It is 3AM. A 7 year old is wheeled into your ED. Mum is exhausted. The triage note says “headache x 2 days.” The child is quiet, curled up, keeping the lights off with one hand over the eyes.

Two doors down there is another 7 year old with a headache. Same triage note. This one is chatty, eating crisps, watching cartoons on a phone at full brightness.

Same complaint. Wildly different stories. One of these might have a bleed. One of these needs paracetamol and a dark room. The whole game tonight is telling them apart without CT scanning every child who walks through the door.

Here is the uncomfortable truth this article is built on: the eye does not see what the mind does not know. If you do not carry the differential in your head, you will not spot the red flag sitting in front of you. So let us load the differential first, then go back to the history and examination knowing exactly what we are hunting for. 🎯

📊 The Spectrum: Three Tiers

Everything that causes a nontraumatic headache in a child fits into a pyramid. Most of what you see lives at the bottom. What can kill lives at the top. Your brain needs both loaded at once.

🔴 TIER 1 ➜ LIFE-THREATENING (rare but lethal)
▪️Infection (meningitis, encephalitis, abscess)
▪️ intracranial haemorrhage
▪️ CO poisoning
▪️ tumour
▪️ malignant hypertension

🟣 TIER 2 ➜ OTHER CONDITIONS (uncommon)
▪️Idiopathic intracranial hypertension
▪️ TMJ dysfunction
▪️ cluster headache

🔵 TIER 3 ➜ COMMON (the overwhelming majority)
▪️Viral illness with fever
▪️ migraine
▪️ tension-type headache
▪️ sinusitis
▪️ pharyngitis
▪️ otitis media

Clinical pearl: The current PECARN and AAN framing is that roughly 0.5% to 1% of children presenting to the ED with headache have an emergent intracranial abnormality such as a tumour, haemorrhage, or stroke.
Yet more than a third get scanned.

🔴 TIER 1: The Stuff That Can Kill

🦠 Infection

Children with a life-threatening infection almost never have headache in isolation. Look for the company it keeps: altered mental status or focal neurologic findings.

Bacterial meningitis 🧫

• The classic picture is the irritable or lethargic child with fever plus neck pain, stiffness, or headache.

• Incidence has dropped since routine Hib (2, 3, 5 and 18 months) and pneumococcal vaccination. Do not let that lull you. Undervaccinated and partially vaccinated children still present.

Viral encephalitis 🧠

• Progressive fever and headache plus abrupt altered sensorium (bizarre behaviour or hallucinations).

• In previously healthy kids, HSV causes roughly 10 to 20% of sporadic encephalitis and it is the one treatable pathogen. Miss it and you miss your window with aciclovir.

• Arboviruses matter too (eg Japanese encephalitis, dengue, West Nile depending on region). Take a travel history.✈️

  

Orbital or cerebral abscess 👁️

• Focal infections give focal signs on ocular and/or neurologic examination.

• Cerebral abscess and cerebral venous thrombosis are uncommon but life-threatening complications of orbital, sinus, ear, and dental infections.

• Higher risk in immunocompromise and congenital heart disease with right-to-left intracardiac shunting.

🩸 Subarachnoid and Intracranial Haemorrhage

• Abrupt onset headache reaching maximal intensity in under one minute (thunderclap).

• May come with hemiparesis and seizures.

• Sickle cell disease is a big one: roughly 1 in 3 headaches in sickle cell children are actually bleeds. Do not treat that headache as routine. 🚩

💨 Carbon Monoxide Poisoning

• Symptoms of mild to moderate CO poisoning are nonspecific and headache is prominent.

• Occult sources: improperly vented heaters, space heaters, automobile exhaust fumes, and house fires. 🔥

• The tell: multiple family members from the same household with headache. If the whole family is symptomatic, think CO, not a virus. A single family unit presenting together is a screaming red flag.

🎗️ Brain Tumour

• Chronic progressive headache (increasing frequency and severity over time) is the classic presentation.

• Reassuring reality: fewer than 3% of children presenting to the ED with headache have a newly diagnosed brain tumour.

• Early symptoms are nonspecific and diagnosis is often delayed. But at the point of diagnosis, the overwhelming majority of these children have other symptoms alongside the headache. Isolated headache with a normal exam is rarely a tumour.

🩺 Hypertension

• Hypertensive crisis can present as headache from hypertensive encephalopathy.

• Look for visual changes (retinal haemorrhage or exudates), altered mental status, and seizures.

• In children, hypertensive encephalopathy is uncommon and usually driven by systemic illness or renovascular disease, not primary hypertension.

• Phaeochromocytoma is exceptionally rare but presents with the memorable quartet: headache, pallor, diaphoresis, and hypertension. 💦

• Actually measure the blood pressure. With the right cuff size. It is the single most skipped vital in a headache workup.

🔵 TIER 3: The Common Stuff (Where Most of Your Night Lives)

🤒 Infection with Fever

• Children with fever, almost always from infection, frequently complain of headache.

• The mechanism is vasodilation. Treat the fever, treat the source, the headache usually follows.

🦠 Viral Meningitis

• A typical presentation in older children and adolescents.

• Hunt for virus-specific clues: rash or mouth lesions point to enterovirus.

😷 Pharyngitis

• Headache, sore throat, and abdominal pain are the classic triad of streptococcal pharyngitis.

• Influenza mimics it and adds headache to the mix.

👂 Other Infections

• Otitis media, sinusitis, and dental infections all cause headache.

• Localised pain plus signs of inflammation give you the diagnosis at the bedside.

🌩️ Migraine

• In EDs presenting with acute headache, migraine is second only to viral illness as the primary cause.

• Typically recurrent and episodic with a pattern parents can describe in their sleep.

• The red flag inside migraine: any child with a significant change in the quality, severity, or timing of a known headache pattern needs re-evaluation for another cause. Do not anchor on “known migraineur.”

Features of migraine in children (how paeds differs from adults): 👇

• Attacks last 2 to 72 hours (adults: 4 to 72). This paediatric duration is the ICHD-3 modification for under-18s.

• Headache is more often bilateral in children. The adult unilateral pattern emerges in late adolescence or early adulthood.

• Photophobia and phonophobia can be inferred from behaviour in young children (turning off the lights, hiding under a blanket).

📋 ICHD-3: Migraine Without Aura

A ➜ At least five attacks fulfilling B through D

B ➜ Headache attacks lasting 4 to 72 hours untreated (2 to 72 hours in children and adolescents under 18)

C ➜ At least two of: unilateral location ▪️ pulsating quality ▪️ moderate or severe intensity ▪️ aggravation by or avoidance of routine physical activity

D ➜ During headache at least one of: nausea and/or vomiting ▪️ OR photophobia and phonophobia

E ➜ Not better accounted for by another ICHD-3 diagnosis

📋 ICHD-3: Migraine With Aura

A ➜ At least two attacks fulfilling B and C

B ➜ One or more fully reversible aura symptoms: visual ▪️ sensory ▪️ speech/language ▪️ motor ▪️ brainstem ▪️ retinal

C ➜ At least three of six: aura spreads gradually over ≥5 min ▪️ two or more symptoms in succession ▪️ each symptom lasts 5 to 60 min ▪️ at least one symptom unilateral ▪️ at least one symptom positive ▪️ aura accompanied or followed within 60 min by headache

D ➜ Not better accounted for by another ICHD-3 diagnosis

😣 Tension-Type Headache

• The International Headache Society term for what used to be called tension, muscle-contraction, stress, or psychogenic headaches.

• Occurs in 10 to 25% of schoolchildren and adolescents, making it about as common as migraine.

• In children it associates with depression, oromandibular dysfunction, and muscle tension from stress.

• The key contrast with tumour: pain is intermittent, recurring, but nonprogressive, and the neurologic exam is unremarkable.

🟣 TIER 2: The In-Between

🦷 Temporomandibular Joint (TMJ) Dysfunction

• Occurs in children with primary and mixed dentition.

• Increased pain while chewing plus point tenderness over the mandibular condyle.

• Headache is more frequent among older children.

🧠 Idiopathic Intracranial Hypertension (IIH)

The classic pentad:

• Papilloedema 👁️

• Increased ICP

• Normal CSF content

• Normal neuroimaging

• Absence of neurologic signs except cranial nerve VI palsy

• No known cause

Who gets it: headache is the most common complaint in children old enough to describe it. Most patients are adolescents and typically female. 🚩 Check the fundi. Every time.

👁️ Trigeminal Autonomic Cephalalgia (Cluster Headache)

• The most common trigeminal autonomic cephalalgia.

• Unilateral, frontal-periorbital, severe, lasting under three hours, but occurring in tight “clusters” over a short period.

• Mostly presents between ages 10 and 20.

• Look for ipsilateral autonomic findings: lacrimation, rhinorrhoea, ophthalmic injection, and sometimes Horner syndrome (ipsilateral miosis, ptosis, facial anhidrosis).

🔍 Now Go Back to the History

Now that you know what you are looking for, the history stops being a checklist and becomes a hunt.

👶 Patient Age

Headaches in children younger than six are more likely to have an underlying cause, most commonly a febrile viral illness. Young age plus headache should lower your threshold to look harder.

⏱️ Mode of Onset

• Thunderclap (”worst headache of my life,” maximal in under a minute) raises intracranial haemorrhage from AVM or aneurysm.

• But in isolation, without neck pain or stiffness, thunderclap is nonspecific in children. Context matters.

🕐 Timing

• Tension-type headaches develop late in the day and rarely wake a child from sleep.

🛏️ Positional

• A headache that presents or worsens on lying down, wakes the child from sleep, or hits soon after waking should raise concern for increased ICP or a space-occupying lesion, especially when paired with morning vomiting. 🚩🚩

📈 Severity and Quality

• Less useful in young children than in adults.

• Use sleep disturbance as a proxy for severity when the child cannot articulate it.

🗺️ If the Child Can Describe It

• Laterality: unilateral suggests migraine or focal infection (sinusitis).

• Location: cluster headaches are usually temporal or retro-orbital.

🔄 Pattern of Previous Headaches (The Money Framework)

1️⃣ ACUTE ➜ single episode, no prior history ➜ think viral illness
2️⃣ ACUTE RECURRENT ➜ consistent features, easily described ➜ think migraine
3️⃣ CHRONIC PROGRESSIVE ➜ increasing frequency and severity over time ➜ 🚩 think brain injury / haemorrhage / SOL
4️⃣ CHRONIC NON-PROGRESSIVE ➜ frequent, persistent, unchanged in character ➜ think tension headache

🚩 Associated Symptoms to Actively Ask About

Fever · neck pain · altered mental status · localised pain · trauma · environmental exposure · changes in visual acuity · autonomic symptoms · aura.

📁 Past Medical History That Changes the Game

Malignancy · coagulopathy · sickle cell disease · cyanotic heart disease. Each one moves a “probably benign” headache up the pyramid.

