Secondary Logo

Journal Logo

Pain

The efficacy of local infiltration analgesia in the early postoperative period after total knee arthroplasty

A systematic review and meta-analysis

Seangleulur, Alisa; Vanasbodeekul, Pramook; Prapaitrakool, Sunisa; Worathongchai, Sukhumakorn; Anothaisintawee, Thunyarat; McEvoy, Mark; Vendittoli, Pascal-André; Attia, John; Thakkinstian, Ammarin

Author Information
European Journal of Anaesthesiology (EJA): November 2016 - Volume 33 - Issue 11 - p 816-831
doi: 10.1097/EJA.0000000000000516

Abstract

Introduction

Local infiltration analgesia (LIA), peri or intra-articular injection, has emerged as an alternative technique for acute pain control after total knee arthroplasty (TKA) because of its efficacy in pain relief, ease of performance, low rate of infection and low rate of local anaesthetic systemic toxicity. 1–4 However, some studies did not find an advantage in pain relief of LIA over placebo.5,6 Inconsistent results were also found among previous meta-analyses. Jiang's7 (n = 16 studies) and Teng's8 (n = 10 studies) analyses supported the use of LIA in pain relief, but Keijsers’9 (n = 7 studies) did not. In addition, the efficacy of adding LIA as an adjunct to regional anaesthetic techniques (i.e. spinal morphine, epidural analgesia and peripheral nerve block) is still unclear. We hypothesised that the use of LIA as a primary technique or an adjunct to regional anaesthesia could improve clinical outcomes when compared with no injection or placebo.

Therefore, we conducted a systematic review and meta-analysis of randomised controlled trials (RCTs) which aimed to compare pain intensity and opioid consumption within 48 h after TKA between LIA, used as a primary or an adjunctive treatment and no injection or placebo. Subgroup analysis was also applied to explore the possible causes of inconsistent results from previous studies and reviews of different injection sites and durations of LIA and use of multimodal drug injection. Furthermore, mobilisation at 24 h, length of hospital stay (LOS) and complications from opioid analgesics and LIA were assessed.

Methods

We reported this systematic review according to the Preferred Reporting Items for Systematic Reviews and Meta-Analyses statement.10 The protocol of this study was not registered.

We searched two databases, Medline via PubMed and SCOPUS, from their inceptions until 14th March 2013, and updated searching in PubMed until 15th September 2015. The search terms included (((((((((’Arthroplasty, Replacement, Knee’ [MeSH Terms] OR ‘knee arthroplasty’) OR ‘total joint arthroplasty’) OR ‘knee replacement’) OR ‘knee joint arthroplasty’) OR ‘knee-arthroplasty’) OR ‘knee-replacement’) OR ‘major knee surgery’) OR ‘TKA’) OR ‘TKR’) AND ((((((((’Injections, Intra-articular’ [MeSH Terms] OR ‘wound infiltration’) OR ‘local infiltration’) OR ‘multimodal analgesia’) OR ‘anesthetic cocktail’) OR ‘intraarticular’) OR ‘intra-articular’) OR ‘intrasynovial’) OR ‘peri-articular’). Manual searching of the reference lists of all reviews and included studies was performed to identify other potentially eligible studies, and this process was performed until no additional articles could be identified.

Inclusion and exclusion criteria

Eligible studies were selected from all potential studies by two independent reviewers (A.S., P.V.), and disagreements were discussed and resolved by A.T. The eligibility criteria were Population: adult patients who underwent primary unilateral or bilateral TKA; Intervention: intraoperative peri or intra-articular knee injection of analgesic drugs with or without postoperative intra-articular catheter placement; Comparator: no injection or placebo (for those with bilateral TKA, LIA was performed on one knee and no injection or placebo on the other); Outcomes: at least one of the following outcomes reported: pain intensity at-rest or at-activity measured at 24 or 48 h, opioid consumption during 0 to 24 and 24 to 48 h, mobilisation at 24 h, LOS, opioid side-effects (nausea and vomiting, rash or pruritus, respiratory depression and urinary retention), local anaesthetic toxicity and infection rate; Methodology: only RCTs published in English. Studies were excluded when data could not be extracted after we contacted authors or measured from the available reported graphs.

Data extraction and outcome measures

All data were independently extracted in duplicate by four reviewers (A.S., P.V., S.P. and S.W.). Disagreements were resolved by discussion among reviewers. Mean, SD and the number of patients/group from each individual study were extracted for continuous outcomes; the number of patients and events were obtained for dichotomous outcomes. When data from initial searching (March 2013) were insufficient, we contacted the authors to request additional data. Data were directly extracted from graphs when there was no report from texts or tables. The reported medians along with range or interquartile were converted to mean and SD.11 When SD was not given, we estimated SD by data simulation assuming normal distribution.

Pain intensity was measured by visual analogue scale (VAS) ranging from 0 to 10. When the included studies reported VAS on a scale of 0 to 100, we converted reported scores to a scale of 0 to 10. Pain scores measured by numerical rating scale were treated like VAS. Opioid consumption referred to the amount (in milligrams) of intravenous morphine or morphine equivalents, which were converted using an opioid analgesic conversion table.12–14 Cumulative dosages within 24 and 24 to 48 h, hereafter called 24 and 48 h, were extracted.

Risk of bias assessment

Risk of bias was independently assessed in duplicate by five reviewers (A.S., P.V., S.P., S.W. and T.A), using ‘Cochrane Collaboration's tool for assessing risk of bias’.15 Any disagreement was resolved by discussion among the reviewers.

Statistical analysis

Mean differences were pooled using unstandardised mean differences for continuous outcomes and risk ratios were pooled for dichotomous outcomes across included studies. Heterogeneity was assessed using Cochrane's Q test and I2 statistics. A random effects model was used for pooling when heterogeneity was present (i.e. P value < 0.1 or I2 ≥ 25%); otherwise, a fixed effect model was applied.

Sources of heterogeneity for pain, opioid consumption and mobilisation were explored by meta-regression and subgroup analysis. Two possible sources could be different sites (peri or intra-articular injection) and durations (intra or postoperation) of injection which categorised the included studies into four groups: intraoperative peri-articular injection; intraoperative peri-articular injection plus postoperative injection or infusion through an intra-articular catheter; intraoperative intra-articular injection; postoperative injection or infusion through an intra-articular catheter. Another possible source, the addition of ketorolac to local anaesthetic for peri-articular injection, was further explored.

