Diagnosis and Treatment of Piriformis Syndrome

Piriformis syndrome has been estimated to cause 6% of sciatica in the past, but contemporary electrodiagnostic and imaging techniques find it significantly more common. Its prevalence is raised in the general U.S. population because it has been underdiagnosed and undertreated. The following outline is intended to present elements of the history and physical findings that suggest piriformis syndrome, and the electrophysiological methods for documenting its presence, injection technique and physical therapy. It includes an empirically-based interactive means to determine the probability that a given case of sciatica is due to piriformis syndrome. Finally, there is a section on outcome, alternative treatments and current clinical research protocols.


The patient with piriformis syndrome typically complains of sciatic pain, tenderness in the buttock, and more difficulty sitting than standing. The pain usually arises from overuse: health club, athletics, heavy work, or prolonged sitting, though traumatic causes are also reasonably common, including automobile accidents, falls, and penetrating wounds.

Physical findings include tenderness of the buttock region, increased pain with adduction, internal rotation and flexion of the affected thigh while the patient is in the contralateral decubitus position (after Solheim), weakened abduction of the flexed thigh (after Pace), and frequently sacroiliac joint derangement and/or iliotibial band syndrome.

EMG findings include positive FAIR-test (increased H-reflex latency with Flexion, Adduction and Internal Rotation), signs of denervation (acute or chronic) in the pattern of the posterior tibial or the peroneal nerves, as opposed to a radicular distribution, reduced CMAPs and particularly sural SNAP on the affected side, and usually normal paraspinal EMG.

While standard MRI findings may describe either an atrophic or hypertrophic piriformis muscle in an affected buttock, newer neuroimaging techniques are more diagnostic and informative for future surgical intervention. See Aaron Filler, et. al.

The FAIR-test compares the sum of M-wave plus H-reflex in the anatomical position with the sum of M-wave plus H-reflex in the Flexed, Adducted, Internally Rotated position.

The patient should be lying in contralateral decubitus, face toward the examiner. Though it is somewhat painful, the physician or assistant should flex, adduct and internally rotate the affected leg by grasping the ankle with one hand, and holding back the hip (to prevent the patient turning prone) with the other. One should be sure that the patient's navel is facing above the horizontal, to avoid the patient inadvertently lying prone, and ending up in abduction instead of adduction.

The physician should use caution not to injure the patient by overvigorous adduction, internal rotation or flexion, in that order. This is a provocative test, and indiscretion can bring about some harm. Both posterior tibial and peroneal nerves are tested, making four M-waves plus H-reflexes for each limb tested. For the posterior tibial nerve, the ground is placed just distal to the popliteal fossa, with active and reference electrodes placed over the soleus muscle in proximal to distal array, respectively. For the peroneal nerve, active and reference electrodes are placed in the same alignment four fingerbreadths distal to the fibular head. H-reflexes are studied in the anatomical position, and in flexion, adduction and internal rotation without moving the electrodes. Each limb serves as its own control. Therefore, if an alternative placement is used, there is a high likelihood of a meaningful result, provided the same montage is maintained for both the anatomical and FAIR H-reflex. In each case the anatomical (M + H) is subtracted from the FAIR (M + H).

H-reflex and M-wave are added because it is likely that the FAIR positioning moves the sciatic nerve's fibres beneath the skin. Still, the sum of the reflex arc, from the soleus muscle up to the level of synapse at T-11 or T-12 and back down to the soleus muscle is a set distance. If the nerve moves proximally, then the sensory limb of the reflex becomes longer, and the M-wave shorter by the same amount. As illustrated below, if the point of stimulation (of the nerve, not the skin) moves from S1 to S2, then the motor fibres' course down to the soleus (M-wave) is lengthened by exactly the same distance that the sensory limb of the H-reflex is shortened, (S1 - S2.)

These factors cancel each other, and the movement of the nerve poses no difficulty unless the motor and sensory nerve conduction velocities differ. Usually this is negligible. Assuming the nerve moves 5 cm. and that the motor nerve conduction velocity is 50 m/s while the sensory nerve conduction velocity is 40 m/s. Distance equals rate X time.

D = RT

D/R = T

0.05 m (divided by) 50m/s = .001s

0.05 m (divided by) 40m/s = .00125s

Difference of H-reflex latency = -.00025s

This is less than 0.5 standard deviation beyond the mean found in studies of normals.

In testing 88 normals, the mean standard deviation was 0.62 ms. In testing 1014 suspected legs, the mean maximum prolongation for all legs judged positive for piriformis syndrome on clinical criteria was 3.275 ms, or 5.28 standard deviations.

The prolongation of the H-reflex is reversible, doing no permanent damage to the nerve fibres, but producing a temporary neuropraxia that confirms significant pressure generated by the piriformis muscle on one or both portions of the sciatic nerve.

