Despite well-established benefits of exercise training and its recommendation as a cornerstone of PCOS management (36), few well-controlled studies have evaluated the benefits of exercise training in PCOS (Table 1). Prior to the prescription of an exercise regime, it is important to establish whether any functional limitations exist that may reduce an individual's ability to exercise. Three studies reported no difference in maximal aerobic capacity (VO2max), an indicator of functional capacity, between overweight women with PCOS and age- and body mass index (BMI)-matched controls (59–61). In contrast, Orio et al. (62) reported impaired maximal and submaximal cardiorespiratory responses in young overweight women with PCOS. The reason for discrepancy between these studies is not entirely clear, but may be related to varying degrees of IR. VO2max is inversely related to markers of IR (59,62), and in the study by Orio et al. fasting insulin concentrations were substantially higher than reported in other studies (20.2 vs. ∼15 IU L−1). Although IR is common in PCOS it is not universally present (63) because of numerous PCOS phenotypes. It is possible that varying degrees of IR between studies could explain discrepancies in the observed findings. Further studies are required that directly compare VO2max in normoinsulinemic and hyperinsulinemic women with PCOS to better understand the interrelationships.
Table 1. Intervention trials investigating the effect of exercise in overweight women with polycystic ovary syndrome
|Randeva et al. (69)||21||Progressive brisk walking programme (12 completers vs. 8 non-completers)||Single arm||6 months||35.5 ± 7.6||↑ Aerobic fitness|
– BMI, insulin, lipids
| ||– FAI|| |
|Aubuchon et al. (71)||32||Supervised exercise sessions and nutrition classes||Single arm||14 weeks||36.4 ± 5.3||↓ Weight, BMI, WC, HC||6 pregnancies (46% who desired)|| || |
|Brown et al. (70)||20||Supervised aerobic exercise programme (n = 8) vs. control (n = 12)||Randomized controlled trial||5 months||Exercise: 37.9 (9.4)‡|
Control: 31.3 (14.9)
|– Insulin, glucose, insulin resistance|
– Weight, BMI, WC, HC, BP
↑ Aerobic fitness
| ||– Testosterone|| |
|Vigorito et al. (75)||90||Structured aerobic exercise programme (n = 45) vs. untrained controls (n = 45)||Randomized controlled trial||3 months||Exercise: 29.3 ± 2.9|
Control: 29.4 ± 3.5
|↓ Insulin, insulin resistance|
↓ BMI, WC, WHR, BP
– blood lipids, glucose
↑ Aerobic fitness
|60% restored normal menstrual cyclicity||– SHBG|
|Palomba et al. (76)||40||Aerobic exercise programme (n = 20) vs. hypocaloric high protein diet (n = 20)||Non-randomized controlled trial (based on preference)||24 weeks||Exercise: 33.1 ± 1.3|
Diet: 33.2 ± 1.4
|↓ Weight (greater in diet group)|
↓ WC, insulin (greater in exercise group)
|Improved menstrual cycles and fertility|
↑ menses frequency, ovulation rate (higher in exercise)
↑ SHBG (greater in exercise group)
↓ FAI (greater in exercise group)
|Bruner et al. (77)||12||Nutritional counselling (n = 5) vs. nutritional counselling and combined aerobic-resistance exercise (n = 7)||Randomized controlled trial||12 weeks||Nutrition: 37.1 ± 5.6 Nutrition + exercise:|
36.2 ± 4.6
|↓ Fasting insulin|
↓ Body fat (exercise only)
| ||– SHBG|
|Thomson et al. (78)||94||Diet (n = 30) vs. diet and aerobic exercise (n = 31) vs. diet and combined aerobic-resistance exercise (n = 33)||Randomized controlled trial||20 weeks||36.1 ± 4.8||↓ Weight, WC|
↓ BP, lipids, glucose
↓ Insulin, HOMA
↓ Body fat (greater in exercise groups)
|49% improved menstrual cyclicity and ovulation||↓ Testosterone|
|Thomson et al. (87)||57||Weight loss programme (diet with and without exercise)||Single arm study||10 weeks||36.2 ± 5.3||↑HRR|
↓ weight, WC
↓ BP, glucose,
↓ insulin, HOMA
| ||↓ Testosterone|