Publication bias was assessed by Funnel plot and Egger test. Either the asymmetric funnel plot or the significant Egger P value (P ≤ 0.05) determined the use of contour-enhanced funnel plot.16 All statistical analyses were performed using STATA version 12.0 (StataCorp. 2011. College Station, Texas, USA). Statistical significance was defined as a P value less than 0.05.

Results

Search results

Thirty-eight studies with 3026 total patients were selected from 1186 relevant studies (Fig. 1). Twenty-eight studies1,4–6,17–40 compared LIA versus no injection or placebo, and 10 studies2,41–49 compared LIA with no injection or placebo when regional anaesthetic techniques and systemic analgesia had already been used (Fig. 1). A summary of the included studies is in Table 1. The additional data of pain scores,6,21,24 opioid consumption6,21,24 and LOS24 were obtained for pooling from contacting the authors, and all included studies could be eventually pooled for at least one of the defined outcomes.

Fig. 1
Fig. 1:
Flow chart of study selection process. IAI, intra-articular injection; LIA, local infiltration analgesia; PAI, peri-articular injection; POI, postoperative injection/infusion; RCT, randomised controlled trial; TKA, total knee arthroplasty.
Table 1
Table 1:
Characteristics of included studies
Table 1
Table 1:
(Continued) Characteristics of included studies

Among 28 studies of LIA versus no injection or placebo, 12 studies4,17–27 applied intraoperative peri-articular injection and four of them4,25–27 added postoperative intra-articular catheter placement (Fig. 1). Intraoperative intra-articular injection was used in 11 studies1,5,6,29–36, and three studies38–40 compared postoperative intra-articular catheter placement with no injection or placebo (Fig. 1). Additionally, two studies compared outcomes among three interventions; that is, peri-articular injection versus peri-articular injection plus postoperative injection/infusion versus placebo for one study and peri-articular injection versus intra-articular injection versus placebo for another.28,37

All studies with intraoperative peri-articular injection without postoperative injection/infusion used multidrug injection with different doses and administration sites except Milani's study in which only local anaesthetic was injected (Table 1).17–24,28,37 Among five studies with intraoperative peri-articular injection plus postoperative injection/infusion, three studies4,27,28 evaluated the combination of ketorolac with local anaesthetic, whereas two studies25,26 used local anaesthetic without ketorolac (Table 1).

Risk of bias

The results of risk of bias assessment were summarised in Table 2. Twenty-six (68.4%) and 22 (57.9%) studies properly reported random sequence generation and allocation concealment respectively. Two-third of studies (n = 25 studies) had adequately blinded both patients and outcome assessors. Outcome assessments were also mostly well defined in 31 studies (81.6%), and the risk of selective reporting bias was very low.

Table 2
Table 2:
Risk of bias assessment of included studies

Local infiltration analgesia versus no injection or placebo

Pain scores

Pain scores at rest were pooled across 26,1,4–6,17–32,34,35,37–40 and 171,4,18–28,32,35,39,40 studies, at 24 and 48 h respectively, whereas 24 and 48-h pain scores at activity were pooled across nine studies.4,17,21–26,28 Adjunctive analgesia with systemic administration of analgesic drugs was similarly used in both intervention and control groups in all studies except two studies in which systemic opioids and NSAIDs were added in LIA but not in control group (Table 1).20,22 From overall pooling, 24 and 48-h VAS were lower in LIA than no injection or placebo with high degree of heterogeneity. The pooled mean differences 24 and 48-h VAS at rest were −1.02 [95% CI (−1.36 to −0.68); I2 = 88.6%] and −0.70 [95% CI (−1.10 to −0.31); I2 = 91.5%] and corresponding mean differences at activity were −1.34 [95% CI (−1.97 to −0.71); I2 = 93.3%] and −0.75 [95% CI (−1.26 to −0.23); I2 = 91.9%], respectively.

Subgroup and sensitivity analysis

Subgroup analysis by sites and durations of LIA injection showed that peri-articular but not intra-articular injection reduced pain at rest with the pooled mean differences of 24 and 48-h VAS −0.89 [95% CI (−1.40 to −0.38); I2 = 92.0%] and −0.67 [95% CI (−1.24 to −0.09); I2 = 93.9%], respectively (Fig. 2a, b). These corresponding pooled mean differences of VAS at activity were −0.84 [95% CI (−1.22 to −0.47); I2 = 78.4%] and −0.37 [95% CI (−0.65 to −0.09); I2 = 66.7%] (Fig. 3a, b).

Fig. 2
Fig. 2:
Comparison of VAS at rest between types of LIA and no injection or placebo. (a) 24-h VAS at rest. (b) 48-h VAS at rest. CI, confidence interval; LIA, local infiltration analgesia; WMD, weighted mean difference.
Fig. 2
Fig. 2:
(Continued). Comparison of VAS at rest between types of LIA and no injection or placebo. (a) 24-h VAS at rest. (b) 48-h VAS at rest. CI, confidence interval; LIA, local infiltration analgesia; WMD, weighted mean difference.
Fig. 3
Fig. 3:
Comparison of VAS at activity between types of LIA and no injection and no injection or placebo. (a) 24-h VAS at activity. (b) 48-h VAS at activity. CI, confidence interval; LIA, local infiltration analgesia; WMD, weighted mean difference.
Fig. 3
Fig. 3:
(Continued). Comparison of VAS at activity between types of LIA and no injection and no injection or placebo. (a) 24-h VAS at activity. (b) 48-h VAS at activity. CI, confidence interval; LIA, local infiltration analgesia; WMD, weighted mean difference.

Sensitivity analysis was performed by excluding two studies where systemic opioids and NSAIDs were added to LIA but were not used for the control group (Table 1).20,22 The results demonstrated that 24-h VAS at rest was significantly reduced but not 48-h VAS with the pooled mean differences of −0.58 [95% CI (−0.92 to −0.24); I2 = 74.3%] and −0.32 [95% CI (−0.79 to 0.15); I2 = 82.6%], respectively.

