally use head-neck-chest braces to ensure stability. We found that children with atlantoaxial fusion often get better fusion in 3 months, and children with occipitocervical fusion often need to extend the time of wearing the brace. Specifically, the external fixation support should be removed after the fusion effect has been confirmed by the review of cervical dynamic radiographs and atlantoaxial CT.

We found that early surgical treatment of atlantoaxial dislocation in children with DS has a high potential for neurological recovery. In our study, 3 out of 5 children had neurological symptoms before surgery. At the last follow-up, the neurological function ASIA grade of all children was grade E. Although dural sac compression and spinal cord edema can be observed in preoperative cervical spine MRI examination, our surgical strategy is mainly based on reduction and fusion, and we do not advocate indirect decompression. Since the AAD in most pediatric DS is reducible, we think additional decompressive surgery is not required after reduction.

In all cases, no preoperative traction was performed, but cranial traction was given under intraoperative anesthesia to evaluate the reduction of the atlantoaxial vertebrae. If it was reducible, posterior atlantoaxial screw fixation could be performed directly; if it was non-reducible, the anterior release was required (17). Previously, anterior decompression was limited to transoral release or odontoid excision, but several new approaches have emerged (18–20), some of which have been successfully applied in DS patients. Goel (18) adopted the techniques of posterior exposure of facet joint, joint surface separation and distraction, and reduction and internal fixation of atlantoaxial rotatory subluxation, which avoided the problems of infection and feeding in the anterior mouth release and achieved good results. Recently, transnasal endoscopic odontoidectomy has achieved good results in patients requiring odontoidectomy (19). We found that reactive soft tissue hyperplasia around the odontoid process, especially behind the odontoid process, is closely related to the progression of spinal cord compression. This reactive soft tissue may be cystic or fibrocartilaginous. Currently, the cause of these soft tissue hyperplasia is not well established. However, we found that the soft tissue of reactive hyperplasia around the odontoid was significantly reduced after fusion stabilization, and the spinal cord compression was significantly improved, indicating that anatomical reduction and stable fusion were more important than indirect decompression.

The following were some situations in which we think occipitocervical fixation (C0-C2) is needed: 1. Children with unstable atlanto-occipital joint (case 1); 2. For children younger than 4 years old, the inherent ligamentous laxity, poor bone quality, insufficient screw holding power, and the narrow operation space between the atlantoaxial vertebrae all hinders the sufficient and effective reduction operation. The choice of occipitocervical fixation has a relatively large operation space and a more accurate reduction effect (case 2); 3. Bone graft surface is lacking for children with dysplasia of the atlas, especially in the absence of the posterior arch. Moreover, the long-term dislocation of the atlantoaxial vertebra results in dysplasia of the lateral mass and concave in the articular surface, which is not conducive to the effective fixation of lateral mass screws. Therefore, occipitocervical fixation is selected to ensure the effect of bone graft fusion (case 2, case 5).

Previously, the evaluation of atlanto-occipital instability (AOI) was often based on Powers ratios, the Wiesel-Rothman technique, and Harris/BAI measurement, but all of them have their shortcomings. More recently, others have reported on the importance of mechanical deformity, described in terms of both the kyphosis of the cranio-cervical junction and the amount of pathological translation. The former is measured as the clival axial angle, or CXA (less than 135 degrees is potentially pathological) and the latter translation is measured from the Harris/BAI measurement in flexion subtracting the Harris measurement in extension: a translation of >4 mm is considered (21). In short, the assessment of AOI is very difficult, especially for children, since the assessment often requires CT and MRI examinations of the dynamic position. Children have a low degree of cooperation and often need to be performed under anesthesia, which make the clinical application difficult. Therefore, more feasible evaluation methods for AOI need to be further studied in the future.

CVJ instability is a major cervical spine problem in children with DS, which can cause trauma at any stage of fetal or child development. Therefore, regular screening may help clarify the cause and prevalence of odontoid abnormalities in DS patients and help prevent neurological deterioration. Screening can be accomplished by careful clinical examination and dynamic radiographs of the cervical spine in flexion and hyperextension (22). Some scholars (23) also recommended two screenings between the ages of 5–10 and 10–15 years old, respectively. MRI of CVJ in all children with DS can be a good choice for differentiating AAD from OsO. If AAI occurs, many authors recommend prophylactic surgery for all DS patients (24, 25). We suggest that children with DS should be screened regularly. Parents of children who have AAI should be educated, and some dangerous activities should be prohibited. If AAD occurs, surgery can be performed.

The small sample size and short follow-up were limitations in this study. Therefore, large randomized controlled trials remain to be done in the future.

Posterior atlantoaxial screw fixation can stabilize the upper cervical spine for DS children with atlantoaxial dislocation, improve the compression of the spinal cord, and promote the recovery of nerve damage. In our study, fusion and nerve recovery were achieved in all children, and the results were very satisfactory. Postoperative head-neck-chest braces are necessary for children, especially those with occipitocervical fusion.