ORIGINAL ARTICLE
Ultrasound Imaging of the Thoracic Spine in Paramedian
Sagittal Oblique Plane
The Correlation Between Estimated and Actual
Depth to the Epidural Space
Aliya Salman, MD, FRCPC, Cristian Arzola, MD, MSc, Uma Tharmaratnam, MD, FRCPC,
and Mrinalini Balki, MD
Background: Ultrasound (US) imaging of the spine has been shown
to be a reliable tool to facilitate lumbar epidural needle placement;
however, its feasibility in thoracic epidural placement is still unknown.
The objective of this study was to assess the accuracy and reliability of
prepuncture US imaging in the paramedian sagittal oblique plane to
estimate the depth to the epidural space and optimum insertion point for
guiding epidural needle placement at the mid-low thoracic level.
Methods: This prospective study included 35 healthy adult patients
who requested thoracic epidural analgesia before their upper abdomi-
nal surgeries. Ultrasound imaging was done in the paramedian sagittal
oblique plane at the desired thoracic level to identify the intervertebral
space, the distance from the skin to the epidural space (US depth [UD])
and the needle insertion point. Subsequently, a staff anesthesiologist
located the epidural space through the predetermined insertion point and
marked the actual distance from the skin to the epidural space (needle
depth [ND]) on the needle with a sterile marker. The agreement between
the UD and the ND was calculated using the Pearson and concordance
correlation coefficients and Bland-Altman analysis with 95% limits of
agreement.
Results: The average patient age was 56 (SD, 14) years, and body mass
index was 28 (SD, 6) kg/m2. The precision of the agreement between
UD and ND estimated by Pearson correlation coefficient was 0.75, and
the accuracy was 0.80, whereas the concordance correlation coefficient
was 0.60 (confidence interval, 0.43Y0.78). The mean UD and ND were
4.3 (SD, 0.96) and 5.0 (SD, 1.2) cm, respectively. The Bland-Altman
analysis showed a mean difference of j0.71 cm (95% limits of agree-
ment, 0.8 to j2.2 cm). There was a significant direct correlation of
the ND with the body mass index (r2 = 0.27, P = 0.008). The mean
number of attempts was 1 (p25-p75 = 1Y2), and the epidural space was
identified with 2 or less redirections in 88% of the cases.
Conclusions: We found a good correlation between the US-estimated
distance to the epidural space and the actual measured needle distance
in our patients. We suggest that our proposed prepuncture US method,
using the paramedian sagittal oblique approach, can be a useful guide
to facilitate the placement of epidural needles at mid-low thoracic levels.
A randomized controlled trial is necessary to confirm the utility of
prepuncture US in thoracic epidural placement.
From the Department of Anesthesia and Pain Management, Mount Sinai
Hospital, University of Toronto, Toronto, Ontario, Canada.
Accepted for publication August 7, 2011.
Address correspondence to: Mrinalini Balki, MD, University of Toronto,
Department of Anesthesia and Pain Management, Mount Sinai Hospital,
600 University Ave, Room 1514, Toronto, Ontario, Canada M5G 1X5
(e-mail: mrinalini.balki@uhn.ca).
The authors have no conflicts of interest to declare.
This study received institutional funding.
Copyright * 2011 by American Society of Regional Anesthesia and Pain
Medicine
ISSN: 1098-7339
DOI: 10.1097/AAP.0b013e31823217e7
542 Regional Anesthesia and Pain Medicine
Copyright © 2011 American Society of Regional Anesthesia and Pain Medicine. Unauthorized reproduction of this article is prohibited.