|Giallauria et al. (88)||124||Aerobic exercise programme (n = 62) vs. untrained controls (n = 62)||Non-randomized controlled trial (based on preference)||3 months||Exercise: 29.2 ± 2.9|
Control: 29.5 ± 3.5
|↓ BMI, WHR|
↓ Insulin and insulin resistance
↑ HRR, aerobic fitness
– Lipids, glucose, BP
| || || |
|Thomson et al. (92)||49||Diet (n = 14) vs. diet and aerobic exercise (n = 15) vs. diet and combined aerobic-resistance exercise (n = 20)||Randomized controlled trial||20 weeks||36.1 ± 4.8||↓ Weight|| || ||↓ Depression scores|
|Liao et al. (91)||23||Self-directed walking programme (12 completers vs. 11 non-completers)||Single arm||6 months||36.9 ± 7.3||– BMI|| || ||↓ Body image distress|
Muscle strength is a marker of exercise tolerance that predicts functional ability and participation in active living tasks (64,65). Only one known study has examined the effects of PCOS on muscle strength (59) and found no differences in isometric or isokinetic knee extensor strength between women with PCOS and healthy controls. However, the study was limited by a small sample size and minimal differences in the degree of IR between the groups. Previous research has shown an inverse relationship between muscular strength and IR in a number of populations, including overweight and obese sedentary women (66) and reduced muscle strength in patients with metabolic disturbances, including type 2 diabetes (67). Hence, further research in PCOS with higher levels of IR is necessary to confirm these previous findings. Wright et al. (68) reported no difference in free-living physical activity levels between women with PCOS and controls suggesting that PCOS may not impact negatively on the ability to tolerate exercise or participate in daily living tasks and the inclusion of exercise programmes in PCOS management should be as well tolerated as in other populations.
Effect of exercise on cardiovascular disease risk factors
In the first known study to investigate the effects of exercise in PCOS, Randeva et al. (69) showed that a 6-month walking programme (n = 12; 120–420 min per fortnight) increased VO2max and decreased waist–hip ratio (WHR), with no changes in BMI, fasting insulin, lipids or FAI compared with women that did not adhere to the programme (n = 9). Participants in the exercising group completed 80% of the target exercise volume, indicating that walking was an achievable form of exercise; however nine failed to take up or complete the programme. In a randomized controlled trial, Brown et al. (70) showed in 20 sedentary women with PCOS that a 5-month supervised aerobic exercise programme (∼228 min per week at 60% VO2max) had no effect on weight, waist circumference (WC) or fasting insulin; but a trend for improved area under the insulin curve (−26%). More recently, an uncontrolled study in 32 women with PCOS participating in supervised group exercise (goal of 200–300 min per week) and nutrition classes showed that 66% of women lost more than 3% of body weight and those that exercised two or more times per week reduced their WC (71). However, on average participants completed only 8.8 weeks of out of the desired 14 weeks, with personal time commitments cited as the most common reason for discontinuing (71). This suggests that higher exercise levels may not be sustainable for these women over the long term. Young women have high perceived time pressure because of work and family commitments, which can be a barrier to healthy eating and physical activity participation (72). This may result in the development of a perception that recommended behavioural changes to achieve optimum lifestyle habits are not feasible (73). Promoting and encouraging activities that fit within the context of young women's daily routines may overcome this barrier, together with providing a supportive environment (74).
