May 15, 2008

Sonoanatomy of the Lumbar Spine in Patients With Previous Unintentional Dural Punctures During Labor Epidurals

By Lee, Yung Tanaka, Motoshi; Carvalho, Jose C A

Background and Objectives: Preprocedural lumbar ultrasound is a valuable tool to assess anatomical landmarks and predict the depth of the epidural space. Variations of the ligamentum flavum sonoanatomy are occasionally observed; however, no literature is available as to their incidence or clinical significance. We hypothesize that abnormal sonoanatomy of the lumbar spine detected by ultrasound can be associated with an increase in unintentional dural punctures. This study was undertaken to determine if the sonoanatomy of the lumbar spine of patients who had documented unintentional dural punctures differs from that of patients with a history of uneventful epidural placement for labor analgesia. Methods: Ultrasound of the entire lumbar spine was performed on 18 patients with a documented history of unintentional dural punctures, and 18 volunteers with a history of uneventful labor epidurals. At each interspace, we studied the quality of the ligamentum flavum (normal or abnormal), the symmetry of the bony structures (symmetric or asymmetric), and the distance from the skin to the ligamentum flavum. These parameters were compared in both groups.

Results: The incidence of asymmetric sonoanatomy and the distance from the skin to the ligamentum flavum was similar in both groups. The incidence of abnormal ligamentum flavum sonoanatomy was higher in the dural puncture group (overall odds ratio for the 5 interspaces was 8.21, 95% confidence interval 3.07-22.0, P

Conclusions: Abnormal sonoanatomy of the ligamentum flavum may represent anatomical variations of this structure, which may be related to an increased incidence of unintentional dural punctures during epidural placements. Reg Anesth Pain Med 2008;33:266-270.

Key Words: Unintentional dural puncture, Epidural anesthesia, Ligamentum flavum, Spinal sonoanatomy, Spinal ultrasound.

Epidural analgesia is the most common technique used to control labor pain. Although it has a high success rate, it is associated with some complications, notably unintentional dural punctures, ranging from 0.6% to 1.6%.1,2 Unintentional durai punctures might be related to the operator's skill or to the individual's anatomical variations. There are limited data in the literature to suggest the relative contribution of either factor.

The presence of a normal ligamentum flavum is fundamental to successfully identifying the epidural space when the loss of resistance technique is used during epidural needle placement. However, many studies have reported the presence of ligamentum flavum midline gaps, and implicated them as a potential cause of failure to recognize the loss of resistance.3,4 It is unknown whether these midline gaps and other ligamentum flavum abnormalities play a clinically significant role by increasing the risk of unintentional durai punctures.

Ultrasound has recently been introduced to facilitate the placement of epidurals, and the sonoanatomy of the lumbar spine is usually quite distinct and easily identified.5,6 Sometimes, however, some of the distinguishing elements of the pattern are not present, notably the ligamentum flavum. The exact significance of these abnormal patterns is unknown. There have been no previous reports correlating abnormal sonoanatomy of the lumbar spine with technical difficulties or complications during epidural insertion.

We hypothesize that abnormal sonoanatomy of the lumbar spine can be detected by ultrasound, and can be associated with an increase in complications of epidural placement. This study was undertaken to determine if the sonoanatomy of the lumbar spine of patients who had documented unintentional durai puncture is different than that of patients with a history of uneventful epidural placement for labor analgesia.


This case-control study was conducted after obtaining approval from the Mount Sinai Hospital Research Ethics Board and signed informed consent from the participants.

For the study group, the cases were selected from a database of 94 patients who had previously experienced a documented unintentional durai puncture during their labor epidurals. For the control group, staff volunteers working on the labor and delivery unit who had received uneventful epidurals in the past were recruited. All participants with known/visible spinal deformities, and those with previous lumbar spine surgery, were excluded from the study. In all the patients, the unintentional durai puncture had occurred during a midline approach.

