Acute Respiratory Distress Syndrome: Adjuncts to Lung-Protective Ventilation

John P. Kress, M.D. [1] and John J. Marini, M.D. [2] , [1] Department of Medicine, Section of Pulmonary and Critical Care, University of Chicago, Chicago, Illinois; [2] Department of Pulmonary and Critical Care Medicine, University of Minnesota, St. Paul, Minnesota.

Semin Respir Crit Care Med. 2001;22(3) 

In This Article

Permissive Hypercapnia

The traditional management of respiratory failure has focused on mechanical ventilation strategies seeking to achieve normal arterial blood gases. Relatively large tidal volumes (10-15 mL/kg) have been utilized to avert atelectasis and to establish acid-base homeostasis with PaCO2O 40mm Hg. Such traditional ventilator strategies often elevate airway pressures to high levels in critically ill patients, particularly those with acute respiratory distress syndrome (ARDS) or obstructive airways disease. Numerous laboratory studies have demonstrated that positive pressure ventilation with high airway pressures and/or large tidal volumes may cause and/or propagate lung injury, because of alveolar over-distension and the repeated reopening of collapsed alveolar units by high end-inspiratory tidal pressures. [1,2,3,4] Recognition of this hazard led to the notion that smaller tidal volumes may be useful in preventing baro-trauma and lung injury. [5] Thus, the idea of ventilating patients with a strategy that accepts hypercapnia ("per-missive hypercapnia"), stems from the hypothesis that the adverse effects of alveolar overdistension and end-expiratory collapse are more detrimental to patient out-come than the adverse effects of intentional respiratory acidosis.

Cellular Effects. At the cellular level, CO2O is able to freely diffuse across the cell membrane, leading to a rapid increase in intra-cellular hydrogen ion concentration when PaCO2O is abruptly increased. [6] A variety of counter-regulatory mechanisms are engaged in response to this decrease in intracellular pH (pHi). Such mechanisms include: (a) physicochemical buffering by intracellular proteins and phosphates, [7] (b) modifications of cellular metabolism to reduce intracellular production of hydrogen ions, and (c) modifications of H + and HCO3O - flux across the cell membrane. [8] Physicochemical buffering occurs instantaneously, whereas the other mechanisms are delayed 1 to 3 hours after the onset of hypercapnia.[9,10] Compensatory mechanisms are very efficient, with numerous animal studies demonstrating only modest increases in pHi, even in the presence of profound respiratory acid-emia.[11,12] Intracellular acidosis may lead to a decrease in glycolysis and an increase in oxidative deamination, which results in amino acid depletion.

Nevertheless, intracellular adenosine triphosphate concentrations are less affected. [13] Other effects of intra-cellular acidosis include inhibition of contractility via actin-myosin interaction, [14] interference with neuronal electrical activity, [15] and inhibition of cellular division. [16] The clinical importance of these responses to intracellular acidosis remains to be determined fully. However, in the absence of hypoxemia, intracellular acidosis appears to be well tolerated, even in critically ill patients.

Cardiovascular Effects. The autonomic nervous system responds to acute hypercapnia with a significant increase in sympathetic activity. [17] Elevated levels of circulating epinephrine and norepinephrine derive from efferent sympathetic nerve fibers and the adrenal gland. [18] It is this indirect sympathetic outpouring that largely mediates the effects of hypercapnia on the circulatory system. Numerous investigators have demonstrated the direct myocardial depressant effects of hypercapnia and intracellular acidosis, [19,20,21] which are largely due to the interference with myofilament response to calcium. [22] Hypercapnia results in coronary vasodilation in the normal heart, though this effect is blunted in ischemic left ventricular failure. [23] The oxygen cost of myocardial contraction is increased in the setting of hypercapnia in isolated dog hearts, [24] though human studies in this area are lacking. It is not clear whether hypercapnia-induced coronary vasodilation may result in preferential perfusion through nondiseased coronary arteries, resulting in a steal phenomenon. Such a response has been noted in coronary artery disease after exposure to certain anesthetic agents, though the clinical impact in that setting seems minimal. [25]

