You are here: Home » The Playbook » Please, Just STOP LIMPING!

Please, Just STOP LIMPING!

She won’t walk“, or “He just looks like he’s limping“.

So many things can be going on — how do we tackle this chief complaint?

You’re dreading a big work-up.  You almost want to tell the kid: please, STOP LIMPING…


S – Septic Arthritis 

This is the most urgent consideration in our differential diagnosis. The hip is the most common joint affected, followed by the knee.  Lab work can be helpful, as well as adjunct use of US of the hip to look for an effusion,  but sometimes, regardless of the results, the joint just has to be tapped to know for sure.

T – Toddler’s fracture

This is usually a torque injury when the wobbling toddler pivots quickly or trips and falls.  Toddler’s fractures occur in children 1 to 3 years of age, most commonly in the distal 1/3 of the tibia.  Sometimes a long-leg cast is needed, but currently there is a new trend in foregoing casting in mild cases.

O – Osteomyelitis

Bacteremia – from any source – can seed into any bone.  It’s not very common, but it happens: approximately 2% of children who present to an ED with limp will have osteomyelitis.  Plain films, ESR, and CRP are a fair screen to start.  For more than the casual concern, MRI is the best modality to evaluate, followed by radionuclide scintigraphy.  Although not the first choice modality, CT can show periosteal changes, such as inflammatory new bone formation or periosteal purulence.

P – Perthes disease

This is the famous Legg-Calvé-Perthes idiopathic avascular necrosis of the hip, usually affecting children from 3 to 12 years. They present with a slow onset pain and with an antalgic gait.  Patients will have trouble with internal rotation and abduction of the hip.  Radiographs may be initially normal.  MRI can show the culprit: decreased perfusion to the femoral head and subsequent necrosis.

L – Limb-Length Discrepancy

Parents may notice that he seems “wobblier” than he should be.  It may be that we are just now appreciating a congenital anomaly.  Get out the paper tape, and measure from the anterior superior iliac spine to the medial malleolus and compare both sides.   Children with limb-length discrepancy need a non-urgent referral to pediatric orthopedics to look for congenital dysplasia of the hip, or other growth abnormalities.  Some are treated with orthotics; surgical options vary.  Epiphysiodesis destroys the growth plate on the unaffected side, which evens out limb growth (if done, in the preteen years).  Other options are limb-lengthening or limb-shortening procedures.

I – Inflammatory

Transient Synovitis.  This is what we want them to have right?  The typical age is between 3 and 6 years, sometimes just after a URI.  To be comfortable with this diagnosis, we should have considered all of the dangerous diagnoses, the child should be well, afebrile, in minimal discomfort, and he should respond almost completely to an NSAID.  He’s the one running up and down the department after treatment – or just from sheer boredom after observation.

M – Malignancy

Primary bone tumors such as Ewing’s sarcoma or osteogenic sarcoma typically affect older children.  Limping, however, may be a presenting symptom of leukemia.  If you have any suspicion of the general wellness of the child, get a screening CBC, and perhaps a peripheral blood smear.  Whatever you do, make sure you get close follow up for these kids that are on your malignancy radar — the blast crisis may not have occurred yet – but it can happen hours to days later.

Plain films are insensitive for leukemic involvement of bone but they may show diffuse osteopenia, or metaphyseal bands – symmetrical high-uptake markings around the joint.  They look like stacks of paper within normal bone – you can see them also in anemia, lead poisoning, and other causes.  Also look for periosteal new bone formation, sclerosis, or lysis.

P – Pyomyositis

This usually presents with vague irritability, pain, and fever, and sometimes after subacute minor trauma.  These children do not appear well.

Also think about simple run-of-the-mill myositis, usually from a viral cause, such as influenza.  Typically the calves are affected and are always tender.  Hydration and supportive therapy are indicated for viral causes.

For bacterial focal pyomyositis, give empiric antibiotics, admit them for a major inpatient workup, and think about early surgical consultation if you think you need sepsis source control.

