PEM Myths

“Give atropine before you intubate a child.”

The logic was sound enough — laryngoscopy is vagally stimulating, children tolerate bradycardia badly, so pre-empt it. The evidence went the other way. Intubation-associated bradycardia in children is short-lived and almost never progresses to a dangerous rhythm or to haemodynamic compromise. Atropine's tachycardia consumes oxygen in a hypermetabolic child you have just spent minutes pre-oxygenating. And the practice never rested on clear evidence in the first place — eminence rather than evidence. Horeczko's line is the one to remember: bradycardia with a pulse is answered with oxygenation and ventilation, and atropine is a medication, not a placemat.

“No opioids in abdominal pain — the surgeons won't be able to assess them.”

This one has a paper trail. Zachary Cope codified it in 1921, and his Acute Abdomen is still read. The 1996 edition softened it to waiting for a responsible surgeon; only in 2001 did an edition call the original an unfortunate dictum. Eighty years. Meanwhile the evidence had been in for a decade — adults in 1992, children in 2005, a Cochrane analysis in 2011 — and all of it says analgesia before surgical review is safe and does not delay diagnosis. Horeczko adds the point that actually matters at the bedside: it may sharpen the examination rather than blunt it, because a child who is not braced against you can localise.

“You can't bolus D5NS.”

The concern behind the rule is real — concentrated dextrose is irritant and sclerosing, which is why infants under one get D10, toddlers and preschoolers get D25, and only the over-eights get D50. But D5NS is the most dilute preparation there is: roughly two ampoules of D50 in a bag of saline. It will not sclerose a vein. The child in front of you is the miserable vomiter who has failed ondansetron and small sips, is dehydrated, and has been running on nothing for hours — a mild starvation ketosis. If you are putting fluid in anyway, breaking the ketosis at the same time costs nothing and makes them feel human faster.

“You gave adrenaline, so now you have to keep them.”

The four-to-six hour observation grew from a fear of biphasic rebound. Horeczko makes an argument here that goes beyond disposition: the prospect of committing a child to half a day in the department changes who gets treated in the first place. Those hives don't look that bad. He only vomited once. Under-treatment of anaphylaxis is the well-documented problem, not over-treatment. The pharmacology gives you a defensible floor: intramuscular adrenaline has an onset around eight minutes and a half-life around forty-five, so two half-lives is ninety minutes, and two hours covers it for a straightforward reaction with a known trigger. Longer for tree nuts, for asthma or atopy, for anyone with a history of severe reactions.

The route is not negotiable. Intramuscular, lateral thigh, whatever lines are already in — 0.01 mg/kg to a maximum of 0.3 mg. Autoinjectors are fine when the dose is about right: 0.15 mg from 15 kg, 0.3 mg from 30 kg. Below 15 kg, draw it up, because an autoinjector into a small thigh can deliver into bone.

“Sickle cell trait is nothing.”

Mostly true, and that is exactly what makes it dangerous. Carriers have overwhelmingly normal haemoglobin A and do not get pain crises going about their day. But under extreme physiological stress — heat, exertion, altitude, critical illness — the small fraction of haemoglobin S can sickle. Horeczko's case is a fifteen-year-old who trained hard in summer heat, arrived looking like straightforward heat exhaustion with acidosis and rhabdomyolysis, was admitted and hydrated, and died within a day. Exertional collapse associated with sickle cell trait carries a 37-fold risk, and NCAA screening dropped the associated death rate from about 1 in 28,000 to 1 in 250,000. Worth asking about, in a flat tone, in any undifferentiated collapse.

“No steroids for paediatric pharyngitis.”

The fear is specific and legitimate: leukaemic tonsillitis exists, steroids are part of leukaemia induction, and treating a cancer by accident delays its diagnosis. That consideration hardened into a prohibition. The evidence has moved — a null result in 2003, a small effect in 2010, clearer benefit in strep-positive children by 2012, and a 2020 Cochrane meta-analysis putting resolution at 24 hours at about 2.4 times more likely. Horeczko's position is the sensible one: make the diagnosis first, do a proper head and neck examination, and if the story is long or recurrent or there is lymphadenopathy and fatigue, chase that instead. Then, for the child who cannot swallow their own secretions, give the dexamethasone.

“Self-adhesive dressings are fine on fingers.”

This is the one to tell people about. Grayson, aged two, caught his right hand under something heavy in his father's garage. The lacerations were repaired well. He kept pulling the dressings off, so a self-adherent elastic wrap went on with entirely good intentions, and the family were told to change it twice a day. They changed it once in ten days. He came back with mummified digits and lost the index and middle fingers of his right hand. These wraps are excellent for temporary haemostasis and dreadful as a take-home dressing; the case reports of paediatric digital necrosis are numerous enough that Horeczko calls the position indefensible. And ischaemia is quiet — the pain stops.

What to do instead, for the child who will not keep a dressing on: petroleum gauze directly on the wound, dry gauze over it, a single layer between that finger and the next to stop maceration, then a fluffy gauze roll wrapped loosely into a mitten, anchored round the base of the thumb and the wrist. The child cannot pick it off, and the next dressing change does not tear away the healing tissue.