There was a statistically significant difference in medically-attended injurious falls between the recreational gymnasts and controls, the former having about 30% less injurious falls during the 20 years’ follow-up. However, the 27% between-group difference in injured fallers did not reach statistical significance. Regarding the fractures and fallers with a fracture, there was no significant group difference either.
Previously, clinical trials have shown that resistance and balance training reduces the risk of falls, especially that of injurious falls [10, 11]. In their recent Cochrane-review, Sherrington et al. reported that exercise reduced the rate of falls by 23% on average. Most effective exercise programs primarily comprised balance and functional exercises, while programs including multiple exercise categories (typically balance and functional exercises, as well as resistance exercises) may be beneficial as well [13]. However, individual trials that directly compared higher with lower doses of similar exercise were too few for definitive conclusions [13]. However, in general exercise programs of a higher dose seem to have larger effects. Tai Chi had a greater impact on the rate of falls when classes were delivered twice rather than once per week[23] and Kemmler et al. [24]found greater effects on the rate of falls using a more intensive program delivered twice a week compared with a low intensity program delivered once a week. In their previous review Sherrington et al. concluded that although the literature does not provide a clear cut-off there is an indication for greater benefits from higher doses of exercise, and they recommended that exercise should be undertaken for at least 2 hours per week on an ongoing basis [25, 26].
Recreational gymnastics was executed at least twice a week. In most exercise trials the frequency was once or twice a week. Our results are in line with these findings. Although recreational gymnastics is rather light in general, it is a combination of strength, agility and balance training and executed at least twice a week. Most likely, this is sufficient to improve physical functioning.
Although the recreational gymnasts in our study had less injurious falls in general, they showed a trend for more upper limb fractures. While this may seem contradictory, this finding is not unique. Physical activity was associated with a reduced risk of hip fractures in the SOF-study, but the risk of wrist fractures was slightly increased among physically active older women [32]. In the Tromsø study, high rates of physical activity were related to a 50% increased relative risk of upper limb fractures [33]. Also, a long-term follow-up study of a Finnish cohort suggested more wrist fractures among physically active postmenopausal women compared to inactive women of similar age [34]. Frequent walking has been associated with increased risk of fractures among older adults [35], but active commuting among middle-aged women was associated with a lower wrist fracture risk [36]. A recent meta-analysis of RCTs examining exercise and fracture risk showed that exercise is generally related to reduced fracture risk [37].
As discussed previously [32, 33], some explanations for the increased risk of upper limb fractures in physically active older persons may be debated. Recreational gymnasts were more agile at baseline and six years later [19, 20], and this benefit may have been maintained in older age. Thus, when slipping or tripping, recreational gymnasts reacted fast enough to extend their arm to absorb the impact energy resulting in more wrist fractures in place of other injuries, such as hip fracture or head injury. Having a feeling of good performance may also predispose not only to higher exposure time to physical activity [36], but also to higher walking speed and more risky behavior. Moreover, when stumbling at faster gait speed the impact force on falling is likely greater and possibly sufficient to fracture the upper limb bones. In this study, recreational gymnasts were about 4 years younger than sedentary controls when sustaining the fall-induced fracture, which may support the notion of faster walking speed while falling. However, instead of avoiding moving outside or declining physical activity, older people should pay attention to safe walking e.g. by using practical footwear, walking poles and/or shoe grip spikes. It is equally important to identify fall risks at home and mitigate them.
The strengths of this study are the long follow-up time and comprehensive evaluation of participants’ specific history in recreational gymnastics (determined in 10-year periods from the age of 16 to 45 years, and 5-year periods thereafter including duration and number of sessions per week, training months and years). The eligibility of controls as being sedentary was verified by their answers to the same questions. In addition, injurious falls were verified from medical records, which increased the reliability of the data. The types of injuries and the treatment provided were well described in the records, including the most severe traumas with exact ICD-codes, but the location (e.g. outdoors/indoors) or circumstances of the fall were not always mentioned.
One limitation of the study is that while the recreational gymnasts had exercised an average of 33 years at baseline in 1997, we had no information about their physical activity, participation in gymnastics, physical performance or mobility status after the 6-year follow-up in 2003. It is possible that participants’ physical activity and performance have declined with aging, chronic diseases, other incident health problems have increased or social issues changed, but we have no measured data to support this phenomenon. Physical activity was not recorded systematically in the medical records, nor was physical performance or functioning. However, we decided not to invite the former participants to the 20-year follow-up measurements because it was expected that only the women in good physical condition would have been able to participate, and this would have biased the analyses.
Another limitation is that medical records were available only for the women living in City of Tampere, and who had contacted the public health care system due to injuries. It is possible that some women may have sought treatment from private health care services. However, in Finland, senior citizens after retirement no longer remain under the domain of occupational health care, but fall within the public health care service system. Practically all 52 women with no mention of an injurious fall in their medical records had contacted the public health care system for some other health reason. Some injurious falls were treated elsewhere when travelling, but the aftercare was carried out in the local health care center or hospital, and these cases could be counted. Also, because we did not have access to medical records of surrounding municipalities, we had to exclude the women living in these neighborhoods. In addition, records of 7 deceased women had already been permanently deleted including the cause of death. However, this missing information was not likely to alter the findings because women with and without available register data were similar in their baseline characteristics, and the proportion of missing data (23%) was relatively small.
We did not have information about falls which did not require contact with the health care system. Most likely, these occurrences were falls with no consequences, or they resulted in mild injuries causing no long-term harm, disability or pain, and therefore the person did not consider it necessary to visit a health center to see a physician. We had permission only to access fall-related injurious data, not to the other health data, e.g. other diseases or medication during the follow-up. Although all participants were relatively healthy at baseline, the possibility that within-group changes in health status were not similar cannot be ruled out. It is well known that higher amounts of physical activity and fitness are associated with better health status in general. Furthermore, at baseline about half of the participants used estrogen replacement therapy, but there was no difference in injurious fallers between previous estrogen users and non-users. Estrogen was not associated with physical functioning at baseline either, but was associated with greater bone mass [19, 20]. However, users of estrogen replacement therapy were equally divided into both groups.
Although the health benefits of regular physical activity are well established, poor exercise adherence and compliance with physical activity recommendations is very common in the elderly population [38]. On the other hand, long-term physical activity at a younger age predicts physical activity in old age [39]. Also, in intervention trials, adherence and compliance remain fairly good as long as the training is supervised, but after the intervention, the adherence tends to return to the baseline. The benefit of recreational gymnastics is that it is offered by several clubs in Finland with low semester costs and exercise sessions are held in residential areas, making the training easily accessible, feasible and safe to perform. Despite being light to moderate by intensity, long-term regular participation may compensate for the benefits of more intensive training, which often is discontinued after a short period. Further research is needed to understand and establish the impact of different exercise programs and fall prevention interventions on fall-related fractures and injuries. Also, novel methods are needed to support older adults to continue physical activity after supervised interventions and promote life-long exercise training habits for maintaining physical functioning, balance and agility in older age.