Rogers described his technique of
spinal fusion in 1942, and since then numerous other techniques have been described but no large series describing the anatomical results has been reported. To assess the technical success of Rogers' technique, to identify factors that contribute to less than ideal anatomical results, and to suggest methods of avoiding potential pitfalls, the anatomical results of Rogers' posterior cervical fusion were compared with what we consider an ideal anatomical result by analysis of the 12-week post-operative flexion/extension radiographs. One hundred and sixty-one Rogers-type posterior cervical fusions using either wire or
Ethibond were performed for flexion
injuries. The 12-week post-operative flexion extension radiographs were assessed for union, fusion of extra levels, residual
kyphosis/listhesis, excessive
lordosis, and hypermobility. Results were related to the presence of associated fractures, using the chi 2 test. Bony union was seen in 100 per cent of cases. Fusion of additional levels occurred in 40 (25 per cent), residual
kyphosis in 54 (34 per cent), listhesis in 14 (9 per cent), and excessive
lordosis in seven (4 per cent). Hypermobility at the adjacent level occurred in 10 (6 per cent), and at a distant level in five (3 per cent). Statistically significant associations occurred between fusion of extra levels and fractures, residual
kyphosis and fractures, excessive
lordosis with the use of wire rather than
Ethibond, and the desired anatomical result with absence of fracture. The Rogers technique is a safe, easy and reliable method of achieving cervical fusion, with a 100 per cent fusion rate at 3 months in this series. However, the intended position of fusion, between 1 degree-5 degrees of
lordosis, with normal alignment, is not always achieved. There is also a high incidence of fusion of levels other than those intended. We believe that the incidence of these problems could be reduced by more attention to surgical detail.