Injury Series: Uncovering the role of hip strength and mechanics in the cause and treatment of patellofemoral pain syndrome

Today we are tackling the most common running injury of all: patellofemoral pain syndrome, alternatively known as runner’s knee, anterior knee pain, and (sometimes erroneously) chondromalacia or chondromalacia patellae.  As is often the case, there is some controversy about the naming.  “Patellofemoral pain syndrome” or PFPS is the broadest and most common term, but is often criticized for being a “wastebasket term”—an umbrella definition for any unexplained anterior (frontal) knee pain.  Historically, it was termed chondromalacia patellae, which is a softening of the cartilage under the kneecap.  It turns out that not all cases of PFPS involve softened cartilage, and not all people with softened cartilage have patellofemoral pain! “Runner’s knee” is another common term but is perhaps the most useless.  It doesn’t tell us anything about the problem, other than that you’re a runner and your knee hurts.  Because there is still lack of a specific injury mechanism, virtually all scientific articles use the term “patellofemoral pain syndrome,” as will I.
Patellofemoral pain syndrome is usually characterized by a pain “behind” or “around” the kneecap that is aggravated by squatting, running (particularly downhill), walking down stairs, and prolonged sitting.  In runners it usually starts as a mild irritation and might go away once you’ve gotten warmed up.  But over time it tends to get worse, being aggravated even by short, easy runs.  There is also usually tenderness around the upper edges of the patella and pain if you contract your quadriceps against resistance (on a leg extension weight machine, for example) or compress the patella downward “into” the knee joint. 
“Patella” is the formal name for the kneecap.  The patella itself is a triangular bone attached on one end to the quadriceps muscles via the quadriceps tendon and attached on the other end to the top of the tibia via the patellar tendon.  The patellar tendon is also a common injury location, and patellar tendonitis was covered in an earlier installment in the Injury Series posts.  Note that it’s important to distinguish PFPS from patellar tendonitis,as the ideal treatments are different! The main function of the patella is to give the quadriceps a mechanical advantage in knee extension.  By raising the tendons away from the center of the joint, the knee can generate more power.  The drawback of this is that the patella transmits virtually all of the force to the lower leg during knee extension.  As such, it may be inherently prone to overuse. 
Adopted from Dixit et al.
The backside of the patella slides back and forth during knee flexion and extension in a groove in the femur aptly called the patellar surface.  The assumption is that the development of patellofemoral pain syndrome is due to improper tracking of the patella within this groove.  It’s easy to visualize how a patella that tracks too far towards or away from the center of the body would rub up against the sides of this groove.  Where this improper tracking comes from is an interesting puzzle, and will be dealt with later.  If you’ve read my other Injury Series articles, you know that a fundamental tenant of my approach to managing injuries is that pain is indicative of real, physical damage.  What’s not clear in the case of PFPS is where that damage is occurring.  As mentioned earlier, the common assumption for a long time was that PFPS was caused by chondromalacia  or a softening of the cartilage on the patellar surface brought about by improper patellar tracking.  While chondromalacia is not an uncommon finding among patients with knee pain, PFPS alone is not predictive of cartilage softening or damage, nor is the severity of PFPS pain, as first demonstrated in 1986 by Lindberg, Lysholm, and Gillquist,1 and confirmed in a 2010 surgical study by Pihlajamaki et al.2  While this is good news, since cartilage is not renowned for its healing abilities, it means that we're not entirely certain where the damage/irritation is occurring.  Essentially, the “syndrome” in “patellofemoral pain syndrome” means that we don’t know exactly why it hurts.
Epidemiology and risk factors: prospective studies
Fortunately, though we may not know quite where the damage is happening, we do have some idea of who gets PFPS and perhaps a clue as to why.  It’s the most common overuse injury in all sorts of recreational athletes, from runners to walkers to basketball players.  Among runners, one recent study found that it accounts for 16.5% of all running injuries.  Women are significantly more likely than men to suffer PFPS; they account for 62% of cases.3 
Because of this gender discrepancy, researchers initially suspected that differences in hip and femur structure in women predisposed them to a higher risk of knee injuries.  The “Q angle” is a factor that measures the “angle” of the femur relative to the pelvis.  Since women on the whole have relatively wider hips than men, their Q angle is also larger.  This model tied inwell with the pronation paradigm, since some studies found Q angles to be related to foot pronation.  However, a well-designed prospective study published in 2000 by Witvrouw et al. found no correlation between Q angle, foot pronation, or leg length differences amongst college-aged recreational athletes who would go on to develop PFPS in a two-year period and those who would not.4  Despite these issues being discarded as possible causes by the scientific community some years ago, they are still quoted as “causes” of PFPS on many websites, and I suspect also by many doctors who don’t keep tabs on the latest research.  In any case, we need to turn elsewhere to unravel the causative factors.
For such a common injury, there is an alarming lack of prospective studies on patellofemoral pain syndrome—especially in runners.  The above study by Witvrouw et al. is one of the better designed ones, but it used active college students, not runners specifically.  In short, that study found students who developed PFPS didn’t spend any more time exercising than those who did not, pointing to “intrinsic” risk factors (vs. “extrinsic” ones like training volume and intensity) as being very important.  The factors that were associated with development of PFPS were poor quad and calf flexibility, poor explosive strength in the legs (measured by a vertical jump test), abnormal quadriceps firing patterns, and a hypermobile patella.  A meta-analysis of prospective studies on PFPS published this month, albeit in a relatively obscure journal, concluded that weak quadriceps were a risk factor as well, which we'd expect given the reasonable success of treating PFPS with quad strengthening exercises.5  While not all of these are easily measurable or even fixable, we can begin to infer some things about PFPS.  Certainly, several of these factors suggest that the development of PFPS is associated with abnormal forces going through the knee joint.  Poor explosive muscle strength, for example, would force the knee joint to absorb more of the impact during running, and inflexible or weak quads might increase forces across the knee joint as well. 
