Postoperative pain continues to be managed suboptimally after noncardiac surgery.1 The American Society of Anesthesiologists’ Task Force on Acute Pain Management recommends the use of multimodal analgesia to manage acute postoperative pain.2 NSAIDs and cyclooxygenase (COX)-2 inhibitors are integral components of multimodal analgesic regimens. NSAID use in the perioperative period has been limited by concerns regarding potential risks of bleeding, gastrointestinal side-effects and acute kidney injury.3–6
It is hypothesised that analgesics given before a nociceptive stimulus are more effective than after the stimulus has occurred.7 This concept of preemptive analgesia was first introduced by Woolf in 1983 and has extended into a clinical practice of administering analgesics prior to surgery.8 A systematic review of preoperative COX-2 inhibitors has been published previously and suggested improved analgesia and patient satisfaction after surgery.9 However, this review included several publications that were subsequently retracted because of fraudulent data and included two COX-2 inhibitors that were withdrawn from the market (rofecoxib and valdecoxib). Celecoxib is one of the most widely used COX-2 inhibitors, but there are no systematic reviews summarising benefits of preoperative celecoxib in noncardiac surgery.
The purpose of this study was to systematically review randomised controlled trials (RCTs) assessing the analgesic effects of preoperative celecoxib in noncardiac surgery.
An internal protocol was developed and circulated among authors prior to conducting the review. This protocol was not published.
We included RCTs that enrolled patients of at least 18 years of age (if not specified then lower end of the standard deviation for the mean age or first quartile for the median age had to be ≥18) who were randomised to receive celecoxib within 4 h of noncardiac surgery. Studies were excluded for the following criteria: animal studies; reviews/meta-analyses; did not report pain as an outcome; or provided epidural analgesia. There were no restrictions with respect to language, year of publication or publication status.
The primary outcome of this systematic review was the 24-h parenteral morphine equivalent consumption. Secondary outcomes included pain scores 24 h after surgery, chronic postsurgical pain (incisional pain present >3 months after index surgery), length of stay in post anaesthesia care unit (PACU), patient satisfaction, intra-operative and postoperative bleeding, nausea, vomiting and myocardial infarction (MI).
A comprehensive search strategy was developed with the aid of a university research librarian prior to conducting the systematic search. An example of a search strategy used for Ovid MEDLINE is depicted in Appendix 1 (supplemental data file, http://links.lww.com/EJA/A78). The following databases were searched: Ovid MEDLINE in-process and other nonindexed citations, Ovid MEDLINE daily and Ovid MEDLINE; EMBASE; CINHAL; Cochrane CENTRAL; Web of Science; OpenGrey; ProQuest ERIC and ProQuest theses and dissertations; and open access theses and dissertations. Databases were searched from inception to September 2014 using individualised search strategies for each database. Informal searches were performed using Google Scholar to obtain any additional studies. An international clinical trial registry (www.clinicaltrials.gov) was searched and the references of relevant trials and reviews were manually searched to identify trials not found in the electronic search. A request was sent to Pfizer, manufacturer of celecoxib, to gather any relevant internal or unpublished studies that they possessed.
Results from the systematic search were initially screened for intra-database and inter-database duplicates. After removing duplicates, a two-stage screening and selection process was employed and standardised screening forms were created, piloted and used for selection at each stage. Title and abstract screening was performed in duplicate by three independent investigators (J.K, C.M. and S.C.) according to the inclusion criteria. Subsequently, three independent investigators (J.K, C.M. and S.C.) reviewed the full text of included citations in duplicate. Study authors were contacted if further clarification was needed. A κ-statistic for agreement was calculated. Consensus on disagreements was met through discussion by all study investigators.
A data collection form was created and piloted using three of the identified studies. Data extraction occurred independently and in duplicate. Study authors were contacted for clarification or if additional information was needed. Data on patient demographics, pain, opioid consumption, nausea, vomiting, length of hospital stay and adverse events were collected.
