Title: Variation in the position of the Conus Medullaris and Dural Sac in adult dogs

Authors: Zohra Khan1, Elizabeth Munro1, Darren Shaw1, Kiterie M E Faller1*

Authors’ affiliation: 1The Royal (Dick) School of Veterinary Studies. The University of

Edinburgh, Easter Bush Campus, Midlothian, EH25 9RG.

Qualifications:

Zohra Khan BVMS MRCVS

Elizabeth Munro VetMB MA DVR MRCVS

Darren Shaw BSc PhD

Kiterie Faller DVM DPhil DipECVN MRCVS

* Corresponding author: Kiterie M E Faller. Kiterie.faller@ed.ac.uk

Word Count: 3174 (excluding title page, abstract, references, figures and tables.)

1

1

2

3

4

5

6

7

8

9

10

11

12

13

14

15

16

ABSTRACT:

Although it has long been stated that the level of spinal cord termination varies depending on the

size of the dog, the evidence for this remains limited. The aim of this study is to investigate the

position of the conus medullaris and dural sac in a population of dogs of varying size.

Magnetic resonance images of the thoraco-lumbo-sacral spine of one hundred and one dogs were

included. The location of conus medullaris and dural sac was determined on sagittal T2-weighted

magnetic resonance images and T1-weighted images respectively by three independent

observers. The body weight and the back length were used as markers of size.

Regression analysis showed that the termination point of the conus medullaris had a statistically

significant relationship with bodyweight (R2=0.23, p<0.05). Although not statistically significant

(p=0.058), a similar relationship was found between conus medullaris and back length (R2=0.21).

No statistically significant relationship was found between the termination point of the dural sac

and bodyweight (p=0.24) or back length (p=0.19).

The study confirms the terminal position of the conus medullaris is dependent on size, with a

more cranial position with increasing size; however, the termination point of dural sac remains

constant irrespective of dog size.

Keywords:

Conus medullaris, dural sac, dog, MRI, spinal cord anatomy, tethered cord syndrome

2

17

18

19

20

21

22

23

24

25

26

27

28

29

30

31

32

33

34

35

36

37

INTRODUCTION:

In mammals, the spinal cord tapers to a point caudal to the lumbar intumescence known as the

conus medullaris (CM). At the CM, the central canal dilates to form the ventriculus terminalis

(terminal ventricle), the parenchyma narrows to form a layer of columnar ciliated ependymal

cells lining the terminal ventricle and a very thin glia layer and thicker pia layer; this is

collectively known as the filum terminale (terminal filament). The dura mater and arachnoid

mater extend beyond the CM forming the lumbar cistern containing cerebrospinal fluid (CSF).

Caudal to the lumbar cistern the dural sac (DS) constricts around the filum terminale forming the

spinal dura mater filament, which anchors at one or two sites on a sacral or coccygeal vertebrae.

(1-3)

During early embryological development, the spinal cord extends along the full length of the

vertebral column. The spinal nerves pass out of the corresponding intervertebral foramina, which

at this stage are directly opposite their point of origin. As the embryo and later the neonate

develop, the vertebral column continues to extend as spinal cord growth slows; this continues to

maturity meaning in the adult animal the spinal cord does not occupy the whole length of

vertebral canal.(1)

 

It is known that the position of the conus medullaris relative to the vertebrae varies between

species.(4) It has long been thought that in larger dogs the cord terminates more cranially when

compared to smaller breed dogs; however this is only based on a single cadaveric study from

1966 involving thirty-eight dogs, most of them being medium sized.(5)

3

38

39

40

41

42

43

44

45

46

47

48

49

50

51

52

53

54

55

56

57

58

59

60

Knowing the position of termination for the CM and DS has clinical applications in terms of site

of CSF sampling or epidural local anaesthesia. Additionally it would help with the identification

of pathological variations in the position of the CM such as tethered cord syndrome: a congenital

disorder described in both humans and dogs characterised by an abnormal caudal traction of the

CM.(6-11)

The aim of the study is to determine the position of the CM and DS termination in a population

of dogs of various sizes and evaluate whether these termination points depend on the size of the

dog. Our hypotheses were that the CM position depends on the size of the dog whilst the end

point of the DS is a species constant.

