Effectiveness of Microdiscectomy for Lumbar Disc Herniation: A Randomized Controlled Trial With 2 Years of Follow-up

Heikki Österman, MD; Seppo Seitsalo, MD, PhD; Jaro Karppinen, MD, PhD; Antti Malmivaara, MD, PhD


Spine. 2006;31(21):2409-2414. 

In This Article


The repeated measures model displayed no statistically significant differences in clinical outcomes between the surgical and the control group over the entire 2-year follow-up. Both groups improved significantly over time. However, at 6 weeks, discectomy patients had less leg pain and 5 of 28 patients reported full global recovery, which suggests a faster initial recovery after discectomy.

Our results suggest a rather small benefit from surgery in comparison to other reports.[7,9,10] The Maine Lumbar Spine Cohort Study noted a greater improvement with surgical than nonsurgical management of disc herniation. On the other hand, the cohort of surgical patients had more severe symptoms initially, and they may be expected to gain more from surgery.[9,10] The only randomized trial comparing discectomy and conservative management reported statistically significantly better 1-year results after discectomy.[7] The outcome measure in the study is problematic, however. Weber had a single outcome measure composed of several parameters of which patient satisfaction was of substantial significance.[7] Only satisfied patients could be graded as good or fair. This kind of outcome is potentially biased if surgical patients, as well as treating physicians, have more favorable expectations about their treatment as we noted in our study. Surgical patients were better satisfied with their treatment even if the clinical outcomes did not differ, which seems to reflect more optimistic expectations.

Our results must be interpreted with caution because of our small sample size and the large number of crossover operations. Because of the slow rate of recruitment, enrollment of patients was discontinued at a stage when the study sample was big enough to show a difference of 15 mm in pain intensity. It is possible that a smaller difference may have existed and gone unnoticed. However, minimal clinically important difference in pain on visual analog scale for lumbar surgery patients is considered to be 20 mm.[17] Actual differences noted in this study were less than 15 mm; hence, increasing the study sample would probably have contributed little to the clinical significance of the findings.

Ten patients in the control group crossed over to surgery because of unrelenting or recurring pain, and one because of progressive motor weakness. The crossover patients are problematic methodically but in a setting like this they are hard to avoid. Some of the patients may have had herniations poorly suited to conservative management, whereas in others more subjective factors may have been important. Most of the operations were performed because of unrelenting pain, which is a highly subjective indication. The reason for the operation notwithstanding, the fact remains that 40% of the control group had surgery in the end. With the intention-to-treat analysis, these patients were analyzed belonging to the control group, which might modify the results. In an effort to control for this, we performed an additional on-treatment analysis where all the discectomy patients were compared with patients remaining in the control group. The on-treatment analysis did not show any statistically significant differences between the groups, which suggests that the crossover operations probably did not have a major effect on the results. Reasons for crossover operations remain controversial, but we think that at least in some cases spinal level of the herniation may be important.

According to the subgroup analysis operative treatment resulted in a superior outcome at L4-L5, whereas no difference was noted at L5-S1. The worst outcome at 2 years was seen in the 7 crossover patients with a L4-L5 herniation. These patients may have had a worse prognosis right from the start, hence the poor response to any care. On the other hand, it is possible that they were poorly suited to conservative treatment and the delay in their operation may have contributed to the poor result.

The better surgical result at L4-L5 is, to our knowledge, a new finding that needs to be verified in further studies. Anatomically and biomechanically, the two motion segments are different and biologically it is plausible that herniations at L4-L5 and L5-S1 could be separate entities. Spinal level may be a more important factor than previously thought, and the issue merits more research.

Patients older than median (37 years) seemed to benefit more from surgery. This finding may be independent, but it may also be linked with spinal level. L4-L5 herniation patients were older than L5-S1 patients. Consequently, the apparent benefit from surgery in the older patients may simply reflect the larger number of L4-L5 herniations in this age group.

The Maine Study suggested that patients with more severe symptoms may benefit more from surgery.[9,10] We excluded patients with intolerable pain as their randomization would have been unethical; thus, our study results are applicable only to patients with tolerable pain. On the other hand, this group of patients is precisely the one where clinical decision-making is often difficult and good-quality evidence needed. Within this context, our subgroup analysis suggested that surgical patients with more than median leg pain did not differ from conservatively treated patients or patients with less pain.

The randomized setting and an acceptable rate of follow-up can be counted as merits of this study. The patients enrolled from elective orthopedic outpatient clinics all had an indication for surgery and their baseline characteristics compared well with previously published operative series.[18] Our study protocol was designed to compare microdiscectomy with natural course of disc herniation; therefore, the conservative treatment regimen was similar in both groups. It has to be noted, however, that the patients in our study were fairly young and they could not have been on a sick leave for more than 3 months. Consequently, our findings must be regarded with caution when dealing with older patients or patients receiving Workers' Compensation for a longer period of time.


Comments on Medscape are moderated and should be professional in tone and on topic. You must declare any conflicts of interest related to your comments and responses. Please see our Commenting Guide for further information. We reserve the right to remove posts at our sole discretion.
Post as: