Article Text
PDF
Abstract
Periarticular local infiltration analgesia (LIA) is integral to multimodal analgesia following total knee arthroplasty (TKA); however, the duration of analgesia using traditional long-acting local anesthetics is often insufficient. LIA with slow-release liposomal bupivacaine may provide extended analgesia, but evidence of efficacy beyond the first 24 hours is conflicting. This meta-analysis compares the effects of periarticular liposomal and plain bupivacaine LIA on day 2 analgesic outcomes post-TKA. Trials comparing liposomal and plain bupivacaine LIA for TKA were sought. The two coprimary outcomes were (1) cumulative oral morphine equivalent consumption and (2) difference in area under the curve (AUC) of pooled rest pain scores on day 2 (24–48 hours) post-TKA. We also evaluated pain and analgesic consumption on day 3 (48–72 hours), functional recovery, length of hospital stay, patient satisfaction; and opioid-related side effects. Data were pooled using random-effects modeling. Seventeen trials (1836 patients) were analyzed. Comparing liposomal versus plain bupivacaine LIA for TKA failed to detect differences in morphine consumption and pain AUC on day 2 postoperatively, with mean differences of 0.54 mg (95% CI −5.09 to 6.18) and 0.08 cm/hour (95% CI −0.19 to 0.35), respectively (high-quality evidence). Secondary outcome analysis did not uncover any additional analgesic, functional or safety advantages to liposomal bupivacaine on postoperative day 2 or 3. Results indicate that liposomal and plain bupivacaine LIAs are not different for extended postoperative analgesic outcomes, including pain control, opioid consumption, as well as functional and safety outcomes on days 2 and 3 post-TKA. High-quality evidence does not support using liposomal bupivacaine LIA for TKA.
- regional anesthesia
- analgesia
- pain
- postoperative
Statistics from Altmetric.com
Background
The duration of moderate to severe pain following total knee arthroplasty (TKA) can persist for up to 7 days1 2 postoperatively and is associated with poor functional recovery,3 satisfaction and quality of life measures,4 along with delayed discharge from hospital and greater risk of persistent postsurgical pain.5 6 Periarticular local infiltration analgesia (LIA) has become an integral component of multimodal analgesia following TKA7; however, the duration of analgesia associated with LIA using traditional long-acting local anesthetics is often insufficient.8
In 2011, the US Food and Drug Administration approved the use of slow-release liposomal bupivacaine for LIA,9 which has since been reported to extend the duration of analgesia up to 72 hours compared with traditional plain long-acting local anesthetics following TKA.10 However, clinical evidence supporting the superiority of the analgesic effects of liposomal bupivacaine over plain local anesthetics for periarticular LIA in the setting of TKA continues to be conflicting.10–21 Indeed, several recent trials11 12 16 19–22 and systematic reviews23–32 have suggested that liposomal bupivacaine may not confer any additional benefit compared with plain bupivacaine for periarticular LIA in TKA; however, these trials have primarily focused on acute analgesic outcomes within the first 24 hours following TKA. In order to understand the relative analgesic benefits of liposomal bupivacaine beyond the first 24 hours postoperatively, we aimed to investigate whether or not liposomal bupivacaine provides extended analgesia by evaluating its incremental effects on postoperative pain and opioid consumption on day 2 (24–48 hours) following TKA in comparison to plain bupivacaine for periarticular LIA.
Methods
The authors adhered to the Preferred Reporting Items for Systematic Reviews and Meta-Analyses statement guidelines in preparing this article.33 We sought randomized controlled trials that evaluated pain severity and analgesic consumption in patients who received periarticular LIA with liposomal bupivacaine, compared with plain bupivacaine on postoperative day 2 following TKA. Relevant studies were evaluated using a predesigned protocol that was registered with the International Prospective Register of Systematic Reviews under the number CRD42020176733.
