Correspondence
Protective Ventilation for the Acute Respiratory Distress Syndrome
N Engl J Med 1998; 339:196-199July 16, 1998
- Article
To the Editor:
Experimental studies have provided unequivocal evidence that suboptimal patterns of mechanical ventilation can injure the lung.1 The studies of patients with the acute respiratory distress syndrome reported by Weg et al. and Stewart et al. (Feb. 5 issue)2,3 challenge this position. Another study reported in the same issue, by Amato et al.,4 suggests that a protective-ventilation strategy designed to minimize massive alveolar collapse and cyclic lung reopening and overdistention during mechanical ventilation improves the outcome.
We question the conclusions of Weg and coworkers. They conducted a retrospective analysis of data collected in a study designed to evaluate the effect of surfactant therapy in patients with the acute respiratory distress syndrome induced by sepsis. The study design was inadequate to evaluate the relation of pneumothorax and other air leaks to mortality in patients with the acute respiratory distress syndrome or to draw conclusions about the safety of high inflation pressures and volumes. Clinical manifestations of barotrauma occur after the first week of ventilation, when inflammation peaks; considering only events occurring during the first five days is inappropriate. In addition, the study did not use sensitive means to detect pneumothorax and other (undefined) air leaks. Moreover, the authors present data only on peak inspiratory pressure, not on end-inspiratory plateau pressure (PPLAT). Peak inspiratory pressure does not reflect peak alveolar pressure, especially during the administration of surfactant, since increased resistive pressures result from the physical properties of the drug. Still, the mean (±SD) peak inspiratory pressure in the group without air leaks (45.8±11.7 cm of water) was moderate, probably reflecting an average PPLAT in the upper 30s. Thus, the low rate of pneumothorax would be anticipated. The actual ventilatory pressures reflect an important element of a protective-ventilation strategy.
Although different peak alveolar pressures were targeted for the control group and the limited-ventilation group in the study by Stewart et al., the two groups had similar PPLAT values (26.8±6.7 vs. 22.3±5.4 cm of water, respectively), which were well below the upper limit recommended by a consensus conference (35 cm of water).5 The positive end-expiratory pressure (PEEP) was also similar in the two groups. Since low peak alveolar pressures were used in all patients, it is understandable that there were no differences in outcome.
In contrast, Amato et al. found that the protective strategy was associated with improved survival at 28 days, a higher rate of weaning from mechanical ventilation, and a lower rate of barotrauma. These results are consistent with experimental evidence and with clinical observations by many physicians. Although disappointing, the absence of a significant difference in the rate of survival to hospital discharge in the two treatment groups highlights the influence of the underlying disease on the outcome. This study supports the use of physiologically targeted, patient-specific ventilatory strategies adjusted on the basis of lung mechanics.
5 ReferencesJean-Daniel Chiche, M.D.
Massachusetts General Hospital, Boston, MA 02114Fabrice Brunet, M.D.
Cochin–Port Royal University Hospital, Paris 75014, FranceMaurice Lamy, M.D.
Liege University Hospital, Liege 4000, BelgiumRobert Kacmarek, Ph.D.
Massachusetts General Hospital, Boston, MA 021141
Dreyfuss, D,
Saumon G . Ventilator-induced lung injury: lessons from experimental studies. Am J Respir Crit Care Med 1998;157:294-323
Web of Science | Medline2
Weg JG ,Anzueto A ,Balk RA , et al. The relation of pneumothorax and other air leaks to mortality in the acute respiratory distress syndrome. N Engl J Med 1998;338:341-346
Full Text | Web of Science | Medline3
Stewart TE ,Meade MO ,Cook DJ , et al. Evaluation of a ventilation strategy to prevent barotrauma in patients at high risk for acute respiratory distress syndrome. N Engl J Med 1998;338:355-361
Full Text | Web of Science | Medline4
Amato MBP ,Barbas CSV ,Medeiros DM , et al. Effect of a protective-ventilation strategy on mortality in the acute respiratory distress syndrome. N Engl J Med 1998;338:347-354
Full Text | Web of Science | Medline5
Slutsky AS . Consensus conference on mechanical ventilation -- January 28-30, 1993, at Northbrook, Illinois. 1. European Society of Intensive Care Medicine, the ACCP and SCCM. Intensive Care Med 1994;20:64-79[Erratum, Intensive Care Med 1994;20:378.]
