Single dose intra‐articular morphine for pain control after knee arthroscopy

Zui Zou; Mao Mao An; Qun Xie; Xiaoyan Y Chen; Hao Zhang; Guan J Liu; Xue Y Shi

doi:10.1002/14651858.CD008918.pub2

. 2016 May 3;2016(5):CD008918. doi: 10.1002/14651858.CD008918.pub2

Single dose intra‐articular morphine for pain control after knee arthroscopy

Zui Zou ¹, Mao Mao An ², Qun Xie ³, Xiaoyan Y Chen ⁴, Hao Zhang ⁵, Guan J Liu ⁶, Xue Y Shi ^1,^7,^✉

Editor: Cochrane Pain, Palliative and Supportive Care Group

PMCID: PMC6517216 PMID: 27140500

Abstract

Background

Knee arthroscopy is a common procedure and is associated with postoperative pain. Intra‐articular (IA) injection of morphine for pain control has been widely studied, but its analgesic effect after knee arthroscopy is uncertain.

Objectives

To evaluate the relative effects on pain relief and adverse events of IA morphine given for pain control after knee arthroscopy compared with placebo, other analgesics (local anaesthetics, non‐steroidal anti‐inflammatory drugs (NSAIDs), other opioids) and other routes of morphine administration.

Search methods

We searched CENTRAL (The Cochrane Library Issue 4, 2015), MEDLINE via Ovid (January 1966 to May 2015), EMBASE via Ovid (January 1988 to May 2015), and the reference lists of included articles. We also searched the metaRegister of controlled trials, clinicaltrials.gov and the World Health Organization (WHO) International Clinical Trials Registry Platform for ongoing trials.

Selection criteria

We identified all the randomised, double‐blind controlled trials that compared single dose IA morphine with other interventions for the treatment of postoperative pain after knee arthroscopy. We excluded studies with fewer than 10 participants in each group, using spinal or epidural anaesthesia, or assessing the analgesic effect of IA morphine on chronic pain.

Data collection and analysis

Two authors independently assessed the quality of each trial and extracted information on pain intensity, supplementary analgesics consumption and adverse events. We assessed the evidence using GRADE (Grading of Recommendations Assessment, Development and Evaluation) and created 'Summary of findings' tables.

Main results

We included 28 small, low quality studies (29 reports) involving 2564 participants. Of 20 studies (21 reports) comparing morphine with placebo, nine studies with adequate data were included in the meta‐analysis. Overall, the risk of bias was unclear. Overall, the quality of the evidence assessed using GRADE was low to very low, downgraded primarily due to risk of bias, small study size, and imprecision.

No statistical difference was found between 1 mg IA morphine and placebo in pain intensity (visual analogue scale (VAS)) at early phase (zero to two hours) (mean difference (MD) ‐0.50, 95% CI ‐1.15 to 0.14; participants = 297; studies = 7; low quality evidence), medium phase (two to six hours) (MD ‐0.47, 95% CI ‐1.09 to 0.14; participants = 297; studies = 7; low quality evidence) and late phase (six to 30 hours) (MD ‐0.88, 95% CI ‐1.81 to 0.04; participants = 297; studies = 7; low quality evidence). No significant difference was found between 1 mg and 2 mg morphine for pain intensity at early phase (MD ‐0.56, 95% CI ‐1.93 to 0.81; participants = 105; studies = 2; low quality evidence), while 4 mg/5 mg morphine provided better analgesia than 1 mg morphine at late phase (MD 0.67, 95% CI 0.08 to 1.25; participants = 97; studies = 3; low quality evidence). IA morphine was not better than local anaesthetic agents at early phase (MD 1.43, 95% CI 0.49 to 2.37; participants = 248; studies = 5; low quality evidence), NSAIDs at early phase (MD 0.95, 95% CI ‐0.95 to 2.85; participants = 80; studies = 2; very low quality evidence), sufentanil, fentanyl or pethidine for pain intensity. IA morphine was similar to intramuscular (IM) morphine for pain intensity at early phase (MD 0.21, 95% CI ‐0.48 to 0.90; participants = 72; studies = 2; very low quality evidence).

Meta‐analysis indicated that there was no difference between IA morphine and placebo or bupivacaine in time to first analgesic request. Eleven out of 20 studies comparing morphine with placebo reported adverse events and no statistical difference was obtained regarding the incidence of adverse events (risk ratio (RR) 1.09, 95% CI 0.51 to 2.36; participants = 314; studies = 8; low quality evidence). Seven of 28 studies reported participants' withdrawal. There were not enough data for withdrawals to be able to perform meta‐analysis.

Authors' conclusions

We have not found high quality evidence that 1 mg IA morphine is better than placebo at reducing pain intensity at early, medium or late phases. No statistical difference was reported between IA morphine and placebo regarding the incidence of adverse events. The relative effects of 1 mg morphine when compared with IA bupivacaine, NSAIDs, sufentanil, fentanyl and pethidine are uncertain. The quality of the evidence is limited by high risk of bias and small size of the included studies, which might bias the results. More high quality studies are needed to get more conclusive results.

Plain language summary

Morphine injections for pain relief after knee arthroscopy

Background

Knee arthroscopy is a surgical procedure on the knee. The surgery is minimally invasive, which means that only a small cut (incision) is needed. An examination, and sometimes treatment, of damage is performed using an arthroscope, which is inserted into the joint through the small incision. Knee arthroscopy is used to assess or treat many orthopaedic (musculoskeletal) conditions, and patients may have pain after surgery. Morphine injected directly into the knee (intra‐articular morphine) to relieve pain has been widely studied, but we do not know how well it works.

Key results and quality of the evidence

In May 2015, this review identified 28 small, low quality studies involving 2564 participants looking at intra‐articular morphine for pain relief after knee arthroscopy. From 9/20 studies we did not find evidence that intra‐articular morphine given at a dose of 1 mg was better than placebo for pain relief. From the limited evidence available we were unable to determine how intra‐articular morphine compared with morphine injected into the muscle (intra‐muscular morphine). There was also low quality evidence for the effects of 1 mg intra‐articular morphine compared with intra‐articular bupivacaine, non‐steroidal anti‐inflammatory drugs (NSAIDs), sufentanil, fentanyl and pethidine, so we were unsure which worked best. We were unable to determine how similar the rates of side effects such as nausea and vomiting were between intra‐articular morphine and placebo. Overall, the quality of the evidence was low.

Future research should focus on finding effective analgesics for knee arthroscopy.

Summary of findings

Background

Description of the condition

Knee arthroscopy is a very common procedure performed as day surgery and is associated with postoperative pain (Joshi 1992). The incidence of moderate or severe pain within an hour after knee arthroscopy is around 60%. Baseline scores are about 50 mm on the visual analogue scale (VAS) 10 minutes after test drug injection. However, pain scores quickly decreased to about 35 mm by 30 minutes and less than 5 mm by 12 hours (Solheim 2006). Women report more pain after knee arthroscopy than men (Rosseland 2004b). Adequate postoperative analgesia may accelerate the patient's return to normal activity.

Description of the intervention

Many trials have tried to find an ideal regimen which provides sufficient postoperative analgesia with fewer adverse events, including intra‐articular (IA) anaesthetics and analgesics for postoperative pain relief. The spinal or intravenous (IV) administration of morphine may cause side effects such as respiratory depression, sedation, dependence, pruritus and urinary retention. However, the adverse events of peripheral IA morphine administration were mild or absent. Morphine may therefore be a promising IA agent that is without obvious central side effects (Sawynok 2003). IA injection of morphine has been widely studied for its simplicity, safety and efficacy.

Opioid binding sites have been identified within synovial tissue, implying that the analgesic effect of morphine may be locally mediated (Khoury 1992). When given at the end of arthroscopic surgery, IA morphine could reduce postoperative pain through peripheral opioid receptors (Stein 1991). It has also been reported to reduce pain through other pathways (Kalso 1997) such as inflammatory reaction (Likar 1998; Stein 1995; Stein 1999). However, different opinions exist as to the postoperative effect of peripheral opioids. The reported dosages of IA morphine in studies vary from 0.5 mg to 5 mg (Joshi 1993). Stein 1991 reported that a dose of 1 mg morphine showed analgesic efficacy, whereas a dose of 0.5 mg did not. Allen 1993 reported that 2 mg morphine showed a better analgesic effect than 1 mg, but no dose response was detected in two other studies (Heine 1994; Laurent 1994). Other controlled trials, however, have failed to show any analgesic effect of morphine compared with placebo (Aasbø 1996; Drosos 2002; Gupta 1999; Ruwe 1995; Soderlund 1997).

In addition to opiates, many other interventions have been widely studied for the reduction of postoperative pain following knee arthroscopy. Local anaesthetics (such as bupivacaine, ropivacaine, carbocaine, lidocaine and prilocaine), non‐steroidal anti‐inflammatory drugs ((NSAIDs) such as ketorolac and tenoxicam) and other interventions (such as magnesium, clonidine, neostigmine and ketamine) have been intensely studied.

