Pain relief for outpatient hysteroscopy

Gaity Ahmad; Sushant Saluja; Helena O'Flynn; Alessandra Sorrentino; Daniel Leach; Andrew Watson

doi:10.1002/14651858.CD007710.pub3

. 2017 Oct 5;2017(10):CD007710. doi: 10.1002/14651858.CD007710.pub3

Pain relief for outpatient hysteroscopy

Gaity Ahmad ¹, Sushant Saluja ¹, Helena O'Flynn ¹, Alessandra Sorrentino ², Daniel Leach ³, Andrew Watson ^4,^✉

Editor: Cochrane Gynaecology and Fertility Group

PMCID: PMC6485917 PMID: 35611933

Abstract

Background

Hysteroscopy is increasingly performed in an outpatient setting. Pain is the primary reason for abandonment of procedure or incomplete assessment. There is no consensus upon routine use of analgesia during hysteroscopy.

Objectives

To assess the effectiveness and safety of pharmacological interventions for pain relief in women undergoing outpatient hysteroscopy, compared with placebo, no treatment or other pharmacological therapies.

Search methods

In September 2016 we searched the Cochrane Gynaecology and Fertility (CGF) Trials Register, CENTRAL, MEDLINE, Embase, PsycINFO, CINAHL and two trials registers (ClinicalTrials.gov and WHO ICTRP), together with reference checking and contact with study authors and experts.

Selection criteria

We included randomised controlled trials (RCTs) comparing use of pharmacological interventions with other pharmacological interventions and pharmacological interventions versus placebo or no treatment.

Data collection and analysis

We used standard methodological procedures expected by Cochrane. Our primary outcome was mean pain score.

Main results

We included 32 RCTS (3304 participants), of which only 19 reported data suitable for analysis. Most studies were at unclear or high risk of bias in most of the domains assessed. The evidence was low or very low quality, mainly due to risk of bias and imprecision. Baseline pain scores were relatively low in all groups.

Analgesic versus placebo or no treatment

Local anaesthetics

Local anaesthetics reduced mean pain scores during the procedure [(SMD) ‐0.29, 95% CI ‐0.39 to ‐0.19, 10 RCTs, 1496 women, I² = 80%, low‐quality evidence)] and within 30 minutes (SMD 0.50, 95% CI ‐0.67 to ‐0.33, 5 RCTs, 545 women, I² = 43%, low‐quality evidence). This translates to a difference of up to 7 mm on a 0‐10 cm visual analogue scale (VAS) during the procedure and up to 13 mm within 30 minutes, which is unlikely to be clinically meaningful. There was no clear evidence of a difference between the groups in mean pain scores after > 30 minutes (SMD ‐0.11, 95% CI ‐0.30 to 0.07, 4 RCTs, 450 women, I² = 0%, low‐quality evidence), or in rates of vasovagal reactions (OR 0.70, 95% CI 0.43 to 1.13, 8 RCTs, 1309 women, I² = 66%, very low‐quality evidence). There was insufficient evidence to determine whether there was a difference in rates of non‐pelvic pain (OR 1.76, 95% CI 0.53 to 5.80, 1 RCT, 99 women, very low‐quality evidence).

Nonsteroidal anti‐inflammatory drugs (NSAIDs)

There was insufficient evidence to determine whether there was a difference between the groups in mean pain scores during the procedure (SMD ‐0.18, 95% CI ‐0.35 to 0.00, 3 RCTs, 521 women, I² = 81%, low‐quality evidence). Pain scores were lower in the NSAIDs group within 30 minutes (SMD ‐0.25, 95% CI ‐0.46 to ‐0.04, 2 RCTs, 340 women, I²=29%, low‐quality evidence) and at over 30 minutes (SMD ‐0.27, 95% CI ‐0.49 to ‐0.05, 2 RCTs, 321 women, I² = 78%, low‐quality evidence). This equates to maximum differences of under 7.5 mm on a 0‐10 cm scale, which are unlikely to be clinically significant. One RCT (181 women) reported adverse events: there was insufficient evidence to determine whether there was a difference between the groups in vasovagal reactions (OR 0.76, 95% CI 0.20 to 2.94, very low‐quality evidence). For other reported adverse events (non pelvic pain and allergic reactions) evidence was lacking.

Opioids

One RCT utilised sublingual buprenorphine and one utilised oral tramadol. Data on pain scores during the procedure were unsuitable for pooling due to inconsistency. Tramadol was associated with a benefit of up to 22 mm on a 0‐10 cm scale (SMD ‐0.76, 95% CI ‐1.10 to ‐0.42, 1 RCT, 140 women). However, the effect estimate for this outcome for sublingual opioids did not support a benefit from the intervention (SMD 0.08, 95% CI ‐0.22 to 0.39, 164 women). Compared with placebo, the pain score within 30 minutes of the procedure was reduced in the tramadol group, with a difference of up to 17mm on a 0‐10cm scale (SMD ‐0.57, 95% CI ‐0.91 to ‐0.23 , 1 RCT, 140 women, low‐quality evidence. There was no clear evidence of a difference between the tramadol and placebo groups at over 30 minutes (SMD ‐0.17, 95% CI ‐0.51 to 0.16, 1 RCT, 140 women, low‐quality evidence). Nausea and vomiting occurred in 39% of the buprenorphine group, and in none of the placebo group (OR 107.55, 95% CI 6.44 to 1796.46)

Analgesic versus any other analgesic

Some comparisons did not report pain scores at all time frames of interest, and none reported data on adverse events.

One RCT (84 women) compared local intracervical anaesthesia versus combined intracervical and paracervical anaesthesia. Pain scores were higher in the group with local intracervical anaesthesia during the procedure (SMD 4.27, 95% CI 3.49 to 5.06, very low‐quality evidence), within 30 minutes (SMD 1.55, 95% CI 1.06 to 2.05, very low‐quality evidence) and at more than 30 minutes (SMD 3.47, 95% CI 2.78 to 4.15, very low‐quality evidence). This translates to a possible benefit in the combined group of up to 12 mm on a 0‐10 cm scale during the procedure. Benefits at longer follow‐up were smaller.

One RCT compared antispasmodic + NSAID versus local paracervical anaesthesia. Pain scores were lower in the NSAID group than in the local anaesthesia group (during procedure: SMD ‐1.40, 95% CI ‐1.90 to ‐0.91; >30 minutes after procedure: SMD ‐0.87, 95% CI ‐1.33 to ‐0.41; 80 women, very low‐quality evidence). This suggests a possible benefit of during the procedure of up to 23 mm on a 0‐10 VAS scale and up to 11 mm >30 minutes after the procedure.

Other comparisons included local intracervical anaesthesia versus combined intracervical, paracervical and topical anaesthesia, and opioid versus NSAIDs. Findings were inconclusive.

Authors' conclusions

There was no consistent good‐quality evidence of a clinically meaningful difference in safety or effectiveness between different types of pain relief compared with each other or with placebo or no treatment in women undergoing outpatient hysteroscopy.

Plain language summary

Pain relief for hysteroscopy as an outpatient

Review question

The purpose of this review was to determine which, if any, pain relief drug is effective at reducing the discomfort experienced with outpatient hysteroscopy.

Background

Hysteroscopy is a diagnostic test undertaken to identify the cause of abnormal uterine bleeding. Hysteroscopy involves fluid or gas being injected via the cervix into the uterus, enabling the visualisation of the cervical canal and uterine cavity with a hysteroscope, which can be painful. There is disagreement regarding the best form of pain relief during the procedure.

Study characteristics

We included 32 randomised controlled trials (RCTs), with 3304 participants, of which only 19 reported data suitable for analysis.

All studies took place in a clinical setting. The age of the participants ranged from 33 to 61 years. Studies took place in Australia, Belgium, Brazil, Canada, China, France, Greece, India, Italy, Spain, Taiwan, UK, USA. Baseline pain scores in all groups were relatively low. The evidence is current to September 2016.

Key results

Quality of the evidence

Most studies were at unclear or high risk of bias in most of the domains assessed. The evidence was low or very low quality, mainly due to risk of bias and imprecision.

Summary of findings

Background

Description of the condition

Outpatient hysteroscopy is an established diagnostic gynaecological procedure that enables a clinician to visualise the uterine cavity and take endometrial biopsies or perform small intracavitary procedures as required. Hysteroscopy involves the use of miniaturised endoscopic equipment to examine the uterine cavity. As one of the principal investigations of abnormal uterine bleeding, hysteroscopy has an integral role in the identification of structural abnormalities of the endometrium (NICE guidelines 2007). Outpatient hysteroscopy can also be employed for procedures such as removal of lost intrauterine devices, endometrial polypectomy and work‐up of reproductive problems (Robinson 2013; Sinha 2007).

The procedure is performed without the need for theatre facilities and regional or general anaesthesia. Advantages to both the healthcare providers and women, include decreased complication rates, shorter recovery time, decreased costs and the scope for a 'see and treat' service (Ma 2016; Saridogan 2010). Although safe, studies of the acceptability of outpatient hysteroscopy have displayed various completion rates, ranging between 77% to 97.2% (Agostini 2003; Critchley 2004; De Iaco 2000; De Jong 1990), with pain cited as the most common cause for failure to complete the investigation (Critchley 2004; Jivraj 2004; Nagele 1997; Paschopoulos 1997). During hysterosalpingography (HSG), pain peaks from the time of instillation of the contrast media until five minutes after the procedure; the pain starts to decrease rapidly between five and 10 minutes after the procedure so that at 30 minutes most women classify it as a 'discomfort' (Owens 1985).

The pain experienced is due to several factors including cervical instrumentation, uterine distension and peritoneal irritation from spill of dilatation media. Pain stimuli from the cervix and vagina are conducted by the pelvic splanchnic nerves whereas pain sensation from intraperitoneal structures, such as the uterine body, is conducted by the hypogastric nerves (Moore 2006). Destruction of the endometrium and endometrial biopsy can cause further pain as they may induce uterine contraction (Zupi 1995). Additional delayed pain is also caused by the release of prostaglandins from the cervical manipulation as well as distension of the uterus. Furthermore, blind cervical dilatation, cervical stenosis or tortuosity may increase chances of uterine lacerations (Jansen 2000; Pasini 2001).

