Immediately after randomization, there were 4 patients who conceived before the start of treatment. They completed the last visit and were included in the primary analysis.
The first live birth occurred 228 days after randomization, and the last live birth occurred 403 days after randomization.
eAppendix. Supplemental Methods
eFigure. Acupuncture Point Localization in the Active Acupuncture and Control Acupuncture Groups
eTable 1. Distribution of Number of Patients by Site and Treatment Arm
eTable 2. Acupuncture Points, Stimulation, Localization, Tissue in Which Needles Are Inserted, and Innervation Area in Active Acupuncture, Which Consists of Two Protocols That Are Alternated Every Other Treatment, and Control Acupuncture
eTable 3. Secondary Reproductive and Metabolic Outcome Changes From Baseline to Last Visit Measures
eTable 4. Secondary Quality-of-Life Outcomes Score Changes From Baseline to Last Visit Measures
eTable 5. Exploratory Outcomes With Regard to Fecundity
Customize your JAMA Network experience by selecting one or more topics from the list below.
Wu X, Stener-Victorin E, Kuang H, et al. Effect of Acupuncture and Clomiphene in Chinese Women With Polycystic Ovary Syndrome: A Randomized Clinical Trial. JAMA. 2017;317(24):2502–2514. doi:10.1001/jama.2017.7217
Does acupuncture alone or combined with clomiphene increase the likelihood of live births among women with polycystic ovary syndrome?
In this randomized clinical trial that recruited 1000 Chinese women with polycystic ovary syndrome, the live birth rate was significantly higher in the group of women who received clomiphene compared with placebo (28.7% vs 15.4%, respectively). However, it was not significantly different between the groups who received active vs control acupuncture (21.8% vs 22.4%, respectively), and there was no significant interaction between active acupuncture and clomiphene.
Acupuncture, alone or with clomiphene, was not effective as an infertility treatment in women with polycystic ovary syndrome.
Acupuncture is used to induce ovulation in some women with polycystic ovary syndrome, without supporting clinical evidence.
To assess whether active acupuncture, either alone or combined with clomiphene, increases the likelihood of live births among women with polycystic ovary syndrome.
Design, Setting, and Participants
A double-blind (clomiphene vs placebo), single-blind (active vs control acupuncture) factorial trial was conducted at 21 sites (27 hospitals) in mainland China between July 6, 2012, and November 18, 2014, with 10 months of pregnancy follow-up until October 7, 2015. Chinese women with polycystic ovary syndrome were randomized in a 1:1:1:1 ratio to 4 groups.
Active or control acupuncture administered twice a week for 30 minutes per treatment and clomiphene or placebo administered for 5 days per cycle, for up to 4 cycles. The active acupuncture group received deep needle insertion with combined manual and low-frequency electrical stimulation; the control acupuncture group received superficial needle insertion, no manual stimulation, and mock electricity.
Main Outcomes and Measures
The primary outcome was live birth. Secondary outcomes included adverse events.
Among the 1000 randomized women (mean [SD] age, 27.9 [3.3] years; mean [SD] body mass index, 24.2 [4.3]), 250 were randomized to each group; a total of 926 women (92.6%) completed the trial. Live births occurred in 69 of 235 women (29.4%) in the active acupuncture plus clomiphene group, 66 of 236 (28.0%) in the control acupuncture plus clomiphene group, 31 of 223 (13.9%) in the active acupuncture plus placebo group, and 39 of 232 (16.8%) in the control acupuncture plus placebo group. There was no significant interaction between active acupuncture and clomiphene (P = .39), so main effects were evaluated. The live birth rate was significantly higher in the women treated with clomiphene than with placebo (135 of 471 [28.7%] vs 70 of 455 [15.4%], respectively; difference, 13.3%; 95% CI, 8.0% to 18.5%) and not significantly different between women treated with active vs control acupuncture (100 of 458 [21.8%] vs 105 of 468 [22.4%], respectively; difference, −0.6%; 95% CI, −5.9% to 4.7%). Diarrhea and bruising were more common in patients receiving active acupuncture than control acupuncture (diarrhea: 25 of 500 [5.0%] vs 8 of 500 [1.6%], respectively; difference, 3.4%; 95% CI, 1.2% to 5.6%; bruising: 37 of 500 [7.4%] vs 9 of 500 [1.8%], respectively; difference, 5.6%; 95% CI, 3.0% to 8.2%).
