Results:
Figure 1 shows the study flowchart. Out of 768 patients intubated during the study period, 287 patients were enrolled and randomized into two groups. Some patients (n=36) were missed despite best efforts of the trainee fellow (BP) due to his other commitments. Table 1 shows admitting diagnosis and indication for PICU admission, while Table 2 shows baseline characteristics. Extubation was deferred in 21 patients for various reasons [neurogenic breathing (5), persistent encephalopathy (5), central apnea (3), neuromuscular weakness (2), seizures (2), pulmonary hemorrhage (1), pulmonary arterial hypertension (1), intracranial bleeding (1) and lung collapse (1)]. Twenty-eight patients (14 in each group) needed reintubation for non-PEAO reasons [persistent encephalopathy (9), neurogenic breathing (6), pulmonary hemorrhage (4), poor respiratory efforts (4), lung collapse (2), generalized spasms (1), seizures (1) and central apnea (1)]. After excluding these, 238 patients were available for per-protocol analysis (Figure 1). More than one-third of patients (36%) had chronic malnutrition, while 41% had acute malnutrition. Median body mass index was ~15Kg/m2. Respiratory failure (50%) was the commonest indication for PICU admission, followed by intracranial hypertension (22%). Baseline characteristics of the enrolled cohort (n=287) were similar to the assessable one (n=238) (data not shown). Among assessable patients (n=238), median duration of intubation and/or hand ventilation in pre-PICU areas was 12h (IQR, 0-24). Sixty-four (27%) patients spent > 24 h in emergency room or pediatric wards before they could get transferred to PICU. Fourty-two (18%) children had multiple airway manipulations (i.e., had to be reintubated for accidental extubation or tube obstruction) before their first elective extubation; 10 patients were reintubated for more than once. It is not that rare in our setting due to inadequate nursing care. As treating team was given liberty to time extubation based on the prevalent clinical situation, patients received a median of four dexamethasone doses prior to extubation instead of the planned five; 36 (15%) patients received three doses or less. Serum albumin levels were not available in 22 patients at extubation. Hypoalbuminemia and clinically manifest edema were present in 75% (162/216) and 26% of patients respectively. Both groups were comparable at baseline.
A total of 78 patients (33%) developed PEAO with no difference between two groups [LD, 41/121 (risk ratio, 0.34) vs HD, 37/117 (risk ratio, 0.32); p=0.71] (Table 3). The risk difference between the LD and HD group is 0.02 (90% CI, -0.07 to 0.12). The confidence interval of risk difference touches the non-inferiority margin of 0.12, hence the overall result is non-significant. Nineteen patients (19/238, 8%) needed reintubation with similar incidence in two groups [LD, 10/121 (8%) vs HD, 9/117 (7.8%); p=0.87]. Among PEAO patients (WCS≥4) not requiring reintubation (n=59; LD, 31 vs HD, 28), mean WCS was similar at different points of post-extubation observation (two-way mixed model ANOVA, interactional p=0.22) [Figure 2 (A)]. Further, there was no difference in time to recovery from PEAO (i.e., time to achieve WCS< 2 irreversibly) between two groups (log-rank test, p=0.90) [Figure 2 (B)]. Even after including 28 patients who were clinically considered to be re-intubated due to non-PEAO reasons (14 in each group), no difference was observed in incidence of either PEAO [LD, 55/135 (41%) vs HD, 51/131 (39%); p=0.76] or reintubation [LD, 24/135 (18%) vs HD, 23/131 (18%); p=0.96].
Post hoc analysis revealed non-significant increase in occurrence of PEAO in the subgroup of children intubated for more than 7 days (n=91) with the low dose regime [LD, 25/46 (54%) vs HD, 16/45 (36%); RR, 1.52; 95%CI, 0.95-2.45; p=0.07]. However, incidence of reintubation was comparable among children intubated for more than 7 days with the two dose regimes [LD, 6/46 (13%) vs HD, 5/45 (11%); p=0.77].
Univariate analysis identified longer (>7 days) intubation, prolong ventilation and higher PRISM III score as significant risk factors for PEAO (Table 3). Multiple airway manipulations prior to the first elective extubation seems to be another risk factor for PEAO (p=0.058; RR, 0.67; 95% CI, 0.45-0.99). Dexamethasone dose (0.25 vs 0.5 mg/Kg/dose) or whether patients received three or more doses prior to extubation did not affect occurrence of PEAO. Model for multivariate analysis included higher PRISM III score, longer (>7 days) intubation and duration of assisted ventilation. The latter two were proven to be independent risk factors for development of PEAO in the studied cohort (Table 3). Also, children intubated for more than 7 days had twice the chance of getting reintubated for PEAO compared to those who had shorter intubation [>7 days, 11/91 (12%) vs< 7 days, 8/147 (5.4%); RR, 2.22; 95%CI, 0.93-5.31; p-value, 0.07].
One patient with expanded dengue syndrome (with thrombocytopenia and coagulopathy) succumbed to a bout of massive hematemesis after 3 days of the last dexamethasone dose. None developed hypertension, hyperglycemia, signs infection or any other event attributable to dexamethasone during 5 days of follow up.