Abstract
h the hamstring muscles play an important role in running, very little is known about the individual contributions of each hamstring muscle (biceps femoris long head, biceps femoris short head, semitendinosus, and semimembranosus) toward running economy. As such, our study examined all the muscles in the hamstring to provide insight into which muscles contribute the most to running economy. Such information can provide insight in designing precise exercise training programs for enhancing running performance. Sec- ondary analysis from our cross-sectional study conducted on 23 male recreational runners examined the relationships between stretch shortening cycle potentiation (via leg press throw), running net VO2(inverse of running economy) (at 11.3 km/h), and maximum cross-sectional area of biceps femoris long head, biceps femoris short head, semitendinosus, and semimembranosus was assessed via magnetic resonance imaging. We obtained significant correlations between the maximum cross-sectional area of the biceps femoris long headand log10running net VO2(r =−0.52;p< 0.05). Our multiple
shortening cycle potentiation (via leg press throw), running net VO2(inverse of running economy) (at 11.3 km/h), and maximum cross-sectional area of biceps femoris long head, biceps femoris short head, semitendinosus, and semimembranosus was assessed via magnetic resonance imaging. We obtained significant correlations between the maximum cross-sectional area of the biceps femoris long headand log10running net VO2(r =−0.52;p< 0.05). Our multiple regression model showed that the maximum cross-sectional area of biceps femoris long headbut not stretch shortening cycle potentiation predicted log10running net VO2(r =−0.52;p< 0.01). We found no other relationship between any other hamstring muscles and log10running net VO2. Our findings provide preliminary evidence of the importance of the biceps femoris long headtoward run- ning economy. This may be due to the preferential activation of efficient slow twitch muscle fibers of the biceps femoris long head. Additionally, we noted that the biceps femoris short head, semitendinosus, and semimembranosus muscles were not related to running economy in recreational male runners. Keywords:knee; gait; tendon; sports; injury 1. Introduction Running economy or inverse of running net maximal oxygen uptake (VO2) is defined as the steady-state oxygen consumption in response to running at a specific speed [1]. Individuals with higher vs. lower running economy consume lower oxygen and can thus run for longer periods of time without exhaustion [1]. Indeed, running economy is a strong indicator of running performance, with prior research showing that much of the variance Sports2025,13, 403 https://doi.org/10.3390/sports13110403
Sports2025,13, 403 2 of 10 of running performance is explained by running economy in middle distance runners [2] and physically active men [1]. Thus, creating evidence-based precise training programs to improve running performance requires insight into running economy. It is also known that running economy is negatively related to physical/muscular fatigue [3]. This is important because physical fatigue can contribute to hamstring injuries in athletes [4,5], creating significant time loss and cost for competitive athletes [6,7]. Thus, an insight into running economy could also help with preventive health care needs of athletes, beside its potential as a prognostic and diagnostic marker of hamstring injuries [8,9]. Notably, biceps femoris (BF), semitendinosus muscles (ST), and semimembranosus (SM) muscles make up the hamstring muscle group. BF, SM, and ST comprise ~40%, ~35%, and ~25% of the hamstring muscle cross-sectional area, respectively [10]. Specifically, BF long headcomprises ~67% of the BF and ~27% of the total hamstring muscle volume [11]. Additionally, the BF long headcarries a stabilizing function of the BF and the hamstring dur- ing the late-swing phase of running [9]. Notably, BF, ST, and SM muscles display differential composition of muscle and tendon amongst them. For example, the musculotendinous unit of the BF is almost 75% muscle and 25% tendon [11,12]. While these measurements are from BF long head, BF short headis not known to