Biomechanical, physiological and anthropometrical predictors of performance in recreational runners

Introduction

Recreational runners engage in the activity with various objectives, ranging from participation in long-distance running events to the improvement of physical conditioning (Scheerder, Breedveld & Borgers, 2015). These objectives can be achieved through changes in biomechanical, physiological, and anthropometrical factors (Foster & Lucia, 2007; Novacheck, 1998; Saunders et al., 2004), which may lead to improved running performance (Saunders et al., 2004; Foster & Lucia, 2007; Tartaruga et al., 2012). However, beyond these primary factors, secondary factors such as anatomical factors (e.g., lower limb size) can also play crucial role in determining running economy, which is clearly important for overall running performance (Foster & Lucia, 2007).

Conceptual models present these variables directly or indirectly (Bassett & Howley, 2000; Saunders et al., 2004). Biomechanical factors, for example, are associated to movement patterns that possibly play a significant role in running performance, such as maximal vertical force, vertical stiffness (K_VERT), leg stiffness (K_LEG), stride frequency, stride length, contact time, aerial time, and vertical oscillation of the center of mass (Boullosa et al., 2020; Gómez-Molina et al., 2017; Saunders et al., 2004). Physiological and anthropometrical factors, such as heart rate corresponding to the onset of blood lactate accumulation, maximal running speed (V_MAX), speed associated with the second ventilatory threshold (V_2VT), body mass, and fat percentage, can also be determinants of performance (Abe et al., 1999; Boullosa et al., 2020; Daniels, 2013; Gómez-Molina et al., 2017).

The variable with a strong predictive power for running performance is V_MAX (Gómez-Molina et al., 2017; Houmard et al., 1991; Noakes, Myburgh & Schall, 1990; Scott & Houmard, 1994; Stratton et al., 2009). Stratton et al. (2009) demonstrated that the V_MAX is strongly correlated with long-distance running performance in both trained and untrained individuals. Similarly, Scott & Houmard (1994) observed this association between V_MAX and running performance among trained men and women. Furthermore, the V_MAX has been utilized as a critical performance marker for trained runners (Lanferdini et al., 2020). Current literature has shown that somatic and exertion variables, including age, body mass index, maximum oxygen uptake, blood lactate concentration at anaerobic threshold, and pulmonary ventilation, accurately estimate V_MAX in more than 4,000 endurance runners (Wiecha et al., 2022). However, many studies have been limited to either trained or elite athletes, thus limiting the generalizability of findings regarding the specific variables determining V_MAX in recreationally active runners through low-cost tests with simple instrumentation (Abe et al., 1999; Gómez-Molina et al., 2017; Lanferdini et al., 2020; McLaughlin et al., 2010). Further, very simple measures as body mass, fat mass, previous performance data of training and races may give important clues on performance prediction in distance runners (Melo et al., 2018, 2022). Allometric models have also been applied to understand the relationship between biomechanics versus running economy and performance (Detoni et al., 2015; Tartaruga et al., 2013).

Measurements of biomechanical, physiological, and anthropometrical variables can be easily determined through low-cost field tests, such as the maximal incremental test and the rectangular running test on a treadmill, based on known protocols for monitoring the individuals’ heart rate and recording gait patterns (Marfell-Jones, Stewart & De Ridder, 2012; Sentija, Vucetic & Markovic, 2007; Morin et al., 2005). These tests enable the estimation of athletes’ maximal aerobic capacity and the observation of their mechanical running patterns during the test, as shown by many studies that performed those respective tests (Tartaruga et al., 2012, 2013; Lanferdini et al., 2020). As a result, professional coaches and researchers can use this information to estimate running performance by determining V_MAX (Stratton et al., 2009). In addition, there are approaches that utilize maximal incremental tests to evaluate running economy (di Prampero et al., 2009). Surprisingly, little information exists on relationship between elastic mechanism variables and metabolic cost in recreational runners (Peyré-Tartaruga et al., 2021). Further, previous study has shown that the elastic mechanism plays a crucial role in the relationship between metabolic cost and speed of running in trained runners (Carrard, Fontana & Malatesta, 2018).

