Shoulder pain is one of the most common reasons for a joint injection and one of the easiest to get wrong, because several different structures produce pain in roughly the same place and the treatment target differs for each.
Getting it right is an examination problem before it is a technical one.
The shoulder is two joints plus the cuff
Gisele J. Girault, MD sets out the anatomy that determines everything downstream. The shoulder comprises two joints — the acromioclavicular joint and the glenohumeral joint — with the rotator cuff muscles surrounding them.
That distinction is not academic. AC joint pain is superficial and localises to the top of the shoulder. Glenohumeral pain is deeper and more diffuse. Subacromial bursitis and cuff tendinopathy produce pain in the lateral upper arm, often worse overhead. Three different targets, three different injections.
Her instruction is unambiguous: “before you do any injections into the shoulder, you need to know the anatomy of the shoulder.”
The syndromes that present here
Girault lists what shoulder pain in practice actually turns out to be:
- Rotator cuff tears
- Rotator cuff tendonitis or tendinosis — the distinction matters, since tendinosis is degenerative rather than inflammatory
- Subacromial bursitis
- Osteoarthritis of the AC joint or the glenohumeral joint
These coexist routinely. A patient with subacromial bursitis frequently has cuff tendinosis, and a degenerate AC joint alongside is common at older ages. Which one is generating the current pain is the question the examination answers.
The examination decides the target
Girault requires “a very thorough shoulder exam looking at range of motion, also palpation of the shoulder.”
What each finding points to:
- Point tenderness over the AC joint, with pain on cross-body adduction — AC joint
- Painful arc through mid-range abduction, worse overhead, with tenderness below the acromion — subacromial bursitis or cuff impingement
- Weakness on resisted testing rather than pain alone — raises the question of a tear
- Global loss of range, including passive external rotation — adhesive capsulitis or glenohumeral arthritis, and a capsular problem rather than a subacromial one
That last one is the distinction most often missed. A stiff shoulder that will not passively externally rotate is not a bursitis, and a subacromial injection will not help it.
Tendonitis and tendinosis are not the same injection decision
Girault distinguishes them explicitly, and the reason is practical.
Tendonitis is inflammatory. Tendinosis is degenerative — disorganised collagen without significant inflammation. Corticosteroid is an anti-inflammatory, so in a genuinely degenerative tendon it is treating a process that is not the dominant problem, and repeated steroid around degenerate tendon carries its own risk.
This is where the regenerative options become a reasonable conversation rather than a fashionable one. The published evidence in tendon is the strongest part of that category — a single PRP injection outperformed corticosteroid in lateral epicondylitis with benefit maintained at two years, and adhesive capsulitis showed steroid better short-term with PRP better long-term. Those results are discussed in regenerative pain medicine.
Which steroid, and when to leave it out
For non-spinal joint work the particulate constraint that governs spinal injection does not apply in the same way, so intermediate-potency options are available. Girault's visual test still holds — triamcinolone and methylprednisolone are cloudy, which is undissolved particles in suspension; dexamethasone is clear because it is non-particulate.
The patient factor that changes the decision most often is diabetes. Where control is poor, her approach is to omit the steroid rather than cancel the injection, since several of these can be performed “with local anaesthetic alone, leaving out the steroid, because steroids can increase blood sugars.”
Before injecting steroid in a diabetic patient she wants haemoglobin A1c and fasting glucose under good control.
Landmark or ultrasound?
Shoulder injections are routinely performed on landmarks, and Girault teaches them that way — based on palpation and anatomical landmarks, with ultrasound or fluoroscopy available where the anatomy is difficult.
The reasonable position is that guidance improves accuracy of placement, which matters most for the glenohumeral joint and least for a subacromial injection in a slim patient with clear landmarks. Accuracy and outcome are not the same measure, and the accuracy evidence is set out in musculoskeletal ultrasound for pain practice.
Afterwards
Warn about steroid flare before the patient leaves. Increased pain over the following 24 to 48 hours occurs in roughly one in ten injections, settles, and is not an emergency — though the office should be called.
Relative rest for a day or two, then return to movement. A shoulder that is not moved stiffens, and in a patient with any capsular component that is the outcome to avoid.
The injection buys a window. What happens in that window — rehabilitation, load management, addressing the cause — determines whether the result lasts.
When injection is not the answer
A shoulder with weakness rather than pain on resisted testing needs the tear question answered before repeated injection. A shoulder with global loss of passive range needs the capsule addressed. And a shoulder injected three times without lasting benefit needs a different diagnosis rather than a fourth injection.
Learn blocks with your hands, not from a page
Empire’s Pain Management Training (THE Pain Show) is accredited for 25.25 AMA PRA Category 1 Credits™, jointly provided by AKH, Inc, and Empire Medical Training. For narrower peripheral work, Joint, Extremity and Non-Spinal Injection Training carries 6.75 credits for the complete in-person hybrid program, and Advanced Musculoskeletal Ultrasound Guided Injections builds the guidance skills above.
Claim Your Prize!