Lumbar spinal stenosis (SpS) is a common reason for spine surgery among older adults in the United States.1 Prior studies have found an advantage for surgery compared with nonoperative treatment; however, these studies included a mixed group with and without degenerative spondylolisthesis.2–4 In prior reports from the SPORT study, as-treated comparisons with careful control for potentially confounding baseline factors showed that patients with SpS who were treated surgically had substantially greater improvement in pain and function out to 4 years than patients treated nonoperatively.5 , 6 In this article, we assess the stability of pain and functional outcomes out to 8 years for patients with SpS.
MATERIALS AND METHODS
SPORT was conducted in 11 states at 13 US medical centers with multidisciplinary spine practices. It included both a randomized cohort and a concurrent observational cohort of patients who declined randomization but met all other inclusion exclusion criteria and were willing to be followed in the same manner as the randomized patients.6–10 This design can allow for improved generalizability.11
All patients had neurogenic claudication and/or radicular leg symptoms and confirmatory cross-sectional imaging showing lumbar SpS at 1 or more levels, and they were judged to be surgical candidates. Patients with degenerative spondylolisthesis were evaluated in a separate cohort.8 , 12 All patients had ongoing symptoms for a minimum of 12 weeks that had not improved sufficiently with nonoperative intervention. The content of pre-enrollment nonoperative care was not prespecified but included physical therapy (68%), epidural injections (56%), chiropractic (28%), anti-inflammatory drugs (55%), and opioid analgesics (27%). Enrollment began in March 2000 and ended in March 2004.
The protocol surgery consisted of a standard, posterior decompressive laminectomy.7 The protocol surgery for degenerative spondylolisthesis included the possibility of bilateral single-level fusion (autogenous iliac crest bone grafting with or without posterior pedicle screw instrumentation).7 The nonoperative protocol was “usual care” recommended to include at least active physical therapy, education/counseling with home exercise instruction, and nonsteroidal anti-inflammatory drugs (NSAIDs) if tolerated.7 , 13 An extensive menu of additional treatment options (e.g., epidural steroids, analgesics, spinal manipulation) was tracked for all patients.
Primary endpoints were the SF-36 (MOS 36-Item Short-Form Health Survey) Bodily Pain and Physical Function scales,14–17 and the AAOS/Modems version of the Oswestry Disability Index (ODI)18 measured at 6 weeks, 3 months, 6 months, and yearly out to 4 years. If surgery was delayed beyond 6 weeks, additional follow-up data were obtained 6 weeks and 3 months postoperatively. Secondary outcomes included patient self-reported improvement, satisfaction with current symptoms and care,19 stenosis bothersomeness,2 , 20 and low back pain bothersomeness.2 Treatment effect was defined as the difference in the mean changes from baseline between the surgical and nonoperative groups (difference of differences).
SF-36 scores range from 0 to 100, with higher scores indicating less severe symptoms; the ODI ranges from 0 to 100, with lower scores indicating less severe symptoms; the Stenosis Bothersomeness Index ranges from 0 to 24, with lower scores indicating less severe symptoms; and Low Back Pain Bothersomeness Scale scores ranges from 0 to 6, with lower scores indicating less severe symptoms.
