The "-itis" names lie. Once the pain is more than a few weeks old there is little inflammation left in the tissue — what you find is a tendon that lost its tight collagen structure and is repairing slower than you're loading it ¹. So you can't anti-inflame your way out. The tendon needs a different signal, and that signal is load ².

The cortisone shot backfires. Across three trials in tennis elbow it looked best at six weeks and worst at a year: one found about 54% recovered with the shot versus 78% with a placebo injection, and adding physiotherapy didn't rescue it ^{3 4 5}. The likely trap: it kills the pain, so you load the tendon straight back into trouble.

Slow loading does the opposite. A daily eccentric wrist exercise added to standard care beat standard care alone ⁶, and progressive loading is the endorsed cornerstone across tendinopathies ⁷. If it hurts too much to move, holding a load still calms the pain for hours ⁸.

The wrist is the exception. De Quervain's, the thumb-side one, responds to a single well-placed shot: about five in six cured at six weeks versus one in seven with a splint alone ⁹. One shot, not three.

Turn the aggravating activity down, don't stop it. Then load the tendon daily.

Holds quiet the pain in a day or two. The loading work usually shows an obvious shift by week three or four ⁶: grip returns, the kettle and the jar stop being a problem, people stop asking about your arm. By months two and three the sport or the job comes back at graded volume. Most people recover within one to two years anyway ¹⁰; the point is to get there faster and skip the shot that leaves a chunk of people worse off.