In the management of cervical spine injuries, it is not always clear when to use halo immobilization alone, surgical fusion alone, or a combination of the two. To investigate the relative effectiveness of each of these approaches, we reviewed the medical records of 140 patients with cervical spine injuries treated with either halo immobilization or surgical fusion, or both. Seventy (50%) of the patients were neurologically intact on admission (two of these were paraplegic from previous injuries). Halo immobilization was used as the primary treatment in 99 patients, and yielded a successful fusion rate of 78%. Within this group, the 26 patients with hyperflexion-anterior subluxation injuries had only a 54% successful fusion rate, while the rate for the 73 with non-flexion injuries was 87% (Chi-square = 11.36; p = 0.0008). Surgical fusion was used as the primary treatment in 41 patients and as a subsequent treatment in the 22 for whom halo immobilization did not bring about fusion. Of these 63 patients treated with surgical fusion, six remained unstable after the surgery; five of these six had sustained a hyperflexion-anterior subluxation. One patient experienced neurologic deterioration after surgical fusion. There were three deaths in the entire series. Excluding fusion failure, complications with halo immobilization were frequent (25%) but usually minor; with surgical fusion, less frequent (6%) but usually more severe. We draw the following conclusions. 1) Halo immobilization brings about satisfactory healing for most fracture types. 2) Both halo immobilization and surgical fusion have relatively high failure rates in the treatment of hyperflexion-anterior subluxation injury, with or without bilaterally locked facets.(ABSTRACT TRUNCATED AT 250 WORDS)