Evidence Exchange

SAVE-ICU: Inhaled sedation in mechanically ventilated patients with acute hypoxemic respiratory failure

This pragmatic, open-label RCT was conducted in 14 Canadian ICUs between 2020 and 2025. Adults ventilated for AHRF (PaO2:FiO2 <300 mm Hg or COVID-19) and sedated for <72 hours were randomized to inhaled isoflurane, with IV adjuncts as needed, or usual IV sedation. The primary outcome was hospital mortality.

Results 780 patients were included in the analysis. About half of the inhaled sedative group also received IV sedatives. Hospital mortality was 42.5% with inhaled vs 43.8% with IV sedation (RR 0.98, 95% CI 0.84 to 1.14). Ventilator, ICU, and hospital-free days and 3-month quality of life and disability were similar. The inhaled group used more rescue hypoxemia therapies (NMB, inhaled vasodilators, prone positioning, ECLS; 63.9% vs 55.0%; difference 8.9%, 95% CI 2.0% to 15.8%) and had fewer delirium- and coma-free days to day 14 (median 2 [IQR −1 to 9] vs 4 [IQR 0 to 10]; difference −0.95 days, 95% CI −1.69 to −0.22).

Important to note

Sedation targets were left to the teams in each ICU. 55.5 to 90.5% of patients in both groups were deeply sedated with a RASS of -4 to -5 on any given day.
This study was powered to detect a drop in mortality from 40% to 30% (a 25% relative reduction) which is quite a large mortality benefit to expect from a sedation strategy.
The severity of illness in the patients included in this study was notably higher than in previous sedation trials, due to the inclusion for patients post-cardiac arrest and after trauma. The authors hypothesize that the 26% of patients enrolled with COVID-19 also raised mortality.

Authors' Conclusion

Among mechanically ventilated adults with AHRF, hospital mortality did not significantly differ between those who received an inhaled isoflurane sedation strategy with intravenous adjuncts as clinically needed compared with usual care intravenous sedation.

Jerath A, et al. JAMA. Published online October 14, 2026. doi:10.1001/jama.2026.18471

Jerath et al, 2026. SAVE-ICU.pdf

REMAP-CAP: Oseltamivir treatment arm

Note: This evidence is summarized from a printprint that has not yet underwent peer review.

REMAP-CAP is an international, open-label adaptive platform trial for patients hospitalized with respiratory tract infections. Patients can be randomized to several treatment interventions at once. This trial uses a Bayesian design with no fixed sample size: scheduled interim analyses update the probability that each treatment is best, and an arm stops once a predefined superiority, inferiority, equivalence or futility threshold is crossed. For the oseltamivir treatment arm, hospitalised patients aged 12 years and older with laboratory-confirmed influenza who were critically ill and receiving organ support (high-flow nasal oxygen ≥30 L/min with FiO2 ≥0.4, non-invasive or invasive ventilation, or vasopressor/inotrope support), were enrolled within 48 hours of starting that support. Prior to REMAP-CAP, oseltamivir had never been compared with no antiviral treatment in a randomized trial of critically ill patients with influenza.

Results Of 442 patients randomised, recruitment stopped early when both oseltamivir arms crossed the predefined inferiority threshold. By day 90, 13.7% had died in the no-antiviral group, compared with 19.8% in the 5-day group and 19.4% in the 10-day group. The adjusted odds ratio for death was 2.13 (95% CrI 1.03–4.52) with 5 days and 2.17 (95% CrI 1.05–4.64) with 10 days, giving posterior probabilities of harm of 98.0% and 98.2%. The probability of futility was 99.5% for each arm, and no secondary outcome (90-day survival, ICU length of stay, respiratory support, organ support, and cardiovasuclar support free days, progression to invasive mechanicalt ventilation, ECMO) or subgroup (immunocompromised, bacterial co-infection, >5 days of symptoms, baseline shock) favoured oseltamivir.

Important to note

The oseltamivir arms were sicker (invasive ventilation 48–53% vs 39%; vasopressors 43–55% vs 38%), though results were adjusted.
Only 79 of the 139 sites were willing to offer no antiviral treatment, through results were adjusted for whether the site offered the no-antiviral arm.
Most patients (89%) had influenza A.
Aspergillus was also noted to be more common with oseltamivir treatment (3.7% vs 0.8%) .

