Reducing risk of injury due to exercise

BMJ 2002; 325 doi: http://dx.doi.org/10.1136/bmj.325.7362.451 (Published 31 August 2002) Cite this as: BMJ 2002;325:451

Stretching before exercise does not help

  1. Domhnall MacAuley, general practitioner (domhnall.macauley{at}ntlworld.com),
  2. Thomas M Best, assistant professor of family medicine and orthopaedic surgery (Tm.best{at}hosp.wisc.edu)
  1. Department of Epidemiology and Public Health, The Queen's University of Belfast BT12 6BJ
  2. University of Wisconsin Medical School, 1300 University Avenue, Madison, Wisconsin, WI 53706, USA

    Papers p 468

    It used to be so simple. Prevention of musculoskeletal injury during exercise meant conditioning, warm up, and stretching. We could not argue with these basic principles—until we began to look for the evidence to support such advice. Stretching is long established as one of the fundamental principles in athletic care. No competition is complete without countless athletes throwing shapes along the trackside, trainers and coaches each favouring their own particular exercises, and locker room experts, kinesiologists, and self appointed specialists inventing new contortions for long forgotten muscle groups. Sport is rife with pseudoscience, and it is difficult to disentangle the evangelical enthusiasm of the locker room from research evidence. But in this issue, Herbert and Gabriel (see p 468) question conventional wisdom and conclude that stretching before exercising does not reduce the risk of injury or muscle soreness.1

    They are not the first group to examine the evidence behind stretching and injury prevention. Shrier, in a systematic review of the literature, identified 293 articles but included only those with a control group.2 Three prospective clinical trials showed that stretching was beneficial, but each included a co-intervention of warm up. A fourth, cross sectional, study found that stretching was associated with fewer groin or buttock problems in cyclists, but only in women. In contrast, five studies, of which three were prospective, found no difference in injury rates, and three suggested that stretching was harmful. Shrier concluded that stretching before exercise does not reduce the risk of injury. If we were to argue that only evidence from randomised controlled trials should be used to determine clinical practice, the conclusions of this review may have been different, but the findings were later supported by a large randomised controlled trial. 3 4 So it seems that stretching to prevent muscle injury and muscle soreness is not supported by evidence from quality clinical research studies. These findings are contrary to what many athletes and coaches believe and what is common practice. On the other hand, these findings may not be too surprising if we consider the complex mechanical properties of biological soft tissues and their response to cyclic loading.5 It may also be that the research evidence is incorrect and that there is some, as yet unproved, benefit.

    Nevertheless, evidence for the value of stretching is only one of a myriad of unanswered questions about musculoskeletal injury. As we begin to examine even the basic principles of acute injury management we find a paucity of research evidence. Much of common practice is based on historical precedent rather than randomised controlled trials, which comprise about 10% of the published literature in sport and exercise medicine.6 Even the most accepted treatments find little support when critically evaluated. For example, the mnemonic “ice”—representing ice, compression, and elevation—has become the mantra of sports physicians and physiotherapists. It is used to guide the early treatment after acute musculoskeletal injury. Empirically, it seems logical that ice should be effective. But how much evidence is there to support the use of ice and what is the optimum clinical strategy for its application? Basic questions such as how long, how often, and for what duration we should apply ice remain unanswered. The advice given in various textbooks varies a lot, and little evidence is available in the form of original research. 7 8 Again, reducing swelling by compression after injury seems appropriate, but when we look more closely we see that various bandages, strappings, and supports offer variable degrees of compression.9 Double tubigrip, the favoured compression bandage of accident and emergency departments, seems to offer little benefit. Early movement gives the best result, so it may be better not to apply any bandage that may restrict ankle movement and simply advise on appropriate exercises.

    Thankfully, there is emerging evidence to guide some aspects of clinical practice.10 Ankle sprain, one of the commonest sporting injuries, has always been difficult to treat. There is now evidence to show that taping or bracing can reduce the incidence of recurrent ankle sprains. The protective effect of taping seems to be limited to people with previous injury, in whom postural control, position sense, and postural reflexes are altered. Furthermore, there is evidence that balance training can improve sensorimotor control in athletes with previous injuries.

    Much of sport and exercise medicine and the management of musculoskeletal injury has developed empirically, with little research evidence. Some of the basic principles of caring for acute injuries of the soft tissues have never been questioned, yet there is often little evidence to support common practice. The culture is changing, and Herbert and Gabriel make a valuable contribution to the debate on stretching.


    1. 1.
    2. 2.
    3. 3.
    4. 4.
    5. 5.
    6. 6.
    7. 7.
    8. 8.
    9. 9.
    10. 10.