When postoperative injection or infusion was additionally administered after intraoperative peri-articular injection in five studies, VAS at rest was reduced with a pooled mean difference of −1.50 [95% CI (−1.92 to −1.08); I2 = 60.5%] at 24 h but not at 48 h (Fig. 2a, b).4,25–28 Nevertheless, the analgesic effect of this technique extended up to 48 h at activity with the corresponding pooled mean differences of −2.69 [95% CI (−2.97 to −2.41); I2 = 0%] and −1.61 [95% CI (−2.30 to −0.93); I2 = 83.4%] (Fig. 3a, b). In addition, the result from pooling based on three of five studies adding ketorolac to local anaesthetic for peri-articular injection indicated a significant reduction of 24-h VAS at rest with the mean difference of −1.77 [95% CI (−2.13 to −1.41)] and quite low degree of heterogeneity (I2 = 28.9%) [see Figure, Supplemental Digital Content 1, http://links.lww.com/EJA/A99, which illustrates meta-analysis of 24-h VAS at rest and subgroup analysis comparing the efficacy of peri-articular injection with multimodal drug injection (local anaesthetic and ketorolac) and single-drug injection (local anaesthetic)].4,27,28

Opioid consumption and side-effects

The 24-h (n = 20 studies)4–6,18–25,27,28,30,31,33–36,40 and 48-h (n = 10 studies)4,18–23,25,27,28 morphine consumptions were significantly lower with LIA than no injection or placebo with the mean differences of −13.98 mg [95% CI (−18.53 to −9.42); I2 = 98.1%] and −6.93 mg [95% CI (−9.78 to −4.08); I2 = 92.6%], respectively.

Subgroup and sensitivity analysis

Both intraoperative peri and intra-articular injection could reduce 24-h morphine consumptions compared with no injection or placebo with the pooled mean differences of −13.03 mg [95% CI (−14.94 to −11.13); I2 = 70.3%] and −5.58 [95% CI (−10.21 to −0.96); I2 = 87.3%], respectively (Fig. 4a). The reduction in morphine consumption of peri-articular injection extended up to 48 h with the pooled mean difference of −7.14 [95% CI (−11.16 to −3.11); I2 = 94.9%] (Fig. 4b). However, after excluding two studies,20,22 systemic analgesia was added in the LIA group but not in the control group, the results showed the reduction in morphine consumption within 24-h [pooled mean difference −12.01; 95% CI (−14.10 to −10.10); I2 = 55.3%] but not for 48-h [pooled mean difference −2.53; [95% CI (−5.73 to 0.68); I2 = 84.7%].

Fig. 4
Fig. 4:
Comparison of morphine consumption between types of LIA and no injection or placebo. (a) Morphine consumption 0 to 24 h. (b) Morphine consumption 24 to 48 h. CI, confidence interval; LIA, local infiltration analgesia; SD, standard deviation; WMD, weighted mean difference.
Fig. 4
Fig. 4:
(Continued). Comparison of morphine consumption between types of LIA and no injection or placebo. (a) Morphine consumption 0 to 24 h. (b) Morphine consumption 24 to 48 h. CI, confidence interval; LIA, local infiltration analgesia; SD, standard deviation; WMD, weighted mean difference.

Postoperative injection or infusion after peri-articular injection significantly reduced 24-h morphine consumption, with a pooled mean difference of −24.46 mg [95% CI (−38.98 to −9.94); I2 = 97.7%]. In addition, the reduction still remained during 24 to 48 h with the pooled mean difference of −7.59 mg [95% CI (−11.92 to −3.25); I2 = 79.3%] (Fig. 4a, b).

A total of 16 studies evaluated side-effects of opioid consumption. LIA reduced the risk of postoperative nausea and vomiting when compared with no injection or placebo with a pooled risk reduction of 0.74 [95% CI (0.56 to 0.98); I2 = 35.0%] across 14 studies,4–6,20–23,28,30,31,34,37,38,40 whereas risk of pruritus (n = 7 studies),4,21–23,28,32,34 respiratory depression (n = 7 studies),5,6,21–23,28,32 and urinary retention (n = 7 studies)21–23,27,28,32,40 did not differ with low heterogeneity (I2 = 0%).

Mobilisation and length of hospital stay

Twelve studies reported 24-h mobilisation with different methods; the most common outcome was range of motion (ROM; n = 7 studies), which were sufficient for pooling.1,4,18,19,22,23,39 The mean difference was heterogeneous (I2 = 94.0%) with a pooled mean difference of 11.83 [95% CI (5.39 to 18.26)]. After subgroup analysis, there was higher ROM with peri-articular injection [pooled mean difference 12.87; 95% CI (8.99 to 16.76); I2 = 64.2%] and with peri-articular injection plus postoperative injection/infusion group, whereas ROM after either intra or postoperative intra-articular injection/infusion did not significantly differ when compared with no injection or placebo (see Figure, Supplemental Digital Content 2, http://links.lww.com/EJA/A99, which illustrates meta-analysis of ROM when LIA is compared with no injection or placebo, and subgroup analysis compares the efficacy of different sites and durations of LIA).1,4,18,19,22,23,39

Nine studies compared LOS between LIA and no injection or placebo.4,18,20,24,25,27,33,39,40 Patients with LIA had shorter LOS than patients without injection with a pooled mean difference of −0.94 [95% CI (−1.69 to −0.23); I2 = 96.0%].

Local infiltration analgesia + regional anaesthetic techniques versus regional anaesthetic techniques

Pain scores

Adding LIA as an adjunct to regional anaesthetic techniques reduced 24-h VAS at rest with a pooled mean difference of −1.10 [95% CI (−1.64 to −0.56); I2 = 86.4%] across nine studies2,41,43–49 but did not reduce pain at 48 h across five studies.41,46–49 The results of pooled 24-h (n = 4 studies)42,47–49 and 48-h (n = 3 studies)47–49 VAS at activity also failed to show the difference in pain reduction (see Figure, Supplemental Digital content 3a-d, http://links.lww.com/EJA/A99, which illustrates meta-analysis of VAS when LIA with regional anaesthetic techniques is compared with regional anaesthetic techniques).

