Pharmacological interventions for pain relief during orthodontic treatment

Aoife B Monk; Jayne E Harrison; Helen V Worthington; Annabel Teague

doi:10.1002/14651858.CD003976.pub2

. 2017 Nov 28;2017(11):CD003976. doi: 10.1002/14651858.CD003976.pub2

Pharmacological interventions for pain relief during orthodontic treatment

Aoife B Monk ¹, Jayne E Harrison ^1,^✉, Helen V Worthington ², Annabel Teague ³

Editor: Cochrane Oral Health Group

PMCID: PMC6486038 PMID: 29182798

Abstract

Background

Pain is a common side effect of orthodontic treatment. It increases in proportion to the amount of force applied to the teeth, and the type of orthodontic appliance used can affect the intensity of the pain. Pain during orthodontic treatment has been shown to be the most common reason for people wanting to discontinue treatment, and has been ranked as the worst aspect of treatment. Although pharmacological methods of pain relief have been investigated, there remains some uncertainty among orthodontists about which painkillers are most suitable and whether pre‐emptive analgesia is beneficial. We conducted this Cochrane Review to assess and summarize the international evidence relating to the effectiveness of analgesics for preventing this unwanted side effect associated with orthodontic treatment.

Objectives

The objectives of this review are to determine:

‐ the effectiveness of drug interventions for pain relief during orthodontic treatment; and  ‐ whether there is a difference in the analgesic effect provided by different types, forms and doses of analgesia taken during orthodontic treatment.

Search methods

Cochrane Oral Health’s Information Specialist searched the following databases: the Cochrane Oral Health Trials Register (to 19 June 2017), the Cochrane Central Register of Controlled Trials (CENTRAL;the Cochrane Library 2016, Issue 7), MEDLINE Ovid (1946 to 19 June 2017), Embase Ovid (1980 to 19 June 2017) and CINAHL EBSCO (Cumulative Index to Nursing and Allied Health Literature; 1937 to 19 June 2017). The US National Institutes of Health Ongoing Trials Register (ClinicalTrials.gov) and the World Health Organization International Clinical Trials Registry Platform were searched on the 19 June 2017 for ongoing studies. We placed no restrictions on language or date of publication when searching the electronic databases.

Selection criteria

We included randomized controlled trials (RCTs) relating to pain control during orthodontic treatment. Pain could be measured on a visual analogue scale (VAS), numerical rating scale (NRS) or categorical scale.

Data collection and analysis

Two review authors independently screened the search results, agreed the studies to be included and extracted information from the included studies regarding methods, participants, interventions, outcomes, harms and results. We planned to resolve any discrepancies or disagreements through discussion. We used the Cochrane 'Risk of bias' tool to assess the risk of bias in the studies.

Main results

We identified 32 relevant RCTs, which included 3110 participants aged 9 to 34 years, 2348 of whom we were able to include in our analyses. Seventeen of the studies had more than two arms. We were able to use data from 12 trials in meta‐analyses that compared analgesics versus control (no treatment or a placebo); nine that compared non‐steroidal anti‐inflammatories (NSAIDs) versus paracetamol; and two that compared pre‐emptive versus post‐treatment ibuprofen for pain control following orthodontic treatment. One study provided data for the comparison of NSAIDs versus local anaesthetic.

We found moderate‐quality evidence that analgesics effectively reduced pain following orthodontic treatment when compared to no treatment or a placebo at 2 hours (mean difference (MD) ‐11.66 mm on a 0 to 100 mm VAS, 95% confidence interval (CI) ‐16.15 to ‐7.17; 10 studies, 685 participants), 6 hours (MD ‐24.27 mm on a VAS, 95% CI ‐31.44 to ‐17.11; 9 studies, 535 participants) and 24 hours (MD ‐21.19 mm on a VAS, 95% CI ‐28.31 to ‐14.06; 12 studies, 1012 participants).

We did not find any evidence of a difference in efficacy between NSAID and paracetamol at 2, 6 or 24 hours (at 24 hours: MD ‐0.51, 95% CI ‐8.93 to 7.92; 9 studies, 734 participants; low‐quality evidence).

Very low‐quality evidence suggested pre‐emptive ibuprofen gave better pain relief at 2 hours than ibuprofen taken post treatment (MD ‐11.30, 95% CI ‐16.27 to ‐6.33; one study, 41 participants), however, the difference was no longer significant at 6 or 24 hours.

A single study of 48 participants compared topical NSAIDs versus local anaesthetic and showed no evidence of a difference in the effectiveness of the interventions (very low‐quality evidence).

Use of rescue analgesia was poorly reported. The very low‐quality evidence did not show evidence of a difference between participants taking ibuprofen and participants taking paracetamol (relative risk (RR) 1.5, 95% CI 0.6 to 3.6). Nor did we find evidence of a difference between groups in likelihood of requiring rescue analgesia when ibuprofen was taken pre‐emptively compared to after treatment (RR 0.8, 95% CI 0.3 to 1.9).

Adverse effects were identified in one study, with one participant developing a rash that required treatment with antihistamines. This was provisionally diagnosed as a hypersensitivity to paracetamol.

Authors' conclusions

Analgesics are more effective at reducing pain following orthodontic treatment than placebo or no treatment. Low‐quality evidence did not show a difference in effectiveness between systemic NSAIDs compared with paracetamol, or topical NSAIDs compared with local anaesthetic. More high‐quality research is needed to investigate these comparisons, and to evaluate pre‐emptive versus post‐treatment administration of analgesics.

Plain language summary

Painkillers for relieving pain caused by orthodontic treatment

Review question

Do painkillers, taken before or after orthodontic treatment, help relieve pain? If so, which painkillers work best?

Background

Pain is a common side effect of orthodontic treatment. The pain resulting from orthodontic treatment may differ depending on the amount of force applied and the type of braces used. It may also change over the first few days following treatment. Pain has been ranked as the worst aspect of treatment and is the most common reason for people wanting to discontinue orthodontic treatment. Painkillers, swallowed or applied directly to the sore areas of the mouth following treatment, are thought to relieve the pain, making brace treatment more comfortable and acceptable. These painkillers are often cheap, readily available, easy to use and do not cause serious side effects.

Study characteristics

Authors working with Cochrane Oral Health carried out a review of existing studies and the evidence is current up to 19 June 2017. This review includes 32 studies published from 1993 to 2016 in which 3110 participants aged 9 to 34 years (2348 of whom were included in the analyses) were randomly allocated to groups to receive:

1) painkillers versus no treatment,

2) painkillers versus a placebo (pretend or 'dummy' medicine),

3) one painkiller versus a different painkiller, or

4) a painkiller taken at different time intervals.

The severity of pain experienced by the study participants was compared. Nearly all the evidence was from adults who received oral painkillers versus no treatment, or one oral painkiller versus another oral painkiller. This evidence fell into two main groups:

1) adults receiving paracetamol; or

2) adults receiving non‐steroidal anti‐inflammatory drugs (NSAIDs).

A small amount of evidence also investigated the effect of local anaesthetic and opioids (tramadol).

Key results

Analgesic versus placebo or no treatment

We found evidence that paracetamol, NSAIDs and local anaesthetic were effective at reducing pain intensity at 2 hours, 6 hours and 24 hours following orthodontic treatment when compared with either a placebo or no treatment group.

NSAID versus paracetamol

We found no clear evidence of a difference between the effect of ibuprofen and paracetamol for reducing pain intensity at 2 hours, 6 hours or 24 hours following either the placement of separators (between teeth) or placement of an initial aligning archwire.

Pre‐emptive NSAID versus post‐treatment NSAID

We found some very low‐quality evidence that ibuprofen taken 1 hour prior to separator placement significantly reduces pain intensity 2 hours afterwards when compared to ibuprofen taken post‐treatment. However, at 6 hours and 24 hours, we detected no clear difference.

NSAID versus local anaesthetic

There was no evidence of a difference between the interventions.

Quality of the evidence

The evidence available for the main outcome of pain relief is of moderate to low quality, whilst the quality of the rest of the evidence was very low. We judged only one study to be at low risk of bias.

Summary of findings

Summary of findings for the main comparison. Analgesic compared to control (placebo or no treatment) for pain relief during orthodontic treatment.

Analgesic compared to control (placebo or no treatment) for pain relief during orthodontic treatment
Population: people undergoing orthodontic treatment Setting: any dental setting Intervention: analgesic  Comparison: control (placebo or no treatment)
Outcomes	*Anticipated absolute effects^ (95% CI)**		Relative effect  (95% CI)	Number of participants  (studies)	Quality of the evidence  (GRADE)	Comments
	Risk with control	Risk with analgesic
Pain assessed with VAS (0 to 100 mm where 0 = no pain)  Follow‐up: mean 2 hours	Mean pain ranged from 10 mm to 73 mm on a VAS	MD 11.66 mm lower  (16.15 mm to 7.17 mm lower)	‐	685  (10 RCTs)	⊕⊕⊕⊝ moderate¹	Effect estimates for 6 and 24 hours were larger. 6 hours ‐24.27 mm (95% CI ‐31.44 to ‐17.11) 24 hours ‐21.19 mm (95% CI ‐28.31 to ‐14.06)
Rescue analgesia	Not reported
Adverse events	Not reported
*The risk in the intervention group (and its 95% confidence interval) is based on the assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI)  CI: confidence interval; MD: mean difference; VAS: visual analogue scale
GRADE Working Group grades of evidence  High quality: we are very confident that the true effect lies close to that of the estimate of the effect  Moderate quality: we are moderately confident in the effect estimate: the true effect is likely to be close to the estimate of the effect, but there is a possibility that it is substantially different  Low quality: our confidence in the effect estimate is limited: the true effect may be substantially different from the estimate of the effect  Very low quality: we have very little confidence in the effect estimate: the true effect is likely to be substantially different from the estimate of effect

NSAID compared to paracetamol for pain relief during orthodontic treatment
Population: people undergoing orthodontic treatment  Setting: any dental setting  Intervention: NSAID  Comparison: paracetamol
Outcomes	*Anticipated absolute effects^ (95% CI)**		Relative effect  (95% CI)	Number of participants  (studies)	Quality of the evidence  (GRADE)	Comments
	Risk with paracetamol	Risk with ibuprofen
Pain  assessed with VAS  (0 to 100 mm where 0 = no pain)  Follow‐up: mean 2 hours	Mean pain ranged from 8.8 mm to 32.4 mm on a VAS	MD 2.92 mm lower on the VAS  (8.48 mm lower to 2.65 mm higher)	‐	664  (7 RCTs)	⊕⊕⊝⊝  low^{1, 2}	Findings at 6 and 24 hours were similar and not statistically significant
Rescue analgesia	91 per 1000	136 per 1000  (55 to 327)	RR 1.5  (0.6 to 3.6)	159  (1 RCT)	⊕⊝⊝⊝  very low³	Study does not report any additional information regarding class, dose or timing of rescue analgesics taken
Adverse events	In one study, one participant (< 1%) in the paracetamol group experienced drug hypersensitivity involving a "red, blotchy and itchy rash", which resolved without treatment within 1 week.		‐	(1 RCT)	⊕⊝⊝⊝  very low³
*The risk in the intervention group (and its 95% confidence interval) is based on the assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI).  CI: confidence interval; MD: mean difference; NSAID: non‐steroidal anti‐inflammatory; RR: risk ratio; VAS: visual analogue scale
GRADE Working Group grades of evidence  High quality: we are very confident that the true effect lies close to that of the estimate of the effect  Moderate quality: we are moderately confident in the effect estimate: the true effect is likely to be close to the estimate of the effect, but there is a possibility that it is substantially different  Low quality: our confidence in the effect estimate is limited: the true effect may be substantially different from the estimate of the effect  Very low quality: we have very little confidence in the effect estimate: the true effect is likely to be substantially different from the estimate of effect

Pre‐emptive ibuprofen compared to ibuprofen post‐treatment for pain relief during orthodontic treatment
Population: pain relief during orthodontic treatment  Setting: dental hospital Intervention: ibuprofen pre‐emptively  Comparison: ibuprofen post‐treatment
Outcomes	*Anticipated absolute effects^ (95% CI)**		Relative effect  (95% CI)	Number of participants  (studies)	Quality of the evidence  (GRADE)	Comments
	Risk with ibuprofen post‐treatment	Risk with ibuprofen pre‐emptively
Pain  assessed with VAS (0 to 100 mm where 0 = no pain)  Follow‐up: mean 2 hours	Mean pain ranged from 19.1 mm to 20.5 mm	MD 11.30 mm lower  (16.27 mm lower to 6.33 mm lower)	‐	41  (1 RCT)	⊕⊝⊝⊝  very low^{1, 2, 3}	Effect no longer seen at 6 or 24 hours (2 studies with 69 participants)
Rescue analgesia	316 per 1000	253 per 1000  (95 to 600)	RR 0.8  (0.3 to 1.9)	41  (1 RCT)	‐
Adverse events	Not reported
*The risk in the intervention group (and its 95% confidence interval) is based on the assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI).  CI: confidence interval; MD: mean difference; RR: risk ratio; VAS: visual analogue scale
GRADE Working Group grades of evidence  High quality: we are very confident that the true effect lies close to that of the estimate of the effect  Moderate quality: we are moderately confident in the effect estimate: the true effect is likely to be close to the estimate of the effect, but there is a possibility that it is substantially different  Low quality: our confidence in the effect estimate is limited: the true effect may be substantially different from the estimate of the effect  Very low quality: we have very little confidence in the effect estimate: the true effect is likely to be substantially different from the estimate of effect

