Interscalene Brachial Plexus Block with Continuous Intraarticular Infusion of Ropivacaine

Abstract

Providing intraarticular analgesia with a continuous infusion of local anesthetic via a disposable infusion pump has gained popularity. Despite the prevalence of this technique, data comparing this method of analgesia to conventional regional anesthesia are not available. We present a prospective study that compared a single-dose interscalene block with a single-dose interscalene block plus continuous intraarticular infusion of local anesthetic. Forty patients scheduled for shoulder arthroscopy were entered in this prospective, double-blinded study. All patients received an interscalene brachial plexus block as their primary anesthetic. Patients were randomly assigned to 1 of 2 groups: 1. interscalene block with 1.5% mepivacaine (40 mL) followed by a postoperative intraarticular infusion of 0.5% ropivacaine at 2 mL/h, or 2. interscalene block with 0.5% ropivacaine (40 mL) followed by a postoperative intraarticular infusion of 0.9% saline (placebo) at 2 mL/h. Postoperative infusions were maintained for 48 h. Visual analog scale pain scores and postoperative oxycodone consumption were measured for 48 h. Visual analog scale scores at rest and with ambulation in the Mepivacaine/Intraarticular Ropivacaine group were reduced when compared with the Ropivacaine/Saline group (rest:P = 0.003, ambulation:P = 0.006). Oxycodone consumption was also decreased (28 ± 21 mg vs 44 ± 28 mg, P = 0.046), respectively. We conclude that a brachial plexus block with 1.5% mepivacaine and a continuous intraarticular infusion of 0.5% ropivacaine at 2 mL/h provides improved analgesia for minor surgery at 24 and 48 h versus a single-injection interscalene block with 0.5% ropivacaine.

Intraarticular injection has been used to provide operative and postoperative pain relief for shoulder surgery. Recently, the provision of prolonged intraarticular analgesia by continuous infusion of local anesthetic via a disposable infusion pump has gained popularity. Savoie et al. ⁽¹⁾ demonstrated improved analgesia and high patient satisfaction with this technique when compared with general anesthesia in patients having shoulder surgery. The technique has the advantages of providing local anesthetic to the operative site, ease of catheter insertion, and simplicity of infusion. However, because of the intraarticular location of the catheter, soft-tissue incisions outside of the joint may not be rendered analgesic.

Interscalene brachial plexus blockade is another method of providing operative and postoperative analgesia for shoulder surgery ⁽²⁾. By using this technique, single-injection neural blockade with long-acting amide local anesthetics decreases opioid requirements for 10 to 18 h ⁽³⁾. This technique can also provide complete anesthesia to the shoulder region. However, patients often experience pain after block resolution.

As a result, perioperative analgesia is frequently provided with an intermediate-acting local anesthetic such as mepivacaine combined with a postoperative intraarticular infusion of local anesthetic. Despite the prevalence of this technique, data comparing this method of analgesia with conventional regional anesthesia are not available.

We designed this study to determine whether a block with mepivacaine could be supplemented by an intraarticular infusion so that analgesia equivalent to a long-acting block could be obtained. This would allow the use of a short-acting local anesthetic and avoid a prolonged motor block but still allow extended analgesia. We therefore compared the postoperative analgesic efficacy of an interscalene brachial plexus block with 1.5% mepivacaine combined with a continuous intraarticular infusion of 0.5% ropivacaine at 2 mL/h to an interscalene brachial plexus block with 0.5% ropivacaine, by using a prospective, placebo-controlled study design.

Methods

This study was approved by our IRB and written informed patient consent was obtained. Forty patients classified as ASA physical status I–III, aged 18 yr or older, participated in this study. All patients were scheduled for unilateral shoulder arthroscopy. Patient exclusion criteria included peripheral neuropathy, chronic opioid use, morbid obesity (twice the ideal weight or heavier than 130 kg), or contraindications to regional anesthesia or a history of allergy to the study drugs.

