256 Perioperative Management of the Patient With Acute Respiratory Distress
Page 1 Syndrome (ARDS)
Avery Tung, M.D. Chicago, Illinois
The term “Acute Respiratory Distress Syndrome (ARDS)” describes a characteristic pattern of diffuse lung
injury resulting in impaired lung mechanics and gas exchange. Because ARDS can have numerous and diverse
causes and a widely variable clinical course, the definition relies largely on specific clinical criteria used to define
the nature of the lung injury. A 1994 consensus definition of ARDS is detailed below (Table 1).
Table 1: Clinical and Laboratory criteria for ARDS:
Bedside findings of tachypnea, dyspnea, crackles
Diminished compliance < (40 cc/cm H2O)
Impaired gas exchange (PaO2/FiO2 < 200)
Diffuse (three or four quadrant) airspace infiltrates on chest radiograph
Exclusion of high pressure pulmonary edema
Mortality and clinical course in ARDS depends more on its cause than on the severity of the lung injury
itself2. As a result, patients with severe lung injury may require a trip to the operating room, albeit at increased risk,
for urgent or emergent surgical procedures unable to be performed in the ICU. Such procedures include incision and
drainage of septic wounds, skin grafting in burn victims, amputation, tracheostomy and exploratory laparotomy.
Because of their distinctive pulmonary abnormalities, patients with ARDS present distinct pre and
intraoperative challenges to the anesthesiologist. In particular, the widened alveolar-arterial pO2 gradient and
dramatically reduced lung compliance exhibited by ARDS patients frequently result in difficulty with oxygenation
and mechanical ventilation, either during transport or in the OR. Ideally, operative procedures in patients with
ARDS are best delayed until lung function improves. If an operating room procedure is necessary, however, three
important features of the intraoperative behavior of these patients are important in reducing the risk of critical,
destabilizing events during anesthesia. First, preoperative ICU management of these patients may differ from that in
other critically ill patients with respect to cardiovascular and respiratory support strategies. Understanding these
differences allows appropriate preoperative evaluation. Secondly, anticipating the effects of intraoperative
anesthetic and fluid management on pulmonary function in ARDS can allow the anesthesiologist to predict the
metabolic and pulmonary consequences of strategic choices made in the OR. Finally, understanding the technical
limitations involved in ventilating ARDS patients on transport or in the OR may help in diagnosing and treating
ventilatory abnormalities in the OR.
B. Preoperative status of the patient:
Although no definitive treatment for ARDS has been developed, significant advances in supportive care for
ARDS have occurred over the past several years. It is now recognized in humans that the homogenous appearance
of ARDS on chest X-Ray is misleading. In reality, the lung with ARDS may be conceptualized as existing in
thirds; one third is consolidated and does not participate in gas exchange, one third is overdistended, and the last
third has relatively normal physiology and gas exchange. Ventilator strategies designed to limit overdistention of
normal lung tissue by restricting tidal volumes and minute ventilation are rapidly becoming widespread, and
approaches to recruiting consolidated lung regions are also being evaluated.
Four strategies with significant anesthetic implications are detailed below:
1. Permissive Hypercapnia. This strategy, perhaps, the most common in modern ICU ventilatory care, seeks to
prevent lung injury from mechanical ventilation by reducing minute ventilation and allowing chronically elevated
pCO2 levels. Typically, tidal volumes are limited by restricting peak inflation pressures to less than 40 cm H2O, and
adjusting the respiratory rate until the pCO2 ranges between 50-60 mm Hg. The resulting hypercapnia and
respiratory acidosis may protect against ventilator induced lung injury, and thus are treated with bicarbonate only if
reductions in pH are severe. Atelectasis and worsened oxygenation resulting from reduced tidal volumes are
usually treated with increases in PEEP.
Although the long-term benefits of such a strategy are well-documented, the value of duplicating this approach
intraoperatively is unknown. While respiratory acidosis are relatively well tolerated in the ICU, the potential for
superimposed metabolic acidoses during anesthesia and surgery increase the difficulty of maintaining hypercapnic
ventilation in the OR
2. ‘Open Lung Ventilation’. This strategy targets both ventilator induced lung injury and consolidated, “inactive”
lung areas. Termed “open lung” because of the physiological goal of preventing alveolar collapse, this approach not
only limits alveolar overdistention by limiting end-inspiratory pressure, but also attempts to prevent alveolar closure
by raising end-expiratory pressure to a level high enough to prevent alveolar deflation. Clinically, ideal end-
expiratory volume is determined at the lower inflection point of the static inspiratory pressure-volume curve of the
respiratory system, and usually occurs at PEEP values of 12-15 cm H2O. Peak inspiratory pressures are then limited
to 35-40 mm H2O, with resulting tidal volumes between 350-400 cc. As expected, hypercapnia is nearly
unavoidable with this technique.
Anesthetic concerns for patients ventilated using an “open lung” approach center on the effectiveness of OR
ventilators in duplicating ICU ventilator settings. Reproducing an open lung ventilatory pattern in the operating
room poses two problems. First, applying >10 cm H20 PEEP can be technically difficult during transport or in an
OR setting. Secondly, with the respiratory rates of 20-24 typically required by this strategy, delivering adequate
tidal volumes can be difficult because of limitations of OR ventilators (see below). Both may result in significant
deterioration in intraoperative gas exchange when compared to ICU ventilators with similar settings. As a result, the
short-term benefit of continuing such a ventilator strategy in the OR should be carefully weighed against the risks of
inadequate intraoperative ventilation.
An important implication of combining increased PEEP with reduced tidal volumes is that PEEP levels
necessary in the ICU with an open lung strategy may not be required to maintain oxygenation in the OR if larger
tidal volumes are used.
3. Prone ventilation. By improving ventilation/perfusion abnormalities, prone positioning consistently improves
oxygenation in 60-90% of patients with ARDS. Technical difficulties in caring for patients in the prone position,
however, as well as reports of worsened oxygenation upon turning prone, largely limit this strategy to a salvage role
in patients with limited PaO2 reserve on maximal ventilatory support. Both the technical difficulty of positioning
prone patients intraoperatively and the transient nature of improvements in PaO2 are relevant to the anesthetic
management of these patients. Patients turned prone may take several hours to reach maximal oxygenation, but can
desaturate in minutes when returned to the supine position. Such patients are extremely unstable, and may not
tolerate transport to and from the OR.
4. Pressure control ventilation: The term ‘pressure control’ refers to a ventilator mode wherein gas flows are
continuously regulated to keep airway pressure constant throughout inspiration. Because a step increase in pressure
occurs at the beginning of the breath, airflow is greatest initially and progressively declines as the lung inflates. Use
of pressure control mode possesses two potential advantages in an ICU setting. First, peak inspiratory pressure
(PIP) is applied throughout the inspiratory phase, improving recruitment of collapsed alveoli and decreasing the
intrapulmonary shunt. Secondly, pressure limited modes protect against barotrauma.
The preoperative use of pressure control ventilation has three important anesthetic implications. First, because of
variability in delivered tidal volumes pressure control mode is usually limited to patients who have failed more
conventional ventilator modes. Transporting such patients to the OR using a manual ventilation system may rapidly
induce atelectasis and arterial oxygen desaturation. Secondly, because of increased flow requirements with pressure
control mode, anesthesia ventilators with or without this mode may not be able to maintain gas exchange despite
similar ventilator settings. Although newer OR ventilators such as the Ohmeda 7900 and Narkomed 6000 offer a
pressure control mode, limitations in maximum deliverable flow rates create a performance “gap” between these
machines and standard ICU ventilators (see below).
Finally, the use of pressure control ventilation prevents detection of progressive endotracheal tube (ETT)
occlusion resulting from inspissated secretions. When ventilated in volume control mode, increases in peak airway
pressure relative to plateau values at relatively low levels of tube occlusion may alert the clinician to a potential
problem. In pressure control mode, however, the clinical signs of endotracheal tube narrowing (primarily decreased
tidal volumes and inspiratory flow rates) are indistinguishable from those of worsening lung compliance. Moreover,
in pressure control mode, clinically detectable decreases in tidal volumes occur at higher degrees of occlusion than
airway pressure changes in volume modes (Fig 1).
