Treating Exacerbations of Asthma With Oral Montelukast in Children

Study design parameters Design: 2-arm, phase 3 double-blind RCT Intervention: Oral montelukast sprinkles or identical placebo Dose escalation: Doses between 2 mg/kg and 3 mg/kg based on pharmacokinetic modeling after each group of 5 participants. Montelukast is a potent LT-receptor antagonist and is FDA approved at a daily oral dose of 4-5 mg for chronic asthma and allergic rhinitis in children. It is also a potent bronchodilator. In randomized controlled trials (RCT) of adults with moderate and severe exacerbations and inadequate response to inhaled albuterol, intravenous (IV) montelukast caused rapid (within 10 minutes) and sustained improvement of lung function in all subjects, without selecting for CCS non-responsiveness. This underscores the potential clinical impact of reducing LT-mediated inflammation during exacerbations. However, the standard 5 mg oral dose does not achieve peak plasma levels (Cmax) comparable to the IV doses used in these RCTs, and IV montelukast is not available. The investigators' preliminary pharmacokinetic (PK) study in children with acute exacerbations demonstrated that 30 mg oral montelukast (mean 1.0 mg/kg) achieves Cmax \>1,700 ng/ml in 40% of participants. This is a Cmax expected after doses used in the adult RCTs above. The primary hypothesis for this research is that the investigators will identify a weight-based oral montelukast dose that reliably achieves Cmax \>1,700 ng/ml by examining 3 escalating dose levels between 2 and 3 mg/kg. Dose escalation will be optimized using computerized, PK-guided dose modeling that will be updated after each group of 5 participants among 45 randomized to receive montelukast. The investigators further hypothesize that high-dose montelukast meaningfully decreases exacerbation severity, measured using the validated Acute Asthma Intensity Research Score (AAIRS). Participants will be randomized to receive montelukast powder USP in Ora-Blend suspending agent or identical placebo. Montelukast (MK) dose will begin at 2 mg/kg. Peak MK plasma concentration (Cmax) will be measured after each group of 5 participants randomized to MK. MK dose for subsequent participant groups will be escalated if Cmax is \<1,700 ng/ml for \>14% of participants at current mg/kg dose level. INCLUSION CRITERIA * Child aged 4 - 12 years with doctor-diagnosed asthma * Presents to the Vanderbilt Children's Hospital (VCH) with an acute asthma exacerbation that is moderate or severe (AAIRS \>= 7) after initial treatment with inhaled albuterol o Per emergency department (ED) protocol, patients with AAIRS \>= 7 have an IV catheter placed, enabling blood draws. * The parent agrees to phone and/or mail follow-up at 2-3 weeks for completion of Screen for Anxiety Related Disorders (SCARED) and side-effect questionnaires. EXCLUSION CRITERIA Any of the following preclude enrollment of a child for this protocol: * Gestational age \< 34 weeks * acute or chronic liver disease * allergy to montelukast * female with any evidence of Tanner stage 2 or greater breast development * gastroesophageal reflux requiring acid-blocking medication * history of anxiety disorder, depression and/or other neuropsychiatric disorder except attention-deficit/hyperactivity disorder (ADHD) * positive on question 1 or 2 of the Columbia Suicide Severity Rating Scale (CSSR-S) * score \>25 on the 82-point Screen for Child Anxiety Related Disorders (SCARED) questionnaire * Patients currently receiving daily montelukast (4 or 5 mg) will not be excluded from study participation. Blood sampling, scoring the Acute Asthma Intensity Research Score(AAIRS), and measurement of airway resistance by impulse oscillometry (IOS, below) will be performed before (time 0) and at approximately 2, 3, 4, 6, 12-18, and 18- 24 hours after drug administration, if the participant is available. Reasons for non-availability include but are not limited to discharge from ED or hospital and participant sleeping, performing another procedure, eating, or being rounded upon by the clinical team. If a participant is to be discharged from the hospital before the final time point, the investigators will attempt to perform the above studies immediately prior to discharge. With the variability of plasma levels at each mg/kg dose in mind, the investigators will use plasma level distribution as the primary statistical measure for this aim. Plasma montelukast assays will be performed after each Group of 5 participants are randomized to receive MK, with a Futility Criterion determining whether subsequent Groups at a Dose Level will be enrolled at that Dose Level or moved to the next one. Interim analyses will be conducted after each Dose Level has been studied. Testing Hypothesis 1: The investigators are primarily interested in estimating the change in plasma concentrations of montelukast over time and will report summary statistics at each time point numerically and graphically. They will also summarize the longitudinal measurements by the Cmax, the proportion of subjects achieving a Cmax of at least 1,700 ng/ml, the time (Tmax) to reaching Cmax, and the time to reaching 1,700 ng/ml (T1700). While the investigators anticipate being able to monitor all subjects for 24 hours, the primary analysis will consider subjects who have at least 4 hours of blood draws to allow minimally sufficient time to achieve Cmax. Power calculation for Hypothesis 1: This research is powered for to test hypothesis 1 and may not be sufficiently powered to test hypotheses 2 and 3. PK studies of oral montelukast at daily doses (4 - 5 mg) used for control of childhood asthma and allergic