*3.5. Subgroup Analysis of Moxibustion*

*3.4. Sensitivity analysis: excluding four trials with a high risk of bias*  Limiting the meta-analysis to the 12 trials with moderate to low risk of bias [10,11,15– 17,19,21,24,26–29] which investigated the effects of acupuncture-type interventions in-Moxibustion is effective in the Asian population (RR = 1.42; 95% CI = 1.21–1.67; random effect model, I<sup>2</sup> = 71%) and in the non-Asian population (RR = 1.20; 95% CI = 1.01–1.43; random effect model, I<sup>2</sup> = 0%) (Figure 6).

cluding moxibustion, acupuncture, and moxibustion plus acupuncture reveal significant

#### effects on correcting fetal malposition (RR = 1.36; 95% CI = 1.23–1.51; random effect model, *3.6. Adverse Events*

I2 = 41%) (Figure 5). The sensitivity analysis for the moxibustion subgroup reveals a result like that of the previous analysis (RR = 1.34; 95% CI = 1.19–1.51; random effect model, I2 = 47%). The NNT is 7 (95% CI = 5–12). Only one trial that evaluated acupuncture versus control shows more cephalic presentation in the acupuncture group (RR = 1.68; 95% CI = 1.11–2.55). The NNT is 4 (95% CI = 2–20). Only one trial reports that moxibustion plus acupuncture had more cephalic presentation relative to control (RR = 1.42; 95% CI = 1.06– 1.90). The NNT is 7 (95% CI = 3–45). Information on adverse events was presented in four trials. Because of the clinical heterogeneity between the included studies, we did not perform a meta-analysis of adverse events. Cardini et al. in 2005 reported adverse events (41.5%) related to moxibustion [25]. Patients had abdominal pain, throat problems, and unpleasant odor with or without nausea. Cardini et al. in 1998 and Vas et al. reported that no adverse events occurred in the moxibustion or control groups [10,15]. Neri et al. observed no adverse effects on participants who received moxibustion plus acupuncture or usual care [26].


**Figure 5.** Sensitivity analysis: Acupuncture-type interventions versus Control; Outcome: Cephalic presentation. **Figure 5.** Sensitivity analysis: Acupuncture-type interventions versus Control; Outcome: Cephalic presentation.


**Figure 6.** Subgroup analysis: Moxibustion versus Control; Outcome: Cephalic presentation. **Figure 6.** Subgroup analysis: Moxibustion versus Control; Outcome: Cephalic presentation.

ticipants who received moxibustion plus acupuncture or usual care [26].

Information on adverse events was presented in four trials. Because of the clinical heterogeneity between the included studies, we did not perform a meta-analysis of adverse events. Cardini et al. in 2005 reported adverse events (41.5%) related to moxibustion [25]. Patients had abdominal pain, throat problems, and unpleasant odor with or without nausea. Cardini et al. in 1998 and Vas et al. reported that no adverse events occurred in the moxibustion or control groups [10,15]. Neri et al. observed no adverse effects on par-

We used Review Manager Software (Version 5.3.5) to evaluate the publication bias. The sample size of most studies was >100 participants with two comparison arms except for Do 2011 [28], Li 1996 [22], and Millereau 2009 [27]. Funnel plots are typically symmetrical for studies with large sample sizes (Figure 7). However, for studies with small sample sizes, no study reported a negative result, which suggests that publication bias is probable in the literature reporting correction of breech presentation with moxibustion and acu-

*3.6. Adverse events* 

*3.7. Publication Bias* 

puncture.