MATERIALS AND METHODS:

Study design:

This is a retrospective, single-centre study performed at the Hospital for Small Animals of the

University of Edinburgh. All patients were client owned dogs and were scanned as part of

diagnostic investigations. Ethical approval was obtained from the ethics board of the University

(VERC 116.17). The clinical database was searched for dogs which required magnetic resonance

imaging (MRI) of the vertebral column between January 2017 and May 2018. The MR images

were acquired with the patients positioned in dorsal recumbency with the pelvic limbs in a

relaxed “frog-legged” position, achieving a neutral view of the lumbosacral vertebral column.

The images were acquired using a 1.5 Tesla field strength MRI scanner (Philips Achieva™,

Philips Medical Systems, Reigate, UK).

4

61

62

63

64

65

66

67

68

69

70

71

72

73

74

75

76

77

78

79

80

81

82

83

Inclusion and exclusion criteria:

We searched for dogs, for which a sagittal T2-weighted and where available a sagittal T1-

weighted MR scan had been acquired of the thoraco-lumbo-sacral vertebral column including at

least the second thoracic vertebrae and part of the sacrum. Only dogs aged greater than 12

months and for which a body weight was available on the patient records were included. Cases

were excluded if a mid-sagittal image was not available, there was a deviation from the normal

canine vertebral formula or presence of vertebral malformations or finally, if a caudal lumbar or

lumbosacral myelopathy was found.

 

Size of dog was determined by two methods: the bodyweight, and the distance between the

second thoracic vertebrae and the seventh lumbar vertebrae was used as a surrogate of the dog

size, independent of body condition. This distance was measured on a single mid-sagittal image

if available or sequential images in larger patients. This length was considered equivalent to the

distance between the most dorsal points of the scapula and the wing of the ilium in the standing

patient,(12) which can easily be measured on the live animal, and was referred to as back length.

 

The termination of the CM was determined from a mid-sagittal T2-weighted image in which the

hyperintense CSF highlights the comparatively hypointense spinal cord. The termination of the

DS was determined from a mid-sagittal T1-weighted image in which the hypointense CSF within

the dural sac is delineated by the comparatively hyperintense epidural fat. The measured level of

termination for both structures was then attributed to the corresponding half of the vertebral body

by drawing a perpendicular line to the long axis of the spinal cord or dural sac (Figure 1). This

was performed by three observers (ZK, KF and EM), a neurology resident, a board-certified

5

84

85

86

87

88

89

90

91

92

93

94

95

96

97

98

99

100

101

102

103

104

105

106

neurologist and a board-certified radiologist; each observer was blinded to the patient size and

the other observers’ results. Individual termination points for each patient were based on two or

three observers agreeing or where there was no agreement, the mean of the three observers’

results. Finally, to assess intra-observer reproducibility, one observer (ZK) graded all images on

a second occasion three months later. Before statistical analyses, the sequential halves of the

vertebral bodies were assigned a numerical value from one (cranial half of fifth lumbar vertebra)

to eight (caudal half of the sacrum).

Statistical analysis:

All statistical analyses were performed using R Statistical Software (R Foundation for Statistical

Computing, Vienna, Austria). The termination points of the CM and DS were assigned a

numerical value between one and eight according to the corresponding vertebral bodies; this

conversion assumed a linear relationship for vertebral body length. To determine the effect of

bodyweight and back length on the termination points of the CM, Poisson regressions were

performed with the residuals being normally distributed. For the termination points of the DS, no

such normality of residuals was observed, and instead ordinal logistic regressions were

performed. Percentage agreement and Fleiss’s Kappa were used to assess the inter-observer

agreement between the three observers for the termination points of the CM and the DS,(13) and

the intra-observer agreement was evaluated using percentage agreement and Cohen’s Kappa. The

guidelines recommended by Landis and Koch were used to interpret the Kappa coefficients.(14)

RESULTS:

6

107

108

109

110

111

112

113

114

115

116

117

118

119

120

121

122

123

124

125

126

127

128

One hundred and one dogs met the inclusion criteria. One hundred dogs had T2- weighted

images of the thoraco-lumbo-sacral spine. One dog was included in the DS analysis but excluded

from CM analysis as all three observers independently agreed that the quality of the T2-weighted

images was too poor to reliably assess the position of the CM. Twenty-four dogs were excluded

for the DS position evaluation (for twenty-three patients, T1-weighted images had not been

acquired, and for one patient, the full length of the sacrum had not been included), leaving 77

dogs for DS analysis. The median bodyweight of the patients was 13.0 kilograms (range and

interquartile range 4.1-67.5; 6.9-28.2 kg) (Figure 2).

