Multilevel Fusion for Cervical Spondylosis: Is a Concomitant Posterior Approach Necessary?

Purpose: Retrospective clinical, radiographic, and outcome analysis comparing anterior-only (Ant) and anteroposterior (AP) fusion for multilevel cervical spondylosis, to determine the effects of a concomitant posterior approach in multilevel anterior fusion for cervical spondylosis in terms of operative parameters, fusion rates, reoperation rates, and functional outcomes. Multilevel cervical fusion reportedly leads to dismal fusion rates. AP fusion is considered a potential solution to this problem.

Methods: 78 patients (55 Ant, 23 AP) who had multilevel fusion for cervical spondylosis between 1998 and 2002 and with minimum 2-year follow-up were identified. Operative, inpatient, and outpatient data were reviewed. Preoperative, postoperative, and follow-up imaging studies were analyzed. Between 2 and 6 years post-op, patients were asked to answer functional outcome questionnaires (SF-36, Neck Disability Index [NDI], and Roland Morris).

Results: The Ant group incurred significantly less blood loss (165 vs 311 cc), shorter operative time (189 vs 380 min) and hospital stay (3 vs 4 days). In contrast, pseudarthrosis rates (38 percent vs 0 percent, p<0 .001) and pseudarthrosis-related reoperations (22 percent vs 0 percent, p equals 0.035) were both significantly reduced in the AP group.

Nevertheless, no significant difference overall reoperation rates was shown (Ant 36.4 30.4 percent). Both groups showed improved lordosis post-op (Anterior 8.6 vs anterior-posterior 9.1 degrees) with change over time (Anterior 0.3 vs anterior-posterior 1.6 degrees). Functional scores similar for using SF-36 NDI.

However, using the Roland Morris index, the Anterior group showed a greater degree of functional impairment (Anterior 8 vs anterior-posterior 1.5, p equals 0.035).

Conclusions: A concomitant posterior fusion significantly reduces the incidence of pseudarthrosis and pseudarthrosis-related reoperations compared to an anterior-only approach. However, no significant reduction in overall reoperation rates was demonstrated. Functional outcome at 2 to 6 years may be better using Roland Morris but is similar using the SF-36 and NDI.