Norwegian registry study, 2252 patients: no subgroup differences from adding fusion to decompression

Authors reported in Annals of internal medicine a cohort study that emulated a target trial of lumbar decompression with fusion versus decompression alone in degenerative lumbar spondylolisthesis. The data came from the Norwegian Registry for Spine Surgery and the nested NORDSTEN-DS randomized trial. Of 2252 patients, 236 took part in the randomized trial doi.orgAnnals of internal medicinedoi.orgOpen source ↗.

Earlier randomized trials showed limited benefits and increased risks from adding fusion to decompression, while the use of fusion has grown in the United States. The new analysis compared the two approaches across patient subgroups.

Patients and methods

The study included adults aged 18 to 80 years with symptomatic back or leg pain lasting at least 3 months. The mean age was 66 years, and 1576 (70%) were women. A total of 1446 patients had decompression alone and 806 had decompression with fusion. Unadjusted baseline Oswestry Disability Index (ODI) scores were 38.9 (SD, 14.9) and 41.4 (SD, 14.7), respectively.

The primary outcomes were ODI score and reoperation at 12 months. Linear regression with inverse probability weighting was used to control for baseline confounding. Prespecified subgroups covered sex, age, smoking, body mass index (BMI), back pain intensity and American Society of Anesthesiologists (ASA) grade.

Results at 12 months

At 12 months, the ODI score was 22.3 (95% CI, 15.5 to 29.1) after decompression alone and 20.0 (CI, 13.9 to 26.2) after decompression with fusion. The mean difference was 2.2 (CI, -1.6 to 6.1) doi.orgAnnals of internal medicinedoi.orgOpen source ↗. Reoperation risk at 12 months did not differ between the groups.

In exploratory analyses, the ODI score was worse with fusion in men with a BMI of 30 kg/m² or higher. Overall, the authors concluded that patient subgroups did not have differential effects from adding fusion to decompression at 12 months.

The authors named possible residual confounding as a limitation of the study.