Abnormal quadriceps firing patterns is a common finding in patients with patellofemoral pain syndrome, and as such, treatments were devised to address this by specifically targeting the muscle that appears to be firing improperly: the vastus medialis (sometimes referred to as the “vastus medialis obliquus” or VMO—whether or not this term is “correct” is debatable).  Unfortunately, several recent studies of muscle activity in a variety of exercises designed to “target” the vastus medialis reveal that there is no greater activation in that muscle than the rest of the quadriceps muscles, and any benefits from these exercises are coming from general quad strength gains, not “VMO activation.”6  So yet again another clinical myth is shot down.  
Turning to retrospective studies for answers
Having exhausted the relatively limited pool of prospective studies and having overturned a few theories already, we must turn to retrospective studies for a clearer picture of who gets PFPS and why.  Before, though, it’s important to note the caveats: since retrospective studies look at people who are already injured, it’s hard to differentiate cause and effect.  For example, if we evaluated a sample of runners with PFPS, we might find that they limped on the injured side while jogging.  But did the limp cause the injury, or (more likely) did the injury cause the limp? Patellofemoral pain syndrome tends to be a chronic issue, and many of the studies we’re about to examine require the participants to have had knee pain for at least a month, so it’s very possible that the injured athletes have adopted what you might call “pain avoidance” behaviors—essentially, a limp. 
That being said, there is a wealth of information from retrospective studies.  By far the most common finding is hip strength deficiency.  Among women, the evidence is perfectly clear, as summed up by a 2009 review titled simply “Women with Patellofemoral Pain Syndrome Have Weak Hip Muscles.”7  Studies dealing specifically with runners are also easy to find.8  However, research looking at men specifically is scant or non-existent.  Despite this, I don’t doubt that men with PFPS will also display hip muscle weakness, for two reasons:
1) A landmark 2005 study by Paul Niemuth et al. at the College of St. Catherine9 found that recreational runners (of both genders) who had suffered any injury tended to have weak hip muscles, both comparing their injured leg to their healthy one and relative to healthy control runners. 
2) While a 2003 study by Irene Davis’ lab at the University of Delaware showed that healthy women runners demonstrate significantly greater hip adduction and internal rotation compared to men,10 the stride patterns of males and females are not qualitatively different—that is, healthy men and women run more or less the same way, despite showing some differences in joint angles at the hip and knee.  Indeed, as we’ll soon see, these gender differences may explain why females are disproportionately represented among PFPS cases. 
If you’ve read my article on anatomic terms of location and motion, you know that the hip can move in six ways: flexion, extension, adduction, abduction, internal rotation, and external rotation, and each of these motions have an associated muscle group to control them.  So, which muscle groups are found to be deficient in athletes with patellofemoral pain syndrome?  As it turns out, almost all of them! Maarten Prins’ systematic review paper states:7
Strong evidence was found for a deficit in hip external rotation, abduction and extension strength, moderate evidence for a deficit in hip flexion and internal rotation strength, and no evidence for a deficit in hip adduction strength compared with healthy controls
Prins goes on to state that there are likely strength deficits in the external rotators and abductors, but not extensors, flexors, adductors, or internal rotators between the injured and uninjured sides.  That is, a runner with PFPS in the right knee will have weaker hip external rotators and abductors on the right side than the left.   
Hip mechanics and knee injury
Model of excessive hip adduction/internal rotation.13
While many studies have measured and confirmed the presence of hip muscle weakness in athletes with PFPS, a few more ambitious researchers have investigated biomechanical factors associated with PFPS.  To do so requires a 3D kinematics laboratory with either a treadmill or runway to measure gait differences.  Everyone agrees that athletes with PFPS demonstrate altered biomechanics compared with healthy controls.  But there is some disagreement on exactly how they differ.  Recall earlier we saw that there is no discernible difference in foot pronation between injured and uninjured runners.  The differences appear at the hip and knee.  One study by Richard Souza and Christopher Powers11 found that athletes with PFPS displayed increased hip internal rotation while running, jumping, and descending a step vs. healthy controls.  This was accompanied by a decrease in hip muscle strength.  On the other hand, a study by John Willson and Irene Davis12 found an increase in hip adduction but a decrease in internal rotation.  The authors discuss these differences and pointed out the very real possibility of the limping-like “pain avoidance” patterns we talked about earlier. 
I believe the initial biomechanical cause of PFPS is abnormal mechanics related to hip adduction and internal rotation, and that the lowered internal rotation seen in some studies is a result of subjects trying to avoid pain in their patella by shifting the stress to other parts of the knee.  My case for this stands principally on three pieces of evidence. 
First among these is the weakness in hip muscles found in athletes with PFPS.  The most-documented differences are those in hip abduction and external rotation strength—the very muscles which should be controlling and limiting hip adduction and internal rotation.  It’s no stretch to propose that hip abductor and external rotator weakness causes excessive hip adduction and internal rotation (for one study which shows such a link, see 11).  Indeed in my previous article on IT band syndrome, I argued that this was the case in ITBS as well.  The connection between muscular strength and hip dynamics is not a 1:1 relation, however.  We’ll discuss the nuances of this shortly.
