First described by Hans Chiari in 1891, Chiari malformation type I (CM-I) is a hindbrain herniation that involves herniation of the cerebellar tonsils through the foramen magnum [1,2]. In 1894, Julius Arnold concurrently described herniation of the fourth ventricle and cerebellum through the foramen magnum, now named Arnold-Chiari Malformation [1]. These malformations are part of a spectrum of congenital hindbrain abnormalities and various types exist based on their primary etiology [3]. The most common type of Chiari malformation is the type I malformation and it is diagnosed on magnetic resonance imaging (MRI) with >5 mm descent of the cerebellar tonsils [3]. Patients with these conditions often present with multiple complaints including cranial nerve deficits, headache, neck pain, motor or sensory dysfunctions, and scoliosis [2].

Scoliosis is a spinal deformity often associated with CM-I. This association was first described in 1988 by Dauser et al. [4] in their report of seven patients with CM-I. The concomitant rate for the two conditions varies greatly with one study of 500 patients with CM-I reporting an 18% scoliosis rate while others report a rate as high as 50% [2,5]. While the exact pathogenesis of scoliosis in CM-I is still unknown, it has been associated with the presence of syringomyelia in patients with Chiari malformations [6]. Studies have reported a 25%–85% rate of scoliosis in CM-I patients with syringomyelia [6–9]. The presence of a syrinx has also been independently associated with scoliosis in a study of pediatric patients undergoing MRI [10]. While variations exist in the reported rates, CM-I with a syrinx is believed to have a higher prevalence of scoliosis than CM-I without a syrinx [3]. Furthermore, several studies have reported that the convex side of a scoliosis curve is often associated with the dominant side of an asymmetrically displaced cerebellar tonsil and the deviated side of an eccentric syrinx [11,12]. These reports suggest that tonsillar impaction, particularly on the dorsum of the cervicomedullary junction, is enough of an impetus to lead to scoliosis in patients with CM-I [3,12].

Several studies have reported cases of scoliosis in patients with CM-I without syringomyelia [6]. Most of these patients were small retrospective case series, with poor follow-up, in CM-I patients where only a handful of patients were found to have scoliosis without syringomyelia [6,13–15].

Posterior fossa decompression with or without spinal fusion is often performed to improve or stabilize scoliosis in patients with syringomyelia. Varying levels of success have been reported with some reporting scoliosis improvement in 38% and stabilization in 23% of patients who underwent a posterior fossa decompression for scoliosis in CM-I with syringomyelia [6–8]. Others have reported that decompression alone was adequate treatment for CM-I patients with syrinx and mild scoliosis less than 20° but patients with a curve greater than 20° required bracing or a spinal fusion in addition to the decompression [16].

While numerous studies report on decompression with or without fusion for CM-I patients with scoliosis there exists a paucity of literature on the risks and complication rates of these procedures. The purpose of this study was to utilize a large publicly-available pediatric database to investigate the risks and complication rates in patients with CM who underwent surgical interventions.

In this retrospective cohort study, we utilized a large publicly available pediatric database to investigate the risks and complications of surgery in patients with Chiari malformation and scoliosis when compared to patients with scoliosis alone or Chiari malformation alone. From a sample size of 90,707 total patients, we discovered that compared to the CM and CMS groups, patients in the Sc only group were significantly older with more medical comorbidities and a higher CCI. They also had significantly more invasive procedures performed than the other two groups. Patients in the Sc group also had a higher rate of fusion and osteotomies than the other two groups. This can be attributed to the fact that deformity correction surgeries often necessitate the need for osteotomies and extensive fusion constructs. The significantly higher percentage of 2+ levels fused in this group is also explained by this.

Patients in the CMS group had a significantly higher surgical rate as well as rate of surgical decompression than the other two groups. This can be attributed to the fact that this group contained patients with a Chiari malformation and scoliosis. The higher surgical rate and higher decompression rate are likely due to the need to address both issues.

The overall complication rate was higher in the Sc group as was the rate of infection related complications. This can be attributed to the extensive incisions that are often necessary for deformity correction surgery. Cardiovascular, respiratory, and dysphagia related complications were greater in the CMS group. This is likely due to the fact that the brainstem and cranial nerves are at risk during posterior fossa decompression surgeries for a Chiari malformation. Delays in surgery, though more common in CM cohort, lead to an increase in complications for all groups. The cause of such delays is itself difficult to assess, though previous analyses have demonstrated that in patients with CM, post-admission sepsis complications have been associated with mean operative delays of at least 3 days, and respiratory complications and anemia have been associated with delays of 8 to 14 days [18].