How the intervention might work

Opioid receptors exist on peripheral terminals of primary afferent neurons, demonstrated functionally and morphologically in Bartho 1990 and Stein 1990. As a μ‐opioid receptor agonist, morphine is effective in producing peripheral analgesia. Morphine activates peripheral opioid receptors, resulting in their interactions with G‐proteins and a decrease in cyclic adenosine monophosphate (cAMP) in sensory nerve terminals, an increase in K⁺ efflux and a decrease of Ca²⁺ entry. Thus, the excitability of the peripheral nerve terminal, the propagation of action potentials and release of neuropeptides are attenuated (Sawynok 2003).

Why it is important to do this review

Several systematic reviews have investigated IA morphine for the control of pain after knee surgery, however, consensus on the analgesic effect is still lacking. A previously published qualitative systematic review (Kalso 1997) showed that IA morphine was effective in reducing postoperative pain intensity and the consumption of rescue analgesics. All knee surgeries were included in the review, however, and it failed to focus on the effect of IA morphine on pain relief after knee arthroscopy. Also, no quantitative analysis of pooled data was performed here, nor in the authors' second systematic review on this topic (Kalso 2002). In another systematic review, variability was found not only between studies but also within one study (patient variability) (Gupta 2001). By calculating the weighted mean difference (WMD) of pain scores between treatment groups, Moiniche 1999 found that there was weak evidence for a reduction of postoperative pain after IA instillation of local anaesthetics. More trials comparing morphine and other interventions for knee arthroscopy were available recently and they added new knowledge to inform clinical practice (Rosseland 2003; Rosseland 2004a; Rosseland 2004b). Evidence is still lacking as to whether IA opioids offer clinically relevant pain relief. In light of the existing controversy, this systematic review aimed to investigate the analgesic effect of single dose IA morphine compared with other interventions in the management of pain control after knee arthroscopy.

Objectives

Methods

Criteria for considering studies for this review

Types of studies

We included studies fulfilling the following selection criteria in this systematic review:

randomised, double‐blind controlled trials;
studies comparing single dose IA morphine with other interventions for the treatment of postoperative pain after knee arthroscopy; and
studies with more than 10 participants in each group.

We excluded the following studies:

studies in which spinal or epidural anaesthesia was used; and
studies whose primary aim was to assess the analgesic effect of IA morphine on chronic pain.

Types of participants

We included participants of either gender, aged 15 years or older, and undergoing knee arthroscopy.

Types of interventions

IA morphine versus any other interventions or administration methods:

IA morphine versus placebo;
IA morphine versus local anaesthetic agents (such as bupivacaine, ropivacaine, carbocaine, lidocaine and prilocaine);
IA morphine versus NSAIDs (such as ketorolac and tenoxicam);
different doses of IA morphine;
IA morphine versus intravenous (IV) or intra‐muscular (IM) morphine;
IA morphine versus other opioids.

Types of outcome measures

We included the following outcomes.

Primary outcomes

Patient‐reported postoperative pain intensity (a 0 ‐ 10 cm VAS score).
Use of supplementary analgesic (number of participants using rescue analgesics, time to first rescue analgesics, analgesic drug counts, patient‐controlled analgesia opioid consumption, etc.).

Secondary outcomes

Adverse events.
Withdrawals.

Search methods for identification of studies

Electronic searches

We searched CENTRAL (The Cochrane Library, Issue 4, 2015), MEDLINE via Ovid (January 1966 to May 2015), EMBASE via Ovid (January 1988 to May 2015). Please see Appendix 1; Appendix 2; Appendix 3 for the database search strategies. Filters designed to limit the searches to RCTs were added to the MEDLINE and EMBASE strategies.

Searching other resources

Reference lists

We sought additional studies from the references of retrieved randomised trials, meta‐analyses and systematic reviews.

Unpublished studies

We searched trials registries for ongoing trials. We searched three web sites: the metaRegister of controlled trials (www.controlled-trials.com/mrct), clinicaltrials.gov (www.clinicaltrials.gov) and the World Health Organization (WHO) International Clinical Trials Registry Platform (ICTRP) (apps.who.int/trialsearch/) up to May 2015.

Language

The search identified all relevant studies irrespective of language. We assessed non‐English language papers and translated as necessary.

Data collection and analysis

Selection of studies

Two authors independently assessed the eligibility of studies included in the review, and another author checked these results. For the studies that were reported several times, we used the first edition and added data from the secondary references to the first study to extract full data. We resolved any disagreement by discussion.

Data extraction and management

Two authors independently extracted data from each identified study and recorded data on a standardised data extraction form. We resolved any disagreement by discussion with a third author when necessary.

Assessment of risk of bias in included studies

Two authors assessed risk of bias for each study independently, based on the methods used to generate allocation sequence, allocation concealment, blinding, follow‐up, selective reporting and group size according to the 'Risk of bias' tool (Kjaergard 2001; Moher 1998; Schulz 1995; Higgins 2011a).

We assessed the methods used to deal with incomplete data as: low risk of bias (< 10% of participants did not complete the study or used 'baseline observation carried forward' analysis, or both); unclear risk of bias (used 'last observation carried forward' analysis); and high risk of bias (used 'completer' analysis).

"Size" was added to the 'Risk of bias' table. We assessed studies as low risk of bias (≧ 200 participants per treatment arm); unclear risk of bias (50 to 199 participants per treatment arm); and high risk of bias (< 50 participants per treatment arm).

We resolved any disagreement by discussion.

Dealing with missing data

We did not contact authors for missing data. We followed intention‐to‐treat principles. In some studies, the exact mean, standard deviation (SD) and standard error (SE) of pain scores were not reported or were difficult to decipher when results were presented in figures. In this situation, two review authors independently estimated the visual analogue scale (VAS) score presented in the figures in each study using Engauge Digitizer 4.1 (Lan 2010; Ma 2012) and achieved an agreement on the mean ± SE or SD (Gupta 2001). When the number of participants enrolled and the number of participants who reported outcomes were different, we noted this in the 'Characteristics of included studies' table. If the exclusions of participants was justifiable, we would use available data from the studies; if withdrawal of participants were not justifiable, we would carry out intention to treat (ITT) analysis.

Assessment of heterogeneity

We tested the heterogeneity between studies using the Chi² test (with P < 0.1 indicating significant heterogeneity) and the I² statistic, which described the proportion of variability due to heterogeneity (Higgins 2003). When P > 0.1, we carried out the meta‐analysis using a fixed‐effect model; otherwise we used a random‐effects model.

Data synthesis

Quantitative analysis

We performed a statistical analysis of outcomes. We merged both dichotomous and continuous data quantitatively in meta‐analysis.

Dichotomous data

For dichotomous data, such as the number of participants who suffered from adverse events, we calculated the risk ratio (RR) using Review Manager software (RevMan) 5.3 (RevMan 2014).

Continuous data

For continuous data, such as postoperative VAS score, analgesic duration of intervention drugs, we calculated mean differences (MDs). We applied a random‐effects or fixed‐effect model to assess outcomes, depending on the statistical heterogeneity among studies.

Qualitative analysis

Different analgesics were used to relieve postoperative pain and many studies did not report the exact doses consumed. Consequently, statistical analyses were not always possible. In situations where the data extracted from the original studies were insufficiently similar, we did not conduct a meta‐analysis but produced a qualitative description of the outcome study by study.

Quality of the evidence

Two review authors independently rated the quality of the outcomes pain intensity, analgesia duration and adverse events. We used the GRADE (Grades of Recommendation, Assessment, Development and Evaluation) system to rank the quality of the evidence using the GRADEprofiler Guideline Development Tool software (GRADEPro GDT 2015), and the guidelines provided in Chapter 12.2 of the CochraneHandbook for Systematic Reviews of Interventions (Higgins 2011b).

The GRADE approach uses five considerations (study limitations, consistency of effect, imprecision, indirectness and publication bias) to assess the quality of the body of evidence for each outcome. The GRADE system uses the following criteria for assigning grade of evidence:

High = further research is very unlikely to change our confidence in the estimate of effect;
Moderate = further research is likely to have an important impact on our confidence in the estimate of effect and may change the estimate;
Low = further research is very likely to have an important impact on our confidence in the estimate of effect and is likely to change the estimate;
Very low = any estimate of effect is very uncertain.

We decreased grade if:

Serious (‐1) or very serious (‐2) limitation to study quality;
Important inconsistency (‐1);
Some (‐1) or major (‐2) uncertainty about directness;
Imprecise or sparse data (‐1);
High probability of reporting bias (‐1).

'Summary of findings' table

We included 'Summary of findings' tables to present the main findings in a transparent and simple tabular format. In particular, we included key information concerning the quality of evidence, the magnitude of effect of the interventions examined, and the sum of available data on the outcomes listed above.

Subgroup analysis and investigation of heterogeneity

We carried out subgroup analysis to assess clinical heterogeneity rather than statistical heterogeneity. We carried out separate analyses for the most important clinical parameters as follows.

We identified three phases to assess postoperative pain: early phase (zero to two hours), medium phase (two to six hours) and late phase (six to 30 hours). We analysed the data from these different phases separately. Data of two hours and six hours were allocated to early phase and medium phase, respectively.
Comparisons: different agents were used as the comparator regimens in various studies. We only quantitatively merged the studies with the same drug categories in meta‐analysis.

Sensitivity analysis

We performed sensitivity analysis to evaluate the effect of methodological characteristics (quality assessment) of studies on the results of the meta‐analysis.

Results

Description of studies

See: Characteristics of included studies; Characteristics of excluded studies.

