Although passive static stretching is commonly used during the warm-up procedures of several sports, the effects of this technique on subsequent muscular performance are currently under debate (3,21,22).
Indeed, recent studies showed that pre-exercise static stretching may negatively affect force and power development (2,13,19,21,23). Additionally, an angle-specific effect of pre-exercise static stretching was reported. Examining isometric knee extension strength at different knee angles, Nelson et al. (17) reported that the negative influence of a pre-exercise bout of passive stretching was more evident at working knee angles near full joint extension. From a practical point of view, this means that in those sports requiring lower limbs to perform maximally at working angles near full knee extension (see, e.g., sprinting from starting block, fencing, rugby, American Football) the negative effects of pre-exercise stretching on muscle performance might be enhanced.
Several sports require athletes to perform from a semisquatted position to favorably exert explosive power during competition (e.g., basketball, volleyball, soccer, and sprinting). Therefore, it is of interest to study the possible detrimental effects of stretching on power muscular performance during sports activities requiring more complex movements (as jumps) at different knee working angles and at various levels of muscle preactivation (as those required to hold different starting positions).
The squat jump (SJ) at different knee starting angles may be a good model for these conditions, because the lower limbs extensor muscles are preactivated to different levels to hold the various starting positions and the jump performance is strongly related to starting knee angle. Furthermore, the SJ position is similar to those adopted by athletes as starting posture prior explosive actions (power position) (11).
Currently available data on the effects of static stretching on SJ were produced at standardized knee angles of 90°-100° and reported conflicting results (4,5,12,19,23,25). This typically corresponds to the starting position maximizing force output (18,20). However, we hypothesized that at other starting angles of SJ, characterized by different state of muscle preactivation and initial elongation, the effects of the stretching on the subsequent power output may be different.
The purpose of this study was therefore to evaluate the acute effects of a prior-exercise static-stretching routine on SJ performance at different knee starting positions. This was done to assess whether the possible acute detrimental effect of stretching on explosive muscle performance is joint angle-specific and dependent on the preactivation state of stretched muscles.
Experimental Approach to the Problem
In the present investigation, we addressed the possible effects of static stretching on lower limbs explosive power with a pre-post assessments of SJ performance. The main outcome variables were vertical peak force (Fp), velocity (Vmax), acceleration (Amax), and maximal power (Pmax). Peak force was chosen to evaluate the effects of stretching mainly on the cross bridges formation, and velocity and acceleration could reveal some possible additional effects of stretching on muscle performance because of neuromuscular or central inhibition. In addition, we studied the relationship between SJ performance and 4 different knee starting angles by a repeated measures analysis of variance (ANOVA; prestretching and poststretching). The starting position that enabled maximal power development was estimated through the analytically reconstructed relationship between maximal power and starting angles. The preactivation state of the quadriceps was estimated calculating the starting knee moment by a biomechanical analysis applied to the images of each starting position during the control session, photographed on the sagittal plane.
Seventeen male subjects (age 23 ± 3 years; height 179 ± 5 cm; and body mass 74 ± 6 kg) participated in this study. They were currently active in recreational or competitive sports (mainly track and field and soccer: None of them were involved in jumping sports, such as high jump, long jump, or volleyball). All the subjects were actively training 3-4 times per week, for 90 ± 30 minutes per training session. Testing was performed during the 8th and the 11th weeks of the regular season in soccer players and during the indoor precompetitive period in track and field athletes. Moreover, they were free of recent lower limbs injury, and they were asked to maintain their normal activity over the whole study duration. Subjects were not allowed to consume coffee, tea, or other stimulants 2 hours before the beginning of the experimental procedure. Participants were preliminarily informed about the possible risks of the experimental procedures. A written informed consent to participate in the study was obtained from all the enrolled subjects before the beginning of the study, and the experimental protocol was preliminarily approved by the local Institutional Review Board. The staff of the soccer and track and field teams also approved athletes' participation in this study.
The experimental procedure is shown in Figure 1. Supervised familiarization sessions with this study testing procedures were undertaken before the study by all participants.
Control and stretching sessions were performed on different days in random order. After a standardized warm-up (8-minute running on a motorized treadmill), subjects were submitted to a control or intervention (i.e., static-stretching) session. During the intervention condition (ca. 10 minutes), subjects were requested to perform lower limbs static-stretching exercises (4 stretches of 30 seconds, with 30-second rest for each muscle) after warm-up. Stretching exercises were performed bilaterally with the end position considered as the point of discomfort. The muscles stretched were the ankle plantar flexors and quadriceps, because of their significant contribution to SJ performance (14). The ankle plantar flexor muscles were simultaneously stretched with the subject in supine position by maintaining the dorsiflexion on the foot. The quadriceps muscles were stretched in a standing position with the dorsal side of the foot placed on a horizontal bar approximately positioned at the buttocks level. The experimenter then pushed the knee backward until the point of discomfort (Figure 2). During the control condition, subjects observed 10-minute rest, before SJ.
