U.S. flag

An official website of the United States government

NCBI Bookshelf. A service of the National Library of Medicine, National Institutes of Health.

Cover of Comparing the Effects of Two Types of Epidural Shots on Pain and Physical Ability in Older Adults with Lumbar Spinal Stenosis

Comparing the Effects of Two Types of Epidural Shots on Pain and Physical Ability in Older Adults with Lumbar Spinal Stenosis

, MD, , MD, PhD, , MS, , PhD, , PhD, , PhD, , PhD, , PhD, and , MD, MPH.

Author Information and Affiliations

Structured Abstract

Background:

Epidural steroid injections (ESIs) are commonly used to treat spinal stenosis symptoms, but evidence regarding their long-term effects is lacking.

Objectives:

The Lumbar Epidural Steroid injections for Spinal stenosis (LESS)-Extended Research (LESSER) study comprised several different but related aims that sought to better understand the impact of epidural corticosteroid injections on spinal stenosis symptoms. We sought to contextualize the results of the LESS trial by exploring whether the outcomes reported in the trial (pain and disability) adequately reflected the most important outcomes to older adults with spinal stenosis seeking treatment with ESI (aim 1) and whether providing patients with data on how much improvement in pain and disability they achieved after treatment with ESI impacted their decision-making regarding further treatment with ESI (aim 2). We also sought to determine the long-term risks and benefits of ESI for spinal stenosis and identify which subgroups of patients were most likely to benefit from ESI for spinal stenosis (aim 3).

Methods:

The LESS trial was a multicenter, double-blind, randomized trial comparing epidural injections of corticosteroid plus lidocaine vs lidocaine alone for imaging-confirmed lumbar spinal stenosis in 400 participants with moderate to severe leg pain and disability. Primary outcomes up to 24 months included the Roland-Morris Disability Questionnaire (RMDQ; range 0-24, with higher scores indicating greater disability) and leg pain intensity (range, 0 = no pain to 10 = pain as bad as you can imagine). Secondary outcomes included opioid use, spine surgery, and crossover rates. LESSER was an extension of the original LESS trial, with 3 additional aims to better understand what outcomes are of importance to patients with spinal stenosis and describe the long-term effects of ESI. Aim 1: We recruited patients aged 50 and older with spinal stenosis symptoms to participate in focus groups about back pain outcomes. Aim 2: We randomized 165 patients enrolled in the LESS trial to receive individualized reports depicting their 0- to 12-month outcomes just before collecting 18-month outcomes (intervention group) or to receive them after the study conclusion at 24 months (control group). We asked participants if they found the report helpful in making decisions about future ESI treatment and compared 24-month pain and disability between randomized groups. Aim 3: We analyzed data through 12 months of the LESSER trial to address questions about long-term effects of ESI (6 weeks was the primary outcome time point in the LESS trial).

Results

  • Aim 1 results: A total of 33 older adults with spinal stenosis from the community participated in 6 focus groups. The 3 highest-rated problems were experiencing pain/discomfort (88% of participants), problems with physical function (85%), and difficulty exercising (73%).
  • Aim 2 results: Most of the 165 nested RCT participants found the individualized reports understandable (70%) and helpful (65%) in making ESI treatment decisions. However, there was no difference in patient-reported outcomes or utilization at 24 months between the patients randomized to receive the individualized report at 18 months and patients who did not receive the report until after completion of the study at 24 months.
  • Aim 3 results: At 6 weeks, both treatment groups improved. The corticosteroid plus lidocaine group saw a 0- to 6-week change in RMDQ score of −4.2 (SD, 5.8) and in leg pain of −2.8 (SD, 3.1); the lidocaine alone group saw a 0- to 6-week change in RMDQ score of −3.1 (SD, 5.3) and in leg pain of −2.6 (SD, 3.0). However, there were no clinically meaningful differences between the 2 groups. At 12 months, both treatment groups maintained improvements in outcomes observed at 6 weeks, with no significant differences between those randomized to corticosteroid plus lidocaine vs lidocaine alone on the RMDQ (adjusted mean difference = −0.4; 95% CI, −1.6 to 0.9; P = .55); leg pain (adjusted mean difference = 0.1; 95% CI, −0.5 to 0.7; P = .75); opioid use (41.4% vs 36.3%; P = .41); or spine surgery (16.8% vs 11.8%; P = .22).

Conclusions:

For lumbar spinal stenosis, treatment with epidural injections of corticosteroid plus lidocaine offers no long-term benefits beyond that of injections of lidocaine alone in terms of self-reported pain and function or reduction in use of opioids and spine surgery. Provision of individualized patient-reported outcomes data may be helpful for patients in making chronic pain treatment decisions, but further research is needed to understand the impact of providing these data on outcomes.

Limitations and Subpopulation Considerations:

Because this was a pragmatic trial, there was inherent heterogeneity in patient characteristics. The results of this trial can be applied only to people aged 50 and older with central lumbar spinal stenosis.

Background

Lumbar spinal stenosis is one of the most common causes of low back pain among older adults and can result in significant disability.1 The cause of spinal stenosis is often multifactorial, and the clinical presentation can be variable. Degenerative changes in the spine such as spondylosis, facet arthropathy, disc degeneration, and scoliosis may all contribute to the development of spinal stenosis. The symptoms of lumbar spinal stenosis range from low back pain to neurogenic claudication with lower extremity pain, weakness, or sensory changes and are often aggravated by walking. Because the causes of spinal stenosis are most frequently degenerative changes, the symptoms can worsen over time. Lumbar spinal stenosis is often associated with poor patient health outcomes and functioning; high resource utilization; and substantial payer, patient, and health system costs.

Despite the prevalence of symptomatic lumbar spinal stenosis among older adults, treatment remains controversial, with limited comparative effectiveness evidence. Common treatments include conservative measures such as nonsteroidal anti-inflammatory drugs, activity modification, and physical therapy, as well as more invasive treatments such as epidural steroid injections (ESIs) and surgery.1 Although surgery provides some benefit to certain individuals with spinal stenosis,2,3 the results of surgical treatment tend to be modest and may be accompanied by short- and long-term postoperative surgical complications.4 Consistent with the general trend of using more minimally invasive treatment approaches in health care, ESIs are being used with increasing frequency as a less invasive, potentially safer, and more cost-effective treatment.5-7

Evidence Gaps

An estimated 25% of all ESIs performed in the Medicare population and 74% of ESIs in the Veterans Health Administration (VA) are for spinal stenosis.8,9 Rates and associated costs of ESIs for spinal stenosis increased nearly 300% over the past 2 decades and it is estimated that more than 2.2 million lumbar ESIs are performed each year for spinal stenosis in the Medicare population,8-11 despite the lack of rigorous randomized controlled trials (RCTs) evaluating efficacy and safety.12 Before the Lumbar Epidural Steroid injections for Spinal Stenosis (LESS) trial results publication, only 6 clinical trials published in peer reviewed journals assessed the efficacy of ESI for spinal stenosis compared with either local anesthetic or saline injections.13-18 These studies were all rated as poor or fair quality in a recent Agency for Healthcare Research in Quality (AHRQ) systematic review, with a variety of methodological issues, including small sample sizes, inadequate blinding, unclear randomization or allocation concealment methods, and failure to report primary outcomes.7 Only 1 of these trials reported improved outcomes with those treated with corticosteroids, but the methodological flaws with these studies and the large variations in methodology limit the conclusions that can be drawn about treatment efficacy. Specifically, these trials did not address long-term outcomes (beyond 3 months); whether outcomes differ by approach (interlaminar or transforaminal); whether any patient, disease, or imaging characteristics predicted outcomes; and whether the corticosteroid itself was associated with systemic side effects.

Research Questions

The overall objective of this multicenter, double-blind RCT was to evaluate the comparative effectiveness of ESI plus local anesthetic vs local anesthetic alone in improving pain and function among older adults with back pain and lumbar spinal stenosis. The LESS trial was funded through the AHRQ. The LESS-Extended Research (LESSER) study was the PCORI-funded continuation of the LESS trial through 24 months. No new patients were enrolled for the LESSER portion of the study.

The LESSER study comprised several different but related aims that sought to better understand the impact of epidural corticosteroid injections on spinal stenosis symptoms. We sought to contextualize the results of the LESS trial by exploring whether the outcomes reported in the trial (pain and disability) adequately reflected the most important outcomes to older adults with spinal stenosis seeking treatment with ESI (aim 1) and whether providing patients with data on how much improvement in pain and disability they achieved after treatment with ESI impacted their decision-making regarding further treatment with ESI (aim 2). We also sought to determine the long-term risks and benefits of ESI for spinal stenosis and identify which subgroups of patients were most likely to benefit from ESI for spinal stenosis (aim 3). At the conclusion of the LESS trial and after publication of our primary results, we also conducted phone interviews and online surveys of spine specialists to determine if clinicians were aware of the trial results and, if so, whether they had changed their clinical practice because of these results. We also sought to understand barriers and facilitators to dissemination and implementation of the LESSER trial results.

Participation of Patients and Other Stakeholders in the Design and Conduct of Research and Dissemination of Findings

We conducted primary stakeholder engagement activities in the LESSER trial through collaboration with several patient advisors on the research team and development of a patient advisory group. The Patient advisory group was developed during the early part of the LESSER project (identified through their participation in the focus groups as part of aim 1) and became actively involved throughout the remainder of the study. In addition, a stakeholder advisory board, which consisted of clinicians, researchers, and health policy experts, supported the AHRQ-funded LESS trial (before this PCORI-funded project) and continued in this capacity for LESSER trial.

Stakeholder Characteristics

Patient Advisors and Patient Advisory Group

During the LESSER project development and early implementation, we partnered with 2 patient advisors, Ken Fila and Barbara Barsy, to refine the study plan related to aims 1 and 2 as well as plans for establishing a back pain patient advisory group to support the ongoing work of the study. The LESSER trial patient advisory group was composed of 9 older adults with spinal stenosis and/or back pain drawn from aim 1 focus group study participants. After completion of the focus group, 1 patient advisory group member (Mahshid Lotfi) transitioned to our primary patient partner when scheduling conflicts required our 2 original advisors to step away from involvement. At the end of each focus group, participants were asked if they were interested in remaining involved and providing additional input on the study, and we began convening quarterly patient advisory group meetings.