Other factors that may impact on pain experienced include the type and size of hysteroscope used, choice of distension media and technique employed. It is recommended that miniature hysteroscopes (2.7 mm, with a 3 mm to 3.5 mm sheath) should be used for diagnostic outpatient hysteroscopy to reduce discomfort (Green‐top Guidelines No. 59; Paulo 2015). Use of a vaginoscopic approach has also been advocated to reduce pain experienced (Cooper 2010b). Systematic reviews have demonstrated varied results for difference in pain experienced based upon distension media used although fewer vasovagal episodes are reported with the use of saline when compared to carbon dioxide (Cooper 2011; Craciunas 2013).

Description of the intervention

There are numerous studies that have reported varied outcomes in terms of the effectiveness of different pharmacological interventions for pain relief during hysteroscopy. We have investigated the effect of any one type of pharmacological intervention (opioids, local anaesthetic, non‐steroidal anti‐inflammatory drugs (NSAIDs) or paracetamol) for pain relief during hysteroscopy compared with another type, or with placebo.

How the intervention might work

De Iaco 2000 observed that 34.8% of women who undergo anaesthesia‐free diagnostic hysteroscopy reported severe pain. Pain is mainly produced when the speculum or tenaculum are placed, with cervical dilation, passage of the hysteroscope through the cervical canal, and distension of the uterus with fluid. Operative procedures that damage the endometrial wall are also painful. The pain experienced can have a negative impact on the woman's ability to fully co‐operate with the procedure. Through successful anaesthesia (including paracervical and intracervical anaesthesia) and application of topical agents in the cervical canal and endometrial cavity, we hope to target more than one site. This will result in pain relief and better tolerance of the procedure.

Why it is important to do this review

A study reported a wide variation in clinical practice in the UK in terms of the use of analgesia for outpatient hysteroscopy (O'Flynn 2011). Generally, while pain relief is common in clinical practice the evidence is uncertain on its effect. The review aims to assess the effectiveness of the various pain relief methods available.

The Royal College of Obstetricians and Gynaecologists recommend that women with no contraindications should consider taking a standard dose of a NSAID at least an hour before their appointment (Green‐top Guidelines No. 59). These guidelines also recommended consideration of the use of topical local cervical anaesthesia during cervical tenaculum use.

With the vast potential for its use in the primary care setting, a consensus upon the safest and most efficacious method of pain relief for outpatient hysteroscopy would be advantageous.

The purpose of the meta‐analysis was therefore to compare the effectiveness and safety of different types of pharmacological interventions for pain relief in women undergoing outpatient hysteroscopy.

Objectives

Methods

Criteria for considering studies for this review

Types of studies

Randomised controlled trials (RCTs) comparing any one type of pharmacological intervention for pain relief during outpatient hysteroscopy with another type, placebo, or no treatment. We excluded non‐randomised studies (e.g. studies with inadequate sequence generation, patient numbers) as they are associated with a high risk of bias. We also excluded cross‐over trials, as the design is not valid in this context.

Types of participants

Adult (aged over 18 years) women attending for an outpatient hysteroscopy.

Types of interventions

We analysed the following interventions.

Analgesics (topical or oral) versus placebo or no treatment
- Opioids versus placebo or no treatment
- Local anaesthetics versus placebo or no treatment
- NSAIDs versus placebo or no treatment
- Paracetamol or similar versus placebo or no treatment.
Analgesics (topical or oral or inhaled) versus other analgesics
- Opioids versus paracetamol
- Opioids versus local anaesthetics
- Opioids versus NSAIDS
- Local anaesthetics versus NSAIDs
- NSAIDs versus paracetamol
- Paracetamol versus local anaesthetics (LA)
- Any analgesic versus any other analgesic

We performed subgroup analysis by route of administration. Dosages and routes used in the studies are displayed in Table 6.

1. Dosing of local anaesthetics.

Study	Route	Dose
Al‐Sunaidi 2007	Intracervical versus intra and paracervical	Local intracervical anaesthesia compared to combined intracervical and paracervical anaesthesia. 0.5% bupivacaine hydrochloride into anterior wall of cervix compared to 0.5% bupivacaine hydrochloride into anterior wall of cervix plus bupivacaine into lateral vaginal fornix at 3 and 9 o’clock at 10 mm depth
Arnau 2013	Endocervical topical cream	3 mL of EMLA cream 5% or 3 mL ultrasound gel applied in the endocervical canal 10 min before surgery with 5 mL needleless syringe via speculum. A subsequent application of either gel was made with a swab at ectocervix level during vaginoscopic approach
Broadbent 1992	Intracervical	Intracervical injection of 10 mL of lignocaine 1% with 1:200 000 adrenaline or normal saline was injected into the cervix at 1, 5, 7, and 11 o’clock
Cicinelli 1997	Transcervical	2 mL of 2% mepivacaine or 2 mL of 0.9% saline injected trans cervically (inserted up the cervical canal to the internal os)
Cicinelli 1998	Paracervical	10 mL of 1.5% mepivacaine or saline injected into junction of cervix and vagina (4 and 8 o'clock positions)
Clark 1996	Topical spray	10 mL of 2% lignocaine gel, placebo gel or no gel administered into cervical canal. Some women received intracervical lignocaine block if determined to need cervical dilatation
Costello 1998	Intrauterine	5 mL of 2% lignocaine or 5 mL or 0.9% saline injected into cervical canal and uterine cavity via hysteroscope
Esteve 2002	Intracervical	Intracervical application at 1, 5, 7 and 11 o'clock positions of 4 x 2 mL ampoules of 2% lidocaine hydrochloride ampoules or saline
Finikiotis 1992	Paracervical versus uterosacral	20 mL of 1% lignocaine paracervical block or 2 mL of 2% lignocaine uterosacral block
Giorda 2000	Paracervical	20 mL of 1% mepivacaine injected paracervically at 3, 5, 7 and 9 o'clock position of the junction of cervix and vagina at least 5 min before the procedure
Kabli 2008	Intracervical versus intracervical and intrauterine	2 mL of 1% lidocaine into anterior wall. Distension medium of either saline only or 18 mL of lidocaine in 250 mL of saline
Kokanali 2013	Intracervical	Intracervical local anaesthesia (10 mL of 1% prilocaine) was applied at the 4 and  8 o’clock position on the posterior lip of the cervix in divided doses
Lau 1999	Paracervical	Paracervical block at 3, 5, 7 and 9 o'clock positions of 10 mL of 2% lignocaine or saline, 5 min prior to procedure
Lau 2000	Transcervical	Transcervical intrauterine instillation of 5 mL of 2% lignocaine or normal saline into the uterine cavity 5 min before procedure
Lukes 2015	Para/intracervical versus intracervical only anaesthetic block	The para/intracervical group received a total of 37 cc of anaesthesia injected at 6 different sites. 2‐3 min prior to the injection, topical 1% lidocaine was applied to the cervix in this group. The intracervical group received a total of 22 cc of anaesthetic given at 3 different injection sites
Makris 2001	Intracervical	1 mL to 3 mL (30 to 90 mg) of mepivacaine 3% or saline administered intracervically 3 min prior to procedure
Mercorio 2002	Intracervical versus NSAID	5 mL mepivacaine 2% intracervically up to the level of the internal os or one tablet of dexketoprofen given 1 h before the procedure
Mohammadi 2015	Transcervical intrauterine lidocaine instillation versus rectal diclofenac	5 mL of 2% lidocaine or the same volume of saline was instilled through the endocervix into the uterine cavity with an 18‐gauge angiocatheter 3 min prior to the procedure
Senturk 2016	Intracervical	The second group was administered with a 1000 mL distention medium containing 18 mL lidocaine per 250 mL (Jetokain ampoule 20 mg 2% Adeka, Samsun, Turkey)
Soriano 2000	Topical spray	30 mg (3 metered doses) of lignocaine or placebo sprayed onto surface of cervix and cervical canal through 360° 5 min prior to procedure
Stigliano 1997	Topical cream versus topical spray	1 cm³ of 5% prilocaine cream onto esocervix and 2 cm³ inserted 3 cm into cervical canal 10 min before procedure or 20 mg of lidocaine spray directed onto esocervix and 20 mg 3 cm into cervical canal immediately before procedure or no intervention
Van den Bosch 2011	Intrauterine	Unspecified volume of Instillagel (contains 2% lidocaine) or Endosgel (does not contain local anaesthetic) warmed to 37°C and instilled via a 2 mm neonatal suction catheter as part of sonography procedure 30 min prior to hysteroscopy
Vercellini 1994	Paracervical	Paracervical block at 3, 5, 7 and 9 o'clock positions of 10 mL 1% mepivacaine 5 min before procedure or no intervention given
Wong 2000	Topical gel	4 mL of 2% lignocaine or placebo gel applied onto cervix before procedure
Zupi 1995	Intrauterine	5 mL of 2% mepivacaine or 5 mL of saline intrauterine via 3 mm catheter