Conclusions and Relevance
Among Chinese women with polycystic ovary syndrome, the use of acupuncture with or without clomiphene, compared with control acupuncture and placebo, did not increase live births. This finding does not support acupuncture as an infertility treatment in such women.
clinicaltrials.gov Identifier: NCT01573858
Polycystic ovary syndrome (PCOS), the most common cause of anovulatory infertility, is characterized by ovulatory dysfunction, hyperandrogenism, and polycystic ovaries.1 It is reported to affect 5% to 10% of women of reproductive age.2
Clomiphene citrate is a first-line, inexpensive treatment to induce ovulation in women with PCOS.3,4 However, clomiphene had a high failure rate of 23.4% without ovulation after 5 months of use, a relatively low cumulative live birth rate of 19.1% after 5 months, and a high multiple-pregnancy rate (7.4%) among 750 women with PCOS in 2014.4 Thus, new or adjuvant treatments would be desirable for this population.5 One such treatment is acupuncture, an integral part of traditional Chinese medicine, which has gained increased popularity.6 There are few studies indicating the prevalence of use of acupuncture among patients seeking infertility treatment. A 2010 prevalence study from the United States including 8 community and academic infertility practices reported that 29% of their patients had used a complementary and alternative medicine treatment for infertility and 22% had tried acupuncture.7 Clinical trials from different countries indicate that acupuncture may improve reproductive function.8-15 However, these acupuncture trials provide insufficient evidence to determine the effectiveness of acupuncture to induce ovulation or treat infertility in PCOS because of failure to report important clinical outcomes such as live birth and limitations in quality and precision.6 Methodological problems include small sample size and insufficient information on allocation concealment. Thus, there is need for a multicenter randomized clinical trial on the use of acupuncture in this condition.
The PCOS Acupuncture and Clomiphene Trial (PCOSAct) investigated the effects of acupuncture and clomiphene on live births among Chinese women with PCOS.
The PCOSAct was a randomized, multicenter, clinical trial undertaken at 21 sites (27 hospitals) in the National Clinical Trial Base of Chinese Medicine in Gynecology from mainland China. The PCOSAct was designed as a 2 × 2 factorial trial to examine the effects of active acupuncture (or control acupuncture) and clomiphene (or placebo) on live births. Details of the study design, rationale for the primary and secondary outcome measures, power analyses, and the statistical analysis plan have been previously published.16 The full trial protocol is available in Supplement 1. The institutional review boards at the local sites approved the protocol, and all patients together with their partners provided written informed consent before joining the study. The study was chaired by a multidisciplinary steering committee and overseen by an international data and safety monitoring board.
All patients fulfilled the diagnostic criteria for PCOS according to the modified Rotterdam criteria1,17: oligomenorrhea or amenorrhea, together with clinical or biochemical hyperandrogenism (modified Ferriman-Gallwey hirsutism score ≥5 in Chinese),18,19 polycystic ovaries, or both. Inclusion and exclusion criteria are summarized in the eAppendix in Supplement 2. Metabolic syndrome was defined by meeting any 3 of the following 5 criteria: (1) waist circumference greater than 88 cm; (2) triglycerides level greater than 150 mg/dL (to convert to millimoles per liter, multiply by 0.0113); (3) high-density lipoprotein cholesterol level lower than 50 mg/dL (to convert to millimoles per liter, multiply by 0.0259); (4) systolic blood pressure greater than 130 mm Hg or diastolic blood pressure greater than 85 mm Hg; and (5) fasting glucose level of 110 to 126 mg/dL (to convert to millimoles per liter, multiply by 0.0555).1 Body mass index (BMI; calculated as weight in kilograms divided by height in meters squared) was categorized as normal if less than 24, overweight if 24 or greater but less than 28, and obese if greater than 28, as defined in Chinese women.20 All couples agreed to have regular intercourse during the study period with the intent of conception, and patients were in good health without major medical disorders. Baseline laboratory, anthropometric, and clinical measurements including scoring of hirsutism, acne, and ultrasonography were performed after an overnight fast. All biochemical assays were performed in a core laboratory (eAppendix in Supplement 2).
