Study Type
Meta-analysis
Population
Women (18 or older) with recurrent implantation failure (RIF) in IVF–embryo transfer
Size
15 RCTs; 1029 participants (Abstract; Section 3.1; Table 1)
Who Was Studied
RIF meant no clinical pregnancy after at least 3 embryo transfers or at least 10 high-quality embryos, or a clinician’s or standard diagnosis such as being under 40 with no clinical pregnancy after at least 4 high-quality embryos over 3 or more cycles. All participants were 18 or older, with mean ages per arm of 28.42 to 35.68 years (Section 2.2; Table 1). All trials were done in China and published in Chinese from 2018 to 2025, with 15 to 61 participants per arm. Studies of people with severe mental disorders were excluded (Sections 2.2, 3.2; Table 1).
Main Findings
Acupuncture or moxibustion added to estrogen or estrogen–progesterone therapy was compared with the same hormone therapy alone (Table 1). Clinical pregnancy: 12 trials with 887 participants; 211/445 vs 114/442 events; RR 1.84 (95% CI 1.53–2.20), I² = 0% (Section 3.4.1; Fig. 3). Live birth: 3 trials with 242 participants; 53/122 vs 22/120 events; RR 2.39 (95% CI 1.59–3.58), I² = 0% (Section 3.4.2; Fig. 4). Endometrial thickness: 12 trials with 879 participants; MD 1.37 (95% CI 0.95–1.80), units not stated, high variation between studies (I² = 86%) (Section 3.4.3; Fig. 5). Also reported: type A endometrium RR 1.67 (95% CI 1.30–2.14), 6 trials, 433 participants, I² = 0% (Section 3.4.4; Fig. 6). Serum estradiol SMD 2.70 (95% CI 0.20–5.21), 4 trials, 264 participants, I² = 98%, with 1 of the 4 trials pointing the other way (Section 3.4.5; Fig. 7). Egger test P = .010 for clinical pregnancy, meaning possible publication bias (Section 3.6).
Limitations
Authors reported: all trials came from China and were published in Chinese, only Chinese and English papers were searched, and trial quality may not be adequate (Section 4.1).
Authors reported: endometrial thickness and estradiol varied widely between studies for reasons they could not explain, and the subgroups were too small to generalize from, so results should be read with caution (Section 4.1).
Lucky Vera observation: acupuncture was added on top of hormone therapy, with no sham or placebo arm and no description of blinding, so placebo and attention effects cannot be separated out. Harms were not reported.
Lucky Vera observation: publication bias was found for clinical pregnancy, but the corrected trim-and-fill estimate is not given (Section 3.6). I could not access the supplementary material (sensitivity analyses, funnel plots, Table S1).
What Does the Study Not Establish
It does not firmly show cause and effect. The trials were randomized, but with no sham control and unclear blinding, placebo or attention effects can’t be ruled out. It does not show that the results apply outside China or to other age groups (Section 4.1). It does not establish whether the changes in thickness and estradiol are clinically meaningful: units are unclear and importance is not established. It does not show which method or treatment length works best, and it gives no safety data.
Author’s Conclusion
The authors say acupuncture and moxibustion were effective for all five outcomes and give a theoretical basis for using them in RIF. They also say larger, multi-centre, rigorous trials are still needed and that the findings should be interpreted with caution (Sections 4.1, 5).
Chen J, Lyu Y, Cheng X, Zhang F. The impact of acupuncture and moxibustion treatment in individuals with recurrent implantation failures: A systematic review and meta-analysis. Medicine (Baltimore). 2025;104(51):e46587. doi:10.1097/MD.0000000000046587