Proximal Hamstring Tendinopathy in Speed-Based Athletes: Tension, Compression and Tolerance

Proximal hamstring tendinopathy is usually explained to athletes in a single sentence: the tendon has been loaded too much. The explanation is incomplete, and its incompleteness matters in practice, because it points towards a single intervention. Reducing running volume changes one input in a condition whose onset appears to depend on several, each of which can be modified independently and none of which compensates for the others. Sprinters, hurdlers and team-sport athletes who cut running volume while leaving the rest of their exposure unchanged often improve for a few weeks, then relapse once speed work returns.

At any moment, the state of the enthesis reflects a balance between the mechanical demand repeatedly placed on it and the demand it can currently tolerate without symptoms. Both sides of that balance have several determinants. Demand comes from running velocity, from the hip angles at which force is produced, from the amount of the day spent in sustained hip flexion, and from the share of the hip extension requirement met by the gluteal group. Capacity comes from the loading history of the tissue, which builds tolerance when exposure is progressive and loses it when exposure is withdrawn. This may partly explain why trials comparing one modality with another have produced modest and largely equivalent results, since whether an intervention helps will depend on whether it addresses the input operating in that particular athlete.

Causation, progressive loading and prevention of recurrence are therefore best treated as one continuous problem. The reasoning that identifies which inputs are operating also specifies the loading programme, and the same programme, held at a maintenance dose, is what lowers the risk of the condition returning.

The Combined Load Problem at the Ischial Enthesis

Tendons are built to transmit tension, but a tendon that wraps around a bony prominence is compressed as well as stretched. That compression appears to drive the tissue changes seen in overuse tendinopathy: fibrocartilaginous metaplasia, an accumulation of large proteoglycans and a breakdown of collagen structure (Cook and Purdam, 2012). At the ischial tuberosity, compression rises as the hip flexes, and when adduction is added to that flexion, shear at the insertion rises as well (Campos-Villegas et al., 2024). This helps to explain why prolonged sitting, deep hip flexion positions and static hamstring stretching are so often reported as aggravating factors, and why compression has become central to the way the condition is currently modelled (Goom et al., 2016).

Sprinting imposes high tensile demands at hip angles that are already compressive. Musculoskeletal modelling of overground locomotion indicates that peak hamstring force and peak musculotendon length both occur during terminal swing. At sprinting speed, terminal swing loading substantially exceeds stance-phase loading, which is not the case in walking or jogging (Schache et al., 2010). The same work indicates that gluteus maximus supplies most of the stance-phase hip extension torque, while the hamstrings dominate terminal swing. Electromyography during maximal overground sprinting shows biceps femoris to be highly active immediately before and after foot contact, while the medial hamstrings are relatively more active during late stance and mid-swing (Higashihara et al., 2015).

Cumulative tensile load and cumulative compressive load are therefore two separate exposures that happen to coincide in this population. A tendon may tolerate either one in isolation and become symptomatic under the combination. This is one reason that weekly running volume, taken on its own, describes only part of what the athlete has been doing.

Established and Potential Contributors of Proximal Hamstring Tendinopathy

Several contributors have enough mechanical support, and enough consistency with clinical observation, to be treated as working assumptions. The first is exposure to high hip flexion under load. Block starts, early acceleration, uphill running, hurdling and deep hinge patterns in the gym all combine tension with compression, and an increase in any of them frequently precedes symptom onset. The second is sustained hip flexion outside training. Prolonged sitting, particularly on hard surfaces and while driving, holds the enthesis under compression for hours at a time and is among the more commonly reported aggravating factors (Goom et al., 2016; Campos-Villegas et al., 2024). Athletes with heavy travel, study or office commitments accumulate a compressive exposure that no training metric records.

The third is a change in the character of high-speed running exposure rather than in its total amount. This might be a shift in the balance between acceleration and maximal velocity work, the reintroduction of block starts or hurdle sessions, a move onto spikes or a different surface, or the addition of hill running. Each of these alters the distribution of hip flexion angles at which force is produced, without necessarily changing the distance covered. The fourth is prior hamstring strain injury. In track and field athletes with a previous unilateral injury, biceps femoris long head activation during late swing was lower on the injured side, together with a more flexed knee and a shorter musculotendon length on that limb (Higashihara et al., 2019). The fifth is capacity. Where the gluteal contribution to stance-phase hip extension is limited, or hamstring strength at long lengths is low, a greater share of the demand may reach the proximal tendon at the angles it tolerates least.

