Criteria abstracted from The Users' Guide to Medical Literature, from the Health Information Research Unit and Clinical Epidemiology and Biostatistics, McMaster University

Highlighted lines and questions below provide links to the pertinent description of criteria in The EBM User's Guide, now available at the Canadian Centres for Health Evidence

Article Reviewed:

Please visit the new Evidence Based Journal Club Reviews

Effect of a Protective-Ventilation Strategy on Mortality in the Acute Respiratory Distress Syndrome.

Amato MBP, Barbas CSV, Medeiros DM, et al.

N Engl J Med 1998; 338: 347-354. [abstract]

Reviewed by Kazuo Obata, MD, St. Louis Children's Hospital

Review posted December 14, 1998

---

I. What is being studied?:

The study objective:

To determine whether a ventilation strategy designed to minimize lung injury can reduce pulmonary complications and mortality in adult patients with the acute respiratory distress syndrome.

The study design:

Randomized prospective control study.

The patients included:

53 adult patients with acute respiratory distress syndrome in 2 intensive care units in Brazil. ARDS was defined as an underlying disease process known to be associated with the acute respiratory distress syndrome along with a Murray lung injury score of 2.5 or higher and pulmonary arterial wedge pressure of less than 16 mmHg.

The patients excluded:

Previous lung or neuromuscular disease, mechanical ventilation for more than one week, uncontrolled terminal disease, previous barotrauma (pneumothorax, pneumomediastinum, or subcutaneous emphysema), previous lung biopsy or resection, an age of more than 70 years or less than 14 years, uncontrollable and progressive acidosis, signs of intracranial hypertension, and documented coronary insufficiency.

The interventions compared:

The protective approach was designed to prevent alveolar collapse and overdistention, and featured permissive hypercapnia. Using only pressure-limited modes of ventilation, a stepwise algorithm was used to maintain the goals as follows: tidal volume < 6ml/kg, respiratory rate < 30/min (depressing respirative drive as need with infusions of fentanyl and diazepam), driving pressure (PIP - PEEP) < 20 cmH2O, and a peak airway pressure < 40 cmH2O. Arterial pCO2 levels of up to 80 mmHg were permitted; slow infusion of sodium bicarbonate was used if the arterial pH < 7.2. PEEP was preset at 2 cm of water above P-flex. (P-flex is the lower inflection point on the static pressure-volume curve. If sharp P-flex could not be determined, PEEP of 16 cmH2O was used empirically). During reduction of ventilatory support, PEEP was then kept constant until the FiO2 < 0.4.

The conventional approach utilized volume-cycled assisted or controlled ventilation, sought to maintain pCO2 at 35 to 38 mmHg independent of airway pressures, and used PEEP to achieve an FiO2 < 0.6. The tidal volume was set at 12 ml/kg, inspiratory flow rates at 50 to 80 L/min, inspiratory pause 0.4 sec, and sedation was used to prevent respiratory rate > 24 /min or pCO2 < 25 mmHg.

The outcomes evaluated:

The primary outcome was the mortality at 28 days. Secondary outcome included survival to hospital discharge, weaning rate from ventilator at 28 days (adjusted for APACHE 2 score), the rate of barotrauma, death in ICU, death after weaning, rates of complications (nosocomial pneumonia, neuropathy), cause of in-hospital death, and the use of paralyzing agents, dialysis, and transfusion.

II. Are the results of the study valid?

Primary questions:

1. Was the assignment of patients to treatments randomized?

Yes. After initial enrollment, all patients underwent respiratory and hemodynamic stablization for at least 30 minutes with a standardized protocol. After this initial control period, patients were then randomized with sealed envelopes and a 1:1 assignment scheme.

2. Were all patients who entered the trial properly accounted for and attributed at its conclusion?

Was followup complete?

Probably yes. However, it is not clearly described whether all of the deaths occurred in hospital. If there were any out-of-hospital deaths after 28 days, these would not be reported by "in-hospital death" or "mortality at 28 days."

Were patients analyzed in the groups to which they were randomized?

Yes. The protective or conventional mechanical ventilation protocols were rigorously maintained until the patient was extubated or died. Although there were five minor protocol violations (1 in conventional-ventilation group and 4 in protective-ventilation group), all patients were analyzed in their assigned groups.

Secondary questions:

3. Were patients, health workers, and study personnel "blind" to treatment?

No. Although the patients presumably were blinded, the study personnel and/or health workers were aware of the treatment assignments. The complexity of the procedures precluded the double-blind protocol as is true of many other studies to compare ventilatory strategies. The extent of the potential bias that this introduces is difficult to assess.

4. Were the groups similar at the start of the trial?

Yes. After stabilization during the 30 minute control period, respiratory, hemodynamic, and laboratory measurements were performed. There were no significant differences between groups with regard to age, duration of previous mechanical ventilation, evidence of organ failure, APACHE 2 score, critical-care score, lung-injury score, ventilator score, respiratory tract infection, incidence of sepsis, PaO2/FiO2 ratio, P-flex, static compliance, or primary diagnosis.