Local anaesthetic compared to placebo or no treatment for outpatient hysteroscopy
Population: women undergoing outpatient hysteroscopy  Setting: outpatient hysteroscopy clinic  Intervention: local anaesthetic  Comparison: placebo or no treatment
Outcomes	*Anticipated absolute effects^ (95% CI)**		Relative effect  (95% CI)	№ of participants  (studies)	Quality of the evidence  (GRADE)	Comments
Outcomes	Risk with placebo or no treatment	Risk with local anaesthetic	Relative effect  (95% CI)	№ of participants  (studies)	Quality of the evidence  (GRADE)	Comments
Pain score during procedure	The mean pain score ranged from 0.86 to 4.3 on a 0‐10 cm VAS	SMD 0.29 SD lower  (0.39 lower to 0.19 lower)	‐	1496  (10 studies)	⊕⊕⊝⊝  low^1,2	This suggests a marginally lower pain score in the intervention group, equating to up to 7 mm on a 0‐10 cm point VAS. This is unlikely to be clinically significant
Pain score within 30 min of procedure	The mean pain score ranged from 1.8 to 6.3 on a 0‐10 cm VAS	SMD 0.5 SD lower  (0.67 lower to 0.33 lower)	‐	545  (5 studies)	⊕⊕⊝⊝  low¹	This suggests a marginally lower pain score in the intervention group, equating to up to 13 mm on a 0‐10 cm point VAS. This is unlikely to be clinically significant
Pain score more than 30 min after procedure	The mean pain score ranged from 0.62 to 1.8 on a 0‐10 cm VAS	SMD 0.11 SD lower  (0.3 lower to 0.07 higher)	‐	450  (4 studies)	⊕⊕⊝⊝  low^3,4	There was no clear evidence of a difference between the groups.
Vasovagal reaction	63 per 1000	45 per 1000  (28 to 71)	OR 0.70  (0.43 to 1.13)	1309  (8 studies)	⊕⊝⊝⊝  very low^1,4	There was no clear evidence of a difference between the groups
Non‐pelvic pain	100 per 1000	164 per 1000  (56 to 392)	OR 1.76  (0.53 to 5.80)	99  (1 study)	⊕⊝⊝⊝  very low^1,3	There was insufficient evidence to determine whether there is a difference between the groups
*The risk in the intervention group (and its 95% confidence interval) is based on the mean risk in the comparison group and the relative effect of the intervention (and its 95% CI).    CI: confidence interval; SMD: standardised mean difference; SD: standard deviation; OR: odds 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

Oral NSAID compared to placebo or no treatment for outpatient hysteroscopy
Population: women undergoing outpatient hysteroscopy  Setting: outpatient hysteroscopy clinic  Intervention: oral nonsteroidal anti‐inflammatory drug (NSAID)  Comparison: placebo or no treatment
Outcomes	*Anticipated absolute effects^ (95% CI)**		Relative effect  (95% CI)	№ of participants  (studies)	Quality of the evidence  (GRADE)	Comments
Outcomes	Risk with placebo or no treatment	Risk with oral NSAID	Relative effect  (95% CI)	№ of participants  (studies)	Quality of the evidence  (GRADE)	Comments
Pain score during procedure	The mean pain score ranged from 2.1 to 5.92 on a 0‐10 cm VAS	SMD 0.18 lower (0.35 lower to 0.00)	‐	521  (3 studies)	⊕⊕⊝⊝  low^1,2,3	There was no clear evidence of a difference between the groups
Pain score within 30 min of procedure	The mean pain score ranged from 0.65 to 2.02 on a 0‐10 cm VAS	SMD 0.25 SD lower  (0.46 lower to 0.04 lower)	‐	340 participants  (2 studies)	⊕⊕⊝⊝  low^1,2	The evidence suggests a benefit in the NSAID group, equivalent to up to 2 mm on a 0‐10 cm VAS. This is unlikely to be clinically significant
Pain score more than 30 min after procedure	The mean pain score ranged from 1.2 to 1.55 on a 0‐10 cm VAS	SMD 0.27 lower  (0.49 lower to 0.05 lower)	‐	321  (2 studies)	⊕⊕⊝⊝  low^1,2,3	The evidence suggests a benefit in the NSAID group, equivalent to up to 7mm on a 0‐10 cm VAS. This is unlikely to be clinically significant
Vasovagal reaction	56 per 1,000	43 per 1,000  (12 to 149)	OR 0.76  (0.20 to 2.94)	181 participants  (1 study)	⊕⊕⊝⊝  very low^1,4	Only 9 events There was insufficient evidence to determine whether there is a difference between the groups
Non pelvic pain	Not calculable: no events in one group		OR 2.93  (0.12 to 72.99)	181 participants  (1 study)	⊕⊕⊝⊝  very low^1,4	Only 1 event There was insufficient evidence to determine whether there is a difference between the groups
Allergic reaction	Not calculable: no events in one group		OR 2.93  (0.12 to 72.99)	181 participants  (1 study)	⊕⊕⊝⊝  very low^1,4	Only 1 event There was insufficient evidence to determine whether there is a difference between the groups
*The risk in the intervention group (and its 95% confidence interval) is based on the mean risk in the comparison group and the relative effect of the intervention (and its 95% CI).    CI: confidence interval; SMD: standardised mean difference; SD: standard deviation; NSAID: nonsteroidal anti‐inflammatory drugs; OR: odds 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

Opioid compared to placebo or no treatment for outpatient hysteroscopy
Population: women undergoing outpatient hysteroscopy  Setting: outpatient hysteroscopy clinic  Intervention: oral or sublingual opioid  Comparison: placebo or no treatment
Outcomes	*Anticipated absolute effects^ (95% CI)**		Relative effect  (95% CI)	№ of participants  (studies)	Quality of the evidence  (GRADE)	Comments
Outcomes	Risk with placebo or no treatment	Risk with opioid	Relative effect  (95% CI)	№ of participants  (studies)	Quality of the evidence  (GRADE)	Comments
Pain score during procedure: oral opioid	The mean pain score was 5.92 mm on a 0‐10 cm VAS	Oral opioid: SMD was 0.76 lower (95% CI 1.10 lower to 0.42 lower)	‐	140 (1 study)	low^1,2	Data from the two studies were unsuitable for pooling due to extreme heterogeneity (I² = 92%) with conflicting directions of effect. There is a suggestion of benefit from oral opioid, equating to a difference of up to 22 mm on a 0‐10 cm VAS. This may possibly be clinically significant
Pain score during procedure: sublingual opioid	The mean pain score was from 3.2 mm on a 0‐10 cm VAS	Sublingual opioid: SMD was 0.08 higher (95% CI 0.22 lower to 0.39 higher)		164 (1 study)	low^1,2
Pain score within 30 min of procedure oral opioid	The mean pain score was 3.27 mm on a 0‐10 cm VAS	SMD ‐0.57 lower  (‐0.91 to ‐0.23 lower)	‐	140  (1 study)	low^1,2	There is a suggestion of benefit from oral opioid, equating to a difference of up to 17 mm on a 0‐10 cm VAS. This may possibly be clinically significant
Pain score more than 30 min after procedure oral opioid	The mean pain score was 0.77 on a 0‐10 cm VAS	SMD 0.17 lower  ( 0.51 lower to 0.16 higher)	‐	140  (1 study)	very low^1,2,3
Nausea and vomiting sublingual opioid	Not calculable as all 31 events were in the opioid group and none in the placebo group		OR 107.55 (6.44 to 1796.46)	164  (1 study)	low¹	There were 4 events in the intervention group of the oral opioid study, but events in the placebo group were not reported
*The risk in the intervention group (and its 95% confidence interval) is based on the mean risk in the comparison group and the relative effect of the intervention (and its 95% CI).    CI: confidence interval; SMD: standardised mean difference; SD: standard deviation; OR: odds 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

Intracervical versus combined intracervical and paracervical anaesthesia
Population: women undergoing outpatient hysteroscopy  Setting: outpatient hysteroscopy clinic  Intervention: localintracervical anaesthesia  Comparison: combined intracervical and paracervical anaesthesia
Outcomes	*Anticipated absolute effects^ (95% CI)**		Relative effect  (95% CI)	No of participants  (studies)	Quality of the evidence  (GRADE)	Comments
	Risk with combined intracervical and paracervical anaesthesia	Risk with local intracervical anaesthesia
Pain score during procedure	The mean pain score was 2.1 (SD 0.2) on a 0‐10 cm VAS	SMD 4.27 SD higher (3.49 higher to 5.06 higher)	‐	84 participants  (1 study)	⊕⊝⊝⊝  very low^1,2	This suggests a significant benefit from combined anaesthesia, equating to up to 12 mm on a 0‐10 cm point VAS. This is unlikely to be be clinically significant
Pain score within 30 minutes of the procedure	The mean pain score was 1.5 (SD 0.3) on a 0‐10 cm VAS	(SMD 1.55, 95% CI 1.06 higher to 2.05 higher)	‐	84 participants (1 study)	⊕⊝⊝⊝  very low^1,2	This suggests a significant benefit from combined anaesthesia, equating to up to 5 mm on a 0‐10 cm point VAS. This is unlikely to be clinically significant
Pain score more than 30 minutes of the procedure	The mean pain score was 1 (SD 0.2) on a 0‐10 cm VAS	(SMD 3.47, 95% CI 2.78 higher to 4.15 higher)	‐	84 participants (1 study)	⊕⊝⊝⊝  very low^1,2	This suggests a significant benefit from combined anaesthesia, equating to up to 8 mm on a 0‐10 cm point VAS. This is unlikely to be clinically significant
Adverse effects	No data available
*The risk in the intervention group (and its 95% confidence interval) is based on the mean risk in the comparison group and the relative effect of the intervention (and its 95% CI). CI: confidence interval; SMD: standardised mean difference; SD: standard deviation; 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

Antispasmodic plus NSAID versus paracervical anaesthesia
Population: women undergoing outpatient hysteroscopy  Setting: outpatient hysteroscopy clinic  Intervention: antispasmodic plus nonsteroidal anti‐inflammatory drugs (NSAID)  Comparison: local paracervical anaesthesia
Outcomes	*Illustrative comparative risks (95% CI)**		Relative effect  (95% CI)	No of participants  (studies)	Quality of the evidence  (GRADE)	Comments
	Paracervical anaesthesia	Antispasmodic plus NSAID
Pain score during procedure	The mean pain score was 5.93 (SD ‐1.26) on a 0‐10 cm VAS	SMD 1.40 SD lower (1.90 lower to 0.91 lower)	‐	80  (1 study)	⊕⊝⊝⊝  very low^1,2	This suggests a beneficial effect in the drotaverine hydrochloride and mefenamic acid group, equating to up to 23 mm on a 0‐10 cm point VAS. This may possibly be clinically significant
Pain score within 30 minutes of the procedure	This outcome was not reported
Pain score more than 30 minutes after the procedure	The mean pain score was 2.53 (SD ‐0.81) on a 0‐10 cm VAS	SMD 0.87 SD lower  (1.33 lower to 0.41 lower)	‐	80  (1 study)	⊕⊝⊝⊝  very low^1,2	This suggests a beneficial effect in the drotaverine hydrochloride and mefenamic acid group, equating toup to11mm on a 0‐10 cm point VAS. This is unlikely to be clinically significant.
Adverse effects	This outcome was not reported
*The risk in the intervention group (and its 95% confidence interval) is based on the mean risk in the comparison group and the relative effect of the intervention (and its 95% CI). CI: confidence interval; SMD: standardised mean difference; SD: standard deviation; 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

Date	Event	Description
19 September 2017	New citation required and conclusions have changed	The conclusion of the review is as follows with the inclusion of 8 new studies: there was insufficient evidence of a clinically meaningful effect when different types of pain relief were compared with placebo or other types of pain relief in women undergoing outpatient hysteroscopy.
19 September 2017	New search has been performed	We have added 8 new studies to this review Hassan 2016a; Hassan 2016b; Kokanali 2013; Lukes 2015; Mohammadi 2015; Senturk 2016; Teran‐Alonso 2014; Van den Bosch 2011) and excluded one previously included study (Guida 2003).

