The bone stress injury model: a new way to deal with stress fractures and stress reactions in runners

An MRI reveals a tibial stress fracture
Traditionally, overuse injuries to the bone in distance runners are divided into two distinct categories: stress reactions and stress fractures.  Runners who develop pain along one of their bones hope desperately that they have the former and not the latter, since the usual prescription for stress fracture recovery is six to eight weeks of no running whatsoever.  The usual restrictions for stress reactions depend on the doctor, but typically involve two to four weeks away from running.  Some doctors, coaches, and runners eschew the term "stress reaction" entirely because, in their view, you either have a stress fracture, or you don't—that's all there is to it.
Normally, the story unfolds something like this: a high school runner develops a sharp, aching, localized pain somewhere along a bone in his lower body.  It doesn't improve much with icing and lowered training volume, so his coach or trainer refers him to a doctor.  The doctor orders an X-ray, examines it, but sees no evidence of calcification, so he orders the runner to ease back into training, but return if pain continues.  The high schooler gives running a shot, but continues to have pain.  The doctor then orders a bone scan or an MRI, which shows bone marrow edema or increased metabolic activity at the location of pain.  This is deemed to be a stress fracture, and the runner is put in a boot, forbidden from running for six to eight weeks, and his season is effectively over.  Sound familiar?
The reason for caution with stress fractures is well-known.  If you have a stress fracture and continue to run recklessly on it, it can worsen and eventually lead to the bone splitting in two—a true fracture.  This can lead to heaps of complications and could end your running career.  There is also a category of "high risk" stress fractures that occur in particular areas like the femoral neck, the navicular, and the sesamoid bones, which are known to have a significant risk for poor healing or nonunion.1 These require even more time off and a much slower return to running.
Problems with the old model
However, doctors and physical therapists are starting to learn what coaches have already picked up on: the traditional approach to low risk stress fractures (as the vast majority are) is inadequate on a number of points.
The case for a new approach to bone injuries in runners was laid out in an exhaustive review article published in October of last year by Stuart Warden, Irene Davis, and Michael Fredericson, three extremely prolific running injury researchers.2 They propose using the term "bone stress injury" or BSI, which is intended to encompass all overuse injuries to bone that runners sustain. 
Under Warden, Davis, and Fredericson's model, bone stress injuries exist on a continuum.  On the most severe end of this spectrum are true stress fractures: a fracture line is observable on an MRI or CT scan, and there is edema (swelling) in the bone marrow and periosteum, the membrane that covers the surface of the bones.  A stress fracture is accompanied by a sharp pain or ache during weight-bearing activity that sometimes persists even when you're resting. 
The next step down the continuum of bone stress injury is the stress reaction: pain and aching during or after weight-bearing that is associated with bone marrow edema (on an MRI) or increased bone remodeling (as imaged by a bone scan), but lacks a visible fracture line. 
Further down the bone stress injury spectrum lies asymptomatic areas of bone remodeling.  As it turns out, if you were to schedule weekly MRIs for a group of high-level runners in heavy training—say, a college cross country team—you would quite often find runners developing transient areas of bone marrow or periosteal edema that would be indicative of a stress reaction, except that they have no pain associated with them, and never develop problems in the area.3 
The biology of bone remodeling
One of the core paradoxes of stress fractures and stress reactions is why they occur in fairly experienced runners.  All medical students can recite Wolff's law—bone responds to stress by becoming stronger.  So, theoretically, running more should lead to stronger bones, not stress fractures.

