Frozen Shoulder in a Diabetic Patient: Corticosteroid, Hydrodilatation, or Neither?
Intra-articular corticosteroid is the best-evidenced fast option for adhesive capsulitis — in a patient whose diabetes makes even a targeted injection a real, if smaller, glycemic event to weigh against a mechanical alternative.
Consuela R., a 61-year-old woman who cooks a full Sunday dinner for her grandchildren every week without fail, has had progressive right shoulder stiffness and night pain for three months, now limiting her ability to reach the top shelf of her own kitchen cabinets — something she describes with real frustration, since that weekly dinner is what she looks forward to most. She has type 2 diabetes managed on metformin and glipizide. Exam confirms adhesive capsulitis in its freezing phase: passive and active external rotation and abduction are both markedly restricted, more than would be expected from pain alone, with a firm end-feel rather than the guarding of a simple rotator cuff strain. Her most recent A1c is 7.8%, reflecting moderately controlled but not tightly optimized diabetes, and she checks her fasting glucose most mornings on her own glucometer.
That A1c changes the calculus more than a patient's diabetes label alone would suggest, because even the more targeted of the pharmacologic options carries a real, if smaller, systemic footprint. Intra-articular corticosteroid injection has the most consistent short-term evidence of any single intervention for adhesive capsulitis — multiple randomized trials, summarized in Buchbinder and colleagues' Cochrane review of corticosteroid injection for shoulder pain, show meaningful pain and function improvement in the freezing phase specifically — but intra-articular steroid still produces measurable transient hyperglycemia in diabetic patients for several days afterward, a well-documented effect even at doses far below an oral course. Hydrodilatation — capsular distension performed under fluoroscopic or ultrasound guidance with saline, local anesthetic, and typically a lower steroid dose than a standalone injection — has its own placebo-controlled support from Buchbinder and colleagues' 2004 trial, which found distension with saline and steroid improved function and reduced pain against a placebo procedure. The comparison that would actually settle Consuela's question is thinner than that: Gam and colleagues' 1998 randomized trial is one of the few to test distension plus glucocorticoid directly against glucocorticoid alone, and the head-to-head literature since has stayed small and underpowered rather than resolving it. What can be said without overclaiming is that the distension arm delivers its steroid at a lower total dose while adding a mechanical effect aimed at capsular volume — which matters here because her limitation is disproportionately about lost motion, not pain intensity. Oral glucocorticoids, meanwhile, offer real but short-lived pain relief in the trial literature at a systemic dose that would meaningfully destabilize her glucose control for the full course, not just a few days — the option every voice in today's discussion agrees to rule out before the conversation even properly starts.
Orthopedic clinic, three months of shoulder stiffness
I'd start with a standard intra-articular corticosteroid injection. It has the most consistent trial evidence of any single intervention for the freezing phase — Buchbinder and colleagues' Cochrane review summarizes multiple randomized trials showing real, meaningful improvement in both pain and function. Yes, it will raise her glucose transiently, typically resolving over several days, but that's a real, bounded cost against the best-evidenced option we have.
I don't dispute that intra-articular steroid is the best-evidenced single option in the literature broadly, and I'm not asking us to avoid glucocorticoids in her entirely.
But her limitation isn't really about pain — she can tolerate the pain. It's the lost motion that keeps her off the top shelf. Buchbinder's own 2004 trial showed distension with saline and steroid beating a placebo procedure on both function and pain, and Gam and colleagues tested distension plus glucocorticoid against glucocorticoid alone back in 1998. I'll concede the head-to-head literature is thin and I'm not going to overclaim it as settled. But distension delivers less total steroid and adds a mechanical effect aimed straight at capsular volume, which is what's actually limiting her — and it reduces, rather than accepts, the glycemic exposure you just described as merely bounded.
Whichever of those two we land on, I think we're all already agreed that oral glucocorticoids are off the table here — a systemic course, at the dose shown to produce real pain relief in the trial literature, would destabilize her glucose for the entire course, not the few days either injectable option costs her.
Given that her own complaint centers on motion rather than pain intensity, I'd side with hydrodilatation — it directly addresses her actual limitation while asking less of her glucose control than a standard injection alone.
Hydrodilatation scheduled under ultrasound guidance, with fasting glucose monitoring advised for one week afterward given the lower but nonzero steroid component. Physical therapy for range of motion begun immediately following the procedure rather than delayed.