If you've Googled epidural steroid injections recently, you may have seen some alarming headlines. A major study in The BMJ (February 2025) and an updated systematic review from the American Academy of Neurology (AAN) both questioned the effectiveness of common interventional spine procedures, including epidural steroid injections (ESIs).
Some headlines were blunt: Epidural injections don't work for back pain.
As an interventional pain physician who performs these procedures — and who has seen them provide significant relief for well-selected patients — I want to set the record straight. Not by dismissing the research, but by explaining what it actually says, what it gets wrong, and why the headline version could genuinely hurt patients.
What the BMJ Study Actually Found
In February 2025, Wang et al. published a systematic review and network meta-analysis (NMA) in The BMJexamining common interventional procedures for chronic spine pain. The accompanying clinical guideline (Busse et al.) recommended against most interventional spine procedures, suggesting that “the evidence does not support their routine use.”
That sounds damning. But here's what happened beneath the surface:
The study lumped vastly different procedures together.The meta-analysis combined epidural steroid injections with radiofrequency ablation, facet joint injections, intradiscal procedures, and trigger point injections — as if they're interchangeable. It pooled data across cervical, thoracic, lumbar, and sacral spine regions. It mixed standard techniques with abandoned and non-standard procedures that are rarely performed in actual clinical practice.
This is like studying “surgery” by combining appendectomies, knee replacements, and heart transplants into one analysis, then concluding “surgery has modest benefits.”
34 medical societies formally disagreed. A multisociety response — signed by organizations including the International Pain and Spine Intervention Society (IPSIS), the North American Spine Society, the American Academy of Pain Medicine, and 31 other professional organizations — identified critical methodological flaws. Their key concerns:
- Extreme heterogeneity: The analysis pooled procedures, conditions, and anatomical regions that are clinically distinct. A sacroiliac joint pulsed radiofrequency treatment and a cervical medial branch neurotomy are not the same procedure.
- Omission of key studies: The strongest randomized controlled trial (RCT) supporting lumbar transforaminal epidural steroid injections was excluded from the analysis.
- Inappropriate aggregation: The authors acknowledged that separating studies by procedure and condition would have left too few studies for meta-analysis — but proceeded to pool them anyway, generating conclusions that no practicing clinician could meaningfully apply.
To be fair, conducting meta-analyses of interventional spine procedures is genuinely difficult. The conditions, techniques, and patient populations vary enormously, and the BMJ authors acknowledged these limitations. But the clinical conclusions drawn from such broadly pooled data should be interpreted with great caution. As the multisociety response stated: “Interventional spine procedures reduce pain, improve function, and help patients return to work.”
What the AAN Review Actually Said
The AAN published its updated systematic review in Neurology(February 2025), replacing its 2007 guidelines. Here's the nuanced conclusion that didn't make headlines:
- ESIs “may be modestly effective in improving pain and disability for up to three months” for cervical and lumbar radicular pain.
- ESIs “might be modestly effective in improving disability for up to six months or more” for lumbar radiculopathy.
- For lumbar spinal stenosis, the evidence was less clear.
Notice the language: “modestly effective.” That's not “don't work.” That's a measured scientific statement acknowledging that ESIs provide meaningful — if not dramatic — improvement. And importantly, the AAN was evaluating ESIs across all patients and all conditions, not the carefully selected patients who benefit most.
The Evidence That Supports ESIs
Here's what the skeptical headlines tend to leave out:
A 2024 systematic review in Frontiers in Neurology analyzed ESI efficacy specifically for sciatica secondary to lumbar disc herniation. The conclusion? Multiple RCTs and high-quality observational studies provide evidence supporting ESI effectiveness for pain reduction compared to placebo.
The NCBI StatPearls review (updated 2024) summarized the evidence across 70 studies: “Evidence for efficacy was good for lumbar disc herniations, fair for spinal stenosis, and poor for failed back surgery syndrome.” That's not “doesn't work” — that's a nuanced picture where the right diagnosis makes all the difference.
The ASIPP (American Society of Interventional Pain Physicians) comprehensive guidelines reviewed the evidence extensively and found Level I–II evidence supporting epidural interventions for disc herniation and radiculopathy, particularly with transforaminal and caudal approaches using fluoroscopic guidance.
