Pathogenesis of Chiari I – Pathophysiology of Syringomyelia

EKG: electrocardiogram IOUS: intraoperative ultrasound NIH: National Institutes of Health PFD: posterior fossa decompression PFDD: posterior fossa decompression and duraplasty SAS: subarachnoid space I focus here on 4 important issues related to the Chiari I malformation and syringomyelia. By what mechanism does a Chiari I malformation produce syringomyelia? What is the pathogenesis of the Chiari I malformation? Are the mechanisms underlying development of a Chiari I malformation and syringomyelia similar? How, exactly, is the pathophysiology related to the goals of surgery? The studies of my colleagues and myself are essentially the only studies of physiology in which direct measurements have been made in patients in the modern era of medicine. Thus, I summarize and integrate much, but not all, the work that my colleagues and I have done in this area at the National Institutes of Health (NIH) and the University of Virginia over the past 30 yr. Why might we want to know the answer to these questions? For esthetic reasons for one. As he stated in the early 1900s, William Osler held that our ambitions as physicians are “To wrest from nature the secrets which have perplexed philosophers in all ages, to track to their source the causes of disease.”1 We also need to address these questions for practical reasons. Since the goals of treatment depend on disease mechanism, doing so would permit us to progress from empiric to logical treatment approaches, and allow us to select the most effective, while the least invasive, treatment. Moreover, there are many causes of syringomyelia at the foramen magnum and below (Table 1). Are we to expect different mechanisms of the origin and progression of syringomyelia for each? That seems most unlikely.TABLE 1: Causes of SyringomyeliaThere are, in general, 2 separate components to the clinical syndrome associated with the Chiari I malformation (Figure 1). Tonsil ectopia can exert external pressure of the cerebellar tonsils on the dura, brainstem, and spinal cord resulting in occipital pain, which may be exacerbated by coughing, sneezing, laughing, or straining. The pressure can also underlie weakness, spasticity, difficulty swallowing, and loss of proprioception. Surgical treatment almost always reverses these symptoms. On the other hand, expansion of a syrinx, which may cause pain, suspended sensory loss, weakness of the extremities, and loss of proprioception, can injure the spinal cord and produce damage that may or may not be reversed with successful treatment of the Chiari I and syringomyelia, residual damage that may underlie failure of the symptoms and signs to disappear after collapse of the syrinx.FIGURE 1: Midsagittal T2-weighted MRI of a young man who has slurred speech difficulty swallowing caused by compression of his lower brainstem and upper cervical segment of the spinal cord by a Chiari I malformation associated with syringomyelia.What is the best therapy? The optimal treatment is the least surgery that is safe and that eliminates the mechanism underlying the symptoms and progression. And that begs the question of what mechanism underlies the pathogenesis of the Chiari I malformation and of development and progression of syringomyelia. PRIOR THEORIES ON THE PATHOGENESIS OF SYRINGOMYELIA ASSOCIATED WITH THE CHIARI I MALFORMATION Until the mid-1990s, there were 3 generally recognized theories. Dr James Gardner first proposed his “water-hammer” theory in the 1950s,2 which states that “partial obstruction of the outflow of CSF [cerebrospinal fluid] from the fourth ventricle directs the systolic pulsations of CSF from the fourth ventricle through” a patent “central canal and transmits” a systolic “water-hammer” pressure wave “into the syrinx causing development and progression of” syringomyelia.3 The second theory was the cranial-spinal pressure dissociation theory proposed by Bernard Williams, who proposed that a block to downward, but not upward, flow of CSF in the subarachnoid space (SAS) prolongs elevation of intracranial pressure over spinal intrathecal pressure and forces fluid from the fourth ventricle down a patent central canal and into the syrinx, producing “communicating” syringomyelia.4 Thus, this proposed mechanism is based on abnormal venous pressures during Valsalva, coughing, lifting, and straining. Both theories require patency of the central canal between the base of the obex, into the spinal cord, and into the upper pole of the syrinx. However, all studies in the MRI era have shown only rare patients with a detectable channel between the fourth ventricle and the syrinx. In the third theory was Ball and Dayan suggested that the tonsils obstruct the rapid rostral movement of CSF from the spinal SAS to the cranial SAS