rhinitis included 12-32 participants. Investigators established sample size for these studies without formal power and sample-size calculations. With these considerations in mind, and in collaboration with the FDA, the sample size for this research has been set at 45 for each arm of this RCT. For this power calculation the investigators consider the probability of successfully finding a dose at which at least 86% of Dose level participants achieve a Cmax \>1,700 ng/ml. The first consideration is the scenario in which at least 13 of the 15 (86.7%) group participants successfully achieves a Cmax \>= 1,700 ng/ml. The probability of successfully identifying this dose will depend on p, the true probability a dose will lead to a Cmax above 1,700 ng/ml. If p=0.90, there is 82% power that the investigators will observe 13 or more successes in 15 subjects; if p=0.95, there is a 96% power that the investigators will observe 13 or more successes. These calculations are based on a one-tailed cumulative binomial distribution function with N=15, but there is the possibility of carryover leading to a Dose Level N of 20, 25, or 30. For N=25 and p=.9, there is a 76% power that the investigators will observe 22 or more successes; if p=0.95, there is 97% probability of observing 22 or more successes. Second, consider the scenario of an inadequate dose that needs to be escalated. In this scenario, the investigators would escalate to the next dose when more than 2 subjects have a Cmax below 1,700 ng/ml after enrolling 5, 10, or 15 subjects. If a given dose is only 50% effective (p=0.5) in achieving a Cmax above 1,700 ng/ml, then there is a 50%, 94%, or 99.6% probability of escalating after 5, 10, or 15 subjects, respectively. Testing Hypothesis 2: The Acute Asthma Intensity Research Score (AAIRS) is a 0 - 16 point (16 most severe) bedside severity score that incorporates multiple domains of exacerbation severity and predicts %-predicted FEV1 across all levels of severity. Additionally, an AAIRS \<3 predicts safe discharge from the ED. The AAIRS is an optimal surrogate of need-for hospitalization and an outcome measure for this research that the investigators can accurately measure before and at precise time points after active drug or placebo dosing. Descriptive analysis of key demographic and clinical variables by randomization group will be performed to evaluate randomization. The primary analysis will be linear regression of change of the AAIRS severity score 4-hours post dosing (outcome) with an indicator for treatment group as the predictor of interest and controlling for baseline (pre-treatment) AAIRS and pre-treatment plasma leukotriene E4 concentration (a measure of systemic leukotriene stress). This model will allow the investigators to estimate the mean difference in the change in AAIRS score, treatment versus control, with corresponding 95% confidence intervals. Controlling for baseline severity using the pre-treatment AAIRS will improve the precision of the estimated treatment effect. Secondary analyses will replace the binary indicator of treatment group with the dose of montelukast received. If the investigators find that the assumptions of the linear model are not appropriate, they will use the proportional odds ordinal regression model with the same outcome and predictors. The proportional odds model generalizes the Wilcoxon rank sum test to a regression setting to estimate an odds ratio corresponding to a treatment effect. Power and sample size considerations are based on a fixed sample size of 45 treatment and 45 control subjects. In prior studies, the investigators found that the standard deviation of the change in AAIRS score after 4 hours of treatment was 2.97 points. If the true difference in the change in AAIRS score between treatment and control groups is 2 points, there will be 89% power to detect a difference between groups. The test will use a two-sided significance level of 0.05. Testing Hypothesis 3: Initial descriptive analysis will include summary statistics and bivariate correlations of 4-hour change of the AAIRS as outcome with Cmax, pre-treatment exacerbation severity, and urinary LTE4 concentration. Bivariate associations will be flexibly fit using loess smooth lines and Spearman's rank correlation. Hypothesis 3 will be tested in a multivariable linear regression model using change of the AAIRS severity score as the outcome with Cmax, pre-treatment AAIRS, and pre-treatment plasma leukotriene 4 concentration as predictors. Regression splines will be included as needed to allow for non-linear associations between predictors and the outcome as needed. As with hypothesis 2, if it is found that the assumptions of the linear model are not appropriate, the investigators will use the proportional odds ordinal regression model with the same outcome and predictors. Power and sample size considerations are based on the 45 montelukast recipients. For bivariate correlations, there will be 80% power to detect correlations of 0.40 or greater (R-square of 0.16). For the multivariable model calculation, if it is assumed that exacerbation severity and leukotriene stress explain 20% of the variability in AAIRS severity score, then there will be 80% power to detect if Cmax explains an additional 12% of the variability in AAIRS severity score (improvement in R-square of 0.12). If exacerbation severity and leukotriene stress only explain 10% of the variability in AAIRS severity score, then there will be 80% power to detect if Cmax explains an additional 14% of the variability in AAIRS severity score (improvement in R2 of 0.14). Calculations are based on a two-sided significance level of 0.05.