No Clinically Important Difference in Pain Scores After THA Between Periarticular Analgesic Injection and Placebo: A Randomized Trial

Abstract

Background

Periarticular analgesic injection (PAI) is being used more commonly for pain relief after orthopaedic surgeries. However, there is conflicting evidence regarding the effectiveness of PAI for post-THA pain relief.

Questions/purposes

In a double-blind, randomized, controlled trial among patients undergoing same-day bilateral THA, with each patient serving as his or her own control, we asked: (1) Did the pain score as measured on a 100-mm VAS differ between the hips that received PAI versus placebo? (2) Were there differences in complications between the treatment and control hips in these patients?

Methods

Over a 1-year period at one center, 45 patients underwent same-day bilateral THA; three were excluded for prespecified reasons, and two declined participation in this randomized, controlled trial, leaving 40 patients (80 THAs) in the study. Patients randomly received PAI in one hip and placebo in the contralateral hip; patients, surgeons, and nurses were blinded in terms of which hip received the PAI and which hip received a placebo saline injection. The PAI solution included ropivacaine, morphine hydrochloride hydrate, methylprednisolone, ketoprofen, and epinephrine. The primary outcome was the VAS for pain at rest 24 hours after THA, measured using a 100-mm horizontal VAS. The VAS score was compared between two groups and assessed to reach the reported threshold values for the minimum clinically important difference (MCID) of 20 mm for the postoperative VAS score. No patients were lost to followup, and there were no missing data for the primary outcome. Complications that occurred during the trial were recorded prospectively with emphasis on infection, wound complications, nerve palsy and allergic reactions to the injections.

Results

There were no clinically important differences between hips treated with the PAI and those treated with the placebo injection at any point. The hips that received PAI had less pain than those receiving placebo 24 hours after THA (16 ± 17 mm versus 22 ± 20 mm; mean difference, 6 mm; 95% confidence interval [CI], 2–9 mm; p = 0.006), but this effect size was below the MCID of 20 mm and thus is unlikely to be clinically important. The hips that received PAI also had better VAS scores in the recovery room (38 ± 29 mm versus 52 ± 33 mm; mean difference 14 mm; 95% CI, 5–23 mm; p = 0.004) and 3 hours after THA than placebo controls (28 ± 22 mm versus 37 ± 24 mm; mean difference 9 mm; 95% CI, 2–16 mm; p = 0.010). Neither of these differences exceeded the MCID and likewise were unlikely to be clinically important. No complications, including surgical site infections, were observed in either group.

Conclusions

Periarticular analgesic injection for pain control after THA did not result in a clinically important reduction in pain at any point examined. Given the expense associated with this PAI mixture and the lack of effectiveness outside this timeframe, we cannot recommend its use. Other mixtures or concentrations of drugs may be helpful in short-stay admissions for THA, but this will require further research.

Level of Evidence

Level I, therapeutic study.

Introduction

Although THA is very successful in terms of patient satisfaction, low complication rates, and cost-effectiveness [10], addressing postoperative pain remains an important challenge [18].

Periarticular analgesic injection (PAI) has attracted attention as an effective analgesic modality with a low prevalence of adverse effects [12, 15]. However, there is conflicting evidence about the benefits of PAI for patients undergoing THA [1-5, 7, 11, 13, 15-17, 22, 25]. Some randomized controlled trials have shown that postoperative pain was improved with this modality [1, 11, 13, 15, 17], whereas others have shown no benefit [2-4, 25]. One potential reason for these conflicting results was that these studies compared different patient cohorts with and without PAI [1-4, 11, 13, 15, 17, 25]. Such inclusion criteria may limit the ability to compare pain relief regimens effectively, because pain thresholds vary between patients [6].

We therefore conducted a double-blind, randomized, controlled trial comparing PAI, including local anesthetics, opioids, corticosteroids, NSAIDs, and epinephrine with saline placebo injections in patients undergoing simultaneous bilateral THA. We asked: (1) Did the pain score as measured on a 100-mm VAS differ between the hips that received PAI versus placebo? (2) Were there differences in complications between the treatment and control hips in these patients?

