Despite preventive strategies against opioid side effects, including postoperative nausea and vomiting (PONV), respiratory depression, urinary retention, constipation, and pruritus, these adverse events remain a great concern in the postoperative period and a major source of discomfort for patients.1
Many adult studies have shown non-opioid analgesics such as ketamine, nonsteroidal antiinflammatory drugs (NSAIDs), acetaminophen, and nefopam to exhibit an opioid-sparing effect and improve postoperative analgesia quality.2–6
A qualitative meta-analysis addressed the efficacy of postoperative morphine in children.1 Among its findings, morphine was only effective when compared with a control therapy. Moreover, it was associated with significant side effects such as nausea and postoperative urinary retention. To decrease the incidence of such undesirable effects, the authors concluded that it is necessary to use balanced analgesia when morphine is used.
NSAIDs have been shown to improve postoperative pain both in adult and pediatric populations.4,6,7 However, the opioid-sparing effects of NSAIDs in children and infants are still the subject of debate. In addition, it is still unclear whether combined administration of opioids and NSAIDs improves postoperative pain relief in children.
Meta-analysis of published studies is a statistical methodology allowing aggregation and quantification of the therapeutic effects from multiple studies. It is particularly interesting when conflicting results arise from multiple studies. The primary goal of the present study was to investigate the efficacy of balanced analgesia using NSAIDs and opioids for the management of postoperative pain in children with a special emphasis on both the quality of postoperative analgesia and the opioid-sparing effects of NSAIDs.
METHODS
Bibliographical Search and Analysis
We conducted this meta-analysis according to the guidelines of the Cochrane Handbook forSystematic ReviewsofIntervention, the QUORUM (Quality of Reporting of Meta-analyses) statements, and general guidelines on this statistical method.3,6,8–13
Research was performed on 2 literature databases: PubMed and Embase. The following queries were used: “name of the NSAIDs and children or infant.” The names used included nonselective NSAIDs: diclofenac, dexibuprofen, dexketoprofen, indometacin, ibuprofen, ketoprofen, ketorolac, lornoxicam, mefenamic acid, meloxicam, nabumetone, naproxen, piroxicam, tenoxicam, tiaprofenic acid, and cyclooxygenase-2 selective drugs (celecoxib, etoricoxib, lumiracoxib, parecoxib, rofecoxib, valdecoxib). In addition, a manual search of the references found in the selected articles (including reviews and meta-analyses) was also performed. The date of the most recent search was April 2011. Articles were analyzed by 4 physicians to verify their relevance: presence of postoperative opioid therapy in both arms and presence of a control group without NSAID medication. Readers also assessed the potential risk of bias as recommended by Cochrane experts13: randomization and allocation concealment (clear description of the method used in sufficient detail to determine whether intervention allocations could have been foreseen in advance of, or during, enrollment), double-blind study, incomplete data report statement (statement about excluded patients and data), and absence of selective reporting (absence of report of studied outcomes). Studies were excluded if one of the cited biases was present, if anesthesia and/or analgesia protocols were not standardized, and if the target population was found to exhibit at least one of the following risks: active bleeding, history of gastroduodenal ulcer, impairment of renal function, or severe asthma. The outcomes analyzed were opioid consumption, pain intensity, PONV, urinary retention, and pruritus. When conflicting results were found, the article was checked twice by 2 anesthesiologists.
Statistical Analysis
Statistical analysis was performed using the Review Manager 5 software (RevMan 5; The Cochrane Collaboration, Oxford, UK). To include a maximum of appropriate studies, nonparametric results were managed by the estimation of the mean and the standard deviation on the basis of the sample size, median, and range according to the method described by Hozo et al.14 The absence of a validated method to convert median and interquartile ranges to means and standard deviation led us to discard all such data. For each outcome from a given study, the Hedge's adjusted standardized mean difference (SMD, continuous outcomes) and the Mantel-Haenszel odds ratio (OR, discrete outcome) were computed. To combine these partial ORs and SMD, all partial ORs were first transformed to partial SMDs using Chinn's formula: lnOR = 1.814 × SMD (ln: logarithm).15 In another analysis, partial SMDs were pooled and overall SMD (and its 95% confidence interval) was generated using the weighted mean (also known as inverse variance method).a A 95% confidence interval of the OR <1 indicates a significant decrease in PONV, urinary retention, or pruritus risk, whereas a 95% confidence interval of the SMD <0 indicates significantly adequate pain control or an opioid dose-sparing effect. Concerning the common cutoff values considered for SMD, the effect of the NSAIDs on the studied outcome was regarded as small when the SMD was more than −0.4, moderate when it was between −0.4 and −0.7, and large when it was less than −0.7. b Heterogeneity was accessed using I2 statistics. This describes the percentage of the variability in effect estimates (OR in present study) that is due to heterogeneity rather than sampling error. According to the Cochrane Reviews guidelines,c an I2 > 40% and a P < 0.1 were considered the definition for heterogeneity and indicated the use of a random effect model in OR computation.3,12 Random effect model assumes that the observed effects are estimating different intervention effects, whereas a fixed effect model estimates a same “true” intervention effect. This difference makes the weight given by each study different for each model. In the random effect model, all studies are equally weighted, whereas in the fixed effect model, each study is weighted according to the number of included patients. In addition, regardless of heterogeneity, subgroup analysis for NSAID efficacy was performed (when at least 2 studies included the considered outcome) according to the type of surgery, the duration of NSAID administration (intraoperative versus postoperative), and the coadministration of paracetamol (at regular timing) for outcomes measured during the first 24 hours postoperatively. The effects of the selectivity of the NSAIDs were also investigated.
