Secondary Logo

Journal Logo

Original papers

Colour Doppler imaging of the interspinous and epidural space

Grau, T.; Leipold, R. W.; Horter, J.; Martin, E.; Motsch, J.

Author Information
European Journal of Anaesthesiology: November 2001 - Volume 18 - Issue 11 - p 706-712



Recently published reports [1–4] have demonstrated the value of ultrasound imaging as a diagnostic tool prior to puncture of the epidural space for neuraxial anaesthesia. The accuracy of ultrasound measurements of the epidural space depth, i.e. the distance from the skin to the flaval ligament, has been confirmed by several authors [5–7]. However, none of the preceding researchers, have explored the usefulness of Colour Doppler imaging (C-DI) in prepuncture diagnostics.

Two of the most feared complications of spinal and epidural anaesthesia techniques – the epidural haematoma and the intravenous (i.v.) application of local anaesthetics – are correlated to the inadvertent injury of a vessel. Vessel depiction by Colour Doppler imaging is one of the most appreciated features of medical ultrasonography. Using ultrasound imaging as a routine prepuncture diagnostic tool in our clinics, we found the puncture for epidural anaesthesia to be facilitated to a surprising extent. With regard to side-effects of epidural anaesthesia caused by vessel injury, Doppler examination seemed a promising enhancement of prepuncture diagnostics.

Nevertheless, the ultrasonic approach to the epidural space is complicated. The surrounding vertebral bones account for absorption and scattering of the ultrasound wave [8]. In this study, we aimed to ascertain the possibility of vessel detection in the L3/4 interspinal space using both B-mode (B-MI) and Colour Doppler imaging. At this, we determined the size and visibility of vessels overlying the lumbar epidural space in this area. We compared the efficacy of two transducer frequencies.


This study was approved by the institutional human research committee. After written informed consent was obtained from all participants, we performed an ultrasonic examination of the L3/4 interspinous region in 20 volunteers (10 males, 10 females). None of the patients had a history of back surgery or disease. Demographic data (gender, age, height, weight) were recorded. The body mass index (BMI) of each participant was calculated by dividing the weight (kg) by the square of the height (m2). The body surface (cm2) was calculated according to the Dubois formula (BS=H0.725 × W0.425 × 71.84) [9]. No volunteer showed evidence of congenital spinal abnormalities such as spina bifida or a history of spinal surgery.

For ultrasonography, we employed a General Electric LQ 400 Ultrasound System equipped with two probes: a 4-MHz and a 7-MHz broad-banded transducer (multifrequency output and sensing). Distances were measured (mm) with the built-in distance and measurement program. All ultrasound scans were performed by the same anaesthetist (TG), who had an experience of over 2000 scans of the epidural space.

We planned to compare four ultrasonic settings (B-MI 4 MHz and 7 MHz; C-DI 4 MHz and 7 MHz) for the quality of vessel depiction in the puncture area. Overall resolution on B-mode images was evaluated according to the distinction of the flaval ligament, the epidural space and the dura mater. No distinction – i.e. just one conjoint echo for all three structures – meant a ‘low’ resolution. When ligamentum flavum and dura were merely distinguishable, but the epidural space could not be seen explicitly, resolution was evaluated as ‘medium’. ‘High’ resolution meant a clear distinction of all three structures.

In ultrasound images, vascular structures can be identified by two different phenomena. In B-mode, mere ‘pulsation’ marks the according area. Colour Doppler imaging displays the blood flow itself using different colours for movement in different directions. Hence we evaluated the perceptibility of vessel pulsation on the one hand and the general visibility of vessels on the other hand.

The visibility of blood vessels was scored as ‘not visible’ and ‘visible’. The perceptibility of vessel pulsation was scored from 1 to 3 (1=no detection of pulsation, 2 = pulsation spots clearly identifiable, but no detectable lumen, 3 = both pulsation and lumen visible). All obtained data are given as mean ± SD unless stated otherwise.

Ultrasound scanning was performed with the volunteer in the sitting position. We examined the L3/L4 interspace only. We used two scanning planes, first placing the transducer in a longitudinal direction in a median position on top of the spinous processes, and then in a 2-cm paramedian position in a parallel direction. We compared all findings to evaluate the best technique for vessel detection.

