Spinal Stenosis: Surgery or Not

1
Spinal StenosisSurgery or Not?

• Suzannah Stout, MD
• April 30, 2008

2
Lumbar Spinal Stenosis

• Narrowing of Intraspinal Canal
• Most Common DJD of spine or trauma
• Disc protrusion
• Stress loading of posterior spine (facets)
• Hypertrophy of facets or ligamentum flavum
• Osteophyte formation
• Later Spondylolisthesis

3
Lumbar Spinal Stenosis
4
Lumbar Spinal Stenosis

• Less Common Causes
• Space-occupying lesions
• Post-traumatic or -surgical fibrosis
• Skeletal Disease (Pagets, ankylosing spondylitis,
  RA)
• Congenital (spina bifida, achondroplasia)

5
Common Sx

• Low Back Pain (often mild)--65
• Neurogenic Claudication--94
• Numbness/tingling, weakness--40-60
• Worst with standing or walking
• Relieved with sitting or lying down
• Those with narrowing found incidentally on
  imaging are often asymptomatic

6
Traditional Treatment

• Physical Therapy (regimens not well studied)
• Corticosteriod Injections (also not well studied)
• Indications for surgery not fully agreed upon
• Most common reason for back surgery in gt65yo

7
2005 Cochrane Review

• 31 RCTs, often with small s
• Lack of long-term outcomes beyond 2-3 yrs
• Many trials were heterogeneous spinal stenosis,
  disc nerve compression, and spondylolisthesis
• Bottom line studies inconclusive for benefit of
  surgery, esp. fusion

8
But then.
9
Study Objective

• Analyze the relative efficacy of surgical vs.
  nonsurgical treatment for spinal stenosis without
  degenerative spondylolisthesis based on patient
  self reported pain, function, and disability
  scales

10
Methods

• 13 US medical centers
• Included
• patients with neurogenic claudication or
  radicular leg sx gt12 weeks
• confirmatory imaging
• Previous PT (68), epidural injections (56),
  NSAIDS or opioids OK
• Excluded
• Spondylolisthesis
• Lumbar instability

11
Methods Interventions

• Non-surgical Therapy usual care but not
  standardized
• PT
• Home exercise instruction
• NSAIDS
• Surgery posterior decompressive laminectomy
• A small amount also received instrumented or
  noninstrumented fusion (5)

12
Methods Outcome Measures

• Primary
• Bodily pain and physical function scores on SF-36
  Survey and modified Oswestry Disability Index
• Secondary
• Pt-reported improvement satisfaction with sx and
  care
• Bothersomeness of stenosis and back pain via
  several standardized scales
• F/U at 6w, 3m, 6m, 1yr, and 2yrs
• Treatment Effect (mean in score SURG) -
  (mean in score NON-SURG)

13
Methods Two Cohorts

• Randomized Cohort
• 289 patients enrolled
• 138 assigned to surgery arm
• 151 assigned to nonsurgical treatment

• Observational Cohort
• 365 patients enrolled
• 219 chose surgery
• 146 chose nonsurgical treatment

14
BUT, patients dont always BEHAVE !
15
Methods Unintended Crossover

• Randomized Cohort
• 138 assigned to surgery --gt only 67 had surg by
  2yrs
• 151 assigned to NON-surg tx --gt 43 had surg by
  2yrs

• Observational Cohort
• 219 chose surgery--gt 96 had undergone surg by
  2yrs
• 146 chose NON-surg tx --gt 22 had surg by 2yrs

16
Methods Statistical Analysis

• Almost like 3 studies Randomized, Observational,
  and Combined
• Demographics/Baseline data Rand vs Obs cohorts,
  Surg vs Nonsurg
• Intention-to-Treat analyzed randomized cohort
• Needed 185/group to detect a 10-point difference
  in 100-point scale
• Time from enrollment

