Comparison of the Analgesic Efficacy between Arthroscopically Placed Continuous Suprascapular Nerve Block and Ultrasound-guided Continuous Superior Trunk Block: A Double-blinded Randomized Controlled Trial

This prospective, double-blinded, randomized controlled study was performed at a single tertiary center in Seoul, Republic of Korea. The study was performed with the approval of the institutional review board of our center (protocol No. 2020-1285; approval date: August 14, 2020), and was registered on the Clinical Research Information Service, a clinical trial registry in Korea (KCT0005500; principal investigator, Won Uk Koh; first registration date: October 20, 2020). This study was conducted in accordance with the original protocol. A trial protocol and all the data could be provided upon request.

One researcher screened for eligibility, and another researcher approached the patients at the general ward or outpatient surgery clinic. After explaining the study’s rationale, written informed consent was obtained from the patients who agreed to participate in this study.

Patients were included who met the following criteria: (1) aged between 19 and 80 yr, (2) had an American Society of Anesthesiologists Physical Status of I/II/III, and (3) were scheduled to undergo a simple arthroscopic shoulder surgery and arthroscopic superior capsule reconstruction during general anesthesia in our center between December 2020 and October 2021. Patients with severe pulmonary diseases, uncontrolled medical or psychologic diseases, a history of trauma on the ipsilateral upper extremity, cervical neuropathies, myelopathies, or any allergies to nonsteroidal anti-inflammatory drugs, opioids, or ropivacaine were excluded. We further excluded patients who required additional surgeries, were pregnant or breastfeeding, or refused to participate in this study.

Since arthroscopic superior capsule reconstructions have a longer surgical time and are associated with a greater degree of postoperative pain, we performed stratified randomization to identify patients who would undergo this surgery. We divided the patients into two strata according to the surgical strategy (i.e., arthroscopic rotator cuff repair and arthroscopic superior capsule reconstruction), and the patients in each stratum were allocated in a 1:1 ratio into two groups, the continuous superior trunk block group and continuous suprascapular nerve block group, by one researcher who was not involved in the block procedure and outcome assessment using a computer-generated randomization program. The researcher prepared the opaque concealed envelopes containing the group assignments. The envelopes were opened and handed to the researchers who performed the block procedures when the patients entered the preoperative block room or operating room.

Regardless of the allocated group, all patients received a preoperative single-shot brachial plexus block by two skilled anesthesiologists. After all the patients had arrived in the preoperative block room or operating room, standard monitoring was performed. Next, the patients were placed in a lateral decubitus position with their ipsilateral shoulder placed upward. A prescan of the brachial plexus in the cervical region was conducted using a high-frequency linear transducer (5 to 18 MHz; SONIMAGE HS1, Konica Minolta, Japan). The skin was disinfected, and sterile drapes were applied. During the scan of the cervical region, the brachial plexus between the anterior and middle scalene muscles was primarily visualized.

In the continuous suprascapular nerve block group, the transducer was moved distally, and the suprascapular nerve branching from the superior trunk was identified. The suprascapular nerve was traced until it was beneath the omohyoid muscle. Next, 1 ml 2% lidocaine was initially infiltrated into the skin for needle insertion, and a 21-gauge, 85 mm stimulating needle (Echoplex, Vygon, France) directed toward the suprascapular nerve in a lateral-to-medial direction. This procedure was performed under ultrasound guidance using an in-plane approach. After confirming the motor response to a nerve stimulation of 0.5 mA, a bolus of 0.2% ropivacaine was injected in accordance with the visualization of the injectate spread (fig. 1A). The planned bolus dose of ropivacaine was between 5 and 10 ml, with a minimum of 5 ml injected around the suprascapular nerve. The injection was ceased upon observation of the drug’s spread to the superior trunk of the brachial plexus to minimize the risk of hemidiaphragmatic paresis. The success of the initial block was assessed 30 min after the block by checking the cold sensation on the deltoid area compared with the contralateral deltoid area; success was defined as a reduction in sensation. After confirmation, the patients were transferred to the operating theater with standard monitoring and underwent general anesthesia. The patients were then placed in a lateral decubitus position, and a routine arthroscopic rotator cuff repair or arthroscopic superior capsule reconstruction was carried out. Once the main procedure was completed, an indwelling catheter (Pajunk E-Cath, PAJUNK GmbH, Germany) for continuous suprascapular nerve block was inserted before closing the portals intraoperatively by surgeon. The vascularized fat pad and soft tissue were removed through the anterolateral working portal under the guidance of the posterolateral viewing portal. The location of the transverse scapular ligament at the medial side of the base of the coracoid process and the coracoclavicular ligament was then confirmed. The transverse scapular ligament and suprascapular nerve were identified using an obturator inserted from the Neviaser portal. The supraspinatus muscle and surrounding soft tissues were retracted by a switching stick or probe inserted via the anterolateral portal. After the arthroscopic visualization of the transverse scapular ligament and suprascapular nerve, a catheter was inserted toward the suprascapular nerve in the suprascapular notch through the Neviaser portal (fig. 1B). The catheter tip was positioned near the suprascapular notch, progressing as far as 1 cm following the nerve pathway. After checking the function of the catheter by injecting 5 ml normal saline and confirming the tip position by viewing the bubble, the catheter was fixed with 2-0 nylon skin tagging.