🩺 The Examination (Where the Diagnosis Usually Hides)

👀 General Appearance

Most children with a serious cause look ill or have altered mental status (confused, lethargic, or comatose). The well-appearing, interactive child is your reassurance, not your CT scanner.

👂 Head and Neck

Examine to rule out trauma, then hunt for the source:

• Otitis media on otoscopy

• Sinusitis: facial tenderness or purulent rhinorrhoea

• Strep pharyngitis: tonsillar swelling, erythema, exudate

• Teeth and gingiva: inflammation or abscess

• TMJ / masseter tenderness: TMJ dysfunction

• Nuchal rigidity: migraine, meningitis, or intracranial haemorrhage

🖐️ Skin

• Neurocutaneous syndromes (neurofibromatosis, tuberous sclerosis) carry brain-lesion risk. Look for café-au-lait spots and ash-leaf macules.

• Bruising in a suspicious distribution raises non-accidental injury. Do not skip the skin. 🚩

🧠 Neurologic Examination (Non-Negotiable)

The majority of children with a serious neurologic cause have abnormalities on neuro exam. This is your highest-yield test and it costs nothing.

• Altered mental status: encephalitis, haemorrhage, elevated ICP, hypertensive encephalopathy.

• Focal or prolonged-postictal seizures: intracranial focus (haemorrhage, tumour).

• Extraocular palsies or nystagmus: elevated ICP or mass compression.

• Motor asymmetry, gait disturbance, poor fine motor coordination: focal intracranial lesion.

• ⚠️ Trap: some children with migraine develop focal deficits (ophthalmoplegia, motor weakness, ataxia). Usually caretakers can confirm this is the child’s typical pattern. First-ever focal deficit is not migraine until proven otherwise.

🔦 Fundoscopy

• Do it on every child being evaluated for headache.

• Hard in young or uncooperative patients, but try.

• Papilloedema, haemorrhages, exudates, and abnormal vessels are markers of serious intracranial pathology.

🖥️ Imaging: When to Pull the Trigger

✅ SCAN when any of these are present:

• 🚩 Focal or gross neurologic deficit: cranial nerve deficit, visual field cut, focal motor weakness, asymmetric reflexes, focal seizure (beyond simple febrile), prolonged altered mental status (>60 min), ataxia, or dysmetria.

• 🚩 Papilloedema or other signs of raised ICP: headache precipitated or worsened by prolonged lying down (especially with vomiting), hypertension with brady- or tachycardia, or progressive vomiting.

• 🚩 Thunderclap / “worst headache”, especially with neck pain or stiffness.

🛑 DO NOT scan when:

• Nontraumatic headache + no worrisome features + normal fundoscopy + normal neuro exam.

• This group has a very low risk of a life-threatening intracranial abnormality. \

• These patients do not require imaging. Radiation and sedation are not free. 📉

💉 Lumbar Puncture: When and How

LP is indicated when:

1. Suspected non-focal infection (meningitis, encephalitis) 🦠

2. Concern for subarachnoid haemorrhage not diagnosed on neuroimaging

3. Measuring opening pressure for suspected IIH (after neuroimaging)

Critical safety point ⚠️: Patients with a focal neuro exam or significant concern for raised ICP / papilloedema need an emergency CT or fast MRI before LP because of herniation risk.

Send the right tubes: CSF cell count and differential, glucose, protein, Gram stain, and bacterial culture. Add opening pressure when IIH is on the table.

🧭 The Diagnostic Algorithm

1. “Worst headache of life” / thunderclap?

   • Yes → work up for subarachnoid haemorrhage / vascular cause.

   • No → next question.

2. History of chronic or recurrent headaches?

   • Yes → typical pattern with no new findings? → migraine / tension / chronic daily headache.

   • No, or new findings → next.

3. Focal neuro exam OR signs of raised ICP (papilloedema; hypertension with brady/tachycardia and progressive vomiting; headache worse lying down, especially with vomiting)?

   • Yes → neuroimaging.

     • Acute process on imaging → tumour, mass, hydrocephalus, SAH/vascular cause.

     • No acute process → focal findings >60 min → migraine variants (prolonged aura, hemiplegic, cerebrovascular, Todd’s). Otherwise → LP with opening pressure → interpret CSF (RBCs → SAH; WBCs → meningitis/encephalitis; raised opening pressure → IIH or cerebral venous thrombosis; normal → pseudopapilloedema).

   • No → Fever?

     • Yes → meningeal signs? → CSF abnormal → meningitis. Otherwise viral syndrome / sinusitis / dental abscess.

     • No → other history/exam abnormal? → exposure (CO poisoning), raised BP (hypertension), focal tenderness (sinus/dental abscess, TMJ). Otherwise migraine / tension headache.

🎯 The Bottom Line

• The differential is your safety net. Load it before you walk into the room. The eye does not see what the mind does not know.

• Isolated headache with a normal neuro exam, normal fundoscopy, and no red flags is reassuring and does not need a CT. Roughly 99% of these children are fine, and radiation is not benign.

• The red flags that should stop you cold: thunderclap with neck stiffness, morning headache with vomiting, headache worse lying down, chronic progressive pattern, focal neuro deficit, papilloedema, age under 6, and the child who simply looks unwell.

• Special populations move the needle: sickle cell (1 in 3 headaches are bleeds), the whole family symptomatic (CO poisoning), immunocompromise and cyanotic heart disease (abscess).

• Fundoscopy and a proper neuro exam are your two highest-yield, zero-cost investigations. Do them every single time.

📚 References

1. Bachur RG, Shaw KN. Fleisher & Ludwig’s Textbook of Pediatric Emergency Medicine. Lippincott Williams & Wilkins, 2015.

2. Saladino RA. Emergency department approach to nontraumatic headache in children. UpToDate.

3. Headache Classification Committee of the International Headache Society (IHS). The International Classification of Headache Disorders, 3rd edition. Cephalalgia 2018; 38(1): 1-211.

4. Lewis DW, Ashwal S, Dahl G, et al. Practice parameter: Evaluation of children and adolescents with recurrent headaches. Neurology 2002; 59: 490.

5. Massano D, Julliand S, Kanagarajah L, et al. Headache with focal neurologic signs in children at the emergency department. J Pediatr 2014; 165: 376.

6. Papetti L, Capuano A, Tarantino S, et al. Headache as an emergency in children and adolescents. Curr Pain Headache Rep 2015; 19: 3.

7. Wilne S, Koller K, Collier J, et al. The diagnosis of brain tumours in children. Arch Dis Child 2010; 95: 534.

8. Seshia SS, Abu-Arafeh I, Hershey AD. Tension-type headache in children: The Cinderella of headache disorders. Can J Neurol Sci 2009; 36: 687.

9. Anttila P. Tension-type headache in childhood and adolescence. Lancet Neurol 2006; 5: 268.

10. Lewis DW. Pediatric migraine. Neurol Clin 2009; 27: 481.

11. PECARN. Stratification of risk for emergent intracranial abnormalities in children with headaches (study protocol). 2023.

You’ve decided to intubate.

Good.

Now the uncomfortable bit: the act of intubating a sick child can kill them faster than the illness you are treating.

Peri-intubation cardiac arrest in children is not a freak event. It is a predictable consequence of putting a hypoxic, hypovolaemic, acidotic child to sleep and removing their own respiratory drive at the exact moment they can least afford it.

This post is about the minutes before you touch a laryngoscope. Get these right and the tube almost places itself. Get them wrong and you are running a code. 🚨

⚠️ Evidence note

Much of paediatric peri-intubation practice rests on physiological reasoning, observational data and extrapolation from adult and neonatal work rather than large paediatric RCTs. Treat the mechanisms and principles here as consensus- and physiology-based teaching. Defer to your local RSI and PICU guidelines for specific drugs, doses and flow settings. I have deliberately not stated fixed doses or flow rates, because these vary by protocol and a wrong number is dangerous.

⚠️ What actually kills them

Three mechanisms. Learn them cold.

1️⃣ Hypoxia 🔵

Children desaturate fast. A small child has a small functional residual capacity (the oxygen reservoir left in the lungs after a normal breath out) and a high oxygen consumption per kilo. Small reservoir plus high demand means the safe apnoea window is short. You paralyse, you lose 30 seconds fumbling, and the SpO₂ is already falling hard.

Exact desaturation times vary widely by age, weight, lung pathology and pre-oxygenation quality. Any quoted “seconds to desaturation” figure is a rough teaching estimate, not a promise.

2️⃣ Hypotension and cardiovascular collapse 🩸

The sick child is often running on their own adrenaline. Their blood pressure is being propped up by sympathetic drive. Your induction agent switches that drive off. Positive pressure ventilation then drops venous return (the blood coming back to the heart) on top of it. The floor falls out.

3️⃣ Acidosis 🧪

A severely acidotic child compensates by breathing hard to blow off CO₂. The moment you paralyse them, that compensation stops. CO₂ climbs, pH drops further, and an already irritable heart arrests.

🫧 Apnoeic oxygenation: buy yourself time

Core principle: even when the child is not breathing, the alveoli keep absorbing oxygen. If you keep oxygen flowing past the vocal cords, it is drawn down into the lungs by bulk flow. This stretches out the safe apnoea window.

How to do it 👇

• 🟢 Pre-oxygenate properly first. Aim for the highest SpO₂ you can get, for a few minutes if time allows. This washes nitrogen out of the lungs and fills the reservoir with oxygen.

• 🟢 Leave nasal cannulae on underneath your mask or NIV interface, and keep them running through the apnoeic period and the attempt itself.

• 🟢 Flow scales with size. [Inference] Precise flow settings are unit- and device-dependent and based on observed practice rather than a single fixed rule. Follow your local guidance.

⚠️ Apnoeic oxygenation buys time. [Inference] Based on the underlying physiology and observed practice, it delays desaturation rather than preventing it. It does not replace good pre-oxygenation or a fast, clean first attempt.

🧍 Positioning by age: geometry wins attempts

You cannot brute-force a bad view. Positioning is free and it works.

👶 Infants (roughly under 2)

Big occiput (the back of the head). Lie them flat and the head flexes forward, folding the airway shut. Fix it:

• 🔹 Shoulder roll to counteract the occiput and line the airway up.

• 🔹 Aim for a neutral to slightly extended head position, not the adult sniffing ramp.

• 🔹 Ear-to-sternal-notch is the target sightline.

🧒 Older children and adolescents

• 🔹 The adult sniffing position increasingly works: neck flexed on body, head extended on neck.

• 🔹 In bigger or heavier adolescents, ramp the upper body so the ear sits level with the sternal notch. This also improves pre-oxygenation.

🩺 Haemodynamic preparation: fix the circulation first

This is the step people skip and it is the one that kills.

Before induction, ask: fluid-down, pressure-down, or both? 🤔

• 💧 Volume: if they are dry or septic, get fluid in or at least have it running and ready. Hypovolaemic child plus induction agent equals collapse.

