This retrospective study evaluated a single surgeon's series of patients treated by multilevel cervical disc excision (two or three levels), allograft tricortical iliac crest arthrodesis, and anterior instrumentation. The objective of this retrospective study was to compare fusion success and clinical outcome between multilevel Smith-Robinson interbody grafting and tricortical iliac strut graft reconstruction, both supplemented with anterior instrumentation in the cervical spine. The incidence of nonunion for cervical discectomy and fusion varies widely depending on the number of disc levels involved, type of bone graft used, and whether the anterior grafting is supplemented with instrumentation. An alternative to multilevel interbody fusion is corpectomy and strut grafting, in which the incidence of nonunion has been reported to be 27% with autograft and 41% with allograft. Sixty-four consecutive patients who underwent allograft tricortical iliac crest reconstruction and anterior cervical plating were studied. The average follow-up was 39 months. There were 38 patients in the discectomy and interbody grafting group and 26 patients in the corpectomy and strut graft reconstruction group. Pseudoarthrosis occurred in 42% of the anterior cervical interbody fusion patients and 31% of the corpectomy patients. Nonunion in two-level interbody fusions occurred in 36% of the patients as compared to 10% for patients with one-level corpectomies; while 54% of patients with three-level interbody fusions and 44% of patients with two-level corpectomies were noted to have pseudoarthrosis. Higher percentages of nonunion were noted in multilevel interbody grafting than in corpectomy with strut grafting and when more vertebral levels were involved. These radiographic and clinical findings underscore the shortcomings of multilevel anterior cervical allograft reconstruction with plating. Corpectomy may be the preferred method when multiple disc levels are fused. In addition, anterior corpectomy affords decompression of significant osteophytes in a safer and quicker manner. In retrospective studies, there is a need for long-term follow-up before accurate statements can be made about the study population.