Stakeholder Advisory Board

The stakeholder advisory board (≥50 members) was composed of clinical and research stakeholders at the participating study sites, as well as other experts and key stakeholders in the field identified as having important perspectives (eg, Department of Labor and Industry, Center for Medical Technology Policy, Consumer's Union) in back pain research. We identified stakeholder advisory board members through existing site partnerships (eg, clinician and researcher representation from each site, typically by the site principal investigator and/or coinvestigator) and previous research collaborations (eg, Washington State Department of Labor and Industry and Center for Medical Technology Policy representatives).

Engagement Activities and Impacts

Throughout the study, we utilized a variety of methods to solicit stakeholder input, with varying intensity in levels of engagement. Our primary patient partner collaborated on all research team, patient advisory group, and stakeholder advisory board meetings and took a lead role on key study activities, most notably in the development of the patient-directed summary of LESSER research results. Early in the study development, patient advisors Ken Fila and Barbara Barsy assisted the study team in thinking through several aspects of study operations to improve participant recruitment and retention. For example, they assisted with developing recruitment strategies for the aim 1 focus groups by modifying information on the recruitment flyer to clarify who to contact and identified new locations for displaying recruitment flyers, which directly resulted in 8 additional focus group participants. Our patient partner, Mahshid Lotfi, assisted with peer reviewed manuscript writing and led the initial draft of the patient-directed summary. Through this work she identified the key points most provocative and relevant to patients and was instrumental in identifying appropriate formats and graphic elements to use in the summary. She also coled [sic] the patient advisory group feedback session that focused on the summary.

Engagement of the patient advisory group was primarily consultative in nature, consisting of quarterly meetings in which ongoing study developments and findings were shared with advisors and specific feedback was solicited on areas of the study. Patient advisors provided significant input into the individualized reports that LESS study participants were randomized to receive as part of aim 2. The input we received from patient advisors informed and improved not only the display of information but also the cover letter providing context for the reports. Patient advisor input significantly improved participant understanding and positive views of the reports.

Additionally, our patient advisors throughout the project helped clarify participant-facing materials that could potentially cause confusion and result in lower participation rates and/or understanding of findings. As we worked through communication and dissemination of study results, patient advisors identified the results most important to patients and how we might best communicate those results. For example, for the patient-focused results summary, patient advisor feedback helped us identify important messages (eg, lidocaine alone may be just as beneficial as steroid and lidocaine, without side effects of steroids, and the approach of the injection may make a difference) and understand how patients might use the information to discuss treatment options with their doctors.

Our work with the stakeholder advisory board was primarily consultative. We invited stakeholder advisors to biannual meetings to hear study updates and to provide input on key challenges. This group in particular, being composed primarily of clinicians, researchers, and policymakers, was able to help us anticipate areas where study results needed additional clarification or would be difficult to implement as changes to clinical practice. For example, 1 stakeholder advisory board meeting helped our team identify domains to further probe in clinician interviews and surveys to gauge study results' penetration into the community. These discussions also shaped our dissemination plans, as we were able to craft messaging that specifically responded to clinician and other stakeholder feedback, including better addressing subgroup analyses to justify broad inclusion criteria and allowable injection techniques, providing education on pragmatic clinical trial design hand-in-hand with study results, and including payers in conversations about implementing study results in clinical practice.

Methods

Aim 1 Methods: Determining Outcomes Most Important to Patients With Spinal Stenosis

Study Design

We designed the aim 1 study to compare commonly used patient-reported outcome (PRO) measures for pain and function (pain Numeric Rating Scale [NRS] and Roland-Morris Disability Questionnaire [RMDQ]) to outcome domains prioritized as most important by older adults with spinal stenosis. We used mixed methods combining quantitative data collected from a preassessment with qualitative data collected from structured focus groups to identify these outcomes.

Recruitment Strategy

For participation in our focus groups, we targeted recruitment of people aged 50 and older with back and/or leg pain consistent with spinal stenosis from local clinics in the University of Washington system as well as senior centers, physical therapy clinics, and patient advocacy organizations in the Greater Seattle region.

Inclusion/Exclusion Criteria

Inclusion criteria included age 50 years or older; pain or other discomfort with walking or prolonged standing that radiates into 1 or both buttocks, hips, or legs; low back pain and/or leg pain or discomfort typically relieved by getting off the feet; and low back pain and/or leg pain or discomfort typically relieved by bending forward or spinal stenosis is diagnosed by a clinician. Our main exclusion criterion was inability to understand spoken and written English.

Focus Group Methods

Once participants were deemed eligible and agreed to participate in the study, we asked them to complete a brief baseline questionnaire. We designed the questionnaire based on the LESS trial to be able to compare the baseline demographics, pain intensity, pain duration, and function of patients recruited for the focus groups with patients who participated in the LESS trial. Each focus group meeting was approximately 90 minutes in length and comprised activities including rating and ranking exercises to identify understand the back pain treatment outcomes most important to patients with spinal stenosis and facilitated discussion about outcomes of importance to patients.

In each focus group, we distributed envelopes containing slips of paper printed with 32 spinal stenosis-related problem areas or domains (eg, difficulty exercising, being a burden to others). We compiled these problem areas through consultation with patient advisors, a review of existing spinal stenosis outcomes literature, and investigator judgment. We asked participants to sort these slips into 3 piles: (1) issues that were extremely important to them, (2) issues that were somewhat important to them, and (3) issues that were not important to them. They could sort as many issues as they wanted into each pile and also add new issues not reflected in those provided. After the sorting exercise, we tabulated the responses and facilitated a conversation about why the participants considered the most highly rated domains important.

In another activity, we showed the participants a video explaining the epidural steroid injection treatment for low back pain. We gave participants the RMDQ that they completed over the telephone during recruitment and asked them to mark the items they would want to see improved related to their back pain to consider having epidural steroid injection treatment. After this activity, we tabulated the responses from the group and again discussed why certain RMDQ items were rated most highly.

Data Analysis

We combined the domain importance ratings from all the focus groups to summarize overall results. The audio recordings of the focus group discussions were transcribed, checked for accuracy against the audio recording, and then entered into a Dedoose (www.dedoose.com) qualitative database to select and code relevant participant statements. Two members of the research team (Edwards, Danielle Lavallee, PhD) selected interview text regarding reasons participants felt certain outcomes domains were important generally and for evaluating ESIs specifically. After text had been selected, 5 team members applied a list of codes based on the 32 domains that the participants sorted by importance. Two team members independently coded each transcript, and discrepancies were resolved through discussion between coder pairs.

For the second activity, we tabulated the number of times that participants endorsed each RMDQ item and the percentage for whom the symptom/impact would be important to change to consider an ESI worthwhile. Additionally, for each participant we calculated the number of RMDQ items that would need to change to consider an ESI worthwhile and plotted this against his or her total RMDQ score.

Aim 2 Methods: Individualized Decision Aids for Patients

Study Design

In aim 2, all LESS participants who had not yet reached the 18-month follow-up time period at the start of this nested RCT trial were considered eligible to participate. Eligible LESS participants who consented to participate in the RCT trial (n = 165) were randomized to receive an individualized report of their PROs either immediately before their 18-month LESS follow-up interview (intervention) or after their 24-month follow-up interview (waitlist control). We used permuted blocked randomization to randomly assign study participants to receive PRO reports at either the 18-month or 24-month follow-up.

Intervention Rationale

The final report included a graphic representation that we tailored to individuals by incorporating multiple time points (baseline, 3 and 6 weeks, 3 and 6 months) of plotted PRO data related to physical limitations and leg pain (Appendix 1); it also revealed the type and number of epidural injections individuals had received as part of their participation in the LESS trial (ie, corticosteroid and lidocaine or lidocaine alone injections). In this manner, the PRO report may have provided external validation about an individual's outcomes in relationship to his or her treatment.

Outcomes

Both groups completed a questionnaire (Appendix 2) about whether they would choose to have a future injection and what issues (up to 3) caused by spinal stenosis they most hoped the epidural injection treatment would improve. We asked whether they would choose an additional injection based on the data presented to them in the report. We defined treatment concordance as participants who reported future injection decisions consistent with their report and treatment discordance as participants who reported future injection decisions inconsistent with their report. We considered clinically meaningful improvement to be 30% or more in both RMDQ and leg pain NRS scores.19,20 We considered treatments concordant with decisions if participants who reported that they would seek an additional injection had clinically meaningful improvement or if participants who reported that they would not seek an additional injection and had <30% improvement in RMDQ and leg pain NRS scores.

At the time they received the decision aid, both groups were also asked about the helpfulness and usefulness of the decision aid, as well as their perception of their physical function and pain based on the PRO report. We captured information on utilization of physical therapy, epidural steroid injections, and spine surgeries from electronic health record data at all study sites.

Statistical Analysis

To assess questionnaire responses, we calculated frequencies (percentages) for all categorical variables and means (standard deviations) for continuous variables. To assess utilization, we used chi-square tests to compare the significance of ever and never use of each type of utilization between the intervention and waitlist control groups. We used t tests to compare the significance of the mean counts of physical therapy visits, epidural steroid injections, and morphine equivalent doses between the intervention and waitlist control groups (no patients had more than 1 lumbar spine surgery in the time periods of interest). We analyzed the usefulness and helpfulness of the decision aid and the most important outcome measures to this population using descriptive statistics. For subsequent utilization of ESIs, this study of 220 patients has adequate statistical power (>80%) to evaluate injection rate differences of 20% or greater between groups and >80% power to detect (small to medium) standardized effect sizes of 0.35 on continuous outcome measures.

