Comparison of ultrasound imaging in transverse median and parasagittal oblique planes for thoracic epidurals: A pilot study.

BACKGROUND AND AIMS: The use of ultrasound (US) scanning to assess the depth of epidural space to prevent neurological complications is established in current practice. In this study, we hypothesised that pre-puncture US scanning for estimating the depth of epidural space for thoracic epidurals is comparable between transverse median (TM) and paramedian sagittal oblique (PSO) planes.
METHODS: We performed pre-puncture US scanning in 32 patients, posted for open abdominal surgeries. The imaging was done to detect the depth of epidural space from skin (ultrasound depth [UD]) and needle insertion point, in parasagittal oblique plane in PSO group and transverse median plane in TM group. Subsequently, epidural space was localised through the predetermined insertion point by 'loss of resistance' technique and needle depth (ND) to the epidural space was marked. Correlation between the UD and actual ND was calculated and concordance correlation coefficient (CCC) was used to determine the degree of agreement between UD and ND in both the planes.
RESULTS: The primary outcome, i.e., the comparison between UD and ND, done using Pearson correlation coefficient, was 0.99 in both PSO and TM groups, and the CCC was 0.93 (95% confidence interval [95% CI]: 0.81-0.97) and 0.90 (95% CI: 0.74-0.96) in PSO and TM groups respectively, which shows a strong positive association between UD and ND in both groups.
CONCLUSION: The use of pre-puncture US scanning in both PSO and TM planes for estimating the depth of epidural space at the level of mid- and lower-thoracic spine is comparable.

INTRODUCTION

Use of thoracic epidurals for intra- and post-operative analgesia for open abdominal surgeries has exponentially increased over the last decade. However, the conventional method, since its inception, has been dependent on anatomical surface landmarks and ‘loss of resistance technique (LOR)’. Commonly used blind technique can lead to dural penetration and more serious neurological injury especially at dorsal levels. Though there have been attempts at technical refinement using ultrasound (US) to improve accuracy and reduce the incidence of inadvertent dural entry, this is still underutilised in routine practice. Various clinical trials have compared the conventional method of identifying epidural space based on anatomical landmark with US imaging.[1]

Grau et al. in 2002 first explored the use of pre-procedure US in thoracic epidural punctures and showed efficacy of thoracic spine US imaging in comparison to magnetic resonance imaging.[2] The meticulous practice of US scanning for central neuraxial block has been well defined in literature.[34] However, in comparison to lumbar spine, the anatomy of midthoracic level (T5–T8) makes imaging much more challenging. The positioning of spinous process with steep angulation, overlying laminae and close vicinity of the epidural space to the spinal cord increases impending neurological complications at mid-thoracic level.[45]

The efficacy of US imaging in determining the depth of epidural space has been extensively documented by comparing the ultrasound depth (UD) in various axes to the actual ‘needle depth (ND)’ in obstetric and non-obstetric populations. The authors of this study trust that this is the first study comparing the two axes of US imaging to identify the epidural space at dorsal level. This study was designed to compare the efficacy of US imaging in the transverse median (TM) and paramedian sagittal oblique (PSO) planes in accurately predicting the depth to the epidural space and the point of needle insertion in thoracic epidural needle placement for open abdominal surgeries.

METHODS

Our study was carried out after approval by the Institutional Ethics Committee (EC/TMC/11/13), over a period of 1 year (September 1, 2014–August 1, 2015). We recruited 32 patients (16 in each group), above 18 years of age, ranging from American Society of Anesthesiologists Physical Status (ASA-PS) 1–3 posted for elective open abdominal surgeries. Written informed consent was taken from each patient before enrolment to the study.

