To probe the etiopathogenisis of adjacent-segment disease by analyzing the imageology data and cl inical neurological function in patients with anterior cervical discectomy and fusion (ACDF) harvested by long-term follow-up. Methods A retrospective study was performed on 52 patients who had undergone ACDF with perfect documents from January 1990 to April 2003. Of the patients, 45 were males and 7 were females with a mean age of 48.5 years (range from 25 to 72 years). There was the fusion of 10 one-levels, 38 two-levels and 4 three-levels. The cervical anterior-posterior and lateral X-ray, CT and MRI examination were performed before the operation. Cl inical neurological function was recorded by the Nurick score, and this score at 6 weeks after the operation was compared with the later follow-up. In the radiological examination, the motion of adjacent vertebrae and osteophyte formation were reviewed on X-ray and CT, and were converted to the semi-quantitative degeneration score according to the Goffin method. The correlation between Nurick score or degeneration score and the age at operation or fusion levels was compared by Spearman correlation coefficients. The cervical canal sizes of adjacent level and remote level on MRI were reviewed and compared with each other by t test. Results The follow-up period was 3 to 10 years, 6.9 years on average. There was difference in the Nurick score between the 6th week after operation (1.07 ± 0.84) and the later follow up (1.92 ± 1.28) by rank test (P lt; 0.05). There was no correlation between the Nurick score change and the age at operation (r = 0.21, P gt; 0.05) or fused levels(r = 0.30, P gt; 0.05) by Spearman correlation coefficients. There was obvious difference in degeneration score between the 6th week after operation (0.73 ± 0.67) and the later follow up (1.58 ± 1.06), (P lt; 0.01). There was no correlation between the degeneration score change and the age at operation (r = 0.35, P gt; 0.05) or fusion levels (r = 0.38, P gt; 0.05) by Spearman correlation coefficients. The cervical canal size reductions were (1.7 ± 1.1) mm at superioradjacent level, (1.2 ± 0.6) mm at inferior adjacent level and (0.30 ± 0.68) mm at remote level. There was obvious difference between superior or inferior and remote level by t test (P lt; 0.01). The adjacent level developed prominent degeneration together with nerve function change after the fusion operation and displayed correlation between degeneration and nerve function change(r = 0.41, P lt; 0.05). Conclusion The adjacent-segment disease after interbody fusion is produced by multiple factors. The natural progression in adjacent disc, biomechanical natural change resulting from interbody fusion, destruction to l igament structure in front of cervical vertebrae by operation, and bone graft model are important factors not to be ignored.