The indirect sympathetic nervous system hyperactivity far outweighs the direct myocardial depressant effects of acute hypercapnia, typically resulting in significant increases of heart rate, stroke volume, and blood pressure, as well as a mild decrease in systemic vascular resistance. Respiratory acidosis may result in pulmonary vasoconstriction, with a resulting increase in pulmonary vascular resistance; this may lead to circulatory instability in patients with right heart dysfunction. [26]

Central Nervous System Effects. The most striking effect of hypercapnia on the central nervous system is altered cerebral blood flow. Cerebral arterioles dilate in response to acute hypercapnia, a response largely mediated by extracellular pH.[27,28] This response is transient, however, and cerebral blood flow returns to baseline after approximately 48 hours of continued hypercapnia. A cerebral steal phenomenon, similar to the coronary steal already mentioned, may occur in some settings, [29] though its clinical importance is not clear. The increase in cerebral blood flow seen with acute hypercapnia leads to an increase in cerebral blood volume and a resulting increase in intracranial pressure.

Hypercapnia has a variable effect on consciousness. At extreme levels, carbon dioxide alters conscious-("CO2O narcosis") [30,31,32] ; however, there are reports of patients remaining fully conscious despite CO2O levels that exceed 100 mm Hg.[33,34] Indeed, most patients subjected to acute hypercapnia are not lethargic. Rather, such patients more typically present with severe agitation, and, at extreme levels of hypercapnia, seizure activity may be seen.[33,35] Whether lower levels of acute hypercapnia may lead to seizures in those with preexisting seizure disorders or other brain injuries is not known.

Clinical Studies of Lung Protective Ventilatory Strategies. As concern over the adverse effects of traditional mechanical ventilation strategies has grown, so has experience using a variety of low airway pressure, low tidal volume (so-called lung protective) ventilation strategies. Hypercapnia is a frequent result of these strategies. Early work by Hickling and colleagues noted a striking reduction of mortality in 50 patients with ARDS (lung injury score > 2.5 [36] ) managed with a ventilation strategy limiting peak inspiratory pressures to less than 40 cm H2O. Tidal volumes as low as 350 mL (5 mL/kg) were used, and PaCO2O levels were allowed to rise to a mean of 62 mm Hg (range 30-129 mm Hg). Only eight of 50 (16%) patients died in the hospital using this approach -- a number significantly lower than the Acute Physiologic and Chronic Health Evaluation (APACHE) II-predicted mortality rate of 39.6%. [37] Mean arterial pH at the time of maximum PaCO2O was 7.29 (range 7.02-7.38). Duration of mechanical ventilation was 8.1 days for survivors and 9.0 days for nonsurvivors. This uncontrolled retrospective study was followed by an uncontrolled prospective study of patients with ARDS from the same institution, [38] again showing a reduction in mortality compared with that predicted by APACHE II score. In this study, the degree of acidemia was more pronounced, with a mean arterial pH of 7.23 at the time of maximum PaCO2O (range 6.79-7.45).

Recently, several prospective studies of ventilatory strategies for ARDS have shed light on the impact of permissive hypercapnia. Amato and colleagues studied 53 patients with severe ARDS, [39] randomized to one of two different ventilator strategies. "Conventional" mechanical ventilation used high tidal volumes (12 mL/kg) and the least positive end-expiratory pressure (PEEP) needed to maintain arterial oxygen partial pressure greater than 80 mm Hg at an acceptable FiO2. The conventional strategy sought to maintain an arterial carbon dioxide level between 35 and 38 mm Hg, regardless of airway pressures. In contrast, a "lung protective" strategy was targeted to prevent both alveolar overdistension and collapse, regardless of arterial carbon dioxide levels. A smaller tidal volume (less than 6 mL/kg), higher PEEP (greater than the lower inflection point of the inflation pressure volume curve), and pressure limited ventilation strategy (driving pressure [plateau airway pressure minus PEEP] less than 20 cm H2O, peak airway pressure less than 40 cm H2O) were used. PaCO2O levels were allowed to rise to 80 mm Hg, and sodium bicarbonate infusions were permitted if arterial pH fell below 7.20. Twenty-eight-day mortality was significantly lower in the lung protective group (38% vs 71%). In addition, weaning rates and incidence of barotrauma were both reduced in the lung protective group. The PaCO2O levels were significantly higher in the lung protective group, with an average level of 58.2 mm Hg 1 hour into the protocol, decreasing to an average level of 50.8 mm Hg by days 2 through 7 of the study. The control group averaged PaCO2O levels between 33 and 36 mm Hg throughout the study period. This was the first randomized, controlled, prospective study to demonstrate that a ventilator strategy that utilized permissive hypercapnia was associated with improved mortality.