NSAID compared to local anaesthetic for pain relief during orthodontic treatment
Patient or population: pain relief during orthodontic treatment  Setting: any dental setting  Intervention: NSAID  Comparison: local anaesthetic
Outcomes	*Anticipated absolute effects^ (95% CI)**		Relative effect  (95% CI)	Number of participants  (studies)	Quality of the evidence  (GRADE)	Comments
	Risk with local anaesthetic	Risk with NSAID
Pain assessed with VAS  (0 to 100 mm where 0 = no pain)  Follow‐up: mean 2 hours	Mean pain was 51 mm	MD 13 mm higher  (3.45 mm lower to 29.45 mm higher)	‐	48  (1 RCT)	⊕⊝⊝⊝  very low^{1, 2}	Nor was a difference between groups seen at 6 or 24 hours
Rescue analgesia	Not measured
Adverse events	Not measured
*The risk in the intervention group (and its 95% confidence interval) is based on the assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI).  CI: confidence interval; MD: mean difference; NSAID: non‐steroidal anti‐inflammatory; VAS: visual analogue scale
GRADE Working Group grades of evidence  High quality: we are very confident that the true effect lies close to that of the estimate of the effect  Moderate quality: we are moderately confident in the effect estimate: the true effect is likely to be close to the estimate of the effect, but there is a possibility that it is substantially different  Low quality: our confidence in the effect estimate is limited: the true effect may be substantially different from the estimate of the effect  Very low quality: we have very little confidence in the effect estimate: the true effect is likely to be substantially different from the estimate of effect

Other pain outcome data
Study
Abtahi 2006	Data from the trial is incomplete and inappropriately presented and cannot be used for pooled analysis. "Ibuprofen and Tramadol groups have experienced less pain than Placebo group and the difference between two medication groups was insignificant. In placebo group pain was higher in chewing than biting."
Arantes 2009	“Pain intensity in these groups (tenoxicam before and after orthodontic activation) was lower than in the placebo group. The difference in pain intensity between the active and the control was greatest at the assessment made 12 h after activation.”
Bayani 2016	"ANOVA demonstrated significant between‐group differences in pain perceived at chewing, biting, fitting front teeth and fitting back teeth at all time points (p<0.001). The results of the Tukey test revealed that at almost all intervals, pain intensity in the placebo group was comparable to that of the LLRL group (p>0.05), and both groups experienced significantly greater pain than subjects in the bite wafer, ibuprofen and LLIL." We can infer that participants in the ibuprofen group experienced less pain than those in the placebo group.
Eslamian 2014	Data from cross‐over trial inappropriately presented (and from authors) cannot be used for pooled analysis. "The mean pain score recorded in benzocaine group was lower than that in the ketoprofen and placebo groups but a significant difference was only observed between benzocaine and the ketoprofen groups during the first two hours using Friedman and Wilcoxon tests (P=0.042). Compared to the control group, both ketoprofen and benzocaine chewing gums significantly decreased pain at all time points except for day 7."  Ketoprofen and benzocaine gums were both significantly effective for orthodontic pain reduction.
Eslamian 2016a	Data from cross‐over trial inappropriately presented and cannot be used for pooled analysis. "The recorded pain scores were subjected to non‐parametric analysis"; "The highest pain was recorded at 2 and 6 hours. Pain scores were significantly different between the three groups (Kruskal–Wallis test, p < 0.01). The overall mean (SD) pain scores for the benzocaine 5%, ketoprofen, and control (placebo) groups were 0.89 (0.41), 0.68 (0.34), and 1.15 (0.81), respectively. The pain scores were significantly different between the ketoprofen and control groups (mean difference = 0.47, p = 0.005)." A significant pain reduction was observed following the use of ketoprofen when tested against a control gel (placebo).
Eslamian 2016b	Data from cross‐over trial inappropriately presented and cannot be used for pooled analysis. "The overall mean value of pain intensity for benzocaine and placebo gels was 0.89 and 1.15, respectively. The Mann–Whitney U‐test indicated that there was no significant difference between overall pain in both groups (mean difference = 0.258 p < 0.21)."  Benzocaine gel caused a decrease in pain perception at 2 h compared with placebo gel. Peak pain intensity was at 2 h for placebo gel and at 6 h for benzocaine gel, followed by a decline in pain perception from that point to day 7 for both gels. The P‐values comparing pain intensity between the groups at each time point were all non‐significant. Probably no differences found.
Ngan 1994	"A repeated measures analysis of variance and post hoc studentized range statistics showed that the placebo group had significantly more discomfort than either the ibuprofen or the aspirin group at all the time intervals tested."
Patel 2011	Although there was a sophisticated analysis undertaken we can only use the fixed‐effect estimates for pain summary score (over 2, 6, bedtime, awaking 24 hours) presented in Table 1 in the paper. These estimates indicated that ibuprofen reduced pain when compared with a placebo with an effect size of ‐7.89 (95% CI ‐13.18 to ‐2.60; P = 0.003), on 0 to 100 mm VAS. There was some evidence of a difference for naproxen sodium (effect size ‐5.04, 95% CI ‐9.94 to ‐0.14; P = 0.04) when compared with placebo, but not for acetaminophen (effect size 3.67, 95% CI ‐8.57 to 1.23; P = 0.14).
Sudhakar 2014	"Overall results showed Group 3 (Aspirin 300 mg) patients experienced very less pain in terms of mild discomfort, closely followed by Group 2 (Ibuprofen 400 mg). Group 1 (Paracetamol 650 mg) patients experienced mild to moderate pain on bed time and next day morning, after which gradually reduced to no pain from 3rd day morning. However, Group 4 (Placebo) patients had a bitter experience of moderate to severe pain at all‑time intervals."
Young 2006	Unable to draw conclusions due to poor reporting of data.

Experimental intervention	Outcome	Number of studies (number of participants)	Effect measure RR, MD, SMD (95% CI)	P‐value for effect	P‐value heterogeneity	I² (%)
2 hours
Paracetamol	Pain VAS	2 (79)	MD ‐11.51 (‐19.15 to ‐3.86)	0.003	0.77	0%
6 hours
Paracetamol	Pain VAS	2 (79)	MD ‐16.00 (‐24.65 to ‐7.34)	0.0003	0.38	0%
24 hours
Paracetamol	Pain VAS	2 (79)	MD ‐21.51 (‐54.10 to 11.09)	0.2	0.04	76%

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
1 2 hours	10	685	Mean Difference (IV, Random, 95% CI)	‐11.66 [‐16.15, ‐7.17]
1.1 Paracetamol versus control	4	107	Mean Difference (IV, Random, 95% CI)	‐11.90 [‐18.36, ‐5.44]
1.2 NSAID versus control	10	578	Mean Difference (IV, Random, 95% CI)	‐11.72 [‐16.93, ‐6.51]
2 6 hours	9	535	Mean Difference (IV, Random, 95% CI)	‐24.27 [‐31.44, ‐17.11]
2.1 Paracetamol versus control	4	107	Mean Difference (IV, Random, 95% CI)	‐19.34 [‐24.80, ‐13.88]
2.2 NSAID versus control	9	428	Mean Difference (IV, Random, 95% CI)	‐25.69 [‐34.47, ‐16.92]
3 24 hours	12	1012	Mean Difference (IV, Random, 95% CI)	‐21.19 [‐28.31, ‐14.06]
3.1 Paracetamol versus control	6	161	Mean Difference (IV, Random, 95% CI)	‐22.09 [‐35.99, ‐8.18]
3.2 NSAID versus control	12	851	Mean Difference (IV, Random, 95% CI)	‐21.05 [‐29.44, ‐12.65]
4 Other pain outcome data			Other data	No numeric data

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
1 2 hours	7	664	Mean Difference (IV, Random, 95% CI)	‐2.92 [‐8.48, 2.65]
2 6 hours	7	664	Mean Difference (IV, Random, 95% CI)	‐5.17 [‐11.71, 1.37]
3 24 hours	9	734	Mean Difference (IV, Random, 95% CI)	‐0.51 [‐8.93, 7.92]
4 Qualitative pain			Other data	No numeric data

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
1 2 hours	1	41	Mean Difference (IV, Random, 95% CI)	‐11.3 [‐16.27, ‐6.33]
2 6 hours	2	69	Mean Difference (IV, Random, 95% CI)	‐8.43 [‐30.37, 13.50]
3 24 hours	2	69	Mean Difference (IV, Random, 95% CI)	‐9.74 [‐47.88, 28.40]

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
1 2 hours	1	48	Mean Difference (IV, Random, 95% CI)	13.0 [‐3.45, 29.45]
2 6 hours	1	48	Mean Difference (IV, Random, 95% CI)	7.0 [‐7.16, 21.16]
3 24 hours	1	48	Mean Difference (IV, Random, 95% CI)	7.0 [‐5.19, 19.19]

Methods	Setting: orthodontics private offices (in Hamedan, Iran) Design: parallel (3 arms) Number of centres: 1 Study duration: not reported
Participants	Inclusion criteria: people who were not experiencing any kind of pain before treatment, or using any kind of analgesic; and who did not have any kidney or liver disease or other contraindication for using the medications being tested in the study  Exclusion criteria: people who had not completed the questionnaires in different times [sic]; people who consumed other analgesics during the study Orthodontic intervention: separator placement Participant sampling: n = 60 recruited and randomized n = 0 lost to follow‐up n = 60 data analyzed (18 male:42, female, mean age 15.43 years ± 3.7) Group 1 (n = 20): 10 male:3 female, mean age 15.75 ± 3.95 years Group 2 (n = 20): 4 male:10 female, mean age 15.05 ± 3.49 years Group 3 (n = 20): 6 male:8 female, mean age 15.5 ± 3.98 years
Interventions	Comparison: NSAID vs placebo, opioid vs placebo, NSAID vs opioid Ibuprofen (800 mg) vs tramadol (100 mg) vs placebo (500 mg starch) provided pre‐emptively and postoperatively following insertion of separators Group 1: 400 mg ibuprofen 1 h before and 5 h after insertion of separators Group 2: 50 mg tramadol 1 h before and 5 h after insertion of separators Group 3: placebo 1 h before and 5 h after insertion of separators
Outcomes	Pain score (VAS) recorded at 2 h, 6 h, 24 h, 2 days, 3 days and 7 days after separator placement
Notes	Conflict of interests/funding: no source of funding or conflict of interest reported Adverse events/harm: participants in the tramadol group reported dizziness, drowsiness, nausea and headache. Data handling by review authors: study reported Group 2 as placebo arm and Group 3 as tramadol; for the purposes of the review, Groups 2 and 3 have been inverted to present tramadol as an intervention and placebo as its control Other information: the paper has been translated from Farsi
*Risk of bias*
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Quote: "The patients who met the inclusion criteria randomly allocated in one of the three groups and received the medication". Comment: inadequate information regarding method of randomisation therefore unable to make a judgement on appropriateness
Allocation concealment (selection bias)	Unclear risk	Comment: inadequate information regarding allocation concealment was carried out therefore unable to make a judgement on appropriateness
Blinding of participants and personnel (performance bias)   All outcomes	Low risk	Participants were not aware of the medication to which they or others had been allocated, as they received similar sealed pockets with A, B or C stickers on them Comment: blinding appears to have been adequate
Blinding of outcome assessment (detection bias)   All outcomes	Unclear risk	Comment: inadequate information regarding method of blinding of assessors, therefore unable to make a judgement on appropriateness
Incomplete outcome data (attrition bias)   All outcomes	Low risk	0/60 dropouts = 0% attrition (100% completed)
Selective reporting (reporting bias)	High risk	The data for outcomes in this study were not reported appropriately. We were unable to extract data for use in pooled analysis.
Other bias	High risk	There was large variation in the gender balance of the groups at baseline, which could indicate selection bias.