Results of the search

We identified 724 studies through the initial electronic search. We scanned these studies and found 35 studies which could not be excluded by scrutiny of the title and abstract only. Of these, seven were further excluded for reasons cited in the Characteristics of excluded studies table. Twenty eight studies (29 reports) satisfied the inclusion criteria comparing morphine with placebo or other analgesics (Figure 1). Studies that met the inclusion criteria contained 2482 enrolled participants, with trial size varying from 33 to 320 participants.

Flowchart showing the stepwise screening of search results

Included studies

Twenty studies (21 reports) enrolled participants undergoing elective arthroscopic knee surgery of multiple procedures such as diagnostic arthroscopy, meniscectomy, excision of plicae, full arthroscopic lateral retinacular release, synovectomy and chondral debridement (Aasbø 1996; Akinci 2003; Alagol 2005; Allen 1993; Bjornsson 1994; Christensen 1996; Follak 2001; Franceschi 2001; Kazemi 2004; Khoury 1992; Kizilkaya 2005; Likar 1995; Likar 1999; Muller 2001; Raj 2004; Richardson 1997; Rosseland 2003a; Ruwe 1995; Solheim 2006; Varkel 1999; Wrench 1996). Six studies (Calmet 2004; De Andres 1998; Dierking 1994; Elkousy 2013; Kanbak 1997; Lyons 1995) only enrolled participants undergoing arthroscopic meniscectomy and one study (Guler 2002) only enrolled participants undergoing arthroscopic anterior cruciate ligament (ACL) reconstruction with hamstring tendons.

Some studies had both placebo and active drug as control, and some had more than one active comparator. Twenty studies included a placebo control, whereas 23 used an active drug in the control group. The active drugs relevant to our studies included bupivacaine (Aasbø 1996; Alagol 2005; Allen 1993; Bjornsson 1994; Calmet 2004; De Andres 1998; Follak 2001; Khoury 1992; Richardson 1997; Ruwe 1995), ropivacaine (Franceschi 2001; Muller 2001), tenoxicam (Alagol 2005; Guler 2002), tramadol (Akinci 2003; Likar 1995), sufentanil (Kazemi 2004), fentanyl (Varkel 1999), pethidine (Lyons 1995), morphine injected IV/IM (Christensen 1996; Dierking 1994; Raj 2004; Richardson 1997; Rosseland 2003a) and different doses of morphine (Allen 1993; Kanbak 1997; Kizilkaya 2005; Likar 1999). Morphine was used from 1 mg to 10 mg. Two hundred and seventy three participants took a 1 mg dose (Allen 1993; Bjornsson 1994; Calmet 2004; De Andres 1998; Kanbak 1997; Khoury 1992; Kizilkaya 2005; Likar 1995; Likar 1999; Muller 2001; Richardson 1997; Wrench 1996), 178 took 2 mg (Alagol 2005; Allen 1993; Dierking 1994; Franceschi 2001; Guler 2002; Likar 1999; Rosseland 2003a; Ruwe 1995), 65 took 3 mg (Aasbø 1996; Kazemi 2004; Varkel 1999), 19 took 4 mg (Likar 1999), 325 took 5 mg (Akinci 2003; Bjornsson 1994; Christensen 1996; Follak 2001; Kanbak 1997; Kizilkaya 2005; Lyons 1995; Muller 2001; Richardson 1997; Solheim 2006), and 39 took 10 mg (Raj 2004).

Pain was rated using VAS in 26 trials, and a verbal rating scale (VRS) was used in six ( Akinci 2003; Christensen 1996; De Andres 1998; Rosseland 2003a; Solheim 2006; Wrench 1996). One study (Bjornsson 1994) assessed pain intensity using a modified VAS score, which was different from the conventional VAS score. In the modified score, "10" corresponded to "severe pain" instead of the conventional ''worst imaginable'' to increase the sensitivity of the scale. We excluded its results of VAS score from meta‐analysis. For those presenting both VAS at rest and VAS with movement, we only abstracted VAS at rest for meta‐analysis. However, the studies did not provide all the outcomes of interest. Some studies gave the central tendency of VAS as median, some presented mean values without SD, and some showed results in figures. We abstracted the data from figures using Engauge Digitizer 4.1. We did not include data presented as median and interquartile range in the meta‐analysis. Different analgesics and treatment regimens were employed for rescue medication, such as tylenol, paracetamol, tramadol, ketorolac, and metamizol, so a meta‐analysis of the rescue medication was not considered feasible.

Excluded studies

We excluded four studies (see Characteristics of excluded studies). One trial used spinal anaesthesia rather than general anaesthesia (Alvarez‐Cabo 1998). Two trials included fewer than 10 participants in the intervention group (Lehrberger 1994; Stein 1991). One trial included participants under 15 years old (De Andres 1993). Three studies were classified in Characteristics of studies awaiting classification because full texts were unavailable through database searching, handsearching or inter‐library loan (Altan 1994; Uzma 1997; VanNess 1994).

Risk of bias in included studies

All the included studies were prospective randomised double‐blind controlled trials (see Characteristics of included studies). We completed a 'Risk of bias' table and results were presented graphically in Figure 2 and summarised in Figure 3.

Risk of bias graph: review authors' judgements about each risk of bias item presented as percentages across all included studies.

Risk of bias summary: review authors' judgements about each risk of bias item for each included study.

Allocation

All of the studies were described as 'randomised', and nine studies were of low risk (seven were randomised using a random table or a list of random numbers, two had randomisation stratified by a computerised allocation schedule). One study was assessed as high risk for its randomisation methods, which allocated patients according to the odd or even account number given on the day of surgery. Randomisation technique was not described in 18 studies, which were assessed as unclear risk. Allocation concealment was of low risk in 10 studies and of unclear risk in 18 studies. Although some studies failed to report the methods used to ensure randomisation and allocation concealment adequately, we included all of the studies in the analysis. We did sensitivity analysis to measure the effects of methodological quality.

Blinding

Twenty‐seven included trials were double‐blind and one was single‐blind. Thirteen trials stated the methods to maintain blindness of participants and personnel and were of low risk. Four studies stated the blinding of outcome assessment and were assessed as unclear risk.

Incomplete outcome data

Twenty‐six of the included studies were assessed as low risk and two studies as high risk. Seven studies (Aasbø 1996; Christensen 1996; Dierking 1994; Likar 1999; Lyons 1995; Raj 2004; Solheim 2006) reported the exact number of participants lost to follow‐up. Five studies (Aasbø 1996; Christensen 1996; Likar 1995; Lyons 1995; Raj 2004) had less than 10% of participants who did not complete the study and two studies (Dierking 1994; Likar 1999) had more than 10% withdrawals. Other studies did not mention dropouts. They included all the participants in the analysis as seen from the results section.

Selective reporting

Five studies (Akinci 2003; Alagol 2005; Calmet 2004; De Andres 1998; Raj 2004) were assessed as low risk and four studies (Guler 2002; Lyons 1995; Ruwe 1995; Wrench 1996) were assessed as high risk, while the other nineteen studies were assessed as unclear risk of selective reporting bias.

Other potential sources of bias

We considered study size as a potential source of bias because most of the included studies were small‐sized. Studies with fewer than 10 participants in each group were excluded. Twenty seven in 28 studies had fewer than 50 participants in each group, which were rated as high risk. These studies were all of small size and thus made the conclusions less robust. Only one study had a group size larger than 50 participants.

Effects of interventions

See: Table 1; Table 2; Table 3; Table 4; Table 5

Summary of findings 1. Morphine compared with placebo for pain control after knee arthroscopy.

Morphine compared with placebo for pain control after knee arthroscopy
Patient or population: patients undergoing knee arthroscopy Settings: inpatients Intervention: 1 mg morphine administered via the knee joint Comparison: placebo (saline) administered via the knee joint
Outcomes	*Illustrative comparative risks (95% CI)**		Relative effect (95% CI)	No of Participants (studies)	Quality of the evidence (GRADE)	Comments
	Assumed risk	Corresponding risk
	Placebo	Morphine
Pain intensity VAS score 0‐10 cm VAS score. Scale from: 0 to 10. Follow‐up: 0‐2 hours	The mean pain score in placebo groups ranged from 2 cm to 5.4 cm	The mean pain score in the 1 mg morphine groups was 0.5 cm lower (1.15 lower to 0.14 higher)		297 (7 studies)	⊕⊕⊝⊝ low^1,2
Pain intensity VAS score 0‐10 cm VAS score. Scale from: 0 to 10. Follow‐up: 2‐6 hours	The mean pain score in placebo groups ranged from 1.7 cm to 4.6 cm	The mean pain score in the 1 mg morphine groups was 0.47 cm lower (1.09 lower to 0.14 higher)		297 (7 studies)	⊕⊕⊝⊝ low^1,2
Pain intensity VAS score 0‐10 cm VAS score. Scale from: 0 to 10. Follow‐up: 6‐30 hours	The mean pain score in placebo groups ranged from 0.8 cm to 4.6 cm	The mean pain score in the 1 mg morphine groups was 0.88 cm lower (1.81 lower to 0.04 higher)		297 (7 studies)	⊕⊕⊝⊝ low^1,2
Analgesia duration Time duration from the end of surgery to the time of first analgesic consumption. Scale from: 0 to 100. Follow‐up: 30 hours	See comment	See comment	Not estimable	124 (3 studies)	⊕⊕⊝⊝ low^3,4	The data were not pooled.
Adverse events Percentage of participants with any adverse event Follow‐up: 30 hours	Study population		RR 1.09  (0.51 to 2.36)	314 (8 studies)	⊕⊕⊝⊝ low^5,6
	103 per 1000	97 per 1000 (51 to 179)
	Moderate
	80 per 1000	76 per 1000 (39 to 142)
The basis for the assumed risk* (e.g. the median control group risk across studies) is provided in footnotes. The corresponding risk (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; RR: Risk Ratio; VAS: Visual analogue scale.
GRADE Working Group grades of evidence High quality: Further research is very unlikely to change our confidence in the estimate of effect. Moderate quality: Further research is likely to have an important impact on our confidence in the estimate of effect and may change the estimate. Low quality: Further research is very likely to have an important impact on our confidence in the estimate of effect and is likely to change the estimate. Very low quality: We are very uncertain about the estimate.