In studies with more than one intervention arm, each one was considered a study and compared with the control group. Finally, to avoid calculation problems related to zero values, a 1 was added to all groups.
Statistical methodologies are available to assess the effects of unpublished studies on the results of meta-analysis (published bias). This type of bias is assessed by plotting the OR, or the logarithm of the OR, against a measurement of the precision of the OR such as the SE of the OR. This plot is named the funnel plot. An asymmetry of the funnel plot could indicate that some studies might not have been published because of their negative results.16,17 This asymmetry can also indicate the heterogeneity of results or the poor methodology of included studies.16 Some published studies, because of specific design, can produce large positive results that can lead to asymmetry in the funnel plots interpreted as “publication bias.” Methodological bias of studies can also exhibit large positive results leading to funnel plot asymmetry.18 According to the Cochrane collaborative guidelines,d publication bias had to be assessed when analysis aggregated at least 10 studies. Two tests were used in the current analysis: the Begg-Mazumdar19 and Egger tests.18 When these tests were significant, asymmetry was assumed to be present and publication bias was highly suspected. Results were expressed as OR (95% confidence interval), I2, P value for I2 statistic.
RESULTS
Two hundred ninety-nine articles were identified using the selected criteria. Analysis allowed the selection of 125 relevant articles. A flowchart of the selection process is shown in Figure 1. Publications with no control groups, cohort or observational studies, studies comparing NSAIDs with other therapeutics, and studies with unexploitable results (all outcomes in the study expressed as median and interquartile ranges; absence of SD of the mean) were excluded from the analysis (Fig. 1). There was no difference in recorded information between assessors.
;)
Figure 1: Meta-analysis flowchart. RTC = randomized controlled trial.
This meta-analysis included 27 articles20–47; 567 patients received NSAIDs versus 418 patients who did not receive these drugs. Among these included studies, 25 evaluated nonselective NSAIDs and 2 compared coxibs (rofecoxib) to control inactive treatment.23,39 The methodological features and outcomes measured in the included studies are summarized in Table 1. Characteristics of studies included in this meta-analysis are summarized in Table 2.
Table 1: Relevant Methodological Features of Included Studies
Table 2: Characteristics of Included Studies
NSAIDs significantly decreased opioid consumption in the postanesthesia care unit (PACU) (SMD = −0.66 [−0.84, −0.48]; I2 = 67%, P < 0.00001; Fig. 2), pain intensity during PACU stay (SMD = −0.85 [−1.24, −0.47]; I2 = 90%, P < 0.00001; Fig. 3), and opioid consumption during the first 24 hours postoperatively (SMD = −0.83 [−1.11, −0.55]; I2 = 80%, P < 0.00001; Fig. 4). However, our study could not find a beneficial effect of NSAIDs on pain intensity during the first postoperative day (SMD = −0.32 [−0.75, 0.10]; I2 = 56%, P < 0.00001; Fig. 5). Concerning the efficacy of NSAIDs in preventing opioids' adverse effects, our study could not find a decrease in PONV incidence during PACU stay (OR = 1.02 [0.73–1.44]; I2 = 0%, P = 0.94; Fig. 6). In contrast, according to our results, NSAIDs decreased PONV incidence during the first 24 hours postoperatively (OR = 0.75 [0.57–0.99]; I2 = 0%, P = 0.69; Fig. 7). Finally, our study did not find a preventive effect of NSAIDs against postoperative urinary retention (4 studies; OR = 0.96 [0.25–3.65]; I2 = 49%, P = 0.12) or pruritus (3 studies; OR = 0.96 [0.34–2.71]; I2 = 0%, P = 0.60).