In the continuous superior trunk block group, the transducer was moved distally until the superior trunk was identified. After skin infiltration, an 18-gauge introducing needle connected to a syringe filled with 0.2% ropivacaine was directed toward the superior trunk. When the motor twitch response to 0.5 mA nerve stimulation was detected, 10 ml 0.2% ropivacaine was injected. Subsequently, an indwelling catheter (Pajunk E-Cath) was inserted in the same location beneath the superior trunk; the placement of the catheter tip was confirmed by injecting 1 to 2 ml normal saline (fig. 1C). The catheter was fixed with skin adhesives. After confirming the success of the block 30 min after the end of the preoperative procedures (in the same manner as the continuous suprascapular nerve block group), the patients were transferred to the operating theater and underwent surgery during general anesthesia.

The catheter insertion site was covered with an opaque dressing so that neither the observer nor the patient was aware of the allocated group.

All patients received the same analgesia protocol, except for differences in the peripheral nerve block technique applied. When the patients were transferred to the postanesthesia care unit (PACU), nerve block patient-controlled analgesia (PCA) was connected to the inserted catheter and initiated. Nerve block PCA was prepared using a mixture of 67 ml 0.75% ropivacaine and 183 ml normal saline (250 ml 0.2% ropivacaine). The pump was set for a bolus dose of 5 ml followed by a basal rate of 5 ml/h with a locking interval of 30 min. During the postoperative hospitalization, the same dose of oral medication was prescribed regularly: 100 mg gabapentin, 45 mg naproxen, and 50 mg tapentadol were started from the night on the day of surgery and given every 12 h thereafter. The patients were encouraged to press the button on the nerve block PCA if they felt breakthrough pain. Even if the patient was still in moderate (numerical rating scale 4 or greater) and severe (numerical rating scale 7 or greater) pain after the bolus of PCA, IV 90 mg diclofenac and 1 mg hydromorphone or 50 mg tramadol was administered by a nurse, respectively.

The primary outcome of this study was the assessment of the worst and resting postoperative pain scores, evaluated using a numerical rating scale at postoperative 24 h by an observer blinded to the allocated group. The worst and resting pain were defined as the worst breakthrough pain experienced during the time frame from the previous evaluation and the mean pain at rest, respectively. The secondary outcomes evaluated were the postoperative pain scores at four time points: postoperative 1 (at PACU), 4, 8, and 48 h. Assessment of the pain scores at postoperative 8 h was planned to omit in the patients who were sleeping.

