The Environmental Protection Agency said yesterday that approximately 35 pesticides approved for use on the basis of allegedly faulty or falsified health studies face suspension from the market within 90 days unless their manufacturers commit themselves to expensive new safety tests.

After a seven-year review, the EPA has invalidated three-fourths of all the chronic health tests conducted on pesticides by Industrial Bio-Test Laboratories of Northbrook, Ill. Officials of Industrial Bio-Test, which is now defunct, are on trial in Chicago. They have denied all charges, saying the government "harassed, abused, misled, bullied, intimidated and coerced" key witnesses to make its case.

IBT's allegedly faulty studies, first discovered in a routine inspection in 1976, were part or all of the basis for "registering"--approving for use--140 pesticides.

Aside from the 35 now facing suspension, an EPA report released yesterday shows that some of the chemicals have been withdrawn by their manufacturers and dozens more are already the subject of additional tests.

Environmentalists criticized the EPA bitterly yesterday for leaving most of the pesticides on the market while the tests are under way, but Edwin L. Johnson, director of the agency's office of pesticide programs, said in a press conference that the EPA had no choice.

"The finding of invalid IBT studies did not create a hazard," he said. "It created a large question . . . . Our statute does not permit us to suspend chemicals simply because of invalid tests."

"If at any time there is valid evidence of a health risk," Johnson said, "the agency will take whatever regulatory action is needed to reduce that risk.

"The IBT situation has not proven to be the hidden public health disaster that some had feared," Johnson said.

Of the 594 allegedly faulty lab tests involving pesticides' implication in cancer, birth defects, genetic mutations or nerve damage, the EPA reported:

* Some 212 tests, or 36 percent, "have been replaced or are in progress."

* Another 38, or 7 percent, "are under discussion for possible replacement."

* And 45, or 7 percent, "are of a type no longer required for registration."

Some 299 tests, the remaining 50 percent, are the subject of the further studies the EPA announced it will now require on roughly 35 pesticides. Any pesticide whose safety rests on one of these tests will face suspension if the manufacturer does not agree within 90 days to retest it.

Yesterday's press conference sought to portray the IBT matter as a problem the EPA would have "wrapped up fairly soon," but Johnson and other officials conceded that they have only asked for, not received, new tests to replace most of the 594 lab studies found defective.

Very few, if any, of the pesticides in question would qualify for registration with the safety data now available, officials said, and Johnson said in an interview that "some of them are going to come back with positive studies that show health effects we didn't know about."

Jay Feldman of the National Coalition Against Misuse of Pesticides said yesterday that the EPA's response to the "tremendous data gaps" in pesticide safety has been too little and too late.

"If the EPA is 'registering' a product," he said, "the public assumes that that product is safe. EPA is doing nothing to tell the public that there is any question about its safety . . . . What the public really needs right now is accurate information so it can make informed choices."

"To a community that is making a decision as to what to spray for gypsy moths, they want to know what the risks are. If a study is in progress, the information is not in, and we advise that if there are alternatives that chemical should not be used," Feldman said.

"At the least, and this is key, the EPA has to go on an aggressive program of informing the public whether the particular product has been registered with faulty data or without data at all."

Jack Early, president of the National Agricultural Chemicals Association, said the manufacturers he represents are "supporting what EPA has been doing."

"If you've got three studies that say a product is all right and one that says there may be some questions, you don't throw out the three that say it's all right," he said. "I'm very comfortable that most of the pesticides will be found to be safe."

Until 1976, IBT was a giant in the independent laboratory business.

Allegedly faulty research discovered that year in a routine inspection by the Food and Drug Administration led to a federal indictment in 1981 of four senior IBT officials.

The officials, who deny any wrongdoing, were charged with concealing test results in four long-term health-effects studies. The U.S. attorney's office in Chicago also accused them of lying about the types of tests conducted, the number of animals used and the length of the studies.

EPA officials said yesterday that the flaws in IBT's research required government scientists--in cooperation with Canada's Health Protection Branch--to "go back to the raw data and observations on each individual mouse and rat in the studies."