(Reg Anesth Pain Med 2011;36: 542Y547)
T horacic epidural analgesia is highly effective in improving
the quality of intraoperative and postoperative pain relief
during thoracic and abdominal surgical procedures. The
advantages include reduction in adverse hemodynamic effects,
minimal motor blockade due to segmental analgesia, avoidance
of pulmonary complications, and accelerated recovery of post-
operative gastrointestinal function.1
The insertion of an epidural catheter at the thoracic level is
technically more challenging as compared with lumbar epidural
catheter placement. There is also a potential for neurologic
complications due to the close proximity of the thoracic epidural
space to the spinal cord.2,3 The conventional technique for tho-
racic epidural space localization continues to be heralded by
several drawbacks because of its ‘‘blind’’ nature, challenging
anatomic landmarks, and dependability solely on tactile feed-
back if a loss-of-resistance (LOR) technique is used. Although
the aforementioned factors and experience/skill of the anesthe-
siologist have been identified as possible sources of failure of
thoracic epidurals, a variety of other causes such as placing the
epidural catheter at an inappropriate level because of unreliable
surface landmarks, maldistribution of drugs delivered into epi-
dural space, and catheter dislodgement may also be related to
inadequate analgesia.4
Ultrasound (US) has recently emerged as a useful tool for
the identification of the intervertebral level, estimation of the
depth to the epidural space, and localization of midline and
interlaminar spaces for lumbar epidural catheter placement.5,6
Grau and colleagues7 explored the possibility of US imaging
technique to improve diagnostic and procedural capabilities in
the thoracic spine and have demonstrated a good correlation of
the US findings with those of magnetic resonance imaging.
Therefore, prepuncture US can be a possible tool to facilitate
identification of the thoracic epidural space.
The objective of this pilot study was to assess the accuracy
of prepuncture US scanning in the paramedian sagittal oblique
plane as a tool for estimating the depth to the epidural space and
the optimal puncture site for epidural needle placement at the
level of mid-lower thoracic spine. We believe this is the first
study evaluating the utility of US imaging in epidural catheter
placement at the thoracic level.
METHODS
After the institutional research ethics board at Mount Sinai
Hospital gave approval, this prospective cohort study was con-
ducted between January 2009 and July 2010. All patients gave
their written informed consent for participation in the study.
& Volume 36, Number 6, November-December 2011
Regional Anesthesia and Pain Medicine & Volume 36, Number 6, November-December 2011 US for Thoracic Epidural Placement

The inclusion criteria were adult patients (American Society
of Anesthesiologists physical status I and II) of both sexes
undergoing elective open abdominal surgeries under general
anesthesia and requesting thoracic epidural analgesia for post-
operative pain relief. Patients with coagulopathy or sepsis and
those with previous spinal surgery, trauma or deformity, or preg-
nancy were excluded.
Intravenous access and routine monitoring, including elec-
trocardiogram, noninvasive blood pressure, and pulse oximetry,
were established. The findings on superficial and deep palpa-
tion of the thoracic spines and interspaces and the visibility of
spinous processes were recorded. Ultrasound scanning was done
by 1 of 2 study investigators with experience in neuraxial scan-
ning (M.B. and U.T.); a staff anesthesiologist who had full ac-
cess to US data inserted the epidurals.
US Scanning
Ultrasound examination was performed using a 5-2 MHz
curved array probe (Zonare Medical Products Canada, Beaverton,
Ontario, Canada). The scanning was performed with the patient
in the sitting position, with the back arched and neck flexed
as for the epidural placement. The scanning was done in the
paramedian sagittal oblique plane by tilting the probe slightly
oblique toward the midline, starting at the level of sacrum. By
sliding the probe cephalad, alternate laminae and the interspaces
were identified, and the interspace corresponding to the mid-
line dermatome based on the surgical incision site was chosen.
If the sonoanatomy was not clear, the interspace above or
below was considered for detailed scanning and epidural place-
ment. Within the interspace, 2 parallel hyperechoic structuresV
the ligamentum flavum (LF)Ydorsal dura mater unit as the outer
band and the vertebral body (VB)Yposterior longitudinal ligamentY
ventral dura mater complex as the inner bandVwere identified
and centered on the screen. The image was frozen for various
measurements (Fig. 1A). Once the best image of the interspace
structures was captured, with the transducer stabilized, the skin
was marked at the midpoints of the cephalad and caudad aspects
and at the midpoints of the right and left aspects of the transducer.
The transducer was removed, and lines were drawn to connect
these marks. The puncture site was determined by the intersec-
tion of these 2 lines (Fig. 2). The plane of the US beam that pro-
duced the best image was noted to provide guidance for the
needle direction at the insertion point in that plane.
