Are pharmaco-statistical methods flawed in the context of between systems extrapolation?

Piet recently published a blog on CPT-PSP where he asks an important question - Extrapolation in Pediatric Medicine Development: Time for a SYSTEMatic Approach?
Obviously, it is an attempt to expand the use of more scientific/pharmacological approaches for extrapolation. But, I was wondering if there are a set of examples that can support the statement (below) in the blog that extrapolation between adults and pediatrics has failed. This is the statement made in the blog -

" Pharmaco-statistical approaches can work well for within-system extrapolation (i.e. predicting a higher-than-tested dose), but can be fundamentally flawed in the context of between-system extrapolation (i.e. predicting pediatric PKPD from adults)"