Immediately after either the control or intervention procedure, subjects performed a series of 2-legged unloaded SJs on a force platform (4 Jump, Kistler, Zurich, Switzerland) at each of 4 different knee starting angles (50°, 70°, 90°, and 110°). The full knee extension was assumed to be 0°, and increasing of knee angle means moving toward full flexion. In this protocol, we used the SJ to (a) avoid force potentiation because of the stretch-shortening cycle (as in countermovement jump) and (b) minimize the contribution of muscle elastic elements to jump performance (3).
Before each jump, the subjects held the selected static squat position for about 1 second with each angle monitored by an electrogoniometer (Biopac Systems Inc., Goleta, CA, USA). For each starting angle were performed 5 SJs interspersed with passive rest. Between the first 3 SJ bouts, subjects recovered 40 seconds and thereafter 90 seconds to avoid cumulative fatigue. The sequence of the different starting angles was not randomized, to minimize the possible confounding effect of the postactivation-potentiation phenomenon (9).
To account for the difference in muscle preactivation during the 4 SJ conditions (i.e., 50°, 70°, 90°, and 110°), the quadriceps starting knee moment was calculated. This assuming that, to maintain the static equilibrium of the whole system, the internal knee moment (mainly because of quadriceps activation) must equilibrate the external knee moment (because of the weight of the part of the body above the knees [head-arms-trunk-thighs]).
Each starting position during the control session has been photographed on the sagittal plane by a digital camera (Coolpix 8400, Nikon, Tokyo, Japan), and the images were then analyzed by an image-analysis software (NIH Scion Image, version 4.0). Each image was calibrated by a known length (the length of the force platform = 920 mm), which was measured at the interception of the sagittal plane of the subject with the horizontal plane of the platform. The following parameters were then calculated:
- (a) the center of mass of the head-arms-trunk complex (CoMHAT), which, according to Dempster (7) was placed at 39.6% of the distance head vertex-hip rotation center;
- (b) the CoM of the thigh (CoMT), which was placed at 40.95% of the thigh length (6);
- (c) the arm of the gravity force at the knee (dW), calculated as the distance between the rotation center of the knee and the gravitational force (WHATT) because of weight of the complex head-arms-trunk-thighs (calculated as 88.6% of the total body weight, according to De Leva ).
The CoM of the complex head-arms-trunk-thighs (CoMHATT) was then positioned on the line linking the CoMs of HAT and thighs, at a distance of 12.5% from CoMHAT, according to the ratio of weights of those segments (WT/WHATT = 0.125).
Finally, the starting knee moment was calculated as Mk = ½WHATT × dW, assuming a symmetric distribution of loads between the 2 legs. The method for starting knee moment calculation is shown in Figure 3.
The best 3 jumps of each series were chosen, on the basis of the following criteria: (a) the absence of a countermovement at the beginning of the SJ (defined as a decrease of at least 3% of the force corresponding to subject weight recorded by the force platform); (b) the maximal vertical height achieved in the jumps without any recognized countermovement.
Curves of force-time, acceleration-time, velocity-time, and power-time were calculated from the ground reaction force records acquired from the force platform, by means of a custom made software (National Instrument Inc., Austin, TX, USA). Acceleration-time curves were calculated by dividing the force-time record by the jumper's body mass, whereas the velocity-time curve was obtained by numerically integrating the acceleration-time curve with respect to time. The power-time curve was then calculated by multiplying force and velocity. Peak values of each curve from time = 0 to the beginning of the jump fly phase (where F = 0) were finally calculated as peak force (Fp), maximal velocity (Vmax), maximal acceleration (Amax), and maximal power (Pmax).
The optimal starting angle of SJ (OSASJ) was estimated for each subject in both control and stretching conditions as follows: (a) Pmax (average of the best 3 jumps) was plotted as a function of knee starting angle; (b) a second-order regression curve was calculated by the least-square method, obtaining the coefficients a, b, and c of the following equation: y = ax2 + bx + c (parabola); (c) the OSASJ was then calculated as the abscissa of the parabola vertex as OSASJ = −b/2a. An OSASJ calculation in a representative subject is shown in Figure 4.