Opioid consumption and side-effects

The mean differences of 24 and 48-h morphine consumption were analysed using data from seven41,43,44,46–49 and five41,46–49 studies, respectively. The results suggested that 24 and 48-h morphine consumptions between groups did not differ with pooled mean differences of −1.36 [95% CI (−2.86 to 0.14); I2 = 33.1%] and −0.23 [95% CI (−2.15 to 1.68); I2 = 62.9%] (see Figure, Supplemental Digital content 4a-b, http://links.lww.com/EJA/A99, which illustrates meta-analysis of opioid consumption when LIA with regional anaesthetic techniques is compared with regional anaesthetic techniques). There was also no difference in the risk of opioid side-effects.

Mobilisation and length of hospital stay

There was no difference in ROM (n = 3 studies42,44,47) and LOS (n = 3 studies46,48,49) with pooled mean differences of 0.15 [95% CI (−1.86 to 2.16); I2 = 0%] and 0.05 [95% CI (−0.10 to 0.20); I2 = 0%], respectively.

Complications

Pooled infection rate between LIA and control group was not significantly different across 20 studies.1,2,4,6,19,21–23,25,27,28,32,39,40,43,45–49 Deep knee infection was reported in four of 735 patients receiving LIA, three of whom had postoperative intra-articular catheter placement.45,48,49

Systemic toxicity of local anaesthetic was evaluated in nine studies.1,2,4,18,24,27,40,42,43 Although in five of these studies high-dose ropivacaine (300 to 400 mg) was used for intraoperative peri-articular injection, no clinical presentation of local anaesthetic systemic toxicity was observed in any study.2,4,24,27,43 Three studies4,24,27 reported free plasma level of ropivacaine, which ranged from 0.032 to 1.346 μg ml−1 and these were all lower than the toxicity threshold concentration of 1.5 μg ml−1.50

Publication bias assessment

Funnel plots were constructed for 24-h VAS for pain at rest between LIA versus no injection or placebo when LIA was used as a primary treatment or an adjunct to regional analgesia. Both plots showed asymmetry, but Egger tests were not statistically significant with P value 0.75 and 0.28, respectively (see Figure, Supplemental Digital Content 5 and 6, http://links.lww.com/EJA/A99, which illustrate the funnel plots of 24-h VAS when LIA is compared with no injection or placebo and when LIA with regional anaesthetic techniques is compared with regional anaesthetic techniques). Contour enhanced-funnel plots suggested the asymmetry might be caused by heterogeneity rather than missing studies.

Discussion

The systematic review and meta-analysis indicate that the use of LIA, compared with no injection or placebo, provides better pain relief, improves ROM, shortens LOS and lowers risk of postoperative nausea and vomiting during the early postoperative period after TKA. In addition, there was no report of local anaesthetic toxicity in any studies and risk of infection was not increased.

We also found better pain control in the peri-articular than the intra-articular group. Peri-articular injection reduced 24-h VAS for pain, whereas intra-articular injection did not. Though both peri and intra-articular injection reduced opioid consumptions in the first 24 h compared with no injection or placebo, greater reduction was shown after peri-articular injection. In addition, the improvement in 24-h ROM was shown after peri-articular injection but not after intra-articular injection. This was consistent with Nakai et al.37 and Perret et al.51 who directly compared peri-articular injection with intra-articular injection and found better pain control in the latter group. This might be explained by the more direct diffusion of analgesic drugs into injured tissues after peri-articular injection. Although peri-articular injection provided a statistically significant difference in pain relief at 24 h, the difference of 24-h VAS for pain between peri-articular injection and placebo was only 0.89, which did not reach the minimum clinically important difference [1.8 and 95% CI (1.6 to 2.0)].52 However, the addition of postoperative catheter placement or ketorolac to local anaesthetic for peri-articular injection could increase the mean differences towards the level of clinical significance with the differences of 24-h VAS 1.50 and 1.77, respectively.

Intraoperative peri-articular injection reduced pain scores and opioid consumption until 48 h. However, the pooled effect was very much influenced by two studies which aimed to evaluate multimodal analgesia by adding adjuncts with systemic opioids and NSAIDs in the LIA but not in the control group.20,22 After excluding these studies, we found the difference of 24-h VAS for pain at rest was diluted approximately 30% and the analgesic effect of peri-articular injection at 48 h disappeared. Contrastingly, a meta-analysis by Xu et al.53 and Fang et al.54 showed a significant effect of intraoperative LIA injection up to 48 h. Two included studies that did not show a difference in VAS at 48 h between LIA and placebo as in our review were reported in Xu's study.32,35 Additionally, two trials which failed to demonstrate the benefit of LIA in our study were not included in their analysis.1,24 Fang et al.54 included 17 RCTs with peri and intra-articular injection but did not consider the effect of a postoperative catheter. Three included studies which had postoperative catheter placement reported the benefits of LIA at 48 h.25,39,48 Moreover, one of the included study which had positive results in pain control after 24 h actually performed LIA after unicompartmental knee arthroplasty.55 However, the effect sizes at 48 h of their studies were quite small and not clinically meaningful.

The efficacy of postoperative injection or infusion through an intra-articular catheter is still controversial. Ganapathy,56 Galimba57 and Dillon58 promoted this technique because of its efficacy in extended pain relief, early ambulation and shorten hospitalization; however, Kehlet,59 Gibbs60 and Andersen61 questioned these benefits and emphasised the risk of knee infection after catheter placement. The benefits of catheter placement were also inconclusive among four RCTs which studied the specific effect of postoperative intra-articular infusion after intraoperative LIA injection.28,62–64 Zhang et al.28 found benefits in reduction of pain and opioid consumption, and improvement of ROM in the first 48 h, whereas the others62–64 did not find any benefit which might be because of a lack of power. Our meta-analysis pooling the effect sizes to increase power supported the efficacy in pain relief of postoperative intra-articular catheter placement after peri-articular injection at 24 h and up to 48 h during activity after surgery. However, this benefit needs to be balanced against the risk of catheter-related infection. For this review, three of four patients suffering from deep knee infection after LIA had intra-articular catheters for postoperative administration, but bias may occur because of unsystematic detection of infection in the included studies. Most studies with postoperative catheter had predefined infection as an interested outcome and might focus more on infection than those without a postoperative catheter.

