Ultrasound-guided thoracic paravertebral injection in dogs: a cadaveric study

Transcription

1 Veterinary Anaesthesia and Analgesia 2017, 44, 636e645 RESEARCH PAPER Ultrasound-guided thoracic paravertebral injection in dogs: a cadaveric study Diego A Portela a, Luis Campoy a, Pablo E Otero b, Manuel Martin-Flores a & Robin D Gleed a a Department of Clinical Sciences, College of Veterinary Medicine, Cornell University, Ithaca, NY, USA b Department of Anesthesiology, College of Veterinary Science, University of Buenos Aires, Buenos Aires, Argentina Correspondence: Luis Campoy, Department of Clinical Sciences, College of Veterinary Medicine, Cornell University, 602 Tower Road, Ithaca, NY , USA. luis.campoy@cornell.edu Abstract Objective To describe ultrasound-visualized anatomy and the spread characteristics of a dye injected in the thoracic paravertebral (TPV) space under ultrasound guidance. Study design Anatomic cadaver study. Animals Seven dog cadavers. Methods One cadaver was used to observe, identify, and describe the relevant TPV anatomy. In the remaining six, the left fifth TPV space was randomly assigned to be injected with either a low volume (LV; 0.05 ml kg 1 ) or high volume (HV; 0.15 ml kg 1 ) of dye. Subsequently, the contralateral side was injected with the alternative volume. Anatomic dissections were conducted to determine the incidence of complete spinal nerve staining (>1 cm circumferential coverage), number of contiguous spinal nerves dyed and the absence or presence of solution in particular locations. Results The ultrasound-visualized anatomy of the TPV space was defined as the intercostal space abaxial to the vertebral body, delimited by the parietal pleura ventrally and the internal intercostal membrane dorsally. The endothoracic fascia divides the paravertebral space into dorsal and ventral compartments. The target nerve was completely dyed in five of six and six of six injections in the LV and HV conditions, respectively. In one LV injection, the nerve was partially dyed. No multisegmental spread affecting contiguous spinal nerves was found in either treatment. Multisegmental spread was found in the ventral compartment of the TPV space, affecting the sympathetic trunk on 3 (0e3) and 3.5 (1e6) vertebral spinal levels in the LV and HV conditions, respectively, but differences between volumes were not significant. No intrapleural, ventral mediastinal or epidural migration was observed. Conclusions and clinical relevance Ultrasoundguided TPV block is a potentially reliable technique. The LV appeared sufficient to dye a single spinal nerve and multiple sympathetic trunk vertebral levels. Multiple TPV injections may be needed to provide adequate thoracic analgesia in dogs undergoing thoracic surgery. Keywords dog, endothoracic fascia, paravertebral space, regional anesthesia, thoracic paravertebral block, ultrasound. Introduction Thoracic paravertebral (TPV) block involves the injection of local anesthetic solution proximate to the spinal nerves as they emerge from the intervertebral foramen into the TPV space (Karmakar 2001). In humans, this peripheral nerve block is as effective as thoracic epidural anesthesia for providing perioperative pain relief in patients undergoing thoracic surgery and is associated with lower rates of complications (Davies et al. 2006). Several techniques have been described for approaching the TPV space in humans (Krediet et al. 2015). The injection of a relatively high volume of local anesthetic into the TPV space in humans can produce multisegmental analgesia (Saito et al. 2001). In contrast, no 636

2 radiographic evidence of the multisegmental spread of contrast solution was observed in dogs when the technique was guided by electrostimulation (Portela et al. 2012). Ultrasound guidance allows a realtime view of the TPV space during injection, which improves the accuracy of the technique and is likely to increase its success rate (O Riain et al. 2010). The aims of this study were: 1) to describe the gross and ultrasound-visualized anatomy of the TPV space in dogs; 2) to describe an ultrasound-guided approach for injection into the TPV space; and 3) to determine the patterns of distribution of two volumes of a dye solution injected into the fifth TPV space in dog cadavers. The study hypotheses were that the TPV space and its boundaries can be identified echographically and that the injection of a clinically sensible volume into the fifth TPV space will show multisegmental spread affecting contiguous thoracic spinal nerves. Materials and methods Phase I: anatomic and echographic study A 22 kg dog cadaver with no known spinal abnormalities was studied. The dorsal and lateral thoracic regions were clipped, and the dog cadaver was positioned in sternal recumbency. The spinal column and associated structures in the