Skip to content

The child with a limp


This post is based on the talk Andrew Tagg was invited to give at BadEMFest20 in Capetown. Unfortunately, world events made that a little challenging.

Humans evolved from Homo erectus to Homo sapiens around 500,000 years ago. Fossils dug up by the Leakeys in Olduvai Gorge, Tanzania, revealed how we moved from shuffling knucklewalkers to the upright hominids we recognise today. Walking on two legs allowed us to use our hands – to communicate, use tools, and keep ourselves safe.

How do we walk?

The normal human gait is comprised of three distinct phases:

  • The contact phase – from heel strike to flat foot
  • The stance phase – from flat foot to heel off the ground
  • The swing phase – the propulsion phase as weight transfers from the toes

The swing phase normally comprises 40% of the gait cycle.

Like Sherlock Holmes or his progenitor, Joseph Bell, if you watch your patient walk, you can get a lot of clues as to the potential underlying pathology before you even exchange a word with them or their parents.

Antalgic gait

This hobbling gait has a normal contact phase and normal propulsion with a decreased stance phase. Essentially, the patient is trying to reduce the time spent bearing weight through the hip joint.

Trendelenburg gait

In this gait disturbance, the hip muscles are too weak, so the unaffected side drops towards the floor. It suggests a biomechanical issue such as ass avascular necrosis or a slipped capital femoral epiphysis.

Circumduction gait

The patient swings their leg out and around to clear the floor. This is often due to a leg-length discrepancy in the delayed diagnosis of developmental dysplasia of the hip (DDH).

Equinus gait

Toe walking, though normal up to the age of three, may suggest that the child is unwilling or uncomfortable engaging their gastrocnemius muscles in walking. This can occur in many conditions.

The rate of presentation to the emergency department with an atraumatic limp is approximately 1.4 per 1000 in children under the age of 14. It occurs more often in boys than in girls (1.7:1), and the median age is 4.4 years. When we consider the potential causes, the list is massive. In this post, we are going to concentrate on both the more common conditions and those diagnoses that we really shouldn’t miss. Whilst any part of the limb might be the cause (from the toes to the hip), this post will focus on causes centred around the hip joint. They account for about 60% of cases of limp. That doesn’t mean you shouldn’t look in their shoes for pebbles or look at their toes, though!

Let’s break it down to some age-specific diagnoses and those that can occur at any age.

Late presentation of developmental dysplasia of the hip (DDH)

Routine screening with Ortolani’s and Barlow’s tests may have missed a case of hip dysplasia. The limb resultant limb shortening can be hard to detect if it is bilateral, but this is one time to get out the tape measure and look for a leg length discrepancy as you measure from anterior superior iliac spine (ASIS) to medial malleolus.

If you don’t have a tape measure to hand, look for a positive Galeazzi sign, suggesting a shorter hip segment on the affected side.

Original by Walt Shumway

Transient synovitis

This diagnosis accounts for most cases of atraumatic limp and may follow an upper respiratory tract infection, though the evidence for causation is poor. It is most common in young boys, aged 4 to 8, and is self-limiting in nature. It’s a diagnosis of exclusion rather than anything else made easier by the well-appearing, afebrile child. It should respond to simple NSAIDs, though it is recurrent in 20% of cases.

Dr. Sathya Subramaniam, Pediatric EM Fellow – Kings County/SUNY Downstate and The POCUS Atlas

According to Viera and Levy, bedside sonography has a sensitivity of 90% and specificity of 100% to detect an effusion. Unfortunately, it doesn’t help with determining the cause.

Occult Trauma

It’s beyond the scope of this post to talk about NAI and long bone injuries in children (especially as Nikki Abela did such a great job at DFTB18), so instead, it’s worth thinking about toddlers’ fractures. These occur due to torsional force on the tibia, accompanied by a fall, and may only be picked up as tenderness over the distal third of the tibia. That first set of X-rays performed in the emergency room may not show any obvious pathology, with signs only becoming obvious after a couple of weeks. This is one of those occasions where ultrasound may be much more helpful, though if clinical suspicion remains, it should be treated with immobilization and close follow-up. Repeat imaging may then show the beginnings of some callus formation, or a radionuclide bone scan may be needed. If they are not toddling and waddling, it is not going to be a toddler’s fracture.

Case courtesy of Dr. Jeremy Jones, From the case rID: 9317


Although unlikely to present as an isolated limp, cerebrovascular events may present as hemiparesis; It’s beyond the scope of this post to go into them in more detail.

Perthes disease

This is idiopathic avascular necrosis of the growing femoral head and typically presents in boys at the younger end of the scale (4 to 8 years old). Because of this, these children are often shorter than their peers, and there is a possible association with hyperactivity. The classical findings on a plain AP x-ray of the pelvis include sclerosis leading to destruction and the eventual collapse and flattening of the femoral head.  As this is a biomechanical problem, they might present with a slow onset antalgic gait with pain on internal rotation and ABduction. Around 20% of cases are bilateral.