The US depth (UD), that is, the depth to the epidural space
or the distance from skin to LF, was measured using built-in
calipers on the US machine (Fig. 1). The depth from skin to
lamina and VB was also recorded. On the optimized image, a
vertical line was drawn with the calipers from skin to the center
of the LF (LF line). Another oblique line (VB line) was also
drawn crossing the visible cephalad end of the VB and LF line
at the level of LF. The angle (>) between the LF line and VB
line was visually noted (Fig. 1B) to determine the cephalocaudad
angle of needle insertion in the paramedian plane with respect
to the horizontal plane, that is, 90 degrees j > with respect to
the patient’s back. The quality of images was graded as good,
fair, or poor by the investigator. The image was considered
good if the visualized structures were sharp with clear demar-
cation of both the LF and VB, fair if they were not sharp but
demarcation was possible, and poor if they were visible but
blurry with poor demarcation of 1 or both structures.8
Epidural Procedure
After sterile preparation, the operator infiltrated the skin
at the US-determined puncture site with 2% lidocaine. The epi-
dural space was located with a 17-gauge (8.9 cm) Tuohy epidu-
ral needle with markings at 1-cm intervals, using the LOR to
air or saline method in the paramedian plane. The needle was
advanced cephalad, reproducing the predetermined angle in the
paramedian sagittal plane directly into the epidural space with-
out purposely contacting the lamina with the Tuohy needle. A
sterile marker was then placed on the needle, as close to the skin
surface as possible to determine the actual distance from the
skin to the epidural space. The insertion angle of the needle
was measured using a sterile spinal needle that was held along
the long axis of the spine and bent to reproduce the epidural
needle direction. This angle was later measured with a protractor.
A 19-gauge uniport, wire-embedded epidural catheter (Arrow
FlexTip Plus; Arrow International, Reading, Pa) was inserted
approximately 5 cm into the epidural space. After the removal of
the epidural needle, the distance from the tip of the needle to the
marker was measured, using a ruler with millimeter markings,
and recorded as the needle depth (ND). After aspiration of the
catheter to exclude intrathecal or intravascular placement, a test
dose of 2% lidocaine 3 mL was administered, followed 3 mins
later by an additional 5- to 8-mL dose to ensure dermatomal loss
of sensation to cold. General anesthesia was then instituted, and

FIGURE 1. A, Ultrasound imaging in the paramedian sagittal oblique approach at T6-7 interspace: hyperechoic laminae have
saw-teeth appearance; interspace between the 2 laminae contains parallel hyperechoic bands representing LFYdorsal dura mater
unit (upper) and the VBYposterior longitudinal ligamentYventral dura mater complex (lower). The depth to the epidural space,
measured from the skin to the ventral border of the LF-dura unit in the longitudinal paramedian approach, was 3.84 cm.
B, Ultrasound image, showing the angle (>) between the LF line (vertical line from skin to the center of the LF) and VB line
(oblique line crossing the cephalad end of the VB and the center of the LF) at the level of LF.

* 2011 American Society of Regional Anesthesia and Pain Medicine 543

Copyright © 2011 American Society of Regional Anesthesia and Pain Medicine. Unauthorized reproduction of this article is prohibited.
Salman et al Regional Anesthesia and Pain Medicine
FIGURE 2. Skin markings are shown. The vertical line marking
the paramedian insertion site (A), the horizontal line marking
the interspace (B), and the skin puncture site at the intersection
of these lines (C).
intraoperative analgesia was maintained by an infusion of
bupivacaine 0.2% with either fentanyl 3 Kg/mL or morphine 20
to 25 Kg/mL at a rate of 6 to 9 mL/min. The attending anes-
thesiologist was instructed to adjust the epidural infusion rate
to manage intraoperative hemodynamic changes.
The primary outcome was the accuracy and precision of
the depth to the epidural space determined by US in the para-
median sagittal oblique plane, as measured by the correlation
between the UD and the ND. The secondary outcomes were the
accuracy of the insertion point as determined by the number of
attempts (reinsertion through a different skin puncture site) or
544
Copyright © 2011 American Society of Regional Anesthesia and Pain Medicine. Unauthorized reproduction of this article is prohibited.
& Volume 36, Number 6, November-December 2011
redirections (changing the needle angle through the same
puncture site); the duration of the US scanning (from placing the
probe on patient’s back until marking the insertion point on the
skin) and the epidural procedure (from local infiltration of skin
to completion of catheter insertion); pain during epidural needle
placement assessed after the completion of catheter insertion
(verbal rating scale, 0Y10, 0 = no pain, 10 = maximum pain); and
procedure complications, such as paresthesia, vascular puncture,
and dural puncture.
