Lung-Protective Ventilation and ARDS Mortality: Part II

Synopsis : Patients with ARDS who were ventilated with a lung protective strategy had the same mortality rates as those ventilated conventionally.

Source: Stewart TE, et al. N Engl J Med 1998;338: 355-361.

Stewart and colleagues performed a study similar to that of Amato et al at eight tertiary referral centers in Canada. A lung protective ventilatory strategy was compared to conventional management in a prospective, randomized, controlled trial.

In the treatment group (n = 60, age = 59 ± 17 years, total number of organ failures 1.4 ± 1.0, and Apache II score 22.4 ± 7.3), the peak airway pressure was limited to 30 cm H2O and the tidal volume to 8 mL/kg. In the control group (n = 60, age = 58 ± 19 years, number of organ failures 1.3 ± 1.2, and Apache II score 21.5 ± 9.5), the peak airway pressure was limited to 50 cm H2O, and the tidal volume was set between 10 and 15 mL/kg. In both groups, volume ventilation with a decelerating flow pattern was used, and PEEP was adjusted to maintain the FIO2 at 0.5 or less, with the SaO2 maintained at 89-93%. In both groups, the target PCO2 was 35-45 mmHg, and hypercapnia was allowed before ventilatory limits could be exceeded. All other aspects of care were consistent between the two groups.

There was no difference in hospital mortality between the two groups (50% treatment vs 47% control). On day 3, peak airway pressure (34.0 ± 11.0 vs 23.6 ± 5.8 cm H2O), tidal volume (10.8 ± 1.0 vs 7.2 ± 0.8 mL), plateau pressure (28.5 ± 7.2 vs 22.2 ± 3.8 cm H2O), and respiratory rate (17.0 ± 6.0 vs 23.1 ± 6.3 breaths/min) all differed between groups. Similar differences existed from day 1 to day 7. In addition, a larger percentage of patients underwent dialysis (13 vs 5) and received paralytic drugs (23 vs 13) in the treatment group. Maximal PaCO2 (54.8 ± 18.8 vs 45.7 ± 9.8 mmHg) was greater in the treatment group, and pH (7.29 vs 7.34) was lower in the treatment group. Differences in PEEP only occurred on day 1 (8.6 ± 3.0 vs 7.2 ± 3.3 cm H2O), and, on all other days, PEEP in both groups was 8-9 cm H2O. No difference between groups was observed in minute ventilation or FIO2 on any day, nor were there differences in the incidence of barotrauma (10% vs 7%).


An abundance of animal data indicate that mechanical ventilation can cause or extend lung injury. High peak alveolar pressure at zero PEEP results in an acute lung injury similar to ARDS in humans (Dreyfuss D. Am Rev Respir Dis 1985;132:880; Dreyfuss D. Am Rev Respir Dis 1988;137:1159; Dreyfuss D. Am Rev Respir Dis 1993;148:1194; Parker JC. Am Rev Respir Dis 1990;142:321; Tsuno K. J Appl Physiol 1990;69:956; Webb HH, et al. Am Rev Respir Dis 1974;110:556) and, also, a more rapid extension of existing lung injury (Dreyfuss D. Am J Respir Crit Care Med 1995;151:1568).

Most importantly, numerous groups have shown that the use of PEEP protects against development of high pressure lung injury (Webb HH, et al. Am Rev Respir Dis 1974;110:556; Dreyfuss D. Am Rev Respir Dis 1988;137:1159; Corbridge TC. Am Rev Respir Dis 1990;142:311; Muscedere JG. Am J Respir Crit Care Med 1994;149:1327). In addition, case series by Hickling and colleagues (Intens Care Med 1990;16:372; Crit Care Med 1994;22:1568), by Lee and associates (Chest 1990;97:430), and by Stocker and associates (Chest 1997;111:1008) have all shown improved mortality in ARDS using a lung protective strategy.

Both the Amato and Stewart studies were based on these data, yet Amato et al show marked improvement in 28-day mortality. Stewart et al show no difference in mortality, with the mortality in both groups tending to be higher than in Amato's treatment group, despite the fact that Amato et al apparently recruited more severely ill patients than Stewart (Amato, number of extra-pulmonary organ failures about 2.6, vs Stewart's total organ failures of 1.4).

The primary reason why these differences in outcome were observed may have been the level of PEEP used, and also the method of selection in Amato's treatment group. Using pressure-volume curves, the treatment group received 16 cm H2O PEEP for the first 36 hours after enrollment and 13 cm H2O PEEP from day 2 to day 7. On the other hand, both Amato's control group and Stewart's control and treatment groups set PEEP based on the relationship among PaO2, PEEP and FIO2, with PEEP levels in all three groups averaging about 7-9 cm H2O through the first seven days of enrollment. The higher mortality in these last three groups may be directly attributable to the inadequate use of PEEP, especially in the first few days of treatment.

Assuming all patients enrolled in both studies actually required PEEP at the level of Amato's treatment group (that is, an average of 16 cm H2O), the lower level of PEEP (that is, 7-9 cm H2O) used other than in Amato's treatment group may have influenced the level of consolidation, and perhaps as a result the development of infection, and thus may be a primary factor responsible for the high mortality in Amato's control group in the three days after enrollment (30% of patients in the control group died in this period).

The second potential reason for the differences in outcome between the two studies may be that Stewart et al protected the peak alveolar pressure in both arms of their study. In spite of the fact that the study protocol called for a VT of 10-15 mL/kg in the control group with pressure limited to 50 cm H2O, and a VT of 8 mL/kg with the pressure limited to 30 cm H2O in the treatment group, little actual difference in these two variables was observed between the groups. Tidal volumes were about 7.0 mL/kg in the treatment group and 10-11 mL/kg in the control group, with plateau pressures in both groups throughout the study approximately 28.6 ± 7.2 cm H2O (highest recorded). As a result, regardless of what the protocol indicated, a lung protective strategy was actually used in both groups. This, along with the fact that PEEP did not differ significantly in Stewart and colleagues' treatment groups, would explain the lack of difference in mortality.

The study by Stewart et al clearly indicates that clinicians today are reluctant to ventilate patients with high airway pressures. The peak pressures of 70-90 cm H2O commonly seen in many ARDS patients who were ventilated in the mid 1980s in most centers are a thing of the past. Today, it is common to limit airway pressures in ARDS.

I do not see a different message in these two studies, and I believe that a lung protective ventilatory strategy should always be used in managing patients with ARDS. We have all been convinced of the detrimental effects of a high peak alveolar pressure. What remains is to convince everyone of the beneficial effects of the use of high PEEP, early in ARDS, based on the actual lung mechanics of the individual patient.