Authors' Conclusion

Treatment with oseltamivir is ineffective and highly likely to increase 90-day mortality in critically ill patients with influenza.

Murthy S, et al; REMAP-CAP Investigators. SSRN preprint, July 27, 2026. doi:10.2139/ssrn.7172531

Does every patient need an Arterial Line?

Link as well to the NEJM Video

Evidence Exchange Author: Dr. Khosravani
NEJM - the world's top medical journal is partnering with Dr. Glaucomflecken! His real name is Will Flanary - he's a ophthalmologist who during COVID made great videos to make people understand and laugh at aspects of medicine.

This is a fun view - released Jan 2026 - by NEJM (New England Journal of Medicine) about the arterial line study! Published October 29, 2025

N Engl J Med 2025;393:1875-1888

Here are the details: Deferring Arterial Catheterization in Critically Ill Patients with Shock, below is directly from the NEJM website
"RESULTS: A total of 1010 patients underwent randomization; 504 patients assigned to the noninvasive-strategy group and 502 assigned to the invasive-strategy group were included in the analyses. A total of 74 patients (14.7%) in the noninvasive-strategy group and 493 (98.2%) in the invasive-strategy group underwent insertion of an arterial catheter. Death within 28 days occurred in 173 patients (34.3%) in the noninvasive-strategy group and 185 (36.9%) in the invasive-strategy group (adjusted risk difference, −3.2 percentage points; 95% confidence interval, −8.9 to 2.5; P=0.006 for noninferiority). Results of per-protocol analyses were similar in the two groups. A total of 66 patients (13.1%) in the noninvasive-strategy group and 45 (9.0%) in the invasive-strategy group had at least 1 day of pain or discomfort related to the ongoing presence of the blood-pressure–monitoring device. Hematoma or hemorrhage related to the arterial catheter occurred in 5 patients (1.0%) in the noninvasive-strategy group and 41 patients (8.2%) in the invasive-strategy group.

CONCLUSIONS: Among patients with shock, results for death from any cause at day 28 indicated that management without early arterial catheter insertion was noninferior to early catheter insertion."

Transfusion Practice

Evidence Exchange Author: Dr. Khosravani
We observe different transfusion practices in the unit.

However, It is best to standardize our approach, because the transfusion/blood bank lab have to be stewarts of this precisious resouce - and the act of transfusion comes with risks to the patient - transfusion reaction, or more severe complications such as TACO and TRALI.
If you see / or have a request that is contrary to the below - please direct the provider to this information and the Choosing Wisely website.

Here is a common one - Pt not actively bleeding? Pt not unstable?
Don’t transfuse more than one red cell unit at a time when transfusion is required in stable, non-bleeding patients.
- Transfuse 1 unit at a time!
  "Indications for red blood transfusion depend on clinical assessment and the cause of the anemia. In a stable, non-bleeding patient, often a single unit of blood is adequate to relieve patient symptoms or to raise the hemoglobin to an acceptable level. Transfusions are associated with increased morbidity and mortality in high-risk hospitalized inpatients. Transfusion decisions should be influenced by symptoms and hemoglobin concentration. Single unit red cell transfusions should be the standard for non-bleeding, hospitalized patients. Additional units should only be prescribed after re-assessment of the patient and their hemoglobin value." - Choosing Wisely Canada
Don’t transfuse plasma to correct a mildly elevated (<1.8) international normalized ratio (INR) or activated partial thromboplastin time (aPTT) before a procedure.
- A mildly elevated INR is not predictive of an increased risk of bleeding. Furthermore, transfusion of plasma has not been demonstrated to significantly change the INR value when the INR was only minimally elevated (<1.8).
- FFP itself has a INR, of somewhere between 1-1.6, or 1-1.4 so you cannot use FFP to "reverse INR"
- Remember if it's for reversing Warfarin - PCC is a better alternative + Must remember to give Vitamin K 10mg IV x 1
- FFP also comes with a lot of volume - so alternatives are for lower volume are: Cirrhosis patient?
  - "For bleeding in cirrhosis patients not on anticoagulants, both fibrinogen concentrate and prothrombin complex concentrate (PCC) are low-volume alternatives to fresh frozen plasma, but neither is routinely recommended as first-line therapy, and the choice depends on the specific clinical scenario and coagulation deficits identified." - from OpenEvidence

Page updated

Report abuse