Warden, Davis, and Fredericson outline the biological process of bone remodeling that resolves this problem.  When exposed to a new stress, cells called osteoclasts start to eat away, or "resorb," small pockets of bone in the area of your body exposed to mechanical stress (e.g. the tibia in a runner). 
Often, this bone resorption process eats away at the area around microscopic cracks in the bone that sometimes occur from exposure to a new stress.  In this case, the direction and extent of the microcrack dictates the amount of bone resorption that occurs.  Bone resorption doesn't always rely on microcracks, however: a review paper by Melanie Franklyn and Barry Oakes cites research in rabbits showing bone resorption after beginning a running and jumping program even in the absence of microcracks.4 This is good news, as it implies that actual bone damage isn't necessary to stimulate bone growth.
Bone remodeling occurs as a two-step process: first, osteoclasts eat away older bone tissue, including areas around microcracks, then osteoblasts fill in the gaps, creating new, stronger bone growth. From Warden, Davis, and Fredericson.
 This bone resorption process is necessary to make room for new bone cell growth—just like the first step to remodeling a kitchen or bathroom is to tear out some walls, your body needs to tear out old bone cells to make room for new ones. 
In the wake of the osteoclasts, another wave of cells called osteoblasts follow, filling in the gaps and eventually solidifying, which creates a newer, stronger matrix of bone tissue.  Because of the geometry of bone shape, even a very small increase in overall bone thickness will lead to a significant increase in resistance to bending and twisting forces, which cause bone injury. 
However, this process takes time.  According to Franklyn and Oakes, the entire process of bone resorption and osteoblast regrowth takes about 90 days to complete.  Warden, Davis, and Fredericson cite a similar figure of 3-4 months.  For a significant portion of this window, a runner is vulnerable to further injury, since bone resorption has occurred, but new bone growth is not yet finished.  If a new injury occurs, it can set off another round of osteoclast-driven bone resorption, which further lowers local bone density, and increases vulnerability to injury again. 
Critically, this model of bone remodeling can explain findings like asymptomatic, transient areas of MRI or bone scan findings indicative of stress reactions in high level runners in training.  It also explains why the greatest risk for stress fracture occurs within the first month of introduction of a new training stress.5 
As long as the rate of bone growth outpaces the rate of bone damage incurred both by exercise and by osteoclast bone resorption, no injury will occur.  However, bone stress injuries, like "shin splints" (medial tibial stress syndrome), stress reactions, and stress fractures, will occur if bone damage outpaces bone growth. 
Implications of bone remodeling on bone stress injuries
The main implication of this new model of bone remodeling is that stress fractures and stress reactions are not binary conditions that you either have or don't have; they represent different points on a spectrum of bone stress injuries.  This means that we can move smoothly along a gradient of bone stress injury severity, both in the "upward" direction of an injury getting worse if we continue to run on it, and in the downward direction of improvement if we take time off.  As the bone healing process goes on, you will be able to handle increasing amounts of bone loading without re-aggravating the injury or causing further damage.  Combine this with the fact that the rate of bone healing will vary significantly from person to person and you can see why it makes a lot of sense to abandon the old model of "positive MRI findings for any bone stress injury means 6-8 weeks off."
The new model of bone stress injury management
The evidence laid out above suggests we take up a new approach to managing low-risk bone stress injuries.  Warden, Davis, and Fredericson describe a new model centered on pain—or, hopefully, a lack thereof—during the recovery period.  If you're not experiencing pain, it means that your level of activity is not loading your bones at a rate that outpaces their ability to recover.  And if you do have pain, either during or after an activity (whether it's merely walking around or doing an interval workout) it means you are exceeding your bone regrowth capabilities and are causing damage, so you need to reduce your bone loading level.  Further, this means that if you increase your bone loading too quickly and experience pain again, it doesn't mean you're back to square one; it just means you need to back off for a few days until pain subsides again.
Pain-mediated management of bone stress injuries in the real world
Finally, we can take a look at what the new approach to managing bone stress injuries should look like in the real world. Let's return to our initial example of a high school runner who presents with sharp, localized aching along a bone.  When he talks to his coach and the athletic trainer, both suspect a stress fracture or stress reaction.  What's the appropriate course of action?
The first step is to rule out a high risk stress fracture.  If there's any possibility that our runner's injury is a high-risk stress fracture, he needs to see an orthopedist and get a definitive diagnosis.  This will probably involve an MRI to determine the location and severity of the injury.  Fortunately, these represent only a small minority of bone stress injuries.
But if it's clear that our runner's injury is not a high-risk stress fracture, do we even need to bother with an X-ray or MRI? Probably not.  If, for example, there is a small area of sharp, localized pain along the medial edge of our runner's tibia, an MRI is probably not going to tell us anything we don't already know.  It's extremely likely that an MRI or bone scan would show signs of stress fracture or stress reaction, and given that recovery time is widely variable from person to person, it doesn't really matter whether it's a stress fracture or a stress reaction; you're dealing with a bone stress injury either way.  
There is some evidence that MR imaging can be used to grade the severity of a stress fracture,6 but Warden, Davis, and Fredericson argue that you should "treat the patient and not the X-ray [or, presumably, the MRI or bone scan]"—further emphasizing the importance of pain monitoring as the primary driver of any recovery plan.  MRI is most useful in evaluating high risk stress fractures and for situations where it is hard to differentiate between a bone injury and a soft tissue injury.  A few examples might be differentiating a quadriceps strain from a femoral stress fracture or differentiating a gluteus medius strain from a sacral stress fracture. 
Continuing with our example, once we've identified that our runner has a bone stress injury to the medial tibia, the next thing we need to do ensure he is having no pain during his day-to-day activities.  On this point, Warden, Davis, and Fredericson are clear: if you are having pain walking around, you need to be in crutches or in a "boot" (a walking cast or aircast) so you can get around without pain.  However, they also emphasize that "progres­sion to unassisted pain-free gait should be sought as soon as possible."

