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

Higher versus Lower Positive End-Expiratory Pressures in Patients with the Acute Respiratory Distress Syndrome.

NHLBI ARDS Clinical Trials Network.

New Engl J Med 2004; 351:327-336. [abstract]

Reviewed by Malik White MD, Mattel Children's Hospital at UCLA Medical Center, Los Angeles, CA

Review posted April 28, 2005

---

I. What is being studied?:

The study objective:

To compare clinical outcomes of patients with acute lung injury and acute respiratory distress syndrome treated with a higher positive end-expiratory pressure (PEEP)/lower FiO2 versus a lower PEEP/higher FiO2 ventilation strategy.

The study design:

Multicenter, prospective, randomized, controlled clinical trial.

The patients included:

Patients (n=549) from 23 ARDS network ICU's over a 28-month period with acute onset of:

1. PaO2/FiO2 ≤ 300 (adjusted for barometric pressure).
2. Bilateral infiltrates consistent with pulmonary edema on frontal chest radiograph.
3. Requirement for positive pressure ventilation through an endotracheal tube.
4. No clinical evidence of left atrial hypertension. If measured, pulmonary arterial wedge pressure < 18 mmHg.

The patients excluded:

Patients whose clinicians that are caring for them refuse to use Volume Assist/Control ventilation for at least 12 hours, age < 13 years, participation in other intervention trials in ALI, ARDS, or sepsis within the previous 30 days, > 36 hours since all inclusion criteria are met, neuromuscular disease that impairs ability to ventilate without assistance (such as C5 or higher spinal cord injury, amyotrophic lateral sclerosis, Guillain-Barre syndrome, and myasthenia gravis), pregnancy (negative pregnancy test for women of child-bearing potential), increased intracranial pressure, tricyclic antidepressant overdose (if most recent blood level elevated or no blood level available), Hgb SS, Hgb SC, or other conditions where hypercapnia would be contraindicated, severe chronic respiratory disease, morbid obesity, burns > 30% total body surface area, malignancy or other irreversible disease or condition for which 6-month mortality is estimated < 50%, bone marrow transplant, lung transplant, not committed to full support, severe chronic liver disease (Child-Pugh Score of 10 -15)

The interventions compared:

Higher PEEP/Lower FiO2 Prioritization Scheme (experimental group) versus Lower PEEP/Higher FiO2 Prioritization Scheme (standard therapy).

The outcomes evaluated:

The primary outcome: Mortality at 60 days after randomization.

Secondary outcomes:

1. Number of Ventilator Free Days (percentage of patients who achieve 48 consecutive hours of unassisted breathing at 28 days after randomization)
2. Number of ICU-free days at 28 days after randomization.
3. Number of Organ Failure Free days at 28 days after randomization, using previously validated definitions for renal, central nervous, coagulation, circulation, and hepatic organ and system failure
4. Number of days between the day of first meeting criteria to commence weaning and day 28 after randomization.
5. Number of days between the day of initially achieving unassisted ventilation and day 28 after enrollment.
6. Incidence of barotrauma (pneumothoraces, pneumatoceles > 2 cm largest diameter, pneumomediastinum).
7. Percentage of patients discharged alive from hospital. Patients alive in hospital at 60 days will be considered to have been discharged alive.
8. Mortality and days of unassisted breathing with pre-randomization PaO/FiO2 < 200.

II. Are the results of the study valid?

Primary questions:

1. Was the assignment of patients to treatments randomized?

Yes. A centralized interactive voice system was used to assign eligible patients and stratification was by hospital.

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

Was followup complete?

Yes. No patients were lost during the follow-up period.

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

Yes. There were no crossovers between groups.

Secondary questions:

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

Patients were presumably blind to treatment; health workers were not. Depending on which group the patient was assigned to, a different PEEP/Fi02 protocol was utilized for medical management.

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

Baseline characteristics reveal 2 significant differences between treatment groups: increased mean age and decreased mean Pa02:Fi02 ratio noted in the higher-PEEP group. The age difference (mean age of 54 in higher-PEEP group vs. 49 in lower PEEP group) may not be clinically significant. The difference in PaO2/FiO2 ratio (mean value was 14 points lower in higher-PEEP group) may indicate that the higher-PEEP group had slightly worse lung disease prior to receiving any treatment. These differences may have predisposed the higher-PEEP group to increased morbidity/mortality from the onset of the study.

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

Unknown; other aspects of the groups' care, e.g., fluids, inotropes, nutrition, were not discussed.

There were two fairly significant protocol changes during the study, both within the High-PEEP group.

A) The first 80 patients in the high-PEEP (treatment) group had "alveolar recruitment" maneuvers performed. The authors of this study felt that the subsequent increase in Pa02 after the recruitment maneuvers was "small and transient" and hence decided to stop using this co-intervention on the remaining 196 patients in the treatment group. In a previous 1998 study (1) comparing a high PEEP/low tidal volume/low peak inspiratory pressure strategy (versus standard ventilator strategy) in patients with ARDS, this "alveolar recruitment" maneuver was performed on all members of the treatment group to improve arteriolar oxygenation. In this 2004 study, the limited application of this co-intervention to almost 30% of the treatment group may have biased the results of the high-PEEP (treatment) group. Perhaps if this "alveolar recruitment" was applied to the entire treatment group, it would have had some small overall beneficial effect and could have improved the outcome of the treatment group.