Criteria for exclusion were spinal surgery, trauma or deformity, coagulopathy, sepsis and abnormal local skin condition. Patients were assigned to either group randomly using computer generated allocation numbers. Patients randomised to Group I underwent US in the parasagittal oblique plane and those in Group II in the TM plane. The pre-puncture scans were done with a 2.0–5.5 MHz curved array probe (LOGIQ e™ portable US Machine from GE Healthcare, UK) on patients in the sitting position with the upper back and neck flexed. All measurements were done using the inbuilt calliper of the US machine. Premedication consisted of oral lorazepam (2 mg) the night before surgery and again at 6 am on the day of surgery with sips of water. Preprocedural US scanning in all the patients was performed by single anaesthesiologist with 8 years of experience in doing USG-guided epidurals. The initial step was localisation of the level at which anaesthesia was desired corresponding to the surgical incision. This was done starting caudally at the sacrum and moving cranially to identify the L5–S1 interlaminar space and sequentially all the cranial levels in the PSO view. To see the sonoantomy at the designated level, the transducer was placed 1–2 cm paramedian and tilted obliquely towards the midline in PSO group and transversely in the plane of the intervertebral space in TM group. The image was saved to facilitate measurement of the skin to the inner aspect of posterior complex depth (the UD of epidural space) by an independent observer in PSO and TM plane [Figures 1 and 2]. An interspace above or below was selected, when the image quality in the chosen dermatome appeared poor. Image quality was rated as either good (sharp visualisation with clear demarcation), fair (visualised but with poor demarcation), or inconclusive (structures not discernible).

Figure 1

Ultrasound scanning in paramedian sagittal oblique plane. L1, L2 – Laminae, LF-PD – Ligamentum flavum-posterior duramater complex, VB-VD – Vertebral body-ventral duramater complex

Figure 2

Ultrasound scanning in transverse median plane. AP – Articular process, TP – Transverse process, LF-PD – Ligamentum flavum-posterior duramater complex

In US imaging, the ligamentum flavum, posterior duramater, and the epidural fat contained therein appear as a single hyperechoic line referred to as the posterior complex, as described by Sahota et al.[6] The anterior duramater, posterior longitudinal ligament and posterior vertebral body were identified as discrete structures. In both groups, once an optimum view was obtained, the image was saved, the transducer was stabilised and the midpoints of the upper and lower and the right and left borders of transducer were marked. The intersection of the resulting horizontal and vertical lines was the site chosen for puncture and catheter placement. The actual puncture and catheter placement was done by an anaesthesia registrar blinded to the UD. The direction of needle insertion was the same as that of the US probe which afforded the best view of the relevant anatomy in either group.

The conventional ‘LOR’ technique was used to determine the skin to posterior complex depth and the needle was duly marked once this was felt. The ND was measured to the nearest millimetre from the tip to the point referred to earlier. The epidural catheter was then placed after excluding intravascular or intrathecal placement. The procedural pain was graded on the visual analogue scale. General anaesthesia and epidural analgesia were then instituted as per institution's protocol.

The outcome measures were the precision of measurement of the skin to posterior complex depth in both the planes (UD-PSO and UD-TM) as compared to the depth measured during actual needle insertion (ND-PSO and ND-TM) as detailed above (primary outcome) and the accuracy of determining the insertion point as gauged by the number of reinsertions (change of puncture site) and needle redirections (secondary outcome). Statistical measures included mean, standard deviation (SD) and interquartile range (25th and 75th percentiles) for continuous data and percentages for discrete variables.

Inter-group comparison of continuous variables was done using the two sample t-test and Fisher exact test for discrete variables. Correlation between the UD and actual ND was calculated and concordance correlation coefficient (CCC) was used to determine the degree of agreement between UD and ND in both the planes.[78] The differences between actual ND and UD estimated depth were graphically plotted against the mean of ND and UD for each patient for both TM and PSO planes (Bland–Altman analysis). Statistical analysis was done using the SAS 9.3 for windows. (Copyright © 2011, SAS Institute Inc., Cary, NC, USA) software.