Aim 3: Long-Term Impact of Epidural Corticosteroid Injections for Spinal Stenosis Symptoms

Overall Study Design

The LESS trial was a double-blind randomized controlled trial conducted at 16 US sites. We developed a formal study protocol before study initiation.21 Originally, the LESS trial was funded by AHRQ through 6 weeks of follow-up. The LESSER trial is the extension of the LESS trial through 2 years of follow-up with a nested RCT conducted during year 2 of follow-up (aim 2) and secondary analyses up to 1 year (aim 3), funded by PCORI. Here, we will describe the design and methods of the LESS trial from inception to the end of 2-year follow-up (Figure 1). Up to 4 injections were provided to participants in the first 3 months of the trial. From 3 months to 24 months, all participants received usual care.

Figure 1. Overall Study Design.

Figure 1

Overall Study Design.

Participants and Recruitment Strategy

Older adults with at least moderate pain and disability related to neurogenic claudication from central spinal stenosis who were referred for an ESI as part of their treatment plan (by either their primary care provider or a spine specialist) were invited to participate in the LESS trial by the treating physician. The site research coordinator then screened interested patients for eligibility. A study physician reviewed each potential participant's imaging studies to ensure the presence of central spinal canal stenosis and rated it as mild, moderate, or severe.

Inclusion/Exclusion Criteria

To be eligible for the LESS trial, participants had to be at least 50 years old; have magnetic resonance imaging or computerized tomography evidence of central lumbar spinal stenosis; have average low back, buttock, and/or lower extremity pain with standing, walking, and/or spinal extension in the past week rated as >4 on a 0-to-10 scale (0 = no pain to 10 = pain as bad as you can imagine); worse pain in buttock/leg than in the back; and RMDQ score ≥7 (range 0-24, with higher scores indicating greater disability). We excluded patients without central canal stenosis or who had spondylolisthesis requiring surgery, history of lumbar surgery, or epidural steroid injections within the prior 6 months. We chose these exclusion criteria based on safety concerns as well as to eliminate known confounders that might influence the results of the study or hinder participation and follow-through with study protocol.

Reasons for Declining Participation

Between April 2011 and June 2013, we screened 2224 patients, of whom 441 were eligible and 400 were randomly allocated to receive either an epidural injection with corticosteroid plus lidocaine (n = 200) or lidocaine alone (n = 200). Most people who were screened and did not participate were ineligible because of prior ESI or back surgery. Only a small percentage of people who were eligible declined to participate (n = 41), although 348 patients declined to be screened for eligibility. The primary reason why patients declined to participate was reluctance to be randomized to a lidocaine only injection and the desire to obtain an “active” treatment. We designed this study as a comparative effectiveness trial with lidocaine only injections as an active comparator, but some patients worried about not obtaining a treatment that would alleviate their pain and declined participation.

Study Setting and Reasons for Choosing the Sites

The LESS trial was funded as part of an AHRQ-sponsored research grant to study older adults with back pain. This grant, led by principal investigator Dr Jerry Jarvik, was called the Back pain Outcomes using Longitudinal Data (BOLD) project. The BOLD project involved 2 separate projects: (1) building a registry designed to capture data from patients enrolled in 3 large integrated health care systems in the United States22,23 and (2) conducting a randomized clinical trial of epidural steroid injections for spinal stenosis (LESS). We conducted these separate projects using infrastructure built within 3 integrated health care systems (Kaiser Permanente Northern California; Henry Ford Hospital, Detroit, MI; and Harvard Vanguard/Brigham and Women's Hospital, Boston, MA). We recruited most patients enrolled in the LESS trial from these integrated health care systems. To meet our recruitment goals, we also recruited patients from 11 clinical sites across the United States representing academic medical centers, a VA hospital, and a private practice. We chose these sites based on the expertise of the physicians, availability of patients, and commitment to participation, and to represent a range of clinical practices, specialties (eg, physical medicine and rehabilitation, anesthesiology and interventional radiology), geographical regions, and patient populations. We followed patients enrolled in the LESS trial (n = 400) for 2 years. Given the extension of the trial follow-up beyond the original 6 weeks funded through AHRQ, we designated the long-term follow-up of the LESS trial as LESSER (aim 3). A subset of these original LESS patients also participated in a nested RCT between year 1 and year 2 of follow-up (aim 2).

Randomization and Blinding

The data coordinating center generated electronically concealed permuted-block randomized assignments for each recruitment site. Two opaque syringes were prefilled for each procedure, 1 with corticosteroid plus lidocaine and the other with lidocaine. The randomization assignment was obtained at the time of the procedure via a password-protected study website by the clinical assistant; this person was not involved with subsequent data collection. The assignment indicated which syringe should be labeled “inject” and which labeled “discard.” The assistant confirmed use of the syringe marked “inject” by the physician. The treating physicians, patients, and research staff who conducted follow-up were blinded to the treatment received. We used the Bang blinding index24 to evaluate the blinding effectiveness within each treatment group, where 0 indicates perfect blinding and +1 indicates perfect treatment guessing.

Interventions and Rationale for Choosing Them

Twenty-six board-certified anesthesiologists, physiatrists, and radiologists with expertise in epidural steroid injections performed the injections. We chose to compare epidural corticosteroid and lidocaine injections to lidocaine alone injections as we were most interested in the specific effects of the corticosteroid medication, commonly believed to be the active agent in epidural injections. LESS investigators trained study physicians to perform the procedures in a standardized manner using fluoroscopic guidance.21 Physicians were instructed to choose the injection level (eg, L5-S1) 1 spinal level below the maximal canal stenosis for the interlaminar injections and at the root level of the greatest symptoms for transforaminal injections, although bilateral and multilevel transforaminal injections were allowed. Consistent with usual practice, physicians could perform multiple injections (using a single syringe) based on the location of the patient's symptoms (ie, bilateral transforaminal injections or 2 separate interlaminar injections performed at 2 spinal levels). For the purposes of analysis, we considered injections performed at the same time a single injection. The physician chose the approach (transforaminal or interlaminar) and corticosteroid based on his or her usual practice, which remained consistent for subsequent injections for each patient. The corticosteroid injectate consisted of 1 to 3 ml of 0.25% to 1% lidocaine followed by 1 to 3 ml of triamcinolone (60-120 mg), betamethasone (6-12 mg), dexamethasone (8-10 mg), or methylprednisolone (60-120 mg). The lidocaine procedure was identical to the corticosteroid plus lidocaine injection except that the injectate was an equivalent volume of 0.25% to 1% lidocaine alone. The 2 syringes drawn for each procedure (only 1 was used based on randomization assignment) contained identical volumes. Therefore, the volume of the lidocaine alone injectate was identical to the volume of corticosteroid and lidocaine injectate for each individual patient.

Follow-up and Outcomes

Our primary outcomes were measures of function and pain. These included the RMDQ, a back pain-specific functional status questionnaire adapted from the generic Sickness Impact Profile.25 It consists of 24 yes/no items, which represent common dysfunctions in daily activities experienced by patients with low back pain. The other primary outcome measure was pain recorded on an 11-point NRS measuring average leg and back pain (separately) in the past week. We completed the primary outcome measures at 3 days, 14 days, 3 weeks, 6 weeks (primary outcome time point), 3 months, 6 to 12 months, and 18 to 24 months postrandomization.

We also determined as secondary outcomes the proportion of patients with moderate (defined as 30% or greater) or substantial (50% or greater) clinically meaningful improvement on each measure.19,26,27 Other secondary outcomes included a rating of average back pain in the past week (0-10 scale) and scores on the Brief Pain Inventory interference scale (0-10 scale, with higher scores indicating more activity interference attributable to pain)28; the Patient Health Questionnaire-8 (0-24 scale, with higher score indicating more depressive symptoms)29; the Generalized Anxiety Disorder-7 scale (0-21 scale, with higher scores indicating more-severe anxiety)30; the EuroQol 5D (EQ-5D; 0-1.0 scale, with lower scores indicating worse quality of life)31; and the Swiss Spinal Stenosis Questionnaire (SSSQ),32 comprising symptoms (1-5, with higher scores indicating worse symptoms), physical function (1-4, with higher scores indicating worse function), and treatment satisfaction (1-4, with higher scores indicating worse satisfaction) subscales. We collected secondary outcome measures at 3 weeks, 6 weeks, 3 months, 6 months, 12 months, 18 months, and 24 months.

At baseline and at 3 weeks and 6 weeks postinjection, we drew blood to measure morning blood cortisol levels to monitor for adrenal suppression following steroid administration. Because patients were blinded to treatment assignment, we checked these levels in all patients regardless of treatment received.

Other measures included patient-completed resource utilization questionnaires to collect information on medication use and use outside the health plan during four 3-week periods through 12 months. Information on use within each health system was available through the electronic health information systems (pharmacy and medical use).

Aim 3 Statistical Analysis

For the 12-month outcome analyses, we used an intention-to-treat (ITT) strategy, with participants analyzed according to their randomly assigned treatment, ignoring any treatment arm crossovers. We calculated treatment effects and confidence intervals from analysis of covariance (ANCOVA) models adjusting for baseline values of the outcome measure and recruitment site, and using a treatment indicator as the predictor of interest. We also adjusted for pain duration at baseline as duration could be associated with outcomes and was longer in the corticosteroid plus lidocaine group (P = .02). We used similar models to evaluate the results separately by injection approach subgroups. We evaluated ITT comparisons of the proportion of participants reporting a 50% or greater improvement in RMDQ score or leg pain intensity with logistic regression models, also adjusting for baseline pain duration, recruitment site, and baseline value of the outcome measure.

We used a chi-square test to assess the rate crossed over to the alternate treatment at 6 weeks between ITT groups. Because participants could cross over and receive the alternate treatment at 6 weeks, the analysis sample for the 1-year ITT estimates of treatment effect included a mixture of participants who received only the original assigned treatment and who received both treatments. Thus, we defined 2 mutually exclusive as-treated groups based on the treatment decision at the 6-week study follow-up: crossed over and adhered to randomized treatment (for which patients may have received a repeated baseline injection, or received no additional injection). Due to the potential for differential crossover patterns between randomized groups and potential biases inherent with treatment self-selection, we used univariate logistic regression models to identify and describe pretreatment demographic factors, clinical characteristics, pretreatment and posttreatment outcome measures, and treatment guesses at 14 days that were associated with patient crossover to the alternate treatment. We adjusted for recruitment site, pain duration, the baseline values of the outcome measure, and any pretreatment variables found to predict crossover (P < .05) independently of the randomized treatment assignment or differentially by treatment assignment (statistical interaction between the variable and randomization assignment).