Figure 2: Forest plot of meta-analysis of the effects of perioperative nonsteroidal antiinflammatory drugs (NSAIDs) on postoperative opioid requirement during postanesthesia care unit stay. The square in front of each study (first author and year of publication) is the standardized mean difference (SMD) for individual trials, and the corresponding horizontal line is the 95% confidence interval (CI). The lozenge back in the figure is the pooled SMD with the 95% CI. Studies with >1 intervention group are shown with asterisks (Author, year of publication* and Author, year of publication**).
Figure 3: Forest plot of meta-analysis of the effects of perioperative nonsteroidal antiinflammatory drugs (NSAIDs) on postoperative pain intensity during postanesthesia care unit stay. The square in front of each study (first author and year of publication) is the standardized mean difference (SMD) for individual trials, and the corresponding horizontal line is the 95% confidence interval (CI). The lozenge back in the figure is the pooled SMD with the 95% CI. Studies with >1 intervention group are shown with asterisks (Author, year of publication* and Author, year of publication**).
Figure 4: Forest plot of meta-analysis of the effects of perioperative nonsteroidal antiinflammatory drugs (NSAIDs) on postoperative opioid requirement during the first 24 hours. The square in front of each study (first author and year of publication) is the standardized mean difference (SMD) for individual trials, and the corresponding horizontal line is the 95% confidence interval (CI). The lozenge back in the figure is the pooled SMD with the 95% CI. Studies with >1 intervention group are shown with asterisks (Author, year of publication* and Author, year of publication**).
Figure 5: Forest plot of meta-analysis of the effects of perioperative nonsteroidal antiinflammatory drugs (NSAIDs) on postoperative pain intensity during the first 24 hours. The square in front of each study (first author and year of publication) is the standardized mean difference (SMD) for individual trials, and the corresponding horizontal line is the 95% confidence interval (CI). The lozenge back in the figure is the pooled SMD with the 95% CI. Studies with >1 intervention group are shown with asterisks (Author, year of publication* and Author, year of publication**).
Figure 6: Forest plot of meta-analysis of the effects of perioperative nonsteroidal antiinflammatory drugs (NSAIDs) on postoperative nausea or vomiting during postanesthesia care unit stay. The square in front of each study (first author and year of publication) is the odds ratio (OR) for individual trials, and the corresponding horizontal line is the 95% confidence interval (CI). The lozenge back in the figure is the pooled OR with the 95% CI. Studies with >1 intervention group are shown with asterisks (Author, year of publication* and Author, year of publication**).
Figure 7: Forest plot of meta-analysis of the effects of perioperative nonsteroidal antiinflammatory drugs (NSAIDs) on postoperative nausea or vomiting during the first 24 hours. The square in front of each study (first author and year of publication) is the odds ratio (OR) for individual trials, and the corresponding horizontal line is the 95% confidence interval (CI). The lozenge back in the figure is the pooled OR with the 95% CI. Studies with >1 intervention group are shown with asterisks (Author, year of publication* and Author, year of publication**).
Regarding the high heterogeneity of these results, subgroup analyses were performed. When analyzing the effects of NSAID administration on opioid consumption (in the PACU and during the first postoperative day) or on pain intensity in the PACU according to the type of surgery (adenotonsillectomy versus orthopedic or general surgery), the timing of NSAID administration (intraoperative versus postoperative), and the coadministration of paracetamol, we could not find any significant influence of these factors (Table 3). Subgroup analysis of pain intensity during the first postoperative day according to the type of surgery or the timing of NSAID administration did not modify this result (Table 3). However, when including the 2 studies in which paracetamol was systematically coadministered with NSAIDs, the coadministration of NSAIDs and opioids was found to moderately decrease pain intensity at 24 hours (SMD = −0.35 [−0.50, −0.20]; I2 = 0%, P = 0.35; Table 3). Subgroup analyses of PONV occurrence during the first postoperative day did not affect results, except when performed in studies focusing on adenotonsillectomy (OR = 0.64 [0.44, 0.94]; I2 = 0%, P = 0.76; Table 3).
Table 3: Subgroup Analysis of the Effects of Surgery (Adenotonsillectomy Versus Orthopedic or General Surgery), Timing of Administration (Intraoperative Versus Postoperative), Systematic Postoperative Administration of Paracetamol on the Analgesic Efficacy of NSAIDs
Finally, subgroup analysis found no influence of nonselective NSAIDs on overall results (Table 3), whereas pooling the 2 studies using selective NSAIDs could not find any effect of these compounds on postoperative pain intensity, opioid consumption, or PONV incidence (data not shown).
