OBJECTIVE To establish a treatment algorithm for reconstructing complex midfacial defects. METHODS Retrospective case series. METHODS University-based teaching hospital. METHODS Thirty-one consecutive patients were treated from 1991 through 1995. The 18 males and 13 females were aged 15 to 90 years (mean age, 58 years). The cause of the defect included neoplasm (n = 27) and trauma (n = 4). Reconstruction consisted of 1 of 4 free flaps: rectus abdominis, radial forearm, fibula, or latissimus dorsi. Aesthetic and functional results were determined by patient questionnaires and physical examinations. METHODS Length of stay, postoperative morbidity and mortality, degree of aesthetic and functional restoration, and detection of tumor recurrence. RESULTS Twenty-seven (87%) of the 31 patients underwent reconstruction with a single major procedure. All of the flaps survived. Postoperative hospital stays averaged 14 days. Late tumor recurrence occurred in 7 (23%) of the 31 patients and was promptly detected. Aesthetic and functional results were rated good or excellent in 77% (24/31) and 87% (27/31) of patients, respectively. Of the 20 patients who underwent alveolar ridge resection, 16 (80%) received dental rehabilitation, 44% of whom received osseointegrated implants into either a bone flap or remaining native bone. Osseointegrated implants were inset during the initial reconstruction 57% (4/7 patients) of the time. CONCLUSIONS For complex midfacial defects, free-flap transfer can be performed with a high degree of success, restoring both appearance and function in most patients. The only instance in which bone is necessary to reconstruct the midface involves those areas in which osseointegrated implants are needed, ie, alveolar ridge (dental implant) and/or orbit (ocular prosthesis). In such cases, the fibula osteocutaneous free flap is the flap of choice. Otherwise, soft-tissue flaps are selected based on wound size.