Effect of prolonged methylprednisolone therapy in unresolving ARDS: A randomized controlled trial
This review may be edited
Posted By: Gordon S. Doig
Posted Date: 2May2001
Title: Effect of prolonged methylprednisolone therapy in unresolving ARDS: A randomized controlled trial
Authors: Meduri, Headley et al
Reference: JAMA 1998;280(2):159-165
Link: Click here for a direct link to the paper. A password may be required for access to fulltext.
Abstract: CONTEXT: No pharmacological therapeutic protocol has been found effective in modifying the clinical course of acute respiratory distress syndrome (ARDS) and mortality remains greater than 50%.
OBJECTIVE: To determine the effects of prolonged methylprednisolone therapy on lung function and mortality in patients with unresolving ARDS.
DESIGN: Randomized, double-blind, placebo-controlled trial.
SETTING: Medical intensive care units of 4 medical centers.
PARTICIPANTS: Twenty-four patients with severe ARDS who had failed to improve lung injury score (LIS) by the seventh day of respiratory failure.
INTERVENTIONS: Sixteen patients received methylprednisolone and 8 received placebo. Methylprednisolone dose was initially 2 mg/kg per day and the duration of treatment was 32 days. Four patients whose LIS failed to improve by at least 1 point after 10 days of treatment were blindly crossed over to the alternative treatment.
MAIN OUTCOME MEASURES: Primary outcome measures were improvement in lung function and mortality. Secondary outcome measures were improvement in multiple organ dysfunction syndrome (MODS) and development of nosocomial infections.
RESULTS: Physiological characteristics at the onset of ARDS were similar in both groups. At study entry (day 9 [SD, 3] of ARDS), the 2 groups had similar LIS, ratios of PaO2 to fraction of inspired oxygen (FIO2), and MODS scores. Changes observed by study day 10 for methylprednisolone vs placebo were as follows: reduced LIS (mean [SEM], 1.7 [0.1] vs 3.0 [0.2]; P<.001); improved ratio of PaO2 to FIO2 (mean [SEM], 262 [19] vs 148 [35]; P<.001); decreased MODS score (mean [SEM], 0.7 [0.2] vs 1.8 [0.3]; P<.001); and successful extubation (7 vs 0; P=.05). For the treatment group vs the placebo group, mortality associated with the intensive care unit was 0 (0%) of 16 vs 5 (62%) of 8 (P=.002) and hospital-associated mortality was 2 (12%) of 16 vs 5 (62%) of 8 (P=.03). The rate of infections per day of treatment was similar in both groups, and pneumonia was frequently detected in the absence of fever.
CONCLUSIONS: In this study, prolonged administration of methylprednisolone in patients with unresolving ARDS was associated with improvement in lung injury and MODS scores and reduced mortality.
Are the Results Valid?
1. Was the assignment of patients to treatments randomized? ( Was allocation concealment maintained?)
Yes. Patients were randomized in blocks of 3 to intially receive steroids or placebo in a 2:1 ratio. Randomization was stratified by site.
2. Were all patients who entered the trial properly accounted for and attributed at its conclusion?
2a. Was followup complete?
Yes. All patients who were randomized to the trial were followed-up.
2b. Were patients analyzed in the groups to which they were randomized?
No. Since this was a cross-over trial, some patients received both 'treatments' which makes analysis and interpretation complex.

The two fundamental assumptions that allow the conduct of cross-over trials are: 1) The disease process under study is chronic and will remain farily stable over the duration of the trial and 2) Sufficient time is provided during the cross-over phase to allow the study drug to 'wash out' of the patients.

3. Were patients, health workers, and study personnel blind to treatment?
Yes. This trial was fairly well blinded.
4. Were the groups similar at the start of the trial?
Yes. Both groups were similar with regards to severity of illness.
5. Aside from the experimental intervention, were the groups treated equally?
What are the Results?
1. How large was the treatment effect?
The primary result emphasized by the authors was a reduction in hospital mortality from 5/8 (62%) in the 'control' group to 2/16 (12%) in the 'steroid' group (p=0.03) HOWEVER this interpretation is misleading. Considering that this is a 'cross-over trial' composed of two 10-day treatment periods, a more conservative interpretation would focus on Stage 1 of the trial only.

During the first 10 day course of treatment (Stage 1), 16 of the 16 patients who received steroids survived whereas 6 of the 8 patients who did not receive steroids survived (p=0.2, Fisher's Exact).

If we wish to consider Stage 2:

Of the 22 patients who survived Stage 1, 4 patients crossed over to steroid treatment, 16 remained on their orginal course of steroids and 2 patients remained as controls. Of the 4 patients crossed over to steroids, only one survived (25%). Of the 16 who remained on steroids, 14 survived and both of the patients who never received steroids survived. During Phase II, of the 20 patients who received steroids, 15 survived (75%) and 2 of the 2 patients who did not receive steroids survived (p=1.0)

Individually, each Stage is non-significant with regards to mortality.

Hmmm.... suddenly doesn't look so convincing, does it???

Regardless of how one views this trial design (cross-over or standard RCT), if a true intention to treat analysis is performed based on initial allocation to steroid or control group, the question being addressed is not that of steroid treatment vs. no steroid. Because a full 50% of the control patients were crossed-over to receive steroids, the inferences drawn pertain to steroid treatment vs. late steroid treatment.
2. How precise was the estimate of the treatment effect?
This trial is extremely small and any confidence intervals generated will be VERY large.
Will the Results Help Me In Caring For My Patients?
1. Can the results be applied to my patient care?
Not at this point in time.
2. Were all clinically important outcomes considered?
No. In the context of a cross-over trial, mortality measured after patients have been through both stages of the trial is inappropriate.
3. Are the likely treatment benefits worth the potential harms and costs?
Unknown at this time.

What other people had to say about:
Effect of prolonged methylprednisolone therapy in unresolving ARDS: A randomized controlled trial

Add Your Comment
[ Back ]

Any questions or comments please contact Gordon S. Doig
Implemented and designed by John Soer and Gordon Doig
Page last modified on Friday August 10, 2001