,
23 tweets,
4 min read
Read on Twitter
Hypoxaemia is a potential contributing factor in peri-procedural cardiac arrest and death. RSIs involve delay of 45-90 secs between administration of drugs and laryngoscopy, during which time patient becomes apnoeic.
There are concerns that bag-mask ventilation between induction and laryngoscopy may increase aspiration risk. Previous studies to investigate this have been limited to healthy anaesthetised volunteers.
Do you routinely bag-mask ventilate critically ill patients between administration of neuromuscular blocking drugs and laryngoscopy?
This was a multicentre, parallel-group, unblinded, pragmatic randomised trial to establish whether or not bag mask ventilation during intubation of critically ill patients prevents hypoxaemia.
Primary outcome: Lowest oxygen saturation observed between induction and 2 mins after intubation. Secondary outcome: Incidence of severe hypoxaemia (sats <80%).
Safety outcome: Sats, FiO2 and PEEP in 6-24 hours after intubation – worst value considered safety outcome to capture objective clinical manifestations of aspiration.
Additionally procedural outcomes: Operator-reported aspiration and new CXR opacity within 48 hours post-intubation.
Inclusion criteria: Age >18, ICU patient; require intubation; induction agents used; to be intubated by someone who routinely intubates in that ICU.
Exclusion criteria: Pregnancy; prisoners; urgency of intubation makes following protocol unsafe; clinician feels ventilation is essential (eg patient hypoxaemic) or contraindicated (eg patient vomiting)
401 patients enrolled; randomised in 1:1 ratio to bag-mask ventilation or no ventilation in permuted blocks of 2, 4 and 6, across 7 ICUs in the USA. Median age 60. ~50% were septic; ~60% had hypoxaemic respiratory failure.
Results: In BMV group, median lowest SpO2 96%, IQR, 87-99; in no ventilation group, median lowest SpO2 93%, IQR, 81-99%. 10.9% in BMV group had severe hypoxaemia vs 22.8% in no ventilation group (RR 0.48; CI 0.30-0.77).
No significant between-group difference in SpO2, FiO2 or PEEP in 24 hours after intubation, nor in-hospital mortality or number of ventilator-free days or days out of ICU.
Operator-reported aspiration in 2.5% of intubations in BMV group and 4% in no ventilation group (P=0.41). New opacity on CXR was 16.4% vs 14.8% (P=0.73).
Lowest SpO2 3.9% higher in BMV group than in no ventilation group; 12% fewer patients had severe hypoxaemia in BMV group than no ventilation group. Therefore BMV ventilation of 9 patients prevents severe hypoxaemia in 1
Low incidence of aspiration, therefore unable to determine if BV increases rate. Would require a trial of 4000 patients to establish a 50% increase in risk. However in this study the rate of aspiration in BMV group was actually lower than in no ventilation group.
However, 7.3% of otherwise eligible patients were excluded because clinicians considered them to have a high aspiration risk.
Strengths: Trial was randomised, group assignment was concealed until enrolment, trial conducted across several sites, and endpoint data collected by an independent observer.
Limitations: Unblinded; prexoygenation technique not controlled (clinicians could choose their technique post-randomisation); non-invasive ventilation not investigated at all; unclear whether results can be generalised to intubations in ED or pre-hospital.
This trial indicates, perhaps unsurprisingly, that ventilating apnoeic patients increases oxygen saturations and reduces severe hypoxaemia. Whether or not this translates to improved patient outcomes will require further investigation.
If you were not routinely using BMV during intubation of ICU patients, will you start doing so?
side note from @dogICUma
For me this is a practice-changing trial. While uncertainty remains, BMV to prevent hypoxaemia seems reasonable to me and was not associated with adverse events.
For me this is a practice-changing trial. While uncertainty remains, BMV to prevent hypoxaemia seems reasonable to me and was not associated with adverse events.
@threadreaderapp unroll please
