Introduction. The entity of hindbrain herniation without myelodysplasia in the very young child has been poorly described. A retrospective analysis of children diagnosed with Chiari I malformation (CM I) before their sixth birthday is presented.
Methods. Since 1985, 31 children with CM I (0.3–5.8) years of age have been diagnosed at University of Iowa Hospitals and Clinics. Their records were reviewed for presenting symptoms, signs, radiographic findings, treatment, complications, and outcome.
Results. The average age at diagnosis was 3.3 years. Sixteen patients were under age 3. Chief presenting complaints included impaired oropharyngeal function (35%), scoliosis (23%), headache or neck pain (23%), sensory disturbance (6%), weakness (3%), and other (10%). Sixty-nine percent of children under age 3 had abnormal oropharyngeal function. Three patients under age 3 (19%) had undergone fundoplication and/or gastrostomy before diagnosis of CM I.
Common physical findings included abnormal tendon reflexes (68%), scoliosis (26%), abnormal gag reflex (13%), and normal examination (13%). Vocal cord dysfunction (26%, all under age 3) and syringohydromyelia (52%) were also seen.
Twenty-five patients were treated surgically at our institution with posterior fossa decompression, duraplasty, and cerebellar tonsillar shrinkage. Three patients were lost to follow-up. Ninety-one percent of patients reported improved symptomatology at last follow-up (mean: 3.9 years). Three patients required reoperation for recurrence of symptoms. Syringomyelia improved in all patients. Scoliosis resolved in 2 of 8 patients, improved in 5, and stabilized in 1. There was no permanent morbidity from surgery.
Discussion. We show that children with Chiari I abnormality are very likely to present with oropharyngeal dysfunction if under age 3, and either scoliosis or headache or neck pain worsened by valsalva if age 3 to 5. These symptoms are very likely to improve after Chiari decompression, which can be done with low morbidity.
Conclusions. Very young children presenting with oropharyngeal dysfunction, pain worsened by valsalva, or scoliosis should prompt the clinician to consider CM I as a possible cause.
Chiari first described different variations of hindbrain herniation through the foramen magnum in 1891, although 2 types are now known to be most common. A type II Chiari malformation (CM II) involves descent of the cerebellar vermis, medulla, and fourth ventricle, and is strongly associated with hydrocephalus and myelodysplasia. Included in the CM II are intrinsic brainstem histopathologic changes that can produce symptoms from bulbar dysfunction. Type I Chiari malformations (CM I) typically involve only descent of the cerebellar tonsils as a result of mesodermal pathology that produces a small posterior fossa. This malformation is often an isolated finding. Both types of Chiari malformations can be easily identified on magnetic resonance imaging (MRI) of the craniovertebral junction.
Although much has been documented regarding the presentation of very young children with CM II, less has been described regarding the effects of CM I on the same age group. The entity of CM I may cause symptoms because of impaction of the tonsils at the foramen magnum leading to cerebellar, brainstem, or cervical cord symptoms, and also because of disturbance of cerebrospinal fluid dynamics. The latter may produce spinal cord cavitation called syringohydromyelia.
At the University of Iowa Hospitals and Clinics, we noticed a trend of very young children with CM I presenting primarily with difficulties in swallowing. Several children had required prolonged treatment with anti-acid medications for esophageal reflux as well as enteral nutrition supplementation for poor feeding. We therefore decided to review our series in detail to best determine how this difficult age group should be diagnosed and treated.
Between 1987 and 2001, 31 patients <6 years of age were identified with CM I. To further characterize the effects of hindbrain herniation on this age group, we retrospectively reviewed these patients’ records and diagnostic procedures. The presentation symptoms, clinical signs, comorbidities, radiographic findings, treatment, complications, and outcome were reviewed. The data were pooled and statistically analyzed for significance when applicable.
Thirty-one patients were identified on reviewing the University of Iowa Pediatric Neurosurgical Database of the senior author (A.H.M.) using 2 search criteria: 1) the diagnosis of CM I, and 2) diagnosis before the sixth birthday. All but 1 patient were believed to have symptoms resulting from the CM I.