The inter-observer percentage agreement for two or more observers was 93.0% and 96.1% for

CM and DS termination points respectively. The intra-observer percentage agreement was 88.0%

and 90.0% for CM and DS termination points respectively. When evaluating the inter-observer

agreement, Fleiss’s Kappa gave a value of 0.54 (confidence interval (CI): 0.47-0.60) for the CM

and 0.37 (CI: 0.28-0.46) for the DS, leading to moderate and fair level of inter-observer

agreement respectively. When evaluating the intra-observer agreement, Cohen’s Kappa gave a

value of 0.83 (CI: 0.74-0.92) for the CM and 0.79 (CI: 0.66-0.93) for the DS, leading to almost

perfect and strong level of intra-observer agreement respectively (Table 1).(14)

Table 1: Inter- and Intra-observer variability in assessing the termination points for both the

conus medullaris and dural sac.

Conus medullaristermination point

Dural sactermination point

Number(n = 100)

Percentage (%)

Number (n=77)

Percentage (%)

Inter- Number of 3 54 54.0 40 51.9

7

129

130

131

132

133

134

135

136

137

138

139

140

141

142

143

144

145

146

147

148

observer variability

observers in agreement

2 39 39.0 34 44.2

0 7 7.0 3 3.9Fleiss’s Kappa (95% Confidence interval) 0.54 (0.47-0.60) 0.37 (0.28-0.46)

Intra-observer

variability

Number of observations in

agreement

2 88 88.0 69 90.0

0 12 12.0 8 10.0Cohen’s Kappa (95% Confidence interval) 0.83 (0.74-0.92) 0.79 (0.66-0.93)

We evaluated the relationship between the size of the dog and the position of the CM and DS

respectively. The size was determined by two methods: the body weight and the back length.

Using a Poisson regression, there was a statistically significant relationship between the

termination point of the CM and bodyweight (McFadden pseudo-R2 = 0.23, p<0.05) (Figure 3

A). Although not statistically significant (p=0.058), a similar relationship was found between

CM and back length (McFadden pseudo-R2 = 0.21) (Figure 3 C). Using an ordinal logistic

regression, no statistically significant relationship was found for the termination point of the DS

and bodyweight (p=0.24) or back length (p=0.19) (Figure 3 B and D).

To further highlight the difference in position of the conus medullaris depending on body size

and to give the reader the expected/most frequent position of CM and DS for a dog of a specific

size, we split dogs into bodyweight categories of less than 5 kilograms (kg), between 5 and 10

kg, between 10 and 25 kg, between 25 and 40 kg and greater than 40 kg; these categories were

chosen according to Hawthorne et al.(15) In the less than 5 kg group the most commonly

observed termination points for the CM were the cranial and caudal halves of the seventh lumbar

vertebra found in 33.3% and 44.4% of dogs respectively. In the 5 to 10 kg group, the most

commonly observed CM termination point was the cranial half of the seventh lumbar vertebrae

8

149

150

151

152

153

154

155

156

157

158

159

160

161

162

163

164

165

166

found in 40.0% of dogs. In the 10 to 25 kg group, the most commonly observed CM termination

points were the caudal half of the sixth and cranial half of the seventh lumbar vertebrae found in

36.7% and 40.0% of dogs respectively. In the 25 to 40 kg group, the most commonly observed

CM termination point was the caudal half of the sixth lumbar vertebrae found in 62.5% of dogs.

Finally, in the over 40 kg group the most commonly observed CM termination point was the

caudal half of the sixth lumbar vertebra, which was found in 85.7% of dogs. Using the regression

equation (Figure 3 A), predicted terminations points for the CM were calculated using the mid-

point for each of the above-mentioned weight categories. The predicted termination points of the

CM aligned accurately with the most commonly observed termination points for this population

of dogs (Table 2). Furthermore, the most commonly observed DS termination points was the

cranial half of the sacrum irrespective of body weight, further confirming the absence of

relationship between body weight and the DS termination point.

Table 2: Most commonly observed conus medullaris and dural sac termination points in our

population. Predicted termination points of CM were calculated using the regression equation.

The number 4-7 in italic refers to our grading system where the sequential halves of vertebral

bodies have been assigned to a number from the caudal half of the sixth lumbar vertebrae to the

cranial half of the sacrum: e.g. 4 represents the caudal half of L6, 5 the cranial half of L7, 6 the

caudal half of L7 and 7 the cranial half of the sacrum).

Conus medullaris Dural sac

9

167

168

169

170

171

172

173

174

175

176

177

178

179

180

181

182

183

184

185

186

Body- weight

(kg)

Most commonly observed

termination point (value

from our grading system)

Percentage % (n)

Predicted termination

point

Body-weight used

for predicted

termination point (kg)

Most commonly observed

termination point (value

from our grading system)

Percentage % (n)

<5Cranial and caudal L7

(5/6)77.7 (7) 5.1 2.5 Cranial

Sacrum (7) 50.0 (4)

5-10 Cranial L7 (5) 40.0 (12) 4.9 7.5 Cranial

Sacrum (7) 68.0 (17)

10-25Caudal L6 and cranial L7 (4/5)

76.7 (23) 4.6 17.5 Cranial Sacrum (7) 81.0 (17)

25-40 Caudal L6 (4) 62.5 (15) 4.1 32.5 Cranial

Sacrum (7) 52.9 (9)

>40 Caudal L6 (4) 85.7 (6) 3.7 50 Cranial

Sacrum (7) 83.8 (5)

DISCUSSION:

In this study, we have shown that the CM position varies between dogs and a relationship exists

between CM position and bodyweight. However, the position of the DS remains constant

independent of dog’s size.

In both humans and canines during embryological development, the spinal cord initially occupies

the full length of the vertebral column; ultrasonographical studies have shown in foetuses of both

species the termination of the conus medullaris migrates cranially in the vertebral column

through gestation.(16, 17) During postnatal development, the vertebral column continues to

elongate and the caudal lumbar, sacral and coccygeal spinal segments progressively become

10

187

188

189

190

191

192

193

194

195

196

197

198

more cranial to their respective vertebrae.(3) To avoid dogs for which further CM migration

could occur, dogs less than twelve months old were excluded from the study population.

However, for some large and giant breed dogs, adult bodyweight and skeletal maturity is

achieved beyond 12 months of age. One study showed that toy, small, and medium breed dogs

reached 99% of their adult weight at approximately 9 or 10 months old, whereas large and giant

breed dogs reached this point at approximately 11 to 15 months old.(15) In our study population

there were only four dogs, which were between 12 and 24 months of age, these were all small

breed dogs weighing between 4.1 and 9.8 kg, consequently the age at MR imaging, and potential

for further growth are unlikely to have affected the results.

In larger dogs, growth and the vertebral column elongation occurs for a longer period than in

smaller dogs; therefore, it is not surprising that the CM ends up being located more cranially in

larger dogs. Conversely to the relatively mobile CM, the spinal dura mater filament attaches at

one or two points to a sacral or coccygeal vertebrae,(2) this anchoring of the DS may provide an

explanation to why the position of the CM varies with bodyweight and the DS termination does

not.

To assess non-invasively the position of the CM and DS on a large population of dogs, we used

MR images. MRI has been extensively used in people to study variation of spinal cord length

with gender, age and race. A human literature review revealed most of these studies use sagittal

T2-weighted and/or T1-weighted images to identify the CM, whilst sagittal T1-weighted images

are most often used to assess DS position. Occasionally, dorsal or transverse views or MR

myelography were analysed.(18-25) Based on this information, we opted to use T2-weighted

11

199

200

201

202

203

204

205

206

207

208

209

210

211

212

213

214

215

216

217

218

219

220

221

images to determine the position of CM as the surrounding CSF is markedly hyperintense

compared to the spinal cord parenchyma, allowing easier identification of the CM. A T1-

weighted image was used for identification of the DS as it gives good contrast between the

hypointense CSF and the hyperintense epidural fat.

