Journal club 10/4: Rest after SBT and antibiotic duration in VAP

The article

Fernandez 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 trialIntensive Care Med. 2017 Sep 22. doi: 10.1007/s00134-017-4911-0. [Epub ahead of print]

What’s the question we’re looking at? Why do we care in CCM?

Does it improve chances of successful extubation if you let the patient “rest” on the vent for a short period after the SBT but before pulling the tube?

Extubation is difficult to get right. Do it too early, and patients require reintubation, which places them at risk for morbidity and mortality. Do it too late, and patients stay on the vent unnecessarily, accruing cost and complications. We use spontaneous breathing trials to help predict success, but we still get it wrong sometimes.

If there is a simple trick which might improve our chances of success, we should use it.


Some (not all) respiratory failure is caused by diaphragmatic weakness. This raises the question of whether a diaphragm fatigued by working at its limits needs to “rest” before it can perform again at full capacity. One type of fatigue, “high frequency contractile fatigue,” probably resolves in 10-15 minutes of rest, but “low frequency contractile fatigue” needs over 24 hours to fully recover. However, half of that recovery occurs within the first hour, so a single hour-long rest theoretically offers a lot of benefit.

What type of study was this?

Multicenter, prospective, randomized controlled trial.

Who was the study population? What were the inclusion/exclusion criteria?

Patients in 17 Spanish Med/Surg ICUs who had been mechanically ventilated for 12+ hours.


  • Age <18 (peds)
  • Tracheostomy (no extubation needed)
  • Overwhelming respiratory secretions (predictor of failure)
  • Inability to follow commands (mental status eval, a common rule-out for extubation)
  • DNR/DNI (hmmm?)
  • Out-of-protocol extubation (no way to perform the intervention)
  • Participation in other trials (confounder)
  • Formal indication for planned extubation to non-invasive ventilation

What was the intervention? What was the control?

Patients were wean screened as usual. If they met criteria for SBT, they were block-randomized by a Web-based system to receive a usual care SBT according to institution protocol/provider preference (usually 30, 60, or 120 minutes of T-piece or PS). If they passed, then they received either:


  • Immediate extubation


  • 1 hour back on the ventilator on full support (whatever settings they had previously), then extubation

They were also stratified (to help multivariate analysis) before SBT into either “high risk” or “low risk” by provider judgment.

All other aspects of the process proceeded per usual care, including reintubation or rescue NIPPV.

What is the PRIMARY outcome?

Reintubation within 48 hours

What were the SECONDARY outcomes?

  • Postextubation respiratory failure
  • ICU and hospital lengths of stay
  • ICU and hospital mortality

Were patients blinded? Providers? Evaluators/analysts?

Patients and clinical staff not blinded. The data was gathered by investigators with no clinical involvement, and the analysis was done by third parties with no involvement with the study.

Were the groups initially similar?

Mostly, including the number of “high risk” patients and the duration of intubation. There were some more medical-type patients, and a few more comorbidities in the intervention group (slightly significant), which would probably bias against the intervention, since these patients tend to be harder to liberate.

T-piece trial done in 87% (control) vs 94% (rest), which is more T-piece trials than we tend to do here.

No differences in technique between groups, except that oddly, significantly more in the control group received 120-minute trials versus other durations.

Other than the therapy under investigation, is there any reason why patients might have been treated differently in the two groups, and was this controlled? If not, what effect might it have?

Since it was unblinded, it is conceivable that the treating team would have handled the groups differently — for instance, if they “believed” in the intervention, giving patients in the intervention group “more of a chance” before reintubating them.

Were there any losses/failures after enrollment, and if so, were they analyzed using intention-to-treat?

None lost after randomization. (Many lost before, due to exclusions and denial of consent.)

However, 3.7% of patients in the intervention group did not tolerate the rest period (agitation after being reconnected to vent), and were thus immediately extubated. Intention-to-treat was used for them.

What are the results?

470 patients randomized, due to “miscalculation,” since their power calculation demanded 1372.

Primary outcome (reintubation in 48 hours)

  • Control: 14%
  • Intervention: 5%
  • P<.001
  • ARR: 9% reduction
  • RRR: 64% reduction
  • NNT: 12 to prevent one reintubation

Most common reason for failure: secretions leading to respiratory failure

The patients who didn’t tolerate reconnection to the vent were all extubated immediately and all did fine. (Intolerance of the rest period may predict a robust patient who will do well? In any case, they were not harmed.)

