If you live with knee osteoarthritis, you probably know the routine. Stiff for the first few steps out of bed. A sharp reminder on the stairs. Some swelling after a long day on your feet. Most people who come to me for knee pain have already tried the usual list: ibuprofen, a brace, physical therapy, maybe a cortisone shot. Some of it helped. Some of it stopped helping.
So when a large, carefully designed study on acupuncture for knee osteoarthritis comes out, I read the whole thing. This one was published in May 2026 in eClinicalMedicine, a journal from The Lancet group. I want to walk you through what the researchers actually did, what they found, where the limits are, and how I think about it as an acupuncturist who treats knee pain every week in Stow and Belmont.
Who Was in the Study
The trial ran at six hospitals in Shanghai between late 2021 and late 2024. The researchers screened 686 people and enrolled 480. To qualify, you had to be:
- Between 50 and 75 years old
- Diagnosed with knee osteoarthritis that showed up on X-ray as mild-to-moderate (what doctors call Kellgren–Lawrence grade 2 or 3)
- Living with knee pain for at least six months, rated above 4 and up to 7 out of 10
- Dealing with at least some knee swelling, confirmed by ultrasound
- Not having had acupuncture in the previous three months
The average participant was 57, and about three out of four were women. One detail I noticed: the average body mass index was about 24, which is in the normal-weight range. That matters, and I’ll come back to it.
People couldn’t use pain pills, injections, or other knee treatments during the study. An anti-inflammatory was allowed as a backup for severe pain, but no one ended up needing it.
What the Treatment Looked Like
Half the group got real electroacupuncture. The acupuncturists, all with at least three years of clinical experience and standardized training, used eight points chosen by expert consensus: Xuehai (SP10), Liangqiu (ST34), Dubi (ST35), Neixiyan (EX-LE4), Zusanli (ST36), Yanglingquan (GB34), Sanyinjiao (SP6), and Xuanzhong (GB39). These will be familiar to any acupuncturist who treats knees. They’re spread around the knee and down the lower leg.
Needles went in about half an inch to an inch and a quarter deep (13–30 mm). The acupuncturist lifted, thrust, and rotated each needle a few times to bring on de qi, that dull, heavy, achy sensation that tells us the needle has engaged. Then a gentle electrical current was connected at a slow 2 Hz rhythm for 25 minutes.
The other half got sham treatment. Blunt needles were pressed against small foam pads on the same points, tapped and twisted to feel like real needling, but never broke the skin. The same electrical device was hooked up with the same settings.
Both groups came in three times a week for six weeks, 18 sessions in all, and were followed until week 30, about six months after the last treatment.
The participants, the people scoring the results, and the statisticians didn’t know who got which treatment. At the end, the researchers asked participants to guess which group they were in, and the guesses were close to random. That’s a good sign the sham was convincing.
What They Found About Acupuncture for Knee Osteoarthritis
The main measure was the WOMAC, a standard knee arthritis questionnaire. It scores pain, stiffness, and how hard everyday things are, like stairs, getting out of a chair, or putting on socks. It runs from 0 to 240, and lower is better. Here’s how the two groups changed over the six weeks of treatment:
| Measure | Real electroacupuncture (start → week 6) |
Sham (start → week 6) |
|---|---|---|
| WOMAC overall score (0–240, lower is better) | 155 → 89 | 155 → 130 |
| Pain rating (0–10) | 6.7 → 3.1 | 6.5 → 5.1 |
| WOMAC stiffness (0–20) | 13.1 → 7.0 | 13.1 → 10.1 |
| WOMAC daily function (0–170) | 109 → 66 | 109 → 92 |
| Improved enough to really notice it | 223 of 240 (92.9%) | 163 of 240 (67.9%) |
In plain terms, the average person getting real acupuncture went from a pain level of about 7 out of 10 to about 3. That’s a big change for someone who’s been hurting for years.
A few other things stood out to me:
- It worked fairly quickly. Pain and function were already clearly better in the real-treatment group by week 3, halfway through. Stiffness took a bit longer and separated by week 6.
- It lasted. The advantage was still there at week 30, about six months after the needles stopped.
