Long-Term Recreational Gymnastics, Estrogen Use, and Selected Risk Factors for Osteoporotic Fractures



The purpose of this cross-sectional study was to examine whether long-term participation in recreational gymnastics or folk dancing or estrogen replacement therapy (ERT) is associated with mechanically more competent bones and improved muscular strength and body balance. One hundred and seventeen healthy, female postmenopausal recreational gymnasts (mean age 62.1 [SD 4.7] years) and 116 sedentary controls (mean age 61.5 [4.6] years) were enrolled in the study. Bone mineral content (BMC) of the distal radius, femoral neck, and trochanter were measured with dual-energy X-ray absorptiometry. BMC of the midshaft and distal tibia and trabecular density (TrD) of the distal tibia were measured with peripheral computed quantitative tomography. Maximal isometric strength, muscular power, cardiorespiratory fitness, and body balance of the participants were also assessed. The cardiorespiratory fitness, muscular strength, and dynamic balance of the recreational gymnasts and folk dancers combined were significantly better than those of the controls, the average group difference ranging from 7.5% (95% confidence interval 5.0–9.9%) in dynamic balance to 12.8% (6.6–19.4%) in dynamic muscular power. ERT was not associated with the fitness indicators, muscular power, or balance, but was significantly associated with the BMC at all the measured bone sites, the mean group difference between estrogen users and nonusers ranging from 6.5% (3.7–9.3%) for the tibial shaft to 11.8% (6.4–17.0%) for the distal radius. Recreational gymnastics, in turn, was significantly associated with higher BMC at the tibia only, the mean group difference being 3.9% (0.9–6.9%) for the tibial shaft and 7.7% (3.7–11.9%) for the distal tibia. Recreational gymnastics was also associated with higher TrD at the distal tibia (5.2%; 1.2–9.2%), whereas estrogen usage did not show such association. The results indicate that ERT seems especially effective in preventing postmenopausal bone loss, whereas recreational gymnastics and folk dancing improve muscular performance and body balance in addition to increased bone mass and bone size in the tibia. All these factors are essential in prevention of fall-related fractures of the elderly.


Osteoporosis-related fractures are a common cause of chronic disability and deaths among elderly women. They are clearly a multietiological disorder, and thus an improvement in one risk factor is unlikely to change the number and incidence of the fractures.(1) Bone mineral mass, geometry, and internal structure are determinants of bone strength and important indicators of fracture risk. Bone mass, for example, accounts for 80–90% of the strength variation of the proximal femur.(2) However, not only bone strength per se but also an individual's propensity to fall determines the overall risk of osteoporotic fractures. With advancing age, body balance and muscular performance deteriorate.(3) Consequently, an individual's daily physical activity is likely to decrease and lead to further impairment in functional capacity, thus initiating a vicious circle.

Commonly recommended means for preventing osteoporosis and related fractures are estrogen replacement therapy (ERT) and calcium and vitamin D supplementation.(4–6) Regarding the last option, bone mass at loaded skeletal sites has been shown to be clearly greater in trained athletes than in their sedentary controls,(7) whereas the benefit is smaller for subjects representing recreational activities, although these too are apparently important.(8–10) Body balance has also been shown to be better in physically active older persons than in their less active counterparts; and therefore it seems possible to prevent age-related deterioration of balance.(11) In addition, treatment including exercise seems to reduce the risk of falling among the elderly,(12) and recent randomized trials using high-impact exercise or intense strength training have not only shown increases in bone mass, but also in muscular power and strength and body balance in premenopausal and older women.(13,14)

The aforementioned findings indicate that exercise may indeed decrease the risk of age-related bone fractures by increasing bone mass and strength and by reducing the risk of falling. It is, however, obvious that exercise can become an essential part of a fracture prevention strategy only if the recommended activity is practiced on a broad scale and it is feasible and safe for the majority of the target population, including people of advanced age.