To identify patient characteristics associated with greater benefits from epidural injections of corticosteroid plus lidocaine than from epidural injections of lidocaine only (ie, corticosteroid effect modifiers), as well as with benefits from either corticosteroid plus lidocaine or lidocaine only (ie, nonspecific prognostic factors), we included as candidate predictors variables previously identified as prognostic of outcomes from surgery for lumbar spinal stenosis or of outcomes across different treatments for back pain. Along with the baseline values of the outcome measures, we selected the following measures for study inclusion as candidate predictors: sociodemographic factors, pain duration, psychosocial variables, health-related variables, and qualitative and quantitative imaging measurements of spinal stenosis severity. We included continuous or ordinal predictor variables on the original scale and report the treatment effect by predictor interaction coefficients on a change-per-unit increase in the predictor as well as the estimated treatment effect at the 25th and 75th percentile of the baseline predictor variable. We examined nonspecific predictors of improvement (ie, patient baseline characteristics that predicted response in both treatment groups, with no significant treatment by baseline characteristic interaction effect) on the outcome measures with a likelihood ratio test using the ANCOVA model with each predictor variable of interest included as a main effect only (without an interaction with treatment).

We based our primary statistical approach to determine effect of epidural corticosteroid injection on cortisol levels on longitudinal analysis of baseline and 3-week measures using generalized estimating equations (GEEs)33 with robust standard errors adjusting for site, an indicator of the randomized treatment assignment, baseline cortisol level, and number of days from baseline to the exact day of follow-up since this did vary based on protocol follow-up windows. Group comparisons of cortisol levels at 3 weeks used available data from all 400 study participants. To examine predictors of cortisol suppression, in separate GEE models we included each predictor and tested the resulting model parameter(s) with a Wald test and robust standard errors.

For the primary outcome of the LESS trial at 6 weeks, we used a Bonferroni-corrected significance threshold of P < .05/2 = .025 in analyses within the injection approach subgroup as this was a prespecified subgroup. All other reported P values are 2 sided and considered significant at <.05. Analyses through 12 months and all secondary analyses were not adjusted for multiple comparisons and should be considered hypothesis generating. We made careful efforts to minimize missing data by requiring double entry of primary outcomes and using automated warning messages for missing data. When individual items were missing from a scale, we calculated the percentage of missing items. If <10%, we imputed values using the mean of the remaining items. If >10%, we considered the scale score missing and unavailable for analysis. Study statisticians (PJH, BAC, and XS) were responsible for statistical design considerations and analyses and used R statistical software (version 3.1.1).

Results

Aim 1: Determining Outcomes Most Important to Patients With Spinal Stenosis

We conducted 7 participant focus groups with a total of 33 participants. Focus group participant characteristics are presented in Table 1. One of our main objectives for this study was to determine if the outcome measures used in the LESS trial (RMDQ, SSSQ, and pain intensity) were reflective of what is important to patients with spinal stenosis. We attempted to recruit a variety of patients for the focus groups who were similar to patients enrolled in the LESS trial. Although we did not compare the focus group participants with the LESS trial participants statistically, a higher percentage of focus group participants had college-level or higher education, were retired, and had a longer duration of pain. We also had fewer African American participants in the focus groups than in the LESS trial, which is reflective of the demographics in the greater Seattle area compared with the LESS trial recruitment locations (our largest recruiting site for the LESS trial is Henry Ford Health System in Detroit, where 50% of the patients enrolled were African American). Also, although the groups were well balanced in terms of baseline levels of disability as measured by the RMDQ, participants in the focus group reported less pain than did those enrolled in the LESS trial. That difference may be explained by the different recruitment approaches used for recruitment into the focus groups vs the LESS trial. The LESS participants were recruited at the time of their first scheduled ESI injection, so they likely experienced more-severe symptoms at the time of enrollment compared with the focus group participants, who were identified through the community and potentially may have received treatment before they were recruited.

Table 1. PCORI LESSER Focus Group Participant Characteristics.

Table 1

PCORI LESSER Focus Group Participant Characteristics.

The domain participants most commonly rated as extremely important was experiencing pain/discomfort, rated as extremely important by 88% of participants overall. This domain was followed by problems with physical function (85%), difficulty exercising (73%), difficulty participating in hobbies and leisure activities (55%), and problems with weakness (52%). Participants added only 1 domain: fear of pain coming at the wrong time. In contrast, participants considered 8 domains not important. The domain participants most commonly rated as not important was problems with alcohol or other drug use, selected by 94% of participants overall. This domain was followed by difficulty living independently (73%), difficulty taking care of family members (70%), problems with anger (67%), problems with use of pain medication (61%), difficulty concentrating and remembering (58%), problems with anxiety and worry (55%), and difficulty with self-care (52%).

In the second activity, participants rated which items that they had endorsed on the RMDQ at baseline would need to improve for them to consider ESI treatment again if they had had it previously. Four RMDQ items were endorsed by at least 50% of participants, and at least 50% of those who endorsed them thought that improvement in the item would warrant ESI treatment: only walking short distances because of back or leg pain (n = 17), walking more slowly than usual because of back or leg pain (n = 3), only standing for short periods of time because of back or leg pain (n = 10), and sleeping less well because of back or leg pain (n = 18). When we plotted participants' RMDQ scores against the percentage reduction in RMDQ scores they said needed to improve to make ESI treatment worthwhile, it suggested that on average a 40.5% improvement in RMDQ score would be necessary to warrant ESI treatment according to these participants (Figure 2).

Figure 2. Percentage Reduction in RMDQ Score Required for ESI Treatment to Be Considered Worthwhile vs RMDQ Score at Screening.

Figure 2

Percentage Reduction in RMDQ Score Required for ESI Treatment to Be Considered Worthwhile vs RMDQ Score at Screening.

Aim 2: Individualized Decision Aids for Patients

For the nested RCT, a subset of the original LESS patients were randomized to receive an individualized report—at either 18 months (intervention group) or 24 months (control group) after treatment—depicting their pain and function outcomes over the first 12 months of the trial in relation to the injections they received. Of these participants, 65% indicated at the time they reviewed the individualized report that it was moderately to extremely helpful in decision-making about whether they would undergo a subsequent injection, compared with about 20% who found it only a little helpful or not at all helpful (Table 2). Of participants, 62% indicated the individualized report was somewhat to extremely easy to understand. In contrast, about 10% of participants indicated that the individualized report was somewhat difficult to extremely difficult to understand; about 10% (16 participants) needed assistance to understand the report (Table 3).

Table 2. Helpfulness of the Individualized Report Regarding Epidural Injection Treatment.

Table 2

Helpfulness of the Individualized Report Regarding Epidural Injection Treatment.

Table 3. Understandability of the Individualized Report Regarding Epidural Injection Treatment.

Table 3

Understandability of the Individualized Report Regarding Epidural Injection Treatment.

For participants who had clinically meaningful improvement after treatment, the control group had more concordant decision-making (ie, their decision about whether to get a subsequent epidural steroid injection was consistent with whether they achieved a clinically meaningful benefit after their previous injection or injections (Table 4) (18%) compared with the intervention group (11%) (Table 4). In contrast, for participants who had little or no benefit from the injections, the intervention group had more concordant decision-making (30%) compared with the control group (17%) (Table 4).

Table 4. Outcome–Treatment Decision Concordance.

Table 4

Outcome–Treatment Decision Concordance.

There were no statistically significant differences in pain outcomes (RMDQ, leg pain, back pain) between the intervention and control groups at the 24-month follow-up (Table 5). There were no statistically significant differences in health care utilization in the 18 to 24 months after initial LESS trial randomization/injection between the intervention and control groups (Table 6).

Table 5. Relationship Between the Individualized Decision Aid and Patient-Reported Outcomes.

Table 5

Relationship Between the Individualized Decision Aid and Patient-Reported Outcomes.

Table 6. Health Care Use of LESS Patients at Follow-up (18-24 Months After Index LESS Visit).

Table 6

Health Care Use of LESS Patients at Follow-up (18-24 Months After Index LESS Visit).

Aim 3: Long-Term Impact of Epidural Corticosteroid Injections for Spinal Stenosis

ITT Results

As part of the LESS trial, between April 2011 and June 2013, we screened 2224 participants, of which 400 were eligible and randomized to receive either an epidural injection of corticosteroid plus lidocaine (n = 200) or an epidural injection of lidocaine alone (n = 200). At 12 months, 87% (174 of 200) and 90% (180 of 200) of randomized participants completed the primary outcome assessment in the lidocaine alone and corticosteroid plus lidocaine groups, respectively (Figure 3).

Figure 3. LESS Trial Enrollment and Follow-up.

Figure 3

LESS Trial Enrollment and Follow-up.

Both treatment groups maintained the improvements in outcomes observed at 6 weeks through 12 months (although the small differences observed between the 2 randomized groups at 3 and 6 weeks were no longer apparent) (Table 7, Figure 4). In addition, the small difference in improvement observed in the interlaminar approach subgroup at 3 weeks was also no longer apparent by 12 months (Figure 4). At 12 months, there were no statistically significant differences between the corticosteroid plus lidocaine and lidocaine alone groups on the RMDQ (adjusted mean difference = −0.4; 95% CI, −1.6 to 0.9; P = .55) or leg pain intensity (adjusted mean difference = 0.1; 95% CI, −0.5 to 0.7; P = .75) (Table 7).

Table 7. RMDQ Score and Leg Pain Intensity and Change From Baseline at Each Study Time Point by Randomized Treatment Group.

Table 7

RMDQ Score and Leg Pain Intensity and Change From Baseline at Each Study Time Point by Randomized Treatment Group.

Figure 4. RMDQ Score and Leg Pain Intensity by ITT Treatment Group and by Approach.

Figure 4

RMDQ Score and Leg Pain Intensity by ITT Treatment Group and by Approach.