Experimental intervention	Outcome	Number of studies (number of participants)	Effect measure RR, MD, SMD (95% CI)	P‐value for effect	P‐value heterogeneity	I² (%)
2 hours
Ibuprofen	Pain VAS	4 (152)	MD ‐12.80 (‐17.59 to ‐8.01)	< 0.00001	0.91	0%
Lumiracoxib		1 (51)	MD ‐5.32 (‐112.20 to 1.56)	0.13	0.73	0%
Piroxicam		1 (45)	MD ‐29.70 (‐48.01 to ‐11.39)	0.001	N/A	N/A
Subtotal		5 (248)	‐11.10 (‐15.78 to‐6.42)	< 0.00001	0.25	23%
6 hours
Ibuprofen	Pain VAS	4 (150)	MD ‐23.41 (‐41.43 to ‐5.39)	0.01	0.0001	86%
Lumiracoxib		1 (51)	MD ‐7.06 (‐16.55 to 2.43)	0.14	0.67	0%
Piroxicam		1 (45)	MD ‐40.50 (‐57.37 to ‐23.63)	< 0.00001	N/A	N/A
Subtotal		5 (248)	MD ‐8.80 (‐13.47 to ‐4.12)	0.0002	0.97	0%
24 hours
Ibuprofen		3 (118)	MD ‐21.85 (‐37.33 to ‐6.37)	0.006	0.07	62%
Lumiracoxib		1 (51)	MD ‐17.58 (‐31.62 to ‐3.54)	0.01	0.57	0%
Piroxicam		1 (45)	MD ‐36.10 (‐51.33 to ‐20.87)	< 0.00001	N/A	N/A
Subtotal		4 (214)	MD ‐23.66 (‐32.81 to‐14.51)	< 0.00001	0.12	42%

Experimental intervention	Outcome	Number of studies (number of participants)	Effect measure RR, MD, SMD (95% CI)	P‐value for effect	P‐value heterogeneity	I² (%)
2 hours
Aspirin	Pain VAS	1 (25)	MD ‐26.20 (‐56.37 to 3.97)	0.09	N/A	N/A
Etoricoxib		1 (30)	MD ‐32.00 (‐38.09 to ‐25.91)	< 0.00001	N/A	N/A
Flurbiprofen		1 (25)	MD ‐21.00 (‐51.34 to 9.34)	0.17	N/A	N/A
Ibuprofen		3 (75)	MD ‐6.02 (‐20.52 to 8.47)	0.42	0.45	0%
Naproxen sodium		2 (55)	MD ‐20.80 (‐39.42 to ‐2.18)	0.03	0.55	0%
Subtotal		4 (210)	MD ‐19.23 (‐29.90 to‐8.56)	0.0004	0.07	47%
6 hours
Aspirin	Pain VAS	1 (25)	MD ‐39.30 (‐69.71 to ‐8.89)	0.01	N/A	N/A
Etoricoxib		1 (30)	MD ‐41.00 (‐46.77 to ‐35.23)	< 0.00001	N/A	N/A
Flurbiprofen		1 (25)	MD ‐22.00 (‐52.86 to 8.86)	0.16	N/A	N/A
Ibuprofen		3 (75)	MD ‐20.87 (‐35.21 to ‐6.52)	0.004	0.91	0%
Naproxen sodium		2 (55)	MD ‐36.87 (‐54.82 to ‐18.92)	< 0.0001	0.84	0%
Subtotal		4 (210)	MD ‐35.42 (‐42.15 to‐28.68)	< 0.00001	0.35	10%
24 hours
Aspirin	Pain VAS	1 (25)	MD ‐41.20 (‐71.99 to ‐10.41)	0.09	N/A	N/A
Etoricoxib		1 (30)	MD ‐54.00 (‐60.34 to ‐47.66)	< 0.00001	N/A	N/A
Flurbiprofen		1 (25)	MD ‐10.40 (‐41.69 to 20.89)	0.51	N/A	N/A
Ibuprofen		6 (420)	MD ‐19.71 (‐33.60 to ‐5.82)	0.005	0.004	71%
Naproxen sodium		2 (55)	MD ‐28.61 (‐66.08 to 8.86)	0.13	0.05	75%
Subtotal		7 (555)	MD ‐26.26 (‐37.81 to‐14.71)	< 0.0001	< 0.00001	85%

Experimental intervention	Outcome	Number of studies (number of participants)	Effect measure RR, MD, SMD (95% CI)	P‐value for effect	P‐value heterogeneity	I² (%)
2 hours
Ibuprofen	Pain VAS	4 (351)	MD ‐3.36 (‐7.00 to 0.28)	0.07	0.46	0%
Meloxicam		1 (127)	MD 1.40 (‐4.89 to 7.69)	0.66	N/A	N/A
Subtotal		4 (478)	MD ‐2.16 (‐5.44 to1.12)	0.20	0.37	6%
6 hours
Ibuprofen	Pain VAS	4 (351)	MD ‐4.53 (‐10.23 to 1.18)	0.12	0.14	44%
Meloxicam		1 (127)	MD 0.50 (‐6.85 to 7.85)	0.89	N/A	N/A
Subtotal		4 (478)	MD ‐3.35 (‐8.05 to1.35)	0.16	0.15	40%
24 hours
Ibuprofen	Pain VAS	4 (351)	MD 1.38 (‐3.22 to 5.98)	0.56	0.71	0%
Meloxicam		1 (127)	MD ‐0.30 (‐11.40 to 10.80)	0.96	N/A	N/A
Subtotal		4 (478)	MD 1.14 (‐3.11 to5.39)	0.60	0.83	0%

Experimental intervention	Outcome	Number of studies (number of participants)	Effect measure RR, MD, SMD (95% CI)	P‐value for effect	P‐value heterogeneity	I² (%)
2 hours
Aspirin	Pain VAS	1 (25)	MD ‐10.90 (‐35.87 to 14.07)	0.39	N/A	N/A
Etoricoxib		1 (30)	MD ‐20.00 (‐28.42 to ‐11.58)	< 0.00001	N/A	N/A
Flurbiprofen		1 (25)	MD ‐5.70 (‐20.82 to 9.42)	0.46	N/A	N/A
Ibuprofen		2 (81)	MD 14.63 (0.77 to 28.50)	0.04	0.97	0%
Naproxen sodium		1 (25)	MD 2.50 (‐24.63 to 29.63)	0.86	N/A	N/A
Subtotal		3 (201)	MD ‐2.230 (‐16.06 to11.61)	0.75	0.002	73%
6 hours
Aspirin	Pain VAS	1 (25)	MD ‐8.70 (‐36.03 to 18.63)	0.53	N/A	N/A
Etoricoxib		1 (30)	MD ‐21.00 (‐27.09 to ‐14.91)	< 0.00001	N/A	N/A
Flurbiprofen		1 (25)	MD 8.60 (‐19.23 to 36.43)	0.54	N/A	N/A
Ibuprofen		2 (81)	MD 4.10 (‐10.28 to 18.48)	0.58	0.88	0%
Naproxen sodium		1 (25)	MD ‐9.00 (‐37.89 to 19.89)	0.54	N/A	N/A
Subtotal		3 (201)	MD ‐5.66 [‐18.97 to7.64]	0.40	0.02	63%
24 hours
Aspirin	Pain VAS	1 (25)	MD 5.10 (‐20.77 to 30.97)	0.70	N/A	N/A
Etoricoxib		1 (30)	MD ‐22.00 (‐27.06 to ‐16.94)	< 0.00001	N/A	N/A
Flurbiprofen		1 (25)	MD 35.90 (9.44 to 62.36)	0.008	N/A	N/A
Ibuprofen		4 (151)	MD ‐1.36 (‐13.04 to 10.33)	0.82	0.21	33%
Naproxen sodium		1 (25)	MD ‐2.60 (‐27.11 to 21.91)	0.84	N/A	N/A
Subtotal		5 (256)	MD ‐0.37 (‐14.41 to13.67)	0.96	< 0.00001	82%

Methods	Setting: private clinic, Brazil Design: cross‐over (3 arms) Number of centres: 1 Study duration: not reported
Participants	Inclusion criteria: people of both sexes; aged 16 to 25 years; orthodontic indication for bilateral retraction of the canine teeth  Exclusion criteria: not reported Orthodontic intervention: activation of canine retraction using NiTi springs Participant sampling: n = 36 recruited, randomized and analyzed (Group A = 36, Group B = 36, Group C = 36)
Interventions	Comparisons: NSAID vs placebo, and pre‐emptive vs post‐treatment analgesia Tenoxicam (20 mg) pre‐emptive vs placebo (lactose placebo) vs tenoxicam post‐treatment; provided to combat pain from canine activation Group A: tenoxicam 45 minutes before treatment, placebo just after, tenoxicam 24 h and 48 h after activation Group B: placebo 45 minutes before treatment, tenoxicam just after, tenoxicam 24 h and 48 h after activation Group C: placebo 45 minutes before treatment, just after treatment, 24 h and 48 h after activation
Outcomes	Pain score (VAS) recorded at 12 h, 24 h, 48 h and 72 h after each activation Pain score (VDS) recorded at 12 h, 24 h, 48 h and 72 h after each activation Amount of tooth movement (mm) recorded at 4 weeks after each activation (not an outcome of this review) Rescue medication paracetamol up to 750 mg 3 times per day
Notes	Conflict of interests/funding: "This study was supported by a grant‐in‐aid for scientific research from FAPESP ‐ Fundacao dr Arparo a Pesqiisa de Estado de Sao Paolo." Adverse events/harm: no rescue analgesia taken. Harm/adverse effects not reported Data handling by review authors: we were unable to extract any usable data from the study. Other information: each retraction procedure consisted of 3 activations, alternating between right and left with a 14‐day wash‐out interval between activations. Participants were randomised to two of the three groups for the left and right side retractions.
*Risk of bias*
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Quote: "The patients were initially randomized into three groups (A, B, and C) using the program for randomisation available at http://www.random.org." Comment: randomisation appears to be adequate
Allocation concealment (selection bias)	Unclear risk	Quote: "The drugs and administration methods had been concealed before the procedures." Comment: inadequate information regarding method of allocation concealment, therefore unable to make a judgement on appropriateness
Blinding of participants and personnel (performance bias)   All outcomes	Unclear risk	Comment: inadequate information regarding method of blinding, therefore unable to make a judgement on appropriateness
Blinding of outcome assessment (detection bias)   All outcomes	Unclear risk	Comment: inadequate information regarding method of blinding, therefore unable to make a judgement on appropriateness
Incomplete outcome data (attrition bias)   All outcomes	Low risk	0/36 dropouts = 0% attrition (100% completion)
Selective reporting (reporting bias)	High risk	Data for outcomes of this cross‐over trial were not reported appropriately. We were unable to extract data for use in pooled analysis.
Other bias	High risk	There were no data presented in the paper about baseline characteristics of the groups.

Methods	Setting: Department of Orthodontics, School of Dentistry, Kerman University of Medical Sciences, Kerman, Iran Design: parallel RCT (5 arms) Number of centres: 1 Study duration: not reported
Participants	Inclusion criteria: moderate (4 mm to 8 mm) crowding according to Little’s irregularity index (Little 1975); requiring extraction of 4 first or second premolars for orthodontic reasons  Exclusion criteria: systemic or periodontal diseases; previous orthodontic therapy; using analgesics, or medication interrupting tooth movement Orthodontic intervention: initial archwire placement Participant sampling: n = 100 recruited and randomized (34 male:66 female, aged 14 to 21 years, mean age 17.6 years); n = 90 returned their completed questionnaire; n = 90 had their data analyzed Number in each arm not reported Participant age and gender data not reported for any arm of the trial
Interventions	Comparison: NSAID versus placebo Ibuprofen (400 mg) vs placebo Group 1: placebo vitamin B6 40 mg tablet immediately after appliance placement and at 8‐h intervals for 1 week, if pain persisted Group 2: ibuprofen 400 mg tablet immediately after appliance placement and at 8‐h intervals for 1 week, if pain persisted Group 3: bite wafer to chew on immediately after appliance placement and at 8‐h intervals for 1 week, if pain persisted Group 4: low‐level red laser received as a single session irradiation from a low power indium‐gallium‐aluminium‐phosphide diode laser immediately after appliance placement Group 5: low‐level infrared laser received as a single session irradiation from a low‐level gallium‐aluminium‐arsenide diode laser immediately after appliance placement Only Groups 1 and 2 are included in the comparisons for this review
Outcomes	Pain score (VAS) recorded at 2 h, 6 h, 10 h (or bedtime), 24 h, 2 days, 3 days and 7 days after archwire placement Pain was recorded during chewing, biting, fitting back teeth together and fitting front teeth together (the latter 3 not being used in this review)
Notes	Conflict of interests/funding: no source of funding or conflict of interest reported Adverse events/harm: not reported Data handling by review authors: data for Groups 1 and 2 have been inverted to reflect ibuprofen as an intervention and placebo as its control. Data from Groups 3, 4 and 5 did not contribute to the analyses. Other information: "The study protocol was reviewed and approved by the Ethics Committee of Kerman University of Medical Sciences."
*Risk of bias*
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Quote: "The subjects were randomly assigned to 5 groups of 20 each according to a computer‐generated random assignment program." Comment: randomisation appears to be adequate
Allocation concealment (selection bias)	Unclear risk	Comment: inadequate information regarding method of allocation concealment, therefore unable to make a judgement on appropriateness
Blinding of participants and personnel (performance bias)   All outcomes	Low risk	Quote: "The subjects assigned to the ibuprofen and placebo medication groups were blinded to their allocation. Furthermore, the investigator who assessed the outcomes and the data analyst were kept blinded to the assignments." Comment: blinding appears to be adequate
Blinding of outcome assessment (detection bias)   All outcomes	Low risk	Quote: "The investigator who assessed the outcomes and the data analyst were kept blinded to the assignments." Comment: blinding appears to be adequateding
Incomplete outcome data (attrition bias)   All outcomes	High risk	59/100 dropouts = 59% attrition (41% completion)
Selective reporting (reporting bias)	High risk	Data for outcomes of this study were not reported appropriately. We were unable to extract data for use in pooled analysis.
Other bias	High risk	No baseline characteristics of groups provided; no sample sizes for the final numbers included in the analysis