All patients were scheduled to receive an interscalene brachial plexus block as their primary anesthetic. Patients were randomly assigned by prerandomized sealed envelopes to the following groups (n = 20 per group):

• Mepivacaine/Intraarticular Ropivacaine (M-IR) group: interscalene block with 1.5% mepivacaine (40 mL)/postoperative intraarticular infusion of 0.5% ropivacaine at 2 mL/h via a disposable pump for 48 h.
• Ropivacaine/Intraarticular Saline (R-IS) group: interscalene block with 0.5% ropivacaine (40 mL)/postoperative intraarticular infusion of 0.9% saline (placebo) at 2 mL/h via a disposable pump for 48 h.

All patients were brought to the preoperative holding area and monitored using standard American Society of Anesthesiologists (ASA) monitors. Patients were sedated with IV midazolam (1–5 mg) and fentanyl (50–250 μg), titrated to moderate sedation (arousable on command). All interscalene blocks were performed using the approach described previously by Winnie ⁽²⁾, using a 50-mm insulated, blunt needle (Braun Medical, Bethlehem, PA) and a nerve stimulator. After an appropriate stimulus was localized with a current <0.5 mA, the study solution (40 mL) was injected in 3-mL increments. An intercostobrachial block was then performed separately using the initial injected anesthetic to provide anesthesia for the possible placement of an anterior portal during shoulder arthroscopy. Attending anesthesiologists, experienced with the technique, performed all blocks.

All episodes of local anesthetic toxicity or hemodynamic change requiring anesthesiologist intervention (increased IV fluids or inotropes) were recorded as adverse events. After evidence of a successful motor (loss of shoulder abduction) block, the patient was taken to the operating room for surgery. Failure to lose shoulder abduction after 30 min was considered a block failure. During surgery, additional infiltration of local anesthetic was not performed. Intraoperative sedation was provided with IV propofol 10–50 μg · kg⁻¹ · min⁻¹, titrated to moderate sedation (arousable to commands). IV ketorolac 30 mg was administered at the conclusion of surgery.

At the conclusion of surgery, a 20-gauge multiorifice epidural catheter was inserted by the surgeon via an arthroscopy portal and secured with a cutaneous adhesive suture. On arrival to the postanesthesia care unit (PACU), the catheter was aspirated for the presence of blood and a disposable infusion pump (Mc- Kinley Medical, Wheat Ridge, CO) containing 96 mL of study drug (0.5% ropivacaine or normal saline according to group assignment) was connected to the intraarticular catheter and infused at 2 mL/h. The patient and evaluator who conducted the postoperative follow-up were unaware of the identity of the infused solution. In addition, a cutaneous cooling pad (Wrapon Polar Pad^®; Breg, Vista, CA) was applied and patients were instructed to use it continuously for 48 h and then three times a day as needed. All patients were directed to take oral naproxen 500 mg twice a day for 4 days. Patients were also given a prescription for oral acetaminophen 325 mg with oxycodone 5 mg to take if necessary. When patients met standard PACU discharge criteria they were discharged home.

Patients were asked to rate their pain by using a visual analog scale (VAS) (0 cm = no pain/10 cm = worst pain imaginable) on arrival in PACU, and then at 12, 24, and 48 h postsurgery. At the same times, patients were asked to rate their pain while ambulating during activities of daily living. Because of the surgical requirement for immobilization, scores were not obtained with shoulder movement. In addition, patients were given a written log to document the amount and time when they took oral analgesics. Data were obtained by a trained nurse via telephone call using a written questionnaire after 48 h.

Descriptive statistics for all outcomes were produced, including mean ± sd. Differences in group demographic characteristics were tested by t-test or contingency-table χ² test for categorical measures. The nonparametric Wilcoxon’s ranked sum test was used to compare total oxycodone use between groups, with the significance level set at α = 0.05. For the VAS pain scores, repeated-measures generalized linear models analysis was used to test time and treatment effects taking into account the repeated scores on individual patients. The PACU-arrival measures were not included in repeated-measures tests of treatment effects because no difference was expected at that time. Pain scores at rest and on movement were analyzed separately. Post hoc tests of treatment difference at 12, 24, and 48 h using the ranked sum test were planned if the overall treatment difference or the time by treatment interaction was significant. In consideration of the two repeated measures analyses (rest and ambulation), the significance level was set at α = 0.025, and at 0.05/3 = 0.017 for the post hoc tests. Analyses were conducted by using SAS/STAT Version 6 software (SAS Institute, Inc., Cary, NC).