Eligibility criteria
Trials that allocated adult patients (≥18 years) undergoing unilateral or bilateral TKA with liposomal bupivacaine LIA were sought. The comparator of interest was plain bupivacaine LIA. Only studies that evaluated extended analgesic benefits by examining postoperative analgesic outcomes on day 2 (or beyond) following TKA were considered. To account for the diversity of clinical practice in this population, we considered infiltration mixtures, including adjuncts (ie, opioids, non-steroidal anti-inflammatory drugs, steroids or adrenergic agonists) as long as these were used in both study arms. Since our time frame of interest was greater than 24 hours postoperatively, we included trials where general anesthesia, single-injection neuraxial anesthesia and/or single-injection peripheral nerve blocks were administered. Trials using continuous neuraxial or peripheral nerve blocks (eg, epidural catheter, adductor canal catheter or femoral nerve catheter) were excluded. The scope of this review was confined to a single administration of periarticular LIA; studies using continuous catheter-based LIA infusions were excluded. Studies available as abstracts were not considered unless the full text was available from authors, and any foreign language studies were translated using an online translator.
Literature search
A systematic search strategy was created by NH and used to search the US National Library of Medicine database (MEDLINE); Cochrane Database of Systematic Reviews; and Excerpta Medica Database databases from inception to 1 January 2020. The search strategy was based on the search generated for MEDLINE (appendix A). The strategy contained key words related to the following: liposomal bupivacaine, TKA, local infiltrative analgesia, local anesthetic infiltration and analgesia. The reference lists of potentially eligible citations were also manually searched to identify additional trials that fulfilled inclusion criteria.
Selection of included studies
Two reviewers (NH and BTS) independently screened the titles and abstracts yielded by the literature search. The full texts of potentially eligible citations were then retrieved and evaluated for inclusion. Any disagreement between the two reviewers was discussed until a consensus was reached. If consensus could not be reached, a third reviewer (FWA) made the final decision.
Data extraction
A data extraction form was created and piloted by an independent reviewer (NH). Data extraction was subsequently carried out in duplicate by two independent reviewers (NH and BTS). Any discrepancies in data extraction were discussed until a consensus was reached. If consensus could not be reached, a third reviewer (FWA) made the final decision. The data extraction form collected information regarding the following: year of publication; participant age; publication year; type of surgery; anesthetic technique (ie, general anesthesia vs spinal anesthesia); type of peripheral nerve block (if applicable); nature and dose of local anesthetic infiltration (whether liposomal or plain); use of adjuncts; preoperative, intraoperative and postoperative analgesic regimens; primary outcome studied; rest and dynamic pain scores at all reported times; analgesic consumption at all reported times; satisfaction with pain relief; functional outcomes; opioid-related side effects; infiltration-related complications; liposomal bupivacaine-related adverse effects; and hospital and postanesthesia care unit (PACU) length of stay. We also extracted information on long-term arthroplasty outcomes, including incidence of persistent postsurgical pain, health-related quality of life, opioid dependence and pain-related disability. The primary source of data was numerical data presented in tables and figures. Data reported in graphical form were extracted with a graph digitizing software (GraphClick, Arizona Software, USA).
Assessment of methodological quality and risk of bias
The methodological quality of included trials was evaluated using the Cochrane Collaboration tool for risk of bias assessment.34 The risk of methodological bias in each study was rated as unclear, low or high risk for the following parameters: random sequence generation, allocation concealment, blinding of study personnel and outcome assessment, selective outcome data reporting and loss to follow-up. In addition, the methodological quality for each outcome pooled across trials was assessed using the Grades of Recommendation, Assessment, Development, and Evaluation35 36 guidelines. The strength of evidence was then rated as being of high quality (⊕⊕⊕⊕), moderate quality (⊕⊕⊕⊝), low quality (⊕⊕⊝⊝) or very low quality (⊕⊝⊝⊝) evidence.
All quality assessments were done in duplicate by two independent reviewers (NH and BTS). Any discrepancies in quality assessment were discussed until a consensus was reached. If consensus could not be reached, a third reviewer (FWA) made the final decision.
Primary and secondary outcomes
As liposomal bupivacaine LIA for TKA has been purported to provide extended analgesia lasting well beyond the first 24 hours postoperatively,10 we primarily focused on day 2 (24–48 hour interval) analgesic outcomes to best capture the potential incremental benefits of liposomal bupivacaine, compared with plan bupivacaine. Earlier investigations have already indicated lack of difference between liposomal and plain bupivacaine LIA on postoperative day 1 (interval of 0–24 hours) following TKA; therefore, we purposefully avoided examining day 1 short-term analgesic outcomes in order to preserve statistical power and discourage any inflated risk of type I error by testing multiple outcomes. Our two coprimary outcomes were designated as (1) cumulative postoperative oral morphine equivalent consumption (mg) and (2) difference in the area under the curve (AUC) of the weighted pooled rest pain scores on postoperative day 2 (ie, interval of 24–48 hours) following TKA. An AUC over an interval was favored over a single time point analysis to account for duration of pain severity and to allow detecting variations in the analgesic efficacy (onset/offset) beyond the first 24 hours postoperative period.