thoracic area were examined ultrasonographically on one side of the cadaver. The identity of the structures observed was confirmed by concurrent dissection of the contralateral side. On the left side, the skin was incised a few millimeters abaxial to the dorsal midline, parallel to the spine. The skin was removed and the transversospinalis and longissimus thoracis muscles were carefully disinserted and reflected. Subsequently, a layer-by-layer dissection was performed to expose the thoracic spinal nerves and to determine the anatomic boundaries of the TPV space. On the right side of the thorax, still images and a continuous video-recording of the relevant ultrasound-visualized anatomy were obtained using a linear-array ultrasound transducer (15-6 MHz; HFL50x; SonoSite, Inc., WA, USA) connected to an ultrasound machine (SonoSite Edge; SonoSite, Inc.). The transducer was initially placed in a longitudinal plane, parallel to the direction of the spine, on the dorsal midline, and then gradually glided to a more parasagittal position. Echographic structures were correlated with anatomic observations on the contralateral dissected hemithorax. The echographic scan was repeated with the transducer in a more oblique position (Fig. 1). Once the investigators (DAP, LC) became familiarized with the anatomy, a 20 gauge, 9 cm Tuohy needle (B Braun Medical, Inc., PA, USA) was advanced inplane in a laterocaudal to craniomedial direction toward the fifth TPV space. When the needle was considered to be in the TPV space, 1 ml of a dye solution [methylene blue (Methylene blue powder; Sigmae Aldrich Inc., MO, USA) diluted in 2% lidocaine (Lidocaine 2%; Hospira, Inc., IL, USA)] was injected. Figure 1 Ultrasound-guided approach to the thoracic paravertebral (TPV) space in dogs. The transducer is positioned on the dorsal aspect of the thoracic region and oriented in an oblique direction to the dorsal midline (dotted line). The needle marks the fifth thoracic spinous process (T 5 ). A Tuohy needle is advanced in an in-plane technique toward the TPV space. 637

3 Phase II: ultrasound-guided injection and dye spread evaluation Six adult dog cadavers of various breeds with a mean (range) weight of 20 (11e21) kg were used. Two injectate solutions were prepared using 1 ml of permanent tissue marking dye (Davison Marking System; Bradley Products, Inc., MN, USA) diluted in 50 ml of 2% lidocaine; one solution was dyed orange and the other yellow. On each cadaver, one fifth TPV space was injected with 0.05 ml kg e1 [low-volume treatment (LV)] of yellow injectate and the other fifth TPV space was injected with 0.15 ml kg e1 [high-volume treatment (HV)] of orange injectate. The LV and HV treatments were randomly assigned to the left side using a closed envelope technique. Subsequently, the contralateral side was injected with the alternative volume. Ultrasound-guided TPV injection Dogs were positioned in sternal recumbency, and the entire thorax was clipped. The thoracic spinous processes were identified by palpation starting at the first (T 1 ) thoracic vertebra. A 20 gauge hypodermic needle was pinned to the dorsal aspect of the fifth thoracic spinous process and used as an anatomic reference during subsequent dissection. The ultrasound transducer was first placed longitudinally close to the dorsal midline and gradually glided laterally until two consecutive transverse processes were included in the ultrasound field of view. Thereafter, the transducer was rotated to a more oblique orientation to include the fifth thoracic transverse process and the sixth rib within the field of view (Fig. 2). Before each injection, the following echographic features were recognized: 1) the fifth thoracic transverse process and sixth rib; 2) the parietal pleura; 3) the internal intercostal membrane, and 4) the TPV space (Fig. 2). Additionally, recognition of the thoracic spinal nerve was attempted. A 20 gauge, 9 cm Tuohy needle connected to a T- port (T Conn; Smith Medical Pty, Ltd, NSW, Australia) was filled with dye solution and inserted using an ultrasound-guided in-plane approach toward the fifth TPV space (Fig. 1). When the tip of the needle was located between the internal intercostal membrane and parietal pleura, the assigned volume of dye solution was injected. The entire procedure was first performed on one side (left) and then repeated on the contralateral (right) side. The echographic visualization of the needle was evaluated as follows: 1) good (the shaft and needle tip can be visualized completely); 2) poor (the tip of the needle is recognized only after gentle in-and-out movements or jiggling of the needle; or 3) absent (the tip of the needle cannot be discerned). The occurrence of ventral displacement of the pleura during injection was recorded. Dissection With the dog in sternal recumbency, the first longitudinal skin incision was made at the dorsal thoracic midline, between the 1st (T 1 ) and 12th (T 12 ) thoracic Figure 2 Ultrasound image of the thoracic paravertebral (TPV) space in a dog between the fifth transverse process (TP) and the sixth rib. 1, TPV space; 2, internal intercostal membrane; 3, parietal pleura; 4, external intercostal muscle; 5, intermuscular plane between the levatores costarum and external intercostal muscles; 6, levatores costarum muscle; Cd, caudal; Cr, cranial; D, dorsal; V, ventral. 638

