Secondary Logo

Journal Logo

Obstetric Anesthesiology: Research Reports

Ultrasound Imaging of the Lumbar Spine in the Transverse Plane: The Correlation Between Estimated and Actual Depth to the Epidural Space in Obese Parturients

Balki, Mrinalini MBBS, MD*; Lee, Yung MD*; Halpern, Stephen MD, MSc, FRCPC; Carvalho, Jose C. A. MD, PhD, FANZCA, FRCPC*

Author Information
doi: 10.1213/ane.0b013e3181a323f6
  • Free
  • Chinese Language Editions


Epidural analgesia is commonly used for pain control in obstetrics. When initiating epidural anesthesia, clinicians rely on the palpation of anatomical landmarks to determine the skin puncture site and on “feel” to identify the epidural space. Obese women pose considerable challenges to the performance of this rather “blind” technique; longer procedure times are common, and higher rates of failures and complications have been reported.1–3

Ultrasonography has been used in a variety of ways to assist epidural needle placements. Grau et al.4–7 have done extensive research on the usefulness of ultrasound imaging to facilitate the placement of neuraxial blocks. However, in these studies, the longitudinal paramedian plane and the paramedian approach were used for imaging and needle placement, respectively; in addition, the obese population was insufficiently represented in their studies. A previous study done at our institution demonstrated that ultrasound imaging in the transverse plane accurately estimated the depth to the epidural space and provided reliable information regarding the landmarks required for the placement of epidural catheters for labor analgesia, via the midline approach, in the general obstetric population.8 However, whether this method has the same accuracy in obese parturients is still unknown.

The objective of this study was to assess the accuracy of prepuncture lumbar ultrasound scanning in the transverse plane as a tool for estimating the depth to the epidural space, and the optimal skin puncture site for epidural needle placement, in obese parturients.


After approval by the Research Ethics Board at Mount Sinai Hospital, this prospective cohort study was conducted from October 2006 to December 2007. The study was registered in the Clinical Trials Registry (NCT0043998). Laboring women requesting epidural analgesia were recruited after written informed consent was obtained. The inclusion criteria were parturients with a prepregnancy body mass index (BMI) ≥30 kg/m2, ASA Physical Status I–III, and full-term singleton pregnancies. The exclusion criteria were a prepregnancy BMI <30 kg/m2, marked spinal deformities, or a history of spinal surgery. Patient recruitment occurred only when the investigators were available. Obesity was defined as per the World Health Organization classification (Class I = BMI 30–34.9 kg/m2, Class II = BMI 35–39.9 kg/m2, and Class III or Morbid = BMI ≥40 kg/m2).9 Patients were sampled from each of these categories.

Ultrasound Scanning

Ultrasound scanning was done before the epidural needle insertion, by one of the investigators, using a portable ultrasound system equipped with a 2–5 MHz curved array probe (Zonare Medical Products Canada, Beaverton, Ontario, Canada). The imaging was performed with the patient in the sitting position with legs flexed and crossed, and the same position was used for the epidural needle placement.

Using the top of the buttock crease as a starting point, the ultrasound probe was moved cephalad in the paramedian longitudinal plane to identify the upper border of the sacrum. Then, by counting the laminae and the interspaces, the L3–4 interspace was identified.

The midline of the spine was then identified in the transverse plane by identifying the spinous process of L3. The distance from the skin to the tip of the spinous process was measured using a built-in caliper. The probe was then moved caudad to capture a view of the interspace containing the ligamentum flavum, the dorsal and the ventral dura mater, the posterior longitudinal ligament and the vertebral body, as well as the articular and the transverse processes (Fig. 1). Because the ligamentum flavum and the dorsal dura mater, along with the epidural space, typically appear as a single unit with this equipment, they were termed the “ligament-dura unit.”

Figure 1.:
Ultrasound imaging in the transverse approach showing the ligamentum-dura unit and the vertebral body/ventral dura. The depth to the epidural space, measured from the skin to the ventral border of the ligamentum-dura unit in the transverse approach at the interspace, was 6.22 cm. LF = ligamentum flavum, DD = dorsal dura, VB = vertebral body, VD = ventral dura, AP = articular process, TP = transverse process.