Date	Event	Description
1 August 2010	Amended	Change in title as per discussion with editorial office
18 July 2010	Amended	The title of the comparison 1b) has been changed to 1b) Local anaesthetics and 1c) NSAIDs separately as requested by the peer review.

#	Query	Results
S33	S18 AND S32	145
S32	S19 OR S20 or S21 or S22 OR S23 OR S24 OR S25 OR S26 OR S27 OR S28 OR S29 OR S30 OR S31	1,075,314
S31	TX allocat* random*	5,255
S30	(MH "Quantitative Studies")	14,862
S29	(MH "Placebos")	9,815
S28	TX placebo*	39,502
S27	TX random* allocat*	5,255
S26	(MH "Random Assignment")	41,596
S25	TX randomi* control* trial*	109,973
S24	TX ( (singl* n1 blind) or (singl n1 mask) ) or TX ( (doubl n1 blind) or (doubl n1 mask) ) or TX ( (tripl n1 blind) or (tripl n1 mask) ) or TX ( (trebl n1 blind) or (trebl n1 mask*) )	852,017
S23	TX ( (trebl* n1 blind) or (trebl n1 mask*) )	147
S22	TX ( (trebl* n1 blind) or (trebl n1 mask*) )	0
S21	TX clinic* n1 trial*	189,469
S20	PT Clinical trial	79,715
S19	(MH "Clinical Trials+")	202,772
S18	S4 AND S17	553
S17	S5 OR S6 OR S7 OR S8 OR S9 OR S10 OR S11 OR S12 OR S13 OR S14 OR S15 OR S16	2,175,158
S16	TX (pain N2 relief)	10,274
S15	TX Benzodiazep*	5,542
S14	TX morphine	7,596
S13	TX (hypnovel or midazolam)	2,717
S12	TX (fentanyl or remifentanil)	5,005
S11	TX opioid*	20,033
S10	TX (lidocaine or lignocaine)	4,758
S9	TX(intrauterine N5 block*)	5
S8	TX an?esthetic*	2,167,399
S7	(MM "Anesthetics, Local") OR (MH "Anesthesia, Local") OR "local anesthetics"	6,181
S6	(MM "Nerve Block") OR "nerve block"	6,264
S5	(MH "Analgesia+") OR "analgesia"	17,262
S4	S1 OR S2 OR S3	1,078
S3	TX(outpatient* N5 gyn?ecolog*)	78
S2	TX (office* N5 gyn?ecolog*)	33
S1	(MM "Hysteroscopy") OR "hysteroscopy"	972

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
1 Pain score	12		Std. Mean Difference (IV, Fixed, 95% CI)	Subtotals only
1.1 Pain score during procedure	10	1496	Std. Mean Difference (IV, Fixed, 95% CI)	‐0.29 [‐0.39, ‐0.19]
1.2 Pain score within 30 min of procedure	5	545	Std. Mean Difference (IV, Fixed, 95% CI)	‐0.50 [‐0.67, ‐0.33]
1.3 Pain score more than 30 min after procedure	4	450	Std. Mean Difference (IV, Fixed, 95% CI)	‐0.11 [‐0.30, 0.07]
2 Failure to complete procedure	7		Odds Ratio (M‐H, Fixed, 95% CI)	Subtotals only
2.1 Cervical stenosis	6	805	Odds Ratio (M‐H, Fixed, 95% CI)	1.23 [0.62, 2.43]
2.2 Pain	2	330	Odds Ratio (M‐H, Fixed, 95% CI)	0.29 [0.12, 0.69]
3 Adverse events	8		Odds Ratio (M‐H, Fixed, 95% CI)	Subtotals only
3.1 Vasovagal reaction	8	1309	Odds Ratio (M‐H, Fixed, 95% CI)	0.70 [0.43, 1.13]
3.2 Non‐pelvic pain	1	99	Odds Ratio (M‐H, Fixed, 95% CI)	1.76 [0.53, 5.80]

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
1 Pain score	2		Std. Mean Difference (IV, Fixed, 95% CI)	Totals not selected
1.1 Pain score during procedure	2		Std. Mean Difference (IV, Fixed, 95% CI)	0.0 [0.0, 0.0]
1.2 Pain score within 30 min of procedure	1		Std. Mean Difference (IV, Fixed, 95% CI)	0.0 [0.0, 0.0]
1.3 Pain score more than 30 min after procedure	1		Std. Mean Difference (IV, Fixed, 95% CI)	0.0 [0.0, 0.0]
2 Failure to complete procedure (due to pain)	1	140	Odds Ratio (M‐H, Fixed, 95% CI)	0.33 [0.01, 8.21]
3 Adverse effects	1	164	Odds Ratio (M‐H, Fixed, 95% CI)	107.55 [6.44, 1796.46]
3.1 Nausea and vomiting	1	164	Odds Ratio (M‐H, Fixed, 95% CI)	107.55 [6.44, 1796.46]

Methods	Randomised, double‐blinded, placebo‐controlled trial. No loss to follow‐up. Single‐centre trial at University of Bari, Italy. 40 women in intervention group and 40 women in placebo group
Participants	Postmenopausal women undergoing diagnostic hysteroscopy and endometrial biopsy. Mean age of women was 59 years in both the intervention and placebo group. No dropouts or exclusions
Interventions	2 mL of 2% mepivacaine or 2 mL of 0.9% saline injected transcervically (inserted up the cervical canal to the internal os) 5 min before the procedure
Outcomes	Mean pain score before the procedure, during the procedure, at endometrial biopsy and 15 min after the procedure. 20 cm VAS
Notes
*Risk of bias*
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	"Randomisation was done by opening sealed envelopes containing computer generated block‐randomisation numbers."
Allocation concealment (selection bias)	Unclear risk	"Randomisation was done by opening sealed envelopes containing computer generated block‐randomisation numbers."
Blinding (performance bias and detection bias)   All outcomes	Unclear risk	"Women were randomly and double blindly assigned."
Incomplete outcome data (attrition bias)   All outcomes	Low risk	No dropouts or exclusions. "Hysteroscopy performed in all patients."
Selective reporting (reporting bias)	Low risk	1 measuring instrument used (20 cm VAS) and data reported in 1 standard manner. All time points stated to have data collected were reported. Did not record change from baseline. Did not convert to dichotomous outcomes. Insufficient evidence of under reporting
Other bias	Low risk	Nothing detected

Methods	Randomised, double‐blinded, placebo controlled trial. No loss to follow‐up. Single‐centre trial at University of Bari, Italy. 36 women in intervention group and 36 women in placebo group
Participants	Postmenopausal women undergoing diagnostic hysteroscopy and endometrial biopsy. Mean age of women was 55 years in the intervention group and 56 in the placebo group. No dropouts or exclusions
Interventions	10 mL of 1.5% mepivacaine or saline injected into junction of cervix and vagina (4 and 8 o'clock positions) 10 min before procedure
Outcomes	Mean pain score before the procedure, during the procedure, at endometrial biopsy and 15 min after the procedure. 20 cm VAS
Notes
*Risk of bias*
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	"Randomisation was done by opening sealed envelopes containing computer generated block‐randomisation numbers."
Allocation concealment (selection bias)	Unclear risk	"Randomisation was done by opening sealed envelopes containing computer generated block‐randomisation numbers."
Blinding (performance bias and detection bias)   All outcomes	Unclear risk	"Women were randomly and double blindly assigned."
Incomplete outcome data (attrition bias)   All outcomes	Low risk	No dropouts or exclusions. "Hysteroscopy performed in all patients."
Selective reporting (reporting bias)	Low risk	1 measuring instrument used (20 cm VAS) and data reported in 1 standard manner. All time points stated to have data collected were reported. Did not record change from baseline. Did not convert to dichotomous outcomes. Insufficient evidence of under reporting
Other bias	Low risk	Nothing detected

Methods	Prospective randomised study. No loss to follow‐up. 60 women in both intervention groups Single‐centre trial at Flinders Medical Centre, Adelaide, Australia
Participants	Women undergoing hysteroscopy referred from GPs and other gynaecologists. Mean ages not stated. No exclusions
Interventions	20 mL of 1% lignocaine paracervical block or 2% uterosacral block
Outcomes	Pain during hysteroscopy. 10 cm VAS
Notes
*Risk of bias*
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	"The patients were randomised according to an odd or even unit record number."
Allocation concealment (selection bias)	Unclear risk	"The patients were randomised according to an odd or even unit record number."
Blinding (performance bias and detection bias)   All outcomes	Unclear risk	Not stated
Incomplete outcome data (attrition bias)   All outcomes	Unclear risk	No exclusions
Selective reporting (reporting bias)	High risk	Data grouped into 3 dichotomous outcomes. 1 measurement scale used. All time points stated to have data collected were reported. Did not record change from baseline
Other bias	Low risk	Nothing detected

