Reconstruction of cartilaginous defects to correct both craniofacial deformities and joint surface irregularities remains a challenging and controversial clinical problem. It has been shown that tissue-engineered cartilage can be produced in a nude mouse model. Before tissue-engineered cartilage is used clinically to fill in joint defects or to reconstruct auricular or nasal cartilaginous defects, it is important to determine whether it will integrate with or adhere to the adjacent native cartilage at the recipient site. The purpose of this study was to determine whether tissue-engineered cartilage would adhere to adjacent cartilage in vivo. Tissue-engineered cartilage was produced using a fibrin glue polymer (80 mg/cc purified porcine fibrinogen polymerized with 50 U/cc bovine thrombin) mixed with fresh swine articular chondrocytes. The polymer/chondrocyte mixture was sandwiched between two 6-mm-diameter discs of fresh articular cartilage. These constructs were surgically inserted into a subcutaneous pocket on the backs of nude mice (n = 15). The constructs were harvested 6 weeks later and assessed histologically, biomechanically, and by electron microscopy. Control samples consisted of cartilage discs held together by fibrin glue alone (no chondrocytes) (n = 10). Histologic evaluation of the experimental constructs revealed a layer of neocartilage between the two native cartilage discs. The neocartilage appeared to fill all irregularities along the surface of the cartilage discs. Safranin-O and toluidine blue staining indicated the presence of glycosaminoglycans and collagen, respectively. Control samples showed no evidence of neocartilage formation. Electron microscopy of the neocartilage revealed the formation of collagen fibers similar in appearance to the normal cartilage matrix in the adjacent native cartilage discs. The interface between the neocartilage and the native cartilage demonstrated neocartilage matrix directly adjacent to the normal cartilage matrix without any gaps or intervening capsule. The mechanical properties of the experimental constructs, as calculated from stress-strain curves, differed significantly from those of the control samples. The mean modulus for the experimental group was 0.74 +/- 0.22 MPa, which was 3.5 times greater than that of the control group (p < 0.0002). The mean tensile strength of the experimental group was 0.064 +/- 0.024 MPa, which was 62.6 times greater than that of the control group (p < 0.0002). The mean failure strain of the experimental group was 0.16 +/- 0.061 percent, which was 4.3 times greater than that of the control group (p < 0.0002). Finally, the mean fracture energy of the experimental group was 0.00049 +/- 0.00032 J, which was 15.6 times greater than that of the control group. Failure occurred in all cases at the interface between neocartilage and native cartilage. This study demonstrated that tissue-engineered cartilage produced using a fibrin-based polymer does adhere to adjacent native cartilage and can be used to join two separate pieces of cartilage in the nude mouse model. Cartilage pieces joined in this way can withstand forces significantly greater than those tolerated by cartilage samplesjoined only by fibrin glue.