Statistics from Altmetric.com
If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.
It is tempting to read the Moen et al1 study and conclude that MRI is not very useful in assessing prognosis of hamstring strains. Previous studies have found a reasonable correlation between MRI severity and clinical outcome,2–4 but no previous study has compared such a multitude of clinical parameters with a similar number of MRI parameters head to head with a blinded methodology. That is the strength of Moen's study.
The clinical parameters proved to be at least as good as the MRI parameters and some were clearly superior. The best predictive MRI parameter found that injuries closer to the ischial tuberosity had a worse prognosis, but as a predictive variable it was almost identical to simply recording where tenderness started in relation to the ischial tuberosity on clinical assessment.1
When might MRI be useful?
With these MRI data, the authors are in a good position to conclude that MRI investigation is not justified for the average non-professional athlete suffering an average hamstring strain. The exception, though, is probably one of the types of hamstring injury excluded from this study—hamstring tendon rupture. If there is strong clinical suspicion of hamstring tendon rupture, for which surgery may be contemplated, then MRI would be justified.
For professional athletes, the debate about use of MRI for hamstring strains is complex. The cost of MRI is not an issue compared with the club's investment in the player and his or her salary. Note that a few MRI parameters in the Moen et al1 study improved the predictive equation, but they did not reach statistical significance. Therefore, MRI results might improve an accurate prognosis to a small degree, but even a study with a moderate sample size could not prove this either way. The real- world likelihood may be that having the MRI information (plus clinical information, as opposed to clinical information alone) might only improve your clinical decision-making (as in accuracy of time to return to sport) by a few percentage points and for a professional athlete even a slight improvement would be desirable. I expect professional athletes will continue wanting MRI done when they injure their hamstrings. Where this study may help a professional athlete is the observation that their self-assessment about injury severity is worth at least as much weight, perhaps more, as the MRI at determining time to return to play.1
Does the injury severity completely dictate time to return to play?
The rejection of the notion that MRI severity is the most important determinant of return to play fits neatly with the return to play model being developed by Shrier's Canadian research team.5 They correctly highlight that anatomical severity and recovery of injury is only one part of the return-to-play equation. The demands on the athlete are equally—if not more—important.6 For example, a goalkeeper in football can return to play more quickly from a hamstring injury than an outfield player with the same severity of injury, because of lower sprinting demands.
The other factor that should not be ignored is the ‘risk appetite’ of the player and coach (and perhaps medical team), given the fact that return to play from a hamstring strain has risks as well as benefits. It is rational to ‘wait an extra week’ early in the pre-season period and equally rational for the player to ‘roll the dice’ during an end-of-season game with a major trophy up for grabs. For a coach, it is rational to take risks with the team's best player and equally rational to prefer one fringe player over another of equal ability if the latter is recovering from a hamstring injury. In a sport like cricket (where no substitutes are allowed) it may be selfish to start a match carrying a hamstring injury, compared to American football (with a huge interchange bench) where the cost to the team of a wrong decision is far less.7 These perspectives highlight the limitations of the MRI—after all, the MRI is only assessing pathoanatomy and none of these other relevant return-to-play factors.
Moving from this scientific study to the real world, it is worth a history lesson in considering what the consequences of MRI have been for the management of hamstring injuries. The Australian Football League (AFL) injury database has been running for 21 years,8 long enough to span an era where very few hamstring strains received MRI (in the early-to-mid-1990s) to the last decade where the majority of hamstring strains have had MRI as part of the management.
The recurrence rates of hamstring strains have dropped by more than half from the 1990s (35–40% recurrence rates) to the 2010s (15% or fewer recurrence rates).8 This association between fewer recurrences and increased use of MRI raises the question of whether and how MRI has contributed. As sports physicians, we would like to believe that advances in rehabilitation and prevention of recurrent hamstring injuries have contributed,9 although there was no drop in primary hamstring strains over the same time period. It can be concluded at the very least that the widespread use of MRI has not made accurate prognosis any less likely. In professional sport, the real-world value of MRI may be less in its ability to unfailingly predict return to play (which it obviously cannot) and more in its psychological effect on players and coaches.
The natural tendency of the player and coach is to pretend that pain is not real and that a tough player should be able to push through it. The MRI, when positive for hamstring strain injury (which is not always the case in low-grade injuries), is an objective reminder to all concerned that injuries are real and that there are risks in trying to play with them. If MRI contributes to professional players and coaches properly understanding this, then they are providing real value.
Competing interests None.
Provenance and peer review Commissioned; externally peer reviewed.