How often can you get a cortisone shot?

This is the most common question I get after giving an injection, usually asked as the patient is standing up. It has a short answer that is mostly right, and a longer one that is more useful.

Where does the three-month rule come from?

Convention, largely. The figure most often quoted — no more than one injection per joint every three months, and no more than three or four in a year — is long-standing clinical practice rather than a threshold that emerged from a trial designed to find it. It is a sensible default built on caution about cumulative steroid exposure, and it is applied fairly uniformly across joints that are not actually equivalent.

That does not make it wrong. A conservative interval that limits total exposure is a reasonable way to practise when the alternative is injecting on demand. But it is worth knowing that the number is a professional norm rather than a measured safety limit, because it explains why you will hear it stated with more confidence than the evidence strictly supports.

What does the evidence say about repeating them?

The most informative study is the one that has changed how many clinicians think about this. It randomised 140 people with knee osteoarthritis and ultrasound-confirmed synovitis to triamcinolone or saline every twelve weeks for two years, and found significantly greater cartilage volume loss in the steroid group — with no significant difference in pain.

Read carefully, that is not an argument against injections. It is a specific argument against injecting on a fixed schedule indefinitely. The comparison was not one injection versus none; it was two years of quarterly injections versus two years of quarterly saline.

What it establishes is that the three-month interval, treated as a routine to be followed rather than a floor not to breach, is not a neutral choice. If a joint is being injected every twelve weeks for years, the question is no longer how often — it is whether this is the right treatment at all.

Does it matter if surgery is coming?

Yes — and this is where the blanket advice is least accurate, because the evidence differs by joint.

A meta-analysis of 28 studies covering more than 480,000 patients found a significant association between preoperative corticosteroid injection and periprosthetic joint infection for hip replacement, with the risk higher when the injection fell within three months of surgery. The same analysis found no association for knee replacement.

Two knee-specific studies point the same way. A single-surgeon series of 442 patients found no relationship between the number or timing of injections and infection or complications after knee replacement. A Medicare claims analysis found injections in the four months before knee replacement were not associated with elevated infection risk at one, three, twelve or twenty-four months.

Either way, the actionable step is identical: tell whoever is planning your surgery that you have had an injection, and when. That one sentence lets them make the call with the right information.

When frequency is the wrong question

Any of these mean the interval is not what needs discussing:

• Each one lasts less than the one before. Shortening duration usually means the underlying problem has outgrown the treatment.
• Nothing else has changed between injections. If the joint returns to identical loading, identical strength and identical activity, the injection is postponing rather than enabling.
• The last one did essentially nothing. A genuine non-response is information — often that the pain is mechanical rather than inflammatory.
• Diabetes that is difficult to control. Corticosteroid raises blood glucose, sometimes substantially and for several days, which changes the risk calculation — and is one reason hyaluronic acid comes up as an alternative.

None of those are absolute contraindications. They are signals that the conversation should be about the plan rather than the interval.