contribute to running economy due to its significantly smaller muscle and tendon size [13]. Additionally, the musculotendinous units of the ST and SM have noticeably greater tendon than the BF [14]. We know that various biomechanical factors such as tendon length, tendon thickness, and physiological factors such as muscle strength and biochemical factors such as transport pump efficien- cies, myosin ATPase efficiency, and mitochondrial respiration efficiency can affect running economy [1,15,16] . In addition, differences in muscle vs. tendon composition of different hamstring muscles could uniquely influence the aforementioned biomechanical and bio- chemical factors and could thus differentially affect running economy. However, no such information currently exists to the best of our knowledge. There are two well-known factors which markedly influence running economy: stretch shortening cycle potentiation
affect running economy [1,15,16] . In addition, differences in muscle vs. tendon composition of different hamstring muscles could uniquely influence the aforementioned biomechanical and bio- chemical factors and could thus differentially affect running economy. However, no such information currently exists to the best of our knowledge. There are two well-known factors which markedly influence running economy: stretch shortening cycle potentiation (SSCP; enhancement of muscle concentric force after an eccentric preloading) and the preferential activation of slow-twitch muscle fibers (PASS) [1]. It is well-established that SSCP is important for the walking and running economy [17,18]. Furthermore, ~35% to ~50% of the energy stored during the eccentric phase is used in the following concentric phase during a stretch shortening cycle activity such as leg press throw [19]. Since SSCP is positively dictated by velocity [19], more type II than type I muscle fibers may hold a greater potential in SSCP biomechanics. Interestingly, the role of PASS in the running economy is not very clear. Indeed, type I muscle fibers, which are known to be more biochemically efficient than fast-twitch muscle fibers due to greater mitochondria, are recruited first during a strenuous task [19]. Notably, maximum muscle cross-sectional area (MCSA) is a known surrogate biomarker of muscle strength, where a higher MCSA corresponds to greater muscle strength [20]. Thus, MCSA and SSCP of a muscle could be two unique factors contributing to running economy. To the best of our knowledge, whether that is true for the muscles of the hamstrings is unknown. We do not know whether different muscle characteristics of BF long head, BF short head, ST, and SM are differentially related with running net VO2. Since it is non-feasible to invasively assess BF long head, BF short head, ST, SM for their muscle size, MRI-estimated MCSA is typically used as the surrogate marker of muscle size [21]. MRI-based MCSA measurements are known to have excellent inter- and intra-rater reliability in men as well as women and thus provides validity to the muscle size estimations [22–26]. Knowledge on the relationships of running net VO2with MCSA of BF, ST, and SM, and SSCP
their muscle size, MRI-estimated MCSA is typically used as the surrogate marker of muscle size [21]. MRI-based MCSA measurements are known to have excellent inter- and intra-rater reliability in men as well as women and thus provides validity to the muscle size estimations [22–26]. Knowledge on the relationships of running net VO2with MCSA of BF, ST, and SM, and SSCP can provide insight into understanding running economy and potentially injury risk in athletes and non-athletes. Thus, the main purpose of our study was to determine whether running net