Two independent reviewers assessed the methodological quality of each study according to the Cochrane Collaboration's tool for assessing risk of bias.
Pain scores, opioid consumption, intraoperative and postoperative bleeding, and patient satisfaction were considered continuous variables and pooled using mean differences; pain scores were converted to a 0–10 visual analogue scale, opioids were converted into morphine equivalents and patient satisfaction ratings were converted to a 0–100 scale. Categorical data (e.g. nausea and vomiting) were summarised using pooled relative risks (RRs) and 95% confidence interval (CI). Statistical analysis was conducted using RevMan 5.2.9 software (Cochrane IKMD, London, UK).
If required data were unavailable from exact numerical reporting, we used approximations based on graphic output. For studies reporting median and interquartile range (IQR) for desired outcomes, these data were excluded from meta-analysis because of inaccuracies in converting to means and standard deviations without knowledge of the distribution of the data. Opioids were converted to parenteral morphine equivalents using established conversion ratios.10–13 When multiple pain scores were present (e.g. pain scores at rest, movement and coughing), pain scores at rest were collected and analysed. Results from a trial were included into a meta-analysis if sufficient data were available for both experimental and control groups. We used standard inverse-variance methods to pool continuous outcomes and the Mantel-Haenszel method to pool RRs for categorical outcomes. The random-effects model was chosen for all outcomes. Heterogeneity between trials was quantified using the I2 statistic.
Two a-priori subgroup analyses were planned to explain heterogeneity: comparing preoperative dosing of celecoxib 200 mg with celecoxib 400 mg, and comparing a preoperative-only dose to preoperative and continued postoperative celecoxib dosing. Subgroups were compared using postoperative 24-h parenteral morphine equivalent consumption and 24-h pain scores. The a priori direction of effect was that preoperative celecoxib 400 mg would be superior to celecoxib 200 mg and that preoperative with continued postoperative dosing would be superior to a preoperative-only dose. A sensitivity analysis on the primary outcome was also planned in which trials that had one or more categories at high risk of bias on the Cochrane risk of bias tool were excluded.
A trial sequential analysis (TSA) was conducted on the primary outcome to identify the effect of random error on the pooled effect estimation. This analysis is similar to an interim analysis of a single RCT to determine if prespecified stopping boundaries are met. In TSA, updating a meta-analysis with new trial information is analogous to interim monitoring and similar boundaries are used to determine if an effect is positive and no further patients are needed. Using the TSA software available from the Copenhagen Clinical Trials group, we undertook an α-spending analysis using the O’Brien-Fleming boundaries. Required information size (sample size needed) was calculated using a maximum type I error of 5% and a maximum type II error of 10%, using a mean reduction of 3 mg of morphine and a variance estimated with data from included trials at low risk of bias. Information sizes were heterogeneity-adjusted using the estimate of diversity.
The systematic search yielded 3434 potentially relevant citations. After removing intra-database and inter-database duplicates, a total of 2403 citations remained for title and abstract screening. Thirty-four articles underwent full-text review. The κ agreement after full-text review was 0.90 (95% CI 0.82 to 0.97). Any disagreements were discussed by all study authors and consensus was achieved. Fourteen articles were excluded after full-text review (Fig. 1).
Twenty trials were included for review, comprising a total of 1445 randomised patients. Of these patients, 747 patients received celecoxib and 748 received placebo or standard care. One trial could only be retrieved in Korean.14 One of the investigators (S.C.) could read Korean and the article was translated into English for a second investigator (J.K.).
Many of the included trials were multiarm RCTs with an additional randomisation arm receiving another preoperative medication (e.g. paracetamol).15–25 Data only for the celecoxib and placebo arms were collected from these trials. One trial26 included and separated two cohorts (anterior cruciate ligament repair or meniscectomy) for randomisation and analysis, and thus these cohorts are separately presented in the quantitative analysis. Furthermore, two trials14,27 included a placebo group with two cohorts for celecoxib 200 mg and celecoxib 400 mg. These cohorts were combined so that the placebo group would not be represented twice in the review; however, these cohorts were kept separate for the subgroup analysis comparing celecoxib 200 and 400 mg.
Twelve studies did not report on funding sources. Three indicated that they were funded by the University,25,27,28 one by the Government,24 two by industry or private companies18,29 and two were funded by both industry and institutional support.30,31
Table 1 depicts the characteristics of the 20 included trials. One trial used a crossover trial design23 whereas the remaining 19 trials used a parallel trial design. Five trials were conducted in the USA,18,25,27,29,31 three in Thailand,17,19,22 two each in Brazil,23,24 Iran21,26 and Turkey,15,20 and one trial each from China,32 Finland,16 Germany,30 Korea,14 Sweden33 and Taiwan.28 Patients in the included trials underwent a variety of surgical procedures including abdominal, oral, orthopaedic, otolaryngological, plastic, spinal and thyroid surgery. Seven trials administered an additional preoperative medication to both the intervention and control groups; one study33 provided paracetamol 1 g and six studies provided midazolam.18,20,25,27,30,31
The overall methodological quality of the included trials was moderate to low (Appendix 2, supplemental data file, http://links.lww.com/EJA/A78). Nine trials were identified as having a high risk of bias because of at least one category in the Cochrane risk of bias tool. The one study23 utilising a crossover trial design did not assess for order effect. Only seven studies reported patients who were lost to follow-up. Two studies did not use an intention-to-treat analysis.16,24
Cumulative opioid consumption over the first 24 h postoperatively was reported in 17 trials. Fourteen trials were amenable to meta-analysis (994 patients). One trial29 reported only means but not standard deviations; however, this trial demonstrated a significantly lower opioid consumption in the celecoxib group. Two trials reported opioid consumption as median and IQR; both demonstrated a significant reduction in opioid consumption in PACU and 24 h after surgery.15,27 The pooled 24-h cumulative parenteral morphine equivalent consumption demonstrated a mean difference of 4.1 mg (95% CI −5.6 to −2.7, I2 = 94%) (Fig. 2).
Acute postoperative pain scores were reported in 18 trials. Only seven, eight, seven and 11 trials were amenable to quantitative review for the time points of PACU, 4, 12 and 24 h postoperatively, respectively. One trial29 did not provide standard deviations in the manuscript required for quantitative pooling; this trial did demonstrate significantly lower pain scores at rest in the celecoxib group 8, 10 and 12 h after surgery compared with placebo. Data from three studies were not included because their outcomes were expressed as median and IQR; differences in pain scores were found in PACU for one trial15 and at home for another trial,25 whereas one trial did not detect a difference.18 In the meta-analysis, there was a significant difference in the celecoxib groups for 24-h pain scores [mean difference −1.02 (on a 0–10 pain scale), 95% CI −1.54 to −0.50, I2 = 99%] (Fig. 3). There was also a significant reduction in pain scores in the celecoxib groups in PACU (mean difference −1.23, 95% CI −1.69 to −0.76, I2 = 87%), at 4 h (mean difference −1.09, 95% CI −1.87 to −0.32, I2 = 99%) and at 12 h (mean difference −0.31, 95% CI −0.50 to −0.13, I2 = 77%) after surgery (Appendix 2, supplemental data file, http://links.lww.com/EJA/A78).
None of the studies assessed the impact on chronic postsurgical pain. One study assessed pain scores during a 1 to 12-month follow-up period and indicated no difference between celecoxib and control group pain scores.33
Four trials reported on the length of PACU stay.18,25,27,31 None of these trials found a significant decrease in the duration of time spent in PACU in the celecoxib group compared with the placebo group. Although there was a trend towards significance, there was no difference in length of PACU stay in the meta-analysis (mean difference −5.14 min, 95% CI −11.10 to 0.82, I2 = 0%) (Appendix 2, supplemental data file, http://links.lww.com/EJA/A78).