CrossRef | Web of Science | Medline
To the Editor:
The study by Amato et al.1 may be methodologically flawed and misleading. In comparing their article in the Journal with their 1995 article, in which the first 28 patients were described,2 we noticed two discrepancies, both of which raise questions about the validity of the study. First, the randomization procedure for the first 28 patients was carried out after stratifying patients with the acute respiratory distress syndrome according to the presence or absence of leptospirosis.1 However, the randomization procedure reported in the Journal article does not include stratification based on the underlying disease. Second, the primary end point for the study of the first 28 patients was an evaluation of the effects of the protective-ventilation strategy on lung recovery, which included the weaning rate, the evolution of pulmonary function, and deaths related to respiratory failure.2 In contrast, the primary end point for the study of all 53 patients was survival at 28 days.1 Although the 28-day mortality was not a stated primary (or secondary) end point in the study of the first 28 patients, it must have been evaluated, because the estimate of the sample size for the entire study was based on an improved survival rate in the protective-ventilation group as determined from the original study. The primary outcome of the study was apparently changed after the interim analysis of the first 28 patients.
Other issues raise further questions. The study was conducted over a period of four and a half years, during which the approximately 270 patients with the acute respiratory distress syndrome were identified, yet only 53 patients (20 percent) were enrolled. Although the exclusion and inclusion criteria were described, the specific reasons for the exclusion of so many patients are unclear, thus raising the question of selection bias.
The mortality rate in the conventional-ventilation group was 71 percent, which is much higher than the rates in other trials involving patients with the acute respiratory distress syndrome. This raises questions about the selection of patients, the care of the patients in the conventional-ventilation group, and the generalizability of the findings to other patients with the acute respiratory distress syndrome. Moreover, as the authors noted, survival to hospital discharge did not differ significantly between the two treatment groups.
Finally, in some patients, measurement of the end-expiratory pressure above the lower inflection point on the static pressure–volume curve (PFLEX) could not be carried out, so for these patients, an empirical level of 16 cm of water was used for PEEP. Do the authors believe that their lung-recruitment strategy could be simulated for most patients with this approach, thus avoiding the use of neuromuscular blockade?
2 ReferencesPolly E. Parsons, M.D.
Denver Health Medical Center, Denver, CO 80204Michael Matthay, M.D.
University of California, San Francisco, San Francisco, CA 94143-06241
Amato MBP ,Barbas CSV ,Medeiros DM , et al. Effect of a protective-ventilation strategy on mortality in the acute respiratory distress syndrome. N Engl J Med 1998;338:347-354
Full Text | Web of Science | Medline2
Amato MB ,Barbas CS ,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;152:1835-1846
Web of Science | Medline
To the Editor:
In their discussion, Weg et al. compare the transpulmonary pressure of 35 to 40 cm of water, suggested by the American College of Chest Physicians' Consensus Conference as an upper limit of pressure during mechanical ventilation,1 with the much higher maximal inspiratory pressures generated in normal men and women. Such a comparison is flawed. During the measurement of maximal inspiratory pressure, lung volume remains nearly constant, so that mouth, alveolar, and pleural pressures all change by similar amounts. This means that although the change in pressure at the mouth is quite large, the transpulmonary pressure (defined as alveolar pressure minus pleural pressure) is quite small. There are many other examples from daily life (e.g., coughing) in which very large changes in pressure occur at the mouth or airway, but the accompanying changes in transpulmonary pressure are small. In contrast, during mechanical ventilation, changes in mouth or airway pressure are usually accompanied by proportionate changes in both lung volume and transpulmonary pressure.
The distinction between mouth pressure (as reflected by the maximal inspiratory pressure) and transpulmonary pressure is crucial: it is transpulmonary pressure rather than mouth pressure that determines, in part, alveolar-wall stress. Excessive alveolar-wall stress is probably the mechanism underlying ventilator-induced lung injury. What remains to be determined is whether current ventilation strategies exceed the transpulmonary-pressure threshold for lung injury, and whether strategies based on limiting transpulmonary pressure affect the outcome in patients with the acute respiratory distress syndrome.
1 ReferencesHarold L. Manning, M.D.
Dartmouth–Hitchcock Medical Center, Lebanon, NH 037561
Slutsky AS . Mechanical ventilation: American College of Chest Physicians' Consensus Conference. Chest 1993;104:1833-1859[Erratum, Chest 1994;106:656.]
CrossRef | Web of Science | Medline
To the Editor:
The method of monitoring airway pressures every 8 hours, as Weg et al. did in their study, and then using one or two of these isolated measurements in the 16 hours before the development of barotrauma is not sufficient, since changes can occur rapidly and often momentarily before barotrauma develops. It would therefore be necessary in such a study to monitor pressures continuously.
One potential cause of barotrauma (i.e., the intrinsic PEEP [auto-PEEP]) was not mentioned or measured. Externally measured PEEP does not identify the intrinsic PEEP level.1
Finally, coughing can create sharp increases in airway pressure that are not always recorded. Detailed information about factors that might increase the tendency to cough (such as respiratory infection) or decrease it (such as decreased consciousness or use of sedatives or muscle relaxants) is essential.
