Model based extrapolation of efficacy to support Baloxavir Marboxil for uncomplicated influenza in children aged <1 year
Sylvie Retout1, Sébastien Jolivet1, Yumi Cleary1, Marie Laure Delporte1
Roche Pharma Research and Early Development, Pharmaceutical Sciences, Roche Innovation Center Basel
Objectives Baloxavir marboxil (Bxm) is an antiviral prodrug that is converted by hydrolysis to the active form baloxavir acid, a cap-dependent endonuclease inhibitor which prevents transcription of influenza virus genomes. Bxm is approved for the treatment and post-exposure prophylaxis of uncomplicated influenza in patients = 1 year in the EU. A phase 3 single arm open-label study was recently completed in 49 non-Asian children aged <1 year (miniSTONE-1; NCT03653364) who received a single oral dose of Bxm based on age (<3 months, 1 mg/kg; =3 months, 2 mg/kg). Pharmacometric analyses were conducted to extrapolate efficacy from adults and to support label extension in children aged <1 year. Methods A population PK (popPK) model and a popPK-Efficacy (time-to-alleviation of symptoms, TTAS) model, which were initially developed to support the development of baloxavir acid in the = 1 year population, were considered for the model-based analyses. First, plasma concentrations of baloxavir acid collected in miniSTONE-1 were pooled with PK concentrations data from 7 Bxm studies performed in pediatric, adolescent and adult patients. The previously developed empirical popPK model, which had identified body weight and race (Asian vs non-Asian) as the main covariates influencing baloxavir acid PK parameters, was updated including a clearance (CL) maturation function in order to describe the variation of the apparent CL with age in pediatric patients <1 year of age. Baloxavir acid concentration-time profiles were simulated in pediatrics aged <1 year with two dosing regimens: (i) Bxm doses similar to the ones administered in miniSTONE-1 or (ii) a 2 mg/kg dose in all patients aged <1 year. Simulated PK exposures were compared between adults and pediatrics <1 year old for both Bxm doses. Second, the performance of the PK-TTAS model, in predicting TTAS in patients aged < 1 year was assessed using Kaplan-Meier visual predictive checks. The model was then used to simulate PK-TTAS profiles in pediatric patients <1 year for the two dosing regimens. Simulations were performed per race (Asian/non-Asian) and age category: <28 days, =28 days to <3 months, =3 months to <1 year old. Results The final popPK model was developed from 1884 patient data for a total of 6666 PK concentrations, including 150 samples from 57 patients <1 year of age. It included a sigmoid Emax model to quantify the clearance maturation with age in infants. The model adequately described the concentration-time profiles of baloxavir acid, including patients <1 year of age, and adults, regardless of race. For both Asian and non-Asian patients, the model-based simulations predicted about 2-fold lower baloxavir acid exposure in <3 months old treated with 1 mg/kg than in 3 months to <1 year old pediatrics treated with 2 mg/kg. The model also demonstrated that similar exposures across all <1 year age groups are achieved with a 2 mg/kg dose. Overall, the exposures anticipated in the population <1 year with the 2 dosing regimens were mostly within the target exposure range in adults although the 1 mg/kg dose provided simulated exposures generally in the lower end of the targeted adult exposure range. Consistent conclusions were obtained by a population physiologically-based pharmacokinetic (popPBPK) model taking into consideration enzyme ontogeny and age-dependent physiological data. The PK-TTAS model was qualified to predict the effect of baloxavir acid on TTAS in pediatrics <1 year. After Bxm administration at the dose of 2 mg/kg in patients aged 3 months to <1 year, the predicted reduction in median TTAS compared to placebo was of approximately 1 day in Asians and of about 1.5 days in the non-Asians, whilst a reduction of about ~0.75 days for Asians and ~1 day for non-Asians was predicted at the dose of 1 mg/kg in pediatric patients < 3 months. In the case of a 2 mg/kg dose to all patients <1 year, the median TTAS was consistently reduced by ~1 day in Asians and ~1.7 days in non-Asian patients over the 3 age categories. Conclusion A model-based extrapolation demonstrated that pediatric to adult exposure matching was achieved in infants. The empirical popPK approach was complemented with a popPBPK modeling, in line with ICH11E and ICH15M guidelines on pediatric drug development. Altogether, this supported the Bxm label request in infants.