“[T]he clinical meaningfulness of group differences in pain scores cannot alone determine the overall clinical importance of that difference: tolerability, safety, convenience, durability of benefits, and cost along with a host of additional factors must be considered including comparisons with alternatives.”

Postoperative pain usually lasts considerably longer than the duration of plain bupivacaine HCl when used for single-injection peripheral nerve blocks, even with unapproved additives such as epinephrine, clonidine, and steroids. A formulation of bupivacaine HCl within liposomes (Exparel, Pacira BioSciences, USA) is approved by the U.S. Food and Drug Administration (Silver Spring, Maryland) for use in interscalene, adductor canal, transversus abdominis plane, and popliteal–sciatic nerve blocks. Pacira proposes that liposomal bupivacaine lasts 48 to 72 h, which far exceeds the duration of plain bupivacaine (usually less than 24 h). However, the extent to which liposomal bupivacaine prolongs analgesia remains controversial.  A trial by Chan et al. in this issue of Anesthesiology helps us understand this potentially transformative local anesthetic. 

Chan et al. evaluated 80 patients having distal radial fracture fixation with a supraclavicular brachial plexus block who were randomized to bupivacaine hydrochloride (20 ml of 0.5%) or a combination of bupivacaine hydrochloride (10 ml of 0.5%) and liposomal bupivacaine (10 ml of 1.3%).  The primary outcome was weighted area under the pain-at-rest curve (AUC) during the initial 48 postoperative hours. As defined by the authors, weighted AUC is AUC divided by the monitoring period, which is generally referred to as time-weighted average. Addition of liposomal bupivacaine improved weighted AUC compared with standard bupivacaine: 0.6 versus 1.4 points (P < 0.001). This result raises two fundamental questions. The first is whether weighted AUC adequately characterizes the temporal nature of pain. And the second is whether the improvement was clinically meaningful.

AUC is widely used to characterize outcomes that vary over time. Often it is the best single summary. For example, area is commonly used to collapse duration and magnitude of desaturation or hypotension into a single value. But a disadvantage is that by collapsing results over time, temporal resolution is lost—resolution that may be particularly important when the underlying clinical question is whether liposomal bupivacaine prolongs analgesia duration. Since liposomal bupivacaine costs about 100 times as much as standard bupivacaine, how much longer the drug lasts is of substantial interest.

A problem with AUC is illustrated in two pivotal liposomal bupivacaine trials that led to Food and Drug Administration approval.  Each randomized patients having bunionectomy or hemorrhoidectomy to infiltration of the surgical site with either liposomal bupivacaine or a placebo. Both reported statistically significant differences in their primary endpoint, AUC pain scores for the first 48 and 72 h, respectively, leading to claims of a 48- to 72-h benefit.  Subsequently, after reviewing the primary data set (fig. 1), the Food and Drug Administration concluded that “the two treatments (bupivacaine liposomal and placebo) differed significantly and clinically only during the first 24 h.”  Their reasoning was that while AUC differed significantly at 48 and 72 h compared with placebo, essentially all the difference accrued during the initial 12 to 24 h, where even plain bupivacaine would be expected to provide substantial analgesia. This example illustrates that while AUC is a reasonable estimate of aggregate effect over time, the metric can be misleading when outcomes vary directionally over time—and thus poorly estimates benefit duration.

Fig. 1.
Pain intensity versus time plot showing the difference in effect on mean ± SD pain with liposomal bupivacaine compared to placebo for (A) hemorrhoidectomy and (B) bunionectomy surgical site infiltration. Note that while the primary outcomes were the area under the curve for the first 72 and 48 h, respectively, and were positive for each, no differences were found at individual time points after 24 h. In other words, although liposomal bupivacaine was not found superior to placebo after the first 24 postoperative hours, the positive primary outcomes implied a duration of 48 to 72 h. Reproduced from Ilfeld. Anesthesiology 2021; 134:283–344.