Participants were randomly allocated in a ratio of 1:1:1:1 into 1 of 4 intervention groups: active acupuncture plus clomiphene, control acupuncture plus clomiphene, active acupuncture plus placebo, and control acupuncture plus placebo (Figure 1) by an interactive online computer program (ResMan Research Manager; http://www.medresman.org) in a central office. The randomization was stratified within 21 participating sites and defined with a random block size of 4 or 8 (concealed from the investigators) by the head of the data coordination committee (H.Z.) (eTable 1 in Supplement 2). The statisticians generated and validated the randomization scheme for the study using the Plan procedure in SAS version 9.3 software (SAS Institute, Inc). We preprinted 1000 labels with barcodes (250 per group) and affixed them to medication bottles. The bottles were distributed to sites and assigned to patients as they were enrolled. This approach ensured that the randomization was stratified by site and that there would be 250 patients per group. The clomiphene and placebo assignments were double-blinded, unknown to patients and study investigators except the data manager. The active and control acupuncture treatments were known only to acupuncturists and the data manager at each site.
For a more detailed description of the study interventions, see the eAppendix in Supplement 2.
The acupuncture treatment was standardized. All patients received acupuncture treatment for 30 minutes twice weekly, with a maximum of 32 acupuncture treatments.
In the active acupuncture protocol, 2 sets of acupuncture points were alternated every other treatment to minimize soreness at the needle placement (eTable 2 and eFigure in Supplement 2). Acupuncture points were located in abdominal muscles and leg muscles (with somatic innervation common to the autonomic innervation of the ovaries and the uterus20) and in the hands and head. When placed, all needles were stimulated by manual rotation until the tingling sensation called de qi was achieved. Needle sensation reflects activation of the afferent nerve fibers projecting to the central nervous system at the spinal and central levels. Needles placed in the hand and head were manually stimulated every 10 minutes. Needles in abdominal and leg muscles were manually rotated and then connected to an electrical stimulator and stimulated with low frequency.
In the control acupuncture protocol, 2 needles were inserted superficially to a depth of less than 5 mm, 1 in each shoulder and 1 in each upper arm at nonacupuncture points, and needles were not stimulated manually when inserted (eTable 2 and eFigure in Supplement 2). Thereafter, the 4 needles were attached to electrodes and the stimulator was turned on to mimic the active acupuncture but with zero intensity, ie, no electrical stimulation.
Patients started with an initial oral dose of 1 pill of clomiphene (50 mg) or placebo from days 3 to 7 of the menstrual cycle. The dosage of oral medication was increased by 1 pill in the absence of ovulation or maintained in the presence of ovulation. The maximum dosage of clomiphene or placebo did not exceed 150 mg per day or 750 mg per cycle.3 The treatment could be repeated for up to 4 cycles.
After the baseline visit, acupuncture treatment and clomiphene treatment were started on day 3 of a spontaneous menstrual period. In patients with irregular cycles without recent menses, withdrawal bleeding was induced by medroxyprogesterone acetate, 5 mg/d for 10 days. Patients were instructed to have regular intercourse every 2 to 3 days and were monitored weekly by urinary human chorionic gonadotropin tests and serum progesterone levels to document pregnancy and ovulation. All treatments were stopped upon a positive pregnancy test. If the patient did not conceive, all measurements were repeated after the last treatment on the third day of menstruation in an ovulatory cycle, or within 1 week after the last treatment in an anovulatory cycle. Once the patient was pregnant, treatments were stopped and the end-of-study visit was performed within 1 week. Pregnant patients were followed up by ultrasonography every second week until fetal heart motion was visible. Pregnant women were then referred to the obstetric unit and follow-up scans were performed at weeks 18 to 24, 32, and 36 or at the discretion of the obstetrician.21-23 Birth outcomes were obtained from obstetrical records.
The primary outcome was live birth at 20 weeks’ or later gestation during the study period.24 Prespecified secondary outcomes included ovulation, conception, pregnancy, pregnancy loss, multiple (twin or triplet) pregnancies, anthropometrics, hirsutism, acne, hormonal changes, quality-of-life scores,25,26 and adverse events. Treatment credibility was also a prespecified secondary outcome but is not reported in this article. Quality-of-life scores included the Polycystic Ovary Syndrome Health-Related Quality of Life Questionnaire (range of 1-7, with higher scores indicating better function), Chinese Quality of Life Instrument (range of 50-250, with higher scores indicating better function), Medical Outcomes Study 36-Item Short Form Health Survey (range of 0-100, with higher scores indicating better function), Zung Self-Rating Anxiety Scale (range of 25-100, with higher scores indicating worse anxiety), and Zung Self-Rating Depression Scale (range of 25-100, with higher scores indicating worse depression). Serious adverse events were defined as any event that was fatal or immediately life threatening; led to severe or permanent disability or required prolonged hospitalization; led to congenital anomalies; led to pregnancy loss after 12 weeks’ gestation; or was thought to be serious by the site investigators.22,23 Exploratory outcomes were singleton live births, twin live births, infant birth weight, sex ratio, pregnancy duration, singleton pregnancy, time to conception, pregnancy loss in the first trimester, pregnancy loss in the second or third trimester, biochemical pregnancy loss, and ectopic pregnancy.