Patients and Methods

This was a single center, randomized, controlled trial in which patients undergoing simultaneous bilateral THA were randomly assigned to receive a periarticular injection in one hip and a placebo injection in the contralateral hip.

The study was approved by the institutional review board, and patients were provided written informed consent. Before participant enrollment, the trial was registered with the University Hospital Medical Information Network (registration number UMIN000016479).

Patients scheduled for simultaneous bilateral THA between February 2015 and January 2016 were eligible for inclusion in this trial. We excluded patients scheduled for THA combined with subtrochanteric shortening osteotomy and patients scheduled for THA combined with implant removal. Other exclusion criteria were allergy or intolerance to one of the study drugs, regular opioid use, renal insufficiency, asthma, and prolonged QT interval on electrocardiography to avoid cardiotoxicity due to local anesthetic-induced inhibition of sodium channels. A total of 45 patients underwent simultaneous bilateral THA during the study period and were eligible for inclusion (Fig. 1). Three patients who underwent simultaneous bilateral THA combined with subtrochanteric femoral shortening osteotomy and two patients who declined to participate in the trial were excluded. The remaining 40 patients (Table 1) were randomly assigned to receive PAI in one hip and placebo in the other hip. No patients were lost to followup after randomization, and there were no missing data regarding the primary outcome. Twenty-three patients received PAI in the right hip, and the remaining 17 patients received PAI in the left hip.

Fig. 1.

Patient recruitment and analysis are shown.

Table 1.

Patient demographics

Participants were informed that we were testing the efficacy of periarticular injection for pain control after THA and that they would be randomly assigned to receive a periarticular injection in one hip and a placebo solution in the other hip.

Randomized numbers ranging from 0 to 99 generated using computer software (Excel 2010; Microsoft, Redmond, WA, USA) were placed in an opaque envelope. Just prior to surgery, unblinded allocating staff (YY, YH, and MY) who did not participate in outcome assessment selected a sealed envelope in the operating suite. Patients with even numbers received PAI in the right hip, and those with odd numbers received the PAI in the left hip.

Injections were performed just after completion of total hip prosthesis implantation and prior to closure. The PAI solution contained 40 mL of 7.5 mg/mL ropivacaine, 0.8 mL of 10 mg/mL morphine hydrochloride hydrate, 40 mg of methylprednisolone, 50 mg of ketoprofen, and 0.3 mL of 1.0 mg/mL epinephrine. These agents were mixed with normal saline to a combined volume of 60 mL. The surgeon injected 20 mL of the mixture into the tensor fascia lata, 20 mL into the gluteus medius, and the final 20 mL into the capsule.

For the placebo solution, we used 60 mL of normal saline with an identical injection protocol.

Patients did not receive any other preemptive multimodal medication.

All patients received 4 mg of lornoxicam orally three times a day after each meal. No other oral medication, intravenous patient-controlled analgesia, or single or continuous peripheral nerve blockade was used for pain control after surgery. For rescue analgesia, 50-mg diclofenac sodium suppositories were used. The mean and SD of the number of rectal suppositories used was 0.3 ± 0.7 during the 24 hours after THA.

All surgeries were performed under general anesthesia by a single surgeon (NH), who used the modified Watson-Jones approach in the lateral decubitus position [20].

A cementless femoral stem and acetabular cup was used in all patients. The prosthesis for each femur was selected based on the femoral morphology. We used the Accolade II (Stryker, Mahwah, NJ, USA) in 30 hips, the TaperLoc® Complete Microplasty (Zimmer Biomet, Warsaw, IN, USA) in 25 hips, the S-ROM® A (DePuy Synthes, Raynham, MA, USA) in 15 hips, the Mainstay (Kyocera, Osaka, Japan) in four hips, the Corail® (DePuy Synthes) in two hips, the Wagner Cone® (Zimmer Biomet) in two hips, and the SL-Plus^™ MIA HA (Smith & Nephew, London, UK) in two hips. Similar prostheses were implanted in the bilateral hips of 34 patients, while different prostheses were used in the remaining six patients.

