Role of Parecoxib Sodium in the Multimodal Analgesia after Total Knee Arthroplasty: A Randomized Double‐blinded Controlled Trial

Abstract

Objective

Total knee arthroplasty (TKA) is an established surgical technique and is the standard treatment for degenerative knee joint diseases. However, severe pain after TKA makes it difficult for many patients to perform early postoperative rehabilitation and functional exercise, which might result in subsequent unsatisfactory recovery of knee joint function and great reduction in patients’ satisfaction and quality of life. Orthopaedic surgeons have tried a large variety of analgesics and analgesic modes to relieve patients’ pain after TKA. There are many analgesic regimens available in clinical practice but all have some deficiencies. Parecoxib sodium, a highly selective inhibitor of cyclooxygenase‐2 (COX‐2), can reduce the synthesis of peripheral prostaglandin to exert the effect of analgesia, and relieve inflammation and prevent central sensitization through inhibition of peripheral and central COX‐2 expression. In addition, it can be used as a preemptive analgesic without affecting platelet aggregation. However, there does seem to be conflicting evidence in the current research as to whether parecoxib sodium can be used successfully as a preemptive analgesic; the effect of preemptive analgesia with parecoxib sodium in multimodal analgesia is still controversial. This research investigated the effects of parecoxib sodium in a preemptive multimodal analgesic regimen.

Methods

Eighty‐eight patients were randomized into two groups. The experimental group received parecoxib (46 patients) and the control group received saline (42 patients), administered 30 min before the initiation of the surgical procedure. A patient‐controlled analgesia (PCA) pump was applied within 48 h after surgery. The visual analogue scale (VAS), drug consumption through the PCA pump, use of salvaging analgesia, range of motion (ROM) of the knee joints, and postoperative complications were observed.

Results

The VAS score in the post‐anesthesia care unit (PACU) of the parecoxib group was significantly lower than that of the control group (P = 0.039). There was no significant difference in the demographic profiles, duration of operation, hemorrhage in surgery, postoperative hemorrhage, postoperative drainage, VAS at different time points, function of knee joints, length of hospital stay, use of salvaging analgesia, and postoperative drug consumption through the PCA between the two groups (P > 0.05).

Conclusion

In preemptive multimodal analgesia regimens, parecoxib sodium can significantly decrease the VAS score in the short term, relieve pain shortly after surgery, and does not increase the incidence of complications. Parecoxib sodium is a safe and effective drug in the perioperative analgesic management for TKA.

Introduction

Total knee arthroplasty (TKA) is an established and successful surgical procedure which is the major treatment for degenerative knee joint diseases. With the popularity and promotion of TKA, increasing numbers of patients with degenerative knee diseases are undergoing this surgery to restore knee function and mobility, as well as to improve quality of life. However, severe pain after TKA makes it difficult for many patients to participate in early postoperative rehabilitation and to exercise, which might result in subsequent unsatisfactory recovery of knee joint function and great reduction in patients’ quality of life. Perioperative pain control has direct influence on postoperative recovery and surgical outcome. An appropriate perioperative analgesic protocol could relieve postoperative pain and allow functional exercising, leading to early rehabilitation.

Therefore, perioperative analgesia is indispensable for satisfactory prognosis. Orthopaedic surgeons have tried a large variety of analgesics and analgesic modes to relieve patients’ pain after TKA. There are many analgesic regimens in clinical practice but all have some deficiencies. In addition to the risks of hematoma, infection, neural damage, and postoperative falls, peripheral nerve block and epidural anesthesia are technically demanding and complicated to manipulate, which also limits the application of postoperative anti‐coagulation. The effects of oral analgesics are uncontrollable and unquantifiable because they are affected by many factors, such as gastrointestinal absorption efficiency, serum concentration distribution, and preoperative fasting. Oral opioid therapy is recommended as the first‐choice medication for postoperative pain control but is associated with several adverse effects, such as respiratory depression, lethargy, renal impairment, nausea, and vomiting1.