Without strong preliminary data on live birth after acupuncture, we chose 10% as the minimal clinically detectable difference that was likely to change clinical practice. We used this minimal clinically detectable difference of 10% in an earlier study.4,27 Assuming a 25% live birth rate with both active interventions, a 15% live birth rate with 1 active and 1 control intervention, and a 5% live birth rate with both control interventions, 80% power at a significance level of .05, and a 10% dropout rate, 1000 women would need to be enrolled.
The study was designed to test 3 primary hypotheses comparing specific combinations of interventions (Supplement 1). However, a more standard approach to a factorial trial is to focus on the main effects of the 2 treatments and their interaction. That approach was more consistent with the hypothesized pattern of live birth rates used in the sample size calculation and offered greater power and precision. Therefore, although it departed from the prespecified statistical analysis plan, we performed logistic or linear regression analysis that included the main effects of the 2 interventions and the interaction between the 2 interventions. Patients who received their randomized treatment (or conceived before treatment) and completed the final visit were included in the primary analysis. Study site was included as a random effect.
Categorical variables were summarized with frequencies and percentages. Their distributions were assessed with Fisher exact test. Continuous variables were reported as means with standard deviations or medians with interquartile ranges. To simplify presentation, group means using the t test were presented because the large sample size ensured the robustness of the t test.
There were missing data in 74 women who withdrew and were excluded from the primary analysis. For missing data in the secondary outcomes, we specifically reported the actual sample size of each variable when it differed from the complete sample size in each intervention group. Kaplan-Meier curves3,4 were used to compare days from randomization to last live birth in the 4 groups and plotted as the cumulative incidence of live birth.
All analyses were conducted using SAS version 9.3 software (SAS Institute, Inc). Two-sided P < .05 indicated statistical significance.
A total of 4645 women with PCOS were screened for eligibility; of them, 3645 were ineligible for various reasons and 1000 eligible women were randomized (250 in each group) (Figure 1). Participants were recruited from July 6, 2012, to November 18, 2014. The last live birth was recorded on October 7, 2015. Dropouts at each stage and the number assessed for the primary end point are presented in Figure 1. There was no significant difference in the dropout rates among the groups. The final visit was attended by 926 women, who composed the analytic population.
The baseline characteristics of the participants are shown in Table 1. The mean (SD) age was 27.9 (3.3) years, and the mean (SD) BMI was 24.2 (4.3). In total, 176 of 998 participants (17.7%) were obese and 196 of 999 (19.6%) met metabolic syndrome criteria. Patients had high adherence to treatment, including 849 of 926 participants (91.7%) for clomiphene and placebo and 895 of 926 (96.7%) for active or control acupuncture.
There were 69 live births (29.4%) in 235 patients receiving active acupuncture plus clomiphene, 66 (28.0%) in 236 patients with control acupuncture plus clomiphene, 31 (13.9%) in 223 patients with active acupuncture plus placebo, and 39 (16.8%) in 232 patients with control acupuncture plus placebo (Figure 2). There was no significant interaction on live births between clomiphene and active acupuncture (P = .39); therefore, the main effects of clomiphene and acupuncture were examined. The live birth rate was significantly higher in the group of women treated with clomiphene than among those who received placebo (135 of 471 [28.7%] vs 70 of 455 [15.4%]; difference, 13.3%; 95% CI, 8.0% to 18.5%) but was not significantly different between the groups treated with active and control acupuncture (100 of 458 [21.8%] vs 105 of 468 [22.4%]; difference, −0.6%; 95% CI, −5.9% to 4.7%) (Table 2).
For the secondary outcomes, the interaction between the 2 interventions was not statistically significant except for the adverse event of back pain. Statistically significant results are discussed herein.