Case courtesy of Assoc Prof Frank Gaillard, From the case rID: 7980

MRI and bone scans are equally sensitive if the plain films are equivocal. Treatment may be surgical or conservative, depending on the degree of bone destruction and the child’s age. Left unchecked, neovascularisation occurs with the collapse of the femoral head.

Though first described in 1897, it was not until 1910 that it was found to be unrelated to tuberculosis. Arthur Legg, Jacques Calvé and Georg Perthes are guilty of its eponymous name.

Transient synovitis

I think we’ve said enough about transient synovitis already.


As well as a stroke, a limp might be the initial feature of an ascending paralysis in Guillain-Barré. We’ll talk about that and ADEM another time.

Slipped Capital Femoral Epiphysis

A slipped capital femoral epiphysis or slipped upper femoral epiphysis is more common in older, overweight boys over the age of 10 years of age. There is often a structural weakness of the physis itself, so it is more common in cases of endocrine dysregulation (such as hypothyroidism) and metabolic conditions (such as renal osteodystrophy). The proximal epiphysis displaces anteriorly and laterally relative to the metaphysis. Involvement of the medial obturator nerve may mean that they present as knee pain rather than hip pain. Early fixation, and we mean in less than 24 hours in unstable cases, can ward off the threat of life-long pain and deformity. Plain AP films may not be enough in the case of subtle slippage. So, if you are suspicious, then you should go ahead and order a lateral film as well.  A lot of other hip pathology is made easier to spot in a frog-leg lateral. Don’t do this if you are suspicious of a SCFE; you might make things much, much worse.

Look for Klein’s line. A line along the superior aspect of the femoral neck should intercept with the epiphysis. Imagine it as an ice cream slipping off the cone on a hot summer day.

Adapted from case courtesy of Assoc Prof Frank Gaillard, From the case rID: 2715
Case courtesy of Dr. Hani Salam, From the case rID: 9298

Some things need to be in our differential no matter the child’s age.

Septic arthritis

The hip is the most commonly infected joint with the haematogenous spread of organisms grown dependent on the local flora. Young children (under the age of 2) and those who are immunosuppressed or asplenic are at higher risk. In Australia, for instance, the commonest organisms grown are Staphylococcus aureus and Group B Streptococcus (especially now that Haemophilus influenzae B infections have been almost wiped out by immunization). In the youngest, patients consider the weirder organisms like Kingella kingae and Salmonella (in patients with sickle cell disease).

The juvenile physis does not prevent the spread from a remote source from entering the epiphysis, so haematogenous spread (rather than direct local invasion) is often the culprit. The hip is classically held in a position of external rotation, ABduction and flexion, perhaps to maximise the joint space and minimise the pain.

It would be great if we could differentiate the serious diagnosis of septic arthritis from something less serious, such as transient synovitis. Step forward Meninder Kocher et al. and their seminal paper.

Kocher MS, Zurakowski D, Kasser JR. Differentiating between septic arthritis and transient synovitis of the hip in children: an evidence-based clinical prediction algorithm. JBJS. 1999 Dec 1;81(12):1662-70.

They looked at 17 years of data for patients who presented to a single tertiary hospital with an acutely irritable hip. This amounted to 282 cases, in all of which only 168 had their hip joint aspirated.  Of these 168,  only 26 had true septic arthritis confirmed by a positive culture of joint aspirate or 50,000 WCC in the aspirate with a positive culture, 9 had a positive joint culture and negative blood cultures, and 3 had only positive blood cultures. 

By performing a multivariate analysis, they then looked for key indicators that differentiated cases of septic arthritis from the more benign irritable hip. Four key factors cropped up. They had a history of fever, inability to bear weight, a WBC greater than 12 x 109/l and an ESR ≥ 40mm/hr. If you have none of these, then the probability of having septic arthritis is 0.1%. But if you have all four, it jumps to 99.8%.

Of course, this is all the primary literature that most people read, but when Kocher tried to validate the study having all four markers gave the chance of having septic arthritis as 93%. When Luhmann et al. attempted to validate the criteria externally, it dropped to 59%. Caird et al. must have realised that not many of us use ESR anymore, so they appended the more common C-reactive protein (CRP). Having all 5 gave a 97.5% positive predictive value, though there were only 14 cases of transient synovitis and 5 of septic arthritis.