Pain intensity versus time plot showing the difference in effect on mean ± SD pain with liposomal bupivacaine compared to placebo for (A) hemorrhoidectomy and (B) bunionectomy surgical site infiltration. Note that while the primary outcomes were the area under the curve for the first 72 and 48 h, respectively, and were positive for each, no differences were found at individual time points after 24 h. In other words, although liposomal bupivacaine was not found superior to placebo after the first 24 postoperative hours, the positive primary outcomes implied a duration of 48 to 72 h. Reproduced from Ilfeld. Anesthesiology 2021; 134:283–344.

To their credit, Chan et al. provide day-by-day outcomes. Pain scores with liposomal bupivacaine were significantly lower (better) on postoperative day 1, with mean values at rest being 0.5 (95% CI, 0.3 to 0.8) versus 1.9 (95% CI, 1.3 to 2.5) points (P < 0.001) on the numeric rating scale.  An even larger difference was detected for pain during movement on postoperative day 1, with mean values being 2.7 (95% CI, 2.0 to 3.3) versus 4.9 (95% CI, 4.2 to 5.6) points (P < 0.001). Importantly, though, there were no significant difference in pain scores at rest or with movement on any subsequent day. Nor were there any differences in opioid consumption, which matters because rescue opioid use can mask an ineffective local anesthetic approach. By the second postoperative day, pain scores no longer differed significantly (with correction for multiple comparisons): 0.9 versus 1.3 points at rest and 2.8 versus 3.7 points with movement. The results of the study by Chan et al. thus provide weak evidence for a prolonged benefit from liposomal bupivacaine.

This brings us to our second question: What difference should be considered clinically meaningful? Here we distinguish between statistical significance (very roughly how likely the results are to be correct) and what magnitude might reasonably matter to patients and clinicians. The first can be calculated; the second is based on myriad factors in addition to patient perception including treatment cost, potential toxicity, and clinical judgment. Both statistical significance and clinical meaning are important since statistically significant findings can be of trivial magnitude (especially in observational research with large data sets) and findings can be clinically meaningful without being statistically significant (as in marginally powered trials).

Determining clinically meaningful differences requires distinguishing individuals from treatment populations, or groups. How to estimate the smallest improvements in analgesia important to individuals is well described (fig. 2).  For example, the minimum improvement in numeric rating scale deemed meaningful for individual patients with acute pain is thought to be 1.0 to 2.0 points. However, trials aggregate results into groups of patients, and what intergroup difference is clinically meaningful remains poorly characterized.  Somewhat confusingly, the smallest change that is important to individual patients cannot be extrapolated to the evaluation of differences between groups in clinical trials.  While covered in detail elsewhere  in general, the minimum meaningful analgesic improvement between groups of patients is smaller than for individual patients. Consequently, if intergroup differences are compared to meaningful thresholds based on individual preferences, worthwhile medications/interventions might inappropriately be dismissed as clinically unimportant.

Fig. 2.
A hypothetical example of an “anchor-based” method to determine the smallest improvement in analgesia that individual patients consider relevant or important. This example is simplified to convey the general concept; reality is (unfortunately) more complex. Adapted from Ilfeld. Reg Anesth Pain Med 2022; 47:395–400 (used with permission).13

A hypothetical example of an “anchor-based” method to determine the smallest improvement in analgesia that individual patients consider relevant or important. This example is simplified to convey the general concept; reality is (unfortunately) more complex. Adapted from Ilfeld. Reg Anesth Pain Med 2022; 47:395–400 (used with permission).

For example, consider a situation in which a decrease of 2 points on the 0 to 10 numeric rating scale is considered clinically meaningful by individuals.  It is easy to imagine a situation in which a substantial fraction of treated patients have pain reductions of 2 points or more but the average reduction across the population (group) remains less than 2 points (fig. 3).  Additional examples of why clinical differences deemed relevant to individual patients cannot be extrapolated to the evaluation of differences between groups in clinical trials include the placebo effect, spontaneous resolution, and regression to the mean (fig. 4).

