Paravertebral blocks have gained in popularity recently with two meta-analyses comparing paravertebral with epidural analgesia. Results of the analyses suggest analgesic equivalence but fewer adverse effects, such as hypotension, nausea and vomiting, urinary retention, and pulmonary complications.1–3 Thus, continuous paravertebral blocks remain an attractive alternative to epidural blocks; however, the optimal approach to paravertebral analgesia is yet to be defined, and many modifications have been described, especially after the advent of ultrasound guidance.

Several techniques have been reported, including loss of resistance by passage through the costotransverse ligament, “walking off” the transverse process, an intercostal approach, nerve stimulation, and placement under direct vision at the time of surgery.1,4–7 It is noteworthy that, with blind techniques, the loss of resistance felt with paravertebral needle placement is much less definite than that with epidural insertion.8 The advent of ultrasound-guided regional anesthesia has offered the potential to improve efficacy and safety via real-time visualization of the paravertebral space, the surrounding structures, and the approaching needle.1 As such, ultrasound has been applied to the traditional paravertebral approaches in an effort to reduce the risk of pleural puncture as well as to ensure better delivery of medications into the paravertebral space. Even with the use of ultrasound, however, the sonographic windows are limited and often difficult to obtain, images are rarely optimal, and the paravertebral space is often small, with needle placement in immediate proximity to the pleura. As such, the risk for pleural puncture, especially with less skilled hands, is still present.

At our institution, we encounter a large volume of patients with rib-related pain resulting from traumatic rib fractures or thoracic surgery. These patients often present with severely compromised pulmonary function and with abnormal coagulation profiles; they would clearly benefit from thoracic regional analgesia, yet the concern for pleural injury and/or bleeding complications may be substantial. Confronting this common dilemma, we began to explore techniques of paravertebral block that might potentially reduce the risk of lung injury, especially in high-risk cases, and possibly offer a decreased risk of either bleeding or significant bleeding consequences.

Recently, a simplified “blind” paravertebral lamina approach has been described by Pfeiffer, followed by a study in mastectomy patients by Jüttner et al.9,10 This approach would logically offer the advantage of lower risk of pleural injury, while a possible inadvertent epidural injection remains a concern. Combining this technique with ultrasound-guidance can facilitate easy identification of the lamina and thereby minimize the risk of epidural injection associated with the blind technique.

Herein, we describe the ultrasound-guided retrolaminar block through our early experience with this approach in three patients. This approach is proposed as a simplified, easy, fast, effective, and safe alternative to other described paravertebral analgesia techniques. Each of the three patients gave written informed consent for publication of the anonymized details related to their cases.

Technique

The thoracic paravertebral space is triangular in shape. It is delineated medially by the vertebral body, the intervertebral disks, and the intervertebral foramina; posteriorly by the superior costotransverse ligament, the transverse process, and the ribs; and anteriorly by the parietal pleura.1 Injection of local anesthetic in this space typically results in unilateral block of several spinal nerves as well as variable unilateral sympathetic nerve blockade.1,11 Analgesia and spread of local anesthetic can be inconsistent, and success rates are typically reported to be less than 90%.12–14

The traditional “blind” approach to the paravertebral space involves tactile identification of a transverse process with the tip of the block needle and then “walking the needle off” in either the cephalad or caudad direction about 1 cm deeper or until a “pop” is felt penetrating through the costotransverse ligament (Fig. 1). With the paravertebral lamina approach, contact with the vertebral lamina is sought as the end point of needle localization9,10 (Fig. 1). Jüttner and Pfeiffer hypothesized that local anesthetic can easily gain access to the paravertebral space through the porous costotransverse ligament when the needle tip is in this lamina position. Since the costotranverse ligament is unlikely to be a porous structure, we describe below our view regarding how local anesthetic solution can gain access to the paravertebral space.