Sports2025,13, 403 3 of 10 VO2was related to the MCSA of the individual hamstring muscles (BF long head, BF short head, ST, and SM) and SSCP of the lower limb in recreational runners. 2. Materials and Methods 2.1. Inclusion and Exclusion Criteria All eligible participants had completed one or more 10 km run or marathon within the previous 6 months of their study participation. In line with the pilot nature of our study, our study population was homogenous for sex. Thus, we enrolled male-only participants aged 24–40 years. All participants were physically healthy, and without any clinical conditions/medications that could affect their participation in our study. Individuals with experience with only short-distance running were not included in our study. 2.2. Participant Testing and Consent We tested our participants (n = 23) over 3 separate visits. The first visit comprised testing anthropometrics and body composition. SSCP and MCSA of BF long head, BF short head, ST, and SM testing occurred during the second visit. Finally, we tested VO2max and running economy during the third visit. All the main outcome measures (mean±SD) are reported in Table. The local Institutional Review Board approved our study. We obtained written informed consent from all the participants before any testing. The data presented in this study are secondary analysis of our previous study [19]. A representative image of our study design is shown in Figure. Our study was conducted in accordance with the Declaration of Helsinki [27]. Figure 1.A representative image of our study design is shown above. MRI, magnetic resonance imaging; SSCP, stretch shortening cycle potentiation; VO 2max, maximal oxygen uptake. Created in BioRender. Yaddanapudi, S. (2025) Table 1.Descriptive statistics. Variables Mean ±SD (n = 23) Age (years) 31.73 ±4.72 Height (m) 1.80 ±0.08 Body mass (kg) 80.13 ±10.79%
Sports2025,13, 403 4 of 10 Table 1.Cont. Variables Mean ±SD (n = 23) Body fat percentage (%) 17.90% ±6.53% Body mass index (kg/m 2 ) 25.21±3.24 MCSA BF long head(mm 2 ) (n = 18) 1906±344.66 MCSA BF short head(mm 2 ) (n = 21) 802.52±210.58 MCSA SM (mm 2 ) (n = 21) 1567.52±346.52 MCSA ST (mm 2 ) (n = 21) 1109.62±252.81 VO2max (mL/kg/min) (n = 22) 54.51 ±7.36 NVO2Run11 (mL/kg/min) 38.00 ±4.82 Log10NVO2Run11 1.58 ±0.05 POTV10 (m/s) (n = 21) 0.07 ±0.06 Abbreviations: MCSA = maximum cross-sectional area; BF = biceps femoris; SM = semimembranosus; ST = semi- tendinosus; NVO2Run11 = net running VO2at ~11 km/h; POTV10 = potentiated velocity at initial 10 ms of the concentric phase during a leg press throw. 2.3. Total Body Fat Percentage All our participants underwent a single full-body dual-energy x-ray absorptiometry (GE Lunar Prodigy ® , Madison, WI, USA; enCORE software, version 1.33) scan in the supine lying position to assess total body fat percentage. The coefficient of variation % for total body fat percentage is <2% in our laboratory. 2.4. MCSA Assessment We used a 3-dimensional volumetric T1-weighted Turbo Field Echo imaging sequence (T1TFE) and T1-weighted Turbo Spin Echo ® imaging sequences (TSE) on a Philips Achieva ® system (3 T) (Philips Medical Systems, Best, The Netherlands) to measure the MCSA of BF long head, BF short head, ST, and SM muscles. All the scans were collected using 1H transmit/receive torso phased-array coil. Next, we obtained scout images followed by a set of unique coronal, sagittal, and axial scans from below the patient’s knee to above the subject’s thigh and below the groin. We collected all the images from the right lower extremity to stay consistent with the lower extremity used for additional muscle biopsy tests (not reported here) conducted in this study. In total, we captured a set of 32 sagittal images (T1TFE, flip angle = 88, 32 contiguous slices, slice thickness = 2 mm, TR = 8.068 milliseconds, TE = 4.60 milliseconds, ETL = 160, acquisition matrix = 160×160, reconstructed matrix = 256×256, FOV = 250×250 mm)
extremity used for additional muscle biopsy tests (not reported here) conducted in this study. In total, we captured a set of 32 sagittal images (T1TFE, flip angle = 88, 32 contiguous slices, slice thickness = 2 mm, TR = 8.068 milliseconds, TE = 4.60 milliseconds, ETL = 160, acquisition matrix = 160×160, reconstructed matrix = 256×256, FOV = 250×250 mm) and 54 axial images (TSE, flip angle = 908, 54 contiguous slices, slice thickness = 5 mm, TR = 800 milliseconds, TE = 15 milliseconds, ETL = 3, acquisition matrix = 153×192, reconstructed matrix = 256×256, FOV = 160×160 mm). We used the axial images for all the MCSA measurements reported in this study. 