In-hospital patient satisfaction was reported in nine trials.15,17–19,22,25,27,29,31 Six trials assessed this using a 0–100 scale, two a 0–10 scale and one used a binary outcome for satisfaction; the two using a 0–10 scale were converted to a 0–100 scale. The seven trials assessing patient satisfaction as a continuous outcome were amenable to quantitative review and did not show a significant difference in the celecoxib group (mean difference 2.95, 95% CI −0.12 to 6.03) (Appendix 2, supplemental data file, http://links.lww.com/EJA/A78). The trial that examined satisfaction as a binary outcome reported a significant result wherein 75 and 58% of patients were satisfied in the celecoxib and placebo groups, respectively (P = 0.01). One trial reported patient satisfaction as median and IQR and did not report a significant difference.15
All of the trials reported on adverse events within the postoperative period. Nine trials assessed postoperative in-hospital nausea; five of these were included in a meta-analysis and demonstrated a significant reduction (RR 0.56, 95% CI 0.36 to 0.89, I2 = 0%). The number needed to treat (NNT) was 11 (Fig. 4). Of the 13 trials reporting on postoperative in-hospital vomiting, 10 were amenable to meta-analysis and also indicated a significant difference between the groups (RR 0.62, 95% CI 0.40 to 0.96, I2 = 0%); the NNT was 20 (Fig. 5). The three trials that were not included in the meta-analysis found comparable rates of vomiting between the celecoxib and placebo groups.21,22,32
Intraoperative and postoperative bleeding events were assessed in seven trials. The pooled estimate for intraoperative bleeding in millilitres of blood loss indicated comparable bleeding in the celecoxib and control groups (mean difference 1.98, 95% CI −15.97 to 19.94, I2 = 39%) (Appendix 2, supplemental data file, http://links.lww.com/EJA/A78).
Perioperative MI was assessed in one trial and found no events in either group.31 One trial comprehensively assessed for 19 different adverse events.29 The authors reported a significantly higher rate of all adverse events in the placebo group (37%) compared with the celecoxib group (18%). They also reported equal incidences of events relating to bronchospasm and bleeding.
Sensitivity analysis was conducted by removing the nine trials14–16,21,22,24,27,28 that were judged to have a ‘high risk’ of bias in at least one category. When these trials were removed, the pooled estimates for postoperative 24-h pain (mean difference −1.46, 95% CI −2.24 to −0.68, I2 = 99%) and 24-h cumulative parenteral morphine equivalent consumption (mean difference −4.49 mg, 95% CI −6.25 to −2.73, I2 = 94%) did not change substantially (Appendix 2, supplemental data file, http://links.lww.com/EJA/A78).
Our a priori subgroup analyses comprised comparing preoperative celecoxib 200 mg with 400 mg and comparing a preoperative-only dose with preoperative and continued postoperative celecoxib dosing. Subgroup analyses were performed on the primary outcome and 24-h cumulative parenteral morphine equivalent consumption and indicated no difference in the subgroups (Appendix 2, supplemental data file, http://links.lww.com/EJA/A78).
The calculated information size (number of patients needed) using a type I error rate of 5% and a type II error rate of 10%, while adjusting for heterogeneity, was 1337 patients. Although our primary outcome was 343 patients short of this required size, α-spending analysis indicated that the cumulative z score of included patients in this meta-analysis was well above the O’Brien-Fleming boundaries (Appendix 2, supplemental data file, http://links.lww.com/EJA/A78). This indicates that the review had sufficient power to detect a minimum difference of 3 mg between the groups.
This study represents the first review of preoperative celecoxib in noncardiac surgery. Our results indicate that preoperative celecoxib is associated with a slight reduction in 24-h postoperative opioid consumption and pain scores. Furthermore, there seems to be no increased analgesic benefit of using a 400 mg dose over a 200 mg dose or continuing administration postoperatively.