These contributors appear to interact rather than simply accumulate, which is one reason that management aimed at a single variable often produces partial results. Reduced gluteal capacity raises the demand transmitted to the hamstrings during stance, and a day spent largely seated lowers the tolerance available for the session that follows. Two athletes with similar training data may therefore arrive at quite different tissue states.

Several further contributors remain plausible without being demonstrated. Anterior pelvic tilt increases hamstring musculotendon length at any given hip angle, so the same sprint kinematics may produce different tendon strain depending on how the pelvis is oriented. Treatment-seeking cohorts skew female, with women making up 145 of the 217 patients in one shockwave therapy review (Arthurs et al., 2026), though such samples describe who presents for treatment rather than who develops the condition. Calcific change may be underrecognised and can be missed on magnetic resonance imaging (Elkousy et al., 2026), and the adjacent sciatic nerve may contribute to the symptom picture. Structural findings are only loosely coupled to symptoms in tendinopathy (Docking et al., 2015), so imaging remains more useful for differential diagnosis than for identifying cause.

"Cumulative tensile load and cumulative compressive load are therefore two separate exposures that happen to coincide in sprinters. A tendon may tolerate either one in isolation and become symptomatic under the combination."

What the Mechanism Implies for Exercise Selection

If the tissue is intolerant of tension combined with compression, then the primary progression variable is hip flexion angle rather than external load on its own. Exercise selection can then be judged on two criteria: which part of the hamstring group an exercise loads, and at which joint angles it does so.

Two studies make this practical. Bourne and colleagues (2017) compared ten common strength exercises using electromyography and functional magnetic resonance imaging. The 45-degree hip extension produced the highest ratio of biceps femoris to medial hamstring activation during the eccentric phase and selectively recruited the long hamstrings, whereas the Nordic hamstring exercise preferentially recruited semitendinosus (Bourne et al., 2017). A second group used high-density electromyography across nine exercises. The straight-knee bridge, upright hip extension and leg curl variants produced the highest overall hamstring activity; hip extension was the only biceps femoris-dominant pattern; and the proximal-to-distal distribution of activity varied by exercise and by contraction mode (Hegyi et al., 2019). Regional differences were pronounced in the Nordic and small in the stiff-leg deadlift (Hegyi et al., 2018).

Hip-dominant patterns therefore load the proximal tendon more directly than knee-flexion-dominant ones, but they also place the hip in more flexion. The exercises with the most direct effect on the tissue are consequently the ones most likely to provoke it when introduced too early. Knee-dominant work at low hip flexion is useful in the early phase because it separates the two variables: leg curl variants generate high overall hamstring activity with the hip near neutral. A systematic review of conservative interventions proposed hip flexion around 110 degrees with knee flexion between 45 and 90 degrees as the position that best targets the proximal tendon (Dizon et al., 2023). Experienced clinicians, by contrast, describe keeping hip flexion exposure low during early and middle rehabilitation, before a graded return to sport (Nasser et al., 2021). The two recommendations are compatible once the first is read as a late objective and the second as an early constraint.

The Principles Behind Progressive Loading

The rationale for progressive loading rests on two observations. The first is that tendon pathology may be better described as a continuum of tissue states than as a binary lesion, which allows an intervention to be matched to the state the tissue is currently in (Cook and Purdam, 2009). The second is that reducing load lowers demand and tolerance together, so rest can settle symptoms while leaving the athlete less able to withstand the exposure that will eventually return.

Meta-analysis of exercise interventions in healthy adults indicates that tendon stiffness and material properties respond to sustained loading. The size of the adaptation depends on loading magnitude rather than on contraction type, and effects are larger in interventions lasting twelve weeks or more (Bohm et al., 2015). These findings come from the Achilles and patellar tendons in asymptomatic participants, so their transfer to this condition is indirect. Two implications nonetheless seem reasonable. Loading magnitude has to rise over time for tolerance to rise with it, which means that a programme still using the load tolerated in week two is unlikely to produce adaptation. The traditional emphasis on eccentric work also appears less important than the emphasis on load, which widens the range of exercises available.

Progression is multi-dimensional as well. Compression is governed by hip flexion angle, tensile magnitude by external load, rate of loading by movement velocity and elastic demand, and cumulative exposure by weekly frequency, with non-training exposure contributing to each of them. Advancing one dimension at a time makes it possible to see what the tissue responded to, whereas advancing two at once is a common reason a programme stalls without an obvious explanation.

"The causes of proximal hamstring tendinopathy in speed-based athletes appear to lie in a set of interacting exposures rather than in a single overload event. Tension peaks in terminal swing and scales with running velocity. Compression at the enthesis scales with hip flexion and accumulates outside training as readily as within it. Gluteal capacity, pelvic orientation and prior injury then modify how that combined load is distributed."

A Staged Loading Progression

The sequence below has five stages, distinguished by hip flexion angle and loading character rather than by week number.

The first stage combines isometric loading in low hip flexion with removal of the compressive exposures that are maintaining symptoms, including prolonged sitting, static hamstring stretching and deep hinge positions. In patellar tendinopathy, isometric contractions produced greater immediate analgesia than isotonic contractions, and the size of that early response correlated with functional improvement at four weeks (Rio et al., 2017). This has not been tested directly at the proximal hamstring, so isometrics are better regarded as a means of enabling loading than as a treatment in themselves. Holds of thirty to forty-five seconds in bridge and isometric hip extension positions, repeated four to five times and performed most days, are a reasonable starting point. The compressive side of this stage is addressed through seat modification and more frequent standing breaks, since these exposures often last longer than training itself.

The second stage introduces heavy slow resistance in restricted hip flexion, with loading magnitude as the progressed variable. Bilateral and then unilateral hip extension patterns are performed with a controlled tempo, three to four sets of six to eight repetitions, two or three times weekly. These build tensile capacity while compression remains low, and the external load rises across weeks so that the stimulus does not plateau. Gluteal strengthening belongs in this stage rather than later, since restoring stance-phase load sharing lowers the demand transmitted to the hamstrings during running.

In the third stage, hip flexion angle becomes the progressed variable while load is temporarily held constant. The progression moves through Romanian deadlifts to increasing depth, 45-degree hip extension, long-lever bridging and single-leg hinge patterns towards the target position, after which magnitude begins rising again at the new angle. This stage is typically the longest and the one most often cut short, because symptoms usually settle well before the tissue tolerates deep flexion under substantial load.

The fourth stage adds energy storage and release, beginning with low-amplitude skipping and bounding before submaximal accelerations, which brings in rate of loading as a separate dimension. The fifth stage is sprinting. Terminal swing tensile demand rises steeply with running speed (Schache et al., 2010), so the move from submaximal to near-maximal running represents a step change, and it can be planned with its own frequency, distance and intensity variables alongside the resistance programme. Block starts and early acceleration combine deep hip flexion with high force, which makes them a late reintroduction despite the intuition that they are less demanding than maximal velocity work. Athletes with a prior strain on the same limb warrant attention to inter-limb loading as well as to local symptoms (Higashihara et al., 2019), and running retraining is recommended by experienced clinicians on limited clinical evidence (Barton et al., 2016).

A published case in a runner describes a comparable progression over twelve weeks. Pain fell from 7 to 1 on a visual analogue scale, and the VISA-H rose from 23 to 53 at six weeks and to 80 at twelve (Campos-Villegas et al., 2024). The VISA-H remains the only instrument validated in this population (Nasser et al., 2023). With a minimum clinically important difference of 22 points (Cacchio et al., 2014), it tracks direction over months rather than week-to-week change.

Preventing Recurrence, and Preventing the First Occurrence

The mechanism behind recurrence is reasonably well understood. Tolerance was raised by a progressive stimulus, and when that stimulus is withdrawn at the point of symptom resolution, tolerance falls back while the sporting demand does not, returning the athlete to the balance that produced the problem. Since tendon adaptation is larger in longer interventions (Bohm et al., 2015), twelve weeks of loading may be closer to the beginning of the adaptive process than to its completion. Experienced clinicians see prevention of recurrence as resting on two things: continuing to strengthen the hamstrings and the kinetic chain after symptoms settle, and managing workload on an ongoing basis (Nasser et al., 2021). Both are commonly abandoned once the athlete feels well.