17
Methods Statistical Analysis

• As-Treated Analysis
• Time
• Surgery time starts at date of treatment
• Nonsurgical changes from baseline (even if
  eventually chose Surg) included here
• Randomized and Observational Cohorts analyzed
  separately and combined
• Predictors of Treatment Received in Randomized
  Cohort

18
Results Patients At Baseline

• Rand Cohort vs Obs Cohort
• All very similar demographically, sx severity,
  and level of stenosis
• Observational Cohort More nerve-root tension and
  less lateral recess stenosis
• Randomized Cohort Two Randomized Groups (Surg vs
  Nonsurg)
• All categories very similar
• Combined Cohorts Surg vs Nonsurg
• Surg younger, more working, more reported
  disability, more with pain worsening, more
  severe stenosis

19
Results Treatment Received

• Nonsurgical Treatment
• Similar, but more in Rand vs Obs Cohort visited
  surgeon and got injections
• Surgery
• Looked at levels decompressed, OR time, blood
  loss, post-op mortality, complications
• Complications dural tear (9), wound infection
  (2), transfusion (7)
• Reoperation by 2yrs in 8 (lt1/2 for stenosis)
• 6 Deaths (vs. 7 in Nonsurg group)

20
Results Treatment Effects

• Intention to Treat (Randomized Cohort)
• Lost power from crossover
• Only statistical significance more change in
  surgery group (8 points) in bodily pain score at
  2yrs
• No statistically significant change in Surg vs
  Nonsurg groups physical function or disability
  index
• At early times (6w, 3mo) physical function
  treatment effect went down

21
Results Treatment Effects

• As-Treated Analysis
• Rand vs Obs Cohorts
• Change in scores from baseline were statistically
  similar in the two groups
• Global Hypothesis Test
• Rand vs Obs Cohorts Surg vs Nonsurg
• Favored surgery in 3 main primary outcomes in
  both groups over all time periods
• Statistically Similar--gtCombined Cohorts

22
Results Treatment Effects

• As-Treated Analysis
• Combined Cohorts Surg vs Nonsurg
• Peak change from baseline was 6months
• Bodily Pain treatment effect of surgery was
  17-point difference at 6mo, 14-points at 2yr
• Physical Function 16 points at 6mo, 11 points at
  2yr
• Disability Index 14 points at 6mo, 11 points at
  2yr
• Secondary Outcomes pt-reported satisfied with
  symptoms and major improvement
• Improvement from baseline in Nonsurg group too

23
(No Transcript)
24
Surg
Nonsurg
Treatment Effect
25
Study Strengths

• Randomized and Observational Cohorts were
  statistically similar at baseline
• Allowed for data to be combined to study both
  cohorts together
• As-treated analysis adjusted for many confounding
  variables
• The reality of patient choice about surgery
• Only looked at Spinal Stenosis (not
  Spondylolisthesis or other disc disease)

26
Study Limitations

• Randomization
• Surgery vs Nonsurgical Treatment never blind
• Self-reported symptoms (less after 6mo?)
• Unintended Crossover
• Limited intention-to-treat analysis
• Combining Cohorts eliminating benefit of
  randomization
• Those who ultimately chose surgery were different
  at baseline
• No standard of nonsurgical treatment

27
Bottom Line

• Little risk of harm in surg vs nonsurg tx
• Both surg and nonsurg tx improved symptom scores
• There is improvement in patient-perceived pain,
  function, disability, and satisfaction
  (although, these patients were worse off from
  the start)
• Advice to patients still try noninvasive tx
  first, but may be helped by surgery

28
Discussion? Questions?

• How can you set up a study to prevent confounding
  but recognize patient choice?

29
References

• Weinstein, JN et al. Surgical versus
  Nonsurgical Therapy for Lumbar Spinal Stenosis.
  N Engl J Med 2008358794-810
• Gibson, JN, Waddell, G. Surgery for degenerative
  lumbar spondylosis. Cochrane Database Syst Rev
  2005 CD001352

30
THANKS FOR LISTENING !