2.5. SSCP Assessment We have previously described SSCP assessment in detail [18]. First, the supine lying leg press machine (Nebula #6000-A ® , Versailles, OH, USA) was attached with a cable and a linear position transducer (model # PT5DC-125-V62-UP-MOPO-C25; Celesco Transducer Products, Inc., Chatsworth, CA, USA) with the linear position transducer aligned to the linear motion of the leg press machine. For all the testing, the angle between the hip and back rest was set at 35 degrees with respect to the floor. Participants kept their feet 0.02m apart with the edge of their posterior heel lining with the bottom edge of the leg press foot plate for all the testing. For all the ballistic (concentric only) leg press throws, the starting position was set at 90 degrees of the knee flexion, whereas we had our participants fully extend their knees as the starting position for SSCP leg press throws. A total of 150% of the respective body weight was used for ballistic leg press throws assessment. We employed
Sports2025,13, 403 5 of 10 finite-difference techniques to calculate the leg press throwing velocity (33). Participants performed 3 SSC followed by 3 ballistic leg press throws with 1 min rest between each successive throw. We chose the velocity values from the best 2 trials and then averaged those for SSC throws and ballistic throws, respectively, to report in this study. The initial 10 ms difference between SSC and ballistic velocity was calculated and used as potentiated velocity metric (POTV10). As previously described [18], coefficient of variation % for POTV10 in our laboratory is <4.4% with an intraclass correlation coefficient of >0.93. Data were collected using a National Instruments (Austin, TX, USA) data acquisition system sampled at 1 kHz and digitally filtered using a low-pass 4th order Butterworth filter with a cut-off frequency of 50 Hz. 2.6. Running Net VO2at 11.3 km/h (NVO2Run11) We used a MAX-II Cart ® metabolic system (Physio-Dyne Instrument Company, Quogue, NY, USA) to collect seated resting and running VO2. Before any data collec- tion, metabolic system calibration was conducted with a 3 L calibration syringe and known composition of standard gases. Seated resting VO2was acquired while participants were seated for 10 min. Immediately after seated resting VO2assessment, participants were asked to run at 11.3 km/h for 10 min. The 11.3 km/h speed was used because of its reported reliability for assessing running economy in young adult male recreational runners [19]. We subtracted the average of the last 5 min of the seated resting VO2from the average VO2 value over the last 5 min of steady state running to obtain net running VO2(NVO2Run11). 2.7. VO2max After running net VO2test, a VO2max test was conducted by having our participants run at increasing intensities for 1 min intervals starting at 9.65 km/h. Two options—increasing speed by 0.8 km/h or grade by 2.5%—were offered to our participants at the end of each successful running minute. Participants were asked to run until their maximum exhaustion. We used the following criteria to assess successful testing of VO2max: (a) plateauing of VO2, (b) respiratory exchange ratio (RER) larger than 1.2, or (c) heart
at 9.65 km/h. Two options—increasing speed by 0.8 km/h or grade by 2.5%—were offered to our participants at the end of each successful running minute. Participants were asked to run until their maximum exhaustion. We used the following criteria to assess successful testing of VO2max: (a) plateauing of VO2, (b) respiratory exchange ratio (RER) larger than 1.2, or (c) heart rate within 10 beats of age predicted maximum. 2.8. Statistical Analysis Descriptives are presented as means±SD in Table. We used kurtosis, skewness, and Shapiro–Wilk tests to test the normality of our data. Using an effect size (f 2 ) of 0.4,α= 0.05, power (1−) = 0.8, number of predictors = 2, and one tailed linear multiple regression fixed model with single regression coefficient model yielding a minimum of 18 participants for our study. A logarithmic conversion of the non-normally distributed NVO2Run11 was used for all the analyses. We used simple Pearson correlations among the variables of interest (MCSA of BF long head, BF long head, ST, SM, POTV10, and log10NVO2Run11). Finally, we used linear regression with log10NVO2Run11 as the dependent variable and POTV10, MCSA of BF long head, BF long head, ST, and SM as the independent variables. Due to tech- nical difficulties, data were not recorded for up to 5 participants across a combination of variables as shown in Table. We set the level of significance for all the two-tailed analyses atp< 0.05.Since it is already known that SSCP is positively related to walking and running economy [28,29], we performed a 1-tailed Pearson correlation test to examine the relationship between log10NVO2Run11 vs. POTV10. All the analyses were run using SPSS software version 30.0 (SPSS Inc., IBM, Chicago, IL, USA). 