Int J Sports Med 2015; 36(07): 604
DOI: 10.1055/s-0035-1554629
Letter to the Editor
© Georg Thieme Verlag KG Stuttgart · New York

Asymmetry after Hamstring Injury in English Premier League: Issue Resolved, or Perhaps Not?

Further Information

Publication History

Publication Date:
15 June 2015 (online)

M. Brughelli1, J.-B. Morin2, J. Mendiguchia3: Asymmetry after Hamstring Injury in English Premier League: Issue Resolved, Or Perhaps Not?

Response to Brughelli et al. Letter to the Editor

Thank you for the opportunity to respond to the letter by Brughelli et al. regarding our recent paper “Asymmetry after hamstring injury in English Premier League: Issue resolved, or perhaps not?”. Brughelli et al. were the first to highlight the issue of asymmetry after hamstring injury, and we wish to thank them for continuing debate in this area. In their letter, Brughelli et al. reiterated the importance of hamstring injuries as a major concern in all codes of football and the need to rigorously investigate this with careful scrutiny of data and thorough review processes. We fully agree. They however (1) question whether this had been the case in our recent paper, and (2) stated that “the data presented are invalid, leading to untrustworthy results”. We beg to disagree with the latter. We will first offer further explanation of specific technical issues that were raised based on the data presented in Figure 2 of our paper, addressing each of the numbered remarks from Brughelli et al. We notice first that the references cited by Brughelli et al. are not based on the some model adopted in our study, which may lead to inaccurate comparisons between data collection methods.

  1. Forces do not return to zero
    The sprint assessment instrument adopted in our study was a non-motorized Woodway® curved treadmill, which differs from the methods adopted in the studies used by Brughelli et al. to reference the present criticism to our work. This fact likely justifies why the vertical force profiles highlighted by the example graph do not return to zero. The nature of this treadmill with its curved anterior and posterior aspects, promotes an earlier contact phase as well a later take off. This, combined with the high speed of movement that was requested of the participants during the sprint protocol leads to dramatically shortened flight phases between contacts, if indeed there is flight at all. Furthermore, because the treadmill belt is mounted on the force transducers, the deceleration of the treadmill belt during flight will lead to a sagittal plane torque on the force transducers, again preventing the force immediately dropping to zero in between steps. After all though, it was the distinct peak vertical forces that were used to identify steps with the synchronously recorded video, and whilst we agree with Brughelli et al. that there are limitations to the frame mounted force transducers this was not of immediate concern for our data analysis.

  2. Horizontal forces do not demonstrate distinct braking and propulsion phase
    The curvature of the treadmill ranges from 0–25% incline (both at front and back), and to accelerate the treadmill one has to mimic uphill running. As shown in the study by Gottschall and Kram [1], running uphill will progressively modify the horizontal ground reaction force profile; increasing the magnitude of the propulsive peak whilst significantly decreasing the braking peak. They reported that an uphill inclines of 3° resulted in 19% less braking force, 6° resulted in 38% less braking force, and a 9° incline resulted in 54% less braking force. For the same inclinations (3°, 6° and 9°), propulsive forces increased by 28, 50 and 75% respectively. Hence, the curved shape of the treadmill justifies a different horizontal force pattern.

  3. Peak horizontal forces at maximum velocity were considerably lower than in previous studies
    Taking into consideration the altered biomechanics due to the curved nature of the treadmill, as described above, we disagree that horizontal forces are different than what has been previously reported. Our results for horizontal force development are in agreement with existing literature where a similar treadmill model was used [2] [3]. In the latter study [3] participants performed 2 trials of a 30 s maximum sprint on a non-motorized treadmill similar to the model used in our study. Results from this study for peak horizontal force ranged between 183–352 N for the first trial and 220–358 N for the second trial. Our findings are in these ranges (196–211 N). The observed force profile further appropriately represented the forces needed to accelerate and decelerate the treadmill throughout the trial, with maximum forces during the first part of the acceleration phase, gradually reducing forces as the treadmill speed reaches its maximum, slightly positive forces to overcome the frictional resistance of the treadmill during the steady state phase, and negative forces to help decelerate the treadmill after that.

  4. Number of observed peaks in horizontal compared to vertical forces
    We agree that the force profile of individual steps for horizontal forces is considerably less clear than for the vertical forces. The short flight phase together with the fact that the treadmill frame was not fixed to the floor are 2 reasons that can help explain this. As both force components were recorded simultaneously, we manually verified the vertical peak forces through synchronous video recording, and then used the horizontal force values at the time of those peaks.

We hope that with these responses to the listed technical issues we have been able to demonstrate the necessary rigour with which the data was processed, analysed, and interpreted. Data analysis was done manually in MS Excel from raw exported data rather than through an automated process. We do not see why programming this process into software, as Brughelli et al. suggest, would improve the rigour of this kind of work. In fact, we would argue that the automation process should be done by an engineer who is duly trained to write software and implement accuracy checks, rather than stimulate sport scientists.

Finally, we would like to reinforce the fact that the technical issues 1, 3 and 4 were actually raised by the reviewers. Importantly though, the reviewers recognised that the aim of our study was not to validate this particular method but to explore whether its application in a club setting can reveal previously reported asymmetries. As we stated in the discussion section, there is a ‘need to rigorously test whether modifications to an assessment protocol eliminate its capacity to actually reveal deficiencies’. As such, we believe that the limitations of the work were duly considered during a thorough review process as is common practise for the IJSM. The phrasing of key messages that form the basis of the translational nature of our work were re-worded based on reviewers’ suggestions, such that at no time it implied that previous findings are rejected. In fact, with our paper we aimed to stimulate care in data collection and data interpretation in an applied setting, and from that perspective we seem to fully agree with Brughelli et al.

  • References

  • 1 Gottschall JS, Kram R. Ground reaction forces during downhill and uphill running. J Biomech 2005; 38: 445-452
  • 2 Mangine GT, Fukuda DH, Towsend JR, Wells AJ, Gonzalez AM, Jajtner AR, Bohner JD, Lamonica M, Hoffman JR, Fragala MS, Stout JR. Sprint performance on the Woodway Curve 3.0 TM is related to muscle architecture. Eur J Sport Sci 2014; epub ahed of print
  • 3 Mangine GT, Hoffmane JR, Gonzalez AM, Wells AJ, Towsend JR, Jattner AR, McCormack WP, Robinson EH, Fragala MS, Fukuda DH, Stout JR. Speed force and power values produced from nonmotorized treadmill test are related to sprinting performance. J Strength Cond Res 2014; 28: 1812-1819