Ultrasound Guided Injection vs. a Blind Needle: How Often It Misses

When a shoulder or spine injection does nothing, the chart usually records that the treatment failed. It rarely records where the needle went. That gap is the subject of Chapter 13 of The Pained Brain, the video above, by Dr. Gurpreet Singh Padda, MD, MBA, MHP, and Dr. KrisJay Fucanan, MD. The short version: an ultrasound guided injection, or a fluoroscopy-guided one in the spine, lets the physician see the target. A needle placed by feel is aimed at where the target usually sits. Those are different acts, and the difference decides what your next treatment will be.

How accurate is an ultrasound guided injection compared with feel?

The most direct test put a camera inside the knee right after a landmark injection. In 162 patients headed for arthroscopy, 72.8 percent of the needles were in the joint. Accuracy climbed from 53.2 percent in the least experienced hands to 87.0 percent in the most experienced, and it dropped in patients with a body mass index of 30 or more.

Pooled across knee studies, ultrasound reached the joint 95.8 percent of the time against 77.8 percent for landmarks. At the hip, the pooled figures were 100 percent against 72. The widest gap sits at the biceps tendon sheath in the front of the shoulder: 86.7 percent with ultrasound, 26.7 percent by feel. In our clinic, that swollen tendon is the structure most often found untouched after a series of blind shoulder shots.

At the subacromial space under the point of the shoulder, pooled accuracy was 65 percent with ultrasound against 70 percent without, no real difference, yet the guided group still reported better pain and function at six weeks.

Then there is the body the landmark is drawn on. The line across the top of the hip bones, used by touch to find the fourth lumbar space, sat where it should in 75.3 percent of people and in only 34 percent of people with obesity. Extra weight moves the map a blind needle relies on.

The miss that nobody feels

The worrying part is that the person holding the needle usually cannot tell which injections missed. In 106 consecutive blind shoulder injections read by a blinded radiologist, providers could not predict their own accuracy regardless of experience, and the experienced ones tended toward overconfidence. At the anterior approach, the experienced provider landed 50 percent of injections against 85.7 percent for the less experienced one.

In a randomized trial in inflamed joints, a trainee with an ultrasound probe was accurate 83 percent of the time against 66 percent for senior rheumatologists working by palpation. The probe users could also judge their own accuracy reliably. The palpators could not.

A missed injection is not coded as a miss. It is coded as a failed treatment, and the next rung is usually a bigger procedure. Fee-for-service payment covers the needle either way, which is part of how patients end up with a ninth epidural and nothing built between them. In a randomized trial of 94 arthritic knees, sonographic guidance raised the responder rate by 107 percent and cut the cost per responder by 58 percent, although the guided arm also used a special syringe.

Why a steroid can hide a miss and a diagnostic block cannot

A corticosteroid spreads through tissue and into the bloodstream, so a near miss can still help. In 105 arthritic knees given landmark injections, 70.5 percent were in the joint, and whether a needle landed made no measurable difference to who responded at three or nine weeks. Across 19 shoulder trials, Cochrane reviewers found guidance improved pain by only 0.5 points on a ten-point scale.

That is exactly why steroid trials are the wrong test of precision. A diagnostic block exists to answer one question: is this joint or nerve generating the pain? If the anesthetic spreads to the neighbors, the answer is contaminated. At the cervical medial branch, a quarter milliliter and a half milliliter both reached the nerve, but the larger volume more than doubled spread to adjacent levels, from 16 to 38 percent. Light sedation more than doubled the rate of positive blocks at a 50 percent relief threshold. Sedation inflates the yes, so our practice performs diagnostic blocks on awake patients, who typically go home within two to four hours. Why a pain generator has to be proven rather than assumed is its own argument.

The blood vessel an aspiration test misses

Before injecting, many physicians pull back on the plunger and look for blood. In 761 lumbosacral transforaminal injections under fluoroscopy, contrast entered a vessel 11.2 percent of the time, and blood in the hub or on aspiration caught only 44.7 percent of those. In 344 lumbar medial branch blocks, live fluoroscopy caught vascular uptake in 11 percent, and digital subtraction imaging found another 7.8 percent that live imaging had missed.