Fig. 1
figure 1

Schematic of retrolaminar block vs traditional paravertebral techniques. The left side of the figure illustrates the retrolaminar block needle placement and presumed distribution of local anesthetic in the paravertebral space. The right side of the figure illustrates a traditional paravertebral block needle placement and distribution of local anesthetics. (Image modified with permission from Visible Human Web Server (http://visiblehuman.epfl.ch) courtesy of Prof. R.D. Hersch, Ecole Polytechnique Fédérale de Lausanne (EPFL), Lausanne, Switzerland.)

Full size image

The ultrasound-guided retrolaminar block is a modification of the lamina approach.9,10 The suggested ultrasound transducer orientation is sagittal paramedian (cephalad-caudad direction), as shown in Fig. 2. The choice of ultrasound probe is based on body habitus and anatomy as well as available equipment, with low-frequency curvilinear transducers recommended when sonographic landmarks and target structures are visualized deeper than 5 cm and high-frequency (> 9 MHz) linear probes recommended when target structures are less than 5 cm deep. A transverse plane orientation of the probe (lateral to medial) could also be considered, though the authors appreciate that predictably avoiding inadvertent epidural injection may be more difficult (Fig. 3). We recommend in-plane visualization of the advancing needle tip. We begin our ultrasound scanning in a paramedian sagittal plane by finding the ribs 5-6 cm lateral to the spinous processes, corresponding with the dermatomal levels of interest. The ribs in this view normally have a round contour appearance, with an appreciable pleural line located in between and typically less than 1 cm deeper. Normal sonographic pleural images are observed in the absence of pneumothorax, pleural effusion, or other pathology. Sliding the scanning probe from lateral to medial, the transverse processes are identified next. The contour of the transverse processes usually appears more rectangular than that of the ribs; it is common to see a “step down” change with lateral to medial scanning when the transverse processes come into view from the ribs. Additionally, the pleural line seen between the transverse processes is deeper and often less defined than that in the rib view. Moreover, in the transverse process view, the costotransverse ligament can be visualized above the paravertebral space. Continuing to scan from lateral to medial, the vertebral laminae come into view. They are visualized as continuous flat hyperechoic structures with regularly interposed small “notches” which represent the laminar interfaces and facet joint areas (please see online supplemental video showing the process in reduced speed real-time scanning). Needle entry begins in plane, caudad, or cephalad until contact with the lamina is achieved under real-time ultrasound guidance. The local anesthetic is then injected with intermittent aspiration, while if resistance is encountered, the needle is retracted by 1 mm and injection is attempted again. The spread of the injectate along the plane created between the lamina and the deep paraspinous muscles is observed and optimized, followed by catheter insertion when continuous analgesia is indicated.

Fig. 2
figure 2

Ultrasonographic and skin surface anatomy landmarks for the vertical probe orientation (recommended) in-plane retrolaminar block. Picture I shows how the probe is held for orientation. Counting from the spinous process of thoracic level 1 to level 5, Picture II shows A) the surface marking of the beginning of the ribs, B) the transverse processes, and C) the lamina from the surface anatomic view. Picture A correlates with marking A on Picture II and shows the ultrasonography view of the ribs. Picture B shows the ultrasonographic view of the transverse processes. Picture C shows the ultrasonographic view of the lamina. Picture D shows the entry of a needle, under ultrasound guidance, coming into contact with the lamina. TP = transverse process; CTL = costotransverse ligament

Full size image
Fig. 3
figure 3

Ultrasonographic and skin surface anatomy landmarks for a possible horizontal probe orientation in-plane retrolaminar block. Picture A shows the probe orientation for a horizontal approach to the retrolaminar block. Picture B shows the spinous process, lamina, and transverse process under ultrasound imaging. SP = spinous process; TP = transverse process

Full size image

Following are clinical case descriptions describing our use of this technique. They represent the first three consecutive patients with multiple rib fractures who underwent this block at our centre.