4 vertebrae. The skin was reflected laterally and the epaxial muscles were exposed and removed. Subsequently, the serratus dorsalis cranialis, spinalis, semispinalis, longissimus thoracis and iliocostalis muscles were dissected and removed. The levatores costarum and the external intercostal muscle of the fifth intercostal space were detached from their caudal insertion on the cranial margin of the sixth rib and folded cranially to expose the internal intercostal muscle. The internal intercostal muscle and the internal intercostal membrane were detached from their insertions to expose the fifth spinal nerve. The staining of each fifth spinal nerve was graded (Appendix A). Subsequently, the contiguous TPV spaces were dissected and the spinal nerves were exposed, as described previously. The same dissection procedure was repeated on the contralateral side of the thorax. When all the TPV spaces from T 2 tot 8 were dissected, the ribs were cut using an osteotome at the costochondral junctions and the pleural cavity was subsequently exposed. The pleural cavity and mediastinum were examined for the presence or absence of dye solution. The sympathetic trunk was also identified and observed for dye staining. Finally, a dorsal laminectomy was performed to identify any epidural migration of injectate. Statistical analysis Data were evaluated for normality using the Shapiro ewilk test. Fisher s exact test was used to test the significance of differences between injectate volumes in the presence or absence (yes/no) of multisegmental distribution of dye and staining of the sympathetic trunk. The ManneWhitney test was used to establish the significance of differences between injectate volumes in the grade of staining on each fifth thoracic spinal nerve. Differences that achieved a p value of 0.05 were considered significant. Results are presented as the median (range). Statistical analyses were performed using GraphPad Prism Version 6.0 (GraphPad Software, Inc., CA, USA) and Statistix Version 9.0 (Analytical Software, FL, USA). Results Phase I Anatomic description The levatores costarum muscles and external intercostal muscle were exposed after the removal of the transversospinalis and longissimus muscle systems. The levatores costarum muscles partially covered the origin of the external intercostal muscle. The aponeurosis of the external intercostal muscle extended from the transverse process to the cranial border of the caudal rib. After its removal, the internal intercostal muscle was exposed. In a lateral to medial direction, the internal intercostal muscle appeared to be fused into a thin membrane, named the internal intercostal membrane, which covered the spinal nerve at the level of the TPV space and formed the roof of the TPV space. The internal intercostal membrane was attached medially to the vertebral body. Here, the ventral branch of the spinal nerve was interposed between the internal intercostal membrane and some residuary deeper fibers of the internal intercostal muscle. The deeper fibers of the internal intercostal muscle were attached to the vertebral body. The endothoracic fascia divided the TPV space into dorsal and ventral compartments. The dorsal compartment was delimited dorsally by the internal intercostal membrane and ventrally by the endothoracic fascia and contained the thoracic spinal nerve and the origin of the rami communicantes nerves. The ventral compartment of the TPV space was delimited by the endothoracic fascia dorsally and the parietal pleura ventrally and contained the rami communicantes and the sympathetic trunk. The rami communicantes passed through the endothoracic fascia. Lateral to the TPV space, the endothoracic fascia was in close contact with the parietal pleura. At this level, the spinal nerve was seen between the internal intercostal muscle and endothoracic fascia on the caudal margin of the respective rib where it became the intercostal nerve. Ultrasonographic description Deep to the epaxial muscles, the transverse processes were observed as hyperechoic structures with a typical acoustic shadow underneath (Fig. 2). When the transducer was glided laterally, the ribs were also recognized as contiguous hyperechoic structures. The parietal pleura was recognized as a hyperechoic line (Fig. 2). A thin hyperechoic line connected the transverse process with the cranial border of the caudal rib and corresponded with the intermuscular plane formed by the levatores costarum muscle and the external intercostal muscle (Fig. 2). The internal intercostal membrane was visualized running in an oblique direction as a thin hyperechoic line between the 639