Methods	Setting: University of Iowa, College of Dentistry, Department of Orthodontics Design: parallel RCT (3 arms) Number of centres: 1 Study duration: not reported
Participants	Inclusion criteria: scheduled to begin comprehensive orthodontic treatment; no prophylactic antibiotic coverage required; no debilitating systemic diseases; currently not using antibiotics or analgesics; no contraindication to the use of ibuprofen; a maximum age of 16 years and a minimum weight requirement of 88 pounds (weight requirement based on Food and Drug Administration–approved over‐the‐counter paediatric dosage labelling guidelines)  Exclusion criteria: none outlined in the Methods, but trialists later excluded participants who did not agree to participate or did not return a completed questionnaire, and those who took additional 'rescue' medication. Orthodontic intervention: separator placement Participant sampling: n = 114 recruited and randomized n = 63 returned their completed questionnaire n = 22 excluded from analysis for taking additional medication (evenly distributed between the 3 groups) n = 41 data analyzed for (aged 9 years 3 months to 16 years 11 months) Group 1 (n = 13) 10 male:3 female, mean age 12.1 + 1.6 years Group 2 (n = 14) 4 male:10 female, mean age 13.5 + 1.9 years Group 3 (n =14) 6 male:8 female, mean age12.8 + 1.5 years
Interventions	Comparisons: NSAID vs placebo, and pre‐emptive vs post‐treatment Ibuprofen (400 mg) vs control (lactose placebo) Group 1: ibuprofen 1 h before placement, followed by ibuprofen 6 h after initial dose Group 2: ibuprofen 1 h before placement, followed by placebo 6 h after initial dose Group3: placebo 1 h before placement, followed by ibuprofen 6 h after initial dose
Outcomes	Pain score (VAS) recorded at 2 h, 6 h, 10 h (bedtime), 17 h (awakening) and 24 h, and 2, 3 and 7 days after separator placement Pain was recorded during chewing (other measures reported in the study but not included in this review included pain during biting, when fitting back teeth together, and when fitting front teeth together).
Notes	Conflict of interests/funding: no source of funding or conflict of interest reported Adverse events/harm: 22 participants excluded from analysis for taking additional medication. No harms reported Data handling by review authors: data presented for analysis is based on Figure 1 in the paper, which showed mean pain scores (mean + SEM) for chewing. The SEM was used to calculate SD. Data from Group 1 did not contribute to the analyses, Groups 2 and 3 data were used for the comparison of pre‐emptive versus post‐treatment analgesia. Other information of note: wide variation in gender at baseline for Group 1: "a wide range of individual variation was noted in the pain levels reported, which resulted in large standard deviations. Another possible explanation is the uneven distribution of male and female patients among the groups".
*Risk of bias*
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Quotes: "Patients were randomly assigned to 1 of 3 experimental groups." "The randomization of which of the three experimental conditions the patients were assigned to was computer generated." Comment: randomisation appears to be adequate
Allocation concealment (selection bias)	Low risk	Quotes: "Patients were randomly assigned to 1 of 3 experimental groups." "The randomization of which of the three experimental conditions the patients were assigned to was computer generated." Comment: allocation concealment appears to be adequate
Blinding of participants and personnel (performance bias)   All outcomes	Low risk	Quote: "The ibuprofen and placebo capsules were identical in appearance. The patient, research assistant, and investigator were blinded to each subject’s experimental group." Comment: adequate method of blinding
Blinding of outcome assessment (detection bias)   All outcomes	Low risk	Quote: "The ibuprofen and placebo capsules were identical in appearance. The patient, research assistant, and investigator were blinded to each subject’s experimental group." Comment: adequate method of blinding
Incomplete outcome data (attrition bias)   All outcomes	High risk	73/114 dropouts = 64% attrition (36% completion). High number of dropouts, but equally distributed across the groups
Selective reporting (reporting bias)	Low risk	Data for outcomes of this review reported appropriately
Other bias	Low risk	No other sources of bias detected.

Methods	Setting: Postgraduate Orthodontic Clinic, Kansas City School of Dentistry, University of Missouri Design: parallel (2 arms) Number of centres: 1 Study duration: not reported
Participants	Inclusion criteria: starting orthodontic treatment that required separators; no orthodontic appliances in the mouth; no debilitating systemic diseases or GI problems; not currently taking analgesics or antibiotics; 9 to 19 years of age; minimum weight of 88 pounds; no prophylactic antibiotic coverage required; literate; English‐speaking; able to swallow pills; and no contraindication to the use of acetaminophen or ibuprofen Exclusion criteria: not specified Orthodontic intervention: separator placement Participant sampling: n = 40 selected and randomized n = 7 dropouts/excluded from analysis (sought treatment elsewhere n = 1; did not complete questionnaire n = 1; excluded from analysis for not taking medication before the appointment n = 5) n = 33 data analyzed No numbers per group were provided and no sex or age data
Interventions	Comparison: NSAID vs paracetamol Ibuprofen (400 mg) vs paracetamol (650 mg); provided pre‐emptively to separator placement Group 1: paracetamol 1 h before placement Group 2: ibuprofen 1 h before placement
Outcomes	Pain score (VAS) recorded immediately after separator placement, at 2 to 3 hours, at bedtime and on rising the next day Pain was recorded during the following activities: chewing teeth not touching (not an outcome of this review) biting back teeth together (not an outcome of this review)
Notes	Conflict of interests/funding: no source of funding or conflict of interest reported. Adverse events/harm: none reported Data handling by review authors: only pain during chewing data are reflected in this systematic review. Other information of note: no SD, baseline or group size data presented in the paper; no reply from authors when emailed and asked to provide further information
*Risk of bias*
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Quote: "A computer‐generated stratified random assignment strategy was used to balance the groups with respect to sex." Comment: method of randomisation appears to be adequate
Allocation concealment (selection bias)	Low risk	Quote: "The tablets were given to the patients in sealed, coded envelopes." Comment: adequate method of allocation concealment
Blinding of participants and personnel (performance bias)   All outcomes	Low risk	Quote: "The investigator (C.B.), the patient, and the parent were blinded to the experimental group." Comment: adequate method of blinding
Blinding of outcome assessment (detection bias)   All outcomes	Low risk	Quote: "The investigator (C.B.), the patient, and the parent were blinded to the experimental group." Comment: adequate method of blinding
Incomplete outcome data (attrition bias)   All outcomes	Low risk	7/40 dropouts = 17.5% attrition (82.5% completion)
Selective reporting (reporting bias)	High risk	Data for outcomes of this study were not reported appropriately. We were unable to extract data for use in pooled analysis.
Other bias	High risk	No baseline characteristics of groups provided; no sample sizes for the final numbers included in the analysis.

Methods	Setting: Dorset County Hospital Dorchester; Royal United Hospitals, Bath, and, Southmead Hospital, Bristol Design: parallel (2 arms) Number of centres: 3 Study duration: February 2004 to December 2005 (23 months)
Participants	Inclusion criteria: age between 12 and 16 years; no history of peptic ulceration, or renal, hepatic, or cardiac impairment; no history of asthma requiring steroid inhalers or unstable asthma in the last year; no history of adverse reactions to ibuprofen or paracetamol; and currently not using analgesics or antibiotics  Exclusion criteria: not specified Orthodontic intervention: separator placement Participant sampling: n = 208 selected n = 21 excluded n = 21 refused to participate n = 187 randomized (Group 1 n = 92; Group 2 n = 95) n = 28 dropouts/excluded from analysis (lost to follow‐up (not returning questionnaire): Group 1 n = 6, Group 2 n = 3; did not fulfil inclusion criteria: Group 1 n = 9, Group 2 n = 10) n = 159 data analyzed Group 1 (n = 77): 28 male:49 female, mean age 13.7 + 1.0 years Group 2 (n = 82): 29 male:53 female, mean age 13.8 + 1.2 years Age: P = 0.38 (independent‐samples t test) Sex: P = 1.00 (Chi² test for association) male 35.8%, female 64.2%
Interventions	Comparison: NSAID vs paracetamol Ibuprofen (400 mg; 2 x 200 mg caplet) vs paracetamol (1 g; 2 x 500 mg caplet); provided pre‐emptively to separator placement, and again post‐treatment Group 1: ibuprofen 1 h before placement, followed by ibuprofen 6 h after initial dose Group 2: paracetamol 1 h before placement, followed by paracetamol 6 h after initial dose
Outcomes	Pain score (VAS) ‐ recorded at 2 h, 6 h, 10 h/bedtime (primary outcome), 24 h and 2, 3 and 7 days (secondary outcome) after separator placement
Notes	Conflict of interests/funding: "We thank the Clinical Trials team in the Pharmacy Production Unit at the Royal Hallamshire Hospital, Sheffield, for supplying the drugs and performing the randomization, and the patients who participated in this trial." Both ibuprofen and paracetamol supplies were cited as being produced by Boots Company, Nottingham, United Kingdom. Adverse events/harm: 18 participants required additional medication. Quote: "During this trial 1 patient experienced a suspected adverse reaction to paracetamol." This was reported in more detail in the secondary paper by McAlinden et al (Bradley 2007): "since one of the trial drugs was also paracetamol we decided to break the randomization code for this patient to determine which drug the patient had received. The trial drug given was found to be paracetamol, suggesting a drug hypersensitivity reaction to paracetamol". Data handling by review authors: study reported Group 1 as the paracetamol arm of trial, and Group 2 as the ibuprofen arm. In order to align with the protocol for this systematic review, the figures for Groups 1 and 2 have been inverted to reflect ibuprofen as an intervention and paracetamol as its control. Other information of note: intention‐to‐treat analysis noted in methods, but only completing participants were analyzed. Trial authors carried out an intention‐to‐treat analysis on the 18 participants who required additional analgesia.
*Risk of bias*
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Quote: "The drugs were supplied according to a restricted randomization method in blocks of 8 to ensure that equal numbers of patients were allocated to each group." Comment: block randomisation carried out therefore it can be assumed that this was adequate
Allocation concealment (selection bias)	Low risk	Quote: "The analgesics were in the form of identical capsules and were stored in sealed, numbered containers. The random allocation sequence was concealed in an envelope and held centrally." Comment: adequate method of allocation concealment
Blinding of participants and personnel (performance bias)   All outcomes	Low risk	Quote: "The investigator, the clinicians, the subjects, and the statistician were all blinded to each subject’s treatment group." Comment: adequate method of blinding
Blinding of outcome assessment (detection bias)   All outcomes	Low risk	Quote: "The investigator, the clinicians, the subjects, and the statistician were all blinded to each subject’s treatment group." Comment: adequate method of blinding
Incomplete outcome data (attrition bias)   All outcomes	Low risk	28/187 dropouts = 15% attrition (85% completion)
Selective reporting (reporting bias)	Low risk	Data for outcomes of this review reported appropriately
Other bias	Low risk	No other sources of bias detected

Methods	Setting: Dentistry School of Universidade Federal Fluminense, (Niterói, RJ, Brazil) Design: parallel (3 arms) Number of centres: 1 Study duration: not reported
Participants	Inclusion criteria: at least 18 years of age; presence of second molars and second bicuspids, since separating elastics had to be fixed on the first molars; no clinical signs of gingival inflammation  Exclusion criteria: use of any medication that could interfere with results > 2 weeks before the procedure; any of the following conditions, screened through a questionnaire: cardiopathies, nephropathies, hepatopathies or GI disorders, diabetes, high cholesterol, blood vessel obstructions, allergy to anti‐inflammatory drugs, intolerance to lactose, pregnancy Orthodontic intervention: separator placement Participant sampling: n = 87 recruited and randomized n = 38 dropouts or excluded from analysis (n = 18 missing or incomplete information; n = 10 discomfort due to elastic, sought treatment elsewhere; n = 6 used analgesic medication during the study; n = 2 lost their diaries and were unwilling to be resubmitted to the intervention) n = 51 data analyzed Group A (n = 17) 4 male:13 female, mean age 24.64 years Group B (n = 17) 4 male:13 female, mean age 22.64 years Group C (n = 17) 5 male:12 female, mean age 22.47 years
Interventions	Comparison: NSAID vs placebo Lumiracoxib (400 mg) vs placebo vs no treatment; provided pre‐emptively to separator placement Group A: lumiracoxib 1 h before separator placement Group B: placebo 1 h before separator placement Group C: no intervention
Outcomes	Pain score (VAS) recorded at 2 h, 6 h and 24 hours and at 2 and 4 days after separator placement
Notes	Conflict of interests/funding: quote: "The authors have reported no conflicts of interest." Adverse events/harm: not reported Data handling by review authors: study reports Group A as placebo, Group B as lumiracoxib and Group C as control  For the purposes of aligning with the protocol for this systematic review, the figures for Groups A and B have been inverted to reflect ibuprofen as an intervention and placebo and no treatment as its controls. Other information of note: we received data for means and standard deviations via correspondence with the author.
*Risk of bias*
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Quote: "Patients were randomly assigned into one of three groups by drawing lots. To ensure similarity in size of the groups, randomisation was stratified in blocks of ten (permuted‐block randomisation)." Comment: block randomisation carried out, therefore it can be assumed that this was adequate.
Allocation concealment (selection bias)	High risk	Quote: "Patients were randomly assigned into one of three groups by drawing lots"; "As each volunteer attended the Department of Orthodontics, in Universidade Federal Fluminense (the University), he was allocated to the group following the last participant had entered. The search was not started with the sample enclosed." Comment: allocation concealment not achieved
Blinding of participants and personnel (performance bias)   All outcomes	High risk	Quote: "The placebo and lumiracoxib capsules were perfectly identical and neither the researchers nor the subjects knew the group of each subject. Patients of the non‐medication group knew about the use of capsules by the other two groups." Comment: adequate method of blinding where appropriate. Not possible to blind participants and personnel to allocated control group
Blinding of outcome assessment (detection bias)   All outcomes	Low risk	Quote: "The VAS was given to a statistician blinded to the study group." Comment: adequate method of blinding.
Incomplete outcome data (attrition bias)   All outcomes	High risk	36/87 dropouts = 41.4% attrition (58.6% completion)
Selective reporting (reporting bias)	Low risk	Raw data not presented in paper, only statistical analysis and significance. Further information regarding data was received via correspondence with the author.
Other bias	Low risk	No other source of bias detected

