There has long been debate about exactly how to design a spontaneous breathing trial. If the trial is made too easy, then patients would be extubated before being ready (leading to re-intubation). Alternatively, if the trial is too hard, that would prolong mechanical ventilation beyond the point of being beneficial.
The figure above shows the two key parameters of a spontaneous breathing trial – the duration and the amount of ventilatory support. For example, a T-piece trial provides zero ventilatory support, making it a very difficult trial (try breathing through a 7.5 mm endotracheal tube – seriously – it’s a bit like breathing through a straw).
Subira et al: Effect of pressure support vs. T-piece ventilation strategies during spontaneous breathing trials on successful extubation
This is a multi-center RCT comparing spontaneous breathing trials on either extreme of the spectrum: either a 30-minute trial with 8 cm of pressure support versus a 120-minute trial with zero support (T-piece)(figure below).1 Prior to randomization, treating physicians were forced to decide whether to administer BiPAP or high-flow nasal cannula after extubation. This protocol applied only to the patient’s first spontaneous breathing trial (patients who failed this spontaneous breathing trial were subsequently managed per usual care).
1153 patients were randomized. Patients were well matched at baseline (table below). Most patients were medical ICU patients.
Primary outcome: 30-minute spontaneous breathing trial for the win.
The primary endpoint was successful extubation, defined as extubation with ability to stay off the ventilator for 72 hours. Patients in the 30-minute group had a substantially higher rate of successful extubation (82% vs 74%, p<0.001, fragility index of 21!). A fragility index of 21 is extremely high, supporting the robustness of this finding.
Riddle #1: Same rate of re-intubation?
The primary outcome is far from the most interesting aspect of this paper. The really fascinating part is the fact that the reintubation rates are exactly the same (11.1% vs. 11.9%). This is counterintuitive, because one would generally expect that the more rigorous test should select for the strongest patients – who would then have a lower rate of extubation failure. A tougher test should have lower sensitivity but higher specificity for selecting patients ready for extubation.
Some potential explanations for this riddle:
- The 120-minute T-piece trial is a pretty tough trial. It’s likely that this trial itself induced diaphragmatic fatigue. So it’s conceivable that the trial selected the strongest patients but, while doing so, weakened them sufficiently that their success rate was lower than what it should have been. This possibility of exhaustion due to a T-piece is supported by a prior trial by Hernandez et al. showing that after a T-piece trial, patients were more likely to extubate successfully if they were placed back on ventilatory support to rest for a while prior to extubation.2 Notably, only 27% of the patients in the 120-minute group were rested on the ventilator after the trial and prior to extubation – which would support the possibility that these patients were fatigued after a prolonged, arduous breathing trial.
- Factors influencing the success of extubation might have nothing to do with respiratory muscle strength (e.g. ability to protect the upper airway). In that case, the specifics of the breathing trial are wholly irrelevant.
Some combination of these factors probably explains the paradoxically equivalent rate of reintubation. Regardless, the conclusion seems the same – a 120-minute T-piece trial is unnecessarily difficult and long.
Riddle #2: Improved mortality, similar ICU length of stay
Patients in the 30-minute group had a lower hospital mortality (10% vs. 15%, p=0.02, fragility index of 3). This mortality difference is believable, because the 30-minute group achieved a higher extubation rate with an identical extubation failure rate. The mortality difference is larger than I would have expected, but it does make sense that timely extubation could put a patient on a pathway towards recovery (whereas ongoing intubation could lead to further iatrogenic complications, such as ventilator-associated pneumonia and delirium).
The riddle, then, is that there was no difference in the median ICU length of stay (table above). There may be two explanations for this:
- Increased survival in the 30-minute group could sometimes increase the ICU length of stay (because patients who otherwise might have died rapidly instead survive and require prolonged ICU care).
- The median length of stay is a statistic which is relatively resistant to change. In order to shift the median, lots of patients need to shift. For example, imagine that patients with the top 25% longest ICU length of stay experienced a doubled length of stay – this won't shift the median value a bit!
Ultimately, the mortality difference is obviously more important than the length-of-stay statistic. This is mostly interesting because it highlights some potential problems with using median length of stay as an endpoint (and it ultimately emphasizes the importance of interpreting all endpoints together in context).