A research assistant, not involved in performing the pro-
cedure, documented the number of attempts, the pain with the
epidural procedure, the time required to perform the procedure,
and complications. The need for intraoperative supplemental par-
enteral opioids to control sympathetic response to pain due to
failure of epidural administration of drugs was documented by
the attending anesthesiologist.
Statistical Methods
Descriptive statistics were calculated using means, SDs,
range, or median and interquartile range (25th and 75th per-
centiles) for continuous data and percentages for discrete vari-
ables. We used the concordance correlation coefficient (CCC) to
determine the degree of agreement between UD and ND.9,10 This
coefficient estimates agreement between 2 methods by measur-
ing the variation of their linear relationship from the 45-degree
line through the origin (perfect agreement). It measures how far
each observation deviates from the line that best fits the data
(precision) and how far this line deviates from the 45-degree line
through the origin (accuracy).11 The precision corresponds to the
Pearson correlation coefficient, which as a sole measure, could
potentially overestimate agreement,12 and the accuracy as a co-
efficient is estimated by the bias correction factor. To visually
evaluate CCC, UD was plotted against ND, and the line that best
fitted the data (Pearson) and the line of perfect agreement were
estimated.
The Bland-Altman analysis was performed to place the
magnitudes of the differences between the 2 measurements in a
more clinical context.13,14 This approach shows the graphical
presentation of agreement, plotting the differences between UD
and ND against the means of the 2 measures for each patient.
In addition, we estimated the 95% limits of agreement for the
differences, which represents differences likely to arise between
the 2 measures with a 95% probability. The assumption of nor-
mal distribution of the differences was checked by the Shapiro-
Wilk W test for normal data.
The required sample size of 29 patients was calculated
using a reference CCC of 0.88, as reported in a previous study at
our institution for lumbar spine US15 to test the null hypothesis
(1-tailed) of poor correlation observed in another study16 be-
tween the UD and the actual ND (correlation coefficient e0.59).
A type I error and type II error of 5% and 20% were assumed,
respectively. We elected to evaluate 35 patients (ie, 20% more)
to compensate for possible protocol violations during the study
period. Statistical analysis was performed with Stata 9.2 for
Macintosh (College Station, Tex).
RESULTS
Thirty-five patients were recruited in the study. Two patients
were excluded from the analysis because of technical prob-
lems during US imaging and incomplete documentation, leaving
33 patients for the analysis of the primary outcome.
Patient characteristics revealed the mean age of 56 (SD,
14) years, height of 168 (SD, 11) cm, and weight of 78 (SD, 19)
kg. The mean body mass index (BMI) was 27.5 (SD, 6.2) kg/m2
(range, 18.4Y47.75 kg/m2). There was a good balance between
* 2011 American Society of Regional Anesthesia and Pain Medicine
Regional Anesthesia and Pain Medicine & Volume 36, Number 6, November-December 2011
TABLE 1. Agreement Between Ultrasound Depth and
Needle Depth
All Patients BMI 930 kg/m2
Parameter (n = 33) (n = 14)
Concordance correlation coefficient 0.60 0.79
95% CI 0.43Y0.78 0.61Y0.97
Precision (Pearson r) 0.75 0.88
Accuracy (bias correction factor) 0.80 0.90
95% limits of agreement, cm j2.2 to 0.8 j2.0 to 1.1
Mean difference, cm j0.71 j0.5
sexes (female/male, 15/18). The anatomic landmarks were visi-
ble (85%) and palpable (94%) in the majority of the subjects,
whereas the quality of the US scans was deemed good in 76%
and fair in 24% of cases. The 3 most prevalent thoracic vertebral
interspaces, T7Y8, T8Y9, and T9Y10, were evenly distributed
between 11, 10, and 9 cases, respectively, whereas T10Y11 and
T6Y7 were used in 1 and 2 cases, respectively.
The mean UD was 4.3 (SD, 0.96) cm, and ND was 5.0
(SD, 1.17) cm. The precision of the agreement between UD and
ND estimated by Pearson correlation coefficient was 0.75, and
the accuracy was 0.80, whereas the CCC calculated was 0.60
(95% confidence interval [CI], 0.43Y0.78) (Table 1). The graph-
ical representation of UD versus ND is depicted in Figure 3.