Cross-training to maintain fitness while you can't run is a great idea, but the same rules apply: if a method of cross-training causes pain, either during your workout or afterwards, you need to hold off until you can do it pain-free.

The biggest change allowed by this new model of bone stress injury is what comes next.  Warden, Davis, and Fredericson write that you can begin reintroducing running as soon as you've been pain-free during and after your everyday activities (including unassisted walking) for five to seven days. 

So, perhaps our high school runner is having pain walking around, so we put him in a boot for five days, and after that, he has no pain when walking normally.  After he's been out of the boot and pain-free (both during and after walking and daily activities) for a week, we can start him on a return-to-running program.

Warden, Davis, and Fredericson's return-to-running program

The reintroduction of running should follow a gradual, step-by-step process that moves from short intervals of easy jogging interspersed with walking breaks towards more typical running speeds and durations.  Warden, Davis, and Fredericson propose a three-phase reintroduction program. 

The first phase is a reintroduction to continuous running at very slow speeds (50% of your usual pace), with each day of running separated by a rest day.  By following a stepwise progression from three repetitions of walk 9min / jog 1min toward 8min walk / 2min jog, 7min walk / 3min jog, and so on, you can progress to 30min of continuous running every other day.  Progressing to the next scheduled session is only allowed if you have no significant pain during or after your run.

The second phase involves progressing back to your usual pace, moving from 30min at 50% of your typical pace to 60%, 70% , and so on until you're doing 30min at your typical pace, but still with a rest day in between.  Presumably "50% of normal pace" means multiplying your usual pace by 1.5 (so 50% of 8:00 mile pace would be 12:00 mile pace).

The final and shortest phase involves progressing to running every day again, by beginning to string together multiple days of consecutive running.  From here on out, you can continue to progress intelligently back to your usual training schedule (though hopefully a little more intelligently than the one that got you hurt in the first place!).

Warden, Davis, and Fredericson's sample return-to-running schedule. Click to enlarge.
Another major departure from traditional thinking is the appropriate response if pain does reoccur, either during or after a run session.  As mentioned earlier, it doesn't mean you're back to square one.  The proper response, according to Warden, Davis, and Fredericson, is to take one or two rest days, then return to the previous run session—the last one you were able to without pain.  Then, assuming you complete that session pain-free again, continue working your way back up.  This process is illustrated in the following flowchart.