B) After 171 patients were enrolled in the study, the primary treatment for the high-PEEP group was changed as the minimum PEEP value was increased from 5 to 12. Analysis of the primary outcome was done on the 171 patients in the treatment group prior to the protocol change, the 378 patients enrolled subsequent to the protocol change, and the total group of 549 patients. The analysis of the 3 groups revealed no significant difference in mortality between the treatment and control arms of the study.

III. What were the results?

1. How large was the treatment effect?

There were 276 patients in the high-PEEP group and 273 patients in the low-PEEP group. The trial was stopped early (after 549 patients were enrolled) because preset "futility" criteria were met. It was determined that had the trial been continued to enrollment of 750 patients, there was a less than 1 percent chance of demonstrating superiority of the high-PEEP strategy.

Relative Risk (RR): Using the unadjusted mortality: RR = 0.275/0.249 = 1.10. In this case it appears as if the treatment (higher PEEP is more harmful than the standard therapy.

The adjusted mortality rates represent estimated mortality rates for the lower-PEEP and higher-PEEP groups if the distribution of specific covariates had been completely balanced between the groups. These covariates identified as "predictors of mortality" were: age, APACHE III score, plateau pressure, missing plateau pressure, number of organ failures, number of hospital days before enrollment in the trial, and the alveolar-arterial oxygen difference in the partial pressure of oxygen. Using the adjusted mortality data: RR = 0.251/0.275 = 0.91. In this case the opposite trend appears as the treatment (higher PEEP) does cause less deaths than the standard therapy.

Relative Risk Reduction (RRR): Using the unadjusted mortality: (1-1.10 = -0.10). This negative value suggests that the treatment (higher PEEP) caused more deaths than the standard therapy. Using the adjusted mortality (1-0.91) x 100 = 9%. The high-PEEP strategy reduced the risk of death by 9%. Again these results are not statistically significant.

Despite significantly improved indices of oxygenation in the first 7 days, e.g., PaO2/FiO2 181 ± 115 in the lower PEEP group vs. 218 ± 85 in the higher PEEP group at day 7 (p < 0.05), there were no significant differences found between the lower-PEEP and higher-PEEP groups with regard to these secondary outcome variables: percent of patients breathing without assistance by day 28, number of ventilator free days from day 1 to day 28, number of days not spent in ICU from day 1 to day 28, percent barotrauma, and number of days without organ failure.

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

The confidence intervals for the relative risk and the relative risk reduction were not indicated but can be calculated. The 95% CI's for the RR of 1.1 are 0.84 and 1.46. Therefore, with 95% certainty, the high PEEP patients were between 0.84 and 1.46 times as likely to die as the low PEEP group. These CI's cross one, indicating a lack of statistical significance, and are rather broad, indicating a lack of precision as well.

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

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

The results of this study indicate that the use of higher PEEP/end-expiratory volumes does not improve mortality in adult patients with ARDS. This study suggests that there is no apparent difference in mortality between a high-PEEP and low PEEP strategy in the treatment of ARDS.

However, although the use of higher PEEP/end-expiratory volumes in this 2004 NHBLI study does not appear to improve mortality, it may improve Pa02/Fi02 ratios and allow lower Fi02 administration. The use of lower FiO2 levels could help minimize oxygen toxicity damage in PICU patients with ARDS. Further studies need to be done to evaluate the relevance of this particular finding.

As for the relevance of this study to the patients in the pediatric intensive care unit, it is often difficult to determine the true applicability of results from studies on adult patients because of the difficulty in extrapolating from adults to children.

2. Were all clinically important outcomes considered?

Particularly relevant to the PICU population, long-term morbidity and quality of life issues (pulmonary function, exercise tolerance, incidence and severity of subsequent reactive airway disease, etc) are important subjects that are not addressed in this article.

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

In this case the treatment (use of a higher-PEEP strategy) is not beneficial and does not appear to reduce mortality. The high-PEEP strategy may in fact allow lower concentrations of oxygen to be administered and may improve the PaO2/FiO2 ratio in patients with ARDS. However, there is not enough evidence in this study to suggest that these benefits are clinically significant. Because of these findings, use of traditional lower-PEEP settings with higher FiO2 levels (and low tidal volumes) appears to be the best approach for patients with ARDS.

As for cost, the article does not indicate differences in length of hospital (or ICU) stay between the 2 treatment groups; therefore the effect of the treatment on the cost associated with length of hospital stay cannot be addressed.

References:

1. Amato MBP et.al. Effect of a protective-ventilation strategy on mortality in the acute respiratory distress syndrome. N Engl J Med 1998;338:347-54 [abstract] [PedsCCM EB Review]

Comments

Back to the EB Journal Club Index