5. Aside from the experimental intervention, were the groups treated equally?

Probably yes. The investigators employed stringent and elaborate algorithms for both respiratory and general care (1).

However, after the first 28 patients had been enrolled, a beneficial effect of the protective approach became evident; an interim analysis was performed after each new block of five patients; the study was stopped appropriately during the fifth interim analysis to avoid subjecting the patients to an unnecessary continuation of the protocol. Knowledge of the result of these interim analyses potentially might have further influenced the care by the unblinded personnel and/or the investigators.

III. What were the results?

1. How large was the treatment effect?

Mortality at 28 days was 38% for the protective-ventilation group and 71% for the conventional-ventilation group. (absolute risk reduction 33%, relative risk reduction 46%).

The other outcomes that showed significant reduction in the protective-ventilation group were the proportion of patients weaned from mechanical ventilation successfully by 28 days (66% treated vs. 29% controls; absolute risk reduction of not weaning was 37%, relative risk reduction 52%) and the frequency of barotrauma (7% treated vs. 41% controls; absolute risk reduction 34%, relative risk reduction 83%).

2. How precise was the estimate of the treatment effect?

Outcome	RRR	95% CI	ARR	95% CI	NNT	95% CI
28 day mortality	46%	11% to 82%	0.328	0.074 to 0.582	3	2 to 13
no weaning at 28 days	52%	17% to 87%	0.368	0.117 to 0.619	3	2 to 9
barotrauma	83%	31% to 100%	0.347	0.129 to 0.565	3	2 to 8

None of the 95% confidence intervals for the RRR, ARR, and NNT (number needed to treat) for the three outcome measures described above cross zero. For example, the true absolute risk reduction of mortality at 28 days from the treatment strategy will range from 7% to 58%, with 95% certainty. The 95% confidence intervals for the NNT suggest that we would only need to use the protective ventilation strategy on 2 to 13 patients to prevent one death at 28 days.

Multivariate analysis also supports this treatment effect; the APACHE 2 scores and the ventilatory group assignment (RR 0.19, 95% CI 0.08, 0.47) were the only base-line factors that significantly influenced the relative risk of death at 28 days.

IV. Will the results help me in caring for my patients?

1. Can the results be applied to my patient care?

Maybe. There are a number of caveats one must consider prior to applying this study to our patients.

The study was done in an adult population, and given the differing underlying diseases and complications of children compared to adults, one must always be skeptical of applying adult clinical trials directly to pediatrics. Furthermore, the control mortality and barotrauma rates in this study were very high compared to current estimates, suggesting that either these patients were more critically ill or overall care was suboptimal.

In addition, although the apparent consistency of patient care, both with regard to the study protocol as well as general care, adds strength to this study's validity, it is unclear whether the study's conclusions are applicable in the real world of less consistent care models.

Finally, there are aspects to following their ventilatory management strategy that are not commonly done or are difficult to perform, such as the creation of a static pressure-volume curve for the determination of P-flex.

2. Were all clinically important outcomes considered?

Not completely. Mortality at 28 days was arbitrarily chosen, although this is a common endpoint for such studies. In addition, the long-term morbidity and mortality are not known. For example, the number of patients readmitted after discharge or those left with significant long term morbidity (e.g., renal failure or persistent respiratory dysfunction) is not reported.

3. Are the likely treatment benefits worth the potential harms and costs?

Maybe. Although they did not reach statistical significance, there were parameters that showed a worse outcome for the protective-ventilation group. These included death after weaning (0% vs 14%), nosocomial pneumonia (46% vs 59%), neuropathy after extubation (0% vs 7%), and dialysis required (21% vs 24%).

Since the total in-hospital death rate did not improve significantly, it is not clear whether this represents a true treatment effect (i.e., somehow the protective strategy put patients at risk for delayed complications) or whether the small study size "allowed" several late deaths to dilute out the treatment effect.

With relatively high PEEP levels, there is always a concern about the influence on hemodynamics. The authors examined this issue elsewhere (2) and found that the cardiac output and oxygen delivery were increased in the protective-ventilation group immediately after the start of the study and slowly decreased approaching the level of the control group. These hyperdynamic effects were independently related to respiratory acidosis but not to the level of PEEP.

The study did not address costs directly, although it would appear that there are few additional costs involved with protective ventilation and the potential savings from more rapid weaning from mechanical ventilation could be substantial.

References

1. Amato MB, 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. [abstract]
2. Carvalho CR, et al. Temporal hemodynamic effects of permissive hypercapnia associated with ideal PEEP in ARDS. Am J Respir Crit Care Med 1997;156:1458-1466. [abstract]

Comments

Back to the EB Journal Club Index