Although peri-articular multimodal drug injection provides effective analgesia after TKA from our and other reviews, there is little evidence to guide the choice between various types of injected drugs, different doses and administration sites.56,65 Most studies commonly used local anaesthetic, ketorolac, opioids, epinephrine and steroids.

Various doses of local anaesthetic were used as an ingredient for peri-articular injection. Our review supports the safety of peri-articular injection with high-dose local anaesthetic. There was no report of local anaesthetic systemic toxicity after 300 to 400 mg of ropivacaine applied in five studies,2,4,24,27,43 and the free plasma concentration in three studies was less than the toxic level.4,24,27 A recent study also supported the safety of high-dose ropivacaine injection by measuring free plasma concentration after high-dose ropivacaine injection and postoperative infusion in 28 elderly patients.3 It is still questionable whether a higher dose of local anaesthetic provides better pain relief than the dosage recommended by the manufacturer.

Ketorolac, a nonselective cyclooxygenase inhibitor NSAID, was commonly used in peri-articular injection. It inhibits peripheral nociceptors activated by peripheral tissue injury and N-Methyl-D-Aspartate-evoked nociceptor discharge. Despite higher concentration of ketorolac at peripheral sites after local infiltration than systemic administration, the peripheral analgesic action of NSAIDS with local infiltration is still inconclusive.66 Our subgroup analysis indicates the benefit of adding ketorolac to local anaesthetic for peri-articular injection, which corresponds to the recent RCTs.67,68 Spreng et al.69 also reported greater pain relief after peri-articular infiltration with ketorolac than intravenous injection.

Although sources of heterogeneity were explored by sensitivity and subgroup analysis of several possible causes, heterogeneity in some outcomes still exists. This could be explained by various drugs and doses of LIA, volume of injection, administration sites which were quite operator dependent and different analgesic adjuncts. The appropriate site and volume in detail for peri-articular injection remains inconclusive. For example, the efficacy of posterior capsule injection to control posterior knee pain remains unclear. Three RCTs70–72 reported no difference in pain relief between posterior capsule injection and sciatic nerve block, but an RCT44 from our review found that there might be no additional benefit in pain relief and accelerated functional recovery from this injection compared with placebo. Another recent study compared the efficacy among posterior capsule injection, sciatic nerve block and placebo, and reported no difference in pain relief among these interventions.73 This was consistent with a previous systematic review that could not demonstrate the benefit of sciatic nerve block over placebo.74 Therefore, posterior knee pain blockade after TKA might not be necessary, and posterior capsule injection or sciatic nerve block is still not a present standard recommendation for pain management after TKA.

The benefit of LIA as an adjunct to regional and systemic analgesia is still unclear. Kehlet and Joshi75 suggested that to accurately assess the benefits of LIA in pain control after TKA, the pooled LIA effect should be compared with contemporary multimodal analgesic regimens rather than conventional methods. Their suggestion is especially relevant for clinical outcomes that are affected by variations in perioperative care and postoperative supplemental analgesic regimens. We therefore separately analysed studies which applied regional anaesthetic techniques from those using conventional methods. The benefit of adding LIA in pain score reduction was shown only at 24 h at rest but not at activity. There was also no difference in opioid consumption, ROM and LOS. This might be explained by the high efficacy of regional anaesthetic techniques, which masked any additional effect of LIA.

During preparation of this manuscript, six meta-analyses7–9,53,54,76 on this topic were published, with varied results because of differences in the definition of LIA, the inclusion and exclusion criteria, search terms, number of pooled studies and review methodology. Keijers et al.9 concluded that the beneficial effect of LIA at 24 h was very small and not clinically relevant. They had included seven RCTs and most of them applied intra-articular injection. Jiang7 and Teng8 pooled the effect sizes of pain score reduction after peri-articular multimodal drug injection across 10 and four studies, respectively, whereas Fang et al.54 included 17 RCTs to assess the efficacy of both peri and intra-articular injection. However, none of the three meta-analyses considered the effects of postoperative catheter insertion and the use of regional anaesthetic techniques as supplemental analgesia. Some included studies in Fang's review did not meet their eligibility criteria; two studies62,63 compared intraoperative intra-articular injection versus intraoperative intra-articular injection with postoperative infusion, and LIA was not performed after TKA in another study.55 Xu et al.53 pooled the efficacy of single-dose LIA without considering the effect of different sites of injection and adjunctive treatment. Some studies which met their eligibility criteria were missed during the selection process, thus selection bias might be present.6,20,30,31,43–46 Marques et al.76 studied the efficacy of LIA injection before wound closure, and subgroup analysis was performed by the use of postoperative catheter placement. However, the effect of different supplemental analgesia regimens was not considered. Three studies which compared LIA with other techniques were included in this review.69,77,78 In addition, only 12 from 23 included studies were pooled for the effect size, and some studies analysed in our review were not included in theirs.17–19,24–26,28,29,37 Conversely, we did not include one study which was in their review because of language limitation.79

Finally, we would like to address some potential limitations of our study. We included only studies published in English, excluding three studies reported in Chinese.79–81 Only a half of the included studies clearly defined random sequence generation and allocation concealment. Most outcomes were subjectively measured, whereas 65% of included studies appropriately blinded both participants and outcome assessors. Our study focused on the efficacy of LIA compared with no injection or placebo in the early postoperative period, so the efficacy of LIA versus other techniques and long-term outcomes, such as functional recovery and knee scores, were not considered in our review.

Conclusion

LIA is effective for acute pain relief, and is associated with a reduction in postoperative nausea vomiting and hospital stay, and improvement of range of movement after TKA. Peri-articular injection reduces pain within 24 h, whereas intra-articular injection does not. When catheter placement for postoperative injection or infusion is used, the analgesic effects can be extended to 48 h, but the risk of catheter-related infection is uncertain. The efficacy of LIA as an adjunct to regional anaesthetic techniques is not demonstrated.

