What are the most common complications after epidural steroid injections and how often do they occur?
Executive summary
Epidural steroid injections (ESIs) most commonly cause transient, minor problems—worse short-term pain, injection-site soreness and temporary numbness—occurring in roughly 2–9% of injections in published series, with a well-cited large series reporting a 2.4% overall minor complication rate (increased pain 1.1%, injection-site pain 0.33%, persistent numbness 0.14%) [1] [2]. Catastrophic outcomes such as spinal cord infarction, epidural hematoma, abscess, or permanent neurologic deficit are documented but rare; their exact frequency is difficult to pin down and varies with technique, steroid formulation and reporting window [3] [4] [5].
1. The everyday harms: what patients actually experience and how often
Large retrospective series and reviews show that the common complications after ESIs are typically minor and self-limited: transient worsening of pain is the most frequent complaint, followed by brief injection-site pain, transient numbness or weakness, vasovagal reactions, flushing, and sleep or glucose changes in some patients, with reported minor complication rates in many studies clustered around 2–9% depending on definitions and follow‑up windows [1] [2] [6] [7].
2. The rare but serious complications that drive headlines
More severe adverse events—epidural hematoma causing cord compression, epidural abscess or meningitis, direct spinal cord or nerve-root injury from needle trauma, and embolization of particulate steroid into arterial branches producing infarction—are well described in case reports and reviews and are responsible for most permanent neurologic harm as documented in specialty reviews and safety analyses [3] [4] [8]. Population studies and claims analyses confirm these events are uncommon, but also show that serious-spinal-adverse-event case ascertainment is constrained by short surveillance windows and low absolute event counts, making precise incidence estimates elusive [5].
3. Technique, steroid type and anatomical level change the odds
Complication rates are not uniform: transforaminal versus interlaminar versus caudal approaches, use of particulate versus non‑particulate steroid formulations, and injections performed above L3 all correlate with different risk profiles in the literature, with some series finding higher intravascular or procedural complication frequencies for lumbar transforaminal procedures and adhesiolysis, and guidance that particulate steroids have been implicated in embolic spinal cord injury [9] [7] [2]. Studies comparing approaches report transforaminal injections may have lower minor‑complaint rates in some series but also pose distinct vascular risks that inform choice of steroid and technique [1] [9].
4. Who is at greater risk and what about long‑term effects
Patient factors and treatment frequency matter: repeated or frequent steroid dosing can contribute to systemic steroid effects such as transient hyperglycemia and, over time, may accelerate bone loss or muscle weakening, so many clinicians limit ESIs to a few per year for most patients [10] [11]. Long‑term and delayed complications—epidural abscess, adhesive arachnoiditis or late neurologic decline—are recognized in case reports and smaller studies but are harder to quantify because many surveillance studies use short follow up windows and case reporting is heterogeneous [8] [5].
5. What the numbers mean for clinical decision‑making
The best available aggregated clinical data present ESIs as a generally safe, well‑tolerated intervention in the short term with minor complication rates commonly around a few percent (e.g., 2.4% minor complications in a 4,265‑injection series) while major adverse events remain rare but potentially devastating and variably reported [1] [2] [3]. That balance explains the consensus seen in major centers: ESIs are used selectively for radicular pain when likely to help, with attention to technique, steroid selection and limiting repeat dosing to reduce cumulative harms [12] [11].