The series consisted of 18 females and 13 males. The age at the time of diagnosis ranged from 0.3 years to 5.8 years. The mean age at diagnosis was 3.3 years, and the median age was 3.0 years. Five patients were lost in follow-up. In the remaining 26 patients, the follow-up from the time of diagnosis ranged from 4 months to 12 years. The average duration of follow-up was 3.9 years.
The chief complaint prompting medical evaluation was most often related to impaired oropharyngeal function (aspiration, regurgitation, choking, dysphagia, abnormal vocal cord function, chronic cough, etc) in those patients less than age 3, whereas scoliosis and headache were more common in those at least age 3 (Fig 1). Only 1 child diagnosed with CM I beyond their third birthday had dysphagia, but this was present from age 2 and was not the chief complaint. Other complaints in the series included head or neck pain worsened by valsalva (42%), gait or motor impairment (26%), sleep apnea or snoring (29%), abnormal movements or postures (19%), sensory disturbances (13%), recurrent respiratory infections (13%), and developmental delay (16%).
Seven patients in the series were taking antacid or other gastrointestinal (GI) medications before diagnosis, and all were under age 3 (23% overall, 44% of those under 3). Three patients under age 3 had undergone multiple GI procedures including fundoplication and/or gastrostomy before diagnosis of CM I. Other testing before Chiari diagnosis included upper GI endoscopy in 9 patients, barium swallow studies in 5 patients, and 3 patients had upper GI series. Comorbidities included hydrocephalus (3 patients), basilar invagination (3 patients), Klippel-Feil syndrome (2 patients), Crouzon’s syndrome, Angelman Syndrome, developmental delay, atlas assimilation, neurofibromatosis, and spondylo-epiphyseal dysplasia. One patient required tracheotomy before diagnosis.
The most common neurologic abnormality at the time of presentation was abnormal deep tendon reflexes in 21 patients. Of these, 9 patients were hyperreflexic in all extremities, 3 were hyperreflexic in the lower extremities only, 2 were hyperreflexic in the upper extremities only, 2 were hyperreflexic but not otherwise specified, 3 were hyporeflexic in all extremities, 1 was hyporeflexic but not otherwise specified, and 1 patient had only upgoing plantar responses documented. Those patients with mixed reflex patterns involving differences between upper and lower extremities all had syringomyelia. Four patients had a diminished gag reflex. Four patients had normal examinations. Eight (26%) of 31 patients had scoliosis, and 6 (75%) of the 8 were over age 3 at diagnosis. All 8 children with scoliosis had syringomyelia. Seven of the 8 patients had abnormalities on neurologic examination. The overall incidence of syrinx for the entire series was 52% (16/31 patients). Five of 16 toddlers under age 3 had syrinx compared with 11 of 15 ages 3 or older (χ2; P = .019). The location of the syrinx was cervical in 1 patient, cervicothoracic in 5, thoracic in 5, and holocord in 5. In general, patients with larger syringes were more likely to have scoliosis.
For those patients with scoliosis, 3 of the 8 patients had 2 curves. Six patients had thoracolumbar curves, and 2 had thoracic. The average Cobb angle of the primary curve was 34° (Fig 2). Two patients had left-convexity primary curves. One patient was treated with a brace for 6 months after posterior fossa decompression, and 1 patient fitted preoperatively is wearing a brace 3 months from surgery. No other patient required additional treatment for scoliosis, and no patient has required a spinal surgery for scoliosis. One patient has had improvement of scoliosis without any treatment. This child was diagnosed at 8 months old after he was noted to have spinal curvature and abnormal head movements. He had hyperreflexia and Babinski’s sign on examination, with a small thoracic syrinx in addition to CM I on MRI. He has had improvement from a 28° Cobb angle at presentation to 11° at 7-month follow-up. Of the remaining 7 scoliosis patients treated with Chiari decompression, the scoliosis resolved in 2 patients, improved in 4 patients, and stabilized in 1 patient. The average Cobb angle at last follow-up was 16° (Fig 2).
The degree of tonsillar descent ranged from the C1 to C3 levels. There was not a correlation between the amount of tonsillar descent and severity of symptoms. Furthermore, no correlation was found between the degree of tonsillar descent in those patients with syrinx as compared with those patients that did not have a syrinx.