At 6 weeks, among participants randomized to lidocaine alone, 45% (n = 90) crossed over and received an injection of corticosteroids plus lidocaine, 10.5% (n = 21) repeated the lidocaine alone injection, and 44.5% (n = 89) had no additional injection (Figure 3). Among participants randomized to corticosteroid plus lidocaine, 30% (n = 60) crossed over to an injection of lidocaine alone, 14.0% (n = 28) repeated the injection of corticosteroid plus lidocaine, and 56.0% (n = 112) had no additional injection. The difference in crossover between those initially randomized to corticosteroid plus lidocaine and those randomized to lidocaine alone was statistically significant (P = .003); however, a single recruitment site was responsible for the differential crossover rate by arm. At this site, the crossover rate from corticosteroid plus lidocaine to lidocaine alone was 15% (9 of 60) and the crossover rate from lidocaine alone to corticosteroid plus lidocaine was 50% (31 of 62). At the other recruitment sites the crossover rates were similar by randomization group (40.1% for lidocaine alone vs 36.4% for corticosteroid plus lidocaine; P > .05).

Crossover Results

Table 8 shows the mean change in RMDQ score and leg pain from 0 to 6 weeks and from 6 to 12 weeks by randomized treatment group and by the treatment decision at 6 weeks. At 6 weeks, participants in both groups who elected to receive no additional study injection showed the greatest improvements in RMDQ score and leg pain. Participants who crossed over showed the least improvement and those who repeated the baseline injection showed levels of improvement in between those of the other 2 groups (Figure 5). Among participants who crossed over at 6 weeks, the 6- to 12-week change in RMDQ score did not differ between the 2 randomized treatment groups, with an adjusted mean difference between groups of similar magnitude (Table 8) (adjusted mean difference = −1.0; 95% CI, −2.6 to 0.7; P = .24) to that observed among participants who received a corticosteroid injection at baseline (Table 6) (0- to 6-week ITT analysis adjusted mean difference = −1.2; 95% CI, −2.3 to − 0.1; P = .03). We observed similar results for leg pain. In both randomization groups, participants who crossed over at 6 weeks continued to have worse long-term outcome trajectories compared with participants who repeated the randomized injection or had no additional study injections at 6 weeks (Figure 5).

Table 8. Change in RMDQ and Leg Pain Outcomes by Randomized Treatment Group and 6-Week Treatment Decision.

Table 8

Change in RMDQ and Leg Pain Outcomes by Randomized Treatment Group and 6-Week Treatment Decision.

Figure 5. RMDQ Score and Leg Pain Intensity by As-treated Group Status.

Figure 5

RMDQ Score and Leg Pain Intensity by As-treated Group Status.

Significant predictors of crossover at 6 weeks included the 6-week RMDQ score, 6-week leg pain rating, and 0- to 6-week change from baseline in RMDQ score and leg pain rating, and treatment guess responses at 2 and 5 weeks (Table 9). A treatment guess of lidocaine at 5 weeks predicted the decision to cross over (P < .001) regardless of initial treatment assignment. No other demographic or clinical characteristics were associated with crossover, and only recruitment site predicted crossover differentially by randomization group. Most participants (87%) reported their reason for crossing over as inadequate pain relief from the assigned treatment.

Table 9. Predictors of Crossover Overall and Differentially Between Treatment Groups.

Table 9

Predictors of Crossover Overall and Differentially Between Treatment Groups.

Concomitant Back-Related Use and Treatments

We obtained study injection counts for all participants. Complete electronic medical record data were available for 209 participants from sites within integrated health systems, whereas we collected utilization diary data from 145 of the remaining 191 participants without available electronic medical record data. The mean number of injections did not differ significantly between randomized groups during any of the follow-up periods. However, the number of injections differed between randomized treatment groups between 6 and 12 weeks due to higher crossover in the lidocaine alone arm (P < .001) and from 12 weeks through 12 months due to more patients receiving additional injections in the corticosteroid plus lidocaine arm (P = .02) (Table 10). Crossover participants, regardless of treatment group, were more likely to receive a subsequent repeat epidural injection after the initial 6-week injection (52.1%) than were participants who did not cross over (37.7%; P = .007).

Table 10. Injections by Randomized Treatment Group.

Table 10

Injections by Randomized Treatment Group.

At baseline and during the first 6 weeks, more participants in the lidocaine alone arm reported taking some opioid medication (48.1% vs 38.0%; P = .06). At 12 months, there were no statistically significant differences between treatment groups in reported use of opioids (36.3% vs 41.4%; P = .41), physical therapy (7.5% vs 9.6%; P = .55), or spine surgery (11.8% vs 16.8%; P = .22).

Predictors of Response to Epidural Injections

Among multiple baseline variables examined, only 1 predicted specific response to corticosteroid at 3 or 6 weeks. Compared with patients who rated their health-related quality of life as high on the EQ-5D Index, patients who rated it as poor benefited more from corticosteroid plus lidocaine than from lidocaine only in terms of leg pain intensity at 6 weeks (interaction coefficient = 2.94; 95% CI, 0.11-5.76; P = .04), but not 3 weeks, and RMDQ disability scores at 3 weeks (interaction coefficient = 4.77; 95% CI, −0.04 to 9.59; P = .05), but not 6 weeks. An interpretation of these estimates is that relative to patients with a high quality of life, patients with the lowest quality of life would experience more improvement in leg pain and disability being treated with corticosteroid plus lidocaine than with lidocaine only (eg, 2.94 points greater improvement in leg pain at 6 weeks and 4.77 points greater improvement in disability at 3 weeks).

Only 2 baseline variables predicted 6-week RMDQ scores independent of treatment assignment. The RMDQ scores (adjusted for recruitment site, treatment group, and baseline RMDQ score) differed significantly across baseline employment status categories (P = .01) and pain duration categories (P = .03; Appendix 4). Compared with patients who were working full time or part time, patients who were retired due to ill health or disability had a mean adjusted RMDQ score at 6 weeks that was 3.2 points (95% CI, 1.4-5.0) higher (worse) and those who were retired for reasons other than disability had a mean adjusted RMDQ score 2.0 points higher (95% CI, 0.8-3.3) (Appendix 3). Compared with patients with pain duration of <3 months, those with pain duration of 3 to 12 months had slightly lower (better) adjusted RMDQ scores (0.7 points), those with pain duration of 1 to 5 years had higher (1.6 points) scores, and those with pain duration of greater than 5 years had slightly higher (0.6 points) scores (Table 10). The Patient Health Questionnaire-8 (P = .02) and the Generalized Anxiety Disorder-7 (P = .01) were significant nonspecific predictors of RMDQ scores at 3 weeks (with greater baseline depression and anxiety associated with greater 3-week disability in both treatment groups), but not 6 weeks (Table 10). At 6 weeks, compared with Caucasians, non-Caucasians had buttock/hip/leg pain ratings 1.2 points lower on average (95% CI, 1.9-0.5), adjusting for treatment, recruitment site, and baseline buttock/hip/leg pain rating (Appendix 5). Lower (worse) baseline EQ-5D VAS health state ratings were associated with higher buttock/hip/leg pain ratings at 6 weeks (P = .03), and higher baseline depression scores were associated with higher buttock/hip/leg pain ratings at 3 weeks (P = .04; Appendix 6).

Treatment-Related Change in Cortisol Suppression

Among the 400 LESS patients randomized, 372 had a pretreatment baseline cortisol measurement and were included in the present analysis. Of these 372 patients, 83% (n = 307) and 78% (n = 292) had a 3-week or 6-week follow-up cortisol measurement, respectively, with 71% (n = 264) having cortisol measurements at both study follow-up time points. Before treatment, patients in the lidocaine only and corticosteroid plus lidocaine groups had an average cortisol level of 11.5 (SD, 4.9) and 12.0 (SD, 7.1), respectively (P = .433). At the 3-week follow-up study time point, cortisol levels of patients treated with corticosteroid plus lidocaine were 2.3 μg/dl lower on average than patients treated with lidocaine only (95% CI, −3.3 to −1.2; P = .003; Appendix 7).

Patients treated with corticosteroid plus lidocaine on average experienced a 14.4% reduction in cortisol at week 3 compared with baseline, whereas patients treated with lidocaine had only an 8.2% increase in cortisol (P = .002). Thirty-two patients (20.3%) treated with corticosteroid plus lidocaine experienced a reduction in cortisol at 3 weeks of 50% or larger compared with just 10 patients (6.7%) treated with lidocaine only (P = .002). Of patients treated with corticosteroid, 12% continued to have >50% cortisol suppression at 6 weeks compared with 4% of patients treated with lidocaine alone, which represents a 3.6-fold increased risk of >50% cortisol suppression (95% CI, 1.6-7.9; P = .002). Five patients (3.2%) in the corticosteroid plus lidocaine group experienced nearly complete cortisol suppression (≥90%) at the 3-week study assessment, whereas we observed no such patients in the lidocaine only arm (P = .061).

In Figure 6, we present summary boxplots of the 3-week percentage change in cortisol by study group and corticosteroid type. Corticosteroid type significantly moderated the observed difference in 3-week cortisol change by treatment group (P < .001). Patients treated with methylprednisolone and triamcinolone had an average 3-week cortisol reduction of 41.0% (P = .005) and 41.6% (P < .001) from baseline, respectively, compared with similar patients treated with lidocaine only, whereas patients treated with betamethasone or dexamethasone were not significantly different from comparable patients in the lidocaine arm.

Figure 6. Cortisol Suppression by Type of Corticosteroid Administered.

Figure 6

Cortisol Suppression by Type of Corticosteroid Administered.

