Occipital-Cervical Fusion and Ventral Decompression in the Surgical Management of Chiari-1 Malformation and Syringomyelia: Analysis of Data From the Park-Reeves Syringomyelia Research Consortium

Abstract

BACKGROUND

Occipital-cervical fusion (OCF) and ventral decompression (VD) may be used in the treatment of pediatric Chiari-1 malformation (CM-1) with syringomyelia (SM) as adjuncts to posterior fossa decompression (PFD) for complex craniovertebral junction pathology.

OBJECTIVE

To examine factors influencing the use of OCF and OCF/VD in a multicenter cohort of pediatric CM-1 and SM subjects treated with PFD.

METHODS

The Park-Reeves Syringomyelia Research Consortium registry was used to examine 637 subjects with cerebellar tonsillar ectopia ≥ 5 mm, syrinx diameter ≥ 3 mm, and at least 1 yr of follow-up after their index PFD. Comparisons were made between subjects who received PFD alone and those with PFD + OCF or PFD + OCF/VD.

RESULTS

All 637 patients underwent PFD, 505 (79.2%) with and 132 (20.8%) without duraplasty. A total of 12 subjects went on to have OCF at some point in their management (PFD + OCF), whereas 4 had OCF and VD (PFD + OCF/VD). Of those with complete data, a history of platybasia (3/10, P = .011), Klippel-Feil (2/10, P = .015), and basilar invagination (3/12, P < .001) were increased within the OCF group, whereas only basilar invagination (1/4, P < .001) was increased in the OCF/VD group. Clivo-axial angle (CXA) was significantly lower for both OCF (128.8 ± 15.3°, P = .008) and OCF/VD (115.0 ± 11.6°, P = .025) groups when compared to PFD-only group (145.3 ± 12.7°). pB-C2 did not differ among groups.

CONCLUSION

Although PFD alone is adequate for treating the vast majority of CM-1/SM patients, OCF or OCF/VD may be occasionally utilized. Cranial base and spine pathologies and CXA may provide insight into the need for OCF and/or OCF/VD.

Graphical Abstract

Graphical Abstract.

ABBREVIATIONS

CXA

clivo-axial angle

FOR

frontal-occipital ratio

MRI

magnetic resonance imaging

OCF

occipital-cervical fusion

PFD

posterior fossa decompression

PRSRC

Park-Reeves Syringomyelia Research Consortium

SM

syringomyelia

VBSC

ventral brainstem compression

VD

ventral decompression

Whereas posterior fossa decompression (PFD) is commonly accepted as the primary treatment for pediatric Chiari type I malformation (CM-1) with or without syringomyelia (SM), the indications for and use of occipital-cervical fusion (OCF) and/or ventral decompression (VD) are less clear. Studies have examined variations in craniovertebral anatomy and their relationship to instability with implications for management.^{1,2,3-10,11-13} Clivo-axial angle (CXA) less than 125°, basilar invagination, medullary kinking, and a diagnosis of Chiari 1.5 malformation¹⁴ have been proposed risk factors associated with OCF treatment in a complex subset of CM-1 patients.^2,14 Though controversial, the presence of a syrinx and syrinx size may also confer additional risk for requiring OCF.^15-21

Patients with CM-1 ± SM may present with symptoms of ventral brainstem compression (VBSC) with associated cranial and cervical neuropathies.^3,8,21 Thus, assessment of symptomatic VBSC is important in choosing a treatment approach.^10,11 Grabb et al^22,23 described an indicator of VBSC, the pB-C2 line, and suggested that this parameter may provide insight into the need for OCF and/or VD; however, pB-C2 may not be sufficient as a stand-alone tool for decision-making.^7,23-25

Current studies of OCF and VD in the setting of CM-1 and SM are limited by modest sample sizes and institutional variability. The objective of this current study was to examine factors associated with the use of OCF and VD in a large group of pediatric CM-1/SM patients treated across the United States.

METHODS

Study Design

Subjects were enrolled across 30 Park-Reeves Syringomyelia Research Consortium (PRSRC) sites both retrospectively (n = 369) and prospectively (n = 268) from January 2011 to February 2017. A total of 30 sites contributed patients. This study was approved by the institutional review boards at the host institution and each participating center. All data were de-identified; no patient consent was required. All patients had tonsillar ectopia ≥5 mm, syrinx diameter ≥3 mm, and ≥1-yr postoperative follow-up from their index PFD procedure. Patients who underwent PFD with or without duraplasty were considered in a single cohort (PFD) (duraplasty status specified in Results). The study cohort was divided into three groups: PFD only (referent), PFD + OCF only (OCF group), and PFD + OCF + VD (OCF/VD group). All patients who underwent VD also had OCF.

