Discussion
Multi-dose pretreatment regime of systemic steroid starting 12-24 hours prior to planned extubation reduces PEAO in adult 6-10as well as in pediatric 1,9 patients. Dexamethasone is preferred over other steroids 1,7,8. Our previous study (period: 2009-10) demonstrated that 24-hour pretreatment with a high dose of dexamethasone prevents PEAO and need of reintubation1. Majority of the then patient population in our unit had multiple risk factors e.g., intubation in uncontrolled situations, longer duration of intubation and hand ventilation in emergency room, multiple airway manipulations and hypoalbuminemic edema. Inspired by the findings of non-superiority of the high dose (versus the low dose) in the adult patients 8 and lack of similar information in the pediatric literature, we conducted the present trial (period: 2017-19). It has not only shown the non-inferiority of the low dose (0.25mg/Kg/dose) pretreatment (versus previously studied high dose, 0.5mg/Kg/dose) in reducing PEAO, but it also demonstrated the changing clinical scenario in the unit during the last decade.
Comparison of clinical characteristics and potential risk factors for PEAO during these two study periods is given in the supplemental table 1. While our patients continue to be similarly malnourished, PRISM III score has increased. Now, we saw more patients with presumptive diagnosis of CNS infections (e.g., meningitis, meningoencephalitis), while patients with respiratory infections (e.g., pneumonia, empyema) has decreased. Frequency of risk factors, as identified in the previous trial (2009-10), were much lesser during the current study period (2017-19). Number of patients intubated in uncontrolled situations (89% vs 76%) got reduced as was duration of intubation and/or hand-ventilation in pre-PICU areas (PER/wards) (median, 0.9 days vs 12 hours). Number of patients needed to stay in PER/wards for> 24 hours prior to PICU transfer (57% vs 27%) and undergoing multiple airway manipulations (35% vs 18%) got reduced by half. Number of patients with >7 days of intubation and/or ventilation also got reduced by a third (56% vs 38%). Serum albumin was higher (median, 2.5 gm/dL vs 2.8 gm/dL) at the time of extubation, while lesser number of patients had hypoalbuminemia (88% vs 75%) and clinical edema (40% vs 27%). All these led to a notable reduction in PEAO (73% vs 33%) during the current study period despite using the same yardstick (i.e., WCS). Reintubation rate remained unchanged over last 10 years (11% vs 8%). Incidence of PEAO and reintubation in our unit seems to be approaching towards what is being reported from PICUs of high-income countries 2,5,14. The two study cohorts, a decade apart, provide a glimpse of journey of a PICU in a tertiary care teaching hospital in a lower-middle-income country.
We could not find any new pediatric trial since we published results of our previous study. Recently, effectiveness of dexamethasone pretreatment in reducing reintubation due to PEAO was reported in a retrospective propensity matched study on infants with respiratory syncytial virus 5. A systematic review and meta-analysis of 11 adult trials 10 proved efficacy of steroid pretreatment in prevention of PEAO and reintubation in select group of at-risk patients. A randomized control trial among at-risk adult patients (cuff-leak volume <110ml) revealed effectiveness of dexamethasone pretreatment in reducing PEAO (placebo group, 30% vs two dexamethasone groups, 9.8% and 7.1% respectively). However, there was no difference between the low dose (5 mg/dose) and the high dose (10 mg/dose) regimes 8 similar to the findings observed in the current study. Low dose dexamethasone reduced the baseline incidence of PEAO from 30% (in the placebo arm) to 9.8%, and further reduction could not be achieved with the higher dose. Similarly, lack of difference in the two dose regimes may be attributable to the significantly reduced incidence of PEAO itself (73% vs 33%), with notable weakening of the associated high-risk factors. However, it is not sure if these findings would be replicable in a clinical setting similar to what we had a decade ago, i.e., very high incidence of PEAO with presence of multiple risk factors in the majority. The study reiterates the situation-specific applicability of the evidence generated from various clinical settings15.
Dexamethasone pretreatment is found to be effective only among at-risk adult patients, and not among unselected ones 10. We also demonstrated a lesser incidence of PEAO with high dose dexamethasone pretreatment in the subgroup of children intubated for more than 7 days [RR, 1.52; p-value, 0.07]. Trials to assess effectiveness of various dexamethasone pretreatment regimes in the select group of high-risk children are advisable in context of reducing incidence of PEAO and plateauing of reintubation in the studied clinical setting. Indices based on pre-extubation airway ultrasound16-19 may help identify high-risk children.
Based on the assumptions used for sample size calculation in ‘Material & Methods’ section, the power of the study is less than 60%. With demonstrated PEAO incidence of 32% and 34% in HD and LD arms respectively in our study, with non-inferiority margin of 12%, one sided α-error of 0.05 and power of 80%, a sample size of 398 per group (796 in total) would be required to prove the non-inferiority of LD compared to HD. With the new assumptions also, the power of study is less than 60%. The study was a non-funded time-bound dissertation project. Due to logistics and time-constraints, we could enroll only 238 patients. Two categories of exclusions may raise concerns of bias— patients with deferred extubation (n=21) and patients who were reintubated for non-PEAO reasons (n=28). Inclusion of the former would be unfair as they were not available for assessment of treatment effect. Results after including those reintubated for non-PEAO reasons (n=28) was no different. Relying on clinical scores to identify and monitor PEAO lacks objectivity; respiratory inductance plethysmography and esophageal manometry (which is not available with us) would have been better 14. Lack of measurement of observer agreement2 and blinding are other concerns. To limit the former, minimum of two clinicians assessed the patients for PEAO and/or need for reintubation. The latter could not have been addressed as dexamethasone doses were drawn from the same commercially available preparation, and we do not have in-house pharmacy. A placebo arm in the trial would have demonstrated the role of dexamethasone pretreatment better, however it would be unethical when our unit’s standard of care includes dexamethasone pretreatment for children intubated for> 48 hours. Since it is a single-center study, external validity may be a concern. Enrolment over a short period of 26 months indicates contemporary nature of the cohort. Stratification of randomization using intubation duration is a strength. Despite many limitations, the trial provides clinically relevant data to the pediatric critical care community working in low- and middle-income countries (LMICs) and experiencing high incidence of PEAO as ours.