Morphine compared with bupivacaine for pain control after knee arthroscopy
Patient or population: patients undergoing knee arthroscopy Settings: inpatients Intervention: 1 mg morphine administered via the knee joint Comparison: bupivacaine administered via the knee joint
Outcomes	*Illustrative comparative risks (95% CI)**		Relative effect (95% CI)	No of Participants (studies)	Quality of the evidence (GRADE)	Comments
	Assumed risk	Corresponding risk
	Bupivacaine	Morphine
Pain intensity VAS score 0‐10 cm VAS score. Scale from: 0 to 10. Follow‐up: 0‐2 hours	The mean pain score in bupivacaine groups ranged from 0.7 cm to 2.3 cm	The mean pain score in the morphine groups was 1.43 cm higher (0.49 to 2.37 higher)		248 (5 studies)	⊕⊕⊝⊝ low^1,2
Pain intensity VAS score 0‐10 cm VAS score. Scale from: 0 to 10. Follow‐up: 2‐6 hours	The mean pain score in bupivacaine groups ranged from 1.2 cm to 3.8 cm	The mean pain score in the morphine groups was 0.45 cm higher (0.47 lower to 1.36 higher)		330 (6 studies)	⊕⊕⊝⊝ low^1,3
Pain intensity VAS score 0‐10 cm VAS score. Scale from: 0 to 10. Follow‐up: 6‐30 hours	The mean pain score in bupivacaine groups ranged from 1.3 cm to 3.7 cm	The mean pain score in the morphine groups was 0.71 cm lower (1.23 to 0.19 lower)		270 (5 studies)	⊕⊕⊝⊝ low^1,4
Analgesia duration Time duration from the end of surgery to the time of first analgesic consumption. Scale from: 0 to 100. Follow‐up: 30 hours	See comment	See comment	Not estimable	162 (3 studies)	⊕⊕⊕⊝ moderate⁵	The data were not pooled.
Adverse events Percentage of participants with any adverse event Follow‐up: 30 hours	Study population		RR 0.68 (0.09 to 5.17)	210 (4 studies)	⊕⊕⊕⊝ moderate⁶
	76 per 1000	49 per 1000 (17 to 135)
	Moderate
	40 per 1000	26 per 1000 (9 to 73)
The basis for the assumed risk* (e.g. the median control group risk across studies) is provided in footnotes. The corresponding risk (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; RR: Risk Ratio; VAS: Visual analogue scale.
GRADE Working Group grades of evidence High quality: Further research is very unlikely to change our confidence in the estimate of effect. Moderate quality: Further research is likely to have an important impact on our confidence in the estimate of effect and may change the estimate. Low quality: Further research is very likely to have an important impact on our confidence in the estimate of effect and is likely to change the estimate. Very low quality: We are very uncertain about the estimate.

Morphine compared with NSAIDs for pain control after knee arthroscopy
Patient or population: patients undergoing knee arthroscopy Settings: inpatients Intervention: 1 mg morphine administered via the knee joint Comparison: NSAIDs administered via the knee joint
Outcomes	*Illustrative comparative risks (95% CI)**		Relative effect (95% CI)	No of Participants (studies)	Quality of the evidence (GRADE)	Comments
	Assumed risk	Corresponding risk
	NSAIDs	Morphine
Pain intensity VAS score 0‐10 cm VAS score. Scale from: 0 to 10. Follow‐up: 0‐2 hours	The mean pain score in the NSAIDs groups ranged from 1 cm to 6.2 cm	The mean pain score in the morphine groups was 0.95 cm higher (0.95 lower to 2.85 higher)		80 (2 studies)	⊕⊝⊝⊝ verylow^1,2,3
Pain intensity VAS score 0‐10 cm VAS score. Scale from: 0 to 10. Follow‐up: 2‐6 hours	The mean pain score in the NSAIDs groups ranged from 2 cm to 2.7 cm	The mean pain score in the morphine groups was 1.00 cm higher (0.12 to 1.88 higher)		80 (2 studies)	⊕⊕⊝⊝ low^1,2
Pain intensity VAS score 0‐10 cm VAS score. Scale from: 0 to 10. Follow‐up: 6‐30 hours	The mean pain score in the NSAIDs groups ranged from 1.3 cm to 2 cm	The mean pain score in the morphine groups was 0.43 cm higher (0.54 lower to 1.39 higher)		80 (2 studies)	⊕⊕⊝⊝ low^1,2
Analgesia duration time duration from the end of surgery to the time of first analgesic consumption. Scale from: 0 to 100. Follow‐up: 30 hours	See comment	See comment	Not estimable	50 (1 study)	⊕⊕⊝⊝ low^4,5	The data were not pooled.
Adverse events Percentage of participants with any adverse event Follow‐up: 30 hours	Study population		RR 0.75 (0.19 to 3.04)	120 (3 studies)	⊕⊕⊕⊝ moderate⁶
	67 per 1000	49 per 1000 (11 to 203)
	Moderate
	40 per 1000	29 per 1000 (6 to 129)
The basis for the assumed risk* (e.g. the median control group risk across studies) is provided in footnotes. The corresponding risk (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; RR: Risk Ratio; VAS: Visual analogue scale.
GRADE Working Group grades of evidence High quality: Further research is very unlikely to change our confidence in the estimate of effect. Moderate quality: Further research is likely to have an important impact on our confidence in the estimate of effect and may change the estimate. Low quality: Further research is very likely to have an important impact on our confidence in the estimate of effect and is likely to change the estimate. Very low quality: We are very uncertain about the estimate.

1 mg morphine compared with 2 mg/4 mg/5 mg morphine for pain control after knee arthroscopy
Patient or population: patients undergoing knee arthroscopy Settings: inpatients Intervention: 1 mg morphine administered via the knee joint Comparison: 2 mg/5 mg morphine administered via the knee joint
Outcomes	*Illustrative comparative risks (95% CI)**		Relative effect (95% CI)	No of Participants (studies)	Quality of the evidence (GRADE)	Comments
	Assumed risk	Corresponding risk
	2 mg / 4 mg/5 mg morphine	1 mg morphine
Pain intensity VAS score 1 mg morphine vs 2 mg morphine 0‐10 cm VAS score. Scale from: 0 to 10. Follow‐up: 0‐2 hours	The mean pain score in the 2 mg morphine groups ranged from 1.8 cm to 3.9 cm	The mean pain score in the 1 mg morphine groups was 0.56 cm lower (1.93 lower to 0.81 higher)		105 (2 studies)	⊕⊕⊝⊝ low^1,2
Pain intensity VAS score 1 mg morphine vs 2 mg morphine 0‐10 cm VAS score. Scale from: 0 to 10. Follow‐up: 2‐6 hours	The mean pain score in the 2 mg morphine groups ranged from 1.6 cm to 3.8 cm	The mean pain score in the 1 mg morphine groups was 0.32 cm lower (1.69 lower to 1.05 higher)		105 (2 studies)	⊕⊕⊝⊝ low^1,2
Pain intensity VAS score 1 mg morphine vs 2 mg morphine 0‐10 cm VAS score. Scale from: 0 to 10. Follow‐up: 6‐30 hours	The mean pain score in the 2 mg morphine groups is 0.45 cm	The mean pain score in the 1 mg morphine groups was 0.55 cm higher (0.30 lower to 1.40 higher)		45 (1 study)	⊕⊕⊝⊝ low^1,2
Analgesia duration 1 mg morphine vs 2 mg morphine Time duration from the end of surgery to the time of first analgesic consumption Follow‐up: 30 hours	See comment	See comment	Not estimable	60 (1 study)	⊕⊕⊕⊝ moderate³	The data were not pooled.
Adverse events 1 mg morphine vs 2 mg morphine Percentage of participants with any adverse event Follow‐up: 30 hours	See comment	See comment	Not estimable	None	Not estimable	The were no data available.
Pain intensity VAS score 1 mg morphine vs 4 mg/5 mg morphine 0‐10 cm VAS score. Scale from: 0 to 10. Follow‐up: 0‐2 hours	The mean pain score in the 5 mg morphine groups ranged from 1.5 cm to 3.5 cm	The mean pain score in the1 mg morphine groups was 0.46 cm higher (0.24 lower to 1.16 higher)		67 (2 studies)	⊕⊕⊝⊝ low^4,5
Pain intensity VAS score 1 mg morphine vs 4 mg/5 mg morphine 0‐10 cm VAS score. Scale from: 0 to 10. Follow‐up: 2‐6 hours	The mean pain score in the 5 mg morphine groups ranged from 1.1 cm to 3.7 cm	The mean pain score in the 1 mg morphine groups was 0.44 cm higher (0.18 lower to 1.05 higher)		67 (2 studies)	⊕⊕⊝⊝ low^4,5
Pain intensity VAS score 1 mg morphine vs 4 mg/5 mg morphine 0‐10 cm VAS score. Scale from: 0 to 10. Follow‐up: 6‐30 hours	The mean pain score in the 5mg morphine groups ranged from 0.28 cm to 3.8 cm	The mean pain score in the1mg morphine groups was 0.67 cm higher (0.08 to 1.25 higher)		67 (2 studies)	⊕⊕⊝⊝ low^4,5
Analgesia duration 1 mg morphine vs 4 mg/5 mg morphine time duration from the end of surgery to the first time of analgesic consumption Follow‐up: 30 hours	See comment	See comment	Not estimable	24 (1 study)	⊕⊕⊕⊝ moderate⁶	The data were not pooled.
Adverse events 1 mg morphine vs 4 mg/5 mg morphine Percentage of participants with any adverse event Follow‐up: 30 hours	See comment	See comment	Not estimable	None	Not estimable	The were no data available.
The basis for the assumed risk* (e.g. the median control group risk across studies) is provided in footnotes. The corresponding risk (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; RR: Risk Ratio; VAS: Visual analogue scale.
GRADE Working Group grades of evidence High quality: Further research is very unlikely to change our confidence in the estimate of effect. Moderate quality: Further research is likely to have an important impact on our confidence in the estimate of effect and may change the estimate. Low quality: Further research is very likely to have an important impact on our confidence in the estimate of effect and is likely to change the estimate. Very low quality: We are very uncertain about the estimate.