5 intercostal muscles and the parietal pleura (Fig. 2). Therefore, the echographic TPV space was defined as the space contained between the internal intercostal membrane and the parietal pleura (Fig. 2). In these dogs, the TPV space was located at a depth of 4e5cm beneath the skin. Phase II Dye injection Twelve injections were performed. The HV treatment was performed on the left side in four dogs and on the right side in two dogs, whereas the LV treatment was performed on the left side in two dogs and on the right side in four dogs. All of the echographic landmarks described above were visualized in all 12 injections. The fifth thoracic spinal nerve was identified bilaterally in two dogs and unilaterally in two dogs (one left side and one right side). Ultrasound visualization of the needle tip in the TPV space achieved scores of 1 in eight injections and 2 in four injections. Ventral displacement of the pleura upon injection was observed in 11 of 12 injections. Dissection Six of six injections in the HV treatment resulted in Grade 2 staining of the fifth thoracic nerve. In the LV condition, five injections resulted in Grade 2 staining and one resulted in Grade 1 staining of the nerve. No thoracic spinal nerve other than T 5 was stained (Fig. 3). No significant differences between the HV and LV treatments were found in the incidence of multisegmental dye migration (p > 0.9). Multisegmental spread was found between the parietal pleura and the endothoracic fascia (Fig. 4a), affecting the sympathetic trunk on several spinal segments (Fig. 4b). The sympathetic trunk was stained for a median (range) of 3.5 (1e6) and 3 (0e3) vertebral bodies in the HV and LV conditions, respectively (p ¼ 0.3) (Fig. 5). In all cases, the dye solution was observed in the dorsal mediastinum, dorsal to the large blood vessels, and neither reached the base of the heart nor crossed the parietal pleura. No evidence of intrapleural or epidural migration was found for any injection. Discussion The present study shows that the anatomic structures defining the TPV space can be identified ultrasonographically using a high-frequency, linear-array transducer in dogs and that a needle can be visualized and advanced into this space using an in-plane technique. Moreover, the tested volumes injected into the fifth TPV space consistently produced staining of the respective spinal nerve; however, neither of the tested volumes stained the contiguous spinal nerves. TPV blocks have been used in human anesthesia for nearly 100 years to provide high-quality perioperative pain relief in surgery involving the thorax and the cranial abdomen (Richardson & L onnqvist Figure 3 Dorsolateral view of the fourth to seventh intercostal spaces in a dog. The epaxial muscle, the levatores costarum, and the external and internal intercostal muscles have been removed. Yellow dye solution can be seen in direct contact with the fifth thoracic spinal nerve. Cd, caudal; Cr, cranial. 640

6 Figure 4 Dissection of the T 5 and T 6 thoracic paravertebral (TPV) spaces in a dog after injection of yellow dye using an ultrasound-guided TPV approach at the level of T 5. (a) The dye solution can be seen at T 6, deep to the endothoracic fascia and not in contact with the spinal nerve. *The deeper residuary fibers of the internal intercostal muscle were separated to expose the endothoracic fascia. (b) Intrathoracic view of the ventral aspect of the vertebral bodies of T 4 to T 7 after the thorax was opened. The dye solution is seen to be contained between the parietal pleura and the endothoracic fascia, in direct contact with the sympathetic trunk. Cd, caudal; Cr, cranial. 1998). In humans, several techniques for approaching the TPV space have been described. Loss of resistance to air injection as the tip of the needle crosses the costo-transverse ligament has been used for this purpose (Eason & Wyatt 1979). However, this technique is considered unreliable because it is subjective (Richardson & L onnqvist 1998). Electrostimulation, in which a square-wave electrical current depolarizes the spinal nerve and elicits the contraction of the corresponding intercostal muscle, has also been used. This technique is claimed to reduce subjectivity in humans (Wheeler 2001; Naja et al. 2007) and dogs (Portela et al. 2012). Using electrolocation, Portela et al. (2012) reported sensitivity of 88.2% (true positives) and a success rate of 83.3% when injecting radiographic contrast solution close to the intervertebral foramen in dogs. However, the same authors reported a 20% incidence of inadvertent intrapleural injection, which suggests that penetration of the parietal pleura is a likely 641