Multimodal analgesia using a combination of analgesic methods throughout the perioperative period to control postoperative pain is recommended over the use of opioids alone1. Non‐steroidal anti‐inflammatory drugs (NSAID) are among the options being trialed to reduce postoperative pain and avoid the adverse effects of opioids2. However, the application of non‐selective cyclooxygenase (COX) inhibitors in the early perioperative period can caused adverse events such as upper gastrointestinal ulceration and bleeding because they inhibit the normal function of gastric mucosal COX‐1. They can also prevent the production of thromboxane A2, which reduces platelet aggregation and thereby prolongs bleeding. Therefore, orthopedic surgeons often have concerns about using them during the perioperative period of TKA. Parecoxib sodium, a highly selective inhibitor of COX‐2, can reduce the synthesis of peripheral prostaglandin to exert the effect of analgesia, and relieve inflammation and prevent central sensitization through inhibition of peripheral and central COX‐2 expression. In addition, it can be used as a preemptive analgesic without affecting platelet aggregation3.

Preemptive analgesia can alleviate stress responses in the perioperative period by protecting peripheral and central nervous systems from sensitization, which is of great importance in the improvement of postoperative prognosis. Preemptive analgesia has been defined as treatment that: (i) starts before surgery; (ii) prevents the establishment of central sensitization caused by incisional injury (covers only the period of surgery); and (iii) prevents the establishment of incisional and inflammatory injuries (covers the period of surgery and the initial postoperative period). However, there does seem to be conflicting evidence in the current research as to whether parecoxib sodium can be used successfully as a preemptive analgesic and the effects of preemptive analgesia with parecoxib sodium in multimodal analgesia regimen is still controversial 4, 5, 6, 7.

As a result, a single‐centered randomized controlled double‐blinded study was designed to observe the visual analogue scale (VAS) score, drug consumption through a patient‐controlled analgesia (PCA) pump, use of salvaging analgesia, range of motion (ROM) of knee joints, and postoperative complications, with the purpose of evaluating the role of parecoxib sodium in multimodal analgesia regimen after TKA, and the safety of this protocol.

Patients and Methods

Inclusion and Exclusion Criteria

This study was approved by the Ethics Committee of Peking Union Medical College Hospital, and written consent was acquired from each subject. A trial registration number (ClinicalTrials.gov identifier: ChiCTR1800014872) was obtained from the Chinese Clinical Trial Registry. Patients were recruited who had received primary unilateral TKA between January 2013 and March 2015 in the Department of Orthopaedics, Peking Union Medical College Hospital.

Patients who met the following criteria were included: (i) hospitalized patients over the age of 18 with intact cognitive function (no gender restriction); (ii) patients diagnosed with osteoarthritis (OA) or rheumatoid arthritis (RA) before surgery, with unsatisfactory response to conservative treatment; (iii) patients with no obvious deformities and with unilateral TKA scheduled; (iv) those with American Society of Anesthesiologists (ASA) score ≤2; (v) patients with a post‐stabilizing prosthesis implanted; and (vi) those in whom a unified configuration of PCA pump was applied postoperatively.

Patients were excluded if they met one of the following conditions: (i) allergy to parecoxib sodium, or any composite of the parecoxib sodium product; (ii) history of severe allergic reaction, especially dermatological manifestations, including Stevens–Johnson syndrome, toxic epidermal necrolysis, erythema multiforme, or known history of allergy to sulfonamides; (iii) active gastrointestinal bleeding or ulceration; (iv) allergy to NSAIDS (including COX‐2 inhibitors); (v) pregnant or breast‐feeding; (vi) severe liver dysfunction (serum albumin <25 g/L, or Child–Pugh score ≥ 10); (vii) inflammatory bowel disease; (viii) congestive heart failure (New York Heart Association [NYHA] II to IV); (ix) ischemic cardiac diseases; (x) disease of peripheral arteries or cerebral vessels; and (xi) severe varus or valgus deformity and flexion deformity.

This study was a single‐centered randomized controlled trial. The patients, surgeons, anesthesiologists, and postoperative evaluation team were all blinded from the classification of the patients until all data were collected. A randomized sequence was generated with RandA1.0 (Pulande, Beijing, China).

Recruitment was stopped based on an estimate of power based on blinded data, which showed that adequate power was reached despite the lower than targeted sample size. According to this power analysis, 42 evaluable patients per treatment group were needed to achieve 80% power to detect a 4.1‐point difference in VAS pain scores (common standard deviation [SD], 6.7) and 3‐mg difference in total PCA drug consumption (common SD, 4.67) between treatment groups at a one‐sided 0.05 alpha level with sequential testing. The sample size of 42 patients per treatment group provided 89% power to detect a 4.1‐point difference in VAS pain scores, 83% power to detect a 3‐mg difference in total PCA drug consumption, and 80% power to detect both.

