Clin. Radiol. (1977) 28, 395-400

C E R V I C A L V E R T E B R A E M E A S U R E M E N T S IN S Y R I N G O M Y E L I A

J. LEE and BERNARD WILLIAMS

From the Midland Centre for Neurosurgery and Neurology, Holly Lane, Warley, West Midlands, B67 7JX

Lateral X-rays of the cervical spine taken under standard conditions in 69 cases of syringo- myelia have been measured to assess the relative sizes of the canals and bodies. One hundred and five cases without cervical spine abnormality provided controls. The expected expansion has been observed in males from C5 to C7 and in females from C3 to C7. The oddity that the expansive process seemed to come higher in females than in males has been noted by other workers.

In both sexes it has been found that diminution of vertebral body size is a more consistent finding than enlargement of the canals and correlation coefficients have confirmed that enlargement of the canal appears to take place at least in part at the expense of the body. Possible factors in the enlargement are discussed.

Oblique measurements taken in 57 syrinx cases and 89 controls showed-the expected change caused by the view of the lower cervical vertebrae being closer to AP and pictures of the upper vertebrae being progressively closer to lateral views. There was a tendency for the bigger oblique 'diameters to occur in these cases with the biggest sagittal diameters and an insignificant tendency for the biggest canals to have a diminished difference between the two oblique measurements. There was no evidence that oblique measttrements could be used to detect slight degrees of enlargement in cases of syringomyelia.

INTRODUCTION

Enlargement of the spinal canal in certain intraspinal lesions was noted by Elsberg and Dyke (1934). Boijsen (1954) noted enlargement in two cases of syringomyelia and Wells, Spillane and Bligh (1959) measured 32 cases, concluding that in 15 the canal was enlarged. They found that enlargement of the canal was commonest in cases with a clinical onset before the age of 30 years. These findings have, in general, been substantiated by subsequent workers (Conway, 1961, Bradley and Banna, 1968; McRae and Standen, 1966; Newton, 1969). Hertel in his paper of 1973 produced measurements which corroborated the views of previous workers; he noted greater enlarge- ment of the sagittal diameter of the canal in cases with spasticity of the lower limbs as well as cases with onset before the age of 30. Hertel's paper, however, indicated one result which seemed to be anomalous and at variance with our expectations as well as that of others. He measured the vertebral bodies as well as the canals and found the sagittal diameter of the vertebral bodies to be significantly enlarged in his cases of syringomyelia. This enlargement of the bodies was in males with late onset of disease and spastic signs. He drew this conclusion; 'the enlarge- ment of the cervical canal consequently does not imply a lesion of the vertebral bodies'. We thought that enlargement of the cervical canal took place, at

least in part, at the expense of the vertebral bodies. It was determined therefore to measure the canals in suitable cases and establish correlation coefficients between the diameters of the bodies and canals to test Hertel's view.

The oblique diameters of the cervical spinal canal are commonly increased in syringomyelia. Oblique me asurements ,are unreliable as a repeatable examination because the degree of obliquity of the mid-spine is not determinable by the radiographer. The more nearly it is a lateral view then the smaller is the diameter, conversely views closer to PA show a larger canal (Fig. 1). Errors due to differing degrees of obliquity we thought might be slight in a statistical study. If the increase in size of the spinal canal is dqe to an expansive process, such as an increase in internal pressure, then it might be expected that it would become more circular. Thus the oblique diameter might increase along with the sagittal diameter and asymmetry may lessen, i.e. the dif- ferences between the measurements of the two oblique views become less. The opportunity has been taken to test this by establishing correlation coefficients.

METHODS

Patients. - Adult patients only were selected, cases with congenital abnormalities of the spine and

396 C L I N I C A L R A D I O L O G Y

severe osteoarthritis or Charcot's joints of the spine were excluded as were extrinsic and solid intrinsic tumours. Only radiographs taken under standard conditions were measured. All cases were proven communicating syringomyelia except for two of the males who had non-communicating syringomyelia secondary to intrinsic tumours. Adult patients from 1973 to 1974, excluding spondylosis, were used as controls. Cases suspected of having cervical spon- dylotic myelopathy were excluded because of the tendency for a narrow spinal canal to be found in this condition.

Subjects were divided into male and female but no clinical category or age of onset subdivisions were attempted. Not every case had obliques done.

The numbers were :-

Males Females Lateral Oblique Lateral Oblique

Syringomyelia 34 27 35 30 Controls 61 51 44 38

Radiography. - Lateral views were taken with the patient erect, both flexed and extended. The film was

Fig. 1 - A typical vertebra showing the way in which oblique canal diameters as measured on X-ray become larger as the view becomes more nearly AP. Measurement b is 15% greater than measurement a.

Fig. 2 - Tracing of a lateral radiograph to show the method of measuring the lateral X-rays to give body size and canal diameter.

placed lateral to the shoulder tip with a film-focus distance of 6 ft (1.83 m). The tube was centred just behind and below the angle of the jaw.

Two oblique views were taken with the chest prone on the X-ray table. The head was turned until the skull was parallel with the table top. The chin was tucked well in. The film-focus distance was 40in (1.02m). The tube was angled at 12 ° and centred on the root of the neck.

Measurement. - In the lateral view, the sagittal diameters of the vertebral bodies and canals of C2-7 were measured.

For the bodies, the mid-point of the anterior and posterior margins were used. The canal diameter was taken between the mid-points of the posterior verte- bral body and the cortical line at the junction of the laminae at the back o f the spinal canal of the corresponding vertebra (Fig. 2).

C E R V I C A L V E R T E B R A E M E A S U R E M E N T S IN S Y R I N G O M Y E L I A 397

d

Fig. 3. - Tracing of an oblique radiograph to show the measurements used. Views of both left and right obliques were measured.

For the oblique views only C 3 - 6 proved to have reliable measuring points at the middle of the internal margins of the pedicle and corresponding laminae. Both right and left views were measured and the difference between the two calculated at each level (Fig. 3).

Measurements were taken in millimetres using a micrometer with the dial turned away from the observer and then read to two places of decimals to eliminate significant digit preference. The diagnosis was not revealed to the observer.

Results. - No corrections have been made for enlargement. The calculated magnification is around 1.4 for the lateral and 1.6 for the oblique measure- ments.

The results are given in graphic form, as means with dispersion recorded as standard deviation. The male and female results for canals are given in Fig. 4 and for body diameters in Fig. 5.

The oblique measurements show an almost linear relationship between diameter and level (Fig. 6). Mean results for bo th sexes are given together with standard deviations. A similar graph for the dif- ferences between the two oblique measurements is given in Fig. 7.

Correlation coefficients were determined be tween all measurements for the four groups at each vertebral level. No correlations were assessed between different vertebral levels but even with this l imitation there are 24 correlation coefficients at each level and therefore only correlation between the canal size and the other measurements have been presented in Table 1.

DISCUSSION

Considering the size of the vertebral bodies in Fig. 5 it is clear that the mean diameter of the vertebral bodies is smaller in the syringomyelia cases than in the controls for bo th sexes and at every level. Diminution in size of the bodies is therefore more consistent than enlargement of the canals. In males at C2, C3 and C4 the canals were not larger in the cases with syrinxes and in females C2 and C3 were not enlarged in the syrinx cases. This coincides with Hertel 's calculations using Student 's test (for P< 0.01); in males he found the significant differences between syrinxes and controls were from C5 to C7 and in females from C4 to C7. The fact that there is one vertebra difference in bo th series, the disease seeming

2 2 -

t 20- 0

x '-= x

~ 1 8 - o o o o o

16- C2

Male control 2.13 Male syrinx 1.83 Female control 1.83 Female syrinx 2.89

Male • - - • control x - - x syrinx o - - o control

Female [ ] - - B syrinx

C3 C4 C5 C6 C7

1.94 1.72 1.68 1.64 1.71 1.97 2.33 2.46 2.38 1.83 Standard 1.61 1.76 1.56 2.32 1.61 deviations 2.24 2.33 2.47 2.35 2.47

Fig. 4 - Canal diameters are increased in the syrinx cases, particularly in the lower three vertebrae. The expansive process seems to come higher in the female. The standard deviations are larger for the syrinx patients than for the controls for all vertebrae except male C2.

398 C L I N I C A L R A D I O L O G Y

2 0 -

E E ._c .8 18- •

x

16-

• X

X le x

o o

o [3 o o

C2 C3 C4 C5 C6 C7

Male control 1.32 1,53 1.61 1.67 1.67 1.52 Male syrinx 1,66 1,80 1.86 1,85 1,68 1.55 Standard Female control 1.36 1.41 1.48 1.70 1.79 1.85 deviations Female syrinx 1.17 1.11 1.14 1.34 1.39 1.32

Male 0 - - • control X - - X syrinx

Female o - - o control D - - J3 syrinx

Fig. 5 - At every vertebral level in both sexes the mean sagittal diameter of bodies is less for the syrinx cases than for the controls. This diminution of body size is therefore more consistent than the enlargement of the canals. The standard deviations show an opposite trend being larger in the syrinx cases for the males and larger in the controls for the females.

2 9 -

E

._ 2 7 -

o 25-

2 3 - C3 C4 C5 C6

Male control 1.66 1.79 2.14 2.24 Male syrinx 1.43 1,85 1.98 2.31 Standard Female control 1.66 1.88 1.93 2.46 deviations Female syrinx 2.17 2.28 2.38 2.38

Male • - - • control x - - X syrinx

Female o - - o control £3--E] syrinx

Fig. 6 - The mean oblique diameters increase from C3 to C6 because of the rotation of the spine. C3 is nearly a lateral view and C6 almost an AP. There is no value in this view in the diagnosis of syringomyelia.

2 -

1 - • •

0 - C3 C4 C5 C6

Male control .95 .97 1.00 1.46 Male syrinx 1.40 1.19 1,15 1.25 Standard Female control ,65 .96 .76 ,74 deviations Female syrinx 1.15 .91 .81 1.32

Fig. 7 - The difference between the two oblique measure- ments does not show consistent lessening in the syrinx cases as would be expected if the spinal canal became significantly more circular. The standard deviation indicates a greater scatter for the syrinxes, possibly because of the greater difficulty in rotation for patients with short necks, basilar invagination, kyphoscoliosis and so on.

to come one vertebra higher in the female, is quite striking (Fig. 4) but no ready explanation is forth- coming and the finding may be a coincidence. The greater enlargement o f the lower part of the cervical spine must be significant however. The disease most commonly starts at this level. Williams (1970, 1973) has pointed out that the size o f the cervical enlarge- ment probably increases its vulnerability to the pressure of fluid within the spinal canal. The suscep- t ibil i ty of the cord at this level, it is suggested, is due to its greater circumference and surface area at this point. The forces acting on the circumference of the cord are greatest here because pressures generate forces which are proport ional to the area over which they act. Thus the inflation of an elongated toy balloon is not a uniform progression, one zone gives way first and when it is dilated adjacent sections give way as the enlargement progresses. In the cord the situation is complicated by the presence of relatively soft grey mat ter whose bulk is increased in the cervical enlargement (Williams, 1973). Whatever the exact reasons, the association between this vul- nerabil i ty and the cervical enlargement is clear. It is notable, however, that although the cord is sig- nificantly larger in the lower cervical region, the normal spinal canal is not, as shown by our Fig. 4 and other workers (Wells et aL, 1959; Hertel, 1973; Bloch-Michel, Benoist and Gallovin, 1973). This would seem to suggest that it is factors within the cord that lead to di latat ion of the canal, rather than factors within the subarachnoid space.

Perhaps the most interesting features came from the correlation coefficients. In the syrinx cases, at every level there is a negative correlation between canals and bodies meaning that the bigger the canal is the smaller the bodies tend to be. At eight of the 12 levels P < 0.05 and over the 69 cases of syringo- myelia at 12 levels P is considerably less than 0.01. This calculation takes no account of the more obvious a correlation which is no t present. That is the expectat ion that a big vertebra from a big person would be expected to have both a bigger body and a bigger canal. We interpret this as proving that the enlargement of the cervical canal in syringomyelia takes place, at least in part, at the expense of the size of the vertebral bodies.

The strong negative correlation between the body sizes and the mean canal diameters is also present, although to a lesser extent, in the control cases. So far as we know this has not been noted previously in the literature. It seems to us evidence that some of the factors causing the final size of the spinal canal operate at the expense of the size of the bodies even in the normal. The well known rather wide variations in canal size, represented at one extreme by the

CERVICAL VERTEBRAE MEASUREMENTS IN SYRINGOMYELIA 399

T a b l e l - Corre lat ions o f Canals

Body Mean oblique Difference between obliques

Male controls C2 -0.201 C3 -0.211 0.119 -0.123 C4 --0.094 0.392 -0.140 C5 0.033 0.498 -0.156 C6 -0.003 0.268 -0.038 C7 0.008

Male syrinxes

C2 -0.423 C3 -0.479 0.585 -0.152 C4 0.505 0.714 -0.198 C5 -0.464 0.709 -0.276 C6 -0.564 0.689 -0.286 C7 -0.316

Female controls C2 -0.490 C3 0.226 0.540 0.216 C4 -0.155 0.535 -0.201 C5 --0.145 0.504 -0.265 C6 0.054 0.452 -0.057 C7 -0.315

Female syrinxes C2 -0.267 C3 -0.203 0.805 0.475 C4 -0.318 0.790 0.243 C5 -0.498 0.712 0.328 C6 -0.511 0.683 0.113 C7 -0.421

narrow canal commonly found with cervical spon- dylotic myelopathy, indicate that some variable factor greater than mere growth is at work. If it is suggested that intraspinal high pressure is in some way responsible for the inverse relationship between body size and canal size in syringomyelia then it seems probable that cerebrospinal fluid pathway pressures are effective in enlarging the spinal canal during normal development. Holtzer (1952)believes that the size of the cord itself is important in the precartilaginous stage. Plainly the size o f the cord and the pressure within CSF pathways outside the cord are not the only factors. The function of the spine is also concerned with weight bearing and mobility. Schwartz (1956) measured growth at vertebral levels during infancy, childhood and adolescence and found the maximum growth to be at level S1, a feature which must bear some relation to the locomotor requirements of the spine. The growth of the cervical spinal canal was greater than that in the thoracic region according to Schwartz because it grows faster. The time of cessation of growth is the same for all

parts of the canal, around 18-25 years. It seems unlikely that internal CSF pressures, cord size and locomotor requirements exhaust the factors at work. There is often great enlargement of the canal in the presence of congenital anomalies that produce neither hydrocephalus nor syringomyelia (Walker, 1944) and also in spina bifida. In the latter case it is not un- reasonable to suppose that the dilatation is secondary to hydrocephalus but we would not go so far as Gardner (1973) who believes that the spina bifida is due to raised CSF pressure. Quite unexplained enlargement is sometimes seen. We are not able to agree with Hertel who writes 'a definite diagnosis of syringomyelia can be made by proving the enlarge- ment of the spinal canal at levels C3 to C7'.

