Among 80 community-dwelling older adults undertaking regular gym exercise in Melbourne, the prevalence of sarcopenia ranged from 3.8–11.3%, varying according to definition. This is similar to values reported in the general community of older Melbournians (11% according to EWGSOP1) . However, it is lower than that reported according to FNIH and EWGSOP1 in community-dwelling older adults in Amsterdam (7.9% and 31.9%, respectively) , hospitals in Italy (24% and 36%, respectively) , or people in aged care in China (31.4% and 32.5%, respectively)  or Australia (40.2% according to EWGSOP1) , suggesting an overall healthy, well-functioning group of individuals, with high self-reported QoL.
Due to changes to the algorithm from EWGSOP1 and lower cut-off points , sarcopenia prevalence in this sample was lower for EWGSOP2 than for EWGSOP1, which supports recent findings that the use of EWGSOP2 will potentially underestimate sarcopenia prevalence compared to EWGSOP1 . Consequently, public spending on sarcopenia would be greater if EWGSOP1 guidelines are followed. Given the association of low muscle mass and function with morbidity and mortality [30–32], underestimating the prevalence of sarcopenia will likely have more dire consequences and higher health care costs in the long run. SARC-F has poor sensitivity but high specificity for sarcopenia definitions including FNIH and EWGSOP1  and may detect severe cases . In our study, SARC-F predicted one case of sarcopenia according to FNIH and EWGSOP1, one case of pre-sarcopenia (EWGSOP1) and two cases of probable sarcopenia (EWGSOP2), but none were sarcopenia confirmed (EWGSOP2) or severe (EWGSOP1/EWGSOP2). HUR gym individuals had a significantly higher SARC-F score, implying that they felt they were less functional, thus at higher risk of sarcopenia than the conventional gym group.
Low HGS and GS or the combination of both are associated with higher functional disability levels in older adults . This study demonstrated low proportions of low HGS (8% and 9% according to FNIH and EWGSOP2, respectively) implying that overall the Melbourne cohort are at a low risk of physical disability based on their physical function. However, when applying EWGSOP1 criteria, more participants (31%) of the sample had a low HGS. Regarding physical performance, only 3% of participants had low GS and poor TUG performance but 34% participants had poor 400 mW performance suggesting that walking tests of longer distances may be more sensitive for detecting low physical performance in older adults. Poor 400 mW performance may imply that exercising only once a week is insufficient to provide significant protection above routine activities of daily living (ADL’s) in community-dwelling older adults. Certainly, it is below current exercise guidelines for resistance exercise . The Position Statement providing recommendations for healthy older adults and those with special considerations for frailty, sarcopenia or other chronic conditions incorporates a combination of resistance training, power and functional training, 2–3 times a week, 1–3 sets of 8–12 repetitions . Alternately, members of the gyms may have begun training due to identified weakness/issues with ADL’s, and thus sarcopenia prevalence may have been higher if it were measured prior to them beginning their gym exercise.
Since EWGSOP2 offers many measurement options for each sarcopenia component, discrepancies in prevalence estimates, depending on the option applied, are to be expected. Phu, Vogrin  demonstrate that sarcopenia prevalence varied depending on EWGSOP2 measures, stating that highest prevalence was reported when using CS for muscle strength and lowest when using TUG for physical performance. In this study, HGS was used for muscle strength. However, sarcopenia prevalence would be lower if it was assessed by CS, as poor CS performance was slightly less common than poor HGS. GS was used to assess physical performance. If TUG had been used, sarcopenia prevalence would also be lower. However, if 400 mW was assessed, it would be higher due to highest proportions of poor 400 mW performance across the sample. This demonstrates the inconsistency of sarcopenia prevalence assessment, even within the same definition. A universally accepted consensus on sarcopenia is necessary for consistent diagnosis and implementation in clinical settings .
Given that there were no significant differences between HUR and conventional gyms for sarcopenia prevalence, it could be inferred that exercise training at both types of gyms operated by Uniting AgeWell have similar effects on sarcopenia. However, associations between the components did differ between the gyms. Regarding muscle strength components, low CS time was weakly associated with higher SARC-F scores for HUR gym participants. Conversely, past research shows high CS time and low HGS are associated with low SARC-F scores . It is not immediate apparent why the CS associations are in the opposite direction in this study in the HUR group. Indeed, the expected (and stronger) association of high CS times with higher SARC-F scores was observed in the conventional gym group, although this did not reach statistical significance, most likely due to the lower number of participants in the conventional gym.
HGS has functional importance in ADL’s, such as opening containers, lifting weights, using tools or holding handrails when ascending stairs . However, no associations were observed for HGS in either gym group. Low GS and SPPB are significantly correlated with low SARC-F scores . Similarly, in our study, both low GS and SPPB were associated with poor SARC-F scores, irrespective of the gym setting. In contrast, despite no difference between the two groups, longer 400 mW times were associated with SARC-F in the conventional gym group only, perhaps suggesting that lower leg cardiorespiratory fitness is a focus of the conventional gym exercises. Indeed, the conventional gym included dynamic exercises (e.g., plyometrics using medicine balls and jumping), which may be more effective for improving cardiovascular fitness and endurance than training with resistance equipment alone.
Slow GS and TUG had consistent associations with poor HRQoL among both HUR and conventional gym participants. Lower SPPB scores (HUR group) and slower 400 mW time (conventional group) were associated with poor HRQoL, consistent with past research showing that lower SPPB and higher 400 mW are significantly correlated with lower physical components of HRQoL (using SF-36) . Although poorer HRQoL (using SarQoL) appears to be more related to muscle function than to muscle mass (Beaudart et al., 2018), our study showed that low lean mass (ALM/h2), along with low physical performance measures (GS, TUG and 400 mW), were associated with poor HRQoL on the mental health dimension (sleeping, worrying and pain) among conventional gym participants. The conventional group also experienced a positive association for GS with HRQoL on the relationships dimension (friendships, isolation and family role). Outdoor mobility is important for engaging in social relations and activities . In support, exercise and social support from friends are both correlated with lower scores of depression, anxiety and perceived stress in older community-dwellers . Further, our results showed that while GS was positively associated with poor HRQoL on the independent living dimension (self-care, household tasks and mobility) in the conventional group, low CS along with low TUG and 400 mW, was negatively associated. This supports past research that GS and CS but not HGS correlate with most subscales of HRQoL (using SF-36) . However, this study’s HUR group did not experience any association for CS with the independent living dimension. Inconsistent associations for CS and missing associations for HGS infer that maybe HRQoL is not a good indicator of muscle strength.
Despite an average of a year of training, the mean AQoL-4D utility score (0.70) was lower than the mean of 0.76 previously reported in this age group . This appeared to be influenced by the HUR gym group, which had a 0.1 lower score (0.66 v 0.76) compared to the conventional gym group, although this did not reach statistical significance (p = 0.063). Again, the dynamic nature of the exercises at the conventional gym extending beyond machine-based exercises may be the explanation for this, as they are important for balance, mobility and falls prevention , although further work would be required to demonstrate this. Differences between the gyms could also reflect site-specific differences, and as such education and socioeconomic status of participants. However, we did not collect that information, thus this would need further investigation.
Due to the cross-sectional design and convenience of the sample, the population may be unrepresentative of the general population, and the unbalanced groups at HUR and conventional gyms may have limited finding significant associations, particularly in the conventional gym group. The study was limited to exercising older adults in Melbourne and results may not be generalised for the broader population. Another limitation is that we do not know what the health status of participants was prior to the starting at the gyms. Apart from gym sessions, participants could take part in regular exercise groups and extra activities were not measured in this study.