Percutaneous nucleotomy is a relatively new technique for treating lumbar disc herniation. There is no agreement as to the volume of disc material to be removed. A long-term study of clinical and radiological data from patients treated by percutaneous nucleotomy was designed to identify the factors associated with favourable and unfavourable outcomes. We studied 42 patients for at least ten years; the mean follow-up was 10.9 years. They were divided into two subgroups to assess the value of preserving the nucleus pulposus in the central area of the disc. The overall success rate for both subgroups was 50%. A decrease in disc height on plain radiography and a decrease in signal intensity on MRI were observed more infrequently in patients in whom the nucleus pulposus in the central area of the disc had been preserved, than in those in whom it had been extensively removed. These adverse radiological findings correlated closely with increased low back pain during the first one to two years after operation and a poorer overall outcome. We conclude that percutaneous nucleotomy is most likely to be successful when the central area of the disc is preserved.