Secondary outcomes

  • Respiratory failure within 48 hours: 24% Control, 10% Intervention (p<.001). Similar number in each group got rescue NIV (about half).
  • ICU and hospital LOS: No difference
  • 48-hour mortality: No difference

Multivariate analysis including age, APACHE II score, admission reason, type/duration of SBT, and time on ventilator: only significant predictors of reintubation were the intervention (rest period), APACHE score, and duration on vent.

The only subgroup differences were that the high-risk patients seemed to benefit the most from rest (low risk group alone was not a significant difference), and the intervention group was more likely to fail due to level of consciousness (??).

Median time on vent: 5 days

Who sponsored the study? Any reason to suspect bias? Conflicts of interest among the authors?

No statements of conflicts or funding, but hard to imagine anyone making money from this.

Our take-home

  • Caveat: T-piece trials are probably the most “tiring” SBT method, and often not our choice here, so benefit may have been exaggerated in their cohort. (But for what it’s worth, type of SBT did not emerge from the multivariate analysis as a significant factor.)
  • Nevertheless, resting patients for about an hour after a successful SBT might improve their chance of successful extubation, especially those considered high-risk. The data is not rock-solid, but this is a benign intervention. We should consider doing it.


The article

Chastre et al. Comparison of 8 vs 15 Days of Antibiotic Therapy for Ventilator-Associated Pneumonia in AdultsJAMA. 2003 Nov 19;290(19):2588-98.


Antibiotic durations are often arbitrary. When it’s given thought, VAP is often treated with prolonged courses, 10-14 days at least. Can we treat for less time? Does it depend on the organism?

What type of study was this?

Randomized, controlled, double-blinded trial in 51 French ICUs.

Who was the study population? What were the inclusion/exclusion criteria?

Adult ICU patients mechanically ventilated 48+ hours with:

  1. A persistent new infiltrate on chest imaging, and
  2. Purulent secretions, fever, or WBC count >10000, and
  3. Positive bronch culture from BAL or protected brush specimen/catheter (>10^3 or ^4 on quantitative cultures), and
  4. Appropriate antibiotics started within 24 hours (i.e. excluded patients if the initial antibiotic choice ended up not covering the organisms).

In other words, true pneumonias with correctly-selected antibiotics.

Excluded patients with pregnancy, enrolled in another trial, low chance of survival (SAPS II >65), neutropenic, AIDS or other immunosuppression, other infection needing antibiotics, or not receiving full curative care.

Importantly: patients with pneumonia within 5 days of initiating ventilation and who had not already received antimicrobials within the prior 15 days were excluded, since their organisms were expected to be so sensitive.

What was the intervention? What was the control?


  • 15 days of antibiotics


  • 8 days of antibiotics

Staff remained blinded until day 8, when they would be told whether to continue antibiotics or not by fax.

Antibotic regimen per clinician preference, but recommended to initially include broad-spectrum beta lactam plus either an aminoglycoside or fluoroquinolone, then narrowing based on culture results.

Rigorous surveillance to look for recurrence of same criteria after antibiotics, in which case bronchoscopy was repeated. Bronch was also done for any unexplained worsening of oxygenation or pressor requirements.

What were the outcomes measured?


  • Death
  • Recurrence of pneumonia, defined the same as the initial criteria for pneumonia (growth from a repeat bronch, either the same or a new organism, along with the other criteria)
  • Antibiotic-free days within first 28 days

Various secondary outcomes including vent-free days, ICU stay, etc.

Powered and analyzed for non-inferiority.

What are the results?

402 patients enrolled

No significant difference in mortality (about 18% in both groups)

No significant difference in rate of pneumonia recurrence (27%) and mean time to recurrence.

No signal with multivariate analysis, except: infections with non-lactose-fermenting Gram negative bacilli had more recurrence (i.e. same organism regrew) in the short-course group (40.6% vs 25.4%; 15.2% absolute hazard, 37.4% relative hazard).

No effects on secondary outcomes including organ dysfunction, vent-free days, ICU stay, etc, including in the subset of patients with NLFGNB infection. No difference in any outcome for MRSA pneumonias.

Significant reduction in antibiotic days, and MDR pathogens developed significantly less often in short-course group (42.1% vs 62.3%, p=.04)

Who sponsored the study? Any reason to suspect bias? Conflicts of interest among the authors?

Public money, no conflicts declared.

Our take-home

  • For routine purposes, treat VAP for about 8 days. This is probably the standard of care now.
  • For NLFGNR, or if your initial antibiotics ended up being inappropriate, consider extending the course.

Leave a Reply

Your email address will not be published. Required fields are marked *