Limping Child

A limp is a presentation, not a diagnosis. It may come from pain, weakness, deformity, limb-length difference, neuromuscular control, inflammatory disease, infection, tumour, trauma or compensation, and the safe clinician does not begin by reciting that list. The first questions are whether this child is sick, whether they can weight bear, whether the limp is painful, and what dangerous diagnosis must not be missed today.

Age. Age is the strongest first organiser. Each band carries its own likely causes, set out under Classification, and the band you name first is the most useful piece of information you already have.

The hip. The hip deserves special attention because children localise pain poorly, and hip pathology may be felt in the thigh or knee. A teenager with knee pain and a normal knee examination may have SUFE; adolescents require a specific hip screen for slipped upper femoral epiphysis, even when the child points to the knee. A younger child with a subtle limp and reduced hip abduction or internal rotation may have Perthes disease.

View larger

Gait. The gait pattern gives the first physiological clue. An antalgic gait shortens stance phase on the painful limb; a Trendelenburg gait reflects hip abductor weakness, hip pain or altered femoral head mechanics; a short-leg gait suggests limb-length difference or pelvic obliquity. Circumduction, toe walking, spasticity, ataxia or foot drop move the differential toward neuromuscular disease.

Mechanism. The mechanism behind the limp differs by pathology. The clinical clue that goes with each is in the pattern-recognition table under Differential Diagnosis.

The same child can be sorted by age, by whether the limp hurts, and by how fast they need to be seen. Each sort answers a different question, and the red flags override all of them.

View larger

History

The history should identify acuity, severity, red flags and the most likely anatomical source.

Examination

Watch first. Observe before touching the child. A child who runs around the room is different from a child who lies still and refuses to move the hip. Name the gait pattern before the child is on the couch; what each pattern means is set out under Pathophysiology.

Then examine from spine to foot. The examination is a map rather than a single joint:

The hip findings that matter. Pain with passive hip motion, especially in a febrile or non-weight-bearing child, is a high-risk finding. Loss of internal rotation in an adolescent suggests SUFE until excluded, and reduced abduction and internal rotation in a school-age child raises concern for Perthes disease.

Investigations should answer the dangerous question still open after history and examination. Do not order the same panel for every child without deciding what you are trying to exclude.

Radiographs

Which films. Choose the projection by the question:

• AP pelvis and lateral hip views for hip, thigh or knee pain when Perthes disease or SUFE is possible; they also assess DDH sequelae or fracture.
• Tibia radiographs (AP and lateral) for toddler fracture; repeat imaging may show periosteal reaction if early films are normal.
• Local radiographs for focal bony tenderness or trauma.
• Skeletal survey through the appropriate safeguarding pathway when non-accidental injury is suspected.

A normal film is not a normal child. Normal early radiographs do not exclude infection, toddler fracture, early Perthes disease or malignancy, so a febrile non-weight-bearing child may have septic arthritis or osteomyelitis with normal first films. If the story remains concerning, escalate rather than discharge on the basis of one normal radiograph.

If SUFE is possible. Make the child non-weight bearing and obtain an AP pelvis and appropriate lateral hip imaging. Do not force a painful frog lateral in a possible unstable slip.

View larger

View larger

View larger

View larger

View larger

Blood tests

FBC, ESR, CRP and blood cultures support infection risk assessment, and stratify septic arthritis and osteomyelitis risk. They are particularly useful in febrile, non-weight-bearing or systemically unwell children. Normal early markers do not overrule a concerning examination.

Ultrasound

Hip ultrasound detects effusion but does not prove pus, and effusion alone does not distinguish transient synovitis from septic arthritis. It is requested when indicated in the febrile or non-weight-bearing child, and it is most useful when interpreted with fever, weight-bearing status, pain with passive motion, inflammatory markers and clinical trajectory.

View larger

MRI

MRI is the escalation test when occult infection, abscess, tumour, stress fracture, discitis, early Perthes disease or unexplained persistent pain remains possible after initial assessment. It is also useful when symptoms are severe and the source is unclear, and it is the answer to the false reassurance of a normal early radiograph.

View larger

View larger

View larger

Septic hip risk stratification

Kocher-style criteria help separate septic arthritis from transient synovitis in a child with an acutely irritable hip, but they are not a substitute for clinical judgement, aspiration when indicated, or senior review.

The criteria are worthless without the probabilities. Counting predictors is not the point; what makes Kocher usable is how steeply risk climbs as they accumulate, and the probabilities are the reason the rule exists:

The shape of the curve. It is not linear: in Kocher's original cohort the jump from two predictors to three was 40 percent to 93 percent, so a single additional criterion moved a child from genuine equipoise to near-certainty. That non-linearity is why the individual thresholds matter and why a half-remembered version of the rule is dangerous, though see the validation caveat below before quoting the absolute numbers.

The modified rule performs better, by a measurable margin. Adding CRP (the Caird version) discriminated septic arthritis from transient synovitis better than the original Kocher rule, with an area under the ROC curve of 0.84 against 0.75 across all joints. It also held up for joints other than the hip (knee, ankle, elbow) at an AUC of about 0.80, which the original rule was never designed for.

Treat the percentages as ordinal, not literal; they did not survive external validation. When Luhmann applied Kocher's algorithm at a different hospital (PMID 15118038, 165 irritable hips), the probability with all four predictors present was 59%, not 99.6%, a forty-point collapse, and a model refitted locally reached only 71%. The authors' conclusion was that neither their own model nor Kocher's "appears to be valid at other institutions". The ranking holds everywhere: more criteria means more risk, and 2 of 4 is genuinely uncertain. The absolute figures do not travel.

Where both rules degrade. Kingella infection and early disease may be indolent, with normal markers in a meaningful minority, so a well-looking toddler with unremarkable bloods is not cleared by a low score. If the child looks septic, has severe pain on passive motion, or deteriorates, escalate whatever the criteria say.

Microbiology and empirical antibiotic cover

When septic arthritis or osteomyelitis is likely, empirical antibiotics should reflect the most probable organism for the child's age and host factors. Staphylococcus aureus is the commonest pathogen overall, but that generalisation breaks down in exactly the age group that limps most: in a series of children aged 6 to 48 months investigated for atraumatic osteoarticular complaints, Kingella kingae accounted for 30 of 40 confirmed infections, the majority, not a footnote. Cover changes with age, vaccination status and comorbidity.

The differential is a set of patterns, each with the decision it forces on the day. The mechanism behind each is under Pathophysiology.

Transient synovitis is a diagnosis of recovery, not a label for every painful hip. A child who remains febrile, cannot weight bear, has rising inflammatory markers, severe passive-motion pain or persistent symptoms needs reassessment for septic arthritis, osteomyelitis, abscess or another diagnosis.

View larger

View larger

View larger

Algorithm

View larger

The unwell or high-risk child. Management runs as follows:

• Escalate early to senior orthopaedic and paediatric teams.
• Keep nil by mouth if septic arthritis or urgent surgery is possible.
• Obtain cultures and inflammatory markers without delaying treatment in a septic child.
• Aspirate and drain the joint when septic arthritis is likely.
• Start empiric antibiotics according to local paediatric infection guidelines after cultures when safe.

The well child with a likely benign cause. Observation is active management, and the plan should sound like one:

• Provide analgesia and activity modification.
• Avoid forced weight bearing.
• Use targeted radiographs when focal tenderness, trauma or hip pathology is possible.
• Review within a defined timeframe if symptoms persist or the diagnosis is not fully secure.
• Give written return triggers: fever, refusal to walk, worsening pain, night pain or systemic symptoms.

Diagnosis-specific pathways. Once the dangerous diagnoses are excluded or confirmed, management follows the diagnosis:

• Toddler fracture: immobilise or support comfort and review for healing.
• Perthes disease: refer for paediatric orthopaedic assessment, hip range monitoring and containment planning.
• SUFE: urgent non-weight bearing, AP pelvis/lateral imaging and operative fixation pathway (in-situ screw fixation for stable slips; consider prophylactic contralateral fixation in younger or endocrine-at-risk children).
• Inflammatory disease: coordinate rheumatology assessment when infection is excluded.
• Malignancy suspicion: urgent paediatric oncology pathway and avoid inappropriate biopsy.

The complications listed here are the price of a missed or delayed diagnosis, early and late.

Early

• Delayed septic arthritis diagnosis with cartilage injury.
• Missed SUFE with progression to unstable slip.
• Missed osteomyelitis with abscess formation.
• Inappropriate reassurance despite malignancy red flags.
• Unrecognised safeguarding concern.

Late

• Avascular necrosis or deformity from missed hip disease.
• Chronic osteomyelitis after delayed infection treatment.
• Growth disturbance from infection or physeal injury.
• Persistent gait abnormality due to missed neuromuscular or structural cause.
• Family loss of trust after vague safety-netting.