Experimental studies underscore the importance of using sufficient end-expiratory pressure when attempting to avoid ventilator-induced lung injury. Using sufficient PEEP to keep unstable lung units open can dramatically reduce the lung damage resulting from high tidal inflation pressures. Whether this factor accounts for the success of the lung protective strategy in Amato's group remains unclear. A National Institues of Health (NIH)-sponsored trial that is currently ongoing aims to answer this question.

Brochard and colleagues evaluated a lung protective strategy in ARDS that targeted very low plateau airway pressures ( =/< 25 cm H2O) and low tidal volumes (<10 mL/kg). They compared this approach with a conventional approach using higher tidal volumes (10-15 mL/kg) with a respiratory rate targeted to maintain eucapnia (PaCO2O 38-42 mm Hg), regardless of airway pressures (limit peak airway pressure of 60 cm H2O). [40] Mean PaCO2O levels were significantly higher in the study group (59.5 vs 41.3 mm Hg) and mean arterial pH levels were lower (7.28 vs 7.40). The difference in plateau airway pressures was statistically significant (25.7 vs 31.7 cm H2O), though it is less clear whether this relatively small difference was clinically important. There were no differences in mortality, duration of mechanical ventilation, incidence of barotrauma, or occur-rence of multiple organ failure between the two groups. The authors suggested that the study defined a "safe zone" for mechanical ventilation, avoiding both the risks of ventilator-induced lung injury and the potential problems seen with extreme hypercapnia. The fact that the control group had relatively low plateau airway pressures despite larger tidal volumes suggests that lung compliance may have been less severely deranged in these control patients than in other studies.

Brower and colleagues compared a small tidal volume approach (5-8 mL/kg, plateau airway pressure <30 cm H2O) to a traditional tidal volume approach (10-12 mL/kg, plateau airway pressure < 55 cm H2O) for patients with ARDS. [41] Similar to the study of Brochard and colleagues, the mean plateau airway pressures in the control group were relatively low (only 30.6 cm H2O, compared with 24.9 cm H2O in the small tidal volume group). The highest average PaCO2O levels were 50.3 mm Hg in the small tidal volume group compared with 40.1 mm Hg in the control group, a difference that was statistically significant. The arterial pH levels were lower in the small tidal volume group (7.34 vs 7.38), but this difference did not reach statistical significance. The small tidal volume, permissive hypercapnia approach did not impact requirements for oxygenation support (PEEP or FiO2), intravenous fluid administration, or use of sedatives, vasoactive drugs, or neuromuscular blocking agents. Again, one may infer from the relatively low plateau airway pressures in the control group and the modest degree of hypercapnia in the small tidal volume group that the patients in this study did not suffer from severe derangements in lung compliance. Whether the results of the studies by Brochard and Brower are applicable to patients with ARDS suffering from more severe abnormalities of lung compliance is not clear. [42]

Stewart and colleagues evaluated patients at risk for the acute respiratory distress syndrome with a PaO2/ FiO2 ratio less than 250 on five cm H2O PEEP. [43] A control group was managed with conventional ventilation using tidal volumes of 10-15 mL/kg and peak inspiratory pressures up to 50 cm H2O. The "lung protective" group was ventilated with a peak inspiratory pressure less than or equal to 30 cm H2O and a tidal volume less than or equal to 8 mL/kg. Tidal volumes and plateau airway pressures were significantly lower in the limited pressure group (tidal volumes 7.2 vs 10.8 mL/kg; plateau airway pressures 28.6 vs 20.0 cm H2O). However, plateau airway pressures were quite low in the control group, again suggesting the patients did not suffer from severe derangements in lung compliance. The authors found no difference in mortality between the two groups, but permissive hypercapnia was more common and more prolonged in the limited ventilation group, with an average PaCO2O of 54.4 mm Hg, compared with 45.7 mm Hg in the control group. There were no differences in secondary outcomes such as barotrauma, highest multiple-organ-dysfunction score, or number of episodes of organ failure. However, a greater number of patients required dialysis for renal failure in the pressure-limitation group. Additionally, in contrast to the study by Brower and colleagues, [41] a greater number of patients in the study group required neuromuscular blocking agents.

Recently, the NIH completed a large, multicenter study (sponsored by ARDS Network) testing the hypothesis that limiting airway pressures and tidal volumes in ARDS decreases lung injury and improves out-come. [44] The study was designed to assess the efficacy of 12 mL/kg versus 6 mL/kg tidal volume ventilation strategies with regard to reducing mortality and morbidity in a large number of patients with acute lung injury and ARDS. All patients received volume-cycled ventilation. Ventilator rate was adjusted up to a maximum of 35 breaths per minute in both groups to achieve an arterial pH of 7.30-7.45, if possible. However, endinspiratory (plateau) airway pressure limits were set at less than 50 cm H2O for the 12 mL/kg group and less than 30 cm H2O for the 6 mL/kg group. An arterial pH as low as 7.15 was tolerated in order to maintain plateau airway pressure goals. This study was stopped early after an interim analysis showed a 25% reduction in mortality in the 6 mL/kg tidal volume group, compared with the 12 mL/kg group (Overall mortality was 31% in the 6 mL/kg group and 39% in the 12 mL/kg group). In spite of using smaller tidal volumes, the study group did not experience respiratory acidosis.

Based on data from recent clinical studies, the following recommendations can be made regarding permissive hypercapnia ventilation strategies in ARDS: (a) Plateau airway pressures greater than 30 cm H2O during tidal ventilation are probably detrimental and should be avoided, even if such a pressure-limiting approach leads to hypercapnia. [45] In circumstances where chest wall compliance is diminished (tense ascites, morbid obesity, chest wall trauma/burns), higher plateau airway pressures may be accepted without increasing the risk of ventilator-induced lung injury. (b) If plateau airway pressures remain under 30 cm H2O despite relatively large tidal volumes (10-12 mL/kg) permissive hypercapnia is not likely to be needed. Such patients are likely suffering from less severe lung injury, at least as gauged by the compliance of the respiratory system. (c) As a means of limiting ventilator-induced lung injury, hypercapnia to PaCO2O levels up to 60 mm Hg and arterial pH levels to 7.20 appear to be well tolerated in patients with ARDS, provided aggressive sedative regimens are used. More extreme levels of hypercapnia and acidemia have not been evaluated in a randomized, controlled manner. Though several noncontrolled studies have reported benefits utilizing more extreme levels of hypercapnia,37,38 such approaches should be employed cautiously. (d) The use of sodium bicarbonate to offset purely respiratory acidosis is not necessary when hypercapnia is modest. Indeed, experimental data suggest that hypercapnic acidosis may protect against certain forms of oxidative cellular injury.[46,47] The use of sodium bicarbonate in more extreme cases of hypercapnia should be considered on an individual basis. (e) The relation of airway pressures to barotrauma is not clear. Amato and colleagues found a significantly higher incidence of barotrauma (42% vs 7%) in those patients subjected simultaneously to higher end-inspiratory and lower end-expiratory airway pressures. [39] Confounding the association between airway pressures and baro-trauma in this study was the significant difference in PEEP levels between study and control groups. Because all other studies already discussed used similar PEEP levels in control and study groups, it is not possible to determine whether the higher incidence of barotrauma in the study by Amato and colleagues reflected the impact of plateau airway pressures, lower PEEP levels, their combination, or some other factor(s). That PEEP is potentially important to consider is reflected by the fact that no other study of ventilatory strategies in ARDS has noted a difference in the incidence of baro-trauma. Weg and colleagues evaluated 725 patients from a prospective, multicenter study of inhaled surfactant in ARDS. [48] They found no association between airway pressures or lung volumes and barotrauma. Their barotrauma incidence (6.9% pneumothorax, 10.6% pneumothorax or other air leaks) was similar to that in all of the other studies already discussed (except for Amato's group), where the barotrauma incidence ranged from 4 to 14%.[40,41,43]


Comments on Medscape are moderated and should be professional in tone and on topic. You must declare any conflicts of interest related to your comments and responses. Please see our Commenting Guide for further information. We reserve the right to remove posts at our sole discretion.
Post as: