The exam rarely asks “what is this fracture?” It asks the things that actually catch people out: the subtle sign that’s easy to miss on the film, the second injury the mechanism should make you look for, the nerve or artery the displacement puts at risk, and what you do next — reduce it yourself, refer it, or fast bleep orthopaedics.

It’s organised the way the exam asks it — for each injury: what gets missed, what the mechanism predicts, the complication to watch for, and where the patient goes next. It’s the high-yield stuff, not the whole textbook.

This film shows: a raised, sail-shaped anterior fat pad and a visible posterior fat pad on a lateral elbow — an effusion, and often the only sign that a fracture is there at all. · J. Heilman MD. CC BY-SA 3.0.

The two most commonly-missed elbow injuries are the undisplaced radial head fracture in the adult and the undisplaced supracondylar fracture in the child — and in both, the effusion is often the only sign on the film. When the fracture line is invisible, the joint fluid gives it away.

• Posterior fat pad = never normal. An anterior fat pad may be a normal finding, but a posterior fat pad is never normal on a true lateral — it carries a roughly 75% occult fracture rate. Treat it as a fracture you cannot yet see.
• Sail sign = haemarthrosis. Displacement of both the anterior and posterior fat pads produces the sail sign, which in acute trauma means blood in the joint.
• Anterior humeral line. On the lateral, the anterior humeral line should intersect the middle third of the capitellum. Failure to do so means the capitellum is displaced — think supracondylar fracture. A subtle posterior shift of the capitellum off this line is the sign people overlook.


Anterior humeral line. Drawn down the anterior humeral cortex it should cross the middle third of the capitellum (left). On the right it passes in front of the capitellum — the alignment clue to a supracondylar fracture. · J. Heilman MD & Kilborn et al./Häggström. CC BY-SA.



Radiocapitellar line. Drawn along the axis of the radial neck it should point to the centre of the capitellum on every view (left). On the right it misses — the radial head is dislocated, so look for the ulna fracture that goes with it. · Kilborn T, Moodley H, Mears S; annotations M. Häggström. S Afr J Radiol 2015. CC BY-SA 4.0.

This film shows: an AP paediatric elbow with the ossification centres visible — the film you match against the child’s age before you call anything a fracture. · JBSR. CC BY 4.0.

CRITOE or CRITOL? Both — they are the same mnemonic. The letters run Capitellum, Radial head, Internal (medial) epicondyle, Trochlea, Olecranon, and then the lateral epicondyle — which some write as E for external epicondyle and others as L for lateral. Same bone, same order, same ages. Use whichever you were taught; just don’t let the spelling convince you there are two different sequences.

CRITOE is not something you learn for the exam — it is something you apply to a paediatric elbow film so you don’t miss a fracture. The ossification centres appear in a fixed order, at roughly ages 1, 3, 5, 7, 9 and 11: Capitellum (1), Radial head (3), Internal (medial) epicondyle (5), Trochlea (7), Olecranon (9), External (lateral) epicondyle (11). Get the centres wrong and you miss a fracture. Watch for:

• Ossification centre mistaken for a fracture. An unfused centre appearing at its expected age is normal, not a fragment. The paediatric olecranon pitfall is the bifid olecranon ossification centre, which can be read as a fracture.
• Fracture mistaken for — or hidden behind — an ossification centre. The classic catch is the medial (internal) epicondyle avulsed and incarcerated in the joint. If the film shows the trochlea (which ossifies after the medial epicondyle in the CRITOE order) but the medial epicondyle appears “absent”, the epicondyle fragment may be sitting inside the joint — the order tells you it should already be there. Find it, or you miss it.
• The lateral condyle fracture is intra-articular. In a child (around 4–10 years) it crosses the joint surface and the growth plate (a Salter-Harris IV), so it is unstable and even a small-looking fragment can displace. Miss it or under-treat it and you risk non-union, cubitus valgus and a late (“tardy”) ulnar nerve palsy — so it is never “just a chip” and needs orthopaedic follow-up.

The mechanism in the stem points to both the main injury and the second injury you need to look for. It is how you catch the injury that is not on the view you have been given.

The second axis is just as testable: once you have the first injury, it tells you what to go looking for.