A limping child is a safety diagnosis until the dangerous causes are excluded. The first decision is not the final diagnosis; it is whether the child is systemically unwell, unable to weight bear, has severe pain, has a septic joint risk or has a non-orthopaedic emergency. The framework below turns those questions into actions.

The limping child is a worldwide presentation, but the diagnostic emphasis and available tools vary by setting. The core safety logic, separating sepsis and red flags from benign causes, is universal.

Global epidemiology

• Limp is a common paediatric emergency presentation; transient synovitis is the single most frequent cause of acute non-traumatic hip pain in children, typically aged 3 to 8 years.
• Septic arthritis of the hip is uncommon but time-critical, with the highest incidence in infants and young children.
• Kingella kingae is now recognised as the leading cause of osteoarticular infection in children under 4 years where molecular diagnostics are used; in settings reliant on conventional culture, a high proportion of these infections are reported as culture-negative.
• Perthes disease peaks at roughly 4 to 8 years with a male predominance; reported incidence varies several-fold between populations and is higher in some northern European cohorts.
• SUFE incidence tracks childhood obesity and is rising in many regions; bilaterality is common, which is why both hips are imaged and monitored.

Side-by-side guidance

Registry and surveillance notes

• Unlike arthroplasty, the limping child has no single implant registry; evidence comes from prospective cohorts (Kocher, Caird), classification studies (Loder, Herring) and national infection surveillance.
• SUFE and Perthes outcomes are tracked through paediatric orthopaedic networks and multicentre studies rather than device registries, which is why classification-based prognosis (lateral pillar, stable/unstable) remains central.

High- versus limited-resource practice

• Where MRI and molecular testing are readily available, occult infection, early Perthes disease and Kingella can be confirmed earlier and aspiration targeted more precisely.
• In limited-resource settings, serial clinical examination, repeat plain radiographs, ultrasound and empirical pathways carry more weight, and a lower threshold for senior review and joint aspiration compensates for diagnostic uncertainty.
• The universal safety rule holds everywhere: a febrile child who will not weight bear is treated as a possible septic joint until excluded, regardless of the imaging available.

• Prediction rules versus clinical judgement. Kocher and Caird criteria stratify risk but were derived in specific hip cohorts and perform variably elsewhere. They support, never replace, aspiration and senior review when a child looks septic.
• The Kingella era. The rise of Kingella kingae, which often produces mild markers and negative cultures, blurs the line between transient synovitis and true infection in toddlers and challenges scores built before molecular testing.
• Imaging escalation thresholds. When to move from radiograph and ultrasound to MRI, and how quickly, is debated and resource-dependent; MRI is the most sensitive escalation test but access and the need for sedation in young children constrain its use.
• Perthes containment. The benefit of operative containment is concentrated in older children with more severe lateral pillar grades; for younger or milder hips, observation may match surgery, so management remains individualised.
• Antibiotics before aspiration. Balancing prompt empirical antibiotics in a septic-looking child against obtaining diagnostic cultures first is a recurring judgement call, guided by how unwell the child is.

Do not start with a long list of diagnoses. Start with the child status: sick or well, weight bearing or not, painful or painless, and age band. This immediately separates septic arthritis and major red flags from the more common benign causes.

Transient synovitis is common, but the diagnosis should include why septic arthritis is unlikely. A well child who can weight bear, has mild symptoms after a viral illness, has improving pain and has reassuring inflammatory markers is very different from a febrile child who refuses to move the hip.

A normal X-ray is not the same as a normal child. Early osteomyelitis, tumour, toddler fracture, stress fracture and early Perthes disease can have subtle or normal radiographs. If the story remains concerning, escalate to MRI, repeat imaging or senior review rather than discharging on the basis of one normal film.

Non-accidental injury is not an accusation; it is a safety diagnosis that must be considered when the history, developmental stage and injury pattern do not match. The orthopaedic role is to recognise the mismatch, document carefully and activate the correct safeguarding pathway.

The discharge plan should sound precise: "This is the most likely diagnosis, these are the dangerous diagnoses we do not currently think are present, this is the follow-up plan, and these symptoms should bring you back urgently."

Common pitfalls

• Calling it transient synovitis before assessing septic arthritis risk.
• Forgetting that hip disease can present as knee pain.
• Relying on ultrasound effusion alone to diagnose septic arthritis.
• Missing SUFE because the child points to the knee.
• Ignoring night pain or systemic symptoms.
• Not considering safeguarding in a non-walking child with fracture.
• Quoting the four Kocher criteria without the probabilities that make them usable.