The delay in H-reflex is not subtle: any nerve conduction slowing brought about by stretching the piriformis muscle over the sciatic nerve and/or its divisions has an effect on both the ascending and descending limbs of the H-reflex arc, doubling any discrepancy that the stretch produces. After flexing the hip to 90 degrees, then as one increases adduction and internal rotation (the angle alpha in the drawing above), the H-reflex response is prolonged proportionally. As amount of adduction and internal rotation (the independent variable) increases, the prolongation of the H-reflex (the dependent variable) increases. There are cases in which flexion, adduction and internal rotation is so painful to the patient that one must look for changes in the dependent variable according to smaller increments and decrements of the independent variable in order to establish that piriformis syndrome is present.

Using the FAIR-test to detect piriformis syndrome has the following sensitivity and selectivity.

Combined Posterior Tibial and Peroneal FAIR test > 3 Standard Deviations (1.86ms)

Piriformis Syndrome*

  Yes No Total
Positive 512 69 581
Negative 26 129 155
Total 736**

Sensitivity = 0.881; Specificity = 0.832
*Clinically Determined
**240 additional cases were unjudged.

Combined Posterior Tibial and Peroneal FAIR test > 2 Standard Deviations (1.24ms)

Piriformis Syndrome*

  Yes No Total
Positive 568 19 587
Negative 49 107 156
Total     743**

Sensitivity = 0.986, Specificity = 0.686
*By Clinical Criteria
**237 additional cases were unjudged.

Given this means of diagnosis, (see Fishman and Zybert, 1992) and the neuroimaging techniques recently developed at UCLA, (see Filler, et. al. 2005) piriformis syndrome is no longer a diagnosis of exclusion. Authors of recent articles using these active diagnostic techniques, have found that piriformis syndrome is a common cause of sciatica, possibly as common or more common than the herniated disc. It is important, therefore, to recognize the clinical signs of piriformis syndrome, and obtain one of the reliable tests for it when the signs are present.

However, if there are signs of denervation of the hamstrings, glutei, tensa fascie lata, and/or the piriformis muscle itself, then, since the nerve supply for these muscles does not travel under or through the piriformis muscle with the sciatic nerve, there are likely to be other causes of these findings, and possibly other causes of the patient's sciatica which may, naturally, coexist with piriformis syndrome in the same patient.

A discussion in the journal Muscle and Nerve included a criticism of our work by Dr. John D. Stewart of McGill University. He cast doubt on our findings since we did not isolate patients that had MRIs without evidence of spinal pathology, nor patients with normal electrophysiological examinations of the paraspinal muscles. (1) We feel that Dr. Stewart's criticism was just, since it would help to distinguish patients in whom there was minimal suspicion of a spinal cause for their sciatica.

Since that time we have looked over the records of those same patients. Here are the statistics:

Piriformis Syndrome / MRI (n=449)

MRI (n) Improved 60% Average Improvement Years of Pain
Positive 129 74.7 62.3 6.9
Negative 320 74 61.5 5.8
Normal MRI + EMG 179 76 61.8 5.9

It appears that patients did very much the same whether their MRIs showed spinal pathology or not. This may be due to the fact that an individual may have more than one diagnosis. They may have piriformis syndrome and spinal pathology. The question is which one is the major cause of their pain. Since in our medical environment the herniated disc is generally recognized and treated first, the majority of the patients that have come to us with these two diagnoses are the ones in which the piriformis syndrome was the major cause of their sciatica.

In researching Dr. Stewart's question, the FAIR-test appears able to determine if the piriformis impingement is an operant cause of sciatica in these cases. The patients with spinal pathology actually did (insignificantly) better than the others, but what may be significant is that they suffered more than one year longer. This is possibly due to the greater period of time clinicians spent treating the lower back in these patient, exactly because their MRIs were positive, but, as it turned out, irrelevant. In part, then, their suffering is a consequence of considering piriformis syndrome a diagnosis of exclusion.


The prolongation of the H-reflex is reversible, doing no permanent damage to the nerve fibres, but producing a temporary neuropraxia that confirms significant pressure generated by the piriformis muscle on one or both branches of the sciatic nerve. There can be denervation, but it is usually mild, preserving adequate nerve fibre supply for reflexes, normal activity and the exuberant outbursts of energy that characterize most people's lives.