I – Iliopsoas Abscess

Children most often will develop a primary abscess from bacteremia from an unresolved infection.   Adults more commonly form secondary abscesses from Crohn’s disease, post-op complications, a vertebral infection, or even a bad chronic urinary tract infection.  Lest you think this is a dramatic presentation, think again: iliopsoas abscesses present with protracted vague symptoms of back, flank, abdominal, or hip pain, and sometimes with fever.  The median time from symptoms to diagnosis in children is a whopping 20+ days, according to one study.  If iliopsoas abscess is starting to get your attention, get the CT or MRI.

N – Neurologic

Not to be alarmist, but children do have strokes; unlike adults, half are hemorrhagic, and half are thromboembolic.  Typically they’ll have some underlying pathology that will alert you, such as a cardiac lesion, sickle cell disease, or a complicated infectious or metabolic history.  The good news is that it won’t just be a limp – you’ll have some other neuro sign or symptom to go after.

Guillain Barré is another thing to consider – early lower extremity weakness may present as a limp or refusal to walk.  Maybe it’s not the hip that should be tapped, but the spinal canal.

Think also about muscular dystrophy or peripheral neuropathy and its possible underlying etiology.

G – Gastrointestinal and Genitourinary

What else could be going on?  Appendicitis may be faking us out here.  Perhaps there is a hernia, or testicular or ovarian torsion, all of which can present as lateralizing pain and not wanting to walk.  Think outside the box.


Phases of Gait

The gait cycle has three phases: contact, stance, and propulsion. Contact is the time from heel strike to just when the foot is flat.  Stance is from the foot being flat to lifting the heel from the ground.  The stance phase is when you bear most of your weight.  The propulsion phase is when your weight transfers to your toes, and you push off.

Gait Cycle_Horeczko_Tim


Abnormal Gaits

trendelenburg-gait_Horeczko_TimAntalgic Gait — “hobbling” gait; normal contact phase, but stance phase is abbreviated; propulsion is normal.  The patient is trying to limit the time spent bearing weight on that side.

Trendelenburg Gait — the affected side’s hip abductor muscles are too weak or painful to stabilize the pelvis; the unaffected side dips to the floor. May be due to superior gluteal neuropathy, or a biomechanical problem, such as in avascular necrosis, congenital dysplasia of the hip, or slipped capital femoral epiphysis.

Circumduction Gait — the patient swings his foot laterally (due to a foot or ankle pathology), or to avoid tripping in limb-length discrepancy.

Stiff-leggged Gait — the patient walks with knees locked, in an attempt to avoid using the gastrocnemius muscles; concerning for myositis.

Equinus Gait — toe-walking, as seen in myositis, also to avoid exacerbating pain from the calves.


Hip Internal Rotation_Horeczko TimLag of Internal Rotation of the Hip

Look for symmetry of internal rotation, or lateralizing pain or “guarding” with range of motion.

Keep the pelvis flush to the bed, and simultaneously rotate the lower extremity laterally, which will cause internal rotation of the hip.

crescent sign_Horeczko_Tim

Avascular necrosis will not allow full internal rotation, since the joint space is narrowed with this maneuver, causing impingement of the sensitive necrotic head of the femur.

Note any lateralized pain, asymmetry, and the angle of internal rotation achieved.

Here, an example of the crescent sign, seen in advanced AVN (Fica stage III)


Kocher Criteria

In their original paper in 1999, Dr Kocher et al. performed a retrospective analysis of children who were being evaluated for a septic joint versus transient synovitis over a 15 year period, in a major referral center.  They came up with four independent predictors of a septic joint, and calculated the probability of septic arthritis based on the number of features present.  In 2004 the same group validated their prediction tool, with a slightly decreased sensitivity and specificity in the validation population.

In short, the Kocher criteria are not perfect, but it’s the best evidence we have at the moment.

The four predictors are:

Inability to walk

Fever of 38.5 °C or greater

ESR > 40 mm/h

WBC > 12,000/mm3   (12 x 109/L)


Bonus mnemonic: Walk FEW: Inability to Walk | Fever | ESR | WBC

The probability of septic arthritis increases with increasing predictor. In this prediction model, each predictor has the same weight.