Statistical methods for the analysis of this trial have been reported in previous publications,6 , 8–10 , 12 , 21 and these descriptions are repeated or paraphrased here as necessary. Data for the randomized cohorts were initially analyzed on an intent-to-treat basis. Because of crossover, subsequent analyses were based on treatments actually received as described previously.5 , 12
Primary analyses compared surgical and nonoperative treatments using changes from baseline at each follow-up, with a mixed-effects longitudinal regression model including a random individual effect to account for correlation between repeated measurements within individuals. In the as-treated analyses, the treatment indicator was a time-varying covariate, allowing for variable times of surgery. Follow-up times were measured from enrollment for the intent-to-treat analyses, whereas for the as-treated analysis, the follow-up times were measured from the beginning of treatment (i.e., the time of surgery for the surgical group and the time of enrollment for the nonoperative group); baseline covariates were updated to the follow-up immediately preceding the time of surgery. This procedure has the effect of including all changes from baseline prior to surgery in the estimates of the nonoperative treatment effect and all changes after surgery in the estimates of the surgical effect. The 6-point sciatica scales and binary outcomes were analyzed via longitudinal models based on generalized estimating equations22 with linear and logit link functions, respectively, using the same intent-to-treat and adjusted as-treated analysis definitions as the primary outcomes. Data for the randomized and observational cohorts were analyzed to produce separate as-treated estimates of treatment effect. These results were compared using a Wald test to simultaneously test all follow-up visit times for differences in estimated treatment effects between the 2 cohorts.23
To evaluate the 2 treatment arms across all time periods, the time-weighted average of the outcomes (area under the curve) for each treatment group was computed using the estimates at each time period from the longitudinal regression models and compared using a Wald test.23 Kaplan-Meier estimates of reoperation rates at 8 years were computed for the randomized and observational cohorts and compared via the log-rank test.24 , 25
Computations were done using SAS procedures PROC MIXED for continuous data and PROC GENMOD for binary and non-normal secondary outcomes (SAS, version 9.1, Windows XP Pro; Cary, NC). Statistical significance was defined as P < 0.05, based on a 2-sided hypothesis test, with no adjustments made for multiple comparisons.
A total of 654 participants with SpS were enrolled out of 1091 eligible for enrollment (289 in the randomized group and 365 in the observational group) (Figure 1). In the randomized group, 138 patients were assigned to surgical treatment and 151 to nonoperative treatment. Of those randomized to surgery, 67% (92/138) received surgery by 2 years, 69% (95/138) by 4 years, and 70% (97/138) by 8 years. In the group randomized to nonoperative care, 43% (65/151) underwent surgery by 2 years, 49% (75/151) by 4 years, and 52% (78/151) by 8 years (Figure 1). In the observational group, 219 patients initially chose surgery and 146 patients initially chose nonoperative care. Of those initially choosing surgery, 96% (211/219) underwent surgery by 2 years and 97% (213/219) by 4 years; this remained unchanged at 8 years. Of those choosing nonoperative treatment, 22% (32/146) had surgery by 2 years, 26% (38/146) by 4 years, and 27% (40/146) by 8 years (Figure 1). The proportion of enrollees for whom data were obtained at the 8-year follow-up visit was 55% in the randomized group and 52% in the OC group, with losses due to dropouts, missed visits, or deaths.
Table 1 shows the summary statistics of participants in the randomized and observational cohorts and according to treatment received. The randomized and observational cohorts were remarkably similar except for their preferences for surgery (P < 0.001), with more randomized group patients unsure of their preference (34% vs. 7%), and fewer randomized group patients definitely preferring either surgery (12% vs. 46%) or nonoperative treatment (13% vs. 24%).
At baseline, patients in the group undergoing surgery within 8 years were younger and had slightly more osteoporosis than those receiving nonoperative treatment. They had worse pain, function, disability, and symptoms than patients in the nonoperative group. Patients in the surgery group were more dissatisfied with their symptoms and, at enrollment, more often rated their symptoms as worsening and definitely preferred surgery. These observations highlight the need to control for baseline differences in the as-treated models. On the basis of the selection procedure for variables associated with treatment, missing data, and outcomes, the final as-treated models controlled for the following covariates: age; sex; compensation; baseline stenosis bothersomeness; income; smoking status; duration of most recent episode; treatment preference; diabetes; joint problem; stomach comorbidity; baseline score (for SF-36, ODI); and center.
In the combined randomized/observational group, nonoperative treatments included physical therapy (46%), visits to a surgeon (49%), NSAIDs (52%), and opioids (40%). More patients in the randomized cohort reported receiving injections (54% vs. 41%; P = 0.017), whereas more observational group patients reported receiving NSAIDs (59% vs. 47%; P = 0.035) and “other” medications (76% vs. 63%; P = 0.01).
Surgical Treatment and Complications
For the combined cohorts, the mean surgical time was 129 minutes, with a mean blood loss of 311 mL (Table 2). The most common surgical complication remained dural tear (9%). The 8-year reoperation rate was 18%, with no significant difference between the randomized and observational groups.