1 ReferencesMichael Y. Shapira, M.D.
Sigal Sviri, M.D.
David M. Linton, M.D.
Hadassah University Hospital, Jerusalem 91120, Israel1
Valta P ,Corbeil C ,Chasse M ,Braidy J ,Milic-Emili J . Mean airway pressure as an index of mean alveolar pressure. Am J Respir Crit Care Med 1996;153:1825-1830
Web of Science | Medline
Author/Editor Response
The authors reply:
To the Editor: Experimental studies have shown that so-called ventilator lung injury in normal and injured lungs is due primarily to excessive overinflation (a tidal volume that is 300 to 700 percent of the normal value, or 1500 to 3500 ml in a 70-kg man). However, even in the elegant review by Dreyfuss and Saumon, discussions of pressures in isolation can be confusing.1 Such excessive experimental volumes far exceed those used in patients with the acute respiratory distress syndrome, even if one allows for decreased normal lung.
Although these issues were of concern to Chiche et al., we acknowledged the limitations of our study, defined air leaks, and addressed the absence of plateau pressures. Pneumothorax and other air leaks were identified by standard criteria on chest films; we are unaware of a more sensitive practical means for use in a large study. Since there were no differences between the patients who received surfactant and those who received placebo, this was not a concern. As Table 1Table 1
Peak Pressures and Pneumothorax in the 716 Patients for Whom Data Were Available. shows, 34.5 percent of patients had peak pressures of 50 cm of water or greater — clearly not evidence of a protective-ventilation strategy. The incidence of pneumothorax was 7.9 percent in the 38 patients with peak pressures of less than 30 cm of water, and 4 percent in the 26 patients with peak pressures of 70 to 110 cm of water (Table 1). The mean time to pneumothorax was 53±37 hours, in contrast to the statement by Chiche et al., “Clinical manifestations of barotrauma occur after the first week of ventilation, when inflammation peaks.” In such a diverse syndrome, the peak of inflammation is quite variable.We agree with Manning that transpulmonary pressure, not mouth pressure, determines overdistention.
Shapira et al. indicate that continuous recording of pressure (and volume) might be ideal, but we are unaware of data indicating that momentary increases in pressure (or volume) are deleterious. Although coughing or “fighting the ventilator” may increase mouth pressure, neither increases transpulmonary pressure. Therefore, they are not a cause of overinflation.
Our data in patients with the acute respiratory distress syndrome indicate that conventional ventilatory pressures and volumes do not appear to affect the lung adversely. Therefore, ventilatory pressures need not be limited to predefined values, such as a static pressure of 35 to 40 cm of water.
1 ReferencesJohn G. Weg, M.D.
University of Michigan Medical Center, Ann Arbor, MI 48109-0024Antonio Anzueto, M.D.
University of Texas Health Science Center at San Antonio, San Antonio, TX 782841
Dreyfuss D ,Saumon G . Ventilator-induced lung injury: lessons from experimental studies. Am J Respir Crit Care Med 1998;157:294-323
Web of Science | Medline
Author/Editor Response
Although we will be the first to welcome any multicenter trial that retests our results, we think the questions raised by Parsons and Matthay are unjustified. First of all, selection bias was very unlikely in our study because throughout the trial, randomization was performed only after enrollment and after we had checked the stability of base-line conditions during the control period and the calculation of the pressure–volume curve. Furthermore, randomization was performed in a stratified manner throughout the trial. We did not mention this in our article because of space constraints.
Because our hospital is a tertiary care center, most of the patients in our respiratory care unit have been transferred from other intensive care units, with previous lung disease and previous barotrauma and often having undergone mechanical ventilation for more than one week. Therefore, the high percentage of patients who were excluded should not cause surprise.
As we clearly stated in our article, the previous study with 28 patients was an interim analysis of the current trial, in which mortality was the planned primary end point. The differences in mortality between the groups were not significant at that time, but we decided to report the outstanding differences observed in the evolution of pulmonary function because of their clinical implications. The focus was on lung physiology.
We were aware that we would have to pay a high price for this unplanned report. To avert the bias introduced by a multiplicity of analyses in the same patients, we had to use very conservative statistical corrections for multiplicity, such as Bonferroni's method and Peto's correction, as used in our last report.1 It is important to stress that the penalty was merely a loss of power in both studies, not an increased chance of a type I error. In fact, it is harder to obtain a positive result when many interim analyses are performed in the context of appropriate statistical corrections.2
With regard to the high mortality rate in the control group (71 percent), we treated only patients in a medical intensive care unit. Eighty percent of our patients had sepsis as a risk factor for the acute respiratory distress syndrome, and 35 percent had coexisting conditions, including cirrhosis, infection with the human immunodeficiency virus, and immune disorders. Not surprisingly, in two recent reports on patients who had both sepsis and acute lung injury, including a report with Matthay as an author, a 70 percent mortality rate was reported.3,4 If one considers only the patients with the acute respiratory distress syndrome who were in a medical intensive care unit — as in our trial — the mortality was even higher (83 percent).4 Therefore, the comparison of our results with those of other trials involving patients with trauma or surgical patients (most of which included patients with acute lung injury plus the acute respiratory distress syndrome, as the study by Stewart et al. did5) is not justified.