A maintenance programme follows from this. Hip-dominant loading at moderate and progressively deeper hip flexion is retained year-round at a reduced volume, perhaps one or two sessions weekly rather than three, with enough load to remain a stimulus. Gluteal capacity is maintained so that stance-phase load sharing does not decline. Hills, blocks and hurdle work are reintroduced gradually at the start of each cycle rather than in concentrated blocks, since the redistribution of hip flexion angles appears to matter more than the total distance. Seated exposure is treated as a training variable during travel-heavy and competition-dense periods. Monitoring is helped by a characteristic of tendon pain, which tends to ease during activity. Pain on rising in the morning and pain after prolonged sitting are therefore earlier indicators than pain during running, and a brief weekly check of those two markers, alongside a load-based provocation test, may detect a change in tolerance before it limits training.

Preventing a first occurrence follows the same logic, with one qualification. Programmes including the Nordic hamstring exercise roughly halve the rate of hamstring strain injuries across sports and age groups (van Dyk et al., 2019), and the exercise belongs in a sprinter's programme for that reason. It does not, however, provide a proximal tendon stimulus, since it preferentially recruits semitendinosus (Bourne et al., 2017; Hegyi et al., 2018), whereas hip extension patterns load the long hamstrings and the enthesis more directly. A prevention programme built largely on the Nordic may therefore reduce strain risk while leaving the proximal tendon comparatively undertrained, which offers a plausible account of cases arising in otherwise well-conditioned athletes. No prospective prevention trial has been conducted in this condition, so these recommendations rest on mechanism and clinical consensus rather than on demonstrated efficacy.

Where Other Interventions Fit

Among the adjuncts, extracorporeal shockwave therapy has the strongest supporting evidence, with systematic reviews reporting improvements in pain and function and good tolerance in athletes who continue training (Rhim et al., 2024; Rau et al., 2025). A randomised trial nonetheless found no difference between six sessions of shockwave and six of individualised physiotherapy up to 52 weeks (Rich et al., 2025). Platelet-rich plasma has performed poorly (Levy et al., 2019), and surgery remains a late option, with results drawn from a refractory cohort of mean age 50 that transfer poorly to sprinters (Lefèvre et al., 2026).

Conclusion

The causes of proximal hamstring tendinopathy in speed-based athletes appear to lie in a set of interacting exposures rather than in a single overload event. Tension peaks in terminal swing and scales with running velocity. Compression at the enthesis scales with hip flexion and accumulates outside training as readily as within it. Gluteal capacity, pelvic orientation and prior injury then modify how that combined load is distributed. Several further contributors are plausible but untested, and treating them as hypotheses when planning an individual case seems the more defensible position.

Management and prevention can be understood as the same programme at different doses. Progressive loading raises the tolerance of the tissue, and it appears to do so in response to magnitude sustained over months rather than to any particular contraction type. Hip flexion angle, external load, rate of loading and non-training exposure are advanced one dimension at a time. Prevention is that same structure kept at a maintenance volume once symptoms have resolved, together with gradual reintroduction of the running exposures that combine tension and compression. Where a case is not responding, it is often more productive to ask which input has been left unchanged than which treatment should be substituted.

---

Frequently Asked Questions

Q: What actually causes proximal hamstring tendinopathy in sprinters?

No single factor appears to account for it. The best-supported mechanical account combines two loads: tension, which peaks during terminal swing and rises steeply with running velocity, and compression of the tendon against the ischial tuberosity, which increases with hip flexion and adduction (Cook and Purdam, 2012; Schache et al., 2010). Contributing exposures typically include increases in acceleration, hill or hurdle work, prolonged sitting outside training, a prior hamstring strain on the same limb, and limited gluteal or long-length hamstring capacity. These appear to interact rather than accumulate independently, and prospective risk-factor studies in this specific condition are lacking, so they are best treated as working assumptions.

Q: Why does sitting aggravate the condition so much?

Sitting places the hip in sustained flexion, which compresses the proximal hamstring tendon against the ischial tuberosity for extended and uninterrupted periods. Unlike training, this exposure is continuous rather than cyclical, it occurs without any accompanying adaptive stimulus, and it is not captured in any training-load metric. For athletes with heavy travel, study or desk commitments, modifying seated exposure with a wedge cushion, more frequent standing breaks and a reclined rather than upright seat position can be as influential as any change made in the gym.

Q: Which exercises target the proximal hamstring tendon most directly, and how should load be progressed?

Hip-dominant patterns reach it most directly. Hip extension exercises selectively activate the long hamstrings and are the only reliably biceps femoris-dominant pattern examined, whereas the Nordic hamstring exercise preferentially recruits semitendinosus (Bourne et al., 2017; Hegyi et al., 2019). Because these patterns also place the hip in greater flexion, they are introduced once low-flexion loading is tolerated. Progression then advances one dimension at a time: external load first at a restricted hip angle, then hip flexion angle at a temporarily held load, then rate of loading. This ordering follows from the finding that tendon adaptation depends on loading magnitude sustained over months rather than on contraction type (Bohm et al., 2015).