Fig 1: Response of the Puritan Bennett 7200 ventilator to ETT occlusion in pressure and volume modes.
(Modes adjusted to TV 750cc, RR 12, compliance 0.4 L/cm H20)
350% Tidal volume changes in
pressure control mode
% original value 300%
250% Airway pressure changes in
volume control mode
0 10 30 50 70 80 90
% ETT occlusion
-Tung et al. Anesthesiology 1999;91:A317
Patients ventilated in pressure control mode may exhibit apparently normal respiratory mechanics despite a degree
of ETT occlusion that would increase PIP to prohibitive levels in volume control mode. As a result, switching
patients with partially occluded endotracheal tubes to an anesthesia ventilator capable of operating only in volume
control mode may raise peak airway pressures to prohibitive levels. To further complicate matters, emergent
occlusion can occur acutely during transport or in the operating room where access to the endotracheal tube can be
difficult. As a result, assessing endotracheal tube patency and devising an appropriate airway plan is important in
patients ventilated in pressure control mode for prolonged periods. If concern about adequacy of ventilation exists,
one approach is to test the ability of the ICU ventilator to function in volume control mode while in the ICU. If
equivalent tidal volumes at similar inspiratory pressures can be delivered in either pressure or volume mode,
significant ETT occlusion is unlikely. Although pressure control mode is available on newer OR ventilators such as
the Ohmeda 7900 or Narkomed 6000, these ventilators do not provide performance equivalent to ICU ventilators
because of fundamental design differences (see below).
Another way to assess for significant ETT occlusion is to examine the flow vs. time waveform in pressure
control mode. Although respiratory mechanics will appear normal with ETT occlusion in pressure control mode,
ventilator-based regulation of flow patterns will produce readily detectable, qualitative changes in the flow vs. time
waveform in the presence of progressive ETT occlusion (see below):
Fig 2: Changes in the flow vs time waveform with ETT occlusion
-Tung et al. Anesth Analg 2000;92S:S144
5. Volume Status: Modern management of ARDS includes aggressively reducing intravascular volume to the
lowest level compatible with adequate organ perfusion to reduce pulmonary edema. As a result, patients with
ARDS may often be volume depleted upon arrival to the operating room, and exhibit exaggerated blood pressure
swings upon loss of central sympathetic tone. Preoperative attention to volume status is thus important prior to
anesthetic induction in these patients.
Table 2. Anesthetic implications of “modern” ARDS care strategies:
TREATMENT ANESTHETIC IMPLICATIONS
Permissive Hypercapnia and • Difficult to duplicate strategy in OR
“Open Lung” Ventilation • PEEP values may not reflect severity of lung injury
• May have compensatory metabolic alkalosis
Prone Ventilation • May not tolerate changes in position
Pressure Control Ventilation • Inadequacy of intraoperative ventilation
• Verify endotracheal tube patency
Reduction of extravascular lung water • Increased pressure lability on induction
C. Intraoperative issues:
1. Anesthetic drug choices
No data directly address the effect of anesthetic drug choices on ARDS. Potential effects of anesthetics on airways
resistance, intrapulmonary shunt, and alveolar fluid clearance may all affect the behavior of an ARDS patient in the
OR. Fortunately, most intravenous and inhaled anesthetics have relatively little effect on airways resistance, with
thiopental and desflurane producing small comparative increases relative to other intravenous and inhaled agents.
Both intravenous and inhaled agents have little effect on shunt during one lung ventilation, and as a result probably
do not significantly increase the A-a O2 gradient in patients with ARDS.
Do inhaled anesthetics worsen lung injury or perpetuate the ARDS state? Studies in rats and rabbits
suggest that administration of inhaled anesthetics may reduce alveolar epithelial fluid clearance and increase
alveolar barrier permeability to protein following oxidant-stress injury. These data raise the possibility that the
optimum anesthetic for patients with certain types of lung injury may not be an inhaled agent. Because of the
relative paucity of patients with ARDS who undergo surgery, however, no clinical studies exist to compare the
effect of inhaled and intravenous anesthetics on the injured lung. As a result, the data currently do not support
favoring intravenous over inhaled agents for anesthetics in patients with ARDS.