Scanning in all planes was recorded on videotape (S-VHS) and as individual images. After a first evaluation during the examination, all findings were reviewed on tape. All data were recorded using tick boxes on standardized forms. Data acquisition and statistics were performed with a AMD© K6 II300 CPU System and a Compaq® Armada© 1550 notebook using a Windows 95b© (Microsoft®) platform. Statistic software was Excel 97© (Microsoft®), Primer® 4.04 and Winstat® 3.1. For calculations, the Chi-square test was used with Yates’ correction where appropriate.


Ultrasound examinations were performed on 20 patients. For their physical characteristics, see Table 1.

Table 1
Table 1:
Demographic data

The 4 MHz scans showed a low resolution in 19 volunteers; only one scan was of medium quality. With the 7-MHz transducer, a low resolution was seen in three scans. Twelve scans showed an acceptable imaging quality (medium resolution) and in five scans a highly resolved image was obtained (Figures 1 and 2).

Figure 1.
Figure 1.:
Comparison of 7 MHz with 4 MHz B-mode imaging regarding the quality of resolution. Overall resolution on B-Mode images was evaluated according to the distinction of the flaval ligament, the epidural space and the dura mater. □ 7 MHz; ▮ 4 MHz. P < 0.001 ; Chi-square 25.9.
Figure 2.
Figure 2.:
Ultrasound image (7 MHz B-mode) showing a vein (diameter 2. 5 mm). Longitudinal paramedian access. The spinous arcs of L3 and L4 are marked as well as the flaval ligament and the dorsal and ventral part of the dura mater.

Using the 4-MHz transducer, both imaging modes were utilizable. While B-mode imaging quality was higher using the 7-MHz transducer, Colour Doppler imaging proved to be unachievable with this frequency. Hence only three ultrasonic settings (B-MI 4 MHz, B-MI 7 MHz and C-DI 4 MHz) were recorded. We selected the best imaging quality of each imaging mode (B-MI 7 MHz and C-DI 4 MHz) and compared their quality of vessel depiction in the puncture area.

Vessel detection was possible in 50% of the B-MI pictures and in all of the Doppler pictures. Using the 7 MHz B-mode, in 10 cases no vessels could be seen. In eight participants, vessels were visible in acceptable quality and in two scans vessel visibility was very good. In 4 MHz C-DI, four scans showed an acceptable visibility and in 16 cases vessel visibility was very good (Figures 3 and 4).

Figure 3.
Figure 3.:
Comparison of the visibility of vessels. The techniques with the best resolution were chosen for each mode: 7 MHz B-mode was compared with 4 MHz Colour Doppler imaging. □=B-MI, 7 MHz; ▮=C-DI, 4 MHz;P < 0.001; Chi-square 22.2.
Figure 4.
Figure 4.:
Ultrasound image (4 MHz Colour Doppler mode) showing several small vessels in the oblique Doppler frame. Longitudinal paramedian access. The spinous arcs of L3 and L4 are marked as well as the dorsal and ventral part of the dura mater and the cauda equina.

The perceptibility of vascular pulsation was generally low using B-MI: only four scans showed pulsating spots, none showed a lumen. In the Colour Doppler mode, mere pulsation could be observed in seven participants. Pulsating vessels (including the lumen) were seen in 13 scans (Figure 5). Using the Colour Doppler mode, the visibility of vessels was evaluated as 2.8 ± 0.4, the perceptibility of pulsation was rated as 2.7 ± 0.5 and was significantly higher than in the B-Mode (P < 0.001).

Figure 5.
Figure 5.:
Comparison of the perceptability of vascular pulsation (B-mode vs. Colour Doppler imaging). The perceptibility of vessel pulsation was scored as follows: low=no detection of pulsation; medium = pulsation spots clearly identifiable, but no detectable lumen; high = both pulsation and lumen visible. □=B-MI; ▮=C-DI;P < 0.001; Chi-square 29.2.

In Doppler images, voluminous veins were the predominantly visible structures, small veins were barely echogenic. All veins showed remarkable changes of diameter according to the movements of the dura mater during respiration. Pulsation of small arteries was best detected in 7 MHz B-mode. Subdural and epidural arteries were visible in this mode only.

In 4 MHz C-DI, vessels with a diameter of less than 0.7 mm could not be demonstrated. In 7 MHz B-mode, vessels with a diameter of approximately 0.5 mm were still detectable. Although image resolution was higher with high frequency ultrasound, overall visibility of the vessels of the L3/4 interspace area was better using 4 MHz C-DI (Figure 6).

