Cephalic version by moxibustion for breech presentation

Abstract

Background

Breech presentation at term can cause complications during birth and increase the chance of caesarean section. Moxibustion (a type of Chinese medicine which involves burning a herb close to the skin) at the acupuncture point Bladder 67 (BL67) (Chinese name Zhiyin), located at the tip of the fifth toe, has been proposed as a way of changing breech presentation to cephalic presentation. This is an update of a review first published in 2005 and last published in 2012.

Objectives

To examine the effectiveness and safety of moxibustion on changing the presentation of an unborn baby in the breech position, the need for external cephalic version (ECV), mode of birth, and perinatal morbidity and mortality.

Search methods

For this update, we searched Cochrane Pregnancy and Childbirth’s Trials Register (which includes trials from CENTRAL, MEDLINE, Embase, CINAHL, and conference proceedings), ClinicalTrials.gov, and the WHO International Clinical Trials Registry Platform (ICTRP) (4 November 2021). We also searched MEDLINE, CINAHL, AMED, Embase and MIDIRS (inception to 3 November 2021), and the reference lists of retrieved studies.

Selection criteria

The inclusion criteria were published and unpublished randomised or quasi‐randomised controlled trials comparing moxibustion either alone or in combination with other techniques (e.g. acupuncture or postural techniques) with a control group (no moxibustion) or other methods (e.g. acupuncture, postural techniques) in women with a singleton breech presentation.

Data collection and analysis

Two review authors independently determined trial eligibility, assessed trial quality, and extracted data. Outcome measures were baby's presentation at birth, need for ECV, mode of birth, perinatal morbidity and mortality, maternal complications and maternal satisfaction, and adverse events. We assessed the certainty of the evidence using the GRADE approach.  

Main results

This updated review includes 13 studies (2181 women), of which six trials are new. Most studies used adequate methods for random sequence generation and allocation concealment. Blinding of participants and personnel is challenging with a manual therapy intervention; however, the use of objective outcomes meant that the lack of blinding was unlikely to affect the results. Most studies reported little or no loss to follow‐up, and few trial protocols were available. One study that was terminated early was judged as high risk for other sources of bias.

Meta‐analysis showed that compared to usual care alone, the combination of moxibustion plus usual care probably reduces the chance of non‐cephalic presentation at birth (7 trials, 1152 women; risk ratio (RR) 0.87, 95% confidence interval (CI) 0.78 to 0.99, I² = 38%; moderate‐certainty evidence), but the evidence is very uncertain about the effect of moxibustion plus usual care on the need for ECV (4 trials, 692 women; RR 0.62, 95% CI 0.32 to 1.21, I² = 78%; low‐certainty evidence) because the CIs included both appreciable benefit and moderate harm. Adding moxibustion to usual care probably has little to no effect on the chance of caesarean section (6 trials, 1030 women; RR 0.94, 95% CI 0.83 to 1.05, I² = 0%; moderate‐certainty evidence). The evidence is very uncertain about the effect of moxibustion plus usual care on the the chance of premature rupture of membranes (3 trials, 402 women; RR 1.31, 95% CI 0.17 to 10.21, I² = 59%; low‐certainty evidence) because there were very few data. Moxibustion plus usual care probably reduces the use of oxytocin (1 trial, 260 women; RR 0.28, 95% CI 0.13 to 0.60; moderate‐certainty evidence). The evidence is very uncertain about the chance of cord blood pH less than 7.1 (1 trial, 212 women; RR 3.00, 95% CI 0.32 to 28.38; low‐certainty evidence) because there were very few data. We are very uncertain whether the combination of moxibustion plus usual care increases the chance of adverse events (including nausea, unpleasant odour, abdominal pain and uterine contractions; intervention: 27/65, control: 0/57), as only one study presented data in a way that could be reanalysed (122 women; RR 48.33, 95% CI 3.01 to 774.86; very low–certainty evidence).

When moxibustion plus usual care was compared with sham moxibustion plus usual care, we found that moxibustion probably reduces the chance of non‐cephalic presentation at birth (1 trial, 272 women; RR 0.74, 95% CI 0.58 to 0.95; moderate‐certainty evidence) and probably results in little to no effect on the rate of caesarean section (1 trial, 272 women; RR 0.84, 95% CI 0.68 to 1.04; moderate‐certainty evidence). No study that compared moxibustion plus usual care with sham moxibustion plus usual care reported on the clinically important outcomes of need for ECV, premature rupture of membranes, use of oxytocin, and cord blood pH less than 7.1, and one trial that reported adverse events reported data for the whole sample.

When moxibustion was combined with acupuncture and usual care, there was very little evidence about the effect of the combination on non‐cephalic presentation at birth (1 trial, 226 women; RR 0.73, 95% CI 0.57 to 0.94) and at the end of treatment (2 trials, 254 women; RR 0.73, 95% CI 0.57 to 0.93), and on the need for ECV (1 trial, 14 women; RR 0.45, 95% CI 0.07 to 3.01). There was very little evidence about whether moxibustion plus acupuncture plus usual care reduced the chance of caesarean section (2 trials, 240 women; RR 0.80, 95% CI 0.65 to 0.99) or pre‐eclampsia (1 trial, 14 women; RR 5.00, 95% CI 0.24 to 104.15). The certainty of the evidence for this comparison was not assessed.

Authors' conclusions

We found moderate‐certainty evidence that moxibustion plus usual care probably reduces the chance of non‐cephalic presentation at birth, but uncertain evidence about the need for ECV. Moderate‐certainty evidence from one study shows that moxibustion plus usual care probably reduces the use of oxytocin before or during labour. However, moxibustion plus usual care probably results in little to no difference in the rate of caesarean section, and we are uncertain about its effects on the chance of premature rupture of membranes and cord blood pH less than 7.1. 

Adverse events were inadequately reported in most trials.

Plain language summary

Moxibustion for turning a baby in breech position

What is the issue?

Breech presentation (when the baby is bottom down) is common in the second trimester of pregnancy. Most babies turn so that their head is down, ready for birth, before the onset of labour; however, some do not. A baby coming bottom or feet first can have more difficulty being born. This can cause problems for the mother and baby, and the baby is more likely to be born by caesarean section. 

Moxibustion is a type of Chinese medicine that may help turn a breech baby. It involves burning a herb (Artemesia spp.) close to the skin at an acupuncture point on the little toe to produce a warming sensation and provide stimulation to the uterus. This can be performed by the mother or by a family member or friend after they have received training on how to administer moxibustion safely.

Why is this important?

Vaginal birth when the baby is in the breech position is possible with experienced doctors and midwives and well‐equipped hospitals, and unplanned vaginal breech births can occur outside of primary care settings. However, not all hospitals can offer vaginal birth to women with breech babies, and birth by caesarean section may be planned. Many healthcare providers and mothers want to avoid caesarean section because it has risks for the current and future pregnancies. We wanted to know whether moxibustion treatment, which can be self‐administered at home by the woman or her family or friends, can help the baby turn so that it is head‐down for birth. 

What evidence did we find?

We searched for evidence to 3 November 2021 for studies that tested moxibustion plus usual care (alone or combined with other treatments, such as acupuncture or postural positions) with usual care, sham moxibustion (moxibustion at a point not relevant to breech presentation; used to blind participants to group allocation), postural positions, or other treatments. We evaluated 13 studies involving a total of 2181 women and their babies. We identified seven new trials. We judged the evidence for most outcomes as low to moderate certainty.

Moxibustion plus usual care probably reduces the number of breech babies at birth more than usual care alone or sham moxibustion plus usual care. There were very few data on the effect of moxibustion plus usual care on the need for external cephalic version (where a doctor attempts to turn the baby). Moxibustion plus usual care probably does not reduce the number of babies born by caesarean section (whether compared to usual care alone or sham acupuncture plus usual care). We are uncertain whether moxibustion plus usual care can reduce the chance of the membranes rupturing early. Moxibustion plus usual care probably reduces the use of oxytocin, a hormone used to begin or improve contractions during labour. There were very few data on the effect of moxibustion plus usual care on how acidic the umbilical cord blood is; we are uncertain about these results because the study results varied, and the results were imprecise. We are very uncertain whether moxibustion plus usual care increases the chance of side effects, because only one study reported side effects according to which treatment women received, and all side effects occurred in the treatment group (27/65 versus 0/57).

The most frequently reported side effects were increased fetal movements, uterine contractions, nausea, headache, and burns from holding the moxibustion stick too close to the skin.

What does this mean?

Starting moxibustion treatment before 37 weeks of pregnancy probably reduces the chance of baby being head‐up at birth, but does not reduce the number of babies born by caesarean section. We need more evidence to determine the risk of side effects of moxibustion.

Summary of findings

Summary of findings 1. Summary of findings table ‐ Moxibustion plus usual care compared to usual care for breech presentation.

Moxibustion plus usual care compared to usual care for breech presentation
Patient or population: breech presentation Setting: hospital setting Intervention: moxibustion plus usual care Comparison: usual care
Outcomes	Anticipated absolute effects* (95% CI)		Relative effect (95% CI)	№ of participants (studies)	Certainty of the evidence (GRADE)	Comments
	Risk with usual care	Risk with moxibustion plus usual care
Non‐cephalic presentation at birth follow‐up: range 1 weeks to 2 weeks	51 per 100	44 per 100 (40 to 50)	RR 0.87 (0.78 to 0.99)	1152 (7 RCTs)	⊕⊕⊕⊝ Moderatea
Need for ECV follow‐up: range 1 weeks to 2 weeks	35 per 100	22 per 100 (11 to 43)	RR 0.62 (0.32 to 1.21)	692 (4 RCTs)	⊕⊕⊝⊝ Lowb,c
Caesarean section follow‐up: range 1 weeks to 2 weeks	53 per 100	50 per 100 (44 to 55)	RR 0.94 (0.83 to 1.05)	1030 (6 RCTs)	⊕⊕⊕⊝ Moderated
Premature rupture of membranes follow‐up: range 1 weeks to 2 weeks	6 per 100	8 per 100 (1 to 62)	RR 1.31 (0.17 to 10.21)	402 (3 RCTs)	⊕⊕⊝⊝ Lowb,e
Use of oxytocin follow‐up: range 1 weeks to 2 weeks	31 per 100	9 per 100 (4 to 19)	RR 0.28 (0.13 to 0.60)	161 (1 RCT)	⊕⊕⊕⊝ Moderatef
Cord blood pH less than 7.1 follow‐up: 2 weeks	1 per 100	3 per 100 (0 to 27)	RR 3.00 (0.32 to 28.38)	212 (1 RCT)	⊕⊕⊝⊝ Lowg
Adverse events follow‐up: range 1 weeks to 2 weeks	0 per 100	0 per 100 (0 to 0)	RR 48.33 (3.01 to 774.86)	122 (1 RCT)	⊕⊝⊝⊝ Very lowh,i
*The risk in the intervention group (and its 95% confidence interval) is based on the assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI). CI: confidence interval; RR: risk ratio
GRADE Working Group grades of evidence High certainty: we are very confident that the true effect lies close to that of the estimate of the effect. Moderate certainty: 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 certainty: our confidence in the effect estimate is limited: the true effect may be substantially different from the estimate of the effect. Very low certainty: we have very little confidence in the effect estimate: the true effect is likely to be substantially different from the estimate of effect.
See interactive version of this table: https://gdt.gradepro.org/presentations/#/isof/isof_question_revman_web_430818774604944466.

^a We rated down by one level for serious imprecision as the optimal information size was not met (OIS = 1086 per group).
^b We rated down by one level for serious inconsistency due to substantial heterogeneity in study results.
^c We rated down by one level for serious imprecision due to wide confidence intervals which included both appreciable benefit and moderate harm. The optimal information size was met (OIS = 269 per group).
^d We rated down by one level for serious imprecision because the optimal information size was not met (OIS = 4356 per group).
^e We rated down by one level for serious imprecision due to wide confidence intervals which included both appreciable benefit and appreciable harm, and because the optimal information size was not met (OIS = 1863 per group).
^f We rated down by one level for serious imprecision. The optimal information size was met (OIS = 51 per group); however, we are uncertain whether this result represents the true effect.
^g We rated down by two levels for serious imprecision due to very wide confidence intervals which included both appreciable benefit and appreciable harm, and the optimal information size was not met (OIS = 680 per group).
^h We rated down by one level for risk of bias due to high risk of bias for two domains (incomplete outcome data, other bias).
ⁱ We rated down by two levels due to very wide confidence intervals which suggested appreciable harm with moxibustion plus usual care compared to usual care. The optimal information size was met (OIS = 15 per group).

Summary of findings 2. Summary of findings table ‐ Moxibustion plus usual care compared to sham moxibustion plus usual care for breech presentation.

Moxibustion plus usual care compared to sham moxibustion plus usual care for breech presentation
Patient or population: breech presentation Setting: hospital setting Intervention: moxibustion plus usual care Comparison: sham moxibustion plus usual care
Outcomes	Anticipated absolute effects* (95% CI)		Relative effect (95% CI)	№ of participants (studies)	Certainty of the evidence (GRADE)	Comments
	Risk with sham moxibustion plus usual care	Risk with moxibustion plus usual care
Non‐cephalic presentation at birth follow‐up: 2 weeks	57 per 100	42 per 100 (33 to 54)	RR 0.74 (0.58 to 0.95)	272 (1 RCT)	⊕⊕⊕⊝ Moderatea
Need for ECV ‐ not reported	‐	‐	‐	‐	‐
Caesarean section follow‐up: 2 weeks	60 per 100	51 per 100 (41 to 63)	RR 0.84 (0.68 to 1.04)	272 (1 RCT)	⊕⊕⊕⊝ Moderateb
Premature rupture of membranes ‐ not reported	‐	‐	‐	‐	‐
Use of oxytocin ‐ not reported	‐	‐	‐	‐	‐
Cord blood pH < 7.1 ‐ not reported	‐	‐	‐	‐	‐
Adverse events	Adverse events were reported in aggregate. 1 woman experienced a burn caused by moxibustion (it was not specified whether this was in the real or sham moxibustion group), and another withdrew because of abdominal pain after moxibustion. Adverse events were gastrointestinal disturbances (such as nausea, vomiting, and heartburn, 2%), dizziness (1.7%), mild hypertensive disorders (1.7%), abdominal pain (1.5%), and fetal hiccups (1.2%). The authors report no significant difference in the rate of these adverse events between the three groups included in the trial (moxibustion plus usual care, sham moxibustion plus usual care, usual care alone). Other adverse events that occurred during the trial were cord pathology (1.5%), failure to progress because of a lack of uterine contractions (1.5%), premature rupture of membranes (1.2%), and oligohydramnios (0.5%).			272 (1 RCT)	‐
*The risk in the intervention group (and its 95% confidence interval) is based on the assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI). CI: confidence interval; RR: risk ratio
GRADE Working Group grades of evidence High certainty: we are very confident that the true effect lies close to that of the estimate of the effect. Moderate certainty: 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 certainty: our confidence in the effect estimate is limited: the true effect may be substantially different from the estimate of the effect. Very low certainty: we have very little confidence in the effect estimate: the true effect is likely to be substantially different from the estimate of effect.
See interactive version of this table: https://gdt.gradepro.org/presentations/#/isof/isof_question_revman_web_430819416198452566.

^a We rated down by one level for serious imprecision because the optimal information size was not met (OIS = 174 per group).
^b We rated down by one level for serious imprecision because the optimal information size was not met (OIS = 478 per group).

Background

Description of the condition

Breech presentation, when the baby is bottom‐down, is common in the mid‐trimester of pregnancy, with the incidence of breech decreasing as the pregnancy approaches term. The incidence of breech presentation at term is reported to be 3% to 4% (Hofmeyr 2021). Breech presentation is associated with placenta praevia, multiple gestation, uterine abnormalities, poor uterine tone, prematurity or unknown causes, and is associated with primigravidae, older mothers, small‐for‐gestational‐age babies, and female babies (Hofmeyr 2021; Toijonen 2020). International clinical guidelines recommend using ultrasound to confirm breech presentation (Tsakiridis 2020).

Management of breech presentation remains controversial, and there is no international consensus (Tsakiridis 2020). External cephalic version (ECV), where the position of the baby is externally manipulated through the mother's abdominal wall to be head down, is an effective method of reducing the number of non‐cephalic presentations at birth and the need for caesarean section when performed after 36 weeks' gestation (Hofmeyr 2015). The benefits from ECV at term have been demonstrated in a Cochrane Review, which reported a reduction in the chance of non‐cephalic birth and caesarean section (Hofmeyr 2015). Complications of ECV are reported to be rare; however, the evidence to assess these complications is insufficient. Recent guidelines from the Royal College of Obstetricians and Gynaecologists, which draw on the findings from Hofmeyr 2015, suggest that the success rate of ECV is approximately 50% (Impey 2017). ECV should be offered at term (from 37 + 0 weeks' gestation), or may be offered at 36 + 0 weeks' gestation for nulliparous women (Impey 2017). Most international clinical practice guidelines recommend ECV (Tsakiridis 2020). 

Breech presentation during labour and birth brings increased risk of trauma and lack of oxygen for the baby (RANZCOG 2021). The Term Breech Trial, a large randomised trial of planned vaginal birth versus planned caesarean section for some breech presentations, reported a greater risk of perinatal mortality and serious neonatal morbidity in planned vaginal birth (Hannah 2000); however, there was no difference in the risk of death or neurodevelopmental delay for babies who were assessed at two years of age (Whyte 2004). Australian data show that almost 90% (89.3%) of breech babies are born by caesarean section (AIHW 2021), which also carries risks for the current and future births (RANZCOG 2021).

For women with breech presentation at birth, vaginal breech birth can be a safe option when certain criteria are met and appropriately skilled and experienced clinicians are available (Morris 2021). However, with between 69% and 100% of breech babies being born by caesarean section, few obstetricians and midwives have the opportunity to develop this skill (Morris 2021). When breech presentation is detected early (before term), women should be informed about the option for ECV (RANZCOG 2021). In facilities that are appropriately equipped to manage a planned vaginal birth, women who meet certain criteria should be offered a planned vaginal birth and counselled on the potential risks and benefits of an attempted vaginal birth (RANZCOG 2021). If a caesarean section is planned for birth of a breech fetus, the Royal Australian and New Zealand College of Obstetricians and Gynaecologists recommends that elective caesarean section be offered at 39 weeks' gestation where no other risk factors are present (RANZCOG 2021). Women should also be advised about the risks and benefits with caesarean section, including risks relating to future births (RANZCOG 2021). 

Description of the intervention

There is increasing interest in the exploration of complementary medicine for symptoms that arise during pregnancy and as preparation for labour. Moxibustion (a treatment method of traditional Chinese medicine) involves the burning of a herb (Artemisia vulgaris or Artemisia argyi Levl et vant.) close to the skin to induce a warming sensation (Turner 1991). Anecdotal evidence since 1991 has suggested that moxibustion to the acupuncture point Bladder 67 (BL67) (Chinese name Zhiyin), located at the tip of the fifth toe, may correct breech presentation (Cardini 1991). Women with a breech presentation may seek treatment with moxibustion from 32 to 38 weeks (West 2003). 

Moxibustion for breech presentation involves holding moxa sticks (approximately 1.5 cm in diameter and 20 cm in length) or burning moxa cones on or over BL67 on both fifth toes. The heat felt should be warm but not uncomfortable. The woman (if she is able) or a family member or friend can be taught how to administer moxibustion. Treatment regimens vary, and there is no accepted protocol for using moxibustion for breech presentation (Schlaeger 2018). A Delphi study design was used by Smith 2014 to generate a consensus for the self‐administration of moxibustion by women with breech presentation. The consensus recommended that moxibustion should commence between 34 and 35 weeks' gestation, using smokeless and odourless moxa sticks. The consensus also recommended that moxibustion should be applied once daily for 30 minutes, for at least 10 days. This protocol has yet to be tested (Schlaeger 2018).

Like most interventions, moxibustion is not free from side effects (Park 2010). It has been suggested that the smoke generated while burning moxa may irritate the respiratory tract, although there is no evidence to support this. The effect of in utero exposure to moxa smoke is unknown (Ewies 2002). There is also a risk of blistering the skin with moxibustion; however, due attention reduces this risk significantly, especially when it is applied by an appropriately trained professional.

How the intervention might work

Moxa sticks have been shown to emit visible light and radiate near‐infrared, short‐wavelength infrared, medium‐wavelength infrared, and long‐wavelength infrared radiation, ranging from 0.8 to 5.6 μm (Zhang 2013). Others have suggested that moxa radiates primarily long‐wavelength infrared radiation, stimulating the superficial skin where heat receptors are located (Pach 2009). The wavelength emitted with moxibustion appears to vary depending on whether moxibustion is direct (applied directly on the skin or placed on a medium, such as thin slices of garlic or ginger) or indirect (using a moxa stick held 1 to 2 cm above the skin), and whether the smoke moxa stick or smokeless moxa stick is used (Zhang 2013). Thermal effects have also been observed on the skin and in the subcutaneous layer (Zhang 2013). Due to the limited skin penetration of long‐wavelength radiation from moxa sticks, thermal effects on internal organs are more likely to arise from reflex mechanisms (Kim 2011; Zhang 2013). 