An indolent course and a non-specific physical examination make this a difficult diagnosis to pick up at first glance. Around 1.5-2% of all children presenting with an atraumatic limp will have osteomyelitis.  Plain films may be unremarkable early in the course of the illness and only show periosteal changes after a week to 10 days. An MRI scan is more sensitive, and if the diagnosis is still challenging, radionuclide scanning might pinpoint the infection. The hip is one of the more common joints affected (25%) followed by the tibia/fibula (25%), then the humerus (13%). Osteomyelitis and septic arthritis can coexist as infection from the bone spreads to the joint capsule.

One case series suggests that 40% of cases are due to transient synovitis, chronic muscle sprain or trauma, accounted for 16%, and no diagnosis was made in 30%. All those diagnoses we must learn for exams are much less common (Perthes’ disease 2%, osteomyelitis 1.5%, toddlers’ fractures 1%, and SCFE 1%). What has not been mentioned, as the incidence is incredibly low, is what many parents worry about – cancer.


Whilst parents may not have heard of a slipped capital femoral epiphysis or Kocher’s criteria, they have heard of cancer, and so we need to address it. Primary benign tumours (unicameral bone cysts, for example, as well as malignant ones, can present as bony hip pain.  Hip pain can also present in haematological malignancies such as acute lymphoblastic leukaemia. The most common malignant bone tumours in infants are osteogenic sarcoma and Ewing’s sarcoma. Both of these are most common in the second decade of life.

Persistent pain, coupled with constitutional symptoms such as night sweats, weight loss and night pain, are highly suspicious and should prompt imaging and blood tests.

Selected references

Please, Just STOP LIMPING from Tim Horeczko and the PEM Playbook

Septic arthritis from PEM Morsels

Adamson J, Waterfield T. Fifteen-minute consultation: The limping child. Archives of Disease in Childhood-Education and Practice. 2019 Jun 29:edpract-2018.

Caird MS, Flynn JM, Leung YL, Millman JE, Joann GD, Dormans JP. Factors distinguishing septic arthritis from transient synovitis of the hip in children: a prospective study. JBJS. 2006 Jun 1;88(6):1251-7.

Fischer SU, Beattie TF. The limping child: epidemiology, assessment and outcome. The Journal of bone and joint surgery. British volume. 1999 Nov;81(6):1029-34.

Flynn JM, Widmann RF. The limping child: evaluation and diagnosis. JAAOS-Journal of the American Academy of Orthopaedic Surgeons. 2001 Mar 1;9(2):89-98.

Herman MJ, Martinek M. The limping child. Pediatrics in review. 2015 May;36(5):184-95.

Hill D, Whiteside J. Limp in children: Differentiating benign from dire causes. Journal of Family Practice. 2011 Apr 1;60(4):193.

Hussain M, Gholipour B, Owen N. A pictorial review of the radiological presentations of the child presenting with an acute limp in the emergency department. Clinical Radiology. 2019 Oct 1;74:e14.

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.

Leet AI, Skaggs DL. Evaluation of the acutely limping child. American Family Physician. 2000 Feb 15;61(4):1011-8.

Lewis D, Logan P. Sonographic diagnosis of toddler’s fracture in the emergency department. Journal of Clinical Ultrasound. 2006 May;34(4):190-4.

Long B, Koyfman A, Gottlieb M. Evaluation and Management of Septic Arthritis and its Mimics in the Emergency Department. Western Journal of Emergency Medicine. 2019 Mar;20(2):331.

McCarville MB. The child with bone pain: malignancies and mimickers. Cancer Imaging. 2009;9(Special issue A):S115.

Mooney III JF, Murphy RF. Septic arthritis of the pediatric hip: update on diagnosis and treatment. Current opinion in pediatrics. 2019 Feb 1;31(1):79-85.

Perry DC, Bruce C. Evaluating the child who presents with an acute limp. BMJ. 2010 Aug 20;341:c4250.

Ryan DD. Differentiating Transient Synovitis of the Hip from More Urgent Conditions. Pediatric annals. 2016 Jun 15;45(6):e209-13.

Sawyer JR, Kapoor M. The limping child: a systematic approach to diagnosis. American family physician. 2009 Feb 1;79(3):215.

Singhal R, Perry DC, Khan FN, Cohen D, Stevenson HL, James LA, Sampath JS, Bruce CE. The use of CRP within a clinical prediction algorithm for the differentiation of septic arthritis and transient synovitis in children. The Journal of bone and joint surgery. British volume. 2011 Nov;93(11):1556-61.

Vieira RL, Levy JA. Bedside ultrasonography to identify hip effusions in pediatric patients. Ann Emerg Med. 2010;55(3) :284-9

Wainwright, AM; Catterall, A (2010). “Chapter 27: Legg–Calvé–Perthes disease: Its name”. In Benson, M; Fixsen, J; MacNicol, M; Parsch, K (eds.). Children’s orthopaedics and fractures (3rd ed.). London: Springer. p. 465



No data was found

Leave a Reply

Your email address will not be published. Required fields are marked *