Predictors of Cortisol Suppression

The average pretreatment cortisol levels for baseline demographic, clinical, and PRO variables are presented in Appendix 8. Baseline cortisol did not differ on prespecified PRO measures of depression, anxiety, or leg pain NRS, nor did they differ significantly across other patient characteristics (age, race, smoking history, diabetes, or spinal stenosis severity). We found baseline cortisol to be inversely associated with body mass index, with an associated average reduction in cortisol of 0.12 μg/dl for each unit increase in body mass index (P = .020). To identify whether these patient characteristics yielded differential 3-week cortisol response when treated with steroids, in separate statistical models we evaluated each predictor with a Wald test for interaction with treatment. Patients with diabetes tended to have a greater amount of cortisol suppression (adjusted treatment effect: 36.2%) when treated with corticosteroids vs lidocaine only than nondiabetics (adjusted treatment effect: 13.3%), although the statistical significance (P = .085) did not reach the nominal level of .05 (Appendix 9). The effect of treatment with steroids on 3-week cortisol changes did not otherwise appear to differ by demographic or patient-level characteristics.

Cortisol Suppression and Adverse Events

There was no clear association between cortisol suppression and adverse event rates. Seven patients with cortisol suppression >50% reported adverse events, but only 1 was potentially related to immune suppression related to cortisol suppression (pneumonia) (Appendix 10).

Discussion

The LESSER trial is the largest clinical trial to date of ESI for any low back pain condition and provides high-quality evidence that has direct implications for clinical care of patients with lumbar spinal stenosis. This trial has filled a substantial evidence gap about short- and long-term effectiveness of these injections for spinal stenosis, heterogeneity of treatment effect (ie, which patient and procedural factors are associated with outcomes), and safety of this commonly used treatment. Further, this study expands the evidence base on incorporating PROs in clinical care.

Outcomes of Importance to Patients With Spinal Stenosis

One of our primary aims was to determine if the outcomes we measured in the LESS trial reflect those most important to older adults with spinal stenosis. Through our focus groups, we identified that the outcomes we measured in the LESS trial for the most part do reflect outcomes of importance to patients. Not surprisingly, pain permeated the discussions of participants' experiences with spinal stenosis. The principal driver behind functional, social, and activity limitations participants with lumbar spinal stenosis experience is pain. Pain appears to be more important than other symptoms of spinal stenosis, such as weakness, numbness, and tingling in the legs. Patients identified as important 1 outcome, difficulty exercising, that was not measured in the LESS trial. Focus group participants noted a commonly experienced challenge and frustration—the recommendation by their physician to exercise to maintain back strength and health yet the inability to exercise due to the pain resulting from their condition, out-of-pocket costs to participate in safe forms of exercise such as yoga or water aerobics, and lack of availability and knowledge about how to safely participate in exercise. This is an important finding for health care providers to recognize in providing patient-centered care.

Although the outcome measures used in the LESS/LESSER trial reflect domains of importance to patients, some uncertainties remain about how to use PROs to improve decision-making about treatment options. In our aim 2 study, we provided LESS trial participants with feedback about how they reported their change in pain and function after receiving epidural injections. Although they found these data understandable and helpful, the impact of these reports on their future decision-making and outcomes was less clear. For participants who had clinically meaningful improvement after receiving epidural injections, the control group (which did not receive the individualized report until after the study conclusion) had more concordant decision-making (ie, participants' decision about whether to get a subsequent epidural steroid injection was consistent with whether they achieved a clinically meaningful benefit after their previous injection or injection) compared with the intervention group. In contrast, for participants who had little or no benefit from the injections, those who received the individualized report 6 months before the study conclusion had more concordant decision-making compared with the control group. These results likely reflect the complexity of graphically representing PROs. Existing work has well documented this challenge,34 especially for older adults.35

Among participants with lumbar spinal stenosis enrolled in a randomized trial comparing epidural injections of corticosteroid plus lidocaine vs lidocaine alone, participants in both groups demonstrated modest improvements in pain-related disability and leg pain at 6 weeks and through 12 months after the initial injection. These findings are consistent with prior observational studies of the natural history of spinal stenosis and with trials of surgical management of spinal stenosis.36-38 ITT analyses demonstrated a consistent but small (∼1 point on the RMDQ) difference favoring the corticosteroid plus lidocaine group at 3 and 6 weeks, but this small difference did not persist beyond 6 weeks. In addition, this small benefit was seen only in patients who received interlaminar approach injections.

Significantly more participants assigned to lidocaine crossed over to the alternate treatment at 6 weeks, but we observed this differential crossover rate at only 1 recruitment site. At this site, only interlaminar injections with either triamcinolone or betamethasone were performed, patients tended to have higher baseline levels of disability and pain, and no participants were offered repeat injections at 3 weeks. It is possible that performing interlaminar injections with these particular corticosteroids or other site-specific differences in clinical practice contributed to this differential crossover rate. Among all participants initially treated with lidocaine who crossed over at 6 weeks, we observed a consistent but not statistically significant 1-point benefit on the RMDQ at 12 weeks compared with patients initially treated with corticosteroids who crossed over at 6 weeks. There was no difference in 12-month pain or function between those who crossed over to corticosteroid injections and those who crossed over to lidocaine alone injections. Most participants who crossed over (87%) reported their reason as inadequate pain relief. The only significant predictors of crossover (other than site) were poor pain and function outcomes at 6 weeks and a treatment guess of lidocaine at 2 and 5 weeks (assessed using the Bang blinding index), regardless of what type of injection was actually received. Participants with poor response to their original injection (in either treatment group) were more likely to cross over and less likely to improve after crossover. Thus, it appears that the decision to cross over reflected perceived lack of efficacy of the first treatment, but that repeated injections of any type were not beneficial.

Participants who did not cross over at 6 weeks improved more between baseline and 6 weeks than those who did cross over, regardless of randomization assignment, and continued to show better outcomes through 12 months. We were unable to identify any baseline patient characteristics or imaging findings that predicted differential response to the corticosteroid vs the lidocaine alone injections. However, patients with baseline depression, anxiety, fear avoidance, and catastrophizing had worse outcomes regardless of initial treatment assignment.

We evaluated safety of epidural corticosteroid injections in the LESS trial by the risk of cortisol suppression (which results in the inability of the body to produce stress hormones necessary during periods of physiologic stress and can lead to infections as well as a wide variety of systemic symptoms). Consistent with prior smaller case series, we found that 9.5% of patients receiving ESI had nearly complete cortisol suppression at 3 weeks. We did not identify any patient characteristics that predicted cortisol suppression; however, the type of corticosteroid used (methylprednisolone and triamcinolone) was a significant predictor of cortisol suppression. Although our current study was not adequately powered to fully assess for a correlation between cortisol suppression and adverse event reporting, we did not identify a relationship between self-reported adverse events and cortisol suppression. Cortisol suppression is often clinically asymptomatic or presents with mild, vague symptoms such as nausea, muscle aches, abdominal pain, and depression, which are often experienced by people with chronic pain. Adrenal insufficiency therefore can go undiagnosed until a rare adrenal crisis occurs, often in the setting of significant stress (eg, surgery or major illness).39 Further research is needed to fully understand the clinical impact of cortisol suppression and to determine how to best monitor for cortisol suppression after epidural injections, particularly in high-risk patients such as those with comorbid conditions on chronic oral corticosteroids and/or who have received recent corticosteroid injections.

Subpopulation Considerations

Although the LESSER trial did not identify significant improvements in patient-reported measures of pain or function for patients undergoing epidural injections of corticosteroid and lidocaine compared with lidocaine alone, there was some improvement on average on these measures in both injection groups, with variability across patients. Therefore, we attempted to identify specific patient characteristics that might predict a beneficial response to an ESI. Our findings do not support the existence of a specific subgroup of patients with lumbar spinal stenosis that is particularly responsive to epidural injections of corticosteroid and lidocaine vs lidocaine alone. Although we examined 21 patient characteristics identified in previous research as potentially predictive of treatment response, including sociodemographic characteristics, spinal canal stenosis severity as judged by the clinician based on MRI or CT findings, and psychological factors, as well as 6 outcomes measured at 2 time points, no baseline characteristic emerged as a clear and strong predictor of a benefit from epidural injection of corticosteroid and lidocaine vs lidocaine alone.

Patients in both treatment groups showed improvements in pain and function. Potential explanations for comparable improvements in both groups include placebo effects, regression to the mean, natural history of spinal stenosis, and other factors present in both study groups, including contact with study personnel or lidocaine. Although lidocaine has been proposed as an active treatment,40-42 any sustained benefit from epidural lidocaine injections remains theoretical. Studies comparing epidural lidocaine to sham injection or other treatments are needed to draw conclusions regarding the effectiveness of lidocaine.

Implementation of Study Results

Since initial publication of the LESS trial results in the New England Journal of Medicine, subsequent papers, and the AHRQ Eisenberg Center clinician summary,43-45 we conducted interviews with LESSER trial investigators and surveys with more than 200 physicians attending the American Academy of Physical Medicine and Rehabilitation and North American Spine Society annual meetings to determine how familiar spine providers are with the LESSER trial results, to understand if providers have changed clinical practice because of the LESSER trial, and to identify barriers and facilitators to incorporating the results of the trial into clinical practice. A consistent finding was that physicians felt that LESS findings need to be distilled and contextualized in light of other evidence about ESI (eg, ESI evidence for other spine conditions). Most physicians surveyed also mentioned referral patterns as a key driver of ESI use and, as such, a barrier to practice change. Many expressed the challenge of managing patient expectations and satisfaction after a referring provider had recommended ESI. The physicians identified the need for tools (eg, decision support tree for referring providers, fact sheet summarizing related studies, patient education materials).

We also engaged our patient advisory group in multiple in-depth discussions about the results of the LESSER trial and how to best educate patients about these results. The group's feedback highlighted some of the challenges that patients face in terms of making evidence-based decisions about use of epidural injections. Group members cited as important barriers (1) their lack of knowledge about the range of treatment options and the comparative effectiveness of these options (including different techniques of injections as well as physical therapy, surgery, and other treatments for spinal stenosis) and (2) difficulties being able to openly discuss treatment options and effectiveness evidence with their providers without being labeled as a “bad” patient or disrespecting the provider and impacting their relationship.

Taken together, we identified from our interviews and stakeholder engagement activities that a multipronged intervention for patients, referring providers, and interventional providers that focuses on improving knowledge about the evidence and providing skills for engaging in shared decision-making may help disseminate and implement the LESSER study findings into clinical practice. We are developing as part of a future grant proposal further dissemination efforts focusing on better educating primary care physicians and patients about the LESS/LESSER trial and how to incorporate their findings into clinical decision-making.

