Special Feature

Managing Critical Hypoxemia in The ICU

By Andrew M. Luks, MD

In the majority of patients who develop the acute respiratory distress syndrome (ARDS), oxygenation can be supported using increased inspired oxygen concentrations (FIO2) or higher levels of positive end-expiratory pressure (PEEP). Occasionally, however, patients develop "critical hypoxemia" whereby arterial oxygen tensions cannot be maintained at adequate levels with these standard techniques. It is in these situations that clinicians often use alternative modalities to support oxygenation, including prone positioning, inhaled pulmonary vasodilators, extracorporeal membrane oxygenation, alternative modes of mechanical ventilation and recruitment maneuvers. The purpose of this special feature is to consider these techniques in greater detail. What should be apparent following this brief review is that despite the apparently sound physiologic principles underlying these strategies, the data supporting their use is marginal, at best. As a result, there can be no clear guidelines about when to use the techniques, and clinicians must base their decisions on the needs of the individual patient.

What Constitutes "Critical Hypoxemia?"

Before discussing the options mentioned above, it is worth considering exactly when such options should be considered in the first place. Unfortunately, the literature does not adequately define "critical" or "refractory" hypoxemia and there are no formally established thresholds for using these non-standard measures. ARDS is deemed to be present when the PaO2/FIO2 ratio is below 200 mm Hg, but clearly not everyone who meets this criterion requires non-standard therapies. It also seems difficult to apply a consistent threshold across all patients because the ability to tolerate a low PaO2 may vary between patients. A 20 year-old with ARDS following trauma might tolerate PaO2 values in the 50 mm Hg range while this value might be problematic in a 70 year-old with underlying cardiac disease. In general, the decision to initiate non-standard therapies should be tailored to the individual patient and should only be considered when there is impaired oxygenation (eg, PaO2 below 50 mm Hg on maximum conventional support) and concurrent evidence of clinical instability or negative effects of hypoxemia (eg, myocardial ischemia, multi-organ dysfunction).

What Else Can Be Done Besides These Measures?

Before initiating non-standard therapies, other interventions warrant consideration. An adequate PaO2 is an important objective, but the more important goal is ensuring adequate oxygen delivery. Oxygen delivery is a function of cardiac output, hemoglobin concentration and arterial saturation, with only a minor contribution from the PaO2. Table 1 on the following page illustrates the impact on oxygen delivery from manipulating each of these variables in a hypothetical patient with baseline poor cardiac output, anemia and impaired oxygenation. This hypothetical data demonstrates that for a similar 33% increase in each of the variables, there is a greater gain in oxygen delivery from manipulating cardiac output and hemoglobin concentration than there would be from increasing the PaO2. As a result, we should be looking to improve these variables before adopting one of the alternative strategies described below.

Table 1: Changes in Oxygen Delivery with Manipulation of Different Parameters

Manipulated Variable
O2 Delivery

The values in this table are for a hypothetical patient with impaired cardiac output, anemia and hypoxemia at baseline. A 33% increase in either cardiac output or hemoglobin concentration increases oxygen delivery more than a corresponding 33% increase in the PaO2.

In addition to focusing on the supply side of the oxygen equation, efforts should also be directed towards minimizing oxygen demand. Treatment of fever and elimination of shivering are warranted, as are efforts to minimize patient triggering on the ventilator. Breath-stacking, for example, can be minimized with aggressive use of sedation. In rare instances, neuromuscular blocking agents can be considered although there is no data demonstrating a mortality benefit from this practice and these agents increase the risk of critical care polyneuropathy. In the event that paralytics are used, the duration of use should be minimized, and train-of-four monitoring should be employed to ensure an appropriate degree of paralysis.