The secondary analgesic outcomes evaluated in this review included cumulative postoperative oral morphine equivalent consumption (mg) on day 3 (48–72 hours). We also examined individual postoperative rest and dynamic pain severity (Visual Analog Scale (VAS)) at 24, 48 and 72 hours; hospital length of stay (hours); and postoperative functional recovery during hospital stay and at 3 months following arthroplasty. Additionally, we assessed postoperative opioid-related side effects (postoperative nausea and vomiting, sedation/respiratory depression, pruritus, hypotension, urinary retention or constipation); infiltration-related complications (ie, hematoma, local anesthetic systemic toxicity and wound dehiscence)8; and liposomal bupivacaine-specific adverse effects (ie, hypesthesia, fever and pruritus)37 during days 1, 2 and 3 postoperatively. Finally, we examined long-term arthroplasty outcomes, including the risk of persistent postsurgical pain, health-related quality of life, opioid dependence and pain-related disability.
Measurement of outcome data
All opioid consumption data were converted to oral morphine equivalents.38 Similarly, all pain severity data were converted to a VAS score of 0–10 cm.39 Measures of patient satisfaction were also converted to a VAS equivalent score (0=least satisfied, and 10=most satisfied).40 Time-to-event data were presented in hours. Scales assessing postsurgical functional recovery were anticipated to be diverse, although measuring the same theme. Therefore, we pooled the various measures according to a prespecified objectivity-based hierarchy (see further). Finally, when a study reported cumulative opioid consumption over an interval exceeding 24 hours, we used the pooled results to estimate the proportion of consumption for our interval of interest.
Statistical analysis
The mean and SD were sought for all continuous outcomes. When these were not available, the median and IQR41 were used to approximate these values. Similarly, if needed, the 95% CI was used to estimate the SD value.42 If no measure of variance was provided, the value of the SD was imputed as a last resort.43 Finally, when needed, dichotomous data were converted to continuous data, and the means and SD were used.44
To permit the evaluation of the effect of liposomal bupivacaine on postoperative functional recovery despite of the anticipated diversity of scales, we a priori planned to report (1) the weighted mean difference (WMD) of all studies that used the same continuous scale; (2) the log(OR) if trials reported continuous outcomes used different scales that measured the same theme; and (3) an OR if all trials reported binary outcomes. If scenario 2 was applicable, the conversion to log(OR) from a standardized mean difference (SMD) was done using the formula log(OR)=SMD (π/√3).45 46 This analysis was performed under the assumption that the mean scores for each group followed a logistic distribution, and that variances were equal between the two groups.
Meta-analysis
For continuous outcomes, pooling was performed using the inverse variance method since we anticipated clinical heterogeneity between studies. For dichotomous outcomes, pooling was performed using the Mantel-Haenszel random-effects model.47 For primary outcomes, a WMD with 95% CI was calculated and a p value of <0.025 was designated as the threshold of statistical significance.
For functional recovery, either a WMD, log(OR) or OR and the 99% CI was calculated, depending on the nature of reported data (as described previously). For the remaining continuous outcomes, a WMD with 99% CI was calculated. For dichotomous outcomes, an OR with 99% CI was calculated. The 99% CI was used for all secondary outcomes to decrease the risk of type I error associated with multiple testing. To that end, we also used a threshold for statistical significance adjusted by the Bonferroni-Holm correction48 to account for the several secondary outcomes.
Data were only statistically pooled when it was available from at least two trials. Qualitative reporting was used for outcomes with data available from less than two trials. No additional post hoc analyses were performed.