Before analysing the relationship between dog size and spinal cord terminations, we had to

evaluate the reproducibility of our grading system, both in terms of intra and inter-observer

agreement. The percentage of agreement are high, however some kappa (κ) values only show fair

agreement. Both percentage agreement and κ coefficients can be used to assess both inter-

observer and intra-observer variability. κ coefficients take into account the agreement may have

occurred by chance, so often the values are less than those achieved by measuring percentage

agreement. In addition, weighted κ index can be used for ordinal or ordered categorical variables

and lead to a stronger agreement in cases where observers choose adjacent categories.(13)

However, a weighted κ indices is not available for more than two observers, so could not be used

to assess inter-observer variability in this study. This may explain why only a fair agreement

between the three observers is obtained for the position of the dural sac despite a high inter-

observer percentage agreement and the fact that even if all observers disagreed, it was never by

more than two grades. It has been suggested that both percentage agreement and κ coefficients

should be used to assess inter-observer agreement in studies where the observers are well trained

and unlikely to guess, whereas when there is higher likelihood of guessing, κ coefficients are

likely to be a more accurate assessment of true agreement.(26)

12

222

223

224

225

226

227

228

229

230

231

232

233

234

235

236

237

238

239

240

241

242

243

244

Being able to more accurately estimate the position of the termination of the spinal cord and its

relationship to the termination of the DS is of great importance when sampling CSF from the

lumbar cisterna and also when considering anaesthesia or analgesia in this region. CSF collection

is most often performed from the lumbar subarachnoid space at the junction between the fifth

and sixth or sixth and seventh lumbar vertebrae; however, the junction between the fourth and

fifth lumbar vertebrae or the lumbosacral junction can also be used.(27, 28) Inaccuracy could

lead to damage of the terminal segment of the spinal cord or failure to obtain a sample. Epidural

injections are commonly performed at the lumbosacral junction as it provides the largest access

to the epidural space. Inaccurate needle placement can lead to subarachnoid or intrathecal

injections, which can be hazardous as the dose of anaesthetic or analgesic agents are much lower

when used intrathecally compared to epidural placement.(29)

The results show there is a statistically significant variation in termination point of the conus

medullaris between small and large breed dogs, however there is variability between individuals

of similar weights. This makes ascertaining the termination point of the spinal cord for a specific

dog difficult. The authors would still always recommend attempting lumbar CSF sampling as

caudal as possible even in very large breed dogs and only if the tap is “dry” should you try more

cranially, therefore as much as possible avoiding iatrogenic trauma to the spinal cord.

A further clinical application is during the diagnosis of tethered cord syndrome, a congenital

disorder described in humans, which is characterised by an abnormal caudal traction of the conus

medullaris. In humans, it can be found in association with ventral positioning of the conus

medullaris within the vertebral canal, lipid infiltration of the filum terminale,

13

245

246

247

248

249

250

251

252

253

254

255

256

257

258

259

260

261

262

263

264

265

266

267

lipomyelomeningocoele, myelomeningocoele, myelocystocoele, meningocoele, split cord

malformations, dermal sinus, anorectal malformations and intramedullary tumours.(10) Tethered

cord syndrome has been reported in dogs in association with congenital vertebral and spinal cord

malformations such as spina bifida, myeloschisis, meningocoele and myelomeningocoele.(6, 7,

9, 11) More recently tethered cord syndrome has been identified as a sole finding in a canine

patient causing progressive pelvic limb lameness.(8) Currently, diagnosis in dogs remains

challenging considering the lack of data of normal anatomical position of the conus medullaris,

therefore being able to estimate the expected position of the CM would further aid diagnosis of

this condition.

This study has the limitations of a retrospective study including the distribution of the body sizes

in the study population and using images not specifically acquired for their purpose.

The median bodyweight of the patients was 13.0 kg, with a range 4.1 to 67.5 kg. The majority of

the study population were less than 25 kg of bodyweight, with 69% of dogs less than 25 kg of

bodyweight, 24% between 25 and 40 kg and 7% greater than 40 kg. This distribution of

bodyweight suggests the conclusions drawn from this study are more applicable to toy, small,

medium and large breed dogs, than for giant breed dogs.