We started by scanning in a plane perpendicular to the skin to obtain the optimum quality image and then changed the angle of the ultrasound probe minimally, only if required, to improve the image quality. Once the best image of the interspace structures was captured, with the transducer kept still, the skin was marked at the midpoints of the right and left aspects of the probe and at the midpoints of the cephalad and caudad aspects of the probe. The probe was removed, and lines were drawn to connect these marks. The puncture site was determined by the intersection of these two lines (Fig. 2).

Figure 2.:
Skin markings are shown. (A) The vertical line marking the midline, (B) the horizontal line marking the interspace, and (C) the skin puncture site at the intersection of the lines.

The angle of the probe providing the best image of the interspace structures was visually noted. The ultrasound depth (UD), i.e., the depth to the epidural space or the distance from the skin to the ventral border of the ligament-dura unit, was measured with the aid of a built-in caliper (Fig. 1). The investigator rated the quality of the structures visualized on the ultrasound image as good, fair, or poor. The image was considered good if the visualized structures were sharp with clear demarcation; fair if they were not sharp but with some demarcation; and poor if they were visible but blurry with poor demarcation.

Epidural Procedure

The epidural needle was inserted by either a senior resident or a fellow who was blinded to the UD. The insertion point information was provided by the investigator by visible skin markings, whereas the angle information was provided verbally.

Before the needle insertion, the patient’s back was palpated by both the investigator and the resident/fellow, and the findings on superficial and deep palpation were documented. After infiltration of the skin with 2% lidocaine, the epidural needle insertion was performed using a 17 gauge (8.89 cm) Tuohy epidural needle with markings at 1-cm intervals.

The needle was introduced through the ultrasound-determined insertion point, at an angle reproducing the direction of the ultrasound beam, as instructed by the investigator. Once the epidural space was located using the loss of resistance to air or saline method, a sterile marker was placed on the needle, as close to the skin surface as possible, to determine the actual distance from the skin to the epidural space. A 19-G 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 by the resident/fellow, using a ruler with millimeter markings, and recorded as needle depth (ND). A test dose of 2% lidocaine 3 mL was administered through the epidural catheter, followed 3 min later by an initial loading dose of 0.125% bupivacaine 10 mL with fentanyl 50 μg.

The primary outcome was the accuracy and precision of the depth to the epidural space determined by ultrasound in the transverse plane, as measured by the correlation between the UD and the ND. The secondary outcomes were: (a) the absolute difference between the two values (ND considered to be the “gold standard”); (b) the accuracy of the insertion point as determined by the need to redirect (different angle) or reinsert the needle (different skin puncture site); (c) the correlation between the BMI and the distance from the skin to the spinous process, as estimated by ultrasound; (d) the correlation between the BMI and the ND; (e) the duration of the ultrasound scanning (from placing the probe on patient’s back until marking the insertion point on the skin) and of the epidural procedure (from local infiltration of skin to completion of catheter insertion); (f) the pain during the epidural needle placement assessed after the completion of the catheter insertion (verbal rating scale [VRS], 0–10, 0 = no pain, 10 = maximum pain); (g) the presence of effective analgesia (VRS <1, absence of unilateral or failed block) at 20 minutes after injection of the loading dose; (h) the time to comfort after the administration of the loading dose; (i) the patient satisfaction with the ultrasound and epidural techniques, assessed after the administration of the loading dose (scale 0–10, 0 = not satisfied, 10 = completely satisfied); and (j) the procedure complications (vascular puncture, dural puncture, paresthesia).

A research assistant, not involved in performing the procedure, documented the time required to perform the ultrasound examination and the epidural procedure, the number of attempts at redirection and reinsertion of the needle, the pain with the epidural procedure, the labor pain before and after the placement of the epidural needle, the time to comfort and the patient’s overall satisfaction with the procedure.