There are several limitations to this study. First, there was significant heterogeneity, as demonstrated by the large I2 for many of the outcomes. Although this heterogeneity probably arises from the differential impact of preoperative celecoxib on different surgical procedures, patient populations and perioperative care regimens, we believe the majority of heterogeneity arises from the small sample sizes of the included studies. Small studies lack precision and are more susceptible to random play of chance; these random variations in individual study effect estimations lead to greater heterogeneity across studies. Although meta-analyses seek to improve precision, random error of included studies can influence pooled results. Although our trial sequential analysis revealed sufficient statistical power for our primary outcome and we used the random-effects model for each of our analyses (providing a more conservative estimate), we suggest careful interpretation of our results given the significant heterogeneity and the fact that the majority of included studies were small. Second, trials in our review were of moderate-to-low quality. Despite our sensitivity analysis demonstrating a continued effect of celecoxib for our analgesic outcomes, every trial in this review had an ‘unclear’ judgement in one of the categories in the Cochrane risk of bias tool. As the presence and role of bias in the included trials is not immediately apparent, the impact it has on our results is also unknown. Third, it has been suggested that mean postoperative morphine consumption is not an appropriate measure of analgesia due its skewed distribution.34 However, we were unable to use another measure given the limited data reported in the included trials.
Although this study suggests potential analgesic benefits of preoperative celecoxib, it is limited in providing insight into adverse events because the majority of included studies focused on efficacy, with limited attention to safety; only seven studies commented on bleeding and only one reported on MI. A previous meta-analysis indicated that administration of NSAIDs after surgery is associated with increased bleeding compared with placebo.3 Although COX-2 inhibitors are theoretically devoid of any platelet inhibiting effect, a notion supported by in-vitro studies,35 there exist concerns that preoperative COX-2 inhibitors would promote bleeding from an open wound. Although our results suggest no increase in bleeding with preoperative celecoxib, only three studies were included in this meta-analysis. Furthermore, although a previous meta-analysis on celecoxib use did not show an association with MI (odds ratio 2.26, 95% CI 1.0 to 5.1),36 there is a signal towards harm. Large registry data suggest that patients with a previous MI are at an increased risk of death and nonfatal MI when taking an NSAID (including a COX-2 inhibitor) 37,38; and that patients with atrial fibrillation on antithrombotic therapy are at an increased risk of cardiovascular events, bleeding and thromboembolism.39 Although much of this data is based on chronic NSAID use in at-risk populations and there are no data suggesting an increase in risk with a single dose of celecoxib, we suggest cautious use of celecoxib in the perioperative period until further safety data are acquired.
In addition, there exists great debate on whether COX-2 inhibitors impair fracture healing because of the essential role of COX in bone metabolism. Both animal and human studies provide conflicting results and much of the clinical research in this area arises from retrospective studies.40,41 Furthermore, it appears that COX-2 expression peaks 5 days after initial fracture and because of the 11-h half-life of celecoxib, it is unlikely that a single dose of celecoxib prior to surgery will significantly impair bone remodelling and fracture healing.
