Frozen Shoulder

Longer than most people want to hear. The classic teaching is 1 to 3 years through freezing, frozen, and thawing phases, but long-term data pushes that timeline out. Wong and colleagues (Physiotherapy 2017) systematically reviewed the natural history and concluded that treatment produced some, but not complete, improvement over one to four years, and that the tidy idea of a clean recovery phase is not well supported. Hand and colleagues (JSES 2008) followed patients on average 4.4 years and found 41% reported some ongoing symptoms, though most were mild. Physiotherapy matched to your phase can shorten the course and improve how much function you recover, but I am careful not to promise a quick fix.

Most people regain enough range for daily life, but returning to exactly the motion you had before is not certain. Hand and colleagues (JSES 2008) followed patients for an average of more than four years and found 59% had normal or near-normal shoulders while 41% reported some ongoing symptoms, most of which were mild. My functional targets are pain-free sleep, full overhead reach for daily tasks, and the ability to dress and groom without compensation. Getting the last 10 degrees of end-range rotation back can take many months of consistent work.

The evidence supports them for short-term pain relief, particularly in the painful freezing phase. Cochrane reviews and multiple randomised trials show that intra-articular corticosteroid injection reduces pain and improves function more quickly than physiotherapy alone in the first 6 to 12 weeks. After that window, differences between groups narrow and long-term outcomes look similar. The best results I see are when an injection calms the pain enough for someone to actually tolerate the mobility work. Benefits beyond 6 months are less certain, and the decision is about timing and access, not a permanent fix.

Night pain is one of the most recognisable features of frozen shoulder and it has two main drivers. The inflamed capsule becomes more sensitised when you lie still and blood flow pools in the tissues, and any roll onto the affected side directly compresses an already irritated joint. I usually coach patients to sleep on the unaffected side with pillows propping the sore arm forward, or on their back with a small towel roll behind the shoulder blade. Night pain tends to ease as the condition moves out of the freezing phase.

For most people, no. The UK FROST trial (Rangan et al., Lancet 2020) compared early structured physiotherapy, manipulation under anaesthesia, and arthroscopic capsular release in over 500 patients. At 12 months, none of the three was clinically superior on shoulder pain and function, and capsular release carried the highest complication rate. Physiotherapy is a reasonable first-line option, and I reserve surgical referral conversations for people who have genuinely plateaued after 6 to 9 months of well-delivered conservative care and still cannot meet their daily needs. Surgery carries its own risks and rehabilitation burden.

Recurrence in the same shoulder is uncommon, which is reassuring. The other shoulder becomes involved in roughly 5 to 34% of cases over several years in observational follow-up, with higher risk if you have diabetes or thyroid disease. There is no proven prevention protocol. The practical steps are keeping both shoulders moving through full range regularly, managing underlying metabolic conditions, and addressing any shoulder pain early rather than guarding it for weeks.

Not in the freezing phase. When the capsule is actively inflamed, aggressive end-range stretching tends to flare symptoms, disturb sleep, and stall progress. I use a simple traffic-light rule: stay in the amber zone of tolerable stretch, avoid the red zone of sharp or lingering pain. Once you move into the stiff phase, the tissue tolerates and often requires more vigorous, sustained end-range loading to regain motion. Matching intensity to the phase you are in matters more than how hard you push on any given day.

Yes, and it is worth knowing. People with diabetes are significantly more likely to develop frozen shoulder, the condition often affects both shoulders, and recovery is generally slower. Zreik and colleagues' 2016 meta-analysis estimated prevalence near 13% in people with diabetes compared to the 2 to 5% typically cited in the general population, with diabetic patients roughly five times more likely to develop the condition. I flag this at assessment because it shifts expectations on timeline, and keeping blood glucose well-controlled is sensible general care even if it will not on its own fix the shoulder.