Since the landmark ARMA trial, use of low tidal volume ventilation (LTVV) at 6 mL/kg predicted body weight (PBW) has become our gold standard for ventilator management in acute respiratory distress syndrome (ARDS) (Brower RG, et al. N Engl J Med. 2000;342:1301). While other studies have suggested that patients without ARDS may also benefit from lower volumes, the recently published Protective Ventilation in Patients Without ARDS (PReVENT) trial found no benefit to using LTVV in non-ARDS patients (Simonis FD, et al. JAMA. 2018;320:1872). Does this mean we let physicians set volumes at will? Is tidal volume (VT) even clinically relevant anymore in the non-ARDS population?
Prior to the PReVENT trial, our practice of LTVV for patients without ARDS was informed primarily by observational data. In 2012, a meta-analysis comparing LTVV with “conventional” VT (10-12 mL/kg IBW) in non-ARDS patients found that those given LTVV had a lower incidence of acute lung injury and lower overall mortality (Neto AS, et al. JAMA. 2012 308:1651). While these were promising findings, there was limited follow-up poststudy onset, and the majority of included studies were based on a surgical population. Additionally, the use of VT > 10 mL/kg PBW has become uncommon in routine clinical practice. How comparable are those previous studies to today’s clinical milieu? When comparing outcomes for ICU patients who were ventilated with low (≤7mL/kg PBW), intermediate (>7, but <10 mL/kg PBW), and high (≥10 mL/kg PBW) VT, a second meta-analysis found a 28% risk reduction in the development of ARDS or pneumonia with low vs high, but the similar difference was not seen when comparing low vs intermediate groups (Neto AS, et al. Crit Care Med. 2015;43:2155). This research suggested that negative outcomes were driven by the excessive VT.
Slated to be the definitive study on the matter, the PReVENT trial used a multicenter randomized control trial design comparing target VT of 4 mL/kg with 10 mL/kg PBW, with setting titration primarily based on plateau pressure targets. The headline out of this trial may have been that it was “negative,” in that there was no difference between the groups in the primary outcome of ventilator-free days and survival by day 28. However, there are some important limitations to consider before discounting LTVV for everyone. First, half of the trial patients were ventilated with pressure-control ventilation, the actual VT settings were 7.3 (5.9 – 9.1) for the low group vs 9.1 (7.7 – 10.5) mL/kg PBW for the intermediate group by day 3, statistically significant differences, but perhaps not as striking clinically. Moreover, a secondary analysis of ARDSnet data (Amato MB, et al, N Engl J Med. 2015;372:747) also suggests that driving pressure, more so than VT, may determine outcomes, which, for most patients in the PReVENT trial, remained in the “safe” range of < 15 cm H2O. Finally, almost two-thirds of patients eligible for PReVENT were not enrolled, and the included cohort had PaO2/FiO2 ratios greater than 200 for the 3 days of the study, limiting generalizability, especially for patients with acute hypoxemic respiratory failure.
When approaching the patient who we have determined to not have ARDS (either by clinical diagnosis or suspicion plus a low PaO2/FiO2 ratio as defined by PReVENT’s protocol), it is important to also consider our accuracy in recognizing ARDS before settling for the use of unregulated VT. ARDS is often underrecognized, and this delay in diagnosis results in delayed LTVV initiation. Results from the LUNG SAFE study, an international multicenter prospective observational study of over 2,300 ICU patients with ARDS, showed that only 34% of patients were recognized by the clinician to have ARDS at the time they met the Berlin criteria (Bellani G, et al. JAMA. 2016;315:788). As ARDS is defined by clinical criteria, it is biologically plausible to think that the pathologic process commences before these criteria are recognized by the clinician.
