Loading ...

At the recent World Congress on Osteoarthritis (OARSI) 2025 Annual Meeting in Incheon, South Korea, speaker after speaker presented on studies that were carefully designed to identify even the smallest benefit of interventions, including drugs, physical therapy, exercise therapy, and even surgery, compared with placebo. And in many cases, the benefit wasn’t there.

It’s not that the interventions didn’t achieve anything. Most did show improvements in pain and even function. But the problem was that the placebo did as well. In control groups, patients given in some cases the barest minimum of care — a pamphlet and advice — and in other cases everything up to and including sham ultrasound and faked surgery showed improvements in pain and function that defied orthopedic explanation.

This isn’t a new problem — a trial published more than two decades ago found no additional benefit of knee arthroscopy in people with knee osteoarthritis compared to placebo surgery. Nor is the problem limited to osteoarthritis; there’s evidence for a strong placebo effect in numerous chronic musculoskeletal conditions, including tennis elbow, shoulder pain, meniscal tears, and low back pain.

Placebo is now understood to be far more than the stereotypical sugar pill; the term is now used to encompass a range of factors under the headings of “contextual” and “nonspecific” effects. Contextual effects describe factors such as the patient themselves, their relationship with their healthcare provider, and the setting in which care is provided, while nonspecific effects relate to features like the natural history of the disease and natural fluctuations in disease severity.

But as the burden of osteoarthritis increases globally with increasing life expectancy and an aging population, more clinicians and researchers are seeking to understand why placebo effects dominate in osteoarthritis and what it means for research and treatment of a condition that affects more than 500 million people worldwide.

Defining Osteoarthritis

The most widely used classification system for the presence and severity of osteoarthritis has been in use for more than 60 years.

In 1957, when rheumatologists Jonas Kellgren and John Lawrence proposed their system to describe and rate radiographic evidence of what they called osteoarthrosis, they looked for a range of pathologic characteristics on x-ray, including the formation of osteophytes and narrowing of the joint cartilage.

If none of these features were present, then the Kellgren and Lawrence grade was 0, which meant no evidence of osteoarthritis. A grade of 2 indicated the presence of disease, albeit of minimum severity, and a grade of 4 was severe disease.

One of the ongoing challenges for clinicians and researchers trying to improve symptoms is that not all people who meet the Kellgren and Lawrence definition of osteoarthritis have symptoms, and not all who have clinical symptoms of osteoarthritis meet the Kellgren and Lawrence definition.

This complicates both research into and treatment of osteoarthritis. It suggests that there are a number of factors contributing to the experience of pain in osteoarthritis that have little or nothing to do with the currently understood disease pathophysiology — the contextual and nonspecific effects. And that makes it challenging to tease out what are genuine treatment effects associated with an intervention from what are placebo effects.

“You’ve got people who are running around with just self-described bone-on-bone knees, and they’re still managing to be out and about and active, and there are people with less severe structural changes who are less active and in more pain,” said rheumatologist Anita Wluka, MBBS, PhD, a consultant rheumatologist at The Alfred Hospital and professor of rheumatology at the University of Melbourne, both in Melbourne, Australia. “There are other factors that contribute to the pain that are not necessarily related to the structural damage, and I think a lot of those fall into that pain perception and things that affect pain perception.”

Furthermore, around 10% of people who receive a total knee replacement report dissatisfaction with the procedure, mostly due to ongoing pain and stiffness. But with a knee replacement, the underlying pathology — supposedly the cause of their symptoms — has been surgically removed, said pain specialist Apkar Apkarian, PhD, director of the Center for Translational Pain Research at Northwestern University in Chicago.

“Now, the nociceptive input is completely gone [but] years later, they’re still unhappy with the surgery, so we should be confident to say that the nociceptive signal is not enough to explain pain,” Apkarian said.

Apkarian argues that osteoarthritis research has focused on the joint for far too long and should instead be taking a closer look at the brain to understand why some feel pain and others don’t, which could then shed light on what role the placebo effect plays in response to treatment in this disease. “The brain circuitry, brain emotional circuitry, brain learning circuitry, all of those are massively, causally engaged in chronic pain, and we need to become serious and start targeting them,” he said.

Persistent Placebo

Which brings us to the placebo effect in osteoarthritis and why so many individuals with the disease report sometimes significant benefits from placebo treatments that are designed specifically not to work.

Neuroscientist Luana Colloca, MD, PhD, director of the Placebo Beyond Opinions Center at the University of Maryland, Baltimore, was giving a talk on the placebo effect at a major US clinic and was struck by the observation that the clinic was able to successfully treat so many seemingly intractable rheumatology cases. She asked for their secret and was told that it wasn’t to do with the intervention, but with the interaction between the clinician and patient. “The therapeutic alliance for a rheumatological disease has such an impact,” Colloca said. “You get many patients that continue to go [to meet the doctor] because of this powerful therapeutic alliance, so definitely [it] is a good trigger to produce placebo effects.”

The therapeutic benefit of an interaction with a health professional is well documented and can be significant. For example, in a study of morphine for postoperative pain, patients were found to need less morphine for pain relief when it was administered by a clinician compared with “hidden” automated administration.

