Mechanical Ventilation in the Emergency Department: A Call to Action in a Resource-constrained Era


  • The authors have no relevant financial information or potential conflicts of interest to disclose.
  • Fuller et al. conducted their study before new consensus definitions of mild, moderate, and severe acute respiratory syndrome (ARDS) replaced the terms ALI and ARDS for pO2/FiO2 ratios 200–300, 100–200, and <100, respectively. This editorial maintains the authors’ use of the prior terminology.
  • A related article appears on page 659.

“Mechanical ventilation and acute lung injury in emergency department patients with severe sepsis and septic shock: an observational study,” by Fuller et al. in this month's Academic Emergency Medicine,[1] deserves a large audience of emergency physicians (EPs), intensivists, respiratory therapists, nurses, and administrators. Airway management and initiation of mechanical ventilation is the most important task EPs perform in the care of the critically ill and injured, and the authors found that EPs in a large academic medical center rarely used lung-protective low tidal volume (Vt) ventilation for patients with severe sepsis, even when these patients had acute lung injury (ALI)* on presentation. Vt varied between 5 and 15 mL/kg and averaged 9 mL/kg; in only 30% of patients was plateau pressure (Pplat or end-inspiratory static pressure) measured. The Acute Respiratory Distress Syndrome Network (ARDSNET) protocol recommends initial Vt between 6 and 8 mL/kg of height-based ideal body weight to maintain plateau pressures below 30 cm/H2O. If Pplat rises above 30 cm/H2O, the protocol recommends reducing Vt from 6 mL/kg to as low as 4 mL/kg and increasing respiratory rate into the mid-30s as tolerated.[2]

Fuller et al. focused on Vt and Pplat, but did not comment on positive end-expiratory pressure (PEEP) or fraction of inspired oxygen (FiO2).[1] Both are integral to the ARDSNET protocol, which recommends that PEEP increase as ALI worsens[2] because repeated alveolar inflation and collapse causes injury, as does overdistension and traction on alveoli adjacent to atelectatic lung.[2, 3] To the extent that added PEEP decreases atelectasis and aerates well-perfused but injured lung segments, it can reduce shunt fraction, which causes most clinically significant hypoxemia.[3] The most conservative ARDSNET PEEP recommendations range from 5 cmH2O with 30% to 40% oxygen to 18 cmH2O with 100% oxygen.[2, 3] Finally, the protocol accepts pH values as low as 7.2, pCO2 as high as 70 mm/Hg, and SaO2 between 88% to 92%, because minimizing ventilator induced lung injury (VILI) lowers mortality in ALI, while efforts to normalize blood gas values do not.[2, 3] In this current study, the overwhelming majority of patients received the same PEEP/FiO2 setting: 5 cmH2O and 100% oxygen.[1]

Although the paper is a retrospective cohort from a single center, it is important because the authors studied a major academic emergency department (ED), where EPs are perhaps most likely to make well-informed clinical decisions. Also important is that the authors used rigorous inclusion criteria and examined ventilation practices in 251 patients over a 5-year period. Clinically, ARDSNET ventilation is one of the few critical care therapies repeatedly shown to reduce mortality,[2, 4] and ED boarding of intensive care unit (ICU) patients is now common.[5, 6] As the authors point out, ventilator settings initiated in the ED often transfer to the ICU,[1] where patients may spend days or weeks before being liberated from mechanical ventilation. Finally, patients intubated in the ED almost always have severe illness and injury accompanied by an inflammatory cytokine cascade that, with VILI, comprises the “two hits” that predisposes to ALI.[2, 3] As with trauma and sepsis, ALI may have a “golden hour” that we are letting slip away.[1]

Some caveats deserve mention. First, as the authors note, ARDSNET ventilation is not standard of care for all emergently intubated patients.[7] Although animal, biomarker, and retrospective human data suggest that low Vt and progressive PEEP settings minimize VILI,[7] prospective trials have not yet provided unequivocal support for ARDSNET ventilation as prophylaxis against ALI.[7] In addition, ARDSNET can conflict with other therapeutic goals. These include optimizing pCO2 and pO2 in patients with high intracranial pressure, stroke, or myocardial ischemia; offsetting potentially life-threatening metabolic acidoses; and treating severe obstructive disease. Patients in the early hypermetabolic phase of critical illness may develop air-hunger with ARDSNET ventilation, and this can cause agitation, ventilator asynchrony, high sedation requirements,[7] and catastrophic early self-extubation.

Even taking these considerations into account, the results of the series by Fuller et al. are disturbing, because the constellation of practices (lack of adherence to recommended Vt in patients presenting with ALI, failure to measure Pplat or change settings, and a nearly universal PEEP/FiO2 formula of 5 cm H2O of PEEP and 100% oxygen[1]) suggests poor management rather than careful deliberation about risks and benefits.

