Mechanistic model based meta-analysis for paediatric respiratory tract infection prophylaxis trial design

Prophylaxis of RTI in children could have a large impact on hospitalizations, antibiotics overuse or asthma inception but options for vaccination are still limited. The bacterial lysate OM-85 (Broncho-Vaxom®) can significantly prevent RTIs in the paediatric population with recurrent RTIs as meta-analyzed previously. Nevertheless, patient and trial heterogeneity are key limitations to designing clinical trials that can confirm the effect for a reasonably large population of patients. We have therefore set up a modelling approach leveraging existing data on OM-85. We use a mechanistic administration and pharmacodynamics response model to rationalize the variability and age-dependency of immunomodulation and an epidemiological-immunological viral infection model to assess the impact of this modulation on RTIs and wheezing in individual patients.  With in silico clinical trials and application of virtual population eligibility criteria on the back of this model, we explore the efficacy of OM-85 as a function of trial design variables and population. Longer follow up times can capture more prevented RTIs or wheezing episodes (larger mean effect) but (in particular wheezing endpoints) suffer from increased inter-patient variability. Patient age and at-risk population have an impact on the mean effect and its variability but their contribution needs to be confronted with other non-identified factors in inter-trial variability. Our in silico analyses point also to optimization potential with respect to the dosing-regimen as longer exposure seems to allow for an increase of the effect size without deteriorating the response heterogeneity.