British Journal of Anaesthesia 105 (6): 857–62 (2010)
Advance Access publication 15 September 2010 . doi:10.1093/bja/aeq246
Is spinal anaesthesia at L2 – L3 interspace safe in disorders of
the vertebral column? A magnetic resonance imaging study
N. Lin 1, J. F. Bebawy 2, L. Hua 3 and B. G. Wang 4*
1
Department of Anaesthesiology, Beijing Tiantan Hospital, Capital Medical University, Beijing 100050, P.R. China
Northwestern University Feinberg School of Medicine, 251 E. Huron St., Suite F5-704, Chicago, IL 60611, USA
3
Biomedical Engineering Institute of Capital Medical University, Beijing 100069, China
4
Department of Anaesthesiology, Beijing Sanbo Brain Hospital, Capital Medical University, Beijing 100093, China
2
* Corresponding author. E-mail: [email protected]
Key points
† The authors have
described the impact of
spinal disorders on the
positioning of conus
medullaris terminus
(CMT).
† Magnetic resonance
image scans of 1047
Chinese patients were
reviewed retrospectively.
† A lower than normal
location of CMT was seen
in females, and in
patients with thoracic
vertebral compression.
† The risk of spinal cord
injury during spinal
anaesthesia at L2 –L3
interspace is increased in
female patients with
thoracic vertebral
compression.
Background. The varying point at which the spinal cord terminates in the lumbar spinal
canal may affect the incidence of spinal cord injuries associated with needle insertion for
spinal anaesthesia, especially in patients with vertebral body or intervertebral disc
disease. This is a complication which has been frequently reported when spinal needle
insertion was performed at higher lumbar spinal levels.
Methods. We retrospectively reviewed magnetic resonance images of the spine in 1047
Chinese patients to determine the conus medullaris terminus (CMT) in patients with and
without vertebral disorders. Patients with tumours in and around the spine and those
with congenital spinal anomalies were excluded from the study. Patients with mixed
vertebral disorders were also excluded.
Results. Our data demonstrate that patients with thoracic vertebral compression fractures
had lower ending points of the CMT than those without (P,0.05), while patients with
lumbar compression fractures did not demonstrate such a correlation. With regard to this
difference, females were significantly at higher risk for a lower CMT than males.
Conversely, lumbar disc disorders such as intervertebral disc extrusion, herniation, or
bulging did not have any significant influence on the level of CMT. Moreover, patients
with spondylolisthesis or scoliosis did not demonstrate an abnormal CMT location.
Conclusions. When performing spinal anaesthesia, anaesthesiologists should be aware of
potential differences of the CMT location, particularly in female patients with thoracic
vertebral compression fractures, who may have a lower CMT than normal, extending to
the level of L2. Performing spinal anaesthesia at the L2–L3 interspace would seem to be
ill-advised in this patient population.
Keywords: conus medullaris; intervertebral disc disorder; spinal cord injury; subarachnoid
block; vertebral disease
Accepted for publication: 9 July 2010
One of the major concerns during needle insertion for spinal
anaesthesia is the location of the conus medullaris terminus
(CMT). Anaesthesiologists remain aware that any manoeuvre
which places the spinal needle in contact with the spinal cord
may lead to serious neurological injury.1 – 6 This risk may be
increased when the CMT is at a lower than expected level
and the L2–L3 interspace is chosen as the needle entry
site. The CMT may be lower in disorders of the vertebral
column. The position of the CMT has been studied previously
in cadavers,7 but a possibly better evaluation of its true position, and one which might translate into clinical practice
more readily, might be magnetic resonance imaging (MRI).
It is widely known that MRI is an extremely accurate and
practical way to evaluate the spinal cord, and a more accurate method of determining the location of the CMT than
cadaveric examination.8 Several retrospective studies9 – 12
have investigated the influences of age, gender, and position
of the body on the CMT previously; however, the influence of
disorders of the vertebral bodies or intervertebral discs has
not been well investigated. Our aim in this study is to investigate the effects of these disorders on the CMT location,
especially with regard to the customary sites of needle insertion during spinal anaesthesia.
Methods
With institutional Ethics Board approval, we retrospectively
studied MRIs of the spine of 1047 Chinese patients (444
males and 603 females), of age ranging from 10 to 79 yr,
in order to determine the location of CMT. These patients
& The Author [2010]. Published by Oxford University Press on behalf of the British Journal of Anaesthesia. All rights reserved.
For Permissions, please email: [email protected]
BJA
Lin et al.