The unifying rule: a single-bone forearm fracture is a dislocation until you have cleared the joint at the other end. An isolated ulna fracture without a checked radiocapitellar line, or a distal radius fracture without a checked DRUJ, is how the second injury gets missed.

The supracondylar fracture is high-yield for two straightforward reasons: it is common — roughly 60% of paediatric elbow fractures — and it has a BOAST standard of its own, which tells you how much there is to get wrong. Peak age 5–8 years, from a FOOSH, and 95–98% are the extension type. The disposition follows the Gartland grade and the neurovascular exam.

Like any classification worth learning, Gartland encodes two things: how stable the injury is and how severe it is — and those are what decide what happens to the patient and what your department has to do next. Expect it in both directions. A stem may hand you “a Gartland II” and expect you to know what follows; or it may hand you the film and expect you to grade it yourself and act on it.

So the ED decision is straightforward: an undisplaced Gartland I with a normal hand goes home with fracture-clinic follow-up; anything displaced is an urgent orthopaedic referral for reduction and fixation; a pulseless hand needs urgent reduction to restore the pulse; and an ischaemic hand, an open injury or compartment syndrome is a surgical emergency. In the ischaemic limb the thing that has to happen is reduction. Ideally orthopaedics do it — these usually need fixing in theatre anyway — so bleep them at once. But that assumes the ortho team can attend immediately, and a compromised limb cannot wait on that: if they cannot come, an experienced ED clinician should attempt the reduction rather than let the limb sit.

The neurovascular exam. Test and document the radial pulse, digital capillary refill, and the individual function of the radial, median (including the AIN, via the OK sign) and ulnar nerves. The extension type threatens the AIN / median nerve and the brachial artery; the flexion type threatens the ulnar nerve. Most neuropraxias resolve spontaneously.

When you backslab, don’t hyperflex. Immobilise at ≤90° of flexion — hyperflexion tightens the antecubital fossa and risks Volkmann’s ischaemic contracture. (Varus malunion later gives the cubitus varus / gunstock deformity.)

The exam tests this both ways — it gives you the injury or the X-ray and asks what you will find on examination, or it gives you the examination finding and asks for the injury or the nerve. Either direction needs the same map. Each injury has a signature nerve, each nerve a clean bedside test — and crucially, the two purely-motor branches (AIN and PIN) have no sensory loss, which is itself the clue.

The distinction the exam leans on is wrist drop versus finger drop: the radial nerve proper gives a true wrist drop with sensory loss; the PIN gives a finger drop with the wrist still extending (radially) and no sensory loss. Similarly, the AIN gives a motor deficit with a normal sensory exam — easy to dismiss as “the child won’t cooperate” unless you specifically test the OK sign. Read it forwards — “posteriorly displaced supracondylar, what will you find?” → a failed OK sign with normal sensation (AIN) — and backwards — “finger drop, wrist still extends, no sensory loss?” → the PIN, so go looking for a Monteggia.

Monteggia. A paediatric ulna fracture with the radial head dislocated — obvious on the lateral once you follow the radiocapitellar line. The ulna fracture is the part that gets reported; the dislocation is the part that gets missed. · Hellerhoff, Wikimedia Commons. CC BY-SA 3.0.

Galeazzi. A distal-third radius fracture with disruption of the distal radioulnar joint. One forearm bone broken should always send you to the joint at the other end. · JETem. CC BY 4.0 (Honda R. 2017).

Essex-Lopresti — the wrist half. The affected wrist (a, b) shows the DRUJ lesion; the unaffected opposite wrist (c, d) is there for comparison, which is often what makes it obvious. The elbow half of this injury is a radial head fracture — this pair is the reason you examine the wrist before you splint one. · Fontana et al., BMC Musculoskelet Disord 2018. CC BY 4.0.

These three eponyms share one feature: a fracture with an easily-missed associated dislocation. The fracture is obvious; the dislocation is what gets missed, and finding it is the exam skill.

• Monteggia = proximal-third ulna fracture + radial head dislocation. Bado I (anterior, ~60%) is commonest; II is posterior; III lateral; IV involves both bones. Check the radiocapitellar line on the elbow film, not just the forearm. The PIN is at risk (~10%, blunt stretch, observed).
• Galeazzi = distal-third radius fracture + DRUJ disruption. Image and inspect the wrist.
• Essex-Lopresti = radial head fracture + interosseous membrane disruption + DRUJ injury. Always examine the wrist in a radial head fracture — miss it and a “simple” radial head fracture becomes a forearm-length instability.