The rates of ovulation, conception, pregnancy, and multiple pregnancy were significantly different between patients treated with clomiphene vs placebo but not significantly different between those receiving active vs control acupuncture. Differences in the effects of clomiphene vs placebo (Table 2) were 32.1% (95% CI, 28.9% to 35.4%) in ovulation rate per cycle (1030 of 1561 [66.0%] vs 569 of 1681 [33.8%]), 23.3% (95% CI, 18.5% to 28.1%) in ovulation rate per woman (439 of 471 [93.2%] vs 318 of 455 [69.9%]), 22.1% (95% CI, 16.2% to 28.1%) in conception rate (214 of 471 [45.4%] vs 106 of 455 [23.3%]), and 14.3% (95% CI, 8.9% to 19.7%) in pregnancy rate (144 of 471 [30.6%] vs 74 of 455 [16.3%]).
The frequency of serious adverse events was very low and did not differ significantly among the groups (Table 3). There were 72 serious adverse events reported. Abnormal vaginal bleeding was less common in patients receiving clomiphene vs placebo (10 of 500 [2.0%] vs 47 of 500 [9.4%]; difference, −7.4%; 95% CI, −10.2% to −4.6%), but dysmenorrhea was more common in patients receiving clomiphene vs placebo (13 of 500 [2.6%] vs 3 of 500 [0.6%]; difference, 2.0%; 95% CI, 0.4% to 3.6%). Patients receiving active acupuncture compared with those receiving control acupuncture more commonly had bruising (37 of 500 [7.4%] vs 9 of 500 [1.8%]; difference, 5.6%; 95% CI, 3.0% to 8.2%) and diarrhea (25 of 500 [5.0%] vs 8 of 500 [1.6%]; difference, 3.4%; 95% CI, 1.2% to 5.6%).
During pregnancy, back pain was reported less frequently in patients receiving active acupuncture plus clomiphene than in those receiving control acupuncture plus clomiphene in the first trimester (1 of 108 [0.9%] vs 8 of 106 [7.5%]; difference, −6.6%; 95% CI, −12.0% to −1.3%) and less frequently in patients receiving active acupuncture plus placebo than in those receiving control acupuncture plus placebo in the second and third trimesters (2 of 51 [3.9%] vs 10 of 55 [18.2%]; difference, −14.3%; 95% CI, −25.8% to −2.8%). Clomiphene was more frequently associated with back pain in patients receiving active acupuncture plus clomiphene than in those receiving active acupuncture plus placebo in the second and third trimesters (22 of 108 [20.4%] vs 2 of 51 [3.9%]; difference, 16.4%; 95% CI, 7.2% to 25.7%). The most common adverse events were gestational diabetes mellitus and preterm labor. Two congenital anomalies were reported. Four newborn deaths were reported from 2 preterm twin deliveries in the group that received control acupuncture plus placebo.
Significantly greater increases in circulating levels of progesterone, total testosterone, estradiol, and sex hormone–binding globulin were found in women receiving clomiphene vs placebo, and these levels were not significantly different between women receiving active vs control acupuncture. The hormone results are presented in eTable 3 in Supplement 2.
There were no significant changes in quality of life among the 4 groups (eTable 4 in Supplement 2).
For the exploratory outcomes, the interaction between the 2 interventions was not statistically significant except for duration of pregnancy. Clomiphene was associated with higher rates of conception, single and twin pregnancies, and single and twin live births among women who ovulated compared with placebo. Active acupuncture compared with control acupuncture was not significantly associated with differences in rates of ovulation, conception, pregnancy, and twin pregnancy. Neither treatment was associated with newborn weight, sex ratio of boys to girls, or pregnancy loss among groups. Other exploratory outcomes are shown in eTable 5 in Supplement 2.
This trial found that clomiphene was superior to placebo for achieving live births among infertile women with PCOS and that active acupuncture provided no additional benefit over control acupuncture. Secondary outcomes of ovulation and pregnancy were more likely to occur after treatment with clomiphene than with placebo, but not with active acupuncture vs control acupuncture. These findings do not support acupuncture alone or combined with clomiphene as an infertility treatment in patients with PCOS.
A Cochrane review6 found insufficient evidence for active acupuncture compared with control acupuncture to treat anovulation and infertility in women with PCOS. Before this study, 2 trials demonstrated that acupuncture was effective for ovulation induction compared with no treatment or regularly meeting with a therapist.13,14 The only sham-controlled trial using a nonpenetrating sham was underpowered for investigating live births,8 but it found ovulation rates between active and control acupuncture similar to those observed in this study. That study used the nonpenetrating Streitberger placebo needle,28 whereas superficial needle placement was used as a control in this study. Both control situations involved fewer needles and placement away from the active acupuncture needle placement sites.