- The effect was large. Researchers use a measure called Cohen’s d to describe how big a difference is. Anything over 0.8 is considered large. For the overall WOMAC score it was 1.21, and for the 0–10 pain rating it was 1.98. Effects that size are uncommon in pain research.
- Almost everyone stayed in the study. Only three people were lost to follow-up out of 480.
- It was safe. Five percent of the real-treatment group reported a side effect. The ones related to treatment were small bruises (four people) and a bit of skin numbness (three people). All were mild and went away on their own, and nobody dropped out because of them.
The MRI Findings: Interesting, but Early
This is the part that’s getting a lot of attention, so I want to be careful with it.
Everyone had a knee MRI at the start and again at week 6. The researchers used an AI-based program to measure very small changes in the joint.
The clearest signal was in inflammation. In the real-treatment group, there was less fluid and swelling in the joint lining. The fat pad just below the kneecap, which often gets inflamed in knee arthritis, also looked calmer. These differences were moderate in size, not huge, but real.
The cartilage findings were much smaller. In four of the eight cartilage measurements, the real-treatment group basically held steady while the sham group lost a tiny amount. The other four showed no difference. The researchers themselves say these effects were very small, the imaging window was only six weeks, and the study wasn’t designed to answer questions about joint structure. They describe the findings as early signals, not evidence that acupuncture changes the course of the disease.
The authors do make one point I found interesting. The sham group felt better too, but their MRIs didn’t show the same changes. That suggests the imaging changes in the real-treatment group weren’t just the result of expecting to feel better. It’s a hypothesis worth testing, not a conclusion.
So I’m not going to tell you acupuncture rebuilds cartilage. What I can say is that this study adds to good evidence that acupuncture helps people with knee arthritis hurt less and move better. It also may be calming inflammation in the joint along the way.
A Fair Word About the Sham Group
You probably noticed that two out of three people in the sham group also improved enough to notice. That’s common in pain research. Coming in three times a week, lying down quietly, being cared for, and expecting to feel better all have real effects on how we experience pain. The researchers even suggest their sham may have worked more like an active treatment than a true placebo.
I don’t see that as a knock against acupuncture for knee osteoarthritis. The real treatment still did clearly better, and the difference held for months. It’s also a reminder that how care is given matters. Time, attention, and a calm space are part of healing, not extras.
What This Study Can’t Tell Us
Every good study has limits, and I think you deserve to know them:
- The acupuncturists knew which treatment they were giving. That’s almost impossible to avoid in needle research, but it’s worth knowing.
- Treatment was free, and people got a travel allowance. That helped keep almost everyone in the study, but it’s different from real life, where cost and time get in the way.
- Some planned measurements didn’t happen. The researchers had planned to look at gait, muscle strength, and inflammation markers in the blood and joint fluid, but COVID disruptions and other practical problems meant they couldn’t collect enough data to report them.
- The walking test was a modest result. On the six-minute walk, the real-treatment group held about steady while the sham group walked a bit less by week 6. The difference favored acupuncture, but it was small compared to the pain and function results.
- The results were much bigger than expected. The researchers planned for a 9-point difference on the WOMAC and found a 40-point difference. That’s encouraging, but results this large should be repeated by other research teams, ideally in other countries and with more varied patients.
- The participants were fairly lean and all had some swelling. We don’t know for sure that the results would be the same for everyone with knee arthritis.
How This Fits With What I See in My Practice
A trial like this has to use the same points on every person so the results can be measured. That’s how good research works. In my practice I take a different approach and treat each person individually. Two people can both have “knee arthritis” and need very different treatments. It depends on whether the knee is hot and swollen or cold and achy, whether the pain is on the inside or the outside of the joint, and how they’re sleeping, digesting, and feeling overall.
Still, this is one of the better trials on acupuncture for knee osteoarthritis I’ve read, and two things in it shape how I talk with patients.
The first is frequency. The researchers chose three sessions a week partly because earlier research found that acupuncture three to five times a week worked better for knee arthritis than once or twice a week. Many people come in once a week, or once every few weeks, and wonder why change is slow. When knee pain is really getting in the way of your life, a more concentrated stretch of treatment at the start can make a real difference. Then we space visits out as things settle.