Analysis found a significant publication bias for the following outcomes: opioid consumption during the first 24 hours (Fig. 8A; Begg-Mazumdar test: Kendall τ= −0.51, P = 0.004; or Egger's test: bias = −1.85, P = 0.04), pain during PACU stay (Fig. 8B; Begg-Mazumdar test: Kendall τ = −0.57, P = 0.0006; or Egger's test: bias = −4.36, P = 0.001). However, bias analysis was not significant for opioid consumption in the PACU (Fig. 8C; Begg-Mazumdar test: Kendall τ = −0.12, P = 0.34; or Egger's test: bias = −1.63, P = 0.09), pain during the 24 hours postoperatively (Fig. 8D; Begg-Mazumdar test: Kendall τ = −0.11, P = 0.58; or Egger's test: bias = −2.79, P = 0.5), PONV during the PACU stay (Fig. 8E; Begg-Mazumdar test: Kendall τ = −0.21, P = 0.35; or Egger's test: bias = −1.36, P = 0.07), or PONV during the first postoperative day (Fig. 8F; Begg-Mazumdar test: Kendall τ = 0.11, P = 0.5; or Egger's test: bias = −0.33, P = 0.7).
Figure 8: Funnel plots of the studies included in the meta-analysis. A, Opioid consumption during the first 24 hours. B, Pain during postanesthesia care unit (PACU) stay. C, Opioid consumption in the PACU. D, Pain during the 24 hours postoperatively. E, Postoperative nausea and vomiting (PONV) during PACU stay. F, PONV during the first postoperative day.
DISCUSSION
The main findings of this meta-analysis can be summarized as follows: coadministering opioids and NSAIDs during the perioperative period in children and infants decreased postoperative opioid requirement (both in the PACU and during the first 24 postoperative hours), pain intensity in the PACU, and PONV during the first postoperative day. However, our study could not find any benefit from NSAID administration on pain intensity during the first 24 postoperative hours, PONV during the PACU stay, urinary retention, or pruritus. Subgroup analyses according to the timing of NSAID administration (intraoperative versus postoperative), type of surgery, or coadministration of paracetamol found that these factors had no influence on the studied outcomes except concerning the reduction of pain intensity and PONV incidence during the first 24 postoperative hours, which were influenced by coadministration of paracetamol and the type of surgery, respectively.
Opioids, especially morphine, are considered the most powerful drugs for relieving moderate to severe postoperative pain. However, a qualitative meta-analysis has shown their lack of superiority, compared with non-opioid drugs or regional analgesia, for pain relief in children.1 Based on these results, postoperative pain in children is best managed using balanced analgesia. The opioid-sparing effect of NSAIDs and their potency in decreasing opioid side effects and improving postoperative pain are still controversial. We reviewed evidence on these issues in a quantitative manner.
Our analysis included studies with various times of NSAID administration, associations with other non-opioid analgesics, and type of surgery, leading to heterogeneous results. Consequently, we performed subgroup analyses according to these 3 factors (summarized in Table 3). We found no impact of these factors on the studied outcomes, except concerning pain intensity and PONV incidence during the first 24 postoperative hours. Pain intensity during the first postoperative day was decreased when paracetamol was coadministered with NSAIDs. This finding suggests, as previously reported, a synergistic action between these 2 drugs on pain relief.4 Reduction of PONV during adenotonsillectomy is also not surprising because PONV incidence is very high during this type of surgery, and therefore a decrease in opioid consumption might be most beneficial on PONV incidence in this subgroup of patients. Interestingly, intraoperative, but not postoperative, NSAID administration decreased PONV incidence during the first 24 postoperative hours. This result is likely to be influenced by the type of surgery rather than by the timing of NSAID administration because most studies involving postoperative NSAID administration were performed in orthopedic or general surgery (Table 3).
Recent experimental studies have found an active age-dependent synaptic plasticity in relation to hyperalgesia after peripheral nerve injury in developing rats.48,49 Moreover, experiences of pain during infancy have been shown to induce a long-lasting sensitization to future pain experiences.50,51 This clearly indicates the high potential of surgery-induced injuries to promote hyperalgesia in children.52 In addition, opioids have been found to induce hyperalgesia in adults and are highly suspected to induce this phenomenon in children and infants.53,54 Hyperalgesia preventive strategies used in adults such as an opioid priming dose or ketamine administration, failed to decrease opioid consumption in the pediatric population.55,56 Consequently, promoting NSAID use for decreasing perioperative opioid consumption might positively influence the development of postoperative hyperalgesia.
A recent meta-analysis showed no efficacy of ketamine's opioid-sparing effect during the first 24 postoperative hours in children.55 The authors hypothesized that this result was the consequence of an inadequate dosage of ketamine. Similarly, a pharmacokinetic study of ketorolac showed differences in drug clearance and elimination half-time, which were greater in the pediatric population compared with adults.57 These findings suggest shortening intervals between drug administration to maintain a systemic analgesic effect. Studies included in our analysis used a drug administration interval of 6 to 8 hours, as validated in adults. However, despite this inadequate administration, NSAIDs remained efficient. Moreover, studies using intraoperative NSAIDs showed an improvement in most studied outcomes, highlighting the strong opioid-sparing effect of these compounds.