“I’ve been saying for 30 years that there’s no benefit of exercise therapy, but there’s no doubt that seeing a caring physio probably helps to a small degree,” said rheumatologist and epidemiologist Rachelle Buchbinder, MBBS, PhD, of Monash University, Melbourne, Australia. That therapeutic benefit can come from addressing the patient’s fears and concerns verbally or, in the case of one study on low back pain, simply accompanying them on a short walk to help them overcome their fear of movement.

That compassion or attention is therapeutic, particularly when it both validates the patient’s symptoms and reassures them that they will get better, said orthopedic surgeon Teppo Järvinen, MD, PhD, of the University of Helsinki, Helsinki, Finland. “That’s what seems to be happening with most chronic musculoskeletal pain conditions and also depression, anxiety — [they] seem to be quite well amenable to just care and attention and empathy.”

Wluka said the more intensive the intervention, the greater the placebo effect. “The more invasive the placebo, the higher the placebo response: The sham surgeries and the joint injections, or if there must be an injection to the bottom versus a tablet,” she said.

For example, the 2013 FIDELITY trial — which Järvinen was involved in — compared arthroscopic partial meniscectomy with sham surgery for degenerative meniscal tear. In it, both the intervention and control groups reported around a 21-23-point improvement in their Lysholm knee score — a measure of pain and function — from a baseline of around 60 points, which was well above the minimum clinically important improvement of 11.5 points. But there was no significant difference between the two groups in how much benefit they derived.

Another recent trial used sham ultrasound and massage on unaffected joints to evaluate the effect of physical therapy for meniscal tears and found both the real and sham physical therapy were associated with significant and similar improvements in pain scores.

Another contributor to the large placebo effect seen with chronic musculoskeletal conditions is that they tend to self-resolve over time, Buchbinder said. “When you’re thinking about a treatment, what you have to consider is that it’s got to be better than natural history,” she said. “That’s why nearly everything we do for acute pain doesn’t work because it can’t beat that really favorable natural history.”

Buchbinder and colleagues published a systematic review and meta-analysis looking at the control arms of clinical trials in tennis elbow and found an exponential rate of resolution over time, such that every 2.5-3 months, half of the remaining symptomatic patients improved partially or completely. By 1 year, 89% of patients reported global improvement of their symptoms.

Unfortunately, persistent tennis elbow symptoms are often used as justification for surgical intervention. “Surgeons commonly say, if you’re not better by a certain time, then you need surgery, and that’s nonsense because the natural history is that you’ll still get better irrespective of how long you’ve had symptoms,” Buchbinder said. “By 1 year, 90% are better anyway, so the chances that [surgery] is going to beat that are pretty tiny.”

Placebo in the Clinic and Clinical Trial

In 2016, a group of researchers conducted a clinical trial of placebo for chronic low back pain. Patients who had had persistent low back pain for at least 3 months were randomly assigned either to an open-label placebo or no additional treatment, on top of whatever treatment they were already having for the condition. The patients who were given a placebo were clearly told it was a pill that contained no active medication.

Astonishingly, those given the placebo reported moderate to large reductions in pain and significantly greater reductions than those not given anything. “They knew it was a placebo, and they wanted to know where they could buy it, so that’s how effective it was,” Buchbinder said.

At OARSI 2025, the question was asked as to whether it was possible to make ethical use of placebo in clinical practice.

Wluka said she would find it difficult to give a patient something she knew was completely inert. “I don’t know how I’d sell that; I would have trouble,” she said. But for something that is relatively benign, for which there is evidence of placebo effect, she said she would frame it as “a lot of people find benefit” rather than “it could help” “because I don’t feel like I’m perjuring myself; I’m giving them an option.”

There could be an advantage to patients taking a benign treatment such as supplements if it then relieves their symptoms enough to enable other interventions that might help. “If you can get them on board moving forward with some medication, but that gives them the capacity to do a bit more exercise or to start working on their muscle strength around the effective joint, then that’s actually going to move them forward,” Wluka said.

But that flexibility should not extend to more invasive interventions, Järvinen said. “We actually do the ultimate violation of privacy and entering bodily cavities of human beings to do allegedly something good, and we end up finding out that we haven’t done a whole lot more than just the placebo or contextual effects,” he said.

One solution could be to identify which patients are more likely to respond to placebo. Apkarian’s research suggests that in general about half of patients respond to placebo, and half don’t. One finding has been that people who practice yoga seem more receptive to placebo, perhaps because they are more in tune with and in control of their body, he said.

The advantage of finding patients who are more receptive to placebo means they could be offered less intensive or invasive treatments, “or can I say to someone else, ‘Oh, you have a really nociceptively driven type of osteoarthritis, and so we should treat your knee properly.’”

With relatively few interventions showing any significant benefit over and above placebo for osteoarthritis, Buchbinder said it’s important to be upfront with patients about the limitations of the range of treatments being offered. “We have to be honest about what we can and we can’t do,” she said. But alongside that is the evidence that most of the time, the condition will be self-limiting. “It’s not like pain from cancer, which is going to progress — these are things that are going to get better.”

None of the expert sources interviewed for this article had any relevant conflicts of interest or financial disclosures.