From Wiki Journal Club
Jump to navigation Jump to search
Friedly JL, et al. "A randomized trial of epidural glucocorticoid injections for spinal stenosis". The New England Journal of Medicine. 2014. 371(1):11-21.
PubMedFull textPDF

Clinical Question

Among patients with lumbar spinal stenosis and associated pain, does an epidural injection of a glucocorticoid plus lidocaine reduce disability and leg pain at six weeks when compared to an injection of lidocaine alone?

Bottom Line

Among patients with lumbar spinal stenosis and associated pain, epidural injection of a glucocorticoid plus lidocaine is not associated with improvement in disability scores or leg pain at six weeks when compared to injection of lidocaine alone.

Major Points

Epidural administration of glucocorticoids is widely used for symptomatic relief in patients with spinal stenosis though the therapy has not been evaluated in a large, randomized trial. A 2013 Cochrane review classified the existing evidence for epidural injections in this disease as "very low quality" and identified an urgent need for clinical trials.[1]

The 2014 ARHQ-funded Lumbar Epidural Steroid Injections for Spinal Stenosis (LESS) trial randomized 400 adults ≥50 years old with lumbar spinal stenosis and associated symptoms to epidural administration of lidocaine or a glucocorticoid+lidocaine in a double-blinded fashion. Patients were excluded if they required surgery for spinal instability or if they had significant neurologic or hip pathology. Groups were balanced except for longer symptom duration in glucocorticoid+lidocaine. Both therapies were associated with significant symptom improvement in both primary outcomes at six weeks (which were change on a validated function score and average leg pain in the prior week) though there was no between-group difference for either outcome. The glucocorticoid+lidocaine group experienced improvement in both components of the primary endpoint at three weeks, though the authors note that this was clinically insignificant. When adjusting for duration of pain, the glucocorticoid+lidocaine group experienced a small improvement over the lidocaine group in disability scores at six weeks. As this analysis was post-hoc and the effect was small, it is of dubious clinical significance.

The glucocorticoid+lidocaine group experienced more adverse events, though they were relatively minor and transient.[2] This group also had greater AM cortisol suppression, suggesting that despite the localized administration of the glucocorticoid, its effects are systemic. The glucocorticoid+lidocaine group had improvement in depressive symptoms and higher satisfaction with treatment than the lidocaine group, which may be related to an unintended effect of the glucocorticoid unrelated to its effects on spinal pathology.[2]

Though the study was not designed to detect a difference in technique, the glucocorticoid+lidocaine group experienced greater improvement in both components of the primary outcome at three weeks with an interlaminar approach. There was no difference with either approach at six weeks.

While there was no sham group to evaluate for clinical benefit of lidocaine, it is unlikely that this medication provides pain relief or improves disability at a prolonged duration as its half-life is in the scale of hours. As such, the clinical improvement observed from baseline in both groups is likely an effect of regression to the mean and/or efficacy of other unmeasured therapies like PT or non-trial medications.


As of April 2015, no guidelines have been published that reflect the results of this trial.


  • Multicenter, double-blind, randomized, controlled trial
  • N=400 (2,224 screened)
    • Lidocaine (n=200)
    • Glucocorticoid+lidocaine (n=200)
  • Setting: 16 US centers
  • Enrollment: 2011-2013
  • Follow-up: 6 weeks
  • Analysis: Intention-to-treat
  • Primary outcomes:
    • Roland-Morris Disability Questionnaire
    • Average rating of leg pain in prior week


Inclusion Criteria

  • Age ≥50 years
  • Pain in low back, buttock and/or lower extremity with standing, walking, and/or spinal extension
    • Average pain rating >4 in prior week (out of 10, 10 indicating the worst possible pain)
  • Roland-Morris Disability Questionnaire (RMDQ) score ≥7 (out of 24, higher score indicates higher physical limitation from back and/or leg pain)
  • Mild-severe lumbar central canal spinal stenosis by MRI or CT
  • Lower extremity symptoms of neurogenic claudication
  • English reading with enough proficiency to complete assessment instruments