My second piece of evidence comes from a masterful review of knee injury from a biomechanical perspective authored by Christopher Powers in 2010.13  In it, he describes the effects of abnormal hip mechanics on the knee:
Knee valgus during a lunge.  From Thijs et al.
Excessive hip adduction and internal rotation during weight bearing has the potential to affect the kinematics of the entire lower extremity. More specifically, excessive hip adduction and internal rotation can cause the knee joint center to move medially relative to the foot. Because the foot is fixed to the ground, the inward movement of the knee joint causes the tibia to abduct and the foot to pronate, the end result being dynamic knee valgus [an inward “leaning” of the knee, see right].  Excessive knee val­gus has been shown to be related to di­minished hip muscle strength and has been implicated in contributing to numerous knee injuries, including ante­rior cruciate ligament (ACL) injury and patellofemoral joint dysfunction.
Additionally, Powers puts to rest the “patellar tracking” myth with a summary of some very clever MRI studies.  Recall earlier that the prevailing opinion in orthopedic circles was that PFPS is a result of abnormal patellar tracking, and as such must be addressed by quadriceps strengthening and perhaps altering the firing patterns of the quadriceps muscles. Using dynamic MRI imaging, a technique which can capture real-time changes inside the body, Powers and his coworkers at UCLA demonstrated that, while the patella does indeed track abnormally in the patellar groove of the femur in patients with PFPS, this only occurs in non-weight bearing motions.  When the patient does a weight-bearing motion, like a squat, the strong contraction of the quadriceps muscles essentially “locks” the patella in place.  Instead, the femur rotates relative to the kneecap! So while on the outside it may appear that the patella is moving, in fact it’s the femur that’s moving!  Running, of course, is also a weight bearing activity, so I expect the same mechanism to be at work there too.
Left: the knee in a neutral, straightened state.  Right: during a weight-bearing squat, the apparent lateral tracking of the patella is actually due to an internal rotation of the femur.  Adopted from Powers et al.
Finally, a recent conference proceeding makes perhaps the best case for the biomechanical roots of PFPS.  The study, conducted by Irene Davis’ lab, is the first prospective study of PFPS in runners that has linked faulty biomechanics to patellar pain.14  In it, a large group of healthy female runners underwent a 3D kinematic evaluation and were followed for several years.  Thirteen runners developed PFPS, and as predicted, they showed significantly greater hip adduction than matched control runners who stayed healthy.  They also had greater hip internal rotation, though the difference did not reach statistical significance.  Hopefully Davis and colleagues have done further analysis (perhaps with additional subjects) and will publish a comprehensive article on their findings soon. 
The evidence building in favor of a connection between hip strength, poor biomechanics, and patellar injury also supports the observed gender disparity in runners with PFPS.  As mentioned earlier, women outnumber men 3:2 among those who suffer from PFPS.  It should be no surprise that women as a whole display both greater hip adduction and internal rotation during running10 as well as lower hip external rotator and abductor strength.15
Correcting biomechanical faults
We should not be too hasty to directly connect muscular weakness with biomechanical faults.  There are two ways to approach a biomechanical problem like excessive hip adduction and internal rotation.  The first that comes to mind is strength training for the hip abductors and external rotators, but this isn’t the only way to correct such faults.  Remember that running (or any dynamic activity) is not simply a function of how strongly various muscles contract, but also when.  The timing of muscular contractions is what a biomechanics researcher would call neuromuscular coordination.  While this doesn’t exist wholly outside of muscular strength, some researcher suggests that it’s possible to have adequate hip strength but inadequate neuromuscular coordination.  This may explain why the connection between hip strength and hip mechanics is a bit murkier than we’d expect, and why some researchers are not finding a strong connection between the two (for a summary, see 16).
Next, we’ll look at two different studies which have taken opposite approaches to correcting PFPS—one which trained its subjects to exert better neuromuscular control of their hips and another which prescribed only a hip strengthening program and no neuromuscular control training.
Treating patellofemoral pain syndrome through gait retraining
The first was conducted by Irene Davis’ lab.17 Ten runners with PFPS were recruited for the study, and all underwent a 3D kinematic evaluation at the outset.  However, instead of prescribing a set of exercises, Davis et al constructed a computer display on a treadmill that gave the runners real-time feedback on their hip adduction.  During a two-week intervention which involved four laboratory visits per week, the subjects ran on the treadmill and were shown their current hip adduction as a function of the stance phase in real time, as well as a “normal” region based on data collected from healthy runners.  They were told to modify their stride to try to move their own line inside the “normal” region.  Over the eight sessions the runners were gradually eased off the real-time feedback—the fourth session, for example, was a 25min run with 25min of real-time feedback, whereas the sixth session was a 30min run with 15min of real-time feedback.  After the last session, the runners were allowed to run on their own for a month, and then returned to the laboratory for another 3D kinematic evaluation to see whether they’d reverted to their old habits. 
The subject's hip adduction curve is the top white line.  From 17
Fortunately, the effects of the intervention were maintained a month after.  As shown by the graph below, the runners maintained a lower hip adduction angle throughout the stance phase as a result of the gait retraining.  More importantly, at the one month follow-up, all ten subjects reported no knee pain.  