One patient was believed to be asymptomatic from hindbrain herniation. This patient was incidentally found to have tonsillar descent along with hydrocephalus. After treatment with ventriculoperitoneal shunting, the patient has done well over a 12-year follow-up period. One patient has had improved scoliosis without surgery. One patient underwent posterior fossa decompression at another institution and was lost to follow-up. Three other symptomatic patients were lost to follow-up before any surgical intervention. The remaining 25 patients underwent surgical treatment. At operation, foramen magnum and posterior fossa craniectomy is made, and limited resection of the posterior arch of the atlas is performed. An intradural exploration is made with lysis of adhesions around the fourth ventricle and shrinkage of the cerebellar tonsils. A cervical fascia duraplasty completes the procedure. Four patients, all of whom had scoliosis and syringomyelia, also received fourth ventricle-cervical subarachnoid shunts. In addition to posterior fossa decompression, posterior occipito-cervical fusion was performed in 6 patients—3 because of preexisting craniocervical instability, and 3 patients that required ventral decompression because of associated comorbidities.
All surgical patients or their parents reported improvement of symptoms after operation. At last follow-up preoperative symptoms resolved in 46%, improved in 46%, and were unchanged in 8% of the 26 patients with follow-up (Fig 3A). Neurologic abnormalities resolved in 31%, improved in 42%, and were unchanged in 27% (including those with normal preoperative examinations). To date, 12 patients of the 16 with preoperative syringomyelia have had postoperative imaging. This showed resolution of the syringomyelia in 1 patient, improvement in 10, and no change in 1 patient (Fig 3B). Despite a lack of radiographic change in syrinx, a 4-year-old patient clinically improved and had resolution of scoliosis initially after posterior fossa decompression (PFD) but developed recurrence of spinal cord symptomatology after 20 months. At reoperation she was found to have bony regrowth with compression at the craniovertebral junction. Subsequent MRI 1 year later revealed marked improvement in the syrinx after the second decompression.
There was no mortality in the series and no permanent morbidity related to surgery. Transient postoperative morbidity included nausea and vomiting (3 patients), superficial wound infection (2), fever (2), aseptic meningitis (1), and pneumonia (1). Three patients required reoperation, one of which was previously mentioned. A 15-month-old female was reoperated on for lysis of intradural adhesions 10 months after initial surgery after recurrence of dysphagia. The third patient initially treated with a combination of ventral and dorsal decompressions and posterior fusion at an outside institution has required multiple subsequent procedures for persistent symptomatology and revision of arthrodesis.
Unfortunately, the children we follow with prolonged gastrostomy tubes before posterior fossa surgery have not yet met requirements for removal. These children have other neurologic comorbidities but have also developed significant behavioral components contributing to their difficulties with oral intake.
Illustrative Case 1
A 33-month-old girl presented with a history of many episodes of aspiration pneumonia from 9 months old for which she was treated with Nissen fundoplication. She developed emesis and first underwent dilation followed by revision of the fundoplication. MRI was performed and showed significant cerebellar tonsillar descent in addition to ventral compression of the cervicomedullary junction (Fig 4). After combination transoral and posterior fossa decompressions, the patients oral intake improved and respiratory difficulties abated.
Illustrative Case 2
A 5-year-old girl presented with painless scoliosis. She had abnormal tendon reflexes on examination and a primary right curve (Cobb angle 30°; Fig 5A) with a compensatory lumbar curve. MRI revealed a septated holochord syrinx. After PFD, the scoliosis improved to 12° (Fig 5B) with 8-year follow-up, and the syrinx has dramatically deflated (Fig 6).
The clinical presentation of young children with CM I differs from that of older children and adults.1–3 Several important characteristics are evident in very young children with symptomatic CM I: 1) children under age 3 had significant oropharyngeal symptoms; 2) children over age 3 were more likely to present with scoliosis or headache worsened by valsalva; 3) all patients with scoliosis had syringomyelia; 4) scoliosis rarely required additional treatment after posterior fossa decompression.