All patients began gait exercises and range of motion training the morning after THA. Therefore, rehabilitation began before measurement of the primary study outcome. Our institution applied the no-restriction protocol of rehabilitation in all patients undergoing primary THA [19]. The protocol allowed patients to move the operated hip through a range of motion without restriction and to load their legs without protected weightbearing.

Outcome Measurements

The primary outcome was pain at rest 24 hours after surgery. The pain scores of both hips were measured 24 hours after surgery. Pain intensity was rated using a 100-mm horizontal VAS, for which 0 mm represented no pain and 100 mm represented extreme pain. The VAS score at rest was also measured in the recovery room, 3 and 6 hours postoperatively, and 3 and 7 days postoperatively. These measurements were performed during hospitalization because the current Japanese universal health insurance system allows more than 7 days of hospitalization for patients undergoing THA.

We prospectively assessed complications with special reference to wound issues, nerve injuries, and allergic reactions to the injected medications.

Sample Size

Prior to patient enrollment, we could not reliably estimate a required sample size because the minimum clinically important difference (MCID) was not clear for the VAS pain score after a THA in a Japanese population. The MCID is defined as the smallest difference perceived as important by the average patient [14]. Provisionally, we used a 20-mm mean decrease in the VAS score as the MCID according to the previous reports investigating the efficacy of PAI during knee surgery in a Japanese population [8, 9, 23, 24]. We calculated that 38 patients would be required for this trial to illustrate a clinically important 20-mm mean decrease in the VAS score, with a two-sided 5% significance level and 80% power. For power analysis, we used an SD of 31 in the VAS score based on our preliminary series of the patients undergoing simultaneous bilateral THA.

Statistical Analyses

To analyze the primary outcome, we compared the mean differences and 95% confidence intervals (CIs) with results of the paired t-test. In addition to statistical significance, mean difference of the primary outcome of pain at rest at 24 hours was assessed to determine whether it met the MCID of 20 mm as described above.

We used the paired t-test and the chi-square test to analyze other continuous variables and categorical variables, respectively. A two-sided p < 0.05 was considered statistically significant. One investigator (ST) performed all statistical analyses with R (The R Foundation for Statistical Computing, Vienna, Austria).

Results

All data for the primary outcome of this study were available. There were no clinically important differences between hips treated with PAI and those treated with placebo injection at any point (Table 2). The hips that received PAI had a lower VAS pain score than the hips that received placebo 24 hours after surgery (16 ± 17 mm versus 22 ± 20 mm; mean difference, 6 mm; 95% CI, 2–9 mm; p = 0.006), but this difference did not reach the level of the MCID, and so was unlikely to be clinically important. The hips that received PAI also had a lower VAS score in the recovery room (38 ± 29 mm versus 52 ± 33 mm; mean difference 14 mm; 95% CI, 5–23 mm; p = 0.004) and 3 hours after surgery (28 ± 22 mm versus 37 ± 24 mm; mean difference 9 mm; 95% CI, 2–16 mm; p = 0.010), but again, these differences did not reach the level of the MCID for pain (Fig. 2).

Table 2.

VAS scores for postoperative pain at rest

Fig. 2.

VAS scores for pain at rest after simultaneous bilateral THA in patients receiving periarticular analgesic injection in one hip and placebo in the other hip are shown. Although the pain score was lower in the hips that received periarticular analgesic injection in the recovery room (38 mm ± 29 mm versus 52 mm ± 33 mm; p = 0.004), 3 hours after surgery (28 mm ± 22 mm versus 37 mm ± 24 mm; p = 0.010) and 24 hours after surgery (16 ± 17 mm versus 22 ± 20 mm; p = 0.006), these differences did not meet the MCID for pain VAS.

There were no complications, including surgical site infections, wound complications, prolonged numbness or weakness, or allergy to the medications in our cohort.

Discussion

Conflicting evidence exists about the effectiveness of PAI for post-THA pain management. We sought to clarify this issue by conducting a prospective, randomized, double-blind trial involving a PAI solution versus placebo in simultaneous bilateral THA. The small differences favoring the PAI group in terms of pain scores did not exceed the MCID for the VAS pain instrument, and so were unlikely to have been clinically important.

