Issues of uncertainty and variability are of significant importance to risk assessment. This paper describes briefly the past and current practices devised to deal with those issues when making predictions with PBPK models, the strengths and limitations of those practices; the currently considered best pratices - and under what conditions they should be used - and finally what is needed (e.g., research, data, software, reporting/documentation standards, etc.) to further improve and expand the use of such models in the future.