Case descriptions

Case 1

A 59-yr-old obese male with a body mass index (BMI) of 33 presented with right-sided chest wall pain after a motor vehicle accident while being intoxicated with alcohol and cannabis. His medical history included coronary artery disease, status post four-vessel coronary artery bypass grafting, hypertension, hyperlipidemia, depression, and benign prostate hyperplasia. He had no known drug allergies. His medications at home included acetylsalicylic acid, lisinopril, metoprolol, citalopram, and finasteride. His computed tomographic (CT) scan showed minimally displaced fractures of the lateral aspect of the right third through right seventh rib without underlying pulmonary contusions, liver parenchymal contusion with possible extravasation, and perisplenic fluid suggestive of hematoma without parenchymal injury. Complete metabolic panels, complete blood counts, and coagulation labs were within normal limits.

Management of his pain was initially attempted with hydromorphone intravenous patient-controlled analgesia (PCA) and lidocaine transdermal patches, but he continued to splint with movement and coughing. Right-sided retrolaminar paravertebral block was subsequently performed on hospital day two. With the patient in the sitting position and under aseptic conditions, the block was performed at the level of the fifth thoracic (T) vertebra under ultrasound guidance using an in-plane approach. Analgesia was achieved with a 0.5% solution of ropivacaine 30 mL, and a catheter was subsequently placed caudad 5 cm beyond the needle tip (needle entry at the cephalad end of the ultrasound probe, aiming caudally). The patient’s pain was satisfactorily controlled with a continuous infusion of 0.125% bupivacaine at a rate of 12 mL·hr−1, and his Dilaudid PCA and lidocaine transdermal patches were discontinued on the same day. Throughout hospitalization, the patient required once daily manual boluses of 0.125% bupivacaine 10-15 mL for breakthrough pain. On hospital day five, the patient’s retrolaminar block catheter was removed, at which point he received satisfactory pain control with oral oxycodone and lidocaine patches. He was discharged the same day.

The patient’s pre-procedure verbal rating scale (VRS) pain score was a 10/10. Post-block assessment two to six hours after placement showed the patient had a 3/10 VRS at rest and 4/10 with activity with physical therapy. At 24 hr, 38 hr, 43 hr, 45 hr, and 48 hr, VRS pain scores were 5/10, 5/10, 5/10, 4/10, 4/10, respectively.

Case 2

A 40-yr-old male with a BMI of 31 presented with multiple right-sided non-displaced rib fractures after suffering a motor vehicle accident secondary to intoxication. His medical history included anxiety. He had no known drug allergies and was not taking any medications. He had a 30 pack year smoking history. His chest CT imaging showed right first to seventh posterior rib fractures without evidence of neuroforaminal extension or pre-vertebral soft tissue swelling, mildly displaced fractures of the right transverse processes of cervical vertebral levels six through T2, and a minimally displaced comminuted left scapular fracture.

His pain was initially managed with intravenous hydromorphone-based PCA, lidocaine patch, acetaminophen as needed, and ketorolac as needed, but he continued to splint with movement and coughing. Retrolaminar paravertebral blocks were subsequently performed on hospital day two with the patient in the sitting position. The blocks were performed under ultrasound guidance at the right T3 and right T6 levels. Anesthesia was achieved with 0.5% ropivacaine 20 mL at each level, and catheters were subsequently placed caudad 5 cm beyond the needle tip (needle entry at the cephalad end of the ultrasound probe, aiming caudad). The patient’s pain was subsequently controlled with a continuous infusion of 0.125% bupivacaine at 10 mL·hr−1, and his hydromorphone PCA and lidocaine patches were discontinued on the same day. On hospital day four, the retrolaminar block catheters were removed. At this point, the patient’s pain was satisfactorily controlled with oxycodone, ketorolac, acetaminophen, and lidocaine patches. He was subsequently discharged to home on hospital day six.

The patient’s pre-procedure pain score was a 10/10 (VRS). Post-procedure VRS pain scores were 1-2/10, 7/10, 5/10, and 0/10 at one hour, 21 hr, 44 hr, and 64 hr, respectively.

