Lumbar discectomy outcomes vary by herniation level in the Spine Patient Outcomes Research Trial

Abstract

Background: The Spine Patient Outcomes Research Trial showed an overall advantage for operative compared with nonoperative treatment of lumbar disc herniations. Because a recent randomized trial showed no benefit for operative treatment of a disc at the lumbosacral junction (L5-S1), we reviewed subgroups within the Spine Patient Outcomes Research Trial to assess the effect of herniation level on outcomes of operative and nonoperative care.

Methods: The combined randomized and observation cohorts of the Spine Patient Outcomes Research Trial were analyzed by actual treatment received stratified by level of disc herniation. Overall, 646 L5-S1 herniations, 456 L4-L5 herniations, and eighty-eight upper lumbar (L2-L3 or L3-L4) herniations were evaluated. Primary outcome measures were the Short Form-36 bodily pain and physical functioning scales and the modified Oswestry Disability Index assessed at six weeks, three months, six months, one year, and two years. Treatment effects (the improvement in the operative group minus the improvement in the nonoperative group) were estimated with use of longitudinal regression models, adjusting for important covariates.

Results: At two years, patients with upper lumbar herniations (L2-L3 or L3-L4) showed a significantly greater treatment effect from surgery than did patients with L5-S1 herniations for all outcome measures: 24.6 and 7.1, respectively, for bodily pain (p = 0.002); 23.4 and 9.9 for Short Form-36 physical functioning (p = 0.014); and -19 and -10.3 for Oswestry Disability Index (p = 0.033). There was a trend toward greater treatment effect for surgery at L4-L5 compared with L5-S1, but this was significant only for the Short Form-36 physical functioning subscale (p = 0.006). Differences in treatment effects between the upper lumbar levels and L4-L5 were significant for Short Form-36 bodily pain only (p = 0.018).

Conclusions: The advantage of operative compared with nonoperative treatment varied by herniation level, with the smallest treatment effects at L5-S1, intermediate effects at L4-L5, and the largest effects at L2-L3 and L3-L4. This difference in effect was mainly a result of less improvement in patients with upper lumbar herniations after nonoperative treatment.

Fig. 1

Spine Patient Outcomes Research Trial (SPORT) intervertebral disc herniation enrollment and randomization. This flow diagram describes the group enrollment. The percentages in the final row of boxes are the percentage of patients seen at the time of the two-year follow-up by the treatment received within two years of enrollment.

Fig. 2

The Short Form-36 (SF-36) bodily pain treatment effects. The data points represent the differences in the change of bodily pain scores between the operative and nonoperative groups at each follow-up time period (that is, the improvement in the operative group minus the improvement in the nonoperative group). The vertical bars represent the 95% confidence intervals.

Fig. 3

Oswestry Disability Index (ODI) treatment effects. The data points represent the differences in the change of ODI scores between the operative and nonoperative groups at each follow-up time period (that is, the improvement in the operative group minus the improvement in the nonoperative group). The vertical bars represent the 95% confidence intervals.