One month after the gait retraining, the runners maintained their lowered hip adduction. From 17
However, while the success of this study was impressive, it’s important to note that Davis et al. used a highly selective process in determining the subjects: 85 runners with PFPS underwent video screening, but of these, only 19 were qualified for the study (incidentally, all were female).  This perhaps means that Davis et al. did not feel that the other 66 runners would benefit from the intervention—maybe their hip adduction was already within the “normal” range.  Furthermore, Davis et al. did not include a control group.  It’s possible that a more traditional rehab program consisting of stretching and strengthening exercises might also result in reduced hip adduction and internal rotation, especially if that program included hip strength.  Finally, real-time 3D kinematic feedback is not an accessible treatment option for most runners.  If gait retraining turns out to be a powerful intervention for PFPS and other injuries, it will need to be made accessible for a broader range of patients.
Treating patellofemoral pain syndrome through strength training
Another way to control excessive hip adduction and internal rotation is through strengthening the hip abductors and external rotators.  Tracy Dierks, a professor at Indiana University is carrying out an ongoing study that intends to do just that.18  The design is fairly simple: upon enrollment in the study, a female runner with PFPS is assigned either to the control group or the experimental group.  The experimental group undergoes a 3D kinematic evaluation, returns home with a set of hip strength exercises to be done for six weeks, after which they undergoes another 3D kinematic evaluation.  The control group undergoes a kinematic evaluation as well, returns home to continue their normal running routine for six weeks, and are evaluated again after.  Afterwards, these control runners are re-enrolled in the experimental group, given the same 6-week exercise routine, and re-evaluated at the end—both to speed up gathering evidence, and because it’d be a bit cruel to provide then no treatment at all! At a 2011 conference, Dierks presented his results to date.  So far, eight runners have completed the study (four experimental, four controls).  The control group showed virtually no change in their running mechanics during the six-week control period, whereas all eight of the runners showed significant changes in mechanics after the six-week strengthening program, and all eight reported their knee pain had reduced from a “7/10” on a treadmill running test to less than 2/10.  The best news about Dierks’ study is that the exercises prescribed were simple and fairly mundane: single-legged squats and resistance band exercises targeting all motions of the hip (flexion, extension, abduction, adduction, internal and external rotation). 
Interestingly, the strength program not only effectively treated the runners’ PFPS; it also improved their running mechanics.  The eight subjects showed better joint control and less pain-avoidance behavior (i.e. less of a “limp”) after the six-week strength program.  In an email to me, Dr. Dierks wrote,

