Dani gave this talk at the Irish Association of Emergency Medicine Annual Scientific Conference in October 2023
Once upon a time, there were two beautiful people, Barbie and Ken. Through the play and make-believe of many children, they enjoyed a romance spanning many decades. Sure, they had their ups and downs (what couple doesn’t?), but after more than 40 years of courtship, Ken proposed to Barbie. After a gorgeous wedding, they soon announced their wonderful news: Barbie was pregnant.
Babie Junior, or Bee for short, was the definition of a cute, delightful baby.
Three days after her first birthday, she started coughing, breathing hard and fast. Soon, she was too breathless to feed, and Barbie and Ken were panic-stricken at how pale and lethargic she looked. What were Barbie and Ken to do? It was a Saturday afternoon, and the GP was closed. You, however, are working an ED shift, and when you see the name “Barbie Junior” on your computer screen, your heart does a flutter. A VIP – how exciting!
When you assess Bee, though, it’s clear she’s not good. Her respiratory rate is 70, her sats are only 86% in air, and her work of breathing is pretty bad – she’s grunting and has subcostal and intercostal recession and a marked tracheal tug. She has widespread wheeze and creps throughout her chest, and you diagnose bronchiolitis. She’s pretty sick, though – a gas shows a respiratory acidosis, confirming acute hypoxemic respiratory failure. She’s not sick enough for intubation, but she’s not great either. What are you to do?
You have three choices:
High-flow nasal cannula oxygen seems to be the in-trend management to support breathing, with the aim of reducing respiratory distress and hopefully preventing disease progression. It can be in started in the ED, and many of us are already using it.
The PARIS trial, looking at almost 1,500 infants with bronchiolitis in Australia and New Zealand, showed that babies on high-flow were less likely to have an escalation in care than babies on simple nasal cannula oxygen. Sounds pretty convincing. Is this the route you want to go?
But, and this is a big but…
Potentially problematic, from the study point of view, but unavoidable in real life, randomisation to standard oxygen or high flow could not be blinded, and clinicians could choose to escalate for any reason – and many did. For those in the standard therapy arm, escalation of care meant high flow. Secondary outcome analysis showed there was no difference in length of hospital stay, ICU stay, or duration of oxygen therapy. And importantly, adverse event rates were both low and equivalent in both groups.
Which brings us to PARIS-2. This time, the group asked a different question – does using early high-flow nasal cannula oxygen instead of standard oxygen therapy reduce hospital length of stay?
Over 1,500 children in EDs, aged 1 to 4 with acute hypoxemic respiratory failure needing hospital admission, across 14 emergency departments in Australia and New Zealand were randomly allocated 1:1 to high-flow oxygen therapy OR standard oxygen therapy (approximately 750 children in each group).
This is what they found:
- Length of hospital stay (time from randomisation to the time of hospital discharge or death), the primary outcome, was significantly longer in the high-flow group at 1.77 days compared to 1.50 days in the standard oxygen group – there was no difference between children with wheeze and those without wheeze
- PICU admissions were higher in the high-flow group (12.5%) compared to standard oxygen group (6.9%)
- The length of oxygen therapy was longer in the high-flow group at 1 day compared to 0.75 days in the standard oxygen group. Why? Perhaps there was a bias in the clinicians’ perception that children in the high-flow group were sicker, causing clinicians to wean slower, and therefore leading to a longer length of oxygen therapy and longer length of hospital stay. Maybe this perceived sickness meant there was a lower threshold to escalate to the PICU.
The bottom line is:
Early high-flow nasal cannula oxygen started in the ED resulted in a longer length of hospital stay than standard oxygen therapy in children aged 1 – 4 with respiratory failure.
Convinced, you admit Bee on standard oxygen therapy. She does not need escalation to high flow and is discharged three days later.
Three days before her second birthday, however, she starts coughing again, this time spiking a fever to 39 C. Barbie and Ken were so grateful for the excellent evidence-based care Bee had received a year earlier that they decided to bypass their GP and bring her straight to ED. Luckily, you’re on shift again.