When controlling for age and invasiveness posterior thoracolumbar fusion had a greater OR for complications in the Sc group. For the CM group had higher complication rates if osteotomies or spinal fusion procedures were added to their decompression surgeries. The increase in complications with the addition of such extensive procedures is expected. Similarly, the increase in postoperative complications in the CMS group when spinal fusion was performed is also understandable. Furthermore, we identified that patients’ CCI is associated with complications and for all three groups higher CCI values were associated with increased OR for complications.

There is a paucity of literature comparing risks and complications after surgery in patients with Chiari malformation with scoliosis, Chiari malformation alone, or scoliosis alone. In one of the few studies published on this, Godzik et al. [19] compared the risks and outcomes of spinal deformity surgery in patients with CM-I associated scoliosis and those with adolescent idiopathic scoliosis. In their single institution study, they identified 36 patients with CMS who underwent instrumented fusion and matched them to 36 patients with AIS [19]. They found similar deformity correction and outcome scores between the groups [19]. While there were no differences in overall complication rate, they reported higher rates of neurological complications and neuromonitoring difficulties in the CMS group [19].

Numerous studies have reported on curve and deformity progression after suboccipital decompression for Chiari malformation associated scoliosis, and few have reported on risks and complications. In their retrospective study of 936 patients who underwent surgical decompression for CM-I, Greenberg et al. [20] identified a 12.5% postoperative surgical complication rate with hydrocephalus as the only comorbidity independently associated with increased surgical risk. In their study of 21 children with CMS who underwent hindbrain decompression, Attenello et al identified thoracolumbar junction scoliosis and failure of syrinx to improve as risk factors for postoperative deformity progression [21]. They also identified large preoperative scoliotic curves as another risk factor [21]. Similarly, Chotai et al. [22] reported a 41.2% improvement in scoliosis associated with CMN-I after posterior fossa decompression and identified preoperative curves >35° and adolescent-onset of scoliosis as risk factors for curve progression and need for corrective surgery. These findings were corroborated by Mackel et al. [23] in their study of 44 patients with CMS that underwent posterior fossa decompression. However, Ozerdemoglu et al. [24] reports that in their series of 105 patients with scoliosis and syringomyelia, syrinx shunting alone did not improve scoliotic deformity in any of the subjects studied. Numerous other studies have also reported curve progression rates ranging from 30%–70% in patients with CMS who underwent hindbrain decompression, and have identified double scoliosis curves and later age of decompression as additional risk factors [9,16,25].

This study is one of few that compares postoperative complications in patients with CMS to CM alone and scoliosis alone. It fills a knowledge gap that currently exists in literature. Specifically, though previous studies have focused primarily on the epidemiologic risk factors of Chiari malformation and concurrent scoliosis, this study uniquely focuses on the operative risks of corrective surgery. Furthermore, the utilization of a high-volume database in conjunction with novel multivariate analysis increases the power of resultant models, thus allowing for more accurate assessments of the aforementioned operative risk of corrective surgery. In their review of existing literature, Kelly et al. [26] reported that decompression surgery for the treatment of CM may itself result in resolution of abnormal spinal curvature, yet noted that many studies have suffered from relatively small sample sizes and concerns of underpowered findings. The present study suggests that due to the increased risk of operative complications in patients with concomitant CM and scoliosis, surgeons should consider primary decompression of CM prior to deformity correction. To our knowledge, our review of national data provides the first large-scale evidence that such prioritization may decrease postoperative complications overall and potentially reduce the rates of complications due to surgical delays in the treatment of concomitant CM and scoliosis in the pediatric population.

However, the present study has both strengths and weaknesses. As this study utilized a nationally available public database of pediatric inpatients it allowed for a large number of patients and their extensively collected data to be analyzed. A limitation of such a database is the inability to obtain additional information from patients and to prospectively plan studies. While large-volume datasets such as KID provide the ability to create high-power analyses, we acknowledge that captured encounters represent hospitalization records and not individual patients [27]. Furthermore, due to inherent limitations of national datasets, we are unable to comment on the ideal timing of scoliosis correction after CM decompression, indicating the need for more granular future studies. The retrospective nature of this study, may also introduce elements of reporting and selection bias. Future investigations on the risks, complications, and outcomes of surgery in patients with Chiari malformations and scoliosis is warranted to fully understand their disease processes.

This study demonstrates that surgery for Chiari malformation-associated scoliosis is associated with increased risk of cardiovascular, respiratory, and dysphagia-related postoperative complications. While the overall complication rates are higher in scoliosis alone patients, patient comorbidities increase the operative risk for when thoracolumbar spinal fusion surgeries are performed. Uniquely, this study also demonstrated that having concurrent scoliosis and Chiari malformation increases operative risk for fusion surgeries despite approach. Being independently inflicted with Chiari malformation also leads to an increased complication rate when fusions are paired with osteotomies.