The results showed that both interventions improved the SPPB score at 12 weeks in prefrail community-dwelling older adults but did not lead to statistically significant residual effects when the training is followed by a period of detraining. After stopping regular exercising, the SPPB declined continuously over time in the muscle strength training group and was at baseline level at 36 weeks follow-up. In the PT group however, the immediate intervention effect did not diminish at follow-ups. The lack of significance of these effects might be explained with the low sample size in this group at 24 and 36 weeks and thus, with a low statistical power (resulting from a high drop-out rate). However, the absolute difference between baseline SPPB and both follow-ups can be considered clinically relevant in this population since it has been shown that a substantial meaningful change in the SPPB score ranges between 0.99 and 1.34 points
. This implies that muscle power training might be more beneficial than traditional muscle strength training for residual effects after exercise stop. It can be also suggested that an intervention period of three months could be sufficient for persisting functional improvements in prefrail older adults.
Influences of different training intensities on detraining effects in older adults with an increased fall risk have previously been shown by Hauer et al.
. They investigated the effects of a high-intensity training program comprising muscle strength, functional and balance exercises in comparison to a low-intensity exercise program in community-dwelling geriatric patients with a history of injurious falls. Although the immediate high-intensity training effects declined with increasing time of detraining, the differences between the groups in most functional performances were still significant two years later.
When looking at the individual components of the SPPB, it has been shown, that the finding in the the main outcome is predominantly reflected by the balance and chair rise scores whereas the gait score remained nearly unchanged over time. The immediate effects in the balance and chair rise components can be explained with the combined use of resistance as well as balance exercises in this study. Interestingly, ongoing effects in the PT group were not only shown in the muscle strength (chair rise) but also in the balance component. This implies that the functional improvement following muscle power training was also induced by an increased standing stability. In a meta-analysis, Steib and colleagues
 compared the effects of muscle power training and progressive muscle strength training in adults aged 65 years or more. Based on their findings, muscle power training is suggested to be more effective for enhancing functional performance in older adults than progressive muscle strength training. They reported that chair rise and stair climbing abilities improved more with muscle power training. Despite the similarity in effectiveness of both interventions after 12 weeks in our study, the present findings on the sustainability of SPPB improvements seem to support the superior effects of muscle power training shown by Steib et al.
Mixed data were shown in previous studies on effects of longer periods without exercising after muscle strength or muscle power training. Henwood and Taaffe
 reported a decline in dynamic and isometric muscle strength as well as muscle power following detraining in healthy older adults, which were previously involved in regular muscle strength and muscle power training programs. They also found residual functional ability effects which were comparable between both training modalities. Another trial
 reported no influence of different muscle strength training intensities (light, moderate and high) on detraining-induced changes in muscle strength in healthy older adults. This implies that in healthy older adults, in relation to muscle strength the training modality or intensity may have no influence on potential detraining effects. However, significantly different detraining adaptations in peak and mean muscle power were shown between low-intensity and high-intensity muscle strength training groups for inactive older men
. Following 8 months of detraining in the low-intensity training group, muscle power declined whereas at the same follow-up measurement there were still significant training effects in the high-intensity training group. Consequently, it is suggested that changes due to longer periods without regular training following previous exercise interventions may depend on the physical precondition or activity status of the study population.
Contrary to other studies in this field however, no effects were found on STS muscle power as well as aLM. Especially the lack of muscle power effects by using the sit-to-stand transfer test in this study does not reflect the shown changes in the chair rise component of the SPPB. For the STS muscle power test, the participants were asked to rise as fast as possible from a chair into a standing position and to stand as still as possible for five seconds on a force plate whereas the SPPB included 5 quickly repeated chair stands. In both tests, the time needed to complete the task is used for data analysis. Thus, it might be possible that the chair rise test of the SPPB with 5 repeated stands is more sensitive to training-induced changes than the STS transfer test with only one stand. Furthermore, it is also likely that the prefrail participants in this study might have been responded differently to training than other study populations.
The absence of significant effects on aLM reflects the findings of previously published studies that reported only small or no training-induced changes of lean body mass in older adults
[26–28]. This can be most probably explained with neurological adaptations resulting in an increased voluntary activation
 as well as rapid motor unit activation and higher firing rates
There are several issues that might have limited the generalizability of our findings and should be considered for practical use. In this study, we used a resistance-training machine with elastic bands. Resistance was adjusted by increasing the tension of pulling forces of the elastic bands. Besides the chair rise, all lower extremity exercises were performed in a standing position. Often, the participants performed single-leg exercises during which they had to stabilize on the non-exercising leg. The increased instability during exercising might have led to inadequate perceptions of exertion and thus, inadequate muscle strength and muscle power training intensities. It is also likely that the high-velocity contractions in the PT group were more challenging for keeping balance than the low-velocity muscle strength exercises. This would also explain the predominant changes in the balance component of the SPPB in the PT group.
Furthermore, during the intervention period the drop-out rate was more than twice as high in the muscle power training group than in the muscle strength training group. In two participants in the muscle power training group, the drop-out was directly related to the exercise intervention. This indicates that the physical demands and perceived exertion of muscle power exercises are exceptionally high and may increase the risk for adverse events in the prefrail population. The explicit use of prefrail community-dwelling older adults in this study is another issue concerning the limited generalizability of our findings. Studies regarding residual effects of muscle power or muscle strength training reported mixed findings
[11–13] suggesting that the physical performance status may play a major role in the duration of physical adaptations during detraining.