In this cross-sectional study, the participants went to a fitness health club for three consecutive weeks to perform two weekly sessions, separated by 48 h of rest. We dedicated the first 2 weeks to familiarization and anthropometric measures. During this period, we emphasized the proper execution technique of each exercise and movement velocity. Afterward, in the first session of the third week, the participants performed the following tests: MBT with 1 kg (MBT-1 kg) and 3 kg (MBT-3 kg), 10-m walking speed (10W), and five-repetition sit-to-stand (STS). After 48 h of rest, the participants performed a second session constituted by a progressive loading test in the leg press and chest press. An experienced researcher involved in the study and two certified senior fitness coaches supervised the procedures to guarantee safety and proper supervision during each exercise. In addition, verbal encouragement was provided during each exercise to motivate the participants to give a maximal effort. Figure 1 illustrates the study design.

We estimated a sample size of twenty-three participants to achieve a power of 80%, considering an alpha level of 0.05, two predictor variables (MP_max and PP_max), and an r ² of 0.38 based on the relationship between leg power and the short physical performance battery reported by Bean et al. (2002) (G*Power v3.1). Therefore, thirty-two older adults from residential care facilities and day centers were recruited to participate in this study (Table 1). We included male and female participants aged 65 years or more, able to walk and stand up from a chair independently, and willing to participate in the study. We excluded participants if they had physical dependency (Barthel Index score <60), cognitive decline (Mini-Mental State Examination [MMSE] cut-off scores: no years of schooling, <15 points; 1–11 years of school, <22 points; and >11 years of school, <27 points (Mendes et al., 2017)), musculoskeletal injuries in the previous 3 months, and terminal illness. The clinicians of the centers conducted the initial screening tests, including the Barthel Index and MMSE. According to the clinicians, all participants had no records of risk factors (e.g., uncontrolled hypertension and arrhythmia) that could prevent them from performing the exercises included in the study. Furthermore, all participants were classified as sedentary since they had no records of participating in regular physical exercise programs in the last 3 months. All participants were informed of the study procedures and signed written informed consent. The Ethical Committee of the University of Beira Interior approved this study (CE-UBI-Pj-2019-019).

We examined the assumption of normality of the data using the Kolmogorov-Smirnov test. We used standard statistical methods to calculate means, standard deviations (SD), 95% confidence intervals (CI), Pearson correlation coefficients (r), the adjusted coefficient of determination (r ²), and the standard error of the estimate (SEE). Quadratic regressions analyzed 1) the load-mean and peak power relationships in the leg press and chest press and identified the MP_max-load (% 1RM), MP_max (W), PP_max-load (% 1RM), and PP_max (W) in the leg press and chest press in women and men, 2) the associations between MP_max and PP_max in the chest press with MBT-1 kg and MBT-3 kg, and 3) the associations between MP_max and PP_max in the leg press with 10W_velocity and STS_power. We used quadratic regressions to analyze the associations between MP_max and PP_max with functional capacity markers due to the curvilinear relationship between muscle power and functional capacity (Bean et al., 2002; Cuoco et al., 2004; Marsh et al., 2006; Byrne et al., 2016). Independent samples t-test analyzed 1) the differences between sexes in absolute mean and peak power values (W) in the leg press and chest press for each relative load, including the MP_max-load and PP_max-load, and 2) the differences between sexes in MP_max-load and PP_max-load in the leg press and chest press. A repeated-measures ANOVA with post hoc Bonferroni tests analyzed the differences between MP_max/PP_max in the leg press and chest press with absolute power values (W) at different relative loads in men and women. Paired samples t-test analyzed 1) the differences between MP_max-load and PP_max-load within resistance exercises, and 2) the differences between resistance exercises in MP_max-load and PP_max-load. We performed the statistical analyses in Microsoft Office Excel^® (Microsoft Inc., Redmond, WA, United States) and SPSS v27 (SPSS Inc., United States) and set the significance level at p < 0.05. We designed the figures in GraphPad Prism v7 (GraphPad Inc., San Diego, CA, United States).

Figure 2 shows the load-mean and peak power relationships in the leg press in older women and men. Men presented higher absolute peak power values than women at 35%–95% 1RM (Figure 2A) and higher absolute mean power values at 30%–100% 1RM (Figure 2B). The PP_max-load in the leg press did not differ between men and women (p = 0.59). In men, the PP_max was not different from peak power values associated with loads at 60%–65% 1RM (p > 0.05), while in women, the PP_max was not different from peak power values associated with loads at 60%–70% 1RM (p > 0.05) (Figure 2A). The MP_max-load in the leg press did not differ between men and women (p = 0.62). In men, the MP_max was not different from mean power values associated with loads at 60%–70% 1RM (p > 0.05), while in women, the MP_max was not different from mean power values associated with loads at 65%–70% 1RM (p > 0.05) (Figure 2B).

