Good morning and thank you very much for the kind invitation to talk to you at this plenary session of the SCCM Congress. I'll be talking to you today about the first ABC trial in children. My name is Dr. Ram Narayan. I work as a pediatric intensivist in London and I'm the chief investigator for the first ABC trial. The trial compared high-flow nasal cannula with continuous positive airway pressure for post-extubation respiratory support in children. A couple of disclosures unrelated to this talk. Invasive mechanical ventilation is one of the most commonly provided intervention in paediatric intensive care. In UK paediatric ICUs, nearly 65% of admissions received invasive mechanical ventilation between 2017 and 2019. It is also common practice to provide non-invasive respiratory support in the post-extubation phase, commonly using CPAP, continuous positive airway pressure or high-flow nasal cannula or HFNC. Both CPAP and high-flow are used commonly. This is indicated by international surveys as well as observational data. Cohort studies indicate that between 10 and 43 percent of extubated children receive non-invasive respiratory support and there is a preference amongst clinicians for the use of high-flow nasal cannula. However, despite the fact that these two interventions are so commonly used, there are no randomized control trials comparing the two in children following extubation. A very recent network meta-analysis from critically ill adults, which included 36 RCTs and 6,800 patients, did not show any difference in reintubation rates between non-invasive ventilation, which included CPAP and high-flow nasal cannula. The first ABC trial was designed as a master protocol to compare high-flow nasal cannula with CPAP in the post-extubation as well as in the acute settings. Today I'll be reporting the results of the first ABC step-down RCT. The master protocol allowed a shared infrastructure and integrated health economic evaluation. The research question for the step-down RCT is as follows. In critically ill children assessed by the treating clinician to require non-invasive respiratory support within 72 hours of extubation, is the first-line use of high-flow nasal cannula at 2 liters per kilogram per minute flow rate non- inferior to continuous positive airway pressure or CPAP at 7 to 8 centimeter water pressure in terms of the time to liberation from all forms of respiratory support, invasive as well as non-invasive. The trial was designed as a pragmatic open-label multi-center parallel group non-inferiority trial. Randomization was on a one-to-one basis to high-flow nasal cannula or to CPAP and was stratified by site and age under 12 months versus over 12 months. An independent data monitoring committee as well as a trial steering committee were constituted and a planned interim analysis when 300 patients had reached 60-day follow-up was cancelled because of rapid recruitment to the study. The consent model adopted was a research without prior consent model in line with previous critical care trials. The inclusion criteria for the trial were a child who was either admitted or accepted for admission to the Pediatric Intensive Care Unit, who was not a preterm baby and was not an adult and was assessed by the treating clinician to require non-invasive respiratory support within 72 hours after extubation. This could be either planned initiation or rescue initiation of non-invasive respiratory support but all it required was a clinical decision to initiate support. There were several exclusion criteria. The key amongst them were receipt of high flow or CPAP for a period of over two hours in the 24 hours prior to randomization and a clinical decision to start a form of non-invasive respiratory support that was not high flow or CPAP such as non-invasive ventilation or negative pressure ventilation. Other exclusion criteria included untreated air leak or agreed not for intubation or limitation of critical care treatment plan in place. The primary outcome for the trial was time to liberation from respiratory support which was defined as the beginning of a 48 hour period during which the child was free of all forms of respiratory support invasive or non-invasive. This did not include the use of supplemental oxygen. This primary outcome was chosen because it considered the effect of re-intubation on time deliberation as well as the fact that it took into account long duration of non-invasive support if the patient was not re-intubated. This primary outcome was supported by parents during pre-trial consultation over re-intubation. Several secondary outcomes were studied such as mortality at PQ discharge at day 60 and day 180, rate of re-intubation at 48 hours, duration of PQ and acute hospital stay, patient comfort assessed using a validated comfort B score, proportion of patients in whom sedation was used during non-invasive respiratory support, parental stress measured using the parental stressor scale PQ which is a validated measure. The non-inferiority margin was set at a hazard ratio of 0.75 580 events provided 90% power with a type 1 error rate of 2.5% one-sided to exclude the non-inferiority margin and in context the hazard ratio of 0.75 in the pilot randomized control trial that was performed before the main trial corresponded to approximate 16 hours of time deliberation from support. To take into account 5% for censoring for death or transfer and 10% refusal or withdrawal of retrospective consent and non-adherence in the per protocol population the sample size was set at 600 patients. The primary analysis consisted of all patients who started any form of respiratory support. A Cox regression model was used and adjusted for age, comorbidities, length of prior invasive ventilation, reason for invasive ventilation, baseline ratio of saturation to fraction inspired oxygen, severity of respiratory distress at randomization and the effect of site using a shared frailty model. The per protocol analysis consisted of all eligible patients who started the allocated respiratory support. A number of subgroups were analyzed including age, comorbidities, length of prior invasive ventilation, reason for invasive ventilation, the ratio of saturations to FiO2 and the severity of respiratory distress. This slide shows the screening, randomization and follow-up of patients recruited to the first ABC step-down trial. Over a nine-month recruitment period at 22 paediatric intensive care units in England, Scotland and Wales, just over 3,000 children were extubated and screened of whom 1,051 fulfilled inclusion criteria and exclusion criteria. Of the 1,051, 600 children were randomized into the trial. 