Lumbar epidural steroid injections for herniation and stenosis: incidence and risk factors of subsequent surgery

Abstract

Background ContextLumbosacral epidural steroid injections (ESIs) have increased dramatically despite a narrowing of the clinical indications for use. One potential indication is to avoid or delay surgery, yet little information exists regarding surgery rates after ESI. PurposeThe purpose of this research was to determine the proportion of patients having surgery after lumbar ESI for disc herniation or stenosis and to identify the timing and factors associated with this progression. Study Design/SettingThis study was a retrospective review of nationally representative administrative claims data from the Truven Health MarketScan databases from 2007 to 2014. Patient SampleThe study cohort was comprised of 179,025 patients (54±15 years, 48% women) having lumbar ESIs for diagnoses of stenosis and/or herniation. Outcome MeasuresThe primary outcome measure was the time from ESI to surgery. MethodsInclusion criteria were ESI for stenosis and/or herniation, age ≥18 years, and health plan enrollment for 1 year before ESI to screen for exclusions. Patients were followed longitudinally until they progressed to surgery or had a lapse in enrollment, at which time they were censored. Rates of surgery were assessed with the Kaplan–Meier survival curves. Demographic and treatment factors associated with surgery were assessed with multivariable Cox proportional hazard models. No external funding was procured for this research and the authors' conflicts of interest are not pertinent to the present work. ResultsWithin 6 months, 12.5% of ESI patients underwent lumbar surgery. By 1 year, 16.9% had surgery, and by 5 years, 26.1% had surgery. Patients with herniation had surgery at rates of up to five-fold to seven-fold higher, with the highest rates of surgery in younger patients and those with both herniation and stenosis. Other concomitant spine diagnoses, male sex, previous tobacco use, and residence a rural areas or regions other than the Northeastern United States were associated with higher surgery rates. Medical comorbidities (previous treatment for drug use, congestive heart failure, obesity, chronic obstructive pulmonary disease, hypercholesterolemia, and other cardiac complications) were associated with lower surgery rates. ConclusionsIn the long term, more than one out of every four patients undergoing ESI for lumbar herniation or stenosis subsequently had surgery, and nearly one of six had surgery within the first year. After adjusting for other patient demographics and comorbidities, patients with herniation were more likely have surgery than those with stenosis. The improved understanding of the progression from lumbar ESI to surgery will help to better inform discussions regarding the value of ESI and aid in the shared decision-making process.