Interpretation of primary outcome results
Results of our primary outcomes were interpreted in light of the minimal clinically important difference.49 For cumulative opioid consumption during the interval of 24–48 hour s, we considered a 30 mg difference in oral morphine50 (or 10 mg intravenous morphine) to be clinically important. Similarly, the AUC analysis was interpreted in light of the minimal clinically important difference of 2 cm/hour for one time interval (ie, 24–48 hours).51 52
Assessment of heterogeneity
The extent of statistical heterogeneity was assessed by calculating the I2 statistic, with values of >50% indicating significant heterogeneity. When this threshold was met for the primary outcomes, an additional metaregression analysis using mixed-effects modeling was conducted to explore the interaction between a priori specified clinical predictors and the treatment effect. An R2 value (coefficient of determination) was calculated to quantify the extent to which each covariate explained the variation in data. An R2=1 meant that the covariate explained all the variability, while an R2=0 meant that the covariate did not explain any of the variability. Metaregression analysis was performed only for covariates when data were available from at least two trials. The covariates considered were (1) type of surgical anesthesia (general vs regional), (2) dose of liposomal bupivacaine (mg), (3) mixing plain and liposomal bupivacaine; (4) inclusion of adjuvants in infiltration solution; (5) use of supplemental peripheral nerve blocks; (6) duration of action of peripheral nerve blocks (single injection vs continuous infusion); (7) use of intrathecal morphine; (8) use of long-acting opioids (ie, morphine) intraoperatively; (9) postoperative analgesic modality (multimodal: combination of opioid and other adjuvants vs unimodal: opioids only); and (10) methodological quality (low and intermediate risk of bias vs high risk of bias). When metaregression could not be performed on a specific covariate (<2 trials), sensitivity analysis was conducted. As there is a concern that evidence from industry-supported trials may introduce bias,53–55 we also planned a sensitivity analysis on this variable to explore the influence of including industry sponsored trials.
Assessment of publication bias
The risk of publication bias was assessed using the Egger’s regression test when data from at least three trials were available for an estimate effect.56
Data management
Forest and funnel plots were generated using Review Manager Software (RevMan V.5.2, Nordic Cochrane Center, Cochrane Collaboration). Metaregression was performed using Comprehensive Meta-Analysis 3.0 (Engelwood, USA).
Results
Our literature search identified 373 unique citations. A total of 351 citations were excluded after title and abstract screening, because of incorrect study design (n=62), incorrect patient population or intervention (n=274), or for being an animal or in vitro study (n=15). The remaining 22 citations had their full texts reviewed. Of these, five were excluded for being duplicate publications (n=3)14 17 57 or having an incorrect comparator group (n=2).58 59 Subsequently, a total 17 full-text randomized trials comparing liposomal versus plain bupivacaine LIA for TKA10–13 15 16 18–22 60–65 were included in this systematic review and meta-analysis. The authors of two studies12 16 provided additional details for this review. The flow diagram for study inclusion is depicted in figure 1.
Study flow diagram.
Study characteristics
The study characteristics and outcomes included in this review are presented in table 1. The 17 trials10–13 15 16 18–22 60–65 included a total of 1836 patients (1862 knees), of which 1810 underwent unilateral TKA10–13 15 16 18–22 60–62 64 65 and 26 underwent bilateral TKA.63 Infiltration with liposomal bupivacaine was performed in 920 knees, while plain bupivacaine infiltration was used in 942 knees. All 17 studies10–13 15 16 18–22 60–65 evaluated the primary outcomes of interest, analgesic outcome on day 2 following arthroplasty. The risk of bias assessment for the included studies is presented in figure 2.
Risk of bias summary.
