Recent discussion in Currents has raised important questions about how ventilation strategies should be interpreted and applied in neurocritical care. Below is a Letter to the Editor from Dr. Joseph Shiber, a response from the study authors, Dr. Julian Daza and Dr. Shaurya Taran, and a concluding Editor’s Note from Jordan Yakoby that brings broader themes into focus. Together, these pieces show how scholarly discussion can sharpen interpretation of emerging evidence and its relevance to practice.

Original Article
This exchange responds to Low-Tidal-Volume Ventilation and Mortality in Patients With Acute Brain Injury, which examined the association between tidal-volume strategy and ICU mortality in mechanically ventilated patients with acute brain injury. In a secondary analysis of a large prospective observational study, the authors reported that patients receiving low-tidal-volume ventilation, defined here as ≤8 mL/kg predicted body weight, had lower adjusted ICU mortality than those receiving higher tidal volumes. Read the original feature here.

Letter to the Editor

To the editor,

I write to address a recent article featured in Currents entitled Low-Tidal-Volume Ventilation and Mortality in Patients With Acute Brain Injury. This article was not a trial, but a re-evaluation of data from a previous study using a revised definition of low tidal volume (TV): 6-8 cc/kg of ideal body weight (IBW). Among acute brain injury (ABI) patients, it has been found repeatedly that there are worse outcomes when 6 cc/kg IBW and below are used compared to 8-11 cc/kg. Furthermore, even among general ICU patients, using a tidal volume of 6 cc/kg has never been shown to provide better outcomes, except for tidal volumes of 12 cc/kg or higher.1-3  In recent years, even the definition of ‘low tidal volume’ (LTV) has changed from 6 or less to 8 cc/kg or less which accurately reflects the study group in the most commonly referenced ARDSnet trial (ARMA 2000). That trial compared 6-8 cc/kg to 12 cc/kg. Importantly, the study protocol allowed up to 8 cc/kg if hypercarbia or tachypnea developed, though the common misperception is that the LTV group was 6 cc/kg.

The ARMA study was criticized heavily at publication for not having a control group, since all patients initially were receiving 10cc/kg before being randomized to either low or high TV groups (total number enrolled 861). Only later was the data released of more than 2500 patients who remained on 10 cc/kg; though they were not enrolled, their progress was followed and results showed the same mortality rate as the LTV group. On more detailed analysis, the crucial factor was the patient’s lung compliance. If compliance was decreased then going down from 10 cc/kg to 6-8 decreased mortality. However, if compliance was preserved then going from 10 to 6-8 actually increased mortality.4

This perception that low tidal volume is the only basis of lung protective ventilation is only slowly changing. For several decades it has been known that if the plateau pressure is kept within a safe range, then the TV is not a pertinent factor. The idea of tailoring not only the TV but the entire minute ventilation (MV) prescription to the individual patient’s unique condition and physiology is the basis of “personalized medicine”.  Additionally, the Drive Pressure (Plateau Pressure – PEEP) has been shown to be the primary independent predictor of outcomes with ARDS in recent trials irrespective of TV.6,7 

The authors of the CHEST  article used a variety of statistical methods to generate their conclusion of a relative risk of 60-day mortality of 40.2% in the LTV group (mean TV was 6.8) versus 59.7% in the high TV group (mean TV 9.0) which sounds more dramatic than the actual 60-day mortality rates of 7.1% versus 8.8% respectively (calculated based on a total 60-day mortality of 8.1% and the known N# of each group, though not provided in the study).  This 1.7% increase in 60-day mortality rates, which also included withdrawal of care, is more likely to be explained by the fact that this was not an RCT but retrospective evaluation of patient data. The high TV group was at increased risk of death by baseline characteristics, such as a mean age 5 years older, 50% higher rates of chronic lung disease and diabetes, 30% higher rate of hypertension, and a 17 point lower initial mean P/F ratio, rather than by the 2.2 ml TV difference since there were no differences between groups in plateau pressure nor drive pressure which again have been shown to be the independent predictors of mortality.