Therefore, we aimed to investigate the biomechanical, physiological, and anthropometrical determinants of V_MAX in recreationally active runners. We hypothesized that anthropometrical factors, such as body mass and fat percentage, along with the physiological variable V_2VT would emerge as the primary predictors of V_MAX, given the high heterogeneity observed among recreationally active runners. Further, our second objective was to relate biomechanical variables of running versus metabolic cost. Thus, we also hypothesized that mechanical variables related to the elastic mechanism would be associated with the metabolic cost. This hypothesis is based on the findings of da Rosa and coworkers, who clearly demonstrated landing-takeoff asymmetries of running optimized in runners with good performance in comparison to runners with lower performance (da Rosa et al., 2019).

Materials and Methods

Maximal incremental test

The protocol began with a 5-min warm-up at 10 km.h⁻¹ on a treadmill set at a fixed 1% inclination. Participants had a 4-min rest between the warm-up and the main test. The initial speed of the test was 6 km.h⁻¹, with an additional 1 km.h⁻¹ added each minute until participants reached exhaustion. The RPE was assessed at the end of each stage (Sentija, Vucetic & Markovic, 2007). The heart rate was recorded every 10 s during the incremental test. The V_MAX and HR_MAX were determined by the highest speed and heart rate achieved during a complete running stage (60 s), respectively. If the last stage was not completed, the V_MAX and HR_MAX would be determined by the speed and HR of the previous complete stage. Subsequently, the runners were instructed to walk for 2 min at 4 km.h⁻¹ for recovery (see Fig. 1).

Using the HR data recorded every 10 s by a sensor (Polar FT1, Kempele, Finland) during the maximal incremental test, the heart rate deflection point for each individual was identified in order to estimate the V_2VT and the HR_2VT (Sentija, Vucetic & Markovic, 2007). The HR_MAX, HR_2VT, V_MAX and V_2VT data obtained during the maximal incremental test were then used to calculate the theoretical metabolic cost.

Results

Table 1 shows the biomechanical, physiological and anthropometrical recreationally active runners variables in mean, standard deviation and 95% confidence interval.

The backward regression model for V_MAX resulted in a highly significant model (F(2,12) = 8.995, R² = 0.62, p = 0.011) and a regression equation of

(7) $$\begin{array}{rl}{\mathrm{V}}_{\mathrm{M}\mathrm{A}\mathrm{X}}=& 58.632+(-0.183\ast \mathrm{f}\mathrm{a}\mathrm{t}\mathrm{p}\mathrm{e}\mathrm{r}\mathrm{c}\mathrm{e}\mathrm{n}\mathrm{t}\mathrm{a}\mathrm{g}\mathrm{e})+(-0.507\ast \mathrm{H}\mathrm{R}2\mathrm{V}\mathrm{T}\mathrm{\_}\mathrm{p}\mathrm{e}\mathrm{r}\mathrm{c})\\ & +(7.959\ast \mathrm{C}\mathrm{o}\mathrm{s}\mathrm{t}\mathrm{\_}\mathrm{t}\mathrm{h}\mathrm{e}\mathrm{o}\mathrm{r})\end{array}$$where HR2VT_perc is the heart rate as a percentage of HR_MAX, and Cost_theor is the theoretical metabolic cost. Specifically, three significant regression coefficients were found: fat percentage (b₁ = −0.183, t = −2.981, p = 0.013), heart rate as a percentage of HR_MAX (b₂ = −0.507, t = 2.464, p = 0.031), and theoretical metabolic cost (b₃ = 7.959, t = 0.841, p = 0.042).