The Bland-Altman analysis showed that the mean difference be-
tween UD and ND was j0.71 cm. The upper and lower 95%
limits of agreement were 0.8 and j2.2 cm, respectively (Fig. 4).
The mean depth estimated by US from skin to lamina was 3.6
(SD, 0.79) cm and to VB was 5.9 (SD, 0.94) cm.
There was a significant correlation of the BMI with the
ND (r2 = 0.27, P = 0.008) as well as with the UD (r2 = 0.41,
P = 0.0003). A secondary analysis of the correlation of
UD and ND in patients with BMI of greater than 30 kg/m2
showed Pearson correlation coefficient of 0.88, accuracy of
0.90, and the calculated CCC of 0.79 (95% CI, 0.61Y0.96).
The mean difference between the UD and ND was j0.5 cm
(95% limits of agreement, 1.1 to j2.0 cm) (Table 1).
During the epidural insertion, the mean number of attempts
was 1 (p25-p75 = 1Y2), requiring 4 and 3 attempts in only 1 and
FIGURE 3. The agreement between UD and ND, both in
centimeters. The correlation line that best fit the observed data
(green line) and the line of perfect agreement (45-degree red line)
are shown.
* 2011 American Society of Regional Anesthesia and Pain Medicine
Copyright © 2011 American Society of Regional Anesthesia and Pain Medicine. Unauthorized reproduction of this article is prohibited.
US for Thoracic Epidural Placement
FIGURE 4. Bland-Altman analysis. The difference between the
UD and the ND is plotted against the mean of UD and ND.
Line of perfect mean agreement (green line), observed mean
agreement (purple dash line), and 95% limits of agreement
(red lines) are shown.
2 subjects, respectively. There was no need to redirect the needle
in 58% of cases, whereas epidural space was identified with 2
or fewer redirections in 88% of the cases. The median angle of
needle insertion (90 degrees j >) in the plane of the US beam
was 50 degrees (p25-p75 = 40Y60 degrees). Patient discomfort
during the procedure presented a mean verbal rating scale of
3 (p25-p75 = 2Y4). The time required for US scanning was 6.3
(SD, 3.5) mins, and for the epidural insertion, 10.6 (SD, 6.2)
mins. Regarding complications during the epidural procedure,
2 patients had transient paresthesia without any neurologic se-
quelae, and 1 patient had unintentional dural puncture but did
not develop a headache. All patients demonstrated dermatomal
sensory loss to cold after the administration of lidocaine bolus.
The hemodynamic changes could be managed by adjusting the
rate of epidural infusions in all patients without the need for
additional parenteral opioids.
DISCUSSION
A good correlation of the estimated UD with the actual
ND, as seen in our study, indicates that US can be a useful tool
to facilitate epidural placement and to predict the needle distance
to the epidural space before its insertion at the thoracic level.
We found that the actual needle length to the epidural space
tended to be more than the US-derived depth from the skin to
the epidural space by a mean of 0.7 cm. Hence, we acknowledge
that US examination does not preclude the need for LOR test-
ing for epidural space localization but in fact provides some
margin of safety while performing the block. The wide limits
of agreement of UD and ND in our study (0.8 to j2.2 cm) could
perhaps be due to variation in the oblique angle of the US
probe in the paramedian sagittal plane required to direct the
beam toward midline for optimal imaging, as well as the accu-
racy of the estimated cephalocaudad angle or variation in the
actual needle trajectory during insertion. Interestingly, we found
a smaller difference between the UD and ND in obese patients
and a superior direct correlation of these variables in this pa-
tient population. This could perhaps be due to the contour of
the US probe that is likely to fit better in the paramedian
groove between the spine and the paraspinal muscles in obese
patients.
