Ralph Nader's Health Research Group said yesterday that a drug called phenformin may have dilled at least 17 persons in the six months since a panel of experts urged the Food and Drug Administration to take it off the market.

The group urged Health, Education and Welfare Secretary Joseph A. Califano Jr. to declare the prescription drug and "imminent hazard to public health," a step that would halt its sale immediately.

Phenformin is widely prescribed for mild diabetes.Ciba-Geigy Corp. sells it as DBI and USV Pharmaceutical Corp. as Meltrol. Annual estimated retail sales are $25 million.

Physicians prescribe phenformin to hold down blood sugar in about 250,000 persons with so-called adult onset diabetes. One million persons with mild diabetes try to control blood sugar with pills in a different chemical family.

A spokesman for Califano said he had no immediate comment. Under the drug laws, he alone has the power to decalre a drug an imminent hazard.

No HEW secretary ever has used the power to end sales of a medicine.

In a seven-sentence statement, Dr. J. Richard Crout, director of the FDA's Bureau of Drugs, said the agency is "considering" removal of phenformin from the market.

The FDA could try to end sales by suspending the permission it gave for the sale of phenformin. Under the law, however, sales could continued unimpeded while the companies contest a suspension, first in the agency, and then in the courts.

Crout told a Senate hearing in July, 1975, that in his "personal opinion as a medical scientist," phenformin "has no role" in the treatment of diabetes and should not be sold. In yesterday's statement, he said, "Patients on phenformin should not abruptly discontinue therapy on the basis of unwarranted publicity about this drug."

The death estimate was made in hand-delivered letter to Califano from Dr. Sidney M. Wolfe, director of the HRG, which is financed by Nader's Public Citizen, Inc.

Wolfe wrote that as of last Oct. 21, when the FDA Endocrinology and Metabolism Drug Advisory COmmittee urged removal of phenfirmin from the market, FDA files contained reports of 93 fatal and 97 nonfatal cases of an adverse reaction to phenformin known as lactic acidosis. THe reports date from 1956, when the drug first went on sale.

But Wolfe said there is general agreement that the number of adverse drug reactions reported to the FDA is "is gross underestimate" of reality. He cited a report in a diabetes journal that just one New York CIty hospital, Bronx Municipal, recorded nine lactic acidosis deaths among phenformin users in two years.

In the six months sincce the October report, 45 new cases have been reporting to FDA, Wolfe said. The number of fatalities among them is unknown.

But lactic acidosis kills about half of those it afflicts, so Wolfe estimated the number of deaths has now risen at least to 110, which is 17 over last October's total.

"Thus, the real number of U. S. Deaths due to phenformin lactic acidosis is probably several hundred or more, only 93 of which are are in FDA files," Wolfe said.

"Each further day of delay in baning phenformin may result in the preventable death of another diabetic," Wolfe continued. "Thus, its continued marketing clearly constitutes an imminent hazard to the public health."

The advisory committee found "a clear and unequivocal association between the use of phenformin and lactic acidosis." Because "not all patients at risk may be identifiable," the panel said, "no modification" of the official labeling "or other restrictive caveats would increase the safe use of the drug.

"Crout said the death rate "is fortunately low and has been long known to FDA and to the medical profession."

He pointed that the FDA strengthened the label warning about lactic acidosis in 1973 and further tightened it two months ago to emphasize the risk and to narrow the number of patients exposed to phenformin.

In his July, 1975, testimony, Crout pledged to the Senate Small Business Monopoly Subcommittee that he would refer the phenformin issue to the advisory panel. A year later, the FDA confirmed to The Washington Post that he had not yet done so.

In 1970, researchers said a government-financed study showed that all of the pills - known as oral hypoglycemics - lost their effectiveness in lowering blood sugar in as little as two years of regular use. And, they said, long-term users suffered fatal cardiovascular complications at a rate of 21/2 times as high as persons who relied on diet alone to control onset diabetes.

In 1975, physicians wrote about 3.8 million prescriptions for phenformin and about 13 million for the chemiscally different pills, such as Orinase.