Figure 3 shows the load-mean and peak power relationships in the chest press in older women and men. Men presented higher absolute peak and mean power values than women at 20%–100% 1RM (Figures 3A,B, respectively). The PP_max-load in the chest press did not differ between men and women (p = 0.09). In men, the PP_max was not different from peak power values associated with loads at 40%–65% 1RM (p > 0.05), while in women, the PP_max was not different from peak power values associated with loads at 55%–60% 1RM (p > 0.05) (Figure 3A). The MP_max-load in the chest press did not differ between men and women (p = 0.41). In men, the MP_max was not different from mean power values associated with loads at 55%–65% 1RM (p > 0.05), while in women, the MP_max was not different from mean power values associated with loads at 55%–60% 1RM (p > 0.05) (Figure 3B).

Table 2 shows differences between the leg press vs. chest press in PP_max-load for men and women (p < 0.01). In addition, there were differences between the leg press vs. chest press in MP_max-load for men and women (p < 0.01).

Table 3 shows differences between PP_max-load vs. MP_max-load in the leg press for men and women (p < 0.01). In addition, there were differences between PP_max-load vs. MP_max-load in the chest press for men and women (p < 0.001).

Figure 4A indicates that the PP_max in the chest press explained 48% of MBT-1 kg variance, while Figure 4B shows that the MP_max in the chest press explained 48% of MBT-1 kg variance. In addition, Figure 4C reveals that the PP_max in the chest press explained 52% of MBT-3 kg variance, while Figure 4D shows that the MP_max in the chest press explained 53% of MBT-3 kg variance.

Figure 5A indicates that the PP_max in the leg press explained 61% of STS_power variance, while Figure 5B shows that the MP_max in the leg press explained 58% of STS_power variance. In addition, Figure 5C reveals that the PP_max in the leg press only explained 2% of 10W_velocity variance, while Figure 5D shows that the MP_max in the leg press only explained 1% of 10W_velocity variance.

The current study aimed to 1) analyze the load-mean and peak power relationships in the leg press and chest press in older women and men, 2) examine the differences between MP_max-load and PP_max-load within resistance exercises, 3) identify the differences between resistance exercises in MP_max-load and PP_max-load, and 4) explore the associations between MP_max and PP_max in the leg press and chest press with functional capacity indicators. The main findings of the current study were: 1) the MP_max-load and PP_max-load in the leg press and chest press are similar between older women and men, 2) the MP_max-load and PP_max-load differ between resistance exercises, meaning that they are exercise-specific, 3) the P_max-load varies in the same resistance exercise depending on the mechanical power variable chosen to measure, 4) the MP_max and PP_max in the chest press similarly explain the variability in MBT-1 kg and MBT-3 kg performance, and 5) the MP_max and PP_max in the leg press similarly explain the STS_power variance; however both mechanical variables could not explain the variability in 10W_velocity performance.

The results of this study showed that the MP_max-load and PP_max-load in the leg press and chest press did not differ between older women and men, which agrees with previous findings, particularly for the peak power values (Strand et al., 2019). This convergence in muscle power production between sexes might be related to the more significant and faster age-related losses of muscle power in men than women during aging (Edwén et al., 2014; Strand et al., 2019). The results also showed that the load-power relationship in older adults is resistance exercise-specific, thus corroborating the results of previous observations (Strand et al., 2019). For example, the PP_max-load in the leg press and chest press was around 60% and 55% 1RM, respectively, which agrees with previous findings (de Vos et al., 2008; Potiaumpai et al., 2016; Strand et al., 2019). On the other hand, the MP_max-load in the leg press and chest press was unknown until the completion of our study. Compared to PP_max-load, the MP_max-load in the leg press and chest press increased to around 66% and 62% 1RM, respectively. Despite its novelty in older populations, these data also indicate that the P_max-load differs between mechanical power variables in older adults, as observed in young adults (Pallarés et al., 2014; Sánchez-Medina et al., 2014; Martínez-Cava et al., 2019). Although most studies with older adults analyzed the P_max-load using the peak power variable (de Vos et al., 2005; Potiaumpai et al., 2016; Ni and Signorile, 2017; Strand et al., 2019), several authors observed higher reliability using mean values than peak values when conducting a progressive loading test in the leg press with this population (Alcazar et al., 2017). However, since no study had yet presented data concerning the MP_max-load in resistance exercises, these results present preliminary evidence for clinicians and researchers who want to collect mean power values to estimate the P_max-load. In addition, these results also alert the importance of defining the mechanical power variable beforehand to be monitored during the intervention to avoid misinterpreting information during its course.