Traditional recreational gymnastics and folk dancing are very popular among elderly Finnish women, and they meet the aforementioned feasibility and safety criteria for physical activity. Therefore, we were interested in examining whether long-term participation in recreational gymnastics or folk dancing is associated with mechanically stronger bones in postmenopausal women and whether their muscular strength and balance are also better than those of their sedentary counterparts. In addition, the effect of ERT on these variables was evaluated, since estrogen use is common among postmenopausal Finnish women and its positive effect on bone mineral has been documented previously.(5,15,16)



Postmenopausal gymnasts and folk dancers were recruited from recreational gymnastic or folk dance clubs while sedentary controls were obtained via a local newspaper advertisement. All the women were healthy, did not smoke, and had a body mass index (BMI) of ≤30. With the exception of ERT, they were not allowed to use drugs affecting bone metabolism. None of the subjects had had severe bone fractures or gynecological operations that may have affected their skeletons.

The exercising women (both recreational gymnasts and folk dancers, commonly referred to as gymnasts in this report) had trained a minimum of 20 years in recreational gymnastics or folk dance clubs. Any who had competed in apparatus gymnastics were excluded, however. Women who participated no more than once a week in light or moderate exercise, that was not gymnastics or folk dancing, were accepted as controls.

All together 233 women (98 recreational gymnasts and 19 folk dancers and 116 controls) were included in the study. Fifteen of the folk dancers had participated in gymnastics, and only four were solely folk dancers. These 233 women were also divided into two subgroups according to their use of estrogen. Of both the gymnasts and referents, 54% used estrogen at the time of the study.

An independent ethics committee for clinical investigations approved the study protocol, and each subject gave her written informed consent before the measurements.

Description of the exercise

Traditional Finnish recreational female gymnastics includes rhythmic, aesthetic, and pliant body movements, the most important original purpose of the training being the promotion of physical, mental, and social well being. The exercise is light to moderate in intensity and emphasizes springy gait and body flexibility. Today it also includes more strenuous exercise causing elevation of heart rate and including muscular strength training, but it still does not have any high impact training as is common in standard aerobics. Finnish folk dancing is, in turn, aerobic dancing with brisk turns and light jumps that create low to moderate impact. The dancing is guided by traditional Finnish folk music and performed in groups.



Information on the participant's health, history of physical activity in childhood, education, occupation and work history, work load, number of children, duration of breast feeding, menstrual and menopausal status, life style (possible earlier smoking and use of alcohol), use of medication, and estrogen therapy was collected in an interview. Data on falls during the preceding year were collected in a separate questionnaire.

The subject's history in gymnastics or folk dancing was determined for 10-year periods from the age of 16–45 years, and for 5-year periods thereafter (average duration of one session, number of sessions a week, training months a year, and training years in each period). The controls answered the same questions in order to verify that they fulfilled the selection criteria of the study.

Information on the participant's total lifetime physical activity was obtained by a standardized questionnaire and an interview. Physical activity other than gymnastics or folk dancing was classified into four categories according to its type and frequency: “high” meant vigorous activity at least twice a week (at least 20 minutes per session causing enhanced breathing and clear elevation of heart rate), “moderate” indicated vigorous physical activity no more than once a week or less demanding activity a few times a week, “low” represented less demanding activity once a week or very light activity several times a week (causing no enhanced breathing and only light elevation in heart rate), and “no activity” was used when a participant reported no mentionable daily physical activity. Each subject's daily walking distance was also measured on 3 days (2 weekdays and a Sunday) with a pedometer (Fitty-3 Electronic, Uttenreucht, Germany). The lifetime occupational physical activity of the participants was assessed in an interview with regard to physical activity level during work, corresponding to multiples of the basal metabolic rate (MET).(17)

Calcium intake

Data for current daily calcium intake was assessed by a 7-day calcium intake diary(18) and calculated by validated Micro-Nutrica software (Social Insurance Institution, Helsinki, Finland). Information on milk consumption at different ages and use of calcium or other dietary supplements was obtained in an interview. If milk was not currently used, the age at and reason for cessation were recorded.