Methods	A prospective, randomised trial. No loss to follow‐up. 36 women in local cervical group and 42 women in combined local cervical and intrauterine anaesthesia group. Single‐centre trial at Academic Teaching Centre, McGill University, Montreal, Canada
Participants	Infertile women undergoing hysteroscopy. Mean age of 37 years in local cervical group and 38 in combined local cervical and intrauterine anaesthesia group. 4 exclusions
Interventions	2 mL of 1% lidocaine into anterior wall. Distension medium of either saline only or 18 mL of lidocaine in 250 mL of saline
Outcomes	Pain score during hysteroscopy and 10, 30 and 60 min after hysteroscopy. 10‐point VAS
Notes	Co‐administration of lorazepam 10 mg orally 30 min before the procedure
*Risk of bias*
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	"Randomisation was done using a computer generated random number table."
Allocation concealment (selection bias)	Unclear risk	"Randomisation was done using a computer generated random number table."
Blinding (performance bias and detection bias)   All outcomes	High risk	Single blinded. "The patients were blinded but the operators were aware of the content of the solution."
Incomplete outcome data (attrition bias)   All outcomes	Low risk	"Three withdrew consent and one could not tolerate speculum examination."
Selective reporting (reporting bias)	Unclear risk	Data presented as medians. 1 measuring instrument used (10 cm VAS). All time points stated to have data collected were reported. Did not record change from baseline
Other bias	Unclear risk	Co‐administration of lorazepam 10 mg orally

Methods	Randomised, single‐blind study. No loss to follow‐up. Single‐centre trial at McGill University Health Centre, Montreal, Canada. 42 women in intervention group and 42 women in placebo group
Participants	Women undergoing hysteroscopy. Study included 84 women. Mean age of women was 36 years in intervention group and 35 in the placebo group. 1 woman dropped out of the study.
Interventions	Local intracervical anaesthesia compared to combined intracervical and paracervical anaesthesia. 0.5% bupivacaine hydrochloride into anterior wall of cervix compared to 0.5% bupivacaine hydrochloride into anterior wall of cervix plus bupivacaine into lateral vaginal fornix at 3 and 9 o’clock at 10 mm depth. Both interventions were performed 5 min before the procedure.
Outcomes	Mean pain score during the procedure, 10, 30 and 60 min after the procedure. 10‐point VAS used
Notes
*Risk of bias*
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	"Randomisation was done using a computer generated random table."
Allocation concealment (selection bias)	Unclear risk	Not described
Blinding (performance bias and detection bias)   All outcomes	High risk	Single‐blinded ‐ outcome assessor blinded
Incomplete outcome data (attrition bias)   All outcomes	Low risk	1 woman dropped out of the study but included in the final analysis (ITT). "A patient could not tolerate speculum examination and the procedure was aborted."
Selective reporting (reporting bias)	Low risk	1 measuring instrument used (10 cm VAS) and data reported in 1 standard manner. All time points stated to have data collected were reported. Did not record change from baseline. Did not convert to dichotomous outcomes
Other bias	Unclear risk	Co‐administration of 10 mg lorazepam 30 min prior to procedure

Methods	Randomised, non‐blinded, controlled study. Single‐centre trial at the Obstetrics and Gynaecology Service, Health Consortium of Terrassa, Barcelona, Spain
Participants	102 consecutive women scheduled for diagnostic or operative hysteroscopy were invited; 10 declined. Following randomisation there was 1 loss from the EMLA (intervention) group due to "deviation from protocol". No mean ages stated, no exclusions
Interventions	"Either 3 mL EMLA cream 5% or 3 mL ultrasound gel was applied in the endocervical canal 10 min before surgery, with a 5‐mL needleless syringe. A subsequent application of either gel was made with a swab at ectocervix level."
Outcomes	10 cm VAS. Pain score within 30 min of procedure. Women who completed the hysteroscopy were asked if they would recommend the procedure to other women, if they had wished to abandon the hysteroscopy and whether they would repeat the procedure if needed.
Notes	This study was added to this review in 2014.
*Risk of bias*
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	"Women were randomized to the EMLA or control group using computer‐generated random numbers."
Allocation concealment (selection bias)	Unclear risk	"Randomization was conducted with sealed envelopes containing computer‐generated randomization numbers."
Blinding (performance bias and detection bias)   All outcomes	High risk	Study is described as non‐blinded in the title, however in the discussion comments that "the main limitation of our study was that it was not double‐blinded, as we were unable to prepare a placebo with identical appearance and texture to the EMLA gel in our laboratory." This suggests that the operator would not be blinded, the participant may have been.
Incomplete outcome data (attrition bias)   All outcomes	Unclear risk	"One of the women in the EMLA group was excluded for protocol violation". Nature of violation not specified
Selective reporting (reporting bias)	Low risk	No pre‐published protocol seen, however all planned outcomes from methods are reported, together with adverse outcomes
Other bias	High risk	All women were given 600 mg ibuprofen (if allergic 1 g paracetamol) and 5 mg diazepam 2 h before procedure and misoprostol was administered only to postmenopausal women

Methods	Randomised, placebo‐controlled trial. No loss to follow‐up. Single‐centre trial at The Minimally Invasive Therapy Unit, University Department of Obstetrics & Gynaecology, The Royal Free Hospital, London, UK. 50 women in the placebo group and 50 women in the intervention group
Participants	100 consecutive women undergoing outpatient hysteroscopy for abnormal uterine bleeding consented to be included in the study. The median age of the women was 43 years. 3 exclusions
Interventions	Intracervical injection of either 10 mL of lignocaine 1% with 1:200 000 adrenaline or normal saline was injected into the cervix at 1, 5, 7, and 11 o’clock
Outcomes	10 cm VAS. Pain score before, during, after and 60 min after the procedure. Data grouped into dichotomous outcomes
Notes
*Risk of bias*
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	"Each patient was randomly allocated to receive either lignocaine 1 % or normal  saline intra cervically. Randomization was performed using a predetermined randomization code in a double blind fashion."
Allocation concealment (selection bias)	Unclear risk	"Each patient was randomly allocated to receive either lignocaine 1 % or normal saline intracervically. Randomization was performed using a predetermined randomization code in a double blind fashion."
Blinding (performance bias and detection bias)   All outcomes	Unclear risk	"Randomization was performed using a predetermined randomization code in a double blind fashion."
Incomplete outcome data (attrition bias)   All outcomes	Low risk	"Hysteroscopy was unsuccessful in three women: two (one from each group) found the procedure too painful, and one patient fainted after the intracervical injection of saline. These women were excluded from further analysis."
Selective reporting (reporting bias)	High risk	Data converted into a dichotomous manner. 1 measuring instrument used (10 cm VAS). All time points stated to have data collected were reported. Did not record change from baseline. No P values stated, which may indicate an element of under reporting
Other bias	Low risk	Nothing detected

Methods	A randomised, double‐blind, placebo‐controlled trial. No loss to follow‐up. Single‐centre trial at Day Unit, Queen Charlotte's and Chelsea Hospital, London. 44 women in the intervention arm, 44 women in the placebo arm and 35 women in the control arm
Participants	Women undergoing outpatient hysteroscopy. Mean age of women 44.5 years in intervention group ad 44.6 and 43.4 in the placebo and control group respectively. 14 exclusions
Interventions	Lignocaine gel, placebo gel or no gel administered. Some women received intracervical lignocaine block if determined to need cervical dilatation
Outcomes	Mean pain score across the whole procedure, at gel administration, at lignocaine injection, cervical dilatation, hysteroscopy and endometrial biopsy. 4‐point descriptive scale used
Notes	Co‐administration in some women of intracervical block.
*Risk of bias*
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	"A total of 88 consecutive women undergoing outpatient hysteroscopy were allocated in a double‐blind manner to one of two groups using a pre determined randomization code."
Allocation concealment (selection bias)	Unclear risk	"A total of 88 consecutive women undergoing outpatient hysteroscopy were allocated in a double‐blind manner to one of two groups using a pre determined randomization code."
Blinding (performance bias and detection bias)   All outcomes	Low risk	"Gel was given prior to the investigation by a different operator from those who carried out hysteroscopy, and the patient, hysteroscopist and other staff were unaware of which gel had been used."
Incomplete outcome data (attrition bias)   All outcomes	Unclear risk	Failed procedure in 14 women. No reasons given
Selective reporting (reporting bias)	Unclear risk	Data grouped into two outcomes. 1 measuring instrument used (4‐point descriptive scale). All time points stated to have data collected were reported. Did not record change from baseline
Other bias	High risk	Co‐administration of intracervical block in some women

Methods	Randomised, double‐blind, placebo‐controlled study. No loss to follow‐up. Single‐centre trial at Royal Hospital for Women, Sydney, Australia. 50 women in intervention group and 50 women in placebo group
Participants	Women undergoing outpatient hysteroscopy with or without endometrial biopsy. Mean age of women was 46 years in the intervention group and 47 in the placebo group. 1 exclusion as no cervical os could be identified
Interventions	5 mL of 2% lignocaine or 5 mL of 0.9% saline injected into cervical canal and uterine cavity via hysteroscope 2 min before hysteroscope was manoeuvred
Outcomes	Mean pain score during the procedure. 10 cm VAS
Notes
*Risk of bias*
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	"Pre determined randomization code administered by the hospital pharmacy."
Allocation concealment (selection bias)	Unclear risk	"Pre determined randomization code administered by the hospital pharmacy."
Blinding (performance bias and detection bias)   All outcomes	Low risk	"The physicians performing the procedure and the patients were blinded to treatment assignment."
Incomplete outcome data (attrition bias)   All outcomes	Low risk	"One patient was unable to undergo hysteroscopy as there was no identifiable cervical os."
Selective reporting (reporting bias)	Low risk	1 measuring instrument used (10 cm VAS) and data reported in 1 standard manner. All time points stated to have data collected were reported. Did not record change from baseline. Did not convert to dichotomous outcomes
Other bias	High risk	All women were instructed to take 2 naproxen tablets 275 mg prior to procedure. As women had received naproxen already, the study was more or less on the pain level of manoeuvring the scope, not introduction of the scope, as the pain medication was infused though the scope which was already in the uterine cavity. This may have introduced bias too. In 2 women, hysteroscopy was abandoned due to pain. In both cases, the randomisation code was broken. In both cases normal saline had been used: following the administration of 5 mL 2% lignocaine the procedure was completed