Q: How can recurrence be prevented once symptoms have resolved?

Recurrence commonly follows the withdrawal of the loading that produced resolution, because tolerance declines when the stimulus stops while the sporting demand does not. A maintenance dose of hip-dominant work at moderate to deep hip flexion, retained year-round at one or two sessions weekly with meaningful load, addresses the tissue directly, and gradual rather than concentrated reintroduction of hills, blocks and hurdle work at the start of each cycle addresses the exposure pattern (Nasser et al., 2021). It is worth noting that the Nordic hamstring exercise, which halves hamstring strain injury rates (van Dyk et al., 2019), does not substitute for hip extension loading of the proximal tendon.

Q: How is progression judged from one stage to the next?

Symptom behaviour over the twenty-four hours following a session is the most useful day-to-day criterion, since pain during loading is expected and pain that escalates the next morning suggests the dose was excessive. Alongside this, capacity markers that scale with the demand being reintroduced are more informative than repeat imaging: isometric force at the relevant hip angles, single-leg bridge repetitions, and tolerance of the previous sprint exposure. The VISA-H tracks the overall trajectory across months, although its minimum clinically important difference of 22 points means it changes too slowly to guide weekly decisions (Cacchio et al., 2014; Nasser et al., 2023).

References

• Arthurs, J.R. et al. (2026) 'Radial pressure wave and focused shockwave therapy for proximal hamstring tendinopathy: the role of patient positioning, a narrative review and clinical commentary', International Journal of Sports Physical Therapy, 21(4), pp. 467–476. doi:10.26603/001c.158527.

• Barton, C.J. et al. (2016) 'Running retraining to treat lower limb injuries: a mixed-methods study of current evidence synthesised with expert opinion', British Journal of Sports Medicine, 50(9), pp. 513–526. doi:10.1136/bjsports-2015-095278.

• Bohm, S., Mersmann, F. and Arampatzis, A. (2015) 'Human tendon adaptation in response to mechanical loading: a systematic review and meta-analysis of exercise intervention studies on healthy adults', Sports Medicine – Open, 1(1), 7. doi:10.1186/s40798-015-0009-9.

• Bourne, M.N. et al. (2017) 'Impact of exercise selection on hamstring muscle activation', British Journal of Sports Medicine, 51(13), pp. 1021–1028. doi:10.1136/bjsports-2015-095739.

• Cacchio, A., De Paulis, F. and Maffulli, N. (2014) 'Development and validation of a new VISA questionnaire (VISA-H) for patients with proximal hamstring tendinopathy', British Journal of Sports Medicine, 48(6), pp. 448–452. doi:10.1136/bjsports-2012-091552.

• Campos-Villegas, C. et al. (2024) 'Clinical progression and load management for proximal hamstring tendinopathy in a long-distance runner: a case report', International Journal of Sports Physical Therapy, 19(5), pp. 609–617. doi:10.26603/001c.116578.

• Cook, J.L. and Purdam, C.R. (2009) 'Is tendon pathology a continuum? A pathology model to explain the clinical presentation of load-induced tendinopathy', British Journal of Sports Medicine, 43(6), pp. 409–416. doi:10.1136/bjsm.2008.051193.

• Cook, J.L. and Purdam, C. (2012) 'Is compressive load a factor in the development of tendinopathy?', British Journal of Sports Medicine, 46(3), pp. 163–168. doi:10.1136/bjsports-2011-090414.

• Dizon, P. et al. (2023) 'Comparison of conservative interventions for proximal hamstring tendinopathy: a systematic review and recommendations for rehabilitation', Sports, 11(3), 53. doi:10.3390/sports11030053.

• Docking, S.I. et al. (2015) 'Structural integrity is decreased in both Achilles tendons in people with unilateral Achilles tendinopathy', Journal of Science and Medicine in Sport, 18(4), pp. 383–387. doi:10.1016/j.jsams.2014.06.004.

• Elkousy, H.A. et al. (2026) 'Endoscopic management of calcific tendinopathy in the proximal hamstring: two case reports', Case Reports in Orthopedics, 2026, 5545748. doi:10.1155/cro/5545748.

