What to Do When Steroids Don’t Work for Back Pain

Epidural steroid injections are one of the most common treatments for back pain that radiates into the legs, but they do not work for everyone. When they fail, the path forward depends on why they failed, which could be anything from a misidentified pain source to a condition that simply does not respond well to steroids. The good news is that a range of alternatives exists, from nerve-targeting medications and physical therapy approaches to interventional procedures and, in some cases, surgery.

Why Steroids Fail for Some People

Steroid injections work by reducing inflammation around compressed or irritated spinal nerves. But that mechanism assumes inflammation is the primary driver of your pain, and that the injection reaches the right spot. Several factors can undermine one or both of those assumptions.

The type of spinal problem matters. Steroids tend to perform better for disc herniations than for spinal canal stenosis, likely because a herniated disc triggers a more localized inflammatory response that steroids can effectively tamp down.¹ Even within disc herniations, outcomes vary. A comprehensive review found that success rates are higher in patients with contained herniations that cause only mild nerve compression, while larger or more severe herniations respond less reliably.²

Another factor that can predict failure is central sensitization, a condition where the nervous system itself becomes hypersensitive and amplifies pain signals even when the original source of irritation has calmed down. Patients with signs of central sensitization have significantly worse outcomes from transforaminal steroid injections, including higher pain scores, more disability, and poorer mental health scores at follow-up compared to those without it.³ If your nervous system has essentially learned to stay in pain mode, an anti-inflammatory injection at the spine may not be enough to interrupt that cycle.

Revisiting the Diagnosis

Before exploring alternative treatments, it is worth asking whether the original diagnosis was correct. Back pain with leg symptoms can stem from several structures: a disc pressing on a nerve, an arthritic facet joint, the sacroiliac joint, or even a problem at a different spinal level than the one that was injected. If the steroid landed at the wrong level or addressed the wrong structure, the treatment did not really fail so much as miss the target.

Diagnostic selective nerve root blocks are sometimes used to try to pinpoint which nerve is responsible. The idea is straightforward: numb one specific nerve root with local anesthetic, and if the pain goes away, you have found the culprit. In practice, though, the accuracy is limited. One prospective study found that when “positive” was defined as at least 70% pain relief, these diagnostic blocks had a sensitivity of only 57% and an accuracy of about 73%.⁴ False negatives happened when the anesthetic did not adequately spread to the nerve, and false positives occurred when the medication leaked into adjacent areas, numbing a nerve that was not actually the pain source. A systematic review came to a similar conclusion, finding that the diagnostic accuracy of selective nerve root blocks is uncertain and that specificity in particular may be low.⁵

What this means practically is that if a steroid injection failed, the answer might be as simple as the injection targeted the wrong nerve root or the wrong structure entirely. Advanced imaging and a fresh clinical evaluation can sometimes reveal a pain generator that was overlooked the first time around, such as a facet joint problem masquerading as a disc issue or a sacroiliac joint contributing to what seemed like radicular pain.

Medications That Target Nerve Pain Differently

When the pain is clearly nerve-related but steroids have not helped, gabapentin and pregabalin are commonly prescribed alternatives. These drugs work differently from steroids. Rather than reducing inflammation, they calm overactive nerve signaling by blocking certain calcium channels in the nervous system.

A study examining gabapentin in patients with radicular pain from both spinal stenosis and disc herniations found that pain scores dropped and walking distance improved after three months of treatment.⁶ Pregabalin, a related drug, has shown similar results. A head-to-head comparison of the two found that both reduced pain in patients with cervical and lumbar radiculopathy, with no significant difference between them.⁷ These medications are not painkillers in the traditional sense and they typically take days to weeks to reach full effect, so patience is necessary. Side effects like drowsiness and dizziness are common, especially early on, but they tend to ease over time.

Other oral medications sometimes considered include duloxetine (an antidepressant that also modulates pain pathways), nonsteroidal anti-inflammatory drugs for their baseline pain-relieving effect, and short courses of muscle relaxants if spasm is contributing. Opioids remain an option for severe, intractable pain but carry well-documented risks of dependence and diminishing returns over time, so most guidelines reserve them as a last resort and for short durations.

Physical Therapy and Manual Approaches

If you have not tried structured physical therapy, or if you tried it before injections and stopped, this is often a good place to circle back. Exercise-based rehabilitation is one of the most broadly supported treatments for chronic low back pain, and it can complement or replace injections depending on the situation.

The McKenzie method, a system of repeated directional exercises and postural correction, has performed well in studies of chronic low back pain. A systematic review comparing it to manual therapy found that McKenzie-treated patients showed greater pain improvement at two to three months, and better disability scores at six and twelve months.⁸ The approach is somewhat unusual among physical therapy methods because it uses the patient’s response to specific movements to classify and direct treatment, which means two people with similar imaging findings might get different exercise prescriptions based on how their pain behaves.

Massage therapy has shown effectiveness for persistent back pain in initial studies, while spinal manipulation offers small clinical benefits roughly equivalent to other commonly used therapies.⁹ Neither is likely to be a standalone solution for someone whose steroid injections have failed, but they can be valuable components of a broader management plan, particularly for reducing muscle guarding and improving mobility enough to engage in exercise.

Radiofrequency Ablation

Radiofrequency ablation uses heat delivered through a needle to disrupt the tiny nerve fibers that carry pain signals from a specific joint or structure. It is most commonly used for facet joint pain and sacroiliac joint pain, both of which can mimic or coexist with disc-related back pain and which do not respond to epidural steroid injections at all since the steroid is delivered to the wrong compartment.

A systematic review and meta-analysis found that radiofrequency ablation may offer longer-lasting pain relief compared to steroid injections for spinal facet and sacroiliac joint pain.¹⁰ In a separate study of patients who underwent radiofrequency ablation, participants experienced an average of roughly 48% pain relief lasting about four and a half months. Adding steroid to the ablation procedure did not significantly improve the outcome, suggesting the heat-based nerve disruption is doing the heavy lifting.¹¹

One important caveat: the treated nerves eventually regenerate, so pain can return months later. Many patients undergo repeat procedures. Radiofrequency ablation is also not appropriate for every type of back pain. It works best when diagnostic nerve blocks have confirmed a specific joint as the pain source, which brings us back to the importance of accurate diagnosis before committing to any procedure.

Epidurolysis for Scar Tissue

Some patients, particularly those who have had prior spinal surgery, develop scar tissue (epidural adhesions) that can trap or compress nerves and block the spread of injected medications. If a steroid injection cannot physically reach the nerve root because scar tissue is in the way, it will not work regardless of whether the diagnosis is correct.

Epidurolysis, sometimes called lysis of adhesions, is a procedure that breaks down this scar tissue, either by threading a catheter through the epidural space or by using a small endoscope. Prospective studies, including randomized trials, have found that the technique improves pain and function in a meaningful fraction of patients. One randomized controlled trial of spinal endoscopic adhesiolysis found significant improvement in 80% of treated patients at three months, dropping to about 48% at twelve months.¹² A larger analysis found that roughly half of patients experienced a positive outcome overall.¹³ This is not a home run by any standard, but for patients with confirmed epidural adhesions who have exhausted other options, it can be a reasonable step before considering major surgery.

Spinal Cord Stimulation

For patients with chronic, intractable radicular pain that has not responded to conservative treatment or simpler procedures, spinal cord stimulation is an option that has generated both enthusiasm and controversy. The device delivers electrical pulses to the spinal cord to interrupt pain signals before they reach the brain.

Some studies report impressive results. One trial of high-dose spinal cord stimulation in patients with non-surgical lumbar radiculopathy found that back pain scores dropped from a baseline of about 7.5 out of 10 to roughly 2.8 at one month and 3.9 at twelve months, with improvements in disability and sleep as well.¹⁴ Newer closed-loop systems that automatically adjust stimulation based on real-time spinal cord feedback have shown promise in individual cases, with one patient going from a pain score of 7 out of 10 down to 1 and dramatically improved walking and standing tolerance.¹⁵

The picture gets murkier with higher-quality evidence, though. A randomized trial comparing burst spinal cord stimulation to placebo stimulation in patients with chronic radicular pain after lumbar surgery found no significant difference between the two groups on disability, leg pain, back pain, daily steps, or time spent upright. About 18% of patients experienced adverse events, including some requiring surgical revision of the implanted device.¹⁶ That is a sobering finding and a reminder that not all spinal cord stimulation technology is equal, and that some of the positive results in unblinded trials may partly reflect placebo effects or other biases.

On cost, a cost-effectiveness analysis found that spinal cord stimulation compared favorably to conventional medical management for chronic low back pain, though the initial device and procedure cost of around $30,000 means it takes roughly two to three years before the investment breaks even, assuming sustained benefit and reduced downstream healthcare use.¹⁷ An earlier analysis in failed back surgery patients estimated the cost per quality-adjusted life-year at about £5,600, with a high probability of being cost-effective compared to both continued medical management and reoperation.¹⁸ In practice, most insurers require a successful temporary trial before approving permanent implantation, which gives you a chance to test whether it works for you before committing.

Platelet-Rich Plasma as an Alternative Injectate

One newer approach flips the injection concept on its head. Instead of replacing the steroid with a different drug, platelet-rich plasma (PRP) uses a concentrated sample of your own blood platelets, which contain growth factors thought to promote healing rather than simply suppress inflammation.

A meta-analysis of randomized controlled trials comparing epidural PRP to epidural steroid injections in patients with lumbar disc disease and radiculopathy found comparable outcomes between the two. Pain scores, disability scores, and overall health improvements were statistically similar at all observed time points, with no increase in adverse events from PRP.¹⁹ A randomized pilot study told a slightly more optimistic story, finding that PRP outperformed steroids at six weeks on pain, disability, and quality of life measures. At six months, 56% of PRP patients still showed improvement compared to 39% in the steroid group.²⁰

PRP is still early in its evidence journey for spinal applications. It is not covered by most insurance plans and can cost several hundred dollars out of pocket per injection. Still, for patients who cannot tolerate repeated steroid injections due to side effects like elevated blood sugar, bone thinning, or adrenal suppression, PRP offers an intriguing alternative that at minimum appears to do no worse.

When Surgery Becomes the Right Call

Surgery is not necessarily the last resort it is often painted as, but it is not clearly superior to injections either, at least for disc herniations. The NERVES trial, a large randomized study comparing microdiscectomy to transforaminal epidural steroid injection in patients with sciatica from a herniated disc, found no statistically significant difference in disability scores between the two groups at 18 weeks or at any point up to 54 weeks.²¹ The estimated advantage of surgery was about 4 points on the disability questionnaire, which fell short of the 10-point threshold considered clinically meaningful. Surgery also carried a higher risk of serious adverse events, including one case of foot drop, and was not cost-effective at standard willingness-to-pay thresholds.²²

That said, these results apply specifically to disc herniations where the question is surgery versus steroid injection. For spinal stenosis, large unstable herniations, or progressive neurological deficits like worsening leg weakness or bowel and bladder changes, surgery may be clearly indicated regardless of whether injections were tried first.

One timing consideration worth knowing about: if surgery does become the plan, the window between your last steroid injection and the operation matters. A study found that patients who received lumbar epidural steroid injections within six weeks before microdiscectomy had nearly double the odds of needing reoperation within six months compared to those who had a longer gap or no injection.²³ The reasons are not fully understood but may involve steroid effects on tissue healing. If you are considering surgery after failed injections, discuss the timing with your surgeon.

The Case for Multidisciplinary Care

Chronic back pain that does not respond to a single treatment modality often responds better to a team approach. Multidisciplinary rehabilitation, which combines physical therapy, psychological support, and medical management under one coordinated plan, has moderate evidence for a meaningful treatment effect on chronic low back pain. Patients with higher disability, greater functional impairment, and psychological complications like anxiety or depression tend to respond best to this type of program.²⁴

The practical barrier is access. These programs require coordinated teams, frequent visits, and infrastructure that not every community has. But even a less formal version of the concept, say, working with both a physical therapist and a psychologist who specializes in pain management while your physician manages medications, captures some of the same philosophy. The core insight is that chronic pain is rarely purely mechanical. By the time steroid injections have failed and you are months or years into the problem, the pain has almost certainly recruited emotional, behavioral, and neurological amplifiers that no single injection or pill can fully address.

How Expectations Shape Outcomes

This is an area that often gets dismissed as “it’s all in your head,” but the research says something more nuanced. A meta-analysis of expectation-based interventions found that they produced a medium-sized effect on pain relief overall, with the strongest effects on acute and procedural pain and smaller but still real effects on chronic pain.²⁵ What patients expect from a treatment genuinely changes how much relief they experience, through measurable changes in how the brain processes pain signals.

This does not mean you should trick yourself into believing something works when it does not. But it does mean that a clinician who takes time to explain what a treatment can realistically accomplish, who builds confidence in the process, and who addresses fear and catastrophic thinking is not wasting time on bedside manner. They are delivering an active ingredient. If your experience with failed steroid injections has left you demoralized and skeptical of everything, that emotional state is itself a treatment barrier worth addressing, whether through cognitive behavioral therapy for pain, a conversation with a provider you trust, or simply getting a second opinion that restores some sense of direction.