Methods	Randomised, unblinded study. No loss to follow‐up. 3‐arm study; 5 mm diagnostic sheath, 5 mm diagnostic sheath with paracervical block and 3.5 mm sheath. 119 women in group 1 and 121 women in groups 2 and 3. Results for groups 1 and 2 only were included within the meta‐analysis. Single‐centre trial in Centro di Fiferimento Oncologico in Aviano, Italy
Participants	Postmenopausal women referred for outpatient diagnostic hysteroscopy. The mean ages of women were 60 years in group 1 and 61 years in group 2. 22 women were excluded.
Interventions	20 mL of 1% mepivacaine injected para cervically at 3, 5, 7 and 9 o'clock position of the junction of cervix and vagina at least 5 min before the procedure
Outcomes	Pain score after the procedure. 10 cm VAS
Notes
*Risk of bias*
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	"Procedures were randomly assigned through a computer randomization list."
Allocation concealment (selection bias)	Unclear risk	Not described
Blinding (performance bias and detection bias)   All outcomes	High risk	Not blinded
Incomplete outcome data (attrition bias)   All outcomes	Low risk	Excluded 22 women and 38 women not randomised (previous hysteroscopy, severe previous vagal reaction, allergy to local anaesthesia, vaginal stenosis, cervical stenosis, gas leakage, bubble formation, large polyps)
Selective reporting (reporting bias)	Low risk	1 measuring instrument used (10 cm VAS) and data reported; data were further subgrouped into those receiving hormonal treatment. All time points stated to have data collected were reported. Did not record change from baseline. Did not convert to dichotomous outcomes but data presented as percentages in the abstract. SE stated rather than SDs
Other bias	Low risk	No other source of potential bias identified

Methods	Prospective, double‐blind RCT on 210 women between 20‐45 years old undergoing diagnostic hysteroscopy. No losses to follow‐up. 70 women allocated to the tramadol group; 70 to the celecoxib group and the remaining 70 to the placebo group Cairo University Hopsitals, Egypt. May 2014‐November 2014
Participants	Women undergoing diagnostic hysteroscopy. Study included 210 after 23 women declined to participate and 12 women were excluded. Mean age of women in tramadol group 29.25 ± 6.39; mean age of women in Clecoxib group 29.52 ± 6.44; mean age of women in placebo group 30.8 (6%). None of the procedures had to be stopped in the tramadol and celecoxib groups however, 1 procedure had to be stopped in the placebo group due to severe, intolerable pain (VAS 0‐10 cm)
Interventions	Women were assigned into 1 of three groups to receive either 100 mg of oral tramadol (Tramaw, Global Napi, Giza, Egypt) or 200 mg of Celecoxib (Celebrexw 200, Pfizer, USA) or placebo in the control group. All women received the medication 1 h before the intervention. The procedure was performed in the lithotomy position. A 30º angle was used to introduce a 2.9 mm rigid hysteroscope with 3.8 mm diagnostic sheath (Karl Storzw, Germany). The vaginoscopic approach was used for insertion of the hysteroscope in all cases (no use of speculum or tenaculum). The hysteroscope was gently introduced into the uterine cavity after visualisation of the cervix and identification of the external os. Saline was used as the distension medium and the maximum pressure was set at 100 mmHg.
Outcomes	Before surgery, women were educated about the VAS, in which 0 indicated no pain and 10 the worst pain imaginable. Women’s perception of pain was assessed for each group during, immediately after and 30 min after the procedure with the use of the score on VAS. Time until no pain was estimated by asking women to report the time when they thought pain had completely gone. All women stayed in the clinic for at least 30 min and for up to 2 h until the time no pain was reported, and all women were pain free before leaving the clinic. Women were also asked to report side effects.
Notes	With regard to Hassan 2016a and Hassan 2016b, the corresponding author confirmed by email (20 September 2017) "that the 2 papers represent 2 different studies with independent data"
*Risk of bias*
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	"An independent person generated the allocation sequence using computer generated random numbers in a 3 block table"
Allocation concealment (selection bias)	Unclear risk	"Drugs were enclosed in sequentially numbered, sealed envelopes and were kept with the attending nurse."
Blinding (performance bias and detection bias)   All outcomes	Low risk	Double blinding. "Neither the patient nor the physician were aware of the drug used."
Incomplete outcome data (attrition bias)   All outcomes	Low risk	No loss to follow‐up
Selective reporting (reporting bias)	Low risk	1 measuring instrument was used (10 cm VAS). All time points stated to have data collected were reported. Did not record change from baseline. Did not convert to dichotomous outcomes
Other bias	Low risk	None detected

Methods	Prospective, double‐blind RCT on 140 women undergoing diagnostic hysteroscopy of reproductive age (12‐49 years). 70 women allocated to the tramadol group; 70 allocated to the placebo group Cairo University Hopsitals, Egypt, from May 2014‐March 2015
Participants	Women undergoing diagnostic hysteroscopy. Study included 140 after 46 women declined to participate and 26 women were excluded. Mean age of women in tramadol group 31.5 (+/‐ 7.4); mean age of women in placebo group 32.3 (+/‐ 8.1). None of the procedures had to be stopped in the tramadol group; 1 procedure had to be stopped in the placebo group due to severe, intolerable pain (VAS 0‐10 cm)
Interventions	Women were assigned to 1 of two groups to receive either 50 mg of oral tramadol (Tramaw, Global Napi, Gisa, Egypt) or placebo in the control group. All women received the medication 1 h before the intervention. The procedure was performed in the lithotomy position. A 30º angle was used to introduce a 2.9 mm rigid hysteroscope with 3.8 mm diagnostic sheath (Karl Storzw, Germany). The vaginoscopic approach was used for insertion of the hysteroscope in all cases (no use of speculum or tenaculum). The hysteroscope was gently introduced into the uterine cavity after visualisation of the cervix and identification of the external os. Saline was used as the distension medium and the maximum pressure was set at 100 mmHg. All procedures done by the same operator with same technique
Outcomes	Women’s perception of pain was assessed for each group during, immediately after and 30 min after the procedure with the use of the score on a VAS. Women were also asked to report side effects.
Notes	With regard to Hassan 2016a and Hassan 2016b, the corresponding author confirmed by email (20 September 2017) "that the 2 papers represent 2 different studies with independent data"
*Risk of bias*
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	"An independent person generated the allocation sequence using computer generated random numbers"
Allocation concealment (selection bias)	Unclear risk	"Drugs were enclosed in sequentially numbered, sealed envelopes and were kept with the attending nurse."
Blinding (performance bias and detection bias)   All outcomes	Low risk	Double blinding. "The nurse, the patient and the physician were not aware of the drug used."
Incomplete outcome data (attrition bias)   All outcomes	Low risk	No loss to follow‐up.
Selective reporting (reporting bias)	Unclear risk	1 measuring instrument was used (10 cm VAS). All time points stated to have data collected were reported. Did not record change from baseline. Did not convert to dichotomous outcomes. Pain outcomes reported as median scores and not means.
Other bias	Low risk	None detected

Methods	Prospective RCT on 200 women undergoing hysteroscopy for abnormal uterine bleeding. No losses to follow‐up. 100 women received local anaesthesia while 100 women did not receive any local anaesthesia Department of Gynaecology of Zekai Tahir Burak Woman's Health Research and Education Hospitals, Ankara, Turkey, May 2013‐September 2013
Participants	Study included 200 women undergoing diagnostic hysteroscopy. Mean age of women in Group 1 (received local anaesthesia) 41.94 ± 9.01; mean age of women in Group 2 (no local anaesthesia) 42.03 ± 8.30 None of the procedures had to be stopped in Group 1 or 2
Interventions	All women were examined between the 5 LH and 10 LH cycle days. A disposable speculum was used to visualise the cervix and the cervix was cleaned with a water solution of octenidine hydrochloride 0.1% and 2‐phenoxyethanol 2%. Intracervical local anaesthesia (10 mL of 1% prilocaine) was applied at the 4 and 8 o’clock positions on the posterior lip of the cervix in divided doses and then the speculum was removed. A rigid, 3‐mm outer diameter Storz® hysteroscope was used for the procedure. Hysteroscopy was performed without the use of a tenaculum and without cervical dilatation by the same physician who knew the group to which the women belonged. Uterine distension was maintained by a steady stream of 1.5% Glycine solution
Outcomes	Before surgery, women were educated about the VAS in which 0 indicated no pain and 10 the worst pain imaginable. Women were observed for 60 min and they were asked to complete a standardised pain evaluation form to evaluate the worst pain experienced during the procedure after 30 min and 60 min of the procedure by a clinician who did not know who did not know the group to which the women belonged.
Notes
*Risk of bias*
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	All the 200 women were randomised using a computer  generated randomisation programme into 2 groups.
Allocation concealment (selection bias)	Unclear risk	Not stated
Blinding (performance bias and detection bias)   All outcomes	High risk	Physician performing the procedure were not blinded to the allocation. However, physicians evaluating the pain score using VAS were blinded to the procedure. Unclear as to whether the women were blinded to the procedure.
Incomplete outcome data (attrition bias)   All outcomes	Low risk	No exclusions reported
Selective reporting (reporting bias)	Low risk	1 measuring instrument used (10cm VAS) and data reported in 1 standard manner. All time points stated to have data collected were reported. Did not record change from baseline. Did not convert to dichotomous outcomes
Other bias	Low risk	No other potential bias detected

Methods	Double‐blinded, randomised, placebo‐controlled trial. No loss to follow‐up. 49 women in intervention group, 50 in placebo group. Single‐centre trial at The Chinese University of Hong Kong
Participants	Women undergoing outpatient hysteroscopy for abnormal uterine bleeding. Mean age of 49 years in intervention and placebo group
Interventions	Paracervical block at 3, 5, 7 and 9 o'clock positions of 10 mL of 2% lignocaine or saline, 5 min prior to procedure
Outcomes	Pain score before the procedure, at grasping the cervix, after injection, at insertion of hysteroscope, distension of uterus, after endometrial biopsy and 30 min after the procedure. 10 cm VAS
Notes
*Risk of bias*
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	"They were randomized into two groups using a computer generated block number and put inside a sealed envelope."
Allocation concealment (selection bias)	Unclear risk	"They were randomized into two groups using a computer generated block number and put inside a sealed envelope."
Blinding (performance bias and detection bias)   All outcomes	Low risk	"The attending doctor, nursing staff and the women were all blinded to the identity of the medication used."
Incomplete outcome data (attrition bias)   All outcomes	Low risk	1 exclusion reported. "Hysteroscopy failed in one woman because of cervical stenosis"
Selective reporting (reporting bias)	Low risk	1 measuring instrument used (10 cm VAS) and data reported in 1 standard manner. All time points stated to have data collected were reported. Did not record change from baseline. Did not convert to dichotomous outcomes
Other bias	Low risk	Nothing detected

PERMALINK

Pain relief for outpatient hysteroscopy

Gaity Ahmad

Sushant Saluja

Helena O'Flynn

Alessandra Sorrentino

Daniel Leach

Andrew Watson

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

1. Dosing of local anaesthetics.

Types of outcome measures

Primary outcomes

Secondary outcomes

Search methods for identification of studies

Electronic searches

Searching other resources

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

Overall quality of the body of evidence: 'Summary of findings' table

Results

Description of studies

Results of the search

3.