Similarly, the backward regression model for theoretical metabolic cost also resulted in a highly significant model (F(5,9) = 15.936, R² = 0.91, p = 0.004) and a regression equation as follows

(8) $$\begin{array}{rl}\mathrm{C}\mathrm{o}\mathrm{s}\mathrm{t}\mathrm{\_}\mathrm{t}\mathrm{h}\mathrm{e}\mathrm{o}\mathrm{r}=& 10.421+(4.282\ast contacttime)+(-3.795\ast aerialtime)\\ & +(-2.422\ast stridelength)+(-1.711\ast stridefrequency)\\ & +(0.107\ast verticaloscillation)\end{array}$$where Cost_theor is the theoretical metabolic cost. Specifically, five significant regression coefficients where found, contact time (b₁ = 4.282, t = 1.780, p = 0.006), aerial time (b₂ = −3.795, t = 6.221, p = 0.002), stride length (b₃ = −2.422, t = 4.142, p = 0.009), stride frequency (b₄ = 1.711, t = 4.551, p = 0.006), and vertical oscillation (b₅ = 0.107, t = 2.963, p = 0.031).

Figures 2A and 2B show the relationship between V_MAX versus V_2VT (R² = 0.86, p < 0.001) and fat percentage (R² = 0.34, p = 0.021), respectively.

Marginal effects of biomechanical variables on theoretical metabolic cost are presented in Fig. 3.

Discussion

The purposes of this study were twofold. First, we sought to verify which physiological (HR_2VT, HR_MAX, theoretical metabolic cost, and V_2VT) and anthropometrical variables (body mass, fat percentage) are determinant of V_MAX in recreationally active runners. Also, we sought to relate biomechanical (K_VERT, K_LEG, maximal vertical force, contact time, aerial time, stride length, stride frequency and vertical oscillation of the center of mass) variables are determinant of metabolic cost in recreationally active runners. Therefore, this is the first study to our knowledge to demonstrate specific physiological (V_2VT) and anthropometrical (fat percentage) variables determining V_MAX in recreationally active runners (Fig. 4). We found that velocity associated with second ventilatory threshold, metabolic cost and fat percentage were predictors of V_MAX. Furthermore, our study provided new insights about the mechanical determinants of metabolic cost of running. We found that simple spatiotemporal variables and vertical oscillation were identified as determinants of the metabolic cost. Interestingly, the elastic mechanisms were not related to metabolic cost. These results reinforce the understanding that estimates of the elastic mechanism based on temporal parameters of running do not appear to be sensitive, as already observed in master athletes (Pantoja et al., 2016; Cavagna, Legramandi & Peyré-Tartaruga, 2008). On the other hand, the role of vertical oscillation in the metabolic cost of running is in line with previous findings in highly trained endurance runners (Tartaruga et al., 2012).

According to our results, V_2VT, theoretical metabolic cost, and fat-mass percentage collectively accounted for 62% of the total variance in V_MAX. These findings agree with several studies performed with recreational, trained and elite runners (Noakes, Myburgh & Schall, 1990; Zacharogiannis & Farrally, 1993; Abe et al., 1999; Bassett & Howley, 2000; Stratton et al., 2009; Melo et al., 2022; Zillmann et al., 2013; Gómez-Molina et al., 2017). Noakes, Myburgh & Schall (1990), considering V_MAX as a performance predictor in trained runners, also associated this role to the V_2VT. Furthermore, Gómez-Molina et al. (2017) and Stratton et al. (2009) substantiate this association by incorporating V_2VT to V_MAX in the multiple linear regression equation, resulting in a significant enhancement in the prediction of the performance output in trained runners. The association between V_2VT and V_MAX can be justified as both variables represent the aerobic and anaerobic maximal capacities of the athletes, respectively (Gómez-Molina et al., 2017). Higher mitochondrial density and enzymatic activity lead to lower metabolic acidosis, allowing runners to cover greater distances or increase running speed without compromising the oxidative metabolism (Bassett & Howley, 2000; Gómez-Molina et al., 2017).