Case 3

A 38-yr-old female with a BMI of 25 presented with right-sided rib fractures 1-12 (except 10 and 11), right clavicular and left first rib fracture, multiple facial fractures, a small subarachnoid hemorrhage, and bilateral sacral ala fractures secondary to a motor vehicle accident. Her medical history included polysubstance abuse and bipolar disorder, and she also had a significant smoking history. The patient had no known drug allergies and was taking lorazepam and Adderall. Computed tomography imaging showed the fractures listed above.

Her pain could not be successfully managed with a hydromorphone PCA with acetaminophen and ketorolac, hence, retrolaminar paravertebral blocks were subsequently performed with the patient in the sitting position. The blocks were performed under ultrasound guidance at the right T5 and T8 levels, and catheters were subsequently placed cephalad 5 cm beyond the needle tip (needle entry at the caudal end of the ultrasound probe, aiming cephalad). Anesthesia was achieved with 0.5% ropivacaine 20 mL followed by a continuous infusion of 0.125% bupivacaine at 10 mL·hr−1. Her hydromorphone PCA was continued for her other injuries, but its use was needed only sparingly after the blocks. On hospital day seven, the block catheters were removed. At this point, the patient’s pain was satisfactorily controlled with oxycodone, ketorolac, and acetaminophen. She was subsequently discharged to a rehabilitation centre on hospital day nine.

The patient’s pre-procedure pain score was 10/10 (VRS). Post-block VRS pain scores two to six hours after the procedure were 2/10 at rest and 4/10 with coughing. At 32 hr, 45 hr, 48 hr, and 52 hr, VRS pain scores were 3/10, 4/10, 2/10, and 2/10, respectively.

Please refer to the Table for a summary of clinical details for the three patients.

Table Summary of the three patients in this case series
Full size table

Discussion

The paravertebral space is not an isolated anatomical compartment but communicates with the adjacent intercostal and epidural spaces.1 In 2006, Pfeiffer et al. postulated that needle placement immediately posterior to the vertebral lamina readily allows access of administered local anesthetic into the paravertebral space. Anatomically, the paravertebral space is a triangular space positioned between the head and neck of the ribs. Its posterior wall is formed by the superior costotransverse ligament, the anterolateral wall is the parietal pleura, and the medial wall is the lateral surface of the vertebral body and disk.15 The internal intercostal membrane is the medial extension of the internal intercostal muscle and is continuous medially with the superior costotransverse ligament.10 The paravertebral space is probably best viewed as an epidural-paravertebral intercostal anatomic complex, as an injection into this complex can result in a number of clinical expressions of the paravertebral block, as has been shown previously.1,14,16–18 It is probably for this reason that dermatomal mapping in clinical studies on the paravertebral block is unable to discern the spread of the local anesthetic as being within the paravertebral, intercostal, or unilateral epidural space.15

The endothoracic fascia divides the paravertebral space into two compartments: one anterolateral (extrapleural compartment [EPC]) and the other posteromedial (subendothoracic compartment [SETC]). The spinal nerves with their ganglia are found in the SETC, whereas the sympathetic ganglia are consistently located within the EPC.15 The anatomic characteristics and position of the endothoracic fascia may influence the spread of the local anesthetics during the paravertebral block, although it may not completely limit the spread of solutions on its inner or outer side, as is the case with true fascial investments.15,19 On the contrary, injections made dorsal to the endothoracic fascia result in a cloud-like spreading pattern with only limited distribution over adjacent segments but with a greater chance for spread of the injectate into the epidural space.15

The concept that local anesthetic can penetrate the paravertebral space from a laminar injection challenges the classical teaching that the paravertebral space is defined posteriorly as a closed space by the costotransverse ligament. It is possible that the local anesthetic trickles through the medial aperture of the superior costotransverse ligament where the dorsal ramus of the spinal nerve exits posteriorly to innervate the paraspinal muscles. It is also possible that the fluid tracks anteriorly through the looser tissues just lateral to the facet joints.

