We already know that nasal high-flow (NHF) apnoeic oxygenation is safe and effective in the controlled environment of elective theatre intubation and that it improves first-attempt success rates in neonates. But does it hold up in the unpredictable world of emergency paediatric intubation?
That’s exactly what Shane George et al. set out to explore in the Kids THRIVE study—investigating whether NHF apnoeic oxygenation could improve intubation outcomes in critically unwell children needing emergency airway management.
The Kids THRIVE study was a pragmatic, multicentre, international, open-label, randomised clinical trial, spanning ten hospitals across Australia, New Zealand, and Switzerland. A total of eleven intensive care units (ten PICUs and one non-maternity NICU) and four emergency departments (EDs) took part in the study.
Given the emergency setting, the trial used a research without prior consent model, with delayed consent-to-continue obtained from the child’s legal guardian as soon as practical.
Who were the patients?
Children under 16 undergoing emergency intubation between May 2017 and October 2022 were eligible for inclusion if consent could be obtained from their parent or legal guardian either prospectively or retrospectively.
Exclusion criteria were:
· an intention for primary nasal intubation
· elective endotracheal tube changes
· intubation required immediately for loss of cardiac output or respiratory arrest
· location of intubation outside the ED or ICU
blocked nasal airway due to anatomical abnormalities or acute facial trauma
What was the intervention?
Patients were randomised in a 1:1 ratio to receive either nasal high-flow or standard care. Children who required multiple intubations could be re-enrolled and re-randomised.
Due to the nature of the intervention, blinding was not feasible. Preoxygenation techniques, anaesthetic agents, and neuromuscular blocking agents were left to the discretion of the treating clinicians.
Intervention group: 2 L/kg/min 100% oxygen via nasal cannula using the Optiflow™ THRIVE system
Comparison group: standard care – defined as no oxygen during apnoea
What were the outcomes measured?
Primary outcomes
intubation with a hypoxaemic event
successful intubation (correct endotracheal tube placement and end-tidal carbon dioxide measurement) at first attempt without hypoxaemia
Hypoxaemia was defined as:
SpO2 of 90% or lower
or an SpO2 saturation difference of more than 10% for patients with cyanotic congenital heart disease with known right-to-left shunt
or where pre-intubation SpO2 was under 90%
Secondary outcomes
number of intubation attempts per patient
number of intubation attempts followed by re-oxygenation, defined as the need for bag ventilation or other oxygen device between intubation attempts
duration of mechanical ventilation
lowest SpO2 throughout the intubation period
ventilation-free days (days free of invasive respiratory support for the first 28 days after randomisation)
length of NICU/PICU stay
length of hospital stay
mortality during hospital admission
minor and major adverse events
What were the results?
1069 intubations in 969 children were randomly assigned to nasal high flow (535 intubations) or standard care (534 intubations). 950 intubations were included in a modified intention-to-treat analysis.
Primary outcomes
The modified intention-to-treat analysis found no statistically significant difference between the nasal high-flow and standard care groups in reducing hypoxaemic events or improving first-attempt intubation success.
Hypoxaemia occurred in 12.8% of intubations in the nasal high-flow group, compared to 16.2% in the standard care group.
Successful intubation on the first attempt without desaturation was achieved in 63.0% of cases in the nasal high-flow group and 59.1% in the standard care group.
In the preplanned per-protocol analysis, there was a significantly lower rate of hypoxaemia in the nasal high-flow group. Hypoxaemia occurred in 10.8% of intubations with nasal high-flow, compared to 16.7% in the standard-care group. Successful first-attempt intubation rates were 64.0% and 58.1%, respectively.
In NICU patients, the nasal high-flow group experienced fewer hypoxaemic events, although there was no difference in first-attempt intubation success between the groups.
Lower rates of hypoxaemia were also observed in the nasal high-flow group for junior intubators in the NICU and in infants under one year of age.
Secondary outcomes
There were no differences between groups in any of the secondary outcomes.
There were 168 minor and 74 major tracheal intubation-related adverse events reported during the study.
In the modified intention-to-treat population, 84 children died, with five deaths occurring within 24 hours of randomisation and 79 during hospital admission. None of these deaths were attributed to the study interventions.
The bottom line
The use of NHF did significantly not improve the rate of first attempt intubation success. When NHF is delivered to patients during apnoea it reduces that rate of hypoxaemic events, but further research is required to understand barriers to it use in the ED and ICU setting.
CASP checklist – How good was the paper?
Does the study address a clearly focused issue?
Yes
Was the cohort recruited in an acceptable way?
Yes. Recruitment took place across ten PICUs, one NICU, and four EDs in Australia, New Zealand, and Switzerland. A wide case mix of children were randomised, and a large proportion of families provided consent.
However, 730 eligible patients were not approached for inclusion in the study, which may have introduced selection bias.
Was the exposure accurately measured to minimise bias?
Patients were randomised in a 1:1 ratio to nasal high flow versus standard care. Blinding was not possible, potentially introducing allocation bias.
As this was a pragmatic trial, lots of decisions were left to the clinical team (e.g., process and level of preoxygenation, choice of intubation drugs). This could potentially influence study outcomes and introduce a risk of ascertainment and selection bias.
Was the outcome accurately measured to minimise bias?
Yes. Primary outcomes were intubation with a hypoxaemic event and successful intubation (correct placement and end-tidal carbon dioxide measurement) at first attempt without hypoxemia.
Intubations were recorded on an iPad allowing for independent review of events and confirmation of accurate data capture.
Have the authors identified all important confounding factors?