Current muscular strength, balance, and physical fitness

The maximal isometric strength of the leg extensors was measured by a strain gauge dynamometer,(19) and the grip strength of both forearms was determined with a standard grip strength meter. The leg-extensor power was evaluated with a vertical counter-movement jumping test, using a contact platform (Newtest, Oulu, Finland) and recording the flying time of the jump.(20) The height of the jump (h) was calculated from the flying time as follows: h = gt2/8, where g is 9.81 m/s2 and t is the flying time in seconds. Dynamic balance was tested by a figure 8 running test, the test being performed by running around two poles placed 10 m apart.(21) One leg standing with the eyes open was used to assess postural balance.(22) Cardiorespiratory fitness (estimated maximal oxygen uptake) was assessed by a standardized 2 km walking test.(23)

Bone densitometry

Bone mineral content (BMC, g) and areal bone mineral density (BMD, g/cm2) of the femoral neck on the dominant side, the trochanter area of the femur, and the distal radius were measured with dual-energy X-ray absorptiometry (DXA) (Norland XR-26; Norland Corp., Fort Atkinson, WI, U.S.A.).(24) The analyses of the BMC measurements were controlled for differences in subjects' heights so that for each subject the analyzed skeletal region represented the same height-adjusted anatomic area. The BMC data were also normalized by dividing the crude BMC value by the height of the region of interest. In our laboratory, the in vivo precision (coefficient of variation) of the normalized BMC measurements is about 1%, depending on the site of measurement.

In addition to the aforementioned DXA measurements, the most loaded bone in gymnastics, the tibia, was evaluated with peripheral quantitative computed tomography (pQCT) (XCT 3000; Norland/Stratec Pforzheim, Germany). The tomographic slices were taken from the midshaft and distal part of the right tibia.(25) For the tibial shaft, the BMC, cortical cross-sectional area (CoA, mm2), cortical density (CoD, g/cm3), and density-weighed section modulus (BSI, mm3) were determined. For the distal tibia, the evaluated parameters were BMC, total cross-sectional area (ToA), trabecular density (TrD, g/cm3), and BSI. In terms of bone's mechanical competence, the parameters pertaining to cross-sectional area (CoA and ToA) reflect the bone's ability to resist compressive loading, whereas the BSI denotes the bone mass distribution around the center of gravity of given bone section (the larger the outer dimensions of bone, the higher the BSI), and thus reflects the bone's ability to resist torsional loading. The TrD describes the volumetric apparent density of the trabecular tissue (i.e., the fraction of the trabecular bone volume occupied by bone mineral), but it should be noted that the TrD cannot distinguish the trabecular number, spacing, and thickness variables from each other within the given volume of the trabecular bone. In our laboratory, the in vivo precision of the pQCT parameters varies between 0.7% (CoD) and 3.5% (BSI) for the tibial shaft and between 0.9% (TrD) and 4.2% (BSI) for the distal tibia.(25)

Statistical analyses

Means and SDs were used as descriptive statistics. As the primary analysis, associations of the exercise and ERT with the bone and fitness variables were examined by an analysis of covariance in which gymnastics and the use of estrogen were the factor variables and body weight and age were the covariates. The percentage differences and their 95% confidence intervals (CIs) between the study groups were adjusted for the covariates. When the 95% CI did not include zero, the difference was regarded as statistically significant at α = 0.05.


The basic group characteristics are presented in Table 1. The controls were somewhat heavier than the gymnasts (69.0 [SD 10.2] kg vs. 66.2 [9.7] kg), and the estrogen users were somewhat younger than the nonusers (60.4 [4.1] years vs. 63.4 [4.8] years). There were no intergroup differences in calcium intake, use of alcohol, earlier smoking, age at menarche and menopause, number of pregnancies, or number of falls and fractures during the preceding year.