More recently, larger studies have evaluated the effects of exercise in women with PCOS. In a randomized controlled study, Vigorito et al. (75) showed in 90 overweight women with PCOS that a 3-month structured aerobic exercise programme (40 min cycling at 60–70% VO2max, 3 times per week) improved fasting insulin (9.0%), BMI (4.5%) and WC (2.9%) but not blood lipid levels compared with non-exercising controls. In an unrandomized, controlled study of 40 obese women with PCOS, Palomba et al. (76) compared 24 weeks of aerobic exercise (30 min cycling, 3 times per week) with a hypocaloric, high-protein diet. Although both interventions reduced body weight, WC, fasting insulin, IR and improved the reproductive hormone profile, greater improvements in WC, SHBG, FAI and insulin levels were reported with exercise, despite greater weight loss in the dieting subjects (76). This suggests that exercise training may offer greater benefits for improving IR and reproductive hormones compared with diet induced weight loss. However, treatments were self-selected in an unrandomized manner, which may have biased the results and limits inferences that can be drawn.
While the study by Palomba et al. (76) assessed the relative effects of diet and exercise, whether there was any additive exercise benefit when combined with energy restriction was not determined. A small, randomized study by Bruner et al. (77) showed that 12 weeks of nutritional counselling alone (n = 5) or combined with aerobic-resistance exercise (n = 7; 3 sessions per week of 30 min walking/cycling and 12 resistance exercises) both reduced fasting insulin levels, with the addition of exercise promoting greater reductions in body fat. More recently, we showed that three different 20-week lifestyle modification treatments, diet only (n = 30; 5000–6000 kJ d−1), diet and aerobic exercise (n = 31; 30–45 min walking/jogging 5 times per week) and diet and combined aerobic-resistance exercise (n = 33; 30–45 min walking/jogging 3 times per week and resistance training twice per week) reduced body weight, blood pressure, triglycerides, cholesterol, glucose and insulin levels, with no differences between treatments (78). However, compared with diet alone, exercise provided more favourable effects on body composition with greater reductions in fat mass (∼45%) and preservation of fat-free mass (∼60%), with no difference between aerobic exercise and combined aerobic-resistance exercise (78). Other studies have demonstrated that preservation of fat-free mass assists in maintaining resting metabolic rate and long-term weight maintenance (44,54,79). Taken together, these data suggest that exercise training during energy restriction for weight loss may have important implications for long-term maintenance of weight loss in PCOS; however, longer-term studies are required to confirm this.
Overall, considerable diversity in the degree of changes in IR in response to exercise treatments between studies was observed. While the heterogeneous nature of exercise interventions, including differences in exercise type, duration, frequency and intensity and intervention length could potentially explain these differences, varying degrees of hyperinsulinemia across study populations could be a contributing factor. Because of the heterogeneous diagnostic criteria of PCOS, numerous phenotypes have been identified and a minority of women (∼16%) have normal insulin levels (1). It is therefore possible that while studies examining women with higher baseline values (∼20 IU L−1) showed greater improvements in fasting insulin compared with a non-exercising controls (75) or a diet group (76), in other studies the relatively low baseline insulin levels may have tempered the insulin response to exercise (77,78). In the study by Bruner et al. (77), while both treatment groups experienced reductions in fasting insulin, initial levels varied substantially between groups (33.6 vs. 16.8 IU L−1). Consequently, it is possible that any additional effect of exercise was masked by the lower baseline insulin level in that group. Similarly, the somewhat lower baseline insulin levels (∼16.1 IU L−1) evident in women examined by Thomson et al. (78) may explain the lack of any differences in improvements in insulin sensitivity between treatments; although the possibility that the potent effects of energy restriction may have overridden and masked any additional effects of exercise cannot be dismissed. These data suggest the importance of specifically examining the effects of exercise in various PCOS phenotypes, which may respond differently to interventions. It is also noteworthy that most studies to date have used indirect methods to assess IR (fasting insulin levels, homeostasis model assessment of insulin resistance) and visceral fat (WC, WHR) that may have influenced the results and the use of more sensitive methods (hyperinsulinemic-euglycaemic clamp or CT) would be appropriate for future studies.