Variables for Analysis

Demographic data and clinical findings at initial treatment and follow-up encounters were collected and recorded in the PRSRC database. Variables included diagnosis date, symptom onset date, and brain, cranial base, or spine abnormalities identified from clinical history and imaging. Radiological images were reviewed by two blinded, trained readers, including a board certified pediatric neuroradiologist. PRSRC standard magnetic resonance imaging (MRI) measures included: tonsillar position measured in mm below McRae's²⁶ line, maximum antero-posterior syrinx diameter in mm, syrinx length measured in vertebral levels, frontal-occipital ratio²⁷ (FOR), basilar invagination (odontoid process above McRae's²⁶ line), pB-C2,²² and CXA.²⁸ OCF and OCF/VD were recorded as upfront (performed before or ≤30 d after PFD) or delayed (performed after 30 d). Clinical course and additional treatments were documented on initial and subsequent encounters.

Statistical Considerations

Analyses were performed using SPSS version 25 (IBM Corporation, Armonk, New York). Variables were summarized as percentages (categorical variables), mean ± standard deviation (parametric variables), or median with interquartile range (non-parametric distributed numerical variables). Differences between groups were detected using the Mann-Whitney U test and one-way Games-Howell post hoc ANOVA test for continuous variables and Fisher's exact test for frequency data. Analyses were conducted at a predetermined alpha level of 0.05, with Bonferroni correction for clinical symptoms and signs.

RESULTS

Patient Characteristics

A total of 637 patients in the PRSRC database met study inclusion criteria (Table 1, Figure). A total of 12 (1.9%) patients underwent OCF, 9 of whom had upfront OCF, and 3 of whom had delayed OCF at 20.07 ± 17.53 mo following their initial PFD. VD was performed in four patients (0.6%), all of whom also had OCF. Individual patient level data are detailed in Table 2. Four selected case illustrations with figures have been published as Supplemental Digital Content (Text, Supplemental Digital Content 1 and Figures, Supplemental Digital Contents 2-5). No patient in the OCF or OCF/VD groups had preoperative hydrocephalus or craniosynostosis. Regarding clinical presentation, in the OCF cohort, 10/12 patients had some combination of extremity weakness, numbness, and gait ataxia and 8/12 had a combination of dysphagia, choking, vocal cord dysfunction, and apnea. For the OCF/VD cohort, 4/4 had a combination of extremity weakness, numbness, or gait ataxia, and 2/4 had choking and apnea.

TABLE 1.

Characteristics of the Initial Study Cohort (N = 637)

Variable	Valuea
Ageb	9.67 [5.98-13.33]
Gender
Male	269 (42.2)
Female	368 (57.8)
Race
Asian	10 (1.6)
White	515 (80.7)
Black or African American	62 (9.7)
Native Hawaiian or Pacific Islander	2 (0.3)
Other or unknown	48 (7.5)
Primary procedure
PFD	132 (20.8)
PFD with duraplasty (PFDD)	505 (79.2)
Revision PFD or PFDD	59
PFD → PFD	1
PFD → PFDD	18
PFDD → PFDD	37
PFDD → PFD	2
OCF onlyc	12
Upfront	9
Delayed	3
VD/OCFc	4
Upfrontd	2
Delayede	2

^{a a}Values are frequency (percent of entire cohort) unless otherwise specified.

^{b b}Values are median (interquartile range).

^{c c}Timing defined as upfront if within 30 d of initial PFD procedure, delayed if after 30 d.

^{d d}Both patients underwent VD on the same day as the index PFD and OCF.

^{e e}One patient underwent upfront OCF on the same day as the index PFD, with VD delayed by 6.9 mo. The other patient underwent OCF and VD on the same day 33.6 mo after index PFD.

FIGURE.

Flowchart depicting the study cohort.

TABLE 2.

Individual Patient Characteristics

Site	Age diagnosis	Gender	Procedurea	Revision	Revision intervalc	OCF ± VD surgery	OCF timingb	OCF intervalc	VD timingb	VD intervalc
1	10	Female	PFDD	–	–	OCF	Upfront	0	–	–
2	11	Female	PFDD	–	–	OCF	Upfront	0	–	–
2	4	Male	PFDD	–	–	OCF	Upfront	0	–	–
2	5	Female	PFD	–	–	VD + OCF	Upfront	0	Delayed	6.9
2	13	Female	PFD	–	–	VD + OCF	Upfront	0	Upfront	0
13	14	Male	PFDD	–	–	OCF	Upfront	0	–	–
14	17	Male	PFDD	–	–	OCF	Upfront	0	–	–
14	17	Male	PFDD	–	–	OCF	Upfront	0	–	–
14	3	Male	PFDD	–	–	OCF	Delayed	45.8	–	–
14	8	Female	PFDD	PFDD	11.7	OCF	Delayed	11.7	–	–
17	14	Female	PFDD	–	–	OCF	Upfront	0	–	–
24	15	Male	PFDD	PFDD	13.17	VD + OCF	Delayed	33.6	Delayed	33.6
25	17	Female	PFDD	–	–	VD + OCF	Upfront	22	Upfront	22
27	13	Female	PFD	–	–	OCF	Upfront	0	–	–
27	7	Female	PFDD	–	–	OCF	Delayed	20.2	–	–
28	11	Female	PFDD	–	–	OCF	Upfront	0	–	–

^aValues presented as either PFD with duraplasty (PFDD) or PFD without duraplasty (PFD) at initial procedure.

^bTiming defined as “upfront” if within 30 d of initial PFD procedure, “delayed” if after 30 d.

^cValues presented in days.

PFDD = PFD with duraplasty.

Clinical Parameters

Table 3 details all clinical characteristics; follow-up data are displayed in Table, Supplemental Digital Content 6. Relative to subjects who did not undergo OCF or VD, the OCF group displayed increased rates of dysphagia (4/12, P = .041) and ataxia (6/12, P = .004). For the OCF/VD group, diplopia (2/4, P = .008) and choking (2/4, P = .013) were increased compared to PFD alone. Neurologic exam findings of weak neck rotation (4/10, P < .001), upper extremity weakness (10/12, P = .042), hyperreflexia (5/12, P = .005), and gait instability (5/8, P = .001) were increased in the OCF cohort over the PFD cohort, whereas for the OCF/VD cohort, only weak gag reflex was significantly increased (2/3, P = .018). With Bonferroni correction, only gait instability and weak neck rotation was found to be significant for OCF (adjusted alpha = .001). As OCF and OCF/VD cases were not performed in all PRSRC centers, further analyses were conducted specifically within centers performing these procedures (Table, Supplemental Digital Content 7).

TABLE 3.

Comparison of Clinical Findings at Initial Treatment^a

Variable	PFD onlya	OCF	P value	OCF/VD	P value
Number in group (n)	621	12		4
Symptoms
Headache	346/621 (55.7)	8/12 (66.7)	.564	2/4 (50.0)	1.000
Neck pain	129/621 (20.8)	4/12 (33.3)	.289	0/4 (0.0)	.586
Back pain	101/621 (16.2)	1/12 (8.3)	.701	0/4 (0.0)	1.000
Balance/Gait Ataxia	87/621 (14.0)	6/12 (50.0)	.004	3/4 (75.0)	.010
Dysphagia	70/621 (11.3)	4/12 (33.3)	.041	0/4 (0.0)	1.000
UE weakness	56/621 (9.0)	1/12 (8.3)	1.000	0/4 (0.0)	1.000
LE numbness	55/621 (8.9)	2/12 (16.6)	.295	1/4 (25.0)	.314
LE weakness	45/621 (7.2)	1/12 (8.3)	.599	0/4 (0.0)	1.000
UE pain	42/621 (6.7)	1/12 (8.3)	.573	0/4 (0.0)	1.000
UE numbness	42/621 (6.7)	1/12 (8.3)	.702	0/4 (0.0)	.127
Seizure	36/621 (5.8)	0/12 (0.0)	1.000	0/4 (0.0)	1.000
LE pain	33/621 (5.3)	1/12 (8.3)	.488	0/4 (0.0)	1.000
Apnea	32/621 (5.2)	2/12 (16.6)	.132	1/4 (25.0)	.195
Choking	29/621 (4.7)	1/12 (8.3)	.445	2/4 (50.0)	.013
Diplopia	22/621 (3.5)	0/12 (0.0)	1.000	2/4 (50.0)	.008
Pulmonary infections	12/621 (1.9)	0/12 (0.0)	1.000	0/4 (0.0)	1.000
Tremor	11/621 (1.9)	0/12 (0.0)	1.000	0/4 (0.0)	1.000
Hoarseness	10/621 (1.6)	0/12 (0.0)	1.000	0/4 (0.0)	1.000
Trunk pain	9/621 (1.4)	0/12 (0.0)	1.000	0/4 (0.0)	1.000
Face numbness	9/621 (1.4)	0/12 (0.0)	1.000	1/4 (25.0)	.063
Vocal cord dysfunction	8/621 (1.3)	1/12 (8.3)	.159	0/4 (0.0)	1.000
Face weakness	6/621 (0.9)	0/12 (0.0)	1.000	0/4 (0.0)	1.000
Shortness of breath	6/621 (0.9)	0/12 (0.0)	1.000	0/4 (0.0)	–
Prior genetic history	46/621 (7.4)	3/12 (25.0)	.058	0/4 (0.0)	1.000
Neurologic exam
LE weakness	544/555 (98.0)	11/12 (91.7)	.228	4/4 (100.0)	1.000
UE weakness	543/557 (97.5)	10/12 (83.3)	.042	3/4 (75.0)	.103
Pinprick	30/251 (11.9)	1/5 (20.0)	.478	0/1 (0.0)	1.000
Gait instability	61/516 (11.8)	5/8 (62.5)	.001	2/4 (0.0)	.074
Hyperreflexia	55/493 (11.2)	5/11 (45.5)	.005	2/4 (0.0)	.067
Light touch deficit	35/422 (8.3)	1/8 (12.5)	.506	0/3 (0.0)	1.000
Nystagmus	42/525 (8.0)	3/12 (25.0)	.071	1/4 (25.0)	.288
Clonus	25/315 (7.9)	2/6 (33.3)	.083	0/1 (0.0)	1.000
Weak Gag Reflex	33/436 (7.6)	1/8 (12.5)	.474	2/3 (66.6)	.018
Babinski	14/233 (6.0)	1/6 (16.7)	.325	0/1 (0.0)	1.000
Hoffman	12/219 (5.5)	0/3 (0.0)	1.000	1/2 (50.0)	.114
Romberg	8/229 (3.5)	0/4 (0.0)	1.000	0/1 (0.0)	1.000
Dysmetria	9/352 (2.6)	1/5 (20.0)	.133	1/2 (50.0)	.056
Disconjugate gaze	12/620 (1.9)	1/12 (8.3)	.222	1/4 (25.0)	.081
Weak neck rotation	9/464 (1.9)	4/10 (40.0)	<.001	0/2 (0.0)	1.000
Facial sensation	7/473 (1.5)	1/11 (0.1)	.169	0/3 (0.0)	1.000
Proprioception	3/253 (1.2)	0/5 (0.0)	1.000	0/4 (0.0)	–
Facial weakness	6/549 (1.1)	0/12 (0.0)	1.000	0/3 (0.0)	1.000
Hoarseness	5/461 (1.1)	0/10 (0.0)	1.000	0/3 (0.0)	1.000
Extraocular palsy	5/620 (0.8)	0/12 (0.0)	1.000	0/4 (0.0)	1.000
Tongue deviation	3/548 (0.5)	1/12 (8.3)	.083	0/2 (0.0)	1.000
eWeak shrug	1/442 (0.2)	0/10 (0.0)	1.000	0/2 (0.0)	1.000