Age changes the management. In adults closed treatment of a Galeazzi fails — the DRUJ redislocates — so it needs ORIF. The paediatric versions are often managed by closed reduction, so do not carry the adult rule across to a child. Disposition either way: refer to orthopaedics for reduction and fixation.

This film shows: a posterior elbow dislocation on the lateral view — the commonest direction, and the one you reduce yourself when there is no associated fracture. · JETem. CC BY 4.0.

The disposition splits cleanly on simple versus complex. Elbow dislocations are most often posterior / posterolateral (80%); roughly half to sixty per cent are simple (no associated fracture).

• Simple posterolateral dislocation → reduce in the ED. Check neurovascular status before AND after reduction (brachial artery, median, ulnar), confirm reduction on a repeat X-ray, splint at ~90° for 5–10 days, then start early motion.
• Complex / unstable → urgent orthopaedic referral, and usually theatre. The terrible triad — elbow dislocation + radial head/neck fracture + coronoid fracture — is unstable because the radial head, coronoid and lateral collateral ligament all contribute to stability. It is an operative injury.

The pulled elbow (radial head subluxation, “nursemaid’s elbow”) is the elbow injury you resolve at triage. It affects children aged 1–5, from axial traction on a pronated, extended forearm, with the annular ligament interposed between the radial head and capitellum. The child holds the arm still and pronated and won’t use it.

• Two techniques. Hyperpronation (hyperpronate the forearm in flexion) or supination-flexion (supinate, then maximally flex, feeling for a click over the radial head).
• Hyperpronation is more successful at the first attempt. A Cochrane review (CD007759) found hyperpronation had a lower first-attempt failure rate (RR 0.35, 95% CI 0.25–0.50, NNT 6) — but the evidence is low quality, and you should not claim it is less painful (the pain evidence was insufficient).
• When not to X-ray. No imaging is needed if the history and exam are classic (the film would be normal); image only if the presentation is atypical or a fracture is suspected.
• Immobilisation. None needed after a first episode; a recurrent case is cast in flexion.

Not every elbow that arrives in the ED has been injured. The atraumatic painful elbow is a distinct presentation, and the first thing the exam wants is the right label — RCEM Learning is explicit that in tennis elbow, “previously known as lateral epicondylitis, the pathological process is thought to be due to failed tendon healing and degeneration rather than acute inflammation”. It is a tendinopathy, or epicondylosis. That is why anti-inflammatory thinking disappoints, and why the honest answer to “what will fix it?” is often “time”.

• Lateral epicondylosis (tennis elbow). Overuse of the forearm extensor tendons; a diffuse ache over the lateral elbow. Test it by palpating the common extensor origin during resisted wrist extension with the forearm pronated.
• Medial epicondylosis (golfer’s elbow). Less common; overuse of the forearm flexor tendons; pain medially. Test it on resisted wrist flexion with the forearm supinated. In 20% there are associated ulnar nerve symptoms — paraesthesia in an ulnar distribution — so examine the nerve rather than assuming it is only a tendon problem.
• What the evidence actually supports. Relative rest, ice and analgesia. RCEM Learning notes the condition is often persistent, that corticosteroid injection reduces pain but the effect does not appear to last beyond six weeks, and that physiotherapy has an established role although results at 12 months are comparable with patients who received no intervention at all. Promising a quick fix is the wrong answer clinically and in the exam.

Olecranon bursitis is the other common atraumatic swelling, and the only one of these where you can do real harm. It usually follows direct trauma — a fall onto the elbow — or repetitive friction, and it may be associated with crystal arthropathy or inflammatory arthritis.