IA morphine compared with IM morphine for pain control after knee arthroscopy
Patient or population: patients undergoing knee arthroscopy Settings: inpatients Intervention: 1 mg morphine administered via the knee joint Comparison: morphine administered intra‐muscularly
Outcomes	*Illustrative comparative risks (95% CI)**		Relative effect (95% CI)	No of Participants (studies)	Quality of the evidence (GRADE)	Comments
	Assumed risk	Corresponding risk
	IM morphine	IA morphine
Pain intensity VAS score 0‐10 cm VAS score. Scale from: 0 to 10. Follow‐up: 0‐2 hours	The mean pain score in the IM morphine groups ranged from 1.4 cm to 2.5 cm	The mean pain score in the IA morphine groups was 0.21 cm higher (0.48 lower to 0.9 higher)		72 (2 studies)	⊕⊝⊝⊝ verylow^1,2,3
Pain intensity VAS score 0‐10 cm VAS score. Scale from: 0 to 10. Follow‐up: 2‐6 hours	The mean pain score in the IM morphine groups ranged from 1cm to 1.7 cm	The mean pain score in the IA morphine groups was 0.14 cm lower (0.93 lower to 0.64 higher)		72 (2 studies)	⊕⊕⊝⊝ low^1,2
Pain intensity VAS score 0‐10 cm VAS score. Scale from: 0 to 10. Follow‐up: 6‐30 hours	The mean pain score in the IM morphine groups is 1.5 cm	The mean pain score in the IA morphine groups was 0.30 cm lower (1.39 lower to 0.79 higher)		39 (1 studies)	⊕⊕⊝⊝ low^1,4
Analgesia duration time duration from the end of surgery to the time of first analgesic consumption. Scale from: 0 to 100. Follow‐up: 30 hours	See comment	See comment	Not estimable	None	Not estimable	The were no data available.
Adverse events Percentage of participants with any adverse event Follow‐up: 30 hours	See comment	See comment	Not estimable	None	Not estimable	The were no data available.
The basis for the assumed risk* (e.g. the median control group risk across studies) is provided in footnotes. The corresponding risk (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; RR: Risk Ratio; VAS: Visual analogue scale.
GRADE Working Group grades of evidence High quality: Further research is very unlikely to change our confidence in the estimate of effect. Moderate quality: Further research is likely to have an important impact on our confidence in the estimate of effect and may change the estimate. Low quality: Further research is very likely to have an important impact on our confidence in the estimate of effect and is likely to change the estimate. Very low quality: We are very uncertain about the estimate.

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
1.1 1mg morphine vs placebo	7		Mean Difference (IV, Random, 95% CI)	Subtotals only
1.1.1 0‐2h	7	297	Mean Difference (IV, Random, 95% CI)	‐0.50 [‐1.15, 0.14]
1.1.2 2‐6h	7	297	Mean Difference (IV, Random, 95% CI)	‐0.47 [‐1.09, 0.14]
1.1.3 6‐30h	7	297	Mean Difference (IV, Random, 95% CI)	‐0.88 [‐1.81, 0.04]
1.2 1mg morphine vs bupivacaine	6		Mean Difference (IV, Random, 95% CI)	Subtotals only
1.2.1 0‐2h	5	248	Mean Difference (IV, Random, 95% CI)	1.43 [0.49, 2.37]
1.2.2 2‐6h	6	330	Mean Difference (IV, Random, 95% CI)	0.45 [‐0.47, 1.36]
1.2.3 6‐30h	5	270	Mean Difference (IV, Random, 95% CI)	‐0.71 [‐1.23, ‐0.19]
1.3 morphine vs NSAIDs	2		Mean Difference (IV, Random, 95% CI)	Subtotals only
1.3.1 0‐2h	2	80	Mean Difference (IV, Random, 95% CI)	0.95 [‐0.95, 2.85]
1.3.2 2‐6h	2	80	Mean Difference (IV, Random, 95% CI)	1.00 [0.12, 1.88]
1.3.3 6‐30h	2	80	Mean Difference (IV, Random, 95% CI)	0.43 [‐0.54, 1.39]
1.4 IA morphine vs IM morphine	2		Mean Difference (IV, Fixed, 95% CI)	Subtotals only
1.4.1 0‐2h	2	72	Mean Difference (IV, Fixed, 95% CI)	0.21 [‐0.48, 0.90]
1.4.2 2‐6h	2	72	Mean Difference (IV, Fixed, 95% CI)	‐0.14 [‐0.93, 0.64]
1.4.3 6‐30h	1	39	Mean Difference (IV, Fixed, 95% CI)	‐0.30 [‐1.39, 0.79]
1.5 1mg morphine vs 2mg morphine	2		Mean Difference (IV, Random, 95% CI)	Subtotals only
1.5.1 0‐2h	2	105	Mean Difference (IV, Random, 95% CI)	‐0.56 [‐1.93, 0.81]
1.5.2 2‐6h	2	105	Mean Difference (IV, Random, 95% CI)	‐0.32 [‐1.69, 1.05]
1.5.3 6‐30h	1	45	Mean Difference (IV, Random, 95% CI)	0.55 [‐0.30, 1.40]
1.6 1mg morphine vs 4mg/5mg morphine	3		Mean Difference (IV, Fixed, 95% CI)	Subtotals only
1.6.1 0‐2h	3	97	Mean Difference (IV, Fixed, 95% CI)	0.46 [‐0.24, 1.16]
1.6.2 2‐6h	3	97	Mean Difference (IV, Fixed, 95% CI)	0.44 [‐0.18, 1.05]
1.6.3 6‐30h	3	97	Mean Difference (IV, Fixed, 95% CI)	0.67 [0.08, 1.25]

Outcome or subgroup title	No. of studies	Statistical method	Effect size
2.1 morphine vs placebo	3	Mean Difference (IV, Random, 95% CI)	Totals not selected
2.2 morphine vs bupivacaine	3	Mean Difference (IV, Random, 95% CI)	Totals not selected
2.3 morphine vs NSAIDS	1	Mean Difference (IV, Fixed, 95% CI)	Totals not selected
2.4 1mg morphine vs 2mg morphine	1	Mean Difference (IV, Fixed, 95% CI)	Totals not selected
2.5 1mg morphine vs 5mg morphine	1	Mean Difference (IV, Fixed, 95% CI)	Totals not selected

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
3.1 morphine vs placebo	8	314	Risk Ratio (M‐H, Fixed, 95% CI)	1.09 [0.51, 2.36]
3.2 morphine vs bupivacaine	4	210	Risk Ratio (M‐H, Random, 95% CI)	0.68 [0.09, 5.17]
3.3 morphine vs NSAIDs	3	120	Risk Ratio (M‐H, Fixed, 95% CI)	0.75 [0.19, 3.04]

*Study characteristics*
Methods	Prospective, randomised double‐blind trial
Participants	107 patients scheduled for elective, diagnostic knee arthroscopy
Interventions	Group 1: 20 ml of bupivacaine 2.5 mg/ml with 3 mg of morphine; Group 2: 20 ml of bupivacaine 2.5 mg/ml; Group 3: 20 ml isotonic saline with 3 mg morphine; Group 4: 20 ml of isotonic saline
Outcomes	VAS, time to first analgesic administered, analgesic consumption
Notes	No additional analgesic effect of IA morphine or bupivacaine
*Risk of bias*
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Quote: The participants were randomly assigned to double‐blind administration of 20 ml of the test drug
Allocation concealment (selection bias)	Unclear risk	Quote: The test drugs were drawn into a syringe by an independent nurse
Incomplete outcome data (attrition bias) All outcomes	Low risk	Comment: no patient was excluded, but 3 in 107 (less than 10%) participants did not complete the questionnaire.
Selective reporting (reporting bias)	Unclear risk	Comment: insufficient data were available
Blinding of participants and personnel (performance bias) All outcomes	Unclear risk	Quote: was given to the surgeon who administrated the injection at the end of the procedure without knowing the contents
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	Comment: not specified
Size	High risk	Comment: < 50 participants per treatment arm.