Bold value indicate statistically significant P values (p < .05).

^aCells show values reported as n of documented cases, with significance tested with Fisher's exact test with PFD only used as referent group. Alpha values set to with Bonferroni correction set to .001. In this table, the PFD Only column included subjects who underwent PFD with or without duraplasty, while the OCF column included subjects treated with PFD/D and OCF, and the VD column included subjects treated with PFD/D, VD, and OCF. UE = upper extremity, LE = lower extremity, DTR = deep tendon reflex.

Initial Radiological Measurements

Radiological and clinical exam findings are summarized in Table 4. Tonsillar position was significantly lower for the OCF group compared with PFD alone (P = .026). There was no difference among groups in syrinx diameter, syrinx length, or pB-C2 distance. For both OCF and OCF/VD groups, the CXA was lower compared to PFD, with mean measurements of 128.8 ± 15.3° (P = .002) and 115.0 ± 11.6° (P = .001) degrees, respectively; however, CXA was not significantly different when comparing OCF and OCF/VD groups (P = .128) Further examination by dichotomizing CXA to ≤125° or >125° showed CXA ≤125° in 60% of the OCF group (P < .001) and 75% of the VD/OCF group (P < .001).² FOR was not different for any of the groups. History of platybasia (3/10, P = .011), Klippel-Feil (2/10, P = .015), or basilar invagination (3/12, P < .001) all were increased within the OCF group, whereas scoliosis (4/12, P = 1.00) was not different vs the PFD-only group. For the OCF/VD cohort, only basilar invagination (1/4, P < .001) was significantly increased. Of note, only 5/12 OCF and 1/4 OCF/VD patients had cervical spine flexion/extension radiographs available for review. Table, Supplemental Digital Content 8 provides analyses for centers performing all procedure types.

TABLE 4.

Initial Radiologic Findings^a

Variable	PFD Only	OCF	OCF Mean Difference	P valuee	OCF/VD	OCF/VD Mean Difference	P valuee	OCF vs OCF/VD P valuef
Number in group (n)	621	12			4
Age of surgeryb	10.09 [7.12-16.74]	11.51 [7.10-16.74]	1.42 ± 1.50	.621	12.59 [7.34-15.79]	2.50 ± 2.54	.634	.930
Radiographic findingsc
Tonsillar Ectopia (mm)	12.5 ± 4.9	17.3 ± 5.4	−4.8 ± 1.6	.026	17.5 ± 5.8	−5.00 ± 2.9	.335	.999
AP Syrinx diameter (mm)	7.5 ± 3.1	7.1 ± 3.0	0.6 ± 0.9	.811	9.5 ± 3.0	−2.1 ± 1.5	.442	.349
Syrinx Levels	8.9 ± 4.6	7.6 ± 5.2	1.2 ± 1.7	.762	10.7 ± 6.1	−1.89 ± 3.5	.864	.735
pB-C2 (mm)	6.9 ± 1.2	8.4 ± 2.2	−1.3 ± 0.7	.164	6.8 ± 0.5	0.1 ± 0.3	.895	.146
pB-C2 ≥ 9 mmd	47/510 (9.2)	3/11 (27.3)	–	.079	0/4 (0.0)	–	1.000	.516
Clivoaxial angle (°)	145.5 ± 12.4	128.8 ± 15.3	16.8 ± 4.4	.008	115.0 ± 11.6	30.8 ± 5.8	.025	.128
Clivoaxial angle ≤ 125°d	31/511 (6.0)	7/11 (63.6)	–	<.001	3/4 (75.0)	–	.001	1.000
FOR	0.3 ± 0.04	0.3 ± 0.02	0.009 ± 0.007	.501	0.3 ± 0.01	0.02 ± 0.005	.226	.668
Spine Pathologyd
Platybasia	12/281 (4.3)	3/10 (30.0)	–	.011	0/3 (0.0)	–	1.000	.528
Scoliosis	230/637 (36.1)	4/10 (40.0)	–	1.000	1/3 (33.3)	–	.127	1.000
Klippel Feil	4/283 (1.4)	2/10 (20.0)	–	.015	0/3 (0.0)	–	1.000	1.000
Basilar invagination	11/281 (3.9)	6/10 (60.0)	–	<.001	3/3 (100.0)	–	<.001	.497

