Abstract Background Today there is some evidence-based medicine support for a positive short-term treatment effect of fusion in chronic low back pain in spondylolisthesis and in nonspecific degenerative lumbar spine disorders. The long-term effect is, however, unknown. Purpose To determine the long-term outcome of lumbar fusion in adult isthmic spondylolisthesis. Study design Prospective, randomized controlled study comparing a 1-year exercise program with instrumented and non-instrumented posterolateral fusion with average long-term follow-up of 9 years (range, 5–13). Patient sample 111 patients aged 18 to 55 years with adult lumbar isthmic spondylolisthesis at L5 or L4 level of all degrees, and at least 1-year's duration of severe lumbar pain with or without sciatica. Outcome measures Pain and functional disability was quantified by pain (VAS), the Disability Rating Index (DRI), the Oswestry Disability Index (ODI) work status, and global assessment of outcome by the patient into much better, better, unchanged or worse. Quality of life was assessed by the SF-36. Methods The patients were randomly allocated to treatment with 1) a 1-year exercise program (n=34), 2) posterolateral fusion without pedicle screw instrumentation (n=37), or 3) posterolateral fusion with pedicle screw instrumentation (n=40). Long-term follow-up was obtained in 101 (91%) patients. Nine patients in the exercise group were eventually operated on. Results Longitudinal analysis: At long-term follow-up pain and functional disability were significantly better than before treatment in both surgical groups. No significant differences were observed between instrumented and non-instrumented patients in any variable studied. In the exercise group the pain was significantly reduced but not the functional disability. Compared with the 2-year follow-up a significant increase in functional disability was observed, as measured by the DRI, but not the ODI, in the surgical group at long term. In the exercise group no significant changes were observed between the 2-year and the long-term follow-up. Cross-sectional analysis: Between the surgical and conservative group no significant differences were observed in any outcome measurement at long-term follow-up except for global assessment, which was significantly better for surgical patients. Of surgical patients 76% classified the overall outcome as much better or better compared with 50% of conservatively treated patients (p=0.015). Quality of life as estimated by the SF-36 at long term was not different between treatment groups in any of the eight domains studied but was considerably lower than for the normal population. Conclusions Posterolateral fusion in adult lumbar isthmic spondylolisthesis results ina modestly improved long-term outcome compared with a 1-year exercise program. Although the results show that some of the previously reported short-term improvement is lost at long term, patients with fusion still classify their global outcome as clearly better than conservatively treated patients. Furthermore, because the long-term outcome of the patients conservatively treated most likely reflects the natural course, one can also conclude that no considerable spontaneous improvement should be expected over time in adult patients with symptomatic isthmic spondylolisthesis. Substantial pain, functional disability and a reduced quality of life will in most patients most likely remain unaltered over many years.