Decisional Context

The LESSER trial results provide clarity about the effectiveness of epidural corticosteroid injections for patients with neurogenic claudication from spinal stenosis and fill an important void in the existing evidence. While the LESSER trial provides valuable evidence about the use of epidural steroid injections, the data are complex and can be applied in a number of different but appropriate ways depending on perspective. Providers (including primary care providers and spine or interventional pain specialists), patients, and payers all have different perspectives about the LESSER trial results and how they should be best applied to clinical practice. For example, patients and providers may choose to reduce, eliminate, or replace the use of epidural injections for spinal stenosis depending on how the data are interpreted. Although each of these interpretations of the LESS trial results is different, each represents a more nuanced and evidence-based approach to treatment that will improve outcomes and reduce unnecessary harm from corticosteroid injections.

Study Results in Context

The LESS trial primary outcome findings46 and recent systematic reviews indicate little short-term benefit from adding corticosteroid to epidural lidocaine injections for stenosis symptoms.7,47,48 The LESS/LESSER study also suggests no differences in long-term benefits of epidural injections with corticosteroid compared with injections with lidocaine alone in terms of pain, function, opioid use, and spinal surgery for spinal stenosis. These findings are consistent with prior observational studies of the natural history of spinal stenosis.36,37 However, the publication of the LESS trial results has been met with significant resistance from interventional pain societies that are concerned our findings will be used to eliminate coverage for these procedures for patients. We have engaged with these societies as well as individual interventionalists, and although some agree with the findings of the trial, most express concern about the interpretation of the results and the potential impact on clinical practice. These interventionalists strongly believe that the demonstrated improvement in pain and function in both groups indicates that these injections are effective, despite the fact that the corticosteroid injections were no more effective than the lidocaine injections. They also express strong preference for accepting their own clinical observations about effectiveness of epidural corticosteroid injections rather than the results of the clinical trial. This strong opposition to incorporating the results of the LESS trial into clinical practice has made efforts to reduce or eliminate the use of these injections extremely difficult.

Since the publication of the LESS trial results, AHRQ sponsored a technology assessment/systematic review of injection therapies for chronic low back pain, and both the states of Oregon and Washington have undertaken re-review of coverage policies of epidural corticosteroid injections. The AHRQ technology assessment concluded that evidence for epidural corticosteroid injections vs placebo interventions for spinal stenosis is limited and shows no differences in outcomes related to pain or function. These conclusions were largely based on the LESS trial results given the paucity of other clinical trial data for spinal stenosis. However, at the time of this review only the LESS trial data through the primary outcome of 6 weeks were published and therefore included in the AHRQ technology assessment. The Washington State Health Technology Assessment committee in 2016 also conducted a similar systematic review, and although similar conclusions were reached, the committee did not change the current coverage recommendations or incorporate the results of the LESS trial into an updated coverage decision. The Oregon state review is currently being reviewed and has incorporated the results of the LESS trial with a preliminary recommendation for noncoverage of epidural corticosteroid injections for spinal stenosis, in large part due to the results of the LESS trial.

Generalizability

We conducted the LESS/LESSER trial at 16 clinical sites across the United States and included patients being treated in academic medical centers, private practice, and a VA Medical Center. The trial also included several integrated health systems with variability in terms of geographic regions, ethnicities, and socioeconomic statuses. As this was a pragmatic trial designed to reflect actual clinical practice as much as possible in terms of patient selection and injection technique, this trial can be generalized to a variety of clinical practices across the United States. However, this trial was limited to participants 50 years and older with neurogenic claudication and central lumbar spinal stenosis. The results may not generalize to younger patients or patients with radicular pain associated with foraminal stenosis or herniated discs.

Study Limitations

The focus groups for aim 1 recruited patients from Washington state only and relied on patient report to determine if patients had spinal stenosis. The aim 2 decision aids were mailed to participants without training on how to interpret the results of the report and were not provided in the context of a clinic visit in which treatment decisions were being made. This may have limited the impact of these reports on outcomes.

Because this was a pragmatic trial (aim 3), there was inherent heterogeneity in patient characteristics and clinicians' procedural techniques (approach, corticosteroid type, and dose). All patients in this trial received an epidural injection of lidocaine, and there was no true sham or noninjected control group. Thus, for the variables that predicted outcome regardless of treatment, we do not know whether these variables were specific predictors of response to an epidural injection of lidocaine, predictors of a placebo response to injection, or predictors of outcomes regardless of any treatment received. The LESS trial was powered to address the effectiveness of corticosteroid vs lidocaine alone at 6 weeks in total and for the subgroup of interlaminar vs transforaminal injections. All other secondary analyses (including the 12-month analyses and all other analyses associated with LESSER) should therefore be considered hypothesis generating rather than definitive. Finally, for statistical inference we could have employed longitudinal data methods that simultaneously analyze the full longitudinal outcomes profile and take into account missing data. Given the low proportion of missing data, the lack of any significant or meaningful predictors of missing data, and the lack of a difference between treatment groups, it is unlikely that the reported results would show more than a clinically unimportant difference in outcomes between treatment groups.

Future Research

Future research efforts should focus on ways to engage patients in shared decision-making about treatments for pain; particularly in the context of limited effectiveness evidence. This can be seen in the mixed impact of the individualized reports depicting their outcomes following injections in helping patients decide whether to pursue additional injections. A better understanding of how to best graphically display and communicate results back to patients to improve future decision-making is needed. Additional research is also needed to understand how data on PROs can be used to help patients make treatment decisions. In addition, more research on ways to engage patients in alternative treatments for spinal stenosis, including exercise, and the comparative effectiveness of these treatments on outcomes is warranted.

Conclusions

In summary, the LESSER study provided several key results with direct implications for clinical care and shared decision-making, including the following: (1) In addition to pain and function, ability to participate in exercise is an important outcome for older adults with spinal stenosis; (2) provision of individualized PRO data is helpful to patients from their perspective in making treatment decisions, but further research is needed to understand the impact of providing these data on outcomes; (3) there are no benefits in pain reduction and improving functional disability of epidural injections with steroid over epidural injections with lidocaine alone for patients with lumbar spinal stenosis beyond 3 weeks; (4) patients with depression, anxiety, and fear avoidance behaviors tend to do worse in terms of self-reported pain and function at all time points, regardless of the treatment they receive43; and (5) repeat injections are of limited value, particularly if the first injection was not associated with improved outcomes. Given the complexity of the results and the current health care landscape, incorporating these key study findings into clinical practice will require a multipronged dissemination and implementation strategy to engage patients, referring providers, and interventional providers in evidence-based shared decision-making about use of epidural injections for spinal stenosis.