Probability of Septic Arthritis (Kocher et al. 1999)

0 Predictor – <0.2 %

1 Predictor – 3%

2 Predictors – 40%

3 Predictors – 93.1%

4 Predictors – 99.6%

Now, remember, this is to be used in children in whom you already have some suspicion of a septic joint.  So, 0 predictors, generally you’re alright.  1 predictor, you may start to worry.  Once you have 2 predictors, your chances jump for 3% to 40%.  You really have to go looking.

The Kocher caveat is that there is no single test or single decision rule that will stop you from investigating if you are concerned enough.  Don’t have too much faith in this imperfect decision tool – we use it because we need somewhere to start.  Treat and push for the aspiration of the hip if you are left in doubt.  Septic arthritis can be devastating if not identified early.



  • Fever and limp? – do not pass go – especially with 2 or more Kocher criteria – or if there is any doubt, tap that joint.
  • Refusal to walk after adequate analgesia? Admit for observation, MRI, further workup, blood cultures, and much more.
  • Remember some developmental concerns to help you to decide whether to continue the investigation urgently as an inpatient, or non-urgently as an outpatient.
  • After all of that, try to get them them to STOP LIMPING!



Jain S et al. A stiff-legged gait: benign acute childhood myositis. CMAJ. 2009 Nov 10; 181(10): 711–713.

Julien L et al. Pyogenic Sacroiliitis in a 13-Month-Old Child: A Case Report and Literature. Medicine (Baltimore). 2015 Oct; 94(42): e1581.

Kocher MS et al. Differentiating between septic arthritis and transient synovitis of the hip in children: an evidence-based clinical prediction algorithm. J Bone Joint Surg Am. 1999 Dec;81(12):1662-70.

Kocher MS et al. Validation of a clinical prediction rule for the differentiation between septic arthritis and transient synovitis of the hip in children. J Bone Joint Surg Am. 2004 Aug;86-A(8):1629-35.

Lindsay D, D’Souza S. A limping child. BMJ. 2016 Feb 9;352:i476.

Navarro López V et al. Microbiology and outcome of iliopsoas abscess in 124 patients. Medicine (Baltimore). 2009;88(2):120.

Rennie LM et al. Benign acute childhood myositis in an Accident and Emergency setting. Emerg Med J 2005;22:686-688

Sapru K, Cooper JG. Management of the Toddler’s fracture with and without initial radiological evidence. Eur J Emerg Med. 2014 Dec;21(6):451-4.

Schuh AM, Whitlock KB, Klein EJ. Management of Toddler’s Fractures in the Pediatric Emergency Department. Pediatr Emerg Care. 2015 Jun 17. [Epub ahead of print]

Stoica Z et al. Imaging of Avascular Necrosis of Femoral Head: Familiar Methods and Newer Trends. Curr Health Sci J. 2009 Jan-Mar; 35(1): 23–28.

Verma S. Pyomyositis in Children. Curr Infect Dis Rep. 2016 Mar;18(4):12.


This post and podcast are dedicated to the estimable yet graciously humble Andrew Tagg BSC(Hons), MBBS, MRCSEd, ACEM for his dedication to #FOAMed, Emergency Medicine, Pediatric Emergency Medicine, and all things caffeinated.  Thank you for your dedication, generosity, and your example.

Don’t Forget the Conference!   #DFTB17   #DFTB




Powered by #FOAMed — Tim Horeczko, MD, MSCR, FACEP, FAAP

Pediatric; Emergency Medicine; Pediatric Emergency Medicine; Podcast; Pediatric Podcast; Emergency Medicine Podcast; Horeczko; Harbor-UCLA; Presentation Skills; #FOAMed #FOAMped #MedEd

  • Thanks as alway Tim for your comprehensive review of common pitfalls and problems.

    If I may I would like to add two additional comments

    1) We are slowly losing the art of focused clinical examination as a investigative tool in itself. Every health care professional who deals with children should really be able to perform an:

    Axial Load: Hold the bottom of the foot and the top of a flexed knee. Apply pressure through the tibia. Don’t look anywhere BUT the child’s face. Even in the already crying child you can detect increased discomfort which demonstrates localised pathology i.e more likely to be lower limb than hip issue.