As for hip kinematics, I did not see a change in hip adduction or hip internal rotation peaks. For peak angles, I saw changes in rearfoot eversion, tibial internal rotation, and knee flexion. But this doesn't necessarily mean that nothing changed at the hip, which is why I also examined joint timing, where I did indeed see changes that involved the femur, which is part of the hip [...] in my pilot study, hip adduction and internal rotation peak values seemed normal. In other words, the peak values were similar to peak values or uninjured runners, or those that we would call normal. But this doesn't mean that the hip is not the primary reason for PFP. I think the biggest contributor is weakness in the hip/core which produces an inability to control the leg throughout an entire gait cycle; not just one point in time (i.e. a peak angle).

This evidence hints that perhaps the peak hip abduction and internal rotation may not tell the whole story.  Could a lack of hip strength result in poor control over the hip, even if the average or peak motion remains the same?

Average or peak joint mechanics may not be the whole story

Some answers on this come from a 2011 study by Reed Ferber, Karen Kendall, and Lindsay Farr at the University of Calgary in Canada.19  They prescribed a short and simple two-exercise hip abductor and external rotator strength program for three weeks to a group of patients with PFPS.  While Ferber et al. only used 2D kinematic analysis and thus we can’t directly compare to 3D studies, their results still highlighted two important points.  The first is that, while the average knee genu valgum angle (a combination of hip adduction and tibial abduction) didn’t change after the strength intervention, the stride-to-stride variability did! In other words, the runners’ gait becomes more consistent after the strength training.  As we might predict, their knee pain levels also decreased significantly.  This illustrates how simply using the average hip adduction or internal rotation angles may lead us astray.  So perhaps some of the runners that were screened out of Irene Davis’ gait retraining study for having normal average or peak hip adduction and internal rotation angles still have abnormal biomechanics which manifest in greater stride-to-stride variability.  Future research needs to take this into account! Secondly, this study also provides evidence that strength training affects neuromuscular control.  How? Well, three weeks isn’t long enough to see any significant changes in muscular strength—most or all of the improvements in biomechanics must have come from the hip muscles being recruited more efficiently, rather than the hip muscles gaining raw strength.  Furthermore, this indicates it is indeed possible to alter neuromuscular coordination and control without a complicated real-time kinematics setup.
It appears that, in some cases, PFPS may be associated with greater stride variability as well.  Strength training appears to lessen the stride-to-stride variability in injured runners, even if the average joint kinematics don't change.  From 19
Other treatments: taping, bracing, and custom orthotics
One of the most common treatments prescribed for PFPS is the use of patellar taping or a patellar brace—anyone who’s been to a road race has probably seen runners sporting PattStraps or neoprene knee sleeves.  While several studies have found patients get short-term relief through taping or bracing the patella, neither taping nor bracing seem to confer any additional benefit when compared to strength exercises for the quads and hips.20  Additionally, these devices are all based around the faulty premise that the patella is tracking improperly relative to the femur—a myth debunked by Powers et al.13  Why do they exhibit any benefit at all? Older studies used a control group which either received no treatment at all or a “sham” treatment of fake ultrasound or similar, and these studies found that patellar taping provided a short-term benefit.  But newer studies which have tested a “properly” applied patellar taping against a loose taping intentionally applied wrong have found little or no difference in pain relief!  Bolgla and Boling explain:20
Interestingly, results from this review indicated that the manner of tape application (i.e. either applied in a corrective manner or applied loosely) may not necessarily influence its beneficial effects.  Therefore, taping may have an important effect on the neuromuscular system rather than actually altering patella movement.
This neuromuscular effect may explain the rising popularity of taping strategies like Kinesio Tape.  With these methods, the tape is so thin and flexible that there’s no way it could restrict abnormal patellar motion to a meaningful extent, yet trainers and athletes report good results! Wearing a patellar strap or getting your patella taped (either with “old fashioned” athletic tape or fancy new stretchy tape) may provide some short term relief, but don't rely on it as a long-term solution.
The reason why a tight, "correct" patellar taping (right) is about as effective as a loose, poorly applied taping may be the same reason why flexible "kinesiology tapes" (left) are becoming more popular: the benefit is not so much in the forces  exerted on the joint, but in the tactile stimulation to the skin.  Images from and
Custom orthotics, though commonly prescribed for patellofemoral pain, show a similar pattern: though they may reduce some pain in the short term, when added to a standard strength protocol, they confer no additional benefit.21  While it’s important to note that some runners may still benefit from foot orthotics, there is no evidence for any systemic effect on patellofemoral mechanics.  As explained in my previous article on the foot pronation paradigm, the function of foot orthotics does not seem to be to “align” the foot and ankle.  Rather, they alter muscle activation patterns, sometimes in a positive manner and sometimes not.  Also keep in mind that faulty hip mechanics may increase foot pronation by increasing tibial abduction/knee genu valgum, so looking to the foot to correct this is probably barking up the wrong tree.
Finding the best solution
The best treatment for patellofemoral pain will probably turn out to be a synthesis of the traditional quadriceps strengthening exercises that have been prescribed for many years, hip strength targeted at the muscle groups that control hip mechanics during running (like the abductors and external rotators), and new techniques that retrain your body to run with better mechanics, all supported by ancillary treatments to reduce short-term pain like icing and taping.  While gait retraining is an exciting new frontier in biomechanics research, there are still many more unknowns: Do runners continue to maintain their “new” gait for several months/years after retraining? Does this new gait prevent new injuries? How much of abnormal running mechanics is a result of inadequate neuromuscular control and how much of it is a result of a lack of muscular strength? These are all very important questions that need to be addressed before gait retraining is used on any large scale as a treatment. 
It's unlikely you have one of these in your basement