The first is a false-negative block: anesthetic that drained into a vein, and a facet joint declared innocent when it was guilty. The second is rare and catastrophic. A particulate steroid that enters a small artery feeding the spinal cord can clump red cells and plug the smallest vessels. The safeguards that followed, radiographic guidance, contrast and non-particulate steroid, now define careful practice, and they are the basis of our ultrasound-guided and fluoroscopic procedures. Across 16,638 consecutive image-guided epidurals at three centers following society guidelines, there were no immediate major adverse events and a dural puncture rate of 0.06 percent. Sacroiliac joint injections carry the same lesson, since a blind needle there was inside the joint 22 percent of the time.

Two blocks before an ablation

Radiofrequency ablation of a facet or sacroiliac nerve can only work if the block that selected you was accurate. In a randomized trial of patients with suspected facet pain, success among those who went on to ablation rose from 33 percent when no block was used to 64 percent when comparative blocks selected them. The same trial carries the honest opposite finding: per patient randomized, skipping blocks was the cheapest way to buy a success, because blocks screen out some people who would have responded. The cost per success with a single block ran nearly three times the no-block figure.

At the knee, before cooled radiofrequency ablation, a prognostic genicular block did not improve outcomes: 58.6 percent success with it and 64.0 percent without. Our practice still blocks twice at the spine, because a coin-toss block followed by an ablation tests the procedure on the wrong patient. How facet joint pain is traced with medial branch blocks walks through that process.

What to ask before your next injection

A failed injection is evidence about a needle before it is evidence about you. Bring these questions:

• Will this be done under imaging, and which kind? Ultrasound watches soft tissue move under the probe; fluoroscopy proves a spinal needle sits where the picture says. What each imaging tool can and cannot show differs by structure.
• Will contrast go in under live imaging before the drug?
• If this is a diagnostic block, will I be awake, and how much volume will you use?
• What is the plan for the weeks of relief?

An injection is a bridge. The relief it buys is a window for slower work, because tendon, muscle and joint tissue rebuild in response to load, and the metabolic terrain they live in shapes how well that repair goes. Humility belongs here too: image guidance has not beaten blind injection in every steroid trial, and novices needed a mean of 28 supervised attempts to become competent with a probe in one cadaver study. What to do once the damaged tissue is visible is the question behind removing a tendon lesion instead of jamming the nerve signal. Every study and number here, with what each one does and does not show, is in the Technical Supplement to Chapter 13, written to be handed to your physician.

Frequently asked questions

Is an ultrasound guided injection more accurate than a regular injection?

At most joints, yes. Pooled knee studies put ultrasound at 95.8 percent accuracy against 77.8 percent for landmarks, and at the biceps tendon sheath the gap was 86.7 percent against 26.7. The subacromial space is the exception, where accuracy was similar. With a steroid, accuracy does not always change the outcome because the drug spreads, but it matters a great deal for diagnostic blocks and biologic injections. Frozen shoulder care shows how an injection fits inside a wider plan.

Why didn’t my cortisone shot work?

There are several possibilities, and a missed target is one of them. A blind needle can land outside the joint or tendon sheath, and the injector usually cannot tell. The shot may also have reached its intended spot while the pain came from a different structure, such as a tendon rather than a bursa. Or the relief was real but brief, which is the nature of steroid. Here is why each injection can buy less time than the one before.

Are fluoroscopy guided injections safer for the spine?

Live imaging with contrast is how a physician detects a needle inside a blood vessel, which aspiration misses more than half the time. In 761 transforaminal injections, contrast showed vessel entry in 11.2 percent. The national safeguards adopted after rare spinal cord injuries call for radiographic guidance, contrast and non-particulate steroid, and across 16,638 image-guided epidurals no immediate major adverse events occurred. This is what an epidural steroid injection actually does to an inflamed nerve root.

What is a diagnostic nerve block?

A diagnostic nerve block places a small volume of local anesthetic on one nerve or joint to test whether it is generating your pain. If the pain drops, that structure becomes the target for treatment such as ablation. Volume and sedation both distort the answer: larger volumes spread to neighboring levels, and light sedation more than doubles the rate of positive results. That is why the patient should be awake. The medial branch nerves that serve the facet joints are a frequent target of this test.

How many diagnostic blocks do I need before radiofrequency ablation?

Guidelines differ. The cervical guideline accepts a single block at 50 percent relief to preserve access to care, while comparative blocks raised success among patients who went on to ablation from 33 to 64 percent in one randomized trial. Before sacroiliac joint fusion, guidelines report two blocks with at least 75 percent relief. Our practice blocks twice. How radiofrequency ablation works, and why the setting matters, is explained here.

