Shoulder Injections, Often Inexactly Placed, Deemed 'Poor Tool' for Pain

Both diagnostic and therapeutic injections for shoulder pain are often placed inaccurately, even when clinicians are convinced they've injected the subacromial bursa rather than other structures, reported Dr. Hans-Erik Henkus of Medisch Centrum Haaglanden, the Hague, and his associates.

They found that only 66% of subacromial injections were accurate in a series of 33 patients, suggesting that the technique is “a poor tool” for diagnosing the source of shoulder pain and “worrisome” as a therapeutic strategy, the researchers said.

Rather than easing pain and restoring function, inaccurate injections worsen both.

Injecting a local anesthetic around the shoulder area is a widely used method for determining the source of the pain and for predicting the success of subacromial decompression surgery. Injecting corticosteroids is done to reduce inflammation and pain. Both strategies are “based on the assumption that these injections can be given with great accuracy,” Dr. Henkus and his associates wrote (Arthroscopy 2006;22:277–82).

But few studies have examined the accuracy of shoulder injections.

The investigators assessed the technique in 33 patients with nontraumatic shoulder pain localized to the deltoid region. The 11 men and 22 women, with an average age of 46 years, were unable to lie on the affected side. Abduction, retroversion, or internal rotation of the glenohumeral joint against resistance provoked further pain.

With the subjects in an upright seated position, an experienced orthopedic surgeon injected all the shoulders with a mixture of bupivacaine, methylprednisolone, and a contrast agent. The surgeon was randomly assigned to approach either anteromedially (16 cases) or posteriorly (17 cases). An MRI was performed immediately to determine the location of the infiltration.

The subacromial bursa was accurately targeted in only 22 (66%) of cases, even though the surgeon was confident that the injections had been accurate in 30 cases (91%) and “doubtful” in only 3.

Three injections infiltrated only the deltoid muscle and subcutaneous tissue, two the glenohumeral joint only, two the acromioclavicular joint only, and three the rotator cuff only.

Even when the subacromial bursa was correctly targeted, many surrounding tissues were infiltrated as well. The rotator cuff was infiltrated 13 times, the deltoid muscle 3 times, and the coracoacromial ligament 2 times.

Pain was reduced and function improved in the cases in which the subacromial bursa alone was injected. However, pain increased and function declined or showed no change when other structures, particularly the rotator cuff, were infiltrated.

“The rotator cuff muscle or tendon was hit in 17 patients,” a “worrisome” incidence that could well have caused rotator cuff rupture, particularly if corticosteroids had been injected, the investigators noted.

Neither the type of approach (anterior or posterior) nor the patient's body mass index had any influence on the accuracy of injection placement.

Given that a single faulty injection to any of a variety of structures could produce either a false-positive or a false-negative result, these findings indicate that “the diagnostic use of local injections in the subacromial bursa [is] a poor tool,” they said.