Table 1 Given values of CMT levels and age decades. T12-U1/3,
upper 1/3 of T12; T12-M1/3, middle 1/3 of T12; T12-L1/3, lower 1/3
of T12; T12-L1 disc, intervetebral disc between T12 and L1; L1-U1/
3, upper 1/3 of L1; L1-M1/3, middle 1/3 of L1; L1-L1/3, lower 1/3 of
L1; L1-L2 disc, intervertebral disc between L1 and L2; L2-U1/3,
upper 1/3 of L2; L2-M1/3, middle 1/3 of L2; L2-L1/3, lower 1/3 of L2;
L2 – L3 disc, intervetebral disc between L2 and L3. The youngest
patient studied was 10 years old, corresponding to a value of 1 for
age decade (10 – 19 yr); the oldest patient studied was 79 yr old,
corresponding to a value of 7 for age decade (70 – 79 yr)
L1-U1
/3
L1-M1
/3
L1-L1
/3
CMT
L2
L3
L4
L5
Fig 1 T2-weighted sagittal MRI of the lumbar spine, demonstrating the method used of dividing the vertebral bodies into three
equal parts for the purpose of classification. The CMT is indicated
by the arrow; the dashed line corresponds to its assigned spinal
level (in this case, the L1 – L2 intervertebral disc).
were outpatients who presented for diagnosis and/or treatment of low back pain, hip pain, and/or lumbar radiculopathic pain. Patients with diseases such as tumour,
infection, ischaemia, or haemorrhage of the spine or spinal
cord, patients with congenital spinal anomalies, and patients
with the spine or spinal cord diseases that made CMT observations difficult were excluded. MRI examinations were
performed as T2-weighted images on 1.5 T Visart (Toshiba,
Tokyo, Japan) or 3.0 T Magnetom Trio (Siemens, Erlangen,
Germany) systems; other technical specifications of the
images obtained included: matrix 256×320, 4 mm slice
thickness, and 0.8 mm interslice gap. All scans were obtained
in the supine position. In each case which was used, the CMT
was clearly visualized on the sagittal sequences.
For the purposes of quantification, we drew a perpendicular line from the ending point of the CMT to the long axis of
the spine and determined the level of the surrounding vertebrae and intervertebral discs. Each vertebral body was
divided into three equal parts (upper 1/3, middle 1/3, and
858
Value
CMT level
Age decade (yr)
1
T12-U1/3
10 – 19
2
T12-M1/3
20 – 29
3
T12-L1/3
30 – 39
4
T12-L1 disc
40 – 49
5
L1-U1/3
50 – 59
6
L1-M1/3
60 – 69
7
L1-L1/3
70 – 79
8
L1-L2 disc
—
9
L2-U1/3
—
10
L2-M1/3
—
11
L2-L1/3
—
12
L2– L3 disc
—
lower 1/3, abbreviated as ‘U1/3’, ‘M1/3’, and ‘L1/3’, respectively); the entire intervertebral disc was considered an
additional separate part (Fig. 1).
CMT levels were then assigned a value ranging from 1 (U1/
3 of T12) to 12 (L2– L3 intervertebral disc) for the purpose of
statistical analysis, which were the highest and lowest CMT
levels observed in this population, respectively. This system
of CMT level correlation to adjacent vertebrae or intervertebral discs as a method of measurement has been reported
previously.10 – 12 13 – 15
On the basis of age, patients were stratified into seven
groups, such that ages 10–19 yr corresponded to a value
of 1, ages 20–29 yr corresponded to a value of 2, and
every decade thereafter corresponded to the next value
number, with value 7 corresponding to patients aged 70 –
79 yr (Table 1).
We performed all statistical analyses utilizing Statistical
Package for the Social Sciences (SPSS) version 11.5
(Chicago, IL, USA). The Kolmogorov –Smirnov tests or the
Skewness –Kurtosis tests were used to describe the distribution; two-sample t-tests were used for analysing variance.
Statistical significance was defined as a P-value of ,0.05.
Results
The location of the CMT in patients with and without various
vertebral and intervertebral disc diseases demonstrated a
normal age distribution among every disease cohort
(Table 2). The mean position of the CMT for each disease
cohort, as designated by its assigned value, is also shown
in Table 2.