Sources

1. Altuntas, Y., Ipek, E., Balkanlı, B., Alibakan, G., & Eren, O. T. (2026). Arthroscopic Assessment of Landmark-Guided Knee Injection Accuracy: A Prospective Observational Study. American Journal of Physical Medicine & Rehabilitation. Epub ahead of print, 2026-06-05. https://doi.org/10.1097/PHM.0000000000003057
2. Aly, A.-R., Rajasekaran, S., & Ashworth, N. (2015). Ultrasound-guided shoulder girdle injections are more accurate and more effective than landmark-guided injections: a systematic review and meta-analysis. British Journal of Sports Medicine, 49(16), 1042–1049. https://doi.org/10.1136/bjsports-2014-093573
3. Malik, M., & Ismail, S. (2019). Accuracy of Tuffier’s Line Identification by Palpation Method: Cross-Sectional Comparative Study Among Obese, Pregnant and Control Groups. Turkish Journal of Anaesthesiology and Reanimation, 48(2), 108–114. https://doi.org/10.5152/TJAR.2019.82346
4. Tobola, A., Cook, C., Cassas, K. J., Hawkins, R. J., Wienke, J. R., Tolan, S., & Kissenberth, M. J. (2011). Accuracy of glenohumeral joint injections: comparing approach and experience of provider. Journal of Shoulder and Elbow Surgery, 20(7), 1147–1154. https://doi.org/10.1016/j.jse.2010.12.021
5. Cunnington, J., Marshall, N., Hide, G., Bracewell, C., Isaacs, J., Platt, P., & Kane, D. (2010). A randomized, double-blind, controlled study of ultrasound-guided corticosteroid injection into the joint of patients with inflammatory arthritis. Arthritis and Rheumatism, 62(7), 1862–1869. https://doi.org/10.1002/art.27448
6. Sibbitt, W. L., Band, P. A., Kettwich, L. G., Chavez-Chiang, N. R., Delea, S. L., & Bankhurst, A. D. (2011). A randomized controlled trial evaluating the cost-effectiveness of sonographic guidance for intra-articular injection of the osteoarthritic knee. Journal of Clinical Rheumatology, 17(8), 409–415. https://doi.org/10.1097/RHU.0b013e31823a49a4
7. Furman, M. B., O’Brien, E. M., & Zgleszewski, T. M. (2000). Incidence of intravascular penetration in transforaminal lumbosacral epidural steroid injections. Spine, 25(20), 2628–32. https://doi.org/10.1097/00007632-200010150-00014
8. El-Yahchouchi, C. A., Plastaras, C. T., Maus, T. P., Carr, C. M., McCormick, Z. L., Geske, J. R., Smuck, M., Pingree, M. J., & Kennedy, D. J. (2016). Adverse Event Rates Associated with Transforaminal and Interlaminar Epidural Steroid Injections: A Multi-Institutional Study. Pain Medicine, 17(2), 239–49. https://doi.org/10.1111/pme.12896
9. Hurley, R. W., Adams, M. C. B., Barad, M., Bhaskar, A., Bhatia, A., Chadwick, A., Deer, T. R., Hah, J., Hooten, W. M., Kissoon, N. R., Lee, D. W., Mccormick, Z., Moon, J. Y., Narouze, S., Provenzano, D. A., Schneider, B. J., van Eerd, M., Van Zundert, J., Wallace, M. S., … Cohen, S. P. (2022). Consensus practice guidelines on interventions for cervical spine (facet) joint pain from a multispecialty international working group. Regional Anesthesia and Pain Medicine, 47(1), 3–59. https://doi.org/10.1136/rapm-2021-103031
10. Cohen, S. P., Williams, K. A., Kurihara, C., Nguyen, C., Shields, C., Kim, P., Griffith, S. R., Larkin, T. M., Crooks, M., Williams, N., Morlando, B., & Strassels, S. A. (2010). Multicenter, randomized, comparative cost-effectiveness study comparing 0, 1, and 2 diagnostic medial branch (facet joint nerve) block treatment paradigms before lumbar facet radiofrequency denervation. Anesthesiology, 113(2), 395–405. https://doi.org/10.1097/ALN.0b013e3181e33ae5

Medically reviewed by Dr. Gurpreet Singh Padda, MD, MBA, MHP, board certified in anesthesiology, pain medicine, interventional pain management, addiction medicine and obesity medicine. Adapted from Chapter 13 of The Pained Brain by Dr. Padda and Dr. KrisJay Fucanan, MD. Last reviewed September 2026.

Dr. Gurpreet Singh Padda, MD, MBA, MHP