There were many limitations in this trial. First, the random allocation was performed in the same institution where the surgical trial was performed. Remote methods of randomization may be more effective than our manner of random allocation [21]. Moreover, we did not survey patients, surgeons, or outcomes assessors regarding whether they could ascertain which hip had been treated and which received the placebo injection. Such a survey could help to confirm the efficacy of blinding. However, in a trial of this size and this nature, we believe our randomization and blinding procedures were adequate. This study was conducted in a single center, and only one experienced surgeon performed surgery, which might limit the generalizability of the results; however, this also provides some uniformity to surgical technique and procedure length. The pain scores of both hips were measured 24 hours after completion of surgery; however, the time of completion was different between the right and left hips. Thus, the measurement time did not correspond exactly to the 24-hour time point. We believe that our randomization to determine which side should receive PAI minimized the importance of this issue. Although the pain score at rest was measured as the primary outcome of this study, the pain score during activity was not evaluated, though we would expect a relationship between the two. Moreover, rehabilitation was started before measurement of the primary outcome. These issues may have distorted these study results. However, the pain at rest would be one determinant for patients’ satisfaction after THA. Although our randomized controlled trial enrolled patients undergoing bilateral simultaneous THA, the difference in procedure complexity between both hips could affect postoperative pain. We believe that random determination of which hip should be treated first would also help to minimize these issues. Allowing patients to use rescue analgesia could mask the effectiveness of the periarticular injection, although we believe that the masking effect would be small because the mean number of rescue analgesia used was relatively low. Finally, the drugs used in the PAI solution generally do not have an effective duration of 24 hours. However, it should be noted that none of the observed differences were clinically important even at much shorter intervals.

Although ours was a no-difference trial in that the intervention studied (PAI after THA) did not produce a clinically important reduction in pain, we note that the study was adequately powered to detect statistical differences in our pain-related endpoints. It is important to realize that statistical differences are not what patients seek; clinical decisions should be made based on reasonable effect sizes, and the effect size favoring PAI here was very small (below the MCID). However, this is not the same as an underpowered trial; our trial was adequately powered, since it detected statistical differences.

Our study did not find compelling evidence that PAI provided pain relief; none of the differences we observed exceeded the MCID for the VAS pain tool, suggesting that any differences observed were too small to be meaningful to the typical patient. The largest effect was in the recovery room, and so any benefits of PAI were not just small, but also quite transient.

The controversy surrounding this topic is evident in multiple studies. Some RCTs compared PAI with no injection or placebo injection with varying conclusions regarding effectiveness (Table 3). Other RCTs compared PAI with other pain relief regimens after THA, again with varying results (Table 4). To our knowledge, this is first RCT including effect size to assess pain relief in this setting. Our unique study design may have been one reason for the significant differences in pain scores without reaching MCID: PAI was injected into one hip and placebo into the other hip in the same patients undergoing same-day bilateral THA. Some drugs, such as opioids and corticosteroids, may have systemic effects. Although the effectiveness of PAI for knee arthroplasty is widely accepted, one RCT using an intervention like that used here, in which PAI was injected into one knee and placebo into the other in patients undergoing same-day bilateral TKA, concluded that there were no differences in pain relief between the two groups [6].

Table 3.

Trial characteristics of randomized controlled studies comparing periarticular injection with no injection or placebo injection for patients undergoing THA

Table 4.

Trial characteristics of randomized controlled studies comparing periarticular injection with other regimens for pain relief after THA

There were no complications attributable to the injections during the study period. Although no previous studies have shown that PAI substantially influences complications other than postoperative nausea [15], more serious complications, including surgical site infections, wound problems, allergies, and nerve palsy may occur. Our study was underpowered to analyze the impact of PAI on complications.

In conclusion, though no obvious complications were observed, there was little clinical benefit to PAI, and the additional cost of PAI precludes a recommendation to use the regimen we described here. Other PAI regimens or dosages may be effective in short-stay settings, but this will require further study.