BJA
Spinal anaesthesia in spinal disorders
Table 2 Frequency distribution and mean CMT of normal, vertebral disease, and disc disease cohorts. The Kolmogorov–Smirnov testing was
used for evaluating the age distribution; P.0.05 for the Kolmogorov– Smirnov test indicates a normal distribution. The t-test was used for
comparing each cohort with the normal population; P,0.05 indicates a significant difference
Number
Frequency (%)
Age
Mean (SD)
CMT
Kolmogorov–
Smirnov
P-value
Mean (SD)
t
P-value
—
Normal
65
6.21
38 (9.7)
0.458
0.985
6.5 (1.85)
—
Intervertebral disc extrusion
31
2.96
48 (14.4)
0.685
0.736
7.0 (2.01)
21.180
0.241
Intervertebral disc herniation
130
12.42
54 (13.9)
0.747
0.632
6.6 (1.96)
20.161
0.872
Intervetebral disc bulging
643
61.41
56 (12.6)
1.354
0.051
6.8 (2.04)
21.355
0.175
Scoliosis
21
2.00
65 (15.2)
0.958
0.317
6.3 (2.37)
20.412
0.681
Spondylolisthesis
56
5.35
62 (11.4)
1.098
0.179
6.8 (1.77)
0.451
0.653
Thoracic vertebral compression fracture
33
3.15
58 (16.0)
0.733
0.656
7.5 (2.56)
22.068
0.044
Lumbar vertebral compression fracture
68
6.49
58 (15.9)
1.059
0.212
6.6 (2.20)
20.440
—
—
—
—
—
Total
1047
100
0.661
—
Beij
WUPEIZHEM
T11
T12
T12
L1
L1
L2
L3
L2
L4
L3
L5
L4
L5
Fig 2 T2-weighted sagittal MRI demonstrating a T12 vertebral compression fracture (left) and a T11 vertebral compression fracture (right), as
shown by the arrows.
Compared with the disease-free cohort [mean CMT¼6.50
(1.85), corresponding to the M1/3 and L1/3 of L1], the position of the CMT in lumbar disc disorders such as lumbar
intervertebral disc extrusion, herniation, and bulging was
not shown to be significantly different. Likewise, for patients
with spondylolisthesis, acquired degenerative scoliosis, and
lumbar vertebral compression fractures, no significant difference compared with the normal population was found. The
mean CMT position value in thoracic vertebral compression
fracture patients [7.5 (2.56), corresponding to the L1/3 of
L1 and the L1–L2 intervetebral disc], however, was significantly different (t¼ 22.068, P,0.05) from that in the
disease-free cohort (Table 2 and Fig. 2). In anatomical
terms, thoracic vertebral compression fracture patients presented with lower positions of the CMT than normal patients
by approximately one-third of a vertebral body length.
On the basis of gender, some significant differences in the
location of the CMT also appeared for certain diseases.
Female patients with thoracic vertebral compression fractures,
lumbar intervertebral disc herniation, and disc bulging had
859
BJA
Lin et al.
Table 3 Diseases of vertebral bodies and discs compared with normal cohort in number and mean position of CMT between males and females.
P,0.05 indicates a significant difference
Male
Number
Female
Mean (SD) CMT
Number
Statistic result
Mean (SD) CMT
t
P-value
Normal
28
6.32 (2.09)
37
6.59 (1.67)
20.585
0.561
Thoracic vertebral compression fracture
18
6.33 (2.33)
15
8.93 (2.12)
23.327
0.002
Lumbar vertebral compression fracture
33
6.18 (1.57)
35
7.06 (2.61)
21.686
0.097
Intervertebral disc extrusion
12
7.08 (1.83)
19
6.89 (2.16)
0.251
0.804
54
6.17 (1.77)
76
6.86 (2.05)
0.108
0.048
273
6.44 (2.08)
370
7.06 (1.98)
0.168
0.000
Intervertebral disc herniation
Intervertebral disc bulging
significantly lower CMT values than their male counterparts
(Table 3). Among all patients with thoracic vertebral compression fractures, females (mean CMT corresponding to the
U1/3 of L2) had a CMT which was one-third of a vertebral
body plus a whole disc width lower than that in males (mean
CMT corresponding to the M1/3 of L1). Those females with
lumbar intervertebral disc herniation or bulging (mean CMT
corresponding to the L1/3 of L1) had a CMT which was
one-third of a vertebral body lower than the males with the
same diseases (mean CMT corresponding to the M1/3 of L1).
Discussion
The administration of anaesthetic agents by a needle into the
thecal sac surrounding the spinal cord is a routine and commonplace technique used by anaesthesiologists. The intervertebral needle insertion site that is usually targeted is the L3–L4
interspace, or less commonly, the L2–L3 interspace. Spinal
anaesthesia is generally considered to be a very safe technique, with the complication of spinal cord injury being rarely
reported,16 but several recent reports1 – 6 highlight cases of permanent damage to the spinal cord after spinal anaesthesia.