during episodic increases in thoracic venous pressure causing sporadic increased spinal CSF pressure (during coughing, sneezing, and straining), which drives the CSF through the spinal cord surface and along extracellular paths to initiate and advance syringomyelia.5 However, there are no compelling physiological studies to refute or confirm any of these 3 proposals. INITIAL CLINICAL AND PHYSIOLOGY STUDY AND ITS IMPLICATIONS In the late 1980s, my colleagues and I began to study the pathophysiology of Chiari I and syringomyelia at the NIH. I had spent the 1975 to 1976 academic year, the year before I began neurosurgical residency at Vanderbilt University, The National Hospital for Nervous Disease, Queen Square, London, much of the time working with Professor Valentine Logue, who had a special interest in syringomyelia. Logue had shown that opening the dura without opening the arachnoid was effective.6 At the NIH, we used a surgical approach in which we opened the dura without opening the arachnoid, so that our measurements during surgery would not be compromised by changes in the physiology produced by loss of CSF. The critical step in that first study was caught on the video of ultrasound during surgery of a young man with a syrinx that extended from the thoracic segments to the very upper cervical portion of the spinal cord, permitting us to monitor it with the ultrasound during craniospinal surgery (see Video, Supplemental Digital Content 1).3 There was prominent pulsatile movement of the walls of the syrinx with each heartbeat. Thus, it initially appeared that Gardner was correct, and we prepared and presented an abstract the annual meeting of the American Association of Neurological Surgeons emphasizing that. Later, when I was giving our weekly conference at the NIH and using this video, it became apparent that the video was showing systolic compression, not expansion, of the upper portion of the syrinx with each heartbeat. As we examine it, it becomes obvious that the upper portion of the syrinx is being compressed, not expanded, during the exaggerated downward movement of the tonsils during systole. Furthermore, there is no expansion of the syrinx during forced inspiration until after the dura is opened, the therapeutic step that produced complete resolution of the Chiari I malformation and syringomyelia in this young man. These observations are incompatible with the theories of Gardner, Williams, and Ball and Dayan. In that initial study, anatomic MRI and phase-contrast MRI showed obstruction of the SAS at the level of the foramen magnum with no visible CSF space behind the cerebellar tonsils and no visible communication of CSF between the fourth ventricle and the central canal of the cord. Phase-contrast MRI demonstrated systolic downward movement of the syrinx fluid. These observations with anatomic and cine MRI are also incompatible with the theories of Gardner and Williams. Studies performed during the era of ventriculography led to important observations on the movement of ventricular and subarachnoid CSF during the cardiac cycle in normal persons. In the 1960s and 1970s, George du Boulay at The National Hospital for Nervous Disease studied the movement of CSF during the cardiac cycle during ventriculography in patients who proved to have normal studies. He showed that when a systolic bolus of blood is delivered to the brain, the third ventricle is rapidly squeezed by the rapid expansion of the cerebral hemispheres.7,8 This produces systolic downward movement of CSF in the spinal SAS. He demonstrated that most of the CSF moving in and out of head during the cardiac cycle is from the basal cisterns (cisterna magna), not from the fourth ventricle, and showed that the cerebellar tonsils are relatively immobile during the cardiac cycle in normal persons, but descend and plug the foramen magnum by their downward movement with each systole in patients with Chiari I malformation. The predominant buffer to prevent abrupt increases in the intracranial pressure during the delivery of a systolic bolus of blood to the brain is the cerebral venous system, but there is normally also about 0.75 mL of CSF that rapidly moves from the basal cisterns of the posterior fossa and into the rostral spinal SAS.7,8 Our observations during our initial clinical study suggested a previously unrecognized mechanism for progression of syringomyelia associated with occlusion of the SAS at the foramen magnum (Figure 2).3 Thus, successful treatment requires only that we eliminate that block to the rapid movement of CSF in the SAS at the foramen magnum. The study indicated that the mechanism of the development and progression of syringomyelia is on the outside, not the inside, of the spinal cord. It also was apparent that the same mechanism may underlie other