The mechanisms for moxibustion in changing breech presentation to cephalic presentation are unclear, and the biological plausibility of the intervention has not been determined. It has been hypothesised that this technique stimulates the production of maternal hormones (placental oestrogens and prostaglandin) and encourages the lining of the uterus to contract which, in turn, stimulates fetal activity (Cooperative 1984; West 2003), thereby increasing the chance that the baby will turn cephalic (Vas 2009). This proposed mechanism of action has not yet been determined. 

Why it is important to do this review

Several techniques have been used to change the presentation prior to term to reduce the incidence of breech presentation at term. Current management options include ECV. External cephalic version is usually performed at or after 37 weeks and appears to reduce the chance of breech births and caesarean delivery when performed before 37 weeks' gestation (Hutton 2015).

Postural techniques are often suggested by doctors and midwives to promote version of the baby. A Cochrane Review found insufficient evidence to support the use of postural management in changing breech presentation and reducing caesarean section rates and Apgar scores less than seven at 1 minute (Hofmeyr 2012). Increasingly, women are seeking other forms of health care to complement their pregnancy care. Moxibustion is one such modality.

We published the first version of this systematic review in 2005 (Coyle 2005), and the second version in 2012 (Coyle 2012). Since that time, new trials have been published. This review examined the evidence supporting the use of moxibustion for breech presentation. This is an update of a review last published in 2012.

Objectives

To examine the effectiveness and safety of moxibustion on changing the presentation of an unborn baby in the breech position, the need for external cephalic version, mode of birth, and perinatal morbidity and mortality.

Methods

Criteria for considering studies for this review

Types of studies

All published and unpublished randomised controlled trial (RCT) or quasi‐RCTs comparing moxibustion with a control group (no moxibustion) or other methods (e.g. ECV, acupuncture).

Types of participants

Women with a singleton breech presentation.

Types of interventions

Moxibustion (entailing application of heat to the acupuncture point BL67) alone or in combination with other techniques, such as acupuncture, acupressure, and postural techniques. Given that women are receiving usual antenatal care while participating in the trials, the intervention tested in the included studies was moxibustion (with or without other interventions) plus usual care. Moxibustion is often combined with other acupuncture therapies in clinical practice. Studies that tested moxibustion alone or in combination with other acupuncture therapies compared to sham/placebo moxibustion, acupuncture or moxibustion were eligible.

Eligible comparators were usual care alone, acupuncture techniques plus usual care, sham or placebo plus usual care, and postural therapy plus usual care. Examples of comparisons eligible for inclusion were: 

• moxibustion plus usual care versus usual care (no moxibustion);

• moxibustion plus usual care versus sham moxibustion plus usual care;

• moxibustion plus usual care versus acupuncture plus usual care;

• moxibustion plus acupuncture plus usual care versus usual care;

• moxibustion plus postural therapy versus postural therapy.

Types of outcome measures

Primary outcomes

• Baby's presentation at birth (non‐cephalic presentation).

• Need for ECV.

Secondary outcomes

• Mode of birth (caesarean section).

• Perinatal morbidity and mortality (e.g. Apgar score < 7 at 5 minutes, intrauterine fetal death, cord blood pH less than 7.1, neonatal intensive care unit (NICU) admission).

• Maternal complications (e.g. premature rupture of membranes, use of oxytocin, placental abruption).

• Maternal satisfaction.

• Adverse events of moxibustion (e.g. burns, respiratory irritation, increased fetal movements, uterine contractions).

Search methods for identification of studies

Electronic searches

For this update, we searched Cochrane Pregnancy and Childbirth’s Trials Register by contacting their Information Specialist (3 November 2021). The Register is a database containing over 27,000 reports of controlled trials in the field of pregnancy and childbirth. It represents over 30 years of searching. For the full current search methods used to populate Cochrane Pregnancy and Childbirth’s Trials Register – including the detailed search strategies for CENTRAL, MEDLINE, Embase, and CINAHL; the list of handsearched journals and conference proceedings; and the list of journals reviewed via the current awareness service – please follow this link.

Briefly, Cochrane Pregnancy and Childbirth’s Trials Register is maintained by their Information Specialist and contains trials identified from:

1. monthly searches of the Cochrane Central Register of Controlled Trials (CENTRAL), which includes centralised searches of ClinicalTrials.gov and the WHO ICTRP;

2. weekly searches of MEDLINE (Ovid);

3. weekly searches of Embase (Ovid);

4. monthly searches of CINAHL (EBSCO);

5. handsearches of 30 journals and the proceedings of major conferences;

6. weekly current awareness alerts for a further 44 journals plus monthly BioMed Central email alerts.

Search results are screened by two people, and the full text of all relevant trial reports identified through the searching activities described above is reviewed. Based on the intervention described, each trial report is assigned a number that corresponds to a specific Pregnancy and Childbirth review topic (or topics) and is then added to the Register. The Information Specialist searches the Register for each review using this topic number rather than keywords. This results in a more specific search set that has been fully accounted for in the relevant review sections (Included studies, Excluded studies, or Studies awaiting classification).

In addition, we searched ClinicalTrials.gov and the WHO International Clinical Trials Registry Platform (ICTRP) for unpublished, planned, and ongoing trial reports (4 November 2021) using the search methods detailed in Appendix 1.

We also conducted comprehensive searches of five English language biomedical databases: we searched MEDLINE, Embase, CINAHL, AMED, and MIDIRS from inception to 3 November 2021 (see Appendix 2). The search strategy was modified from the previous version to overcome changes in database settings. Results were exported to reference management software. Duplicates were removed, titles and abstracts were screened, and full‐text articles were reviewed for eligibility.

Searching other resources

We consulted university departments of complementary medicine, nursing, and midwifery to locate other published works, conference headings, and text works. We also searched bibliographies of relevant papers. We did not apply any language restrictions.

Data collection and analysis

For the methods used when assessing the trials identified in the previous versions of this review, see Appendix 3. For this update, we used the following methods for assessing the trials identified by the updated search.

Selection of studies

Two review authors (MC and CS) independently assessed the identified studies for eligibility, except the report by Do 2011, for which CS was an author. The eligibility of this study was assessed independently by review authors MC and BP.

Where there was uncertainty about the eligibility of the study, the full text was retrieved. We contacted the original author for further information if necessary. Any disagreements were resolved through discussion or by consulting the third review author if required. We provided reasons for the exclusion of excluded studies.

Screening eligible studies for trustworthiness

All studies meeting our inclusion criteria were evaluated by at least two review authors against predefined criteria to select studies that, based on the available information, were deemed to be sufficiently trustworthy to be included in the analysis. The Cochrane Pregnancy and Childbirth Group have developed a Trustworthiness Screening Tool (CPC‐TST), which includes the following criteria.

Research governance

• Are there any retraction notices or expressions of concern listed in the Retraction Watch Database relating to this study?

• Was the study prospectively registered (for those studies published after 2010)? If not, was there a plausible reason?

• When requested, did the trial authors provide/share the protocol and/or ethics approval letter?

• Did the trial authors engage in communication with the Cochrane Review authors within the agreed timelines?

• Did the trial authors provide individual patient data upon request? If not, was there a plausible reason?

Baseline characteristics

• Is the study free from characteristics of the study participants that appear too similar (e.g. distribution of the mean and standard deviation (SD) that were excessively narrow or excessively wide, as noted by Carlisle 2017)?

Feasibility

• Is the study free from characteristics that could be implausible? (e.g. large numbers of women with a rare condition (such as severe cholestasis in pregnancy) recruited within 12 months). In cases with (close to) zero losses to follow‐up, is there a plausible explanation?

Results

• Is the study free from results that could be implausible (e.g. massive risk reduction for main outcomes with small sample size)?

• Do the numbers randomised to each group suggest that adequate randomisation methods were used (e.g. is the study free from issues such as unexpectedly even numbers of women ‘randomised’ including a mismatch between the numbers and the methods; if the authors say ‘no blocking was used’ but still end up with equal numbers; or if the authors say they used ‘blocks of 4’ but the final numbers differ by 6)?

• We did not include studies assessed as being potentially ‘high risk’ in the review. Where a study was classified as ‘high risk’, we attempted to contact the study authors to address any possible lack of information/concerns. In cases where we could not obtain contact details for the study authors, or where there was still inadequate information, the study remained in the ‘awaiting classification’ section, and the reasons and communications with the author (or lack thereof) were described in detail.

Abstracts

• We planned to only include data from abstracts if, in addition to the trustworthiness assessment, the study authors confirmed in writing that the data to be included in the review have come from the final analysis and will not change. If such information is not available/provided, the study will remain in the ‘awaiting classification’ section (as above).

See Figure 1 for details of how we applied the trustworthiness screening criteria.

1.

Applying the Cochrane Pregnancy and Childbirth Trustworthiness Screening Tool (TST).

Data extraction and management

We used the standard data extraction template of the Cochrane Pregnancy and Childbirth Group. For eligible studies, three review authors (MC, CS, and BP) extracted data relating to participants, methods, interventions, outcomes, and results using the template. Any discrepancies were resolved through discussion or by consulting the third review author if required. Data were entered into RevMan Web and checked for accuracy (RevMan Web 2022). Where information regarding the study characteristics or results was unclear, we attempted to contact authors of the original reports to provide further details. We sent one request for clarification, but did not receive a reply. We requested the trial report for one trial that was identified through the clinical trial register search, and this was shared by the authors. 

Assessment of risk of bias in included studies

Three review authors (MC, CS, and BP) independently assessed the risk of bias for each included study using the criteria outlined in the Cochrane Handbook for Systematic Reviews of Interventions (Higgins 2021). Any disagreements were resolved through discussion or by involving the third review author.

(1) Random sequence generation (checking for possible selection bias)

For each included study, we described the method used to generate the allocation sequence in sufficient detail to allow an assessment of whether it should produce comparable groups. We assessed the method as:

• low risk of bias (any truly random process, e.g. random number table; computer random number generator);

• high risk of bias (any non‐random process, e.g. odd or even date of birth; hospital or clinic record number);

• unclear risk of bias.   

(2) Allocation concealment (checking for possible selection bias)

For each included study, we described the method used to conceal allocation to interventions prior to assignment and assessed whether the intervention allocation could have been foreseen in advance of, or during recruitment, or changed after assignment. We assessed the methods as:

• low risk of bias (e.g. telephone or central randomisation; consecutively numbered, sealed, opaque envelopes);

• high risk of bias (e.g. open random allocation; unsealed or non‐opaque envelopes, alternation; date of birth);

• unclear risk of bias.   

(3) Blinding of participants, personnel, and outcome assessment (checking for possible performance bias)

For each included study, we described the methods used, if any, to blind study participants and personnel from the knowledge of which intervention a participant received. We considered studies to be at low risk of bias if they were blinded, or if we judged that lack of blinding would be unlikely to affect the results. We assessed the methods as:

• low, high, or unclear risk of bias for participants and personnel;

• low, high, or unclear risk of bias for outcome assessors.

(4) Incomplete outcome data (checking for possible attrition bias due to the amount, nature, and handling of incomplete outcome data)

For each included study, we described the completeness of data, including attrition and exclusions from the analysis. We stated whether attrition and exclusions were reported and the numbers included in the analysis at each stage (compared with the total randomised participants), reasons for attrition or exclusion where reported, and whether missing data were balanced across groups or were related to outcomes.  Where sufficient information was reported, or was supplied by the trial authors, we re‐included missing data in our analyses. We assessed methods as:

• low risk of bias (e.g. no missing outcome data; missing outcome data balanced across groups);

• high risk of bias (e.g. numbers or reasons for missing data imbalanced across groups; ‘as treated’ analysis done with substantial departure of the intervention received from that assigned at randomisation);

• unclear risk of bias.

(5) Selective reporting (checking for reporting bias)

For each included study, we described how we investigated the possibility of selective outcome reporting bias and what we found. We assessed the methods as:

• low risk of bias (e.g. where it was clear that all the study’s prespecified outcomes and all expected outcomes of interest to the review have been reported);

• high risk of bias (e.g. where not all of the study’s prespecified outcomes have been reported; one or more reported primary outcomes were not prespecified; outcomes of interest were reported incompletely such that they could not be used; the study failed to include results of a key outcome that would be expected to have been reported);

• unclear risk of bias.

(6) Other bias (checking for bias because of problems not covered by items 1 to 5)

For each included study, we described any important concerns we had about other possible sources of bias. We assessed whether each study was free of other problems that could put it at:

• low risk of other bias;

• high risk of other bias;

• unclear risk of other bias.

Measures of treatment effect

Dichotomous data

For dichotomous data, we presented results as summary risk ratio (RR) with 95% confidence intervals (CI). 

Continuous data

For continuous data, we used the mean difference (MD) with 95% CI if outcomes were measured in the same way between trials. We planned to use the standardised mean difference (SMD) to combine trials that measured the same outcome using different methods.  

Unit of analysis issues

Due to the nature of the intervention and types of participants, it is unlikely that a cross‐over or cluster‐randomised trial has been performed. However, in the event that this occurred, cross‐over trials would be subject to within‐participant analysis. We planned that if we identified cluster‐randomised trials, we would adjust their sample sizes using the methods described in the Cochrane Handbook for Systematic Reviews of Interventions (Higgins 2021) employing an estimate of the intracluster correlation coefficient (ICC) derived from the trial (if possible), or from a similar trial or from a study of a similar population. If we used ICCs from other sources, we would report this and conduct sensitivity analyses to investigate the effect of variation in the ICC. If we identified both cluster‐randomised trials and individually randomised trials, we would synthesise the relevant information. We considered it reasonable to combine the results from both when there was little heterogeneity between study designs, and interaction between the effect of the intervention and the choice of randomisation unit was thought to be unlikely.

We identified no cross‐over or cluster‐randomised trials.

Dealing with missing data

We examined studies to determine whether intention‐to‐treat analysis was used. We did not impute missing data, but we did report the proportion lost to follow‐up and analysed per protocol.

For included studies, we noted the levels of attrition. We explored the effect of including studies with high levels of missing data in the overall assessment of treatment effect through sensitivity analysis, where possible. The denominator for each outcome in each trial was the number randomised minus any participants whose outcomes were known to be missing.

Assessment of heterogeneity

We identified and measured heterogeneity by visually inspecting the overlaps of the CIs for the results of individual studies. If there was poor overlap, this was suggestive of statistical heterogeneity, and we included a more formal Chi² test. We assessed statistical heterogeneity in each meta‐analysis using the Tau², I², and Chi² statistics. We considered heterogeneity to be substantial if the Tau² was greater than zero, and either the I² was greater than 30% or there was a low P value (less than 0.10) in the Chi² test for heterogeneity. We measured inconsistency across trials in the meta‐analysis using I², which describes the percentage of total variation across studies due to heterogeneity rather than to chance (Higgins 2021).

Assessment of reporting biases

Had there been 10 or more studies in the meta‐analysis, we would have investigated reporting biases (such as publication bias) using funnel plots. We would have assessed funnel plot asymmetry visually and used formal tests for funnel plot asymmetry. We planned to use the test proposed by Egger 1997 for continuous outcomes, and the test proposed by Harbord 2006 for dichotomous outcomes. If asymmetry was detected in any of these tests or was suggested by a visual assessment, we would perform exploratory analyses to investigate it.

Data synthesis

We carried out statistical analysis using RevMan Web software (RevMan Web 2022). Where moxibustion was used in combination with another intervention, we made a separate comparison due to incompatibility of the interventions. We used fixed‐effect meta‐analysis to combine data where it was reasonable to assume that studies were estimating the same underlying treatment effect (i.e. where trials were examining the same intervention, and the trials’ populations and methods were considered to be sufficiently similar). If there was clinical heterogeneity sufficient to expect that the underlying treatment effects would differ between trials, or if substantial statistical heterogeneity was detected, we used random‐effects meta‐analysis to produce an overall summary if an average treatment effect across trials was considered to be clinically meaningful. The random‐effects summary was treated as the average range of possible treatment effects, and we discussed the clinical implications of treatment effects differing between trials. If the average treatment effect was not clinically meaningful, we did not combine trials. When we used random‐effects analyses, the results were presented as the average treatment effect with 95% CIs, and the estimates of Tau² and I².

Subgroup analysis and investigation of heterogeneity

We planned to carry out the following subgroup analyses for the primary outcomes (baby's presentation at birth and need for ECV) for studies that: 

• included (a) multiparous women, (b) nulliparous women, and (c) both primiparous and multiparous women;

• administered moxibustion (a) before 34 weeks' gestation, (b) at or after 34 weeks' gestation, and (c) both before and after 34 weeks' gestation.

We investigated substantial heterogeneity using subgroup analyses and sensitivity analyses. We considered whether an overall summary was meaningful; if it was, a random‐effects analysis was undertaken. For fixed‐effect inverse variance meta‐analyses, we assessed differences between subgroups by interaction tests. For random‐effects and fixed‐effect meta‐analyses using methods other than inverse variance, we assessed differences between subgroups by inspection of the subgroups’ CIs; non‐overlapping CIs indicated a statistically significant difference in treatment effect between the subgroups. The small number of studies in some subgroups meant that the results of subgroup analyses were not meaningful. 

Sensitivity analysis

We planned that if heterogeneity could not be explained through subgroup analysis, we would analyse the data using a random‐effects meta‐analysis. Where studies were at high risk of bias, we assessed the likely magnitude and direction of the bias, and whether it was likely to influence the findings of the review. We planned to perform sensitivity analysis to explore the effect of the level of bias. Sensitivity analyses were not conducted.

Summary of findings and assessment of the certainty of the evidence

For this update, we assessed the certainty of the evidence using the GRADE approach (Schünemann 2013). We assessed the certainty of the body of evidence relating to the following outcomes for two clinically important comparisons:

• moxibustion plus usual care versus usual care;

• moxibustion plus usual care versus sham moxibustion plus usual care.

We selected the following clinically important outcomes to include in summary of findings tables:

• non‐cephalic presentation at birth;

• need for ECV;

• caesarean section;

• premature rupture of membranes;

• use of oxytocin;

• cord blood pH less than 7.1;

• adverse events.

We used GRADEpro GDT to import data from RevMan Web to create the summary of findings tables (GRADEpro GDT; RevMan Web 2022). We produced a summary of the intervention effect and a measure of certainty for each of the above outcomes using the GRADE approach. The GRADE approach uses five considerations (study limitations, consistency of effect, imprecision, indirectness, and publication bias) to assess the certainty of the evidence for each outcome. The evidence can be downgraded from 'high certainty' by one level for serious (or by two levels for very serious) limitations, depending on assessments for risk of bias, indirectness of evidence, inconsistency, imprecision of effect estimates, or potential publication bias.

Results

Description of studies

See Characteristics of included studies; Characteristics of excluded studies; Characteristics of studies awaiting classification.

Results of the search

In updating the review, we reassessed the two studies previously listed as awaiting classification (Li 1996; Millereau 2009). We were unable to obtain additional information to confirm the eligibility of the report by Li 1996 (specifically, disaggregated outcomes for women with breech presentation, transverse lie and occipitoposterior presentation), and the study was excluded. The methods and results sections of the Millereau 2009 study were translated, and the translation was verified by a person fluent in French. We judged the study to be eligible for inclusion, and included it in this update. Two ongoing studies from the previous version of the review were completed and results published (Smith 2009, see Do 2011; Vas 2008, see Vas 2013). We included both trials in this update. 

We assessed 26 new study reports from Cochrane Pregnancy and Childbirth and 4290 reports from our additional searches (see Figure 2). We assessed the full‐text articles of 59 new reports for eligibility, of which four require additional information to confirm eligibility (see Characteristics of studies awaiting classification) (Albaladejo 2017; Casas 2017; Yang 2007; Yang 2010).

2.

Flow chart of study selection.

This updated review includes 2181 women from 13 studies (Bue 2016; Cardini 1998; Cardini 2005; Coulon 2014; Do 2011; Guittier 2009; Higashihara 2021; Millereau 2009; Neri 2004; Neri 2007; Tsujiuchi 2017; Vas 2013; Yang 2006). 

Screening eligible studies for trustworthiness

Of the eight eligible studies identified from the original search, we judged that two studies required further information to assess trustworthiness (Chen 2004; Lin 2002). We have moved these studies to 'awaiting classification'. 

For four newly identified studies (Albaladejo 2017; Casas 2017; Chen 2007; Yang 2008), we had concerns about the plausibility of the results (see Characteristics of studies awaiting classification). We plan to contact the investigators to request further information, the outcome of which will be included in the next update to this review. Four of the new studies were prospectively registered (Do 2011; Higashihara 2021; Tsujiuchi 2017; Vas 2013), and two were retrospectively registered (Bue 2016; Coulon 2014). Recruitment for retrospectively registered studies commenced in 2003 and 2006, respectively. We considered the studies to be otherwise trustworthy. One study demonstrated no evidence of prospective registration (Albaladejo 2017), but details of ethical approval were provided.

Included studies

Design

All trials used a parallel design. Three trials included three arms (Higashihara 2021; Neri 2007; Vas 2013), while the remaining studies were two‐arm trials. In the three‐arm trial by Higashihara 2021, two treatment arms were included: one treatment arm used a moxibustion stick that produces smoke as it burns, and the other used a smokeless moxibustion stick. We considered these two treatment arms to be sufficiently similar that they could be merged for data analysis of clinical outcomes. Neri 2007 compared moxibustion with acupuncture and with the combination of moxibustion plus acupuncture. Vas 2013 included two control groups: one received sham moxibustion plus usual care, and the other received usual care alone.