Acknowledgements related to this article

Assistance with the study: we would like to thank Dr Yun-Li Zhu, Dr Clifford W. Colwell Jr and Dr Constant A. Busch for providing additional data from their studies for pooling.

This study is a part of Dr. Alisa Seangleulur's training in Ph.D (Clinical Epidemiology), Faculty of Medicine Ramathibodi Hospital, Mahidol University, Bangkok, Thailand.

Financial support and sponsorship: none.

Conflicts of interest: none.

Presentation: none.

References

1. Fajardo M, Collins J, Landa J, et al. Effect of a perioperative intra-articular injection on pain control and early range of motion following bilateral TKA. Orthopedics 2011; 34:e33–e36.
2. Koh IJ, Kang YG, Chang CB, et al. Additional pain relieving effect of intraoperative peri-articular injections after simultaneous bilateral TKA: a randomized, controlled study. Knee Surg Sports Traumatol Arthrosc 2010; 18:916–922.
3. Brydone AS, Souvatzoglou R, Abbas M, et al. Ropivacaine plasma levels following high-dose local infiltration analgesia for total knee arthroplasty. Anaesthesia 2015; 70:784–790.
4. Essving P, Axelsson K, Kjellberg J, et al. Reduced morphine consumption and pain intensity with local infiltration analgesia (LIA) following total knee arthroplasty. Acta Orthop 2010; 81:354–360.
5. Browne C, Copp S, Reden L, et al. Bupivacaine bolus injection versus placebo for pain management following total knee arthroplasty. J Arthroplasty 2004; 19:377–380.
6. Rosen AS, Colwell CW Jr, Pulido PA, et al. A randomized controlled trial of intraarticular ropivacaine for pain management immediately following total knee arthroplasty. HSS J 2010; 6:155–159.
7. Jiang J, Teng Y, Fan Z, et al. The efficacy of peri-articular multimodal drug injection for postoperative pain management in total knee or hip arthroplasty. J Arthroplasty 2013; 28:1882–1887.
8. Teng Y, Jiang J, Chen S, et al. Peri-articular multimodal drug injection in total knee arthroplasty. Knee Surg Sports Traumatol Arthrosc 2013; 22:1949–1957.
9. Keijsers R, van Delft R, van den Bekerom MP, et al. Local infiltration analgesia following total knee arthroplasty: effect on postoperative pain and opioid consumption: a meta-analysis. Knee Surg Sports Traumatol Arthrosc 2013; 23:1959–1963.
10. Liberati A, Altman DG, Tetzlaff J, et al. The PRISMA statement for reporting systematic reviews and meta-analyses of studies that evaluate healthcare interventions: explanation and elaboration. J Clin Epidemiol 2009; 62:e1–e34.
11. Hozo S, Djulbegovic B, Hozo I. Estimating the mean and variance from the median, range, and the size of a sample. BMC Med Res Methodol 2005; 5:13.
12. Wilsey BL, Fishman SM. Benzon HT, Raja SN, Molley RE, et al. Minor and short-acting opioids. Essential of pain medicine and regional anesthesia. Philadelphia, PA: Elseiver; 2005. 107.
13. Hurley RW, Wu CL. Miller RD. Acute postoperative pain. Miller's anesthesia. Philadephia, PA: Churchill Livingstone; 2010. 2774.
14. Inturrisi CE, Lipman AG. Fishman SM, Ballantyne JC, Rathmell JP. Opioid analgesics. Bonica's management of pain. Baltimore, MD: Lippincott Williams & Wilkins; 2010. 1174–1175.
15. Higgins JP, Altman DG, Gotzsche PC, et al. The Cochrane Collaboration's tool for assessing risk of bias in randomised trials. BMJ 2011; 343:d5928.
16. Peters JL, Sutton AJ, Jones DR, et al. Contour-enhanced meta-analysis funnel plots help distinguish publication bias from other causes of asymmetry. J Clin Epidemiol 2008; 61:991–996.
17. Milani P, Castelli P, Sola M, et al. Multimodal analgesia in total knee arthroplasty: a randomized, double-blind, controlled trial on additional efficacy of peri-articular anesthesia. J Arthroplasty 2015; 30:2038–2042.
18. Leownorasate M, Ruangsillapanu N. Postop pain and blood loss in total knee arthroplasty: an RCT using peri-articular injection with diclofenac-based multimodal drugs. J Med Assoc Thai 2014; 97:1332–1337.
19. Niemelainen M, Kalliovalkama J, Aho AJ, et al. Single peri-articular local infiltration analgesia reduces opiate consumption until 48 h after total knee arthroplasty. A randomized placebo-controlled trial involving 56 patients. Acta Orthop 2014; 85:614–619.
20. Lamplot JD, Wagner ER, Manning DW. Multimodal pain management in total knee arthroplasty: a prospective randomized controlled trial. J Arthroplasty 2013; 28:1882–1887.
21. Chen Y, Zhang Y, Zhu YL, et al. Efficacy and safety of an intra-operative intra-articular magnesium/ropivacaine injection for pain control following total knee arthroplasty. J Int Med Res 2012; 40:2032–2040.
22. Fu PL, Xiao J, Zhu YL, et al. Efficacy of a multimodal analgesia protocol in total knee arthroplasty: a randomized, controlled trial. J Int Med Res 2010; 38:1404–1412.
23. Fu P, Wu Y, Wu H, et al. Efficacy of intra-articular cocktail analgesic injection in total knee arthroplasty: a randomized controlled trial. Knee 2009; 16:280–284.
24. Busch CA, Shore BJ, Bhandari R, et al. Efficacy of peri-articular multimodal drug injection in total knee arthroplasty. A randomized trial. J Bone Joint Surg Am 2006; 88:959–963.
25. Kazak Bengisun Z, Aysu Salviz E, Darcin K, et al. Intraarticular levobupivacaine or bupivacaine administration decreases pain scores and provides a better recovery after total knee arthroplasty. J Anesth 2010; 24:694–699.
26. Andersen LØ, Husted H, Otte KS, et al. High-volume infiltration analgesia in total knee arthroplasty: a randomized, double-blind, placebo-controlled trial. Acta Anaesthesiol Scand 2008; 52:1331–1335.