Methods	Setting: Orthodontics Department of Shahid Beheshti University, School of Dentistry and a private clinic in Tehran Design: cross‐over (3 arms) Number of centres: 2 Study duration: not reported
Participants	Inclusion criteria: no pain at the onset of study; in previous sessions complained of pain > 50 based on the VAS; 6 mm to 8 mm crowding; no use of analgesics during the study period; no history of renal or liver disease or any other contraindication for the use of understudy medications (Comment: we have assumed a typographical error was made in the paper and that the trial authors meant 'liver' disease when they wrote 'river' disease.)  Exclusion criteria: did not sign the consent form; used analgesics and anti‐inflammatory drugs during the study; did not complete the questionnaire; did not use or used benzocaine or ketoprofen chewing gums improperly Orthodontic intervention: mid‐treatment adjustments Participant sampling: n = 30 recruited and randomized n = 4 dropouts/excluded from analysis n = 26 data analyzed Group A (n = 26) 12 male:14 female, mean age 18.07 + 3.19 years Group B (n = 26) 12 male:14 female, mean age 18.07 + 3.19 years Group C (n = 26) 12 male:14 female, mean age 18.07 + 3.19 years
Interventions	Comparisons: NSAID vs placebo, local anaesthetic vs placebo, and NSAID vs local anaesthetic Ketoprofen chewing gum vs benzocaine chewing gum vs placebo chewing gum provided postoperatively following appliance adjustments Group A: ketoprofen chewing gum every 8 h for 3 days after treatment Group B: benzocaine chewing gum every 8 h for 3 days after treatment Group C: placebo chewing gum every 8 h for 3 days after treatment
Outcomes	Pain score (VAS) ‐ recorded at 2 h, 6 h, 24 hours, 10 am and 6 pm on day 2, 10 am and 6 pm on day 3, and 7 days after appliance adjustment
Notes	Conflict of interests/funding: not reported Adverse events/harm: not reported Data handling by review authors: study did not allocate intervention to specific group labels. For the purposes of this systematic review, Group A is the ketoprofen arm, Group B the benzocaine arm and Group C the control arm.  Other information of note: data for means and standard deviations were received via correspondence with the author. Orthodontic intervention involved in the study was described as "fixed orthodontic treatment". This was clarified through correspondence with the author as involving retie of an 0.016" or 0.018" NiTi archwire. 4‐week washout period allowed between interventions
*Risk of bias*
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Quote: "Thirty patients were randomly divided into three groups of 10"; "In the first session patients 1‐10 receive ketoprofen, 11‐20 receive benzocaine, 21‐30 receive placebo gums. In the next visit 1‐10 benzocaine, 11‐20 placebo, 21‐30 ketoprofen; and in the last visit 1‐10 placebo, 11‐20 ketoprofen, 21‐30 benzocaine". Comment: inadequate information and methods regarding method of randomisation, therefore unable to make a judgement on appropriateness.
Allocation concealment (selection bias)	Low risk	Quote: "Thirty patients were randomly divided into three groups of 10"; "random allocation by a third person who was responsible for explaining to the patients". Comment: appeared to be adequate method of allocation concealment.
Blinding of participants and personnel (performance bias)   All outcomes	Low risk	Quote: "The placebo chewing gum was manufactured with the same shape and packaging as the experimental gums. Patients and those administering the gums among patients were blinded to the type of chewing gums". Comment: adequate method of blinding
Blinding of outcome assessment (detection bias)   All outcomes	Low risk	Quote: "Questionnaires were analyzed by a statistician blinded to the group allocation of patients". Comment:adequate method of blinding
Incomplete outcome data (attrition bias)   All outcomes	Low risk	4/30 dropouts = 13.3% attrition (86.7% completion). However the reason for drop out was unclear.
Selective reporting (reporting bias)	High risk	Unclear data presented in paper and received from correspondence with the author. We were unable to use this study in pooling for meta‐analysis.
Other bias	Low risk	No other source of bias detected

Methods	Setting: Orthodontics Departmentof Shahid Beheshti University, School of Dentistry, Tehran, Iran Design: cross‐over (2 arms) Number of centres: 1 Study duration: 14 months (January 2011 up to March 2012)
Participants	Inclusion criteria: signed informed consent of the included subjects or their legally authorised representatives, no current chronic or acute pain in the oral cavity, positive history of pain after first orthodontic appliance activation, and subject was in the levelling and alignment phase of treatment  Exclusion criteria: taking analgesics or antibiotics, history of severe kidney or liver disease, and history of allergic reaction to local anaesthetic drugs or a history of methemoglobinemia Orthodontic intervention: mid‐treatment adjustments Patient sampling: n = 30 recruited and randomized n = 5 dropouts/excluded from analysis (n = 1 refusal to continue, n = 2 imprecision in obeying the study protocol, n = 2 use of other pain relieving drugs during the study (exclusion of these participants is a bias)) n = 25 data analyzed (8 male:17 female, mean age 19.5 years range 15 to 25 years) Group A (n = 13) data for age and gender not presented Group B (n = 12) data for age and gender not presented
Interventions	Comparison: Local anaesthetic vs placebo Local anaesthetic vs placebo gel; provided postoperatively following appliance adjustments Group A: 5% benzocaine gel twice a day (at 10 am and 10 pm) for 3 days Group B: placebo gel twice a day (at 10 am and 10 pm) for 3 days
Outcomes	Pain score (VAS) ‐ recorded at 2 h, 6 h, 24 h and day 2 at 10 am and 6 pm, day 3 10 am and 6 pm and 7 days after appliance adjustment
Notes	Conflict of interests/funding: not reported Adverse events/harm: not reported Data handling by review authors: The data presented for the analysis were based on Figure 1, which showed mean pain scores (mean + SEM). The SEM was used to calculate SD.  Other information of note: 4‐week washout period allowed between interventions. We clarified that the orthodontic intervention involved in the study involved retie of an 0.016" or 0.018" NiTi archwire through correspondence with the author.
*Risk of bias*
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Quote: "Randomization and sequence generation were carried out with the aid of random number generator in SPSS software". Comment: adequate method of randomisation
Allocation concealment (selection bias)	Low risk	Quote: "Allocation concealment was achieved with sequentially numbered letters in opaque envelopes containing gel tube numbers." Comment: adequate method of allocation concealment
Blinding of participants and personnel (performance bias)   All outcomes	Low risk	Quote: "Adequate method of allocation concealment. The tubes were essentially the same, and neither the subjects nor the clinicians knew whether the tubes contained placebo or benzocaine." Comment: adequate method of blinding
Blinding of outcome assessment (detection bias)   All outcomes	Unclear risk	Comment: inadequate information regarding method of blinding of outcome assessment, therefore unable to make a judgement on appropriateness.
Incomplete outcome data (attrition bias)   All outcomes	High risk	5/30 dropouts = 16.7% attrition (83.3% completion). Participants were excluded as they used other pain relieving drugs ‐ this is a source of bias as these were probably the participants experiencing the greatest pain.
Selective reporting (reporting bias)	High risk	Data for outcomes of this review were not reported appropriately. The cross‐over structure of the data was ignored in the trialists' presentation and analysis of the data.
Other bias	Low risk	No other source of bias detected.

Methods	Setting: Orthodontic clinic of Mashhad University of Medical Sciences in Iran Design: parallel (5 arms) Number of centres: 1 Study duration: not reported
Participants	Inclusion criteria: female orthodontic patients between 13 and 18 years of age, scheduled for fixed orthodontic treatment, with no systemic diseases and not receiving analgesic therapy; with moderate crowding (4 mm to 8 mm) according to Little’s irregularity index (Little 1975)  All participants needed extraction of the 4 first premolars for orthodontic reasons, and the extractions were scheduled to be finished at least 2 weeks before the placement of the orthodontic appliances.  Exclusion criteria: none specified Orthodontic intervention: initial archwire placement Patient sampling: n = 50 recruited and randomized and analyzed Group 1 (n = 10) female only, age data not reported Group 2 (n = 10) female only, age data not reported Group 3 (n = 10) female only, age data not reported Group 4 (n = 10) female only, age data not reported Group 5 (n = 10) female only, age data not reported
Interventions	Comparison: NSAID vs placebo Ibuprofen (400 mg) vs placebo vs chewing gum vs soft viscoelastic bite wafer vs hard viscoelastic bite wafer; provided after initial archwire placement Group 1: ibuprofen immediately after archwire placement and at 8‐h intervals for a week if pain persisted Group 2: placebo immediately after archwire placement and at 8‐h intervals for a week if pain persisted Group 3: gum chewed for 5 minutes immediately after archwire placement and at 8‐h intervals for a week if pain persisted Group 4: soft bite wafer chewed or bitten down on for 5 minutes at 8‐h intervals for a week if pain persisted Group 5: hard bite wafer chewed or bitten down on for 5 minutes at 8‐h intervals for a week if pain persisted
Outcomes	Pain score (VAS) ‐ recorded at 2 h, 6 h, bedtime, 24 h and 2, 3 and 7 days after placement of initial archwires Pain was recorded during the following activities: chewing biting (not an outcome of this review) fitting front teeth together (not an outcome of this review) fitting posterior teeth together (not an outcome of this review)
Notes	Conflict of interests/funding: not reported Adverse events/harm: not reported. “None had used any analgesics” Data handling by review authors: study reports Group 1 as placebo arm and Group 2 as ibuprofen arm. For the purposes of aligning with this systematic review's own protocol, the figures for Groups 1 and 2 have been inverted to reflect ibuprofen as an intervention and placebo as its control. Data for Groups 3, 4 and 5 have not been included for the purposes of this review. Other information of note: only pain during chewing data were reflected in this systematic review.
*Risk of bias*
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Quote: "The subjects were randomly assigned to 1 of 5 parallel groups in a 1:1:1:1:1 ratio according to their clinical entrance number and a random number table." Comment: block randomisation carried out, therefore it can be assumed that this was adequate.
Allocation concealment (selection bias)	Unclear risk	Quote: "The subjects were randomly assigned to 1 of 5 parallel groups". Comment: inadequate information regarding method of allocation concealment, therefore unable to make a judgement on appropriateness.
Blinding of participants and personnel (performance bias)   All outcomes	Low risk	Quote: "The subjects in these 2 groups (ibuprofen & placebo) were blinded about the drug that they took." Comment: adequate method of blinding was carried out where appropriate for Group 1 and Group 2. Not possible to blind participants and personnel regarding allocation to Groups 3, 4 or 5.
Blinding of outcome assessment (detection bias)   All outcomes	Unclear risk	Inadequate information regarding method of blinding of outcome assessment.
Incomplete outcome data (attrition bias)   All outcomes	Low risk	0/50 dropouts = 0% attrition (100% completion)
Selective reporting (reporting bias)	Low risk	Data for outcomes of this review were reported appropriately.
Other bias	Low risk	No other source of bias detected.