Fig. 3.
A hypothetical example of extrapolating the smallest change that is important to individual patients to the evaluation of group differences. This example is simplified to convey the general concept; reality is (unfortunately) more complex. Adapted from Ilfeld. Reg Anesth Pain Med 2022; 47:395–400 (used with permission).13

A hypothetical example of extrapolating the smallest change that is important to individual patients to the evaluation of group differences. This example is simplified to convey the general concept; reality is (unfortunately) more complex. Adapted from Ilfeld. Reg Anesth Pain Med 2022; 47:395–400 (used with permission). 

Fig. 4.
A hypothetical example of how the placebo effect decreases the smallest important difference between two groups. This example is simplified to convey the general concept; reality is (unfortunately) more complex. The placebo effect can influence the group of participants receiving active treatment; however, the effect is far greater on the group receiving placebo. NRS, numeric rating scale. Adapted from Ilfeld. Reg Anesth Pain Med 2022; 47:395–400 (used with permission).13

A hypothetical example of how the placebo effect decreases the smallest important difference between two groups. This example is simplified to convey the general concept; reality is (unfortunately) more complex. The placebo effect can influence the group of participants receiving active treatment; however, the effect is far greater on the group receiving placebo. NRS, numeric rating scale. Adapted from Ilfeld. Reg Anesth Pain Med 2022; 47:395–400 (used with permission). 

Do the between-group differences of 1.4 to 2.2 reported by Chan et al. indicate a meaningful analgesic improvement to the participants? We believe that they do because (1) the minimum improvement in numeric rating scale deemed meaningful for individual patients with acute pain has been estimated to be 1.0 to 2.0 points and (2) clinically relevant differences between group means tend to be smaller than for individual patients thus, the results suggest that the benefit to trial participants was clinically meaningful on postoperative day.  Consistent with this theory is the observation that during the first 24 h, only 28% of participants given liposomal bupivacaine experienced moderate or severe pain (numeric rating scale, 4 or more) compared with 68% for the bupivacaine-only group. Furthermore, liposomal bupivacaine was associated with lower Overall Benefit of Analgesia Scores during the first 48 h (although not on individual days).

While the analgesic improvement from liposomal bupivacaine on postoperative day 1 suggests that the patients themselves valued the difference, determining overall clinical importance requires considering additional factors. For example, acetaminophen produces minimal—if any—benefits after knee and hip arthroplasty  yet remains included in every Enhanced Recovery After Surgery (ERAS) protocol because there are few reasons to not include it. Conversely, an analgesic with a greater risk of side effects or toxicity will usually require a greater analgesic benefit to be considered warranted. These examples demonstrate that the clinical meaningfulness of group differences in pain scores cannot alone determine the overall clinical importance of that difference: tolerability, safety, convenience, durability of benefits, and cost along with a host of additional factors must be considered including comparisons with alternatives.  Fortunately, the available evidence suggests there to be little difference between unencapsulated and liposomal bupivacaine for most of these considerations leaving cost as the primary differentiating factor. Due to differing clinical scenarios and financial priorities—among other issues—various clinicians, administrators, payers, and researchers will likely draw differing conclusions on the value of adding liposomal bupivacaine.

Conclusions

Chan et al. evaluated liposomal bupivacaine for supraclavicular brachial plexus blocks (a use that is not currently Food and Drug Administration approved). Based on 48-h pain score–weighted AUC (time-weighted average), they conclude that “addition of liposomal bupivacaine in the supraclavicular brachial plexus block reduced postoperative pain at rest in the first 48 h after distal radial fracture surgery compared to standard bupivacaine alone.” Analysis of daily pain scores reveals a more nuanced picture in which benefit was largely restricted to the first day and was of a clinically meaningful magnitude to the participants. There was considerably less benefit on the second postoperative day. Whether the observed benefit justifies the additional cost of encapsulated bupivacaine will depend on circumstances. We also note that the amount and duration of analgesia from liposomal bupivacaine are context sensitive and will presumably differ for other nerve blocks and infiltration applications. We look forward to additional trials comparing liposomal and plain bupivacaine for various uses.