NEW YORK (Reuters Health) - Judging by the symptoms of people with knee arthritis, one-third of knee replacement surgeries may be inappropriate, according to a new study.

"We found that some patients undergo total knee replacement when they have very low grade symptoms or minor knee arthritis," lead author Daniel L. Riddle from Virginia Commonwealth University in Richmond told Reuters Health in an email.

The number of total knee replacement surgeries done each year more than doubled between 1991 and 2010, leading some to question whether the procedure is overused. This is hard to prove, though, since symptoms like pain tend to drive the decision to proceed to surgery, and different people perceive pain differently.

To get a better idea of what is going on, Riddle and his team analyzed data from a study of about 4,800 people in the U.S. with knee osteoarthritis or at high risk of the condition. During the study period, 205 of them had total knee replacement surgery.

The researchers used criteria for knee replacements developed by Dr. Antonio Escobar of Hospital de Basurto in Vizcaya, Spain and his colleagues. Surgeries were judged to be appropriate, inappropriate or inconclusive considering factors like a person's range of motion, pain and arthritis severity.

Based on prior studies, the researchers expected to find that about 20 percent of surgeries were inappropriate. What they found, though, was that about 34 percent of patients had total knee replacements that were deemed to be inappropriate.

For the most part, these patients had symptoms that were moderate at worst and joint damage that was not widespread.

Less than half of knee replacements - 44 percent - were classified as appropriate, according to findings published in Arthritis & Rheumatology. That left about 22 percent in the inconclusive category, which included patients with severe symptoms who were either younger than 55 years old or had less joint damage and normal mobility.

Does this mean U.S. surgeons are performing too many knee replacements on patients who don't need them? Not necessarily. For one thing, Riddle said, "the scientific content and the standard at the time (this system) was developed is clearly different from that in the U.S. in 2014."

Doctors and patients need to do a better job deciding when it's the right time for a knee replacement, if ever, he added.

"I would encourage patients to gather and share information with their family physician and surgeon to determine if they are good candidates for the procedure," Riddle said.

"The key issues, in addition to a reasonably healthy medical status, are the extent of pain, extent of compromised function and extent of knee osteoarthritis," Riddle explained. "We have very good prognostic data now and we know that persons with high levels of psychological distress, minor knee osteoarthritis, serious (other health conditions) and multiple joint arthritic disease are at greatest risk of poor outcome."

The authors say that research should now focus on developing a system to separate inappropriate from appropriate knee replacements that is based around U.S. patients.