Results

Forty patients completed the study protocol, 20 in the M-IR group and 20 in the R-IS group. There were no differences in age, height, weight, sex, or ASA classification between groups (M-IR/M-IS): age (yr) 37 ± 12/38 ± 14, height (cm) 171 ± 20/176 ± 10, weight (kg) 82 ± 17/82 ± 16, sex (m/f)10:10/13:7.

There were no episodes of acute (during injection) or chronic (after discharge) local anesthetic toxicity or hemodynamic instability. There was one failed interscalene block that required reinjection. One catheter fell out before arrival in the PACU. A second catheter clotted with blood before start of the infusion. Two patients scheduled for arthroscopy were converted to more extensive open procedures. These five patients had their study protocols reentered into the randomization pool. Time spent in the PACU was similar between groups. One patient in the M-IR group returned 24 h after surgery with redness at the catheter insertion site. There were no additional signs of localized or systemic signs of infection; however, the catheter was still removed as a precaution. This patient was included in subsequent data analysis.

On arrival to the PACU, mean VAS scores at rest and with ambulation were 0.2 ± 0.9 in group M-IR and (Figs. 1 and 2). In the repeated-measures VAS pain modeling, the time/treatment interactions were not significant in any model, indicating that the difference between groups was similar over all times. After eliminating the interaction term, the treatment effect on VAS pain score was significant with both rest and ambulation (P = 0.003 and 0.006, respectively). Time was also significant in both models, indicating a difference in pain with time. After adjustment for multiple comparisons in the tests at separate times, the M-IR group showed significantly less pain at 24 h (Wilcoxon’s P = 0.005 at rest, P = 0.005 with ambulation), and the difference at 48 h approached significance (P = 0.023 at rest, P = 0.043 with ambulation).

Figure 1: Mean visual analog scale (VAS) pain scores at rest ± sd (0 = no pain/10 = worst pain imaginable). Patients reported scores every 12 h postoperatively for the first 48 h after receiving an interscalene brachial plexus block with 1.5% mepivacaine and a continuous intraarticular infusion of 0.5% ropivacaine at 2 mL/h (M-IR) or an interscalene brachial plexus block with 0.5% ropivacaine and a continuous intraarticular infusion of saline (R-IS). The treatment effect on VAS pain score at rest was P = 0.003.

Figure 2: Mean visual analog scale (VAS) pain scores with ambulation ± sd (0 = no pain/10 = worst pain imaginable). Patients reported scores every 12 h postoperatively for the first 48 h after receiving an interscalene brachial plexus block with 1.5% mepivacaine and a continuous intraarticular infusion of 0.5% ropivacaine at 2 mL/h (M-IR) or an interscalene brachial plexus block with 0.5% ropivacaine and a continuous intraarticular infusion of saline (R-IS). The treatment effect on VAS pain score with movement was P = 0.006

0.5 ± 1.2 in group R-IS (P = 0.8), indicating the effectiveness of the initial brachial plexus block in both groups

The time until first oxycodone use was similar between groups, 788 ± 365 min (M-IR) and 607 ± 377 min (R-IS) (P = 0.6). The R-IS group required more total oxycodone than the M-IR group (44 ± 28 mg vs 28 ± 21 mg) (Wilcoxon P = 0.046) (Fig. 3). This demonstrated significantly improved analgesia and a decreased opioid requirement.

Figure 3: Total oxycodone use after 48 h in patients receiving an interscalene brachial plexus block with 1.5% mepivacaine and a continuous intraarticular infusion of 0.5% ropivacaine at 2 mL/h (M-IR) or an interscalene brachial plexus block with 0.5% ropivacaine and a continuous intraarticular infusion of saline (R-IS). Horizontal bars indicate mean values.

Discussion

The results of this study demonstrate that a small-dose continuous infusion of local anesthetic into the shoulder joint provides improved analgesia when compared with a brachial plexus block with long-acting amide local anesthetic alone. This improved analgesia also resulted in a 30% reduction in mean opioid consumption (P = 0.046) in the group receiving the intraarticular infusion of local anesthetic.

