When an Epidural Doesn’t Work for Back Pain

Roughly half of patients who receive an epidural steroid injection for back pain do not get meaningful relief. In one study tracking outcomes over three weeks, about 53% of patients never reached a 50% reduction in pain at any point after the injection. The reasons range from misidentified pain sources to physical barriers inside the spine that block the medication from reaching the right spot. Understanding why an epidural fails is often more useful than simply repeating it.

How Often Epidurals Work and How Quickly You Should Know

The timeline of response tells you a lot. A study of 108 patients tracked daily pain scores after epidural steroid injections and found that among those who ultimately responded well, nearly three-quarters reported at least 50% pain relief on the very first day. Another fifth of responders first noticed meaningful improvement by day four. Only a handful of patients who were going to respond took longer than two weeks to feel it.¹ In practical terms, if you have had no improvement after two to three weeks, you are very likely in the non-responder group.

The strength of evidence also depends on what is actually causing your back pain. For disc herniations that produce leg pain, caudal epidural injections delivered through a catheter left in place for two days have moderate-quality evidence supporting their effectiveness over three, six, and twelve months. For other common conditions like spinal stenosis, discogenic low back pain without herniation, or chronic pain after back surgery, the evidence drops to low quality.² An epidural is more likely to fail when the underlying condition is one for which the evidence was thin to begin with.

When the Pain Source Is Not What It Appears to Be

An epidural steroid injection targets inflammation around spinal nerves. If your pain is not primarily caused by that inflammation, no amount of steroid in the right spot will fix the problem. One increasingly recognized scenario involves patients with diabetes. Diabetic peripheral neuropathy can produce leg pain and numbness that looks, on the surface, identical to symptoms caused by a pinched nerve in the lower back. Even when an MRI shows a disc bulge or narrowing that could theoretically explain the symptoms, the actual culprit may be nerve damage from diabetes itself. A recent study found that in diabetic patients with mild, chronic lower-extremity pain, the real source of symptoms was sometimes neuropathy rather than the spinal problem visible on imaging, and in many cases both conditions were contributing at once, which substantially reduced how well the injection worked.³

This is a broader issue than diabetes alone. Back pain has many possible generators: facet joints, sacroiliac joints, muscles, ligaments, and the discs themselves can all produce pain that gets blamed on nerve compression. An epidural steroid injection is a targeted anti-inflammatory treatment for nerve root irritation. When the pain is coming from a structure the injection was never designed to treat, the failure is not really a failure of the epidural but a failure of diagnosis.

Physical Barriers Inside the Epidural Space

Even when the diagnosis is correct and the needle lands in the right general area, the medication sometimes cannot reach the inflamed nerve. The epidural space is not the open, uniform channel that anatomy textbooks often depict. A connective tissue structure called the plica mediana dorsalis runs down the back of the spinal canal and can act as a wall dividing the epidural space along the midline. In an imaging study of 30 patients, this tissue band was visible in 80% of cases. Among those patients who had it, almost four out of five showed contrast dye pooling more heavily on one side of the midline than the other.⁴ The asymmetry was statistically linked to the presence of this tissue band.

The clinical consequence is straightforward: if the medication pools on the wrong side, the inflamed nerve root on the opposite side gets little or none of it. This is one plausible explanation for the unilateral block problem, where an epidural works for pain on one side of the body but does nothing for the other. Case reports have confirmed that the plica mediana dorsalis can function as either a complete or partial membrane, creating unpredictable variation in how the injected solution distributes.⁵ Your doctor may do everything perfectly and the medication still misses its target because of anatomy nobody can see without imaging.

Whether the steroid reaches the front of the spinal canal, near where the nerve root exits, also matters. When researchers compared contrast dye spread patterns across different injection approaches, they found that having dye reach the ventral (front) side of the epidural space was associated with significantly better pain improvement at three months.⁶ That said, one retrospective review of transforaminal injections found that patients could get immediate pain relief regardless of the contrast pattern, so the relationship between spread and outcome is not perfectly predictable.⁷

Does the Injection Route Matter?

Epidural steroid injections can be performed through three main approaches: caudal (through the tailbone), interlaminar (between the vertebral bones from the back), and transforaminal (through the bony opening where the nerve exits the spine). The transforaminal route has a reputation for being the most precise because it delivers medication directly alongside the target nerve root. In one study comparing all three routes for lumbar disc problems, the transforaminal approach showed improvement in about 90% of patients at one year, compared with roughly 74% for caudal and 78% for interlaminar. The transforaminal route was significantly more effective at both six and twelve months.⁸

However, other research has not confirmed that advantage. A separate study found no meaningful difference between transforaminal and interlaminar routes in pain scores, disability scores, or functional outcomes at one, three, and six months. The authors concluded that despite the transforaminal route’s theoretical advantage as a selective nerve block, its real-world results at six months were comparable to the interlaminar approach.⁹ The evidence here is genuinely mixed, which means that switching from one injection route to another after a failed attempt is reasonable but not guaranteed to change the outcome.