*Study characteristics*
Methods	Prospective double‐blind randomised trial
Participants	75 patients having elective arthroscopic surgery of the knee, included diagnostic arthroscopy, meniscectomy, and minimal debridement or loose body removal, or both
Interventions	3 groups ‐ IA tramadol 50 mg (tramadol group), IA morphine 5 mg (morphine group), or IA normal saline (control group), injected through the arthroscope with the study drug supplied in a coded syringe
Outcomes	VRS, supplemental analgesic requirements, incidence of side effects
Notes	There was no significance with respect to the pain scores postoperatively except the first VRS pain score when the participants arrived at the PACU. Nausea and vomiting were reported in 32% of participants in the control group and 24% of participants in both the morphine and the tramadol groups (P>0.05). Somnolence was noted in 24% of the control group, 16% of the morphine group, and 8% of the tramadol group (P=0.10)
*Risk of bias*
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Quote: a computerized allocation schedule
Allocation concealment (selection bias)	Low risk	Quote: coded syringes containing the study drugs
Incomplete outcome data (attrition bias) All outcomes	Low risk	Comment: no participants withdrew from the study
Selective reporting (reporting bias)	Low risk	Comment: all the outcomes in methods section were reported. And VRS score of continuous time course were reported
Blinding of participants and personnel (performance bias) All outcomes	Low risk	Quote: the coded syringes were prepared by an anaesthesiologist not involved in the administration of the drug, patient care, or data collection
Blinding of outcome assessment (detection bias) All outcomes	Low risk	Quote: A blinded observer recorded pain, supplemental analgesic requirements, and incidence of side effects
Size	High risk	Comment: < 50 participants per treatment arm

*Study characteristics*
Methods	Randomised, placebo‐controlled, double‐blinded study
Participants	150 patients undergoing arthroscopic knee surgery, multiple arthroscopic procedures
Interventions	Group N received 500 μg neostigmine, Group M received 2 mg morphine, Group T received 20 mg tenoxicam, Group C received clonidine, Group B received 100 mg bupivacaine and Group S received saline 20 ml.
Outcomes	VAS, duration of analgesia
Notes	Neostigmine, clonidine, tenoxicam, morphine and bupivacaine decreased postoperative pain intensity and reduced analgesic consumption when compared with placebo. The most effective drugs that are administered intra articularly are neostigmine and clonidine among the five drugs we studied. Tenoxicam provided longer analgesia when compared with morphine and bupivacaine, postoperatively
*Risk of bias*
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Quote: a table of random numbers
Allocation concealment (selection bias)	Unclear risk	Comment: not specified
Incomplete outcome data (attrition bias) All outcomes	Low risk	Comment: no participants withdrew from the study
Selective reporting (reporting bias)	Low risk	Comment: all the outcomes in methods section were reported and VAS score of continuous time course were reported
Blinding of participants and personnel (performance bias) All outcomes	Low risk	Quote: The solutions were prepared in 20‐ml volume by an anaesthesiologist and administered by a surgeon who was blinded to the contents of the syringe
Blinding of outcome assessment (detection bias) All outcomes	Low risk	Quote: After extubation, participants were observed in the postanaesthesia care unit for 1 h and in the clinic of orthopaedic surgery by an anaesthesiologist who was blinded to the groups
Size	High risk	Comment: < 50 participants per treatment arm

*Study characteristics*
Methods	Randomised, double‐blind trial
Participants	120 patients ASA 1‐2, need arthroscopy surgery
Interventions	Group 1: bupivacaine Group 2: 1 mg morphine Group 3: 2 mg morphine Group 4: morphine & bupivacaine
Outcomes	VAS, duration
Notes
*Risk of bias*
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Quote: use a random table to ensure randomisation
Allocation concealment (selection bias)	Unclear risk	Comment: not specified
Incomplete outcome data (attrition bias) All outcomes	Low risk	Comment: no participants withdrew from the study
Selective reporting (reporting bias)	Unclear risk	Comment: insufficient data were available
Blinding of participants and personnel (performance bias) All outcomes	Unclear risk	Comment: not specified
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	Comment: not specified
Size	High risk	Comment: < 50 participants per treatment arm

*Study characteristics*
Methods	Two‐stage, randomised, double‐blind, controlled trial
Participants	149 patients ASA 1‐2, part 1: 78 part 2: 71
Interventions	Morphine 1 mg, saline, bupivacaine, morphine & bupivacaine 5 mg morphine IA, saline, 5 mg morphine IM
Outcomes	VAS (severe, sensitivity)
Notes	VAS scores similar
*Risk of bias*
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Comment: not specified
Allocation concealment (selection bias)	Unclear risk	Comment: not specified
Incomplete outcome data (attrition bias) All outcomes	Low risk	Comment: no participants withdrew from the study.
Selective reporting (reporting bias)	Unclear risk	Comment: insufficient data were available
Blinding of participants and personnel (performance bias) All outcomes	Unclear risk	Comment: not specified
Blinding of outcome assessment (detection bias) All outcomes	Low risk	Quote: After the operation, a nurse who was blind to the method asked the patient to assess the severity of pain
Size	High risk	Comment: < 50 participants per treatment arm

*Study characteristics*
Methods	Prospective, double‐blind, randomised study
Participants	80 consecutive patients were studied who had been diagnosed with torn meniscus and recommended to have arthroscopic surgery
Interventions	Group 1 participants (n = 20) received postoperative injection of 60 mg IA ketorolac, Group 2 participants (n = 20) 10 cc IA bupivacaine 0.25% Group 3 participants (n = 20) 1 mg IA morphine diluted in 10 cc saline, and Group 4 participants (n = 20, controls) only 10 cc saline
Outcomes	VAS, analgesic duration
Notes	The best analgesic effect was IA ketorolac
*Risk of bias*
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Comment: not specified
Allocation concealment (selection bias)	Unclear risk	Comment: not specified
Incomplete outcome data (attrition bias) All outcomes	Low risk	Comment: no participants withdrew from the study.
Selective reporting (reporting bias)	Low risk	Comment: all the outcomes in methods section were reported and VAS score of continuous time course were reported
Blinding of participants and personnel (performance bias) All outcomes	Unclear risk	Comment: not specified
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	Comment: not specified
Size	High risk	Comment: < 50 participants per treatment arm

*Study characteristics*
Methods	A controlled, randomised and double‐blind study
Participants	61 ASA 1‐2, having elective arthroscopic surgery of the knee (61 participants recruited, 58 participants completed trial)
Interventions	Group 1: (n = 29) 5 mg morphine IA, Group 2: (n = 29) 5 mg morphine IM
Outcomes	VRS and VAS, analgesic consumption
Notes	The clinical analgesic effect of 5 mg morphine given intra‐articularly is equal to 5 mg morphine given intra‐muscularly. The occurrence of villous synovitis seems to be of no clinical importance concerning the local effect of morphine.
*Risk of bias*
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Comment: not specified
Allocation concealment (selection bias)	Unclear risk	Comment: not specified
Incomplete outcome data (attrition bias) All outcomes	Low risk	Comment: 3/61 participants excluded, < 10% of participants did not complete the study
Selective reporting (reporting bias)	Unclear risk	Comment: insufficient data were available
Blinding of participants and personnel (performance bias) All outcomes	Unclear risk	Comment: not specified.
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	Quote: all participants were assessed by an observer blinded to the randomisation of the patient
Size	High risk	Comment: < 50 participants per treatment arm

*Study characteristics*
Methods	Randomised double‐blind, placebo‐controlled trial
Participants	78 patients having elective arthroscopic meniscectomy
Interventions	Group 1 (n = 25): 0.25% bupivacaine (50 mg) (IA), Group 2 (n = 27): 1 mg morphine, Group 3 (n = 26): morphine bupivacaine mixture, Group 4 (n = 25): normal saline
Outcomes	VAS, VRS, side effect, time to first request of analgesics
Notes	Side effects occurred in 13.7% of the participants, with urinary retention being the most common (n = 8, 10%). In this regard there were no significant differences. These results show no significant differences between the groups
*Risk of bias*
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Quote: a random number table
Allocation concealment (selection bias)	Unclear risk	Comment: not specified
Incomplete outcome data (attrition bias) All outcomes	Low risk	Comment: no participants withdrew from the study
Selective reporting (reporting bias)	Low risk	Comment: all the outcomes in methods section were reported and VRS score of continuous time course were reported
Blinding of participants and personnel (performance bias) All outcomes	Low risk	Quote: The surgeon injected 20 ml solution into the knee joint. The patient, the anaesthesiologist in charge and the surgeon were blind to the solution injected. Participants were informed pre‐operatively by a blinded observer about all pain scores
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	Comment: not specified
Size	High risk	Comment: < 50 participants per treatment arm

