…or is it? The media are lapping up the press release associated with this study: Researchers explode the myth about running injuries. The title above is the title of the study as it appears in the British Journal of Sports Medicine. I added the question mark at the end for a reason. Does the study live up to the hype that it is generating in the media? Lets start with the abstract:
Foot pronation is not associated with increased injury risk in novice runners wearing a neutral shoe: a 1-year prospective cohort study
Rasmus Oestergaard Nielsen, Ida Buist, Erik Thorlund Parner, Ellen Aagaard Nohr, Henrik Sørensen, Martin Lind, Sten Rasmussen
Br J Sports Med doi:10.1136/bjsports-2013-092202
Objective: To investigate if running distance to first running-related injury varies between foot postures in novice runners wearing neutral shoes.
Design: A 1-year epidemiological observational prospective cohort study.
Setting: Denmark.
Participants: A total of 927 novice runners equivalent to 1854 feet were included. At baseline, foot posture on each foot was evaluated using the foot-posture index and categorised into highly supinated (n=53), supinated (n=369), neutral (n=1292), pronated (n=122) or highly pronated (n=18). Participants then had to start running in a neutral running shoe and to use global positioning system watch to quantify the running distance in every training session.
Main outcome measure A running-related injury was defined as any musculoskeletal complaint of the lower extremity or back caused by running, which restricted the amount of running for at least 1 week.
Results: During 1 year of follow-up, the 1854 feet included in the analyses ran a total of 326 803 km until injury or censoring. A total of 252 participants sustained a running-related injury. Of these, 63 were bilateral injuries. Compared with a neutral foot posture, no significant body mass index-adjusted cumulative risk differences (RD) were found after 250 km of running for highly supinated feet (RD=11.0% (−10% to 32.1%), p=0.30), supinated feet (RD=−1.4% (−8.4% to 5.5%), p=0.69), pronated feet (RD=−8.1% (−17.6% to 1.3%), p=0.09) and highly pronated feet (RD=9.8% (−19.3% to 38.8%), p=0.51). In addition, the incidence-rate difference/1000 km of running, revealed that pronators had a significantly lower number of injuries/1000 km of running of −0.37 (−0.03 to −0.70), p=0.03 than neutrals.
Conclusions: The results of the present study contradict the widespread belief that moderate foot pronation is associated with an increased risk of injury among novice runners taking up running in a neutral running shoe. More work is needed to ascertain if highly pronated feet face a higher risk of injury than neutral feet.
This is a really impressive, and generally, well done and analysed large prospective study that classified the feet of 927 runners into different foot types using the Foot Posture Index; gave them all the same running shoe (Supernova Glide 3) and then followed them for a year and 252 of them got an injury.
The reported incidence of injury in each of the groups that they used to classify the feet were:
As you can see from the table, the incidence of injury was lower in the pronated group (p=0.03) and appeared highest in the highly pronated group, but that was not statistically significant, though it was trending that way. This was relative to the neutral group which they defined as a Foot Posture Index (FPI) between 0 and 7. The pronated group was defined as an FPI of 7 to 10 and highly pronated as greater than 10.
Does the study support all the media hype of “Researchers explode the myth about running injuries“?
Anyone familiar with the FPI, knows that any FPI value above 0 is pronated, so almost all the feet in their ‘neutral’ group (except those with an FPI of 0) will be pronated!! So no, the study does not explode the myth as most of the neutral group had a pronated foot!! (personally, and anecdotally, I and all the people I know who have worked with the FPI and have discussed this, generally consider 0-4 to be a reasonable normal range and anything above 4 to be pronated. The study used anything above 7 to be pronated 
).
Even if the results were correct, it does not support the hype in the press release. The evidence from the most recent meta-analysis of all the available studies show that there is a small, but statistically significant risk from a pronated foot for overuse injury. If another meta-analysis was done to include the above study claiming to show no relationship between pronation and injury and the other recent new study that found there was, then the overall results of the meta-analysis is probably unlikely to change from the small, but statistically significant risk from ‘overpronation’.