Celecoxib exerts its analgesic effect by binding to the COX-2 isoform which is responsible for prostaglandin synthesis during inflammation.42 COX-2 is also expressed in the brain and spinal cord, which is thought to explain its role in central sensitisation of pain.42 NSAIDs inhibit both COX-1 and 2 isoforms, whereas celecoxib has a 375-fold selectivity for the COX-2 isoform.6 Our results indicate that preoperative celecoxib in noncardiac surgery may be associated with a decrease in opioid consumption. Although these findings were statistically significant, the clinical significance is not readily apparent. In itself, a 4-mg decrease in morphine equivalents may not be meaningful; however, there seems to be some benefit from this reduction. Preoperative celecoxib resulted in lower rates of nausea and vomiting after surgery, an effect probably mediated through a reduction in opioid consumption.43,44 Furthermore, the reduction in opioid consumption may become more meaningful if celecoxib is combined with other preemptive medications. Previous studies suggest that preemptive multimodal regimens that include NSAIDs and COX-2 inhibitors allow for a greater reduction in opioid consumption and improved analgesia.45,46
In addition, preoperative celecoxib resulted in a 1-point difference on a 0–10 pain scale. It has been suggested that a 1 to 2-point decrease on a 0–10 scale denotes a minimally important difference.47 Consequently, although preoperative celecoxib results in a minimally important reduction in pain scores, it is not clear whether this difference is important in the clinical setting. Similarly, celecoxib reduced nausea and vomiting by 44 and 38%, respectively, but the NNT was 11 for nausea and 20 for vomiting. Although these NNT values are good relative to other commonly used medications in clinical practice, they demonstrate that only 5–10% of those treated will experience this benefit.
Our study suggests that celecoxib may be an effective preemptive analgesic. Despite experimental evidence suggesting the superiority of preemptive analgesia over postoperative analgesia, clinical evidence has not been conclusive. A meta-analysis found 20 RCTs that compared preincisional NSAIDs with postincisional NSAIDs.48 Four of the 20 studies indicated that preincisional NSAIDs allowed for reduced pain scores and morphine consumption. However, when compared with postincisional NSAID use, there were no differences in pain scores.48 Since publication of this meta-analysis, several trials have indicated lower postoperative pain with preemptive NSAIDs over postoperative administration.49,50 Nonetheless, the literature is inconclusive and it provides little insight into preemptive efficacy of COX-2 inhibitors. Further research is needed on whether administering celecoxib prior to surgery is more effective than doing so after surgery.
Overall, preoperative celecoxib appears to provide a modest reduction in postoperative morphine consumption, pain, nausea and vomiting after noncardiac surgery. The decision to use preoperative celecoxib in clinical practice must be viewed in the light of limited safety data in the perioperative context, and the fact that results from this study had significant heterogeneity and were derived small from mainly trials. Nonetheless, this study suggests that celecoxib may be a potentially useful adjunct in preoperative multimodal regimens to improve postoperative pain and recovery.
Acknowledgements relating to this article
Assistance with the study: we acknowledge the assistance of the study authors Drs Pilatti, Karst, Boonriong and Ittchaikulthol for providing additional information and data to be included in our systematic review.
Financial support and sponsorship: none.
Conflict of interest: none
1. Apfelbaum JL, Chen C, Mehta SS, Gan TJ. Postoperative pain experience: results from a national survey suggest postoperative pain continues to be undermanaged. Anesth Analg