Statistical Analysis

Statistical analyses were done using Minitab 14.2 (Minitab, State College, PA). Descriptive statistics, including the mean, standard deviation (sd), and range, were calculated for the continuous data, and percentages were calculated for the discrete data. The precision of measurement, i.e., the correlation and mean difference (with 95% confidence interval [CI]) between the ND and the UD was calculated using the Pearson correlation coefficient (r) and the paired t-test, respectively. The accuracy was measured using the concordance correlation coefficient. To graphically represent the data, the UD was plotted against the ND, and regression analysis was used to determine the equation of the line of best fit. ND − UD was plotted against (ND + UD)/2 (Bland-Altman analysis) to display the upper and lower 95% limits of agreement.10

The sample size was based on a previous study from our institution.8 In that study, the correlation coefficient between the ND and UD was 0.88, the mean absolute difference was 0.01 cm and the sd was 0.3 cm. We assumed a similar correlation and a standard deviation of 0.5 cm. Using 80% power and an α error of 0.05, a sample size of 46 patients enabled us to demonstrate a correlation of at least 0.80 and an absolute difference of 0.2 cm between the measurements.


Fifty-four parturients were recruited, 46 of whom were considered for the analysis. Eight patients were excluded for reasons such as missing documentation (n = 1), missing consent (n = 1), or a prepregnancy BMI determined to be <30 kg/m2 after recruitment (n = 6). The mean maternal age was 32 ± 7 yr, and the mean height was 164 ± 8 cm. The range of prepregnancy body weight was 72–219 (mean 102 ± 28) kg, the prepregnancy BMI was 30–79 (median 36) kg/m2, and the BMI at delivery was 33–86 (median 40) kg/m2. Based on their prepregnancy BMI, 46% of the patients had Class I obesity, 30% had Class II obesity, and 24% had morbid/Class III obesity.

The Pearson correlation coefficient (r) between the UD and the ND was 0.85 (95% CI: 0.74–0.91, r2 = 71.4, P < 0.001). The concordance correlation coefficient was 0.79 (95% CI: 0.71–0.88). The graphical representation of the UD versus the ND, demonstrating the line of best fit and the regression equation, is shown in Figure 3. The Bland-Altman analysis is shown in Figure 4. The upper 95% confidence limit was 1.3 cm, and the lower limit was −0.7 cm.

Figure 3.:
The solid line is the regression analysis showing ultrasound depth (UD) versus needle depth (ND). The equation for the line of best fit is: ND = 0.022 + 1.04 UD. The dotted line is the line of identity.
Figure 4.:
The Bland-Altman analysis: The difference between the needle depth (ND) and the ultrasound estimated depth (UD) is plotted against the mean depth (ND + UD)/2. The dotted line is the line of identity. The solid lines represent the mean difference (0.3 cm) and 95% CI.

The mean ND was 6.6 ± 1.0 cm (range 4.5–8.5 cm), whereas the UD was 6.3 ± 0.8 cm (range 4.7–8.2 cm); the difference between the measurements was statistically significant (0.3 ± 0.5 cm, 95% CI: −0.1 to −0.4, P = 0.002).

We found a direct correlation between the BMI with the ND (r2 = 0.34, P < 0.0001) and with the UD (r2 = 0.19, P = 0.003). The mean distance from the skin to the tip of the spinous process, as determined by ultrasound, was 2.7 ± 0.9 cm (range 1.3–5.1 cm). This skin-spinous process distance correlated with the BMI (r2 = 0.23, P = 0.001), UD (r2 = 0.64, P < 0.001), and ND (r2 = 0.58, P = 0.004).

On examination of the back, the spinous processes could not be visualized or easily palpated in any of the patients but could be palpated on deep palpation in 63% of patients. With regards to the image quality, the visualization of the ligament-dura unit was good in 63%, fair in 28%, and poor in 9% of the patients. The spinous process image was good in 70%, fair in 28% and poor in 2% of the patients.

The epidural needle placement was done without reinsertions in 76% of the patients, and there was no need to redirect the needle in 67% of the parturients. The maximum number of reinsertions (all at the same intervertebral space) was three, and 93% of the catheters were successfully placed in three or fewer redirection attempts through the same puncture site. None of the patients had an accidental dural puncture or paresthesia. A single vascular puncture occurred with the needle in one patient, and the epidural needle was reinserted. The median VRS pain score during epidural placement was 1 (range 0–7), whereas the median patient satisfaction was rated as 10 (range 6–10). The mean duration of the ultrasound examination was 4.5 ± 2.9 min, and the duration of the epidural procedure was 8.3 ± 7.3 min. Effective pain relief was obtained in all the patients within 10.3 ± 7.5 min.