3. Results Our population description is provided in Table. Our participants, all males, ranged from 24 to 40 years with an average body mass index of ~25.2 kg/m 2 . As shown in Table,
Sports2025,13, 403 6 of 10 the MCSA of BF long headand log10NVO2run11was negatively related (r =−0.51;p= 0.03); however, we did not find any relationships between the MCSA of BF short head,ST and SM and log10NVO2Run11 (r = 0.21,−0.34,−0.20;p= 0.37, 0.13, 0.37, respectively). Table demonstrates a negative relationship between POTV10 and log10NVO2Run11 (r =−0.44; p= 0.05). Our multiple linear regression yielded a significant negative relationship only be- tween MCSA of BF long headand log10NVO2Run11 (r 2 = 0.31;p= 0.02,β−coefficient = 0.54) as shown in Table. Table 2.Correlation table. MCSA POTV10 LogNVO 2Run11 MCSA SM r = 0.07 r = −0.20 p= 0.76 p= 0.37 MCSA ST r = 0.17 r = −0.34 p= 0.47 p= 0.13 MCSA BF short head r =−0.02 r = −0.21 p= 0.95 p= 0.37 MCSA BF long head r =−0.07 r =−0.56 p= 0.79 p= 0.03 POTV10 r = 1 r =−0.44 p= 0.05 * Abbreviations: MCSA, maximum cross-sectional area; SM, semimembranosus; ST, semitendinosus; POTV10, potentiated velocity at 10 ms into the concentric phase of the leg press throw; LogNVO2Run11; logarithmic converted value of net running VO2at 11 km/h. * Trend toward significance. Bolded correlation between MCSA BF long headand LogNVO2Run11 indicate statistical significance. Table 3.Multiple regression model relating Log 10NVO 2Run11 with BF long headMCSA and POTV10. Model Summary R 2 = 0.31. Variable Unstandardized B B-Coefficients Partial r 95% CI (Lower CI–Upper CI) Sig. (p-Value) Constant 53.46 41.72–65.20 <0.001 POTV10 −17.71 −0.23 −0.27 −52.88–17.47 0.30 MCSA BF long head −0.01 0.54 −0.55 −0.01–−0.001 0.02 Abbreviations: CI, confidence interval; POTV10, potentiated velocity at 10 ms of concentric phase of; MCSA, maximum cross-sectional area; BF, bicep femoris;β-coefficients display changes in SD in dependent variable per SD change in independent variable. Boldedp-value indicates statistical significance between MCSA BF long head and LogNVO2Run11. 4. Discussion The main finding of our study was the negative relationship between net running VO2 and BF long headcross-sectional area. Since running economy is the inverse of the net running VO2, our results show that cross-sectional area of BF long headis a positive contributor to running economy in male recreational runners. To the best
BF long head and LogNVO2Run11. 4. Discussion The main finding of our study was the negative relationship between net running VO2 and BF long headcross-sectional area. Since running economy is the inverse of the net running VO2, our results show that cross-sectional area of BF long headis a positive contributor to running economy in male recreational runners. To the best of our knowledge, this is the first study to preliminary illustrate the importance of the MCSA of BF long headtoward running economy in male recreational runners. In fact, MCSA of BF long head,independent of POTV10, predicted log10NVO2Run11 in our study. Importantly, we noted that BF short head, ST, and SM did not play a role in running economy [19]. Since poor running economy may contribute to fatigue and fatigued muscles may increase incidence of injury, it is probable that undersized/weaker BF long headfosters not only poor running economy but may also explain, in part, reduced endurance performance and increased risk of hamstring injury [4]. Thus, our preliminary findings have direct and indirect clinical implications for running performance enhancement and hamstring injury prevention.
Description
The study investigates the relationship between biceps femoris long head size and running economy in male recreational runners.