*Study characteristics*
Methods	Double‐blind, randomised study
Participants	40 healthy patients undergoing elective arthroscopic meniscectomy
Interventions	18 participants received IA morphine and 15 participants IM morphine 2 mg
Outcomes	VAS
Notes
*Risk of bias*
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Comment: not specified
Allocation concealment (selection bias)	Unclear risk	Comment: not specified
Incomplete outcome data (attrition bias) All outcomes	High risk	Comment: 7 in 40 of participants did not complete the study, used 'completer analysis'
Selective reporting (reporting bias)	Unclear risk	Comment: insufficient data were available
Blinding of participants and personnel (performance bias) All outcomes	Low risk	Quote: Study medication was drawn from coded ampoules to ensure the double‐blind nature of the study
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	Comment: not specified
Size	High risk	Comment: < 50 participants per treatment arm

*Study characteristics*
Methods	Randomised
Participants	82 patients underwent partial meniscectomy, chondral debridement, or both
Interventions	10 mg morphine versus 10 cc bupivacaine
Outcomes	VAS score, side effects
Notes	Quasi‐randomised trial
*Risk of bias*
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	High risk	Comment: group assignment was determined by an odd or even account number given on the day of surgery
Allocation concealment (selection bias)	Unclear risk	Comment: insufficient data were available
Incomplete outcome data (attrition bias) All outcomes	Low risk	Comment: eighteen participants (11 in the morphine group and 7 in the bupivacaine group) had data available through hospital discharge but no data in the subsequent 24 hours. The remaining 64 participants (36 in the morphine group and 28 in the bupivacaine group) had post‐anaesthesia care unit data, as well as follow‐up VAS and medication use data through 24 hours postoperatively.
Selective reporting (reporting bias)	Unclear risk	Comment: insufficient data were available. The time point of VAS recording was not clear
Blinding of participants and personnel (performance bias) All outcomes	Unclear risk	Comment: single‐blind. The participants were blinded to group assignment, but the investigators were not
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	Comment: not specified
Size	High risk	Comment: < 50 participants per treatment arm

*Study characteristics*
Methods	Prospective, randomised, double‐blind study
Participants	320 patients, subdivided into 4 groups of 80, who underwent arthroscopic knee surgery
Interventions	Each of the 4 groups received a different solution: 15 ml of bupivacaine 0.5%, 5 mg of morphine in 15 ml of isotonic saline solution, 15 ml of bupivacaine 0.5% with epinephrine 0.0005% 15 ml of isotonic saline solution (control group)
Outcomes	VAS, consumption of analgesics and time point of first analgesic application
Notes	Bupivacaine 0.5% with epinephrine 0.0005% was found to be the most effective
*Risk of bias*
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Comment: not specified
Allocation concealment (selection bias)	Unclear risk	Comment: not specified
Incomplete outcome data (attrition bias) All outcomes	Low risk	Comment: no participants withdrew from the study
Selective reporting (reporting bias)	Unclear risk	Comment: insufficient data were available
Blinding of participants and personnel (performance bias) All outcomes	Unclear risk	Comment: not specified
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	Comment: not specified
Size	Unclear risk	Comment: 50 to 199 participants per treatment arm

*Study characteristics*
Methods	Randomised, double‐blind
Participants	90 patients scheduled to undergo elective knee arthroscopy
Interventions	Group 1 received ropivacaine 75 mg, Group 2 received 2 mg morphine, Group 3 received 20 ml of saline solution.
Outcomes	VAS, duration, side effects
Notes
*Risk of bias*
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Comment: not specified
Allocation concealment (selection bias)	Low risk	Quote: the 3 different analgesics were administered in a double‐blinded randomised fashion from a coded syringe into the joint space
Incomplete outcome data (attrition bias) All outcomes	Low risk	Comment: no participants withdrew from the study
Selective reporting (reporting bias)	Unclear risk	Comment: insufficient data were available
Blinding of participants and personnel (performance bias) All outcomes	Low risk	Quote: the 3 different analgesics were administered in a double‐blinded randomised fashion from a coded syringe
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	Comment: not specified
Size	High risk	Comment: < 50 participants per treatment arm

*Study characteristics*
Methods	Prospective, randomised, controlled, double‐blinded
Participants	35 patients, included partial or total meniscectomy and repair of ruptured ligaments
Interventions	Group 1: n = 11, NS intra‐articularly Group 2: n = 11, 1 mg morphine intra‐articularly Group 3: n = 13, 5 mg morphine intra‐articularly
Outcomes	VAS, analgesic consumption, side effects
Notes
*Risk of bias*
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Comment: not specified
Allocation concealment (selection bias)	Unclear risk	Comment: not specified
Incomplete outcome data (attrition bias) All outcomes	Low risk	Comment: no participants withdrew from the study
Selective reporting (reporting bias)	Unclear risk	Comment: insufficient data were available
Blinding of participants and personnel (performance bias) All outcomes	Unclear risk	Comment: not specified.
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	Quote: VAS score recorded by a blinded observer
Size	High risk	Comment: < 50 participants per treatment arm

PERMALINK

Single dose intra‐articular morphine for pain control after knee arthroscopy

Zui Zou

Mao Mao An

Qun Xie

Xiaoyan Y Chen

Hao Zhang

Guan J Liu

Xue Y Shi

Abstract

Background

Objectives

Search methods

Selection criteria

Data collection and analysis

Main results

Authors' conclusions

Plain language summary

Summary of findings

Background

Description of the condition

Description of the intervention

How the intervention might work

Why it is important to do this review

Objectives

Methods

Criteria for considering studies for this review

Types of studies

Types of participants

Types of interventions

Types of outcome measures

Primary outcomes

Secondary outcomes

Search methods for identification of studies

Electronic searches

Searching other resources

Reference lists

Unpublished studies

Language

Data collection and analysis

Selection of studies

Data extraction and management

Assessment of risk of bias in included studies

Dealing with missing data

Assessment of heterogeneity

Data synthesis

Quantitative analysis

Dichotomous data

Continuous data

Qualitative analysis

Quality of the evidence

'Summary of findings' table

Subgroup analysis and investigation of heterogeneity

Sensitivity analysis

Results

Description of studies

Results of the search

1.

Included studies

Excluded studies

Risk of bias in included studies

2.

3.

Allocation

Blinding

Incomplete outcome data

Selective reporting

Other potential sources of bias

Effects of interventions

Summary of findings 1. Morphine compared with placebo for pain control after knee arthroscopy.

Summary of findings 2. Morphine compared with bupivacaine for pain control after knee arthroscopy.

Summary of findings 3. Morphine compared with NSAIDs for pain control after knee arthroscopy.

Summary of findings 4. Different doses of morphine for pain control after knee arthroscopy.

Summary of findings 5. IA morphine compared with IM morphine for pain control after knee arthroscopy.

Primary outcomes: patient‐reported postoperative pain intensity and use of supplementary analgesic

1. Morphine versus placebo

4.

2. Morphine versus local anaesthetics

5.

3. Morphine versus NSAIDs

*Study characteristics*
Methods	Prospective, double‐blind study
Participants	45 patients who were ASA physical status I and II and scheduled for arthroscopic knee surgery
Interventions	Sufentanil 5 µg (group S), morphine 3 mg (group M) or normal saline 20 cc as placebo (group p), intra‐articularly at the end of arthroscopic knee surgery
Outcomes	VAS, consumption of rescue medication
Notes
*Risk of bias*
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Comment: not specified
Allocation concealment (selection bias)	Low risk	Quote: The contents of these syringes were unknown to anesthesiologist and surgeon who performed the study. The codes were not revealed until completion of the study.
Incomplete outcome data (attrition bias) All outcomes	Low risk	Comment: no participants withdrew from the study
Selective reporting (reporting bias)	Unclear risk	Comment: insufficient data were available
Blinding of participants and personnel (performance bias) All outcomes	Low risk	Quote: the content of syringe in unknown to surgeon and anaesthesiologists
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	Comment: not specified
Size	High risk	Comment: < 50 participants per treatment arm

*Study characteristics*
Methods	Double‐blind, randomised
Participants	72 patients scheduled for arthroscopic knee surgery either for diagnostic purposes or for partial meniscectomy
Interventions	Group C (n = 18), 20 ml of IA isotonic saline and 5 mg morphine in 5 ml of isotonic saline IV; Group M5 (n = 17) received 5 mg morphine in 20 ml of isotonic saline intra‐articularly and 5 ml isotonic saline IV; Group M1 (n = 18) received 1 mg morphine in 20 ml of isotonic saline intra‐articularly and 5 ml of isotonic saline IV; Group M1M (n = 19) received 1 mg morphine plus 40 mg methylprednisolone in 20 ml of isotonic saline intra‐articularly and 5 ml of isotonic saline IV. Injected the solution through the arthro‐scope
Outcomes	VAS, analgesic consumption
Notes	The analgesic effect of morphine given intra‐articularly is dose dependent and that combination of methylprednisolone with morphine has an additive effect on analgesia
*Risk of bias*
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	By a computer‐generated randomisation sequence
Allocation concealment (selection bias)	Unclear risk	Comment: not specified
Incomplete outcome data (attrition bias) All outcomes	Low risk	Comment: no participants withdrew from the study
Selective reporting (reporting bias)	Unclear risk	Comment: iInsufficient data were available
Blinding of participants and personnel (performance bias) All outcomes	Low risk	Quote: without knowing the contents measured and recorded by a anaesthesiologist who is blind to the group allocation
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	Comment: not specified
Size	High risk	Comment: < 50 participants per treatment arm