Swallowing Difficulties in Children Under 3
Abnormal oropharyngeal function can manifest as cough, stridor, dysphagia, abnormal vocal cord movement, gastroesophageal reflux (GER), aspiration, recurrent respiratory infection, feeding intolerance, poor weight gain, failure to thrive, or regurgitation. In our series, abnormal function was present in 11 (69%) of 16 children under age 3 at diagnosis of CM. This is an important finding that is not obvious in the current literature.
The correlation between neurogenic dysphagia, vocal cord paresis, respiratory impairment, or growth retardation is well-known for patients with CM II.4–13 Both compression of the brainstem and intrinsic brainstem abnormalities are believed to be responsible for these symptoms.4 Less has been described regarding CM I and oropharyngeal difficulties, especially in this young age group.14–21 Most series combine both adult and pediatric age patients, with little reported on the prevalence of oropharyngeal symptoms and the difficulty or delay in final diagnosis. Those that have mentioned this correlation name a few patients or case reports.
Cai and Oakes5 state that in general dysphagia in CM I is differentiated from that in CM II because of longer duration of symptoms and subtle loss of swallowing function. Gendell et al8 described cricopharyngeal achalasia in 6 very young patients but all had type II CM. Genitori et al2 reported a series of CM I patients including 16 children with “brainstem compression,” 3 of whom had swallowing difficulties and 3 with apnea.
Itoh et al17 published a case report of a 9-month-old infant who developed anorexia and failure to gain weight followed by quadriplegia and respiratory distress. The patient then developed recurrent respiratory infections until posterior fossa decompression was performed at 17 months. The child subsequently improved. Milhorat et al18 had a large series of predominantly adult patients where dysphagia was present in 43%, sleep apnea in 38% and dysarthria in 31%. Nohria and Oakes19 reported a combined adult and pediatric series with 8% of patients having vocal cord palsy. Dure et al14 reported 2 newborns with apnea and stridor, 1 of whom resolved and the other improved after posterior fossa decompression. Pollack et al11 reported a series of 46 patients (CM I in 6, CM II in 40) with neurogenic dysphagia present in 15 patients. Only 2 of the 6 patients with CM I were children. A 15-month-old and a 3-year-old had “severe” dysphagia, which was defined as either recurrent pneumonitis or weight loss/failure-to-thrive. The 3-year-old had undergone tracheotomy, gastrostomy, and Nissen fundoplication. After posterior fossa decompression, both patients became asymptomatic with normal postoperative swallow studies and the tracheostomy was removed, as was the gastrostomy. The authors concluded that most individuals have acquired brainstem impairment causing neurogenic dysphagia, and patients with myelodysplasia with initial symptoms before 6 months old had poorer results. They recommended swallowing studies in all patients with dysphagia and warned that recurrence of symptoms could indicate recurrent compression or syrinx.11
In a separate presentation of 5 patients <3 years old, Pollack et al12 reported 3 patients diagnosed with CM I after evaluation of dysphagia. The other 2 patients had CM II. Additional anomalies were present in 3 patients. All patients had GER. After posterior fossa decompression, all 3 patients with incomplete relaxation of the upper esophageal sphincter had resolution on manometry studies. Esophagography showed normalization in 2 patients, whereas 1 patient had persistent narrowing of the upper esophageal sphincter despite clinical improvement. One patient had persistent vocal cord paresis. An additional patient noted resolution of dysphagia but had persistent dysphonia.12
Other authors have failed to see oropharyngeal impairment attributable to CM I. Choi et al6 did not see vocal cord paresis in 18 patients ranging from birth to 24 years of age. The series of Nagib consisted of 16 pediatric patients with 10 under age 6. Half of those 10 had sleep apnea, but there were no cases of dysphagia. In contrast, 4 patients over age 6 did have dysphagia.3
There are numerous causes of oropharyngeal dysfunction in young children in addition to CM. The differential diagnosis for very young patients with regurgitation includes anatomic obstruction such as esophageal webs or masses, malrotation, gastroenteritis, formula intolerance, increased intracranial pressure, inborn errors of metabolism, Reye’s syndrome, obstructive uropathy, hepatobiliary problems, and pancreatitis. It is important to consider reflux in patients who present with hoarseness, cough, aspiration, recurrent wheezing or pneumonia, chest pain, apnea, choking, difficulty swallowing, and failure to thrive.22,23
The pediatric literature describes 3 general classifications of GER, namely physiologic, functional, and pathologic. Physiologic and functional reflux are believed to be benign and require no diagnostic testing or intervention. Pathologic GER is frequent regurgitation resulting in complications such as failure to thrive or aspiration, requiring further diagnostic evaluation and treatment.22,23
It thus becomes necessary to entertain a neurologic cause for chronic, unexplained, and refractory reflux, failure to thrive, or respiratory difficulties in the population under 3 years old. Imaging of the brain and posterior fossa is part of this evaluation, and usually involves sedation or general anesthetic in this age group. However, the benefits of the study outweigh the small risks incurred.