7 Figure 5 Spread characteristics of a high (0.15 ml kg e1 ) or low (0.05 ml kg e1 ) volume of a dye solution injected at the fifth thoracic paravertebral (TPV) space using an ultrasound-guided TPV approach in six dogs. The columns represent staining of the sympathetic trunk. The circles represent the spinal levels of thoracic spinal nerves dyed. complication of the use of electrolocation to achieve a TPV block in dogs (Portela et al. 2012). Ultrasound guidance in TPV blocks has been described and used in humans to increase the rate of success of TPV blocks and to decrease the incidence of unintended injections outside the TPV space (O Riain et al. 2010; Krediet et al. 2015). The present study demonstrates that the echographic anatomy of the TPV space in dogs is broadly similar to that in humans, except that the thoracic spinal processes and epaxial muscles are relatively tall in some dogs. This anatomic characteristic requires the needle to travel deeply and at a steep angle relative to the ultrasound transducer in order to approach the TPV space at this level. High-frequency, linear-array ultrasound transducers produce highresolution images but provide relatively low penetration; therefore, the deep structures may be difficult or impossible to recognize. In the present study, all the TPV spaces were < 6 cm deep, but the tip of the needle was poorly visualized in 4 of the 12 injections. Tissue disturbance produced by needle jiggling helped distinguish the tip of the needle when tip visualization was suboptimal. The deep location of the TPV space and the narrow window between the transverse process and the adjacent rib can greatly limit the path in which the needle can be advanced and yet be visible under the ultrasound beam (double-fulcrum effect). In order to resolve this problem, a modification of the parasagittal longitudinal approach has been described in which the target TPV space is offset from the center of the ultrasound field to allow the needle to penetrate at a deeper angle (Abdallah & Brull 2014). Under these circumstances, a microconvex, high-frequency transducer may be a better choice than the linear array used in the present study. To preserve the echographic anatomy and yet use an in-plane technique, a slight oblique orientation of the ultrasound transducer was applied. With this small modification, the transverse process of T 5 and the sixth rib could be clearly identified in the same field of view. Moreover, the parietal pleura could be seen as a bright, thick hyperechoic line between these two bony structures. The internal intercostal membrane, which is intimately attached to the internal intercostal muscle, could be recognized as a thin hyperechoic line between the parietal pleura and the intercostal muscles. In order to approach the TPV space, the needle was advanced through the skin, the epaxial muscles, levatores costarum muscles, the external intercostal muscle and the internal intercostal muscle/internal intercostal membrane. When the tip of the Tuohy needle pierced the internal intercostal membrane, a subtle pop could be felt. The injectate should be deployed when the tip of the needle is located between the internal intercostal membrane and the parietal pleura. Ventral displacement of the pleura during injection has been 642

8 suggested to confirm that the injection is being delivered in the correct place (O Riain et al. 2010). In the study reported here, ventral displacement of the pleura was observed in 11 of the 12 injections; the injection was deemed completely successful (Grade 2) in these 11 injections. In the single instance in which displacement was not evident, dissection showed that the nerve was only partially stained (Grade 1). A greater number of cases are needed to confirm that ventral displacement of the pleura during injection is a reliable predictor of successful injection. The volumes tested in the present study appeared sufficient to stain the target spinal nerve. No significant differences emerged between the high and low volumes in terms of staining quality. However, no dye solution was found to affect the contiguous thoracic spinal nerves at either volume used, which suggests that migration to contiguous spinal nerves is unlikely with this approach after a single injection. The apparent absence of the spread of injectate to contiguous spinal nerves in dogs contrasts with the findings of studies in humans, which report the occurrence of multisegmental distribution and analgesia following single injections (Marhofer et al. 2013). Cadaveric studies conducted in humans also show the presence of dye solution in contiguous TPV spaces