• 💉 Vasopressors ready: have a push-dose or peripheral infusion drawn up and in your hand or running before you induce any child you think is fragile. Reaching for it after the arrest is too late.

• ❤️ Choose the agent for the haemodynamics. [Inference] Agent choice and dosing are context- and guideline-specific. The general principle from observed practice is to favour agents with a smaller cardiovascular hit and to reduce dose in the shocked child. Follow your local RSI guideline for the actual drugs and doses.

🚩 Resuscitate before you intubate. If you have time to optimise the circulation, take it. A slightly delayed tube in a stable child beats an immediate tube in an arresting one.

✅ Checklist culture: why the boring card saves lives

Under stress your working memory collapses. You will forget something. The checklist is not bureaucracy. It is a cognitive offload. 🧠

The pre-intubation checklist covers, at minimum:

• 🟩 Team: roles assigned. Who intubates, who pushes drugs, who does external laryngeal manipulation, who times and watches the sats.

• 🟩 Airway kit: correct tube size plus one above and one below, working laryngoscope (check the light), bougie or stylet, suction on and reachable, working bag-valve-mask.

• 🟩 Suction, oxygen, monitoring: waveform capnography attached and working before the attempt. No trace, no confirmation.

• 🟩 Drugs: induction and paralytic drawn, labelled, dosed to weight. Push-dose pressor ready if fragile.

• 🟩 Plan A, B, C: say the failed-airway plan out loud before you start. Everyone hears it.

• 🟩 Access: at least one reliable IV or IO, patent and flushing.

🗣️ A verbalised, shared checklist beats a silently ticked one. The point is shared situational awareness, not the paperwork.

🎯 The bottom line

The tube is the easy part. The optimisation is the medicine. 💊

Before you induce any sick child:

1. 🫧 Oxygenate hard, then keep oxygen flowing through the apnoea.

2. 🧍 Position for the age. Fix the geometry.

3. 🩸 Resuscitate the circulation. Pressor and fluid ready, not hoped for.

4. ✅ Run the checklist out loud, as a team.

Do these four and the peri-intubation arrest largely designs itself out.

Intubation is not a treatment - it’s is a trade. You trade a failing or threatened airway and inadequate gas exchange for a controlled one and in return you accept the risks of laryngoscopy, positive pressure, sedation, haemodynamic collapse on induction and everything else that follows downstream in the PICU.

The decision is therefore never “can I intubate this child” but “does the benefit of taking the airway now exceed the cost of doing so, and the cost of waiting.”

Junior doctors get this wrong in two directions.
They intubate children who would have turned the corner on non-invasive support and they delay in children who were always going to need a tube, losing the controlled window and ending up doing it as a crash procedure in a peri-arrest state.

Both errors come from treating intubation as a numbers decision driven by a single blood gas or saturation reading. It is not. It is a trajectory decision.

🧭 This post gives you the frameworks to make that call.

🩺 The two questions (in oder)

Every intubation decision collapses into two sequential questions. Ask them in this order. Do not skip the first. 👆

🚨 Question 1: Is this child too sick to NOT intubate?

This is the override question. If the answer is yes, you stop deliberating and you secure the airway. There is no trial of non-invasive support, no further blood gas, no waiting for the registrar. These are the absolute indications, and they are mostly clinical, not biochemical.

🤔 Question 2: If not absolutely indicated, can we avoid it?

This is the harder, more common conversation. Most children in the resus bay are not in category one. They are working hard, they are tiring, their numbers are drifting, and you have a genuine choice. This is where trajectory, reversibility, and the failure of a non-invasive trial decide the outcome.

Figure 2.1: The two-question intubation spine

🚨 Question 1: The absolute indications

These are the situations where deliberation itself is the error. Commit them to memory as clinical patterns, not numbers.

🔥 Failure or imminent loss of airway patency.
A child who cannot protect or maintain their own airway needs it secured regardless of gas exchange.

The relevant question is whether the airway will still be patent in ten minutes, not whether it is patent now. A partially obstructed airway that you can still intubate is a far better proposition than a completely obstructed one that you cannot.

💤 Apnoea or inadequate respiratory drive.
Central depression from raised intracranial pressure, status epilepticus, poisoning or primary CNS pathology.

If the child is not breathing adequately and the cause is not instantly reversible (opioid reversal, glucose, a brief postictal pause that resolves), you ventilate.

🫁 Failure of oxygenation or ventilation despite maximal non-invasive support. This is not a single low saturation.

It is persistent, refractory hypoxaemia or hypercapnic acidosis that has not responded to escalating non-invasive measures.

The trajectory and the response to treatment matter more than the absolute value.

🧠 The need to control physiology that cannot be controlled awake.

Severe traumatic brain injury requiring tight CO2 control, the child requiring deep sedation for an emergent procedure or imaging in an unstable state, status epilepticus requiring agents that abolish respiratory drive or significant burns requiring transfer where the airway will predictably deteriorate en route.

⏩ Anticipated clinical course.

This is the most under-weighted indication and the one juniors miss most. A child who is currently compensating but whose underlying disease will predictably worsen, especially before or during transfer, should be intubated in the controlled environment of the resus bay rather than in the back of a retrieval ambulance.

🤔 Question 2: Can we avoid it?

Most of your decisions live here. The child is working hard, the numbers are not catastrophic, and you have a real choice. Three things decide it: reversibility, trajectory, and the response to a time-bound trial.

🔁 Reversibility

Is the underlying problem something you can fix faster than the child will tire? Bronchospasm responds to bronchodilators and steroids over minutes to hours. Pulmonary oedema responds to diuresis and pressure offloading.

A supraventricular tachycardia driving respiratory distress responds to adenosine in seconds. If the cause is reversible on a timescale shorter than the child’s respiratory reserve, you have grounds to support rather than intubate.

The inverse is the trap.

A fixed or progressive disease process (a large effusion, worsening ARDS, an evolving neurological catastrophe) will not be talked out of failing by CPAP. Non-invasive support here only delays the inevitable and burns the controlled window.

📈 Trajectory beats snapshot

A single observation tells you almost nothing. Two observations separated by time tell you almost everything.

The child whose work of breathing, gas exchange, and conscious level are stable or improving over thirty to sixty minutes of treatment is winning. The child going the other way is losing, and the rate of change predicts how much time you have.

🧗 Paediatric reserve and the cliff

Children compensate well and then fail fast.
This is a recurring theme of this series, established physiologically in Part 1: high oxygen consumption, low functional residual capacity relative to closing capacity, and a chest wall that contributes little mechanical advantage.

The practical consequence for this decision is brutal.

The clinical signs of compensation (tachypnoea, recruitment of accessory muscles, tachycardia) can look stable right up until they are not, and the transition from “working hard but holding” to “arrested” is shorter than in an adult.

The dangerous signs are the ones that indicate compensation is ending: a falling respiratory rate in a child who has been tachypnoeic and is now tiring, a rising then falling conscious level, bradycardia, and silent chest.

🚩 Do not wait for them to appear.

Figure 2.2: The compensation cliff (why trajectory, not snapshot - decides what happens next)

⏱️ The time-bound trial

If you elect to support rather than intubate, you are running a trial. A trial without a stop rule is not a clinical plan, it is hope. Before you start non-invasive support, define three things out loud and write them down:

1. 🎯 What you expect to improve. Work of breathing, gas exchange, conscious level. Be specific.

2. ⏰ By when. A defined re-assessment point. Minutes, not “later.” For most acute respiratory failure this is a short interval, on the order of an hour or less, with continuous observation in between.

3. 🛑 What failure looks like. The threshold that converts the trial into an intubation. Worsening on any of your chosen parameters, no improvement by the re-assessment point, rising sedation requirement to tolerate the interface, or a falling conscious level.

The failure of a non-invasive trial is itself a recognised indication for intubation. The point of declaring the stop rule in advance is that it removes the sunk-cost reasoning that keeps tired teams persisting with a failing trial because they have already invested in it.

The child does not care how much CPAP you have already given.

💯 They care whether it is working.

A note on non-invasive support as a bridge versus a destination: it is reasonable as a bridge while a reversible cause is treated, and it is reasonable as a ceiling of care decided in advance.

It is dangerous when it drifts into being neither, used to avoid a decision rather than to enact one.

⚠️ The crash-intubation trap

💀 The worst intubation is the one you are forced into.

When you wait past the controlled window, you intubate a child who is hypoxic, acidotic, haemodynamically marginal and physiologically intolerant of the apnoea and the induction agents.

Desaturation is faster because the reserve is already gone.

Induction agents drop a blood pressure that was already being propped up by endogenous catecholamines.

The first attempt matters more because there is no buffer for a second.

Junior doctors instinctively frame early intubation as escalation and waiting as conservatism.

In a child who is going to need it, that framing is inverted.

Waiting is the high-risk choice.

⚠️ Acting early while you still have reserve, help, and a controlled environment is the conservative one.

Figure 2.3: Integrated bedside flow

🩺 What this looks like at the bedside

🚨 Override check first. Failing airway, absent drive, refractory to max support, or physiology you can’t control awake? → secure the airway now, stop reading gases.

🤔 Then the avoidability check. Not reversible faster than the child will tire? → intubate now while it’s controlled. Reversible? → run a time-bound trial with a target, deadline, and stated failure rule, and watch the trajectory, not a single number.

🧗 The paediatric truth underneath it all: the child looks like they’re coping until just before they aren’t. Decompensation signs arrive late, the cliff is steep, and the controlled window closes quietly. ⏳

🔑 The one reframe to keep: in a child who’s going to need a tube, early intubation is the conservative choice and waiting is the gamble. Harm comes from being forced to act too late, not from acting too early.

🔜 Coming next

Part 3 takes the decision you have just made and turns it into a safe procedure: rapid sequence induction in children.

Drug selection and weight-based dosing, the physiologically difficult paediatric airway, apnoeic oxygenation and pre-oxygenation that actually buys you time, and how induction agents behave in the shocked child.

The decision to intubate is only as good as the first attempt that follows it. 🎯

This series is educational and written for junior doctors building a structured understanding of paediatric ventilation. It does not replace local PICU, anaesthetic, and retrieval-service protocols, senior clinical judgement, or formal airway training.

Everything about managing a child’s airway and lungs is different.

Not slightly different. Fundamentally different.

The margins are thinner. The physiology is less forgiving. And the anatomy is working against you in ways that adult practice never prepares you for. 😤

This Spoon Fed PEM covers the four things that matter most at the bedside:

🔺 Airway geometry 🦴 Chest wall compliance 🔥 Oxygen consumption 💨 Functional residual capacity

Get these wrong and you are behind before you have even touched the ventilator.

Let’s break them down. ⬇️

1️⃣ The Airway Is a Funnel, Not a Cylinder 🔺

In adults, the narrowest point of the airway is the glottis (the vocal cords). Below the cords, the trachea is a roughly uniform cylinder. Simple enough.

The narrowest point is the cricoid ring, sitting below the cords. The airway tapers from the laryngeal inlet down to the cricoid like a funnel. 🫠

Why does this matter?

