Surgical Intervention for CMI

First-line surgical therapy for patients with CMI in the absence of hydrocephalus is PFD via midline suboccipital craniectomy, with appropriate removal of the posterior arch of the atlas with or without dural opening (Figs. 58-2 through 58-4). PFD attempts to reestablish bidirectional CSF flow across the craniocervical junction. This is accomplished through the expansion of the posterior fossa subarachnoid space, thus decompressing the cerebellar tonsils and brain stem. PFD may also eliminate the craniospinal CSF pressure differential that is postulated to contribute to syrinx formation.⁵ For the remainder of this discussion, the acronym PFD, when used alone, refers to bony suboccipital decompression with dural scoring or splitting but without frank opening of both layers of the dura mater. Most commonly, PFD is undertaken with duraplasty (PFDD) with or without cerebellar tonsil coagulation or resection. Other intradural interventions, such as fourth ventricular stenting and syrinx shunting, have been largely abandoned or are reserved for second- or third-line therapies.^6–9

FIGURE 58-2 Positioning of the patient.

FIGURE 58-3 Skin incisions for PFD. The upper incision is used to harvest a periosteal graft for duraplasty.

FIGURE 58-4 Schematic drawing of the PFD. Sutures are shown retracting the dura, and the forceps are shown spreading apart the left and right cerebellar tonsils for identification of the floor of the fourth ventricle.

Studies Directly Comparing Surgical Techniques

Despite significant literature describing the treatment of CMI, no randomized trials comparing PFD against PFDD have been completed. Among the studies that directly compare the two techniques, only one meta-analysis and two surveys have been completed.^7,9,13

Durham and Fjeld-Olenec¹³ published a meta-analysis of studies that directly compare cohorts of pediatric patients who underwent PFD with cohorts who were treated with PFDD. A total of seven studies met their inclusion criteria.^{18,19,25–29} The authors concluded that patients who undergo duraplasty are less likely to require reoperation (2.1% vs. 12.6%) for persistent or recurrent symptoms but are more likely to suffer CSF-related complications (18.5% vs. 1.8%). There was no statistical difference in clinical outcomes between the two groups, specifically with regard to symptom improvement and syringomyelia. In summary, rates of clinical improvement were 65% in the PFD patients and 79% in the PFDD patients. Rates of radiologic syrinx improvement were influenced by small numbers in some studies but were 56% in the PFD patients and 87% in those undergoing PFDD. The authors appropriately acknowledged that their conclusions were limited by the patient selection methods of the studies they examined. Among the seven papers, five studies used intraoperative ultrasound to help determine whether or not to perform a dural opening.^{18,19,26,28,29} The inherent subjectivity of this technique limits the degree to which the findings of each work may be generalized. Additionally, no study included a randomization or blinding process.

Using expert opinion as an indicator of current practice, Haroun et al.⁷ reported that 25% of survey respondents would perform PFD for children with symptomatic CMI, 32% recommended PFDD, and 55% recommended further intradural manipulations (some respondents chose more than one intervention). International survey data published by Schijman and Steinbok⁹ demonstrated that 76% of pediatric neurosurgeons always open the dura mater for treating CMI.

Studies of PFD without Dural Opening

There is significant literature retrospectively examining the results of PFD and PFDD. With regard to outcomes, three general categories can be assessed: clinical improvement, syrinx resolution, and scoliosis progression.

Electrophysiologic evidence suggests that neurologic improvement can be achieved without a dural opening. Groups from Columbia University and Ohio State University found that improved conduction of nerve impulses through the brain stem occurs after bony decompression rather than after dural opening.^30–32 As mentioned previously, multiple groups have used intraoperative ultrasound findings to aid their decision-making with regard to dural opening in children with CM-I.^10,19,29,33 Yeh et al.²⁹ found that factors that were associated with adequate decompression without duraplasty included age of less than 1 year. Factors that were more likely to be associated with the need for duraplasty included spinal symptoms (motor, sensory, or scoliosis) and a greater magnitude of tonsillar descent.