Will a Cortisone Shot Delay My Knee Replacement?

Researchers have proposed two mechanisms. A needle crosses the skin and the lining of the joint, which breaches the joint’s physical barrier. Cortisone also suppresses the local immune response while the medication remains in the knee.⁴ Neither mechanism has been directly proven as the cause in patients. The evidence on this question comes from comparing infection rates, not from measuring what happens inside the joint.

The 3-month window

The study most often cited is a 2019 analysis in The Journal of Bone & Joint Surgery. After adjusting for age, sex and other health conditions, the odds of infection within 6 months of surgery were 1.21 times higher with a cortisone shot in the prior 3 months and 1.55 times higher with a gel shot. Injections given more than 3 months before surgery showed no added risk.²

Pooled analyses point the same way. Across 14 studies and about 370,000 patients, a cortisone shot within 3 months raised the odds of infection by 26%.⁵ An analysis of hip and knee replacements found a 39% rise within 3 months and no significant rise from 3 to 6 months.⁶ A European review reported a 52% rise within 3 months.⁷

What that means in absolute numbers

A 26% relative increase sounds larger than it is. If the baseline infection rate is about 1 in 100, a 26% increase raises it to about 1.26 in 100, or roughly one additional infection for every 400 knee replacements. That is my own arithmetic from the pooled figure, and the baseline rate varies widely across databases. One study with a low baseline reported a larger gap: 1.6% with a cortisone shot within 90 days compared with 0.41% without, or about one additional infection for every 85 patients.⁸ Even at the upper end of these estimates, the large majority of patients who had a recent shot did not develop an infection.

Is there a safe number of weeks?

Three months is the most common cutoff, partly because researchers tend to group data into 3-month blocks. Within that window, the risk appears to be highest closest to surgery. One large database study found higher infection risk with an injection within 4 weeks of surgery, and no added risk in the group injected 4 to 6 weeks beforehand.⁹ Another study of gel shots found the highest risk within 1 month, some increase from 1 to 2 months, and no difference beyond 2 months.¹⁰

The studies that disagree

Not every study finds this link. A study of Medicare patients found 2-year infection rates of 0.75% with a cortisone shot within 3 months and 0.77% with no injection.¹¹ A 2025 study of 43,219 patients found no significant difference with an injection in the prior 90 days,³ and a national cohort from England found no rise in complications after knee replacement in patients who had earlier cortisone shots.¹² One 2023 meta-analysis found a link for hip replacement but none for knee replacement.¹³

Gel shots, repeat shots and the other knee

Gel shots do not appear to avoid the problem. In the 2019 study, gel shots within 3 months carried at least as much risk as cortisone.² A 2025 study of 38,546 patients found three times the odds of infection at 3 months after surgery when a gel shot was given within 1 month.¹⁰

The medication may not be the only factor. A 2025 database study looked at procedures without cortisone, such as a fluid draw or a platelet-rich plasma (PRP) injection. Infections requiring surgery became more common the closer the procedure was to the knee replacement: 7.6% within 1 month, 4.6% within 3 months and 2.2% within 6 months, compared with 1.4% without one.⁴ These rates are higher than in most other studies, so I would not rely on the exact figures, but they suggest that the needle itself may matter.

The number of shots does not appear to matter as much as the timing. A Medicare study of 371,511 knee replacements found no link between the number of injections in the prior 2 years and infection within 90 days,¹⁴ and a smaller study found the same for injections given more than 90 days before surgery.¹⁵

A shot in the other knee is a different question. A 2026 study compared 195 patients who had a cortisone shot in the opposite knee on the day of their knee replacement with 195 who did not, and found no significant difference in complications.¹⁶ That group is too small to detect a difference in an event as rare as joint infection, but the finding is reassuring for patients with arthritis in both knees.

What this means for you: if a knee replacement is likely in the next few months, the timing of any injection is worth discussing before you have it. If you have already had a recent shot, the added risk is small, and you have time to plan around it.

• These are associations, not proof of cause. A patient who needs a shot shortly before surgery may differ from one who does not, in ways that billing records do not capture. Propensity matching helps, but it cannot account for what was never recorded.
• The data cannot assess how the injection was given. Skin preparation and sterile technique are not in insurance records, and these factors could plausibly matter as much as the timing.
• Some conclusions go further than the data. One meta-analysis concludes that cortisone injection “is not a safe procedure” for patients expected to have a knee replacement, although its overall result (odds ratio 1.13) was of borderline significance and its signal was limited to the 3-month window.⁵ Conversely, the 2025 study reporting no added risk had 93.6% of its injections given 61 to 90 days before surgery, so it says little about injections in the last 2 months.³
• The 4-week finding is narrower than its headline. The study titled as showing no benefit from waiting past 4 weeks looked only as far as 4 to 6 weeks before surgery. It did not test a longer wait.⁹
• The type of steroid has not been settled. One study found significant increases with methylprednisolone and betamethasone but not with triamcinolone or dexamethasone.⁸ The dexamethasone group had only 54 patients. I would not choose a medication on the basis of this one study.

The evidence suggests that the timing of an injection should be part of the surgical plan. That is how I approach the conversation with patients who are weighing an injection against a knee replacement.

1. I ask about every injection

Patients often receive injections from more than one office. Before scheduling surgery, I need to know which knee was injected, what was injected and the date. This includes cortisone, gel, PRP and any fluid drawn from the knee. Your records from the other office are helpful, and an approximate date is better than none.

2. I ask how close you are to surgery before an injection

If a patient has reached the point of considering a knee replacement and expects to go ahead soon, a shot may buy a few weeks of relief and also push back the surgery date. If surgery is a year or more away, or may not be needed, an injection remains a reasonable part of non-surgical care, and these timing concerns carry much less weight.

3. We plan the months in between

When a patient has had a recent shot and is ready for surgery, the evidence favors a wait of about 3 months where that is practical. Physical therapy, activity changes and appropriate oral medication can help during that time. The individual decision also weighs how much pain and disability the patient has, and that discussion is specific to each patient and their other risk factors, such as dental health, diabetes and smoking. In the 4-week study, tobacco use further increased the risk of infection.⁹

Who this evidence fits, and who it doesn’t

Closest fit: a patient having a first total knee replacement who has had a cortisone or gel shot in the same knee and has flexibility in scheduling.

Less certain fit: a patient planning a partial knee replacement or a hip replacement, since these studies focused on total knee replacement or combined different joints; a patient whose shot was in the other knee; and a patient with a higher baseline risk of infection, for whom the same relative increase means more in absolute terms.