Two studies call into question whether many people with arthritis are needlessly undergoing one of the most common operations in America: arthroscopic knee surgery.

One finds that surgery is no better than medication and physical therapy for relieving the pain and stiffness of moderate or severe arthritis. The other reveals that tears in knee cartilage — which often prompt such surgeries — are very common without causing symptoms.

Experts said the new studies and other evidence show arthroscopic knee surgery still has a place, such as after a recent injury, but shouldn't be done routinely for osteoarthritis.

"A lot of people would prefer physical therapy and their doctors would as well," said Dr. E. Anthony Rankin, spokesman for the American Academy of Orthopedic Surgeons.

The studies were published in Thursday's New England Journal of Medicine.

Nearly 27 million Americans have osteoarthritis, a top cause of disability. Symptoms usually begin after age 40 and progress as a breakdown of cartilage on the end of bones causes them to rub together at joints, leading to stiffness and pain and limiting movement.

One popular solution is arthroscopic surgery, in which a scope with a miniature camera is inserted through a tiny incision and surgeons operate through other tiny cuts. Surgeons typically smooth damaged cartilage surfaces on the bone's ends and flush out bone chips. This allows quicker healing than traditional surgery, but can still trigger blood clots, infection and nerve or blood vessel damage.

About 1 million arthroscopic knee surgeries are done in this country each year, costing roughly $7,000, depending on the location, when done as an outpatient procedure.

A large study in 2002 found the operation was no better than sham surgery, but the procedure remains popular.

The new study, done at the University of Western Ontario in Canada, included nearly 200 patients with moderate-to-severe osteoarthritis in the knee. Researchers gave half medication, weekly physical therapy for three months and instruction on twice-daily exercises to do at home. The other half got those treatments, plus arthroscopic surgery.

Medications included anti-inflammatory drugs, Tylenol, glucosamine or chondroitin supplements and injections of a knee lubricant. Therapy included exercises to strengthen knee muscles and hot and cold packs.

After two years, both groups of patients reported the same pain levels, physical function and overall quality of life.

Study co-author Dr. Brian Feagan, a professor at the Ontario university, concluded the procedure benefits a minority of patients, those with milder symptoms or large meniscus tears. He predicted the study would change practice.

In the other study, researchers at Boston University School of Medicine and elsewhere examined MRI knee scans from 991 randomly selected people, aged 50 to 90, from Framingham, Mass. About one-third had a tear or other damage to the meniscus — a pair of cartilage pads that act as shock absorbers between the upper and lower leg bones — with prevalence increasing with age. But nearly two-thirds who had these knee cartilage tears had no pain or stiffness in the prior month.

Some experts think the increased use of MRIs to diagnose knee problems could be leading to unnecessary surgeries or at least referrals to orthopedic surgeons.

Surgery for that problem "probably isn't terribly helpful compared to just medication and physical therapy," said Dr. C. David Geier Jr., a spokesman for the American Orthopaedic Society for Sports Medicine.

He said there may be a misperception that doing arthroscopic surgery on arthritis patients could prevent the need for major knee surgery later.

"This might change that opinion a little bit," Geier said.

Rankin said the studies show arthroscopic surgery benefits patients with milder cases, who have had arthritis symptoms for weeks or months, not years, particularly those with a recent knee injury, and those with a major cartilage tear. It isn't for people with advanced arthritis, who may need a partial or total knee replacement, Rankin said, but many family doctors don't realize this.

In an editorial, Dr. Robert G. Marx of the orthopedic surgery department at the Hospital for Special Surgery in New York wrote that treatment decisions must be individualized and that arthroscopic surgery for osteoarthritis has been associated with worse outcomes than therapy. He said it's still appropriate for specific patients where the primary cause of pain isn't osteoarthritis.

"I think everybody should be more cautious about going to surgery," said Dr. Patience White, chief public health officer at the Arthritis Foundation. "We need to understand what is the best approach because the costs are so high."

Many U.S. health insurance plans don't cover arthroscopic knee surgery for people with mild and moderate arthritis, according to Robert Zirkelbach, spokesman for the trade group America's Health Insurance Plans. He said the new findings could lead more health plans to tighten coverage rules.

The Medicare program did so back in 2003, and the number of procedures it paid for plunged 77 percent within three years.