Exclusion Criteria

  • Cognitive impairment preventing informed consent or accurate collection of data
  • Fibromyalgia
  • Chronic widespread pain
  • LE amputation
  • Parkinson's disease
  • Head injury
  • Stroke
  • "Other neurologic conditions"
  • Spinal instability requiring surgery
  • Severe osteoporosis with multiple compression fractures or a fracture at the level of the stenosis
  • Metastatic cancer
  • Alcohol or illegal drug abuse
  • Possible pregnancy or other condition precluding fluoroscopy use
  • Hip joint pathology or pain with internal hip rotation
  • Local or systemic infection
  • Allergy to local anesthetic, steroid, or contrast

Baseline Characteristics

From the lidocaine group. Groups were similar except where specified. Comparisons are lidocaine vs. glucocorticoid+lidocaine.

  • Demographics: Age 68 years, female 52%, married or living with partner 56%
    • Race/ethnicity: White 69%, Black 26%, other 4.5%, Hispanic 3%
    • Educational: High school or less 33%, some college 28%, undergraduate completed 16%, professional/graduate completed 23%
    • Employment: Full or part time 36%, retired but not disabled 44%, retired and disabled 12%, other 9%
  • PMH: Current smoker 16%, diabetes on insulin 7.5%
  • Baseline health data: 30 kg/m2
  • Lumbar stenosis-specifics:
    • Physician rating of severity: Mild 22%, moderate 48%, severe 30%
  • Back pain-specifics:
    • Duration (P=0.02):
      • <3 months: 20.1% vs. 12.0%
      • 3-12 months: 31.2% vs. 29.5%
      • 1-5 years: 21.1% vs. 33.5%
      • >5 years: 27.6% vs. 25.0%
    • Rating of pain relief expectation: 7.8 (out of 10, with 10 indicating that the epidural injection would be extremely helpful in resolving the pain)
    • RMDQ score: 16
    • Numerical pain score: 7.2


  • Patients were randomly assigned to a group:
    • Lidocaine
    • Glucocorticoid+lidocaine
  • Injections were either interlaminar or transforaminal and were fluoroscopically-guided


Presented as lidocaine vs. glucocorticoid+lidocaine.

Primary Outcomes

Roland-Morris Disability Questionnaire (RMDQ) at 6 weeks
Out of 24, higher score indicates higher physical limitation from back and/or leg pain.
12.5 vs. 11.8
Change from baseline: -3.1 vs. -4.2 (adjusted difference -1.0; 95% CI -2.1 to 0.1; P=0.07)
Average rating of leg pain in prior week at 6 weeks
Out of 10, 10 indicating worst pain possible.
4.6 vs. 4.4
Change from baseline: -2.6 vs. -2.8 (adjusted difference -0.2; 95% CI -0.8 to 0.4; P=0.48)

Secondary Outcomes

RMDQ at 3 weeks
13.1 vs. 11.7
Change from baseline: -2.6 vs. -4.4 (adjusted difference -1.8; 95% CI -2.8 to -0.9; P<0.001)
Average rating of leg pain in prior week at 3 weeks
5.0 vs. 4.4
Change from baseline: -2.2 vs. -2.9 (adjusted difference -0.6; 95% CI -1.2 vs. -0.1; P=0.02)
Questionnaires at 6 weeks
BPI: No difference
Symptoms and physical function: No difference
Satisfaction with treatment: 54% vs. 67% (P=0.01)
EQ-5D: No difference
GAD-7: No difference
PHQ-8: More improvement in depression symptoms with glucocorticoid+lidocaine (average treatment effect -1.0; P=0.007)