Setting

Most trials recruited women from hospital clinics. Three studies were conducted in Italy (Cardini 2005; Neri 2004; Neri 2007); two each in mainland China (Cardini 1998; Yang 2006), France (Coulon 2014; Millereau 2009), and Japan (Higashihara 2021; Tsujiuchi 2017); and one each in Australia (Do 2011), Denmark (Bue 2016), Spain (Vas 2013), and Switzerland (Guittier 2009). Moxibustion was administered by staff in four studies (Coulon 2014; Guittier 2009; Neri 2004; Neri 2007), by staff for the first session then self‐administered by the patient or a family member or friend in one study (Millereau 2009), and was self‐administered by the patient or a family member or friend in seven studies (Bue 2016; Cardini 1998; Cardini 2005; Do 2011; Higashihara 2021; Tsujiuchi 2017; Vas 2013). Information about who administered the treatment was not available for one study (Yang 2006). 

Participants

Breech was confirmed by ultrasound in 10 studies (Bue 2016; Cardini 1998; Cardini 2005; Coulon 2014; Do 2011; Guittier 2009; Higashihara 2021; Millereau 2009; Neri 2004; Vas 2013), while the method of confirmation was not reported in the three remaining studies. Four trials recruited primiparous women only (Cardini 1998; Cardini 2005; Neri 2007; Tsujiuchi 2017); this information was not available for one study (Yang 2006). The gestational age at which women were recruited varied across trials, from 28 weeks', in Yang 2006, to 36 + 5 weeks' gestation, in Do 2011. Treatment extended past 37 weeks' gestation in two trials (Do 2011; Guittier 2009).

Interventions

Women in all trials received usual care, regardless of group allocation. In this updated review, we revised the descriptions of interventions accordingly. Interventions tested in the included trials were:

• moxibustion plus usual care (Bue 2016; Cardini 1998; Cardini 2005; Do 2011; Guittier 2009; Higashihara 2021; Millereau 2009; Neri 2007; Vas 2013);

• moxibustion plus acupuncture plus usual care (Coulon 2014; Neri 2004; Neri 2007; Tsujiuchi 2017);

• moxibustion plus postural therapy plus usual care (Yang 2006). 

Comparators were usual care alone (Bue 2016; Cardini 1998; Cardini 2005; Do 2011; Guittier 2009; Millereau 2009; Neri 2004; Tsujiuchi 2017; Vas 2013), sham moxibustion plus usual care (Vas 2013), acupuncture plus usual care (Neri 2007), laser placebo plus usual care (Coulon 2014), patient education plus usual care (Higashihara 2021), and postural therapy plus usual care (Yang 2006).

Treatment duration ranged from 14 minutes, in Tsujiuchi 2017, to 30 minutes, in Cardini 1998 and Cardini 2005, with the majority of studies applying moxibustion for 20 minutes. Treatment frequency was daily (Cardini 2005; Bue 2016; Guittier 2009; Millereau 2009; Vas 2013), once or twice daily (Cardini 1998; Higashihara 2021), twice daily (Do 2011; Tsujiuchi 2017; Yang 2006), on alternate days (Coulon 2014), or twice weekly (Neri 2004; Neri 2007). The intervention period ranged from seven days (Cardini 1998; Cardini 2005; Millereau 2009; Yang 2006) to two weeks (Bue 2016; Cardini 1998; Cardini 2005; Guittier 2009; Neri 2004; Vas 2013), with two studies providing treatment for seven or 14 days (Cardini 1998; Cardini 2005). One study provided treatment until breech presentation was changed to cephalic presentation (Tsujiuchi 2017).

Outcomes

All studies reported on at least one of the prespecified outcomes. The following studies reported data for the outcomes selected for our summary of findings tables.

• Non‐cephalic presentation at birth: 12 studies (Bue 2016; Cardini 1998; Cardini 2005; Coulon 2014; Do 2011; Guittier 2009; Higashihara 2021; Millereau 2009; Neri 2004; Neri 2007; Vas 2013; Yang 2006).

• Need for ECV: six studies (Cardini 1998; Coulon 2014; Do 2011; Guittier 2009; Higashihara 2021; Tsujiuchi 2017).

• Caesarean section: 10 trials (Bue 2016; Cardini 1998; Coulon 2014; Do 2011; Guittier 2009; Higashihara 2021; Millereau 2009; Neri 2004; Tsujiuchi 2017; Vas 2013).

• Premature rupture of membranes: five studies (Cardini 1998; Cardini 2005; Coulon 2014; Do 2011; Higashihara 2021).

• Use of oxytocin: one study (Cardini 1998).

• Cord blood pH less than 7.1: two studies (Guittier 2009; Higashihara 2021).

• Adverse events: nine studies (Bue 2016; Cardini 1998; Cardini 2005; Coulon 2014; Do 2011; Higashihara 2021; Neri 2004; Tsujiuchi 2017; Vas 2013), although not all studies reported data in a way that could be reanalysed.

Other relevant outcomes were non‐cephalic presentation at the end of treatment (8 studies; Bue 2016; Cardini 1998; Cardini 2005; Coulon 2014; Higashihara 2021; Neri 2004; Tsujiuchi 2017; Vas 2013), Apgar score less than seven at 5 minutes (7 studies; Bue 2016; Cardini 1998; Cardini 2005; Do 2011; Guittier 2009; Higashihara 2021; Vas 2013), successful ECV (6 studies; Bue 2016; Cardini 1998; Coulon 2014; Higashihara 2021; Tsujiuchi 2017; Vas 2013), maternal satisfaction and adherence to treatment (4 studies; Cardini 1998; Cardini 2005; Do 2011; Guittier 2009), fetal heart rate and movements (4 studies; Cardini 1998; Cardini 2005; Do 2011; Neri 2007), intrauterine fetal death (3 studies; Cardini 1998; Coulon 2014; Higashihara 2021), NICU admission (2 studies; Do 2011; Higashihara 2021), duration of stay (Guittier 2009), placental abruption (Cardini 2005), and pre‐eclampsia (Tsujiuchi 2017).

Some studies described clinical outcomes as adverse events (e.g. premature rupture of membranes, Cardini 2005; Do 2011). These were analysed as clinical outcomes rather than adverse events. Several studies that reported safety of moxibustion did not report the number of events by group (Bue 2016; Cardini 1998; Neri 2004; Vas 2013), or only reported the number of adverse events in the intervention group (Tsujiuchi 2017). Consequently, statistical analysis of adverse events was not possible for these studies. 

Study dates

Two studies either did not report the study dates or they were not available (Neri 2007; Yang 2006). One study was conducted in the 1990s (Cardini 1998); eight studies were conducted in the 2000s (Bue 2016; Cardini 2005; Coulon 2014; Do 2011; Guittier 2009; Millereau 2009; Neri 2004; Vas 2013); and two studies were conducted after 2010 (Higashihara 2021; Tsujiuchi 2017). 

Sources of trial funding

The source of funding was not reported for seven studies (Bue 2016; Coulon 2014; Guittier 2009; Millereau 2009; Neri 2004; Neri 2007; Yang 2006). The remaining seven studies received funding support from healthcare services, public, or educational sources.

Trial authors' declarations of interest

In eight studies, the authors declared that they had no competing interests (Bue 2016; Cardini 2005; Coulon 2014; Do 2011; Guittier 2009; Higashihara 2021; Tsujiuchi 2017; Vas 2013). Assessment of declarations of interest was not available for one study (Yang 2006), for which data were extracted by a third party in the original review (Coyle 2005). No such declarations were made in the remaining studies.

Sample size

The sample sizes of included studies ranged from 14, in Tsujiuchi 2017, to 406, in Vas 2013. Eight studies included more than 100 participants (Bue 2016; Cardini 1998; Cardini 2005; Coulon 2014; Guittier 2009; Neri 2004; Vas 2013; Yang 2006). The mean and median sample size was 162 participants. Nine studies reported conducting a sample size calculation (Bue 2016; Cardini 1998; Cardini 2005; Coulon 2014; Guittier 2009; Higashihara 2021; Millereau 2009; Neri 2004; Vas 2013).

Excluded studies

Thirty‐nine of the 59 full‐text publications that were assessed for eligibility were excluded for the following reasons.

• 29 studies did not meet the inclusion criteria:  

  • 13 studies were not clinical studies (Anon 2019; Anon 2021; Balk 2005; Brill 2003; Ernst 1999; Ernst 2005; Lee 2008; Mitchell 2008; Smith 2013; van den Berg 2005; van den Berg 2010a; White 2010; Wong 1999);

  • 9 studies were not RCTs or quasi‐RCTs (Hayashida 1989; Li 1996; Miranda Garcia 2017; Pardeshi 2020; Rosim 2017; Sarsmaz 2021; van den Berg 2006; van den Berg 2010b; Wen 1979);

  • 3 studies did not use moxibustion as the intervention (Fu 1990; Langer 2016; Sananes 2016);

  • 2 studies did not report on our outcomes of interest (García‐Mochón 2016; Neri 2002);

  • 2 studies did not report results data according to group allocation (Guittier 2008; Miranda Garcia 2011).

• 5 studies were included in the 2012 version of this review (Cardini 1998; Cardini 2005; Guittier 2009; Neri 2004; Neri 2007).

• 4 were duplicate reports: 1 of Do 2011 (Dahlen 2011), 2 of Vas 2013 (Vas 2012; Vas 2014), and 1 was a duplicate of an article excluded after title and abstract screening (Baccetti 2012).

• 1 clinical trial registration was for a trial that did not proceed (ISRCTN98448646).

For details of excluded studies, see Characteristics of excluded studies.

Risk of bias in included studies

All studies were described as randomised, but only one study was judged to be of low or minimal bias on all seven risk of bias domains (Cardini 1998). The risk of bias assessments for individual studies are illustrated in Figure 3 and summarised in Figure 4. 

3.

Methodological quality summary: review authors' judgements about each methodological quality item for each included study.

4.

Methodological quality graph: review authors' judgements about each methodological quality item presented as percentages across all included studies.

Allocation

Sequence generation

Group allocation was determined by a computer‐generated sequence in seven trials (Cardini 1998; Cardini 2005; Do 2011; Guittier 2009; Millereau 2009; Neri 2004; Vas 2013), by a random number table in two trials (Bue 2016; Coulon 2014), and by date of admission in one trial (Yang 2006). Sequence generation was unclear in the Neri 2007 and Tsujiuchi 2017 trials. The quasi‐randomised trial by Higashihara 2021 allocated the first 20 participants to the first treatment group, the second 20 participants to the second treatment group, and the final 20 participants to the control group.

Allocation concealment

Group allocation was concealed by central randomisation in one trial (Neri 2004), by sealed envelopes in six trials (Bue 2016; Cardini 2005; Coulon 2014; Do 2011; Guittier 2009; Vas 2013 trials), and by drawing of lots in one trial (Millereau 2009). We judged the study by Higashihara 2021 to be at high risk of bias as there was no allocation concealment. The potential for bias of allocation concealment was unclear in the remaining trials.

Blinding

Blinding of participants and personnel

It was not feasible to blind the participant or therapist in the included studies; however, we judged that the outcomes and outcome measurements were not likely to have been influenced by the lack of blinding for most outcomes. One study reported subjective outcomes (e.g. maternal satisfaction) that could have been influenced by the lack of blinding (Do 2011); we assessed this study as at high risk for blinding of participants and personnel. For three studies, it was unclear whether the lack of blinding could have influenced some outcomes (Higashihara 2021; Neri 2007; Tsujiuchi 2017).

Blinding of outcome assessors

The outcome assessor was blinded to group allocation in the Cardini 2005 and Vas 2013 trials, while the obstetrician was blinded in the Millereau 2009 trial. The lack of blinding of outcome assessors was unlikely to have influenced most outcomes reported in the included studies. It was not clear whether the analyst was blinded to group allocation in any of the included studies.

Incomplete outcome data

No losses to follow up were reported in four trials, which were assessed as low risk (Guittier 2009; Millereau 2009; Tsujiuchi 2017; Yang 2006). Intention‐to‐treat analysis was performed in seven trials (Bue 2016; Cardini 1998; Cardini 2005; Coulon 2014; Do 2011; Guittier 2009; Vas 2013). In two trials (Higashihara 2021; Neri 2007), the number of withdrawals was small and balanced across groups, and was therefore considered unlikely to have influenced the outcome results. In the Cardini 1998 trial, nine women in the treatment group withdrew from the treatment but were not exclusions. Data were available for all women randomised (low risk of bias). 

In the Cardini 2005 trial, 14 women in the treatment group discontinued treatment but were not exclusions. One woman in the observation group was lost to follow‐up. We sought additional data from the authors concerning the use of ECV by group. Twelve women in the treatment group underwent ECV, eight of which were successful. Nine women in the control group underwent ECV, four of which were successful. The trial was interrupted when interim analysis revealed poor compliance and a high number of treatment interruptions (high risk of bias). In the Neri 2004 trial, eight women in the treatment group and six in the control group withdrew from the trial and were not included in the analyses (high risk of bias). 

Selective reporting

Seven trials were registered prospectively, Do 2011; Higashihara 2021; Tsujiuchi 2017; Vas 2013, or retrospectively, Bue 2016; Coulon 2014; Guittier 2009, in clinical trial registries. The studies by Do 2011 and Higashihara 2021 reported all specified outcomes and were judged as at low risk of bias. The trial protocol and trial registration were available for Vas 2013; however, differences in outcomes were noted between these and the trial report, therefore we judged this study as at unclear risk of bias. The trial registration for Tsujiuchi 2017 specified outcomes that were not included in the trial report (unclear risk of bias). We assessed the three studies that were retrospectively registered as at unclear risk of bias.

The trial reports of four trials that were published before protocols and prospective registration were required included all expected outcomes (Cardini 1998; Cardini 2005; Neri 2004; Neri 2007); however, since no trial registration or trial protocol was available, these studies were still assessed as at unclear risk of bias. The study protocols and trial registrations for the Millereau 2009 and Yang 2006 trials were unavailable (unclear risk of bias).

Other potential sources of bias

Sample size calculations were performed in nine trials (Bue 2016; Cardini 1998; Cardini 2005; Coulon 2014; Guittier 2009; Higashihara 2021; Millereau 2009; Neri 2004; Vas 2013). The Cardini 2005 trial was interrupted when interim analysis showed poor compliance and high numbers of treatment interruptions. We assessed this study as at high risk of bias. The treatment frequency was changed in Guittier 2009 after the first 12 women were recruited (after review of cardiotocographic data). It is unclear whether this could have influenced the results. 

Insufficient information precluded a determination of whether other potential sources of bias existed for four trials (Bue 2016; Millereau 2009; Tsujiuchi 2017; Yang 2006). We detected an anomaly in outcome results in the Bue 2016 trial. The source of funding was not specified in the Millereau 2009 trial. We observed unequal participant numbers in each group in the Neri 2007 trial, but it was unclear whether this resulted from a failure of randomisation. Tsujiuchi 2017 reported a difference in the age of participants, but it is unclear whether this could have influenced the outcome results.

Effects of interventions

See: Table 1; Table 2

We performed meta‐analyses where possible. It was not possible to perform the planned subgroup analyses for all comparisons because of the small number of studies or lack of meta‐analysis. We did not conduct subgroup analyses for parity and timing of the intervention due to the small number of studies that would have been included in the subgroups. See also Table 1; Table 2.

1. Moxibustion plus usual care versus usual care

Primary outcomes

1.1 Non‐cephalic presentation at birth

Seven trials (1152 women) that compared moxibustion with no treatment reported on the baby's presentation at birth (Bue 2016; Cardini 1998; Cardini 2005; Do 2011; Guittier 2009; Millereau 2009; Vas 2013). A meta‐analysis using a fixed‐effect model showed that moxibustion plus usual care probably reduces the chance of a non‐cephalic presentation at birth more than usual care alone (risk ratio (RR) 0.87, 95% confidence interval (CI) 0.78 to 0.99, I² = 38%; moderate‐certainty evidence; see Analysis 1.1).

1.1. Analysis.

Comparison 1: Moxibustion plus usual care versus usual care, Outcome 1: Non‐cephalic presentation at birth

1.2 Need for ECV

Four trials (692 women) reported on the need for ECV (Bue 2016; Cardini 1998; Do 2011; Guittier 2009). A meta‐analysis using a random‐effects model showed uncertainty about the average effect of moxibustion on the need for ECV compared with usual care alone because the CIs included both appreciable benefit and moderate harm (RR 0.62, 95% CI 0.32 to 1.21, I² = 78%; low‐certainty evidence; see Analysis 1.2). The results of this meta‐analysis should be interpreted with caution due to statistical heterogeneity and the comparatively extreme effect observed in Cardini 1998.

1.2. Analysis.

Comparison 1: Moxibustion plus usual care versus usual care, Outcome 2: Need for ECV

Secondary outcomes

1.3 Non‐cephalic presentation at the end of treatment

Four trials (853 women) assessed fetal presentation at the end of treatment (Bue 2016; Cardini 1998; Cardini 2005; Vas 2013). Women who received moxibustion plus usual care were no more or less likely to have a non‐cephalic presentation at the end of treatment than women who received usual care alone (RR 0.80, 95% CI 0.61 to 1.05, I² = 81%; see Analysis 1.3). Substantial statistical heterogeneity limits our confidence in this result, and it seems likely that the result is influenced by the comparatively extreme effect observed in Cardini 1998.

1.3. Analysis.

Comparison 1: Moxibustion plus usual care versus usual care, Outcome 3: Non‐cephalic presentation at the end of treatment

1.4 Caesarean section

Six trials reported on the number of births by caesarean section (Bue 2016; Cardini 1998; Do 2011; Guittier 2009; Millereau 2009; Vas 2013). One study reported the percentage of caesarean sections (Bue 2016), and the number of women was calculated for reanalysis. Among the 1030 women included in the meta‐analysis, moxibustion plus usual care probably has little to no effect on the likelihood of caesarean birth compared to usual care alone (RR 0.94, 95% CI 0.83 to 1.05, I² = 0%; moderate‐certainty evidence; see Analysis 1.4). Clinical heterogeneity may exist because of differences in the treatment regimens, therefore the results of this meta‐analysis should be interpreted with caution.

1.4. Analysis.

Comparison 1: Moxibustion plus usual care versus usual care, Outcome 4: Caesarean section

We noted an anomaly in the number of caesarean and vaginal births in the study by Bue 2016, where the total number of caesarean and vaginal births in the control group (n = 106) exceeded the number of participants allocated to that group (n = 98). We contacted the trial authors but did not receive a response by the time of completion of the review. We conducted sensitivity analysis excluding the data from that study. The magnitude and direction of the effect was similar to the result for the main analysis (RR 0.95, 95% CI 0.83 to 1.08, I² = 5%).

1.5 Premature rupture of membranes

Three trials (402 women) reported on the outcome premature rupture of membranes (Cardini 1998; Cardini 2005; Do 2011). Studies varied in the frequency and duration of treatment. A meta‐analysis using a random‐effects model showed uncertainty about the average effect of moxibustion plus usual care on the risk of premature rupture of membranes because there were very few data (RR 1.31, 95% CI 0.17 to 10.21, I² = 59%; low‐certainty evidence; see Analysis 1.5).

1.5. Analysis.

Comparison 1: Moxibustion plus usual care versus usual care, Outcome 5: Premature rupture of membranes

1.6 Preterm birth

Two trials (290 women) assessed preterm birth (Do 2011; Vas 2013). Meta‐analysis showed no statistical difference in the number of preterm births between women who received moxibustion plus usual care and those who received usual care alone (RR 1.17, 95% CI 0.39 to 3.54, I² = 24%; see Analysis 1.6). The wide CIs suggest that the true effect may include considerable benefit and considerable harm.

1.6. Analysis.

Comparison 1: Moxibustion plus usual care versus usual care, Outcome 6: Preterm birth

1.7 Use of oxytocin

One trial (260 women) reported on the use of oxytocin for women who had a vaginal birth (161 of 260 women had a vaginal birth) (Cardini 1998). Compared with usual care alone, moxibustion plus usual care probably reduces the use of oxytocin before or during labour for women who had vaginal births (7/81 in the intervention group, 25/80 in the control group; RR 0.28, 95% CI 0.13 to 0.60; moderate‐certainty evidence; see Analysis 1.7).

1.7. Analysis.

Comparison 1: Moxibustion plus usual care versus usual care, Outcome 7: Use of oxytocin

1.8 Apgar score less than seven at 5 minutes

Five trials (962 women) that compared moxibustion plus usual care with usual care alone assessed neonatal Apgar score less than seven at 5 minutes (Bue 2016; Cardini 1998; Do 2011; Guittier 2009; Vas 2013). Two studies reported no instances of this outcome in either the treatment or control group (Do 2011; Vas 2013), and therefore did not contribute to the meta‐analysis result (since the effects were not estimable). Meta‐analysis showed no difference between groups in the chance of the neonate having an Apgar score less than seven at 5 minutes (RR 0.33, 95% CI 0.09 to 1.22, I² = 45%; see Analysis 1.8). The fixed‐effect meta‐analysis showed moderate statistical heterogeneity; however, we were unable to identify specific clinical differences between the studies.