Click here for full-scale version

Adjusting the return-to-running schedule

Warden, Davis, and Fredericson freely admit that their proposed return-to-running schedule is no more than an example.  They write that some runners will be able to use more aggressive schedules, and some may need an even more gradual approach.  Coming up with your own return-to-running plan will also be dependent on what your immediate goals are—an elite runner with a major championship in six weeks will likely elect to take a very different approach than a recreational runner who trains mainly for fun or fitness.

I've prepared two additional sample return-to-running schedules to illustrate how you might return to training in different situations after suffering a bone stress injury.  Both assume you've been pain-free during walking and normal daily activities for at least a week. 

The first schedule outlines a return to running during the "off-season" when you have no important competitions in the near future and when your main priorities are returning to full training and maintaining your fitness. 

Click to enlarge.
In this schedule, cross-training sessions (preferably aqua-jogging, since it is mechanically similar to running) serve as a replacement for the workouts you would be doing if you were healthy, and run sessions are used to work back towards daily continuous runs. 

Click to enlarge.

The second schedule depicts how you might return to training when a major competition is imminent.  This type of return is more aggressive and prioritizes returning to doing race-specific workouts as soon as possible, plus being able to race reasonably well at your major competition.  As such, it is a riskier plan, and you may find that pain has returned after your race.  In this case, you'll need to take more time off and restart a return-to-running plan to fully recover. In this plan, cross-training replaces hard workouts initially, but as you work your way back towards doing actual running workouts, cross training replaces easy runs instead.  Each week has several days of cross-training or time off in an attempt to allow maximum bone healing following the running workouts, which are reintroduced much earlier than usual. 

As a rule of thumb, your last one or two workouts before your major competition should be the same or almost the same volume as the race itself, at least for races 5k and under.  If you can't get through three miles of intervals without pain (or serious fatigue from lost fitness!), you won't be able to run well in your race.  These last two sessions can serve as a test-run to see whether your injury will hold up in a race.  If these are not successful, you should cancel your plans to compete. 

When following a more aggressive plan, it's important to emphasize that you must follow the rules outlined earlier.  You may only progress to the next run session if you had no significant pain during or after the previous one.  If you do experience pain, you need to take one or two days off, then re-do the last session you completed without pain.  If you're ever unsure of whether running on an injury is a good idea, you should consult your doctor, since no chart, plan, or scientific paper can ever take every aspect of your particular situation into consideration.

Again, these are just examples.  The ideal return-to-running program is tailored to you and is guided by avoiding any significant pain during or after any running session.  Don't over-glamorize aggressive return-to-running plans—they often involve setbacks and sometimes require more time off after the major competition to return to full health.  If you aren't in a rush to get back into running, there's nothing wrong with taking it nice and slow.


There are a few important takeaways from the new model of bone stress injuries.  First among these is the relative unimportance of distinguishing between a "stress reaction" and a "stress fracture."  If you have sharp, localized pain along a bone in your lower body, it's very likely a bone stress injury.  If you and your doctor can confidently rule out a high-risk bone stress injury, there's no real need for imaging of any kind (be it X-rays or MRI).  When the situation is unclear, it's probably best to skip the X-ray and go straight for the MRI: it's by far the best way to pinpoint what the problem is. 

If you confirm that you have a low-risk bone stress injury, you should work with your doctor, trainer, physical therapist, or coach to adopt a return-to-running plan that is dictated by your ability to handle activity without pain during or afterwards.  If you can't walk without pain, you need to be on crutches or in a boot.  Once you can walk normally and go about your daily activities pain-free for 5-7 days, you're ready to begin reintroducing running in a gradual fashion, progressing when you don't have pain, and backing off if you do. 

Finally, don't forget to investigate the reason you got a bone stress injury in the first place.  Simple training errors might have been the main cause, but there's a whole host of issues that can contribute to increased risk for bone stress injury.  You can read about them in my other articles on medial tibial stress syndrome and tibial stress fractures.