27. Vendittoli PA, Makinen P, Drolet P, et al. A multimodal analgesia protocol for total knee arthroplasty. A randomized, controlled study. J Bone Joint Surg Am 2006; 88:282–289.
28. Zhang S, Wang F, Lu ZD, et al. Effect of single-injection versus continuous local infiltration analgesia after total knee arthroplasty: a randomized, double-blind, placebo-controlled study. J Int Med Res 2011; 39:1369–1380.
29. Shen SJ, Peng PY, Chen HP, et al. Analgesic effects of intra-articular bupivacaine/intravenous parecoxib combination therapy versus intravenous parecoxib monotherapy in patients receiving total knee arthroplasty: a randomized, double-blind trial. Biomed Res Int 2015; 2015:450805.
30. Guará Sobrinho H, Garcia JBS, Vasconcelos JW, et al. Analgesic efficacy of the intra-articular administration of S(+)- ketamine in patients undergoing total knee arthroplasty. Rev Bras Anestesiol 2012; 62:665–675.
31. Garcia JBS, Barbosa Neto JO, Vasconcelos JW, et al. Analgesic efficacy of the intra-articular administration of high doses of morphine in patients undergoing total knee arthroplasty. Rev Bras Anesthesiol 2010; 60:1–12.
32. Tanaka N, Sakahashi H, Sato E, et al. The efficacy of intra-articular analgesia after total knee arthroplasty in patients with rheumatoid arthritis and in patients with osteoarthritis. J Arthroplasty 2001; 16:306–311.
33. Ritter MA, Koehler M, Keating EM, et al. Intra-articular morphine and/or bupivacaine after total knee replacement. J Bone Joint Surg Br 1999; 81:301–303.
34. Klasen JA, Opitz SA, Melzer C, et al. Intraarticular, epidural, and intravenous analgesia after total knee arthroplasty. Acta Anaesthesiol Scand 1999; 43:1021–1026.
35. Mauerhan DR, Campbell M, Miller JS, et al. Intra-articular morphine and/or bupivacaine in the management of pain after total knee arthroplasty. J Arthroplasty 1997; 12:546–552.
36. Badner NH, Bourne RB, Rorabeck CH, et al. Intra-articular injection of bupivacaine in knee-replacement operations. Results of use for analgesia and for preemptive blockade. J Bone Joint Surg Am 1996; 78:734–738.
37. Nakai T, Tamaki M, Nakamura T, et al. Controlling pain after total knee arthroplasty using a multimodal protocol with local peri-articular injections. J Orthop 2013; 10:92–94.
38. Ikeuchi M, Kamimoto Y, Izumi M, et al. Local infusion analgesia using intra-articular double lumen catheter after total knee arthroplasty: a double blinded randomized control study. Knee Surg Sports Traumatol Arthrosc 2013; 21:2680–2684.
39. Gomez-Cardero P, Rodriguez-Merchan EC. Postoperative analgesia in TKA: ropivacaine continuous intraarticular infusion. Clin Orthop Relat Res 2010; 468:1242–1247.
40. Ong JC, Chin PL, Fook-Chong SM, et al. Continuous infiltration of local anaesthetic following total knee arthroplasty. J Orthop Surg (Hong Kong) 2010; 18:203–207.
41. Chinachoti T, Lungnateetape A, Raksakietisak M. Peri-articular infiltration of 0.25% bupivacaine on top of femoral nerve block and intrathecal morphine improves quality of pain control after total knee arthroplasty: a randomized double-blind placebo controlled clinical trial. J Med Assoc Thai 2012; 95:1536–1542.
42. Joo JH, Park JW, Kim JS, et al. Is intra-articular multimodal drug injection effective in pain management after total knee arthroplasty? A randomized, double-blinded, prospective study. J Arthroplasty 2011; 26:1095–1099.
43. Koh IJ, Kang YG, Chang CB, et al. Does peri-articular injection have additional pain relieving effects during contemporary multimodal pain control protocols for TKA?: a randomised, controlled study. Knee 2012; 19:253–259.
44. Krenzel BA, Cook C, Martin GN, et al. Posterior capsular injections of ropivacaine during total knee arthroplasty: a randomized, double-blind, placebo-controlled study. J Arthroplasty 2009; 24:138–143.
45. Mullaji A, Kanna R, Shetty GM, et al. Efficacy of peri-articular injection of bupivacaine, fentanyl, and methylprednisolone in total knee arthroplasty: a prospective, randomized trial. J Arthroplasty 2010; 25:851–857.
46. Yuenyongviwat V, Pornrattanamaneewong C, Chinachoti T, et al. Peri-articular injection with bupivacaine for postoperative pain control in total knee replacement: a prospective randomized double-blind controlled trial. Adv Orthop 2012; 2012:107309.
47. Han CD, Lee DH, Yang IH. Intra-synovial ropivacaine and morphine for pain relief after total knee arthroplasty: a prospective, randomized, double blind study. Yonsei Med J 2007; 48:295–300.
48. Goyal N, McKenzie J, Sharkey PF, et al. The 2012 Chitranjan Ranawat award: intraarticular analgesia after TKA reduces pain: a randomized, double-blinded, placebo-controlled, prospective study. Clin Orthop Relat Res 2013; 471:64–75.
49. Reeves M, Skinner MW. Continuous intra-articular infusion of ropivacaine after unilateral total knee arthroplasty. Anaesth Intensive Care 2009; 37:918–922.
50. Scott DB, Lee A, Fagan D, et al. Acute toxicity of ropivacaine compared with that of bupivacaine. Anesth Analg 1989; 69:563–569.
51. Perret M, Fletcher P, Firth L, et al. Comparison of patient outcomes in peri-articular and intraarticular local anaesthetic infiltration techniques in total knee arthroplasty. J Orthop Surg Res 2015; 10:119.
52. Todd KH, Funk KG, Funk JP, et al. Clinical significance of reported changes in pain severity. Ann Emerg Med 1996; 27:485–489.
53. Xu CP, Li X, Wang ZZ, et al. Efficacy and safety of single-dose local infiltration of analgesia in total knee arthroplasty: a meta-analysis of randomized controlled trials. Knee 2014; 21:636–646.