types of syringomyelia—for instance, post-traumatic primary spinal syringomyelia, syringomyelia associated with a block of the spinal SAS by an arachnoid cyst, arachnoid web, etc. Thus, we had a hypothesis that we examined with a series of later studies.FIGURE 2: Schematic drawings of the foramen magnum region. Sagittal view, illustrating the proposed mechanism of origin and progression of syringomyelia associated with Chiari I malformation of the cerebellar tonsils. A, In normal subjects during systole (right), as the brain expands with the reception of blood, CSF moves from the fourth ventricle into the cisterna magna; a larger volume of fluid passes from the basal cisterns to the subarachnoid space of the upper portion of the spinal canal. The magnitude of the systolic pressure wave that is conveyed to the CSF in the spinal canal dissipates with increasing distance inferiorly. During diastole (left), CSF flows rostrally across the foramen magnum. B, With obstruction to the rapid to-and-fro movement of CSF in the SAS across the foramen magnum (ventrally by anterior displacement of the brain stem and posteriorly by the impacted cerebellar tonsils) during systole (center) and diastole (left), brain expansion during systole is accommodated by abrupt caudal movement of the tonsils. The piston-like effect of this movement on the partially isolated spinal subarachnoid space imparts an accentuated systolic pressure wave to the spinal subarachnoid CSF. This acts on the surface of the upper segments of the spinal cord, abruptly constricting the syrinx, propelling the fluid in it inferiorly, and increasing bulk movement of CSF into the cord. C, With surgical decompression of the foramen magnum and tonsils, occlusion of pulsatile CSF flow across the foramen magnum is eliminated, normal physiology is restored, and syringomyelia disappears. Reprinted with permission from Oldfield et al.3PROSPECTIVE STUDY OF THE PATHOPHYSIOLOGY ASSOCIATED WITH CHIARI I AND SYRINGOMYELIA We then investigated that hypothesis with a prospective study of the physiology of patients with Chiari I and syringomyelia, a study that was masterfully lead by my colleague, Dr John Heiss.9 For this we examined several aspects of the craniospinal anatomy and physiology during the cardiac cycle (Table 2). We studied 20 patients with syringomyelia and Chiari I malformation—before, during, and after surgery, and 18 normal subjects All pressure measurements were collected digitally, q30 ms. Pressure measurements, cine MRI, and intraoperative ultrasound (IOUS) were linked in time with the electrocardiogram (EKG). All patients received surgical decompression of tonsils and cisterna magna to produce free flow of CSF between the cranial and spinal SAS during the cardiac cycle.TABLE 2: Components of Prospective Study of the Pathophysiology of Syringomyelia Associated With Chiari I MalformationWe assessed obstruction of CSF flow at the foramen magnum with 3 approaches (Table 2). In the Queckenstedt test, a lumbar puncture is performed and the jugular veins are compressed in a controlled fashion while measuring the rate of rise of pressure in the lumbar CSF. A block of CSF flow slows the rate of rise in lumbar pressure (Figure 3).FIGURE 3: Queckenstedt test. Pressure recordings show partial occlusion of CSF flow. The rate of rise in intrathecal pressure is shown in a healthy volunteer (left panel) and in a patient with Chiari I malformation and syringomyelia (right panels) before surgery and before the craniocervical junction is decompressed during surgery. After surgery, the rise in intrathecal pressure becomes normal. icp = intracranial pressure; lum = lumbar; s = seconds. Reprinted with permission from Heiss et al.9There was no CSF space posteriorly at the foramen magnum in the but after the surgery that space was in the 18 normal For the there was a between the before surgery in patients and and between the before and after surgery. There was pressure and pressure in the cervical subarachnoid CSF space to the normal (Figure These to normal after surgery. The same was so for CSF pressures in the lumbar of the of a prospective clinical study of the pathophysiology of syringomyelia with Chiari I malformation. A, The pressures and pressures in the cervical and lumbar CSF are in patients in 18 healthy before surgery and to normal after surgery. B, in which the of rise in intrathecal pressure during in patients with Chiari I malformation and syringomyelia with of healthy are After surgery, the rise in intrathecal pressure becomes normal. C, showing that craniospinal was compromised before surgery and to normal after surgery. was by the of CSF by the in intrathecal pressure of that showing the syrinx at by using with pressure recordings the syrinx and cervical SAS. The syrinx during cardiac systole