1.8. Analysis.

Comparison 1: Moxibustion plus usual care versus usual care, Outcome 8: Apgar score < 7 at 5 minutes

1.9 Intrauterine fetal death

One trial (260 women) reported on intrauterine fetal death (Cardini 1998). One instance of intrauterine fetal death was reported in the control group (RR 0.33, 95% CI 0.01 to 8.11; see Analysis 1.9). As such, there is little information about the effect of moxibustion plus usual care on this outcome.

1.9. Analysis.

Comparison 1: Moxibustion plus usual care versus usual care, Outcome 9: Intrauterine fetal death

1.10 Placental abruption

One trial (122 women) reported on placental abruption (Cardini 2005). One case of placental abruption was seen in the control group (RR 0.29, 95% CI 0.01 to 7.05; see Analysis 1.10). As such, there is little information about the effect of moxibustion plus usual care on this outcome.

1.10. Analysis.

Comparison 1: Moxibustion plus usual care versus usual care, Outcome 10: Placental abruption

1.11 Duration of hospital stay (days)

One trial (212 women) found no difference between groups in duration of hospital stay (days) (mean difference (MD) −0.10 shorter, 95% CI −0.55 shorter to 0.35 longer; see Analysis 1.11) (Guittier 2009). 

1.11. Analysis.

Comparison 1: Moxibustion plus usual care versus usual care, Outcome 11: Duration of hospital stay (days)

1.12 Cord blood pH less than 7.1

The evidence from one trial (212 women) is very uncertain about the average effect of moxibustion plus usual care on the likelihood of cord blood pH less than 7.1 compared to usual care alone (random‐effects model; RR 3.00, 95% CI 0.32 to 28.38; low‐certainty evidence; see Analysis 1.12) (Guittier 2009). Given that there were very few data, there is very limited information about the true effect of moxibustion plus usual care on this outcome.

1.12. Analysis.

Comparison 1: Moxibustion plus usual care versus usual care, Outcome 12: Cord blood pH less than 7.1

1.13 NICU admission

In one trial (20 women) (Do 2011), there were very few data and therefore very limited information about the true effect of moxibustion plus usual care on the number of neonates admitted to NICU (RR 5.00, 95% CI 0.27 to 92.62; see Analysis 1.13).

1.13. Analysis.

Comparison 1: Moxibustion plus usual care versus usual care, Outcome 13: NICU admission

1.14 Maternal satisfaction and adherence to treatment

Four trials reported on the acceptability of moxibustion and adherence to treatment (Cardini 1998; Cardini 2005; Do 2011; Guittier 2009). Meta‐analysis was not possible for these outcomes; therefore, a narrative synthesis was performed instead. In one study (Do 2011), women found moxibustion acceptable, not painful, and with few side effects, although some said that was 'hard to judge'. In total, 93% of participants adhered to the treatment protocol of twice‐daily moxibustion. There were few reports of uterine contractions after moxibustion. In the second study (Guittier 2009), women found moxibustion acceptable, and it was viewed favourably. Most women reported little or no pain (96%); the mean pain score during sessions was 8 mm (range 0 to 50) on a 100‐millimetre visual analogue scale. Nine women in the treatment group of the Cardini 1998 trial did not adhere to treatment: one did not comply with the schedule, and eight withdrew from the treatment (three on advice of their obstetrician). Recruitment for the Cardini 2005 trial was stopped early due to poor adherence to the treatment schedule (after 123 of the recruitment target of 260 participants were randomised).

1.15 Adverse events

One trial (122 women) reported adverse event data by study group (Cardini 2005). Women in the treatment group reported adverse events relating to treatment, including nausea, unpleasant odour (with or without throat problems), and abdominal pain from contractions (27 events, all in the treatment group, some of which resulted in withdrawal from the study; RR 48.33, 95% CI 3.01 to 774.86; very low‐certainty evidence; see Analysis 1.14). The evidence is very uncertain about adverse effects relating to moxibustion plus usual care.

1.14. Analysis.

Comparison 1: Moxibustion plus usual care versus usual care, Outcome 14: Adverse events

Other trials reported data in a way that did not permit reanalysis because the number of events were either not reported at all or not reported according to group allocation, or were assessed only in the intervention group. Meta‐analysis was not possible for these studies, and a narrative synthesis was performed instead. 

One trial reported adverse effects (Cardini 1998); however, actual data were not reported. The main discomfort reported by women in both groups in the Cardini 1998 trial was pressure and tenderness in the epigastric region or the hypochondria (epigastric crushing) attributed to the head of the breech baby pressing against maternal organs. The study also reported a higher mean number of fetal movements with moxibustion plus usual care than with usual care alone (P < 0.001). 

In the Guittier 2009 study, few women reported uterine contractions during the intervention period (the study reported that 93% had no contractions or fewer than 10 contractions per day, although it was not clear whether this was among all participants). Fifty‐three per cent of women perceived increased fetal movements. None of the participants reported a disturbing side effect during treatment; most women found moxibustion treatment 'pleasant or relaxing'. 

Bue 2016 reported that 92 of the 102 women allocated to the moxibustion group experienced general symptoms, such as increased fetal activity, uterine contractions, bad smell, headache and nausea. The authors reported that 32 women had fewer than 10 days of moxibustion treatment; however, it was not clear whether this was related to the adverse events. Adverse events in the control group were not reported.

One trial (272 women) that compared the combination of moxibustion plus usual care with sham moxibustion plus usual care and with usual care alone measured adverse events (Vas 2013). Adverse events were not reported according to group allocation. The study authors reported that there were no severe adverse events. One woman experienced a burn caused by moxibustion (it was not specified whether this was in the real or sham moxibustion group), and another withdrew because of abdominal pain after moxibustion. Adverse events included gastrointestinal disturbances (such as nausea, vomiting, and heartburn, 2%), dizziness (1.7%), mild hypertensive disorders (1.7%), abdominal pain (1.5%), and fetal hiccups (1.2%). The authors reported no significant difference in the rate of these adverse events between the three groups that were included in the trial.

Other adverse events that occurred in the trial were cord pathology (1.5%), failure to progress because of a lack of uterine contractions (1.5%), premature rupture of membranes (1.2%), and oligohydramnios (0.5%).

In the study by Do 2011, five women who received moxibustion reported more fetal movements than normal; these data were not collected for the control group. 

2. Moxibustion plus usual care versus sham moxibustion plus usual care

Primary outcomes

The single study that reported this comparison did not report the other primary outcome, need for ECV.

2.1 Non‐cephalic presentation at birth

One trial (272 women) that compared moxibustion plus usual care with sham moxibustion plus usual care reported the number of non‐cephalic presentations at birth (Vas 2013). Moxibustion plus usual care probably reduces the chance of a non‐cephalic presentation at birth compared to sham moxibustion plus usual care (RR 0.74, 95% CI 0.58 to 0.95; moderate‐certainty evidence; see Analysis 2.1). 

2.1. Analysis.

Comparison 2: Moxibustion plus usual care versus sham moxibustion plus usual care, Outcome 1: Non‐cephalic presentation at birth

Secondary outcomes

The single study that reported this comparison did not report on maternal satisfaction.

2.2 Non‐cephalic presentation at the end of treatment

One trial (272 women) assessed the number of non‐cephalic presentations at the end of treatment (Vas 2013). Moxibustion plus usual care reduced the chance of a non‐cephalic presentation at the end of treatment compared to sham moxibustion plus usual care (RR 0.79, 95% CI 0.67 to 0.92; see Analysis 2.2).

2.2. Analysis.

Comparison 2: Moxibustion plus usual care versus sham moxibustion plus usual care, Outcome 2: Non‐cephalic presentation at the end of treatment

2.3 Caesarean section

In one trial (272 women) (Vas 2013), moxibustion plus usual care probably results in little to no effect on the chance of caesarean section compared to sham moxibustion plus usual care (RR 0.84, 95% CI 0.68 to 1.04; moderate‐certainty evidence; see Analysis 2.3).

2.3. Analysis.

Comparison 2: Moxibustion plus usual care versus sham moxibustion plus usual care, Outcome 3: Caesarean section

2.4 Preterm birth

Four women in each group of one trial (272 women) had a preterm birth (Vas 2013). There was no statistical difference between the two groups for this outcome (RR 1.00, 95% CI 0.26 to 3.92; see Analysis 2.4).

2.4. Analysis.

Comparison 2: Moxibustion plus usual care versus sham moxibustion plus usual care, Outcome 4: Preterm birth

2.5 Apgar score less than seven at 5 minutes

One trial (272 women) assessed the number of neonates with an Apgar score less than seven at 5 minutes (Vas 2013). None of the babies born to trial participants had an Apgar score less than seven at 5 minutes (see Analysis 2.5).

2.5. Analysis.

Comparison 2: Moxibustion plus usual care versus sham moxibustion plus usual care, Outcome 5: Apgar score < 7 at 5 minutes

2.6 Adverse events

Vas 2013 reported adverse events in aggregate for all groups. The adverse events are described in the comparison moxibustion plus usual care versus usual care (see Section 1.15).

3. Moxibustion plus usual care versus acupuncture plus usual care

Primary outcomes

The single study that reported this comparison did not report on the other prespecified primary outcome, need for ECV, or on any of the prespecified secondary outcomes.

3.1 Non‐cephalic presentation at birth

There is very little evidence about the effect of moxibustion plus usual care compared to acupuncture plus usual care on the number of non‐cephalic presentations at birth (1 trial, 25 women; RR 0.25, 95% CI 0.09 to 0.72; see Analysis 3.1) (Neri 2007).

3.1. Analysis.

Comparison 3: Moxibustion plus usual care versus acupuncture plus usual care, Outcome 1: Non‐cephalic presentation at birth

4. Moxibustion plus usual care versus patient education plus usual care

4.1 Non‐cephalic presentation at birth

There is very little evidence from one trial (60 women) about the effect of moxibustion plus usual care compared to patient education plus usual care on the chance of non‐cephalic presentation at birth (RR 0.82, 95% CI 0.48 to 1.38; see Analysis 4.1) (Higashihara 2021).

4.1. Analysis.

Comparison 4: Moxibustion plus usual care versus patient education plus usual care, Outcome 1: Non‐cephalic presentation at birth

4.2 Need for ECV

There is very little evidence from one trial (60 women) about the effect of moxibustion plus usual care compared to patient education plus usual care on the need for ECV (RR 0.30, 95% CI 0.08 to 1.13; see Analysis 4.2) (Higashihara 2021).

4.2. Analysis.

Comparison 4: Moxibustion plus usual care versus patient education plus usual care, Outcome 2: Need for ECV

Secondary outcomes

The single study that reported this comparison did not report on maternal satisfaction.

4.3 Non‐cephalic presentation at the end of treatment

There is very litte evidence from one trial (60 women) about the effect of moxibustion plus usual care compared to patient education plus usual care on the chance of non‐cephalic presentation at the end of treatment (RR 0.63, 95% CI 0.42 to 0.96; see Analysis 4.3) (Higashihara 2021).

4.3. Analysis.

Comparison 4: Moxibustion plus usual care versus patient education plus usual care, Outcome 3: Non‐cephalic presentation at the end of treatment

4.4 Caesarean section

There is very little evidence about the effect of moxibustion plus usual care compared to patient education plus usual care on the rate of caesarean section (1 trial, 60 women; RR 0.77, 95% CI 0.49 to 1.20; see Analysis 4.4) (Higashihara 2021).

4.4. Analysis.

Comparison 4: Moxibustion plus usual care versus patient education plus usual care, Outcome 4: Caesarean section

4.5 Premature rupture of membranes

There were no cases of premature rupture of membranes in the one trial (60 women) that compared moxibustion plus usual care with patient education plus usual care (see Analysis 4.5) (Higashihara 2021). As such, there is very little information about the effect of moxibustion plus usual care compared to patient education plus usual care for this outcome. 

4.5. Analysis.

Comparison 4: Moxibustion plus usual care versus patient education plus usual care, Outcome 5: Premature rupture of membranes

4.6 Preterm birth

In the one trial (60 women) that compared moxibustion plus usual care with patient education plus usual care (Higashihara 2021), there were no cases of preterm birth (see Analysis 4.6). As such, there is very little information about the effect of moxibustion plus usual care for this outcome.

4.6. Analysis.

Comparison 4: Moxibustion plus usual care versus patient education plus usual care, Outcome 6: Preterm birth

4.7 Apgar score less than seven at 5 minutes

There is very little evidence about the effect of moxibustion plus usual care compared to patient education plus usual care on the number of neonates with an Apgar score less than seven at 5 minutes (1 trial, 60 women; RR 1.54, 95% CI 0.07 to 36.11; see Analysis 4.7) (Higashihara 2021).

4.7. Analysis.

Comparison 4: Moxibustion plus usual care versus patient education plus usual care, Outcome 7: Apgar score < 7 at 5 minutes

4.8 Cord blood pH less than seven

There were no instances of cord blood pH less than seven in one trial (60 women) (see Analysis 4.8) (Higashihara 2021). As such, there is very little information about the effect of moxibustion plus usual care on this outcome.

4.8. Analysis.

Comparison 4: Moxibustion plus usual care versus patient education plus usual care, Outcome 8: Cord blood pH less than 7

4.9 Intrauterine fetal death

In one trial (60 women) (Higashihara 2021), no instances of intrauterine fetal death were reported in either group (see Analysis 4.9); consequently, there is very little information about the effect of moxibustion plus usual care on this outcome. 

4.9. Analysis.

Comparison 4: Moxibustion plus usual care versus patient education plus usual care, Outcome 9: Intrauterine fetal death

4.10 NICU admission

One trial (60 women) provided very few data, hence there is very limited information about the true effect of moxibustion plus usual care on the likelihood of having a baby admitted to NICU compared to women who received patient education plus usual care (RR 1.54, 95% CI 0.07 to 36.11; see Analysis 4.10) (Higashihara 2021).

4.10. Analysis.

Comparison 4: Moxibustion plus usual care versus patient education plus usual care, Outcome 10: Admission to NICU

5. Moxibustion plus acupuncture plus usual care versus usual care

Primary outcomes

5.1 Non‐cephalic presentation at birth

There is very little evidence about the effect of moxibustion plus acupuncture plus usual care compared to usual care on the chance of non‐cephalic presentation at birth (1 trial, 226 women; RR 0.73, 95% CI 0.57 to 0.94; see Analysis 5.1) (Neri 2004).

5.1. Analysis.

Comparison 5: Moxibustion plus acupuncture plus usual care versus usual care, Outcome 1: Non‐cephalic presentation at birth

5.2 Need for ECV

There is very little evidence about the effect of moxibustion plus acupuncture plus usual care compared to usual care on the need for ECV (1 trial, 14 women; RR 0.45, 95% CI 0.07 to 3.01; see Analysis 5.2) (Tsujiuchi 2017).

5.2. Analysis.

Comparison 5: Moxibustion plus acupuncture plus usual care versus usual care, Outcome 2: Need for ECV

Secondary outcomes

The studies that reported this comparison did not report on maternal satisfaction or perinatal morbidity or mortality.

5.3 Non‐cephalic presentation at 33 weeks

There is very little evidence about the effect of moxibustion plus acupuncture plus usual care compared to usual care on the chance of non‐cephalic presentation at 33 weeks' gestation (2 trials, 254 women; RR 0.73, 95% CI 0.57 to 0.93, I² = 0%; see Analysis 5.3) (Neri 2004; Tsujiuchi 2017).

5.3. Analysis.

Comparison 5: Moxibustion plus acupuncture plus usual care versus usual care, Outcome 3: Non‐cephalic presentation at the end of treatment

5.4 Caesarean section

There is very little evidence about whether moxibustion plus acupuncture plus usual care reduces the chance of caesarean section more than usual care alone (RR 0.80, 95% CI 0.65 to 0.99, I² = 0%; see Analysis 5.4) (Neri 2004; Tsujiuchi 2017).

5.4. Analysis.

Comparison 5: Moxibustion plus acupuncture plus usual care versus usual care, Outcome 4: Caesarean section

5.5 Preterm birth

One trial (14 women) reported no cases of preterm birth in either group (RR not estimable; see Analysis 5.5) (Tsujiuchi 2017).

5.5. Analysis.

Comparison 5: Moxibustion plus acupuncture plus usual care versus usual care, Outcome 5: Preterm birth

5.6 Pre‐eclampsia

One trial (14 women) provided very few data; hence, there is very limited information about the true effect of moxibustion plus acupuncture plus usual care on the chance of developing pre‐eclampsia compared to usual care alone (RR 5.00, 95% CI 0.24 to 104.15; see Analysis 5.6) (Tsujiuchi 2017).

5.6. Analysis.

Comparison 5: Moxibustion plus acupuncture plus usual care versus usual care, Outcome 6: Pre‐eclampsia

5.7 Adverse events

The Neri 2004 trial explored maternal heart rate and blood pressure, and fetal heart rate immediately after the intervention was applied. No fetal or maternal cardiovascular changes were detected, nor were any preterm uterine contractions reported.

6. Moxibustion plus acupuncture plus usual care versus acupuncture plus usual care

Primary outcomes

The single study that reported this comparison did not report on the other prespecified primary outcome, need for ECV.

6.1 Non‐cephalic presentation at birth

There is very little evidence about the effect of moxibustion plus acupuncture plus usual care compared to acupuncture plus usual care on the number of non‐cephalic presentations at birth (1 trial, 24 women; RR 0.54, 95% CI 0.27 to 1.06; see Analysis 6.1) (Neri 2007).

6.1. Analysis.

Comparison 6: Moxibustion plus acupuncture plus usual care versus acupuncture plus usual care, Outcome 1: Non‐cephalic presentation at birth

Secondary outcomes

The study that reported this comparison did not report on mode of birth, perinatal morbidity or mortality, maternal complications, or maternal satisfaction.

6.2 Adverse events

The authors of one trial (25 women) reported that the combination of moxibustion plus acupuncture plus usual care reduced fetal heart rate and fetal movements more than acupuncture plus usual care (Neri 2007). No data were reported according to group allocation that could be analysed.

7. Moxibustion plus acupuncture plus usual care versus moxibustion plus usual care

Primary outcomes

The single study that reported this comparison did not report on the other prespecified primary outcome, need for ECV.

7.1 Non‐cephalic presentation at birth

There is very little evidence about the effect of moxibustion plus acupuncture plus usual care versus moxibustion plus usual care on the number of non‐cephalic presentations at birth (1 trial, 29 women; RR 2.14, 95% CI 0.66 to 6.97; see Analysis 7.1) (Neri 2007).

7.1. Analysis.

Comparison 7: Moxibustion plus acupuncture plus usual care versus moxibustion plus usual care, Outcome 1: Non‐cephalic presentation at birth

Secondary outcomes

The study that reported this comparison did not report on mode of birth, perinatal morbidity or mortality, maternal complications, or maternal satisfaction.

7.2 Adverse events

The authors of one trial (25 women) found that the combination of moxibustion plus acupuncture plus usual care reduced fetal heart rate and fetal movements more than moxibustion plus usual care (Neri 2007). No data were reported according to group allocation that could be analysed.

8. Moxibustion plus acupuncture plus usual care versus laser placebo plus usual care

Primary outcomes

8.1 Non‐cephalic presentation at birth

There is very little evidence about the effect of moxibustion plus acupuncture plus usual care compared to laser placebo plus usual care on the number of non‐cephalic presentations at birth (1 trial, 328 women; RR 1.04, 95% CI 0.86 to 1.27; see Analysis 8.1) (Coulon 2014).

8.1. Analysis.

Comparison 8: Moxibustion plus acupuncture plus usual care versus laser placebo plus usual care, Outcome 1: Non‐cephalic presentation at birth

8.2 Need for ECV

There is very little evidence about the effect of moxibustion plus acupuncture plus usual care compared to laser placebo plus usual care on the need for ECV (1 trial, 328 women; RR 1.13, 95% CI 0.89 to 1.43; see Analysis 8.2) (Coulon 2014).

8.2. Analysis.

Comparison 8: Moxibustion plus acupuncture plus usual care versus laser placebo plus usual care, Outcome 2: Need for ECV

Secondary outcomes

The studies that reported this comparison did not report on maternal satisfaction.

8.3 Non‐cephalic presentation at 37 + 2 weeks' gestation

There is very little evidence about the effect of moxibustion plus acupuncture plus usual care compared to laser placebo plus usual care on the number of women with non‐cephalic presentation at the end of treatment, 37 + 2 weeks' gestation (1 trial, 328 women; RR 1.13, 95% CI 0.98 to 1.32; see Analysis 8.3) (Coulon 2014).

8.3. Analysis.

Comparison 8: Moxibustion plus acupuncture plus usual care versus laser placebo plus usual care, Outcome 3: Non‐cephalic presentation at 37 + 2 weeks gestation

8.4 Caesarean section

There is very little evidence about the effect of moxibustion plus acupuncture plus usual care compared to laser placebo plus usual care on the rate of caesarean section (1 trial, 328 women; RR 1.10, 95% CI 0.81 to 1.49; see Analysis 8.4) (Coulon 2014).

8.4. Analysis.

Comparison 8: Moxibustion plus acupuncture plus usual care versus laser placebo plus usual care, Outcome 4: Caesarean section

8.5 Premature rupture of membranes

There is very little evidence about the effect of moxibustion plus acupuncture plus usual care compared to laser placebo plus usual care on the chance of premature rupture of membranes (1 trial, 328 women; RR 0.94, 95% CI 0.50 to 1.77; see Analysis 8.5) (Coulon 2014).