Opioids are considered one of the major causes of PONV.1 Our study showed NSAIDs to statistically decrease PONV during the first 24 postoperative hours (OR = 0.75 [0.57–0.99]). However, this effect was probably clinically irrelevant while the high boundary of the 95% confidence interval narrows to 1. In addition, regardless of the considered subgroup, NSAIDs did not decrease PONV during the first postoperative day, except when studies performed during adenotonsillectomy were pooled (Table 3). These findings might reflect an insufficient NSAID opioid-sparing effect observed during the postoperative period and an increased incidence of PONV in children after adenotonsillectomy.58–60 Finally, according to our results, neither urinary retention nor pruritus incidence was influenced by NSAID administration. This also supports the insufficient opioid-sparing effect of these compounds. However, the results in the small number of studies reporting these 2 outcomes should be interpreted with caution.
Our study could not find any efficacy of the selective NSAID rofecoxib on all evaluated outcomes. In addition, removing the 2 studies from analyses did not affect the overall results (Table 3).23,39 Using selective NSAIDs for postoperative pain management in children is rare. Moreover, recent controversies about the involvement of these drugs in ischemic heart disease in adults61,62 have limited their use.
Meta-analysis quality relies on 3 important factors: (1) the quality of selected studies, (2) their heterogeneity, and (3) bias detection. Concerning article quality, all selected studies were randomized and double-blind (authors maintained the quality of their studies by choosing blinded evaluators) and fulfilled all the quality criteria recommended by experts in meta-analysis.13 In addition, both anesthesia protocols and postoperative evaluation were standardized in intervention and control groups. Concerning heterogeneity and bias, selected studies were heterogeneous because of anesthesia and analgesia protocols. Factors such as administration of intraoperative analgesics and timing of administration of the NSAIDs varied among studies. All statistical analyses were performed after assessment of statistical heterogeneity, and subgroup analysis examined study design variations, resulting in decreasing a possible statistical heterogeneity (Table 3). Finally, as recommended by the Cochrane collaborative guidelines for meta-analysis and many other experts in this statistical methodology, when heterogeneity was still found despite subgroup analysis, a random effect model was used.12 All efforts have been made to include the maximal appropriate studies in this meta-analysis (mean and SD estimation and SMD transformations). However, using 2 statistical tools, we found evidence of a significant publication bias in 2 important outcomes: opioid consumption during the first 24 hours, and pain during PACU stay. Although these biases may result from the heterogeneity of the selected studies,16 our findings, especially those of opioid consumption, could suggest that some studies that had found negative results may not have been published.
In conclusion, our study found that the coadministration of NSAIDs and opioids decreased both postoperative opioid requirement and side effects and increased pain relief. Consequently, in the absence of contraindications, the use of NSAIDs must be considered both for managing postoperative pain and for decreasing opioid consumption and side effects in children. However, regarding the possible presence of a published bias, these results must be interpreted with caution. Further studies are still required to define more precisely the use of NSAIDs in this setting.
DISCLOSURES
Name: Daphne Michelet, MD.
Contribution: This author helped design the study, conduct the study, and write the manuscript.
Attestation: Daphne Michelet has seen the original study data, reviewed the analysis of the data, and approved the final manuscript.
Name: Juliette Andreu-Gallien, MD, PhD.
Contribution: This author helped conduct the study, analyze the data, and write the manuscript.
Attestation: Juliette Andreu-Gallien has seen the original study data, reviewed the analysis of the data, and approved the final manuscript.
Name: Tarik Bensalah, MD.
Contribution: This author helped design the study and conduct the study.
Attestation: Tarik Bensalah has seen the original study data, reviewed the analysis of the data, and approved the final manuscript.
Name: Julie Hilly, MD.
Contribution: This author helped design the study, conduct the study, and write the manuscript.
Attestation: Julie Hilly has seen the original study data, reviewed the analysis of the data, and approved the final manuscript.
Name: Chantal Wood, MD.
Contribution: This author helped design the study, conduct the study, and write the manuscript.
Attestation: Chantal Wood has seen the original study data, reviewed the analysis of the data, and approved the final manuscript.
Name: Yves Nivoche, MD, PhD.
Contribution: This author helped design the study, conduct the study, analyze the data, and write the manuscript.
Attestation: Yves Nivoche has seen the original study data, reviewed the analysis of the data, and approved the final manuscript.
Name: Jean Mantz, MD, PhD.