• Goom, T.S.H. et al. (2016) 'Proximal hamstring tendinopathy: clinical aspects of assessment and management', Journal of Orthopaedic & Sports Physical Therapy, 46(6), pp. 483–493. doi:10.2519/jospt.2016.5986.

• Hegyi, A. et al. (2018) 'Region-dependent hamstrings activity in Nordic hamstring exercise and stiff-leg deadlift defined with high-density electromyography', Scandinavian Journal of Medicine & Science in Sports, 28(3), pp. 992–1000. doi:10.1111/sms.13016.

• Hegyi, A. et al. (2019) 'High-density electromyography activity in various hamstring exercises', Scandinavian Journal of Medicine & Science in Sports, 29(1), pp. 34–43. doi:10.1111/sms.13303.

• Higashihara, A. et al. (2015) 'Differences in activation properties of the hamstring muscles during overground sprinting', Gait & Posture, 42(3), pp. 360–364. doi:10.1016/j.gaitpost.2015.07.002.

• Higashihara, A. et al. (2019) 'Hamstring muscles' function deficit during overground sprinting in track and field athletes with a history of strain injury', Journal of Sports Sciences, 37(23), pp. 2744–2750. doi:10.1080/02640414.2019.1664030.

• Lefèvre, N. et al. (2026) 'Partial tenotomy and reattachment for chronic proximal hamstring tendinopathy: functional outcomes, return to sport, and patient satisfaction', American Journal of Sports Medicine, 54(9), pp. 2213–2219. doi:10.1177/03635465261453077.

• Levy, G.M., Lucas, P. and Hope, N. (2019) 'Efficacy of a platelet-rich plasma injection for the treatment of proximal hamstring tendinopathy: a pilot study', Journal of Science and Medicine in Sport, 22(3), pp. 247–252. doi:10.1016/j.jsams.2018.08.001.

• Nasser, A.M. et al. (2021) 'Proximal hamstring tendinopathy: expert physiotherapists' perspectives on diagnosis, management and prevention', Physical Therapy in Sport, 48, pp. 67–75. doi:10.1016/j.ptsp.2020.12.008.

• Nasser, A. et al. (2023) 'Towards development of a core outcome set in proximal hamstring tendinopathy: a systematic review of measurement instruments and their clinimetric properties', Musculoskeletal Science and Practice, 66, 102774. doi:10.1016/j.msksp.2023.102774.

• Rau, O.R. et al. (2025) 'Extracorporeal shockwave therapy for tendinopathies around the hip and pelvis: a systematic review', HSS Journal, 22(2), pp. 197–207. doi:10.1177/15563316251332189.

• Rhim, H.C. et al. (2024) 'Use of extracorporeal shockwave therapies for athletes and physically active individuals: a systematic review', British Journal of Sports Medicine, 58(3), pp. 154–163. doi:10.1136/bjsports-2023-107567.

• Rich, A. et al. (2025) 'Physiotherapy compared with shockwave therapy for the treatment of proximal hamstring tendinopathy: a randomized controlled trial', American Journal of Sports Medicine, 53(14), pp. 3396–3407. doi:10.1177/03635465251391134.

• Rio, E. et al. (2017) 'Isometric contractions are more analgesic than isotonic contractions for patellar tendon pain: an in-season randomized clinical trial', Clinical Journal of Sport Medicine, 27(3), pp. 253–259. doi:10.1097/JSM.0000000000000364.

• Schache, A.G. et al. (2010) 'Hamstring muscle forces prior to and immediately following an acute sprinting-related muscle strain injury', Gait & Posture, 32(1), pp. 136–140. doi:10.1016/j.gaitpost.2010.03.006.

• van Dyk, N., Behan, F.P. and Whiteley, R. (2019) 'Including the Nordic hamstring exercise in injury prevention programmes halves the rate of hamstring injuries: a systematic review and meta-analysis of 8459 athletes', British Journal of Sports Medicine, 53(21), pp. 1362–1370. doi:10.1136/bjsports-2018-100045.

Antonio Robustelli is the founder of Omniathlete. He is an international high performance consultant and sought-after speaker in the area of Sport Science and Sports Medicine, working all over the world with individual athletes (including participation in the last 5 Olympics) as well as professional teams in soccer, basketball, rugby, baseball since 24 years. Currently serving as Faculty Member and Programme Leader at the National Institute of Sports in India (SAI-NSNIS).