The mean oblique measurements presented in Fig. 6 show a virtually straight line linking the points for any group of patients and this confirms what has been said in the introduction that the lower vertebrae are less rotated and therefore more nearly AP and that the higher vertebrae progressively approach the lateral. The mean oblique canal diameters are smaller in our male syringomyelia patients than control and almost identical in the two groups in the females. This may be connected with altered mobility in these cases, factors which could affect both those with and without enlarged canals. The correlation coe f f i c i en t s

however prove the obvious conclusion that the biggest oblique measurements occur in the cases with the biggest sagittal diameter and these correlation coefficients are highly significant and noticeably larger in the syrinx groups.

The difference between left and right oblique diameters in general had a negative correlation co- efficient with the sagittal canal size, the exception being the female syringomyelia patients. Thus the bigger the canal the less the difference, which appears to support our original belief that the canal as it enlarges becomes more nearly circular. There is, however, no evidence that the oblique diameter enlarges to a greater extent than the sagittal and these measurements do not seem to have any significance for any individual case in which a slight degree of cervical canal enlargement might be sought.

Acknowledgements. - We are grateful to Dr C. P. Moxon and the X-ray staff of the Midland Centre for Neurosurgery and Neurology for their help.

REFERENCES

Bloch-Michel, M. Benoist, M. & Gallovin, A. (1973). Mesure du diam~tre sagittal du canal rachidien a l'6tage cervical. Technique radiologique, l~tude statistique effectu6e chez le sujet normal. Revue du Rhumatisme et des Maladies Osteo-Articulares. 40, 553-556.

400 CLINICAL R A D I O L O G Y

Boijsen, E. (1954). The cervical spinal canal in intraspinal expansive processes. ActaRadiologica, Stockholm. 42, 101-115.

Bradley, W, G. & Banna, M. (1968). The cervical dural canal. A study of the 'tight dural canal' and of syringomyelia by prone and supine myelography. British Journal of Radiology. 4 1 , 6 0 8 - 6 1 4 .

Conway, L. W. (1961), Radiographic studies of syringo- myelia, Transactions of the American Neurological Asso- ciation. 86 ,205-206 .

Elsberg, C. A, & Dyke, C. G., (1934). The diagnosis and localization of tumours of the spinal cord by means of measurements made on the x-ray films of the vertebrae, and the correlation of clinical and x-ray findings. Bulletin of the Neurological lnstitute o[ New York. 3,359 394.

Gardner, W. J. (1973) The dysraphic states - from syringo- myelia to anencephaly. Exeerpta Medica. Amsterdam.

Hertel, G. (1973). The width of the cervical spinal cord and the size of the vertebral bodies in syringomyelia. European Neurology. 9, 168-182.

Holtzer, O. (1952). Experimental analysis of development of spinal column. 1. Response of pre-cartilage cells to size

variations of spinal cord. Journal of Experimental Zoology. 121 ,121-147 .

Newton, E. J. (1969). Syringomyelia as a manifestation of defective fourth ventricular. Journal of The Royal College of Surgeons of London. 44~ 194-213.

McRae, D. L. & Standen, J. (1966). Roentgenologic findings in syringomyelia and hydromyelia. American Journal of Roentgenology. 98, 695-703.

Schwartz, G. S. (1956). Width of the spinal canal in the growing vertebra with special reference to the sacrum. American Journal of Roentgenology. 76, 476-481.

Walker, A. E. (1944). Dilatation of the vertebral canal associated with congenital anomalies of the spinal cord. American Journal of Roentgenology. 52 ,571-582 .

Wells, C. E. C., Spillane, J. D. & Bhgh, A. S. (1959). The cervical spinal canal in syringomyelia. Brain. 82, 23-40 .

Williams, B. (1970). Current concepts of syringomyelia. British Journal of Hospital Medicine. 4 ,331-342 .

Williams, B. (1973). Syringomyelia produced by intra- medullary tluid injections in dogs. Journal of Neurology, Neurosurgery and Psychiatry. 36 ,467-477 .