54. Fang R, Liu Z, Alijiang A, et al. Efficacy of intra-articular local anesthetics in total knee arthroplasty. Orthopedics 2015; 38:e573–e581.
55. Weston-Simons JS, Pandit H, Haliker V, et al. Intra-articular local anaesthetic on the day after surgery improves pain and patient satisfaction after Unicompartmental Knee Replacement: a randomised controlled trial. Knee 2012; 19:352–355.
56. Ganapathy S. Wound/intra-articular infiltration or peripheral nerve blocks for orthopedic joint surgery: efficacy and safety issues. Curr Opin Anaesthesiol 2012; 25:615–620.
57. Galimba J. Promoting the use of peri-articular multimodal drug injection for total knee arthroplasty. Orthop Nurs 2009; 28:250–254.
58. Dillon JP, Brennan L, Mitchell D. Local infiltration analgesia in hip and knee arthroplasty: an emerging technique. Acta Orthop Belg 2012; 78:158–163.
59. Kehlet H, Andersen LØ. Local infiltration analgesia in joint replacement: the evidence and recommendations for clinical practice. Acta Anaesthesiol Scand 2011; 55:778–784.
60. Gibbs DM, Green TP, Esler CN. The local infiltration of analgesia following total knee replacement: a review of current literature. J Bone Joint Surg Br 2012; 94:1154–1159.
61. Andersen LØ, Kehlet H. Analgesic efficacy of local infiltration analgesia in hip and knee arthroplasty: a systematic review. Br J Anaesth 2014; 11:360–374.
62. Nechleba J, Rogers V, Cortina G, et al. Continuous intra-articular infusion of bupivacaine for postoperative pain following total knee arthroplasty. J Knee Surg 2005; 18:197–202.
63. Williams D, Petruccelli D, Paul J, et al. Continuous infusion of bupivacaine following total knee arthroplasty: a randomized control trial pilot study. J Arthroplasty 2013; 28:479–484.
64. Ali A, Sundberg M, Hansson U, et al. Doubtful effect of continuous intraarticular analgesia after total knee arthroplasty: a randomized double-blind study of 200 patients. Acta Orthop 2015; 86:373–377.
65. Dahl JB, Møiniche S. Relief of postoperative pain by local anaesthetic infiltration: efficacy for major abdominal and orthopedic surgery. Pain 2009; 143:7–11.
66. Romsing J, Moiniche S, Ostergaard D, et al. Local infiltration with NSAIDs for postoperative analgesia: evidence for a peripheral analgesic action. Acta Anaesthesiol Scand 2000; 44:672–683.
67. Andersen KV, Nikolajsen L, Haraldsted V, et al. Local infiltration analgesia for total knee arthroplasty: should ketorolac be added? Br J Anaesth 2013; 111:242–248.
68. Kim TW, Park SJ, Lim SH, et al. Which analgesic mixture is appropriate for peri-articular injection after total knee arthroplasty? Prospective, randomized, double-blind study. Knee Surgery Sports Traumatol Arthrosc 2015; 23:838–845.
69. Spreng UJ, Dahl V, Hjall A, et al. High-volume local infiltration analgesia combined with intravenous or local ketorolac+morphine compared with epidural analgesia after total knee arthroplasty. Br J Anaesth 2010; 105:675–682.
70. Tanikawa H, Sato T, Nagafuchi M, et al. Comparison of local infiltration of analgesia and sciatic nerve block in addition to femoral nerve block for total knee arthroplasty. J Arthroplasty 2014; 29:2462–2467.
71. Gi E, Yamauchi M, Yamakage M, et al. Effects of local infiltration analgesia for posterior knee pain after total knee arthroplasty: comparison with sciatic nerve block. J Anesth 2014; 28:696–701.
72. Mahadevan D, Walter RP, Minto G, et al. Combined femoral and sciatic nerve block vs combined femoral and peri-articular infiltration in total knee arthroplasty: a randomized controlled trial. J Arthroplasty 2012; 27:1806–1811.
73. Safa B, Gollish J, Haslam L, et al. Comparing the effects of single shot sciatic nerve block versus posterior capsule local anesthetic infiltration on analgesia and functional outcome after total knee arthroplasty: a prospective, randomized, double-blinded, controlled trial. J Arthroplasty 2014; 29:1149–1153.
74. Abdallah FW, Brull R. Is sciatic nerve block advantageous when combined with femoral nerve block for postoperative analgesia following total knee arthroplasty? A systematic review. Reg Anesth Pain Med 2011; 36:493–498.
75. Kehlet H, Joshi GP. Systematic reviews and meta-analyses of randomized controlled trials on perioperative outcomes: an urgent need for critical reappraisal. Anesth Analg 2015; 121:1104–1107.
76. Marques EM, Jones HE, Elvers KT, et al. Local anaesthetic infiltration for peri-operative pain control in total hip and knee replacement: systematic review and meta-analyses of short- and long-term effectiveness. BMC Musculoskelet Disord 2014; 15:220.
77. Andersen KV, Bak M, Christensen BV, et al. A randomized, controlled trial comparing local infiltration analgesia with epidural infusion for total knee arthroplasty. Acta Orthop 2010; 81:606–610.
78. Essving P, Axelsson K, Aberg E, et al. Local infiltration analgesia versus intrathecal morphine for postoperative pain management after total knee arthroplasty: a randomized controlled trial. Anesth Analg 2011; 113:926–933.
79. Zhang J, Jiang Y, Shao JJ, et al. Effect of peri-articular multimodal drug injection on pain after total knee arthroplasty. J Clin Rehabilitative Tissue Eng Res 2007; 11:8678–8682.
80. Man ZT, Li W, Zhang W, et al. Local injection of bupivacaine following total knee arthroplasty: analgesia effect and safety evaluation. J Clin Rehabilitative Tissue Eng Res 2010; 14:4774–4777.
81. Qian WW, Weng XS, Fei Q, et al. [Application study of peri-articular multimodal drug injection in total knee arthroplasty]. Zhonghua Yi Xue Za Zhi 2010; 90:2593–2596.

Supplemental Digital Content

© 2016 European Society of Anaesthesiology