Steroid injections to treat arthritis pain may worsen joint damage

"Cortisone injections for hip and knee pain are more dangerous than was thought," reports The Telegraph.

The headline follows a study looking into the effects of steroid injections (also called corticosteroid injections) in the hip and knee.

Steroid injections in joints are intended to relieve joint pain by reducing inflammation. In the UK, you may be offered this type of injection if you have moderate to severe pain from osteoarthritis.

Doctors in the US reviewed results from a group of 459 patients who had a steroid injection into their hip or knee joint during 2018. They found 8% of patients had problems after an injection, including worsening pain and breakdown of cartilage in the joint.

They warn that steroid injections are "perhaps not as safe as we thought" and say that people should be warned about the possibility that a steroid injection might make their joint symptoms worse.

This research raises questions about how thoroughly the safety of steroid injections into hip and knee joints has been assessed, and about whether these injections are suitable for everyone.

However, this study has too many limitations to form any reliable conclusions. Large-scale, long-term studies are needed to give us an accurate picture of the possible risk of joint damage after injections.

Find out more about treatments for osteoarthritis

Where did the story come from?

The researchers who carried out the study were from Boston University School of Medicine in the US.

The study was published in the peer-reviewed medical journal Radiology on an open-access basis and is free to read online.

It was covered by several media outlets. Metro focused on potential risks to elite athletes, warning they "may suffer devastating long-term effects" from steroid injections. However, the treatment is used far more widely to treat joint pain in older people with osteoarthritis, and we do not know whether any of the people in the study were elite athletes.

The Telegraph included a section on rheumatoid arthritis, which is a different disease from that focused on in the study.

Mail Online provided a mostly accurate and balanced report, including expert comment that potential complications affected a minority of people.

What kind of research was this?

This was not a formal cohort-study, as the researchers did not do "before and after" scans of all patients, but only arranged for a scan of patients who returned to the hospital for further investigation or treatment after an earlier injection.

This means we cannot put much trust in the figures, as the study was not done according to a standard research protocol.

What did the research involve?

Doctors reviewed patient notes and scan (radiography) results, where available, from 459 patients who had received steroid injections in 2018 (307 hip injections and 152 knee injections). They looked for signs of worsening joint damage on radiography or MRI scans after the injections had been done.

They considered signs of accelerated progression of osteoarthritis, bone damage below the joint surface, complications from death of bone cells, and destruction of the joint, including bone loss.

The doctors described the joint problems they found in their patients after steroid injections, and reviewed the previous studies published on the subject.

What were the basic results?

Of the 459 people who had injections:

  • 218 had no follow-up joint scans
  • 30 of 307 (10%) people who received injections into their hip had a joint problem afterwards
  • 6 of 152 (4%) people who received steroid injections into their knee had a joint problem afterwards

The most common problem was accelerated progression of osteoarthritis (26 people), indicated by reduced joint space between the bones because of cartilage loss.

The researchers also found bone damage below the joint surface (4 people), complications from death of bone cells (3 people), and destruction of the joint through bone loss (3 people).

The 36 patients (8%) who had joint problems after the injection ranged from 37 to 79 years of age and had between 1 and 3 injections. Most (72%) had moderate osteoarthritis of the knee or hip.

The researchers said that they did not know whether the problems they found were caused by the injections, or if they were already happening but not obvious on scans when the injections were given.

How did the researchers interpret the results?

The researchers said there was insufficient research into the use of steroid joint injections in the hip or knee to be sure of their safety.

They said: "We believe that certain patient characteristics… should lead to careful reconsideration of a planned IACS [a corticosteroid injection into a joint]."

The characteristics they mention included pain that cannot be explained by radiographic images, no sign of osteoarthritis, or signs of mild osteoarthritis on scans. They also suggest more people should have radiographic or MRI images taken before joint injections, to be sure they do not have existing bone weakness that could be made worse by the injection.


Steroid injections into joints can be a useful way to help to manage joint pain for people with conditions such as osteoarthritis. They have been widely used for years. This study raises questions about how well their safety has been assessed in past research, and about whether they are suitable for everyone.

The study does not tell us that the injections were the cause of the joint damage found. As the researchers say, it is possible that the joint damage was already underway, but not visible at the time the injections were done.

The patients already had osteoarthritis or joint pain when they were referred to the hospital where the study was done. We do not know what would have happened to their joints if they had not received the injections.

The study does not give us an accurate picture of the possible risk of joint damage after steroid injections. This is because not everyone in the study had images taken before and after injections, and images taken after injections were only taken when and if the patient returned to the hospital, not at a set time according to a study protocol.

However, the study does demonstrate that some people who have joint injections go on to have joint damage at a faster rate than expected, which might be linked to the injection. We need large-scale, long-term studies to find out whether the injections are the cause of the damage, and if so, how common this problem is.

NHS Attribution