Two big reasons:

🔸 Reason 1: An ETT that passes the cords easily can still jam at the cricoid or create a pressure point there. The cords are not the final checkpoint in kids.

🔸 Reason 2: Even 1mm of mucosal oedema has a catastrophic effect on resistance.

🧮 The Maths That Should Scare You

In a 4mm infant airway, 1mm of circumferential swelling reduces the cross-sectional area by roughly 75% and increases resistance by a factor of 16.

That is Poiseuille’s law in action:

Resistance ∝ 1/r⁴

Resistance is inversely proportional to the fourth power of the radius.

A 1mm reduction is nothing in an adult. It is everything in an infant. 😳

🧠 Don’t Forget the Other Anatomical Curveballs

🎯 Clinical Pearl

You have almost no room for error with tube sizing in small children. Any condition causing airway swelling (croup, post-extubation oedema, burns) can become life-threatening fast. A half-size-too-large tube creates subglottic injury. A half-size-too-small tube creates an unacceptable leak. Size matters. A lot. 🫣

2️⃣ The Chest Wall Is Too Compliant 🦴

An adult rib cage is ossified and relatively rigid. It holds its shape under negative pleural pressure during inspiration. It does its job.

A neonate or infant rib cage? Cartilaginous. Highly compliant. And the ribs are oriented more horizontally. 😬

💡 What Actually Happens

When the diaphragm contracts and generates negative intrathoracic pressure, the compliant chest wall gets sucked inward instead of holding its shape.

This is chest wall retraction and it means the infant is wasting diaphragmatic work on deforming its own rib cage rather than inflating the lungs.

The infant is literally fighting its own chest wall to breathe. 😩

🪣 The Missing Bucket Handle

The diaphragm is also at a mechanical disadvantage. The ribs insert more horizontally, so the “bucket handle” effect that adults rely on for expanding thoracic volume is minimal in infants.

💪 The Diaphragm Fatigues Fast

This is the part that really gets you. ⚡

The infant diaphragm fatigues much faster. And once that diaphragm fatigues? There is very little accessory muscle reserve to compensate. The tank is empty. 🪫

🎯 Clinical Pearl

An infant in respiratory distress who is retracting heavily is telling you their system is failing at the mechanical level. They cannot sustain that work. Waiting for blood gas deterioration means you have waited too long. By the time the gas tells you there is a problem, the child has been losing the fight for a while. Act on clinical signs. Not numbers. 🚨

3️⃣ Oxygen Consumption Is Double 🔥

Here is a number that should change how you behave during every paediatric intubation:

That is a two-fold difference per kilogram. The engine is running hotter in kids. All the time. 🏎️💨

⏱️ The Brutal Consequence

The time from apnoea to desaturation is dramatically shorter.

90 seconds.

That is your entire window. From the moment ventilation stops to the moment the sats start crashing. There is no plateau. There is no grace period. 📉

The metabolic rate is higher. CO₂ production is higher. The ventilatory demands relative to body weight are greater. The system is running at near-maximum capacity even at baseline.

🎯 Clinical Pearl

Paediatric intubation cannot be treated as a leisurely procedure. You preoxygenate. You have your plan, your backup plan and your equipment ready. There is no time for troubleshooting mid-attempt. First-pass success is not a flex. It is a physiological necessity. 🎯

4️⃣ FRC Is Dangerously Close to Closing Capacity 💨

This is the one that catches people out. It is the most conceptually tricky of the four but it might be the most clinically important.

📖 Definitions First

Functional Residual Capacity (FRC) = the volume of gas remaining in the lungs at the end of a normal passive exhalation. It is your oxygen reserve. It is also what keeps alveoli from collapsing between breaths.

Closing Capacity (CC) = the lung volume at which small airways begin to close during exhalation.

🧩 The Critical Relationship

In adults: CC sits well below FRC during tidal breathing. Airways remain open throughout the respiratory cycle. Comfortable margin. ✅

In neonates and infants: CC is at or above FRC. 😱

This means that during normal tidal breathing, small airways are closing and alveoli are collapsing with every single breath.

The infant is living on the edge of atelectasis at baseline. 🪨

Now add anything that further reduces FRC:

• 🛏️ Supine positioning

• 🫃 Abdominal distension

• 💊 General anaesthesia

• 💉 Muscle relaxation

• 🌬️ Positive pressure ventilation with inadequate PEEP

...and you push the patient into frank airway closure and shunt.

🎛️ This Is Why PEEP Matters So Much

PEEP is not optional in paediatric ventilation. It is the thing keeping small airways open.

But it is a tightrope:

🎯 Clinical Pearl

PEEP is non-negotiable. Recruitment is an active process. Anything that reduces FRC (positioning, abdominal pathology, anaesthesia) will cause immediate clinical deterioration. Set it. Respect it. Monitor it. Don’t wing it. 🎚️

🧵 Pulling It All Together: A Cheat Sheet

These four physiological facts drive almost every decision in paediatric ventilation. Here is what each one is telling you at the bedside:

🔺 Airway Geometry Says...

• Tube sizing is critical

• Cuffed tubes need careful pressure monitoring

• Post-extubation stridor is a real and predictable threat

• A half-size-too-large tube → subglottic injury

• A half-size-too-small tube → unacceptable leak

🦴 Chest Wall Compliance Says...

• The infant cannot generate or sustain the work of breathing you might tolerate in an adult

• Early intervention matters

• Waiting for exhaustion is not a strategy

🔥 Oxygen Consumption Says...

• Apnoeic windows are measured in seconds, not minutes

• Preoxygenation buys you less time than you think

• Every failed intubation attempt carries a higher cost

💨 FRC Physiology Says...

• PEEP is non-negotiable

• Recruitment is an active process

• Anything that reduces FRC will cause immediate clinical deterioration

⚡ The Bottom Line

Paediatric patients have less reserve, less margin and less time.

If you approach them with the same timeline and the same thresholds you use in adult practice, you will be too late.

Every second counts. Every millimetre counts. Every cmH₂O of PEEP counts. 🫁

🔜 Next in the Series

Part 2: Modes of Ventilation in Paediatrics

Pressure control vs volume control, when SIMV still has a role and why HFOV exists.

Stay tuned. 👀

Spoon Fed PEM: bite-sized, high-yield paediatric emergency medicine education.

If you found this useful, share it with a colleague who needs it. 🩺💛

You’re halfway through a packed night shift. A 9-year-old with a history of oppositional defiant disorder arrives by ambulance with police escort. He’s biting, spitting, throwing things. Staff are scared. Parents are crying. The waiting room is watching.

What do you do?

📈 The Elephant in the Resus Bay: This Problem is Getting Bigger

Paediatric mental and behavioural health (MBH) visits to the emergency department have more than doubled between 2009 and 2019. Up to 12% of children seen in the ED for mental health complaints receive intramuscular medication for acute agitation, and 5 to 10% are physically restrained.

In 2021, the American Academy of Pediatrics (AAP), the American Academy of Child and Adolescent Psychiatry (AACAP) and the Children’s Hospital Association jointly declared a national emergency in child and adolescent mental health.

😮 Over 79 paediatric deaths have been attributed to restraint-related events in the United States alone.

This is not a niche topic. And we need to get it right.

🧩 Tier 0: Understand WHY the Child Is Agitated

Before reaching for midazolam, take a breath. 🫁

The AAEP (American Association for Emergency Psychiatry) BETA Consensus Statement is crystal clear: agitation is a symptom, not a diagnosis. Just like pain points to a source, agitation points to an underlying driver.

Common aetiologies in children include:

• 🧠 Psychiatric: Psychosis, mania, severe anxiety, PTSD, oppositional defiant disorder (ODD), conduct disorder

• 🧪 Toxic: Substance intoxication or withdrawal (yes, even in kids)

• 🤕 Medical/Organic: Hypoglycaemia, head injury, sepsis, hypoxia, pain, urinary retention

• ♿ Neurodevelopmental: Autism spectrum disorder (ASD), attention deficit hyperactivity disorder (ADHD), intellectual disability with sensory overwhelm

• 💔 Situational: Fear, loss of control, unfamiliar environment, separation from carers

🟢 Tier 1: Environmental Modification (Your First Move)

Before anyone says a single word, change the environment. The ED itself can be a massive agitation trigger.

What the Evidence Says 📚

The 2024 ACEP Pediatric Emergency Medicine Committee review by Saidinejad et al. specifically highlights how the ED environment worsens agitation:

• 💡 Intense, harsh lighting

• 📢 Loud, chaotic noise

• 🌡️ Poorly regulated temperature

• 🚫 Lack of familiar objects

• 👥 Crowding and constant interruptions from staff

• 🍽️ No access to food or drink

• 😣 Physical discomfort (restraints, IV lines, monitors)

Your Environmental Checklist ✅

🟡 TIER 2: Verbal De-escalation (Your Most Powerful Tool) 🗣️

This is the cornerstone of agitation management. The AAEP Project BETA De-escalation Workgroup describes a 3-step approach:

The 3-Step Framework

Step 1: Engage 🤝

• Approach calmly, at eye level

• Introduce yourself

• Use a soft, slow, low-pitched voice

• Maintain a non-threatening posture (hands visible, no crossed arms)

• Respect personal space (at least one arm’s length)

Step 2: Collaborate 🧩

• Ask what’s bothering them

• Listen actively without interrupting

• Validate their feelings (”I can see you’re really upset”)

• Offer choices (”Would you like to sit on the bed or in the chair?”)

• Include the child in decision-making

Step 3: De-escalate 🌊

• Set clear, simple limits (”I need you to stop throwing things so we can help you”)

• Offer concrete alternatives (”You can squeeze this stress ball instead”)

• Negotiate achievable goals (”If you can sit down for 5 minutes, I’ll get you that juice”)

• Praise any positive behaviour immediately

What Does This Look Like in Real Life? 🎬

Imagine a 12-year-old girl with known PTSD, brought in after a self-harm episode, now screaming and pulling at her IV line.

❌ Wrong approach: “Stop that! You need to calm down! Security!”

✅ Right approach: Kneel to her level. Soft voice. “Hey. I’m Dr. K. I can see this is really scary for you. Can you tell me what’s making things worse right now? ... The IV is hurting? OK. Let me see if we can take a break from that. Would you like some water while we figure this out together?”

📌 Clinical Pearl: A case report by Croce and Mantovani (2017) demonstrated how a consistent, dedicated provider using verbal and behavioural de-escalation prevented the need for physical restraint in a child who had been restrained multiple times in prior ED visits.

⚠️ TIER 2.5: Agitation Scoring and Team Readiness

Before escalation spirals, quantify the risk and rally your team.

Personalised De-escalation Plans 📋

One of the most impactful quality improvement strategies from Hoffmann et al. (2023) was the introduction of personalised de-escalation plans for high-risk patients. These plans document:

• Known triggers (noise? touch? specific people?)