Table Table 1.. Means (SD) of the Basic Characteristics of the Study Groups
original image

Past and current physical activity

The gymnasts had clearly more leisure time physical activity than the controls. The years of high activity since 16 years of age (including the gymnastics) were 37.4 (10.0) and 4.7 (11.0), respectively.

For the gymnasts, the duration of the gymnastic or folk dance activity averaged 33.6 (11.7) years, the current mean number of exercise sessions being 1.9 times a week. The current training hours per year averaged 88, ranging from 77 to 97 between the age of 16 and the current age of the subjects. In addition to gymnastics, common types of moderate to high activities reported by the gymnasts were brisk walking, jogging, biking, swimming, dancing, skiing, and, to a smaller degree, racket games such as tennis and golf.

In the control group, 56 women had practiced gymnastics or folk dancing on average 3.2 (6.1) years. Forty-eight of them reported that they had done gymnastic exercises during a period of less than 10 years, mostly during young adulthood. Only 4 of these 48 women had started gymnastics rather recently (during their retirement period) but had exercised no more than a couple of years. The remaining eight controls had done gymnastics or folk dancing as a hobby for more than 10 years, but all of them had stopped this activity more than 20 years before the measurements. Seventeen additional controls reported that they had had short periods of moderate to vigorous endurance activities previously, mainly walking, skiing, biking, or swimming. The remaining 43 reported no leisure time exercise or only light outdoor activities causing no sweating (such as walking, biking, skiing, and gardening).

The measured current daily walking distance did not differ between the groups (Table 1), neither did the occupational work load index.

Current muscular strength, balance, and physical fitness

The current physical fitness, balance, and muscular strength of the participants and the age-adjusted effects of gymnastics, estrogen use, and their interaction on these variables are described in Table 2. None of the outcome variables showed a statistically significant interaction between physical activity and ERT, and therefore the results have been described separately in terms of gymnastics and estrogen.

Table Table 2.. Mean (SD) of the Fitness Variables and the P Values for Age-Adjusted Effects of Gymnastics (PG), Use of Estrogen (PE) and Their Interaction (PG*E)
original image

The gymnasts did better than their controls in all the fitness tests except grip strength (Fig. 1). The mean intergroup difference was 10% (95% CI, 5–15%) for isometric muscular strength of the leg extensors and 13% (7–20%) for leg extensor power as indicated by jumping capacity. The gymnasts were also able to keep their body balance better and had an 8% (5–10%) better time in the figure 8 running test. Fifty-one percent of the gymnasts were able to complete the 1-minute static balance test in contrast to 42% of the controls, the difference not being statistically significant, however. The mean intergroup difference in the estimated maximal oxygen uptake was 13% (7–18%) in favor of the gymnasts.

Figure FIG. 1..

Means and 95% confidence intervals of the age adjusted group differences (%) in the fitness tests between the gymnasts and their sedentary controls (open bars) and between the estrogen users and nonusers (solid bars). VO2max = maximal oxygen uptake.

The mean grip strength was 5% higher (1–9%) in the estrogen users than in the nonusers: otherwise there were no differences between the estrogen groups (Fig. 1). Forty-nine percent of the estrogen users were able to complete the 1-minute static balance test in contrast to 44% of the nonusers, the difference being nonsignificant.

Current bone status

The bone characteristics and the age and weight-adjusted effects of gymnastics and estrogen use and their interaction on bone variables are described in Table 3. None of the outcome variables showed an interaction between physical activity and ERT, and therefore the results are described separately in terms of gymnastics and estrogen.