Improvement of Piriformis Patients with Positive FAIR-tests

NERVE Improved 50% with PT (n) Improved 50% with surgery (n) Clinicians Seen (n) Years with Pain (n) Affected Sural SNAP Reduced 50% (n)
P Tib. 71.2% (345) 70%  (43 ) 6.5 (345) 6.6 (345) About 30%
Peron. 71.0% (320) 70% (23) 6.6 (320) 5.8 (320) About 30%
Total 71.1% (665) 70% (66) 6.55 (665) 6.206 (665) About 30%

If overuse and trauma are the chief causes of piriformis syndrome, spasm and scarring are the main mechanisms by which pain is produced. Leg-length discrepancy, contracted iliopsoas muscle, prolonged sitting, poor seating and over-enthusiastic exercise can be countered with heel-lift, iliopsoas stretching and gait training, breaks in sitting, better chairs and vehicular seats, and cutting down on exercise programs, respectively. For simple spasm, the following program of physical therapy has brought 79% of the 1014 legs treated a minimum of 50% improvement or more.(See Fishman, et. al., 2001).

Improvement of Piriformis, Non-piriformis and Surgical Patients

  Weight (n) Age (n) Average Months Followup (n) Average Improvement (n)
Piriformis Present 150.99 (671) 54 (671) 10.2 (671) .71 (671)
Piriformis Absent 149.57 (332) 57 (331) 9.9 (339) .55 (339)
Surgery 149.12 (42) 49 (42) 16 (42) .66 (42)

These exercises require a physical therapist or physician, and should not be attempted otherwise. They are dangerous and will harm people who have advanced osteoarthritis of the hip, sacroiliac joint derangement, lumbar hypermobility, hip or knee replacements, and a number of other conditions. For patients with piriformis syndrome and these other conditions, physicians must devise and supervise the safe practice of different exercises, often modifying what follows.

Physical Therapy for Piriformis Syndrome

1. Place patient in contralateral decubitus and Flexed Adducted Internally Rotated (FAIR) position:**
2. Ultrasound to piriformis muscle, while leg is placed in flexion, adduction and internal rotation (FAIR): 2.25-2.5 watts/cm2 for 10-14 minutes. Beware of patients with any hypoaesthesia or anaesthesia due to neurological or surgical causes in the dorsal lumbocarsal region. Beware of cavitization in post-laminectomy patients.**
3. Wipe off Ultrasound gel.
4. Hot packs or cold spray at the same location x 10 minutes.
5. Stretch the piriformis muscle for 10-14 minutes by applying manual pressure to the muscle's inferior border, being careful not to press downward, rather directing pressure tangentially, toward the ipsilateral shoulder.****
6. Myofascial release at lumbosacral paraspinal muscles.
7. McKenzie exercises.
8. Use lumbosacral corset when patient in the FAIR position.*****
Duration: 2-3 times weekly for one-two months.

*Patients usually require 2-3 months of biweekly therapy for 60-70% improvement.
**Because it is painful, patients often subtly shift to prone. This must be avoided because it works to place the affected leg in abduction, not adduction, greatly reducing the stretch placed on the piriformis muscle.
***Cavitation is unreported in more than 20,000 treatments
****Unless explicitly stated, therapists may tend to knead or massage the muscle, which is useless or worse. The muscle must be stretched perpendicular to its fibres, in a plane that is tangent to the buttock at the point of intersection of the piriformis muscle and the sciatic nerve, but approximately 3-4 of an inch deep to the buttock, (i.e., just below the gluteus maximus).
*****This is useful in patients that have had laminectomies, or in others to prevent lumbar hypermobility.

Calculate the Likelihood of Piriformis Syndrome

This brief interactive table is intended to estimate the likelihood that an individual' will recover in treating them for piriformis syndrome via our protocol. It is likely that this is therefore the likelihood that they have piriformis syndrome, but since we have come to define piriformis syndrome using these very criteria, that would be circular. What is not a matter of definition is that people's answers to these questions turns out to be a strong b basis for predicting whether treating them for piriformis syndrome will help them significantly. This table does not diagnose anyone, but gives the added probability that each symptom provides.

The table is based on a careful study of more than 1000 cases of sciatica, most of which had no evidence of spinal injury, or had already had spinal surgery which did not benefit them.

Once again, it is intended as a means for a clinician or patient to determine the value of pursuing the diagnosis of piriformis syndrome further. It is not a means to diagnose in itself. One needs to see a person "in the flesh," do a physical exam, review a good history, and perform tests to arrive at a diagnosis one can trust enough to act upon.

Now what percentage of the people with sciatica have piriformis syndrome in the first place? No one knows. We certainly don't, since our population of patients is significantly skewed toward those people that have it! Twenty years ago doctors at the Mayo Clinic estimated piriformis syndrome to account for 6% of sciatica, (Hallin, 1983) but the incidence of the diagnosis in the county of the Mayo Clinic has gone up five-fold since then. (Fishman,2003) Comparing the number of MRIs done annually the United States with the number of lumbar surgeries, a neurosurgeon at Cedars-Sinai Hospital in Beverly Hills, California estimated that piriformis syndrome may be as common a cause of sciatica as herniated disc. (Filler, 2005) All we can say is that based on current data, an individual's chance of having piriformis syndrome is greater or lesser depending upon the factors we have isolated so far, and offer tests to confirm or exclude it.