Fettes and Wildsmith17 report and hypothesize as to the possible causes and mechanisms of direct spinal cord injury, including patient factors, equipment factors, and technique.
Vertebral body and intervertebral disc disorders, such as
vertebral compression fractures and disc herniations, are
found commonly in general population, and MRI is not routinely performed on these patients, as it is often unnecessary
and expensive. As this study shows, however, some of these
patients, namely those with thoracic vertebral compression
fractures (especially females), may be at risk for having a
CMT which is low enough within the spinal canal to be traumatized directly or indirectly by a spinal needle inserted at
the L2– L3 interspace. The results of this retrospective analysis, which focused on the anatomical inconsistency of the
CMT in various spinal disease states in 1047 Chinese patients
who underwent spinal MRI, may serve as an alert to those
anaesthesiologists who perform subarachnoid block at the
L2– L3 interspace in the susceptible population.
To date, this study is the largest analysis of the variation of
the CMT location in the vertebral body and intervertebral disc
disease population, and the first to describe and discuss the
860
possible effects of such diseases on the CMT position. We did
not identify a statistically significant difference in the CMT
level between healthy people and those with lumbar spinal
disorders such as lumbar vertebral compression fractures,
spondylolisthesis, scoliosis, or intervertebral disc diseases.
The CMT position in our study did significantly extend somewhat caudally in patients with thoracic vertebral compression fractures (approximately one-third of one vertebral
body width, on average).
Some investigations have reported a positive correlation
between the CMT level and increasing age. Lirk and colleagues10 and Soleiman and colleagues11 postulated that
older people in general have a lower CMT position than
younger people. But, Soleiman and colleagues hypothesized
that age might not be the key variable per se, but rather
that the deterioration of the spine that accompanies increasing age might be the determining patient factor. The spinal
column is surrounded and supported by many structures,
and decreasing spine length may be attributed to pathological changes of the intervertebral discs, perispinous ligaments, or vertebral bodies themselves. Indeed, spinal
deterioration due to these changes might be a substantial
factor which results in a lower CMT than what is normal.
The variable effect of lumbar vs thoracic vertebral compression fractures which we found would seem to make
some anatomical sense, since compression at the thoracic
level, where the spinal cord is present, should cause a downward movement of the CMT, whereas this effect should be
less apparent with lumbar compression below the level of
the CMT. As for disorders of the lumbar vertebral discs,
tearing of annular fibres and loss of hydration of the
nucleus pulposus may result in disc space narrowing, rendering the length of the spine slightly shorter. However, we
found no downward movement of the CMT in this population
(gender excluded), which also makes some anatomical sense
for the reasons mentioned above. Of note, most cases of
spondylolisthesis within that cohort occurred at the L4–L5
or L5–S1 levels, which by the same reasoning would also
not be expected to affect the CMT position significantly.
Our results indicate that gender may affect the CMT position
in various disease states, and hence could be an important
factor. Analysis between males and females with thoracic
vertebral compression fractures, lumbar intervertebral disc
BJA
Spinal anaesthesia in spinal disorders
herniations, and bulging discs displayed a significant difference, with females having a lower CMT position in each case.
Females with thoracic vertebral compression fractures,
having a CMT one-third of a vertebral body plus the entire
width of an intervertebral disc lower than their male counterparts (corresponding to the U1/3 of the L2 vertebral body, on
average), could clearly be at risk for spinal cord injury if the
L2– L3 interspace is chosen for spinal anaesthesia.
Demiryürek and colleagues12 studied 639 patients and
found a difference between males and females, stating that
the spinal cord tends to extend further caudally in females.
Others,13 14 18 however, observed no difference in spinal cord
length based on gender. These observations were made in
the general population, and in the absence of an examination
and comparison of complex patient factors, such as disease
states. At this point in time, the effect of gender alone on
spinal cord length is still unclear. On the basis of our data,
however, it seems that the CMT position in females is more
prone to the effects of spinal disease than in males.
It is well established that vertebral body pathology is
often the result of changes in bone mineral density, and
that females seem more vulnerable in this regard. Felson
and colleagues19 evaluated the association of weight and
bone mass in elderly male and female patients, and found
that differences in weight displayed much less of an effect
in males than in females on bone mineral density.
It should be noted that the effect of spinal column disease
on the CMT position may be different with lumbar, thoracic,
or cervical disorders, and the absence of any high thoracic
or cervical imaging in our study is a limitation of our analysis.
Secondly, despite a normal age distribution in each disease
cohort, our sample population was overall elderly and consisted of outpatients with pain complaints but generally
normal spinal anatomy. Thus, our findings may reflect this
patient population instead of the general population.