Table Table 3.. Means (SD) of the Bone Characteristics and the P Values for the Age and Body Weight Adjusted Effects of Gymnastics (PG), Use of Estrogen (PE) and Their Interaction (PG*E)
original image

Gymnastics was associated with the tibial BMC only (Fig. 2). The mean difference was 3.9% (0.9–6.9%) at the tibial shaft and 7.7% (3.7–11.9%) at the distal tibia. In contrast, at all the measured bone sites the BMC values were higher for the estrogen users than the nonusers, the mean group difference ranging from 5.1% (1.2–8.8%) for the distal tibia to 11.8% (6.4–17.0%) for the distal radius (Fig. 2). The BMD results were similar to the BMC results (Table 3).

Figure FIG. 2..

Means and 95% confidence intervals of the age and weight adjusted group differences (%) in the bone mineral content between the gymnasts and their sedentary controls (open bars) and between the estrogen users and nonusers (solid bars).

The tibial shaft and distal tibia were larger in terms of cross-sectional area in the gymnasts than in controls, the mean difference being 3.7% (1.0–6.5%) for the tibial shaft CoA and 5.1% (0.9–9.4%) for the distal tibia ToA. The TrD of the distal tibia was on average 5.2% higher (1.2–9.2%) in the gymnasts than in the controls. The CoD values were similar between the groups. The association between the gymnastics and tibial BSI was not statistically significant for the distal tibia or tibial shaft (6.6% [–2.5 to 16.5%] and 1.9% [–1.5 to 5.4%], respectively).

When the estrogen users were compared with the nonusers, the mean difference between the groups was statistically significant for the tibial shaft CoA (4.2% [1.5–6.8%]) and CoD (2.5% [1.8–3.1%]). In contrast to the association with gymnastics, the difference in the BSI was statistically significant for both the tibial shaft and distal tibia, the mean differences being 4.0% (0.5–7.3%) and 22.7% (15.1–29.5%), respectively, in favor of the estrogen users.


This study showed that long-term recreational gymnastics or folk dancing was associated with improved body balance and improved muscle strength and power. This type of exercise was also significantly associated with high bone mass, trabecular density, and bone area at the tibia, the most loaded bone. ERT was not associated with the fitness indicators, dynamic muscle power, or body balance, but it was significantly associated with the higher bone mass in general.

The findings of improved body balance and improved muscle strength and power are in agreement with those of earlier studies.(12,14,26) One leg static standing test describes general balance while the figure 8 running test describes more appropriately the dynamic balance and agility needed in physical activity. Jumping test, as a measure of dynamic power of leg extension, complements the information regarding the dynamic balance. In particular, better dynamic balance may translate to decreased propensity to falls and fractures among elderly people. However, in this respect the evidence for the value of resistance, endurance, or flexibility training is still defective.(11)

The literature provides some evidence that estrogen replacement therapy may also be beneficial to the postural balance of elderly women.(27) However, in our study the estrogen replacement therapy was not associated with significantly improved balance or muscular strength, the grip strength excluded (Fig. 1). This finding agrees with the results of many previous studies, in which estrogen replacement therapy was not found to considerably increase muscle strength,(28–30) nor did it improve balance or reduce the number of falls(28) among middle-aged or older women. Our observation that the grip strength (but not the other strength indices) of the estrogen users was somewhat better than that of the nonusers is noticeable, however, although there is no conclusive explanation for this. Despite the above noted studies,(28–30) literature provides data that withdrawal of estrogen at menopause leads to a decreased muscle mass, a change that may be prevented by estrogen replacement therapy.(31) Consequently, it is possible that in our study the maintained estrogen levels by the estrogen treatment may have prevented or reduced the rate of decrease in muscle strength, and thus, maintained the muscle strength better in estrogen users than nonusers. To support this argument, in addition to the significant association in the grip strength, there was some, although nonsignificant, trend for better values in estrogen users in the other strength variables, too (Table 2).