RESULTS

The study group consisted of 32 patients of whom one patient from the TM group (Group II) was excluded in the final analysis due to technical issues during imaging. There were no statistical differences between the PSO and the TM groups for age (P = 0.64), sex (P = 0.85), weight (P = 0.54), height (P = 0.48) and body mass index distribution (P = 0.76) [Table 1]. The mean UD was 3.64 (SD: 0.72) cm in PSO group, and 3.71(SD: 0.74) cm in TM group, while mean ND was 3.8 (SD: 0.77) cm in PSO group and 3.88 (SD: 0.77) cm in TM group.

Table 1

Demographics

The primary outcome, i.e., the comparison between UD and ND, done using Pearson correlation coefficient, was 0.99 in both PSO and TM group, and the CCC was 0.93 (95% confidence interval [95% CI]: 0.81–0.97) and 0.90 (95% CI: 0.74–0.96) in PSO and TM groups respectively, which showed a strong positive association [Figure 3]. The mean difference between UD and ND arrived at by the Bland–Altman analysis was 0.16 cm in both groups [Table 2]. The 95% limits of agreement were 0.08 (upper) to 0.24 (lower). Total number of patients with redirections of epidural needle is 13 (81%) in PSO group and 8 (53%) in TM group (P = 0.13) and with reinsertions is 3 (18.75%) in PSO group and 0 (0%) in TM group (P = 0.22). This difference was not statistically significant. There was no significant difference in image quality in two groups [Table 3]. The mean pain score during procedure was 1.31 in PSO group and 1.13 in TM group, which were comparable in both groups.

Figure 3

The differences between actual needle depth and ultrasound depth (y-axis) estimated depth graphically plotted against the means of the needle depth and ultrasound depth for each patient (x-axis) for both transverse median and paramedian sagittal oblique planes (Bland–Altman analysis)

Table 2

Bland-Altman analysis: Actual needle depth versus estimated ultrasound depth

Table 3

Characteristics of the ultrasound imaging and neuraxial procedures

DISCUSSION

Our results indicate that there is a high degree of conformity between the depth to the posterior complex estimated on US and the actual depth as measured on needle insertion. US can, therefore, be a useful adjunct during insertion of thoracic epidural catheters, particularly at dorsal levels. Our data show that total number of redirections and reinsertions of epidural needle are higher in PSO group as compared to TM group but due to limitations of sample size we could not draw definite conclusions regarding the difference between the two groups. Higher number of redirections and reinsertions in PSO group can be attributed to the fact that we could not reproduce the oblique angle of the US beam and match it with actual needle trajectory during needle insertion. Furthermore, the difference could be a consequence of epidural puncture in the transverse plane being the more commonly used technique in routine practice. In our study, the actual ND is more than the US-estimated depth by a mean of 0.16 cm, which shows that US scan can likely guard against unintended dural puncture. Thus LOR testing along with US imaging go hand in hand to avoid evitable complications while performing the epidural block.[9]

The use of US in comparing the depth of epidural space to the actual ‘ND’ and a good correlation of UD, with ND in different axes for obstetric and non-obstetric populations has been described in literature.[3101112131415] We can compare our results with a similar study [16] which showed a good association of the UD, and actual ND in PSO plane at mid and low dorsal levels. Another group compared two axes in obese pregnant women with US and both the PSO and TM scan were performed at a single lumbar level in all patients [6], whereas our study group was heterogeneous and patients were randomised to either PSO or TM group and the procedure was performed at the dorsal level. The limitations of our study are small sample size and the inability to direct the needle along a trajectory identical to that of the US beam. This could account for the differences between UD and ND. The solution to this could be real time scanning during needle insertion. In addition, all scans were done by a single operator and therefore inter-rater variability at different levels of experience and proficiency could not be assessed.

CONCLUSION

Pre-puncture ultrasonographic estimates of the depth to the epidural space obtained in the PSO plane are comparable to those obtained in the TM plane at thoracic level. Larger studies with the incorporation of real-time sonography would be helpful in evaluating the potential advantages of pre-puncture US scanning.

Financial support and sponsorship

Nil.

Conflicts of interest

There are no conflicts of interest.