Bee sits quietly in Ken’s lap, looking at you warily. Her respiratory rate is up a bit at 40, but her oxygen sats are okay at 96%. Her heart rate is normal. Although she’s plastic, she has equal chest wall rise and focal creps at the right base, with only mild respiratory distress. She doesn’t look that sick.
You diagnose community-acquired pneumonia and tell Barbie and Ken that Bee’s well enough to have some antibiotics at home – and you’d like to give her amoxicillin. But how much will you prescribe?
You have four options:
- Three days low dose
- Seven days low dose
- Three days high dose
- Seven days high dose
Lucky for us, PERUKI looked into this exact question.
The CAP-IT trial recruited children aged six months to 5 years with a diagnosis of community-acquired pneumonia on discharge from the emergency department or ward. Children were eligible if they had had a cough within 4 days, fever within 2 days, signs of difficulty breathing or focal chest signs. They were excluded if they’d had recent antibiotics, underlying chronic lung disease, a complicated LRTI or bilateral wheeze.
They recruited over 800 children, who were randomised to three days of low-dose amoxicillin, three days of high dose, seven days of low dose or seven days of high dose amoxicillin. Families and clinicians were blinded using a clever placebo for days 4 to 7 of the short course groups that looked, smelled and tasted like amoxicillin.
- No difference in retreatment with antibiotics for community-acquired pneumonia within 28 days between groups.
- Longer cough by two days for the 3-day courses (12 days vs 10 days).
- Side effects were similar, although more likely to develop a rash in long course.
- Compliance was better in the 3-day groups (98% v 92%)
The main limitation was that lots of the children had viral lower respiratory tract infections. Pragmatically, it can be difficult to differentiate between viral and bacterial LRTIs, even with a chest x-ray.
The bottom line is:
A short course of 3 days of amoxicillin is as good as a longer course or a higher dose. There are a couple of extra days of cough, but does this matter? When balanced against compliance and side effects, and importantly, antimicrobial stewardship, I’d argue not.
The SAFER trial, a randomised-controlled trial looking at short-course antimicrobial therapy for paediatric CAP, had similar findings. 281 children between 6 months and ten years were randomised to 5 or 10 days of high-dose amoxicillin. Shorter courses were found to be non-inferior to longer courses for treating CAP. To add to this, the SCOUT-CAP RCT showed that shorter 5-day courses had a higher chance of a desirable outcome (symptom resolution and fewer adverse events) with less antimicrobial resistance than longer 10-day courses.
You discharge Bee with three days of low-dose amoxicillin. Like the good girl she is, Bee takes her antibiotics with a smile on her face and makes a good recovery.
Just two weeks later, though, you see Bee’s name on the ED computer screen again. This time, she’s in with an injury. She was playing in the garden. Ken was mowing the lawn, paying close attention to the borders and little attention to Bee. He heard a cry and turned to see her sprawled on the patio. Giving her cuddles, he managed to console her, but she was reluctant to bear weight. She looks at you with eyes brimming with tears as she limps into the cubicle.
An x-ray confirms a toddler’s fracture. But what will you do?
You have four options:
- Above knee cast
- Below knee cast
Our approach to managing toddler’s fractures mostly comes from retrospective case studies and only two small prospective case series. These had small sample sizes and no blinding. Toddler’s fractures heal quickly, with excellent outcomes. So, what’s the best way to get toddlers back on their feet as quickly as possible?
The initial study design was to randomise toddlers with toddler fractures to either an above-knee cast or nothing – an observation group, which we’ll call non-immobilisation. Because parents struggled with the idea of being randomised to a treatment, after a year, the team allowed parents to choose which treatment they preferred the sound of. Shared decision-making at its finest.