Bold value indicate statistically significant P values (p < .05).

^aCells show mean ± standard deviation (P-value from one-way ANOVA) for continuous variables and frequency (%, P-value from Fisher's exact test) data, comparing procedure to PFD only.

^bValues are reported median [inner quartile range].

^cValues are reported as mean ± standard deviation.

^dValues reported as n (%) of documented cases.

^ePost hoc analysis with Games-Howell tests, values reported as mean difference as compared with PFD only, significance.

^fPost hoc analysis with Games-Howell tests, values reported as significance between OCF and OCF/VD mean differences.

Radiologic Follow-up

A total of 359 (57.3%) PFD-only patients, 9 (75%) OCF patients, and 3 (75%) OCF/VD patients had radiologic measurements collected more than 30 d after surgery (Table 5). Mean follow-up was 3.53 ± 1.41 yr. There was a small but significant increase in pB-C2 observed in the PFD alone group on follow-up imaging (preoperative = 6.88 ± 1.24 mm and postoperative = 7.01 ± 1.25 mm, P = .045). There was also a small but significant decrease in CXA in the PFD alone group on follow-up imaging (preoperative = 145.48 ± 12.43° and postoperative = 143.21 ± 11.89° mm, P = .013). Not surprisingly, the OCF and OCF/VD groups still had significantly lower CXA (125.38 ± 20.46°, P = .007, and 121.0 ± 21.95°, P = .033, respectively) compared with the PFD group. When comparing pre- and postoperative groups, only the PFD-only group showed significant improvement for syrinx diameter (P < .001) or syrinx length in levels (P < .001), though these analyses were limited because of the small sample sizes of the OCF and OCF/VD groups. Table, Supplemental Digital Content 6 provides detailed follow-up on individual clinical signs and symptoms.

TABLE 5.

Postoperative Radiographic Follow-up^a,b

Variable	PFD only	P value	OCF	P value	OCF/VD	P value
Radiographic findings c
AP Syrinx Diameter (mm)
Preoperative	7.49 ± 3.09		7.10 ± 3.00		9.50 ± 3.00
Postoperative	4.06 ± 3.04	<.001	5.33 ± 3.56	.252	4.80 ± 3.03	.233
Change	−3.76 ± 3.76		−5.00 ± 3.90		−3.50 ± 7.78
Syrinx length (levels)
Preoperative	8.94 ± 4.63		7.60 ± 5.19		10.67 ± 6.11
Postoperative	7.04 ± 4.28	<.001	8.00 ± 5.18	.956	8.50 ± 0.71	.564
Change	−1.86 ± 4.03		2.16 ± 6.37		−0.50 ± 4.94
pb-C2 (mm)
Preoperative	6.88 ± 1.24		8.40 ± 2.17		6.75 ± 0.50
Postoperative	7.01 ± 1.25	.045	7.44 ± 1.51	.057	6.83 ± 1.84	.724
Change	0.60 ± 8.10		−1.00 ± 1.87		0.33 ± 1.16
Clivalaxial angle(°)
Preoperative	145.48 ± 12.43		128.80 ± 15.27		115.00 ± 11.55
Postoperative	143.21 ± 11.89	.013	125.38 ± 20.46	.829	121.00 ± 21.95	.703
Change	−2.15 ± 12.76		−3.75 ± 11.93		3.67 ± 9.29

Bold value indicate statistically significant P values (p < .05).

^aCells show mean ± standard deviation (P-value from asymptotic two-tailed Mann-Whitney U test) for continuous variables that are statistically tested within each group from preoperative and postoperative.

^bPostoperative follow-up refers to the initial radiographic imaging variables taken at least 30 d from primary procedure.

^cValues are reported as mean ± standard deviation.

AP = anteroposterior; OCF = occipito-cervical fusion.