Still, people with the lowest probablilty may, of course, have piriformis syndrome, just as horses with long odds sometimes win the race! People that currently have no spots may develop chicken pox! In the case of piriformis syndrome, as in just about everything in medicine, the history and physical, performed by a conscientious and informed clinician are the golden standard of care, and the royal road to diagnosis.

Just answer the questions below: the increase or decrease in odds represents the increased or decreased probablility that treating the individual for piriformis syndrome will be helpful in relieving their pain at least 50%.*

Years of pain:
Positive Fair Test?
Tender Piriformis / sciatic nerve junction?
Solheim Sign?
Pain is worse when sitting (vs. standing)?
Female gender / high weight?
Illiotibial Band Syndrome?
Non-sciatic Denervantion?
Peroneal Polyphasics?
Non-sciatic Polyphasics?
Relative Risk:

* Based on thorough history and physical and approximately 10 month follow-up of more than 1000 sciatica-sufferers.

Definitions and explanations:

Years of pain: Please use digits and decimals.

Positive FAIR-test: Nerve conduction test performed by physiatrists and neurologists in which the H-reflex is prolonged at least 1.86 ms when the leg with sciatica is stretched in flexion, adduction and internal rotation.

Overuse: More than 3 miles/day running, high-performance athletics such as tennis, fencing, pole-vaulting, or sitting for hours at a time are examples.

Tender piriformis/sciatic junction: Pain with digital or other pressure just below the mid-buttock. Occasionally this may induce or exaggerate the sciatica.

Solheim's sign: Pain with adduction of the flexed thigh, which stretches the piriformis muscle against the sciatic nerve. We named after the Norwegian neurosurgeon that described it.

Pain is worse when sitting than standing: This may take 15 minutes to 1 hour to appear.

Female gender/high weight: Obese females, with body mass indices above 30, had lower recovery rates than others. This is the only negative factor on the list.

Iliotibial band syndrome: Pain and tenderness along the lateral edge of the affected thigh. A physician, physical therapist or chiropractor might help confirm this diagnosis.

Injection: 20 mg triamcinolone acetonide with 1.5 cc of 2%Lidocaine was used in our studies. We now use EMG guidance, and may use .5% Marcaine or up to 300 units of Botulinum neurotoxin Type A, or 12,500 units of Botulinum neurotoxin Type B, all of which improve the positive influence of this factor beyond what is represented here.

Non-sciatic denervation: EMG findings of nerve loss in muscles that are not innervated by the sciatic nerve improved people's responsiveness for debatable reasons.

Peroneal polyphasics: EMG findings of chronic denervation with recovery may reflect a long-standing condition. Because of its underdiagnosis, piriformis may be more likely to be untreated for longer than other causes of sciatica with denervation.

Non-sciatic polyphasics: EMG findings of chronic nerve loss in muscles that are not innervated by the sciatic nerve improved people's responsiveness. The explanation for this is not clear.

Yoga Therapy for Piriformis Syndrome

1. Janusirsasana
2. Matsyendrasana
3. Parivrtta Trikonasana
4. Parivrtta Parsvakonasana
5. Garudasana
6. Gomukhasana
7. Kapotasana

There are many substitutes and fine points to thelong term yoga treatment of piriformis syndrome. Those interested find further information in Yoga for Back Pain. Fishman LM, Ardman CA. W.W. Norton 2006.

Injection of 20 mg. Triamcinolone Acetonide with 1.5cc of 2% lidocaine or 0.5% marcaine has been successful in relieving pain in the short run and enhancing patients' recovery rates and overall improvement. BOTOX injections do not take effect for five to ten days, but have been shown to enhance overall improvement most reliably in the 2-6 month range. A detailed description of these injections will appear here shortly. Physicians can e-mail loren@sciatica.org for further information.

Characteristics Favoring a Successful Outcome

  p-value odds Ratio 95% C.I.*
Positive Fair Test .001 2.225 1.42-3.60
Overuse .001 2.05 1.35-3.12
Tender Piriformis/sciatc .003 1.97 1.20-2.82
Solhiem Sign .005 1.84 2.25-4.90
Sitting worse than standing .007 1.77 1.17-2.68
Male Gender .017 1.67 1.10-2.50
Illiotibial Band Syndrome .028 1.67 1.12-7.38
Injection .030 1.55 1.05-2.31

* Confidence interval

Values are from logistic model (univariate analyses) and stepwise regression: receiver operating characteristic (ROC) = 6.92. Stepwise logistic regression model CI was larger but comparable. n = 651 patients, 514 successful outcomes ( 50% improvement or more.)