Forty-four toddlers participated, 34 (77%) in the preference arm and 10 (23%) in the randomized cohort – 10 in a cast and 24 not immobilised (the higher number in the non-immobilised group was because once parents were allowed to choose which group their child was in, more parents chose not to immobilise).
- 92% of toddlers who weren’t immobilised were back to normal walking by four weeks, compared to only 50% of those with a cast – this is a big difference.
- Over 90% of patients in both groups were back to normal walking at week 8.
- In a cute twist of patient-centred outcomes, the team looked at initial play scores. Toddlers who weren’t immobilised had a higher play score at diagnosis (probably because no heavy cast impeded their play) than toddlers in a cast.
- Healing was excellent in both groups.
- Regarding complications – four of the ten toddlers in a cast had cast issues, and there were no complications in the non-immobilised group.
- At eight weeks, 95% of parents in the non-immobilisation group said they’d choose the same treatment again, compared to 80% of parents of toddlers in a cast.
The bottom line:
This small study suggests that toddlers with toddler’s fractures will be back on their feet quicker, back to play faster, with fewer complications, and happier parents, without any immobilisation.
The PERUKI group are taking this a step further in the TWIST study. A scoping review is currently underway, with a plan to launch a randomised controlled trial. Watch this space.
There’s a definite vogue to minimise immobilisation in children with fractures. The FORCE study randomised 900 children with a distal radius buckle fracture to either rigid immobilisation (with either a cast or splint depending on normal local practice) or a crepe bandage (the focus group were very clear that they needed at least a bandage!)
- There was no statistically significant difference in pain scores on day three.
- 11% of participants who were offered a bandage crossed over to the rigid immobilisation group, mainly due to pain.
- Parents in the rigid immobilisation group were more satisfied on day one, but there was no difference by six weeks.
- Both treatment options led to a similar number of missed school days – around one and a half.
- There was a small difference in analgesia use. 83% of the bandage group had painkillers, compared to 78% in the rigid immobilisation group on the first day, though there was no significant difference down the track.
The bottom line:
It’s safe to treat distal radius buckle fractures with the offer of a soft bandage and immediate discharge.
So – knowing TWIST is coming soon and knowing the importance of shared decision-making, you discuss the treatment options with Barbie and Ken. They like the sound of an earlier return to play and opt for observation without immobilisation. Bee toddles off, blowing you a kiss as she leaves the department.
Several years pass before you cross paths with Barbie, Ken and Bee again. Bee remains a sweet, happy girl, the perfect daughter. Barbie and Ken don’t need to seek further healthcare advice until one awful day. Bee, aged 12, was walking home from school. At the crossing by her house, the red man turned to green, and unaware of the danger of a speeding car, Bee started to cross. The driver of a 57 Chevy Classic didn’t see the red light. Disaster strikes when the car hits Bee, hurtling her into the air.
As it so happens, you’re on shift again.
The trauma team swoops into action. The primary survey confirms isolated head injury – Bee’s drowsy, and a formal GCS is 12 – E3, V4, M5. She has a right parietal-occipital region haematoma, and you think you might be able to feel a step – you’re pretty convinced she has a palpable skull fracture. Pupils are equal and reactive. There’s no evidence of any other injury.
It’s clear Bee needs a scan of her brain, and given her mechanism, you’d like to include her c-spine. You tell the team you’re worried about an intracranial bleed, but here’s a question – are you going to give tranexamic acid?
You’ve two options:
- Yes – give TXA
- No – don’t give TXA
The 2013 CRASH 2 international multicentre RCT compared TXA with placebo in >20,000 adult trauma patients. The team chose a primary outcome of 28-day mortality and found that TXA, when given within 3 hours of injury, reduced mortality without causing harm, so it’s been widely adopted into trauma management.
The international PATCH (Prehospital Antifibrinolytics to Trauma Coagulopathy and Haemorrhage) study published this June compared a prehospital TXA with a placebo in more than 1300 adult patients. The TXA group had lower mortality, both at 24 hours and six months. However, the primary outcome, a quality of life score at six months, the Glasgow Outcome Scale Extended (GOS-E), was the same in both groups (53.7% in the TXA group vs 53.7% in the placebo group).