In our study, gymnasts' bone mass was generally higher than the controls', but the group difference was significant for the tibia only (Fig. 2). This finding is in agreement with the existing knowledge that a training effect is usually seen at the loaded bone sites only (9,32) and that both gymnastics and folk dancing especially load the lower limbs. This activity of our participants was moderate in intensity and did not include high impacts on bones, although it had lasted for decades before the measurements. The controls, in turn, were not involved in these kinds of activities. However, healthy Finnish women of this age often do low-intensity household chores and take part in nonsporting recreational activities. Our subjects were probably no exception, and therefore the true group difference in the actual daily loading of bones was smaller than could be anticipated by the questionnaire evaluation of physical activity. This possibility could account for the observed small intergroup differences in the bone characteristics despite clear differences in the recorded physical activity and fitness.

Our findings also speak in favor of the explanation that healthy bones are well adapted to normal daily activity and associated loading, and the need for further improvement in bone mass or strength is marginal. In other studies the bone mass of physically active premenopausal women has also been similar to that of sedentary controls independent of whether the bones were loaded at work (carrying mail or newspapers)(33) or during leisure time sports and recreational activities.(8,9) Therefore, if the goal is to increase bone mass by physical activity, more strenuous skeletal loading is likely to be needed.

The positive effect of estrogen replacement therapy on bone mineral has been shown in many studies.(4–5,15,16) Also in this study, the bone mass of the estrogen users was systematically greater than that of the nonusers (Fig. 2), and, in general, estrogen usage was associated with better bone characteristics than recreational gymnastics was (Table 3). The only exception was the gymnasts' higher tibial BMC, this finding especially arising from larger distal tibia with denser trabecular tissue. The estrogen use was associated with a 5–12% benefit in BMC. In epidemiological studies this amount of BMC increase has been shown to decrease the risk of osteoporotic fractures considerably (about 50%), and, in fact, such fracture risk reductions have been seen in estrogen users vs. nonusers. (2,34) The different bone mass distribution in the estrogen users' tibiae, as indicated by significantly higher BSI values both at tibial shaft and distal tibia (Table 3), suggests beneficial effects of estrogen not only on BMC but also on bone structure.

Both mechanical loading and estrogen increase bone turnover, and in theory it is possible that estrogen can even add to the bone's osteogenic response to loading. If so, exercise regimens designed to increase bone mass should be more bone-loading in nature to achieve the same result in the absence of estrogen than when it is present.(35) In our study, the bones of the estrogen users had higher mineral mass irrespective of physical activity. Probably, the magnitude of the customary loading between the groups did not differ enough in terms of osteogenicity of the loading. In some studies (36,37) exercise has been found to enhance the known bone-conserving effect of estrogens, but, in general, the magnitude of the possible synergistic effects of exercise and estrogen on bone tissue have remained unknown. The same is true regarding the effects of estrogen and exercise on cardiorespiratory fitness and muscular performance. In agreement with some earlier studies(36,37) we found an indication of additive (Table 3), but not synergistic, effects of estrogen and exercise on these parameters. Additional, well controlled intervention is needed to determine whether a combination of exercise and estrogen has synergistic effects on bone tissue, various aspects of physical performance and body balance, and, if so, to what extent.

In summary, long-term regular participation in traditional Finnish female gymnastics and folk dancing was associated with improved bone variables at the most loaded distal tibia, and improved cardiovascular fitness, muscular strength and body balance in postmenopausal women. Although the tibia is not the most common bone site for osteoporotic fractures, this result supports the concept of beneficial effects of exercise on bone tissue, and opens possibilities to develop an exercise program which would be effective in preventing age-related bone loss in other skeletal sites as well. Estrogen replacement therapy, in turn, was not associated with fitness or balance but showed a clear positive association with all bone mineral content values and estimated bone strength. Thus, regarding the risk factors of osteoporotic fractures, estrogen replacement therapy seems especially beneficial to bone, while long-term recreational gymnastics to muscular performance and body balance. All these factors together are essential for prevention of osteoporotic fractures.


The authors thank the Juho Vainio Foundation, Helsinki, Finland for the financial support.