How to Examine the Hip in the Clinical Exam: OA, FAI and the Trendelenburg Test

The hip examination comes up as a short case in almost every orthopaedic clinical exam, and candidates tend to drop marks in the same places. They describe the gait without saying which side, measure leg length before squaring the pelvis, call internal rotation "restricted" without giving a number, and present a positive FADIR as if it were the diagnosis. What examiners want is a fixed script: watch the gait first, then look, feel, measure movement with a goniometer, choose special tests that answer a specific question, and finish by naming the pattern. This post sets out that script, explains what each test can and cannot tell you, and ends with a model answer for the commonest short case, hip osteoarthritis.

What is the hip examination station testing?

Much of the diagnosis is made before you touch the joint. Organise the station around three questions. Is the pain coming from the hip at all? Is this a hip that has lost motion, or one that has too much? What would change management? Every finding you report should help answer one of them.

Set-up. Expose both legs to the underwear, with the iliac crests and lower back visible and socks off, and offer a chaperone. You need a goniometer, a tape measure and blocks of known thickness. Examine standing first, then supine on a firm couch, side-lying for the lateral structures and prone for extension and the posterior hip. In a seven-minute station, a workable budget is 60 seconds for gait and standing, 75 for supine inspection and leg length, 45 to feel, 90 to move (including the Thomas test), 100 for special tests and 30 for the neurovascular check, spine and knee. That leaves 20 seconds for the closing sentence.

The other hip is your control. Normal right and left hips have similar ranges of movement, so record every finding as a difference between the two sides rather than as an absolute number.

Look: start with the gait

Watch the patient walk away, turn and come back. Name the pattern and the side:

• Antalgic: stance on the painful side is shortened. This is the commonest finding.
• Trendelenburg: the pelvis drops on the swing side while the weak leg bears weight.
• Compensated Trendelenburg (gluteal lurch): the trunk lurches over the stance leg, which shortens the lever the abductors have to work against. The drop and the lurch are the same deficit, uncompensated and compensated.
• Short-leg: the pelvis and shoulder dip towards the short side as it takes weight.
• Stiff-hip: the limb circumducts because the hip cannot flex during swing.

A stick belongs in the hand opposite the painful hip. With the patient standing, put a thumb on each iliac crest to check for pelvic obliquity. Look for increased lumbar lordosis (the clue to a fixed flexion deformity), gluteal wasting, scars and sinuses.

On the couch, an osteoarthritic hip rests in slight flexion, adduction and external rotation because of anterior capsular and adductor contracture. An acutely effused hip lies in flexion, abduction and external rotation instead, the position of greatest capsular volume. In an older patient who cannot bear weight after a fall, fixed external rotation with shortening is a fractured neck of femur until proven otherwise.

Leg length comes only after the pelvis is square. Check that both anterior superior iliac spines (ASISs) are level before you pick up the tape.

The Galeazzi test locates the shortening. With the knees flexed to 90° and the heels together, a knee that sits lower and further back points to the femur, while one that is only lower points to the tibia. Blocks under the short leg give the functional answer, but the method is imprecise: the mean error is about half a centimetre. Treat clinical differences under 1 cm as within measurement error.

Feel: temperature, then the three territories

Run the back of your hand over each groin and trochanter. A warm, tender, irritable hip is septic until it has been aspirated. Do not stress-test it.

Feel the femoral pulse at the mid-inguinal point, halfway between the ASIS and the pubic symphysis. Then press the flat of your hand deep, just lateral to the pulse and over the femoral head, while rolling the leg with your other hand. Crepitus, a click or guarding here comes from inside the joint. Palpate the adductor longus origin at the pubic body. With the patient lying on the good side, map the greater trochanter: its tip, the posterior border and the gluteus medius insertion above it. In 65 patients with lateral hip pain, trochanteric palpation was 80% sensitive for MRI-confirmed gluteal tendinopathy, so a trochanter that is not tender makes the diagnosis unlikely.