Additional Analyses

Post-hoc analysis of the primary outcomes at 6 weeks, adjusted for duration of pain, change since baseline
RMDQ: Greater reduction in glucocorticoid+lidocaine (adjusted difference -1.2; 95% CI -2.3 to -0.1; P=0.03)
Average rating of leg pain in prior week: No difference (adjusted difference -0.3; 95% CI -0.9 to 0.3; P=0.32)
Morning cortisol level
At 3 weeks
<3 ug/dL: 0.7% vs. 9.8% (P<0.001)
<10 ug/dL: 42.4% vs. 60.1% (P=0.001)
At 6 weeks
<3 ug/dL: 0% vs. 3.3% (P=0.06)
<10 ug/dL: 37.8% vs. 53.3% (P=0.01)
Glucocorticoids used
Celestone: 30.0% vs. 31.0%
Depomedrol: 29.0% vs. 30.5%
Dexamethasone: 17.0% vs. 17.0%
Kenalog: 24.0% vs. 21.5%
Transforaminal approach
30.5% vs. 28.5%
Bilateral injections: 50.8% vs. 42.1%
Multilevel injections: 31.1% vs. 24.6%

Subgroup Analysis

For subgroup analyses, the authors defined statistical significance as <0.025.

Change in RMDQ score by approach
Week 3: -2.9 vs. -5.4 (adjusted difference -2.5; 95% CI -3.7 to -1.3; P<0.001)
Week 6: -3.3 vs. -4.8 (adjusted difference -1.4; 95% CI -2.8 to -0.1; P=0.04)
Week 3: -1.8 vs. -1.8 (adjusted difference -0.1; 95% CI -1.7 to 1.6; P=0.94)
Week 6: -2.6 vs. -2.4 (adjusted difference -0.3; 95% CI -1.9 to 1.8; P=0.95)
Change in average leg pain in prior week
Week 3: -2.4 vs. -3.2 (adjusted difference -0.9; 95% CI -1.5 to -0.3; P=0.005)
Week 6: -2.8 vs. -3.1 (adjusted difference -0.3; 95% CI -1.0 to 0.4; P=0.37)
Week 3: -2.0 vs. -1.9 (adjusted difference -0.0; 95% CI -0.9 to 0.9; P=0.99)
Week 6: -2.0 vs. -2.0 (adjusted difference 0.1; 95% CI -0.9 to 1.0; P=0.89)

Adverse Events

Total number of adverse events
34 vs. 58 events (P=0.02)
Interlaminar approach: 14/139 vs. 32/143 (P=0.02)
Transforaminal approach: 20/61 vs. 26/57 (P=0.27)
≥1 adverse event
15.5% vs. 21.5% (P=0.08)
Procedure-related: 9.5% vs. 15.0% (P=0.09)
Hospitalization and/or surgery
4 vs. 5 events
Number of specific events
Excessive pain: 7 vs. 5
Headache: 3. vs. 8
Fever and/or infection: 2 vs. 10
Dizziness and/or lightheadedness: 4 vs. 4
Numbness and/or tingling: 4 vs. 5
CV and/or lung problems: 2 vs. 6
Falls: 2 v. 4
Facial flushing: 0 vs. 3
Skin irritation: 4 vs. 2
Leg swelling: 1 vs. 2
Dural puncture: 1 vs. 1
Other: 4 vs. 11


  • Low rate of enrollment from screened population
  • No sham injection so unable to detect lidocaine-derived benefit
  • Unclear if leg pain experienced was radicular or diffuse in nature and if related subgroup analysis would have shown different responses with glucocorticoid administration[2]
  • Interventions included both transforaminal and interlaminar approaches with variable quantities of medication delivered in each case and four different glucocorticoids utilized
  • Potential variation in technique amongst numerous providers and study sites may confound the results
  • Duration of follow-up for outcomes assessment at 6 weeks may be longer than the typically anticipated duration of symptom relief; thereby inappropriately emphasizing the diminishing effects of symptom relief in both intervention arms


  • Agency for Healthcare Research and Quality
  • Authors with multiple financial disclosures

Further Reading