when cervical subarachnoid pressure Reprinted with permission from Heiss et the spinal CSF partially This was examined by measuring for an obstruction to free flow of CSF between the head and lumbar CSF using (Figure and for (Figure by measuring the lumbar pressure before and after mL of CSF and the pressure mL of CSF As we had there was a block of the pressure to the lumbar CSF and the was compromised before to normal after surgery. the upper portion of syrinx compressed by downward movement of tonsils during For the was linked to the while pressures from cervical and lumbar and syrinx. It demonstrated compression of the upper portion of the syrinx during systole (Figure Video, Supplemental Digital Content 2). cine MRI, we assessed CSF movement in the syrinx and in the SAS at the foramen magnum before and after surgery. That demonstrated systolic downward movement of the syrinx but no of CSF flow in the SAS at the level of the foramen magnum. We then used the same to examine the pathophysiology of syringomyelia to surgery in patients with Chiari I and syringomyelia and that the with MRI, cine MRI, and the pressure measurements were to the pathophysiology in Chiari I patients without This that the step of treatment in patients with surgery be surgical decompression of the SAS at the level of the foramen there are that that approach is to as etc. PATHOPHYSIOLOGY OF SYRINGOMYELIA Since the underlying of syringomyelia associated with the Chiari I malformation is obstruction of the SAS at the level of the foramen and primary spinal syringomyelia is associated with block of the but at a lower spinal we that primary spinal syringomyelia is caused by an We studied patients with spinal the spinal SAS using the same of that we had used in the studies. syringomyelia patients with healthy cervical subarachnoid pressure was pressure to lumbar CSF the block was and spinal CSF was showed that with each pressure compressed the surface of the spinal cord the to obstruction of the subarachnoid The are that a spinal subarachnoid block increases spinal subarachnoid pressure the of the spinal a pressure across the segment of the and underlies and to the of studies that examined the pathophysiology of syringomyelia associated the Chiari I This indicated that a pathophysiology is associated with syrinx in patients with Chiari I malformation and primary spinal syringomyelia. OF THE SYRINGOMYELIA ASSOCIATED WITH CHIARI I MALFORMATION AND SYRINGOMYELIA With syringomyelia, the of the cord are with the syrinx, and the and to be the by which and the extracellular space of spinal and the after intrathecal This is with Ball and CSF the cord through these to produce syringomyelia by with the subarachnoid However, the of” our studies that it is caused pressure in the spinal In the normal physiological bulk flow of CSF into the in the spinal cord, brain, and brainstem on the of pulsatile not increased venous The by moves into the of the cervical segments of the spinal cord into the lower segments after intrathecal that bulk flow of CSF into the cord is also related to the of pulsatile pressure The predominant of the cervical segments of the spinal cord Chiari I malformation is linked the that the pulsatile pressure in the spinal subarachnoid space in the upper portion of the cervical canal and are with increasing down the studies demonstrated that to the central canal of the spinal cord of after intrathecal but this not after the pulsations by partially the Thus, and into the spinal cord from without by the forces of and bulk flow by pressure in the CSF. The of our that the exaggerated pressure and pressure in the SAS in patients with syringomyelia associated with a Chiari malformation or primary spinal syringomyelia is an important of the development and progression of syringomyelia. Thus, the of all our studies were with our We in that patients with increased intracranial increased intracranial pressure and CSF from not have syringomyelia there is The development of a syrinx requires increased pressure and pressure in the spinal CSF and at least a partial obstruction of the SAS with a partially isolated spinal CSF This the of syringomyelia in patients with no Chiari but with obstruction of the SAS at the foramen initially by Dr and his colleagues as the Chiari The same mechanism to other and at the foramen magnum and that obstruct the spinal SAS and produce partial of the SAS below the level of as foramen magnum and posterior fossa that produce spinal arachnoid associated with syringomyelia (Table 1). PATHOGENESIS OF CHIARI I MALFORMATION What is the pathogenesis of the Chiari I malformation? studies over the past have demonstrated that the posterior fossa is for the volume it in but not all, patients with a Chiari I malformation. In the initial study in and