The appropriate control group in acupuncture trials is debated because the nonpenetrating and superficial acupuncture needle techniques are not completely inert.29 The ovulation rates in the active and control acupuncture groups in this trial were higher than rates in trials with a no-intervention control group.13,14 Thus, the acupuncture procedure, unrelated to needle placement or stimulation, may have placebo effects.
There was no statistically significant difference between active and control acupuncture in clinical or biochemical variables or serious adverse events. There were 2 congenital anomalies. The incidence was similar to recent trials evaluating the effectiveness of clomiphene, metformin, letrozole, or berberine for live births in women with PCOS.3,4,24 Although 4 neonatal deaths occurred, these were in the group receiving control treatments and were due to the loss of 2 preterm twin pregnancies. Patients receiving active acupuncture had a higher incidence of skin bruising and diarrhea compared with control acupuncture. However, back pain during pregnancy was less frequent in the active acupuncture groups.
Clomiphene is the first-line choice for ovulation induction for anovulatory infertility in PCOS.5 The ovulation rate per cycle with clomiphene seems not to be affected by ethnicity or race, although obesity has a negative effect. In previous studies, ovulation rates have been 49.0% in obese white women,4 56.2% in lean or overweight Indian women,30 and 59.0% in obese Malaysian women,31 compared with 66.0% in the Chinese women with a mean BMI of 24.2 in this study.
Pregnancy loss in the control acupuncture and clomiphene group was 34.9%. The pregnancy loss was higher than in other large multicenter trials in women with PCOS, ie, 25.8% in the PPCOS I study3 and 29.1% in the PPCOS II study.4,32,33 The reason may be increased detection of early conception with subsequent biochemical pregnancies or early pregnancy loss in this trial.
Strengths of this study include the factorial design, adequate power, similar withdrawal rates among groups, and high adherence to treatment.
This study had several limitations. One limitation was the fixed acupuncture protocol. Personalized acupuncture treatment might have been more effective. In traditional Chinese medicine, acupuncture is often combined with individualized herbal mixtures, which were prohibited in the protocol to avoid confounding. Another limitation is the deviation of the statistical approach with factorial analysis from the serial 2-way comparisons that were prespecified in the protocol.
Among Chinese women with PCOS, the use of acupuncture with or without clomiphene, compared with control acupuncture and placebo, did not increase live births. This finding does not support acupuncture as an infertility treatment in such women.
Corresponding Author: Xiao-Ke Wu, MD, PhD, Department of Obstetrics and Gynecology, First Affiliated Hospital, Heilongjiang University of Chinese Medicine, Harbin 150040, China (email@example.com).
Accepted for Publication: May 24, 2017.
Author Contributions: Dr X.-K. Wu had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis. Drs X.-K. Wu and Stener-Victorin share first authorship. Drs X.-K. Wu and H. Zhang contributed equally to this work.
Concept and design: X.-K. Wu, Stener-Victorin, T.-X. Wu, Liu, Ng, Legro, H. Zhang.
Acquisition, analysis, or interpretation of data: X.-K. Wu, Stener-Victorin, Kuang, H.-L. Ma, J.-S. Gao, Xie, Hou, Hu, Shao, Ge, J.-F. Zhang, Xue, Xu, Liang, H.-X. Ma, H.-W. Yang, Li, Huang, Sun, Hao, Du, Z.-W. Yang, X. Wang, Yan, Chen, Fu, Ding, Y.-Q. Gao, Zhou, C. C. Wang, Ng, Legro, H. Zhang.
Drafting of the manuscript: X.-K. Wu, Stener-Victorin, H.-L. Ma, H. Zhang.
Critical revision of the manuscript for important intellectual content: X.-K. Wu, Stener-Victorin, Kuang, H.-L. Ma, J.-S. Gao, Xie, Hou, Hu, Shao, Ge, J.-F. Zhang, Xue, Xu, Liang, H.-X. Ma, H.-W. Yang, Li, Huang, Sun, Hao, Du, Z.-W. Yang, X. Wang, Yan, Chen, Fu, Ding, Y.-Q. Gao, Zhou, C. C. Wang, T.-X. Wu, Liu, Ng, Legro, H. Zhang.