The second is swelling. Everyone in this study had some swelling in the knee, and the MRI changes centered on inflammation. If your knee tends to puff up after activity, this research is especially relevant to you. If there is swelling I will often not use a local treatment in the area of the swelling and instead use a distal treatment to reduce swelling first. Often this distal treatment, known as Saam Acupuncture, can often reduce not only swelling but pain symptoms.
Where Nutrition Fits In
I’m also a licensed dietitian, and with knee osteoarthritis I rarely think about needles alone.
Remember that the people in this study averaged a normal body weight. Many of my patients here in Massachusetts are carrying more weight than that, and every step puts several times your body weight through the knee. For people carrying extra weight, even a modest loss takes meaningful pressure off the joint. Beyond weight, what you eat affects inflammation throughout your body. I usually look at vegetable and fiber intake, the kinds of fats you eat, enough protein to support the muscles around the knee, and blood sugar control.
Acupuncture can calm pain and swelling so you can move more. Good nutrition and stronger muscles help that improvement stick. They work better together than either one does alone.
Is Acupuncture for Knee Pain Right for You?
Acupuncture for knee osteoarthritis may be a good fit if:
- You have knee arthritis and pain that limits your walking, stairs, or sleep
- Medications aren’t helping enough, or you’d like to rely on them less
- Your knee tends to swell, stiffen, or ache after activity
- You’re doing physical therapy and want something to calm the pain so you can keep up with it
Acupuncture isn’t a replacement for a medical evaluation. If your knee is suddenly hot, red, very swollen, or you can’t put weight on it, see your doctor first. And if you take blood thinners, let me know so we can plan accordingly.
Frequently Asked Questions
Does acupuncture actually work for knee osteoarthritis?
Good-quality research says it can help. In this 480-person trial, real electroacupuncture lowered average pain from about 7 out of 10 to about 3 after six weeks, compared with about 5 for sham treatment. The benefit was still there six months after treatment ended. An earlier large trial published in Arthritis & Rheumatology in 2021 also found lasting benefits from intensive acupuncture for knee osteoarthritis.
How many acupuncture sessions do I need for knee pain?
In this study, people had three sessions a week for six weeks, 18 in total. Pain and function were noticeably better by week 3. In practice, I tailor the plan to each person and adjust how often we meet based on how you respond.
What is electroacupuncture, and does it hurt?
Electroacupuncture adds a gentle electrical current to acupuncture needles. Most people feel a light tapping or pulsing sensation. It shouldn’t be painful, and the intensity can be turned down at any time. In this study, the only treatment-related side effects were minor bruising and brief skin numbness.
Can acupuncture rebuild knee cartilage?
There’s no good evidence for that yet. This study found small, early MRI signals that cartilage held steady and that joint inflammation went down, but the researchers say those findings need to be confirmed in longer studies built to measure joint structure. The strongest evidence is for less pain and better function. I also use infrared red light therapy to help rebuild cartilage. It helped rebuild mine.
Where can I get acupuncture for knee pain near me?
I offer acupuncture for knee osteoarthritis and other joint pain in two locations. I see patients at InnerSpring WellBeing in Stow, MA, on Sundays, Tuesdays, Thursdays, and Fridays, and at the Om Namo Center in Belmont, MA, on Wednesdays.
Want Help With Your Knee?
If knee arthritis is keeping you from walking, gardening, keeping up with your grandkids, or just getting through the day without thinking about your knee, I’d be glad to talk with you. We’ll look at the whole picture, including acupuncture, movement, and nutrition, and put together a plan that fits you.
Call or text me at 617-913-5970 or book an appointment to schedule a visit in Stow or Belmont.
Sources:
Hang M, Shao Y, Lu W, et al. Electroacupuncture for knee osteoarthritis: a multicentre randomised controlled trial assessing symptomatic and structural efficacy. eClinicalMedicine. 2026;96:103982. doi.org/10.1016/j.eclinm.2026.103982
Tu JF, Yang JW, Shi GX, et al. Efficacy of intensive acupuncture versus sham acupuncture in knee osteoarthritis: a randomized controlled trial. Arthritis & Rheumatology. 2021;73(3):448–458.
Graca S. EA for Knee OA: Encouraging Research. Acupuncture Today. November 2026. acupuncturetoday.com
This article is for general education and isn’t a substitute for medical advice. Please talk with your healthcare provider about your specific situation.