Since I doubt many readers of this article have a 3D kinematics laboratory at their disposal (like the one pictured to the right, the UVA Speed Clinic), the best program I can recommend for home treatment of patellofemoral pain syndrome is a rehab protocol that focuses on the factors we’ve found to be associated with PFPS: poor quad and calf flexibility, weak quadriceps, and weak hip muscles—particularly those which limit excessive adduction and internal rotation: the hip abductors and external rotators.  As usual, I’m not alone in recommending the addition of hip strength to the standard PFPS rehab protocol.  A 2011 review of recent developments in treatment for patellofemoral pain syndrome concluded:

The findings of the current review showed that hip strengthening exercise can benefit individuals with PFPS. Moderate evidence supports the use of hip abductor and external rotator strengthening, which may be further enhanced with the inclusion of exercises targeting hip flexion and hip extension.  Although all works prescribed exercise for strengthening effects (i.e., 3 sets of 10 to 15 repetitions), emerging evidence has suggested the need to address muscle endurance. Therefore, clinicians should consider exercise dosage focusing on higher repetitions (i.e. 3 sets of 20 to 30 repetitions), especially in patients who participate in more demanding activities like running and jumping.20
If you want a "scientifically approved" program, below is a selection of various rehab protocols from several papers that have successfully used them to treat patients with PFPS.  Two are fairly complicated; the third is quite simple.  Incidentally (and perhaps paradoxically), the first two "complicated" programs were designed for the general population, whereas the simple two-exercise program was designed for runners.  22 23 19

 Click on the images to enlarge
The best course of action is not entirely clear.  Do most runners really need a complicated set of ten or twelve exercises to carry out every day which changes every week? Based on the impressive short-term success of Ferber et al.’s extremely simple program of two exercises, I have a hunch the answer is “no.”  Since there’s no consensus on exactly what constitutes a “good” rehab protocol, it’s hard to recommend any one particular program.  I suspect most injured runners won’t want to wait around, however, so here’s a rough idea of what most programs include:
1) Hip abduction and external rotation: either standing with resistance band or laying on side
Click to enlarge
2) Hip flexion and extension: either standing with resistance band or on stomach/back

Click to enlarge
3) Single-leg mini-squats: usually off a ~6in step, facing forward and/or to the side
From 23
4) Straight leg raises or other isolated quadriceps strength exercise
5) Single-leg balance, possibly on an unstable surface (wobble board, foam mat, bosu ball)

Most programs recommend doing these exercises every day, typically three sets of anywhere from 10-30 repeats each (or a few minutes for balancing exercises).  Starting at perhaps two sets of 15 repeats and building to three sets of 30 over several weeks is not a bad idea.  I would also recommend adding quadriceps, calf, and possibly hamstring stretching—about two or three sets of 20-30 seconds each, based on the evidence above that people with poor quad and calf flexibility are at higher risk for developing PFPS. 

Please note that THIS IS NOT A COMPREHENSIVE REHAB PROGRAM! Unfortunately, since there have not been any "home run" studies that use runners, have a simple strength protocol, and show good results, I can't recommend any one particular set of exercises.  The above is only a SAMPLE of exercises commonly found in the literature.  I have high hopes for a few studies that are still "in the pipes," and I will update this article as soon as there's a good study with a rehab protocol that I can recommend.  Until then, I recommend working with a good physical therapist who is familiar with runners if at all possible to strengthen your quads and hips.  You could also "adopt" one of the three programs detailed above, taken from the scientific literature.