A total of 88 patients were recruited in this study and randomly divided into a parecoxib group (46 patients) and a control group (42 patients) (Fig. 1).

Figure 1.

Patient flowcharts.

Surgical Procedures

The operations were completed by one senior surgeon at our hospital. All patients underwent the operation using the patellar medial approach under general anesthesia with the airbag tourniquet setting on 100 mm Hg above systolic pressure. The coronal and sagittal alignments were determined by intra‐medullary localization with valgus 6‐degree osteotomy in the femur and extra‐medullary localization with 3‐degree retroversion in the tibia. The external rotation of the femur was carried by the prosthesis (GII, Smith & Nephew, USA) and was re‐checked during the surgery. After automatic alignment with flexion and extension of the knee joint, the external rotation of the external rotation of the tibia baseplate was rechecked with guide rod to make sure that the proximal end of the rod was located at the medial 1/3 while the distal end pointed to the middle of the first and second metatarsal bones. Posterior cruciate ligament resection and bone cement fixation were performed. After the above procedure was carried out, the negative pressure drainage was performed in the joint cavity, as well as pressure dressing of the wound. On the first day after the operation, if the amount of drainage was less than 100 mL, the drainage tube and pressure dressing were removed. All patients performed continuous passive motion (CPM) function training, ankle pump training, and active straight‐leg raising training postoperatively and could get down to the ground 2–3 days after surgery.

Postoperative Analgesia Regimens

In the parecoxib group, 40 mg of parecoxib was administered 30 min before surgery. After the surgery, 40 mg of parecoxib dissolved in 2 mL of saline was administrated every 12 h. In the control group, 2 mL of pure saline was administrated at the same time as for the parecoxib group.

The PCA pump was applied in both groups. There was no background or loading infusion, with bolus of 1 mg of morphine at each press. Other analgesia agents were also administrated when necessary and were added to the total quantity after conversion to morphine equivalent.

Variables Observed

1. The VAS score at rest and in motion. Time of evaluation: before operation, after operation in PACU, 6 h after operation, 12 h after operation, 24 h after operation, 2 days after operation, 3 days after operation, 5 days after operation, and 14 days after operation.

2. Total PCA drug consumption. Morphine consumed = 1 mg × times of pressure.

3. Quantities of salvage analgesia drugs consumed. All analgesics were converted as morphine equivalent.

4. Assessment of knee joint function. The ROM of the knee joint was measured on postoperative days 1, 3, and 5. The active straight‐leg raising time, postoperative weight‐bearing time, and length of hospital stay were collected.

Safety Assessment

1. The vital signs of patients were monitored after the operation (twice or three times per day within 3 days after the operation, and daily after postoperative day 3).

2. Liver function, renal function, and coagulation function were evaluated at postoperation days 1, 3 and 5 and every 3 days after day 5.

3. Occurrence of complications, including nausea, vomiting, rashes, gastrointestinal bleeding, cardiovascular events, and delayed wound healing.

Statistical Analysis

SPSS 10.0 was used to analyze the data. Quantitative data were displayed as mean ± SD and compared with a paired t‐test. Median and non‐parameter tests were applied when the data did not show a normal distribution. Enumeration data were compared with the χ²‐test. The difference was considered significant if the P‐value was less than 0.05.

Ethical Approval

The procedures performed the study involving human participants were in accordance with the ethical standards of the institutional and/or national research committee and with the 1964 Declaration of Helsinki and its later amendments or comparable ethical standards.

Results

Demographic Profiles and Other Preoperative Data of the Patients

There was no difference in gender, age, height, weight, body mass index (BMI), and preoperative diagnosis between the two groups. The VAS score, the ROM of knee joints, and comorbidity did not differ between the two groups either (Table 1).

Table 1.

