The purpose of this investigation was to establish whether the tissue response and migration of polyethylene debris differed at noncemented smooth and porous implant surfaces. This was accomplished through 3 separate but closely related studies: (1) a canine cylindrical implant model with smooth and porous surfaces exposed to polyethylene debris; (2) a canine total hip arthroplasty model analyzing the interface between bone and femoral implants with various porous-coating configurations; and (3) a histologic analysis of autopsy-retrieved, human, noncemented hip prostheses with noncircumferential porous coating. The cylindrical implant model involved the placement of split cylinders, 1/2 porous and 1/2 smooth, into the distal femur and proximal tibia of 4 dogs. Four control implants and 10 test implants (chronically exposed to simulated polyethylene debris with a mean size of 4.7 microns) were examined histologically as long as 30 weeks after surgery. The canine hip study involved the study of 54 noncemented hip prostheses at periods of 1, 6, and 24 months. The prostheses possessed 4 different porous surface configurations: 1 with circumferential porous coating, 2 with noncircumferential coating, and 1 without porous coating. The human retrieval analysis involved the study of 7 cadaveric femora (age, 6 months-5 years) implanted with a straight titanium-alloy prosthesis possessing proximal pads of titanium fiber metal on the anterior, posterior, and medial aspects. With all implants in all 3 studies, there was the common finding of bone ingrowth at the porous implant surface and a fibrous interface or periprosthetic cavity around the portion of the implant that was smooth surfaced. The periprosthetic cavity typically was encapsulated by a thin continuous shell of trabecular bone. In addition, polyethylene debris was found to have preferentially migrated along the smooth implant surfaces. In the longer-term canine and human hip retrievals, polyethylene particles in the micron size range were present within histiocytes, whereas larger particles as much as 100 microns were found within foreign-body giant cells. Of importance for the implants from all 3 studies, with the exception of some pronounced cavities on the lateral aspect of the human hip prostheses, the periimplant cavities around the smooth surfaces were not detectable radiographically. This study clearly established a fundamental principle of relative barriers to particulate debris migration. Smooth implant surfaces are more susceptible than porous surfaces to the development of a fibrous tissue filled periimplant cavity and the subsequent migration of polyethylene wear debris.(ABSTRACT TRUNCATED AT 400 WORDS)