The extent of ipsilateral hemidiaphragmatic paresis was evaluated to check the effect of the nerve block on the phrenic nerve. The diaphragmatic excursion was measured using a curved ultrasound transducer (2 to 5 MHz; SONIMAGE HS1). The diaphragmatic excursion was measured at four time points: before the nerve block, 30 min after the block, at PACU and at postoperative 24 h. Patients were maintained in the upright sitting position, and the curved probe was placed in the subcostal area between the midclavicular and anterior-axillary planes using the liver or spleen as an acoustic window.¹²  After the diaphragm was visualized in the two-dimensional mode, patients were asked to breathe deeply, and the movement of the diaphragm was measured using the M-mode. The maximal value of the diaphragmatic movement was recorded. We defined the complete and partial paresis as the postblock diaphragmatic excursion less than 25% and 25 to 75% of the preblock value.¹³ 

In addition, the use of rescue IV analgesics within postoperative 24 h was recorded in the electronic record system. Specifically, the amount of IV opioids used was converted to a morphine equivalent dose and documented. Postoperative sensory and motor deficits of the affected side forearm and hand were also investigated at PACU, postoperative 4, 8, 24, and 48 h. Sensory function was assessed in the C5/6 skin segment (lateral aspect of the forearm and first digit) and was recorded as absent if there was a loss of sensation to a cold stimulus, partial recovery if there was a reduced sensation, and complete recovery if the sensation was complete. Motor function was assessed by asking the patients to take the following actions: radial nerve, thumb up; and median nerve, fist clenching. Motor function was recorded as absent if all actions were impossible, partial recovery if some actions were difficult, and complete recovery if all actions were possible. At the same time, postoperative nausea and vomiting were also evaluated. We further obtained other clinical outcomes, including the change of hand grip, quality and quantity of night sleep, patient satisfaction, and postoperative Quality of Recovery score (QoR-15) at postoperative 24 h.¹⁴  In addition, we investigated the incidence of secondary block failure, which was defined as the occurrence of a block of the catheter, catheter dislodgement, or leakage from the catheter insertion site. All the outcome assessments were performed by a blinded observer who was not involved in the block procedure and anesthesia, and if the patient was discharged before postoperative 48 h, the outcome variables at that time point were reported via a telephone call.

This study was designed as a noninferiority trial. For the sample size calculation, we defined the noninferiority margin as 1, considering the minimal clinically important difference with an alpha value of 0.025 and a power of 80%.⁷  The SD of numerical rating scale scores at postoperative 24 h was set as 1.7, based on our previous study.⁸  Therefore, a minimum sample size of 46 patients was required for each group. We enrolled 50 patients considering a 5% dropout rate. SAS, version 9.4 (SAS Institute Inc., USA) was used for data analysis.

In the analysis of the primary outcome, the upper margin 97.5% CI was calculated. P values comparing the numerical rating scale scores at postoperative 24 h were calculated by a noninferiority test. The noninferiority hypothesis was statistically assessed using the Wald-type test with a 97.5% one-sided CI and margin of noninferiority. Continuous variables are presented as the mean ± SD and compared using a t test or Kruskal–Wallis test. Categorical variables are presented as numbers (percentages) and analyzed using a chi-square, Fisher exact test, or Mann–Whitney test. Moreover, we used the general linear model (covariance pattern model in the linear mixed model) to explore the effects of the time-by-group interaction on the pain score by considering the correlation between the observations within the same patients. A generalized estimating equations model with a logit link function was used to account for the natural correlation with measurements and the clustering effect on the same patient for the analysis of categorical outcomes. The correlations among the categorical outcomes are modeled as exchangeable (compound symmetry) correlation structure. Since the portion of missing data was insignificant, and the missing data were considered not associated with the value per se, a pairwise deletion (also known as available-case analysis) was performed without any special processing.

A total of 107 patients scheduled to undergo arthroscopic shoulder surgery for a rotator cuff tear were screened for eligibility. Among these patients, 100 were finally enrolled, and the enrollment was halted upon reaching the target sample size. Ultimately, 98 patients were included in the final analysis (fig. 2). The baseline characteristics of the patients and surgery are shown in table 1. The volume of ropivacaine used for single-shot was 10 ml in all patients in the continuous superior trunk block group, while it ranged from 8 to 10 ml in continuous suprascapular nerve block group (9.9 ± 0.5 ml, mean ± SD).