Author Response: Tidal Volume and Clinical Outcomes in Acute Brain Injury: The Devil in the Details

Julian Daza MD PhD^1,2, Shaurya Taran MD^2,3,4

¹ Division of General Surgery, Department of Surgery, University of Toronto, Toronto, ON
² Institute of Health Policy, Management, and Evaluation, University of Toronto, Toronto, ON
³ Department of Medicine, University Health Network, Toronto, ON
⁴ Interdepartmental Division of Critical Care Medicine, University of Toronto, Toronto, ON

Corresponding author:

Shaurya Taran
ORCID: 0000-0001-7639-0365
Toronto Western Hospital
Office 411-L, 2^nd Floor McLaughlin
399 Bathurst St, M5T 2S8
Toronto, ON, Canada
shaurya.taran@mail.utoronto.ca

We thank Dr. Shiber for his thoughtful engagement with our work.¹ The management of mechanical ventilation in acute brain injury (ABI) remains an area of clinical equipoise, and we welcome constructive dialogue. Several points raised in the letter, however, warrant clarification.

1. Tidal volumes in ABI: clarifying the definition

Neither the present analysis¹ nor the parent study² used a “revised definition” of low tidal volume (Vt) of 6 – 8 mL/kg predicted body weight (PBW). We compared a treatment strategy of Vt ≤ 8 mL/kg predicted body weight (PBW) to a strategy of > 8 mL/kg PBW and selected the 8 ml/kg PBW threshold because contemporary surveys and cohort studies previously found that tidal volumes close to (or lower than) 8 ml/kg PBW are already common in ABI patients regardless of ARDS status.^3-5 This was also recently corroborated by the VENTIBRAIN study, in which median Vt at ICU admission among 2,095 ABI patients across 26 countries was 6.5 mL/kg PBW (interquartile range, 5.7 – 7.3ml/kg PBW).⁶

2. Vt ≤ 6 mL/kg has not been “repeatedly” shown to worsen outcomes in ABI

The letter states that outcomes are “repeatedly” worse when 6 mL/kg PBW or lower is used in ABI. We respectfully disagree. Evidence for harm is based largely on the PROtect Lung in Acute Brain Injury (PROLABI) trial, which compared a fixed strategy of Vt = 6 mL/kg PBW to ≥ 8 mL/kg PBW.⁷ In this study, the composite of mortality, ARDS, and ventilator dependence at day 28 was higher in the low Vt group. While PROLABI provides preliminary data that aggressive Vt reduction may not be prudent, caveats of the trial include its modest sample size (the trial was stopped after enrolling 36% of its target), a composite primary outcome including ARDS diagnosis (with low interrater reliability), and uncertain physiologic rationale for the observed differences.⁸ In the VENTIBRAIN observational study, lower tidal volumes were also linearly associated with higher ICU mortality, but a specific Vt threshold was not evaluated, and the findings are applicable to a cohort where patients were already receiving very low Vt (50% of patients had Vt<6.5ml/kg PBW at ICU admission).⁶ We agree that available data tempers enthusiasm for aggressive Vt reduction in the ABI population outside of well-defined indications (e.g., concomitant ARDS). However, additional data are needed to substantiate the claim that  ≤ 6 mL/kg has been “repeatedly found” to be harmful.

3. Driving pressure: important, but with caveats

Current analyses linking driving pressure (ΔP) to mortality are largely based on observational data. The study by Amato et al.⁹ was a re-analysis of nine previously published ARDS RCTs, not a trial in itself. To date, there is no randomized trial among patients with respiratory failure demonstrating that targeting ΔP improves survival, and current ARDS guidelines do not address using ΔP to guide ventilation strategies.^10,11 Moreover, ΔP represents the ratio of Vt to respiratory system compliance (ΔP = Vt / Crs). Thus, higher ΔP may reflect worse compliance rather than a causal driver of injury, and the mediation analysis employed in the Amato study does not decouple this relationship. That said, we agree there is a strong mechanistic rationale to limit ΔP in clinical practice, including in patients with ABI. Recent observational data suggest associations between higher ΔP and worse outcomes in neurocritical care cohorts,^12,13 but definitive interventional data are not available.

4. Baseline imbalances and time-varying confounding

The letter argues that baseline differences (age, comorbidities, P/F ratio) account for the mortality signal from our analysis. These variables were explicitly incorporated into stabilized inverse probability of treatment weights within a marginal structural model, in which exposure groups were weighted on measured baseline and time-varying covariates. No observational study fully eliminates confounding; we acknowledge this clearly. However, the suggestion that the results reflect crude baseline differences does not accurately represent the analytic strategy employed.

