Guided bone regeneration (GBR) is employed to encourage the formation of new bone in osseous defects by restricting the infiltration of soft tissues. While a variety of membranes have been evaluated for this surgical procedure, the non-resorbable material of choice is currently expanded polytetrafluoroethylene (e-PTFE). A new alkali-cellulose membrane produced by a biotechnological process has been developed as an alternative to e-PTFE for GBR. In this study, the biocompatibility of this novel alkali-cellulose membrane and e-PTFE was compared using tissue culture and an in vivo GBR model. In vitro both materials supported the attachment, migration and differentiation of osteoblast-like cells in culture for up to 3 weeks. The in vivo model was based upon full-thickness transcortical bone defects in the mandibular rami of Sprague-Dawley rats. The right rami were used as controls, contralateral defects being covered bucally and lingually with either e-PTFE or alkali-cellulose membranes. Pathological and histomorphometric analysis was undertaken at 4 and 10 weeks post-implantation. Bone regeneration associated with alkali-cellulose membranes was predominantly endochondral in type in contrast to e-PTFE which induced direct bone formation (intramembranous ossification). The amount of new bone formed in defects was similar for both types of membrane, but alkali-cellulose membranes induced significantly greater inflammatory response; characterized by lymphocytes, macrophages and multinucleated giant cells. Degradation and possible exposure of individual cellulose fibres may account for the poor performance of alkali-cellulose membranes in vivo. This animal and in vitro study indicates that when choosing a non-resorbable membrane for GBR, e-PTFE membranes are likely to perform better than those produced from alkali-cellulose.