‘Meta-analysis of randomised controlled trials (RCTs) involving acupuncture for labour pain shows acupuncture to be more effective than comparison treatments in analgesic consumption' but the overall evidence is still limited, calling for further investigations

Authors


Authors’ Reply

Sir,

We thank Citkovitz et al.1 and Shiflett & Schwartz2 for stimulating the debate on the conclusion of our recent systematic review of acupuncture for labour pain,3 and for challenging its conclusions. Drs Shiflett and Schwartz2 state that we ignored substantial evidence that acupuncture is effective in reducing labour pain. Simple vote-counting by Figure 3 of our meta-analysis would definitely result in a positive conclusion, nine out of 17 outcomes favouring acupuncture compared with various controls. Yes, women receiving acupuncture did require less meperidine and other analgesic methods compared with those on the conventional analgesia arm (based on three studies), as stated in the abstract and summary of the main findings in our discussion. During 30 minutes of electroacupuncture (EA) treatment, women reported significantly less pain than those receiving placebo EA or no treatment at all, but the effect was not maintained after EA was discontinued; acupuncture still reduced pain 30 minutes after the cessation of EA treatment in only one study.2 Therefore, women felt some pain relief whilst they were being given acupuncture, but the effect did not persist after the needles were removed.

Drs Shiflett and Schwartz2 also state that, if we had removed the study by Ziaei and Hajipour5 from our analysis of acupuncture versus no treatment (Figure 3C in our review), more impressive evidence would have emerged. In fact, all three EA trials showed that acupuncture was better than no treatment by 2 hours (1 hour after acupuncture: mean difference, −10.44; 95% CI [−19.09, −1.79]; P = 0.02; 2 hours after acupuncture: mean difference, −9.13; 95% CI [−14.79, −3.48]; P = 0.002). However, these analyses were associated with a considerable heterogeneity of 93 and 77%, respectively. The trial by Zhang6 appears to deviate from the other two studies, and it is of note that this study was assessed as having a lower risk of bias compared with the other two.4,7

Citkovitz et al.1 also criticised us for not mentioning that acupuncture may be a safe and potentially beneficial treatment option in terms of reducing the duration and augmentation of labour. In eight of 10 trials reporting any of such secondary outcomes, we could not find any notable effect of acupuncture; duration of labour was reduced in three studies, whereas five trials found no such difference; less oxytocin was given to the acupuncture group in three studies, but this was not the case in the other two trials. A recent, independent review on this topic also found inconclusive evidence, calling for further investigations.6

To conclude, we maintain that, based on the findings of our review, the evidence for acupuncture as a treatment for labour pain is limited and unconvincing. Given that our aim was ‘to critically evaluate the evidence’,3 we believe our conclusions to be correct and uncritical conclusions to be misleading.

Ancillary