Preoperative data of the patients

Characters	Parecoxib group (n = 46)	Control group (n = 42)	Statistics	P‐value
Gender (male/female)	12/34	12/30	χ2 = 0.068	0.815
Diagnosis (OA/RA)	44/2	41/1	χ2 = 0.258	1.000
Comorbidity (with/without)	18/28	18/24	χ2 = 0.126	0.829
Age (mean ± SD, years)	66.64 ± 7.27	66.12 ± 8.34	t = 0.314	0.754
Height (mean ± SD, cm)	161.93 ± 7.24	161.57 ± 7.46	t = 0.232	0.817
Weight (mean ± SD, kg)	68.20 ± 11.34	70.73 ± 10.04	t = −1.104	0.273
BMI (mean ± SD, kg/m2)	26.09 ± 4.50	27.13 ± 3.77	t = −1.171	0.245
VAS at rest (mean ± SD)	1.72 ± 1.96	1.83 ± 2.02	t = −0.273	0.786
VAS in motion (mean ± SD)	5.39 ± 2.05	5.38 ± 1.65	t = 0.026	0.979
ROM, extension (mean ± SD, °)	5.22 ± 12.24	6.71 ± 17.36	t = −0.471	0.639
ROM, flexion (mean ± SD, °)	105.54 ± 21.35	104.76 ± 23.76	t = 0.163	0.871

BMI, body mass index; OA, osteoarthritis; RA, rheumatoid arthritis; ROM, range of motion; VAS, visual analogue scale.

Data Related to Surgical Procedures

There was no significant difference in operative duration, operative blood loss, postoperative drainage, and allogeneic transfusion between the two groups (Table 2).

Table 2.

Data related to surgical procedures

Characters	Parecoxib group (n = 46)	Control group (n = 42)	Statistics	P‐value
Operative duration (mean ± SD, min)	94.07 ± 18.13	91.38 ± 17.06	t = 0.713	0.478
Operative blood loss (mean ± SD, mL)	34.91 ± 35.54	43.14 ± 55.27	t = −0.838	0.404
Drainage (mean ± SD, mL)	309.57 ± 259.11	317.26 ± 212.56	t = −0.151	0.880
Transfusion (with/without)	8/38	5/37	χ2 = 0.525	0.556

Postoperative Visual Analogue Scale Score

The VAS score in the post‐anesthesia care unit (PACU) at rest was significantly lower in the parecoxib group than that of the control group (t = −2.099, P = 0.039). At other times, the VAS score did not show significant differences between the two groups either at rest or in motion (Figs 2 and 3).

Figure 2.

Postoperative visual analogue scale (VAS) score of patients at rest. The VAS score of the parecoxib group in the post‐anesthesia care unit was significantly lower than that of the control group (t = −2.099, P = 0.039).

Figure 3.

The postoperative visual analogue scale (VAS) score in motion. There was no significant difference between the two groups.

Postoperative Function, Patient‐controlled Analgesia Consumption, Salvaging Analgesia and Complications

There was no significant difference in active straight‐leg raising time, postoperative weight‐bearing time, ROM of the knee joint, drug consumption through the PCA pump, using salvaging analgesia, and length of hospital stay between the two groups. Complications included nausea, vomiting, rashes, dizziness and headache, and the occurrence did not differ significantly between the two groups either (Table 3). The adverse effects were resolved with appropriate treatments. One patient in the parecoxib group developed atrial fibrillation and one patient in the control group was somnolent after the operation. Primary healing was achieved in all wounds. The vital signs of patients were stable, with no abnormal liver, renal or coagulation function detected postoperatively.

Table 3.

Postoperative data of the patients (mean ± SD)

Characters	Parecoxib group (n = 46)	Control group (n = 42)	Statistics	P‐value
Active straight‐leg raising
Time (mean ± SD, days)	1.91 ± 1.41	2.17 ± 1.38	t = −0.852	0.397
Weight bearing
Time (mean ± SD, days)	3.15 ± 0.79	3.24 ± 0.85	t = −0.492	0.624
ROM, extension (mean ± SD, °)
POD1	4.07 ± 5.73	6.07 ± 10.39	t = −1.134	0.260
POD3	3.37 ± 5.28	4.52 ± 5.72	t = −0.985	0.327
POD5	3.04 ± 4.53	3.81 ± 5.50	t = −0.715	0.476
ROM, flexion (mean ± SD, °)
POD1	57.39 ± 21.10	60.48 ± 27.58	t = −0.592	0.555
POD3	85.54 ± 11.56	82.14 ± 18.05	t = 1.042	0.301
POD5	104.89 ± 14.63	99.17 ± 22.19	t = 1.440	0.153
Consumption of PCA
Pump (mean ± SD, mg)	48.86 ± 38.35	51.33 ± 34.56	t = −0.316	0.752
Postoperative salvaging
Analgesia (with/without)	7/39	14/28	χ2 = 3.966	0.078
Length of hospital
stay (mean ± SD, days)	8.11 ± 2.42	8.74 ± 3.37	t = −1.012	0.314
Complications (with/without)
Nausea and vomiting	8/38	9/33	χ2 = 0.230	0.788
Rashes	2/44	1/41	‐	‐
Arrhythmia	1/45	0/42	‐	‐
Headache and dizziness	3/43	2/40	‐	‐
Somnolence	0/46	1/41	‐	‐
Impaired wound healing	0/46	0/42	‐	‐
Sum	14/32	13/29	χ2 = 0.003	1.000

PCA, patient‐controlled analgesia pump; ROM, range of motion.

