Acupuncture and clomiphene for Chinese women with polycystic ovary syndrome (PCOSAct): statistical analysis approach with the revision and explanation

Abstract

Background: Polycystic ovary syndrome (PCOS) is the most common endocrinopathy of reproductive-aged women. Clomiphene is regarded as the first-line medical treatment for ovulation induction in PCOS patients and acupuncture is often used as an alternative and complementary treatment for fertility issues such as those associated with PCOS. The efficacy of acupuncture alone or combined with clomiphene still lacks strong supporting evidence. Factorial 2 × 2 designs can be used for the evaluations of two treatments within a single study, to test the main effects of acupuncture and clomiphene and their interactions.

Methods: PCOSAct was designed to test the effect of clomiphene and acupuncture by three two-group comparisons in the original protocol. However, the trial was designed as a standard factorial trial and the factorial analysis approach for analyzing the data that were actually obtained during the trial was found to be more appropriate and more powerful than the three two-group comparisons described in the original protocol, so the statistical analysis approach and different datasets of PCOSAct in the primary publication were accordingly changed.

Discussion: Although the statistical analysis approach used in the primary publication deviated from the statistical analysis planned in the study protocol, focusing on the main effects of the two interventions and their interactions was a more standard approach to a factorial trial and proved to be more suitable and consistent with the characteristics of the trial data. Statistically, the revision is more powerful and precise and should be more useful to the journal and the readers.

Trial registration: Chinese clinical trial registry, ChiCTR-TRC-12002081 . Registered on 20 March 2012. Clinicaltrials.gov, NCT01573858 . Registered on 4 April 2012.

Keywords: Factorial design; Infertility; Polycystic ovary syndrome; Statistical analysis approach.

Conflict of interest statement

Ethics approval and consent to participate

This trial was approved by the Ethics Committee of the First Affiliated Hospital, Heilongjiang University of Chinese Medicine and then was individually approved by local ethics committees in the participating sites. For detailed information, please see Table 1.

Consent for publication

Not applicable.

Competing interests

The authors declare that they have no competing interests.

Publisher’s Note

Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Figures

Fig. 1

The main effect analyses of clomiphene and acupuncture. a The active acupuncture + clomiphene and the control acupuncture + clomiphene groups (Group A and Group B) were combined and compared with the combination of the active acupuncture + placebo and control acupuncture + placebo groups (Group C and Group D) to test the main effect of clomiphene. b In case of the active acupuncture main effect, the active acupuncture + clomiphene and active acupuncture + placebo groups (Group A and Group C) were compared with the combination of the control acupuncture + clomiphene and control acupuncture + placebo groups (Group B and Group D)

Fig. 2

Revised CONSORT diagram in the PCOSAct. Adopted from JAMA 2017;317(24):2502–14