• Septic or not? Septic bursitis is most commonly Staphylococcus aureus, arriving directly through broken skin. Look for the systemically unwell patient: pyrexia, cellulitis and axillary lymphadenopathy, on top of the local swelling, tenderness and erythema. Bloods may show a raised ESR, CRP and white cell count, and bursal aspiration with microscopy and Gram stain confirms a septic or crystal-induced cause.
• The trap is the needle. RCEM Learning is blunt: “caution is required when considering aspiration of olecranon bursae since fistula formation is a widely recognised complication of the procedure.” Aspirate to answer a question, not out of habit.
• Treatment. Septic bursitis needs antibiotics — intravenous if the patient is systemically unwell — as per local policy.
• Don’t mistake it for septic arthritis. Bursitis sits superficially over the olecranon and range of movement is typically preserved; a hot elbow that will not move is a joint problem until proven otherwise.

Finally, the tendon ruptures. Resisted elbow flexion and extension stress the biceps and triceps tendons respectively, and increased pain suggests a musculotendinous strain. What matters more is the patient with grossly reduced flexion or extension strength alongside pain and localised swelling or bruising — that pattern, in RCEM Learning’s words, “might suggest a rupture of the biceps tendon or avulsion fracture of the triceps insertion”. Weakness, not pain, is the finding that changes the diagnosis.

Some complications convert a discharge into an admission or an urgent orthopaedic referral. These are the recognition points that change the disposition.

• Compartment syndrome / Volkmann’s ischaemic contracture. The cardinal sign is pain on passive stretch, out of proportion; the setting that precipitates it is a hyperflexed cast after a supracondylar fracture. Suspicion mandates immediate reassessment and intervention — not analgesia and review.
• Tardy (delayed) ulnar palsy. Develops years later from a lateral condyle non-union with cubitus valgus (~10%). The relevance in the ED is recognising that a “minor” lateral condyle fracture is the seed of a late nerve palsy, so it is not a trivial injury.
• Incarcerated medial epicondyle fragment. After an elbow dislocation reduces, the medial epicondyle fragment is trapped in the joint in ~15% — an operative problem, and the reason to scrutinise the post-reduction film (CRITOE again).
• Essex-Lopresti — the wrist. A radial head fracture with wrist pain is interosseous membrane + DRUJ disruption until excluded; missing it leaves a forearm-length instability.

The whole guide resolves into one decision: who owns this injury next? Map each pattern to its disposition.

Match the immobilisation to the injury:

• Collar & cuff = gravity traction — the arm hangs, aiding alignment; use for humeral and radial head injuries.
• Backslab (above-elbow) at ≤90° flexion — for olecranon and supracondylar fractures (never hyperflex the supracondylar).
• Broad arm sling = support without traction — general support of the forearm.

The exam hands you an elbow scenario and asks for the next step. It’s a sequence, and each step maps to something above:

• 1. Read the mechanism — predict the injury and the second injury (FOOSH by age; ulna → radiocapitellar line; distal radius → DRUJ; radial head → wrist).
• 2. Apply the tools to spot the subtle sign — fat-pad and alignment lines in the adult; CRITOE in the child (the incarcerated medial epicondyle, the bifid olecranon).
• 3. Name the nerve or complication — from the pattern (AIN vs ulnar vs PIN vs radial proper; Volkmann’s; tardy ulnar palsy).
• 4. Make the disposition call — reduce it yourself, refer it, or fast bleep orthopaedics. Remember: in the exam you are the consultant, so “ask a senior” is never the answer — escalate to orthopaedics.

The commonest trap is stopping at the diagnosis. Naming the supracondylar fracture earns nothing if you then hyperflex the cast, fail to document the AIN, or send a pink pulseless hand for an exploration that BOAST says it does not need.

If you want the method behind building revision around how this specific exam works — rather than how you revised for everything else — see how I rebuilt my guideline revision for the FRCEM Final, and our FRCEM Final SBA technique guide for reading the lead-in. For another high-yield “spot the second injury” topic, see our paediatric cardiac arrest in special circumstances guide.

Sources: RCEMLearning — Traumatic Elbow Injuries; Orthobullets (supracondylar, radial head, Monteggia, Galeazzi, lateral condyle, medial epicondyle, terrible triad); BOAST — Supracondylar Fractures of the Humerus in Children (British Orthopaedic Association); Cochrane review CD007759 (manoeuvres for pulled elbow); StatPearls (PIN syndrome, wrist drop, radial nerve injury, claw hand). Doses and thresholds are a high-yield revision guide — always confirm against current guidance.