But what about the evidence of TXA in isolated head injury? The CRASH 3 trial looked specifically at this. TXA was compared with placebo in >12,000 adult trauma patients with a GCS of 12 (or confirmed intracranial haemorrhage on CT), with no significant other non-head injury bleeding. 18.5% of patients in the TXA group died from their head injuries, compared to 19.8% in the placebo group – this was not statistically significant. However, subgroup analysis threw up some interesting results. When patients were separated into those with a GCS <9 or GCS 9-12, TXA was shown to significantly improve outcomes in patients with GCS 9 or above (a mild to moderate head injury). The risk of death from a mild to moderate head injury was reduced when TXA was given (a decent relative risk reduction of 0.78). Giving TXA earlier (<3h) conferred a greater benefit.
But what about the evidence in children?
PED-TRAX, the Paediatric Trauma and Tranexamic Acid Study was a retrospective review of >750 paediatric trauma admissions in Afghanistan. They found that the more injured a child was (higher ISS, hypotension, acidosis or coagulopathy), the more likely they were to receive TXA. Interestingly, even though kids who received TXA were much more severely injured, the children who received TXA had better survival outcomes than those who didn’t, with no complications.
Adding to this evidence in favour of TXA, a secondary analysis of 2022’s MATIC (MAssive Transfusion In Children), an international prospective observational study of 450 children with life-threatening bleeding needing >40mL/kg of blood products. Children who received TXA or other antifibrinolytics had better survival outcomes at 6 and 24 hours— another win for TXA.
However, Kornelson et al.’s 2022 systematic meta-analysis of TXA in paediatric trauma found that TXA only seemed to improve survival in children injured in combat and didn’t improve survival in injured civilian children. A strike against TXA.
There’s something on the horizon, though. The PECARN TIC-TOC trial, a large-scale, prospective, multicentre, double-blinded RCT, is randomising severely injured children to receive either low dose TXA (15 mg/kg bolus + 2 mg/kg/h infusion), high dose TXA (30 mg/kg bolus dose + 4 mg/kg/h infusion) or normal saline placebo bolus + infusion.
The TIC-TOC team will be looking at immediate survivability and longer-term sequelae, including quality of life at one week, one month, three months and six months, as well as longer-term neurocognitive follow-up, taking the evidence beyond immediate survival and into the quality of life realm.
The bottom line:
Early TXA improves survival in adults with isolated mild to moderate head injury, but it’s not clear if this is true in children. The TIC-TOC trial will give us more answers.
You decide, on balance, not to give Bee tranexamic acid. Her CT shows a skull fracture and a small subdural that does not need neurosurgical intervention. She’s admitted for 48 hours of observation and is discharged with follow-up.
And that’s where our involvement with Barbie, Ken and Bee ends for now.
We’ve learned a lot:
Early high-flow nasal cannula oxygen leads to longer, not shorter, length of stay and time on oxygen than standard oxygen therapy in children aged 1 to 4 with hypoxic respiratory failure.
In children with community-acquired pneumonia, a short course of 3 days of low-dose amoxicillin is as good as longer courses or higher doses, with the only disadvantage being a couple of extra days of cough.
Although evidence is limited, toddlers mobilise faster, with fewer complications, and happy parents when they are not immobilised for their fracture. The PERUKI TWIST RCT will give us more answers.
Although early tranexamic acid improves survival in adults with isolated mild to moderate head injury, it’s not clear if this is true in children. The PECARN TIC-TOC RCT will shed some more light on the use of TXA in paediatric trauma.
So here’s your call to action – find one paper that will change your practice per month. It doesn’t have to be in a journal – it can be from a blog post or social media. And then tell one clinical friend. Next time you see that condition – think back on the paper and see if it changes your practice.