Contribution: This author helped design the study and write the manuscript.
Attestation: Jean Mantz has seen the original study data, reviewed the analysis of the data, and approved the final manuscript.
Name: Souhayl Dahmani, MD, PhD.
Contribution: This author helped design the study, conduct the study, analyze the data, and write the manuscript.
Attestation: Souhayl Dahmani has seen the original study data, reviewed the analysis of the data, approved the final manuscript, and is the author responsible for archiving the study files.
This manuscript was handled by: Peter J. Davis, MD.
a See http://www.cochrane-handbook.org/(Section 9.4.6). Accessed October 6, 2011.
Cited Here
b See http://www.cochrane-handbook.org/(Section 12.6.2). Accessed October 6, 2011.
Cited Here
c See http://www.cochrane-handbook.org/(Section 9.5.2). Accessed October 6, 2011.
Cited Here
d See http://www.cochrane-handbook.org/(Section 10.4.3.1). Accessed December 2010.
Cited Here
REFERENCES
1. Duedahl TH, Hansen EH. A qualitative systematic review of morphine treatment in children with postoperative pain. Paediatr Anaesth 2007;17:756–74
2. Evans MS, Lysakowski C, Tramer MR. Nefopam for the prevention of postoperative pain: quantitative systematic review. Br J Anaesth 2008;101:610–7
3. McDaid C, Maund E, Rice S, Wright K, Jenkins B, Woolacott N. Paracetamol and selective and non-selective non-steroidal anti-inflammatory drugs (NSAIDs) for the reduction of morphine-related side effects after major surgery: a systematic review. Health Technol Assess 2010;14:1–153, iii–iv
4. Maund E, McDaid C, Rice S, Wright K, Jenkins B, Woolacott N. Paracetamol and selective and non-selective non-steroidal anti-inflammatory drugs for the reduction in morphine-related side-effects after major surgery: a systematic review. Br J Anaesth 2011;106:292–7
5. Elia N, Tramer MR. Ketamine and postoperative pain: a quantitative systematic review of randomised trials. Pain 2005;113:61–70
6. Marret E, Kurdi O, Zufferey P, Bonnet F. Effects of nonsteroidal antiinflammatory drugs on patient-controlled analgesia morphine side effects: meta-analysis of randomized controlled trials. Anesthesiology 2005;102:1249–60
7. Standing JF, Savage I, Pritchard D, Waddington M. Diclofenac for acute pain in children. Cochrane Database Syst Rev 2009;4:CD005538
8. Moher D, Cook DJ, Eastwood S, Olkin I, Rennie D, Stroup DF. Improving the quality of reports of meta-analyses of randomised controlled trials: the QUOROM statement. Quality of Reporting of Meta-analyses. Lancet 1999;354:1896–900
9. Hatala R, Keitz S, Wyer P, Guyatt G. Tips for learners of evidence-based medicine. 4. Assessing heterogeneity of primary studies in systematic reviews and whether to combine their results. CMAJ 2005;172:661–5
10. Normand SL. Meta-analysis: formulating, evaluating, combining, and reporting. Stat Med 1999;18:321–59
11. Martell BA, O'Connor PG, Kerns RD, Becker WC, Morales KH, Kosten TR, Fiellin DA. Systematic review: opioid treatment for chronic back pain—prevalence, efficacy, and association with addiction. Ann Intern Med 2007;146:116–27
12. Riley RD, Higgins JP, Deeks JJ. Interpretation of random effects meta-analyses. BMJ 2011;342:d549
13. Lundh A, Gotzsche PC. Recommendations by Cochrane Review Groups for assessment of the risk of bias in studies. BMC Med Res Methodol 2008;8:22
14. Hozo SP, 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
15. Chinn S. A simple method for converting an odds ratio to effect size for use in meta-analysis. Stat Med 2000;19:3127–31
16. Sterne JA, Egger M, Smith GD. Systematic reviews in health care: investigating and dealing with publication and other biases in meta-analysis. BMJ 2001;323:101–5
17. Sterne JA, Egger M. Funnel plots for detecting bias in meta-analysis: guidelines on choice of axis. J Clin Epidemiol 2001;54:1046–55
18. Egger M, Davey Smith G, Schneider M, Minder C. Bias in meta-analysis detected by a simple, graphical test. BMJ 1997;315:629–34
19. Begg CB, Mazumdar M. Operating characteristics of a rank correlation test for publication bias. Biometrics 1994;50: 1088–101