Upfront vs Delayed Treatment Timing

Additional analyses regarding timing of either OCF (Table 6) or OCF/VD (Table 7) demonstrated no statistically significant differences in preoperative radiologic, clinical, or historical parameters between upfront and delayed approaches. Out of the 5 patients with delayed OCF (3 received OCF and 2 received OCF + VD), the mean time to OCF was 20.07 ± 17.53 mo. The mean time to delayed OCF/VD was 20.27 ± 18.85 mo for 2 patients, with 1 patient having an upfront OCF and the other delaying VD by 33.6 mo.

TABLE 6.

Comparison Between the Timing of Patients Who Underwent Occipital Fusion Only^a

Variable	Upfront	Delayed	P value
Number in group	9	3
Radiographic findings
Tonsillar Ectopia (mm)
Preoperativeb	15.4 ± 7.6	19.7 ± 7.4	.138
Syrinx diameter (mm)
Preoperative	6.7 ± 2.1	8.0 ± 5.0	.564
Postoperative	3.5 ± 3.7	8.0 ± 2.8	.159
Improvement	4.3 ± 2.5	2.5 ± 0.7	.355
pb-C2 (mm)
Preoperative	9.1 ± 2.2	6.7 ± 0.6	.062
Postoperative	7.7 ± 1.9	7.0 ± 1.0	.789
Improvement	1.7 ± 1.9	−0.3 ± 0.6	.058
Clival angle (°)
Preoperative	123.7 ± 13.4	140.7 ± 14.5	.110
Postoperative	125.2 ± 7.2	137.3 ± 22.1	.456
Improvement	8.0 ± 11.8	−3.3 ± 10.0	.294

^aPreoperative refers to measurements before index PFD or OCF.

^bPostoperative measures for tonsillar ectopia were not recorded due to the post-surgical absence of the opisthion. Cells show mean ± standard deviation (P-value from asymptotic two-tailed Mann-Whitney U test) for continuous variables.

TABLE 7.

Comparison Between the Timing of Patients Who Underwent Occipital Fusion/VD^a

Variable	Upfront	Delayed	P value
Number in group	2	2
Radiographic findings
Tonsillar Descent (mm)
Preoperativeb	14.5	20.5	.439
AP Syrinx Diameter (mm)
Preoperative	10.0	9.0	.683
Postoperative	3.0	8.0	–
Improvement	9.0	2.0	–
pb-C2 (mm)
Preoperative	7.0	6.5	.317
Postoperative	6.0	8.0	.157
Improvement	1.0	−1.0
Clival angle (°)
Preoperative	115.0	115.0	1.000
Postoperative	116.5	125.0	1.000
Improvement	1.5	8.0

^aPreoperative refers to measurements made before index PFD and OCF/VD.

^bPostoperative measures for tonsillar ectopia were not recorded because of the post-surgical absence of the opisthion. Cells show mean (P-value from asymptotic two-tailed Mann-Whitney U test) for continuous variables between upfront and delayed.

DISCUSSION

Key Findings

This is the first multicenter study to examine the use and timing of OCF and VD in the management of CM-1/SM patients. Our findings show that OCF ± VD was recommended for a minority of patients with CM-1/SM. Neurological exam deficits of gait instability and weak neck rotation were associated with performing OCF, whereas no clinical findings were associated with the performance of VD (with robust statistical correction). A prior history of basilar invagination was associated with OCF and OCF/VD. These data build upon previous work evaluating tonsillar position, syrinx size, CXA, and pB-C2 in the surgical management of CM-1 and SM.^{2,16,23,29-31} In this study, tonsil position and CXA were different in those undergoing OCF, whereas pB-C2 was not. When dichotomizing these variables as previously described,^2,22 pB-C2 ≥ 9 mm did show a higher (but nonsignificant) incidence in the OCF group, and presence of CXA ≤125° was significantly higher in both OCF and OCF/VD groups.

Interpretation of Findings

CXA has garnered interest as a radiological parameter in formulating risk of craniocervical instability and ventral cord and brain stem compression,² and our data extend these findings. Henderson et al³² suggest that OCF may be considered in patients with clinical symptoms, Chiari 1 or 0, and a CXA less than 135°. They also indicate that VD may be considered when CXA cannot be re-established to greater than 150°, and there are continued signs of ventral compression.³²

Although the OCF/VD group had lower CXA values (115.0 ± 17.5°) than the PFD (145.3 ± 12.7°) and OCF (128.8 ± 15.3°) groups, overall rates of OCF and VD in our purely pediatric study cohort were strikingly low. Improvement in CXA was modest for OCF/VD (121.0 ± 22.0°) and not observed in OCF alone (125.4 ± 20.5°). Variability in CXA improvement following fusion may be related to perioperative reduction or whether operative fusion was undertaken in Situ (eg, for fixed deformities/cervical anomalies, rather than predominantly hereditary connective tissue disorders, as in Henderson et al³³). In addition to symptomatic presentation, reducibility of basilar invagination may also dictate whether OCF is adequate to restore craniocervical alignment and improve ventral medullary compression.

