Exercise Training and Reproductive Outcomes in Women With Polycystic Ovary Syndrome

A Pilot Randomized Controlled Trial

Jamie L. Benham; Jane E. Booth; Bernard Corenblum; Steve Doucette; Christine M. Friedenreich; Doreen M. Rabi; Ronald J. Sigal


Clin Endocrinol. 2021;95(2):332-343. 

In This Article


We evaluated feasibility of a randomized trial evaluating HIIT and CAET versus control on reproductive outcomes. Recruitment and retention were adequate to conduct such a trial, but adherence to exercise and especially to OPK-testing, failed to meet pre-specified success criteria.

Exercise adherence has not been widely reported in PCOS exercise trials.[10] One trial[26] reported 60% adherence. Another[27] only included the 14/16 (87.5%) participants who completed ≥75% of prescribed exercise sessions in their analysis, but did not specifically report exercise adherence. In our trial, we found median adherence of 68% (IQR 53%, 86%), slightly below our pre-defined feasibility criterion of >70%. In both exercise groups, exercise adherence trended down over time, similar to other exercise intervention studies in women.[28,29] Exercise adherence is influenced by many participant factors including perceptions about exercise, perceived supportive behaviour and autonomous motivation.[30] Low exercise adherence in our trial could be secondary to barriers including lack of time, fear of injury, physical limitations or lack of confidence with respect to exercise as noted in other studies.[14,15] This study illustrates the importance of a feasibility study. Proposed lifestyle interventions for women with PCOS must be appropriate and easily integrated into their lives to encourage adherence.

HIIT and CAET were compared in a previous 8-week trial involving participants without PCOS with overweight or obesity; no differences were reported between training groups in adherence or exercise enjoyment.[31] In our study, we noted a negatively skewed distribution of adherence for participants randomized to CAET where half the participants completed >80% of the prescribed sessions, perhaps suggesting a subgroup of participants were adopters of CAET. Conversely, positive skew was observed for the HIIT group, especially during the final three months of the intervention. While exploratory, these different patterns suggest there may be differences between the training regimens affecting adherence, and may also reflect differences in participant characteristics within and between groups.

Reported attrition in previous PCOS exercise trials was 6%-54%, with most having attrition ≥25%.[15] This pilot trial's relatively low attrition demonstrates interest among women with PCOS in exercise research participation.[32] Unfortunately, low OPK-testing adherence caused inadequate sensitivity; ovulations could have been missed as the LH surge typically would not be detectable by the OPK test for longer than 24–48 h. OPKs have been used in research protocols involving women not actively trying to conceive;[33,34] however, OPK adherence was not reported. While OPKs are practical, easy to use and non-invasive in ovulation monitoring for family planning,[35] this trial suggests prolonged daily OPK use may not be sustainable in individuals not seeking conception. To capture ovulation accurately in a research setting, other techniques such as twice-monthly serum progesterone levels may be necessary.

It is unclear how menstrual regularity is affected by exercise in women with PCOS; some trials demonstrated improvements while others found no change.[10] One exercise trial[7] found participants with reproductive function improvements had significantly more weight loss than those who did not. We found participants from each group experienced improvements in menstrual cycle regularity, with no significant intergroup differences. The participants with increased menstrual cycle regularity had hyperandrogenism (PCOS phenotype A/B), and overweight or obesity; 50% lost weight during the intervention. This finding raises two questions that require further investigation: (1) is the response to exercise related to PCOS phenotype (are phenotypes with hyperandrogenism more responsive to exercise); and (2) does the reproductive response to exercise vary by baseline BMI and/or weight loss?

This study has several strengths including being the first randomized trial to compare effects of HIIT and CAET on reproductive outcomes. It included a three-month run-in-phase to determine baseline reproductive function including ovulations and menstrual cycle lengths in women with PCOS. We included normal-weight women as well as women with elevated BMIs making our study findings more generalizable than previous exercise trials in PCOS[7,27] that only included participants with elevated BMIs. We also included rigorous and direct measurements for the anthropometric, metabolic and hormonal outcomes.

Since this study was a pilot trial to evaluate feasibility and inform planning of future trials, the sample size was inadequate to evaluate efficacy of HIIT and CAET on reproductive health. In addition, low OPK-testing adherence precluded clear conclusions regarding ovulation rates. Finally, participants may have changed their dietary intake, which we did not assess and could not control for in the analyses that might have influenced these results.

In conclusion, while acceptable participant recruitment and relatively low attrition indicate women with PCOS are interested in exercise research participation, future trials should address barriers to exercise adherence, and a different, more feasible and sustainable ovulation testing method should be used. HIIT and CAET did not result in statistically significant improvements in reproductive health outcomes, specifically menstrual cycle and luteal phase length. HIIT and CAET were both effective at improving anthropometrics, insulin resistance and lipids in women with PCOS. Our findings indicate there may be differential anthropometric and cardiometabolic effects of exercise based on exercise type. Further, larger studies are needed to evaluate impacts of these two exercise interventions on reproductive, anthropometric and cardiometabolic outcomes as well as to explore associations of PCOS phenotype and BMI to exercise response, to inform evidence-based clinical practice guidelines for the management of PCOS in reproductive-aged women.