Thirdly, the number of degenerative scoliosis cases which
we were able to capture was too small to analyse the differences between males and females. Expansion of this and
any other disease cohort would improve the overall analysis.
Other patient factors, such as height and weight, were summarily excluded from this analysis.
In conclusion, our data suggest that it is important for the
anaesthesiologist to give ample consideration to the site of
spinal needle insertion when spinal column disease is
present. This is especially true for female patients with thoracic
vertebral compression fractures, who may be at an increased
risk for spinal cord injury if spinal anaesthesia is attempted
at the L2 –L3 interspace level. Interestingly, all published
reports of the spinal cord damage with spinal anaesthesia to
date involve female patients. Clearly, however, other
complex factors may play a role, such as needle insertion
level, equipment used, and other patient factors.8
Acknowledgements
We are most grateful to the Imaging Center of Neuroscience
of Beijjng Tiantan Hospital for providing MRI images for this
study and to Dr Antoun Koht and Dr Ming Lu for reviewing
the manuscript and providing valuable guidance.
Conflict of interest
None declared.
Funding
This work was supported by the Department of Anaesthesiology of Capital Medical University in Beijing, China, and by the
Department of Anaesthesiology of Northwestern University
in Chicago, USA.
References
1 Absalom AR, Martinelli G, Scott NB. Spinal cord injury caused by
direct damage by local anaesthetic infiltration needle. Br J
Anaesth 2001; 87: 512–5
2 Wenger M, Hauswirth CB, Brodhage RP. Undiagnosed adult diastematomyelia associated with neurological symptoms following
spinal anaesthesia. Anaesthesia 2001; 56: 764–76
3 Reynolds F. Damage to the conus medullaris following spinal
anaesthesia. Anesthesiology 2001; 56: 238– 47
4 Hamandi K, Mottershead J, Lewis T, et al. Irreversible damage to
the spinal cord following spinal anesthesia. Neurology 2002; 59:
624– 6
5 Parry H. Spinal cord damage. Anaesthesia 2001; 56: 290
6 Ahmad FU, Pandey P, Sharma BS, et al. Foot drop after spinal
anesthesia in a patient with a low-lying cord. Int J Obstet
Anesth 2006; 15: 233– 6
7 Huang SJ, Wu RM, Pu HQ, et al. Anatomical observation and clinical estimation of the position of adult spinal cord ending. Anat
Res 2003; 25: 194– 6
8 Wang GH, Wang KG, Jiang LM, et al. MR imaging of the
variation in position of the tip of the conus medullaris in the
Chinese population. Chin J Med Imaging Technol 2003; 19:
1193 – 4
9 Malas MA, Salbacak A, Buyukmumcu M, et al. An investigation
of the conus medullaris termination level during the period of
fetal development to adulthood. Kaibogaku Zasshi 2001; 76:
453– 9
10 Lirk P, Messner H, Deibl M, et al. Accuracy in estimating the
correct intervertebral space level during lumbar, thoracic and cervical epidural anaesthesia. Acta Anaesthesiol Scand 2004; 48:
347– 9
11 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
12 Demiryürek D, Aydingöz U, Aksit MD, et al. MR imaging determination of the normal level of conus medullaris. J Clin Imaging
2002; 26: 375– 7
13 Asif S, Sarah JDB, John W. The variation of position of the conus
medullaris in an adult population: a magnetic resonance imaging
study. Spine 1998; 23: 1452–6
14 Wang XJ, Song LL, Wei YQ, et al. MRI study of the position of the
conus medullaris and the dural sac: an analysis of 156 normal
adults. J Clin Radiol 2003; 22: 101–4
861
BJA
15 Wang KG, Wang GH, Zhang T, et al. MR imaging of the variation in
position of the tip of the conus medullaris in an adult population.
HeiBei Med 2002; 8: 866–7
16 Fischer B. Complications of regional anaesthesia. Anaesth Intensive Care Med 2007; 8: 151– 4
17 Fettes PDW, Wildsmith JAW. Somebody else’s nervous system. Br
J Anaesth 2002; 88: 760–3
862
Lin et al.
18 Kim JT, Bahk JH, Joohan S. Influence of age and sex on the position of the conus medullaris and Tuffier’s line in adults. Anesthesiology 2003; 99: 1359–63
19 Felson DT, Zhang Y, Hannan MT, Anderson JJ. Effects of weight
and body mass index on bone mineral density in men and
women: the Framingham study. J Bone Miner Res 1993; 8:
567– 73