Sports and recreation injuries are now known to be a significant public health problem. Lower limb injuries sustained during childhood and adolescence are associated with increased morbidity, including early development of osteoarthritis and long-term pain and disability;1 ,2 ultimately interfering with work, sports participation and a healthy level of physical activity.

In March 2010, we published a systematic review and meta-analysis (literature search conducted October 2008) on the effectiveness of neuromuscular training for prevention of sports injuries in athletes.3 Seven high-quality studies involving young male and female athletes (12–24 years) were included. Participants were engaged in organised sports, including basketball, volleyball, soccer, team handball, hockey and floorball. The pooled analyses revealed that multi-intervention exercises (comprising balance and agility training, stretching, plyometrics, running exercises, cutting and landing technique, strength training) significantly reduced the relative risk of lower limb injuries (relative risk reduction (RRR)=39%, 95% CI  23% to 51%), acute knee injuries (RRR=54%, 95% CI 24%  to 72%) and ankle sprain injuries (RRR=50%, 95% CI 21% to 69%).

What is new since then?

Since the publication of our review, five key trials have been published that aimed to reduce the incidence of lower limb injuries in young basketball or soccer players (table 1).

These studies are well designed with respect to important key criteria of internal validity and risk of bias, such as randomisation, allocation concealment, blinding of outcome assessors, statistical adjustment (eg, clustering and previous injury) or training compliance. A common problem across studies is that control interventions are not described explicitly. Subjects in the control group are customarily described as maintaining their routine warm-up. It is conceivable that a large variety of routine warm-up exists and that different warm-up routines might have different effects. A more detailed description of control intervention could improve our judgement about the similarity of studies. Multi-intervention training was provided as warm-up (without special equipment) before training and practice, including a variety of combinations of balance training, agility, stretching, plyometrics, running erxercises, cutting and jumping/landing technique, ‘core stability’ and strength training. Most studies placed a special emphasis on core stability and proper knee alignment.4–6 One trial combined the training programme with educational elements addressed at players, parents and the team's leaders to increase injury risk awareness.7 Players were usually followed up for one league season (approximately 8 months).

The results of these studies show consistent evidence that neuromuscular training programmes can reduce the relative risk of all injuries in adolescent and young adult athletes by 32–38%.4 ,5 Lower limb injuries were reduced by 32–65%, but statistical significance was not reached in all studies.4 ,5 ,8 A protective effect has been demonstrated on acute (38–56% risk reduction)4 ,8 as well as on overuse/gradual onset injuries (53–65% risk reduction).5 ,8 Injuries occurring in athletes that used neuromuscular training tended to be less severe.5 ,7 Furthermore, a specific preventive effect on knee injuries (77–90% risk reduction)7 including anterior cruciate ligament (ACL) injuries (64% risk reduction)6 as well as ankle injuries (50–66% risk reduction)4 ,8 was shown.

How do neuromuscular training programmes work?

On the basis of effective models for prevention of sports injury, preventive strategies should target modifiable factors that increase the risk of injuries. It is widely believed that sports injury risk might be associated with variables of strength, proprioception, coordination and neuromuscular control. Exercise programmes aimed at enhancing these abilities are thus often recommended for prevention.9 However, studies using multivariate statistical approaches and adequate sample sizes to determine the predictive value of proposed neuromuscular risk factors are still rare. The training programmes referenced in this editorial do not aim to enhance specific abilities, and instead are designed to improve neuromuscular function in general. Specific aims are, for example, ‘to improve strength, awareness and neuromuscular control during static and dynamic movements’ or ‘to provide strengthening exercises aimed at achieving an improved motion pattern that produces less strain to the knee joint’. The evidence is also lacking to demonstrate that training programmes aimed at improving neuromuscular function actually induce such improvements which, in turn, reduce injury risk.

Who is the target of neuromuscular training?

All training programmes outlined above have been investigated in specific groups of athletes considered to be vulnerable to specific injuries (ie, selective prevention, eg, female soccer players and ACL injury) rather than targeting athletes with specific impairments, that is, athletes who are actually identified as having an increased injury risk based on individual assessment of neuromuscular function. It is conceivable that athletes with a low risk of injury (possibly due to ‘healthy’ neuromuscular function) are less likely to benefit from preventive training, resulting in a large number needed-to-treat (NNT). However, approaches acting at the individual level might be more difficult to implement and more cost-intensive. In the future, it might be worthwhile to compare these two different approaches (selective vs indicated application) in terms of effectiveness and efficiency.

What is the practical significance?

In addition to the RRR, the absolute risk reduction (ARR) and its inverse, the NNT, are meaningful approaches to report the benefit of neuromuscular training programmes over control interventions. The additional value of the ARR is that it reflects the baseline risk of sports injury, that is, the risk of sustaining an injury without neuromuscular training. Waldén et al6 reported that the risk of sustaining an ACL injury was reduced by 64% in relative terms and 7% in absolute terms. The corresponding NNT to prevent one ACL injury was 14. However, the absolute rate reduction remained statistically borderline and the NNT should thus be treated with caution. The NNT for acute lower extremity injuries was 46, 32 for severe injuries and 191 for non-contact ACL injuries.5 ,8 This corresponds to ARRs of 2%, 3% and 0.5%, respectively. Clearly, a large RRR can be associated with much smaller reductions in the absolute risk and this association is determined by the general incidence of specific types of injuries (ie, the smaller the basis rate the larger is the discrepancy). This begs the question: how important are the findings referenced above in terms of practical relevance? The answer to this question should consider factors such as possible risks, cost–benefit estimates and implementation issues as to the preventive training. Assuming that such programmes are relatively safe, efficient (eg, no individual screening, team-based training and no costs for equipment) and, most importantly, can be implemented and sustained on a wide scale, then these findings might be important. However, the extent to which the results from efficacy or effectiveness studies might be extrapolated to the real sport world remains to be further elucidated.10 The studies evaluated here focused on competitive young athletes engaged in ball sports only, and all were engaged either in school teams or clubs organised by professional sports leagues. From a public health perspective, it is especially relevant to ascertain whether the preventive effects of neuromuscular training also apply to other popular high-risk sports (eg, hockey, skating, skiing and snowboarding), age groups (eg, children and adults) and athletic performance levels (recreational and novice athletes).