Baseline predictors of placebo response in exercise-induced bronchoconstriction (EIB): pooled regression analysis...

Exercise-induced bronchoconstriction (EIB) can be variable in its presentation and severity. Evaluating patterns of placebo response and patient-related factors driving placebo response could facilitate more efficient clinical trials for EIB.

Data were pooled from three randomized, double-blind, crossover trials evaluating single-dose montelukast 10 mg or placebo in patients (N = 160) 15-45 years of age with EIB, defined as maximum % fall in forced expiratory volume in one second (FEV₁) ≥20% after two screening exercise challenges. Serial exercise challenges were performed at 2, 8.5-12, and 24 h postdose. The authors evaluated the distribution and variability of placebo response. They also evaluated possible drivers of response, analyzing all baseline patient demographic and prerandomization screening visit pulmonary function data as single covariates in a simple univariate regression model for maximum % fall in FEV₁ while on placebo at 2 h postdose. All covariates with p values <.1 were entered into both stepwise forward and backward regression procedures to select the "best" model.

Placebo response was variable, and showed a significant non-normal distribution (p < .001). Significant predictors of a greater response to placebo included: higher screening FEV₁ % predicted (p <.001), smaller maximum % fall in FEV₁ in screening (p < .001), shorter time to recovery in screening (p = .007), more asthma-related health care visits in the previous year (p = .004), older age (p = .001), less frequent asthma awakenings in the previous month (p = .003), and less frequent asthma symptoms in the past year (p = .011).

Predictors of a larger placebo response were generally markers of less severe asthma and/or EIB. This may be related to EIB variability, spontaneous improvement, or the extent of placebo response relative to the outcomes in less severe patients.