Discussion

The idea of preemptive analgesia was first proposed by Woolf in Harvard Medical School8, with the purpose to block the pain sensation center before harmful stimulation occurs, to inhibit the center sensitization and to raise the threshold of pain sensation. Recently, there have been many studies published on the preoperative use of local anesthesia agents, including opiates and NSAIDS9, 10, 11, 12, but the value of preemptive analgesia is still controversial13, 14, 15.

It has been proved that synthesis of COX‐2 increases after inflammatory stimulation, and increases the production of prostaglandins, which induces inflammatory pain. Meanwhile, COX‐2 receptor generalizes widely in the peripheral and central nervous system, and, consequently, COX‐2 may also sensitize the peripheral and central receptor and, thus, induce pain. As a pro‐drug of valdecoxib, parecoxib is transformed into valdecoxib in the liver, which is an inhibitor of COX‐2 and could, therefore, reduce pain by inhibiting synthesis of prostaglandins. Previous studies have indicated that parecoxib relieved post‐TKA pain effectively and reduced the consumption of opiates16, 17, 18.

The essentials of using parecoxib as a preemptive analgesia lie with changed to depend on the time and duration of the administration, as well as the analgesia effect of the drug itself. According to the definition, preemptive analgesia is not simply applying analgesics before the first injury occured19. The idea is to prevent the sensitization of the peripheral and central nervous system. To realize that goal, analgesics should be administrated before harmful stimulation. As for the duration of administration, the analgesics should cover the time from the harmful stimulation to central nervous excitation. The operation is the first phase of sensitization, and the inflammation after the operation is the second phase. The effect of preemptive analgesia is related to the intensity of harmful stimulation. In previous studies, the time of administration varied significantly. Martinez et al. gave the drug before anesthesia induction13, while Bajaj et al. injected the analgesics 30 to 45 min before anesthesia20. Based on the pharmacokinetics of parecoxib sodium, the serum concentration peaks at 30 min or 1 h after intravenous or intramuscular administration. Consequently, the controversy over the analgesic effect of parecoxib may be caused by the differences in defining “preemptive”. In our study, we administered the first dose 30 min before administering the first dose 30 min intravenously before operation, in expectation of achieving satisfactory blood concentration of parecoxib sodium before noxious stimulation. As for the schedule and intensity of administration, we administered 40 mg of parecoxib sodium twice a day for no more than 3 days after the operation. Opiates and other analgesics were added when necessary.

In our study, we found that the VAS score at rest is significantly lower in the preemptive group than that of the control group (t = −2.099, P = 0.039), while there was no significant difference in the VAS score after the operation. The range of motion, active straight‐leg raising time, postoperative weight‐bearing time, and length of hospitalization were similar in both groups, which was also found in previous studies. Preemptive analgesia decreases the VAS score of patients early after surgery, and, therefore, reduces the usage of opiates, as well as the adverse effects associated with opiates, including nausea, vomiting, dizziness, and constipation. In the long term, preemptive analgesia did not affect pain and knee joint function. Several studies focused on inflammation indicators such as hs‐CRP after surgery, and showed significant decreases 1 week after the operation compared to control groups10.

Conclusion

Preemptive analgesia with parecoxib sodium decreases the VAS score in the early postoperative period and reduces the consumption of opiates and associated adverse effects. However, there was no long‐term difference between the two groups.

Limitations

There are some limitations of this study. First, we studied only single‐dose parecoxib sodium without including a dose‐dependent control group, so we could not detect the most effective and appropriate dose of parecoxib sodium for patients undergoing TKA. Second, the sample size of our study was smaller than in other similar studies. We could not exclude the potential bias caused by the gender distribution of both groups, with many more female patients included than male patients. Further studies could recruit more cases to increase the detection power provided by the sample size and exclude sources of potential bias. Third, all patients had short‐term follow‐up until they were discharged; therefore, we could not verify the VAS pain score or joint function of the knee after a long period postoperatively.