Rebound Hypoxemia in Mechanically Ventilated Patients With COVID-19 Completing a Standard 10-Day Course of Corticosteroid Therapy.

Abstract

Background: Rebound hypoxemia may occur after cessation of corticosteroid therapy for COVID-19 pneumonitis. We aimed to determine the incidence of this phenomenon in mechanically ventilated patients with COVID-19 completing corticosteroid therapy. Methods: We conducted a retrospective observational cohort study across 2 tertiary ICUs from September to December 2021. We included all adult patients receiving invasive mechanical ventilation on completion of a 10-day course of dexamethasone for COVID pneumonitis. Our primary outcome was change in P_aO2/F_IO2 ratio in the 7 days following cessation of dexamethasone. Secondary outcomes included duration of ventilation, frequency of corticosteroid recommencement, and mortality. Results: We studied 88 subjects. Median age was 61 years (interquartile range [IQR] 51-67), and median duration of mechanical ventilation was 14 days (IQR 11-26). On completion of dexamethasone, 62 subjects (70%) remained off corticosteroid therapy for the subsequent 7 days. Of these, 63% (n = 20) had a stable P_aO2/F_IO2 ratio on day 7 (mean change 197 ± 42 mm Hg to 258 ± 83 mm Hg). Subjects who had a reduction in P_aO2/F_IO2 ratio >20 mm Hg by day 7 (mean change 214 ± 68 mm Hg to 152 ± 78 mm Hg) were more likely to die in hospital (P < .001). These subjects had a higher C-reactive protein level at time of steroid cessation (204 ± 87 mg/L, P = .17), which remained persistently elevated (206 ± 89 mg/L, P = .01) on day 7. The 30% of subjects who recommenced steroids required a longer duration of ventilation (13.5 vs 24.5 d, P = .002) but showed no differences in ICU mortality (19% vs 27%, P = .43). Conclusions: Rebound hypoxemia and steroid recommencement were common occurrences in mechanically ventilated subjects with COVID-19. Rebound hypoxemia was associated with higher mortality. Steroid recommencement was associated with longer duration of mechanical ventilation but no significant difference in mortality.