8.5. Analysis.

Comparison 8: Moxibustion plus acupuncture plus usual care versus laser placebo plus usual care, Outcome 5: Premature rupture of membranes

8.6 Preterm birth

There is very little evidence about the effect of moxibustion plus acupuncture plus usual care compared to laser placebo plus usual care on the chance of preterm birth (1 trial, 328 women; RR 1.10, 95% CI 0.48 to 2.52; see Analysis 8.6) (Coulon 2014).

8.6. Analysis.

Comparison 8: Moxibustion plus acupuncture plus usual care versus laser placebo plus usual care, Outcome 6: Preterm birth

8.7 Intrauterine fetal death

There is very little evidence about the effect of moxibustion plus acupuncture plus usual care compared to laser placebo plus usual care on the chance of intrauterine fetal death (1 trial, 328 women; RR 3.00, 95% CI 0.12 to 73.11; see Analysis 8.7) (Coulon 2014).

8.7. Analysis.

Comparison 8: Moxibustion plus acupuncture plus usual care versus laser placebo plus usual care, Outcome 7: Intrauterine fetal death

8.8 Adverse events

In one trial (328 women) (Coulon 2014), more adverse events were reported among women who received moxibustion, acupuncture, and usual care than in those who received laser placebo plus usual care (RR 29.00, 95% CI 1.74 to 482.13; see Analysis 8.8). Adverse events among women who received moxibustion, acupuncture, and usual care were seven cases of painful burns, four cases of blisters or burns, two cases of nausea and vomiting, and one case of vasovagal syncope. 

8.8. Analysis.

Comparison 8: Moxibustion plus acupuncture plus usual care versus laser placebo plus usual care, Outcome 8: Adverse events

9. Moxibustion plus postural therapy plus usual care versus postural therapy plus usual care

Primary outcomes

The single study that reported this comparison did not report on the other prespecified primary outcome, need for ECV, or on any of the prespecified secondary outcomes.

9.1 Non‐cephalic presentation at birth

There is very little evidence about the effect of moxibustion plus postural therapy plus usual care compared to postural therapy plus usual care on the chance of a non‐cephalic presentation at birth (1 trial, 206 women; RR 0.50, 95% CI 0.27 to 0.92; see Analysis 9.1) (Yang 2006).

9.1. Analysis.

Comparison 9: Moxibustion plus postural therapy plus usual care versus postural therapy plus usual care, Outcome 1: Non‐cephalic presentation at birth

Discussion

Summary of main results

This updated review includes 13 studies and 2181 women. Several new high‐quality trials have been published since the previous version of the review (Coyle 2012), and the conclusions have changed. Most studies included in this review compared moxibustion plus usual care with usual care alone.

Moxibustion plus usual care versus usual care

Moderate‐certainty evidence showed that the combination of moxibustion plus usual care probably reduces the number of non‐cephalic presentations at birth compared with usual care alone, but probably has little to no effect on the likelihood of caesarean section (see Table 1). Birth by caesarean section can occur for various reasons that are not related to fetal presentation; however, none of the included studies reported the reason for caesarean section. Moderate‐certainty evidence shows that moxibustion plus usual care probably reduces the use of oxytocin more than usual care alone. The evidence is very uncertain about the effect of moxibustion plus usual care on the need for ECV. Furthermore, the evidence is very uncertain about the effect of moxibustion plus usual care on the chance of premature rupture of membranes or cord blood less than 7.1 because there were very few data for these outcomes; therefore, there is very little idea of the effect on this outcome. We are very uncertain about the likelihood of adverse events with moxibustion plus usual care (very low–certainty evidence).

Moxibustion plus usual care versus sham moxibustion plus usual care

Moderate‐certainty evidence shows that moxibustion plus usual care probably reduces the chance of non‐cephalic presentation at birth more than sham moxibustion plus usual care, but probably results in little to no effect on the likelihood of caesarean section (see Table 2). None of the included studies that compared moxibustion plus usual care with sham moxibustion plus usual care reported on the need for ECV, premature rupture of membranes, use of oxytocin, cord blood pH less than 7.1, or adverse events in a way that permitted reanalysis. 

Other comparisons

There is very little evidence from other comparisons about the effect of moxibustion plus usual care on the chance of non‐cephalic presentation at the end of treatment. Furthermore, meta‐analysis of two studies provides very little evidence about the effect of the combination of moxibustion, acupuncture, and usual care on non‐cephalic presentation at birth and at the end of treatment and the chance of caesarean birth. 

Clinical and statistical heterogeneity

We found considerable diversity in the comparisons, which reflects the various ways in which moxibustion is combined with other techniques in clinical practice. Many analyses were based on single studies, and for many comparisons, there is insufficient evidence to draw any conclusions for clinical practice. 

Most meta‐analyses exhibited levels of statistical heterogeneity that might not be important. However, clinical heterogeneity (e.g. parity; timing of the intervention; and variations in usual care, particularly relating to ECV) may exist among the included studies. 

Safety

Many studies that reported adverse events did not report events according to group allocation, nor were assessments made of causality. The most frequently reported adverse events were increased fetal movements, uterine contractions, nausea, headache, and burns from holding the moxibustion stick too close to the skin. In one study for which adverse events could be analysed, 27 adverse events occurred in the treatment group (65 women), and none occurred in the control group (57 women).

Overall completeness and applicability of evidence

All studies reported at least one of the two primary outcomes (presentation at birth and need for ECV); 12 of the 13 studies reported presentation at birth, and six of the 13 studies reported the need for ECV. However, only one of the included studies reported on all prespecified primary and secondary outcomes (Cardini 1998).

This review included studies published in languages other than English, including one that was identified through clinical trial registers but not found in searches of the biomedical databases. It is possible that studies published in languages other than those that we were able to translate may exist, and the inclusion of such studies may alter the findings. The included studies were conducted in Australia as well as East Asian and European countries, with more studies from European countries than in the previous version of the review (Coyle 2012). We identified no eligible studies that were conducted in North America, South America, or Africa. 

All studies used moxibustion as an adjunct to usual care; however, usual care appeared to vary among the included studies. For example, some studies offered ECV to all women with non‐cephalic presentation at the end of treatment, while others did not. We were unable to compare what constituted usual care among the included studies, but differences may exist that could alter the meta‐analysis results. Furthermore, some studies used postural techniques in the treatment or control groups. The most recent Cochrane Review on postural management for breech presentation found that there was insufficient evidence for postural management of breech presentation (Hofmeyr 2012), but most of the studies included in our review that used postural techniques were conducted prior to the Hofmeyr 2012 review. Whether these techniques were part of usual care in the included studies was unclear.

The frequency of moxibustion treatment in the included studies was generally consistent (daily or twice‐daily treatment) and reflects the approach used in clinical practice. We noted differences in the interventions (e.g. moxibustion alone or combined with acupuncture or postural techniques) that also reflect clinical practice. In studies that assessed maternal satisfaction with treatment, women found moxibustion to be acceptable; however, there is limited evidence on maternal satisfaction. Furthermore, some women withdrew from trials due to adverse events, so women should be informed about the potential risks of moxibustion for turning breech babies. The findings of the main comparisons are likely to be easily translated into clinical practice.

Certainty of the evidence

We assessed most included studies as at low risk of bias for the domains random sequence generation and allocation concealment. Blinding of participants and personnel (particularly the therapist) is challenging in trials of manual therapies like moxibustion (Zhao 2014). Two studies attempted to blind participants to group allocation: one used sham moxibustion to the acupuncture point SP1 Yinbai (Vas 2013), and one used laser placebo to BL67 (Coulon 2014). None of the studies reported having assessed the success of these techniques in blinding participants.

The included studies mostly reported objective outcomes, such as presentation at birth, meaning that a lack of blinding was unlikely to have influenced outcomes. However, some studies reported subjective outcomes such as maternal satisfaction, fetal movements, and adverse events. It is possible that these outcomes may have been influenced by the lack of blinding (i.e. judged unclear or high risk, depending on the information reported).

As the included trials were generally conducted in a hospital setting where birth outcomes were available through hospital or maternal health records, few studies were assessed as high risk of bias for incomplete outcome data. We found that many studies conducted prior to 2015 did not have published protocols or trial registrations; this finding is not unique. Furthermore, for trials that were registered in international clinical trial registers, some differences were noted in the outcomes specified in the trial registration and those described in the trial report. We identified no other obvious sources of bias for six of the 13 included studies. 

For the clinically important comparisons (moxibustion plus usual care versus usual care; moxibustion plus usual care versus sham moxibustion plus usual care), the certainty of the evidence ranged from moderate to very low. We downgraded the certainty of the evidence for imprecision when results were from a single study, where the optimal information size was not met, or where there were wide confidence intervals, and for inconsistency when statistical heterogeneity was detected. Many analyses were based on single small studies and were underpowered.

Potential biases in the review process

A potential source of bias in this review may originate from the search strategy. The search strategy utilised is more likely to detect English language publications. Efforts have been made to assess publications in languages other than English; indeed, several trials published in other languages were included in this update. However, we did not search any Chinese, Korean, or Japanese language biomedical databases (i.e. those databases more likely to include relevant studies). It is likely that relevant publications published in other languages have not been identified.

We translated the methods and results sections and any additional information relating to competing interests and funding source for articles published in Spanish, French, and Japanese. We were able to have the translations verified by people fluent in these languages. Our work in assessing studies published in other languages was restricted to those three languages. Studies published in other languages may exist, and their inclusion in meta‐analysis could alter the conclusions of this review. 

The risk of bias and trustworthiness assessments for each study were undertaken independently by two review authors, with any disagreements discussed until a consensus was reached. Such judgements are inherently subjective, despite independent assessment. Furthermore, assessments of the certainty of the evidence (GRADE) require a degree of interpretation. It is possible that other researchers may reach different assessments of potential biases, trustworthiness, and certainty of the evidence.

Caroline Smith is an author on the Do 2011 trial. Meaghan Coyle and Brian Peat assessed the eligibility of that trial and conducted data extraction (including risk of bias assessment), and Meaghan Coyle conducted the data analysis and interpretation. 

Agreements and disagreements with other studies or reviews

The findings from this updated review support the conclusions from other published reviews. As comparatively few RCTs have examined whether moxibustion is effective for changing breech presentation, the published systematic reviews generally rely on the same body of evidence. Earlier systematic reviews concluded that moxibustion was beneficial in changing non‐cephalic presentation at delivery (Li 2009; Vas 2009; Zhang 2013), although the authors acknowledged that considerable heterogeneity was detected.

A more recent systematic review included many of the studies included in this review as well as some that were excluded (because of concerns about trustworthiness and lack of clarity about the number of women with breech presentation) (Liao 2021). Liao and colleagues used the revised Cochrane risk of bias tool (RoB 2), which produced different assessments of potential sources of bias compared to this review. Similar to our review, Liao and colleagues found that moxibustion was beneficial in changing breech presentation at delivery, although we found lower levels of statistical heterogeneity. The benefits of moxibustion in changing breech presentation at delivery were seen in subgroup analyses of studies that included Asian women and studies that included non‐Asian women, which may suggest that ethnicity is not a factor in the clinical effects of moxibustion. 

The current review includes studies published in languages other than English that have not been included in other published systematic reviews; thus, it provides the most up‐to‐date and comprehensive assessment of trustworthy evidence. 

Authors' conclusions

Implications for practice.

Moderate‐certainty evidence shows that the combination of moxibustion and usual care probably reduces the number of non‐cephalic presentations at birth than usual care alone, but probably does not reduce the chance of caesarean section. Birth by caesarean sections can occur for reasons unrelated to fetal presentation, and none of the included studies reported the reason for caesarean section. The evidence is very uncertain about the effects of moxibustion plus usual care on the need for external cephalic version because the confidence intervals are compatible with a wide range of effects that encompass both appreciable benefit and moderate harm. Similarly, the evidence is very uncertain for the effects of moxibustion plus usual care on the chance of premature rupture of membranes or cord blood pH less than 7.1 because there are very few data. Moderate‐certainty evidence from one study shows that moxibustion plus usual care probably reduces the use of oxytocin before or during labour. The updated analysis of newly included studies provides a clearer picture of the role of moxibustion for women with breech babies.

We are less certain about the effectiveness of combining moxibustion with other techniques, such as acupuncture and postural techniques, as many analyses were based on one or two studies. Further evidence is needed on the effectiveness of these combinations, which are common in clinical practice. However, given that moxibustion has been shown to be effective in reducing the chance of non‐cephalic presentation at birth, the clinical need for combining interventions remains unclear. Women can easily self‐administer moxibustion at home, and it seems that combining moxibustion with other interventions (in the absence of evidence of effectiveness) may unnecessarily increase the burden and cost of treatment. Clinicians should use their professional judgement for whether additional treatments would benefit each patient.

One study that assessed the acceptability of moxibustion reported that the intervention was acceptable to women. More evidence is needed on the acceptability of moxibustion. 

Adverse events were reported with moxibustion, but many studies did not report the details of adverse events in a way that could be reanalysed. An important consideration is the potential respiratory irritation that may occur with moxibustion smoke. Practitioners and women self‐administering moxibustion should be properly trained and ensure appropriate ventilation is available when performing moxibustion to reduce the potential respiratory adverse events from the procedure. Smokeless moxibustion sticks may reduce this risk. In clinical practice, patients could be informed about the potential for burns, increased fetal movements, uterine contractions, nausea and headaches, which seem to be the most common adverse events.

Implications for research.

Since the previous version of this review, several high‐quality trials have been published. The current evidence shows that the combination of moxibustion and usual care probably reduces the chance of non‐cephalic presentation at birth. While this finding is based on moderate‐certainty evidence, it is possible that future randomised controlled trials will change the direction and magnitude of the treatment effect.

One study included in this review used sham moxibustion to a theoretically irrelevant acupuncture point (SP1) as a control. This approach has been common in acupuncture and moxibustion clinical trials, but it remains controversial. The study that used this control was published in 2013, a time when stimulating theoretically irrelevant acupuncture points was an acceptable sham procedure. Such an approach is used less frequently in clinical trials of acupuncture‐related therapies since the introduction of standardised reporting criteria (MacPherson 2010), and the growing understanding that such procedures are not inert controls (Zhang 2015). The assumption that SP1 is irrelevant for women with breech babies needs to be tested.

Future research may consider reporting the outcome 'failure to achieve spontaneous cephalic birth' as the primary outcome, since this is a useful way to study the desired outcome of any intervention for breech presentation. Furthermore, studies must report adverse event data according to group allocation to allow accurate assessment of the potential risks of moxibustion. Detailed reporting of the nature of the event, severity, and assessment of causality will allow practitioners and patients to make informed choices about their care. In addition, further research is needed on the acceptability of moxibustion for changing breech presentation to cephalic.

What's new

Date	Event	Description
9 May 2023	New citation required and conclusions have changed	New trials have been published since the previous version of this review. In this update, we added seven new trials, excluded one previously included trial due to concerns about trustworthiness, and included one trial previously awaiting classification. Based on the results of analyses of the new and previously included studies, the conclusions of this review have changed.
9 May 2023	New search has been performed	Search updated to November 3, 2021.

History

Protocol first published: Issue 4, 2002
Review first published: Issue 2, 2005

Date	Event	Description
21 February 2012	New citation required and conclusions have changed	There is now some evidence to support the use of moxibustion (in combination with other methods) to change a breech presentation (as well as to improve other outcomes relevant to this review). However, the results should be interpreted with caution because of statistical and clinical heterogeneity.
31 August 2011	New search has been performed	Search updated. We identified 11 new trials. We included six new trials (Cardini 2005; Chen 2004; Guittier 2009; Lin 2002; Neri 2007; Yang 2006), and excluded two trials (Huang 1990; Manyande 2009). Two trials are ongoing (Do 2011; Vas 2008), and one is awaiting classification (Millereau 2009). We excluded one trial that was previously awaiting classification (Wu 1994).
14 April 2008	Amended	Converted to new review format

Appendices

Appendix 1. Search methods for ICTRP and ClinicalTrials.gov

Each line was run separately

ClinicalTrials.gov

Advanced search

cephalic version | Interventional Studies | moxibustion

Interventional Studies | Breech Presentation | moxibustion

ICTRP

moxibustion AND version

moxibustion AND breech

Appendix 2. Additional search strategy

Institutional access to CINAHL and AMED was via the same platform; therefore, the same search strategy was used for both databases. The search strategy was updated to accomodate changes in the database settings.

MEDLINE search strategy:

1.  "breech presentation"[MeSH Terms] OR ("breech"[All Fields] AND "presentation"[All Fields]) OR "breech presentation"[All Fields] OR ("labour presentation"[All Fields] OR "labor presentation"[MeSH Terms] OR ("labor"[All Fields] AND "presentation"[All Fields]) OR "labor presentation"[All Fields])

2. "acupunctural"[All Fields] OR "acupuncture"[MeSH Terms] OR "acupuncture"[All Fields] OR "acupuncture therapy"[MeSH Terms] OR ("acupuncture"[All Fields] AND "therapy"[All Fields]) OR "acupuncture therapy"[All Fields] OR "acupuncture s"[All Fields] OR "acupunctured"[All Fields] OR "acupunctures"[All Fields] OR "acupuncturing"[All Fields]

3. "moxibustion"[MeSH Terms] OR "moxibustion"[All Fields] OR "moxa"[All Fields]

4. "artemisia"[MeSH Terms] OR "artemisia"[All Fields] OR ("artemisia"[All Fields] AND "vulgaris"[All Fields]) OR "artemisia vulgaris"[All Fields] OR ("artemisia"[MeSH Terms] OR "artemisia"[All Fields] OR "artemisias"[All Fields] OR "artemisiae"[All Fields])

5. "version"[All Fields] OR "version s"[All Fields] OR "versions"[All Fields] OR ("version, fetal"[MeSH Terms] OR ("version"[All Fields] AND "fetal"[All Fields]) OR "fetal version"[All Fields] OR ("version"[All Fields] AND "fetal"[All Fields]) OR "version fetal"[All Fields])

6. 1 or 5

7. 2 or 3 or 4

8. 6 and 7

Embase search strategy:

1. breech

2. ‘breech presentation’/exp OR ‘breech presentation’

3. version

4. 'moxibustion'/exp OR 'moxibustion' OR 'moxa'

5. 'acupuncture'/exp OR 'acupuncture'

6. 'artemisia'/exp OR ('artemisia' AND 'vulgaris')

7. 4 OR 5 OR 6

8. 1 OR 2 OR 3

9. 7 AND 8

CINAHL and AMED search strategy:

1. acupuncture [All Text] OR moxibustion [All Text] OR moxa [All Text] OR artemisia vulgaris [All Text]

2. breech [All Text] OR version [All Text]

3. 1 AND 2

We no longer have online access to search MIDIRS directly. Instead, a search was conducted of the RMIT University Library catalogue using the keywords 'moxibustion', 'moxa', 'acupuncture' and 'artemisia vulgaris', each of which was combined with the ISSN of MIDIRS. No results were retrieved for any combination.

Appendix 3. Data collection and analysis methods used in previous versions of this review

We evaluated the appropriateness of trials for inclusion. Where there was uncertainty about inclusion of the study, the full text was retrieved. We contacted the original author for further information if necessary. If there was disagreement between authors about the studies to be included that could not be resolved by discussion, we sought assistance from the third reviewer. Reasons for excluding trials were stated.

Following an assessment for inclusion, we assessed the methodology of the trial. We extracted the data onto hard copy data sheets. Meaghan Coyle and Caroline Smith extracted the data and assessed quality.

We assessed included trials according to the following five main criteria:

• adequate concealment of treatment allocation (e.g. opaque/sealed numbered envelopes, sequentially numbered envelopes, central randomisation)

• adequate method of allocation to treatment (e.g. by computer randomisation, random number tables, lot drawing, coin tossing, shuffling cards, throwing dice)

• adequate documentation of how exclusions were handled after treatment allocation to facilitate intention‐to‐treat analysis

• adequate blinding of outcome assessment (e.g. outcome assessor and data analyst blinded)

• losses to follow up (trials with losses greater than 25% will be excluded from the analysis).

We assigned quality scores to each trial using the criteria described in section six of the Cochrane Reviewers' Handbook (Clarke 2003).
(A) was used to indicate a trial which has a high level of quality in which all the criteria are met
(B) was used to indicate that one or more criteria were partially met or if it is unclear if all the criteria were met and
(C) was used if one or more criteria were not met.

In addition, we assigned quality scores to each trial for completeness of follow up and blinding of outcome assessment as follows:
For completeness of follow up:
(A) less than 3% of participants excluded
(B) 3% to 9.9% of participants excluded
(C) 10% to 19.9% of participants excluded
(D) 20% or more of participants excluded
(E) unclear.

For blinding of assessment outcome:
(A) double blind, neither investigator nor participant knew or were likely to guess the allocated group
(B) single blind, either the investigator or the participant knew the allocation, or the trial is described as double blind but side‐effects of one or other treatment meant that it is likely that for a significant proportion (7% to 20%) of participants could correctly identify their treatment allocation
(C) no blinding, both investigator and participant knew (or were likely to guess) the allocated treatment
(D) unclear.

