A new diabetes pill that appeared headed for federal approval can double the risk for deaths, heart attacks and strokes, according to an independent analysis rushed online Thursday because of public safety concerns.

The drug muraglitazar (search), developed by Bristol-Myers Squibb and Merck & Co. to be sold under the name Pargluva (search), was endorsed by a Food and Drug Administration panel last month. It is a treatment for Type 2 diabetes, the most common form of the condition that occurs most often in adults who are overweight.

Researchers with the Cleveland Clinic (search) analyzed the data the FDA made public before the panel vote and found that patients taking Pargluva faced double the risk of death, heart attack or stroke, compared with those on dummy pills or a similar drug.

If the analysis is correct, the drug could have meant a "public health catastrophe," said one of the researchers, Dr. Steven Nissen, a prominent heart specialist with Cleveland Clinic.

The analysis was published Thursday morning on the Journal of the American Medical Association's Web site.

The drug's makers said earlier this week that they had received an "approvable" letter from the FDA that also asked for more safety data on the drug's cardiovascular effects.

"It is beyond me why individuals who are supposed to be overseeing the safety of the public would take a chance when it's not necessary," said Dr. Catherine DeAngelis, JAMA's editor in chief. "It's not like there are not other drugs that can be used" for diabetes.

An FDA spokeswoman said agency officials do not comment on pending drug applications. Calls seeking comment from Merck and Bristol-Myers Squibb were not immediately returned.

The non-insulin drug is designed to lower blood sugar levels and increase levels of "good" cholesterol in patients with Type 2 diabetes, the most common form of the disease, which affects about 18 million Americans.

The analyzed data involved 3,725 patients who were given Pargluva, a similar drug called pioglitazone, or dummy pills in a variety of studies lasting from 24 weeks to 104 weeks.

Deaths, heart attacks or strokes occurred in 35 of the 2,374 Pargluva patients versus nine of 1,351 patients in a combined group on the other drug or dummy pills, the analysis found. Slightly higher risks for mini-strokes and heart failure also were found among Pargluva patients.

DeAngelis said the Cleveland Clinic analysis shows much more study is needed, and she likened the situation to what occurred with Merck's Vioxx. Merck removed the painkiller from the market last year because of evidence linking it with cardiac problems. Critics contend the FDA did not adequately address safety concerns raised about Vioxx.

Co-author Nissen also called for more studies.

"If our analysis is correct, then this could be a public health catastrophe of the magnitude of what we saw with Vioxx," Nissen said. "We had an opportunity in this particular case to stop this from happening before it was ever approved."

Nissen has done consulting work for a number of drug companies, including Merck and makers of other diabetes treatments, but said he does not accept fees for that work.

DeAngelis said JAMA fast-tracks research studies for online publication only once or twice yearly.

"We worked like heck to get this one out," she said.

The advocacy group Public Citizen also has voiced concerns about approving Pargluva because of safety concerns.