Study characteristics and outcomes of interest assessed in included studies
The technical details and solutions used for infiltration, as well as the postoperative analgesic regimens administered, are summarized in table 2. Arthroplasty was performed under spinal anesthesia in 11 studies,11 13 16 18–21 60 61 63 64 under general anesthesia in 2 studies,10 15 either of the two in 2 studies,12 62 under total intravenous anesthesia in 1 study,22 and 1 study15 did not specify the surgical anesthetic used. As for use of supplemental peripheral nerve blocks, four studies16 22 64 65 combined spinal anesthesia with nerve blocks, with adductor canal block used in three studies16 64 65 and a combined femoral–sciatic blocks in one study.22
Local anesthetic techniques for liposomal bupivacaine and analgesic regiments of included studies
The timing of periarticular infiltration was preimplantation in 16 trials10 12 13 15 16 18–22 60–65 and postimplantation in one.11 The tissues targeted by periarticular infiltration included the knee capsule, periosteum, surrounding soft tissue, tendons and ligaments, suprapatellar pouch, arthrotomy site and subcutaneous tissue. The dose of liposomal bupivacaine ranged between 133 and 532 mg, while the volume ranged between 40 and 120 mL10–13 15 16 18–22 60–65; and four trials22 60 61 64 did not detail the liposomal bupivacaine dose. The adjuncts added to the solution included epinephrine in four trials,11–13 15 ketorolac in three trials12 13 15 and morphine in one trial.15 Ten trials mixed liposomal and plain bupivacaine.12 13 15 16 18–20 61 63 64
Primary outcomes
Cumulative morphine consumption on day 2
Fourteen studies11–13 15 16 18–22 61 62 64 65 inclusive of 1392 patients (liposomal: 636, plain: 756) reported oral morphine equivalent consumption during day 2 following arthroplasty. Liposomal bupivacaine was not different from plain bupivacaine for oral morphine equivalent consumption, with a WMD of −0.54 mg (95% CI −5.09 to 6.18) (p=0.85, I2=46%) (figure 3). The results were not characterized by significant heterogeneity (I2 <50%) threshold, mitigating the need for any metaregression or sensitivity analysis. However, preplanned sensitivity analysis to explore any potential biases introduced by including industry-sponsored studies suggested that the magnitude of the estimate of effect was robust when these trials were excluded.12 13 15 18 The risk of publication bias was low (p=0.06), and the quality of evidence was rated high.
Forest plot of cumulative oral morphine equivalent consumption at postoperative day 2 (24–48 hours) for liposomal bupivacaine versus plain bupivacaine. Pooled estimates of the weighted mean difference are shown with 95% CIs. Pooled estimates are represented as diamonds and lines represent the 95% CIs.
AUC of rest pain on day 2
The weighted means of the pooled rest pain scores for the interval of 24–48 hours were calculated and used to estimate the mean difference between liposomal and plain bupivacaine in the AUC (95% CI) for rest pain. This estimate was derived from 1520 patients (liposomal: 775, plain: 745) at 24 hours,10–13 15 16 19 20 22 60–62 64 65 and from 1313 patients (liposomal: 631, plain: 682) at 48 hours.10–13 15 16 19 20 22 61 62 64 65 The mean difference in AUC for rest pain scores for day 2 following arthroplasty was 0.08 cm/hour (95% CI −0.19 to 0.35) (p=0.56), indicating absence of difference between liposomal and plain bupivacaine (figure 4). Sensitivity analysis by eliminating industry-sponsored trials suggested that the magnitude of the estimate of effect was robust when these trials were excluded.10 12 13 15 The quality of evidence for each time point included in the analysis was rated high.
Graphical representation of the area under the curve of the pooled weighted mean pain scores at rest as measured by the Visual Analog Scale (0–10 cm) over the time interval of 24–48 hours for liposomal bupivacaine versus plain bupivacaine.
Secondary analgesic outcomes
Morphine consumption on day 3
Five studies,11 19 21 22 62 inclusive of 482 patients (liposomal: 244, plain: 238), evaluated oral morphine equivalent consumption during the interval of 48–72 hours. Liposomal bupivacaine was not different from plain bupivacaine for this outcome (table 3). The quality of evidence was rated as high.
Secondary endpoint results
Postoperative rest pain severity at individual time points
Rest pain severity scores were not different between liposomal and plain for any of the time points examined on days 1, 2 and 3. Compared with plain bupivacaine, liposomal bupivacaine did not improve rest pain at 24 hours (data from 1520 patients),10–13 15 16 19 20 22 60–62 64 65 48 hours (data from 1313 patients)10–13 15 16 19 20 22 61 62 64 65 and 72 hours (data from 661 patients)11 13 19 22 62 postoperatively (table 3). The quality of evidence was high.
Postoperative dynamic pain severity at individual time points
Compared with plain bupivacaine, liposomal bupivacaine did not improve dynamic pain control at 24,21 22 63 4821 22 63 and 7222 hours postoperatively (table 3). The quality of evidence was low due to heterogeneity and scarcity of data.
Length of PACU stay
None of the studies included in the review evaluated this outcome.
Length of hospital stay
Ten studies,11–13 15 19 20 22 60 64 65 inclusive of 1271 patients (liposomal: 524, plain: 747) reported time to hospital discharge. Overall, liposomal bupivacaine was not different from plain bupivacaine for this outcome (table 3). The quality of evidence of the pooled estimate was rated as high.
Functional recovery during hospital stay
Seven studies,13 15 16 19–22 inclusive of 769 patients (liposomal: 315, plain: 454) reported functional recovery during hospital stay. A variety of scales were used to evaluate knee function; these included (1) Knee Society Score,13 (2) knee range of motion,19 21 22 (3) activity measure for Post-Acute Care Basic Mobility Inpatient Score,16 (4) time to ambulation15 and (5) total distance walked,20 in this order based on objectivity. The use of continuous measurements, exclusively, facilitated pooling and calculating the log(OR). Overall, liposomal bupivacaine was not different from plain bupivacaine for this outcome (table 3). The quality of evidence of the pooled estimate was rated as high.
Postoperative function at 3-month follow-up
Only one study12 reported this outcome, precluding statistical pooling. Qualitatively, this study did not detect any difference between liposomal and plain in functional outcomes for at 3 months following arthroplasty.
Patient satisfaction
Two studies,11 62 inclusive of 174 patients (liposomal: 86, plain: 88) reported patient satisfaction. The use of liposomal bupivacaine was not associated with improved satisfaction compared with plain bupivacaine (table 3). The quality of evidence was low due to heterogeneity and scarcity of data.
Secondary safety outcomes
Opioid-related side effects
Seven studies10 13 18–20 62 65 inclusive of 829 patients (liposomal: 403, plain: 426) reported the incidence of opioid-related side effects; however, none of the studies stratified these side effects according to incidence on days 1, 2 and 3 following knee arthroplasty. Additional (not preplanned) post hoc analysis of these data performed irrespective of the time of incidence suggested that liposomal bupivacaine was not different from plain bupivacaine for this outcome (table 3). Quality of evidence was rated moderate because of pooling over different follow-up periods.
LIA-related complications
Eight studies inclusive of 965 patients (liposomal: 477, plain: 488) reported the incidence of LIA-related complications; however, none of the studies stratified these complications according to incidence on days 1, 2 and 3 following knee arthroplasty. Additional (not preplanned) post hoc analysis of these data performed irrespective of the time of incidence suggested that liposomal bupivacaine was not different from plain bupivacaine for this outcome (table 3). Quality of evidence was rated moderate because of pooling over different follow-up periods.
Liposomal bupivacaine adverse effects
Two studies inclusive of 277 patients (Liposomal: 174, Plain: 103) reported the incidence of liposomal bupivacaine adverse effects; however, none of the studies stratified these complications according to incidence on days one, two and three following TKA. Additional (not pre-planned) post-hoc analysis of these data performed irrespective of the time of incidence suggested that liposomal bupivacaine was not different from plain bupivacaine for this outcome (table 3). Quality of evidence was rated low because of pooling over different follow-up periods and scarcity of data.
Secondary long-term analgesic outcomes
None of the studies included in this review evaluated the incidence of chronic postsurgical pain, quality of life, long-term opioid dependence or pain-related disability.