Prone Ventilation

Prone ventilation is thought to benefit patients by causing favorable changes in regional ventilation and perfusion as well as potentially aiding in secretion clearance and redistribution of extravascular lung water. Studies demonstrate that the technique improves oxygenation but there is still no proof that it improves mortality. Earlier randomized controlled trials,1, 2 which failed to show a mortality benefit, were criticized due to the short duration of proning and high tidal volumes employed in the studies. A more recent trial,3 in which patients were proned for an average of 17 hours per day, did show a trend toward decreased mortality but this result was not statistically significant and the mortality rates in the proned group (43%) were still higher than those reported for low tidal-volume ventilation in the ARDSnet trial (30%). Logistical issues must also be considered in the decision to prone a patient. The technique is time and labor intensive, nursing staff lose easy access to lines and tubes, and patients are at increased risk for pressure sores as well as aspiration, among other potential problems. Specialized beds, such as the RotoProne system, are available to facilitate proning and minimize complications but come at a potentially high price. In our region, hospitals pay, on average, $500-700 per day for use of the bed, a high cost in light of the lack of an established mortality benefit.

Inhaled Nitric Oxide

By generating localized pulmonary vasodilation in areas that receive adequate ventilation, inhaled nitric oxide (NO) is felt to improve ventilation-perfusion matching and, as a result, arterial oxygenation. As with prone positioning, the therapy has been demonstrated in randomized trials4,5 and a meta-analysis6 to improve oxygenation but there is still no evidence of improved outcomes such as decreased mortality or shortened duration of mechanical ventilation. There are often overlooked risks with the therapy such as the potential for methemoglobinemia, a complication that would, ironically, further impair oxygen delivery, and there remain no reliable markers for predicting which patients will experience improved oxygenation with this treatment. The biggest downside of the therapy, however, is its high cost; our institution currently pays $131/hr for the first 96 hours of use up to a maximum of $12,576; each additional month costs an additional $12,576. Although it has been reported7 that use of inhaled NO is not associated with increased cost relative to standard treatment, one must recognize that because the therapy is not FDA-approved for this indication, these costs are not reimbursed by insurance or Medicare and the hospitals must absorb the expense.

Inhaled Prostacyclin

Inhaled prostacyclin works by the same mechanism as inhaled NO but has not been studied as extensively. The available data demonstrate improvements in oxygenation and pulmonary hemodynamics comparable to those of nitric oxide,8,9 but important differences exist between the two therapies. Inhaled prostacyclin is roughly one-tenth the cost of inhaled nitric oxide, but this cost benefit is offset by the fact that administration is substantially more difficult. There is no reliable delivery mechanism as exists with nitric oxide; the medication is inactivated by room temperature and light and, most importantly, the glyceine diluent used to prepare the inhaled form has been reported to cause tracheitis and has the potential to clog valves in the ventilator circuit. Long-term safety data is also lacking, as is evidence of a mortality benefit from the therapy.

Non-Conventional Modes of Mechanical Ventilation

Multiple modes of mechanical ventilation including pressure control, inverse ratio, airway pressure release and high frequency oscillatory or jet ventilation, have been proposed as alternative strategies when oxygenation cannot be supported using conventional volume control approaches. Space considerations do not permit analysis of the data from trials on all of these modes of ventilation but the general message that comes from these studies is that they fail to demonstrate consistent improvements in oxygenation or benefit in terms of mortality or other important patient outcomes. When one also considers the lack of familiarity clinicians and respiratory therapists might have with these modes, the lack of good data makes it difficult to recommend any of these strategies as viable alternatives.

Recruitment Maneuvers

Increased levels of continuous positive airway pressure (CPAP) for short durations have been used in an effort to open collapsed alveoli and improve oxygenation. Available studies10 on these maneuvers demonstrate improved oxygenation but the benefits appear to be of short duration (less than 60 minutes). It is unclear whether adding additional PEEP after the recruitment maneuver leads to sustained benefit, as the one study to address this question11 did not examine oxygenation more than 60 minutes after the maneuver was completed. A further complicating issue with routine use of recruitment maneuvers is the lack of information regarding best practices for these maneuvers including their frequency, magnitude and duration.

