Large bone defects remain a clinical challenge because they do not heal spontaneously. 3-D printed scaffolds are a promising treatment option for such critical defects. Recent scaffold design strategies have made use of computer modelling techniques to optimize scaffold design. In particular, scaffold geometries have been optimized to avoid mechanical failure and recently also to provide a distinct mechanical stimulation to cells within the scaffold pores. This way, mechanical strain levels are optimized to favour the bone tissue formation. However, bone regeneration is a highly dynamic process where the mechanical conditions immediately after surgery might not ensure optimal regeneration throughout healing. Here, we investigated in silico whether scaffolds presenting optimal mechanical conditions for bone regeneration immediately after surgery also present an optimal design for the full regeneration process. A computer framework, combining an automatic parametric scaffold design generation with a mechano-biological bone regeneration model, was developed to predict the level of regenerated bone volume for a large range of scaffold designs and to compare it with the scaffold pore volume fraction under favourable mechanical stimuli immediately after surgery. We found that many scaffold designs could be considered as highly beneficial for bone healing immediately after surgery; however, most of them did not show optimal bone formation in later regenerative phases. This study allowed to gain a more thorough understanding of the effect of scaffold geometry changes on bone regeneration and how to maximize regenerated bone volume in the long term.