5 ReferencesMarcelo B.P. Amato, M.D.
Carmen S.V. Barbas, M.D.
Carlos R.R. Carvalho, M.D.
Hospital das Clínicas, CEP 05344-000 São Paulo, Brazil1
Peto R ,Pike MC ,Armitage P , et al. Design and analysis of randomized clinical trials requiring prolonged observation of each patient. I. Introduction and design. Br J Cancer 1976;34:585-612
CrossRef | Web of Science | Medline2
Geller NL ,Pocock SJ . Interim analyses in randomized clinical trials: ramifications and guidelines for practitioners. Biometrics 1987;43:213-223
CrossRef | Web of Science | Medline3
Doyle RL ,Szaflarski N ,Modin GW ,Wiener-Kronish JP ,Matthay MA . Identification of patients with acute lung injury: predictors of mortality. Am J Respir Crit Care Med 1995;152:1818-1824
Web of Science | Medline4
Zilberberg MD ,Epstein SK . Acute lung injury in the medical ICU: comorbid conditions, age, etiology, and hospital outcome. Am J Respir Crit Care Med 1998;157:1159-1164
Web of Science | Medline5
Stewart TE ,Meade MO ,Cook DJ , et al. Evaluation of a ventilation strategy to prevent barotrauma in patients at high risk for acute respiratory distress syndrome. N Engl J Med 1998;338:355-361
Full Text | Web of Science | Medline
Author/Editor Response
On the basis of our findings, we concluded that routine limitation of peak inspiratory pressure and tidal volume, as described, in all patients at high risk for the acute respiratory distress syndrome does not reduce mortality.1 It remains to be determined whether mortality can be reduced by limiting transpulmonary pressures (hence, lung distention) in patients who would be at risk for lung overdistention with the use of conventional ventilatory strategies. In addition, other methods of lung protection (lung recruitment and maintenance of end-expiratory lung volume) may also offer an advantage in terms of reducing mortality.
On the basis of Amato and colleagues' observation of reduced mortality at 28 days,2 some would argue that the method of lung protection they used is beneficial, as the final statement in the letter by Chiche et al. seems to imply. These results are interesting but should be interpreted cautiously because of the high mortality rate in the conventional-ventilation group. The value of a one-time measurement of the lower inflection point may also be limited, since it is difficult to perform the measurement safely in patients with the most severe hypoxia, lung injury evolves with time,3 and incorrect data can be obtained if the contribution of the lung is not separated from that of the chest wall and abdomen.4 Further studies assessing the protective effects of lung-recruitment strategies are required.
4 ReferencesThomas E. Stewart, M.D.
Maureen O. Meade, M.D.
Wellesley Central Hospital, Toronto, ON M4Y 1J3, CanadaArthur S. Slutsky, M.D.
Mount Sinai Hospital, Toronto, ON M5G 1X5, Canada1
Stewart TE ,Meade MO ,Cook DJ , et al. Evaluation of a ventilation strategy to prevent barotrauma in patients at high risk for acute respiratory distress syndrome. N Engl J Med 1998;338:355-361
Full Text | Web of Science | Medline2
Amato MBP ,Barbas CSV ,Medeiros DM , et al. Effect of a protective-ventilation strategy on mortality in the acute respiratory distress syndrome. N Engl J Med 1998;338:347-354
Full Text | Web of Science | Medline3
Matamis D ,Lemaire F ,Harf A ,Brun-Buisson C ,Ansquer JC ,Atlan G . Total respiratory pressure-volume curves in the adult respiratory distress syndrome. Chest 1984;86:58-66
CrossRef | Web of Science | Medline4
Ranieri VM ,Brienza N ,Santostasi S , et al. Impairment of lung and chest wall mechanics in patients with acute respiratory distress syndrome: role of abdominal distension. Am J Respir Crit Care Med 1997;156:1082-1091
Web of Science | Medline
- Citing Articles (2)
Citing Articles
1
Joyce Da Silva Bevilacqua, Roberto Masaishi Yoshikawa. (2007) Monitoring lungs with electrical impedance tomography. Inverse Problems in Science and Engineering 15:4, 325-337
CrossRef2
Carlos Roberto Ribeiro Carvalho, Eduardo P. Bethlem. (2002) Pulmonary complications of leptospirosis. Clinics in Chest Medicine 23:2, 469-478
CrossRef