Additionally, they eliminated all those wearing inserts or foot orthotics from the study. It would be reasonable to assume that those wearing these would probably have been in the pronated group and probably be at greater risk of developing an injury (hence they needed the foot orthotics!). This presumably left the group of pronated feet that were included in the study at less risk for injury in the study (as they did not need foot orthotics). Can you see how this might have introduced some sort of bias into the results?
Furthermore, no data was presented on the reliability of the 4 testers in their assessment of the FPI. We have no idea what the intratester or intertester reliability was or if there was consistency between those doing the testing.
Moreover, was it reasonable to give the pronated feet a neutral shoe and then determine the injury rate? Surely if the prevailing paradigm (flawed as it may be!) is to give a pronated foot a motion control shoe, would it not have been more reasonable to do that? In the real retail world, that is probably what would have happened. (However, I not sure this would actually make any difference).
As an aside, the 252 injuries that the study found is a prevalence of 27%; certainly a lot less than than the 70% figure you read so often in the crankosphere blogosphere and at the lower end of 20 to 70% you read in the scientific literature.
As always, I go where the evidence takes me until convinced otherwise and I will stick to the higher level of evidence from a meta-analysis and not the over-hyped media release.
Nielsen, R., Buist, I., Parner, E., Nohr, E., Sorensen, H., Lind, M., & Rasmussen, S. (2013). Foot pronation is not associated with increased injury risk in novice runners wearing a neutral shoe: a 1-year prospective cohort study British Journal of Sports Medicine DOI: 10.1136/bjsports-2013-092202
What FPI values did they use for supinated and highly supinated?
Supinated was considered -1 to -3
Highly Supinated was considered > -3
Which suggests a bias/ attempted deception from the authors. Is it just me, or is the current trend toward publishing a paper associated with a media fanfare just crass?
P.S. it should be nice to see the data re-analysed using the FPI boundaries which have previously been employed within the literature. Moreover, repeated with a “pronation control shoe”.
Exactly! The study a few months ago that showed the exact opposite had no press release and no media hype and no marketing on twitter and no writes ups on blogs, except what I did!
How a peculiarity (here the pronation), present from ever for a specific person can suddenly become the cause of a new pathology. “Pronation” was used since long time by health professional to justify their medical intervention : “corrective foot orthoses”… hope we will be more rigours in the future, for any type of medical intervention.
PS : I will analyze deeply the MA of Tong soon… because me too I go where the evidence takes me 🙂
The most recent rigorous meta-analysis of all the available good studies has shown that “overpronation” is a small, but statistically significant risk factor for injury. Another soon to be published systematic review found the same thing – this is what the evidence says. Do you have evidence from a meta-analysis or systematic review that says something different?
Even if the above study is right, why cherry pick it over the other most recent study that showed the exact opposite?
I have been saying for >15 yrs that ‘overpronation’ was not a problem as there were more studies showing it was not a problem than there were studies showing it was a problem. However, I have recently had to start changing my views as more studies appeared and the pooling of the data in the meta-analysis are showing it is a small, but statistically significant risk factor.
don’t change to fast 🙂
I just finish to do a deep analyses of the study and there some very important point that make me saying again : Foot posture/index/pronation is NOT associate with pathology!
Hi Craig!
Followed a link on twitter here and have been lost for the last hour in your awesome site. There is lots of sensible information compared to what you can read elsewhere.
I agree with you above and with the comment about how much publicity this study is getting. The other study you linked to that showed the opposite results is interesting. It is interesting that it is not getting the publicity online, especially on twitter that the above study is getting. It is even more interesting about seeing who those are who are choosing to tweet and retweet the study and ignore the other study and the meta-analysis you mentioned. Are they just promoting an agenda and only the studies that support that agenda?
