Liposomal Bupivacaine versus Dexamethasone: Comment

A recently published article in Anesthesiology reported results from a randomized trial that compared the analgesic effectiveness of interscalene brachial plexus block with liposomal bupivacaine to standard bupivacaine with perineural dexamethasone.1 This article by Kim et al.1 concluded that the two study drugs provided similar analgesia with no differences in opioid consumption. However, the article contains a statistical error that led to an incorrect conclusion on the primary outcome measure, and the relationship between pain intensity scores and opioid rescue medication was not fully considered.In this blinded study, 112 adult patients undergoing ambulatory arthroscopic shoulder surgery were randomized to receive an interscalene brachial plexus block with either an admixture of 10 ml (133 mg) liposomal bupivacaine and 5 ml 5% bupivacaine or an admixture of 15 ml 0.5% bupivacaine and 4 mg dexamethasone. The primary outcome was the average numerical rating scale pain scores at rest during 72 h. The primary analysis evaluated whether liposomal bupivacaine was noninferior to bupivacaine with dexamethasone at a margin of 1.3 points.2,3 Additional outcomes included opioid consumption, patient satisfaction, and duration of sensory and motor block at predefined time points up to 7 days postsurgery.The article reported that the mean ± SD numerical rating scale pain score during the first 3 postoperative days was 2.4 ± 1.9 in the liposomal bupivacaine group and 3.4 ± 1.9 in the bupivacaine with dexamethasone group, with a mean difference of −1.1 (95% CI, −1.8 to −0.4; P < 0.0001 for noninferiority). The article concluded that liposomal bupivacaine was not superior to bupivacaine with dexamethasone (one-sided P = 0.998). However, this conclusion and associated P value are incorrect because the upper bound of the 95% CI for the difference between groups excludes 0 in favor of liposomal bupivacaine.4 The P value of 0.998 actually corresponds to a test of superiority (with a null value of 0) for bupivacaine with dexamethasone over liposomal bupivacaine. The correct P value for a one-sided test of superiority of liposomal bupivacaine is 0.002, which is statistically significant. Therefore, the conclusion that the treatments provided similarly effective analgesia is not supported, because liposomal bupivacaine demonstrated both noninferiority and superiority to bupivacaine with dexamethasone. Notably, a treatment demonstrates superiority to a comparator when the CI excludes 0, even when a noninferiority study design is used (fig. 1).5,6The analyses presented in the article by Kim et al.1 did not consider the important relationship between pain intensity scores and rescue medication. Consider, for example, a clinical trial of two analgesics in which patients randomized to receive the less effective analgesic study drug required more opioid rescue medications to achieve satisfactory pain control. The difference between the treatment groups in pain intensity scores over time will be attenuated owing to greater rescue medication use in the group that received the less effective study drug. Therefore, assessments of analgesic effectiveness must consider both pain intensity scores and the amount of rescue medication that supplemented the study drug to achieve those scores.The trial used a stepwise approach to opioid rescue pain medication based on patient-reported pain severity (i.e., tramadol for mild or moderate pain and oxycodone or intravenous hydromorphone for severe pain). Kim et al.1 noted that there were no significant differences in opioid consumption at specific time points (i.e., in the postanesthesia care unit and on postoperative days 1, 2, 3, 4, and 7). However, performing repeated statistical tests for each individual postoperative day reduced the statistical power to detect between-group differences. Furthermore, given the primary objective to evaluate the ability of the study drugs to provide prolonged analgesia during 72 h, it is more appropriate to consider cumulative opioid consumption through 72 h as well. The cumulative postoperative opioid consumption during the first 72 h, the same time window as the primary outcome measure, was 33% lower for patients who received liposomal bupivacaine than for those who received bupivacaine with dexamethasone. Therefore, liposomal bupivacaine achieved a statistically significant reduction in pain intensity scores during the primary study period while, during that same time, fewer opioids were used. It naturally follows that the statistically significant effect on pain intensity scores in favor of liposomal bupivacaine would have been larger than the observed difference of –1.1 if the statistical analysis had accounted for differential use of rescue medication.In conclusion, the randomized controlled trial reported by Kim et al.1 was a well-conducted evaluation of two promising treatment options to provide extended analgesia with an interscalene brachial plexus block. Although postsurgical pain was generally well-controlled with both interventions, the conclusions regarding the key study outcomes were inaccurate. In fact, the trial demonstrated that liposomal bupivacaine provided superior pain control with lower opioid consumption in the postoperative period.An erratum has been published regarding Kim et al.: Anesthesiology 2022; 136:434–47, and a corrected version of the manuscript has been posted online.Dr. DiGiorgi, Dr. Winston, and Dr. Yu are employees of Pacira BioSciences, Inc. (Parsippany-Troy Hills, New Jersey) and may own stock or stock options in the company.