following single injections (Cowie et al. 2010; Albokrinov & Fesenko 2014). In those studies, the pattern of distribution was always evaluated from inside the thoracic cavity. A similar pattern of distribution was observed in the present study when the thoracic cavity was opened. Interestingly, when the TPV spaces were dissected from the dorsal aspect, leaving the parietal pleura and the endothoracic fascia in situ, the contiguous spinal nerves were not found to be affected by the dye solution. Unfortunately, in the aforementioned human studies, this pattern of distribution was always evaluated using imaging or by dissecting the TPV spaces from inside the thorax, which disrupts the endothoracic fascia and makes it impossible to assess whether the dye had migrated as a result of multisegmental distribution or as a result of the dissection itself. Naja et al. (2004) suggested that the endothoracic fascia seemed to play an important role in the pattern of distribution of solutions injected into the TPV space and reported that injections performed superficial to the endothoracic fascia produced a localized spread of injectate, but injections performed deep to the endothoracic fascia elicited a multisegmental spread. As previously mentioned, the endothoracic fascia divides the paravertebral space into a dorsal and a ventral compartment (Stopar Pintaric et al. 2012). The ventral compartment of the TPV space communicates with the contiguous ventral compartments of the TPV space and with the dorsal mediastinum. In the study presented here, it is likely that the injection disrupted the integrity of the endothoracic fascia; therefore, it is possible that the dye deposited at the target spinal nerve may have migrated toward adjacent thoracic spaces following only the ventral compartment of the TPV space. In fact, when the contiguous TPV spaces were dissected, the dye solution was seen to be contained between the endothoracic fascia and parietal pleura and not in direct contact with the contiguous spinal nerves. Laterally, away from the neuraxis, the parietal pleura is in close proximity to the endothoracic fascia. The dye solution was able to spread laterally between these two thin layers. The intercostal nerves are located deep to the internal intercostal muscle and are separated from the parietal pleura by the endothoracic fascia. Therefore, the dye solution was never in direct contact with the intercostal nerves. However, although no direct contact was observed between the injected solution and the intercostal nerves, it is possible that a local anesthetic solution might diffuse through the endothoracic fascia, affecting the intercostal nerves. This diffusion may explain the multisegmental analgesic effect reported in human volunteers (Marhofer et al. 2013). It is important to note that some injectate was found in the dorsal mediastinum between the parietal pleura and endothoracic fascia. This suggests that if excessive volumes were to be injected at this location, the excess volume might take the path of least resistance into the mediastinum. Because the side effects of mediastinal injection are undefined, the current group proposes that a volume of 0.05 ml kg e1 of local anesthetic should be sufficient to produce a reliable blockade of a single thoracic spinal nerve. Moreover, this volume is in agreement with the findings of a previous study in which 0.05 ml kg e1 was suggested when nerve stimulation was utilized to locate the thoracic spinal nerve (Portela et al. 2012). However, the clinical utility of this volume is yet to be confirmed. In the present study, no intrapleural migration was observed. A 20% incidence of intrapleural injection was recorded when nerve stimulation was used to approach the TPV space (Portela et al. 2012). The absence of this complication in the present study may 643

9 have resulted from the use of ultrasound guidance to enable visualization of the parietal pleura during injection, which will have reduced the likelihood of pleural puncture. Incidentally, the results identified the migration of dye into the dorsal mediastinum and staining of the sympathetic nerve trunk in several spinal segments. There was no significant difference between the highand low-volume injections in the extent of this migration. The clinical relevance of unilateral or multisegmental sympathetic trunk blockade must be evaluated. The present authors speculate that this blockade may provide some visceral analgesia in dogs undergoing intrathoracic surgery. In conclusion, the in-plane, parasagittal, oblique ultrasound-guided TPV block may potentially represent a reliable and accurate technique for delivering regional anesthesia in dogs. The present results suggest that a volume of 0.05 ml kg e1 per nerve will provide sufficient distribution of local anesthetic to result