The neurosurgical subset of fibromyalgia: Part II: Radiological observations—patients vs. controls..

Michael J. Rosner, Jorge Flechas, Royce K. Bailey

Introduction: Clinical observations suggest the upper spinal cord and brainstem may be involved in a subset of patients with FM like symptoms. Objective: Test the hypothesis that there will be radiological differences related to the posterior fossa and/or cervical canal between control and the surgical subset of FMS patients. Methods: 37 patients carrying the diagnosis of ‘fibromyalgia’ were selected based upon an abnormal tilt table examination defining cardiovascular/autonomic dysfunction (NMH/POTS) and abnormalities identified from the neurological examination. Protocol magnetic resonance (MR) scans of the craniocervical junction region were accomplished for each patient in sagittal, coronal and transaxial planes; cerebrospinal fluid (CSF) flow via phase contrast MR was assessed.  Specific grading of tonsillar descent (midline and paramedian), obex height, tonsillar impaction within the foramen magnum, vertebral artery invagination of the brainstem, and AP diameter of the foramen magnum was carried out.  The spinal cord and canal diameters were measured at each disk level with electronic calipers.  These results were compared with those of asymptomatic control patients.  Systat 7.0 and Excel were used for data analysis; all values X + SD.

Results:                           
                                         FM Subset    Controls    P Value
Tonsil-Right  (mm)              2.7 + 3.1     0.46+2.4      0.006
Tonsil-Left (mm)                2.6 + 2.8      0.50+2.4      0.008
Tonsil-Midline (mm)           2.1 + 3.2     -1.7+3.7     <0.001
Impaction in fore mag         2.2 + 1.0      1.1+0.96     0.005
VA impaction brainstem     1.2 + 0.6      0.24+0.56   <0.001 
C3 Canal  (mm)                 11.2 + 2.1    12.8+1.9        0.04
C4 Canal  (mm)                 11.0 + 2.7    12.9+1.8        0.03
C5 Canal  (mm)                 10.7 + 2.7    12.2+1.8        0.07

Conclusion:  FM-NMH/POTS patients with abnormal neurological examinations represent a population that is radiographically distinct from asymptomatic individuals. These differences relate to the brainstem and cervical spinal cord as suggested by neurological examinations.  However, the overlap of standard deviations is such that simple review of MR scans will not discriminate between these groups independently of the neurological exam.  INTRODUCTION: We have observed subtle radiographic changes on craniocervical MR scans of patients with neurological findings related to the brainstem and upper cervical spinal cord sufficient to explain clinical symptoms and physical findings.  However, in most cases, these MR scans were reasonably read as “normal” by radiologists, leading the referring physician to seek other than structural etiologies for the patient’s complaints.  This is an important error not to make, and we wish to establish a set of observations which might provide a core of information to help make an accurate radiological diagnosis in this group of patients. HYPOTHESES: The cerebellar tonsils will be lower in FMS patients when compared with controls.  Similar hypotheses will be tested for differences between FMS patients and controls with regard to: Tonsillar impaction into the foramen magnum, vertebral artery impaction into the brainstem, and the AP diameter of the cervical canal at each level.  Hypothesis statement and testing was of the form:

                H1: μ1 ≠ μ0

                H0: μ1 = μ0

METHODS: Protocol MR craniocervical scans were obtained on 37 patients who carried the diagnosis of fibromyalgia syndrome (FMS) and who had also had a tilt table test positive for neurally mediated hypotension (NMH) and/or positional orthostatic tachycardia syndrome (POTS).  Assessments were recorded on a data sheet which included:

Distance above or below the lower limit of the foramen magnum (in mm) for the:

                ____        Midline Tonsil                                                          
                ____        Right paramedian tonsil                          

                ____        Left paramedian tonsil                                            
                ____        Obex

Relation of the upper cervical cord and/or medulla to the vertebral arteries:

Right      ____          0 = “CSF” signal completely separates cord from vertebral artery
                                1 = Vertebral artery abuts cord
                                       without distortion
                                2 = Vertebral artery invaginates
                                       (< 33%) or with mass effect
Left          ____        3 = Vertebral invaginates > 33% but < 67%
                                4 = Vertebral invaginates > 68% of diameter
                                9 = Cannot determine

Relationship of cerebellar tonsils within foramen magnum (transaxial views):

    0 = No tonsil in foramen magnum
    1 = Portion of one or both tonsils in foramen
          magnum 2 = Tonsils mold to cord/medulla < 180°
    3 = Tonsils mold > 180° of cord/medulla
    4 = Only trace CSF signal within foramen magnum
    9 = Cannot determine

Measurements (in mm) of the cervical cord at C2 through C7: These measurements were made perpendicularly to the cord and centered at each disc space to facilitate reproduction of these same measurements on subsequent scans.  These results were compared to similar measurements obtained from 40 asymptomatic control subjects previously recruited for a separate study and supported by the National Fibromyalgia    Research Association.