Dlouhy, Dahdaleh, and Menezes³⁴ provided an algorithm for treatment of craniovertebral junction pathology, including patients with CM-1. According to these authors, cervicomedullary compression and/or craniovertebral junction instability may be subgrouped, with reducible pathology receiving posterior fusion alone, and irreducible compression or craniovertebral junction instability receiving dorsal decompression (dorsal encroachment) or VD (ventrolateral encroachment). Craniovertebral junction pathologies that remain unstable after decompression should then receive posterior fusion, but this should not occur until decompressions alone have proven insufficient.

In agreement with recent studies,^2,10 our data suggest that pB-C2 distance may not be associated with the occurrence of OCF or VD. VBSC may occur in 10% of CM-1 patients, typically presenting as lower cranial nerve neuropathies.³⁵ Interestingly, sleep apnea, a common symptom of medullary compression was not significant in our dataset, though formal sleep studies were neither required nor routinely collected. Goel¹³ has suggested that atlantoaxial instability may be causative in Chiari malformation, advocating for treatment with atlantoaxial stabilization without foramen magnum decompression. Our study was not designed to evaluate this assertion; rather, our study reflects the current practice of pediatric neurosurgery many centers across the country.

In this study, a history of spinal pathology was more frequent in those undergoing OCF, whereas basilar invagination was associated with OCF/VD. Klippel-Feil syndrome, which was found to be more common in the OCF group (2/12, P = .015), has an estimated occurrence of 3% to 5% of patients with CM-1, and recent linkage analyses implicate growth differentiation factors.³⁶ Klippel-Feil syndrome is also associated with a variety of craniovertebral junction anomalies, such as basilar invagination, platybasia, and C1-C2 instability,³⁷ further supporting our findings. In some cases, complex craniovertebral anomalies may be associated with clinical syndromes, many of which have a known genetic basis.^38-40 Our OCF cohort demonstrated a significantly higher rate of genetic findings (25% for OCF, 7.8% for PFD only, and 0.0% for OCF/VD). Plans for further detailed examination of the genetic influences in the PRSRC cohort is underway, as are prospective gene discovery projects for CM-1 and SM.

Limitations

We acknowledge several limitations. The minimum length of follow-up (1 yr) may not be long enough to capture some cases of delayed OCF and VD. The number of patients who underwent OCF or OCF/VD represent a small minority of the entire cohort, limiting the generalizability of the analyses. There is inherent institutional bias or center effect (and institutional experience) in offering these craniovertebral procedures (8 centers performed OCF, whereas 3 performed OCF/VD), and there may be bias towards OCF at certain centers or referral bias to these centers to surgeons who are more likely to perform these procedures. Our database did not capture CM-1 without SM who may have undergone OCF or OCF/VD. Further, there is an inherent bias introduced with upfront procedures, as those with delayed OCF or OCF/VD may have been offered surgery based on the failure of PFD alone to resolve the clinical symptomatology. There was overlap in radiological parameters in upfront procedures; for example, 118 patients in the PFD-only cohort had tonsillar ectopia exceeding the 17.3 and 17.5 mm mean tonsillar ectopia observed in the OCF and OCF/VD groups, respectively. Overlap of results between the PFD-only and OCF/VD cohorts may affect generalizability.

Despite defined PRSRC criteria, clinical history and physical examination data points in this multicenter, mixed retrospective/prospective study are likely to create variability in the data. Indeed, data on 61.5% of the study subjects were collected retrospectively and subject to sampling bias. Imaging was not standardized and certain radiographic (eg, cervical flexion-extension radiographs) and adjunctive tests (eg, sleep studies and swallowing studies) were applied variably across the PRSRC. Rigorous testing of lower cranial nerve function (eg, gag), important in assessing VBSC, may have been inconsistently applied. Clinical and radiological follow-up followed each institution's usual clinical practice and thus varied across centers. Genetic testing for CM-1-related connective tissue disorders was (and remains) limited, which likely limited the number of patients reported to have such disorders in our database. Prospective studies may be able to better define the patient population to be treated with OCF or OCF/VD.

We have investigated a limited number of clinically important and well-characterized craniovertebral morphometrics, which were agreed upon by PRSRC investigator consensus at the consortium's inception. In keeping with other recent findings in the literature,^1,2,23 a more comprehensive analysis of additional factors, including brain/spinal cord parameters (eg, obex position/Chiari 1.5), bony parameters (eg, atlanto-occipital assimilation), and posterior fossa volumes is planned in the near future. Similarly, of significant interest are the natural history and outcomes of patients who have complex craniovertebral junction anatomy but who did not undergo these advanced procedures.

CONCLUSION

The vast majority of CM-1/SM patients do not undergo OCF or OCF/VD. Clinical symptoms and signs of lower cranial nerve dysfunction and VBSC were common in patients undergoing OCF and OCF/VD. Patients with significantly greater radiological findings of tonsillar ectopia and lower CXA more often received OCF/VD, and patients with genetic anomalies more frequently received OCF, but most patients with a high degree of tonsillar herniation and a low CXA do not require fusion. Judicious application of OCF and OCF/VD is warranted in the care of pediatric patients with CM-1/SM until more rigorous understanding of outcomes based on uniform indications is established.