Conclusion: patellofemoral pain is best treated through a combination of hip and quad strengthening (for now)
We’ve seen (as usual) that much of the prevailing wisdom you’ll find out on the internet about patellofemoral pain syndrome is incorrect.  It does not seem to be primarily a patellar tracking issue that can be controlled through taping, strapping, and VMO activation.  It’s not associated with anatomic factors like Q-angle and foot pronation, the source of pain is not always a softening of knee cartilage, nor is its sole cause weak quadriceps.  Instead, the most important risk factors are quad strength, quad and calf flexibility, hip muscle strength (particularly abductors and external rotators), and excessive hip adduction and internal rotation during running.  Women are at greater risk of developing PFPS, likely because they tend to run with greater hip adduction and internal rotation and also have weaker hip abductors and external rotators than men.
Patellofemoral pain syndrome seems to be a biomechanical issue: excessive hip adduction and internal rotation (or possibly poor control of the hip joint) cause an apparent shift in patellar tracking due to internal rotation of the femur.  While the exact connection between muscular strength, neuromuscular coordination, and abnormal biomechanics has yet to be revealed, there are several promising studies that indicate hip strength and neuromuscular control is at least as important as quadriceps strength in addressing the root cause of patellofemoral pain syndrome. The most effective home treatment program will consist of a daily routine of both hip and quad strength, possibly supplemented by quad, calf, and hamstring stretching and ancillary treatments like patellar taping.  As with many injuries, foot orthotics may provide relief to some people, but do not confer any additional benefit when added to a standard hip and quad strength program.  Sometime in the near future we may see gait retraining emerge as a standardized treatment option easily available at physical therapy offices, but until then, hip and quad strength will likely be the mainstay for correcting faulty hip biomechanics in runners.