20. Antila H, Manner T, Kuurila K, Salantera S, Kujala R, Aantaa R. Ketoprofen and tramadol for analgesia during early recovery after tonsillectomy in children. Paediatr Anaesth 2006;16: 548–53
21. Bean-Lijewski JD, Hunt RD. Effect of ketorolac on bleeding time and postoperative pain in children: a double-blind, placebo-controlled comparison with meperidine. J Clin Anesth 1996;8:25–30
22. Hiller A, Meretoja OA, Korpela R, Piiparinen S, Taivainen T. The analgesic efficacy of acetaminophen, ketoprofen, or their combination for pediatric surgical patients having soft tissue or orthopedic procedures. Anesth Analg 2006;102:1365–71
23. Joshi W, Connelly NR, Reuben SS, Wolckenhaar M, Thakkar N. An evaluation of the safety and efficacy of administering rofecoxib for postoperative pain management. Anesth Analg 2003;97:35–8
24. Kokki H, Hendolin H, Maunuksela EL, Vainio J, Nuutinen L. Ibuprofen in the treatment of postoperative pain in small children: a randomized double-blind-placebo controlled parallel group study. Acta Anaesthesiol Scand 1994;38:467–72
25. Kokki H, Nikanne E, Tuovinen K. I.v. intraoperative ketoprofen in small children during adenoidectomy: a dose-finding study. Br J Anaesth 1998;81:870–4
26. Kokki H, Homan E, Tuovinen K, Purhonen S. Peroperative treatment with i. v. ketoprofen reduces pain and vomiting in children after strabismus surgery. Acta Anaesthesiol Scand 1999;43:13–8
27. Kokki H, Tuomilehto H, Tuovinen K. Pain management after adenoidectomy with ketoprofen: comparison of rectal and intravenous routes. Br J Anaesth 2000;85:836–40
28. Kokki H, Salonen A. Comparison of pre- and postoperative administration of ketoprofen for analgesia after tonsillectomy in children. Paediatr Anaesth 2002;12:162–7
29. Korpela R, Silvola J, Laakso E, Meretoja OA. Oral naproxen but not oral paracetamol reduces the need for rescue analgesic after adenoidectomy in children. Acta Anaesthesiol Scand 2007;51:726–30
30. Mather SJ, Peutrell JM. Postoperative morphine requirements, nausea and vomiting following anaesthesia for tonsillectomy: comparison of intravenous morphine and non-opioid analgesic techniques. Paediatr Anaesth 1995;5:185–8
31. Maunuksela EL, Olkkola KT, Korpela R. Does prophylactic intravenous infusion of indomethacin improve the management of postoperative pain in children? Can J Anaesth 1988;35:123–7
32. Maunuksela EL, Ryhanen P, Janhunen L. Efficacy of rectal ibuprofen in controlling postoperative pain in children. Can J Anaesth 1992;39:226–30
33. Morton NS, O'Brien K. Analgesic efficacy of paracetamol and diclofenac in children receiving PCA morphine. Br J Anaesth 1999;82:715–7
34. Munro HM, Walton SR, Malviya S, Merkel S, Voepel-Lewis T, Loder RT, Farley FA. Low-dose ketorolac improves analgesia and reduces morphine requirements following posterior spinal fusion in adolescents. Can J Anaesth 2002;49:461–6
35. Nikanne E, Kokki H, Tuovinen K. IV perioperative ketoprofen in small children during adenoidectomy. Br J Anaesth 1997;78:24–7
36. Oztekin S, Hepaguslar H, Kar AA, Ozzeybek D, Artikaslan O, Elar Z. Preemptive diclofenac reduces morphine use after remifentanil-based anaesthesia for tonsillectomy. Paediatr Anaesth 2002;12:694–9
37. Romsing J, Ostergaard D, Walther-Larsen S, Valentin N. Analgesic efficacy and safety of preoperative versus postoperative ketorolac in paediatric tonsillectomy. Acta Anaesthesiol Scand 1998;42:770–5
38. Rugyte D, Kokki H. Intravenous ketoprofen as an adjunct to patient-controlled analgesia morphine in adolescents with thoracic surgery: a placebo controlled double-blinded study. Eur J Pain 2007;11:694–9
39. Sheeran PW, Rose JB, Fazi LM, Chiavacci R, McCormick L. Rofecoxib administration to paediatric patients undergoing adenotonsillectomy. Paediatr Anaesth 2004;14:579–83
40. Sims C, Johnson CM, Bergesio R, Delfos SJ, Avraamides EA. Rectal indomethacin for analgesia after appendicectomy in children. Anaesth Intensive Care 1994;22:272–5
41. Sutters KA, Levine JD, Dibble S, Savedra M, Miaskowski C. Analgesic efficacy and safety of single-dose intramuscular ketorolac for postoperative pain management in children following tonsillectomy. Pain 1995;61:145–53