We entered data directly from the published reports into the Review Manager version 4.2 software (RevMan 2002) for the original review (Coyle 2005) and Review Manager version 5.1 for the version published in 2012 (Coyle 2012), with double data entry performed by the co‐author (Caroline Smith). Where data were not presented in a suitable format for data entry, or if data were missing, we sought additional information from the trialists by personal communication in the form of a letter or telephone call.

Due to the nature of the interventions, double blinding of assessments was not possible; therefore, we considered for inclusion studies where there was no double blinding of assessments. An intention‐to‐treat analysis would have enhanced the quality of the trial but was not present in all retrieved trials. We included only studies in which data were reported by allocated group. Where moxibustion was used in combination with another intervention, a separate comparison was made due to incompatibility of the interventions. We performed statistical analysis using Review Manager software (RevMan 2002; RevMan 2011). For dichotomous data, we calculated relative risks and 95% confidence intervals calculated. The I² statistic was used to detect statistical heterogeneity, with a score of greater than 50% indicating significant heterogeneity.

Data and analyses

Comparison 1. Moxibustion plus usual care versus usual care.

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
1.1 Non‐cephalic presentation at birth	7	1152	Risk Ratio (M‐H, Fixed, 95% CI)	0.87 [0.78, 0.99]
1.2 Need for ECV	4	692	Risk Ratio (M‐H, Random, 95% CI)	0.62 [0.32, 1.21]
1.3 Non‐cephalic presentation at the end of treatment	4	853	Risk Ratio (M‐H, Random, 95% CI)	0.80 [0.61, 1.05]
1.4 Caesarean section	6	1030	Risk Ratio (M‐H, Fixed, 95% CI)	0.94 [0.83, 1.05]
1.5 Premature rupture of membranes	3	402	Risk Ratio (M‐H, Random, 95% CI)	1.31 [0.17, 10.21]
1.6 Preterm birth	2	290	Risk Ratio (M‐H, Fixed, 95% CI)	1.17 [0.39, 3.54]
1.7 Use of oxytocin	1	161	Risk Ratio (M‐H, Fixed, 95% CI)	0.28 [0.13, 0.60]
1.8 Apgar score < 7 at 5 minutes	5	962	Risk Ratio (M‐H, Fixed, 95% CI)	0.33 [0.09, 1.22]
1.9 Intrauterine fetal death	1	260	Risk Ratio (M‐H, Fixed, 95% CI)	0.33 [0.01, 8.11]
1.10 Placental abruption	1	122	Risk Ratio (M‐H, Fixed, 95% CI)	0.29 [0.01, 7.05]
1.11 Duration of hospital stay (days)	1	212	Mean Difference (IV, Fixed, 95% CI)	‐0.10 [‐0.55, 0.35]
1.12 Cord blood pH less than 7.1	1	212	Risk Ratio (M‐H, Fixed, 95% CI)	3.00 [0.32, 28.38]
1.13 NICU admission	1	20	Risk Ratio (M‐H, Fixed, 95% CI)	5.00 [0.27, 92.62]
1.14 Adverse events	1	122	Risk Ratio (M‐H, Fixed, 95% CI)	48.33 [3.01, 774.86]

Comparison 2. Moxibustion plus usual care versus sham moxibustion plus usual care.

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
2.1 Non‐cephalic presentation at birth	1	272	Risk Ratio (M‐H, Fixed, 95% CI)	0.74 [0.58, 0.95]
2.2 Non‐cephalic presentation at the end of treatment	1	272	Risk Ratio (M‐H, Fixed, 95% CI)	0.79 [0.67, 0.92]
2.3 Caesarean section	1	272	Risk Ratio (M‐H, Fixed, 95% CI)	0.84 [0.68, 1.04]
2.4 Preterm birth	1	272	Risk Ratio (M‐H, Fixed, 95% CI)	1.00 [0.26, 3.92]
2.5 Apgar score < 7 at 5 minutes	1	0	Risk Ratio (M‐H, Fixed, 95% CI)	Not estimable

Comparison 3. Moxibustion plus usual care versus acupuncture plus usual care.

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
3.1 Non‐cephalic presentation at birth	1	25	Risk Ratio (M‐H, Fixed, 95% CI)	0.25 [0.09, 0.72]

Comparison 4. Moxibustion plus usual care versus patient education plus usual care.

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
4.1 Non‐cephalic presentation at birth	1	60	Risk Ratio (M‐H, Fixed, 95% CI)	0.82 [0.48, 1.38]
4.2 Need for ECV	1	60	Risk Ratio (M‐H, Fixed, 95% CI)	0.30 [0.08, 1.13]
4.3 Non‐cephalic presentation at the end of treatment	1	60	Risk Ratio (M‐H, Fixed, 95% CI)	0.63 [0.42, 0.96]
4.4 Caesarean section	1	60	Risk Ratio (M‐H, Fixed, 95% CI)	0.77 [0.49, 1.20]
4.5 Premature rupture of membranes	1	0	Risk Ratio (M‐H, Fixed, 95% CI)	Not estimable
4.6 Preterm birth	1	0	Risk Ratio (M‐H, Fixed, 95% CI)	Not estimable
4.7 Apgar score < 7 at 5 minutes	1	60	Risk Ratio (M‐H, Fixed, 95% CI)	1.54 [0.07, 36.11]
4.8 Cord blood pH less than 7	1	0	Risk Ratio (M‐H, Fixed, 95% CI)	Not estimable
4.9 Intrauterine fetal death	1	0	Risk Ratio (M‐H, Fixed, 95% CI)	Not estimable
4.10 Admission to NICU	1	60	Risk Ratio (M‐H, Fixed, 95% CI)	1.54 [0.07, 36.11]

Comparison 5. Moxibustion plus acupuncture plus usual care versus usual care.

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
5.1 Non‐cephalic presentation at birth	1	226	Risk Ratio (M‐H, Fixed, 95% CI)	0.73 [0.57, 0.94]
5.2 Need for ECV	1	14	Risk Ratio (M‐H, Fixed, 95% CI)	0.45 [0.07, 3.01]
5.3 Non‐cephalic presentation at the end of treatment	2	254	Risk Ratio (M‐H, Fixed, 95% CI)	0.73 [0.57, 0.93]
5.4 Caesarean section	2	240	Risk Ratio (M‐H, Fixed, 95% CI)	0.80 [0.65, 0.99]
5.5 Preterm birth	1	0	Risk Ratio (M‐H, Fixed, 95% CI)	Not estimable
5.6 Pre‐eclampsia	1	14	Risk Ratio (M‐H, Fixed, 95% CI)	5.00 [0.24, 104.15]

Comparison 6. Moxibustion plus acupuncture plus usual care versus acupuncture plus usual care.

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
6.1 Non‐cephalic presentation at birth	1	24	Risk Ratio (M‐H, Fixed, 95% CI)	0.54 [0.27, 1.06]

Comparison 7. Moxibustion plus acupuncture plus usual care versus moxibustion plus usual care.

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
7.1 Non‐cephalic presentation at birth	1	29	Risk Ratio (M‐H, Fixed, 95% CI)	2.14 [0.66, 6.97]

Comparison 8. Moxibustion plus acupuncture plus usual care versus laser placebo plus usual care.

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
8.1 Non‐cephalic presentation at birth	1	328	Risk Ratio (M‐H, Fixed, 95% CI)	1.04 [0.86, 1.27]
8.2 Need for ECV	1	328	Risk Ratio (M‐H, Fixed, 95% CI)	1.13 [0.89, 1.43]
8.3 Non‐cephalic presentation at 37 + 2 weeks gestation	1	328	Risk Ratio (M‐H, Fixed, 95% CI)	1.13 [0.98, 1.32]
8.4 Caesarean section	1	328	Risk Ratio (M‐H, Fixed, 95% CI)	1.10 [0.81, 1.49]
8.5 Premature rupture of membranes	1	328	Risk Ratio (M‐H, Fixed, 95% CI)	0.94 [0.50, 1.77]
8.6 Preterm birth	1	328	Risk Ratio (M‐H, Fixed, 95% CI)	1.10 [0.48, 2.52]
8.7 Intrauterine fetal death	1	328	Risk Ratio (M‐H, Fixed, 95% CI)	3.00 [0.12, 73.11]
8.8 Adverse events	1	328	Risk Ratio (M‐H, Fixed, 95% CI)	29.00 [1.74, 482.13]

Comparison 9. Moxibustion plus postural therapy plus usual care versus postural therapy plus usual care.

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
9.1 Non‐cephalic presentation at birth	1	206	Risk Ratio (M‐H, Random, 95% CI)	0.50 [0.27, 0.92]

Characteristics of studies

Characteristics of included studies [ordered by study ID]

Bue 2016.

Study characteristics
Methods	Parallel‐group RCT of moxibustion plus usual care versus no moxibustion plus usual care
Participants	200 women were recruited to the study, which was conducted in obstetrics departments in Holstebro and Herning, Denmark. Women were randomised to receive either moxibustion plus usual care (n = 102) or usual care alone (no moxibustion; n = 98). Inclusion criteria were women with a singleton pregnancy with fetal breech presentation verified by ultrasound at 32 weeks' gestation. Exclusion criteria were women with vaginal bleeding in the second and third trimester, placental insufficiency, cervical shortening, PROM, pre‐eclampsia, high blood pressure, pelvic insufficiency, low placental position, placenta praevia, uterine malformations, previous corrective uterine surgery, and known fetal morbidity.
Interventions	Intervention group: moxibustion plus usual care. Moxibustion was applied to BL67 for 15 to 20 minutes each day, preferably in the evening. The intervention was intended to be applied daily for 2 weeks, and the partner was instructed on how to apply the moxibustion. Control group: usual care (no moxibustion)
Outcomes	Fetal presentation after 3 weeks, fetal presentation at birth, caesarean section, safety
Notes	Study dates: October 2003 to January 2010 Sample size calculation: performed. There was no loss to follow‐up. Funding sources: not specified Authors' declarations of interest: declared, no competing interests Stratification by parity. ITT analysis for all women randomised, PP analysis for all women who performed 3 days of moxibustion. Women were offered ECV "according to current guidelines". Oblique presentation was considered cephalic.
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Quote: 'The randomisation was performed from a list of random numbers from the Geigy Scientific Tables, with even and odd numbers indicating moxibustion or not'.
Allocation concealment (selection bias)	Low risk	Quote: 'randomisation result was pre‐packed in sealed, non‐transparent envelope'.
Blinding of participants and personnel (performance bias)	Low risk	Neither participants nor personnel were blinded. The lack of blinding of personnel was considered unlikely to have influenced outcomes, including those that required clinical decision‐making.
Blinding of outcome assessment (detection bias)	Low risk	The study was not blinded, but the lack of blinding was considered unlikely to influence outcomes, including those that required clinical decision‐making.
Incomplete outcome data (attrition bias) All outcomes	Low risk	Some loss to follow up, but both ITT and PP analyses were conducted.
Selective reporting (reporting bias)	Unclear risk	The study reported all outcomes specified in the trial registration; however, the trial was registered retrospectively in 2014, after recruitment had ceased.
Other bias	Unclear risk	Baseline characteristics were similar between the groups. There was an inconsistency in the number of vaginal births and caesarean sections. The number of women in each group differed by four, but it is not clear whether this is due to change.

Cardini 1998.

Study characteristics
Methods	Parallel‐group RCT of moxibustion plus usual care versus no moxibustion plus usual care
Participants	260 women were recruited to the trial in Jiangxi province, China. Women were randomised to receive either moxibustion plus usual care (n = 130) or usual care alone (no moxibustion; n = 130). Inclusion criteria were primigravidae, at 33 weeks' gestation, breech presentation diagnosed by ultrasound, normal fetal biometry. Exclusion criteria were pelvic defects, previous uterine surgery, uterine malformation, fibromyoma greater than 4 cm diameter, fetal malformation, twin gestation, tocolytic therapy during pregnancy, risk of premature birth, pathological pregnancy that contraindicated inclusion in the trial, e.g. intrauterine growth retardation, gestosis, serious infections, placenta praevia, polyhydramnios, oligohydramnios.
Interventions	All women had breech presentation confirmed by ultrasound in the 33rd week of gestation, within 24 hours of randomisation. All women were asked to record fetal movements once daily for 7 days, during a 1‐hour period.  Intervention group: moxibustion plus usual care. The first 87 women randomised to the treatment group were instructed to administer moxibustion to BL67 bilaterally once daily for 30 minutes (15 minutes each side) for 7 days, while the subsequent 43 women were instructed to administer moxibustion twice daily for 30 minutes (15 minutes per side). Participants were given an instruction session on how to administer the moxibustion. After 7 days, fetal position was checked via fetal heartbeats and abdominal palpation, with ultrasound being used only when the other techniques were uncertain. If version had not occurred, women were advised to continue another week of treatment providing the woman agreed and there were no adverse effects. Women with breech presentation at 35 weeks could undergo ECV. Control group: usual care (no moxibustion)
Outcomes	The primary outcomes were the number of cephalic presentations at the 35th week and at birth, and fetal motor activity. Secondary outcomes included ECV, compliance with treatment, observation of possible adverse effects in the intervention group and adverse events in both groups, number of cephalic versions with 2 different dosages of moxibustion, number and causes of caesarean deliveries, spontaneous and induced deliveries, and Apgar score at 5 minutes.
Notes	Study dates: April 1995 to August 1996 Sample size calculation: performed Funding sources: funded by the Centro di Orientamento Educativo, Milan, Italy, and the Commission des Communauts Europtennes, Brussels, Belgium Authors' declarations of interest: not reported
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Computer‐generated.
Allocation concealment (selection bias)	Unclear risk	Sealed, numbered envelopes.
Blinding of participants and personnel (performance bias)	Low risk	Neither participants nor personnel were blinded. The lack of blinding of personnel was considered unlikely to have influenced outcomes, including those that required clinical decision‐making.
Blinding of outcome assessment (detection bias)	Low risk	The study was not blinded, but the lack of blinding was considered unlikely to influence outcomes, including those that required clinical decision‐making.
Incomplete outcome data (attrition bias) All outcomes	Low risk	There was complete follow‐up of all women randomised.
Selective reporting (reporting bias)	Unclear risk	No trial protocol or trial registration available.
Other bias	Low risk	No imbalances at randomisation. The study appears free of other sources of bias.

Cardini 2005.

Study characteristics
Methods	Single‐blind RCT of moxibustion plus usual care versus no moxibustion plus usual care
Participants	123 women were recruited to the trial through 6 hospital obstetric departments in northern and central Italy. Women were randomised to receive either moxibustion plus usual care (n = 65) or usual care alone (no moxibustion; n = 58). Inclusion criteria were non‐Chinese, nulliparous women with a singleton pregnancy at 32 to 33 weeks + 3 days, in good health with normal fetal biometry, and breech presentation confirmed by ultrasound within the previous 24 hours. Exclusion criteria were refusal of randomisation or treatment, defective pelvis, previous uterine surgery, recognised fetal/uterine malformation or fibroid greater than 4 cm in diameter, twin pregnancy, current or previous tocolytic therapy, and any other pregnancy complications.
Interventions	All women were instructed to record active fetal movements for 2 x 1‐hour periods per day for 2 weeks. Presentation was assessed by ultrasound 2 weeks after randomisation. Women with persistent breech were offered ECV after the 37th week if they were at a participating centre where ECV was routinely performed.  Intervention group: moxibustion plus usual care. Women allocated to the treatment group were trained in moxibustion treatment. Moxibustion was administered to BL67 bilaterally for 30 minutes (15 minutes per side), daily for 7 days. Presentation was assessed by ultrasound after 7 days. Women with persistent breech presentation were instructed to continue moxibustion for an additional 7 days. Control group: usual care (no moxibustion)
Outcomes	The primary outcome was the proportion of cephalic presentations in the 35th week. Secondary outcomes included: number of cephalic presentations at birth, fetal motor count, compliance with treatment, and adverse events by allocated group.
Notes	Study dates: March 2001 to February 2003 Sample size calculation: performed Funding sources: funded by the University of Turin. East‐West Italia (Rome) supplied materials.  Authors' declarations of interest: declared, no competing interests Data for 14 participants who withdrew due to adverse events were included in adverse event analysis.
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Randomisation by computer‐generated random numbers.
Allocation concealment (selection bias)	Low risk	Sealed opaque envelopes containing group allocation.
Blinding of participants and personnel (performance bias)	Low risk	Neither participants nor personnel were blinded. The lack of blinding of personnel was considered unlikely to have influenced outcomes, including those that required clinical decision‐making.
Blinding of outcome assessment (detection bias)	Low risk	The study was not blinded, but the lack of blinding was considered unlikely to influence outcomes, including those that required clinical decision‐making.
Incomplete outcome data (attrition bias) All outcomes	High risk	1 woman in the control group was lost to follow up, and was considered a positive result (cephalic version) in the intention‐to‐treat analysis. 14 women in the intervention group discontinued treatment, due to the treatment being unpleasant or associated with physical disturbances (e.g. unpleasant odour, with or without nausea and throat problems, abdominal pain with contractions).
Selective reporting (reporting bias)	Unclear risk	No trial protocol or trial registration available.
Other bias	High risk	The trial was interrupted when interim analysis revealed poor compliance and a high number of treatment interruptions.

Coulon 2014.

Study characteristics
Methods	Parallel‐group RCT of moxibustion plus acupuncture plus usual care versus laser placebo plus usual care
Participants	328 women were recruited from the obstetric department of Hopital Jeanne de Flandre, Universite Lille 2, France. Women were randomised to receive either moxibustion plus acupuncture or usual care (n = 164) or laser placebo plus usual care (n = 164). Inclusion criteria were women aged at least 18 years, able to speak and understand French, and provide consent in writing. Women with an ultrasound‐confirmed breech presentation and a gestational age of 33 + 4 to 35 + 4 weeks were eligible. Exclusion criteria were multiple pregnancy, uterine malformation, premature rupture of members, placenta praevia, third‐trimester vaginal bleeding, fetal malformation, admission for threatened preterm delivery, contraindication to vaginal delivery, and inability to attend for acupuncture every 48 hours.
Interventions	Fetal presentation was checked by clinical examination at the beginning of each session. Ultrasound was used when there was doubt about the presentation. If the presentation had become cephalic, no further study treatment was given. Intervention group: moxibustion plus acupuncture plus usual care. Moxibustion (using a smokeless moxa stick) and acupuncture were administered by a midwife to BL67 for 20 minutes (10 minutes per side) every 48 hours, unless version was confirmed. The moxa stick was held above the acupuncture point and on the acupuncture needle; each side was heated alternately until the patient perceived the heat as too strong. 6 sessions were administered in total. Control group: inactivated laser placebo, administered by a midwife, plus usual care. The inactivated laser probe was placed on BL67, without pressure, for 20 minutes (10 minutes per side) every 48 hours. The simulation characteristics were confirmed by a laser study centre (CERLA ‐ atom, molecule and laser physics laboratory, UMR University ‐ CNRS 8523, Villeneuve d’Ascq, France). The collimating lens was removed from the probe, and it emitted only 150 microwatts after inactivation. 6 sessions were administered in total.
Outcomes	Primary outcome: rate of breech presentation at 37 + 2 weeks' gestation (or at birth if born before 37 + 2 weeks) Secondary outcomes: percentage of successful ECV, rate of breech presentations at delivery, caesarean rate at delivery, safety, type of breech at inclusion, number of sessions, presentation at each session, rate of preterm delivery, rate of PROM, mode of labour onset
Notes	Study dates: October 2006 to May 2010 Sample size calculation: conducted, based on the study Neri 2004 Funding sources: no financial disclosures reported Authors' declarations of interest: declared, no conflicts of interest
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Quote: 'The randomization allocation sequence was generated by the study statistician from tables of random numbers for two equal arms (one‐to‐one), in blocks of eight, stratified by parity (primiparous or multiparous)'.
Allocation concealment (selection bias)	Low risk	Quote: 'opaque, white, sealed, and consecutively numbered envelopes'.
Blinding of participants and personnel (performance bias)	Low risk	Single‐blind study, attempted to blind participants. The lack of blinding of personnel was considered unlikely to have influenced outcomes, including those that required clinical decision‐making.
Blinding of outcome assessment (detection bias)	Low risk	The outcome assessors were not blind to group allocation; however, the lack of blinding was considered unlikely to influence the outcomes, including those that required clinical decision‐making.
Incomplete outcome data (attrition bias) All outcomes	Low risk	A small number of women were ineligible after randomisation (treatment group: 2/164, 1.2%; control group: 5/164, 3.5%), and a small number of women did not receive treatment (treatment group: 6/164, 3.7%; control group: 5/164, 3.1%). ITT analysis was conducted.
Selective reporting (reporting bias)	Unclear risk	All outcomes specified in the trial registration were reported. Additional outcomes were reported that were not specified in the trial registration. The trial was retrospectively registered.
Other bias	Low risk	No differences in baseline characteristics.

Do 2011.