In addition, increases in running speed during long-distance events are common, mainly at the beginning and in the final sprint of the race (Maroński, 1996). However, a higher fat percentage in recreational runners is not favorable to increase the V_MAX values. Notably, very low-fat mass has been observed in African elite runners, who exhibit both low muscle volume and a low-fat percentage, contributing to a lower total body mass. This characteristic allows them to reach higher speeds throughout long-distance running events (Zillmann et al., 2013). This perspective aligns with previous research by Hoogkamer et al. (2016), where they demonstrated that a 100 g increase in body mass results in approximately a 1% decrease in the 3,000 m running performance. Concerning long-distance events, Zillmann et al. (2013) supported the findings of the present study, demonstrating a negative association between fat percentage and V_MAX. Therefore, the fat percentage seems to be determining factor for optimal running performance in this population, as V_MAX is directly affected.

The multiple linear regression test did not show an association between variables related to the elastic mechanism and metabolic cost in the present study. Initial variations in stride length appear to be fundamental for achieving higher speeds, while stride frequency becomes crucial at high speeds (Novacheck, 1998; Peyré-Tartaruga et al., 2021). However, the negative association between stride length and contact time may be confounded by variations in speed (Gómez-Molina et al., 2017), as changes in these variables are known to occur to increase running speed. Therefore, our results may not have shown associations due to the relatively slow speed used in the present study (10 km.h⁻¹).

The spring-mass model during running can be characterized by two major variables associated with the leg and the vertical stiffness of the center of mass (K_LEG and K_VERT, respectively) (Morin et al., 2005), and these variables appear to be influenced by speed and performance level (Carrard, Fontana & Malatesta, 2018). Rogers et al. (2017) evaluated V_MAX and K_LEG during the 50 m sprint, along with tendon stiffness, K_VERT (during unilateral and bilateral drop jump), running economy on a treadmill (12, 14, 16 and 18 km.h⁻¹), K_LEG at 14 km.h⁻¹ and velocity associated with the maximal oxygen consumption in 11 highly trained male runners. They found large and moderate correlations between the values of K_LEG (during the 50 m sprint) and K_VERT (in unilateral jump) with the running economy at 14 km.h⁻¹ (Rogers et al., 2017), supporting the concept that tendon stiffness in the lower limb could influence the global stiffness models (running and jump), thereby running performance.

Regarding performance level, Burns et al. (2021) found that speed influences the biomechanical behavior and spring-mass characteristics in middle-distance elite (international/Olympic) versus trained (regional) with an average speed of 6.90 and 6.06 m.s⁻¹ for 1,500 m, respectively. The authors noted spatiotemporal and spring-mass parameter differences across running speeds between 10 to 18 km.h⁻¹. Elite runners demonstrated greater stability in contact time, longer aerial time, and a lower duty factor than well-trained runners (Burns et al., 2021). The spring-mass characteristics in both groups showed that F_MAX and K_VERT increased with speed, while K_LEG showed minimal changes, with elite runners exhibiting high values (Burns et al., 2021).

Contrary to these findings, our study did not reveal associations between variables related to the elastic mechanism and metabolic cost in recreationally active runners. It suggests that individuals engaging in leisurely exercise may adapt their step frequency and length to minimize vertical oscillation of the center of mass (Tartaruga et al., 2012) and leg deformation (Morin et al., 2007), optimizing the task. Therefore, it could potentially counterbalance any impairment in F_MAX during each step due to a lower level of training (da Rosa et al., 2019), without significant effects on K_VERT and K_LEG at 10 km.h⁻¹.