42. Sutters KA, Shaw BA, Gerardi JA, Hebert D. Comparison of morphine patient-controlled analgesia with and without ketorolac for postoperative analgesia in pediatric orthopedic surgery. Am J Orthop (Belle Mead NJ) 1999;28:351–8
43. Tuomilehto H, Kokki H, Tuovinen K. Comparison of intravenous and oral ketoprofen for postoperative pain after adenoidectomy in children. Br J Anaesth 2000;85:224–7
44. Tuomilehto H, Kokki H. Parenteral ketoprofen for pain management after adenoidectomy: comparison of intravenous and intramuscular routes of administration. Acta Anaesthesiol Scand 2002;46:184–9
45. Vetter TR, Heiner EJ. Intravenous ketorolac as an adjuvant to pediatric patient-controlled analgesia with morphine. J Clin Anesth 1994;6:110–3
46. Viitanen H, Tuominen N, Vaaraniemi H, Nikanne E, Annila P. Analgesic efficacy of rectal acetaminophen and ibuprofen alone or in combination for paediatric day-case adenoidectomy. Br J Anaesth 2003;91:363–7
47. Watters CH, Patterson CC, Mathews HM, Campbell W. Diclofenac sodium for post-tonsillectomy pain in children. Anaesthesia 1988;43:641–3
48. Baccei ML. Modulation of developing dorsal horn synapses by tissue injury. Ann N Y Acad Sci 2010;1198:159–67
49. Lebel A, Becerra L, Wallin D, Moulton EA, Morris S, Pendse G, Jasciewicz J, Stein M, Aiello-Lammens M, Grant E, Berde C, Borsook D. fMRI reveals distinct CNS processing during symptomatic and recovered complex regional pain syndrome in children. Brain 2008;131:1854–79
50. Peters JW, Schouw R, Anand KJ, van Dijk M, Duivenvoorden HJ, Tibboel D. Does neonatal surgery lead to increased pain sensitivity in later childhood? Pain 2005;114:444–54
51. Wollgarten-Hadamek I, Hohmeister J, Demirakca S, Zohsel K, Flor H, Hermann C. Do burn injuries during infancy affect pain and sensory sensitivity in later childhood? Pain 2009;141: 165–72
52. Dufton LM, Konik B, Colletti R, Stanger C, Boyer M, Morrow S, Compas BE. Effects of stress on pain threshold and tolerance in children with recurrent abdominal pain. Pain 2008;136: 38–43
53. Heger S, Maier C, Otter K, Helwig U, Suttorp M. Morphine induced allodynia in a child with brain tumour. BMJ 1999;319:627–9
54. Sammartino M, Garra R, Sbaraglia F, De Riso M, Continolo N. Remifentanil in children. Paediatr Anaesth 2010;20:246–55
55. Dahmani S, Michelet D, Abback PS, Wood C, Brasher C, Nivoche Y, Mantz J. Ketamine for perioperative pain management in children: a meta-analysis of published studies. Paediatr Anaesth 2011;21:636–52
56. McDonnell C, Zaarour C, Hull R, Thalayasingam P, Pehora C, Ahier J, Crawford MW. Pre-treatment with morphine does not prevent the development of remifentanil-induced hyperalgesia. Can J Anaesth 2008;55:813–8
57. Lynn AM, Bradford H, Kantor ED, Seng KY, Salinger DH, Chen J, Ellenbogen RG, Vicini P, Anderson GD. Postoperative ketorolac tromethamine use in infants aged 6–18 months: the effect on morphine usage, safety assessment, and stereo-specific pharmacokinetics. Anesth Analg 2007;104:1040–51
58. Gross D, Reuss S, Dillier CM, Gerber AC, Weiss M. Early vs late intraoperative administration of tropisetron for the prevention of nausea and vomiting in children undergoing tonsillectomy and/or adenoidectomy. Paediatr Anaesth 2006;16:444–50
59. Kaan MN, Odabasi O, Gezer E, Daldal A. The effect of preoperative dexamethasone on early oral intake, vomiting and pain after tonsillectomy. Int J Pediatr Otorhinolaryngol 2006;70:73–9
60. Liechti M, Feurer R, Gross D, Schmitz A, Stutz K, Gerber A, Weiss M. Prevention of postoperative nausea and vomiting in children following adenotonsillectomy, using tropisetron with or without low-dose dexamethasone. J Anesth 2007;21:311–6
61. Moodley I. Review of the cardiovascular safety of COXIBs compared to NSAIDS. Cardiovasc J Afr 2008;19:102–7
62. Moore RA, Derry S, McQuay HJ. Cyclo-oxygenase-2 selective inhibitors and nonsteroidal anti-inflammatory drugs: balancing gastrointestinal and cardiovascular risk. BMC Musculoskelet Disord 2007;8:73