Ask the patient to show you where it hurts. Pain from inside the joint is deep and hard to point to. The patient often cups the hand over the hip, thumb behind the trochanter and fingers in front reaching towards the groin. This is the C-sign, and it means intra-articular pain even though the hand sits on the side of the hip. One finger on the trochanter suggests the gluteal tendons. Buttock pain points to the posterior capsule, the sacroiliac joint or the lumbar spine.

Move: in degrees, on both sides

Ask for active movement first, then move the hip passively. Watch the patient's face, and use your other hand on the pelvis to catch the point where the movement becomes pelvic rather than hip movement. For flexion, that hand goes flat under the lumbar spine. For abduction and adduction, it rests on the opposite ASIS. Measure rotation in the 90/90 position, remembering that swinging the foot outwards rotates the hip internally. Record each movement as an arc compared with the other side, for example "internal rotation in flexion 10° on the left against 35° on the right".

Osteoarthritis follows the capsular pattern. Internal rotation goes first and is lost most, followed by abduction. Flexion is preserved longest, and loss of extension shows as a fixed flexion deformity. Birrell and colleagues studied new primary-care attenders aged 40 and over with hip pain. Restricted internal rotation was the plane most predictive of radiographic osteoarthritis and flexion the least. Restriction in any one plane was 86% sensitive for moderate disease and 100% for severe disease, but specificity was only 54% and 42%. Restriction in all three planes reversed the trade-off, with specificity of 93% and 88%. Sutlive's preliminary five-item prediction rule includes passive internal rotation of 25° or less. When four or more items were present, the positive likelihood ratio was 24.3 (95% CI 4.4–142.1) in 72 patients, but the rule has not been validated.

FAI loses motion against a hard stop. In FAI, internal rotation in 90° of flexion and flexion itself are reduced, both with a hard, painful end-point, and the abductors are usually normal. In Clohisy's prospective series of 52 symptomatic hips, mean flexion was 97° and mean internal rotation in flexion only 9°. The opposite phenotype has rotation arcs beyond 60°, apprehension and hypermobility. This points to adult hip dysplasia or microinstability, not impingement.

How do you perform the Trendelenburg test?

Stand or kneel behind the patient with the back exposed and a thumb on each posterior superior iliac spine or iliac crest. Ask the patient to stand on the leg being tested and bend the other knee so that the foot clears the floor. Offer your hands for balance. Watch the pelvis, not the shoulders, and hold the position for up to 30 seconds. The test is positive if the pelvis drops on the unsupported side. If the trunk lurches towards the stance leg, that is the compensated sign. A level pelvis that drifts down at 15 or 20 seconds is a positive test of fatigue, so say when the drop happened. Test both sides.

What it means. The abductors cannot produce enough force to keep the pelvis level. Possible causes include pain inhibition from arthritis, a gluteus medius or minimus tendon tear, superior gluteal nerve palsy, coxa vara or a shortened femoral neck lever, and neurological disease. The test shows which side is failing, not why. After a lateral-approach arthroplasty, think of abductor avulsion or nerve injury.

Where the test misleads. Hardcastle and Nade noted that at least four methods of performing the test had been described. They examined 50 normal subjects and 103 patients with hip or spinal disorders and defined a standard method. False positives came from pain, poor cooperation and the rib cage impinging on the iliac crest. False negatives came from the patient using trunk and leg muscles above and below the pelvis, or leaning beyond the standing hip. Do not allow the patient to lean.

How accurate it is. In Bird's study of 24 women with greater trochanteric pain syndrome, assessed with MRI, Trendelenburg's sign was the most accurate of three physical signs for a gluteus medius tear: sensitivity 72.7%, specificity 76.9% and intraobserver kappa 0.676. The same posture can be used with a different endpoint. In Grimaldi's 65 patients, reproduction of lateral hip pain within 30 seconds of single-leg stance had a specificity of 100% and a positive likelihood ratio of about 12 for MRI-confirmed gluteal tendinopathy. That moved a nominal 50% pre-test probability to 98%. The sample was small, so quote the figures with that caveat. In the Trendelenburg test you watch the pelvis; in the single-leg stance test you ask about pain.