On the other hand, a critical association was found between HR_2VT and V_MAX in the present study. Indeed, this finding is in line with previous results where Abe et al. (1999) observed a strong correlation between heart rate during the onset of blood lactate accumulation and V_MAX in elite long-distance runners with similar performance levels. The heart rate found by Abe et al. (1999) may correspond to HR_2VT in the present study, given that the exercise intensity was close to 4 mM, corresponding to the blood lactate values at the second ventilatory threshold approximately. Additionally, the values for HR_2VT and HR_MAX are consistent with previous findings (Bunc et al., 1995; Hofmann et al., 1994, 1997; Vucetić et al., 2014; Zacharogiannis & Farrally, 1993). Moreover, our study found no association between HR_MAX and V_MAX, also in line with the results reported by Gómez-Molina et al. (2017) and Noakes, Myburgh & Schall (1990). A similar previous study has sought to relate the determinants of running velocity (Wiecha et al., 2022). However, the runners used in Weicha’s study were at a higher level than ours, and using the approach recommended by McKay et al. (2022), their athletes are considered trained athletes (tier 2) and ours recreationally active (tier 1).

The primary limitation of this study was the analysis of the biomechanical variables at a fixed speed of 10 km.h⁻¹. Conceptual models suggest variations in biomechanical responses with increasing running speed, including step length, step frequency, contact time, and aerial time (Novacheck, 1998; Peyré-Tartaruga et al., 2021). Another limitation was the absence of blood lactate concentration evaluation during the maximal incremental test. This evaluation would have brought HR_2VT and V_2VT closer to the actual second ventilatory threshold, offering better associations with results in the literature.

Conversely, our study provides a practical and valuable approach for analyzing the determinants of running performance and evaluating training development in recreationally active runners. The results can offer insights for coaches working with amateur and professional running groups, particularly those guiding inexperienced runners aiming to improve their health indicators or seeking information through science to enhance their running practice. Coaches should prioritize attention to the fat percentage, especially when using speed to control intensity (Daniels, 2013), preventing injuries from weight overload. Additionally, coaches can focus on improving V_2VT to enhance the aerobic capacity of recreationally active runners, often through high-intensity models (Silva et al., 2017). This procedure, in turn, can contribute to improved running performance and the development of new biomechanical features (Daniels, 2013). Therefore, this study provides valuable information for achieving optimal running performance (V_MAX).

Further, refining the heart rate ratio model (Castagna, Krustrup & Póvoas, 2022) gives a simple equation for estimating the metabolic cost during running. The only input variables were the maximal, rest and exercise heart rate. Using combustion enthalpy parameters (Peyré-Tartaruga & Coertjens, 2018), we are able to obtain a single equation for determining the metabolic cost of running. While these findings are encouraging, a couple of assumptions of the model have to be kept into account. Conditions where the cardiac output is not a key constraint of performance, as at supramaximal intensities (higher than V_MAX), the model presented is limited (Lepretre, Koralsztein & Billat, 2004). Indeed, at supramaximal intensities, glycolytic metabolism precludes the estimation of metabolic cost using only aerobic pathways.

Considering the established interactions between physiological, anthropometrical, and biomechanical variables, we suggest exploring non-linear approaches in future studies. Although not the most suitable for the present research questions, future studies could employ these methods to address the complexity of running performance. Network models, for instance, could offer valuable insights, even taking into account critical aspects as sex effects on the relationships studied here (e.g., Manchado-Gobatto et al., 2022; Pereira et al., 2018). Furthermore, future studies may include additional variables contributing to determining V_MAX. Specifically, investigating biomechanical variables during the maximal incremental test, such as external and internal mechanical work, mechanical efficiency (Peyré-Tartaruga et al., 2021; Peyré-Tartaruga & Coertjens, 2018), and muscle and tendon mechanical properties, could offer further insights into the complex interplay affecting V_MAX.

Conclusions

In conclusion, our study demonstrates that V2VT, metabolic cost, and fat-mass percentage can explain 62% of the variability in VMAX. Also, 90% of the variability in metabolic cost is accounted for by simple spatiotemporal variables (contact and aerial time, stride frequency and length, and vertical oscillation). Therefore, we concluded that recreationally active runners with higher V_2VT, lower metabolic cost, and a lower fat percentage will likely present better V_MAX performance.