1.            Lindberg, U.; Lysholm, J.; Gillquist, J., The correlation between arthroscopic findings and the patellofemoral pain syndrome. Arthroscopy: The Journal of Arthroscopic and Related Surgery 1986, 2 (2), 103-107.
2.            Pihlajamaki, H.; Kuikka, P.; Leppanen, W.; Kiuru, M.; Mattila, V., Reliability of clinical findings and magnetic resonance imaging for the diagnosis of chondromalacia patellae. The Journal of Bone and Joint Surgery, American Volume 2010, 92 (4), 927-934.
3.            Taunton, J.; Ryan, M.; Clement, D.; McKenzie, D.; Lloyd-Smith, D.; Zumbo, B., A retrospective case-control analysis of 2002 running injuries. British Journal of Sports Medicine 2002, 36, 95-101.
4.            Witvrouw, E.; Lysens, R.; Bellemans, J.; Cambier, D.; Vanderstraeten, G., Intrinsic Risk Factors For the Development of Anterior Knee Pain in an Athletic Population: A Two-Year Prospective Study. The American Journal of Sports Medicine 2000, 28 (480-489).
5.            Pappas, E.; Wong-Tom, W., Prospective Predictors of Patellofemoral Pain Syndrome. Sports Health 2012, 4 (2), 115-120.
6.            Smith, T. O.; Bowyer, D.; Dixon, J.; Stephenson, R.; Chester, R.; Donell, S. T., Can vastus medialis oblique be preferentially activated? A systematic review of electromyographic studies. Physiotherapy Theory and Practice 2009, 25 (2), 69-98.
7.            Prins, M. R.; van der Wurff, P., Females with patellofemoral pain syndrome have weak hip muscles: a systematic review. Australian Journal of Physiotherapy 2009, 55, 9-15.
8.            Ireland, M. L.; Willson, J. D.; Ballantyne, B. T.; Davis, I. S., Hip Strength in Females With and Without Patellofemoral Pain. Journal of Orthopaedic & Sports Physical Therapy 2003,  (33), 671-676.
9.            Niemuth, P. E.; Johnson, R. J.; Myers, M. J.; Thieman, T. J., Hip Muscle Weakness and Overuse Injuries in Recreational Runners. Clinical Journal of Sports Medicine 2005,  (15), 14-21.
10.          Ferber, R.; McClay Davis, I.; Williams Iii, D. S., Gender differences in lower extremity mechanics during running. Clinical Biomechanics 2003, 18 (4), 350-357.
11.          Souza, R. B.; Powers, C. M., Differences in hip kinematics muscle strength and muscle activation between subjects with and without patellofemoral pain. Journal of Orthopaedic & Sports Physical Therapy 2009, 39 (1), 12-19.
12.          Willson, J. D.; Davis, I. S., Lower extremity mechanics of females with and without patellofemoral pain across activities with progressively greater task demands. Clinical Biomechanics 2008, 23 (2), 203-211.
13.          Powers, C. M., The Influence of Abnormal Hip Mechanics on Knee Injury: A Biomechanical Perspective. Journal of Orthopaedic and Sports Physical Therapy 2010.
14.          Davis, I.; Noehren, B. In Prospective Study of the Biomechanical Factors Associated with Patellofemoral Pain Syndrome, American Society of Biomechanics Meeting, Palo Alto, CA, Palo Alto, CA, 2007.
15.          Leetun, D. T.; Ireland, M. L.; Willson, J. D.; Ballantyne, B. T.; Davis, I. M., Core Stability Measures as Risk Factors for Lower Extremity Injury in Athletes. Medicine & Science in Sports & Exercise 2004, 36 (6), 926-934.
16.          Heiderscheit, B. C., Lower extremity injuries-is it just about hip strength? Journal of Orthopaedic & Sports Physical Therapy 2010, 40 (2), 39-41.
17.          Noehren, B.; Scholz, J.; Davis, I., The effect of real-time gait retraining on hip kinematics, pain and function in subjects with patellofemoral pain syndrome. British Journal of Sports Medicine 2010, 45 (9), 691-696.
18.          Dierks, T. A.; Phipps, R. L.; Cardinal, R. E.; Altenburger, P. A., The Effect Of Hip Muscle Strengthening On Pain and Running Mechanics In Females With Patellofemoral Pain. Medicine & Science in Sports & Exercise 2011, 43 (5), 93.
19.          Ferber, R.; Kendall, K. D.; Farr, L., Changes in Knee Biomechanics After a Hip-Abductor Strengthening Protocol for Runners With Patellofemoral Pain Syndrome. Journal of Athletic Training 2011, 46 (2), 142-149.
20.          Bolgla, L. A.; Boling, M. C., An Update For The Conservative Management Of Patellofemoral Pain Syndrome: A Systematic Review Of The Literature From 2000 To 2010. The International Journal of Sports Physical Therapy 2011, 6 (2), 112-125.
21.          Collins, N.; Crossley, K.; Beller, E.; Darnell, R.; McPoil, T.; Vicenzino, B., Foot orthoses and physiotherapy in the treatment of patellofemoral pain syndrome: randomised clinical trial. British Journal of Sports Medicine 2009, 43 (3), 163-168.
22.          Boling, M. C.; Bolgla, L. A.; Mattacola, C. G.; Uhl, T. L.; Hosey, R. G., Outcomes of a Weight-Bearing Rehabilitation Program for Patients Diagnosed With Patellofemoral Pain Syndrome. Archives of Physical Medicine and Rehabilitation 2006, 87 (11), 1428-1435.
23.          Dolak, K. L., Hip Strengthening Prior to Functional Exercises Reduces Pain Sooner Than Quadriceps Strengthening in Females With Patellofemoral Pain Syndrome: A Randomized Clinical Trial. Journal of Orthopaedic and Sports Physical Therapy 2011.

Updated on 3/21: added comments made by Dr. Tracy Dierks regarding his pilot study

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.

3 thoughts on “Injury Series: Uncovering the role of hip strength and mechanics in the cause and treatment of patellofemoral pain syndrome”

  1. Thanks so much for this - most thorough treatment of PFPS I have seen. I had just noticed the internal rotation/adductor issue in my "bad" leg and suspected it had to be connected to my knee pain. Have yet to try it, but strongly believe that improved gate and biomechanics are going to help much more than VMO strengthening alone!!

  2. Thank you very much for this extensive and wel documented post. Do you think the same principles, findings and exercises could be applied for a cyclist?

  3. Thank you for this article. Suffering from this at the moment. In the process of returning from a femoral stress fracture in my right knee area when pain in my right knee emerged. I was 2 weeks into running (6 miles then 10 miles per week walk/jogging)when Symptoms emerged two days ago. Pretty sure its this injury as it hurts when I extend my knee or squat. Figuring it can't be another stress fracture as I am running very low volume and just returned to training. Was having some slight issues in my opposite leg the week before and I think some sort of imbalance is the root of the issue. Anyway, I tried to run today but after 5 minutess the pain was intolerable. Been doing the strength exercises and icing since then and will continue to do so. My main questions are
    1) Can I run through some pain or is 6 weeks off recommended?
    2) Does icing help or harm healing?
    3) Is cross training harmful to healing?
    4) What is the typical time frame to recovery?
    Thanks again I am a real big fan of your writing.


Leave a Comment

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