Obesity obscures the anatomical landmarks necessary for facile epidural space localization. The presence of adipose tissue, and the pregnancy-induced softening of the soft tissues and ligaments, may increase the false-positive rate when identifying the epidural space by the loss-of-resistance technique. The combination of these factors accounts for a higher incidence of technical difficulty, more puncture attempts, a higher failure rate, and/or an increased potential for accidental dural puncture.11 Our study suggests that ultrasound imaging in the transverse plane may optimize epidural catheter placement in obese parturients by providing a reasonable estimation of the depth to the epidural space and the skin puncture site. A randomized, controlled trial is, however, necessary to confirm the utility of ultrasound in this patient population.

The experience of performing a preprocedural ultrasound examination before initiating a neuraxial technique in a range of obese obstetric patients has not been described in the literature. Wallace et al.12 found that ultrasonography reliably estimated the depth to the epidural space in obese parturients; however, they measured the distance from the skin to the lamina in the longitudinal paramedian plane, rather than the distance from the skin to the ligamentum flavum in the transverse plane. Grau et al.13 studied the usefulness of ultrasonography in epidural catheter placements in patients with presumed difficult epidural punctures, 30% of whom were obese parturients; however, they also used the longitudinal paramedian approach for scanning. Moreover, the mean BMI of their patients at delivery was only 34 ± 5 kg/m2.

Recently, a group at our hospital validated an ultrasound scanning method for epidural catheter placement in the transverse plane, but obese parturients represented only 15% of the study population.8 We believe that scanning in the transverse approach provides a more appropriate assessment of the midline structures when a midline epidural technique is to be performed. Using the transverse approach, we found that the ultrasound-estimated depth to the epidural space correlated highly with the actual ND in obese patients. Although the difference between the UD and the ND was statistically significant, it was small and may not have any clinical relevance in obese patients. The regression analysis shows that the estimation of the depth by ultrasound diverges slightly from the ND as the ND increases. In addition, the concordance correlation coefficient is less than the Pearson correlation coefficient, reflecting a slight deviation from perfect agreement. Using the regression equation generated by our data, the UD underestimates the ND by about 0.48 cm at a ND of 10 cm. This may be attributed to the greater subcutaneous tissue compression by the ultrasound probe with increasing BMI.

The Bland-Altman analysis shows that, although the limits of agreement are somewhat wider than in the normal population,8 the estimation of the depth by ultrasound is still clinically useful because the absolute agreement was estimated within 1 cm in all but three patients. In each of these, the UD underestimated the ND. There were only two patients in whom the ultrasound overestimated the ND by more than 0.5 cm. We acknowledge that ultrasound examination does not preclude the need for loss-of-resistance testing for epidural space localization.

Previous studies have demonstrated a correlation between the distance from the skin to the lumbar epidural space with the BMI in a mixed population consisting of obese and nonobese parturients.14–18 Our results, involving only obese subjects, are consistent with those studies. The mean depth to the epidural space from the skin has been described as being in the range of 4.6–5.3 cm14,16; however, we found that in obese women, the epidural space is even deeper, with a mean of 6.6 cm (range 4.5–8.5 cm). The depth to the epidural space was more than 8 cm in only 17% of the patients. Thus, it seems appropriate to use a standard needle to identify the epidural space in the majority of obese women, unless the ultrasound predicts a depth of more than 8 cm.

In obese parturients, the initial failure rate for epidural catheter placement can be very high (42%), and multiple attempts at catheter placement are common.19 Perlow and Morgan20 noted that 74.4% of morbidly obese parturients needed more than a single attempt, and 14% needed more than three attempts for successful epidural placement. None of our patients required more than three attempts. There were no failed or unilateral blocks, indicating that we had a good level of success for the epidural catheter placement using the ultrasound-determined insertion point. A randomized, controlled trial is necessary, however, to determine whether the use of ultrasonography will improve outcomes in this patient population.