451 were eligible but did not undergo randomization for several reasons, mainly because of being missed or identified too late or a clinical decision not to include the patient in the trial. Of the 600 patients who were randomized, 299 were randomized to high-flow and 301 were randomized to CPAP. Following randomization, consent was obtained in 291 high-flow patients and 296 CPAP patients. Of those patients, 10 did not start any respiratory support in the high-flow group and 24 did not start any respiratory support in the CPAP group. So the primary analysis set for the trial consisted of 281 patients who started any form of respiratory support on the high-flow group and 272 patients who started any form of respiratory support in the CPAP group. As you can see, the majority of patients who started respiratory support started the allocated treatment, 272 out of 281 in the high-flow group and 252 out of 272 in the CPAP group. The follow-up was nearly complete and time to liberation was evaluable in the vast majority of patients. So 258 patients in the high-flow group and 260 patients in the CPAP group contributed primary outcome data. The rest were censored. The baseline characteristics between the two groups were quite similar. The age was three months. Male patients represented nearly 60% of the population. At least one comorbidity was present in nearly 60% of the groups. There were some differences in the main reason for invasive ventilation. More children with bronchiolitis were represented within the CPAP group and more children were ventilated for cardiac reasons within the high-flow group. In both groups, the duration of prior invasive ventilation was around 89 hours. The nature of post-extubation non-invasive respiratory support was planned in nearly two-thirds of the patients in both groups and rescue in about 20% of both the groups. This slide shows that the physiology and clinical characteristics at randomization were quite similar between the two groups. This includes the comfort B score which indicates a measure of discomfort in the patient. Clinical management of the patients was as expected as per trial recommended algorithms in both the high-flow and the CPAP groups and as can be seen on the left the high-flow group the median flow rate that was provided was 100% of the recommended starting rate based on weight and on the CPAP group pressure was median of 7 centimetres of water with range between 6 and 8 centimetres of water as recommended by the trial protocol. These graphs show the primary outcome for the primary analysis set and for the protocol analysis. On the left is the primary analysis set which shows that the median time to liberation for high-flow nasal cannula was 50.5 hours whereas for CPAP it was 42.9 hours and following adjustment the adjusted hazard ratio was 0.83 and the lower limit of the 95% confidence interval was 0.7 which fell below the non-inferiority margin of 0.75. Similar findings were seen in the per protocol analysis where the median time to liberation was 50.5 hours for high-flow and 42.9 hours for CPAP with an adjusted hazard ratio of 0.82 and a 95% confidence interval lower limit of 0.68 falling below the non-inferiority margin of 0.75. A number of secondary outcomes were examined and as you can see there were no significant differences in key outcomes such as re-intubation at 48 hours, 13% in high-flow and 11% in CPAP, the mean comfort score whilst on randomized treatment, sedation used during non-invasive respiratory support, mean parental stress score, mean duration of stay in the PICU, mean duration of acute hospital stay and PICU mortality. There were however significant differences in mortality at day 60 and day 180 which as you can see there represent proportions of 4% in the high-flow and 1.2% in the CPAP at 60 days and 5.6% versus 2.4% at 180 days the significance of which is unclear and clearly these findings are not powered to show a difference. The subgroup analysis looking at age, comorbidities, duration of prior invasive ventilation, main reason for invasive mechanical ventilation, nature of respiratory support in terms of planned or rescue and baseline ratio of saturations to FiO2 did not indicate any subgroups where there was an interaction between the treatment allocation and the subgroup. So in conclusion high-flow compared to CPAP did not meet the preset non-inferiority criterion for time to liberation from respiratory support in extubated children. These findings were seen in the main analysis and they were consistent in the protocol analysis and in subgroup analysis. The high proportion of patients who were extubated and received planned non-invasive respiratory support nearly 40% of extubated patients indicates the need for more protocolized approaches to post-extubation support in children which in the absence of evidence is variable practice. I'd like to thank the entire first ABC trial team, the oversight committees, the clinical trials unit at ICNHRC and all the staff there, the staff at all participating sites and the patients and families who consented to being part of the trial because without them this trial would not have been possible. Thank you very much for the opportunity to talk to you today about the first ABC trial.

ABC trial

high-flow nasal cannula

CPAP

respiratory support

randomized controlled trial