The current research also demonstrated that the P_max-load range in the leg press (∼60%–70% 1RM) and chest press (∼40%–65% 1RM) was narrower than those observed for younger populations when using, for example, the squat or bench press exercises (∼30%–70% 1RM) (Soriano et al., 2015; Soriano et al.,2017). These differences might be attributed to the progressive reduction in size and number of fast-twitch muscle fibers in the lower and upper limbs with aging, which negatively affects the elbow and knee extensor’s power capacity (Metter et al., 1997; Candow and Chilibeck, 2005; Korff et al., 2014). In addition, as observed in our data, the P_max-load range in the leg press was narrower than the chest press, which might be associated with the higher muscle power production losses in the lower limbs than in the upper limbs during aging (Macaluso and De Vito, 2004; Candow and Chilibeck, 2005). According to the literature, a significant reduction in physical activity with age and greater use of the upper limbs than the lower limbs to perform the activities of daily living (e.g., using arms to help to stand up from a chair) might contribute to higher decreases in lower limb’s power than upper limb’s power (Macaluso and De Vito, 2004; Candow and Chilibeck, 2005). Therefore, these results suggest a broad spectrum of relative loads to maximize the upper-limb muscle power and a narrow range of relative loads to maximize the lower-limb muscle power in older adults. Nevertheless, future research should analyze if training only with the P_max-load improves older adults’ muscle power to a greater extent than a broader range of relative loads.

The regression analysis showed that the MP_max and PP_max in the chest press could similarly explain the MBT-1 kg and MBT-3 kg performance. These data reinforce the influence of the MBT as an indicator of muscle power and functionality in older adults (Harris et al., 2011). Furthermore, although the relationship between the chest press power and functional capacity in older adults is scarce, earlier findings demonstrated a correlation between the chest press peak power and self-reported functional status (lower scores representing better functional status) (r = −0.35) in older women (Foldvari et al., 2000). Consequently, considering the associations between chest press muscle power with MBT, it can be suggested that the MBT seems an essential indicator of the capacity to perform the activities of daily living independently in older adults, such as lifting and carrying groceries and boxes, opening jars, rising from a chair with the help of the arms, and even catching oneself to prevent a fall (Adams et al., 2001; Candow and Chilibeck, 2005; Harris et al., 2011). Based on this information, clinicians, sport-related professionals, and researchers can administer the MBT test to analyze the upper-limb muscle power capacity and derive information regarding the functional ability of older adults.

As for the regression analysis in the lower limbs, the MP_max and PP_max in the leg press could similarly explain the variability in the STS_power performance. These results reinforce the substantial impact of lower-limb muscle power on explaining the variability during sit-to-stand transitions in older adults (Byrne et al., 2016). However, neither MP_max nor PP_max in the leg press could explain the variance in 10W_velocity performance. These results were surprising and unexpected since previous research found that leg press power could explain the variance in walking speed performance in older adults (Bean et al., 2002; Puthoff and Nielsen, 2007). Nevertheless, it is essential to note that the latter investigations that assessed the relationships between leg power and walking performance were conducted with mobility-limited older adults, unlike our study. Therefore, the impaired physical condition might have influenced the relationship between lower-limb muscle power with walking performance. Interestingly, research with community-dwelling older adults with similar maximal walking velocity values as our participants (1.6–2.0 m s⁻¹) found that hip and ankle muscle strength were better predictors of maximal walking speed than leg strength (Uematsu et al., 2014; Muehlbauer et al., 2018). Therefore, it is essential to consider that hip and ankle strength might better account for the variance in walking performance than leg strength in older adults without mobility impairments (Muehlbauer et al., 2018). Nevertheless, future large-scale research is necessary to determine the influence of leg, hip, and ankle power and strength on maximal walking performance in older adults with and without mobility limitations.

Of note, the range of r ² values observed in our study is in line with previous research (Byrne et al., 2016), which indicates that a large part of the variance in functional capacity is to be explained by other outcomes (Puthoff and Nielsen, 2007). For example, aerobic endurance, balance, flexibility, agility, and even the fear of falling might explain the variance in functional capacity in older adults (Puthoff and Nielsen, 2007). Therefore, future research should examine, along with lower and upper-limb muscle power, what physiological and psychological indices play a significant role in explaining the variability in functional capacity in older adults.

The current study presents several limitations that we need to address. Firstly, a cross-sectional design does not allow us to establish causal relationships between muscle power with functional capacity in the tested population. In this perspective, future longitudinal studies with older adults should examine the effects of resistance training on muscle power and functional capacity and determine their relationships to support causal links. Secondly, although the sample size calculation determined that twenty-three participants were needed to obtain a statistical power of 80%, the actual number of participants is insufficient to generalize the results to other older adults. In addition, considering that our participants were functionally independent, caution should be taken when generalizing these results to mobility-limited older adults. Finally, including physiological and psychological outcomes would be helpful to examine if, along with lower and upper-limb muscle power measures, they could increase the capacity to explain the remaining part of the variance in functional capacity in older adults. Therefore, future research should consider the limitations mentioned above and conduct large-scale, longitudinal, and experimental studies to examine the physiological and psychological mechanisms that better explain the variability in functional capacity in older adults.