Each of the participants underwent a Ll-Sl spine ultrasound scan with the aid of a portable MicroMaxx Ultrasound System equipped with a 2-5 MHz curved array probe (Sonosite Canada Inc, Markham, Ontario, Canada). All three investigators had been using ultrasound routinely for the placement of epidurals for at least 6 months. Images were obtained with the participants in the sitting position. The scanning consisted of a longitudinal approach to locate the interspaces L5- S1 to Ll-2, and a transverse approach for each of the interspaces. First, a paramedian longitudinal scan was performed by positioning the ultrasound probe vertically and perpendicularly to the long axis of the spine. The probe was initially placed over the sacrum, 3 cm to the left of the midline. At this point, a continuous hyperechoic line representing the ultrasound image of the sacrum was visualized. The probe was then moved cephalad until a hyperechoic saw-like image representing the articular processes and the interspaces was seen on the ultrasound screen. The exact level of each of the interspaces was pen-marked on the skin. Once the interspaces were determined, a transverse scanning of each individual interspace was performed. The spinous process, corresponding to the midline of the spine, was identified as a hyperechoic signal immediately underneath the skin, and continuing as a long vertical triangular hypoechoic acoustic shadow. The probe was then moved cephalad or caudad to capture a view of an acoustic window (interspace), which contained the ligamentum flavum, dorsal dura, durai sac, ventral dura, posterior longitudinal ligament, and vertebral body. Within the interspace, on the midline, a hyperechoic band corresponding to the ligamentum flavum and the dorsal dura was visualized. A second hyperechoic band, parallel to and deeper than the first band, corresponded to the ventral dura, posterior longitudinal ligament, and vertebral body unit. In addition to these midline structures, paramedian hyperechoic structures corresponding to the transverse and articular processes on each side were visualized on the same screen.

For each interspace, the best possible image of the transverse approach was captured and saved. At least 2 out of the 3 investigators were present and agreed upon the best images for each of the patients enrolled. Upon completion of all the cases, the 3 investigators reviewed and analyzed each image obtained for the presence of normal ligamentum flavum, and symmetry of the structures. The conclusion was based on at least 2 investigators agreeing on the images. The investigators were not blinded to the study groups or to each other's assessments.

The primary outcome of this study was the presence of an abnormal ligamentum flavum, as seen in the transverse approach. Sonoanatomy of the ligamentum flavum was described as either normal or abnormal. It was considered normal if a hyperechoic midline band, aligned with the hyperechoic signals of the transverse processes, was present and continuous (Fig 1). Conversely, it was defined as abnormal if a hyperechoic midline band was either absent or grossly discontinuous (Fig 2). When the ligamentum flavum was described as normal, the image was frozen and the ligamentum flavum depth was measured using a built-in caliper. This measurement was a secondary outcome of our study. The other secondary outcome was the symmetry of structures in relation to the midline. At each interspace, symmetric anatomy was described if the distances measured from the left and right articular processes and the transverse processes to the midline appeared to be the same, and the size and location of the articular and transverse processes also appeared to be similar on each side (Fig 1). When one of the above conditions was not satisfied, the interspace was considered asymmetric (Fig 3).

Fig 1. Normal lumbar spine sonogram with symmetric anatomy and typical ligamentum flavum (LF). AP, articular process; TP, transverse process; VB, vertebral body.

Statistical Analysis

For abnormal and asymmetric outcomes, exact (mid-P adjusted) P values and 95% confidence intervals (CI) of the odds ratios (OR) were calculated using StatXact Version 3 software (Cytel Software Corp, Cambridge, MA). The overall OR for each of these binary outcomes (with 95% CI and P value) was estimated by modeling the data using a generalized estimating equation (GEE) approach and an autoregressive (AR(I)) covariance structure to model the within- subject correlation using Proc GENMOD in SAS Version 9.1.3 (SAS Institute, Gary, NC).

Skin-ligamentum flavum depth data were log transformed to make the variances at each interspace independent of the means, so relative differences between groups are estimated. SAS Proc MIXED was used to model the data using a mixed linear model with an autoregressive covariance structure and heterogeneity between groups, as there was some evidence of a greater variance in the control group. At each interspace, the unequal variances f test was used to test for differences between groups using the SAS Proc TTEST. Fig 2. Abnormal lumbar spine sonogram with discontinuous ligamentum flavum (LF). AP, articular process; TP, transverse process; VB, vertebral body.

Fig 3. Abnormal lumbar spine sonogram with asymmetric articular processes (AP) and only a partial view of the vertebral body (VB). LF, ligamentum flavum; TP, transverse process.


For the study group, the information was sent to 94 potential participants by regular mail, but only 41 could be reached by telephone. Of the 41 reached, 24 agreed to participate but only 18 came to the appointment. For the control group, 18 staff volunteers were recruited. Patients were recruited between March 22 and July 23, 2007.

The mean +- SD for age, height, and weight in the study and control groups respectively were 35 +- 4 years and 42 +- 10 years; 160 +- 13 cm and 162 +- 5 cm; and 59 +- 12 kg and 61 +- 9 kg.