The Thomas test: unmasking a fixed flexion deformity

With the patient supine, put your hand flat under the lumbar spine. Flex the opposite hip until the lordosis flattens onto your palm, and ask the patient to hold that knee. If the tested thigh lifts off the couch and cannot be pressed flat, there is a fixed flexion deformity. Measure the angle between the thigh and the couch, then repeat on the other side, because a bilateral deformity looks symmetrical.

The commonest false negative comes from not flexing the opposite hip far enough, so the lordosis still hides the deformity. A soft couch gives false positives. If the knee of the tested leg hangs flexed, the tight structure may be rectus femoris rather than the capsule. The evidence for the Thomas test covers reliability only. No study has measured its sensitivity or specificity, so report the deformity in degrees and do not attach an accuracy figure to it.

FADIR, FABER and the provocation tests: what does a positive mean?

FADIR. With the patient supine, flex the hip and knee to about 90°. Adduct the hip 15–20° across the midline, then rotate it internally by swinging the ankle outwards. Vary the flexion between 80° and 100° to find the position that hurts. Ask the patient to point: only the familiar deep groin pain counts. Lateral or posterior pain does not.

Reiman's meta-analysis pooled nine studies and found FADIR sensitivity of 0.94 to 0.99, but a diagnostic odds ratio of only 5.71 to 7.82. The authors concluded that the test has screening accuracy only. In Clohisy's series, 88% of symptomatic hips were painful on the test. A negative result still does not clear a convincing history. In Troelsen's 18 hips that had undergone periacetabular osteotomy for dysplasia, checked against MR arthrography, the impingement test had a sensitivity of only 59% and a negative predictive value of 13%. Most painful hips hurt in the FADIR position, including those with arthritis, dysplasia and synovitis.

Reiman's earlier meta-analysis of hip tests found only one exception to their weak diagnostic accuracy: the patellar-pubic percussion test for occult fracture, with pooled sensitivity of 95% and specificity of 86%. Every impingement manoeuvre is a screen.

FADIR does not diagnose FAI. The Warwick Agreement defines FAI syndrome as a motion-related clinical disorder of the hip with a triad of symptoms, clinical signs and imaging findings. It was agreed by 22 panel members and one patient from nine countries. A cam on a radiograph without symptoms is not a syndrome, and a positive FADIR without corroborating imaging is not a diagnosis. Our post on hip arthroscopy for FAI and labral repair covers what happens after the triad is complete.

What are the common candidate errors?

• Describing gait without a side. "Antalgic gait" is incomplete; "shortened stance phase on the right" is a finding.
• Measuring leg length before squaring the pelvis. Every number taken after that is wrong, and examiners watch your thumbs on the ASISs.
• Calling movement "restricted" without numbers. The diagnosis lies in the side-to-side difference, especially in internal rotation.
• Missing the fixed flexion deformity because the opposite hip was never flexed enough to flatten the lordosis.
• Calling a wobble a positive Trendelenburg. Give the abductors their 30 seconds, watch the pelvis and say when it dropped.
• Labelling FAI from a positive FADIR, or pushing through guarding and reporting the test as negative.
• Quoting untraceable numbers. Pairings such as "77% sensitive, 41% specific" for FABER do not trace to a verifiable study. Name the study or leave the number out.
• Calling lateral hip pain "bursitis". In Bird's MRI series, trochanteric bursal distension appeared in only two of 24 patients and never without gluteus medius pathology.
• Forgetting the neighbours. The spine refers pain to the groin and the hip refers pain to the knee.

Completing the examination. Say the closing sentence without being prompted: "To complete my examination I would examine the lumbar spine and the knee, perform a neurovascular assessment of both lower limbs, examine the other hip, and request a weight-bearing AP pelvis and a lateral radiograph of the hip." Add MRI or MR arthrography where labral pathology would change management, and ESR, CRP and aspiration where infection is possible. The full station, with every test card and its evidence, is in the comprehensive hip examination guide. For how to structure clinical stations in general, see mastering the clinical examination for orthopaedic exams.

In the exam

Examiners want a synthesis rather than a list of tests: a pattern, the measurements that support it and the investigations you would request.