Angie
Thanks. I would not quite go that far. The above study was released with a press release with a catchy title and distributed via Eureka Alert which gives wide and good distribution. The media love this sort of thing and fell for it. If you were not familiar with the FPI, then you would not realize that most of the feet in the neutral group were pronated. If you did not read the small print of the actual publication, then you would not realize that those pronated feet with the highest risk for injury (ie orthotics wearers) were removed from the study.
The previous study I linked to was released with nothing – it just showed up in one of my alerts and did not get much mileage at all, so most if not all of the media would not even had known about it. In fact, I have only even seen it mentioned in two places – here on this blog and on Podiatry Arena – its pretty much been ignored by everyone else. If the media had done there homework, rather than parrot the press release headline about the myth, they would have found the meta-analysis and given more weight to that than just “one study”… go figure!
There have been plenty of studies on how the media portray medical research and how badly they do it, so there is nothing to be surprised at here (I wrote about that here on the debacle of running shoes causing knee OA). Fortunately we only talking about running injuries and not the lieing cranks from the vaccine’s causing autism movement or the alternative cancer woo pushers or something that actually involves life or death (that the media still gets wrong so often as well).
Just read the Teyhen study you liked. Seems to me the only significant finding was that highly pronated individuals are at greater risk of knee injury. In the results they point out that “musculoskeletal injury rates were similar between participants with pronated (53.8%), supinated (52.7%), and normal (47.8%) foot types.” So injuries as a whole did not differ much. According to Table 2 overall injury rates were between 48-55% for all of the foot type groups.
My bigger issue with the Nielsen study is that they are using a static measure of pronation, even if it is more complex than just looking at arch height. How well is FPI correlated with kinematic measures of pronation in running. Do we even know? We know, for example, that there is little correlation between navicular drop and dynamic pronation measures.
There is no or limited data that has linked single static measures of “pronation” to dynamic function.
Two recent studies have shown a link between the Foot Posture Index (a static measure that combines a number of observations) and dynamic function:
Viv Chuter showed that the FPI was strongly correlated to dynamic function http://www.biomedcentral.com/content/pdf/1757-1146-3-9.pdf
Christian Barton showed a moderate correlation between FPI and dynamic function http://link.springer.com/article/10.1186/1757-1146-4-10
I would not necessarily say I like the Teyhen study. My point is that its results are different to the above (assuming we accept the FPI cut-off; and the elimination of the probably or allegedly high risk ‘pronators’ who wore foot orthotics) and its not getting the media attention as they did not put out a sexy press release!. When there are conflicting studies, we need to rely on the pooling of data in meta-analyses – when that is done, there is a small, but statistically significant risk from “overpronation” and injury. We should not be cherry picking the studies that show there is a risk or cherry pick those studies that show there is no risk, depending on the point one is trying to make.
Anyone who has heard me lecture will know that I have been preaching for >15yrs that “overpronation” is not a problem, as that was what the bulk of the evidence was saying at the time. More recently some studies have reported it is a problem, but I still thought it wasn’t a problem as that is what the bulk of the studies said. I now have to change my tune somewhat as the formal meta-analysis that pooled data from the good studies I linked to above (and another upcoming systematic review) are telling us it is a risk,
Unfortunately neither of those FPI papers looked at running. Here’s one on navicular drop vs. dynamic movement in running: http://www.ncbi.nlm.nih.gov/pubmed/19648718
Two things that often frustrate me when these types of studies and discussions appear are that:
1) there is a difference between “risk of injury” in the general sense and risk of specific types of injury. A measure may be found to increase risk of knee injury, as in the Teyhen study, but overall injury risk doesn’t appear to be elevated in pronated individuals. So in that sense the Teyhen study can be interpreted in two ways. Pronated individuals are not at greater injury risk in general, but they may have increased risk of knee trouble.
2) There is the issue of injury risk given the type of behavior performed. Is injury risk for “pronated” walkers the same as for “pronated” runners? Or pronated kids? Or pronated elderly people? When looking at meta-analyses it’s important to consider the thrust of the pool of studies included. Were they looking at risk for runners, as Nielsen did, or pooling studies that looked at many different kinds of study subjects.