1.         Murray, S. R.; Reeder, M. T.; Udermann, b. E.; Pettitt, R. W., High Risk Stress Fractures. Pathogenesis, Evaluation, Treatment. Comprehensive Therapy 2009, 32 (1), 20-25.
2.         Warden, S. J.; Davis, I. S.; Fredericson, M., Management and Prevention of Bone Stress Injuries in Long-Distance Runners. Journal of Orthopaedic & Sports Physical Therapy 2014, 44 (10), 749-765.
3.         Bergman, A. G.; Fredericson, M.; Ho, C.; Matheson, G. O., Asymptomatic Tibial Stress Reactions: MRI Detection and Clinical Follow-up in Distance Runners. American Journal of Roentgenology 2004, 183 (3), 635-638.
4.         Franklyn, M.; Oakes, B., Tibial stress injuries: aetiology, classification, biomechanics and the failure of bone. In An international perspective on topics in sports medicine and sports injury, Zaslav, K. R., Ed. Intech: 2012; pp 509-534.
5.         Beck, B. R., Tibial Stress Injuires-An Aetiological Review for the Purposes of Guiding Management. Sports Medicine 1998, 26 (4), 265-279.
6.         Fredericson, M.; Bergman, A. G.; Hoffman, K. L.; Dillingham, M. S., Tibial stress reaction in runners: Correlation of Clinical Symptoms and Scintigraphy with a New Magnetic Resonance Imaging Grading System. American Journal of Sports Medicine 1995, 23 (4), 472-481.

About the Author

John J Davis, PhD

I have been coaching runners and writing about training and injuries for over ten years. I've helped total novices, NXN-qualifying high schoolers, elite-field competitors at major marathons, and runners everywhere in between. I have a Ph.D. in Human Performance, and I do scientific research focused on the biomechanics of overuse injuries in runners. I published my first book, Modern Training and Physiology for Middle and Long-Distance Runners, in 2013.

14 thoughts on “The bone stress injury model: a new way to deal with stress fractures and stress reactions in runners”

  1. I appreciate this article. I have been dealing with a tibial stress fracture for 14 months now. Seven months after the original symptoms began I suffered a complete fracture in the tibia. The fracture was non-displaced, but was treated with an IM nail. I am now eight months post surgery and still having some issues. My ortho is perplexed with the healing, as I have followed his recommendations. The timeline of injury, surgery and running is long, however I have discovered something today. The month before my surgery I was placed on the Rx medication Actonel - 150mg once a month - after a bone density scan showed Osteopenia. In reading your article today I have found the Bisphosphontes may inhibit bone healing. I was wondering if you had any experience of research on this subject?

  2. Indeed, in their article, Warden, Davis, and Fredericson mention that bisphosphonates could have an impact on bone healing. These drugs work by inhibiting the action of osteoclasts, the cells that eat away the old bone tissue. The idea is that this slows the process of loss of bone density, but the authors write that "preclinical studies showing that bisphosphonates delay BSI healing by suppressing intracortical bone remodeling and damage removal" and they cite these studies as evidence:
    Maybe bone strengthening and healing can ONLY occur when osteoclasts are allowed to eat away old or damaged bone tissue before the osteoblasts come in to fill in the gaps. Perhaps you can take these abstracts to your doctor and ask if this had anything to do with your delayed healing issues. Sorry to hear about your troubles! Good luck!

  3. This is really interesting, thank you for presenting a different approach in layman's terms. I am currently dealing with a suspected BSI with very perplexing MRI findings and presentation. The MRI revealed no (visible) cortical disruption and no periosteal edema, but moderate to severe bone marrow edema throughout both femurs. No pain with fulcrum, hopping, or other load+twisting motions done by a sports med doc. No pain walking either. Original presentation was knee pain at joint line... I've been trying to dig up information or even anecdotes about marrow lesions/edema and bone remodeling as I don't think they doc's prescription of crutches for 4 weeks is going to be the right answer and I have no idea how this happened. Have you come across anything about marrow edema without periosteal involvement? Neither the sports doc nor my GP seem concerned and are treating it like a basic stress fracture. Note: I do not have a history of BSI/sfx or female athlete triad issues. Thanks.