• Successful calming strategies from prior visits

• Preferred medications and routes

• Parent/carer input

This information should be recorded, communicated during handovers, and updated with each visit.

🔴 TIER 3: Chemical Restraint (When All Else Has Failed)

Let’s be absolutely clear. 🛑

Chemical (pharmacological) restraint should only be considered after environmental modification and verbal de-escalation have been genuinely attempted and have failed. This is not optional. This is the standard of care.

The AAEP BETA Consensus Principles:

1. Agitation management should be individualised, multidisciplinary and collaborative

2. The aetiology of agitation should drive the choice of treatment

3. Medication should be ONE part of a comprehensive strategy

4. Consider an extra dose of a home medication first (it’s familiar, better tolerated, fewer drug interactions)

5. Avoid over-sedation (a child who cannot participate in their own assessment cannot be properly evaluated)

🔴🔴 TIER 4: Physical Restraint (The Absolute Last Resort)

Physical restraint in a child is traumatic. For the child. For the parents. For the staff. And it can be lethal.

Key Principles from the Evidence 📚

• Physical restraint should be used only when there is imminent danger of harm to the patient or others, and all other strategies have been exhausted

• It should be applied using the least restrictive means for the shortest possible duration

• The patient must be continuously monitored during restraint (vital signs, airway, circulation to extremities, emotional state)

• Restraint is never therapeutic. It is a safety intervention only

Disparities Matter ⚖️

This is important and often overlooked. Evidence shows disparities in restraint use across racial and ethnic lines. Black and brown youth are more likely to be physically restrained compared to white youth, even after controlling for acuity. Training in trauma-informed, culturally responsive care and the use of standardised agitation pathways can help reduce these disparities.

Training Gaps Are Real 🎓

A study by Dorfman and Kastner (2004) found that a large percentage of emergency medicine and paediatric emergency medicine residency programmes do not teach trainees about the appropriate application of restraints or the clinical situations warranting their use. This has to change.

👨‍👩‍👧 Talking to Parents: The Conversation That Matters Most

This is the part many clinicians dread. But done well, it transforms the encounter.

Before Escalation 🕐

• Introduce yourself and your role clearly

• Explain the situation in plain language: “Your child is very distressed right now and is at risk of hurting themselves or others. We are going to try everything we can to help them calm down safely.”

• Involve parents as partners: Ask about triggers, calming strategies, medications

• Set expectations: “If the de-escalation strategies don’t work, we may need to use medication to help them feel calmer and safer. I will keep you informed every step of the way.”

During Restraint (Physical or Chemical) 💬

• Narrate what is happening and why: “We’re giving your child a medication that will help them relax. This is to keep them safe.”

• Validate their distress: “I know this is incredibly hard to watch. You’re not failing your child. This is a medical situation and we are here to help.”

• Offer presence or privacy: Some parents want to stay; some need to step out. Both are OK. Ask them.

• Designate a team member to stay with the family and answer questions in real time

After the Event 🌅

• Debrief with the family: Explain what happened, what medications were given, and what the plan is going forward

• Screen for caregiver distress: Parents who witness their child being restrained can develop secondary traumatic stress

• Document the de-escalation plan for future visits

• Connect with outpatient resources: Mental health follow-up, crisis plans, school-based supports

  ---

🧠 Key Takeaways for the Busy ED Clinician

1. 🔍 Always look for a treatable cause before labelling behaviour as “psychiatric”

2. 🏠 Modify the environment first: Dim lights, reduce noise, offer comfort items and food

3. 🗣️ Invest in verbal de-escalation: This is your most powerful and safest tool

4. 📊 Use agitation scores and personalised de-escalation plans to standardise care

5. 💊 When medication is needed, choose based on aetiology: Prioritise home medications, use the least sedating effective option, and never combine olanzapine with benzodiazepines

6. 🛑 Physical restraint is the absolute last resort: Monitor continuously, use the least restrictive means, and aim for the shortest duration

7. 👨‍👩‍👧 Communicate with parents proactively, honestly and compassionately: They are partners in care, not bystanders

8. 📝 Document everything: Triggers, strategies tried, medications given, and plans for future visits

9. ⚖️ Be aware of disparities: Standardised pathways help ensure equitable care

10. 💪 Debrief your team: This work is emotionally taxing. Staff wellbeing matters too

Enjoyed this post? 🙌 Share it with your ED team. Bookmark it for your next shift.
And remember: the best restraint is the one you never need to use.

Stay safe out there. 💛🩺

TLDR;

A bump on the back or side of a young baby’s head (under 3 months old) is the single biggest red flag for a skull injury. A bump on the forehead is less concerning.

This was confirmed again in a small Saudi study of 77 babies. Most fell less than a metre onto hard floors.

The standard decision tool most emergency departments use didn’t pick up the at-risk babies in this age group. The location of the bump on the head did.

Only 23 babies were scanned so the numbers are too small to draw firm conclusions. Doctors scanned the ones they were already worried about, which skews the results.

The good news: serious outcomes were rare. Only 2.6% had a bad clinical outcome. No baby needed brain surgery. No deaths.

A bump on the back or side of the head in a very young baby should either get a scan or get watched closely for at least 4 to 6 hours.

Finding a skull fracture on a scan doesn’t automatically mean the baby is in danger. Most went home fine. But you’ve now added radiation to a tiny brain and a lot of worry for the parents.

Falls from bed height onto tile or marble (common in Malaysian homes) are enough to fracture a very young baby’s skull. Don’t dismiss them as trivial.

🍼 The Clinical Setup

A 6-week-old baby lands in your Paediatric Emergency Department. Rolled off the bed. Maybe 60 centimetres onto a tile floor. There is a parietal scalp swelling. GCS is 15. No vomiting. Fontanelle soft. Baby looks well.

You pull up PECARN in your head. No high-risk features. Non-frontal scalp haematoma sits in the intermediate-risk box. PECARN says: observe versus CT, shared decision-making.

But wait. ✋

Your gut says the rule does not quite fit here. And a new paper from Saudi Arabia is trying to answer the question you have been asking yourself every time one of these tiny humans rolls through your doors.

📄 The Paper at a Glance

Alghannam et al. ran a retrospective cohort study at King Fahad Medical City in Riyadh. They looked at all infants under 3 months presenting within 24 hours of minor head trauma over 4.5 years (January 2021 to May 2025).

🔑 Key details:

• Excluded suspected non-accidental trauma and anyone with GCS below 14

• Ended up with 77 infants (their target sample size was 384... they got 20% of what they needed 😬)

• Primary outcome: Radiologic TBI (any acute intracranial abnormality or skull fracture on CT)

• Secondary outcome: Clinically important TBI (ciTBI): death, neurosurgery, intubation beyond 24 hours, or hospital admission of 2 nights or more

⚠️ These are two very different things. PECARN was designed to predict ciTBI. This paper’s primary outcome casts a much wider net, including isolated skull fractures that may not change a single thing about management.

🔢 The Numbers That Matter

Of 77 infants, only 23 (30%) got a CT. This was at clinician discretion, not protocolised. The other 54 were observed and discharged without imaging.

Of the 23 scanned:

🩻 18 (78%) had radiologic TBI. Mostly skull fractures with or without subgaleal haematomas.

That headline number looks alarming... until you realise that clinicians cherry-picked the ones they were most worried about. This is textbook verification bias. 📊

The clinical outcomes?

• ✅ ciTBI occurred in only 2 out of 77 (2.6%)

• ✅ No infant needed neurosurgery

• ✅ No infant went to ICU

• ✅ No deaths

• ✅ Even among the 18 with CT-confirmed findings, two-thirds were sent home from the ED

The fractures were there on imaging. But clinically? Most babies did fine.

📉 The Fall Height Reality Check

Falls accounted for 80.5% of injuries. Here is the part that should make you pause:

🏠 82% of those falls were from below 0.9 metres. That is below the PECARN threshold for “severe mechanism.”

Yet 78% of radiologic TBI cases followed these low-height falls.

For those of us practising in Southeast Asia where our homes have hard tile, marble or ceramic floors... this literally hits different 🏡

⭐ THE STAR FINDING: Non-Frontal Scalp Haematoma

This is the finding the paper hangs on. And it is the one that actually matters. 🎯

Among the 18 CT-positive infants:

• 15 (83%) had a non-frontal scalp haematoma (parietal, temporal or occipital)

Among the 5 CT-negative infants:

• ZERO had one 🚫

Complete separation. Every single infant with a non-frontal scalp haematoma who was scanned had something on CT.

📊 The odds ratio was 50 (95% CI: 2 to 1179). That confidence interval is comically wide. The true effect could be modest or enormous. With only 23 imaged infants and 5 negatives, you cannot pin it down more precisely.

But the direction is rock-solid. This is not a new finding. It keeps being confirmed:

Across every study that has looked at this age group, non-frontal scalp haematoma keeps showing up as the strongest clinical predictor of radiologic TBI. 🏆

🦴 But WHY Non-Frontal Specifically?

The anatomy tells the story:

🔵 Frontal bone = thicker, more resilient, better protected

🔴 Parietal and temporal bones = thinner, less overlying tissue, more vulnerable to fracture

🔴🔴 Temporal bone sits over the middle meningeal artery, so temporal fractures carry a higher risk of epidural haemorrhage 🩸

A haematoma over the back or side of a young infant’s head is a completely different animal from a forehead bump.

🚨 The PECARN Problem in This Age Group

59 of 77 infants (77%) were classified as PECARN low-risk. Of these, 18 were scanned and 16 came back positive.

The authors report this as 27.1% radiologic TBI in the low-risk group.

PECARN classified them as low-risk because non-frontal scalp haematoma alone is an intermediate feature that prompts observation rather than mandatory CT. In older infants, that is reasonable. In a 6-week-old, this paper suggests the calculus is different. ⚖️

Neither age nor PECARN risk classification was significantly associated with CT findings (p>0.05). The formal risk stratification tool did not discriminate. The scalp exam did. 🔍

🔬 Honest Critique: What Is Wrong With This Paper

Credit where it is due: the authors were unusually transparent about their limitations. They admitted underpowering, explained why their regression model failed and reported exact confidence intervals rather than hiding behind p-values. That honesty is worth acknowledging. 👏

But the problems are real:

🔴 Sample size. 77 infants. 23 scanned. 5 CT-negative. The statistical analysis rests on a denominator of 5. One additional CT-negative infant with non-frontal scalp haematoma would demolish the OR of 50. This generates hypotheses, not reliable effect estimates.

🔴 Verification bias. Only 30% were imaged. The decision to scan was driven by the same clinical features being tested as predictors. Non-frontal scalp haematoma appears to predict TBI partly because clinicians who saw it were more likely to order the CT that would find the fracture. The association is real, but the magnitude is inflated.

