Critique and update of Janda's Prone Hip Extension - making it useful

Audience: Therapists and patients with too much time on their hands Purpose:  Provide a mild critique of the utility of the prone hip extension test



The prone hip extension test (or prone leg extension - PLE) is a very common clinical test in use for more than 20 years.  Two influential clinicians have advocated its use although for slightly different reasons.  Both Vladmir Janda and Shirley Sahrmann have described its use for decades.  This blog will focus more on the clinical rationale that Janda proposed.

Briefly, the test asks the patient to lie prone on a plinth, hands palm up at the side of the patient.  The patient is then asked to lift a leg off of the table approximately 6 inches.

It has been suggested that there exists an ideal movement pattern when lifting the leg off of the table.  It is proposed (Rehabilitation of the Spine,  by Craig Liebenson pg 214 -great book by the way) that his hip hyperextension movement is performed to analyse the hip extension that occurs during the terminal stance phase of gait.

Deviations from the ideal pattern while prone are indicators of the dysfunction that could be occurring during gait or possible other tasks (e.g. an individual might hinge at the spine during a squat rather than the hinge because of a weak/inhibited gluteus maximus aka. gluteal amnesia).

The crux of this test is that there is an ideal or optimal way to lift the leg off of the bed.  Deviations from this pattern would therefore indicate dysfunction and dysfunction would then be the cause of all manner of nasty stuff.

Janda proposed that a normal pattern of muscle activation timing occurs during the PLE.  The ideal order of muscle firing being:

1. Gluteus Maximus

2. Hamstrings

3. Contralateral erector spinae

4. Ipsilateral erector spine


Deviations from this pattern or alterations from abnormal recruitment patterns (both amplitude of muscle activation and timing) may result in the observations of the following assumed dysfunctions:

1.  Hinging does not occur at the hip joint rather it occurs in the lower spine.  This is assumed to occur because of weakness or inhibition in the Gluteus Maximus.  The patient therefore gets thigh extension from an anterior pelvic tilt rather than the hip actually extending.

2. Flexion of the knees occurs suggesting hamstring dominance

3. An observable (or palpable) delay or absence of Gluteus Maximus firing.  (aka, hey where is your butt?)

4. rotation in the lumbar spine

5.  initiation of hip movement by periscapular muscle activation


The research, tinged by opinion, on this test

Full disclosure, I use this test with my patients yet I have also published research questioning its validity.  My initial research in 2004 suggested that a delay in the firing of the gluteus maximus was common in all subjects and the range of timing across muscles is probably something that can not be seen by the human eye.  In other words, the timing is so close together between 20 and 500 milliseconds that I doubt you can see what muscle turns on first.  Especially, with fat being distributed differently over the different regions.


My research was nothing brilliant or new. Researchers before me had this found this as well (Pierce and Lee , Vogt et al 1997) and many researchers after me (Oh et al 2007, Sakamoto et al 2009)

Some interesting work has even looked at changing the timing and the amplitude of muscle activity (Chance-Larson et al 2010, Lewis et al 2009) but they still showed that the Gluteus Maximus fires last BUT it can be taught to fire sooner BUT without actually changing the order of activation.  NEAT...and I don't know if it is relevant.


So, the research suggests that the originally suggested firing order may not accurate.

The question now is "Should we then conclude that the test is invalid?".

My personal opinion is that we can't make this conclusion yet.    I even argued this in a paper I published in 2006 which is often cited to discredit the PLE test (kind of the opposite of what I was going for in my discussion but stuff happens).  In this study, my girlfriend at the time (now wife) sprained her ankle.  She had previously been a subject of mine in the 2004 study.  I therefore, had data on her in an uninjured state.  I immediately rushed her to my lab, through electrodes on her butt (something we already did for fun, just kidding), hamstrings and back and tried to see if this ankle injury influence muscle activation timing over the course of 8 weeks (click here for the study).  We found nothing, the variability was huge and it was a mess.

But I was not ready to throw out the test.  Here is why...

Most research to date only looks at EMG timing.  It does not really look at amplitude of muscle activation and certainly does not look at the force contributions from the various muscles to the movement.  Nor does any of the research look at Kinematics (e.g. what the movement looks like).  And it is with this information that I think  people use this test.  I don't think it is valid to look at a the posterior chain and conclude that the glutes are slightly delayed after the hamstrings.  But I do think you might be able to say that the Gluteus Maximus is not turning on at all. For example, the GMax might be remaining completely flaccid on one side and be hard as rock on the other.  This type of obvious difference might be something that is relevant.

What about alterations in form?

We can't discount the idea that a movement pattern that sees the patient spinal hinge or anteriorly tilt their pelvis a great deal rather than merely extending at the hip is not dysfunctional.  I am not saying that it is dysfunctional, just that no one has proven that it is not (they haven't proven that it is but I am sure a biomechanical - tissue overload argument could be made and is made).  This research needs to be done, but it is much a larger question.  This is the question of whether there is an optimal way to move and if you don't move in a certain pattern then you are dysfunctional and therefore more prone to injuries.  This a huge and difficult question.

How might the test be Useful?