Extracorporeal Membrane Oxygenation (ECMO)

Of all the therapies proposed for the management of critical hypoxemia, this is the one with perhaps the weakest data supporting its use. Only two randomized trials12,13 have examined this technique and both failed to show a mortality benefit, although it should be noted that these studies were conducted before low tidal-volume ventilation became standard therapy for ARDS. More recent studies have been published reporting benefit from the therapy but these have all been either observational14 or uncontrolled trials,15 usually from a single institution, and, as a result, provide little useful input into whether this therapy is really of benefit. In some of these reports, the claims of benefit simply do not fit the data. Lewandowski et al,15 for example, documented a 55% survival in their ECMO treated group in an uncontrolled prospective trial and concluded in their abstract "ARDS can be successfully treated with the clinical algorithm and high survival rates can be achieved." They make this claim even though the non-ECMO group had a much higher survival rate of 89%. There is also no data at present as to whether venoarterial or venovenous methods are more appropriate for patient management. Data collection from a randomized, controlled trial (CESAR trial) examining the use of ECMO in ARDS has been completed in the United Kingdom but the results have yet to be published. Until that data is available, ECMO should not be used for critical hypoxemia in the adult population.


Multiple strategies have been proposed for supporting oxygenation in patients with critical hypoxemia. Unfortunately, the data supporting the use of such strategies is limited and none of the modalities has been associated with a mortality benefit. As a result, any decision to employ these techniques will have to take into consideration the needs of the individual patient with a careful weighing of the risks and benefits of the proposed strategy.


  1. Guerin C, et al. Effects of systematic prone positioning in hypoxemic acute respiratory failure: A randomized controlled trial. JAMA. 2004;292:2379-2387.
  2. Gattinoni L, et al. Prone-Supine Study Group. Effect of prone positioning on the survival of patients with acute respiratory failure. N Engl J Med. 2001;345:568-573.
  3. Mancebo J, et al. A multicenter trial of prolonged prone ventilation in severe acute respiratory distress syndrome. Am J Respir Crit Care Med. 2006;173:1233-1239.
  4. Taylor RW, et al. Inhaled Nitric Oxide in ARDS Study Group. Low-dose inhaled nitric oxide in patients with acute lung injury: A randomized controlled trial. JAMA. 2004;291:1603-1609.
  5. Dellinger RP, et al. Effects of inhaled nitric oxide in patients with acute respiratory distress syndrome: Results of a randomized phase II trial. Inhaled nitric oxide in ARDS study group. Crit Care Med. 1998;26:15-23.
  6. Adhikari NK, et al. Effect of nitric oxide on oxygenation and mortality in acute lung injury: Systematic review and meta-analysis. BMJ. 2007;334:779.
  7. Angus DC, et al. NO-06 Investigators. Healthcare costs and long-term outcomes after acute respiratory distress syndrome: A phase III trial of inhaled nitric oxide. Crit Care Med. 2006;34:2883-2890.
  8. Zwissler B, et al. Inhaled prostacyclin (PGI2) versus inhaled nitric oxide in adult respiratory distress syndrome. Am J Respir Crit Care Med. 1996;154:1671-1677.
  9. Walmrath D, et al. Direct comparison of inhaled nitric oxide and aerosolized prostacyclin in acute respiratory distress syndrome. Am J Respir Crit Care Med. 1996; 153:991-996.
  10. Brower RG, et al. Effects of recruitment maneuvers in patients with acute lung injury and acute respiratory distress syndrome ventilated with high positive end-expiratory pressure. Crit Care Med. 2003;31:2592-2597.
  11. Lim CM, et al. Effect of alveolar recruitment maneuver in early acute respiratory distress syndrome according to antiderecruitment strategy, etiological category of diffuse lung injury, and body position of the patient. Crit Care Med. 2003;31:411-418.
  12. Morris AH, et al. Randomized clinical trial of pressure-controlled inverse ratio ventilation and extracorporeal CO2 removal for adult respiratory distress syndrome. Am J Respir Crit Care Med. 1994;149:295-305.
  13. Zapol WM, et al. Extracorporeal membrane oxygenation in severe acute respiratory failure. A randomized prospective study. JAMA. 1979;242:2193-2196.
  14. Hemmila MR, et al. Extracorporeal life support for severe acute respiratory distress syndrome in adults. Ann Surg. 2004;240:595-605; discussion 605-607.
  15. Lewandowski K, et al. High survival rate in 122 ARDS patients managed according to a clinical algorithm including extracorporeal membrane oxygenation. Intensive Care Med. 1997;23:819-835.