Figure 6.
Figure 6.:
Ultrasound image: overview on a 4 MHz B-mode picture presenting interspinous and epidural vessels.


Transducers of current real-time scanners emit sound waves of between 3.5 and 7.0 MHz in repetitive arrays, which scan the local anatomy in thin slices and are reflected back. As we used two probes – 4 MHz and a 7-MHz broad-banded transducer – the higher frequency consequently led to a higher resolution in the B-mode. Nevertheless, Colour Doppler imaging with 7 MHz probes led to no usable results: the Doppler signal quality was notably lower. This may be explained as follows: the energy content in the beam diminishes progressively as the ultrasonic wave propagates through tissue, and this attenuation is proportional to the frequency, too [8]. According to the high extent of energy absorption and of wave scattering in the tissue preceding the epidural space, the comparatively slowly acquired 7-MHz Doppler scans were unreadable. Furthermore, the ultrasonic approach to the epidural space is complicated by the surrounding vertebrae. The ultrasound beam enters the spinal column through a narrow ‘acoustic window’ between the vertebrae, respectively, and the spinous processes [1]. This effect is the main reason for a reduction of visual quality in ultrasound pictures and this factor is also responsible for the reduction of overview [4]. In Colour Doppler scanning the ‘weakened’ beam was nearly perpendicular to the direction of blood flow in most of the vessels along the spine. Ideally, the angle of insonation should be < 60°. Above this angle, error in determination of Doppler shift and blood flow velocity are magnified and the sonographer receives an incorrect impression of ‘no flow’ in the vessel.

In the B-mode image, small vessels were not identified as such. They were recomposed as ‘pulsating spots’, whose pulse frequency was in time with the volunteer’s heart rate. We interpreted those findings as vessels with a diameter below the axial and lateral resolution of our ultrasound equipment. The heartbeat creates alternating pressure on the tissue around vessels that can be observed as ‘pulsation’ even around low-diameter vessels.

By using 4-MHz Colour Doppler imaging, vessels of > 1.0 mm diameter neighbouring the epidural space could be identified. In 7 MHz B-mode, vessels were detectable from a diameter of 0.5 mm. Nevertheless, Colour Doppler imaging provided a better overall visibility of vessels and vessel pulsation. The basis of Doppler ultrasonography is the fact that reflected or scattered ultrasonic waves from a moving interface will undergo a frequency shift, the magnitude and the direction of which is quantified by the receiver. The integration of real-time imaging with pulsed wave Doppler allows for vessels to be identified. As the diameter of the vessel and the angle of interrogation can be measured, it even allows for an estimation of propagation speed and volume of flow. The higher resolution of the 7 MHz B-mode images could not counterbalance these advantages (Figure 3).

The diameters of conventional epidural catheters are roughly the same size or larger (> 0.8 mm) than the smallest vessels identified with either of the techniques. It seems save to assume that most vessels large enough for accidental cannulation can be located prior to or maybe, using real-time ultrasonographic guidance, even during epidural puncture.

Over the last decade, various techniques for the detection of catheter misplacement have been recommended. In 1990, the Leighton group considered the injection of air to be an effective indicator of intravenously (i.v.) located epidural catheters [10,11]. This has been reconsidered for the common multiorifice catheters in a more recent publication [12]. In 1993, Wulf and colleagues suggested epidurography for the control of catheter positioning [13]. This technique requires the injection of radio-opaque dye and is still rarely used [14,15]. Tomczak and colleagues described magnetic resonance epidurography with gadolinium-DTPA as superior to conventional and CT epidurography [16].

In 1995, Trojanowski and Murray developed a simple, noninvasive test to distinguish between intravascular, intrathecal and extra-epidural placement [17]. Three years later, Colonna-Romano and colleagues [18] described the injection of a small amount of epinephrine to identify misplaced catheters. In the most recent publications, Tsui and colleagues used nerve stimulation to determine epidural catheter placement [19–21]. Consequently, several, mostly invasive, tools are available to avoid i.v. injection of local anaesthetics after inadvertent venous puncture. No research, however, has so far been performed on diagnostic measures to avoid vessel injury in the first place.

Utrasound imaging prior to neuraxial anaesthesia is a new field and Colour Doppler imaging provides many new additional aspects for detecting the vessels. We think that in the depiction of vascular structures of the vascularization of the interspinous and epidural spaces Colour Doppler imaging will be extremely valuable. For diagnostic evaluation further studies are warranted and planned.