Statistical analysis: H.-L. Ma, Xie, H. Zhang.
Obtained funding: X.-K. Wu.
Administrative, technical, or material support: X.-K. Wu, Kuang, H.-L. Ma, J.-S. Gao, Hou, Hu, Shao, Ge, J.-F. Zhang, Xue, Xu, Liang, H.-X. Ma, H.-W. Yang, Li, Huang, Sun, Hao, Du, Z.-W. Yang, X. Wang, Yan, Chen, Fu, Ding, Y.-Q. Gao, Zhou, C. C. Wang, T.-X. Wu.
Supervision: X.-K. Wu, Stener-Victorin, Liu, Ng, Legro, H. Zhang.
Conflict of Interest Disclosures: All authors have completed and submitted the ICMJE Form for Disclosure of Potential Conflicts of Interest. Dr Legro reported serving as a consultant to Bayer, Kindex, Euroscreen, Millendo, Takeda, Clarus, Sprout, and AstraZeneca; serving as a site investigator and receiving grants from Ferring; and receiving grants from AstraZeneca. No other disclosures were reported.
Funding/Support: This work was supported by grants 201107005 and 200807002 from the National Public Welfare Projects for Chinese Medicine, JC200804 from the Heilongjiang Province Foundation for Outstanding Youths, 2011TD006 from the Intervention for PCOS Based on Traditional Chinese Medicine Theory–TianGui Disorder, and JDZX2012036 and 2015B009 from 2009 through 2016 at the First Affiliated Hospital, Heilongjiang University of Chinese Medicine from the National Clinical Trial Base in Chinese Medicine Special Projects, by the National Key Discipline of Chinese Medicine in Gynecology from 2009 through 2016, by the Heilongjiang Province “Longjiang Scholar” Program (Drs X.-K. Wu, Stener-Victorin, and Legro), and by the Chinese “Thousand Talents Plan” scholarship (Drs Legro and H. Zhang).
Role of the Funder/Sponsor: The funding agencies had no role in the design and conduct of the study; collection, management, analysis, and interpretation of the data; preparation, review, or approval of the manuscript; and decision to submit the manuscript for publication.
Group Information: The PCOSAct Study Group includes the authors plus the following individuals who contributed to data collection at local sites: Jin-Ying Fu, MD, Henan Province Hospital of Chinese Medicine, Zhengzhou, China; Chang-Ling Zhu, MD, Wenzhou Zhongshan Hospital, Wenzhou, China; Xiao-Hong Wang, MD, Xuzhou City Hospital of Chinese Medicine, Xuzhou, China; Yang Xia, MD, First Teaching Hospital of Tianjin University of Traditional Chinese Medicine, Tianjin, China; Yan-Qiu Du, MD, Daqing Longnan Hospital, Daqing, China; and Hui-Fang Cong, MD, Second Affiliated Hospital, Heilongjiang University of Chinese Medicine, Harbin, China. The following individuals contributed to administrative resource supports: Song-Jiang Liu, MD, and Gui-Yuan Wang, MSc, First Affiliated Hospital, Heilongjiang University of Chinese Medicine, Harbin, China; Shu-Lai Li, MD, Tanggu District Maternal and Children’s Hospital, Tianjin, China; Ke-Qiu Zhang, MD, Xuzhou Maternity and Children’s Hospital, Xuzhou, China; and Jian-Hua Shen, MD, Liwan District Hospital of Chinese Medicine, Guangzhou, China.
Additional Contributions: Data and safety monitoring board members of this trial included Robert W. Rebar, MD (chair), American Society for Reproductive Medicine, Birmingham, Alabama; Esther Eisenberg, MD, Vanderbilt University Medical Center, Nashville, Tennessee; Wei-Liang Weng, MD, China Academy of Traditional Chinese Medicine, Beijing, China; Su-Lun Sun, Beijing University of Chinese Medicine, Beijing, China; Wei Zou, MD, Heilongjiang University of Chinese Medicine, Harbin, China; and Zi-Dan Chen, MSc, China University of Mining and Technology, Beijing, China. They ensured the safety of study patients and the integrity of the research data independently; they received consulting fees. We thank the Reproductive Medicine Network Steering Committee of the National Institutes of Health for sharing the protocol and case report forms from the Pregnancy in Polycystic Ovary Syndrome II study.