  4. Hmm. That's very puzzling. According to this paper on using MRI to grade the severity of bone stress injuries, even relatively mild stress fractures should have some cortical involvement. Did MRI rule out knee joint issues like meniscus tear or a soft tissue problem like pes anserius bursitis? If so I'd see a different orthopedist that has a lot of experience with athletic injuries. He or she may have seen something similar in the past.

  5. Thanks for the reply! I'm pretty well versed in sports injuries/anatomy/reading MRIs, and I still find my own scenario puzzling. A previous knee MRI ruled out any obvious soft tissue injury; mild thickening and increased signal of medial retinaculum. Xrays at 3 and 6 weeks post onset of serious pain showed some serious cortical thickening postero-medial midshaft, but it looked like old bone, no cloudy callus formation and no change between 3 & 6 weeks. I'm wondering if the marrow edema is just serious bone remodeling in an endurance athlete. I've seen some very small studies that have revealed transient bone marrow lesions in asymptomatic runners.

  6. I have been dealing with pain in the lower inside of my shin area and a burning/tingling sensation in the bottom of my foot on the same leg. Had a MRI in Nov. 2015, showed a stress reaction in the medial tibia area, took time off, went back to running in Feb. started with pain again. Had another MRI about 3 weeks ago, showed a sprained ankle and a healed sf that I didn't know I had. Still not sure what's causing the pain in the leg and burning/tingling sensation in the foot. Dr suspected tarsal tunnel, but unsure. Any idea what would cause the tenderness/pain in shin area & foot sensations. all on left side. did a little running in back yard last night and pain was felt around my left knee area, so I'm wondering if it's not in the knee instead of shin area. Thanks.

  7. Hi Anonymous,
    Are you still having issues? I just came across your comment. If they have ruled out Tarsal Tunnel, and you are having knee pain (is it on the medial or lateral side of your knee) - it could perhaps be Saphenous nerve irritation in your medial quad/adductor. Have a physio take a look at your hip flexor and quad length!

  8. Agreed that this is by far the best article I have read too and have read a ton. I have a fibula stress fracture that is taking forever to heal as well. Two steps forward, one step back. After five months I am still just trying to be able to walk. Latest stretch was ten days before I felt some soreness again and this is about 3,000 steps a day only. Back to crutches for a few days and then try again. You've given me hope that I am getting nearer to being able to do all my outdoor activities again, THANK YOU!

  9. Quick question. Does a compression sleeve in any way for a bone stress injury?
    Thanks for a great detailed article but still in layman terms

  10. Have you seen much evidence related to a high risk stress fracture? I have an anterior mid-tibia stress fracture. First diagnosed Feb 2016 with x-ray showing the "dreaded black line". Time off, returned to tennis, pain again, more time off etc. X-rays from Aug and Nov 2016 unchanged and then a Dec x-ray showed slight improvement. Only thing I can contribute to that is I began Vit D and resumed regular exercise without any running, jumping, tennis. Then Feb 2017 X-ray showed no additional improvement. Can walk 10 miles with no pain but the minute I start to run it gets sore to the touch. Stays that way for a couple of days and then is fine. Very frustrating. Any advice?

  11. Nicely summarized. I would say most sports medicine physicians will end up getting an MRI to exclude other diagnosis. It was mentioned but if you have the triad or poor energy intake your progress will most not be linear in a positive direction. Well written.

  12. Very interesting and helpful article thank you. I have suffered a number of bone stress injuries and am trying to improve my bone density. Multi-directional explosive jumps/hops are commonly recommended - 20-30 2x day. I am wondering if adding these in with a run program would also temporarily weaken the bone/put me at risk for another injury as it remodels/adapts to this stress, or are these safe to add as the number of loads is so minimal? Thanks

    • Good question! I think you'd want to hold off on adding in plyometrics - those are more of a long-term fix, given that it takes ~6 months or so for real bone strength adaptations to really come online. Once you are back to your usual training is probably when you'd want to (gradually) start doing plyometrics. Or better still, start them before you get hurt!


Leave a Comment

Did you know I have a book? Check it out here!