    Even when the axial load test is negative. its important the remember

    Tibial Torsion: In the same position as axial load provide a twisted force on the tibia. This is a very subtle test for a fracture of the tibia which may only become apparent with distress on this movement.

    2) Kocher’s is the classic exam question, but as you rightly point out Tim, is only relevant if you have a predisposing gut feeling that infection may be an underlying cause. More importantly it depends on blood tests actually being performed. What we lack in the paediatric practice is a clear guide for when bloods are needed as well as what you should do if those bloods are abnormal (or normal!). Presently, and please do correct me if I am wrong, we don’t have a sense of what the current prevalence of septic arthritis is in infants, especially in an enhanced post-immunisation era. The pre-test possibility a key component in decision making.

    • Tim Horeczko


      Thank you very much for your always excellent comments and advice.

      Yes — a careful exam can help us to avoid non-productive testing — or more importantly, focus our efforts into hypothesis-driven testing.

      Thank you very much for a perfect review of the physical exam for toddler’s fracture; I teach these maneuvers routinely.

      So glad you brought this up — it’s hard to pick-and-choose what makes it into a (hopefully close to) 30-min program with a long differential. You honor The Playbook and benefit our community with your input and fantastic points.

      Re: Kocher’s criteria, etc. You are feeling our pain, my friend. I think of it as a cognitive paradigm, and the best evidence we have, but you’re right — you have to consider the diagnosis first. The default for me is tap the joint, but I do not want to be dogmatic about this — there are many shades of grey with presentation of limp, and sometimes an NSAID for analgesia and a brief observation paint a completely different picture.

      As far as neonates and infants go, frankly they worry me when we talk about missed joint infections. Sometimes they are identified because they’re bacteremic/febrile/ill and they receive antibiotics and they either miraculously recover without our knowledge of the source or they continue to display fever, prompting further workup (usually with suspicion for osteomyelitis). What troubles me most is the infant, say, with sacroiliitis:

      These are the best references I have for septic arthritis in the infant:

      More recently:

      As you can see, both are observational/descriptive and retrospective. We have no idea of prevalence in general, and definitely not in the post-pneumococcal era. A brave new (but really the same) world.

      I guess the moral of the story here is: “what would Damian do?” (WWDD?). A good history, a through exam, (maybe a test or two!), and when in doubt: observation!

      Thanks, my friend — I really appreciate your taking the time to share with all of us.

      • Thank you Tim. I am sure I no better an arbiter of decision making than you are!

        I feel we should debate the brave (but really the same) world at some point. The challenge over time is that the risk of serious infection is probably lower but the fear of missing something is probably much higher. Can we really be sure that we aren’t being over cautious in children, and potentially inflicting unnecessary investigations, to find diseases of increasingly decreasing prevalence…? As always look forward to your excellent topic reviews… viva la #FOAMed

        • Tim Horeczko


          Thank you for the reality check.

          Yes, it’s a tough balance — suspicion versus obligation — which will differ depending on the consequences of missing a particular diagnosis. The specter of overdiagnosis and overtreatment is real.

          I also have to confess that my population has relatively poor access to care and is underserved. If we don’t figure it out in the ED right-here-and-now, it may not happen…

          I teach the conservative route, to offer a starting paradigm; after all, ” a plan is something to deviate from.”

          Another strength of #FOAMed: perspective. Thanks very much, Damian, for enriching the conversation!

          • its a good point which is often over looked in many conversations/debates/blogs and academic papers. Is the population you see the same as mine?

            You are spot on about being agile to patient groups who may only get one shot at getting effective treatment. You could argue regardless of what Bayesian modelling may tell you – the risks of no, or poor safety netting, may alter post test probabilities enough for you to alter your risk profiles. Will ponder this… Thanks for the food for thought 🙂

  • Pingback: – Emergency Medicine EducationPEM Playbook - Please Stop Limping - - Emergency Medicine Education()