Scoliosis in Children 3 to 5
It has previously been shown that syringomyelia in young children is strongly associated with scoliosis and CM I.31 It has also been shown that after decompression of the CM, both scoliosis and syringomyelia improve.2,14, 24,25,35, 36 Our present findings provide additional data to support those concepts, and extend them to this very young patient population.
Scoliosis in very young children has serious implications. Patients who develop large-curve scoliosis before age 5 are at risk for life-threatening cardiopulmonary complications.26 Infantile or juvenile scoliosis may be idiopathic, neurogenic, attributable to skeletal abnormalities, or other causes. The challenge for clinicians is to distinguish among causes to determine optimal treatment. Not all children with scoliosis require imaging of the posterior fossa and spinal column, but those patients with neurologic deficit, rapid progression of the scoliosis, or other concerning features merit a MRI.
In our series, all children with scoliosis had CM I by definition attributable to inclusion criteria. The published association of scoliosis and CM I ranges from 0 to 100%14,18, 19,27–38 of cases depending on the series focus, but is commonly believed to range from 20% to 40%.
All of our patients with scoliosis were found to have syringomyelia. Syrinx is found in 4% to 58% of scoliosis patients and is commonly found in those patients with scoliosis who have abnormalities on neurologic examination.30 However, up to 20% of scoliosis patients with normal examinations harbor syringomyelia.30 In children with infantile or juvenile scoliosis, the percentage of patients with syrinx can be much higher than older children.
The incidence of syrinx in patients with CM I is reported to be in the range of 19% to 75%.2,5, 14,18,19, 31,34,35, 39 Our incidence of 52% illustrates similarities in the very young age group. A significant improvement was noted in the size of the syrinxes following posterior fossa decompression, both with and without fourth ventricle to subarachnoid shunts. In 1 patient, clinical improvement was noted without change in the syrinx. This patient went on to develop recurrent symptoms necessitating reexploration.
Incidental CM I
This series reports 31 patients with CM I. Only 1 of the 31 children did not have symptoms referable to the malformation. This patient, as described above, was found to have hydrocephalus and was successfully treated with a ventriculo-peritoneal shunt. Unfortunately, a follow-up MRI was not obtained to follow the position of the tonsils.
Thus, we have not encountered a truly “incidental” CM I in our clinic in this age group. This fact is obviously dependent on referral patterns and likely is not representative of the entire population of children under age 6, however. As a tertiary care specialty clinic, nearly all of the children had undergone MRI and been diagnosed with CM I before referral to our clinic. Therefore, a referral bias is present.
The literature on asymptomatic CM is expanding. Genitori et al2 found 12 of 16 patients 0 to 2 years old and 2 of 7 patients 3 to 5 years old with CM I that were asymptomatic and remained so, although follow-up was only 1 to 12 months. Unfortunately, the authors did not quantify the degree of tonsillar descent. Also of note is the fact that 70% of 27 asymptomatic patients for the entire series had comorbidities, including hydrocephalus (33%), cranio-facial syndromes (19%), epilepsy (15%), and occult spinal dysraphism (4%).2
Similarly, our series had a significant incidence of comorbidities (39%). These complicated decisions regarding possible posterior fossa decompression for CM, but the fact that only 2 patients did not experience symptomatic improvement and only 3 patients did not experience improvement on neurologic examination after surgery shows that surgical decompression can be of benefit.