Methods	Setting: 2 private practices in Lee’s Summit, Missouri Design: parallel (3 arms) Number of centres: 2 Study duration: not reported
Participants	Inclusion criteria: started orthodontic treatment that required banding of posterior teeth and placement of 2 or more separators; able to swallow analgesic pills; English speaking; 9 to 17 years of age; minimum weight requirement of 88 pounds based on mg/kg paediatric dosage recommendations  Exclusion criteria: existing orthodontic or space maintenance appliances; contraindication to the use of acetaminophen or ibuprofen; taking antibiotics or analgesics; cognitive impairment, or any systemic disease that in the assessment of the investigator might impact pain perception Orthodontic intervention: separator placement Patient sampling: n = 35 enrolled n = 9 dropouts/excluded from analysis (Group 1 = 4, Group 2 = 2, Group 3 = 3 all failed to return questionnaires) n = 26 data analyzed for: Group 1 (n = 7) 1 male:6 female, mean age 12.7 + 1.3 years Group 2 (n = 10) 3 male:7 female, mean age 13.0 + 1.6 years Group 3 (n = 9) 5 male:4 female, mean age 12.6 + 1.8 years
Interventions	Comparisons: NSAID vs placebo, NSAID vs paracetamol, and paracetamol vs placebo Ibuprofen (400 mg) vs paracetamol (650 mg) vs placebo (640 mg avicel); provided pre‐emptively and post‐treatment for separator placement Group 1: ibuprofen 1 h before placement and 6 h after initial dose Group 2: paracetamol 1 h before placement and 6 h after initial dose Group 3: placebo 1 h before placement and 6 h after initial dose
Outcomes	Pain score (VAS) ‐ recorded immediately before and after, 2 h, 6 h, bedtime, 24 h after separator placement Pain was recorded during the following activities: chewing teeth not touching (not an outcome of this review) biting (not an outcome of this review)
Notes	Conflict of interests/funding: not reported Adverse events/harm: not reported Data handling by review authors: study reports Group 1 as placebo arm and Group 3 as ibuprofen arm. For the purposes of aligning with this systematic review's own protocol, the figures for Groups 1 and 3 have been inverted to reflect ibuprofen as an intervention and placebo as its control. Other information of note: only pain during chewing data were reflected in this systematic review.
*Risk of bias*
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Quote: "Computer generated random patient coding and group allocation was utilized". Comment: randomisation appears to be adequate
Allocation concealment (selection bias)	Low risk	Quote: "The random allocation assignments were concealed and inaccessible to the investigator." Comment: allocation concealment appears to be adequate
Blinding of participants and personnel (performance bias)   All outcomes	Low risk	Quote: "The ibuprofen, acetaminophen, and placebo tablets were compounded by a licensed pharmacist (O’Brien Pharmacy, Kansas City, MO) according to specifications and were all provided in identical white opaque capsules. Medications and placebo tablets were packed and distributed in sealed, coded envelopes"; "Subjects, patients and investigator would be blinded to group allocation". Comment: adequate method of blinding
Blinding of outcome assessment (detection bias)   All outcomes	Low risk	Quote: "The random allocation assignments were concealed and inaccessible to the investigator"; "Subjects, patients and investigator would be blinded to group allocation". Comment: adequate method of blinding
Incomplete outcome data (attrition bias)   All outcomes	High risk	9/35 dropouts = 25.7% attrition (74.3% completion)
Selective reporting (reporting bias)	Low risk	Data for outcomes of this review were reported appropriately.
Other bias	High risk	Large gender variation at baseline, more females in all groups indicating sampling bias.

PERMALINK

Pharmacological interventions for pain relief during orthodontic treatment

Aoife B Monk

Jayne E Harrison

Helen V Worthington

Annabel Teague

Abstract

Background

Objectives

Search methods

Selection criteria

Data collection and analysis

Main results

Authors' conclusions

Plain language summary

Summary of findings

Summary of findings for the main comparison. Analgesic compared to control (placebo or no treatment) for pain relief during orthodontic treatment.

Summary of findings 2. NSAID compared to paracetamol for pain relief during orthodontic treatment.

Summary of findings 3. Pre‐emptive ibuprofen compared to ibuprofen post‐treatment for pain relief during orthodontic treatment.

Summary of findings 4. NSAID compared to local anaesthetic for pain relief during orthodontic treatment.

Background

Description of the condition

Description of the intervention

How the intervention might work

Opioids

NSAIDs

Paracetamol

Local anaesthetic

Why it is important to do this review

Objectives

Methods

Criteria for considering studies for this review

Types of studies

Types of participants

Inclusion criteria

Exclusion criteria

Types of interventions

Active interventions

Controls

Types of outcome measures

Primary outcomes

Secondary outcomes

Search methods for identification of studies

Electronic searches

Searching other resources

Language

Unpublished studies

Data collection and analysis

Selection of studies

Data extraction and management

Assessment of risk of bias in included studies

1.

2.

Measures of treatment effect

Unit of analysis issues

Dealing with missing data

Assessment of heterogeneity

Assessment of reporting biases

Data synthesis

Subgroup analysis and investigation of heterogeneity

Sensitivity analysis

Results

Description of studies

Results of the search

3.

Included studies

Characteristics of the trial designs and settings

Design

Setting

Funding

Conflict of interest

Characteristics of the participants

Orthodontic intervention

Characteristics of the interventions and comparisons

Opioid

Paracetamol

NSAIDs

Local anaesthetic

Duration

Characteristics of the outcomes

Methods	Setting: Graduate Orthodontic Clinic at the University of Kentucky College of Dentistry and the full‐time and part‐time faculty practices Design: parallel (2 arms) Number of centres: unclear Study duration: not reported
Participants	Inclusion criteria: orthodontic treatment included full, fixed orthodontic appliances (braces); males and females; periodontal tissues were in good health; no systemic disease that would compromise normal healing (e.g. diabetes); no medications being taken at the time of the study  Exclusion criteria: none specified Orthodontic intervention: initial bracket placement without archwire Patient sampling: n = 80 randomized (Group 1 n = 40; Group 2 n = 40) n = 10 dropouts/excluded from analysis (lost to follow‐up: not returning questionnaire = 7; did not require pain relief = 3) n = 70 data analyzed for: Group 1 (n = 35) 18 male:17 female, mean age 23.8 + 10.3 years Group 2 (n = 35) 16 male:19 female, mean age 25.2 + 8.6 years Age P = 0.5 Sex: P = 0.6
Interventions	Comparison: local anaesthetic wax vs placebo Benzocaine wax vs placebo; provided post‐treatment for topical use, not applied until 24 h after treatment Group 1: orthodontic menthol wax, medicated with 20% benzocaine Group 2: unmedicated orthodontic wax (without menthol)
Outcomes	Pain score (VAS) ‐ recorded at baseline, 1 h, 17 h, 29 h, 41 h, 53 h after baseline
Notes	Conflict of interests/funding: not reported. The benzocaine wax product received a patent by the United States Patent and Trademark Office (Patent No.6,074,674) Adverse events/harm: not reported Data handling by review authors: only intervention study data included in this review. Study did not allocate intervention to specific group labels. For the purposes of this systematic review, Group 1 has been allocated as the benzocaine arm and group 2 as the control arm. Although multiple time points measured, intervention not taken until 24 h after visit, therefore data from this study has not contributed to analyses in this systematic review. Other information of note: much of the detail in the paper related to the pilot (e.g. the inclusion criteria) prior to the RCT. Both are presented collectively in the paper.
*Risk of bias*
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Quote: "The medicated and unmedicated waxes were randomized and were contained in numerically coded cases." Comment: inadequate information regarding method of randomisation, therefore unable to make a judgement on appropriateness.
Allocation concealment (selection bias)	Low risk	Quote: "The medicated and unmedicated waxes were randomized and were contained in numerically coded cases." Comment: allocation concealment appears to be adequate
Blinding of participants and personnel (performance bias)   All outcomes	Low risk	Quote: "The patients received dental wax without knowing whether the anaesthetic was incorporated into the wax". Comment: described as double‐blind, both identically prepared, however, "neither the benzocaine nor the menthol" was included in the placebo. Intervention and control possessed different tastes, but patients unlikely to recognise whether active drug or not, or to discuss with other participants.
Blinding of outcome assessment (detection bias)   All outcomes	Unclear risk	Quote: "Randomised double‐blind prospective RCT". Comment: described as double‐blind, but inadequate information provided regarding how blinding was carried out, therefore unable to make a judgement on appropriateness.
Incomplete outcome data (attrition bias)   All outcomes	Low risk	10/80 dropouts = 12.5% attrition (87.5% completion)
Selective reporting (reporting bias)	Low risk	Data for outcomes of this review were reported appropriately.
Other bias	Low risk	No other source of bias detected.

Methods	Setting: Department of Orthodontics & Dentofacial Orthopedics, Hitkarini Dental College & Hospital, Jabalpur, India Design: parallel (3 arms) Number of centres: 1 Study duration: not reported
Participants	Inclusion criteria: at least 13 years of age and not older than 20 years; beginning orthodontic treatment for the first time; reporting no contraindications or adverse reactions related to ibuprofen and piroxicam; not using any antibiotics; and meeting a minimum weight requirement of 88 pounds, as per Food and Drug Administration approved‐over‐the‐counter paediatric dosage labelling guidelines; required to provide written informed consent for participation in the study.  Exclusion criteria: none specified Orthodontic intervention: separator placement Patient sampling: n = 90 randomized (Group 1 n = 30; Group 2 n = 30; Group 3 n = 30) n = 0 dropouts/excluded from analysis n = 90 analyzed (45 male: 45 female) Group 1 (n = 30) 15 male:15 female, mean age 14.7 + 3.4 years  Group 2 (n = 30) 15 male:15 female, mean age 14.2 + 2.8 years Group 3 (n = 30) 15 male:15 female, mean age 15.1 + 3.6 years
Interventions	Comparisons: NSAID vs placebo vs NSAID Ibuprofen (400 mg) vs piroxicam (20 mg) vs placebo; provided 1 h pre‐emptively Group 1: ibuprofen 1 h before separator placement Group 2: placebo 1 h before separator placement Group 3: piroxicam 1 h before separator placement
Outcomes	Pain score (VAS) ‐ recorded at 2 h, 6 h, bedtime, 24 h and 2, 3 and 7 days after separator placement Pain was recorded during the following activities: chewing biting (not an outcome of this review) fitting front teeth together (not an outcome of this review) fitting posterior teeth together (not an outcome of this review)
Notes	Conflict of interests/funding: not reported Adverse events/harm: not reported. “None of them had resorted to the usage of any kind of ‘rescue medication.’” Data handling by review authors: study reports Group 1 as placebo arm of trial, Group 2 as ibuprofen arm and Group 3 as piroxicam arm. For the purposes of aligning with this systematic review's own protocol, the figures for Groups 1 and 2 have been inverted to reflect ibuprofen as an intervention and piroxicam and placebo as its controls. Additional information received through correspondence with the author. Other information of note: only pain during chewing data were reflected in this systematic review.
*Risk of bias*
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Quote: "30 patients were randomly assigned to the three experimental groups". Comment: inadequate information regarding method of randomisation, therefore unable to make a judgement on appropriateness.
Allocation concealment (selection bias)	Unclear risk	Quote: "30 patients were randomly assigned to the three experimental groups". Comment: inadequate information regarding method of allocation concealment, therefore unable to make a judgement on appropriateness.
Blinding of participants and personnel (performance bias)   All outcomes	Unclear risk	Quote: "The investigational drug pharmacy at the institute dispensed the drugs so that the investigator would be blinded to the experimental group." Comment: described as double‐blind, but inadequate information regarding how blinding of participants was carried out, therefore unable to make a judgement on appropriateness.
Blinding of outcome assessment (detection bias)   All outcomes	Low risk	Quote: "The investigational drug pharmacy at the institute dispensed the drugs so that the investigator would be blinded to the experimental group." Comment: adequate method of blinding.
Incomplete outcome data (attrition bias)   All outcomes	Low risk	0/90 dropouts =0% attrition (100% completion)
Selective reporting (reporting bias)	Low risk	Data for outcomes of this review were reported appropriately.
Other bias	Low risk	No other source of bias detected.

Methods	Setting: San Raffaele Hospital, Madrid Design: parallel (2 arms) Number of centres: 1 Study duration: not reported
Participants	Inclusion criteria: male and female mobile patients; aged between 18‐40 years; fitted with a brace; moderate or severe pain in the mouth region measured using the VAS; at least 1 of the 2 signs of inflammation (oedema or hyperaemia) of a moderate intensity, or severe on a point scale from 0 to 3; written consent provided by the patient  Exclusion criteria: none specified Orthodontic intervention: initial archwire placement Patient sampling: n = 120 patients selected and randomized n = 0 dropouts/excluded from analysis n = 120 data analyzed for: Group 1 (n = 60) gender and age data not reported Group 2 (n = 60) gender and age data not reported
Interventions	Comparison: NSAID vs NSAID Ketoprofen (160 mg) vs benzidamine chloride (22.5 mg); mouthwash provided postoperatively Group 1: ketoprofen 10 ml in 100 ml of water twice a day (after breakfast and the evening meal) for up to 7 days Group 2: benzidamine chloride 15 ml twice a day (after breakfast and the evening meal) for up to 7 days
Outcomes	Primary outcome: pain score (VAS) ‐ recorded at baseline, on days 1, 2, 3, 4, 5, 6, and 7, 1 h after breakfast and 1 h after the evening meal Secondary outcome: the seriousness of the following signs – oedema and hyperaemia were measured on a graduated scale of 0‐3 (3 being intense pain). The pain was measured by the person conducting the experiment, on examination of the oral cavity when the brace was initially fitted, on the second visit and on the third and final visit. Resolution of any signs of inflammation was deducted from the data produced regarding seriousness, following the same marking procedure (complete remission of inflammation, a good improvement, slight improvement, no effect) 0 = a complete remission.
Notes	Conflict of interests/funding: not reported Adverse events/harm: not discussed Data handling by review authors: although multiple time points were measured, data were only presented for 4 days. Therefore, data from this study have not been used for this systematic review. No information is provided relating to dropouts, so we have assumed that all participants returned questionnaires and contributed to the final analysis. Other information of note: this paper was translated from Italian, with additional correspondence from the author.
*Risk of bias*
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Quote: "Randomised study in the single caesium for parallel groups". Comment: inadequate information regarding method of randomisation, therefore unable to make a judgement on appropriateness.
Allocation concealment (selection bias)	Unclear risk	Quote: "Randomised study in the single caesium for parallel groups". Comment: inadequate information regarding method of allocation concealment, therefore unable to make a judgement on appropriateness.
Blinding of participants and personnel (performance bias)   All outcomes	Unclear risk	Quote: "The study was carried out in 120 patients submitted to orthodontic therapy by oral route, under single blind conditions." Comment: described as single‐blind, but inadequate information regarding how blinding was carried out, therefore unable to make a judgement on appropriateness.
Blinding of outcome assessment (detection bias)   All outcomes	Unclear risk	Quote: "The study was carried out in 120 patients submitted to orthodontic therapy by oral route, under single blind conditions." Comment: described as single‐blind, but inadequate information regarding how blinding was carried out, therefore unable to make a judgement on appropriateness.
Incomplete outcome data (attrition bias)   All outcomes	Low risk	0/120 dropouts = 0% attrition (100% completion)
Selective reporting (reporting bias)	High risk	Data for outcomes at all time points recorded were not reported, only data for 4 days were available.
Other bias	High risk	There are no data presented in the paper concerning baseline characteristics of the groups.