Heart rate recovery (HRR), defined as the reduction in heart rate (HR) 1 min after a maximal exercise test, is an emerging powerful independent predictor of CVD and all cause mortality (80–82). HRR is inversely related to IR (83) and is impaired in young overweight women with PCOS (84–86). However, few studies have investigated the effect of lifestyle modification on HRR to determine whether it is a modifiable risk factor. We recently showed that HRR was improved in overweight and obese women with PCOS following a 10-week weight loss programme incorporating diet with and without exercise (87). The improvement in HRR was related to weight loss and reductions in WC, with no independent effect of exercise (87). In a non-randomized controlled study, Giallauria et al. (88) reported significant improvements in HRR, BMI, WHR, fasting insulin, IR and aerobic fitness following a 3-month exercise training programme consisting of 30 min cycling at 60–70% VO2max, three times per week compared with untrained controls (exercise = 62; control = 62). The researchers suggested that the exercise-induced improvement in HRR was mediated by reductions in BMI and improvements in insulin sensitivity, rather than increased fitness (88). From these studies it is evident that HRR is impaired in women with PCOS and lifestyle management can improve HRR in this population, which is likely mediated by reductions in body weight and IR.
Effect of exercise on reproductive function
Several studies have examined the health effects of exercise as part of general lifestyle modification programmes (20,30,32); however few studies have investigated specific effects of exercise training in PCOS on reproductive outcomes. A preliminary uncontrolled study by Aubuchon et al. (71) reported that six women became pregnant during the 14-week study or within 3 months of completion, which was 46% of women desiring pregnancy. There were no observed differences in body composition changes or eating behaviours between pregnancy occurrence, and no other measures were taken (71). Vigorito et al. (75) showed in previously anovulatory women with PCOS that a 3-month aerobic training programme restored normal menstrual cyclicity in 60% (27/45) of women. Similarly, Palomba et al. (76) reported that 24 weeks of dieting or aerobic exercise improved menstrual cyclicity and ovulation in overweight women with PCOS, with no observed differences in quantity or cycle length between treatments. However, menses frequency and ovulation rates were higher in the exercising group with a trend for higher pregnancy rates and greater improvements in hormonal profile (76). Body weight, BMI, WC, fasting insulin, IR, testosterone, SHBG and FAI improved in those that ovulated, while no changes were observed in any parameters in women who remained anovulatory (76). The investigators hypothesized that improved insulin sensitivity was the primary factor involved in ovarian function restoration given that the exercise group, which experienced greater improvements in insulin sensitivity, experienced the greatest ovulation improvements. More recently, we reported that 49% of women with PCOS improved ovulation and/or menstrual cyclicity following 20-weeks of an energy-restricted diet alone or combined with aerobic exercise or aerobic-resistance exercise, with no difference between treatments (78). Importantly, while there were no differences between treatments in the number of menstrual cycles during the 20-week intervention period, women in the combined diet and exercise groups experienced more ovulatory cycles than the diet only group. Overall, exercise studies have shown improvements in menstrual cyclicity and/or ovulation in ∼50% of PCOS women, comparable to that observed with other weight loss interventions. Improvements in insulin and hormonal profile appear to play important mediating roles for improving reproductive function, although this is not exclusively the case. Further research is needed to explore the role of exercise and weight loss in improving reproductive function. Future research is also needed to investigate the effects of exercise during pregnancy as it may provide benefits such as limiting complications and preventing excess weight gain (89,90).
Effect of exercise on psychological well-being
Currently, only two studies have investigated the effect of exercise on psychological outcomes in PCOS. A small, non-randomized study in overweight and obese women with PCOS reported that a 6-month self-directed brisk walking programme improved body image distress scores (91). Recently, Thomson et al. (92) observed similar improvements in HRQOL and depression in overweight women with PCOS after 20 weeks following an energy-restricted diet with and without exercise (aerobic only or combined aerobic-resistance exercise). Separate studies also show improvements in self-esteem, depression and anxiety following lifestyle modification programmes aimed at improving fertility in obese infertile women that include information about increasing exercise (20,32,93). While these studies show promising results, further research is needed to evaluate the effects of exercise on numerous psychological outcomes in overweight women with PCOS.