To What Extent Do Different Criteria Influence 3-Month Fusion Evaluation in Anterior Cervical Arthrodesis Trials?

OBJECTIVES

Multiple imaging criteria are available for assessing fusion following anterior cervical discectomy and fusion (ACDF). In clinical trials, the 3-month postoperative follow-up serves as a critical timepoint for evaluating the efficacy of interventions on accelerating the fusion process. This study aims to determine how applying different fusion criteria influences the conclusions of a comparative analysis.

METHODS

Patients aged 18 or older who underwent ACDF with allograft or beta-tricalcium phosphate artificial bone between C3 and C7 were reviewed from 1 April 2023 to 30 September 2023. Fusion rates between the two grafts at three-month follow-up were compared under different criteria. Fusion status was judged by CT or dynamic radiographs, or their combinations. Cut-offs of dynamic indicators included angle changes of 4°, 3°, and 2°, and interspinous motion of 3, 2, and 1 mm. Criteria were applied singly, combined in pairs, or combined in groups of three, leading to a total of 31 criteria. Student's t-test and Chi-squared test were employed, and Cohen's kappa coefficient and phi coefficient were calculated.

RESULTS

Ninety-eight segments were included. Twenty-five criteria yielded higher fusion rates for artificial bone, with 7 out of 25 reaching statistical significance (p < 0.05). The remaining six criteria led to a reversed result, but none reached significance (p > 0.05). The agreement and correlation between CT and dynamic criteria were poor (kappa and phi < 0.200). In contrast, the agreement and correlation between two dynamic indicators were better, and even being close to moderate (kappa = 0.398, phi = 0.398) between 3° and 2 mm.

CONCLUSION

Changes in fusion criteria affected result significance but did not produce conflicting conclusions. There was a significant disagreement between the results under CT and dynamic radiographs criteria. Thresholds of 3° or 2 mm can be optimal choices for dynamic criteria.