Our findings are comparable to those of US scanning at
lumbar levels. Various studies in obstetric and nonobstetric
patients have demonstrated an excellent correlation of the UD,
545
Salman et al Regional Anesthesia and Pain Medicine
measured in paramedian sagittal oblique or transverse median
planes, with the ND at lumbar levels (r = 0.79Y0.98).5,6,14Y16
Superior visibility of the sonoanatomic structures, because of
better acoustic windows in both paramedian sagittal oblique and
transverse median planes, and wider intervertebral spaces allow
easy passage of the epidural needle via both the approaches at
lumbar levels.14,17,18 However, at thoracic levels, imaging is
much more difficult because of extreme caudad angulation of
the spinous processes and overlapping laminae, which make the
interlaminar spaces very narrow and difficult to access.7,15 More-
over, a high number of occluded interlaminar spaces and exis-
tence of supraspinous ligament ossification at thoracic levels are
likely to impair the visibility of LF in some cases.19
Grau et al,7 in their study comparing the visibility with US
versus magnetic resonance imaging, demonstrated that para-
median sagittal oblique plane provides the best access to US
imaging of the thoracic spine. We believe that scanning in this
plane is more relevant to clinical practice as it also represents the
direction of the commonly practiced thoracic epidural needle
insertion via paramedian route. Studies have shown high corre-
lations of the computed tomography derived and the actual
depths to the epidural space; however, computed tomography
depth has shown to overestimate the actual epidural ND.20,21
Moreover, these modalities are expensive and not available as
bedside tools, and hence do not have clinical application during
epidural insertion.
We have demonstrated a high success rate of our US-guided
technique, with 88% catheter insertions accomplished on the
first attempt. Patient discomfort during epidural placement was
described as minimal, and there was evidence of adequate intra-
operative analgesia. Conventionally, thoracic epidural block is
performed by introducing the needle 1 to 2 cm lateral to the spi-
nous process until the lamina is encountered, and the needle is then
‘‘walked off’’ the superior edge of the lamina into the epidural
space.22 This requires needle redirections and may lead to patient
discomfort due to periosteal irritation. With the US-guided tech-
nique, the LFYdura mater unit can be directly visualized through
the interlaminar windows. Hence, our modified block technique
consisted of directly inserting the needle through the prede-
termined puncture site and testing for LOR into the epidural space
without walking off the lamina. This could have minimized the
degree of bony pain and discomfort during epidural needle in-
sertion in our patients. Such an experience of performing pre-
procedural US examination before initiating a neuraxial block at
the thoracic level has not been described before in the literature.
The complications recorded during this study included
transient paresthesia (n = 2) and dural puncture (n = 1). Although
the incidence of paresthesia during thoracic epidural placement
is unknown, an incidence of up to 47% has been described
during epidural needle placement in the lumbar region as part
of the combined spinal epidural technique.23 In the patient
with unintentional dural puncture, the visibility of the sono-
anatomic structures was fair, and US overestimated the ND by
0.5 cm. This underscores the need for further optimizing the US
imaging technique in patients with poor visibility of thoracic
spine structures.
The limitations of our study include the small sample size
and inability to precisely measure the angle of the US beam or
determine the accurate needle trajectory, which can also account
for the differences between UD and ND. Future advances in the
US technology will likely overcome this problem. We did not
blind the epidural performer to US findings, as we wanted to
test the feasibility of this technique of using US information
for epidural catheter placement; however, we considered LOR
as the end point for needle insertion to eliminate any bias.
546
Copyright © 2011 American Society of Regional Anesthesia and Pain Medicine. Unauthorized reproduction of this article is prohibited.
& Volume 36, Number 6, November-December 2011
In conclusion, this pilot study suggests that US can reliably
determine the skin puncture site and the depth to the epidural
space at the thoracic level and optimize the placement of mid-
low thoracic epidural catheters. However, a randomized con-
trolled trial is necessary to confirm the utility of US in this
patient population and aid thoracic epidural placement in an-
ticipated difficult cases.
ACKNOWLEDGMENT
The authors thank Kristi Downey (research assistant) for
organizing the database for this study.
REFERENCES
1. Groeben H. Epidural anesthesia and pulmonary function. J Anesth.
2006;20:290Y299.
2. Kao MC, Tsai SK, Tsou MY, Lee HK, Guo WY, Hu JS. Paraplegia
after delayed detection of inadvertent spinal cord injury during
thoracic epidural catheterization in an anesthetized elderly patient.
Anesth Analg. 2004;99:580Y583.
3. Mayall MF, Calder I. Spinal cord injury following an attempted
thoracic epidural. Anaesthesia. 1999;54:987Y998.
