Thoracic paravertebral block results in an ipsilateral somatic motor and sensory nerve block of multiple contiguous thoracic dermatomes above and below the site of injection.1 Locating the paravertebral space can be technically difficult in part because it requires location of the transverse process by blind needle placement and has an overall failure rate of >10%.2 Failure to identify the transverse process results in several needle redirections causing pain and discomfort and increases the potential risk of pneumothorax.
The use of ultrasound offers the capability to place a catheter in the paravertebral space with real-time image guidance. Sonographic measurements obtained using ultrasound scanning of vertebral transverse processes and parietal pleura can give an accurate measurement of the depth from the skin to the paravertebral space.3 Herein, we describe a new technique for ultrasound paravertebral block. An initial cadaver study was performed to provide a description of the sonographic anatomy of the paravertebral space, and a subsequent clinical observation assessed patients' pain and satisfaction after this ultrasound-guided paravertebral block.
An embalmed male cadaver (age at death 75 yr, height 175 cm, and body mass index 31.1) was selected for study of ultrasound characteristics of the paravertebral space. With the cadaver prone, 5 thoracic paravertebral regions on each side were scanned using a linear array transducer of 10 MHz (SonoSite Titan, Bothell, WA). For each space, the examiner attempted to identify the transverse process, the superior costotransverse ligament, and the paravertebral space, and using real-time imaging, inject dye or place an epidural catheter into 1 of these targets.
A blinded anatomist then performed a dissection of the paravertebral region at each level, in a layer-by-layer fashion. For each paravertebral space, the locations of the dye and catheters were identified, photographed, and referenced with the intended target structures.
With institutional ethical approval and written, informed patient consent, 10 ASA physical status I and II patients undergoing mastectomy or wide local excision ± lymph node sampling ± axillary clearance were studied.
While in a seated, kyphotic posture, baseline pain was assessed using a verbal rating scale before midazolam up to 0.05 mg/kg was administered for anxiolysis as clinically indicated.
The third thoracic vertebral level was identified by palpating and counting down from vertebra prominens (C7) and using a 38-mm broadband (5-10 MHz) linear array transducer placed initially at a point 2.5 cm lateral to the tip of the spinous process in a vertical orientation, obtaining a sagittal paramedian view of the transverse process, superior costotransverse ligament, and underlying pleura (Fig. 1). The parietal pleura was identified as a bright structure running deep to the adjacent transverse processes, distinct from the deeper lung tissue, which could be seen to shimmer and move with patient respirations. The superior costotransverse ligament, less distinct, could be seen as a collection of homogeneous linear echogenic bands alternating with echo poor areas running from 1 transverse ligament to the next.
The midpoint of the transducer was aligned midway between the 2 adjacent transverse processes, local anesthesia infiltrated at its lower border, and an 18-gauge Tuohy needle introduced in a needle-in-plane approach in a cephalad orientation. The paravertebral space was entered midway between the 2 transverse processes avoiding bony contact. The tip of the needle was advanced under direct vision to puncture the costotransverse ligament. Saline (3 mL) was then injected deep to the superior costotransverse ligament to (a) demonstrate the position of injectate deep to the ligament, and (b) allow easier passage of the catheter, to a distance of 2-3 cm beyond the needle tip.
The time taken for the block was recorded, from the initial scanning point up to the point of securing the catheter. An initial test dose of 3 mL of 2% lidocaine with 1:200,000 epinephrine was followed by 0.25% bupivacaine (0.3 mL/kg), administered over 10 min, after recording arterial blood pressure.
Before induction of anesthesia and 20 min after paravertebral block, sensory block was assessed by bilateral application of pinprick to the chest wall in the midclavicular lines, and patients' pain and satisfaction related to the paravertebral block were assessed using a verbal rating scale pain score and a satisfaction score (0 = dissatisfied, 10 = satisfied), respectively.
General anesthesia was induced with fentanyl 1-2 μg/kg and propofol 1-3 mg/kg and maintained with sevoflurane in 50% oxygen/nitrous oxide. Tracheal intubation was facilitated using atracurium 0.5 mg/kg. Patients received paracetamol 1 g and sodium diclofenac 100 mg per rectum. In the event of block failure (defined as sustained heart rate or mean arterial blood pressure increase >10% of preincisional value for 5 min or longer), patients received morphine 0.1 mg/kg.
Patients were transferred to the recovery room, and a paravertebral infusion was prescribed (0.25% bupivacaine at 5 mL/h), continued for up to 24 h postoperatively. Patients were assessed for pain and presence of a sensory block in the recovery room and on the first postoperative day, where patients rated their overall perioperative experience using a satisfaction scale (0 = completely dissatisfied, 10 = completely satisfied).
In the cadaver, 9 of 10 paravertebral spaces and targeted structures within were identified as transverse process, superior costotransverse ligament, and paravertebral space. A clear image of the 10th space was not achieved. Figure 2 demonstrates a typical paravertebral space and Figure 1 depicts the corresponding ultrasound image of the paravertebral space. Needles were successfully placed into the intended targets in 8 of 9 attempts.
In the clinical study, 5 women underwent wide local excision with axillary sampling, and the remaining 5 patients had a radical mastectomy and axillary clearance. Nine women (median [range] age 54.5 [42-76] yr and weight 70 [60-83] kg) had paravertebral catheters placed preoperatively. The 10th patient was withdrawn because of a fainting episode during the block performance. The mean (sd) block time (from initiating scan until completion of catheter fixation) was 523 (211) s.
Sixty-six percent of patients had either complete or partial sensory loss measured at mean (sd) 20 (4.8) min after block performance, increasing to 100% in the recovery room. The quantitative number of levels blocked was not recorded. Three of nine patients (1 wide local excision and 2 mastectomy), all of whom had sensory block preoperatively, were administered morphine intraoperatively, with a median (range) postoperative pain score of 0 (0-8) (Table 1). Two of nine (1 wide local excision and 1 mastectomy) used their patient-controlled analgesia device, with a median (range) morphine consumption of 6.5 (4-9) mg over the 24-h study period.
In the anatomical portion of this study, we obtained clear sonographic images of the thoracic paravertebral space, linking them directly to their anatomical structures at dissection. Based on these findings, we subsequently described a real-time ultrasound-guided paravertebral block, which was successful in a clinical pilot study. The technique resulted in a successful block in 8 of the 9 patients in whom it was attempted.
A consistent ultrasonographic appearance was demonstrated during the study, which was thought to be a characteristic sign of accurate placement of the needle in the paravertebral space. As the tip of the needle was advanced under direct vision to puncture the superior costotransverse ligament, injection of saline into the paravertebral space led to an anterior displacement of the parietal pleura (Fig. 3).
A technical difficulty with this technique is potential loss of image of the needle tip as it is advanced. This is due to the acute angle the needle must take to enter between adjacent transverse processes. Tissue disturbance may facilitate tracking of the needle tip in these circumstances.
Obvious limitations in both the anatomical and clinical elements of this study were the limited numbers studied and the lack of a control group. Further work will be required in a blinded comparative study of the traditional loss-of-resistance technique with our ultrasound-guided block.
1. Karmakar MK. Thoracic paravertebral block. Anesthesiology 2001;95:771–80
2. Eason MJ, Wyatt R. Paravertebral thoracic block—a reappraisal. Anaesthesia 1979;34:638–42
3. Pusch F, Wildling E, Klimscha W, Weinstabl C. Sonographic measurement of needle insertion depth in paravertebral blocks in women. Br J Anaesth 2000;85:841–3