Thousands of deaths or serious medical problems occurred in patients with the recalled devices, which included external heart defibrillators, brain shunts and implanted pumps that deliver cancer drugs, the researchers said.
Device makers say the new data are flawed and conflict with previous reports.
For their analysis, the researchers looked at the FDA's list of high-risk devices that were recalled from 2005 through 2009.
Of the 113 highest-risk recalled devices, 71 percent, or 80 devices, had been approved through the less stringent regulation. Only 19 percent, or 21 devices, were approved under a stricter process for brand new products that involves inspections and human testing. Eight were registered with the FDA but exempt from regulation.
"Because so many medical devices are not being held to a higher safety standard, people are dying who wouldn't otherwise die and who don't have to die, and people are being harmed who don't have to be harmed," said study co-author Diana Zuckerman, president of the National Research Center for Women & Families, a Washington-based health advocacy group.
FDA spokeswoman Karen Riley said the study's findings aren't new and noted that the recalls it highlighted represent a small portion of the more than 19,000 devices cleared through the less strict standard during those years.
The report appears in Monday's Archives of Internal Medicine. An editorial says the analysis shows that "millions of Americans may be at risk for device-related injuries and recalls from high-risk devices that were cleared by FDA without any supporting clinical trial data."
"The public deserves better protection," the editorial said.
The researchers didn't include a tally of all deaths and injuries linked with the recalled devices. They noted that in 2006 alone, the FDA received reports of 2,830 potential device-related deaths and more than 100,000 injuries. More than half of the deaths were linked with devices approved under the less rigid process.
AdvaMed, a medical device makers trade group, said the analysis conflicts with other reports -- including one with industry ties -- showing devices cleared through that process have a good safety record.
Under the FDA's 510(k) process, new devices can be approved without human testing if they are similar to devices already on the market. A stricter category, the pre-market approval process, requires scientific evidence, typically extensive testing that is similar to what is required of new medicines.
The 510(k) process is "a short-cut backdoor approach" stemming from a 1976 law, when there weren't many implanted devices meant to sustain life, said study co-author Dr. Steven Nissen, chairman of cardiovascular medicine at the Cleveland Clinic.
He and Zuckerman are among FDA critics who've urged the agency to subject such devices to more rigorous testing, which device makers oppose.
The FDA has said it is awaiting input from the Institute of Medicine, an independent government advisory group, before making a final decision on revisions to the device approval process.