The trusted source for
healthcare information and
Managing ARDS: Steroid Saga Sequel
Abstract & Commentary
By Saadia R. Akhtar, MD, MSC, Idaho Pulmonary Associates, Boise, is Associate Editor for Critical Care Alert.
Dr. Akhtar reports no financial relationship to this field of study.
Synopsis: This multicenter, randomized, double-blind, placebo-controlled trial finds that early initiation of low-dose, prolonged glucocorticosteroid therapy for ARDS results in improved lung injury scores and a greater likelihood of successful extubation by day 7. Other clinical outcomes may also be favorably impacted.
Source: Meduri GU, et al. Chest. 2007;Apr;131(4):954-963.
Systemic and pulmonary inflammation are key in the development and progression of the acute respiratory distress syndrome (ARDS). Lack of improvements in inflammatory markers and in the lung injury score (LIS) by day 7 are associated with higher mortality.1 Some small studies have suggested that low-to-moderate-dose, prolonged corticosteroid infusion may reduce inflammatory markers, LIS, and duration of mechanical ventilation.2 With this foundation, Meduri et al set out to determine whether early initiation of low-dose prolonged glucocorticosteroid replacement for ARDS leads to improvement in LIS (1 point reduction) and greater success of extubation by day 7.
A randomized double-blind placebo-controlled study was conducted at 5 hospitals in Memphis, Tennessee, over 5 years. Adult patients meeting usual criteria for ARDS within 72 hours of initiation of mechanical ventilation were eligible. Once enrolled, patients were randomized 2:1 to receive methylprednisolone vs placebo: a bolus dose of 1 mg/kg was followed by daily doses of 1 mg/kg/day from day 1-14 (or until extubation, whichever came first), and then a taper (0.5mg/kg/day for 1 week, 0.25mg/kg/day for 4 days, and 0.125mg/kg/day for 3 days). Low-tidal-volume ventilation per the ARDSnet protocol was utilized once the data from the ARDS Network's initial trial became available. Screening for ventilator-associated pneumonia was performed prior to initiation of drug, then every 5-7 days until extubation. Neuromuscular blockers were not permitted. Finally, patients whose LIS did not improve between study day 7-9 were removed from the study and treated unblinded with 2mg/kg/day methylprednisolone. Intention-to-treat analysis was performed, and other standard statistical techniques were employed. Interim data analyses were done after enrollment of every 30 patients.
Over 5 years, 500 patients were screened. Of these, 91 (63 methylprednisolone, 28 placebo) met inclusion criteria and agreed to participate. At baseline, patients in the 2 groups were similar in all ways (including precipitating factor for ARDS) except that there were a larger number of subjects in the placebo group with catecholamine-dependent septic shock. At day 7, there were marked, statistically significant differences between the corticosteroid and placebo groups in 1-point reduction in LIS (69.8% vs 35.7%) and successful liberation from mechanical ventilation (54% vs 25%). A variety of other clinical parameters were similarly impacted: C-reactive protein level, PaO2/FIO2, multiple organ dysfunction score, length of ICU stay and ICU mortality (hospital and post-discharge mortality up to 12 months) were not significantly different.
Interestingly, patients whose initial short cosyntropin stimulation test revealed relative adrenal insufficiency had less improvement in both study arms. There were less nosocomial infections in the methylprednisolone group and no difference in number of patients who developed neuromuscular weakness. Five patients in the control group and 10 in the methylprednisolone group were transitioned to open-label glucocorticosteroids when LIS did not improve significantly at days 7-9.
Short-term high-dose corticosteroid infusions do not improve and in fact may worsen outcomes in patients with ARDS.3 As noted above, some small studies have suggested that low-to-moderate-dose prolonged glucocorticosteroid infusion reduces inflammatory markers, LIS and duration of mechanical ventilation.2 Such work motivated a recent large multicenter study of low-dose glucocorticosteroids for late ARDS (beginning at day 7 or later, for a total of 21 days of therapy). Investigators found that there was no overall mortality benefit at 60 or 180 days; furthermore, there was the suggestion of worse outcome in those patients started on glucocorticosteroids at day 14 or later but a trend towards improved survival in those patients started on glucocorticosteroids between days 7-13.4 Thus, interest remains in using glucocorticosteroids early in ARDS. Supportive evidence for this was also provided by a post hoc analysis of a study using glucocorticosteroids for septic shock, which suggested an association between ARDS and mortality benefit with glucocorticosteroids.5
Meduri et al's investigation of methylprednisolone infusion beginning (at 1mg/kg/day) within 72 hours of onset of ARDS and tapering off over 2-4 weeks finds a marked difference in the primary endpoints of LIS and successful extubation by day 7. Although these are interesting and exciting results, they are not, as the authors themselves acknowledge, definitive but rather hypothesis-generating for a larger clinical trial. That is, whether these findings signify that long-term outcomes such as mortality will be similarly improved with this treatment in most patients with ARDS remains unknown. The critical care community has learned from multiple prior experiences that surrogate markers of long-term outcomes may not be reliable or accurate predictors. This current data should be assessed with that information in mind.
This work has limited generalizability as the patient population was fairly specific. This was a regional rather than national or international cohort. In addition, as with the recent trial by the ARDSnet investigators,4 a large percentage of screened patients were excluded from enrollment by the predefined study criteria. The study was not powered to assess important clinical outcomes such as hospital length of stay or mortality (the placebo group in particular was quite small). Furthermore, the results and analysis are complicated by several factors. There was a significant baseline difference between the treatment arms in numbers of patients with catecholamine-dependent septic shock.
Based on other evidence about glucocorticosteroids in septic shock, we must question the conclusions about the basis of the outcome differences in this study. Due to the time period during which this trial was conducted, some patients did not receive a low-tidal-volume ventilation protocol, and there was no consistent weaning protocol in place. Both of these practices may have impacted the magnitude of difference in outcomes between the 2 treatment groups. The fact that some patients were taken off the study to receive glucocorticosteroids may change the results in a variety of unpredictable ways. (I am intrigued and surprised by the fact that the investigators used a study of 24 patients to take current study patients off protocol for failure to improve LIS at day 7-9 .)
Thus, contrary to Annane's recommendation (in the editorial accompanying this paper)6 that all patients with ARDS (except those at > 2 weeks) be treated with prolonged glucocorticosteroids as Meduri et al have proposed, I suggest caution until other larger studies examine this protocol. The findings are promising but I believe additional evidence is needed to more clearly guide our usage of glucocorticosteroids in patients with ARDS. Those physicians who decide to use the therapy recommended by Meduri et al now must at least carefully monitor their patients for infection and neuromuscular weakness, aggressively manage blood glucose, and use low-tidal-volume ventilation.