At 8 years, there were 27 deaths in the nonoperative group compared with 35 expected on the basis of age-/sex-specific mortality rates; there were 39 deaths in the surgery group compared with 53 expected. The hazard ratio based on a proportional hazards model adjusted for age was 0.26 (95% confidence interval, 0.45–1.2; P = 0.26). All 66 deaths were independently reviewed, and 45 were judged not to be treatment related. One patient died 9 days after surgery of a myocardial infarction. The death was judged probably related to treatment by the Dartmouth-Hitchcock Medical Center review and not related to treatment by the external review. Twenty deaths were from “unknown” causes, with median (min, max) days from enrollment/surgery of 2297 (501, 2856).
Main Treatment Effects
The intent-to-treat analyses of the randomized cohort with SpS showed no statistical differences between surgery and nonoperative care based on overall global hypothesis tests for differences in mean changes from baseline for either the primary (Figure 2) or secondary outcomes (Figure 3).
In the as-treated analyses of the randomized group, the advantage for surgery seen at 4 years diminished over time to the point that there were no longer any discernible differences between the treatment groups after 5 years, although the overall comparison between treatment groups remained significant (Figure 2). Interestingly, the advantage for surgery was maintained out to 8 years in the observational cohort. The Wald test comparing the as-treated randomized and observational treatment effects during all time periods was statistically significant only for ODI and of borderline significance for Bodily Pain (P = 0.08 for Bodily Pain; P = 0.36 for Physical Function; and P = 0.02 for ODI).
Among the secondary outcomes, differences seen in the as-treated groups at 4 years were generally maintained at 8 years, however, there was a steady decline over time in the proportion of surgical patients rating their result as a “major improvement” compared with baseline; this occurred in the observational cohort as well despite relatively stable results on symptom severity and functional status measures in this cohort during the same time period.
Loss to Follow-up
At the 8-year follow-up visit, data were obtained for 53% of initial enrollees, with losses due to dropouts, missed visits, or deaths. Table 3 summarizes the baseline characteristics of those lost to follow-up compared with those retained in the study at 8 years. Those who remained in the study at 8 years were somewhat younger; more likely to be college educated, married, and working at baseline; had fewer comorbidities; and had a shorter duration of symptoms, less severe symptoms, and less severe radiographical stenosis at baseline. These differences were small but statistically significant. Table 4 summarizes the short-term outcomes during the first 2 years for those retained in the study at 8 years compared with those lost to follow-up. Those lost to follow-up had worse outcomes on average; however, this was true in both the surgical and nonoperative groups with nonsignificant differences in treatment effects. The long-term outcomes are therefore likely to be somewhat overoptimistic on average in both groups, but the comparison between surgical and nonoperative outcomes seems to be unbiased despite the long-term loss to follow-up.
In patients presenting with signs and symptoms of image-confirmed SpS persisting for at least 12 weeks, the intention-to-treat analysis found no significant difference between surgical or nonoperative treatments. However, as has been reported previously, these results must be viewed in the context of substantial rates of nonadherence to the assigned treatment; this mixing of treatments generally biases treatment effect estimates toward the null.6–10 , 12 , 21
As-treated analyses continue to show an overall advantage for surgery; however, in the randomized group, the as-treated results for surgical and nonoperative treatments converged after 5 years. The advantage for surgery in the observational group at 4 years was maintained out to 8 years. This is the only divergence in outcomes seen so far between the randomized and observational results. This might have to do with greater baseline differences in the observational groups than in the as-treated randomized groups. Although the as-treated analysis of the randomized group loses the strong protection against confounding supplied by randomization, the as-treated randomized groups were much more similar at baseline than the observational groups. Thus, the long-term results in the as-treated randomized group are somewhat less likely to be confounded by baseline differences, suggesting that the advantage of surgery in SpS likely does diminish over time.
Comparisons With Other Studies
SPORT represents by far the largest study of its kind and the largest study to isolate SpS from stenosis secondary to degenerative spondylolisthesis. Its cohort was recruited from 13 centers in 11 US states, making it the most generalizable study of stenosis. The characteristics of the participants and the short-term outcomes of SPORT as previously reported are comparable with studies both of isolated SpS and of mixed cohorts of patients with and without degenerative spondylolisthesis with stenosis.2–4
Few studies have long-term outcomes to compare with SPORT's, and most of these include mixed cohorts of stenosis with and without some degree of degenerative spondylolisthesis. The Maine Lumbar Spine Study (MLSS)2 cohort had outcomes for 8 to 10 years. In terms of sciatica bothersomeness, long-term results for the MLSS showed persistent statistically significant benefit for the surgical group (treatment effect = −9.4; P = 0.02) similar to the SPORT observational cohort. However, the MLSS found no difference between the treatment groups at 8 to 10 years in the percent satisfied with current symptoms (55% surgical vs. 49% nonoperative; P = 0.52) whereas SPORT showed a persistent advantage for the surgical group in both the randomized (56% surgery vs. 33% nonoperative; P < 0.001) and observational (64% vs. 34%; P < 0.001) cohorts.
Two other studies have followed mixed cohorts of patient with stenosis with and without degenerative spondylolisthesis, each with 10 years worth of data. One reported on the Amundsen et al 26 , 27 original stenosis cohort and the other was the Slatis et al 28 long-term study of a mixed stenosis cohort. Amundsen et al 27 report a small, persistent advantage for the surgery group at 10 years. Slatis et al 28, reporting the 6-year results of a randomized controlled trial of moderate SpS, found a narrowing of the advantage for surgery at 6 years, but with a persistently significant advantage when viewed during the entire time period, similar to the SPORT as-treated randomized results.
There was little evidence of harm from either treatment. In the interval between 4 and 8 years, there have not been any cases of paralysis in either the surgical or nonoperative group, and there was no statistical difference in morbidity between the surgical and nonoperative groups. The 8-year rate of reoperation for recurrent stenosis was 10%, and the overall reoperation rate increased from 13% at 4 years to 18% at 8 years, compared with 23% at 8 to 10 years in the MLSS. The 6-month perioperative mortality rate was extremely low at 0.2%.
In the as-treated analysis combining the randomized and observational cohorts of patients with SpS, those treated surgically showed significantly greater improvement in pain, function, satisfaction, and self-rated progress during 8 years than patients treated nonoperatively, but with convergence in outcomes between treatment groups after 5 years in the randomized cohort. Preferential loss to follow-up of patients with worse baseline characteristics and early outcomes in both treatment groups could lead to overestimates of long-term outcomes but not likely bias treatment effect estimates.
- Intent-to-treat analyses in the randomized cohort showed no differences between treatment groups; however, 70% of those randomized to surgery and 52% of those randomized to nonoperative had undergone surgery by 8 years.
- In the as-treated analyses, outcomes converged in the randomized cohort with no difference between treatment groups at 6 to 8 years whereas a persistent advantage for surgery was seen in the observational cohort through 8 years.
- At 8 years, outcomes data were available from 53% of the original enrollees.
- Loss to follow-up of patients with worse early outcomes in both treatment groups could lead to overestimates of long-term outcomes but not likely treatment effect estimates.
1. Weinstein JN, Lurie JD, Olson PR, et al. United States' trends and regional variations in lumbar spine surgery
: 1992–2003. Spine 2006;31:2707–14.
2. Atlas SJ, Deyo RA, Keller RB, et al. The Maine Lumbar Spine Study, part III: 1-year outcomes
of surgical and nonsurgical management of lumbar spinal stenosis
. Spine 1996;21:1787–94; discussion 94–5.
3. Atlas SJ, Keller RB, Robson D, et al. Surgical and nonsurgical management of lumbar spinal stenosis
: four-year outcomes
from the Maine Lumbar Spine Study. Spine 2000;25:556–62.
4. Malmivaara A, Slatis P, Heliovaara M, et al. Surgical or nonoperative
treatment for lumbar spinal stenosis
? A randomized controlled trial. Spine 2007;32:1–8.
5. Weinstein JN, Tosteson TD, Lurie JD, et al. Surgical versus nonoperative
treatment for lumbar spinal stenosis
four-year results of the Spine Patient Outcomes
Research Trial. Spine (Phila Pa 1976) 2010;35:1329–38.
6. Weinstein JN, Tosteson TD, Lurie JD, et al. Surgical versus
nonsurgical therapy for lumbar spinal stenosis
. N Engl J Med 2008;358:794–810.
7. Birkmeyer NJ, Weinstein JN, Tosteson AN, et al. Design of the Spine Patient outcomes
Research Trial (SPORT
). Spine 2002;27:1361–72.
8. Weinstein JN, Lurie JD, Tosteson TD, et al. Surgical versus
nonsurgical treatment for lumbar degenerative spondylolisthesis
. N Engl J Med 2007;356:2257–70.
9. Weinstein JN, Lurie JD, Tosteson TD, et al. Surgical vs nonoperative
treatment for lumbar disk herniation: the Spine Patient Outcomes
Research Trial (SPORT
) observational cohort. JAMA 2006;296:2451–9.
10. Weinstein JN, Tosteson TD, Lurie JD, et al. Surgical vs nonoperative
treatment for lumbar disk herniation: the Spine Patient Outcomes
Research Trial (SPORT
): a randomized trial
. JAMA 2006;296:2441–50.
11. Tunis SR, Stryer DB, Clancy CM. Practical clinical trials: increasing the value of clinical research for decision making in clinical and health policy. JAMA 2003;290:1624–32.
12. Weinstein JN, Lurie JD, Tosteson TD, et al. Surgical compared with nonoperative
treatment for lumbar degenerative spondylolisthesis
. Four-year results in the Spine Patient Outcomes
Research Trial (SPORT
) randomized and observational cohorts. J Bone Joint Surg Am 2009;91:1295–304.
13. Cummins J, Lurie JD, Tosteson TD, et al. Descriptive epidemiology and prior healthcare utilization of patients in the Spine Patient Outcomes
Research Trial's (SPORT
) three observational cohorts: disc herniation, spinal stenosis
, and degenerative spondylolisthesis
. Spine 2006;31:806–14.
14. McHorney CA, Ware JE Jr, Lu JF, et al. The MOS 36-item Short-Form Health Survey (SF-36), part III: tests of data quality, scaling assumptions, and reliability across diverse patient groups. Med Care 1994;32:40–66.
15. Stewart AL, Greenfield S, Hays RD, et al. Functional status and well-being of patients with chronic conditions. Results from the Medical Outcomes
Study. JAMA 1989;262:907–13.
16. Ware J, Sherbourne D. The MOS 36-item Short-Form Health Survey. Med Care 1992;30:473–83.
17. Ware JJ. SF-36 Health Survey: Manual and Interpretation Guide. Boston, MA: Nimrod Press; 1993.
18. Fairbank JC, Pynsent PB. The Oswestry Disability Index. Spine 2000;25:2940–52; discussion 52.
19. Deyo RA, Diehl AK. Patient satisfaction with medical care for low-back pain. Spine 1986;11:28–30.
20. Patrick DL, Deyo RA, Atlas SJ, et al. Assessing health-related quality of life in patients with sciatica. Spine 1995;20:1899–908; discussion 909.
21. Weinstein JN, Lurie JD, Tosteson TD, et al. Surgical versus nonoperative
treatment for lumbar disc herniation: four-year results for the Spine Patient Outcomes
Research Trial (SPORT
). Spine 2008;33:2789–800.
22. Diggle PJ, Heagerty P, Liang K-L, et al. Analysis of Longitudinal Data. Oxford, England: Oxford University Press; 1994.
23. Fitzmaurice G, Laird N, Ware J. Applied Longitudinal Analysis. Philadelphia, PA: John Wiley & Sons; 2004.
24. Kaplan EL, Meier P. Nonparametric estimation from incomplete observations. J Am Stats Assoc 1958;53:457–81.
25. Peto R, Peto J. Asymptotically efficient rank invariant test procedures. J R Stat Soc Ser A (Gen) 1972;135:185–207.
26. Amundsen T, Weber H, Lilleas F, et al. Lumbar spinal stenosis
. Clinical and radiologic features. Spine (Phila Pa 1976) 1995;20:1178–86.
27. Amundsen T, Weber H, Nordal HJ, et al. Lumbar spinal stenosis
: conservative or surgical management? A prospective 10-year study. Spine (Phila Pa 1976) 2000;25:1424–35; discussion 35–6.
28. Slatis P, Malmivaara A, Heliovaara M, et al. Long-term results of surgery
for lumbar spinal stenosis
: a randomised controlled trial. Eur Spine J 2011;20:1174–81.
Keywords:© 2015 by Lippincott Williams & Wilkins
spinal stenosis; degenerative spondylolisthesis; randomized trial; surgery; nonoperative; SPORT; outcomes