in blockade of the target spinal nerve. However, clinical studies will be necessary to determine the number of spinal segments that should be blocked to provide adequate analgesia in a dog undergoing an intercostal thoracotomy. It appears that the endothoracic fascia plays an important role in facilitating the multisegmental spread of the injectate and blockade of the sympathetic nerve trunk. Acknowledgements The authors would like to thank Pamela Schenck (Department of Biomedical Sciences, College of Veterinary Medicine, Cornell University) for providing the cadavers used in this study. Authors' contributions DAP, LC and PEO: conceived and participated in the design of the study and data analysis. DAP and LC: participated in the execution of the study and data collection. DAP, LC, MM-F and RDG: contributed to the writing of the manuscript. Conflict of interest statement This manuscript has not been previously published and is not under consideration for publication elsewhere. The authors of this typescript meet to the journal s criteria for authorship as outlined in the instructions for publication enclosed in the journal. The authors whose names are listed certify that they have NO affiliations with or involvement in any organization or entity with any financial interest (such as honoraria; educational grants; participation in speakers bureaus; membership, employment, consultancies, stock ownership, or other equity interest; and expert testimony or patentlicensing arrangements), or non-financial interest (such as personal or professional relationships, affiliations, knowledge or beliefs) in the subject matter or materials discussed in this manuscript. References Abdallah FW, Brull R (2014) Off side! A simple modification to the parasagittal in-plane approach for paravertebral block. Reg Anesth Pain Med 39, 240e242. Albokrinov AA, Fesenko UA (2014) Spread of dye after single thoracolumbar paravertebral injection in infants. A cadaveric study. Eur J Anaesthesiol 31, 305e309. Cowie B, McGlade D, Ivanusic J, Barrington MJ (2010) Ultrasound-guided thoracic paravertebral blockade: a cadaveric study. Anesth Analg 110, 1735e1739. Davies RG, Myles PS, Graham JM (2006) A comparison of the analgesic efficacy and side-effects of paravertebral vs epidural blockade for thoracotomy e a systematic review and meta-analysis of randomized trials. Br J Anaesth 96, 418e426. Eason MJ, Wyatt R (1979) Paravertebral thoracic block e a reappraisal. Anaesthesia 34, 638e642. Karmakar MK (2001) Thoracic paravertebral block. Anesthesiology 95, 771e780. Krediet AC, Moayeri N, van Geffen GJ et al. (2015) Different approaches to ultrasound-guided thoracic paravertebral block: an illustrated review. Anesthesiology 123, 459e474. Marhofer D, Marhofer P, Kettner SC et al. (2013) Magnetic resonance imaging analysis of the spread of local anesthetic solution after ultrasound-guided lateral thoracic paravertebral blockade: a volunteer study. Anesthesiology 118, 1106e1112. Naja MZ, Ziade MF, El Rajab M et al. (2004) Varying anatomical injection points within the thoracic paravertebral space: effect on spread of solution and nerve blockade. Anaesthesia 59, 459e463. Naja ZM, Al-Tannir MA, Zeidan A et al. (2007) Nerve stimulator-guided repetitive paravertebral block for thoracic myofascial pain syndrome. Pain Pract 7, 348e351. O Riain SC, Donnell BO, Cuffe T et al. (2010) Thoracic paravertebral block using real-time ultrasound guidance. Anesth Analg 110, 248e251. Portela DA, Otero PE, Sclocco M et al. (2012) Anatomical and radiological study of the thoracic paravertebral space in dogs: iohexol distribution pattern and use of the nerve stimulator. Vet Anaesth Analg 39, 398e408. Richardson J, L onnqvist P (1998) Thoracic paravertebral block. Br J Anaesth 81, 230e

10 Saito T, Den S, Cheema SP et al. (2001) A single-injection, multi-segmental paravertebral block-extension of somatosensory and sympathetic block in volunteers. Acta Anaesthesiol Scand 45, 30e33. Stopar Pintaric T, Veranic P, Hadzic A et al. (2012) Electron-microscopic imaging of endothoracic fascia in the thoracic paravertebral space in rats. Reg Anesth Pain Med 37, 215e218. Wheeler LJ (2001) Peripheral nerve stimulation endpoint for thoracic paravertebral block. Br J Anaesth 86, 598e599. Received 28 March 2016; accepted 21 May Available online 16 February 2017 Appendix A. Grading system for the distribution of dye staining on fifth thoracic spinal nerves after ultrasound-guided injection of a permanent tissue marking dye solution into the fifth thoracic paravertebral space. Grade Description 0 Fail: nerve free of dye. 1 Partial: the nerve was stained for less than 1 cm length or the dye was not affecting the entire circumference of the nerve. 2 Complete: the entire circumference of the nerve was dyed for at least 1 cm of its length. 645