🔴 Post-hoc power analysis. The authors report achieved power of 0.995. Post-hoc power is a function of the p-value. If your p is below 0.05, your post-hoc power will always look good. It tells you nothing you did not already know from the CI. This should not appear in modern papers. 📝

🔴 Single centre, retrospective, almost entirely one nationality. King Fahad Medical City is a tertiary referral centre in Riyadh, 98.7% Saudi nationals. The population, the referral patterns and the clinical decision-making may not generalise.

🇲🇾 What This Means For Malaysian Practice

The injury patterns in this paper are familiar to anyone working in a Malaysian Paediatric Emergency Department. Babies falling off beds onto tile floors, rolling off adult laps, tumbling from low surfaces. Our homes have hard floors, not carpet. A 60-centimetre fall onto marble or ceramic is our bread and butter. 🍞

PECARN is the decision tool most of us use. It has never been validated in a Malaysian population. There is no local clinical decision rule for paediatric head trauma.

🎯 Practical Takeaways for the ED

1️⃣ Document Scalp Haematoma Location (Every. Single. Time.)

Writing “scalp swelling” or “haematoma present” is not enough. ❌

Frontal versus parietal versus temporal versus occipital MATTERS.

It changes the PECARN risk category and, based on this and prior data, it should change your imaging threshold. Make this a habit. Make it a department standard.

2️⃣ Non-Frontal Scalp Haematoma + Under 3 Months = Take It Seriously

It does not mandate CT in every case. But it means you need a clear plan:

✅ Either scan

✅ Or observe closely with serial assessments and documented safety-netting

✅ Observation period: 4 to 6 hours minimum ⏰

✅ Discharge advice in a language the family actually understands 🗣️

3️⃣ Finding a Fracture ≠ Finding a Bad Outcome

Two-thirds of CT-positive infants in this study went home. ciTBI was 2.6%. No neurosurgery.

If you scan and find a linear skull fracture in a well-appearing baby, you have not necessarily changed management. You have changed:

• 😟 The parents’ anxiety level

• 📋 Your documentation burden

• ☢️ The radiation dose delivered to a tiny developing brain

Whether that trade-off is worth it is a question worth discussing in your department.

4️⃣ POCUS May Have a Role 🔊

For centres where CT is down the corridor, it may not change the workflow much. But for colleagues in district hospitalswhere CT means a transfer, POCUS screening for skull fractures could be a useful triage adjunct:

• ✅ See a fracture line on ultrasound → tips the balance toward transfer and CT

• ✅ Do not see one → observe with more confidence

5️⃣ Low-Height Falls Are NOT Low-Risk in This Age Group ⚡

The PECARN threshold for severe mechanism is 0.9 metres. In an infant under 3 months with a thin, incompletely ossified skull, a 60-centimetre fall onto a hard surface can produce a fracture.

Our Malaysian floors are unforgiving. Factor that into your assessment. 🏠

🎯 The Bottom Line

Non-frontal scalp haematoma is the strongest clinical predictor of radiologic TBI in infants under 3 months. This is not a new finding. It is a finding that keeps being confirmed across multiple studies, multiple countries and multiple study designs.

It should be the primary examination finding driving your imaging decision in this age group. 🧭

PECARN remains the best framework we have. But it was not built for neonates and it shows. In the youngest infants, the intermediate-risk features deserve more weight than the algorithm formally assigns them.

ciTBI is uncommon. Clinical outcomes are generally good even when you find things on CT. The goal is not to CT every infant with a bump on the head. The goal is to identify the rare serious injury without irradiating the many who do not need it. ⚖️

The honest answer? We do not have a validated tool for this specific age group. Until we do, the combination of:

🔹 PECARN as a starting framework
🔹 A focused scalp examination with attention to haematoma location
|🔹 A careful mechanism history
🔹 Good clinical judgment

...is the best we can offer.

📚 Reference

Alghannam I, Alrashidi A, Alghamdi M, et al. Outcomes and risk factors of traumatic brain injury in infants under 3 months with minor head trauma: a retrospective study from a tertiary care center in Riyadh, Saudi Arabia. BMC Emergency Medicine. 2025;25:255. doi:10.1186/s12873-025-01415-5

Supporting literature:

• Kuppermann N, Holmes JF, Dayan PS, et al. Identification of children at very low risk of clinically-important brain injuries after head trauma: a prospective cohort study. Lancet. 2009;374:1160-1170.

• Dayan PS, Holmes JF, Atabaki S, et al. Association of traumatic brain injuries with vomiting in children with blunt head trauma. Lancet Child & Adolescent Health. 2022.

• Alonso-Cadenas JA, et al. Clinical predictors of intracranial injury in infants younger than 3 months. European Journal of Pediatrics. 2023.

On April 1, 2026, the MV Hondius set sail from Ushuaia, Argentina, carrying 88 passengers and 59 crew on a polar expedition through Antarctica and the remote islands of the South Atlantic. It was supposed to be the trip of a lifetime.

Five days out, a 70-year-old male passenger developed fever, headache and mild diarrhoea. Then respiratory distress. Then he was dead by April 11. His body was removed when the ship docked at St. Helena. His wife, 69, became symptomatic days later and was evacuated to South Africa, where she too died.

By May 4, 2026, the WHO confirmed seven cases including three deaths. The ship, now stranded off the coast of Cape Verde with 150 people aboard, was refused permission to dock. Passengers were confined to cabins. Health officials in full PPE were photographed leaving the vessel. Three critically ill patients were airlifted to the Netherlands and Germany on May 5.

On May 6, 2026, the strain was confirmed: Andes virus, a hantavirus capable of rare human-to-human transmission through close contact.

The MV Hondius expedition cruise ship at sea

🎬 ICYMI 2: It Killed Gene Hackman’s Wife a Year Ago

In February 2025, Betsy Arakawa, the 65-year-old wife of Oscar-winning actor Gene Hackman, died from hantavirus pulmonary syndrome (HPS) at their home in Santa Fe. Hackman, 95 and living with advanced Alzheimer’s disease, died approximately one week later from cardiovascular disease. Their bodies and one of their three dogs were found together by maintenance workers two weeks after Arakawa’s death.

The investigation revealed Arakawa was last seen alive at a CVS pharmacy on February 11, 2025. She called a physician the next day but never made it to the appointment. She died at home from HPS. Rodent entry points were found on the property.

This case thrust hantavirus into the global spotlight.
And now, just over a year later, it is headlining international news again on the MV Hondius.

🦠 WHAT IS HANTAVIRUS?

Hantaviruses belong to the family Hantaviridae, genus Orthohantavirus. They are enveloped, single-stranded RNA viruses named after the Hantaan River in South Korea, where the first pathogenic strain was identified during the Korean War in the 1950s.

There are two major clinical syndromes:

A Recent Discovery

HPS was first described in 1993 during the infamous Four Corners outbreak in the American Southwest, when a cluster of previously healthy young Navajo people developed fatal respiratory failure.
The CDC identified a novel hantavirus (later named Sin Nombre virus, literally “the virus with no name”) carried by deer mice.

Transmission electron micrograph of Sin Nombre hantavirus]

🐀 HOW DO YOU CATCH IT?

The natural reservoir is rodents. Different hantaviruses are carried by different rodent species:

🐭 Deer mouse (Peromyscus maniculatus) → Sin Nombre virus (North America)
🐭 Pygmy rice rat (Oligoryzomys longicaudatus) → Andes virus (South America)
🐭 Striped field mouse (Apodemus agrarius) → Hantaan virus (Asia)
🐭 Bank vole (Clethrionomys glareolus) → Puumala virus (Europe)

Routes of transmission:

🔸 Inhalation of aerosolised rodent urine, faeces or saliva (the BIG one)
🔸 Direct contact with rodent excreta through broken skin or mucous membranes
🔸 Rodent bites (rare)
🔸 Ingestion of contaminated food or water

Can it spread person-to-person?

For most hantaviruses: NO.

The major exception is Andes virus, the strain now confirmed on the MV Hondius. Andes virus has documented (though rare) person-to-person transmission through close contact. This is what makes the current cruise ship outbreak so alarming. The WHO confirmed on May 6, 2026 that human-to-human spread likely occurred among cabin-mates and spouses aboard the ship.

🏥 CLINICAL FEATURES: WHAT DOES HPS LOOK LIKE?

Yikes. This is where it gets ugly.

⏱️ Incubation Period

📋 Phase 1: Prodrome (3 to 5 days)

Looks like everything else in the ED. That is the problem.

🔹 Fever (often high, >38.5°C)
🔹 Myalgia (especially large muscle groups: thighs, hips, back)
🔹 Headache
🔹 Nausea, vomiting and diarrhoea
🔹 Malaise and fatigue
🔹 Dizziness

❌ What you WON’T see: Coryza, sore throat, rash.
This is NOT a typical upper respiratory infection.

📋 Phase 2: Cardiopulmonary Phase (rapid onset, 24 to 48 hours)

This is where patients crash. Hard. Fast.

🔴 Sudden onset of cough and dyspnoea
🔴 Non-cardiogenic pulmonary oedema (the lungs flood)
🔴 Bilateral infiltrates on CXR
🔴 Tachycardia progressing to arrhythmias
🔴 Hypotension progressing to cardiogenic shock
🔴 ARDS

The pathophysiology is vascular. Hantavirus infects pulmonary capillary endothelium, causing massive microvascular leak.

The lungs fill with fluid while the circulation loses volume. Patients drown from the inside while their blood pressure tanks.

📋 Phase 3: Recovery (days to weeks)

Survivors typically become polyuric during convalescence and improve rapidly. But getting to this phase is the hard part.

🧪 LAB FINDINGS: THE DIAGNOSTIC CLUES

Here is where the smart clinician can catch this before the patient spirals.

The classic HPS triad on bloods:

Other lab abnormalities:

🔸 Elevated LDH (often >1000 IU/L)
🔸 Elevated AST and ALT
🔸 Hypoalbuminaemia
🔸 Mildly elevated creatinine
🔸 Leukocytosis with left shift
🔸 Elevated PT (prothrombin time >14 seconds at admission was predictive of mortality in paediatric cases)

Definitive diagnosis:

🔬 Hantavirus-specific IgM (usually positive at symptom onset)
🔬 IgG antibodies (rise over time)
🔬 RT-PCR for hantavirus RNA
🔬 Immunohistochemistry on tissue
🔬 ELISA and immunofluorescent assays

👶 HANTAVIRUS IN CHILDREN: WHAT PAEDIATRICIANS NEED TO KNOW

This is the bit most people miss. Hantavirus in kids is rare but man is it deadly.

The numbers:

From the largest US study (1993 to 2018, published in Pediatrics 2023):

🔹 Of 719 HPS patients, only 22 (3.0%) were aged 0 to 12 years and 47 (6.5%) were aged 13 to 18
🔹 Less than 10 to 15% of all cases are diagnosed in children
🔹 But the case fatality rate in children is 32 to 34%
🔹 Children with HPS die more quickly than adults
🔹 Elevated haematocrit and creatinine are specifically associated with death in children

Paediatric clinical features (from the Sin Nombre paediatric case series):

In 13 patients aged 10 to 16 years:

🔸 100% had fever, headache and cough/dyspnoea
🔸 90% had nausea or vomiting
🔸 80% had myalgia
🔸 67% had tachypnoea at admission
🔸 33% had hypotension at admission
🔸 92% developed full HPS
🔸 67% of HPS patients required mechanical ventilation
🔸 ECMO was used in 2 patients (1 survived)

Why is it underdiagnosed in children?