Study characteristics
Methods	Parallel‐group RCT of moxibustion plus usual care versus no moxibustion plus usual care
Participants	20 women were recruited from the John Hunter Hospital, a tertiary public hospital in urban Newcastle, Australia. Women were allocated to receive moxibustion plus usual care (n = 10) or usual care alone (no moxibustion; n = 10). Inclusion criteria were pregnant women aged at least 18 years, with a singleton breech presentation confirmed by ultrasound at 34 to 36 + 5 weeks' gestation. Exclusion criteria were twin pregnancy, risk of premature birth, heart or kidney disease affecting the mother, placenta praevia, a history of antepartum haemorrhage, intrauterine growth restriction, hypertensive disease, isoimmunisation, previous uterine operations, uterine anomaly, prelabour rupture of the membranes, multiple pregnancy, fetal congenital abnormality, contraindication to vaginal delivery, and fetal death in utero.
Interventions	Women with confirmed breech 10 days after the end of treatment were offered ECV. Intervention group: moxibustion plus usual care. Moxibustion was applied to BL67 for 20 minutes (10 minutes each foot), twice daily for 10 days. Women or their partner (or both) were trained to provide the treatment at home. Control group: standard care (no moxibustion) according to the protocols of the John Hunter Hospital (weekly visits with midwives and obstetricians; no additional care, although individual clinicians may provide advice to women)
Outcomes	Cephalic presentation at birth, need for ECV, mode of birth, perinatal morbidity and mortality, maternal complications, maternal satisfaction with treatment, uterine contractions, fetal movements, intervention adherence, and adverse events
Notes	Australian New Zealand Clinical Trials Registry (accessed 8 February 2010) Study dates: December 2009 to June 2010 Sample size calculation: not conducted (feasibility study) Funding sources: University of Western Sydney ExPo Authors' declarations of interest: declared, no conflicts of interest
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Computer‐generated randomisation using balanced, variable blocks.
Allocation concealment (selection bias)	Low risk	Opaque, sealed, sequentially numbered envelopes.
Blinding of participants and personnel (performance bias)	High risk	Neither participants nor personnel were blinded. The lack of blinding was considered unlikely to influence the primary outcomes; however, the lack of blinding was likely to influence patient‐reported, subjective outcomes.
Blinding of outcome assessment (detection bias)	Low risk	Not specified, although the lack of blinding is unlikely to influence objective outcomes.
Incomplete outcome data (attrition bias) All outcomes	Low risk	Outcome data were available for all women. ITT analysis was conducted.
Selective reporting (reporting bias)	Low risk	All outcomes specified in the trial registration were reported.
Other bias	Low risk	None identified.

Guittier 2009.

Study characteristics
Methods	Parallel‐group RCT of moxibustion plus usual care versus no moxibustion plus usual care
Participants	212 women were recruited from the maternity unit of a Swiss university hospital. Women were randomised to receive moxibustion plus usual care (n = 106) or usual care alone (no moxibustion; n = 106). Inclusion criteria were women between 34 to 36 weeks' gestation with a single fetus in breech presentation verified by ultrasonography. Exclusion criteria were women with uterine malformation, placenta praevia, and transverse lie.
Interventions	Breech presentation was confirmed prior to randomisation. Fetal position was assessed by ultrasound prior to administration of moxibustion. ECV was offered if breech presentation persisted at 37 weeks' gestation. Intervention group: moxibustion plus usual care. Moxibustion was administered at the hospital by trained staff 3 times per week for the first 12 women allocated to the moxibustion group, while the remaining 94 women received moxibustion 3 times per week at the hospital and self‐administered on the days they did not attend the hospital. Moxibustion was administered to the point BL67 for 10 minutes each side. A maximum of 14 sessions was scheduled over 2 weeks.  Control group: usual care (no moxibustion)
Outcomes	Primary outcome: cephalic presentation at delivery or before ECV if attempted Secondary outcomes: cephalic presentation at delivery, including after successful ECV; mode of delivery; perineal injury or complications of delivery (from trial report); proportion of neonates with Apgar scores less than 7 at 5 minutes (from trial report); arterial cord blood pH less than 7.10 (from trial report); neonatal complications (from trial report); and total duration of hospital stay for the woman and neonate (from trial report)
Notes	Sample size calculation and ITT analysis performed. Study dates: October 2004 to December 2008 Sample size calculation: performed Funding sources: not reported Authors' declarations of interest: declared, no conflicts of interest The additional outcomes from the trial report that were not specified in the trial registration all showed no difference between groups. Hence, the inclusion of these outcomes is unlikely to have biased the findings of our review.
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Computer‐generated randomisation sequence in blocks of 4, 6 and 8. Stratified according to women's intention to attempt ECV if required at 37 weeks.
Allocation concealment (selection bias)	Low risk	Sealed and consecutively numbered opaque envelopes.
Blinding of participants and personnel (performance bias)	Low risk	Neither participants nor personnel were blinded. The lack of blinding of personnel was considered unlikely to have influenced outcomes, including those that required clinical decision‐making.
Blinding of outcome assessment (detection bias)	Low risk	The study was not blinded, but the lack of blinding was considered unlikely to influence outcomes, including those that required clinical decision‐making.
Incomplete outcome data (attrition bias) All outcomes	Low risk	There was complete follow‐up of all women randomised.
Selective reporting (reporting bias)	Unclear risk	Additional outcomes reported in the trial report that were not specified in the retrospective trial registration.
Other bias	Unclear risk	No imbalances at randomisation. The study appears free from other sources of bias. First 12 women in the moxibustion group received treatment three times per week, remaining women received moxibustion three times per week plus self‐administered on non‐visit days.

Higashihara 2021.

Study characteristics
Methods	Parallel‐group, 3‐arm quasi‐RCT of smoke moxibustion plus usual care versus smokeless moxibustion plus usual care versus patient education plus usual care
Participants	63 women were recruited from 2 perinatal medical centres, a maternity hospital and an obstetrics and gynaecology clinic in Tokyo, Japan. Women were sequentially allocated to groups; the study planned to allocate the first 20 women to the smoke moxibustion plus usual care group, the next 20 women to the smokeless moxibustion plus usual care group, and the final 20 women to the control group. Inclusion criteria were pregnant women aged 18 years and above with ultrasound‐confirmed singleton breech presentation between 33 and 35 weeks' gestation. Women were also required to have normal progression of pregnancy and normal fetal biometry. Exclusion criteria were maternal heart or kidney disease; obstetric complications (pregnancy with multiples of twins and beyond; risk of preterm birth (preterm uterine contractions, initial dilatation or shortening of the cervix with a score of 4 on the Bishop scale, tocolytic therapy), uterine fibroids > 4 cm, placenta praevia, hypertensive disorders of pregnancy or PROM); contraindication to vaginal delivery (previous uterine surgery, uterine malformations or bone pelvic defects); fetal conditions (intrauterine growth restriction, fetal malformation, or chromosomal disorder); conditions to avoid in interventions (pregnant women or siblings of the fetus diagnosed with bronchial asthma or a pulmonary problem and are treated; allergies to Artemesia vulgaris; pregnant women or siblings of the fetus who have symptoms of coughing, respiratory discomfort from smoke; prior moxibustion treatment to achieve fetal version).
Interventions	In Japan, obstetricians perform an ultrasound at most appointments. In the third trimester of pregnancy, when the baby is in the breech presentation, an ultrasound is performed every time to inform and educate the pregnant woman about the fetal position. The pregnant woman thus knows the position of the fetal spine.  Intervention group 1: smoke moxibustion (conventional moxa stick) plus usual care. Moxibustion was self‐administered to BL67 for 20 minutes (10 minutes each foot), once or twice daily for 10 to 14 days. Moxibustion was applied 1.5 cm to 3 cm away from the skin. Intervention group 2: smokeless moxibustion (carbonized, odourless moxa stick) plus usual care. Moxibustion was self‐administered to BL67 for 20 minutes (10 minutes each foot), once or twice daily for 10 to 14 days. Moxibustion was applied 1.5 cm to 3 cm away from the skin. Control group: patient education plus usual care. Women received education leaflets about adequate sleep, balanced diet, exercise, rest, stress reduction, "important aspects of uterine contraction", and "lying down positions that they have to assume depending on the location of the fetal spine". Women were also instructed not to have any moxibustion treatment but to spend their time naturally until the next medical check‐up after 10 to 14 days.
Outcomes	Primary outcome: rate of cephalic presentation, determined by ultrasound, 10 to 14 days from the commencement of the intervention Secondary outcomes: rate of cephalic presentation at birth, mode of birth, maternal outcomes (number of ECV), wellbeing of the mother and child (premature birth, PROM at < 37 weeks, Apgar score < 7 at 5 minutes, umbilical cord blood pH < 7.1, admission to NICU, and intrauterine fetal death) Additional outcomes reported in the trial registration were gestational age at birth, birthweight, number of self‐reported fetal movements, and side effects.
Notes	Study dates: March 2016 to January 2017 Sample size calculation: performed, based on Do 2011 Funding sources: Japan Society for the Promotion of Science (JSPS) KAKENHI, Grant/Award Numbers: 16H05589, 17H06985; The Japan Academy of Midwifery Grant‐in‐Aid for Encouragement of Scientists; Yamaji Fumiko Nursing Research Fund Authors' declarations of interest: declared, no competing interests The 2 treatment groups were considered to have similar clinical effects, so results from the 2 groups were merged for analysis.
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	High risk	Participants allocated sequentially: ‘First, participants were recruited for the SM group until the required number of 20 people was secured. Second, participants were recruited for the SLM group until the required number of 20 people was secured. Third, participants were recruited for the control group until the required number of 20 people was secured’.
Allocation concealment (selection bias)	High risk	No allocation concealment.
Blinding of participants and personnel (performance bias)	Unclear risk	No blinding, but the lack of blinding was considered likely to influence objective outcomes. It is possible that the lack of blinding affected reporting of subjective outcomes, such as fetal movements.
Blinding of outcome assessment (detection bias)	Unclear risk	Not reported.
Incomplete outcome data (attrition bias) All outcomes	Low risk	Low rate of attrition, which was similar between groups. The low rate of attrition was considered unlikely to influence the outcomes (treatment group [smoke moxa]: 2/22, 10.0% excluded from analysis; treatment group [smokeless moxa]: 0/20 excluded from analysis; control group: 1/21, 5.0% excluded from analysis).
Selective reporting (reporting bias)	Low risk	Outcomes reported in prospective trial registration were reported. Additional variables were specified in the trial registration (e.g. birthweight, gestational age at birth); howevere, these were not considered clinical outcomes.
Other bias	Low risk	No differences between groups in baseline characteristics.

Millereau 2009.

Study characteristics
Methods	Parallel‐group RCT of moxibustion plus usual care versus no moxibustion plus usual care
Participants	68 women were recruited from the Saint‐Nazaire hospital centre in Loire Atlantique, France. The hospital centre is a multidisciplinary general hospital with 1100 beds, featuring a level 2B maternity unit with 44 beds and 7 obstetrician gynecologists. Inclusion criteria were women of legal age who were pregnant with a single fetus and who planned to give birth at the hospital. Breech was confirmed by ultrasound in the 32nd week of gestation and at the obstetrical consultation at the 8th month of pregnancy. Exclusion criteria were multiple pregnancies, planning to give birth outside the department, refusal to participate (patient or obstetrician), and women who were expected to have attempted ECV.
Interventions	Intervention group: moxibustion plus usual care. The smokeless moxibustion stick was held close to BL67 until a sensation of tolerable heat was felt. The first moxibustion session was administered by the acupuncturist, and subsequent treatments were either self‐administered by the woman or were administered by the partner or a third party. Moxibustion was performed for 15 to 20 minutes, preferably between 3 and 5 p.m., once daily for 1 week.  Control group: usual care (no moxibustion)
Outcomes	Primary outcome: presentation at delivery Secondary outcomes: mode of delivery, birthweight
Notes	Study dates: January 2006 to April 2008 Sample size calculation: conducted, based on previously published data (study not specified) Funding sources: not specified Authors' declarations of interest: not specified
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Quote: 'The draw was performed using a permutation table of four random numbers. A stratification was made on parity with a “primiparous” group and a “multiparous” group'.
Allocation concealment (selection bias)	Low risk	Drawing of lots.
Blinding of participants and personnel (performance bias)	Low risk	The participant was not blind to group allocation, but the obstetrician was blind to group allocation. The lack of blinding of personnel was considered unlikely to have influenced outcomes, including those that required clinical decision‐making.
Blinding of outcome assessment (detection bias)	Unclear risk	Not reported.
Incomplete outcome data (attrition bias) All outcomes	Low risk	No attrition.
Selective reporting (reporting bias)	Unclear risk	No trial protocol or trial registration identified.
Other bias	Unclear risk	Funding source not specified, insufficient reporting.

Neri 2004.

Study characteristics
Methods	Parallel‐group RCT of moxibustion plus acupuncture plus usual care versus no moxibustion plus usual care
Participants	240 women were recruited to the trial in Italy. Women were randomised to receive either moxibustion plus acupuncture plus usual care (n = 120) or usual care alone (no moxibustion; n = 120). Inclusion criteria were Caucasian (understood to be white), singleton pregnancy, between 33 and 35 weeks' gestation, breech presentation confirmed by ultrasound within 24 hours of randomisation. Exclusion criteria were greater than 35 weeks' gestation, previous uterine surgery (including caesarean section), uterine malformations, risk of premature birth, twin pregnancy, fetal malformations or chromosomal abnormalities, abnormal fetal biometry, renal/cardiac maternal disease, transverse or oblique lie.
Interventions	All women had breech presentation confirmed by ultrasound and were randomised within 24 hours. ECV was not offered to those participating in the study. Intervention group: moxibustion plus acupuncture plus usual care. Women allocated to the moxibustion plus acupuncture plus usual care group received twice‐weekly treatments for 2 weeks, if necessary. To determine whether to continue with treatment, an ultrasound was performed before each session to determine presentation. Acupuncture needles were inserted bilaterally into BL67, and de qi sensation was sought. Once obtained, needle stimulation ceased and needles were left in situ for 20 minutes. Moxibustion was applied for 20 minutes, with the side of moxibustion stimulation alternated every 2 minutes. After the first session, a non‐stress test was performed for 30 minutes, and maternal blood pressure and heart rate were monitored.  Control group: usual care (no moxibustion)
Outcomes	Primary outcome: presentation at birth Secondary outcomes: incidence of caesarean section, presentation at the end of treatment, and adverse events relating to treatment
Notes	Study dates: May 2000 to June 2002 Sample size calculation: performed Funding sources: not specified Authors' declarations of interest: not specified
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	The trial used a computer‐generated sequence for allocation.
Allocation concealment (selection bias)	Low risk	Central randomisation.
Blinding of participants and personnel (performance bias)	Low risk	Neither participants nor personnel were blinded. The lack of blinding of personnel was considered unlikely to have influenced outcomes, including those that required clinical decision‐making.
Blinding of outcome assessment (detection bias)	Low risk	The study was not blinded, but the lack of blinding was considered unlikely to influence outcomes, including those that required clinical decision‐making.
Incomplete outcome data (attrition bias) All outcomes	High risk	8 women from the treatment group and 6 women from the control group withdrew from the trial. In the treatment group, 5 women withdrew due to uterine contractility observed by their obstetrician, 2 due to the onset of mild hypertension and 1 did not adhere to the treatment. In the control group, 2 women withdrew due to threatened preterm labour and 4 refused to be assigned to the observation group after disclosure of their group allocation.
Selective reporting (reporting bias)	Unclear risk	No trial protocol or trial registration available.
Other bias	Low risk	No imbalances at randomisation. The study appears free from other sources of bias.

Neri 2007.

Study characteristics
Methods	Parallel‐group RCT of moxibustion plus usual care versus acupuncture plus usual care versus moxibustion plus acupuncture plus usual care
Participants	41 primiparous women were recruited through the acupuncture clinic of an obstetric department in Italy. Women were randomised to receive either moxibustion plus usual care (n = 16), acupuncture plus usual care (n = 10), or acupuncture plus moxibustion plus usual care (n = 15). Inclusion criteria were women with a singleton pregnancy of 33 to 36 weeks' gestation, with fetus in frank position. Exclusion criteria included previous uterine surgery, twin pregnancy, low level of amniotic fluid, maternal renal or cardiac disease, and fetal malformation or chromosome abnormality.
Interventions	Women were subject to a computerised cardiotocograph to measure fetal activity for 20 minutes before, during, and for 20 minutes after treatment. Acupoint stimulation was elicited during an active period (defined as a period with presence of fetal movements and accelerations of fetal heart rate, and variability of fetal heart rate greater then 10 beats per minute). It was unclear when fetal presentation was assessed. Group 1: moxibustion plus usual care. Moxibustion was administered twice per week, 20 minutes each time. Group 2: acupuncture plus usual care. Acupuncture was administered twice per week, 20 minutes each time. Group 3: moxibustion plus acupuncture plus usual care. Moxibustion and acupuncture were administered twice per week, 20 minutes each time.
Outcomes	The primary outcomes were fetal heart rate, fetal movements, and cephalic version.
Notes	Study dates: not reported Sample size calculation: not reported Funding sources: not reported Authors' declarations of interest: not reported
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Method of group allocation was not reported.
Allocation concealment (selection bias)	Unclear risk	Method of group concealment was not reported.
Blinding of participants and personnel (performance bias)	Unclear risk	Neither participants nor personnel were blinded. It is not clear whether the lack of blinding may have influenced the outcomes.
Blinding of outcome assessment (detection bias)	Unclear risk	The study was not blinded, and it was not clear whether the lack of blinding would have influenced outcomes.
Incomplete outcome data (attrition bias) All outcomes	Low risk	2 women (1 from the moxibustion group and 1 from the moxibustion plus acupuncture group) were excluded post‐randomisation due to not meeting the baseline cardiotocographic criteria. The review authors deemed this low risk. No ITT analysis was conducted.
Selective reporting (reporting bias)	Unclear risk	No trial protocol or trial registration available.
Other bias	Unclear risk	While there were unequal numbers of participants per group, it is difficult to determine whether this was due to a failure of randomisation or whether this was due to change.

Tsujiuchi 2017.

Study characteristics
Methods	Parallel‐group RCT of moxibustion plus acupuncture plus usual care versus usual care
Participants	14 women were recruited from the obstetrics and gynaecology clinic at Serie Hospital, Kanagawa Prefecture, Japan. Women were randomised to receive either moxibustion plus acupuncture plus routine care (n = 5) or usual care (no moxibustion; n = 9). Inclusion criteria were primiparas aged between 20 and 45 years who had been diagnosed with a breech presentation during the 30th week of gestation (30 + 0 to 30 + 6). The exclusion criterion was pregnant women who were being treated with moxibustion and acupuncture.
Interventions	Intervention group: moxibustion plus acupuncture plus usual care. Moxibustion was self‐administered or administered by the woman's family at home using pedestal, stick‐on indirect moxa (Yamamasa Chosei Moxibustion Hard) for SP6 and moxa stick for BL67. Moxibustion was applied for 7 minutes, twice daily until the breech position was changed. Acupuncture (with 0.6‐millimetre press needles) was administered to SP6 Sanyinjiao for 15 minutes and BL67 for 10 minutes. Treatment was performed in the half‐sitting position.  Acupuncture treatment was provided once per week for 3 weeks (between 30 and 32 weeks' gestation). Control group: usual care (no moxibustion or acupuncture)
Outcomes	Primary outcome: rate of cephalic version after treatment Secondary outcomes: delivery outcomes (caesarean delivery and vaginal delivery), obstetric abnormalities and adverse acupuncture and moxibustion events in the intervention group (preterm birth, rupture of membranes, mood illness, moxa burns, etc.) Additional outcome from trial registration: "Questionnaire of general malaise"
Notes	Study dates: September 2013 to August 2015 Sample size calculation: not specified Funding sources: part of the study was funded by a research grant from the 2011 and 2012 Oriental Therapy Trial Foundation. The trial registration reports funding from the Foundation for training and licensure examination in amma‐massage‐acupressure, acupuncture, and moxibustion. Authors' declarations of interest: none declared
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Not reported.
Allocation concealment (selection bias)	Unclear risk	Sealed envelopes.
Blinding of participants and personnel (performance bias)	Unclear risk	Trial registration reports that the study was 'double blind—all involved are blinded'. However, it is not clear how the control group could be blind to group allocation. The effect of the lack of blinding on the 'Questionnaire of general malaise' is not clear.
Blinding of outcome assessment (detection bias)	Unclear risk	Trial registration reports that the study was 'double blind—all involved are blinded'. However, it is not clear how the control group could be blind to group allocation. The effect of the lack of blinding on the 'Questionnaire of general malaise' is not clear.
Incomplete outcome data (attrition bias) All outcomes	Low risk	Data available for all women.
Selective reporting (reporting bias)	Unclear risk	The results for the 'Questionnaire of general malaise' were not reported.
Other bias	Unclear risk	The mean age of the intervention group was higher than the control group. No power analysis was reported, and the reporting was unclear.

Vas 2013.