Finer points: Details of the spontaneous breathing trial
It’s worth taking a quick look at the criteria these authors used to define failure of a spontaneous breathing trial:
- Mental status change (Agitation/anxiety or reduced level of consciousness)
- Increased work of breathing (use of accessory muscles, facial expression suggesting stress, severe dyspnea)
- Hypoxemia (saturation <90% with FiO2 > 50%)
- Tachypnea (respiratory rate >35 breaths/minute)
- Hemodynamic instability or arrhythmia (e.g. heart rate >140 b/m, 20% increase in heart rate from baseline, or systolic blood pressure outside 90-180 mm).
It’s notable that the rapid-shallow breathing (RSBI) index wasn’t used here. This suggests that patients may be extubated despite a high rapid-shallow breathing index if they otherwise look well and meet the above criteria.
consideration of this trial in the context of existing literature
A review article recently summarized all trials comparing different techniques for spontaneous breathing trials:3
It’s clear why confusion about spontaneous breathing trials persists: this literature is a mess. Most of these studies are underpowered. Studies compare a hodgepodge of different ventilator modes, making it different to integrate all of these studies in a meaningful fashion.
Among these prior studies, two are directly comparable to the new study from Subira et al, in that they compared a T-piece vs. pressure support at 7 cm or 8 cm. These studies deserve review, to see if Subira et al is an outlier or if it is consistent with prior literature:
Esteban et al. 1997: This was an RCT involving 484 patients which compared a 2-hour T-piece trial versus a 2-hour trial on 7 cm of pressure support.4 More patients failed the T-piece trial (22% vs. 14%, p=0.03). Among patients who passed a spontaneous breathing trial, the rate of reintubation was exactly the same between groups (19%). This is precisely the same result Subira et al found: Fewer patients pass the T-piece trial, but somewhat paradoxically the reintubation rate is the same across both groups of patients (see Riddle #1 above).
Matic et al. 2004: This was an RCT involving 260 patients which compared a 2-hour T-piece trial versus a 2-hour trial on 8 cm of pressure support.5 Patients randomized to the T-piece arm were more likely to fail the breathing trial (27% vs. 20%) and eventually spent more time on ventilation and in the ICU. This study is underpowered, yet generally consistent with the studies reviewed above.
how should we practice now?
RCTs will never answer all clinical questions – which leaves us to extrapolate between studies using our best judgement. The challenge of applying evidence-based medicine is applying it in a fashion which is neither too restrictive nor too loose.
- Too restrictive: We cling rigidly to the exact protocols used in RCTs, fearing to deviate from them at all (evidence-restricted medicine).
- Too loose: If we extrapolate too much, then our practice may differ dramatically from the interventions in RCTs (evidence-inspired medicine).
Based on the available evidence, the following conclusions seem reasonable:
- T-piece trials are too difficult; they should generally be avoided.
- 30 minutes seems like a reasonable length for a breathing trial.
One question which remains is the ideal settings for the pressure support trial. Automatic Tube Control (ATC) is a mode which adjusts pressure levels to match the resistance of the endotracheal tube – this theoretically should be superior to a fixed amount pressure. Furthermore, whether or not to add PEEP remains unclear (the studies discussed above didn’t use PEEP, but some have).
Another question which remains is the effect of routine, prophylactic application of high-flow nasal cannula following extubation. Several studies have shown that routine use of high-flow nasal cannula following extubation reduces the risk of reintubation.6–8 In Subira et al, only ~15% of patients received prophylactic post-extubation high-flow nasal cannula. Theoretically, routine post-extubation high-flow nasal cannula might allow for an even easier breathing trial (e.g. 8 cm pressure support / 5 cm PEEP) to facilitate prompt extubation of more patients, while maintaining a low re-intubation rate.
A final question which this study doesn’t address is weaning patients off airway pressure release ventilation (APRV). This is a different beast altogether, which is beyond the scope of this post.
Regardless of these remaining questions, Subira et al is undoubtedly the highest quality and most modern evidence available on this topic. The study has a multi-center design, allowing it to become the largest RCT ever performed on this topic. With great power comes statistically robust conclusions (which are, reassuringly, consistent with prior studies).