🤔 The prodrome looks like every other febrile illness in a child
🤔 Clinicians simply don’t think of it
🤔 Low overall incidence means low index of suspicion
🤔 Many cases likely occur at home in rural settings and are never tested
🤔 Family exposure clusters may be missed if only the adult is diagnosed

💊 TREATMENT: THERE ISN’T ONE (SORT OF)

Intravenous ribavirin was tested for HPS. It did not work.

Treatment is aggressive supportive care, ideally in an ICU:

The management approach:

🔹 Early recognition (this is the single biggest factor in survival)
🔹 Judicious fluid resuscitation (tricky because of capillary leak; overzealous fluids worsen pulmonary oedema)
🔹 Vasopressors for shock
🔹 Mechanical ventilation for respiratory failure
🔹 ECMO for refractory cardiopulmonary failure
🔹 Broad-spectrum antibiotics until diagnosis is confirmed (because the differential includes bacterial sepsis)
🔹 Avoid overhydration (the lungs are already leaking; more fluid = more oedema)

📸 IMAGING

Chest X-ray findings in HPS:

🔸 Bilateral interstitial infiltrates progressing to alveolar infiltrates
🔸 Pulmonary oedema (non-cardiogenic pattern)
🔸 Pleural effusions (common)
🔸 Rapid progression over hours
🔸 Normal heart size (distinguishing it from cardiogenic pulmonary oedema)

The CXR can go from near-normal to complete whiteout in 24 hours.

🌍 GLOBAL DISTRIBUTION

Hantaviruses are found worldwide, but different strains dominate different regions:

Global distribution map of hantaviruses

What about Southeast Asia and Malaysia? 🇲🇾

Seoul virus (carried by brown rats, Rattus norvegicus) is present globally including in Southeast Asia.
While HPS is not typically reported in this region, HFRS cases and seroprevalence studies suggest that hantavirus circulation occurs.

The differential includes:

🔸 Influenza / COVID-19 / RSV
🔸 Bacterial pneumonia / sepsis
🔸 Leptospirosis
🔸 Scrub typhus
🔸 Severe Dengue
🔸 Melioidosis (in our Southeast Asian context)
🔸 Plague (pneumonic)

📚 REFERENCES AND FURTHER READING

1. WHO Disease Outbreak News. Hantavirus cluster linked to cruise ship travel, Multi-country. 4 May 2026. https://www.who.int/emergencies/disease-outbreak-news/item/2026-DON599

2. CDC. Clinician Brief: Hantavirus Pulmonary Syndrome (HPS). Updated March 2025. https://www.cdc.gov/hantavirus/hcp/clinical-overview/hps.html

3. Thorp L, Fullerton L, Whitesell A, Dehority W. Hantavirus Pulmonary Syndrome: 1993 to 2018. Pediatrics. 2023;151(4):e2022059352.

4. Ramos MM, Overturf GD, Crowley MR, et al. Infection with Sin Nombre hantavirus: clinical presentation and outcome in children and adolescents. Pediatrics. 2001;108(2):E27.

5. Pinto Junior VL, et al. Hantavirus pulmonary syndrome in children: case report and case series from an endemic area of Brazil. Rev Inst Med Trop Sao Paulo. 2019;61:e65.

6. Jonsson CB, Figueiredo LT, Vapalahti O. A global perspective on hantavirus ecology, epidemiology, and disease. Clin Microbiol Rev. 2010;23(2):412-441.

7. StatPearls. Hantavirus Pulmonary Syndrome. Updated April 2024. https://www.ncbi.nlm.nih.gov/books/NBK513243/

8. Duchin JS, et al. Hantavirus Pulmonary Syndrome: A Clinical Description of 17 Patients with a Newly Recognized Disease. NEJM. 1994;330(14):949-955.

Written for clinicians, trainees and anyone who wants to understand the news better.

🧩 1. What Even IS Paediatric Sepsis Now? Hello, Phoenix.

The old SIRS-based definition is retired. The 2024 Phoenix Sepsis Definitions Taskforce gave us a data-driven, organ-dysfunction-centred framework and the 2026 SSC guidelines are built on it.

📌 Phoenix Sepsis Definition (2024)

Sepsis = Suspected infection + Phoenix Sepsis Score ≥ 2 points, indicating life-threatening organ dysfunction in the respiratory, cardiovascular, coagulation or neurological systems

Septic Shock = Sepsis + ≥ 1 cardiovascular dysfunction point: severe age-appropriate hypotension, venous or arterial lactate > 5 mmol/L, or need for vasoactive medication

The Phoenix score covers 4 organ systems:

• 🫁 Respiratory — SpO2/FiO2 ratio, ventilator support (0–3 points)

• ❤️ Cardiovascular — blood pressure, lactate, vasoactive medications (0–6 points)

• 🩸 Coagulation — platelets, INR, D-dimer, fibrinogen (0–2 points)

• 🧠 Neurological — GCS, pupillary response (0–2 points)

Threshold: Phoenix score ≥ 2 = Sepsis

The score ≥ 2 threshold identifies children at the highest risk of mortality — irrespective of SIRS criteria, which have notoriously poor specificity. SIRS is dead. Know Phoenix.

🔍 2. Recognition and Screening — The Uncomfortable Truth

Don’t misread this. The panel is NOT saying screening is bad. They are saying that adding a screening tool on top of solid clinical protocols hasn’t been proven to add value. Your clinical protocols for triage and identification of the sick child? Keep them sharp. This is NOT permission to be passive.

💪 STRONG recommendation: Hospitals must implement a performance improvement program with standard operating procedures for the treatment of children with sepsis or septic shock. (Certainty: Low — but this is non-negotiable operationally.)

⚡ ED Bottom Line: No validated screening tool to mandate, but your department must have a written sepsis protocol. If you don’t have one, this guideline is the push to build it.

🧪 3. Diagnostics — What to Measure Before You Treat

Blood Cultures

✅ Good Practice Statement: Obtain blood cultures before initiating antimicrobial therapy — provided this does not substantially delay treatment. Unchanged from 2020.

The key word is “substantially.” Do not delay antibiotics chasing a blood culture.

Lactate

💪 Recommended: Measure blood lactate as part of the initial evaluation and management of children with probable sepsis or suspected septic shock. (Certainty: Very low — but clinically essential.)

Lactate ≥ 5 mmol/L now defines cardiovascular dysfunction in the Phoenix score. This is not just a resuscitation endpoint. It is part of your diagnostic criteria.

Molecular / Rapid Pathogen Testing

⚠️ Insufficient evidence — NEW: No recommendation for or against routine molecular testing for pathogen detection in children with probable sepsis or suspected/confirmed septic shock.

💧 4. Fluid Resuscitation — Context Is Everything

This is where the 2026 guidelines earn their keep. Fluid resuscitation is now explicitly stratified by healthcare system capacity.

Setting: ICU Available + Septic Shock

🔵 Conditional: Administer up to 40–60 mL/kg total fluid, in 10–20 mL/kg boluses, over the first hour of initial resuscitation. (Certainty: Low)

Setting: No ICU Available + Sepsis WITHOUT Hypotension

🚫 STRONG recommendation AGAINST fluid bolus: Start maintenance fluids only. (Certainty: HIGH)

Setting: No ICU Available + Septic Shock WITH Hypotension

🔵 Conditional: Administer up to 40 mL/kg total in 10–20 mL/kg boluses over the first hour. (Certainty: Low)

🚨 NON-NEGOTIABLE: Reassess after EVERY bolus.

Stop if shock resolves OR if any signs of fluid overload appear. Signs that should stop your boluses include clinical pulmonary oedema and new or worsening hepatomegaly. Reassess using: heart rate, blood pressure, capillary refill time, level of consciousness, urine output.

What fluid to use?

🔵 Conditional: Crystalloids over albumin for initial resuscitation. (Certainty: Moderate)

🔵 Conditional: Balanced or buffered crystalloids (Hartmann’s solution, lactated Ringer’s) over 0.9% saline. (Certainty: Very low)

If balanced crystalloids are not available, 0.9% saline is an acceptable alternative.

💊 5. Antimicrobials — Time Is Tissue

The timing recommendations are clearer than ever. The 2026 SSC draws a sharp line between septic shock and probable sepsis without shock.

💪 STRONG recommendation: For children with suspected septic shock — start antimicrobials as soon as possible, ideally within 1 hour of recognition. (Certainty: Very low — clinical urgency drives this, not just RCT data.)

🔵 Conditional: For probable sepsis without shock — allow a time-limited rapid investigation period. If concern is substantiated, start antimicrobials ideally within 3 hours of recognition. (Certainty: Very low)

✅ Good Practice Statement: Empiric broad-spectrum therapy covering all likely pathogens. If immunocompromised or high MDR (multidrug-resistant organism) risk: use empiric multi-drug therapy.

⚡ The Clock Rule:

Shock = antibiotics within 1 hour ⏱️

Sepsis without shock = antibiotics within 3 hours ⏱️⏱️⏱️

Blood cultures first — but NEVER delay antibiotics to get them.

💉 6. Vasoactive Medications — Do Not Wait for a Central Line

🔵 Conditional — NEW 2026: For children with septic shock requiring vasoactive medications, initiate via peripheral venous access rather than delaying until central venous access is obtained. (Certainty: Very low)

⚠️ Insufficient evidence: No recommendation on whether epinephrine or norepinephrine is preferred as first-line vasopressor in children with septic shock.

⚠️ Insufficient evidence: No recommendation on whether to start vasoactives before or after 40 mL/kg of bolus fluid therapy in septic shock.

🔊 7. POCUS — The New Kid in the Guidelines

🔵 Conditional — NEW 2026: For children with sepsis or septic shock, use cardiac and lung point-of-care ultrasound (POCUS) to guide resuscitation over not using POCUS — IF local training and resources allow. (Certainty: Low)

[🖼️ IMAGE EMBED: POCUS cardiac views in paediatric emergency] Search: “paediatric POCUS sepsis cardiac views” — RadioGraphics and ACEP have excellent free images

This is a big deal. POCUS lets you see IVC collapsibility, cardiac contractility and lung pathology in real time — titrating fluids and vasopressors with eyes instead of guesswork. The guidelines acknowledge the research gap and call for more outcome data, but they want you to use it if you have the skill.

The conditional language matters — this is not a mandate. It is an endorsement of capability.

🌬️ 8. Oxygen Targets — Conservative Is the New Gold

🔵 Conditional — NEW 2026: For children with sepsis or septic shock following resuscitation, titrate supplemental oxygen to a conservative SpO2 target of 88–92%, NOT a liberal target of > 94%. (Certainty: Moderate — one of the strongest evidence grades in the 2026 document.)