Discussion
This systematic review and meta-analysis provides high-quality evidence undermining the role of liposomal bupivacaine for periarticular LIA following TKA. Our primary analysis of day 2 analgesic outcomes revealed no difference between liposomal and plain bupivacaine in both AUC of rest pain and analgesic consumption during the interval of 24–48 hours. Furthermore, the effect of liposomal bupivacaine was not different from that of plain bupivacaine for postoperative rest and dynamic pain severity, as well as analgesic consumption at 3 days following TKA. Moreover, liposomal bupivacaine also failed to reduce opioid-related side effects, shorten hospital length of stay, enhance safety or improve functional outcomes. Importantly, our findings appear to contradict clinical evidence from an industry-supported phase IV multicenter double-blind trial involving 16 US hospitals,18 which reported a 78% reduction in opioid consumption in the first 48 hours, a 13.6% improvement in the AUC pain scores over the interval of 12–48 hours, and a 10% rate of opioid-free knee arthroplasty (up to 3 days postoperatively). Although we had selected similar outcome measures for the present quantitative review, our systematic data pooling and evidence synthesis could not reproduce these findings.18 One plausible explanation for the differences between our findings herein and the aforementioned trial are several methodological and statistical concerns pertaining to the latter that have been elucidate in a recent letter to the editor.66 Finally, our results are also consistent with those of recent trials comparing liposomal versus plain bupivacaine in the setting of field blocks67 and peripheral nerve blocks.68 69
Our findings are novel and clinically pertinent. While other reviews articles24–31 have attempted to answer the same question, earlier efforts to pool data from these same trials24–31 have been undermined by a (1) paucity of available evidence; (2) pooling results of observational studies with those of randomized trials24–28 31; (3) inclusion of continuous periarticular infusions29 31; (4) choice of primary outcomes focusing on day one analgesia24–26 28 30; (5) inability to refute the claim18 that infiltration with liposomal bupivacaine provided a superior AUC of rest pain beyond 12 hours following TKA; and (6) consistently high degree of statistical heterogeneity,24 26–28 31 undermining the precision, validity and generalizability of findings. With such limitations, it is not surprising that some earlier systematic reviews suggested that using liposomal bupivacaine for periarticular infiltration improved pain control27 28 30 and reduced opioid consumption31 beyond 24 hours. In contrast, focusing on day 2 analgesic outcomes allowed us to generate consistent high-quality definitive evidence marked by precision and low heterogeneity. This novel evidence indicates that using liposomal bupivacaine for periarticular infiltration in TKA does not resolve the problem of mismatch between pain trajectory and duration of analgesia.
Our review has several notable strengths. First, given our exhaustive search strategy, we were uniquely positioned to produce the most inclusive knowledge synthesis that included 17 trials. Second, and as a result of this, we were able to generate high quality evidence for several clinically important outcomes which have not been previously investigated, such as AUC pain analysis and analgesic consumption during days 2 and 3 postoperatively. Third, our primary outcomes were characterized by a large number of included patients (>1000), allowing us to generate precise effect estimates. Fourth, the majority of our outcomes were associated with a low level of heterogeneity, which increases the external validity of our findings. Finally, through the calculation of 99% CI for all secondary outcomes and very stringent thresholds of statistical significance, we were able to reduce the risk of type I error and multiple testing bias.
Our review also has limitations. First, due to inconsistent reporting, we were unable to provide estimates of effect for several safety outcomes, including liposomal bupivacaine adverse events. Second, we were also unable to evaluate important long-term outcomes such as pain-related disability, persistent pain, opioid dependence and health-related quality of life. Third, we were unable to draw any conclusions for liposomal bupivacaine compared with local anesthetics other than plain bupivacaine (eg, ropivacaine) because none of the published studies examined ropivacaine. Fourth, our AUC analysis included only two time points, which may limit its overall accuracy and external validity. Fifth, we cannot completely exclude the possibility of small study effect, but the homogeneity of results undermines this possibility. Finally, given that heterogeneity was below our predefined cut-off in our primary outcomes, we did not conduct metaregression analysis to evaluate potential predictors of treatment effect, such as inclusion of adjuncts in the infiltration solution and coadministration of peripheral nerve blocks.
Conclusions
Our results demonstrate periarticular infiltration with liposomal bupivacaine in knee arthroplasty is not different from using plain bupivacaine. High-quality evidence indicates lack of difference in pain control and opioid consumption on days 2 and 3 postoperatively. Further, liposomal bupivacaine does not appear to improve any postoperative analgesic, functional or safety outcomes. These findings do not support using liposomal bupivacaine for LIA following TKA.
References
Footnotes
Twitter @nasir418, @Faraj_RegAnesth
Correction notice This article has been corrected since it published Online First. The title has been corrected.
Contributors All authors provided equal contribution to warrant authorship.
Funding The authors have not declared a specific grant for this research from any funding agency in the public, commercial or not-for-profit sectors.
Competing interests RB receives research time support from the Evelyn Bateman Cara Operations Endowed Chair in Ambulatory Anesthesia and Women’s Health, Women’s College Hospital, Toronto.
Patient consent for publication Not required.
Provenance and peer review Not commissioned; externally peer reviewed.
Data availability statement Data are available upon reasonable request. All data relevant to the study are included in the article or uploaded as supplementary information.