To investigate the importance of timing of LTVV in ARDS, Needham and colleagues performed a prospective cohort study in patients with ARDS, examining the effect of VT received over time on the outcome of ICU mortality (Needham DM, et al. Am J Respir Crit Care Med. 2015;191:177). They found that every 1 mL/kg increase in VT setting was associated with a 23% increase in mortality and, indeed, increases in subsequent VT compared with baseline setting were associated with increasing mortality. One may, therefore, be concerned that if we miss the ARDS diagnosis, the default to higher VT at the time of intubation may harm our patients. With or without clinician recognition of ARDS, LUNG SAFE revealed that the average VT for the patients with confirmed ARDS was 7.6 (95% CI 7.5-7.7) mL/kg PBW. While this mean value is well within the range of lung protective ventilation (less than 8 mL/kg PBW), over one-third of patients were exposed to larger VT. A recently published study by Sjoding and colleagues showed that VT of >8 mL/kg PBW was used in 40% of the cohort, and continued exposure to 24 total hours of these high VT was associated with increased risk of mortality (OR 1.82 (95% CI, 1.20–2.78) (Sjoding MW, et al. Crit Care Med. 2019;47:56). All three studies support early administration of lung protective ventilation, considering the high mortality associated with ARDS.
Before consolidating what we know about empiric use of LTVV, we also must highlight the important concerns about LTVV that were investigated in the PReVENT trial. Over-sedation to maintain low VT, increased delirium, ventilator asynchrony, and possibility of effort-induced lung injury are some of the potential risks associated with LTVV. While there were no differences in the use of sedatives or neuromuscular blocking agents between groups in the PReVENT trial, more delirium was seen in the LTVV group with a P = .06, which may be a signal deserving further exploration.
Therefore, now understanding both the upside and downside of LTVV, what’s our best approach? While we lack prospective clinical trial data showing benefit of LTVV in patients without ARDS, we do not have conclusive evidence to show its harm. Remembering that even intensivists can fail to recognize ARDS at its onset, default utilization of LTVV, or at least lung protective ventilation of <8 mL/kg PBW, may be the safest approach for all patients. To be clear, this approach would still allow for active physician decision-making to personalize the settings to the individual patient’s needs, including the use of higher VT if needed for patient comfort, effort, and sedation needs. Changing the default settings and implementing friendly reminders about how to manage the ventilator has already been shown to be helpful for the surgical population (O’Reilly-Shah VN, et al. BMJ Qual Saf. 2018;27:1008).
We must also consider the process of health-care delivery and the implementation of best practices, after considering the facilitators and barriers to adoption of said practices. Many patients decompensate and require intubation prior to ICU arrival, with prolonged boarding in the ED or medical wards being a common occurrence for many hospitals. As such, we need to consider a ventilation strategy that allows for best practice implementation at a hospital-wide level, appealing to an interprofessional approach to ventilator management, employing physicians outside of critical care medicine, respiratory therapists, and nursing. The PReVENT trial had a nicely constructed protocol with clear instructions on ventilator adjustments with frequent plateau pressure measurements and patient assessments. In the real world setting, especially in a non-ICU setting, ventilator management is not as straightforward. Considering that plateau pressures were only checked in approximately 40% of the patients in LUNG SAFE cohort, active management and attention to driving pressure may be a stretch in many settings.
Until we get 100% sensitive in timely recognition (instantaneous, really) of ARDS pathology augmented by automated diagnostic tools embedded in the medical record and/or incorporate advanced technology in the ventilator management to avoid human error, employing simple defaults to guarantee a protective setting in case of later diagnosis of ARDS seems logical. We can even go further to separate the defaults into LTVV for hypoxemic respiratory failure and lung protective ventilation for everything else, with future development of more algorithms, protocols, and clinical decision support tools for ventilator management. For the time being, a simpler intervention of setting a safer default is a great universal start.
Dr. Mathews and Dr. Howell are with the Division of Pulmonary, Critical Care, and Sleep Medicine, Department of Medicine; Dr. Mathews is also with the Department of Emergency Medicine; Icahn School of Medicine at Mount Sinai, New York, NY.