The automated randomisation process used block stratification by age, operator experience, and recruitment location to ensure an even distribution of potential confounding factors between groups. Randomisation was conducted using sealed envelopes to maintain allocation concealment.
However, as with any randomised clinical trial, particularly in a heterogeneous paediatric population, there is always the possibility of unidentified or unaccounted confounding factors that could influence the results.
Was the follow-up of the subjects complete and accurate?
Follow-up continued until patient discharge. Ventilation-free days were defined as the number of days without respiratory support for all episodes requiring an endotracheal tube within the first 28 days after randomisation, with a maximum of 28 days recorded.
Fewer than 10% of patients were withdrawn after randomisation, with similar numbers in both groups.
Compliance with the study intervention was high, exceeding 95%.
What were the results?
The modified intention-to-treat analysis found no statistically significant differences between the nasal high-flow and standard care groups in reducing hypoxaemic events or improving first-attempt intubation success.
However, the per-protocol analysis showed a significant reduction in hypoxaemic events in favour of nasal high-flow.
There were no differences in secondary outcomes between the groups.
Do you believe the results?
Yes
Can the results be applied to a local population?
Yes. This was the largest prospective trial to date investigating emergency intubation in paediatric patients, including those in PICU, NICU, and ED settings.
However, the generalisability of the findings to other countries or healthcare centres may be limited due to differences in resource availability and variations in clinical practice..
Do the results fit with the other evidence available?
Yes. Study results regarding the incidence of successful first-attempt intubations were similar to those of a large registry study (NEAR4kids) of North American PICUs and a recent meta-analysis.
What did the authors conclude, and what can we take away from this study?
The authors concluded that in children under 16 years of age, the likelihood of experiencing a hypoxaemic event or achieving successful first-attempt intubation without oxygen desaturation was comparable between those receiving nasal high-flow during the apnoeic phase of emergency intubation and those receiving standard care.
However, while the per-protocol analysis found that nasal high-flow was associated with fewer hypoxaemic events, it did not improve the rate of successful first-attempt intubation without oxygen desaturation.
In NICU patients, the nasal high-flow group experienced fewer hypoxaemic events compared to standard care, raising the question of whether this population may particularly benefit from nasal high-flow use.
Finally, the study reported a higher-than-expected rate of intubation attempts abandoned despite the absence of oxygen desaturation. This raises an important question—were these intubations stopped due to structured training guidelines, such as a time limit or a specific SpO₂ target? Understanding the reasons behind these abandoned attempts could provide further insights into optimising paediatric intubation practices.
From Shane George
The use of apnoeic oxygenation when intubating in the critical care setting has become increasingly popular as a technique that may improve intubation success rates and prevent hypoxaemia during the intubation attempt. This practice has been largely translated from adult practice, with limited data to confirm its effectiveness in children. Nasal high-flow (NHF) oxygen during apnoea for intubation has been demonstrated to be effective in neonates and operating theatre settings but has not previously been studied in unwell children in the critical care setting.
The Kids THRIVE study is the largest, prospective, randomised clinical trial in paediatric intubation published to date. Its pragmatic design ensured that a broad range of clinical presentations were included across the ED and ICU settings. At first glance, the study outcomes appear to imply that there is no benefit of NHF on either hypoxaemia or intubation success. However as with all things, there is a more nuanced interpretation of the data presented required.
When designing the study we pre-planned a per-protocol analysis expecting that there would a proportion of children who didn’t receive the intervention they were allocated to. This is an essential inclusion for pragmatic studies, particularly those introducing a new technique or procedure into an existing practice or in time-critical interventions. While the intention-to-treat (ITT) analysis remains the gold standard for reporting of clinical trials as it represents the effect of an intervention in real-world clinical practice (i.e. despite the best intentions sometimes things don’t go to plan), the per protocol analysis gives us insight into whether the intervention would work if it was used in the manner intended (i.e. under ideal circumstances).
It’s not surprising to demonstrate in our study that if you don’t put the NHF on the patient it isn’t as effective. We have demonstrated that if NHF is used as intended the rate of hypoxaemia falls from 16.7% to 10.8% (p = 0.017). What we need to understand better is what the barriers are to applying the NHF during intubation, something we will focus on in secondary analyses using the captured video recordings.
One of our unexpected findings was that around 17% of intubations in both groups were abandoned before the child experienced a significant physiological deterioration. Previously published data have suggested that failed first attempt intubation is associated with higher rates of adverse events on subsequent intubations. In general, those intubation attempts were associated with abnormal vital signs at the point of failure, so further analysis of this cohort is needed to inform clinical practice. The high rate of abandoned attempts prior to oxygen desaturation was not planned for in the sample size calculations, and thus reduces the power of the study to detect a significant difference. This finding may also explain why we demonstrated a lower frequency of severe desaturation (SpO2 < 80%) in our cohort when compared to other registry datasets.
Translating results of trials where the mITT and per-protocol analyses yield conflicting statistical significance is challenging for clinicians. The interpretation needs to include other factors such as potential for harm, cost, difficulty in implementing the intervention and potential patient benefits. Given the low risk of harm and potential for benefit to the patient, it will be system and clinical factors which will influence clinicians on whether to provide apnoeic oxygenation or not.
This study was a massive undertaking, and I sincerely thank all the participating sites who embraced this study and recruited patients during often high-stress situations, and also throughout the COVID-19 pandemic. It is a remarkable achievement and I thank the many people who have contributed to the study. I sincerely look forward to seeing how this data is incorporated into clinical practice.