If not otherwise stated, data are expressed as mean ± SD. All parameters were normally distributed (Kolmogorov-Smirnov test). Sample size was calculated from the pooled estimate of within-group standard deviations derived by preliminary data. The detection of a 5% decrease in SJ performance (SJ height and maximal muscular power, p = 0.05, 2 sided) because of stretching application (with β = 0.20 and α = 0.05) would require a sample of 7 and 10 subjects for experimental group, respectively. A 2-way ANOVA for repeated measures (followed by a Fisher Least Significant Difference post hoc test) was used to test the null hypothesis of no effects of stretching on performance parameters at each different knee starting angle, with stretching and angles as main factors. A 1-way ANOVA was used to test the null hypothesis of no changes of starting knee moment with respect to knee starting angles. A paired Student t test was used to verify the hypothesis of no differences between OSASJ in control and stretching conditions. The test reliability for the dependent variables has been calculated by evaluating the coefficient of variation (calculated as [SD/mean]100) of each parameter in the 5 SJ series executed at the 4 different angles in the control condition, according to Hopkins et al. (10). The results have been added to Table 2. Finally, 1 hour after the end of each control procedure, the subjects were required to perform a last series of SJs at a 90° knee starting angle. The Intraclass Correlation Coefficient between the intraprotocol and the postprotocol 90° SJ series was finally calculated.
The alpha level for significance was set at p ≤ 0.05.
The estimated weight of head-arms-trunk-thighs (WHATT), DW, and starting knee moment in the 4 starting positions of SJ (control session) are shown in Table 1. Starting knee moment significantly increased (p < 0.01) with knee starting angles.
The squat jump height (SJH) significantly increased with knee starting angles in both control and stretching conditions (p < 0.01). Squat jump height values were however significantly lower in the stretching condition (p = 0.02) (Table 2). The Fp significantly decreased with increasing knee starting angles in both control and stretching conditions (p < 0.01). However, Fp values were significantly lower in the stretching condition (p < 0.01), especially at 50° and 90° knee angles, with significant interaction between main factors (p = 0.005) (Table 2). The Amax decreased with increasing knee starting angles in both the control and stretching conditions, being significantly lower after stretching at 50° only (p < 0.001), with significant interaction between main factors (p < 0.05) (Table 2).
The Vmax increased with knee starting angles in both the control and stretching conditions (p < 0.01) and was significantly lower in the stretching condition (p < 0.001) especially at lower knee angles (50°, p < 0.01 and 70°, p = 0.02) (Figure 5A).
The Pmax was obtained at 90° knee starting angle in both control and stretching conditions and was significantly lower (p < 0.01) after stretching, especially at lower starting angles (50°, p < 0.01, 70° and 90°, p = 0.06) (Figure 5B), with significant interaction between main factors (p < 0.05).
The estimated OSASJ was significantly higher in the stretching condition (p = 0.024, Figure 6). The coefficient of variation calculated for each SJ series during the control procedure never exceeded 5.0% (Table 2). The ICC between interprotocol and postprotocol SJH (SJ series al 90° starting angle) during the control procedure was 0.89.
The main finding of this study was the angle-dependent detrimental effect of pre-exercise static stretching on all SJ performance variables, such an effect being higher at lower knee angles (50° and 70°).
This is similar to what was previously reported by McNeal et al. and by Young et al., who showed a lower SJ performance (3-4%) after static stretching (15,25). However, conflicting results were reported by other authors, who found no effect of pre-exercise routine on successive SJ performance (3,12,19,24). Rubini et al. (21) in their review claimed that differences in SJ performance were due to differences in the stretching routines used in those studies. Specifically in those studies that reported post-static-stretching impairment in SJ performance hamstring muscles were not exercised (21).
Therefore, it may be hypothesized that the stretching applied to hamstring muscles caused a reflex activation of the quadriceps, thus masking the possible detrimental effects of stretching on subsequent force development. From a practical point of view, this possible contribution of hamstring muscle stretching to the global effects of stretching on jump performance will deserve future experiments, as stretching of both flexor and extensor muscles is commonly practiced in many sports activities. However, to minimize this possible confounding effect, in our protocol, we decided to stretch only the quadriceps and the triceps surae, but not the hamstrings.
