They might as well call it magic. One clinic’s Web site dubs platelet-rich plasma (PRP) therapy “revolutionary.” Other purveyors call it a “breakthrough” and “a paradigm shift.”

The treatment involves injecting concentrated platelets, taken from the patient’s own blood, into the site of injury to speed recovery, and stories of PRP’s use by high-profile athletes have turned it into a lucrative business.

Tennis star Maria Sharapova turned to it for a shoulder injury and Hines Ward used it to overcome a knee injury in time to help the Pittsburgh Steelers win the 2009 Super Bowl. The PRP market was valued at $45 million in 2009, and it is expected to grow to $126 million by 2016.

Yet the question remains: Does it work? The evidence remains questionable. In April, the Cochrane Collaborative, a group of independent medical experts, examined 19 trials of PRP for eight conditions and concluded that the evidence is insufficient to recommend PRP for musculoskeletal soft-tissue injuries.

A PRP treatment begins with a blood draw from the patient. Next, the blood is spun to separate platelets from the other blood cells. The process creates a solution of concentrated platelets and growth factors that is injected into the injured area to promote healing.

Utah Jazz's Enes Kanter, of Turkey, shoots as Los Angeles Clippers' Blake Griffin defends during an NBA basketball game in Salt Lake City. Kanter underwent platelet-rich plasma treatment to help with tendinitis in his right knee. The injection is designed to stimulate recovery in the 22-year-old's joints. (Rick Bowmer/AP)

“We don’t know if it helps people,” says Frederick Azar, president of the American Academy of Orthopaedic Surgeons. “The clinical evidence is controversial.” Some studies show a benefit, some don’t, he says.

A randomized, double-blind study published last year found that PRP outperformed a placebo for osteoarthritis of the knee. A review published last year noted increasing numbers of clinical studies of PRP, turning up both positive and negative evidence for its effectiveness, but concluded that there is still little evidence that PRP works better than a placebo in most cases. Similarly, a review published last month concluded that there is “strong evidence” that PRP is not effective for treating tennis elbow. A randomized, double-blind trial published as a letter in the New England Journal of Medicine last month concluded that PRP was no better than a placebo for hamstring strain.

The Cochrane review found that most studies are small and not definitive. The wide variety of methods used to deliver PRP and the absence of a standard methodology may explain why studies have reached such different results, Azar says: They’re comparing different ways of delivering the treatment.

Despite the lack of standardization and definitive evidence, PRP is used to treat a wide range of maladies, including chronic tendon injuries, acute muscle and ligament injuries, osteoarthritis of the knee, and to speed recovery from surgery and fractures.

If PRP hasn’t been validated in studies, why is it in such widespread use? Because even without conclusive evidence, those who provide PRP believe in it. “We don’t have enough information to show that it works, but we believe it has great potential to work,” Azar says. “There’s little downside or risk that would prohibit us from offering this.”

Some patients might point to the cost — which can range from $500 to $1,200 for a single treatment — as a downside. Most insurance companies won’t pay, since the treatment is experimental.

But that doesn’t stop doctors and clinics from marketing PRP to patients, some of whom are desperate to return to their previous activities. “This is one of those things that falls between the cracks,” says Karen Maschke, a bioethics scholar at the Hastings Center in Garrison, N.Y., because PRP doesn’t fall under the Food and Drug Administration’s jurisdiction. As a result, patients have to rely on their physician to “do no harm,” she says.

“Orthopedic medicine has a long history of pushing procedures and devices that don’t fall under FDA regulation,” Maschke says. “They’re engaging in entrepreneurial medicine.” Procedures such as PRP can generate a lot of income for doctors and clinics, and if the practitioners providing the treatment really believe in it, they may think they have a right to offer it, even if there’s not strong evidence to support it, Maschke says.

When asked whether it’s ethical to market an unvalidated treatment, Azar replied, “I’m comfortable with it. My son has had a couple of these injections, and I’m comfortable because the downside from a patient-safety standpoint is minimal.” He doesn’t see the cost as a problem. “If the patient is fully informed and they want to do it and they want to pay for it, I think it is ethical,” Azar says.

“Usually, I won’t offer it upfront,” says Bryan Murtaugh, a sports medicine physician at MedStar National Rehabilitation Network in Washington. He calls PRP “part of the spectrum of treatment options” and says he normally recommends other approaches first, such as physical therapy, medication, activity modification, bracing or even steroid injections.

If those don’t work, “I always tell patients that it’s a treatment option,” Murtaugh says. “There is a part of it that’s still experimental: We don’t know what the best preparation of the platelets are — the volume, the frequency, how many treatments each patients should get.”

Patients may not get this cautious message by perusing MedStar’s Web site, which describes PRP as a “paradigm shift” and refers to “convincing evidence of its efficacy” for certain problems.

Murtaugh readily acknowledges that “studies have shown mixed results.” He can’t say for sure how effective PRP is. “It’s really hard to quantify anecdotally,” he says, and since he doesn’t systematically track its outcomes in his patients, he can make only ballpark estimates of what proportion of his patients benefit from the procedure. “I’d say at least two-thirds to 75 percent might see a partial response. Some people get a complete response, and a percentage don’t get any improvement,” he says. “I wouldn’t say it’s perfect science.”