You right. Share the concerns and the way studies used.
Some other issues:
1. As I have stated many times, its the parameters that are chosen to measure (whether its a study on risk factors or correlation to dynamic function). Navicular drop is just a measure of changes in arch height in the saggital plane. You can have a large amount of navicular drop with no eversion of the calcaneus – is that a “pronated foot”? Similarly, you can have a large amount of medial midfoot bulging and forefoot abduction with no change in navicular drop or calcaneal eversion – – is that a “pronated foot”? Etc, etc It all depends on the orientation of joint axes as to how much of which one you get compared to another (ie the concept of planal dominance) – just check Inman’s original work on the massive variability in STJoint axis position – its not difficult to conceptualize the relative movement of some of the putative measure of ‘pronation’ relative to each other based on the variability of joint axes positions.
2. What injuries? Even if pronation is a “global” risk factor, its obviously not in a risk factor for them all! Without checking, the only prospective risk factor study on specific injury I can recall at the moment is Shaun White and Ben Yates one on MTSS – they showed a higher FPI doubled the risk for MTSS (there are probably others, but the brain is currently fried from a head cold and can’t recall them).
3. Overuse injury occurs when the cumulative load in the tissues exceeds what the tissues can tolerate. “Pronation” has nothing to do with loads. We showed a very poor link between the FPI and the force needed to supinate the foot … ie the forces that are needed to stop the ‘pronation’ —- it is those forces that damage the tissues (motion does not damage tissues, forces do; as I often say “kinetics hurt, kinematics doesn’t“). You can have a severley pronated foot, but the forces to supinate it are low (ie post tib muscles does not have to work hard and the windlass mechanism can work easy and efficiently) –> up comes the arch and foot supinates easy –> are they at risk for injury? At other extreme, you see feet that are only mildly pronated, that I can not supinate with two hands (ie supination resistance is very high) –> imagine the forces going through the tissues in that foot –> are they at risk for injury? – they probably at a very very high risk (… in reality these people don’t run as its too hard) — its all about forces and loads in the tissues and our work shows a poor correlation between “pronation” and forces in the tissues. No prospective risk factor studies have looked at these issues. We have done a number of experiments and cross sectional studies on it (some published and some not yet). I am planning a more complete blog post on the concept of supination resistance….. one day!
4. Personally and anecdotally, I do not even consider “pronation” or any of the parameters associated with it – its all about the loads going through the tissues and strategies needed to reduce them. As I said above, there is a poor correlation between the FPI and the forces. So the focus clinically is on things like supination resistance, windlass establishment forces, etc. However, there is no prospective risk factor data or outcome studies on this … yet….so we have to rely on the biological plausibility, theoretical coherence and consistency with available evidence principles. The work we have done on this was fueled by the work showing the lack of correlation between ‘overpronation’ and injury/dynamic function.
Just on the comment: “Or pronated kids? – “as an overuse injury occurs when the cumulative load in the tissues exceeds what the tissues can tolerate” – the tissues in kids can better tolerate the loads than an older person, so even if pronation is a risk factor, then there is a pretty good chance that its not in kids (UNLESS, presumably, the loads associated with that overpronation, ie supination resistance, are higher than what the tissues can tolerate).
Just recalled this study from last week. Nothing to do with running, but they looked at 3D kinematics between symptomatic and asymptomatic flatfoot in kids – they found no differences …. why did they not measure the kinetics? … ie the forces in the tissues!!!! Forces damage tissues and hurt. The study just confirmed that kinematics does not hurt …. another wasted opportunity.