2. Nitric oxide (NO)
The use of nitric oxide (NO) as supportive therapy in ARDS has rapidly expanded in the 1990s. Administered
as an inhaled gas, NO selectively vasodilates pulmonary capillaries in ventilated areas of the lung, diverting blood
flow away from non-ventilated areas of the lung. Systemic vasodilatory effects of NO are limited by rapid binding
of NO to circulating hemoglobin. In patients who respond (approximately 60%), pulmonary vascular resistance is
reduced and oxygenation improved, allowing reduced ventilatory support and lower FiO2 levels.
Continuous administration of NO via inhalation is necessary to maintain this effect. Acute interruptions in NO
delivery can result in acute hypoxemia or hypotension, with potentially dangerous consequences. Anesthetic
management of these patients thus requires the technical ability to administer NO without interruption both during
transport and in the operating room. Although NO delivery is relatively straightforward in an ICU setting,
administering NO intraoperatively introduces additional technical issues.
• The respiratory irritant nitrogen dioxide, NO2, is formed from the spontaneous reaction of NO and O2, and is
toxic at 1-2 ppm. Conversion of NO to NO2 is accelerated by increased contact time between NO and oxygen,
increased NO concentration, and a high FiO2. Monitoring NO2 levels is thus necessary in an OR environment.
• Because abrupt termination of NO has been associated with cardiovascular collapse, NO levels should be
carefully monitored during all phases of the perioperative period. Most available monitors for NO have slow
response times (30 sec to full equilibration), which should be taken into account.
• Special equipment is required to administer the appropriate dose of nitric oxide into the circuit during mechanical
ventilation. The INO-Vent, manufactured by Ohmeda-Datex, meters NO into the inspiratory limb based on the
assumption that inspiratory limb flow contains no NO. Use of the INO-Vent with an anesthesia ventilator,
however, may not produce accurate results because most anesthesia machines reintroduce exhaled gases which
may contain NO into the inspiratory limb.
For manual ventilation, NO can be bled into the O2 inlet port of the manual ventilation apparatus. In such a
configuration, the NO concentration may change with changes in oxygen flow, minute ventilation, or NO gas flow
and should be monitored. In addition, such a system should be flushed between uses to avoid buildup of NO2.
3. Mechanical Ventilation in the OR:
Despite being powered from the same gas and electricity sources, OR and ICU-type ventilators differ
substantially in their performance characteristics. This difference is primarily manifested in the inability of the OR-
type ventilator to deliver high gas flows at high airway pressures
Fig 2: Pressure/flow characteristics of ICU (Siemens) and anesthesia ventilators
-From Marks et al. Anesthesiology 1989;71:403-8
Three reasons exist for this discrepancy. First, the gas-driven bellows common to OR ventilators are
intrinsically less powerful than the spring-plate driven bellows or direct manifolds used in ICU ventilators.
Secondly, OR ventilators preserve the exhaled gas mixture for rebreathing, and use a CO2 absorber to recirculate
exhaled gas back into the inspiratory limb. The internal, compressible volume of the resulting circuit, as much as
1.5-2L, then “soaks up” some of the gas delivered by the bellows. In comparison, the internal volume of an ICU
ventilator may be as little as 15-20 cc, allowing more of the gas delivered during inspiration to reach the patient.
Finally, resistance and compliance characteristics of the CO2 absorber act to impede delivery of tidal volumes at
inspiratory flows exceeding 50-60 l/min. Newer ventilators such as the Ohmeda 7900, the Narkomed 6000 and the
Siemens KION systems deliver improved performance, primarily by reducing internal compliance. The Narkomed
6000 even replaces the traditional oxygen-driven bellows of older machines with a mechanical, microprocessor-
At peak inspiratory pressures (PIP) of 50 cm H20 or greater, even newer Ohmeda or Narkomed ventilators are
only able to deliver inspiratory flows of 50 l/min (Fig 2). In either volume or pressure control modes, this limitation
results in reductions in delivered tidal volumes with increased PIP or respiratory rate. Patients who exceed these
flow and pressure levels while in the operating room will exhibit a progressive rise in pCO2, development of
autopeep if respiratory rates are increased to compensate, the onset of a respiratory acidosis, and eventually,
hypoxemia from decreased tidal volumes. Such patients may require an ICU-type ventilator in the OR to maintain
adequate oxygenation and CO2 removal (Table 3).