The retrospective study design and high mortality in the non-ALI group limits one's ability to make causal inferences. That said, it is clear that 22 patients with ALI in the ED were subject to ventilation known to increase their risk of death.[1, 2, 4, 7] Patients with high body mass index (BMI) were more likely to develop ALI than other patients in this series. This is important, because high BMI per se does not increase one's chances of developing ALI.[8] Calculating Vt from actual body weight yields higher Vt than would the height-based ARDSNET protocol,[2] and this may explain the observation by Fuller et al. of increased ALI risk in high-BMI patients. Any injury from high Vt was probably compounded by the universal 5 cmH2O PEEP setting, which ignores known effects of chest wall compliance in the obese and probably left most high-BMI patients with zero or even negative effective PEEP.[9]

While short periods of ventilation with 100% oxygen have not been proven to cause harm, they may worsen shunt/atelectasis and cause free radical damage, in this case without therapeutic justification. Perhaps more important, patients ventilated with 100% oxygen almost always have SaO2 at or near 100%. In the absence of frequent arterial blood gas measurements, this masks developing lung pathology that clinicians would want to address before patients get so sick that they desaturate on pure O2. Because of these masking effects, these patients’ true FiO2 requirements remain unknown, but at least some with severe sepsis must have required more than 40% oxygen and, with that, more than 5 cmH2O of PEEP.

Why was mechanical ventilation performed so poorly in this major academic ED? There are several possibilities. First, EPs are not trained to provide ongoing care for critically ill patients, and even intensivists with such training often fail at this task.[4] Although the ARDSNET protocol is easy to read from an excellent summary card (, implementing it and integrating it with other goals is much more difficult.[7] Emergency medicine is a broad field; while many non–fellowship-trained EPs are experts in resuscitation and initiation of critical care, others are competent but less well versed. In contrast to office or consulting practices, anything can and does come through the door, and staffing needs are such that almost every EP manages critically ill patients.

Emergency physicians attending to patients coming through that door are compelled to pay more attention to trauma notifications, STEMIs, and stroke codes than to the insidious process of VILI. Burdensome nondirect patient care tasks currently force EPs to spend the majority of their clinical time away from the bedside.[10] Fewer EDs now care for more patients, and ED nursing ratios are a fraction of those in the ICU. Because of this, ventilator settings may be relegated to an afterthought once intubation and initial stabilization are established for a patient who may have been near death only minutes before. The data in this paper[1] suggest that this is what occurred, and the authors and their leadership should be applauded for bringing it to light. From here, we must work toward a solution.

Systemwide resource fixes are not forthcoming. We live in an era of hospital closings and budget cuts, and the United States already has seven times as many ICU beds per capita as the UK,[11] so the supply is unlikely to increase anytime soon. Unfortunately, dedicated ED-ICU staff and space is not economically or logistically viable. While we may sense that ICU care will not meaningfully affect a particular patient's downward spiral, no regulatory means to allocate scarce ICU resources will emerge in the near future—the recent “death panel” scare created by opponents of the Affordable Care Act offered a sobering lesson for would-be policymakers. Only the advance of palliative care and other measures that offer desirable alternatives to aggressive resuscitation at life's end will reduce the demand for ICU care.

Finally, most EPs cannot find additional time to learn how to titrate ventilator settings with everything else they must know and do. Although Fuller et al. point to early goal-directed therapy (EGDT/surviving sepsis)[12] as a model for improving ventilator management,[1] severe sepsis resuscitation is more front loaded and vital sign based than ARDSNET[2, 12] and therefore more familiar to EPs. In addition, EGDT is not translated into practice as frequently as Fuller et al. imply,[5] and it is even less likely that EPs will succeed with the complex demands of the ARDSNET protocol, given that intensivists often fail[4] with more training and resources.

The failure of physicians acting independently to implement evidence-based mechanical ventilation has forced ICU teams to develop successful strategies that may provide a model for the ED.[13, 14] ICU teams have created nurse- and respiratory therapist–led protocols that promote compliance with use of low Vt,[13] appropriate sedation,[14] and liberation from mechanical ventilation.[14] The latter protocols require physician-independent titration of both sedation/analgesia and ventilator settings.[14]

At present, emergency medicine critical care (EMCC) specialists can become board certified in internal medicine-critical care and certified in neurocritical care. Anesthesia and surgical critical care subspecialties are likely to follow, and there are many other EMCC physicians who trained before accreditation was possible. These EP-intensivists are especially well prepared to develop clinical pathways with ED-based respiratory therapists that trigger adherence to ARDSNET ventilation or other protocols as indicated. Because nurse:patient staffing ratios in the ED do not match those of the ICU, such protocols must include sedation/analgesia order sets that respiratory therapists can suggest to EPs so they can address agitation and asynchrony. As with any cultural change, early intense meetings between EP-intensivists or EPs with expertise in mechanical ventilation, ED medical and nursing leadership, EPs, nurses, and respiratory therapists must be followed by regular, shorter meetings that provide feedback and facilitate a two-way exchange of information and ideas[14] (following the familiar Shewart cycle: Plan, Do, Check, Act[15]). Such an intervention adds little additional resource demand because it concentrates policy formation and clinical action into a few very capable hands. Such a solution may make a dent in the problem highlighted by Fuller et al. and provide a model for addressing the larger problem of increased mortality in ED/ICU boarders.[6] By calling attention to this important issue in their own institution, Fuller et al. have courageously started a conversation that may benefit many critically ill patients in the United States and beyond.

The author acknowledges Drs. Michael Hockstein, Michael Khilkin, Todd Slesinger, and Brian Wright.