During investigations of thoracolumbar myelopathy, commonly obtained MR images include the

thoracic, lumbar and sacral segments and T1-weighted and T2-weighted sequences. In two

human studies assessing the relationship between CM, body mass and age, the use of tranverse

and dorsal sequences helped better identify the termination points of the caudal spinal cord.(19,

14

268

269

270

271

272

273

274

275

276

277

278

279

280

281

282

283

284

285

286

287

288

289

290

25) However, in our hospital, transverse images of the caudal spinal cord do not tend to be

acquired in absence of pathology in that area. A study in human subjects assessing the impact of

age and gender on the CM and DS terminations also used an MR myelography sequence to

identify the endpoint of the dural sac. This sequence accurately differentiates the epidural fat

from the cerebrospinal fluid;(24) however, currently we limit the use of this specific technique

for investigation for spinal arachnoid diverticula. A true mid-sagittal image is required for

accurate identification of the terminations points of both the CM and DS. However, when

reviewing the retrospective images, the area of interest, which prompted the MRI acquisition, is

often not the lumbosacral region so positioning for this region may have not be optimal. A

prospective study including equal proportions of dogs of various sizes, with fine slice sagittal,

tranverse T1-weighted and T2-weighted MR sequences and an MR myelography sequence

between the fifth lumbar and first coccygeal vertebrae with optimal positioning of this region

would allow more accurate identification of the conus medullaris and termination of the dural

sac.

In conclusion, the results of the study confirm the widespread belief that the position of the

conus medullaris is dependent on bodyweight, with a more cranial position with increasing

bodyweight. However, the termination of dural sac remains similar despite increasing

bodyweight.

Competing interests:

None declared.

15

291

292

293

294

295

296

297

298

299

300

301

302

303

304

305

306

307

308

309

310

311

312

313

16

314

315

316

317

REFERENCES:

1. Uemura EE. Spinal Cord. In: Fundamentals of canine neuroanatomy and

neurophysiology. Ames, Iowa: Wiley Blackwell; 2015:99-119.

2. Marin-Garcia P, Gonzalez-Soriano J, Martinez-Sainz P et al. Spinal cord central canal of

the German shepherd dog: morphological, histological, and ultrastructural considerations.

J Morphol. 1995;224:205-12.

3. Fletcher TF. Spinal Cord and Meninges. In: Evans HE, Miller ME, De Lahunta A,

editors. Miller's anatomy of the dog. Fourth ed. St. Louis, Missouri: Elsevier Saunders;

2013:589-610.

4. De Lahunta A, Habel RE. Vertebral Column and Spinal Cord. In: Applied veterinary

anatomy. Philadelphia: Saunders; 1986:202-21.

5. Fletcher TF, Kitchell RL. Anatomical studies on the spinal cord segments of the dog. Am

J Vet Res. 1966;27:1759-67.

6. Acevedo Naranjo CM, Martínez MdP, Ruíz Sierra IC et al. Tethered spinal cord

syndrome in an english bulldog puppy. A case report. Revista Colombiana de Ciencias

Pecuarias. 2008;21:87-96.

7. Cloquell A, Mateo I, Munoz A. Spina bifida, myelomeningocele and tethered cord as the

cause of incontinence and paraphimosis in a dog. Diagnosis by Computed Tomography

and myelography. Clinica Veterinaria de Pequenos Animales. 2012;32:81-6.

8. De Decker S, Gregori T, Kenny P et al. Tethered cord syndrome associated with a

thickened filum terminale in a dog. J Vet Intern Med. 2015;29:405-9.

17

318

319

320

321

322

323

324

325

326

327

328

329

330

331

332

333

334

335

336

337

338

9. Fingeroth J, Johnson G, Burt J et al. Neuroradiographic diagnosis and surgical repair of

tethered cord syndrome in an English bulldog with spina bifida and myeloschisis. J Am

Vet Med Assoc. 1989;194:1300-2.

10. Hertzler DA, DePowell JJ, Stevenson CB et al. Tethered cord syndrome: a review of the

literature from embryology to adult presentation. Neurosurgical focus. 2010;29:E1.

11. Shamir M, Johnston D, Rochkind S. Surgical treatment of tethered spinal cord syndrome

in a dog with myelomeningocele. Vet Rec. 2001;148:755–6.

12. Evans HE, De Lahunta A, Miller ME. Skeleton. In: Miller's anatomy of the dog. Fourth

ed. St. Louis, Missouri: Elsevier Saunders; 2013:589-610.

13. Hallgren KA. Computing inter-rater reliability for observational data: an overview and

tutorial. Tutorials in quantitative methods for psychology. 2012;8:23-34.