Methods	Setting: University of Florida orthodontic clinic Design: parallel (3 arms) Number of centres: 1 Study duration: 37 months
Participants	Inclusion criteria: at least 13 and not older than 30 years of age; not pregnant; beginning orthodontic treatment for the first time; orthodontic treatment required the placement of at least 1 separator in each of the 4 quadrants; no contraindications or adverse reactions to ibuprofen or almonds; and written informed consent to participate.  Exclusion criteria: none specified Orthodontic intervention: separator placement Patient sampling: n = 51 enrolled n = 0 dropouts/excluded from analysis n = 51 data analyzed for: Group A (n = 16) 6 male:10 female, mean age 17.6 ± ‐5.0 years Group B (n = 17) 10 male:7 female, mean age 14.9 ± 2.7 years Group C (n = 18) 5 male:13 female, mean age 16.4 ± 3.6 years
Interventions	Comparison: NSAID vs placebo Ibuprofen (400 mg) vs placebo; provided pre‐ and post‐treatment to separator placement, or post‐treatment, or both Group A: ibuprofen 1 h before placement, 3 h after and 7 h after placement Group B: placebo 1 h before placement, ibuprofen 3 h and 7 h after placement Group C: placebo 1 h before placement, placebo 3 h and 7 h after placement
Outcomes	Pain score (VAS) ‐ recorded at pre‐treatment expectation of pain, 2 h, 6 h, 10 h,/bedtime, 17 h,/awakening, 24 h and 2, 3 and 7 days after separator placement Pain was recorded during the following activities: chewing biting (not an outcome of this review) fitting front teeth together (not an outcome of this review) fitting posterior teeth together (not an outcome of this review) Masticatory efficiency test (masticatory performance index (not an outcome of this review) Expectation of pain (VAS) (not an outcome of this review) Affective states (State and Trait Anxiety Inventory (STAI))
Notes	Conflict of interests/funding: none reported Adverse events/harm: not discussed Data handling by review authors: only pain during chewing data required for this systematic review, however it was unclear from the published data what the VAS measurements presented ‐ Table III labelled as 'chewing' but the Discussion stated that the data for chewing was not included. Mean VAS data for 24 h showed values of over 10 cm, despite a 10 cm VAS being used. Therefore, data for 24 h has been excluded for the purposes of this review. For the purposes of this review, data from Group B have not been used, data from Group A and C have been used for the comparison of NSAID versus placebo. Other information of note: pain during chewing data as presented in Table III is reflected in this systematic review.
*Risk of bias*
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Quote: "They were randomly assigned to 1 of 3 groups stratified by sex." Comment: inadequate information regarding method of randomisation, therefore unable to make a judgement on appropriateness.
Allocation concealment (selection bias)	Unclear risk	Quote: "They were randomly assigned to 1 of 3 groups stratified by sex." Comment: inadequate information regarding method of allocation, therefore unable to make a judgement on appropriateness.
Blinding of participants and personnel (performance bias)   All outcomes	Unclear risk	Quote: "Placebo‐controlled, double‐blind, parallel arm, prospective study". Comment: described as double‐blind, but inadequate information regarding how blinding of participants and personnel was ensured, therefore unable to make a judgement on appropriateness.
Blinding of outcome assessment (detection bias)   All outcomes	Unclear risk	Quote: "Placebo‐controlled, double‐blind, parallel arm, prospective study". Comment: described as double‐blind, but inadequate information regarding how blinding of outcome assessment was ensured, therefore unable to make a judgement on appropriateness.
Incomplete outcome data (attrition bias)   All outcomes	Low risk	0/51 dropouts = 0% attrition (100% completion)
Selective reporting (reporting bias)	High risk	Study protocol not available. Individual VAS scores for biting, chewing, fitting front teeth, and fitting back teeth were not recorded. 10 cm VAS used, but measurements show values of > 10 cm at 24 h for all 3 groups.
Other bias	Low risk	No other sources of bias detected.

Methods	Setting: Orthodontic Clinic of Dental School at Shiraz University of Medical Sciences, Iran Design: parallel (3 arms) Number of centres: 1 Study duration: not reported
Participants	Inclusion criteria: patients needing separator placement to begin orthodontic treatment in the maxillary arch; aged 15 years or older; were informed and signed the written informed consent; not using antibiotics, analgesics, anti‐inflammatory, anti‐coagulative, diuretics, oral anti diabetics, lithium, cyclosporine, or methotrexate; no need for antibiotic prophylaxis; no chronic systemic disease or clotting disorders; not reporting contraindication for NSAIDs; not pregnant or nursing  Exclusion criteria: none specified, but excluded participants who took additional analgesics Orthodontic intervention: separator placement Patient sampling: n = 349 assessed for eligibility (Group 1 = 107, Group 2 = 107, Group 3 = 107) n = 28 excluded (12 did not meet inclusion criteria, 16 decided not to participate) n = 321 enrolled and randomized n = 16 dropouts (Group 1 = 5, Group 2 = 7, Group 3 = 4 lost to follow‐up) n = 64 excluded from analysis (Group 1 = 26, Group 2 = 24, Group 3 = 14 did not complete questionnaire correctly (n = 46)/took additional analgesics (n = 18)) n = 241 data analyzed for: Group 1 (n = 76) 21 male:55 female, mean age 22.1 + 3.2 years Group 2 (n = 76) 19 male:57 female, mean age 21.7 + 3.5 years Group 3 (n = 89) 21 male:68 female, mean age 21.2 + 3.8 years
Interventions	Comparisons: NSAID vs NSAID, and NSAID vs paracetamol Ibuprofen (400 mg) vs paracetamol (650 mg) vs meloxicam (7.5 mg); provided pre‐emptively to separator placement Group 1: ibuprofen 1 h before separator placement Group 2: paracetamol 1 h before separator placement Group 3: meloxicam 1 h before separator placement
Outcomes	Pain score (VAS) ‐ recorded mmediately, and at 2 h, 6 h, 24 h and 48 h after separator placement Pain was recorded during the following activities: chewing rest (not an outcome of this review) fitting posterior teeth together(not an outcome of this review)
Notes	Conflict of interests/funding: this work was supported by the Vice‐Chancellery of Shiraz University of Medical Science (2168). The authors declare that they have no competing interests. Adverse events/harm: not reported Data handling by review authors: study does not allocate intervention to specific group labels. For the purposes of this systematic review, Group 1 has been allocated as the ibuprofen arm, Group 2 as the paracetamol arm and group 3 as the meloxicam arm. Although referred to as acetaminophen in the study, this group has been referred to as paracetamol for the purposes of this review. Other information of note: only pain during chewing data were reflected in this systematic review.
*Risk of bias*
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Quote: "The block randomization method was used with block length 9, and number of repetition for each group n = 3, to allocate subjects in each group. This method was used separately for each sex group to provide groups with equal numbers of male and female." Comment: randomisation appears to be adequate
Allocation concealment (selection bias)	Low risk	Quote: "The block randomization method was used with block length 9, and number of repetition for each group n = 3, to allocate subjects in each group. This method was used separately for each sex group to provide groups with equal numbers of male and female." Comment: inadequate information regarding method of allocation, therefore unable to make a judgement on appropriateness.
Blinding of participants and personnel (performance bias)   All outcomes	Low risk	Quote: "In each group, all tablets were covered by identical gelatin cover, so the investigators, the patients, and the statistician were all blind to the treatment groups." Comment: blinding appears to be adequate
Blinding of outcome assessment (detection bias)   All outcomes	Low risk	Quote: "In each group, all tablets were covered by identical gelatin cover, so the investigators, the patients, and the statistician were all blind to the treatment groups." Comment: blinding appears to be adequate
Incomplete outcome data (attrition bias)   All outcomes	High risk	80/321 = 25% dropouts (75% completion)
Selective reporting (reporting bias)	Low risk	Data for outcomes of this review were reported appropriately.
Other bias	Low risk	No other source of bias detected.

Methods	Setting: Ohio State University Orthodontic Clinic Design: parallel (arms) Number of centres: 1 Study duration: not reported
Participants	Inclusion criteria: not specified  Exclusion criteria: not specified Orthodontic intervention: separator placement, 2nd stage of initial archwire placement Patient sampling: n = ? selected n = ? excluded/refused to participate 84 patients randomized; 28 per group. 34 female:43 male. n = 7 dropouts/excluded from analysis; Group 1 = 5, Group 2 = 0, Group 3 = 2. n = 77 separators and 56 archwires data analyzed for: Group A separators (n = 23) 15 male:8 female, mean age 16.1 ± 5.5 years Group A archwire (n = 17) 11 male:6 female, mean age data not available Group B separators (n = 28) 16 male:12 female, mean age 15.4 ± 6.9 years Group B archwire (n = 17) 10 male:7 female, mean age data not available Group C separators (n = 26) 12 male:14 female, mean age 18.1 ± 7.3 years Group C archwire (n = 22) 11 male:11 female, mean age data not available
Interventions	Comparisons: NSAID vs placebo, NSAID vs aspirin, and aspirin vs placebo Ibuprofen (400 mg) vs aspirin (650 mg) vs placebo(beta‐lactose); provided immediately after separator and initial archwire placement Group A: ibuprofen immediately after placement Group B: aspirin immediately after placement Group C: placebo immediately after placement
Outcomes	Pain score VAS) ‐ recorded at 2 h, 6 h, 24 h and 2, 3 and 7 days after separator or initial archwire placement Pain was recorded during the following activities: chewing biting (not an outcome of this review) fitting front teeth together (not an outcome of this review) fitting posterior teeth together(not an outcome of this review)
Notes	Conflict of interests/funding: none reported Adverse events/harm: not discussed Data handling by review authors: VAS data were recorded for chewing, biting, fitting or putting front teeth together, and fitting on the back teeth. However data were combined for analysis in the study and these have been used for this review, as the individual data were not available. Other information of note: different initial archwires and appliances were used, "11 had upper and lower arch Begg fixed appliances fitted with 0.016‐inch base arch wires. In the remaining 45 patients, 0.022‐inch edgewise brackets were bonded to the upper and lower arches, and all brackets fully engaged with an initial 0.0175‐inch Response archwire."
*Risk of bias*
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Quote: "Randomly divided into three groups". Comment: inadequate information regarding how randomisation was ensured, therefore unable to make a judgement on appropriateness.
Allocation concealment (selection bias)	Unclear risk	Comment: inadequate information regarding method of allocation, therefore unable to make a judgement on appropriateness.
Blinding of participants and personnel (performance bias)   All outcomes	Low risk	Quote: "The capsules appeared identical and were coded by number with the code kept by the pharmaceutical company who encapsulated the placebo and medications. The code was not revealed until the study was completed." Comment: blinding appears to be adequate
Blinding of outcome assessment (detection bias)   All outcomes	Low risk	Quote: "The capsules appeared identical and were coded by number with the code kept by the pharmaceutical company who encapsulated the placebo and medications. The code was not revealed until the study was completed." Comment: blinding appears to be adequate
Incomplete outcome data (attrition bias)   All outcomes	Low risk	7/84 = 8% dropouts (92% completion)
Selective reporting (reporting bias)	High risk	Data for outcomes of this study were not reported appropriately. We were unable to extract data for use in pooled analysis.
Other bias	Low risk	No other sources of bias detected.