Funding: Landesforschungsprogramm/Forschungsförderungsprogramm der Universität Heidelberg 373/1999. The colour illustrations were enabled by AstraZeneca, Wedel, Germany.


1 Grau T, Leipold RW, Conradi R, Martin E, Motsch J. Ultraschall und Periduralanästhesie. Technische Möglichkeiten und Grenzen einer diagnostischen Untersuchung des Periduralraums. Anaesthesist 2001; 50: 90–101.
2 Grau T, Leipold RW, Conradi R, Martin E, Motsch J. Ultrasound imaging facilitates localization of the epidural space during combined spinal – epidural anesthesia. Reg Anaesth 2001; 26: 64–67.
3 Grau T, Leipold RW, Horter J, Conradi R, Martin E, Motsch J. The lumbar epidural space in pregnancy: visualisation by Ultrasonography. Br J Anaesth 2001; 86: 798–804.
4 Grau T, Leipold RW, Conradi R, Martin E, Motsch J. Paramedian access to the epidural space: the optimum window for ultrasound imaging. J Clin Anaesth 2001; 13: 213–217.
5 Cork RC, Kryc JJ, Vaughan RW. Ultrasonic localization of the lumbar epidural space. Anesthesiology 1980; 52: 513–516.
6 Currie JM. Measurement of the depth of the epidural space using ultrasound. Br J Anaesth 1984; 56: 345–347.
7 Bonazzi M, de Gracia LB. Individuazione ecoguidata dello spazio epidurale lombare. Minerva Anestesiol 1995; 61: 201–205.
8 Kossoff G. Basic physics and imaging characteristics of ultrasound. World J Surg 2000; 24: 134–142.
9 Du Bois D, Du Bois EF. A formula to estimate the approximate surface area if height and weight be known. 1916. Nutrition 1989; 5: 303–311.
10 Leighton BL, Gross JB. Air: an effective indicator of intravenously located epidural catheters. Anesthesiology 1989; 71: 848–851.
11 Leighton BL, Norris MC, DeSimone CA, Rosko T, Gross JB. The air test as a clinically useful indicator of intravenously placed epidural catheters. Anesthesiology 1990; 73: 610–613.
12 Leighton BL, Topkis WG, Gross JB et al. Multiport epidural catheters: does the air test work? Anesthesiology 2000; 92: 1617–1620.
13 Wulf H, Kibbel K, Mercker S, Maier C, Gleim M, Crayen E. Radiologic position control of epidural catheters (epidurography). An instrument of quality assurance for regional analgesia. Anaesthesist 1993; 42: 536–544.
14 Diez Rodriguez-Labajo A, Reinoso-Barbero F, Sanabria P, Rodriguez E, Suso B. Usefulness of radiologic monitoring of epidural catheters using epidurography. Rev Esp Anestesiol Reanim 1998; 45: 416–420.
15 Johnson BA, Schellhas KP, Pollei SR. Epidurography and therapeutic epidural injections: technical considerations and experience with 5334 cases. Am J Neuroradiol 1999; 20: 697–705.
16 Tomczak RJ, Seeling W, Mergo P, Rieber A, Aschoff A, Brambs HJ. Magnetic resonance epidurography with gadolinium-DTPA. Eur Radiol 1998; 8: 1452–1454.
17 Trojanowski A, Murray WB. A test to prevent subarachnoid and intravascular injections during epidural analgesia. S Afr Med J 1995; 85: 531–534.
18 Colonna-Romano P, Nagaraj L. Tests to evaluate intravenous placement of epidural catheters in laboring women: a prospective clinical study. Anesth Analg 1998; 86: 985–988.
19 Tsui BC, Gupta S, Finucane B. Detection of subarachnoid and intravascular epidural catheter placement. Can J Anaesth 1999; 46: 675–678.
20 Tsui BC, Gupta S, Finucane B. Determination of epidural catheter placement using nerve stimulation in obstetric patients. Reg Anesth Pain Med 1999; 24: 17–23.
21 Tsui BC, Guenther C, Emery D, Finucane B. Determining epidural catheter location using nerve stimulation with radiological confirmation. Reg Anesth Pain Med 2000; 25: 306–309.

ANAESTHETIC TECHNIQUES, spinal anaesthesia, epidural anaesthesia; DIAGNOSTIC TECHNIQUES, NEUROLOGICAL, spinal puncture; ULTRASONOGRAPHY, Doppler, colour

© 2001 European Academy of Anaesthesiology