The worst and resting pain scores at postoperative 24 h, the primary outcome of this study, are demonstrated in table 2. In the noninferiority test of the primary outcome with a noninferiority margin of 1, the worst numerical rating scale scores at postoperative 24 h in the continuous suprascapular nerve block group were shown to be inferior compared to those in the continuous superior trunk block group (5.2 ± 1.9 and 4.3 ± 2.1, respectively; mean difference, 0.9; 95% CI, 0.1 to 1.7; P = 0.370). The resting numerical rating scale scores at postoperative 24 h in the continuous suprascapular nerve block group were inferior compared to those in the continuous superior trunk block group (2.1 ± 1.2 and 1.6 ± 1.3, respectively; mean difference, 0.5; 95% CI, 0.0 to 1.0; P = 0.029). The results were not changed after adjusting the stratification variable (surgical strategy; Supplemental Table 1, https://links.lww.com/ALN/D235).

All individual pain scores are presented in figure 3, while the worst and resting numerical rating scale scores at each time point are shown in Supplemental Figure 1 (https://links.lww.com/ALN/D236). The general linear model detected a significant effect of the group-by-time interaction on the worst numerical rating scale scores (P = 0.042; Supplemental Table 2, https://links.lww.com/ALN/D235). Significant differences were observed in the worst numerical rating scale score at PACU, postoperative 4 h, and at postoperative 24 h between the two groups (P = 0.011, 0.021, and 0.036, respectively); the continuous superior trunk block group further showed a significantly lower worst numerical rating scale score than the continuous suprascapular nerve block group. However, no significant group-by-time interaction was observed in the analysis of resting numerical rating scale scores (P = 0.451).

All diaphragmatic excursions, regardless of right or left side, were measurable, and the percent changes in diaphragmatic excursion after the block, relative to the baseline diaphragmatic excursion, are depicted in figure 4. There were significant differences in the absolute values of diaphragmatic excursion and the incidences of partial and complete hemidiaphragmatic paresis between the two groups (Supplemental Table 3, https://links.lww.com/ALN/D235). The incidences of dyspnea and Horner’s syndrome were lower in the continuous suprascapular nerve block group, while the patient satisfaction level was lower and the amount of ropivacaine via PCA was greater. No desaturation occurred in any patients in either group, and no differences were observed in other clinical outcomes (Supplemental Table 3, https://links.lww.com/ALN/D235).

Analysis of the incidence of postoperative nausea and vomiting showed that the group had no significant effect (P = 0.891 and P = 0.265, respectively). In the analysis of the sensory and motor function using the generalized estimating equations model, no significant group-by-time interaction effects were observed. However, the group effect showed a significant impact on motor function (P = 0.006), while it was not significant in the analysis of sensory function (P = 0.105). The comparison of motor function at each time point was demonstrated in Supplemental Table 4 (https://links.lww.com/ALN/D235).

In this study, the continuous suprascapular nerve block did not provide a noninferior postoperative analgesic effect compared to the continuous superior trunk block in patients undergoing arthroscopic shoulder surgery. However, no incidence of complete hemidiaphragmatic paresis was seen in the continuous suprascapular nerve block group.

Conventional interscalene brachial plexus block is an effective analgesic technique for arthroscopic shoulder surgery. In particular, using catheters with this technique led to better postoperative analgesia and opioid-sparing effects than a single-injection interscalene brachial plexus block.^15–17  However, the incidence of hemidiaphragmatic paresis after this technique was reported to be up to 100%.¹⁸ 

Several peripheral nerve block techniques were used as an alternative to the conventional interscalene brachial plexus block to reduce the risk of hemidiaphragmatic paresis.⁹  Superior trunk block was shown to reduce the risk of hemidiaphragmatic paresis and to have noninferior analgesia compared to the interscalene block;⁶  however, although a significant reduction in the incidence of hemidiaphragmatic paresis was observed after the superior trunk block, hemidiaphragmatic paresis was still induced 30 min after the block in 76.3% of the cases.⁶  These results are consistent with those of our study, in which approximately two thirds of the patients in the continuous superior trunk block group developed complete or partial hemidiaphragmatic paresis. Another study showed that superior trunk block was noninferior to the interscalene block in terms of its analgesic effect with significantly rare hemidiaphragmatic paresis.⁷  However, complete hemidiaphragmatic paresis occurred in about 5% of patients in this study.⁷  Furthermore, these two studies only evaluated the effect of a single-shot superior trunk block and could not guarantee the safety of continuous superior trunk block in terms of hemidiaphragmatic paresis.

Additionally, more distal approaches to the brachial plexus are being explored, including retroclavicular blocks and selective suprascapular nerve blocks. Retroclavicular block has demonstrated relatively little effect on diaphragm function, while effectively blocking the suprascapular nerve.¹⁹  Selective suprascapular nerve block has also exhibited efficacy in postoperative analgesia for shoulder surgery.²⁰  However, in our study, superior trunk block showed a more potent analgesic effect than suprascapular nerve block. The higher worst numerical rating scale scores observed at PACU and 4 h postoperatively in the continuous suprascapular nerve block group may be highly associated with the effect of a single-shot block before the surgery. In this current study, the subomohyoid suprascapular nerve block technique was used as a single-shot block. The analgesic effect of the subomohyoid technique has been previously shown to be noninferior compared to interscalene brachial plexus block.²¹  The discrepancy between this study and ours may be partially explained by the extent of injectate spread in the superior trunk. Abdallah et al. used 15 ml 0.5% ropivacaine for the suprascapular nerve block,²¹  whereas we injected only 8 to 10 ml 0.2% ropivacaine for the suprascapular nerve block to avoid the spread of the local anesthetics to the adjacent superior trunk.

The strength of our study was the comparative study of the analgesic effects of continuous suprascapular nerve block and continuous superior trunk block in patients undergoing arthroscopic shoulder surgery. Continuous suprascapular nerve block was found to be inferior compared to continuous superior trunk block at postoperative 24 h based on the numerical rating scale scores. Also, the amount of nerve block PCA used was greater in the continuous suprascapular nerve block group. Although the major nerve innervating the shoulder is the suprascapular nerve, other nerves, including the axillary nerve, lateral pectoral nerve, musculocutaneous nerve, and long thoracic nerve, are also involved in postoperative pain. Therefore, the continuous superior trunk block, which targets the proximal portion of the brachial plexus before it branches off into small nerves, may be more effective than the continuous suprascapular nerve block. Moreover, a previous study reported that while most sensory branches of suprascapular nerve typically split off from the main nerve around the transverse scapular ligament, 48.2% of cases exhibited proximal branching of the sensory nerve before the ligament.²²  Considering this anatomical characteristic, the effect of a suprascapular nerve catheter placed under the transverse scapular ligament might be inferior to that of the superior trunk block. However, the mean differences in the worst and resting pain scores at postoperative 24 h between the two groups were less than 1. Myles et al. suggested that the minimal clinically important difference for the visual analogue scale (0 to 100) was 9.9.²³  Thus, based on the minimal clinically important difference, the difference in pain score at postoperative 24 h between the two groups was considered not clinically important. Moreover, a systematic review article identifying the minimal clinically important difference in orthopedic surgeries demonstrated that the median minimal clinically important differences during movement and rest were 18 and 15 mm, respectively, when using the visual analogue scale.²⁴  In addition, no differences were observed in sleep quality, sleep hours, or QoR-15 scores between the two groups. The lack of differences in these parameters could be interpreted as no clinically important difference between the two groups.

We observed that continuous suprascapular nerve block rarely induced hemidiaphragmatic paresis. In our study, only one patient developed partial hemidiaphragmatic paresis at postoperative 24 h in the continuous suprascapular nerve block group. Notably, this patient did not exhibit any clinical symptoms or signs of hemidiaphragmatic paresis. In contrast, two thirds of patients in the continuous superior trunk block group demonstrated partial or complete hemidiaphragmatic paresis at postoperative 24 h. These findings highlight the potential benefits of continuous suprascapular nerve block in preserving phrenic nerve function. Therefore, despite the inferior analgesic efficacy compared to the continuous superior trunk block, we propose that the insertion of the suprascapular nerve block catheter with arthroscopic assistance should be the preferred analgesic method to mitigate the risk of hemidiaphragmatic paresis. Auyong et al. also showed the advantages of continuous suprascapular nerve block on pulmonary function and diaphragmatic movement; however, they noted that hemidiaphragmatic paresis was not completely avoided using suprascapular catheters (Supplemental Table 5, https://links.lww.com/ALN/D235).²⁵  We assumed that the differences between our study and that of Auyong et al. were derived from differences in the location of the catheter tip. Auyong et al. placed the catheter tip adjacent to the suprascapular nerve beneath the omohyoid muscle, suggesting that they evaluated the effect of the continuous anterior suprascapular nerve block technique.²⁵  Meanwhile, we investigated the effect of continuous posterior suprascapular nerve block, in which the catheter tip was located adjacent to the suprascapular nerve in the suprascapular notch. Local anesthetics injected beneath the omohyoid muscle have a higher risk of spreading to the proximal brachial plexus and phrenic nerve in a retrograde direction due to the small distance between the subomohyoid region and the phrenic nerve.

