Plasma and BAL Cytokine Response to Corticosteroid Rescue Treatment in Late ARDS

Abstract

In late ARDS, a persistent and exaggerated inflammatory response causes recurrent injury to the alveolocapillary barrier and amplification of intra-alveolar fibroproliferation. When ARDS patients fail to improve, corticosteroid (CS) rescue treatment frequently leads to rapid improvements in lung function. We tested the hypothesis that response to CS treatment is related to suppressing the inflammatory response by comparing changes in lung function to inflammatory cytokine (IC) levels in the plasma and BAL. Blood samples were obtained on days 1, 3, 5, and 7 of ARDS, and on days -5, -3, 0 (initiation of treatment), +3, +5, +7, +10, and +14 of CS treatment. Bilateral BAL was obtained on day 1 of ARDS, before administration of CS treatment, and at weekly intervals. We analyzed changes in IC levels during CS treatment in relation to improvements in lung injury score (LIS), indices of endothelial permeability, and final outcome. We also analyzed data to identify timing to a significant reduction in plasma IC levels and predictors of response. Nine patients entered the study. CS treatment was initiated 15 +/- 9 days into ARDS. Improvement in LIS (> 1-point reduction) was rapid (< 7 days) in five, delayed (< 14 days) in two, and absent in two. Baseline plasma and BAL IC levels in study patients were similar to a previously reported comparison group of 12 ARDS nonsurvivors. No significant changes in plasma and BAL IC levels were observed before CS administration. Following initiation of CS treatment, significant reductions in plasma tumor necrosis factor-alpha and interleukin 6 (IL-6) levels were seen by day 7 in both rapid and delayed responders (p = 0.03). IL-1 beta was significantly reduced by day 5 (p = 0.04) in rapid responders and by day 10 (p = 0.03) in delayed responders. In responders, improvement in LIS and BAL albumin paralleled reduction in plasma and BAL IC levels. At initiation of treatment, rapid responders had significantly lower tumor necrosis factor-alpha and IL-6 levels. Nonresponders had a significantly higher plasma IL-6 level on days 1 to 3 of ARDS (p = 0.004) and lower ratio of arteriolar oxygen tension to inspired oxygen concentration at initiation of treatment (p < 0.01). In patients with late ARDS and a low likelihood of survival, prolonged corticosteroid rescue treatment was associated with a reduction in plasma and BAL IC levels and parallel improvements in indices of endothelial permeability and LIS.