A 2025 analysis of 72 RCTs involving 7,701 patients examined epidural steroid injections specifically for lumbar disc herniation and found meaningful evidence of efficacy.
The Key the Headlines Missed: Patient Selection
Here's the truth that every interventional pain physician knows: ESIs work when they're given to the right patients for the right indications.
An epidural steroid injection for a patient with an acute lumbar disc herniation compressing a nerve root, causing radiating leg pain, with correlating MRI findings? That's a well-indicated procedure with good evidence of benefit.
The same injection for a patient with chronic nonspecific low back pain, no radiculopathy, and no clear structural cause? That's poorly indicated, and no, it probably won't help much.
The studies that show modest or no benefit often fail to distinguish between these two scenarios. They lump well-selected and poorly-selected patients together, diluting the real treatment effect.
It's also worth noting that many older studies — including some pooled in the BMJ review — used blind (non-image-guided) injection techniques. Modern best practice requires fluoroscopic or CT guidanceto confirm precise needle placement. Studies using image-guided techniques consistently show better outcomes than blind approaches. When you read a study showing ESIs “don't work,” ask whether the injections were performed with image guidance — it matters enormously.
In my practice, patient selection is everything. Before I recommend an ESI, I need:
- A clear diagnosis — disc herniation, foraminal stenosis, radiculopathy
- Correlating symptoms — leg pain following a specific nerve distribution
- Supporting imaging — MRI findings that match the clinical picture
- Failed conservative treatment — physical therapy, medications, and time have been given a fair chance
When these criteria are met, I see patients get significant relief — often enough to participate in physical therapy, return to work, and avoid surgery.
What ESIs Actually Do (and Don't Do)
Let me be honest about what epidural steroid injections are:
They are not a cure.They don't fix a herniated disc. They don't reverse spinal stenosis. They reduce inflammation around irritated nerves, which reduces pain, which creates a window for healing and rehabilitation.
They are a bridge.The goal is to reduce pain enough that patients can engage in physical therapy, stay active, and allow their body's natural healing processes to work. Many disc herniations improve on their own — but the pain during that healing period can be disabling. ESIs manage that pain.
They work best as part of a multimodal plan. An injection alone, without physical therapy, activity modification, and addressing underlying factors, is a missed opportunity. The injection opens the door; the rehab walks through it.
They are not for everyone. Patients with central spinal stenosis (as opposed to foraminal stenosis), failed back surgery syndrome, or chronic axial low back pain without radiculopathy are less likely to benefit. Knowing who not to inject is as important as knowing who to inject.
Why This Matters for You
If you're a patient considering an epidural steroid injection, here's my advice:
Don't let headlines make medical decisions for you.
The studies raising questions about ESIs have legitimate methodological concerns, and the clinical reality is more nuanced than any headline suggests.
Ask your doctor about patient selection.
A good pain physician should be able to explain why they believe an ESI will help you specifically — not just that “we try this for back pain.”
Expect a multimodal approach.
If your doctor recommends an ESI and nothing else — no physical therapy, no follow-up plan — get a second opinion.
Understand the goal.
ESIs are about creating a window of relief that enables recovery. They're a tool, not a standalone treatment.
The evidence supports epidural steroid injections for well-selected patients with radicular pain from disc herniation and foraminal stenosis. The evidence is weaker for spinal stenosis and nonspecific back pain. A blanket statement that “ESIs don't work” is not supported by the totality of the literature — and 34 medical societies agree.
A Note on Transparency
As an interventional pain physician, I perform ESIs as part of my practice. I want to be upfront about that. I've tried to present the evidence fairly — including its limitations — because my goal is to help you make informed decisions, not to sell you on a procedure. The 34 medical societies that responded to the BMJ study also include organizations whose members perform these procedures, which is relevant context. Ultimately, the best evidence we have supports ESIs for specific, well-defined conditions — not as a blanket treatment for all back pain.
Written by Dr. Varun Patibanda, M.D., D.A.B.P.M. This article is for educational purposes and does not constitute medical advice. Always consult your healthcare provider about your specific condition and treatment options.