References

1.
Harrast MA. Epidural steroid injections for lumbar spinal stenosis. Curr Rev Musculoskelet Med. 2008;1:32-38. [PMC free article: PMC2684150] [PubMed: 19468896]
2.
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;91:1295-1304. [PMC free article: PMC2686131] [PubMed: 19487505]
3.
Tokuhashi Y, Ajiro Y, Umezawa N. Outcomes of posterior fusion using pedicle screw fixation in patients > or = 70 years with lumbar spinal canal stenosis. Orthopedics. 2008;31:1096. [PubMed: 19226091]
4.
Weinstein JN, Tosteson TD, Lurie JD, et al. Surgical versus nonsurgical therapy for lumbar spinal stenosis. N Engl J Med. 2008;358:794-810. [PMC free article: PMC2576513] [PubMed: 18287602]
5.
Manchikanti L, Pampati V, Falco FJ, Hirsch JA. Growth of spinal interventional pain management techniques: analysis of utilization trends and Medicare expenditures 2000 to 2008. Spine (Phila Pa 1976). 2013;38:157-168. [PubMed: 22781007]
6.
Manchikanti L, Pampati V, Falco FJ, Hirsch JA. An updated assessment of utilization of interventional pain management techniques in the Medicare population: 2000-2013. Pain Physician. 2015;18:E115-E127. [PubMed: 25794210]
7.
Chou R, Hashimoto R, Friedly J, et al. Epidural corticosteroid injections for radiculopathy and spinal stenosis: a systematic review and meta-analysis. Ann Intern Med. 2015;163:373-381. [PubMed: 26302454]
8.
Friedly J, Chan L, Deyo R. Increases in lumbosacral injections in the Medicare population: 1994 to 2001. Spine (Phila Pa 1976). 2007;32:1754-1760. [PubMed: 17632396]
9.
Friedly J, Nishio I, Bishop MJ, Maynard C. The relationship between repeated epidural steroid injections and subsequent opioid use and lumbar surgery. Arch Phys Med Rehabil. 2008;89:1011-1015. [PubMed: 18503793]
10.
Manchikanti L, Singh V, Boswell MV. Interventional pain management at crossroads: the perfect storm brewing for a new decade of challenges. Pain Physician. 2010;13:E111-E140. [PubMed: 20309388]
11.
Manchikanti L, Pampati V, Falco FJ, Hirsch JA. Assessment of the growth of epidural injections in the medicare population from 2000 to 2011. Pain Physician. 2013;16:E349-E364. [PubMed: 23877459]
12.
Ammendolia C, Stuber K, de Bruin LK, et al. Nonoperative treatment of lumbar spinal stenosis with neurogenic claudication: a systematic review. Spine (Phila Pa 1976). 2012;37:E609-E616. [PubMed: 22158059]
13.
Manchikanti L, Singh V, Cash KA, Pampati V, Damron KS, Boswell MV. A randomized, controlled, double-blind trial of fluoroscopic caudal epidural injections in the treatment of lumbar disc herniation and radiculitis. Spine (Phila Pa 1976). 2011;36:1897-1905. [PubMed: 21897343]
14.
Manchikanti L, Singh V, Cash KA, Pampati V, Damron KS, Boswell MV. Effect of fluoroscopically guided caudal epidural steroid or local anesthetic injections in the treatment of lumbar disc herniation and radiculitis: a randomized, controlled, double blind trial with a two-year follow-up. Pain Physician. 2012;15:273-286. [PubMed: 22828681]
15.
Manchikanti L, Singh V, Cash KA, Pampati V, Datta S. Preliminary results of a randomized, equivalence trial of fluoroscopic caudal epidural injections in managing chronic low back pain: part 3—post surgery syndrome. Pain Physician. 2008;11:817-831. [PubMed: 19057628]
16.
Fukusaki M, Kobayashi I, Hara T, Sumikawa K. Symptoms of spinal stenosis do not improve after epidural steroid injection. Clin J Pain. 1998;14:148-151. [PubMed: 9647457]
17.
Apáthy A, Penczner G, Licker E, et al. The value of caudal epidural steroids in the treatment of lumbar neural compression syndromes. J Neurol Orthop Med Surg. 1991;112:181-184.
18.
Nam HS, Park YB. Effects of transforaminal injection for degenerative lumbar scoliosis combined with spinal stenosis. Ann Rehabil Med. 2011;35:514-523. [PMC free article: PMC3309239] [PubMed: 22506167]
19.
Jordan K, Dunn KM, Lewis M, Croft P. A minimal clinically important difference was derived for the Roland-Morris Disability Questionnaire for low back pain. J Clin Epidemiol. 2006;59:45-52. [PubMed: 16360560]
20.
Lauridsen HH, Hartvigsen J, Manniche C, Korsholm L, Grunnet-Nilsson N. Responsiveness and minimal clinically important difference for pain and disability instruments in low back pain patients. BMC Musculoskelet Disord. 2006;7:82. [PMC free article: PMC1635558] [PubMed: 17064410]
21.
Friedly JL, Bresnahan BW, Comstock B, et al. Study protocol-Lumbar Epidural steroid injections for Spinal Stenosis (LESS): a double-blind randomized controlled trial of epidural steroid injections for lumbar spinal stenosis among older adults. BMC Musculoskelet Disord. 2012;13:48. [PMC free article: PMC3349603] [PubMed: 22458343]
22.
Jarvik JG, Comstock BA, Heagerty PJ, et al. Back pain in seniors: the Back pain Outcomes using Longitudinal Data (BOLD) cohort baseline data. BMC Musculoskelet Disord. 2014;15:134. [PMC free article: PMC4021204] [PubMed: 24755158]
23.
Jarvik JG, Comstock BA, Bresnahan BW, et al. Study protocol: the Back pain Outcomes using Longitudinal Data (BOLD) registry. BMC Musculoskelet Disord. 2012;13:64. [PMC free article: PMC3403933] [PubMed: 22554166]
24.
Bang H, Ni L, Davis CE. Assessment of blinding in clinical trials. Control Clin Trials. 2004;25:143-156. [PubMed: 15020033]
25.
Roland M, Fairbank J. The Roland-Morris Disability Questionnaire and the Oswestry Disability Questionnaire. Spine (Phila Pa 1976). 2000;25:3115-3124. [PubMed: 11124727]
26.
Dworkin RH, Turk DC, Wyrwich KW, et al. Interpreting the clinical importance of treatment outcomes in chronic pain clinical trials: IMMPACT recommendations. J Pain. 2008;9:105-121. [PubMed: 18055266]
27.
Ostelo RW, Deyo RA, Stratford P, et al. Interpreting change scores for pain and functional status in low back pain: towards international consensus regarding minimal important change. Spine (Phila Pa 1976). 2008;33:90-94. [PubMed: 18165753]
28.
Tan G, Jensen MP, Thornby JI, Shanti BF. Validation of the Brief Pain Inventory for chronic nonmalignant pain. J Pain. 2004;5:133-137. [PubMed: 15042521]
29.
Kroenke K, Strine TW, Spitzer RL, Williams JB, Berry JT, Mokdad AH. The PHQ-8 as a measure of current depression in the general population. J Affect Disord. 2009;114:163-173. [PubMed: 18752852]
30.
Spitzer RL, Kroenke K, Williams JB, Lowe B. A brief measure for assessing generalized anxiety disorder: the GAD-7. Arch Intern Med. 2006;166:1092-1097. [PubMed: 16717171]
31.
Brazier J, Jones N, Kind P. Testing the validity of the Euroqol and comparing it with the SF-36 health survey questionnaire. Qual Life Res. 1993;2:169-180. [PubMed: 8401453]
32.
Stucki G, Daltroy L, Liang MH, Lipson SJ, Fossel AH, Katz JN. Measurement properties of a self-administered outcome measure in lumbar spinal stenosis. Spine (Phila Pa 1976). 1996;21:796-803. [PubMed: 8779009]
33.
Liang KY, Zeger SL. Longitudinal data analysis using generalized linear models. Biometrika. 1986;73(1):13-22.
34.
Bantug ET, Coles T, Smith KC, et al. Graphical displays of patient-reported outcomes (PRO) for use in clinical practice: what makes a pro picture worth a thousand words? Patient Educ Couns. 2016;99:483-490. [PubMed: 26603445]
35.
McNair AG, Brookes ST, Davis CR, Argyropoulos M, Blazeby JM. Communicating the results of randomized clinical trials: do patients understand multidimensional patient-reported outcomes? J Clin Oncol. 2010;28:738-743. [PubMed: 20065187]
36.
Benoist M. The natural history of lumbar degenerative spinal stenosis. Joint Bone Spine. 2002;69:450-457. [PubMed: 12477228]
37.
Micankova Adamova B, Vohanka S, Dusek L, Jarkovsky J, Bednarik J. Prediction of long-term clinical outcome in patients with lumbar spinal stenosis. Eur Spine J. 2012;21:2611-2619. [PMC free article: PMC3508241] [PubMed: 22772352]
38.
Lurie JD, Tosteson TD, Tosteson A, et al. Long-term outcomes of lumbar spinal stenosis: eight-year results of the Spine Patient Outcomes Research Trial (SPORT). Spine (Phila Pa 1976). 2015;40:63-76. [PMC free article: PMC4288009] [PubMed: 25569524]
39.
Charmandari E, Nicolaides NC, Chrousos GP. Adrenal insufficiency. Lancet. 2014;383:2152-2167. [PubMed: 24503135]
40.
Bicket MG, Gupta A, Brown C, Cohen SP. Epidural injections for spinal pain: a systematic review and meta-analysis evaluating the “control” injections in randomized controlled trials. Anesthesiology. 2013;119:907-931. [PubMed: 24195874]
41.
van Eijs F, Geurts J, van Kleef M, et al. Predictors of pain relieving response to sympathetic blockade in complex regional pain syndrome type 1. Anesthesiology. 2012;116:113-121. [PubMed: 22143169]
42.
Cohen SP, Kapoor SG, Rathmell JP. Intravenous infusion tests have limited utility for selecting long-term drug therapy in patients with chronic pain: a systematic review. Anesthesiology. 2009;111:416-431. [PubMed: 19602955]
43.
Turner JA, Comstock BA, Standaert CJ, et al. Can patient characteristics predict benefit from epidural corticosteroid injections for lumbar spinal stenosis symptoms? Spine J. 2015;15:2319-2331. [PubMed: 26096484]
44.
Suri P, Pashova H, Heagerty PJ, et al. Short-term improvements in disability mediate patient satisfaction after epidural corticosteroid injections for symptomatic lumbar spinal stenosis. Spine (Phila Pa 1976). 2015;40:1363-1370. [PubMed: 26010037]
45.
Agency for Healthcare Research and Quality (AHRQ). A randomized trial of epidural glucocorticoid injections for spinal stenosis: a brief summary of findings for clinicians. Published June 2016. https:​//effectivehealthcare​.ahrq.gov/sites​/default/files/pdf​/glucocorticoid-injections_clinician.pdf [PubMed: 27583322]
46.
Friedly JL, Comstock BA, Turner JA, et al. A randomized trial of epidural glucocorticoid injections for spinal stenosis. N Engl J Med. 2014;371:11-21. [PubMed: 24988555]
47.
Shaughnessy AF. Epidural steroid not better than placebo injection for sciatica and spinal stenosis pain and function. Am Fam Physician. 2016;93:315-316. [PubMed: 26926820]
48.
Rabinovitch DL, Peliowski A, Furlan AD. Influence of lumbar epidural injection volume on pain relief for radicular leg pain and/or low back pain. Spine J. 2009;9:509-517. [PubMed: 19398387]

Acknowledgment

Research reported in this report was [partially] funded through a Patient-Centered Outcomes Research Institute® (PCORI®) Award (#CE-12-11-4469) Further information available at: https://www.pcori.org/research-results/2013/comparing-effects-two-types-epidural-shots-pain-and-physical-ability-older

Original Project Title: Long Term Outcomes of Lumbar Epidural Steroid Injections for Spinal Stenosis
PCORI ID: CE-12-11-4469
ClinicalTrials.gov ID: NCT02260401

Suggested citation:

Friedly JL, Bauer Z, Comstock B, et al. (2019). Comparing the Effects of Two Types of Epidural Shots on Pain and Physical Ability in Older Adults with Lumbar Spinal Stenosis. Patient-Centered Outcomes Research Institute (PCORI). https://doi.org/10.25302/4.2019.CE.12114469

Disclaimer

The [views, statements, opinions] presented in this report are solely the responsibility of the author(s) and do not necessarily represent the views of the Patient-Centered Outcomes Research Institute® (PCORI®), its Board of Governors or Methodology Committee.

Copyright © 2019. University of Washington. All Rights Reserved.

This book is distributed under the terms of the Creative Commons Attribution-NonCommercial-NoDerivs License which permits noncommercial use and distribution provided the original author(s) and source are credited. (See https://creativecommons.org/licenses/by-nc-nd/4.0/

Bookshelf ID: NBK599886PMID: 38315786DOI: 10.25302/4.2019.CE.12114469

Views

  • PubReader
  • Print View
  • Cite this Page
  • PDF version of this title (2.9M)

Other titles in this collection

Related information

Similar articles in PubMed

See reviews...See all...

Recent Activity

Your browsing activity is empty.

Activity recording is turned off.

Turn recording back on

See more...