Of particular interest was the improved analgesia despite the fact that ropivacaine delivery to the joint was only 10 mg/h. This improvement in pain control was obtained with the presence of incisions in the soft tissue for trochar placement and the inclusion of extensive operative procedures, such as thermal capsulorrhaphy within the shoulder.

Caregivers frequently laud regional anesthesia for the intense pain relief provided by neural blockade but complain that postoperative management can be difficult when there is the sudden onset of pain after block resolution. In this study, the patients in both groups received multi-modal analgesic therapy during the postoperative period to avoid this difficulty ^(4,5). This included a regional anesthetic for the procedure, nonsteroidal antiinflammatory drugs (ketorolac and naproxen), cryotherapy, as well as oxycodone and acetaminophen. Despite this broad analgesic regimen, the group receiving the long-acting amide local anesthetic block without intraarticular local anesthetic infusion had a mean VAS score of 5, 5, and 3 at 12, 24, and 48 h, respectively, which we consider to be suboptimal.

The most obvious explanation for the improved pain relief in the intraarticular group is the effect of local anesthetic on the articular and capsular surfaces of the shoulder. Because it is unlikely that 2 mL/h of local anesthetic reached the soft-tissue incisions, incisional pain seems to be less of a factor in pain after shoulder arthroscopy than does intraarticular pain. It also can be suggested that the reduced pain provided by intraarticular ropivacaine infusion can be attributed to avoidance of the rapid resolution of neural blockade and sudden development of pain from a previously analgesic site.

We designed this study to determine whether a block with mepivacaine could be supplemented by an intraarticular infusion so that analgesia equivalent to a long-acting block could be obtained and side effects avoided. We hypothesized that the amount of analgesia provided by a small-dose intraarticular infusion would be minimal. As a result, the R-IS group was expected to have increased postoperative analgesia. The magnitude of extended analgesia provided by the intraarticular infusion was unexpected. A future investigation suggested by these findings would be to compare a ropivacaine block with and without an intraarticular infusion of local anesthetic to quantify the efficacy of postoperative intraarticular infusion using a different local anesthetic for intraoperative anesthesia.

Currently, there are few randomized data examining the efficacy of ambulatory continuous intraarticular infusions and, in particular, there are no studies examining this mode of treatment when combined with regional anesthesia and multi-modal analgesia ⁽⁶⁾. Savoie et al. ⁽¹⁾ demonstrated improved analgesia and reduced analgesic requirements with a continuous intraarticular infusion of 0.25% bupivacaine (2 mL/h for 48 h) in patients undergoing subacromial decompression under general anesthesia. In that study, specific multimodal interventions were not described and a long-acting interscalene block was not used. However, they too found a significant analgesic and opioid-sparing effect from postoperative intraarticular infusion.

The fact that the method of catheter insertion is simple, quick, and can be accomplished under direct visualization via the arthroscope makes the technique easy to perform. In addition, the disposable infusion pump is simple, has no mechanical parts, and requires no intervention by patient or staff. The limitations of the system are the inability to tailor infusion rates or deliver a bolus of local anesthetic. Determining the exact amount of infused solution is difficult. Furthermore, because there is no high-pressure alarm or an ability to flush the catheter, identifying and correcting an occlusion can be difficult in the ambulatory setting ⁽⁷⁾. Although unlikely, the potential for infection within the joint also exists. In addition, despite the simplicity of the available disposable infusion pumps, the cost of the system remains an issue. Available models can range in price from $95–$175, depending on acquisition agreements. Randomized cost/benefit data have not been established.

The results of this study confirm the efficacy of ambulatory intraarticular local anesthetic infusion for arthroscopic shoulder surgery. The data demonstrate the increased analgesic effect that can be obtained even when combining it with an intermediate-acting regional anesthetic and a broad multi-modal pain-relief strategy. Comparing this technique with a continuous brachial plexus block would be interesting. Given the results of this study, further data defining the minimal local anesthetic infusion rate and the technique’s efficacy in more invasive surgeries would be useful.