Subscription is FREE for qualified healthcare professionals in the US.

How Safe Is Epidural Steroid Injection? Examining Drug-Related Factors

While the frequency of serious complications due to ESI is undefined, certain pharmaceutical characteristics of corticosteroids (ie, particulate vs nonparticulate) are thought to contribute to overall risk.
Page 1 of 2

Epidural steroid injection (ESI) has been a treatment modality for low back pain (LBP) and neck pain, with and without radiculopathy, for decades.1 This procedure involves administration of corticosteroids into the epidural space around the spinal cord and nerves. LBP is the second most common symptom-related reason for physician office visits in the United States, and ESI has become the most commonly performed intervention to treat LBP.1,2 The beneficial effects of ESI are proposed to include decreased production of inflammatory mediators, depressed nerve conduction, and dilution of local concentrations of endogenous inflammatory cytokines.1,3 Although the majority of controlled studies of ESI for LBP and neck pain document a moderate and transient benefit from ESI,1 there is controversy over whether these benefits are long-term.2

While considered minimally invasive, ESI can cause serious adverse events, including neurologic injury, embolization, bleeding, and infection.1 Risk depends in large part on technique, including interlaminar versus transforaminal injection, which theoretically have different risk of dural puncture. Examples of serious adverse events include inadvertent injection of drug into a vein or artery, which can result in debilitating or fatal infarction of the brainstem or spinal cord.4 However, it also can be related to characteristics of the corticosteroid used in ESI.5 Other drug-related concerns with ESI were highlighted by the 2012 outbreak of fungal meningitis in patients who received contaminated corticosteroids produced at a compounding pharmacy.6

Frequency of Serious Adverse Events

These rare, but potentially devastating, complications of ESI have an undefined incidence rate because clinical trials generally have not been well-designed and epidemiologic safety data have not been routinely collected.1 Available review articles highlight the infrequent nature and difficulty of quantifying the risk of severe adverse events, despite large sample sizes. For example, one 7-year retrospective review in which 4,265 ESIs were administered to 1,857 patients reported no major complications.7 The majority of data describing disastrous complications from ESI come from case reports.8-21 Findings from evidence reviews have been nonspecific, stating only that the risk of major complications is low and that causative factors mostly are related to physical technique (eg, needle placement) and the formulations of injected steroids.1,22

In the absence of trials designed to detect rare adverse events due to ESI, some useful data have come from a national survey of physician members of the American Pain Society (APS).23 Of 1,340 surveyed physicians, 21.4% responded, revealing 78 total ESI complications in patients receiving cervical transforaminal ESI. Serious complications included 16 brain infarcts, 12 spinal cord infarcts, and 2 combined brain/spinal cord infarcts; death occurred in 13 cases. Corticosteroids were used in 70 of the cases with reported complications. The reported corticosteroids included triamcinolone, betamethasone, and methylprednisolone, which were associated with 10%, 11%, and 79% of complications, respectively.

Pharmaceutical Characteristics of Corticosteroids

While the frequency of serious complications due to ESI is undefined, certain pharmaceutical characteristics of corticosteroids are thought to contribute to overall risk.1,24 For example, the size of aggregated drug molecules may affect the likelihood of embolization following inadvertent injection into vertebral or foraminal arteries.25 Corticosteroids considered more likely to aggregate and potentially embolize are termed particulate corticosteroids. In contrast, those less likely to embolize are termed non-particulate corticosteroids. The latter contain particles smaller than red blood cells and exhibit low propensity to aggregate or pack; they are therefore advocated by some as the corticosteroids of choice in ESI to reduce embolization risk.24,25 Among corticosteroids commonly used in ESI, particulate corticosteroids include methylprednisolone and triamcinolone.23,25,26 Dexamethasone generally is agreed to be non-particulate, while the classification of betamethasone has varied.23,27 Corticosteroids reported in the APS survey all were considered particulate by the authors.23

Microscopic analysis of steroid preparations has assessed pharmaceutical characteristics of corticosteroids used in ESI.25 Particles of triamcinolone and betamethasone equaled or exceeded the diameter of red blood cells and exhibited aggregation, with the largest particles exceeding the median red blood cell size by 12 times, indicating potential for embolization. Those of methylprednisolone acetate were smaller than red blood cells and did not aggregate, although they were densely packed, leaving potential for embolization. Because of this potential, methylprednisolone is generally categorized as particulate. Particle characteristics were most favorable in dexamethasone, which showed sizes 5 to 10 times smaller than red blood cells and no aggregation. The absence of case reports associating dexamethasone with ESI complications substantiates this finding.23,28 In fact, injection of dexamethasone directly into the carotid artery of 8 rats did not cause any injury, whereas injection with methylprednisolone acetate (Depo-Medrol) or methylprednisolone sodium succinate (Solu-Medrol) in 15 of 19 rats caused cerebral hemorrhage (Figure 1).2

Trials comparing outcomes of particulate versus non-particulate steroids have been limited, with inconsistent safety outcomes.30-34 Of 5 retrospective and prospective studies, 2 reported no complications—1 study of 159 patients receiving dexamethasone or triamcinolone for cervical transforaminal epidural injection, and another study of 60 patients receiving dexamethasone or methylprednisolone in lumbar translaminar epidural injection (though 1 patient lost to follow-up inadvertently received intrathecal injections).30,31

Other pharmaceutical characteristics presumed to increase risk of injury include the presence of the neurotoxic preservatives polyethylene glycol (PEG) and benzyl alcohol.1 However, controversy exists regarding which is the offending agent.35 Concentrations of PEG in corticosteroid preparations approximate 2.8% to 3% of the solution; animal studies have shown changes in nerve conduction only at concentrations of 20% and greater.4 Additionally, benzyl alcohol is present in some preparations at concentrations of 0.9%, though injection of up to 9% produced only transient changes in animal models.36 Some available corticosteroid preparations contain either or both of these preservatives—some preparations of methylprednisolone acetate contain PEG or benzyl alcohol, whereas some preparations of betamethasone, triamcinolone, and dexamethasone sodium succinate contain benzyl alcohol but not PEG.37-40

Last updated on: April 14, 2015
First published on: March 1, 2014