4. Motamed C, Farhat F, Rémérand F, Stéphanazzi J, Laplanche A,
Jayr C. An analysis of postoperative epidural analgesia failure
by computed tomography epidurography. Anesth Analg. 2006;103:
1026Y1032.
5. Chin KJ, Karmakar MK, Peng P. Ultrasonography of the adult
thoracic and lumbar spine for central neuraxial blockade. Anesthesia.
2011;114:1459Y1485.
6. Balki M. Locating the epidural space in obstetric patients-ultrasound
a useful tool: continuing professional development. Can J Anesth.
2010;57:1111Y1116.
7. Grau T, Leipold RW, Martin E. Ultrasound imaging of the thoracic
epidural space. Reg Anesth Pain Med. 2002;27:200Y206.
8. Balki M, Lee Y, Halpern S, Carvalho JC. Ultrasound imaging of
the lumbar spine in the transverse plane: the correlation between
estimated and actual depth to the epidural space in obese parturients.
Anesth Analg. 2009;108:1876Y1881.
9. Lin L. A concordance correlation coefficient to evaluate reproducibility.
Biometrics. 1989;45:255Y268.
10. Lin L, Hedayat AS, Sinha B, Yang M. Statistical methods in
assessing agreement: models, issues, and tools. J Am Stat Assoc.
2002;97:257Y270.
11. King TS, Chinchilli VM. A generalized concordance correlation
coefficient for continuous and categorical data. Stat Med.
2001;20:2131Y2147.
12. Bland JM, Altman DG. Statistical methods for assessing agreement
between two methods of clinical measurement. Lancet. 1986;1:307Y310.
13. Bland JM, Altman DG. Measuring agreement in method comparison
studies. Stat Methods Med Res. 1999;8:35Y60.
14. Cepeda MS, Africano JM, Polo R, et al. Agreement between percentage
pain reductions calculated from numeric rating scores of pain
intensity and those reported by patients with acute or cancer pain.
Pain. 2003;106:439Y442.
15. Arzola C, Davies S, Rofaeel A, Carvalho JCA. Ultrasound using the
transverse approach to the lumbar spine provides reliable landmarks
for labor epidurals. Anesth Analg. 2007;104:1188Y1192.
16. Chin KJ, Perlas A, Singh M, et al. An ultrasound-assisted approach
facilitates spinal anesthesia for total joint arthroplasty. Can J Anesth.
2009;56:643Y650.
17. Borges B, Wieczoreck P, Balki M, Carvalho JC. Sonoanatomy of the
lumbar spine of pregnant women at term. Reg Anesth Pain Med.
2009;34:581Y585.
* 2011 American Society of Regional Anesthesia and Pain Medicine
Regional Anesthesia and Pain Medicine & Volume 36, Number 6, November-December 2011 US for Thoracic Epidural Placement

18. Grau T, Leipold RW, Conradi R, Martin E, Motsch J. Efficacy of
ultrasound imaging in obstetrical epidural analgesia. J Clin Anesth.
2002;14:169Y175.
19. Murata H, Sakai T, Goto S, Sumikawa K. Three-dimensional
computed tomography for difficult thoracic epidural needle placement.
Anesth Analg. 2008;106:654Y658.
20. Kao MC, Tsai SK, Chang WK, et al. Prediction of the distance
from skin to epidural space for low-thoracic epidural catheter
insertion by computed tomography. Br J Anaesth. 2004;92:
271Y273.
21. Carnie J, Boden J, Smith FG. Prediction by computerised tomography
of distance from skin to epidural space during thoracic epidural
insertion. Anaesthesia. 2002;57:690Y709.
22. Bernards CM. Epidural and spinal anesthesia. In: Barash PG,
Cullen BF, Stoelting RK, Cahalan MK, Stock MC, eds. Clinical
Anesthesia. 6th ed. Philadelphia, PA: Lippincott Williams & Wilkins;
2008:927Y954.
23. Ahn HJ, Choi DH, Kim CS. Paraesthesia during the
needle-through-needle and the double segment technique for
combined spinal epidural anaesthesia. Anaesthesia. 2006;61:634Y638.

* 2011 American Society of Regional Anesthesia and Pain Medicine 547

Copyright © 2011 American Society of Regional Anesthesia and Pain Medicine. Unauthorized reproduction of this article is prohibited.