I am very hesitant to look at how someone moves and say that it is a dysfunctional pattern - especially if there is no pain.  But a painful PLE might be relevant and useful clinically under certain circumstances.  We just need to Reconceptualize the test.  Here is an example:

A runner complains of low back pain while running.  When running it looks like they are very upright and leaning backwards.  They also have a long stride and therefore a low stride rate.  When they arch backwards it hurts. Passively and actively they don't have much hip extension.  When the do the prone leg extension test it hurts and it seems like they arch their back a great deal.  This is easily observed and doesn't require some magical EMG sensors to built into your eyes to see a muscle activation.  These are BIG ROCKS hitting you in the face.  This might be a case where the pattern is relevant.

Why and how is the pattern relevant clinically?

The painful pattern is relevant because:

1. The movement is kinematically and symptomatically similar to the painful functional task (running).

2. The painful PLE might give us insight into a HABIT of movement that is painful.  Its odd, but people often keep doing the same movement/posture adoption even when it hurts

3.  The lack of hip extension might decrease their movement options so they have to move in a pattern that is already sensitized...they can't/don't move differently to give themselves a break.

Clinical Decision Making & Treatment

We can try to do things to desensitize the pattern. Lets not view the PLE as faulty but just sensitive.  The movement has become coupled with pain.   Desensitizing the painful PLE and doing things to desensitize the painful running gait would go hand in hand.  What is possible as well is that you just desensitize this for the short term and perhaps they can go back to running the same way in the future.  So, how do you desensitize it?

Desensitizing the pattern

1. Running Changes: EASY-PEASY- 1. shorten their stride and maybe try leaning forward and if all else fails very temporarily have them posteriorly tilt their pelvis (they will hate this but frame it only as temporary).  Shortening the stride might be enough.  Do FORM FARTLEKS of 30 seconds of the knee running style and 3 minutes of the old.  Keep them running because this is meaningful.  As long as they don't flare up keep them active.

2. Load locally - my guess is that spine stability exercise "work" because you load locally and this has an analgesic and cognitive influence.  You confront with strength.  Load is good.

3. Sensitivity/Symptom Modification: Modify the prone leg extension in a billion ways (it doesn't matter how) and have them extend that hip with decreased pain.  Learn a new habit.  Maybe we are uncoupling pain from that movement.  Then go back and try to do the movement again and see if it hurts less.  Sometimes it carries over.  If it doesn't no worries.  Keep doing the symptom modifying exercise hourly for a few days.

4. Maybe stretch the hip flexors - can't believe I'm saying it.  It won't hurt and this is a case where the impairment may be relevant.

5. Load Globally: General hip, spine and leg exercises help sometimes.  We don't need a fancy kinematic explanation.

I think there are a few take home points

1.  The research against the test suggests that for the majority of people a delay in Gluteus Maximus is the norm.

2.  There is no research* investigating whether alterations in form (i.e. kinematics) that are different than the ideal (e.g. the spine hinges or twists) are normal and harmless variations on how to move or are indeed painless dysfunctions that are related to current or future injuries.  *When I say research, I mean a direct link.  I recognize that many esteemed biomechanists and clinicians can provide a very plausible biomechanical or clinical rationale (hence, I admit I use this test occasionally)

3. Activation patterns and even amplitude can be changed. One biomechanical model suggests that changing the activation patterns and amplitudes can influence anterior hip joint health.  See my post here on the great research of Dr. Cara Lewis.

4. "Dysfunctional" patterns may not be absolute.  They are probably more relevant in specific cases.


Some relevant points that I did not even touch on

1. You can criticize the test and say that an open chain hip extension test is not at all relevant to what occurs the hip extension during walking.  I have lab EMG research showing that the Glut Max barely fires during walking and this is supported by Dr Lieberman's (of barefoot running fame) EMG work on the role of the Gluteus Maximus during running (see the paper here). Hip extension during terminal stance may even be mostly passive in that the hip extensors work more to slow the hip down and then accelerate during terminal swing and early stance rather than provide any extensor thrust backwards. I believe this is why Dr. Sahrman advocates testing hip extension with the starting point of having the hip in 30 degrees of flexion.  However, even if the criticism is watertight it does not mean that the test is useless.  Just that the rationale behind its use may need to evolve.

2.  We should question rather the PLE is related to the movement that occurs during our functional tasks.  While it is true that a lack of hip extension during running is compensated for by an increase in anterior pelvic tilt it has also been shown that our static tests for decreased hip extension (e.g. the Thomas test) do not correlate with the amount of hip extension that occur during running.  This area would really be a great PhD and another post.  It is the idea that many of our clinical tests of function might not even be appropriate surrogates for tests of dynamic function. For example, dysfunctions during a single leg squat (e.g. knee valgus, femoral internal rotation) may not consistently correlated with the same dysfunction during running or even be related to Gluteus Medius strength (reference to follow)

3. We can't conclude that the gluteus maximus does not become weak or inhibited.  Maybe we just don't have the best test.  Other researchers (Stu McGill) have advocated testing its function with the supine bridge.


As always, more research is needed but no one wants to do it and no one will pay for it. Hence, I am no longer a researcher.


All the best,


Greg Lehman

Physiotherapist, Chiropractor and reformed biomechanist