The differences between the two block techniques lie in the personnel responsible for catheter insertion and the timing of insertion. In the case of the suprascapular nerve block catheter, it is placed intraoperatively by the surgeon, allowing for direct visualization of the suprascapular nerve and potentially increasing the accuracy of catheter placement. However, the intraoperative placement of the catheter does not have a preemptive analgesic effect. Therefore, to maximize the analgesic effect of the peripheral nerve block for shoulder surgery, preoperative single-shot peripheral nerve block should be performed in addition. On the other hand, the superior trunk block catheter is typically placed preoperatively by an anesthesiologist using ultrasound guidance, which allows for a potential preemptive effect. However, due to the possibility of multiple position changes after the catheter placement, there may be a higher chance of the catheter displacement.

This study had several limitations. First, we did not perform a pulmonary function test, which was originally planned to evaluate the clinical effects of the block. As this study was conducted during the COVID-19 pandemic, we omitted this test for the safety of the patients and medical staff. Second, our study protocol could not prevent hemidiaphragmatic paresis during the immediate postoperative period. The single-shot anterior suprascapular nerve block induced hemidiaphragmatic paresis in some patients, although we attempted to avoid the retrograde spread of the local anesthetics. The volume of local anesthetics for the single-shot block technique was determined by visualization using an ultrasound in the range of 8 to 10 ml in this study. In a previous cadaver study, Maikong et al. investigated the phrenic-sparing volume of the anterior suprascapular nerve block and showed that the maximum effective volume of local anesthetic was 4.2 ml.²⁶  Therefore, this volume of local anesthetics could be used in patients with a high risk of postoperative pulmonary complications. Alternatively, a single-shot posterior suprascapular nerve block combined with catheterization using arthroscopy may be another option for hemidiaphragmatic paresis prevention.¹²  Last, in this study, we could not accurately investigate the incidence of secondary block failure. Although there is currently no definitive way to identify secondary block failure, a previous study confirmed catheter function during the postoperative period by visualizing the perineural spread of the local anesthetic via color Doppler.²⁷  However, we were unable to confirm catheter function due to the dressing used for blind assessment. Considering that a previous study reported that a secondary failure rate of the interscalene catheter failure was 11%, the incidence of secondary failure in our study might be underestimated.²⁸  Nevertheless, since these conditions were equally applied to both study groups, their impact on our results is likely negligible.

In conclusion, our study demonstrated that arthroscopically assisted continuous suprascapular nerve block showed statistically inferior analgesic effects with minimal effect on phrenic nerve function in patients undergoing arthroscopic shoulder surgery. We suggest that the continuous suprascapular nerve block technique can be considered in patients with a high risk of postoperative respiratory complications.

Support was provided solely from institutional and/or departmental sources.

The authors declare no competing interests.

Full protocol available at: koh9726@naver.com. Raw data available at: koh9726@naver.com .

Supplemental Digital Content 1. Supplemental Tables, https://links.lww.com/ALN/D235

Supplemental Digital Content 2. Supplemental Figure 1, https://links.lww.com/ALN/D236

Supplemental Video File 1. Subomohyoid suprascapular nerve block, https://links.lww.com/ALN/D237

Supplemental Video File 2. Superior trunk block and catheterization, https://links.lww.com/ALN/D238