demonstrated the of the posterior fossa in patients with Chiari I Later, using Dr and his colleagues demonstrated to normal persons, the of the posterior fossa volume to the volume was in but not all, patients with a Chiari I The same is so in other an of the were in the to have a to the of the That also the volume of the posterior fossa and lead many of to have Chiari I malformation and only the patients with Chiari I malformation in the study had a posterior fossa below the normal there also be mechanisms by which of the tonsils in the foramen magnum and a Chiari is produced other a posterior A it became apparent that many patients have much for the posterior fossa producing a very posterior fossa on MRI, have space in the posterior We studied this and showed that patients with Chiari I there are 2 and with in of patients (Figure when the volume is related to the posterior fossa volume in of the volume of the posterior fossa is by in the that to have a posterior at surgery and syringomyelia were in patients with the The of the which of the that the pathogenesis of Chiari I malformation in many Thus, in these patients there be other mechanism, that in of the tonsils in the foramen on MRI between and types of patients with Chiari I malformation. the in the or of SAS by in the posterior fossa of and of Chiari I malformation in these 2 of patient with the and of Chiari I malformation. The to the and prominent CSF space the cerebellar in the and Reprinted with permission from et is a Chiari I malformation is a malformation of the cerebellar tonsils, as has been For instance, the patient shown in has a I of the cerebellar tonsils. it is from the effect of a is not In a prospective clinical study of patients with a Chiari I most of had syringomyelia, we used MRI to examine the of the after After surgery, the abnormal of the tonsils, the Chiari I to normal a of surgery (Figure By 3 to after surgery, the tonsils became the and ectopia by This in essentially all patients with a Chiari I malformation after successful surgery to the SAS to normal pulsatile CSF flow across the foramen Midsagittal T2-weighted MRI and MRI in a patient with of the cerebellar tonsils A caused by a The abnormal anatomy of the cerebellar tonsils is to the anatomy of patients to have a Chiari I Sagittal MRI before and after decompression with Surgical of the obstruction of the SAS at the foramen magnum in of the abnormal of the cerebellar tonsils, of syringomyelia, and of the prominent at the This in essentially all patients after successful the Chiari I malformation is a the cerebellar tonsils to normal after the of the tonsils is The Chiari is a of systolic of the tonsils in the foramen magnum with each which over with a of It is not a as it is not THE INITIAL CHIARI I MALFORMATION WITH SYRINGOMYELIA is the only treatment for Chiari I malformation. The is to symptoms associated with of the cerebellar tonsils swallowing and prevent loss of in patients with at the foramen and to eliminate the mechanism underlying the of optimal from several surgical it has or associated with be in to treatment associated with Thus, surgery is to of but with of the central system, as or of the cerebellar tonsils or subarachnoid of the of the fourth craniocervical decompression and with or without opening the arachnoid, is successful in almost all of a of surgery, might much space is at the level of the foramen magnum for successful of the observations that the space is very For instance, Chiari I in young without treatment as the the brain in early There are also that of an in of the at the level of the foramen magnum associated with is for a Chiari I and syringomyelia after of the at the foramen magnum for successful surgery may be a of a That any that eliminates the obstruction of pulsatile movement of CSF in the SAS is successful so many different are effective, as all of work eliminate that In surgical approaches have been in which decompression a or decompression with a partial is These have been successful in to of all are for but are not in to who require surgery in which the dura is opened (Table The of decompression a or decompression with a partial as the initial of are that of associated with opening the dura and are the patient has a and it these that in patients with syringomyelia, the of patients with most of successful surgery, had a much lower of resulting in a of the syrinx in in to patients who received in in no who received surgery to of the patients who were with 3: of after and and in With Chiari I the of the 2 approaches in demonstrated need for of clinical and a of of the syrinx in patients who received a However, that at the of a of CSF and and a of Thus, the for using posterior fossa decompression the initial are that it the of the and it is and these the of that surgery