The major advantage of this technique, if performed properly, is minimizing or even eliminating the risk of pneumothorax.9 Additionally, the risks of nerve root damage and inadvertent injection into a dural sleeve, an intervertebral foramen, or the epidural or intrathecal spaces should also be decreased. Furthermore, this technique requires less skill, as the exact identification of the paravertebral space, a task often challenging even in experienced hands, is not needed. Nevertheless, using a “blind” approach to perform this procedure leaves significant room for potentially dangerous needle misplacement. We point out that severe complications can be as high as 5% with the blind traditional paravertebral technique.7,20

In this report, we suggest that utilizing ultrasound-guidance will further simplify and significantly improve the lamina approach described by Pfeiffer. Ultrasound imaging of the vertebral lamina is easy and reproducible. With an “in-plane” approach, real-time visualization of the advancement of the block needle to bone contact, the spread of the injectate, and placement of the catheter are relatively simple and predictable tasks. Since the needle does not physically enter the paravertebral space, we suggest the name ultrasound-guided retrolaminar block (URB).

This new technique was introduced to clinical routine at our institution by the authors. Since its introduction, we have treated more than 20 patients undergoing a wide variety of surgeries or thoracic trauma with a high rate (100%) of successful regional anesthesia, defined by at least a 50% reduction in pain scores with no major serious or irreversible side effects. Furthermore, we found that the blood pressure of all patients treated with the URB remained unchanged during onset and maintenance of the nerve block.

It is well recognized that solution injected into the paravertebral space via various traditional methods is not limited only to the paravertebral space but also spreads predictably to the intercostal and epidural space— approximately 40% epidural spread in a cadaveric model and up to 70% in an in vivo radiological model.1,13,14,21–23 Of interest, Cowie recently showed that more solution covers a greater number of intercostal segments with an in-plane lateral to medial ultrasound guided-approach when compared with the paravertebral segments.1 Additionally, spread of solution was greater in terms of distance from the injection site in the intercostal space when compared with the paravertebral space.1

The actual distribution of the injectate with the ultrasound-guided retrolaminar block is not known and will be the subject of further studies. These studies will most likely rely on direct anatomic investigation, as clinical studies that rely on dermatomal mapping after paravertebral blockade are unable to distinguish among local anesthesia in the paravertebral, intercostal, or epidural spaces.1,2,24,25 Since it is likely that the analgesic effects of the paravertebral block are due to some combination of paravertebral, intercostal, or epidural spread, and even systemic local anesthetic effects, it would be interesting and important to know if different block techniques result in different injectate distribution. In these cases of multiple rib fractures presented here, another factor could also play an important role in the local anesthetic spread, i.e., the traumatic injury and possibly uneven disruption of ribs, ligaments, and fascias, possibly facilitating access of the local anesthetic to the true paravertebral space. Thus far, our retrolaminar experience from patients with different indications for regional anesthesia, such as surgical pain after mastectomy, thoracotomy, laparotomy, is still insufficient to draw meaningful comparisons. Both anatomical and clinical studies are needed.

At the present time, we find ourselves using slightly higher bolus solutions, about 20 mL of solution, and the same rate of continuous infusions when compared with our traditional placement of paravertebral blocks. Our daily management of the lamina approach and the traditional approach has been similar. The observation that we are using slightly higher bolus solution volumes was coincidental in our review of these cases, as we do not have sufficient data or clinical experience to recommend adjustment of local anesthetic volumes as used in the traditional paravertebral block approach.

When compared with the traditional paravertebral techniques, the ultrasound-guided retrolaminar block described here is predictably performed in a more superficial tissue plane, with target, needle trajectory, and injection point that are relatively easy to visualize and farther away from the pleura. Though our first cases were uncomplicated, it is too early to draw conclusions regarding safety and advantages of this block; however, it seems logically designed to decrease the risks of injury to the pleura and the deep paravertebral structures.

These initial clinical experiences lead us to suggest that the ultrasound-guided retrolaminar block is feasible in patients with multiple rib fractures. It is easy to perform while providing analgesia that is effective, thereby providing a viable alternative to more traditional approaches to paravertebral blockade.