In the 6 aforementioned studies, SJs were performed at 90° (most papers) or 100° starting angles. In the present protocol, we decided to perform SJs at other different knee starting angles. Indeed, the conflicting results yielded in previous studies might also depend on the choice of the 90° knee starting angle selected in these experiments. In fact, it has been proposed that the negative effects of stretching on muscle performance is most apparent at knee working angles near full extension (16). In this study, we found that at starting angles near full extension of lower limbs, the detrimental effect of stretching on jumping performance is more evident. Conversely, at 90° knee angle, the differences between stretched and nonstretched lower limb performance are less relevant, being significant only for Fp. From a methodological point of view, this suggests that the use of lower angles of SJ (e.g., 50° and 70°) may help to emphasize the effects of stretching on explosive muscular power output.
Starting knee moment increased with knee starting angles, suggesting that the preactivation state of the quadriceps rises with knee angles. Therefore, peak force, maximal velocity, and power output during jumping performance are reduced by stretching especially at those knee starting angles at which the overall muscular activation, estimated by starting knee moment, is lower. Thus, the inhibitory effects of stretching on SJ performance not only appear to be knee angle-specific but may also be related to the initial quadriceps moment. One possible limitation of this study is that our model assumes that all of the moment about the knee is due to quadriceps activity. This is not completely true, because of a possible co-contraction of the hamstrings and other muscles. However, from a practical viewpoint, we considered this assumption sufficiently valid in the case of this static analysis, even though it provides only an approximation of the real knee moment.
A possible hypothesis to explain these data is that the increased compliance of the musculo-tendonous unit induced by stretching may have effectively reduced force transmission to the bones (14). If this was the case, such an effect may be more apparent when the muscle and tendon complex is less preactivated (i.e., the global tension of the whole musculotendonous system is lower), as in the case of lower knee starting angles. Accordingly, this also results in a change of the optimal angle of power output development (OSAsj), which was shifted toward higher angles (about +8°, Figure 6).
Overall, these data suggest that the stretched muscles need to be more preactivated than nonstretched muscles to give a similar maximal power output at the same working angle. A possible practical application of this finding is that after a stretching session, the knee might start from a more flexed position to obtain the maximal power output. It is well acknowledged that during the knee extension exercise, the anterior cruciate ligament is loaded at low knee angles and increases as the knee angle decreases. Thus, after stretching, the more squatted position necessary to obtain the maximal power output during knee extension could better preserve the anterior cruciate ligament from elongation injuries.
Our results do not allow us to clarify the mechanisms underlying the effects of stretching on jumping performance. However, it may be hypothesized that the increased length of the musculotendonous complex induced by the stretching technique may have caused the muscle fibers to work in the ascending (i.e., in a suboptimal curve portion) rather than in plateau (i.e., in the optimal curve portion) limb of the force-length curve (8). In addition, this work did not investigate whether the stretch-induced decreases in force and power could be attributed to impairments in neural output to the muscles. However, altered torque/fascicle length relationship may also have influenced the neural activation patterns. Avela et al. (1) suggested that the increased compliance of the muscle could decrease the resting discharge of the muscle spindles, leading to disfacilitation of the α-motoneuron pool. Nevertheless, some studies provided evidence that such an effect is not likely to be long lasting. For example, Fowles et al. (8) reported a significant decrease in the motor unit activation after static stretching that lasted 30 minutes, whereas the mechanical effects of stretching on muscle power output persisted for 1 hour. As our experiment lasted at least 45 minutes from the stretching bout application, it is unlikely that our results may have been affected only by the possible neuromuscular inhibition.
In conclusion, this study addressed the effects of stretching on SJ performance in 4 different starting positions (corresponding to different initial quadriceps moments), to see whether the effects of stretching on complex explosive muscle performance (vertical jumps) depend on the starting position and on the preactivation state of the extensor muscles. We found that, at lower knee angles (50° and 70°), the effects of stretching seem to be detrimental to performance, whereas at higher angles (90° and 110), such effects are negligible.
This study has implications for all those complex sport tasks requiring the knee joint to maximally perform at some critical angles lower than 90°. This may be the case for leg movement in swimming (e.g., front crawl), cycle sprinting, rugby union (e.g., scrum position), fencing (e.g., lunge), and many positions of Greco-Roman wrestling. From a practical viewpoint, this suggests that for certain power activities, the practice of static stretching during the warm-up procedures may be detrimental for subsequent muscular performance.
The authors sincerely thank Dr. Andrea Bosio, Dr. Eloisa Limonta, and Ing. Massimiliano Sacchi for their valuable technical assistance, and all the study participants.
The authors declare no professional relationships with companies or manufacturers who will benefit from the results of the present study and state that the results of the present study do not constitute endorsement of the product by the authors or the NSCA.