The incidence of asymmetric sonoanatomy and abnormal ligamentum flavum, and the ligamentum flavum depth as measured by ultrasound, are presented in Table 1. Both groups were similar with respect to the incidence of asymmetric sonoanatomy, and the distance from the skin to the ligamentum flavum. No patient had a distance from the skin to the ligamentum flavum less than 3 cm (range 3.42-7.21 cm). The incidence of abnormal ligamentum flavum sonoanatomy was significantly higher in the group of patients that had experienced durai punctures, especially at the levels of L4-5 and L3-4 (overall OR 8.21, 95% CI 3.07-22.0, P


Cork et al.7 and Currie8 were the first to observe a strong correlation between the depth of the epidurai space as estimated by ultrasound and the actual depth as confirmed by the epidural needle insertion. Further to those publications. Grau et al.6,9-12 and Arzola et al.5 also used ultrasound to assist epidurals, and proved its efficacy and consistency for localizing the site of needle insertion, and predicting the depth of the epidural space. All ultrasound estimations of the epidural space depth in the aforementioned studies relied on the presence of a clear ultrasound image of the ligamentum flavum. However, the image of this fundamental structure is sometimes grossly abnormal or entirely missing. There have been no previous reports on either the incidence of these abnormal patterns, or on the correlation of these abnormal patterns with technical difficulties or complications.

Table 1. Incidence of Asymmetric Anatomy and Abnormal Ligamentum Flavum and Skin-Ligamentum Flavum Distances at Different Interspaces

The present study compared the lumbar sonoanatomy of patients, who had previous documented unintentional durai punctures, to that of volunteers, who had uneventful labor epidurals. The results clearly indicated a higher incidence of abnormal sonoanatomy in the durai puncture group. While 72% of the study group sonograms showed abnormal patterns at level L4-5, only 17% of the control group images were abnormal at the same interspace. Significant discrepancies between the 2 groups were also observed at L3-4 and L5- S1.

The differences between the groups with respect to asymmetric sonoanatomy, and depth of the ligamentum flavum, conversely, were not significant. Asymmetric spine anatomy and "shallow" epidural space are commonly associated with a higher incidence of unintentional durai punctures, although a systematic correlation has never been established. In the general population, the distance from the skin to the epidural space, on the midline, varies from 4.23 cm to 4.93 cm on average, depending on the spinal level.13 The skin to ligamentum flavum distances measured in the present study were consistent with those findings.

Ligamentum flavum midline gaps have been documented in many anatomical and radiological studies, and are believed to be a cause of failure to recognize the loss of resistance to air or saline.3,4 Anatomical studies involving the direct dissection of embalmed cadavers showed the presence of ligamentum flavum midline gaps in variable frequencies at cervical and high thoracic,14 low thoracic,3 and lumbar4 levels. Hogan investigated cadavers by cryomicrotomy and also found, at the lumbar levels, a variable incidence of ligamentum flavum midline gaps.15 Lirk et al. were the first to document the frequency of lumbar spine ligamentum flavum midline gaps.4 Their results showed a frequency of up to 22.2%, with a higher incidence in the upper lumbar interspaces. In contrast to their results, our study showed a higher incidence of abnormal ligamentum flavum sonoanatomy at the lower lumbar interspaces. At this point in time, we can only suggest that the abnormal sonoanatomy does not necessarily reflect midline gaps, but rather something else to be further clarified.

Modern advances in medical imaging, such as computerized tomography and magnetic resonance imaging,16-18 have made it possible to evaluate the lumbar spine anatomy, ligamentum flavum included, without the potential for causing artifacts, such as in cadaver dissection. Further studies comparing images obtained by ultrasound and these other methods should be performed to clarify our results.

This study has limitations. First, the investigators were not blinded to the groups. We tried to minimize this factor by having all 3 investigators review the images. The final decision for each image was based on at least 2 agreeing on the assessment. second, this is the first study to investigate abnormal sonoanatomy patterns of the lumbar spine, and we are uncertain about the clinical relevance of our findings. Moreover, we are unable to rule out technical artifacts inherent in the imaging method as the cause of our results. In any case, there is a clear difference between normal and abnormal sonoanatomy, and this is what is clinically accessible at the bedside to guide clinicians.

In our opinion, lumbar spine ultrasound is a very useful clinical tool to facilitate epidural needle placement. It may be possible that, by detecting abnormal sonoanatomy, we can decrease complications such as unintentional durai punctures. Further prospective studies are necessary to prove our hypothesis. Until that happens, we suggest that clinicians should choose the interspace with the clearest sonoanatomy whenever ultrasound is being used to facilitate epidural placement.