Your Nervous System Can Undermine the Injection

Sometimes the problem is not where the steroid lands but how your nervous system processes pain. Central sensitization is a condition in which the spinal cord and brain become hypersensitive, amplifying pain signals even after the original tissue injury has partially healed. Think of it like a volume knob that is stuck on high. A study of patients undergoing transforaminal epidural injections for chronic lumbar radiculopathy found that those showing signs of central sensitization had significantly higher pain scores at every follow-up time point compared to patients without it. The sensitized patients also had worse disability and poorer mental health outcomes after the injection.¹⁰ Similar findings have been reported for cervical epidural injections, where central sensitization worsened pain, disability, and quality-of-life outcomes.¹¹

Psychological factors weigh in too, and not in the dismissive “it’s all in your head” sense. A pilot study found that patients with higher levels of negative emotions at baseline reported pain about one point higher on a ten-point scale at twelve weeks compared to those with lower negative affect. On the flip side, patients with stronger cognitive resilience reported pain about one point lower at six months.¹² A scoping review of the broader literature confirmed that poor psychosocial functioning is associated with inferior outcomes after epidural steroid injections for back pain.¹³ Depression, anxiety, catastrophizing, and fear-avoidance beliefs all tilt the odds against a good response. These are not character flaws; they are measurable neurobiological states that change how pain signals are processed and how inflammation resolves.

The Steroid Itself Is Probably Not the Weak Link

You might wonder whether your doctor used the wrong steroid. Corticosteroids used in epidural injections come in two broad categories: particulate formulations that linger in tissues longer, and nonparticulate formulations that dissolve quickly. The intuition would be that longer-acting drugs work better. A systematic review found no statistically significant difference between the two types in pain reduction or functional improvement for either cervical or lumbar injections.¹⁴ One head-to-head trial actually found that the nonparticulate steroid dexamethasone was not only noninferior but in some measures superior to particulate steroids in both pain relief and functional improvement at two months.¹⁵

Particulate steroids carry a small but real risk of catastrophic complications if they accidentally enter a blood vessel and block blood flow to the spinal cord. Given that the efficacy difference is negligible, most guidelines now recommend nonparticulate steroids as the first-line choice.¹⁴ So while the type of steroid used could matter in rare scenarios, it is unlikely to explain why your epidural did not work.

The Risks of Repeating Injections That Are Not Helping

When one epidural does not work, it is tempting to try another. Many insurance plans actually require a series of injections before approving surgery, a policy that has drawn criticism from physicians who argue it can delay necessary operations while exposing patients to repeated procedural risks.¹⁶ Each injection delivers a dose of corticosteroid that enters your bloodstream and affects the whole body, not just the spine.

One well-documented risk of cumulative steroid exposure is bone loss. A study of postmenopausal women who received frequent epidural steroid injections, averaging about fourteen injections with a cumulative dose of roughly 400 milligrams of triamcinolone, found lower bone mineral density in the hip compared with women who had fewer injections. The rates of osteoporosis and osteopenia were significantly higher in the frequent-injection group.¹⁷ Blood sugar spikes in diabetic patients, adrenal suppression, and weight gain are other systemic effects that accumulate with repeated injections. If the first one or two did nothing, continuing to repeat them has diminishing justification and growing cost to your body.

What to Try After a Failed Epidural

The path forward depends on why the injection failed. If the diagnosis is uncertain, the failure itself becomes diagnostic information. An epidural that does nothing for leg pain supposedly caused by a disc herniation may signal that the pain is coming from somewhere else, whether a joint, a muscle, or a neuropathic process unrelated to the disc.

For patients whose pain genuinely originates from a disc herniation with confirmed nerve root compression, surgery becomes a consideration after conservative measures fail. A systematic review of indications for surgery found that imaging-confirmed nerve root compression and severe or refractory pain were the two most consistent triggers for surgical referral. The most commonly used threshold was failure of conservative therapy after roughly four to six weeks, combined with neurological progression or worsening imaging findings.¹⁸

Physical therapy, particularly when matched to your specific movement pattern, deserves another look even if you have tried it before. Research has shown that patients who did poorly with generic exercise programs improved when they were given direction-specific exercises tailored to their individual pain response.¹⁹ A therapist trained in mechanical assessment can identify whether bending, extending, or side-shifting reliably changes your symptoms and build a program around that. Not all physical therapy is equivalent, and a failed round of generic strengthening exercises does not mean physical therapy itself has failed.

Spinal Cord Stimulation for Refractory Cases

For patients who have exhausted epidural injections, oral medications, and conventional physical therapy without adequate relief, spinal cord stimulation is a technology that delivers mild electrical pulses to the spinal cord to interrupt pain signals before they reach the brain. A long-term follow-up study of 77 patients with chronic, refractory low back and radicular pain found that results were good in about 64% of cases and fair in another 22%.²⁰ The catch is that patient selection matters enormously; the researchers emphasized that strict selection criteria were essential to good outcomes. Most programs require a trial period with temporary leads before committing to a permanent implant.

Case reports have documented patients with failed back surgery syndrome who had no relief from oral medications or repeated injections but regained the ability to walk and return to work after spinal cord stimulator implantation, followed by intensive rehabilitation.²¹ Stimulation is not a cure and does not fix the underlying structural problem. It works by changing how the brain perceives the signals coming from the spine. For the right patient, that can be the difference between disability and functional life, but it is a last-resort tool, not a first step after one failed epidural.