5. Pseudo-population mortality estimates are not raw event rates

The letter compares weighted mortality estimates from our analysis (40.2% vs 59.7%) to crude mortality rates (7.1% vs 8.8%) reported in the parent study.² This comparison is methodologically inappropriate. The 40.2% and 59.7% values represent counterfactual cumulative incidence estimates in a weighted pseudo-population under sustained adherence to a specific Vt strategy (always ≤ 8 mL/kg PBW vs always > 8 mL/kg PBW), with censoring for ARDS and ICU discharge. These are not raw observed event proportions; rather, they are model-based estimates derived from stabilized weights. Conflating these distinct quantities risks misunderstanding the causal estimand.

6. Randomized trials do not automatically guarantee causal validity

The letter implies that because our study is observational, its findings are inherently inferior to those of randomized trials. While randomization is a powerful design tool, causal inference is not guaranteed by RCT status alone. Among other challenges, causal claims from RCTs may be limited by underpowering, early termination, loss to follow-up, protocol non-adherence, and restricted inclusion criteria limiting generalizability. Conversely, well-conducted observational studies using appropriate causal inference methods can approximate target trial frameworks under explicit assumptions. When conducted carefully, observational analogues can yield very close estimates to the RCTs they seek to emulate.¹⁴

In conclusion, we agree that optimal ventilatory targets in ABI remain uncertain. We do not advocate indiscriminate reduction of Vt in this population. Rather, our findings suggest that avoiding Vt > 8 mL/kg PBW may be associated with lower ICU mortality in a predominantly non-ARDS ABI cohort—a practice that already seems to be widely employed.

Editor’s Note

The article Low-Tidal-Volume Ventilation and Mortality in Patients With Acute Brain Injury examined the association between tidal-volume strategies and mortality in mechanically ventilated patients with acute brain injury (ABI). In this secondary analysis of an prospective observational study, the authors reported that tidal volumes ≤8 mL/kg predicted body weight were associated with lower ICU mortality compared with tidal volumes >8 mL/kg during the first seven days of ventilation.

The points from the Letter to the Editor and the authors’ response highlight an important overarching clinical management issue in neurocritical care, namely, whether lung-protective strategies derived from ARDS populations can or should be applied broadly to neurocritical care patients, whose physiologic priorities, such as intracranial pressure, carbon dioxide management, and cerebral perfusion, may differ.

Takeaway Points for the Neurocritical Care Community

Several practical insights emerge from this exchange:

1. The body of evidence is still developing. Optimal ventilatory strategies in patients with acute brain injury remain uncertain, particularly in those without ARDS.
2. Causal-inference techniques cannot fully eliminate confounding.
3. Avoiding clearly high tidal volumes may be prudent. The current analysis suggests a potential association between tidal volumes >8 mL/kg and worse outcomes.
4. Ventilation should remain individualized. Factors such as lung compliance, driving pressure, gas exchange targets, and cerebral physiology must all be considered in neurocritical care.
5. Randomized trials are still needed. The discussion underscores the need for ongoing prospective studies specifically designed to evaluate lung-protective ventilation strategies in ABI populations.

The Bottom Line

This scholarly exchange highlights the evolving tension between lung-protective ventilation principles and neurophysiologic priorities in patients with acute brain injury. While the study adds evidence suggesting potential harm from higher tidal volumes, the debate reinforces that the optimal ventilatory strategy in ABI remains unresolved and likely requires individualized, physiology-guided care until stronger prospective evidence emerges.

Acknowledgements

We thank Dr. Shiber for his thoughtful critique of the study and for highlighting important considerations in the interpretation of ventilatory strategies in patients with acute brain injury. Scholarly discourse of this nature strengthens the scientific process by encouraging careful examination of methodology, physiology, and clinical implications.

We also thank the study authors for their detailed and collegial response, which further clarifies the analytic approach and contextualizes their findings. Exchanges such as this exemplify the constructive dialogue that advances neurocritical care practice and research. Such discourse also contributes to the mission of Currents in examining emerging evidence and advancing the practice of neurocritical care.

Jordan Yakoby, EdD, DNP, ACNP-BC, CCRN, CNE, FNYAM, FCCM, FNCS
Editor-in-Chief, Currents