Methods	Setting: Dental faculty of Tehran University of Medical Sciences Design: parallel (3 arms) Number of centres: 1 Study duration: not reported
Participants	Inclusion criteria: participants started orthodontic treatment that required separators; had no systemic or GI diseases, not taking analgesics or any other drugs; no contraindication to the use of either acetaminophen (paracetamol) or liquefied ibuprofen; weight above 40 kg; first molar was without decay, filling or periodontal problem (this criterion was checked through clinical observation, probing and panoramic radiographs)  Exclusion criteria: none specified, however, later stated they would exclude people who had taken additional analgesics Orthodontic intervention: separator placement Participant sampling: n = 101 randomized n = 12 dropouts/excluded from analysis (did not take drugs correctly = 8; did not complete questionnaire = 3) n = 89 data analyzed for: Group 1 (n = 29) 13 male:16 female, mean age 15.6 ± 4.17 years Group 2 (n = 32) 14 male:18 female, mean age 15.8 ± 3.49 years Group 3 (n = 28) 12 male:16 female, mean age 15.3 ±3.15 years
Interventions	Comparisons: NSAID vs placebo, paracetamol vs placebo, and NSAID vs paracetamol Ibuprofen (400 mg) vs paracetamol (650 mg) vs placebo; provided pre‐emptively to separator placement Group 1: ibuprofen 1 h before separator placement and every 6 h until 24 h (5 doses) Group 2: paracetamol 1h before separator placement and every 6 h until 24 h (5 doses) Group 3: placebo 1h before separator placement and every 6 h until 24 h (5 doses)
Outcomes	Pain score (VAS) ‐ recorded immediately, and at 2 h, 6 h, bedtime and 24 h after separator placement
Notes	Conflict of interests/funding: not reported Adverse events/harm: not reported Data handling by review authors: study did not allocate intervention to specific group labels. For the purposes of this systematic review, Group 1 has been allocated as the ibuprofen arm, Group 2 as the paracetamol arm and Group 3 as the placebo arm. Although referred to as 'acetaminophen' in the study, we have referred to this group as paracetamol for the purposes of this review. Gender data presented as a percentage, calculated as values for the purposes of this review. Other information of note: no baseline information provided about groups before drop out. Mean age for Group 2 differed throughout study. For the purposes of the review, the data from Table 1 have been used.
*Risk of bias*
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Quote: "To divide the patients into three groups, block randomization method was used. Each block contained three coded pockets (acetaminophen, liquefied ibuprofen, and placebo) and consisted of one sex (male or female)." Comment: inadequate information regarding how randomisation was carried out, therefore unable to make a judgement on appropriateness
Allocation concealment (selection bias)	Low risk	Quote: "The random allocation and coding of drugs was performed by an operator outside the study and was concealed in an envelope." Comment: allocation concealment appears to be adequate
Blinding of participants and personnel (performance bias)   All outcomes	Low risk	Quote: "To ensure that the patients were blind to the experimental group, the analgesics and placebo were placed in identical capsules"; "In each group, the male to female ratio was equal, and the patient and the operator were blind of the kind of drug." Comment: blinding appears to be adequate
Blinding of outcome assessment (detection bias)   All outcomes	Low risk	Quote: "The patients were asked to put each questionnaire in a pocket and seal it after marking the scale"; "randomized triple blinded clinical trial". Comment: blinding appears to be adequate
Incomplete outcome data (attrition bias)   All outcomes	Low risk	12/101 = 12% dropouts (88% completion)
Selective reporting (reporting bias)	Low risk	Data for outcomes of this review were reported appropriately.
Other bias	Low risk	No other sources of bias detected.

Methods	Setting: orthodontic consultation and dental treatment unit, Ibn Rochd Hospital Center, Casablanca, Morocco Design: parallel (2 arms) Number of centres: 1 Study duration: not reported
Participants	Inclusion criteria: no drug treatment during the study; good oral hygiene; good general health; adults stratified by age group with presenting malocclusion requiring orthodontic treatment; consent provided  Exclusion criteria: contra‐indication to the use of paracetamol or ibuprofen; taking medication including short‐term anti‐inflammatory analgesics or long‐term corticosteroids; dropouts; non‐compliance Orthodontic intervention: initial archwire placement Participant sampling: n = 56 randomized and analyzed: Group A (n = 27) Group B (n = 29) Overall: 17 male:39 female; age < 15 years = 21.4%: aged > 15 years = 78.6%
Interventions	Comparison: NSAID vs paracetamol Ibuprofen (600 mg; 2 x 300 mg/day for 5 days) vs paracetamol (2 g; 4 x 500 mg/day for 7 days) Group A: ibuprofen provided immediately post‐treatment, daily oral dose thereafter Group B: paracetamol provided immediately post‐treatment, daily oral dose thereafter
Outcomes	Pain score (VAS) ‐ recorded at 2 h, 6 h, 24 h and 2, 3 and 7 days after placement of initial archwire
Notes	Conflict of interests/funding: not reported Adverse events/harm: not reported Data handling by review authors: additional information received through correspondence with the author. Other information of note: original paper translated from French to English.
*Risk of bias*
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Quote: "The randomisation was carried out by a computer algorithm; random block was performed by the software". Comment: computer‐generated block randomisation carried out, therefore appears to be adequate
Allocation concealment (selection bias)	Low risk	Quote: "The distribution of study subjects was determined by a biostatistician who gave us the list of participants to the study". Comment: allocation concealment appears to be adequate
Blinding of participants and personnel (performance bias)   All outcomes	High risk	Quote: "Double‐blind was impossible because the patient could read the tablet trade name". Comment: blinding of the drugs was not carried out, therefore it was assumed that the researcher supplying the intervention and the participants were not blinded.
Blinding of outcome assessment (detection bias)   All outcomes	Low risk	Quote: "Drug distribution was done with a single‐blind method, single investigators were unaware of the drug". Comment: blinding appears to be adequate
Incomplete outcome data (attrition bias)   All outcomes	Low risk	0/56 = 0% dropout (100% completion)
Selective reporting (reporting bias)	Low risk	Data for outcomes of this review were reported appropriately.
Other bias	Unclear risk	Baseline information on gender and age not given by group.

Methods	Setting: Orthodontic Clinic, Dental School, University of Brescia, Italy Design: parallel (3 arms) Number of centres: 1 Study duration: not reported
Participants	Inclusion criteria: healthy; 12 to 16 years of age; oral membrane lesions from 2 mm‐6 mm caused by wearing a fixed brace; available to participate in the study  Exclusion criteria: already treated for orthodontic or systemic pain during the last month; syndromes or mental retardation; unavailable to participate to the study; not suffering from anxiety according to parents’ rating; no history of dental treatment refusal Orthodontic intervention: within 1 month of having a fixed or removable appliance fitted Participant sampling: n = 60 selected, randomized and analyzed: (30 male:30 female; < 14 years n = 30; > 14 years n = 30) Group 1 (n = 20) 10 male:10 female, mean age 14 + 2 years Group 2 (n = 20) 10 male:10 female, mean age 14 + 2 years Group 3 (n = 20) 10 male:10 female, mean age 14 + 2 years
Interventions	Comparison: NSAID vs placebo Flurbiprofen (10 ml 0.25% mouthwash 3 times daily for 7 days) vs placebo (10 ml mouthwash 2‐minute rinse duration 3 times daily for 7 days) vs control Group 1: flurbiprofen; provided postoperatively after separator placement Group 2: placebo; provided postoperatively after separator placement Group 3: control, no treatment
Outcomes	Pain score (VAS) ‐ recorded at baseline, 3 and 7 days after separator placement
Notes	Conflict of interests/funding: not reported Adverse events/harm: "2 cases of reduced taste sensation with Flurbiprofen which did not cause discontinuation"; "No local or systemic ADRs were reported". Data handling by review authors: additional information received through correspondence with the author. Although multiple time points measured, due to variations in the time points of interest to this review, data from this study have not been used for this systematic review. Other information of note: original paper translated from Italian to English. All participants had pre‐existing ulceration and had started orthodontic treatment within the last month.
*Risk of bias*
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Quote: "Separated randomization lists furnished by the statistical department of our university, conceived with a variable block size of 3, 6 and 9, and stratified for lesion type (vestibular ulcers, lower labial fraenum lesions, keratinized mucosa lesions, aphthous ulcers, decubitus ulcers), age (more than 14 and less than 14 years) and gender (5x2x2=20 strata), in order to obtain homogeneous and comparable group." Comment: randomisation appears to be adequate
Allocation concealment (selection bias)	Low risk	Quote: "Allocation concealment was obtained identifying patients with a progressive numeration from 1 to 60, after a casual names draw, and groups with a letter from A to C, assigned by lot." Comment: allocation concealment appears to be adequate
Blinding of participants and personnel (performance bias)   All outcomes	High risk	Quote: "Single blinding: patients were not aware of the treatment received". Comment: adequate blinding of participants, however no blinding of personnel.
Blinding of outcome assessment (detection bias)   All outcomes	Low risk	Quote: "The statistician who analyzed outcome data were not blind regarding study aims but he was blind regarding treatment assigned to every single patient." Comment: blinding appears to be adequate
Incomplete outcome data (attrition bias)   All outcomes	Low risk	0/50 dropouts = 0% attrition (100% completion)
Selective reporting (reporting bias)	Low risk	Data for outcomes of this review were reported appropriately.
Other bias	Low risk	No other sources of bias detected.

Methods	Setting: Institution of IASP Education ‐ Institute for Support in Dentistry, Londrina, Parana, Brazil Design: parallel RCT (6 arms) Number of centres: 1 Study duration: not reported
Participants	Inclusion criteria: aged between 18 and 28 years; classified as healthy during the clinical history; able to provide written consent; not under prophylactic antibiotic coverage; absence of systemic disease with the use of drugs; no antibiotics or analgesics for a minimum of 15 days prior to the study; no contraindication for the use of the proposed study drugs; no extraction of teeth at least 2 weeks prior to installation of fixed orthodontic appliance; Index of Orthodontics Need (IOTN) Dental Health Component score 3 or above Exclusion criteria: history of hypersensitivity to the study drug (normal or idiosyncratic reaction to the drug); evidence of organ dysfunction or clinically significant deviation from normal; history of any psychiatric illness that might compromise the ability to provide written consent; history of GI, liver disease, renal, cardiovascular, pulmonary, neurologic or haematologic, diabetes, or glaucoma; consumption of more than 20 cigarettes per day or difficulty in refraining from smoking during the study period; history of drug or alcohol abuse; pregnant or lactating; participated in any similar clinical study during 6 weeks preceding the study. Participants were also excluded if they suffered an adverse reaction or allergic reaction clearly related to the drugs; were diagnosed with a systemic disease unrelated to the drugs during the study, that would require concomitant therapy; or did not fulfil the requirements of the Protocol, including rules related to the use of drugs, alcohol or lack of co‐operation during the study. Orthodontic intervention: initial archwire placement Patient sampling: n = 180 recruited and randomized n = 0 excluded n = 180 analyzed Group 1 (n = 30) 15 male:15 female, age data not reported Group 2 (n = 30) 15 male:15 female, age data not reported Group 3 (n = 30) 15 male:15 female, age data not reported Group 4 (n = 30) 15 male:15 female, age data not reported Group 5 (n = 30) 15 male:15 female, age data not reported Group 6 (n = 30) 15 male:15 female, age data not reported
Interventions	Comparisons: NSAID vs control (lactose placebo), and NSAID vs NSAID Group 1: diclofenac sodium 50 mg 1 h before placement Group 2: lumiracoxib 400 mg 1 h before placement Group 3: dexamethasone 4 mg 1 before placement Group 4: nimesulide 100 mg 1 h before placement Group 5: ibuprofen 400 mg 1 h before placement Group 6: placebo 1 h before placement
Outcomes	Pain score (VAS) ‐ recorded at 2 h, 12 h, 24 h and 2, and 7 days after initial archwire placement Pain was recorded during the following activities: chewing biting (not an outcome of this review) fitting back teeth together (not an outcome of this review) fitting front teeth together (not an outcome of this review)
Notes	Conflict of interests/funding: no source of funding reported Adverse events/harm: no harms reported Data handling by review authors: this paper was translated from Portugese. The data presented for the analysis are based on Figure 1 which shows mean pain scores (mean + SEM). The SEM was used to calculate SD, but we were unable to calculate SD for lumiracoxib at 24 h from Figure 1. Data from Group 2 and Group 6 have been inverted to reflect lumiracoxib as an intervention and placebo as its control. Data from Group 3 did not contribute to the analyses. Other information of note: “This study was conducted in accordance with the guidelines established by Resolution 196/96 of the National Council of Health Ministry of Health published on October 10, 1996 and the Code of Professional Ethics Dental, according to Resolution 179/93 CFO ... This study was approved by the Ethics Committee of the Faculty of Dentistry and Dental Research Center São Leopoldo Mandic under protocol 06/333”.
*Risk of bias*
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Quote: "Patients were randomly assigned to 1 of 3 experimental groups"; "The randomization of which of the three experimental conditions the patients were assigned to was computer generated". Comment: described as randomized, but inadequate information regarding method or randomisation to make a judgement on appropriateness.
Allocation concealment (selection bias)	Unclear risk	Comment: inadequate information regarding method of allocation concealment
Blinding of participants and personnel (performance bias)   All outcomes	Unclear risk	Quote: "The drugs were divided into different numeric codes". Comment: inadequate information regarding method of blinding
Blinding of outcome assessment (detection bias)   All outcomes	Unclear risk	Quote: "The study was double‐blind." Comment: inadequate information regarding method of blinding
Incomplete outcome data (attrition bias)   All outcomes	Low risk	0/180 dropouts = 0% attrition
Selective reporting (reporting bias)	Low risk	Data for outcomes of this review were reported appropriately.
Other bias	Unclear risk	No data presented regarding age of participants in groups which could have resulted in sampling bias.