2. Apfelbaum J, Ashburn MA, Connis RT, et al. Practice guidelines for acute pain management in the perioperative setting: an updated report by the American Society of Anesthesiologists Task Force on Acute Pain Management. Anesthesiology
3. Maund E, McDaid C, Rice S, et al. Paracetamol and selective and nonselective nonsteroidal anti-inflammatory drugs for the reduction in morphine-related side-effects after major surgery: a systematic review. Br J Anaesth
4. Schafer AI. Effects of nonsteroidal anti-inflammatory therapy on platelets. Am J Med
5. Brater DC. Effects of nonsteroidal anti-inflammatory drugs on renal function: focus on cyclooxygenase-2-selective inhibition. Am J Med
6. Hawkey CJ. COX-2 inhibitors. Lancet
7. Dahl JB, Møiniche S. Preemptive analgesia. Br Med Bull
8. Woolf C. Evidence for a central component of postinjury pain hypersensitivity. Nature
9. Straube S, Derry S, McQuay HJ, Moore RA. Effect of preoperative Cox-II-selective NSAIDs (coxibs) on postoperative outcomes: a systematic review of randomized studies. Acta Anaesthesiol Scand
10. Scott LJ, Perry CM. Tramadol: a review of its use in perioperative pain. Drugs
11. Berdine HJ, Nesbit SA. Equianalgesic dosing of opioids. J Pain Palliat Care Pharmacother
12. Kumar N, Rowbotham D. Piritramide. Br J Anaesth
13. Svendsen K, Borchgrevink PC, Fredheim O, et al. Choosing the unit of measurement counts: the use of oral morphine equivalents in studies of opioid consumption is a useful addition to defined daily doses. Palliat Med
14. Kim D-Y, An Y-M, Moon J-P, Choi J-G. Effect of preoperative administration of celecoxib on postoperative pain after laparoscopic cholecystectomy. Korean J Anesthesiol
15. Alanoglu Z, Ateş Y, Orbey BC, Türkçapar AG. Preoperative use of selective COX-II inhibitors for pain management in laparoscopic nissen fundoplication. Surg Endosc
16. Al-Sukhun J, Al-Sukhun S, Penttilä H, et al. Preemptive analgesic effect of low doses of celecoxib is superior to low doses of traditional nonsteroidal anti-inflammatory drugs. J Craniofac Surg
17. Boonriong T, Tangtrakulwanich B, Glabglay P, Nimmaanrat S. Comparing etoricoxib and celecoxib for preemptive analgesia for acute postoperative pain in patients undergoing arthroscopic anterior cruciate ligament reconstruction: a randomized controlled trial. BMC Musculoskelet Disord
18. Issioui T, Klein KW, White PF, et al. The efficacy of premedication with celecoxib and acetaminophen in preventing pain after otolaryngologic surgery. Anesth Analg
19. Ittichaikulthol W, Prachanpanich N, Kositchaiwat C, Intapan T. The postoperative analgesic efficacy of celecoxib compared with placebo and parecoxib after total hip or knee arthroplasty. J Med Assoc Thail
20. Karamanlioğlu B, Arar C, Alagol A, et al. Preoperative oral celecoxib versus preoperative oral rofecoxib for pain relief after thyroid surgery. Eur J Anaesthesiol
21. Kashefi P, Honarmand A, Safavi M. Effects of preemptive analgesia with celecoxib or acetaminophen on postoperative pain relief following lower extremity orthopedic surgery. Adv Biomed Res
22. Niruthisard S, Earsakul A, Bunburaphong P, Chinda P. Preoperative pregabalin and/or celecoxib for pain management after total knee arthroplasty under intrathecal morphine: a randomized controlled trial. Asian Biomed
23. Pilatti GL, André dos Santos F, Bianchi A, et al. The use of celecoxib and dexamethasone for the prevention and control of postoperative pain after periodontal surgery. J Periodontol
24. Steffens JP, Santos FA, Pilatti GL. The use of etoricoxib and celecoxib for pain prevention after periodontal surgery: a double-masked, parallel-group, placebo-controlled, randomized clinical trial. J Periodontol
25. Watcha MF, Issioui T, Klein KW, White PF. Costs and effectiveness of rofecoxib, celecoxib, and acetaminophen for preventing pain after ambulatory otolaryngologic surgery. Anesth Analg
26. Mardani-Kivi M, Mobarakeh MK, Haghighi M, et al. Celecoxib as a preemptive analgesia after arthroscopic knee surgery; a triple-blinded randomized controlled trial. Arch Orthop Trauma Surg
