An international health agency has withdrawn a controversial finding that legally acceptable amounts of benzene may cause cancer. The withdrawal came after a federal official and chemical industry representatives urged the group not to make such assessments.

Richard Adamson, director of the National Cancer Institute's division of cancer prevention, asked the International Agency for Research on Cancer (IARC) to delay any attempt to evaluate the amount of suspected carcinogens that should be considered unsafe. Adamson acted after discussing the issue with officials from Exxon, Shell Oil and the Chemical Manufacturers Association.

The IARC recently published a scientific paper on benzene that deleted a key reference to a government study--approved by its own panel of scientists--in which 17 factory workers died after exposure to tiny amounts of benzene. The findings are important because these workers were said to be exposed to the current legal limit of 10 parts of benzene for every million parts of air.

Benzene, which is known to cause leukemia at high levels, is widely used in gasoline, paints, solvents and thousands of chemical products. An estimated 500,000 workers are exposed to small amounts of benzene, and a warning by the respected IARC could rekindle the debate over whether these levels are safe. A recent attempt to make the benzene limit even stricter was struck down in court in 1980.

Rep. David R. Obey (D-Wis.), who has been investigating the matter, said he believes that NCI officials pressured the IARC into erasing the benzene finding from its report without consulting the scientists who wrote it. A spokesman for NCI, which provides $1.5 million a year to the IARC, strongly denied the allegation.

One of the scientists on the IARC panel, Dr. Phil Landrigan, said he was "surprised and chagrined to see that a critical portion of the benzene risk assessment had been altered."

Adamson told an IARC official in April that the group should not try to determine what levels of carcinogens are unsafe--a task known in scientific jargon as a quantitative risk assessment. Such an effort, Adamson wrote, "is fraught with scientific and societal difficulties" and "involves national policy decisions . . . . I wish to make sure that no discredit comes to the NCI or IARC as a result."

In January, three months after IARC scientists drafted their report on benzene, Adamson met with the chemical industry and oil company officials. NCI acknowledged that they discussed criticisms of the benzene study and the IARC's move toward risk assessment.

An internal memo by the Chemical Manufacturers Association, which sharply criticized the benzene report, said: "Dr. Adamson of NCI has instructed their representatives . . . to recommend that IARC not engage in risk assessments . . . . Dr. Adamson understands the regulatory impact of risk assessment by IARC."

The IARC's draft report noted that 17 tire factory workers had died from leukemia after exposure to what a federal government study estimated was the legal limit of 10 parts per million of benzene. The final report deleted this reference, saying only that up to 170 workers in the study had died after exposure to a much higher benzene level of 100 parts per million.

"We did not try to influence the content of the benzene monograph," said NCI spokesman Paul Van Nevel. "There's nothing in Adamson's letter to IARC that relates to benzene specifically. He's talking about risk assessments in general."

He added that Adamson "disavows" the chemical industry's memo on NCI's involvement. "It's totally and completely inaccurate," Van Nevel said. "He can't figure out how they got those impressions."

Dr. Lorenzo Tomatis, an IARC official, told Obey that NCI had urged him to be cautious in evaluating small amounts of carcinogens, but that "no pressure was ever exerted on IARC by NCI's staff" to change the benzene study.

Obey, however, said: "It is difficult for me to believe that the sudden strong interest of the cancer institute in opposing quantitative risk assessment was not directly related to the preparation of the benzene document."

A Shell official who works with the chemical industry said he was pleased with the IARC's action because the study of factory workers actually reflected exposure to much higher levels of benzene.