💚 Why conservative oxygen?

Hyperoxia causes pulmonary vasoconstriction, oxidative stress and free radical damage. In a post-resuscitation child, targeting SpO2 88–92% reduces oxygen toxicity without evidence of harm. This shifts standard practice significantly from the “100% O2 always good” era.

This applies after resuscitation — not during active shock management.

❌ 9. Etomidate — Still on the Banned List

🔵 Conditional — against use: For children with sepsis or septic shock requiring intubation, do NOT use etomidate. (Certainty: Low)

Etomidate suppresses cortisol synthesis via adrenal inhibition, potentially worsening the adrenal crisis component of sepsis. Ketamine remains your RSI agent of choice for the haemodynamically compromised child. Etomidate was already out in most paediatric guidelines — 2026 doubles down.

💊 10. Corticosteroids — Nuanced, Not Blanket

🔵 Conditional — against: For children with septic shock where haemodynamic stability can be restored with fluid resuscitation and vasoactive therapy: do NOT use IV hydrocortisone. (Certainty: Low)

⚠️ Insufficient evidence: For children with fluid-refractory, catecholamine-resistant septic shock: no recommendation for or against IV hydrocortisone.

✅ Good Practice Statement: Stress-dose corticosteroids must be administered for suspected or documented adrenal insufficiency. Routine steroids are not recommended — but adrenal insufficiency changes everything.

⚡ Bottom line on steroids:

Responds to fluids and vasopressors → no steroids.

Catecholamine-resistant + suspected adrenal insufficiency → give stress-dose steroids.

Refractory shock without clear adrenal insufficiency → clinical judgment, no firm guidance.

🧠 11. Other Metabolic and Adjunctive Therapy

⚠️ Insufficient evidence: No recommendation on sodium bicarbonate for metabolic acidaemia in sepsis.

🔵 Conditional — against: Do not use routine levothyroxine in children with septic shock in a sick euthyroid state. (Certainty: Low)

⚠️ Insufficient evidence: No recommendation on targeting normal calcium levels versus tolerating hypocalcaemia.

⚠️ Insufficient evidence: No recommendation on targeting normothermia versus permissive fever.

🏥 12. Post-Sepsis: The Chapter the 2020 Guidelines Never Had

This is a landmark addition to the 2026 guidelines. For the first time, the SSC paediatric guidelines explicitly address post-sepsis morbidity and long-term follow-up. This is not just an ICU problem. It is an ED problem, a family problem and a health system problem.

📌 Among children who survive ICU-level sepsis treatment, 30–40% will face lasting health consequences. We are discharging children who look better but may not be better.

Post-sepsis sequelae to recognise:

• 🧠 Neurocognitive impairment — executive dysfunction, learning difficulties

• 😰 Psychological sequelae — PTSD in children AND caregivers

• 💪 Physical deconditioning and functional limitation

• ❤️ Organ-specific complications: cardiac, pulmonary, renal

💚 Clinical Implication

Children discharged after a sepsis admission need a structured follow-up plan. This is increasingly the ED’s job to initiate — the referral letter, the warning to the paediatrician, the conversation with the family. The 2026 guidelines put this on the map for the first time.

🚫 The “Don’t Do” List — Quick Reference Table

RecommendationStrength

❌ Fluid bolus in sepsis WITHOUT hypotension (no ICU available)
❌ Etomidate for intubation in sepsis (Conditional against)
❌ Hydrocortisone if haemodynamically stable on fluids + vasopressors (Conditional against)
❌ Levothyroxine in sick euthyroid state (Conditional against)
❌ Liberal SpO2 targets (>94%) post-resuscitation(Conditional against)
❌ Delaying vasopressors until central line obtained (Conditional against)
❌ Albumin over crystalloids for initial resuscitation (Conditional against)

⚡ The Spoon Fed Takeaway

Phoenix ≥ 2 = sepsis. Lactate ≥ 5 + cardiovascular dysfunction = septic shock.

Shock → antibiotics within 1 hour. Sepsis without shock → antibiotics within 3 hours.

Fluids: 10–20 mL/kg boluses, reassess after each, stop at signs of overload.

No ICU + no hypotension = NO boluses. Strong recommendation. High certainty.

Vasopressors via peripheral IV. Do not wait for a central line.

SpO2 88–92% after resuscitation. Conservative oxygen is the new standard.

Avoid etomidate. Think about post-sepsis follow-up before every discharge.

📚 Key References and Resources

• Full guideline (Intensive Care Medicine):
  Published in Pediatric Critical Care Medicine and Intensive Care Medicine | Endorsed by 14 organisations | Weiss S et al. Pediatr Crit Care Med. 2026 Mar 23. doi: 10.1097/PCC.0000000000003927 - https://doi.org/10.1007/s00134-026-08360-2

• SCCM interactive recommendation filter: https://www.sccm.org/clinical-resources/guidelines/guidelines/surviving-sepsis-campaign-international-guidelines-for-the-management-of-sepsis-and-septic-shock-in

• Official paediatric quick guide infographic (free PDF):https://mcusercontent.com/acff250cf3d0418bbe904af42/files/640adb17-9ef2-6c34-398e-b5a43624a87b/SSC_Children_Quick_Guide_Infographic_2026.01.pdf

• Phoenix Sepsis Criteria — open access review: https://www.ncbi.nlm.nih.gov/pmc/articles/PMC11239205/

• ESICM 2026 SSC launch page (all infographics, adult and paediatric): https://www.esicm.org/2026-surviving-sepsis-campaing-guidelines/

• PubMed citation: https://pubmed.ncbi.nlm.nih.gov/41870559/
  

  Thanks for reading Spoon Fed PEM! This post is free so feel free to share it.

  Share

🧠 Why This Is a Real Problem

The normal thymus is the single most misidentified structure on paediatric chest radiographs. It is large, it is anterior and it sits exactly where pathology likes to hide. Miss it in one direction and you have an unnecessarily admitted infant with a battery of investigations. Miss it in the other direction and a mediastinal mass walks out the door.

The good news is that the thymus is generous with its clues. Once you know the signs, they are hard to unsee.

The thymus is soft. Pathology is not.
This single fact explains almost every distinguishing sign on the CXR.

📅 Know Your Age Window First

The thymus is only a diagnostic problem in the first few years of life. After that it involutes, becomes fatty and disappears from the film.

• 🔵 0 to 2 years: Thymus largest relative to body size. Most prominent and most commonly misread.

• 🟡 2 to 5 years: Visible but less prominent. Still a relevant differential.

• ⚪ After puberty: Involuted. An anterior mediastinal opacity now is NOT normal thymus.

⚠️ AGE RULE

If a child is over 6 and you are seeing an anterior mediastinal opacity, stop calling it thymus. Think lymphoma, teratoma or thymoma until proven otherwise.

⛵ The Four Classic Signs

⛵ Sail Sign

The right lobe projects laterally and inferiorly forming a triangular opacity like a sail. Inferior angle points inferolaterally. Your most important clue.

⛵ Sail Sign — Radiopaedia Teaching CaseClassic triangular right-sided thymic opacity with convex lateral border and inferolateral apex in an infant.

View on Radiopaedia: Thymic Sail Sign→

🌊 Wave Sign

Anterior ribs indent the pliable thymus creating a scalloped lateral border. Only a soft structure deforms around ribs. Consolidated lung and masses do not.

🌊 Wave Sign — Radiopaedia Teaching CaseScalloped lateral border of the thymus produced by anterior rib indentation. Diagnostic of a pliable soft-tissue structure.

View on Radiopaedia: Thymic Wave Sign→

🔔 Notch Sign

A step at the junction of the thymic shadow and cardiac silhouette on the frontal view. Separates thymus above from heart below.

👁️ See-Through Sign

Pulmonary vascular markings remain visible through the thymic opacity. Consolidation obliterates them. The thymus does not.

📐 Always Get a Lateral View

If the frontal view is ambiguous, the lateral CXR will usually resolve the question.

• ✅ Normal thymus fills the anterior retrosternal space

• ✅ The posterior clear space is preserved — thymus does not extend behind the heart

• ✅ Trachea is not displaced posteriorly

• ❌ Posterior mediastinal extension = not normal thymus

📐 Lateral CXR — Normal Thymic ShadowThymus confined to the anterior mediastinum with clear posterior space preserved and no tracheal displacement.

View on Radiopaedia: Normal Lateral Thymic Shadow→

🚩 Red Flags — This Is NOT Thymus

🔄 The Thymic Rebound Phenomenon

This one catches experienced clinicians.
The thymus undergoes acute stress involution during illness, surgery or steroid exposure. On recovery, it rebounds and can appear larger than before.

Classic scenario: a child completes chemotherapy. Follow-up CXR shows a large anterior mediastinal opacity. Relapse? No. Rebound thymic hyperplasia — and it is actually a good sign.

✅ REBOUND RULE

If an opacity appears or enlarges on a recovery CXR after illness or chemotherapy and has classic thymic morphology, this is rebound. Do not escalate without specialist input.

🔄 Thymic Rebound — Radiopaedia Teaching CaseThymus enlarges dramatically after recovery from illness or completion of chemotherapy. Easily mistaken for relapse.

View on Radiopaedia: Thymic Rebound→

🧩 5-Step Bedside Framework

1 Age check

Under 5 years? If not, thymus is not your answer.

2 Look for the sail sign

Triangular right-sided opacity with inferolateral apex. Present = strongly favour thymus.

3 Look for the wave sign

Scalloped lateral border from rib indentation. Present = almost certainly thymus.

4 Get the lateral view

Posterior clear space preserved? No posterior extension? No tracheal displacement? = Thymus confirmed.

5 Clinical overlay — non-negotiable

Well child with mild symptoms = reassure and discharge. Unwell child with high fever or distress = treat for pathology and re-image at 24 to 48 hours regardless of what the film shows.

✅ Bottom Line

What to take into your next shift

• 🔵The normal thymus is visible and prominent in children under 5

• ⛵Sail sign and wave sign are your primary tools. Learn them cold.

• 📐The lateral view is not optional when the diagnosis is in doubt

• ✋No mass effect, no air bronchograms, no tracheal shift = thymus until proven otherwise

• 🔻The thymus shrinks with stress and illness. Growing while the child is acutely unwell = not thymus.

• 🔄Thymic rebound after recovery or chemotherapy can be dramatic. Do not panic.

• 🚨A sick child with a mediastinal opacity always needs clinical correlation. Radiology does not trump the bedside.

REFERENCES

1. Donnelly LF et al. Diagnostic Imaging: Pediatrics. Elsevier.

2. Swischuk LE. Emergency Imaging of the Acutely Ill or Injured Child, 4th Ed. Lippincott Williams and Wilkins.

3. Radiopaedia.org — Thymic sail sign, wave sign, notch sign, thymic rebound. radiopaedia.org