Craig,
I agree completely with this: “it’s all about forces and loads in the tissues.” Injury management is about modifying forces applied to the injured tissues to stop the injury process. This could be through orthoses, footwear change, form work, soft tissue work, strengthening work, etc. I think all of these interventions essentially are means to the same end – modifying force application such that it is manageable by the target (injured) tissue. As is human nature, we all like to debate which intervention is best 🙂
My issue with pronation as it relates to running is the shoe prescription process whereby pronation is still viewed in a lot of places as something that needs to be corrected. The shoe store in my town still uses one of those machines that scans the foot and spits out shoe suggestions based on arch shape. This kind of thing really bothers me. Pronation is a factor to consider if an individual has an injury that might be caused or excerbated by excessive pronation (e.g., MTSS), it’s not a diagnosis that needs to be treated. Now I sound like Ian Griffiths 🙂
Pete
Just to add further to the mix; there are probably up to a 100 different causes of “overpronation” (I mentioned some here http://www.runresearchjunkie.com/the-nonsensical-understanding-of-overpronation/ (muscle weakness is one of the less common causes) – each cause may or may not have different risk profile with injury, so its probably not appropriate to lump them all in together in a risk factor study.
Some of the causes are easily dealt with by the “anti-pronation” features of running shoes; most of them will not be, so its probably not appropriate to lump all causes of overpronation into the one study of running shoes.
The Redmond et al paper states that normal (+/- 1 SD) FPI is 1 to 7 (mean 4), based on n=619. See Table 2: http://www.jfootankleres.com/content/1/1/6
Thanks Hilton; therein is the problem. The FPI of 7 is 1 SD from the mean but show any clinician a picture of a FPI of 5, 6 or 7 and they will almost all probably say ‘very pronated’ …. I don’t know hat the answer is. It not until you get down to a 4 will they probably start saying “more normal” than “pronated”. Probably more importantly, they would be considered “overpronated” at the retail level and prescribed (either rightly or wrongly) a stability or motion control shoe.
My point is that in the absence of any other validated cutpoints, it’s perfectly reasonable to use distribution-based cutpoints to define foot type categories based on the FPI. I therefore think your criticism of this aspect of the paper is unjustified, as any other cutpoints would be essentially arbitrary.
I am not an expert in this area. Wouldn’t the ovepronators who have to wear foot orthotics be at a great chance of getting an injury. From what I read the study removed those wearing foot orthotics from the results. Surely those overpronators would have been in the classification that are overpronators? Does this make the study nonsense?
Hylton makes a good point. I suggest if they had have used a FPI of 0-4 then you’d have people arguing from the otherside that that the cut off doesn’t reflect the literature.
I am no where near on your level of experience or understanding but would have thought that simply assessing pronation or over pronation is far too simplistic, hence the confusion and residing debate. As you have mentioned, there is the issue of kinetics but also injury Hx, structural anomalies, systemic illness, proximal and distal movement dysfunction, age, exercise programming parameters (if related to sports performance). Perhaps “over pronation” is implicated as suggested by the meta-analysis but it’s not really as big a deal as sometimes made out to be. It’s over pronation in a specific. context/s that perhaps is when it comes into play?
I have been preaching for years that overpronation was not a problem. I have changed my mind recently due to the meta-analysis showing its a small but still statistically significant risk factor.
We in the lunch break o fa seminar I teaching at right now. This paper was one that was discussed. The big issue most had was the point Angie made a few posts above – they eliminated the orthotic wearing pronators who would have been hypothetically at greatest risk!
Yeah I agree, in light of the Meta-analysis and the exclusion of orthotic wearers the value of this research is somewhat reduced. Are you aware of an ammendment to the meta-analysis with the inclusion of this study in the pooling of data?
Not sure. I do believe that there is another systematic review ‘in press’ that does include this paper. Don’t know if it will also include the Teyhen paper due to its publication status…
So would you recommend that those with flat feet and/or those who pronate or “over” pronate use stability shoes as opposed to neutral shoes, or simply go with what feels better on when testing them at the store? Thank you.
Based on the current evidence, it is hard to base running shoe prescription based on ‘pronation’.