1. Avela, J, Kyrolainen, H, and Komi, PV. Altered reflex sensitivity after repeated and prolonged passive muscle
stretching. J Appl Physiol
86: 1283-1291, 1999.
2. Behm, DG, Bambury, A, Cahill, F, and Power, K. Effect of acute static stretching on force, balance, reaction time, and movement time. Med Sci Sports Exerc
36: 1397-1402, 2004.
3. Church, JB, Wiggins, MS, Moode, FM, and Crist, R. Effect of warm-up and flexibility treatments on vertical jump performance. J Strength Cond Res
15: 332-336, 2001.
4. Cornwell A, Nelson AG, and Heise, P. Acute effects of passive muscle
stretching on vertical jump performance. J Hum Movement Stud
40: 307-324, 2001.
5. Cornwell, A, Nelson, AG, and Sidaway, B. Acute effects of stretching on the neuromechanical properties of the triceps surae muscle
complex. Eur J Appl Physiol
86: 428-434, 2002.
6. de Leva, P. Adjustments to Zatsiorsky-Seluyanov's segment inertia parameters. J Biomech
29: 1223-1230, 1996.
7. Dempster, WT. Free-body diagrams as an approach to the mechanics of human posture and locomotion. In: Biomechanical Studies of the Musculoskeletal system
. E.F.G., eds. Springfield, IL: Charles C. Thomas, 1961.
8. Fowles, JR, Sale, DG, and MacDougall, JD. Reduced strength after passive stretch of the human plantarflexors. J Appl Physiol
89: 1179-1188, 2000.
9. Hodgson, M, Docherty, D, and Robbins, D. Post-activation potentiation: underlying physiology and implications for motor performance. Sports Med
35: 585-595, 2005.
10. Hopkins, WG. Measures of reliability in sports medicine and science. Sports Med
. 30: 1-15, 2000.
11. Johnson, J. The Power Snatch: How to Use It and Why. NSCA J
5: 14-18, 1983.
12. Knudson, D, Bennett, K, Corn, R, Leick, D, and Smith, C. Acute effects of stretching are not evident in the kinematics of the vertical jump. J Strength Cond Res
15: 98-101, 2001.
13. Kokkonen, J, Nelson, AG, and Cornwell, A. Acute muscle
stretching inhibits maximal strength performance. Res Q Exerc Sport
69: 411-415, 1998.
14. Luhtanen, P and Komi, RV. Segmental contribution to forces in vertical jump. Eur J Appl Physiol Occup Physiol
38: 181-188, 1978.
15. McNeal, JR and Sands, WA. Static stretching reduced power production in gymnasts. Technique
(Nov/Dec): 5-6, 2001.
16. Nelson, AG, Allen, JD, Cornwell, A, and Kokkonen, J. Inhibition of maximal voluntary isometric torque production by acute stretching is joint-angle specific. Res Q Exerc Sport
72: 68-70, 2001.
17. Nelson, AG, Guillory, IK, Cornwell, C, and Kokkonen, J. Inhibition of maximal voluntary isokinetic torque production following stretching is velocity-specific. J Strength Cond Res
15: 241-246, 2001.
18. Pincivero, DM, Salfetnikov, Y, Campy, RM, and Coelho, AJ. Angle- and gender-specific quadriceps femoris muscle
recruitment and knee extensor torque. J Biomech
37: 1689-1697, 2004.
19. Power, K, Behm, D, Cahill, F, Carroll, M, and Young, W. An acute bout of static stretching: effects on force and jumping performance. Med Sci Sports Exerc
36: 1389-1396, 2004.
20. Rassier, DE, MacIntosh, BR, and Herzog, W. Length dependence of active force production in skeletal muscle
. J Appl Physiol
86: 1445-1457, 1999.
21. Rubini, EC, Costa, AL, and Gomes, PS. The effects of stretching on strength performance. Sports Med
37: 213-224, 2007.
22. Smith, CA. The warm-up procedure: to stretch or not to stretch. A brief review. J Orthop Sports Phys Ther
19: 12-17, 1994.
23. Taylor, DC, Brooks, DE, and Ryan, JB. Viscoelastic characteristics of muscle
: passive stretching versus muscular contractions. Med Sci Sports Exerc
29: 1619-1624, 1997.
24. Unick, J, Kieffer, HS, Cheesman, W, and Feeney, A. The acute effects of static and ballistic stretching on vertical jump performance in trained women. J Strength Cond Res
19: 206-212, 2005.
25. Young, WB and Behm, DG. Effects of running, static stretching and practice jumps on explosive force production and jumping performance. J Sports Med Phys Fitness
43: 21-27, 2003.