*Study characteristics*
Methods	Prospective randomised, double‐blind controlled trial
Participants	135 patients > 18 yrs, knee arthroscopy
Interventions	9 groups, each has 15 participants 1A NS 2A morphine 5 mg 3A morphine 1 mg 4A mor + ropivacaine 5A ropivacaine 20 mg 1B NS 2B morphine 5 mg 3B morphine 1 mg 4B mor + ropivacaine 5B ropivacaine 20 mg
Outcomes	VAS, tramadol consumption
Notes
*Risk of bias*
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Comment: not specified
Allocation concealment (selection bias)	Unclear risk	Comment: not specified
Incomplete outcome data (attrition bias) All outcomes	Low risk	Comment: no participants withdrew from the study
Selective reporting (reporting bias)	Unclear risk	Comment: insufficient data were available
Blinding of participants and personnel (performance bias) All outcomes	Unclear risk	Comment: not specified
Blinding of outcome assessment (detection bias) All outcomes	Low risk	Comment: outcome assessor was blind
Size	High risk	Comment: < 50 participants per treatment arm

*Study characteristics*
Methods	Two prospective, randomised, double‐blind clinical trials
Participants	Trial 1: n = 106 Trial 2: n = 48, various kinds of knee surgery
Interventions	Trial 1: 1A 1 mg morphine in 20 ml normal saline IA, 1B 20 ml normal saline IA, 1C 20 ml 0.5% bupivacaine IA Trial 2: 20 ml normal saline IA plus 5 mg morphine in 10 ml normal saline IV, 2B 1 mg morphine in 20 ml normal saline IA plus 10 ml normal saline IV 2C 5 mg morphine in 20 ml normal saline IA plus 10 ml normal saline IV
Outcomes	VAS, analgesia consumption
Notes	5 mg IA was the most effective analgesic following knee arthroscopy
*Risk of bias*
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Comment: not specified
Allocation concealment (selection bias)	Unclear risk	Comment: not specified
Incomplete outcome data (attrition bias) All outcomes	Low risk	Comment: no participants withdrew from the study
Selective reporting (reporting bias)	Unclear risk	Comment: insufficient data were available
Blinding of participants and personnel (performance bias) All outcomes	Low risk	Quote: solution was injected into the knee by the surgeon, who was blinded to its contents
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	Comment: not specified
Size	High risk	Comment: < 50 participants per treatment arm

*Study characteristics*
Methods	A randomised, double‐blind controlled clinical study
Participants	40 patients (59 recruited, 40 included in analysis), various kinds of knee surgery
Interventions	Saline 10 ml with morphine 2 mg (n = 19), without morphine (n = 21)
Outcomes	VAS, VRS, analgesic time and consumption
Notes	Only 70% of 57 participants had pain of moderate to severe intensity within 1 h after an arthroscopic procedure of the knee joint under general anaesthesia. IA injection of saline 10 ml and saline 10 ml with morphine 2 mg were both associated with pain relief
*Risk of bias*
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Quote: list of random numbers
Allocation concealment (selection bias)	Low risk	Quote: by a person not involved in the study. Block size and randomisation codes were not revealed to the investigators until all measurements and calculations had been entered into the database.
Incomplete outcome data (attrition bias) All outcomes	Low risk	Comment: no participants withdrew from the study
Selective reporting (reporting bias)	Unclear risk	Comment: insufficient data were available
Blinding of participants and personnel (performance bias) All outcomes	Unclear risk	Quote: maintain blinding of both participants and examiner throughout the study
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	Comment: not specified
Size	High risk	Comment: < 50 participants per treatment arm

*Study characteristics*
Methods	A double‐blind, randomised trial
Participants	124 patients elective outpatient arthroscopic knee surgery
Interventions	Group 1 (n = 23) received 15 ml of 0.5% bupivacaine and 5 ml of normal saline Group 2 (n = 26) received 15 ml of normal saline and 2 mg of morphine in 5 ml of normal saline Group 3 (n = 25) received 15 ml of 0.5% bupivacaine with 1 mg morphine in 5 ml of normal saline Group 4 (n = 22) received 15 ml of 0.5% bupivacaine with 2 mg morphine in 5 ml of normal saline Group 5 (n = 28) received 20 ml of normal saline
Outcomes	VAS, supplemental analgesic, weightbearing status
Notes	Our results failed to show any beneficial effect of morphine used for postoperative analgesia, either alone or in combination with bupivacaine
*Risk of bias*
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Comment: not specified
Allocation concealment (selection bias)	Low risk	Quote: Syringes were coded by the pharmacy with their contents unknown to the investigators
Incomplete outcome data (attrition bias) All outcomes	Low risk	Comment: no participants withdrew from the study
Selective reporting (reporting bias)	High risk	Comment: no exact data were provided. Only one figure showed the results. No exact P value was provided
Blinding of participants and personnel (performance bias) All outcomes	Unclear risk	Comment: not specified
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	Comment: not specified
Size	High risk	Comment: < 50 participants per treatment arm

*Study characteristics*
Methods	Randomised, double blind
Participants	40 patients (60 recruited, 40 included in analysis), scheduled for day‐case knee arthroscopy, such as meniscectomies, removal of loose body, or shaving and lavage of degenerated cartilage
Interventions	IA injection was given through a 20‐gauge catheter IA saline 1 ml (placebo) or IA morphine 5 mg and to immediate or delayed removal of IA catheter
Outcomes	VAS, VRS, time to and consumption of rescue analgesic drugs, side effects
Notes
*Risk of bias*
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Quote: by using a list of random numbers
Allocation concealment (selection bias)	Low risk	Quote: by the senior author who did not deal with the participants. Block size and randomisation codes were not revealed to the investigators until all measurements and calculations had been entered into the database
Incomplete outcome data (attrition bias) All outcomes	Low risk	Quote: None lost to follow‐up during hospital day and 9 participants lost to follow‐up during 1 and 2 day. Used all the included participants' data in the analysis
Selective reporting (reporting bias)	Unclear risk	Comment: insufficient data were available
Blinding of participants and personnel (performance bias) All outcomes	Low risk	Quote: To maintain blinding of both participants and examiner throughout the study, syringes for each patient were prepared in the morning of surgery by a nurse not involved in the treatment or assessment of the participants
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	Comment: not specified
Size	High risk	Comment: < 50 participants per treatment arm

*Study characteristics*
Methods	Double blind, randomised
Participants	69 patients, arthroscopic surgery
Interventions	Group I (n = 23) fentanyl in 20 ml saline; Group II (n = 24) 3 mg morphine in 20 ml saline; Group III (n = 22) 20 ml saline
Outcomes	VAS, consumption of rescue medication
Notes	Better postoperative analgesia was achieved with 50 Hg IA fentanyl than with 3 mg IA morphine
*Risk of bias*
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Comment: not specified
Allocation concealment (selection bias)	Low risk	Quote: using a sealed envelope technique
Incomplete outcome data (attrition bias) All outcomes	Low risk	Comment: no participants withdraw from the study
Selective reporting (reporting bias)	Unclear risk	Comment: insufficient data were available
Blinding of participants and personnel (performance bias) All outcomes	Low risk	Quote: by an observer blinded to patient group assignment
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	Comment: not specified, double blind, probably done
Size	High risk	Comment: < 50 participants per treatment arm

*Study characteristics*
Methods	A randomised double‐blind placebo‐controlled trial
Participants	60 patients (ASA 1 or 2) scheduled to undergo day‐case arthroscopic knee surgery
Interventions	(i) 0.9% saline 20 ml, n=20 (ii) morphine 1 mg in 20 ml of 0.9% saline 20 ml, n=19 (iii) buprenorphine 30 jig in 20 ml of 0.9% saline 20 ml, n=19
Outcomes	VRS, consumption of rescue medication
Notes	There were no differences in pain scores among groups, first 24 h postoperatively
*Risk of bias*
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Comment: not specified
Allocation concealment (selection bias)	Low risk	Quote: these solutions were prepared by an anaesthetist who was not one of the investigators
Incomplete outcome data (attrition bias) All outcomes	Low risk	Comment: no participants withdrew from the study
Selective reporting (reporting bias)	High risk	Comment: no exact data were provided. Only one figure showed the results. No exact P value was provided
Blinding of participants and personnel (performance bias) All outcomes	Low risk	Quote: These solutions were prepared by an anaesthetist who was not one of the investigators; the patient and the investigator were blinded to the identity of the test solution
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	Comment: not specified
Size	High risk	Comment: < 50 participants per treatment arm

Study	Reason for exclusion
Alvarez‐Cabo 1998	Participants received spinal anaesthesia
De Andres 1993	Some participants were younger than 15 years old
Lehrberger 1994	Group size. Number of participants in group who completed the study less than ten
Stein 1991	Group size. Number of participants in group who completed the study less than ten