Funding

This publication was made possible through the generous support of Sam and Betsy Reeves, the Spears and O’Keefe families, and the many other contributors to the Park-Reeves Syringomyelia Research Consortium. Support was also provided by the Eunice Kennedy Shriver National Institute of Child Health and Human Development of the National Institutes of Health under Award Number U54 HD087011 to the Intellectual and Developmental Disabilities Research Center at Washington University.

Disclosures

The authors have no personal, financial, or institutional interest in any of the drugs, materials, or devices described in this article. Dr Limbrick has received research funds and/or research equipment for unrelated projects from Medtronic Inc, Karl Storz Inc, and Microbot Medical Inc. Dr Limbrick has received philanthropic equipment contributions for humanitarian relief work from Karl Storz Inc and Aesculap Inc.

Supplemental Digital Content 1. Text. Case descriptions. Selected cases to illustrate scenarios in which patients required upfront/delayed OCF and upfront/delayed OCF/VD.

Supplemental Digital Content 2. Figure. Upfront OCF in a 14-yr-old female with a history of Goldenhar syndrome and Klippel-Feil. Shown is a sagittal T2-weighted MRI demonstrating tonsillar ectopia of 18.6 mm, CXA of 120.2 degrees, pB-C2 of 7.0 mm, basilar invagination, and syrinx diameter of 3.6 mm. She underwent PFD with duraplasty with upfront occiput-C4 fusion and experienced symptomatic relief. No further treatment was necessary.

Supplemental Digital Content 3. Figure. Delayed OCF in a 7-yr-old girl with a history of scoliosis and symptoms of dysphagia and gait ataxia. A, T1-weighted sagittal MRI from her initial MRI demonstrating VBSC with cerebellar tonsillar ectopia of 14 mm, CXA 124°, and pB-C2 of 7.0 mm. B, She initially underwent PFD with duraplasty but later developed sleep disordered breathing, and imaging performed 16 mo after her surgery demonstrated a reduction in CXA to 113°. A total of 20 mo after her initial PFD, she underwent occiput-C2 fusion with symptomatic improvement.

Supplemental Digital Content 4. Figure. Upfront OCF and VD in a 17-yr-old with a history of headaches and gait ataxia. T2-weighted sagittal MRi demonstrating an 8.0 mm holocord syrinx, cerebellar tonsillar ectopia 13.0 mm, pB-C2 7.0 mm, CXA of 101°, and brainstem compression. The patient underwent PFD with duraplasty with occiput-C3 fusion and VD performed 22 d later. She remained stable 36 mo following surgery.

Supplemental Digital Content 5. Figure. Delayed OCF and VD in a 15-yr-old male who presented for evaluation of suboccipital headaches and was found to have nystagmus, disconjugate gaze, gait instability, dysmetria, and hyperreflexia. A, T1-weighted, contrast-enhanced sagittal image from his initial, preoperative MRI demonstrated severe cerebellar ectopia of 26 mm, syrinx diameter of 6.0 mm, CXA of 117°, and a pB-C2 of 7.0 mm. He underwent PFD with duraplasty but returned for follow-up 12 mo postoperatively reporting new headaches and sensory symptoms that were exacerbated by coughing or sneezing. B, He underwent revision PFD with duraplasty without relief. He subsequently developed intractable nausea and vomiting, upper extremity tingling, increased hyperreflexia, and wide-based gait. He underwent VD and occiput-C3 fusion 33 mo after his initial surgery. Afterward, his symptoms stabilized, and his sensory disturbances improved.

Supplemental Digital Content 6. Table. Clinical symptoms with radiologic follow-up.

Supplemental Digital Content 7. Table. Subset analysis of initial clinical symptoms and signs of centers that performed OCF or VD.

Supplemental Digital Content 8. Table. Subset analysis of centers who performed OCF or VD with initial radiologic findings.

COMMENT

Park-Reeves is a high quality, multi-institutional database analyzing a vexing issue in pediatric neurosurgery, the need to standardize treatment for Chiari malformation. In this article the authors present a nice review looking at a large cohort of Chiari patients, and analyzing those patients who required occipital cervical fusion with or without ventral decompression. It should be noted that the rates for requiring fusion are extremely low, at less than 2% of the entire population, and there seemed to be some degree of center bias, with only a few institutions in this cohort having performed most of the procedures. There is also substantial overlap in presentation, so many children with symptoms and radiographic findings that might have been fused at one center were not fused at others, reminding us that surgeons need to analyze each of their patients carefully as opposed to assuming there is a standardized algorithm for performing a fusion. In summary, a very well done study that reminds us the rate of fusion should be extremely low after Chiari decompression.

Mark Proctor

Boston, Massachusetts