to the dura be in of patients in this On the other hand, the of the to the subarachnoid space to examine for a subarachnoid or arachnoid arachnoid after opening the dura, the arachnoid is or and that the of with the associated with the of and need surgery It also seems that the of the syrinx and the of the tonsils to a normal and is on several with to posterior fossa decompression and duraplasty there would be a of during surgery the least of surgery that each patient requires to the subarachnoid space at the foramen magnum. That it would be it be during surgery of of with a partial opening of the dura, or and opening the dura, but not the arachnoid, is in so that the surgery at the least that is to be ultrasound has been used for this to examine the distance between the of dura and the posterior of the cerebellar tonsils and to the of the cerebellar tonsils and the spinal cord a syrinx, and to decompression or partial the space has not been MRI, and of CSF and also be for during each step of surgery space between the posterior of the cerebellar tonsils and the of and space at the foramen magnum for normal pulsatile movement of CSF in the subarachnoid space at the level of the foramen by using phase-contrast In a prospective clinical study, we used intraoperative anatomic and cine MRI to the subarachnoid at the level of the foramen The of intraoperative MRI during surgery, when the was for surgery, us into during surgery that had the space based on the space between the posterior of the tonsils and the dura and CSF flow behind the tonsils on cine we the at that in of the 18 of patients all later in which the dura was opened or had no or of their to the that surgery was in several of this study suggested that the of intraoperative for Chiari I malformation in general, by ultrasound or intraoperative MRI, is by the CSF flow across the foramen magnum when the patient is for surgery and that posterior fossa decompression with or without a partial is not a for that the arachnoid or the portion of the cerebellar tonsils, and to the CSF what is with decompression and duraplasty The of these approaches to be to approaches that the dura, with or without opening the arachnoid and of a without subarachnoid In most the goals of surgery are with However, or of the cerebellar tonsils can lead to and subarachnoid and initiate a that is to It is the patient to to surgery, as the mechanism has not been eliminated, of or an that the SAS posterior to the cerebellar tonsils from In these a second is successful by the arachnoid during craniocervical with or without an to intrathecal the arachnoid is opened, through the can also but not the and the are for in which increased intracranial pressure associated with is Surgical failure can also from that unrecognized at surgery, as an arachnoid the SAS or a the of the fourth ventricle, with during which the only of the anatomy is with The of surgical can be using clinical and MRI of the cervical and posterior fossa are performed at 2 to 3 and after surgery to CSF have been at the foramen and for of the abnormal of the cerebellar tonsils, or the of a and to an associated syrinx is surgical treatment in or of disease and in many In a prospective clinical study, we used MRI to monitor patients with Chiari I malformation and syringomyelia (Figure time to syrinx was surgical After surgery of patients had symptoms in before partial or complete of the syrinx on Thus, most patients with syringomyelia after decompression for but have residual symptoms. syrinx of of the spinal that the pathophysiology has been reversed by treatment of the of of the syrinx on On a syrinx that is before surgery A is by after surgery B, but then becomes by 3 and after surgery. Reprinted with permission from et It is that all the observations in our several clinical studies have been with the mechanism during the initial in the I 4 important issues related to the Chiari I malformation and syringomyelia. By what mechanism does a Chiari I malformation produce syringomyelia? The mechanism of the origin and progression syringomyelia is on the outside, not the inside, of the spinal cord. What is the pathogenesis of the Chiari I malformation? Are the mechanisms of development of a Chiari I malformation and syringomyelia similar? The Chiari I and other as posterior fossa that the tonsils in the foramen magnum and produce anatomic are a of of the tonsils in the foramen magnum. It is an not a The Chiari I malformation and the origin and progression of syringomyelia are the same of the cerebellar tonsils in the foramen magnum. How, exactly, is the mechanism related to the goals of surgery? treatment when the obstruction to pulsatile subarachnoid CSF flow at the foramen magnum is treatment with any that The has no or interest in any of the or in this