Included studies

Interventions

Participants

Outcomes

Excluded studies

Risk of bias in included studies

Allocation

Sequence generation

Allocation concealment

Blinding

Incomplete outcome data

Selective reporting

Other potential sources of bias

Effects of interventions

Summary of findings for the main comparison. Local anaesthetic compared to placebo or no treatment for outpatient hysteroscopy.

Summary of findings 2. Oral NSAID compared to placebo or no treatment for outpatient hysteroscopy.

Summary of findings 3. Opioid compared to placebo or no treatment for outpatient hysteroscopy.

Summary of findings 4. Intracervical versus combined intracervical and paracervical anaesthesia.

Summary of findings 5. Antispasmodic plus NSAID versus paracervical anaesthesia.

1. Analgesic versus placebo or no treatment

1.1 Local anaesthetic versus placebo or no treatment

1.1.1. Primary outcome: pain score

1.1. Analysis.

4.

Pain score during the procedure

Methods	Randomised, placebo‐controlled study. No loss to follow‐up. 80 women in intervention group. 84 in placebo group. Single‐centre trial at Wu Ho‐Su Memorial Hospital, Taiwan
Participants	Women undergoing office hysteroscopy. 15 exclusions. Mean age of 41 years in intervention group and 40 in placebo group
Interventions	0.2 mg of buprenorphine or placebo given under tongue 40 min prior to procedure
Outcomes	Pain score during the procedure. 10 cm VAS used
Notes
*Risk of bias*
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	"The patients were randomized by computer generated numbers."
Allocation concealment (selection bias)	Unclear risk	Not described
Blinding (performance bias and detection bias)   All outcomes	Unclear risk	Not described
Incomplete outcome data (attrition bias)   All outcomes	Low risk	"The procedure failed in 15 patients because of cervical stenosis."
Selective reporting (reporting bias)	Low risk	1 measuring instrument used (10 cm VAS) and data reported in 1 standard manner. All time points stated to have data collected were reported. Did not record change from baseline. Did not convert to dichotomous outcomes. The P values were described as not significant but were not actually reported, which may suggest an element of under reporting
Other bias	Low risk	Nothing detected

Methods	Randomised, comparative treatment trial conducted by 5 private obstetrics and gynaecology practices in the USA with 1 investigator per site. 3 women lost to follow‐up. 40 premenopausal women randomised to combined para/intracervical anaesthetic block protocol of 37 cc local anaesthetic administered at 6 different injection sites in association with application of topical 1% lidocaine gel or intracervical only anaesthetic block protocol of 22 cc administered at 3 different injection sites without topical anaesthesia
Participants	Women undergoing hysteroscopy for removal of intrauterine polyps and myomas using the MyoSure device. Study included 40 women. 19 randomised to the para/intracervical group and 21 to the intracervical group. Mean age of women was 44.2 ± 7.7 in the combined para/intracervical block group and 41.8 ± 7.5 n the intracervical group. All randomised subjects underwent the procedure
Interventions	The para/intracervical block group received a total of 37 cc of anaesthetic at 6 different sites. Topical 1 % lidocaine was applied to the cervix, with a set time of 2‐3 min before the injection of anaesthetic, for para/intracervical group only. The intracervical group received a total of 22 cc of anaesthetic administered at 3 different sites
Outcomes	The main outcome measure was composite score for procedure‐related pain, which incorporated individual pain scores during: 1) the cervical block injection; 2) cervical dilatation; 3) uterine distension; 4) the tissue resection
Notes	This study used the Wong Baker scale, a 0‐10 cm measure
*Risk of bias*
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Women were randomised on the day to combined para/intracervical block or an intracervical block group in a 1:1 ratio, using a computer‐generated randomisation scheme
Allocation concealment (selection bias)	Unclear risk	The randomisation sequence was provided to sites, using sealed, sequentially labelled opaque envelopes
Blinding (performance bias and detection bias)   All outcomes	High risk	Single‐blinded ‐ women were blinded to the assignment because they were not told how many injections each group would be receiving
Incomplete outcome data (attrition bias)   All outcomes	Unclear risk	All randomised women underwent the procedure however, 2 women failed to meet the inclusion and exclusion criteria and 1 woman had a major protocol deviation. As a result the final analysis included 17 women in the para/intracervical block and 20 in the intracervical block.
Selective reporting (reporting bias)	Low risk	1 measuring instrument, the Wong‐baker face rating scale was used. This scale provides a scale ranging from 0‐10 cm, with 0 indicating no pain with 10 indicating maximum pain. All time points stated to have data collected were reported. Did not record change from baseline. Did not convert to dichotomous outcomes
Other bias	Unclear risk	Women were told to take 800 mg of ibuprofen the night before the procedure. 1 h before the procedure the women received 10 mg of diazepam and 10 mg of hydrocodone/acetaminophen, followed by an intramuscular injection of 30 mg ketorolac and 0.4% of atropine

Methods	A randomised, placebo‐controlled study. No loss to follow‐up. 200 women were prospectively randomised into two groups, the study group (n = 100) and the control group (n = 100). Single‐centre trial at the hysteroscopy unit of the First Department of Obstetrics and Gynecology of the University of Athens
Participants	Women undergoing diagnostic hysteroscopy, with or without endometrial biopsy. Mean age of 35.4 years in the intervention group and 36.1 in the placebo group
Interventions	1 mL to 3 mL (30 to 90 mg) of mepivacaine 3% or saline administered intracervically 3 min prior to procedure
Outcomes	Pain score during the procedure, 30 min after the procedure and 60 min after the procedure. 11‐point scale used
Notes	Varying dose of mepivacaine given
*Risk of bias*
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	"200 patients were prospectively randomised into two groups."
Allocation concealment (selection bias)	Unclear risk	"200 patients were prospectively randomised into two groups."
Blinding (performance bias and detection bias)   All outcomes	Unclear risk	Not stated
Incomplete outcome data (attrition bias)   All outcomes	Unclear risk	"No patient from either group needed to be hospitalised. All patients left the hospital within 90 min after the end of the procedure."
Selective reporting (reporting bias)	High risk	1 measuring instrument used (11‐point scale). Data were reported in a graphical form as well as numerical form. However, these data could not be accurately included within the review as the P values and SDs were described as not significant but were not actually reported, which may indicate an element of under reporting. All time points stated to have data collected were reported in the graph, but could not be interpreted accurately. Did not record change from baseline. Did not convert to dichotomous outcomes
Other bias	Unclear risk	Varying dose of mepivacaine administered

Methods	A randomised prospective study. No loss to follow‐up. A total of 148 women received dexketoprofen tablets and 150 women received local anaesthetic drug. Single‐centre trial at Menopause Clinic, University ‘Federico II’ of Naples, Via Pansini 5, Naples, Italy
Participants	305 consecutive postmenopausal women were referred to outpatient Menopause Clinic for hysteroscopy because of uterine bleeding. Mean ages not stated. Postmenopausal women
Interventions	The women were randomly allocated to receive either local infiltration of the cervix by injecting of 5 mL mepivacaine 2% intracervically up to the level of the internal os or 1 tablet of dexketoprofen given 1 h before the procedure
Outcomes	Pain experienced during hysteroscopy and at 30, 60 and 120 min after the procedure  using an 11‐point VAS
Notes
*Risk of bias*
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	"Randomisation was achieved with sealed envelopes containing computer generated block randomisation numbers."
Allocation concealment (selection bias)	Unclear risk	"Randomisation was achieved with sealed envelopes containing computer generated block randomisation numbers."
Blinding (performance bias and detection bias)   All outcomes	Unclear risk	Not stated
Incomplete outcome data (attrition bias)   All outcomes	Low risk	"Seven women were excluded from the study. Five of these thought to be too anxious to tolerate hysteroscopy under local anaesthesia or pharmacological sedation, two because previous conisation."
Selective reporting (reporting bias)	High risk	Data presented in a graphical form only with no numerical values. 1 measuring instrument used (10 cm VAS). All time points stated to have data collected were reported. Did not record change from baseline
Other bias	Low risk	Nothing detected