27. Recart A, Issioui T, White PF, et al. The efficacy of celecoxib premedication on postoperative pain and recovery times after ambulatory surgery: a dose-ranging study. Anesth Analg
28. Ekman EF, Wahba M, Ancona F. Analgesic efficacy of perioperative celecoxib in ambulatory arthroscopic knee surgery: a double-blind, placebo-controlled study. Arthroscopy
29. Sun T, Sacan O, White PF, et al. Perioperative versus postoperative celecoxib on patient outcomes after major plastic surgery procedures. Anesth Analg
30. Cheng P, Lim MJ, Onsiong MK, et al. Celecoxib premedication in postoperative analgesia for laparoscopic cholecystectomy. Acute Pain
31. Karst M, Kegel T, Lukas A, et al. Effect of celecoxib and dexamethasone on postoperative pain after lumbar disc surgery. Neurosurgery
32. Meunier A, Lisander B, Good L. Effects of celecoxib on blood loss, pain, and recovery of function after total knee replacement: a randomized placebo-controlled trial. Acta Orthop
33. Huang Y-M, Wang C-M, Wang C-T, et al. Perioperative celecoxib administration for pain management after total knee arthroplasty – a randomized, controlled study. BMC Musculoskelet Disord
34. Moore R, Mhuircheartaigh R, Derry S, McQuay H. Mean analgesic consumption is inappropriate for testing analgesic efficacy in postoperative pain: analysis and alternative suggestion. Eur J Anaesthesiol
35. Noveck R, Laurent A, Kuss M, et al. Parecoxib sodium does not impair platelet function in healthy elderly and nonelderly individuals: two randomised, controlled trials. Clin Drug Investig
36. Caldwell B, Aldington S, Weatherall M, et al. Risk of cardiovascular events and celecoxib: a systematic review and meta-analysis. J R Soc Med
37. Olsen AMS, Gislason GH, McGettigan P, et al. Association of NSAID use with risk of bleeding and cardiovascular events in patients receiving antithrombotic therapy after myocardial infarction. J Am Med Assoc
38. Olsen AMS, Fosbol EL, Lindhardsen J, et al. Duration of treatment with nonsteroidal anti-inflammatory drugs and impact on risk of death and recurrent myocardial infarction in patients with prior myocardial infarction: a nationwide cohort study. Circulation
39. Lamberts M, Lip GYH, Hansen ML, et al. Relation of nonsteroidal anti-inflammatory drugs to serious bleeding and thromboembolism risk in patients with atrial fibrillation receiving antithrombotic therapy: a nationwide cohort study. Ann Intern Med
40. Dodwell ER, Latorre JG, Parisini E, et al. NSAID exposure and risk of nonunion: a meta-analysis of case-control and cohort studies. Calcif Tissue Int
41. Geusens P, Emans PJ, de Jong JJ, Van den Bergh J. NSAIDs and fracture healing. Curr Opin Rheumatol
42. Vane JR, Bakhle YS, Botting RM. Cyclooxygenases 1 and 2. Annu Rev Pharmacol Toxicol
43. Oderda GM, Evans RS, Lloyd J, et al. Cost of opioid-related adverse drug events in surgical patients. J Pain Symptom Manage
44. Philip BK, Reese PR, Burch SP. The economic impact of opioids on postoperative pain management. J Clin Anesth
45. Trabulsi E, Patel J, Viscusi EC, et al. Preemptive multimodal pain regimen reduces opioid analgesia for patients undergoing robotic-assisted laparoscopic radical prostatectomy. Urology
46. Michaloliakou C, Chung F, Sharma S. Preoperative multimodal analgesia facilitates recovery after ambulatory laparoscopic cholecystectomy. Anesth Analg
47. Dworkin RH, Turk DC, McDermott MP, et al. Interpreting the clinical importance of group differences in chronic pain clinical trials: IMMPACT recommendations. Pain
48. Moiniche S, Kehlet H, Berg JA. Qualitative and quantitative systematic review of preemptive analgesia for postoperative pain relief. Anesthesiology
49. Bernhardt MK, Southard KA, Batterson KD, et al. The effect of preemptive and/or postoperative ibuprofen therapy for orthodontic pain. Am J Orthod Dentofacial Orthop
50. Norman PH, Daley MD, Lindsey RW. Preemptive analgesic effects of ketorolac in ankle fracture surgery. Anesthesiology
Supplemental Digital Content
© 2016 European Society of Anaesthesiology