One limitation of our study was the inability to precisely measure the angle of the ultrasound beam. Because of the increased depth to the epidural space in obese women, a small-angle change at the skin surface results in a significant change in the needle trajectory, leading to the need for redirection or reinsertion. It can also contribute to the differences between UD and ND.

In summary, we suggest that ultrasonography is a useful tool to predict the depth to the epidural space in obese parturients. In these patients, ultrasound can also reliably determine the skin puncture site. Future advances in ultrasound technology will likely make this procedure even simpler, more accurate, and widely applicable.


The authors acknowledge Kristi Downey (Research Assistant) for organizing the database and recruiting patients for this study.


1.Gerig HJ, Kern F. Success and failure rate in peridural anesthesia. A 1-year study. Reg Anaesth 1985;8:25–32
2.Le Coq G, Ducot B, Benhamou D. Risk factors for inadequate pain relief during epidural analgesia for labor and delivery. Can J Anaesth 1998;45:719–23
3.Konrad C, Schupfer G, Wietlisbach M, Gerber H. Learning manual skills in anesthesiology—is there a recommended number of cases for anesthetic procedures? Anesth Analg 1998;86:635–9
4.Grau T, Leipold RW, Horter J, Conradi R, Martin E, Motsch J. The lumbar epidural space in pregnancy—visualization by ultrasonography. Br J Anaesth 2001;86:798–804
5.Grau T, Bartusseck E, Conradi R, Martin R, Motsch J. Ultrasound imaging improves learning curves in obstetric epidural anesthesia: a preliminary study. Can J Anaesth 2003;50:1047–50
6.Grau T, Leipold RW, Conradi R, Martin E, Motsch J. Efficacy of ultrasound imaging in obstetric epidural anesthesia. J Clin Anesth 2002;14:169–75
7.Grau T. The evolution of ultrasound imaging for neuraxial anesthesia. Refresher course outline. Can J Anaesth 2003;50:R1–R8
8.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;104:1188–92
9.WHO, World Health Organization. Obesity: preventing and managing the global epidemic. Report on a WHO consultation, WHO technical report series 894. Geneva: WHO, 2000
10.Bland JM, Altman DG. Measuring agreement in method comparison studies. Stat Methods Med Res 1999;8:135–60
11.Saravanakumar K, Rao SG, Cooper GM. Obesity and obstetric anaesthesia. Anaesthesia 2006;61:36–48
12.Wallace DH, Currie JM, Gilstrap LC, Santos R. Indirect sonographic guidance for epidural anesthesia in obese pregnant patients. Reg Anaesth 1992;17:233–6
13.Grau T, Leipold RW, Conradi R, Martin E. Ultrasound control for presumed difficult epidural puncture. Acta Anaesthesiol Scand 2001;45:766–71
14.Clinkscales CP, Greenfield MLVH, Vanarase M, Polley LS. An observational study of the relationship between lumbar epidural space depth and body mass index in Michigan parturients. Int J Obstet Anesth 2007;16:323–7
15.Palmer SK, Abram SE, Maitra AM, von Colditz JH. Distance from the skin to the lumbar epidural space in an obstetric population. Anesth Analg 1983;62:944–6
16.Watts RW. The influence of obesity on the relationship between body mass index and the distance to the epidural space from the skin. Anaesth Intensive Care 1993;21:309–10
17.Hamza J, Smida M, Benhamou D, Cohen S E. Parturient’s posture during epidural puncture affects the distance from skin to epidural space. J Clin Anesth 1995;7:1–4
18.Segal S, Beach M, Eappen S. A multivariate model to predict the distance from the skin to the epidural space in an obstetric population. Reg Anesth 1996;21:451–5
19.Hood DD, Dewan DM. Anesthetic and obstetric outcome in morbidly obese parturients. Anesthesiology 1993;79:1210–8
20.Perlow JH, Morgan MA. Massive maternal obesity and perioperative cesarean morbidity. Am J Obstet Gynecol 1994;170:560–5
© 2009 International Anesthesia Research Society