• Brachial Plexus
• Ultrasonography
• Pain, Postoperative
• Pharmacology

We appreciate the opportunity to address the concerns of Winston et al, as their unique perspective is certainly germane to this ongoing discussion.1 To clarify, our study2 does not suggest that liposomal bupivacaine has no advantage over plain bupivacaine. Rather, our aim was to assess the magnitude and value of this advantage in actual clinical practice. As such, we would like to respond to each of the authors’ concerns within the context of this goal.

First, we disagree that our non-inferiority margin is inappropriately large, as it was selected to be equivalent to approximately a single dose of oxycodone (5 mg) per day for the interval of our primary outcome (72 hours). Our typical clinical practice is to prescribe oral oxycodone for pain following surgery. From a resource use standpoint, 5 mg is the relevant minimum value, as the drug is usually dispensed in 5 mg tablets. Because prior studies have used oral morphine equivalents to represent their opioid consumption, we simply converted this dose to that unit of measurement for the purposes of reporting our findings and comparing with other similar work. In addition, most clinical practices employ multimodal analgesia in conjunction with regional anesthetic techniques to reduce opioid consumption following surgery. Our goal was to assess the effectiveness of liposomal bupivacaine in the context of actual clinical practice. While we believe that there might have been some academic interest in a smaller non-inferiority margin or the exclusion of multimodal analgesia, a study designed in this fashion would lack meaningful clinical relevance.

Second, we acknowledge that our study may have been underpowered, a limitation we described within the article. As the result of the formal hypothesis test did not show non-inferiority, we deliberately made no inferential statements about the relative effectiveness of liposomal bupivacaine versus plain bupivacaine. We reported the point estimate of the difference in 72-hour opioid consumption as a description of our findings. Moreover, we agree that it is inappropriate to draw conclusions solely based on point estimates and also advocate for the use of confidence intervals (CIs) in data interpretation.3 Interestingly, in their letter, the authors only note the possibility of a reduction in 30.8 Morphine Milligram Equivalents (the upper limit of the 95% CI) when using liposomal bupivacaine; however, an increase of 6.9 Morphine Milligram Equivalents (the lower limit of the 95% CI) is equally consistent with our findings.

Third, we agree with the authors that there is most certainly a relationship between pain scores and opioid consumption. These differences were highlighted in both the Results and Discussion sections of our article. While it was true that the pain scores in the plain bupivacaine group were numerically higher than those in the liposomal bupivacaine group, the difference only reach significance (with no adjustments for multiple testing) at a single time point. Furthermore, the magnitude of the significant difference was 0.9 on the 11-point numerical rating scale, which cannot be reasonably interpreted as clinically important.

In summary, our study was designed to be consistent with the concepts of the triple aim, which advances healthcare by improving the patient experience, enhancing quality of care and reducing costs.4 As such, we base the interpretation of our findings on the application of these principles.