Study characteristics
Methods	Parallel‐group, 3‐arm RCT of moxibustion plus usual care versus sham moxibustion plus usual care versus no moxibustion plus usual care
Participants	406 women were recruited from 58 primary healthcare centres belonging to the Andalusia (Spain) Public Health System in the provinces of Sevilla, Huelva, Cádiz, and Málaga. Women were randomised to receive moxibustion plus usual care (n = 136), sham moxibustion plus usual care (n = 136), or usual care alone (no moxibustion; n = 134). Inclusion criteria were women with a non‐cephalic presentation confirmed by physical examination and ultrasound. Women were aged at least 18 years, had gestational age 33 to 35 weeks estimated by ultrasound, normal fetal biometry, and had not previously used moxibustion for fetal version. Exclusion criteria were multiple pregnancy, bone pelvic defects, previous uterine surgery, fetal malformation or chromosomal disorder, uterine malformations, risk of preterm birth (preterm uterine contractions and/or dilatation or shortening of the cervix with a score of 4 on the Bishop scale), uterine fibroids > 4 cm, tocolytic therapy, maternal heart or kidney disease.
Interventions	Moxibustion (real or sham) was administered at home by a family member who had received training on its application from the midwife. Moxibustion (real or sham) was applied for 20 minutes, alternating feet when the heat became uncomfortable, once per day for 2 weeks. Women were advised to stop treatment if they noticed a change in the fetal position or experienced any adverse effects. Intervention group: moxibustion plus usual care. Moxibustion was applied to BL67, with the moxa stick held 1.5 cm to 3 cm above the skin. Control group 1: sham moxibustion plus usual care. Moxibustion was applied to SP1 Yinbai, which is thought to be an non‐active point for breech presentation. Control group 2: usual care (no moxibustion). Usual care was provided according to the clinical guidelines of the Andalusian Healthcare System Treatment Guide, which includes recommendations for knee‐chest postural management for women with abnormal fetal presentation after the third quarterly ultrasound and caesarean section when non‐vertex presentation, confirmed by ultrasound, is maintained until full term.
Outcomes	Primary outcome: cephalic presentation at birth Secondary outcomes: cephalic version at the end of treatment and mode of birth Additional outcomes from published trial protocol: rate of cephalic presentation at 38 weeks (determined by ultrasound), number of days of treatment until version occurs, blinding  Additional outcomes from trial registration: rate of cephalic presentation at week 38 of gestation, determined by echography; number of days of treatment received until version occurs; rate of caesarean sections. Security variables: increased rates of fetal cardiac rhythm change, followed up until birth; risk of premature birth; changes in Apgar scores at 1 minute and at 5 minutes The additional outcomes from the published trial protocol were not outcomes prespecified in this review. The additional outcomes from the trial registration that were prespecified in this review showed no difference between groups for risk of preterm birth, and none of the trial participants' babies had an Apgar score less than 7 at 5 minutes. We considered the additional outcomes reported in the trial protocol and trial registration to be unlikely to influence the findings of the review.
Notes	Study dates: April 2008 to December 2010 Sample size calculation: performed, based on Vas 2009 Funding sources: Andalusian Public Health System (grant number PI‐0053/2007). ("The funder had no role in the design or conduct of the study, analysis or interpretation of the data, or in the decision to publish.") Authors' declarations of interest: declared, no conflicts of interest
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Central randomisation performed by a statistician who was not involved in the study. Quote: '1:1:1 scheme in blocks of 12 and stratified by centre.'
Allocation concealment (selection bias)	Low risk	Sealed, opaque envelopes were prepared for each participating healthcare centre.
Blinding of participants and personnel (performance bias)	Low risk	Participants in the moxibustion and sham moxibustion groups were blind to group allocation. The control group was not blinded, and the lack of blinding of personnel was considered unlikely to have influenced outcomes, including those that required clinical decision‐making.
Blinding of outcome assessment (detection bias)	Low risk	Most of the data were recorded from the Pregnant Woman's Health Passport or the Child's Health Passport (completed by hospital staff with no connection to the study).
Incomplete outcome data (attrition bias) All outcomes	Low risk	Low rate of attrition; similar numbers between groups (treatment group: 0/136, 0% lost to follow up; sham moxibustion group: 3/136, 2.2% lost to follow up; control group: 2/134, 1.5% lost to follow up). ITT analysis performed.
Selective reporting (reporting bias)	Unclear risk	Differences noted in outcomes reported in trial protocol, trial registration and trial report.
Other bias	Low risk	No differences in baseline characteristics.

Yang 2006.

Study characteristics
Methods	Randomised trial of moxibustion plus knee‐chest therapy plus usual care versus knee‐chest therapy plus usual care
Participants	206 women with breech presentation were recruited from the Sixth Hospital of Zhongyuan Oil Field, Shandong province, China. Women were randomly allocated to receive either moxibustion plus knee‐chest therapy plus usual care (n = 103) or knee‐chest therapy plus usual care (n = 103). Women were between 28 and 34 weeks' gestation. No exclusion criteria were reported. While it was not reported whether all women had a singleton pregnancy, this was assumed.
Interventions	Intervention group: moxibustion plus knee‐chest therapy. Moxibustion was applied to BL67 for 15 to 20 minutes twice daily, with 7 days constituting 1 course. The duration of the intervention was 1 week.  Control group: knee‐chest therapy plus usual care. The details of knee‐chest therapy were not reported.
Outcomes	The primary outcome was cephalic version.
Notes	Study dates: not available Sample size calculation: not performed Funding sources: not available Authors' declarations of interest: not available
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	High risk	Group allocation by date of admission.
Allocation concealment (selection bias)	Unclear risk	The method of concealment was not reported.
Blinding of participants and personnel (performance bias)	Low risk	Neither participants nor personnel were blinded. The lack of blinding of personnel was considered unlikely to have influenced outcomes, including those that required clinical decision‐making.
Blinding of outcome assessment (detection bias)	Low risk	The study was not blinded, but the lack of blinding was considered unlikely to influence outcomes, including those that required clinical decision‐making.
Incomplete outcome data (attrition bias) All outcomes	Low risk	There was complete follow‐up for all women.
Selective reporting (reporting bias)	Unclear risk	Study protocol unavailable; unable to determine whether all outcomes were pre‐specified.
Other bias	Unclear risk	There is insufficient information to determine whether other bias is present.

BL67: Bladder 67
ECV: external cephalic version
ITT: intention‐to‐treat
NICU: neonatal intensive care unit
PP: per protocol
PROM: premature rupture of membranes
RCT: randomised controlled trial

Characteristics of excluded studies [ordered by study ID]

Study	Reason for exclusion
Anon 2019	The report was not a clinical study.
Anon 2021	The report was not a clinical study.
Baccetti 2012	Duplicate
Balk 2005	The report was not a clinical study.
Beer 1995	The study was not a randomised or quasi‐randomised controlled trial.
Brill 2003	The report was not a clinical study.
Cardini 1991	The study was not a randomised or quasi‐randomised controlled trial.
Dahlen 2011	Duplicate report of Do 2011
Ernst 1999	The report was not a clinical study.
Ernst 2005	The report was not a clinical study.
Fu 1990	The study did not test moxibustion.
García‐Mochón 2016	The article reports the cost‐effectiveness of moxibustion from Vas 2013. Cost‐effectiveness was not one of the prespecified outcomes of this review.
Guittier 2008	The report describes a subset of data from the treatment group of Guittier 2009.
Hayashida 1989	The study was not a randomised or quasi‐randomised controlled trial.
Huang 1990	The study was not a randomised or quasi‐randomised controlled trial.
ISRCTN98448646	Per the investigator, the trial did not proceed.
Kanakura 2001	The study was not a randomised or quasi‐randomised controlled trial.
Langer 2016	This abstract was for an excluded study (Sananes 2016).
Lee 2008	The report was not a clinical study.
Li 1996	Results were not reported separately for different presentation types.
Manyande 2009	The study was not a randomised or quasi‐randomised controlled trial.
Miranda Garcia 2011	The study was not a randomised or quasi‐randomised controlled trial.
Miranda Garcia 2017	The study was not a randomised or quasi‐randomised controlled trial.
Mitchell 2008	The report was not a clinical study.
Neri 2002	The study was not a randomised or quasi‐randomised controlled trial.
Pardeshi 2020	The study was not a randomised or quasi‐randomised controlled trial.
Qin 1989	The trial did not test moxibustion.
Raben 1999	The study was not a randomised or quasi‐randomised controlled trial.
Rosim 2017	The study was not a randomised or quasi‐randomised controlled trial.
Sananes 2016	The trial did not test moxibustion.
Sarsmaz 2021	The study was not a randomised or quasi‐randomised controlled trial.
Smith 2013	The report was not a clinical study.
van den Berg 2005	The report was not a clinical study.
van den Berg 2006	The study was not a randomised or quasi‐randomised controlled trial.
van den Berg 2010a	The report was not a clinical study.
van den Berg 2010b	The study was not a randomised or quasi‐randomised controlled trial.
Vas 2012	Duplicate
Vas 2014	Duplicate report of Vas 2013
Wagner‐Pankl 1990	The study was not a randomised or quasi‐randomised controlled trial.
Wen 1979	The study was not a randomised or quasi‐randomised controlled trial.
White 2010	The report was not a clinical study.
Wong 1999	The report was not a clinical study.
Wu 1994	The study was not a randomised or quasi‐randomised controlled trial.

Characteristics of studies awaiting classification [ordered by study ID]

Albaladejo 2017.

Methods	Parallel‐group RCT of moxibustion plus knee‐chest position plus usual care versus sham moxibustion plus knee‐chest position plus usual care
Participants	88 women were recruited to the study and randomised to receive moxibustion plus knee‐chest position plus usual care (n = 47) or sham moxibustion plus knee‐chest position plus usual care (n = 41). The study was a multicentre trial conducted in the Sexual and Reproductive Health Care Service (ASSIR) of 9 primary care centres in the Baix Llobregat region of Barcelona, Spain. Inclusion criteria were third trimester women (between 32 and 36 weeks' gestation) with non‐cephalic presentation confirmed by ultrasound; women were older than 18 years and had a low‐risk pregnancy. Exclusion criteria were multiple pregnancy, pelvic malformations, kidney or heart disease, risk factors that favour ending the delivery with caesarean section (2 or more previous caesarean sections, previous maternal pathology, severe intrauterine fetal growth retardation, fetal malformations, uterine abnormalities, etc.), risk of preterm birth (contractions and/or onset of cervical dilation), and/or language difficulties.
Interventions	Intervention group: moxibustion plus knee‐chest position plus usual care. Moxibustion was self‐administered to BL67 for 20 minutes (alternating foot every 2 minutes), twice daily for 1 week (14 sessions in total). If cephalic version had not occurred after each treatment, women were placed in the knee‐chest position for 10 minutes. Control group: sham moxibustion (moxibustion at a neutral point: SP1 Yinbai, lateral to the external angle of the nail of the first toe) plus knee‐chest position plus usual care. The treatment technique was the same as for the intervention group. Women whose baby had not turned after treatment were placed in the knee‐chest position for 10 minutes.
Outcomes	Fetal presentation (ultrasound confirmed and Leopold’s manoeuvres) at the end of treatment, fetal presentation at birth, gestational age at delivery, date of delivery, mode of birth, need for ECV, birthweight, safety
Notes	Study dates: April 2014 to September 2015 Sample size calculation: conducted Funding sources: not specified Authors' declarations of interest: not specified 10/47 women in the treatment group and 4/41 women in the control group had fewer than 7 treatments. We are awaiting further information about trial registration.

Casas 2017.

Methods	Parallel‐group RCT of moxibustion plus usual care versus acupressure plus usual care
Participants	395 women were recruited to the trial, which was conducted in the Galician Health Service, Spain. Women were randomised to moxibustion plus usual care (n = 163) or acupressure plus usual care (n = 183); an additional 49 non‐randomised women were included in a control group (usual care alone). The non‐randomised group was excluded from this review.  Inclusion criteria were women aged over 18 years with a single fetus in breech presentation (confirmed by ultrasound). Women were between 34 + 0 and 34 + 6 weeks' gestation and did not meet the exclusion criteria. Exclusion criteria were twin pregnancy, kidney or heart disease, defects of the pelvic bone structure, previous uterine surgery, pathological pregnancy (intrauterine growth restriction, pre‐eclampsia, severe infections, polyhydramnios, oligohydramnios, placenta praevia, isoimmunisation, etc.), threatened preterm delivery with the need for tocolytics during the current pregnancy, fetal malformations, fibroids > 4 cm and uterine malformations.
Interventions	Women with breech presentation at the end of treatment (35 weeks' gestation) were offered ECV. The treatment protocols for moxibustion and acupressure were developed by the research team based on previous research. All women whose fetus remained in breech presentation at the end of treatment were candidates for usual treatment, which consisted of expectant management with the option for ECV at 37 weeks' gestation. Intervention group: moxibustion plus usual care. Moxibustion was applied by a midwife to BL67 (bilaterally) using a moxa stick once daily for up to 10 days, at the same time each day. The moxa stick was held 1.5 cm to 2.5 cm from the skin to induce a sensation of intense heat. Control group: acupressure plus usual care. Acupressure to BL67 was applied using a small metallic sphere that remained in place for 10 days.
Outcomes	Presentation at end of treatment and birth, need for ECV, outcome of ECV, mode of delivery, perinatal morbidity and mortality, maternal satisfaction, adverse events
Notes	Study dates: September 2013 to December 2016 Sample size calculation: not reported Funding sources: funded by the Carlos III Health Institute, Ministry of Economy and Competitiveness with a grant such as Health Research Project (PI12/01128) Authors' declarations of interest: not specified We are awaiting further information about trial registration.

Chen 2004.

Methods	Randomised trial of moxibustion plus lifting‐buttocks‐side‐lying position compared with lifting‐buttocks‐side‐lying position
Participants	142 women with breech presentation were randomised to the trial through a hospital in Zhejiang Province, China. Inclusion criteria was 30 to 34 weeks' gestation; no other inclusion or exclusion criteria were reported.
Interventions	Intervention group: 73 women were allocated to the moxibustion plus lifting‐buttocks‐side‐lying position. The details of lifting‐buttock‐side‐lying position were not explained. Moxibustion was applied to BL67 for 10 to 15 minutes once daily, with 3 days constituting 1 course.  Control: 69 women were allocated to the lifting‐buttocks‐side‐lying position only group. The duration of the intervention was unclear, but the average length of treatment was 4 days.
Outcomes	The primary outcome was cephalic presentation either confirmed by ultrasound or during labour. Details of outcome measurement were unclear.
Notes	No sample size calculation performed. We are awaiting further information about randomisation and plausibility of the results.

Chen 2007.

Methods	Randomised controlled trial of moxibustion plus usual care versus knee‐chest postural therapy plus usual care
Participants	Inclusion criteria were not specified.
Interventions	Intervention group: moxibustion plus usual care. Moxibustion was performed daily for 20 minutes; 1 course is 5 days.  Control group: knee‐chest postural exercises plus usual care. Knee‐chest postural exercises were performed daily; 1 course is 5 days.
Outcomes	Effective rate (number of cephalic, occipitoanterior presentations after treatment)
Notes	No sample size calculation was performed. We are awaiting further information about randomisation and plausibility of the results.

Lin 2002.

Methods	Randomised trial of moxibustion plus knee‐chest therapy plus usual care versus knee‐chest therapy plus usual care
Participants	122 women between 30 and 37 weeks' gestation with breech presentation were randomised to the trial. No inclusion or exclusion criteria were reported. While it was not reported whether all women had a singleton pregnancy, this was assumed. Women were recruited from a hospital in Hubei Province, China.
Interventions	Intervention: moxibustion plus knee‐chest therapy plus usual care. 63 women received moxibustion to BL67 for 15 minutes, with 2 days constituting 1 course.  Control: knee‐chest therapy plus usual care. 59 women were allocated to receive knee‐chest therapy alone. The details of knee‐chest therapy were not reported.
Outcomes	The primary outcome was cephalic presentation either confirmed by ultrasound or during labour. Details of outcome measurement were unclear.
Notes	No sample size calculation was performed. The duration of the intervention is unclear. We are awaiting further information about randomisation and plausibility of the results.

Yang 2007.

Methods	Multicentre randomised controlled trial of moxibustion plus usual care versus knee–chest therapy plus usual care
Participants	Inclusion criteria: women with breech presentation or transverse lie diagnosed by ultrasound. Aged between 20 and 36 years, between 30 and 34 weeks' gestation, informed consent and voluntary participation.  Exclusion criteria: pregnancy diseases such as gestational hypertension and gestation diabetes, reproductive organ tumours, pelvic stenosis, polyhydramnios, ascending cord around the neck, fetal parietal diameter greater than 8.5 cm or placenta attached to the anterior wall of the uterus Setting: 3 hospitals (Chengdu Hospital of Integrated Traditional Chinese and Western Medicine Hospital; Panzhihua Central Hospital; Chengdu Military Region General Hospital)
Interventions	Intervention group: 147 women received moxibustion to BL67. After breech presentation is confirmed by ultrasound, apply moxibustion for 15 minutes, twice daily. 7 days is 1 course of treatment. Stop treatment if fetal presentation turns to cephalic presentation. If baby turns after 1 course but then reverts to breech, repeat another course of treatment and observe. Once cephalic presentation is achieved, observe every 2 weeks until the baby is born. If the baby doesn’t turn after 1 course of treatment, stop treatment and observe every 2 weeks until born.  Comparator group: 149 women received knee–chest therapy. After breech presentation is confirmed by ultrasound, the woman should assume the knee–chest position for 15 minutes, twice daily. 7 days is 1 course of treatment. Stop treatment if fetal presentation turns to cephalic presentation. If baby turns after 1 course but then reverts to breech, repeat another course of treatment and observe. Once cephalic presentation is achieved, observe every 2 weeks until the baby is born. If the baby doesn’t turn after one course of treatment, stop treatment and observe every 2 weeks until born.
Outcomes	Cephalic presentation at birth, fetal heart rate monitoring before and after each treatment
Notes	The data appear to be plausible. We are awaiting further information about randomisation and plausibility of the results.

Yang 2008.

Methods	Parallel‐group trial of moxibustion, acupuncture, postural techniques, and usual care versus postural techniques and usual care
Participants	120 women with ultrasound‐confirmed breech presentation and gestational age consistent with menstrual history
Interventions	Intervention: moxibustion plus acupuncture plus postural techniques plus usual care. Acupuncture to BL67, once de qi sensation obtained, apply moxibustion for 15 to 20 minutes, followed by postural techniques (knee‐chest therapy) for 15 minutes. 7 days is 1 course of treatment. Control: postural techniques plus usual care. Knee‐chest therapy for 15 minutes; 7 days is 1 course of treatment.
Outcomes	Effective rate (cephalic presentation confirmed by ultrasound after 2 weeks of treatment)
Notes	We are awaiting further information about randomisation and plausibility of the results.

Yang 2010.

Methods	Randomised controlled trial of the combination of acupuncture, moxibustion, knee–chest therapy, and usual care versus knee–chest therapy plus usual care
Participants	Inclusion criteria: ultrasound‐confirmed breech, gestational age consistent with menstrual history Exclusion criteria: multiple births, fetal malformations, pelvic stenosis, cephalopelvic failure and other factors not conducive to changing fetal position Setting: obstetrics outpatient department of People's Hospital of Pinglu District, Shuozhou City
Interventions	Intervention group: 60 women received acupuncture plus moxibustion to BL67. Acupuncture was applied on BL67 bilaterally. Once de qi sensation was achieved, moxibustion was applied on top of the acupuncture needles for 15 to 20 minutes, followed by knee‐chest therapy for 15 minutes each time. Both acupuncture and moxibustion were applied daily, and 7 days of treatment was 1 course. Comparator group: 60 women performed knee‐chest techniques daily for 15 minutes each time. 7 days of treatment was 1 course.
Outcomes	Effective rate (cephalic presentation confirmed by ultrasound after 2 weeks of treatment)
Notes	We are awaiting further information about randomisation and plausibility of the results.

BL67: Bladder 67
ECV: external cephalic version
RCT: randomised controlled trial

Differences between protocol and review

We have updated the Methods to reflect the latest Cochrane Handbook for Systematic Reviews of Interventions (Higgins 2021). We have carefully reviewed and revised the naming of comparisons to reflect the fact that all trial participants received usual care. 

We changed the outcome 'adverse events' to 'adverse events of moxibustion' in this update. 

Contributions of authors

Meaghan Coyle undertook additional searching to identify papers relevant to this review update. Meaghan Coyle, Caroline Smith, and Brian Peat selected trials for inclusion and performed data extraction and assessment of the quality of included trials. Meaghan Coyle performed statistical analysis and interpretation of the data for the review update. Caroline Smith and Brian Peat assisted with interpreting data and commented on drafts of the review update.

For the original protocol and review, Meaghan Coyle conceptualised and took the lead in writing the protocol and review; performed initial searches of databases for trials; was involved in selecting trials for inclusion; performed data extraction and assessment of the quality of included trials; and was responsible for statistical analysis and interpretation of the data.

Caroline Smith selected trials for inclusion in the original review; performed data extraction and assessment of quality of included trials; interpreted the data; and commented on drafts of the original protocol and review. Brian Peat commented on the original protocol and review.

Sources of support

Internal sources

• The University of Adelaide, Australia

   In‐kind support

• The University of South Australia, Australia

   In‐kind support

• Monash University, Australia

   In‐kind support

• Western Sydney University, Australia

   In‐kind support

• RMIT University, Australia

   In‐kind support

External sources

• Australian Department of Health and Ageing, Australia

Declarations of interest

Meaghan Coyle is a researcher in a health profession (Chinese medicine), but does not see patients as a health professional, and reports no other conflicts of interest. 

Caroline Smith is an investigator of a pilot study, Do 2011, to examine the effect of moxibustion on breech presentation. Assessment of study eligibility, data extraction, and risk of bias assessment for this study were conducted independently by Meaghan Coyle and Brian Peat, who were not involved in the conduct of the trial. 

Brian Peat: reports no conflicts of interest. 

New search for studies and content updated (conclusions changed)