Last reviewed: September 29, 2026

Exercise, weight loss and pain medicines have the best evidence. Most injections and newer procedures have little proof that they work.

According to the American Academy of Orthopaedic Surgeons, exercise is strongly recommended, whether supervised, at home or in water. So are education and self-management programs, anti-inflammatory pills and gels, and acetaminophen. Sustained weight loss for people who are overweight, a cane, a brace, and cortisone shots for short-term relief have moderate support. Acupuncture has limited support. Narcotics, including tramadol, are not recommended, and neither are routine gel (hyaluronic acid) shots or a scope to clean out an arthritic knee.

Weight loss lightens the load on the knee. In a study of 142 overweight older adults with knee arthritis, each pound lost was linked to about 4 pounds less force on the knee with every step when walking. In an 18-month trial of 454 overweight adults, those who lost about 11% of their weight through diet and exercise had less pain and better function than those who only dieted or only exercised. A review of trials found function improved once weight dropped by more than 5%.

Non-surgical treatment helps, though usually less than surgery. In a Danish trial, 100 people who qualified for a knee replacement either had one or received 12 weeks of exercise, education, diet advice, shoe insoles and pain medicine. Both groups improved, the surgery group more. Within a year, 26% of the non-surgery group chose to have the replacement. The surgery group had four times as many serious complications.

Injections are less proven. Platelet-rich plasma (PRP) did better than gel shots in trials comparing the two, but the evidence for it is limited. People given stem cell or bone marrow injections improved, but these did no better than other injections or placebo. According to the FDA, no stem cell product is approved to treat osteoarthritis or knee pain.

Newer procedures have mixed results. Burning the small nerves around the knee with radiofrequency, including the cooled version (Coolief), relieved pain for up to 12 weeks compared with a fake procedure in pooled trials, but not at 6 months or 1 year, and did not improve function. In one trial of 151 people, cooled radiofrequency relieved pain better than a cortisone shot at 6 months. Blocking small blood vessels around the knee (genicular artery embolization) worked no better than a fake procedure in a trial of 58 people.

In general:

  • Ask about a structured exercise program before deciding on surgery.
  • Ask your doctor whether anti-inflammatory pills or gel are safe for you.
  • Before paying for stem cell or other injections, ask what trials show.
  • If non-surgical treatment is no longer enough, a replacement remains an option.

Drafted with AI assistance from published sources, reviewed and edited by Babak Sheikh, MD.

Sources:

This is general information about what published research shows. It is not medical advice and it is not about your situation. Reading it does not create a doctor-patient relationship with me. I am a retired orthopedic surgeon and I do not see patients. Talk with your own doctor before changing anything about your care. If something feels suddenly or seriously wrong, get medical care promptly rather than waiting.

This page is provided as is. I make no guarantee that it is accurate, complete or current, and medical evidence changes. Use of it is at your own risk.

Published by Clinical Evidence Media LLC. Full disclaimer and how these answers are researched: Disclaimer and how these answers are made