Methods	Double‐blind RCT. 70 nulliparous women with primary infertility undergoing diagnostic hysteroscopy. 18 lost to follow‐up. 44 women in lidocaine group (9 lost to follow‐up in this group so final analysis of 35 women) and 44 (9 lost to follow‐up so final analysis of 35 women) women in diclofenac group. Center of Reproductive Medicine, Dr. Shariati Hospital, July 2012‐April 2013
Participants	Women undergoing hysteroscopy. Study included 70 women after excluding 24 due to not meeting inclusion criteria or being lost to follow‐up. Mean age of women in lidocaine group 29.3 ± 4.4; mean age of women in diclofenac group 30.8 ± 4.4. 9 women in the lidocaine group needed sedation with propofol. 5 women in the diclofenac group needed sedation and 4 women needed an invasive procedure in the diclofenac group
Interventions	Women were assigned into 1 of two groups to receive either 100 mg of rectal diclofenac or 5 mL of 2% intrauterine lidocaine. In the gynaecologic ward, 30 min before transferring to the operating room, the staff gave 100 mg rectal diclofenac to women in the diclofenac group. The study drugs (lidocaine or saline) were prepared by the anaesthesia staff who gave it to the surgeon who was blinded to allocation. Then 5 mL of 2% lidocaine or the same volume of saline was instilled through the endocervix into the uterine cavity with an 18‐gauge angiocatheter. The angiocatheter was left in place for 3 min before it was withdrawn while women were in the Trendelenburg position to limit backflow and to allow the anaesthetic to take effect
Outcomes	Before surgery, women were educated about the NRS, 0 indicated no pain and 10 the worst pain imaginable. Pain scoring was performed during insertion of the hysteroscope, during visualisation of the intrauterine cavity, and during extrusion of the hysteroscope
Notes
*Risk of bias*
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Randomisation was performed using computer‐generated codes. Sealed envelopes containing the information of the randomisation code were kept by the staff not involved in the study. The envelope was transferred to a specific member of the gynaecologic staff
Allocation concealment (selection bias)	Unclear risk	Sealed envelopes containing the information of the randomisation code, generated by a computer were given to a member of the gynaecology staff
Blinding (performance bias and detection bias)   All outcomes	Low risk	Double blinded ‐ outcome assessor and the women were blinded
Incomplete outcome data (attrition bias)   All outcomes	Unclear risk	Of the 88 women randomised, 18 were lost to follow‐up (9 needed propofol in the lidocaine group); 5 needed sedation with propofol and 4 needed an invasive procedure in the diclofenac group
Selective reporting (reporting bias)	Low risk	1 measuring instrument was used (NRS 0–10) and the data were reported in 1 standard manner. All time points stated to have data collected were reported. Did not record change from baseline. Did not convert to dichotomous outcomes
Other bias	Low risk	Nothing detected

Methods	Randomised, placebo‐controlled trial. 49 women were randomised to the active drug, and 46 to placebo. No loss to follow‐up. Single‐centre trial at Minimally Invasive Therapy Unit and Endoscopy Training Centre, University Department of Obstetrics and Gynaecology, The Royal Free Hospital, London, UK
Participants	The mean age of the women was 43 to 49 years. Women attending for diagnostic hysteroscopy
Interventions	Active mefenamic acid 500 mg or placebo tablets 1 h before the procedure
Outcomes	Pain scores during, immediately after (0 min), 30 min and 60 min after outpatient hysteroscopy. 10 cm VAS
Notes
*Risk of bias*
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	"Active (mefenamic acid 500 mg) and placebo tablets, which were identical in appearance, were packaged in coded bottles, randomisation being provided by the manufacturer."
Allocation concealment (selection bias)	Unclear risk	"Active (mefenamic acid 500 mg) and placebo tablets, which were identical in appearance, were packaged in coded bottles, randomisation being provided by the manufacturer."
Blinding (performance bias and detection bias)   All outcomes	Unclear risk	"Active (mefenamic acid 500 mg) and placebo tablets, which were identical in appearance, were packaged in coded bottles, randomisation being provided by the manufacturer."
Incomplete outcome data (attrition bias)   All outcomes	Low risk	Hysteroscopy was unsuccessful in 4 women (2 in each group): 2 refused because they were too anxious and 2 procedures had to be abandoned because of pain
Selective reporting (reporting bias)	High risk	Data presented graphically with no numerical values. 1 measuring instrument used (10 cm VAS). All time points stated to have data collected were reported. Did not record change from baseline. Odds ratio of experiencing pain calculated
Other bias	High risk	Some women given local anaesthesia as well as intervention

Methods	Double‐blind RCT. 206 women suspected of having endometrial polyp, abnormal uterine bleeding or diagnosis of myoma and infertility. No losses to follow‐up. The total 206 women comprised 68 (33%) in the placebo group, 76 (36.89%) in the lidocaine group and 62 (30.09%) in the indomethacin group Bakirköy Dr. Sadi Konuk Training and Research Hospital. July 2014‐March 2015
Participants	Women undergoing diagnostic hysteroscopy. Study included 206 women. Mean age of women in placebo group 45.62 ± 9.04; mean age of women in lidocaine group 45.54 ± 11.73; mean age of women in indomethacin group 44.71 ± 9.97
Interventions	The control group was administered with a 1000 mL distention medium containing 18 mL serum physiologic per 250 mL and a rectal placebo. The second group was administered with a 1000 mL distention medium containing 18 mL lidocaine per 250 mL (Jetokain ampoule 20 mg 2% Adeka, Samsun, Turkey) and rectal placebo. The third group was administered with rectal  indomethacin suppository (Endol 100 mg supp. Deva, Istanbul, Turkey) 45 min before the procedure and with a 1000 mL distention medium containing 18 mL serum physiologic per 250 mL
Outcomes	Before the procedure, women were educated about the NRS in which 0 indicated no pain and 10 the worst pain imaginable. Pain scoring was performed during insertion of the hysteroscope, during visualisation of the intrauterine cavity and 10 min after the procedure
Notes
*Risk of bias*
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Women were selected using random‐number tables
Allocation concealment (selection bias)	Unclear risk	Not stated
Blinding (performance bias and detection bias)   All outcomes	Low risk	The hysteroscopist and woman were blinded
Incomplete outcome data (attrition bias)   All outcomes	Low risk	There was no loss to follow‐up
Selective reporting (reporting bias)	Unclear risk	1 measuring instrument was used (NRS 0–10) and the data were reported in 1 standard manner. All time points stated to have data collected were reported. Did not record change from baseline. Did not convert to dichotomous outcomes. Pain score outcomes were reported as medians
Other bias	Low risk	No other potential risk identified

Methods	Randomised trial. No loss to follow‐up. Single‐centre trial at outpatient gynaecology department of All India Institute of Medical Sciences, New Delhi, India. 40 women in each intervention group with a total of 120 women
Participants	A total of 120 women with a medical indication for hysteroscopy and endometrial biopsy (infertility, abnormal uterine bleeding) were recruited. The mean ages were 36.28 years in group 1, 37.75 in group 2 and 36.82 years in group 3
Interventions	Group 1 women received fixed‐dose oral tablet containing drotaverine (80 mg) with mefenamic acid (250 mg) 1 h prior to the procedure. Group 2 women received paracervical block with 10 mL of 1% lignocaine solution injected at 3 and 9‐o'clock position at the junction of the cervix and vagina in divided doses 5 min prior to the procedure. Group 3 women received intravenous sedation with diazepam (0.2 mg/kg body weight) and pentazocine (0.6 mg/kg body weight) 10 min prior to the procedure. This group was not included in the review
Outcomes	The worst pain experienced during the procedure and the degree of their discomfort after 30 min and 60 min of the procedure using a 10 cm VAS
Notes	Only 2 of the 3 comparison groups were eligible for this review
*Risk of bias*
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	"This study was an open‐label randomised trial where all the 120 patients were randomised using a predetermined computer‐generated randomisation code into 3 groups."
Allocation concealment (selection bias)	Unclear risk	No details reported
Blinding (performance bias and detection bias)   All outcomes	High risk	"This study was an open‐label randomised trial."
Incomplete outcome data (attrition bias)   All outcomes	Low risk	"Of the 120 patients recruited, procedure was performed successfully in all, and at no point anyone was excluded from the study."
Selective reporting (reporting bias)	Low risk	1 measuring instrument used (10 cm VAS) and data reported in 1 standard manner. All time points stated to have data collected were reported. Did not record change from baseline. Did not convert to dichotomous outcomes
Other bias	Low risk	Nothing detected

Methods	Double‐blinded RCT. 62 in intervention group, 56 in placebo group. None lost to follow‐up. Single‐centre trial at Hopital Hotel‐Dieu de Paris, Paris, France
Participants	Women undergoing diagnostic hysteroscopy for abnormal uterine bleeding or infertility. Mean age of 41 years in intervention group and 40 in placebo group
Interventions	30 mg (3 metered doses) of lignocaine or placebo sprayed onto surface of cervix and cervical canal through 360° 5 min prior to procedure. 3 exclusions
Outcomes	Pain score during the procedure. 10 cm VAS (pain)
Notes	Unequal number in groups. No re‐inclusion of exclusions into analysis
*Risk of bias*
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	"Women were assigned to receive either lidocaine spray or placebo according to a computer generated randomisation code."
Allocation concealment (selection bias)	Unclear risk	"Lidocaine and placebo were packaged in identical bottles and could not be differentiated."
Blinding (performance bias and detection bias)   All outcomes	Low risk	Double‐blinded. "Lidocaine and placebo were packaged in identical bottles and could not be differentiated."
Incomplete outcome data (attrition bias)   All outcomes	Low risk	3 women excluded. "Two women did not fill out the questionnaire properly"; "one woman the diagnostic hysteroscopy was not done due to cervical stenosis."
Selective reporting (reporting bias)	Low risk	1 measuring instrument used (10 cm VAS) and data reported in 1 standard manner. All time points stated to have data collected were reported. Did not record change from baseline. Did not convert to dichotomous outcomes
Other bias	Unclear risk	Unequal numbers in intervention and placebo groups ‐ could happen by chance

Methods	Randomised, prospective study. No loss to follow‐up. 88 women in prilocaine cream group and 92 women in lignocaine spray group. Study refers to 165 in control group, but these women not part of RCT. Single‐centre trial at Castrovillari Hospital, Naples, Italy.
Participants	Women attending for diagnostic hysteroscopy. No mean ages given
Interventions	1 cm³ of 5% prilocaine cream onto esocervix and 2 cm³ inserted 3 cm into cervical canal 10 min before procedure or 20 mg of lidocaine spray directed onto esocervix and 20 mg 3 cm into cervical canal immediately before procedure or no intervention
Outcomes	4‐point pain scale. Placement of tenaculum, progression through cervical canal and evaluation of uterine cavity
Notes
*Risk of bias*
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	"180 patients were included in the study and randomly allocated to group A or B"
Allocation concealment (selection bias)	Unclear risk	"180 patients were included in the study and randomly allocated to group A or B."
Blinding (performance bias and detection bias)   All outcomes	Unclear risk	"180 patients were included in the study and randomly allocated to group A or B."
Incomplete outcome data (attrition bias)   All outcomes	Unclear risk	Unclear if any women excluded
Selective reporting (reporting bias)	Unclear risk	1 measuring instrument used (4‐point descriptive scale) and data reported in 1 standard manner. All time points stated to have data collected were reported. Did not record change from baseline. Data converted into two dichotomous groups
Other bias	Low risk	Nothing detected