The authors acknowledge Mr. Robert K. Parkes, B.Math, M.Sc., Senior Biostatistician, Samuel Lunenfeld Research Institute, Mount Sinai Hospital, for the statistical analysis.


1. Pan PH, Bogard TD, Owen MD. Incidence and characteristics of failures in obstetric neuraxial analgesia and anesthesia: A retrospective analysis of 19,259 deliveries. Int J Obstet Anesth 2004:13:227-233.

2. Okell RW, Sprigge JS. Unintentional durai puncture: A survey of recognition and management. Anaesthesia 1987;42:1110-1113.

3. Lirk P, Colvin J, Steger B, Colvin HP, Keller C, Rieder J, Kolbitsch C, Moriggl B. Incidence of lower thoracic ligamentum flavum midline gaps. Br J Anaesth 2005; 94:852-855.

4. Lirk P, Moriggl B, Colvin J, Keller C, Kirchmair L, Rieder J, Kolbitsch C. The incidence of lumbar ligamentum flavum midline gaps. Anesth Analg 2004;98: 1178-1180.

5. 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; 104:1188-1192.

6. Grau T, Leipold RW, Conradi R, Martin E, Motsch J. Efficacy of ultrasound imaging in obstetric epidural anesthesia. J Clin Anesth 2002; 14:169-175.

7. Cork R, Kryc J, Vaughan R. Ultrasonic localization of the lumbar epidural space. Anesthesiology 1980;52: 513-516.

8. Currie JM. Measurement of the depth to the extradural space using ultrasound. Br J Anaesth 1984;56: 345-347.

9. Grau T, Leipold RW, Conradi R, Martin EO, Motsch J. Ultrasound imaging facilitated localization of the epidural space during combined spinal and epidural anesthesia. Reg Anesth Pain Med 2001;26:64-67.

10. Grau T, Leipold RW, Horter J, Conradi R, Martin EO, Motsch J. Paramedian access to the epidural space: The optimum window for ultrasound imaging. J Clin Anesth 2001;13:213-217.

11. Grau T, Leipold RW, Horter J, Conradi R, Martin EO, Motsch J. The lumbar epidural space in pregnancy: Visualization by ultrasonography. Br J Anaesth 2001; 86:798-804.

12. Grau T, Leipold RW, Conradi R, Martin E. Ultrasound control for presumed difficult epidural. Acta Anaesth Scand 2001;45:7660 - 771.

13. Harrison GR, Clowes NWB. The depth of the lumbar epidural space from the skin. Anaesthesia 1985;40: 685-687.

14. Lirk P, Kolbitsch C, Putz G, Colvin J, Colvin HP, Lorenz I, Keller C, Kirchmair L, Rieder J, Moriggl B. Cervical and high thoracic ligamentum flavum frequently fails to fuse in the midline. Anesthesiology 2003;99:1387-1390.

15. Hogan QH. Epidural anatomy examined by cryomicrotome section. Influence of age, vertebral level, and disease. Reg Anesth 1996;21:395-406.

16. Harrison GR. Topographical anatomy of the lumbar epidural region: An in vivo study using computerized axial tomography. Br J Anaesth 1999;83:229-234.

17. Westbrook JL, Renowden SA, Carrie LES. Study of the anatomy of the extradural region using magnetic resonance imaging. Br J Anaesth 1993;71:495-498.

18. Grenier N, Kressel HY, Schiebler ML, Grossman RL Dalinka MK. Normal and degenerative posterior spinal structures: MR imaging. Radiology 1987;165:517-525.

Yung Lee, M.D., Motoshi Tanaka, M.D., and Jose C. A. Carvalho, M.D., Ph.D., F.A.N.Z.C.A., F.R.C.P.C.

From the Department of Anesthesia and Pain Management (Y.L., M.T., J.C.A.C.), and Obstetrics and Gynecology (J.C.A.C.), Mount Sinai Hospital, University of Toronto, Toronto, Ontario, Canada.

Accepted for publication December 19, 2007. Financial source: Departmental funds. Disclosures: None of the authors has any conflict of interest relative to this project.

Reprint requests: Jose C. A. Carvalho, M.D., Department of Anesthesia and Pain Management, Mount Sinai Hospital, 600 University Avenue, Room 781, Toronto, Ontario M5G 1X5, Canada. E-mail: [email protected]

(c) 2008 by the American Society of Regional Anesthesia and Pain Medicine.



Copyright Churchill Livingstone Inc., Medical Publishers May/Jun 2008

(c) 2008 Regional Anesthesia and Pain Medicine. Provided by ProQuest Information and Learning. All rights Reserved.