Thus, this study should change practice. It establishes a 30-minute pressure support trial with 8cm/0cm as both safe and effective. It makes sense to use this protocol while awaiting further studies that may refine our practice further.
- Subira et al have performed the largest ever multi-center RCT investigating how to perform a spontaneous breathing trial. This is a methodologically sound study which should change practice.
- Use of an easier breathing trial (30-minutes on 8 cm of pressure support with 0 cm PEEP) was compared to a harder breathing trial (120-minutes with zero support). The 30-minute trial increased extubation success and reduced mortality (without increasing the risk of extubation failure).
- This study cannot prove that 8 cm pressure support with 0 cm PEEP is superior to all other modes of weaning (e.g. automatic tube compensation). However, using Subira’s protocol for breathing trials is currently a sensible approach until further data emerges.
- The Case of the False Imprisonment (Rory Spiegel, EMNerd) – Slightly different perspective on this study – a bit like the 30,000 foot view – with which I wholly agree and which you need to read.
- Other great studies by this research group reviewed on PulmCrit:
- Rethinking extubation by Scott Aberegg (Medicine Evidence Blog) – Trial of Extubation & Unknown performance of weaning parameters.
- 1.Subirà C, Hernández G, Vázquez A, et al. Effect of Pressure Support vs T-Piece Ventilation Strategies During Spontaneous Breathing Trials on Successful Extubation Among Patients Receiving Mechanical Ventilation: A Randomized Clinical Trial. JAMA. 2019;321(22):2175-2182. https://www.ncbi.nlm.nih.gov/pubmed/31184740.
- 2.Fernandez M, González-Castro A, Magret M, et al. Reconnection to mechanical ventilation for 1 h after a successful spontaneous breathing trial reduces reintubation in critically ill patients: a multicenter randomized controlled trial. Intensive Care Med. 2017;43(11):1660-1667. https://www.ncbi.nlm.nih.gov/pubmed/28936675.
- 3.Burns K, Soliman I, Adhikari N, et al. Trials directly comparing alternative spontaneous breathing trial techniques: a systematic review and meta-analysis. Crit Care. 2017;21(1):127. https://www.ncbi.nlm.nih.gov/pubmed/28576127.
- 4.Esteban A, Alía I, Gordo F, et al. Extubation outcome after spontaneous breathing trials with T-tube or pressure support ventilation. The Spanish Lung Failure Collaborative Group. Am J Respir Crit Care Med. 1997;156(2 Pt 1):459-465. https://www.ncbi.nlm.nih.gov/pubmed/9279224.
- 5.Matić I, Majerić-Kogler V. Comparison of pressure support and T-tube weaning from mechanical ventilation: randomized prospective study. Croat Med J. 2004;45(2):162-166. https://www.ncbi.nlm.nih.gov/pubmed/15103752.
- 6.Hernández G, Vaquero C, González P, et al. Effect of Postextubation High-Flow Nasal Cannula vs Conventional Oxygen Therapy on Reintubation in Low-Risk Patients: A Randomized Clinical Trial. JAMA. 2016;315(13):1354-1361. https://www.ncbi.nlm.nih.gov/pubmed/26975498.
- 7.Hernández G, Vaquero C, Colinas L, et al. Effect of Postextubation High-Flow Nasal Cannula vs Noninvasive Ventilation on Reintubation and Postextubation Respiratory Failure in High-Risk Patients: A Randomized Clinical Trial. JAMA. 2016;316(15):1565-1574. https://www.ncbi.nlm.nih.gov/pubmed/27706464.
- 8.Maggiore S, Idone F, Vaschetto R, et al. Nasal high-flow versus Venturi mask oxygen therapy after extubation. Effects on oxygenation, comfort, and clinical outcome. Am J Respir Crit Care Med. 2014;190(3):282-288. https://www.ncbi.nlm.nih.gov/pubmed/25003980.
- IBCC chapter:Guide to APRV for COVID-19 - April 8, 2020
- PulmCrit Theoretical Post – The COVID Severity Index (CSI 1.0) - April 2, 2020
- PulmCrit wee – Why the SCCM/AARC/ASA/APSF/AACN/CHEST joint statement on split ventilators is wrong. - March 29, 2020