Other observations included:
Coronal views done in thin section (about 3 mm) through the foramen magnum and perpendicular to the plane of the foramen magnum to better define the relationship of the cerebellar tonsil to the lateral medulla.
Lateral shift and/or distortion of the brainstem within the foramen magnum and posterior fossa.
Absence or near absence of CSF spaces posterior to the cerebellum, compression of inferior cerebellar folia.
Abnormalities of CSF flow though the posterior fossa and foramen magnum.

Results:
                               FM Subset  Controls P Value*  Reject H0

Tonsil Relation to Foramen Magnum

Tonsil-Right (mm)      3.6 ± 3.9   0.86 ± 2.9    0.010    YES

Tonsil-Left (mm)        3.3 ± 3.7   0.50 ± 2.4    0.008    YES

Tonsil-Midline (mm)   2.9 ± 3.9    -1.0 ± 4.2    0.005    YES

Impaction in FM          2.5 ± 1.0     1.2 ± 1.1    0.002    YES

VA impaction brainstem

Right Vertebral        1.1 ± 0.8      0.6 ± 0.9     0.170       NO

Left Vertebral          1.2 ± 0.7      0.3 ± 0.6     0.0002     YES

 

Canal Measurements

C1 Canal  (mm)      15.1 ± 1.7     16.5 ± 1.5    0.045      YES

C2 Canal  (mm)      11.7 ± 1.7     13.0 ± 1.5    0.001      YES

C3 Canal  (mm)      11.1 ± 1.7    13.0 ± 1.5    0.001       YES

C4 Canal  (mm)      10.8 ± 2.2    13.1 ± 1.8    0.002       YES

C5 Canal  (mm)      10.6 ± 2.4    11.9 ± 1.6    0.068          ?

C6 Canal  (mm)      11.5 ± 2.0    11.8 ± 1.2     0.597        NO

C7 Canal  (mm)      13.5 ± 1.7    13.6 ± 1.0    0.768       NO

*  All ‘p’ values ‘two-tailed’, separate variance

Adequate control data did not exist to compare other aspects of the scans between the two groups.  However, the scan of the FMS-NMH/POTS patient typically will show:

Poor to absent CSF circulation posterior to the cerebellum
Poor CSF flow posterior to the cervical cord
A ‘meniscus’ sign below the cerebellar tonsils
Distortion and subtle shift of the brainstem on transaxial views of the foramen magnum
Compression of the cerebellar folia in the inferior cerebellum
Absence of the subarachnoid space posterior to the cerebellar hemispheres
Others

DISCUSSION:  The cerebellar tonsils in FMS-NMH/POTS patients descend further through the foramen magnum than do those of control subjects.  The population difference is about 3 mm lower for the affected patients than in controls.  The absolute descent is greatest for the right cerebellar tonsil, though the variability is great. The tonsils typically both descend into the foramen magnum and mold to some degree about the brainstem while the typical control patient shows only minimal presence of the tonsils in the foramen magnum. This will reduce the relative CSF space within the foramen magnum and increase the pressure gradient needed to move CSF through the foramen during systole and diastole. The mid and upper cervical segments fail to enlarge to a normal degree and    remain significantly narrower than in control        patients. The  canal from C5 rostrally is about 1.5-2.0 mm smaller than control subjects.  The narrower canal and tighter foramen magnum will increase the liability of the patient to hyperextension injuries. These are particularly prominent in “whip-lash” like events, during some surgeries, falls, and some forms of forceful forward bending.  The subtle  nature of  these differences, though

though real, mean that accurate diagnosis must be based upon a very thorough and critical neurological exam anatomically related to the upper spinal cord and/or brainstem (part I).  Merely using the MR scan, or worse, its radiological interpretation, as a screen for the presence of compression will usually be misleading. 

CONCLUSIONS:
Patients with the diagnosis of fibromyalgia and NMH/POTS represent a population radiographically separate from the asymptomatic control population.
These distinctions relate to the posterior fossa/brainstem and/or the cervical spine and cord:
The cerebellar tonsils are lower
The foramen magnum is more crowded
The upper cervical segments (C1-C5) are narrower
These relationships correspond anatomically to the neurological examination findings presented in part I.
The findings are in the direction of developmental cervical stenosis and/or the hypoplastic posterior fossa, either of which can be associated with compression of the brainstem and or cervical spinal cord and which are amenable to surgical decompression when symptoms are severe enough.
However, the overlap of means and the standard deviations of the two populations is such that it is not possible:

To identify surgical candidates by screening MR scans of the neck and posterior fossa independently of the neurological history and examination.
The radiologist and others will read most scans as being ”within normal limits.”