14. Landis JR, Koch GG. The measurement of observer agreement for categorical data.

Biometrics. 1977;33:159-74.

15. Hawthorne AJ, Booles D, Nugent PA et al. Body-weight changes during growth in

puppies of different breeds. J Nutr. 2004;134:2027S-30S.

16. Zalel Y, Lehavi O, Aizenstein O et al. Development of the fetal spinal cord: time of

ascendance of the normal conus medullaris as detected by sonography. J Ultrasound

Med. 2006;25:1397-401.

17. Amer MS, Hassan EA, Torad FA et al. Sequential Canine Neonatal Spinal

Ultrasonography from Birth till Spinal Ossification. Pakistan Veterinary Journal.

2015;36:6-10.

18. Saifuddin A, Burnett SJ, White J. The variation of position of the conus medullaris in an

adult population: a magnetic resonance imaging study. Spine. 1998;23:1452-6.

18

339

340

341

342

343

344

345

346

347

348

349

350

351

352

353

354

355

356

357

358

359

360

361

19. Binokay F, Seydaoğlu G, Erman T et al. Relationship between the levels of normal conus

medullaris and body mass index in the Turkish adult population. Neurosurgery

Quarterly. 2013;23:81-4.

20. Demiryürek D, Aydingöz Ü, Akşit MD et al. MR imaging determination of the normal

level of conus medullaris. Clin imaging. 2002;26:375-7.

21. Kim J-T, Bahk J-H, Sung J. Influence of age and sex on the position of the conus

medullaris and Tuffier's line in adults. Anesthesiology. 2003;99:1359-63.

22. Malas M, Seker M, Salbacak A et al. The relationship between the lumbosacral

enlargement and the conus medullaris during the period of fetal development and

adulthood. Surgical and Radiologic Anatomy. 2000;22:163-8.

23. Macdonald A, Chatrath P, Spector T et al. Level of termination of the spinal cord and the

dural sac: a magnetic resonance study. Clin Anat. 1999;12:149-52.

24. Soleiman J, Demaerel P, Rocher S et al. Magnetic resonance imaging study of the level

of termination of the conus medullaris and the thecal sac: influence of age and gender.

Spine. 2005;30:1875-80.

25. Wilson DA, Prince JR. MR imaging determination of the location of the normal conus

medullaris throughout childhood. Am J Roentgenol. 1989;152:1029-32.

26. McHugh ML. Interrater reliability: the kappa statistic. Biochemia medica. 2012;22:276-

82.

27. Chrisman CL. Cerebrospinal Fluid Analysis. Vet Clin North Am Small Anim Pract.

1992;22:781-810.

28. Lorenz MD, Coates JR, Kent M. Confirming a Diagnosis. In: Handbook of veterinary

neurology. 5th ed. St Louis, Missouri: Elsevier Saunders; 201:75-92.

19

362

363

364

365

366

367

368

369

370

371

372

373

374

375

376

377

378

379

380

381

382

383

384

29. Valverde A. Epidural analgesia and anesthesia in dogs and cats. Vet Clin North Am Small

Anim Pract. 2008;38:1205-30.

20

385

386

387

388

FIGURE LEGENDS

Figure 1: Sagittal (A) T2-weighted and (B) T1-weighted magnetic resonance images of the

lumbo-sacral vertebral column. The termination points of (A) the conus medullaris and (B) the

dural sac were attributed to the corresponding half of the vertebral body by drawing a line

perpendicular to the long axis of the spinal cord. In this example the CM ends in the cranial

portion of L7 and the DS in the cranial part of the sacrum; all observers independently agreed to

this.

Figure 2: Histogram representing the distribution of the bodyweight within the study population.

Figure 3: Relationship between (A) the termination of the conus medullaris and bodyweight, (a

statistically significant relationship was found using Poisson regression,

y=exp(1.65−0.007*BW)); (B) the termination of the dural sac and bodyweight (no relationship

was found using an ordinal logistic regression); (C) the termination of the conus medullaris and

back length, (a statistically significant relationship was found using Poisson regression,

y=exp(1.87−0.01*(T2−L7 length))) and (D) the termination of the dural sac and back length (no

relationship was found using an ordinal logistic regression).

21

389

390

391

392

393

394

395

396

397

398

399

400

401

402

403

404

405

406