Table 3: Indications for use of a critical care-type ventilator in the Operating Room
1. Preoperative use of pressure control ventilation
2. High airway pressures (PIP >50 cm H20) and high inspiratory flows (>50 l/min)
3. High preoperative PEEP values or increased A-a gradient (PaO2/FiO2 <100)
4. Coexisting expiratory obstruction (autopeep)
5. Expectation of large fluid shifts.
1. On ARDS:
• Bernard GR, Artigas A, Brigham KL et al. The American-European consensus Conference of ARDS:
Definitions, mechanisms, relevant outcomes, and clinical trial coordination. Am J Respir Crit Care Med
• McIntyre RC, Pulido EJ, Bensard DD, Shames BD, Abraham E. Thirty years of clinical trials in acute respiratory
distress syndrome. Crit Care Med 2000;28:3314-3331.
2. On permissive hypercapnia and the ‘Open Lung’ Strategy:
• Amato MBP, Barbas CSV, Medeiros DM et al. Beneficial effects of the “open lung approach” with low
distending pressures in acute respiratory distress syndrome: A prospective randomized study on mechanical
ventilation. Am J Respir Crit Care Med 1995;152:1835-46
• The Acute Respiratory Distress Syndrome Network. Ventilation with lower tidal volumes as compared with
traditional tidal volumes for acute lung injury and the acute respiratory distress syndrome. N Engl J Med
3. On Prone positioning, Pressure Control mode and Endotracheal tube patency:
• Dupont H, Mentec H, Cheval C et al. Short term effect of inhaled nitric oxide and prone positioning on gas
exchange in patients with severe acute respiratory distress syndrome. Crit Care Med 2000;28:304-8
• Tung A, Morgan S. Tidal volume changes in pressure control mode do not predict progressive endotracheal tube
occlusion. Anesthesiology 1999;91:A317
• Tung A, Morgan S. The flow vs. time waveform distinguishes between decreased compliance and increased
airway resistance in pressure control mode. Anesth Analg 2000;92S:S144.
• Villafane MC, Cinnella G, Lofaso F et al. Gradual reduction of endotracheal tube diameter during mechanical
ventilation via different humidification devices. Anesthesiology 1996;85:1341-9
4. On effects of anesthetic agents in ARDS
• Wang JY, Winship SM. The effects of propofol, isoflurane, and sevoflurane on oxygenation during one-lung
ventilation. Anesth Analg. 1999;89:259.
• Eames WO, Rooke GA, Wu RS, Bishop MJ. Comparison of the effects of etomidate, propofol, and thiopental on
respiratory resistance after tracheal intubation. Anesthesiology. 1996 Jun;84(6):1307-11.
• Nielsen VG, Baird MS, McAdams ML, Freeman BA. Desflurane increases pulmonary alveolar-capillary
membrane permeability after aortic occlusion-reperfusion in rabbits: evidence of oxidant-mediated lung injury.
Anesthesiology. 1998 Jun;88:1524-34.
5. On the limitations of OR ventilators:
• Marks JD, Schapera A, Kraemer RW et al. Pressure and flow limitations of anesthesia ventilators.
• Biery DR, Marks JD, Schapera A et al. Factors affecting perioperative pulmonary function in acute respiratory
failure. Chest 1990;98:1455-62.
• Katz JA, Kallet RH, Alonso JA, Marks JD. Improved flow and pressure capabilities of the Datex-Ohmeda
SmartVent anesthesia ventilator. J Clin Anesth. 2000 Feb;12:40-7.
6. On Nitric Oxide:
• Branson RD, Hess DR. Inhaled nitric oxide: Delivery systems and monitoring. Resp Care 1999;44:281-307.
• Christenson J, Lavoie A, O'Connor M, Bhorade S, Pohlman A, Hall JB. The incidence and pathogenesis of
cardiopulmonary deterioration after abrupt withdrawal of inhaled nitric oxide. Am J Respir Crit Care Med.