A recent survey of the pediatric section of the American Association of Neurologic surgeons showed a large consensus favoring the conservative management of asymptomatic CM I patients. Periodic neurologic examinations and reimaging were favored by 83% of respondents to follow patients with CM I or syringomyelia. On the other hand, all surveyed favored surgical intervention for symptomatic patients. The majority (61%) also recommended surgery in cases of radiographic syrinx progression without clinical change.40 Case reports do exist of both symptomatic and asymptomatic patients having radiographic improvement with conservative management.41,42
The degree of tonsillar descent is a factor in diagnosing a patient with CM I. Mikulis et al43 have shown that the tonsils for normal children in the first decade of life average 1.5 mm below the foramen magnum and suggested that 95% of these subjects have tonsillar tips within 6 mm below the foramen magnum. This is greater descent than the usually quoted 3- to 5-mm criteria that is “abnormal.”44, 45 The true incidence of asymptomatic CM I is unknown. Therefore, it is important to focus on the overall clinical situation of the patient, including symptomatology, detailed neurologic and physical examinations, and ancillary diagnostic tests, in addition to the radiographs when making management decisions.
This series illustrates that children with CM I under age 3 are likely to present with symptoms of oropharyngeal dysfunction. Children ages 3 to 5 are likely to present with either scoliosis or headaches or neck pain. Posterior fossa decompression is an effective and safe treatment for very young children with either isolated CM I or Chiari and syringomyelia. Furthermore, scoliosis responds well to posterior fossa decompression in patients with the triad of CM I, syringomyelia, and scoliosis.
- Received February 20, 2002.
- Accepted July 8, 2002.
- Reprint requests to (A.H.M.) Department of Neurosurgery, University of Iowa Hospitals and Clinics, 200 W Hawkins Dr, Iowa City, IA 52242. E-mail:
- ↵Fukushima T, Matsuda T, Tsuchimochi H, et al. Symptomatic Chiari malformation and associated pathophysiology in pediatric and adult patients without myelodysplasia. Neurol Med Chir (Tokyo).1994;34 :738– 743
- ↵Choi SS, Tran LP, Zalzal GH. Airway abnormalities in patients with Arnold-Chiari malformation. Otolaryngol Head Neck Surg.1999;121 :720– 724
- ↵Gendell HM, McCallum JE, Reigel DH. Cricopharyngeal achalasia associated with arnold-chiari malformation in childhood. Child Brain.1978;4 :65– 73
- Papasozomenos S, Roessmann U. Respiratory distress and arnold-Chiari malformation. Neurology.1981;31 :97– 100
- ↵Putnam PE, Orenstein SR, Pang D, Pollack IF, Proujansky R, Kocoshis SA. Cricopharyngeal dysfunction associated with Chiari malformations. Pediatrics.1992;89 :871– 876
- ↵Ruff ME, Oakes WJ, Fisher SR, Spock A. Sleep apnea and vocal cord paralysis secondary to type I Chiari Malformation. Pediatrics.1987;80 :231– 234
- ↵Werlin SL. Dysphagia. In: Hoekelman RA, ed. Pediatric Primary Care. 3rd ed. St Louis, MO: Mosby; 1997:918–922
- ↵Mardjetko SM: Infantile and juvenile scoliosis. In: Bridwell KH, DeWald RL, eds. The Textbook of Spinal Surgery. Philadelphia, PA: Lippincott-Raven Publishers; 1997:401–422
- ↵Sun PP, Harrop J, Sutton LN, Younkin D. Complete spontaneous resolution of childhood Chiari I malformation and associated syringomyelia. Pediatrics.2001;107 :182– 185
- ↵Sun JC, Steinbok P, Cochrane DD. Spontaneous resolution and recurrence of a Chiari I malformation and associated syringomyelia. Case report. J Neurosurg.2000;92(suppl) :207– 210
- ↵Barkovich AJ, Wippold FJ, Sherman JL, Citrin CM. Significance of tonsillar position on MR. AJNR Am J Neuroradiol.1986;7 :795– 799
- Copyright © 2002 by the American Academy of Pediatrics