A federal agency will ban production of polyester children's sleepwear treated with a flame-retardant chemical called Tris because of scientists' fear that it may cause thousands of cases of cancer each year, government sources said yesterday.

At least three of the five members of the Consumer Product Safety Commission will vote for the production ban at an open meeting Thursday, the sources said.

But additional actions are possible, the sources said. One is to prohibit sale of the estimated 18 million to 22 million Tris-treated garments now in supply chain. These garments are valued at between $72 million and $88 million.

A prohibition on sales could force some sleepwear manufacturers out of business, industry spokesmen told the commission last Thursday. Since 1972, the manufacturers have relied mainly on Tris to comply with the Flammable Standards Act.

Commissioner Lawrence M. Kushner said last month that information on Tris animal studies done by the National Cancer Institute and from agency tests "has convinced me that, if no strongly contradictory evidence is forthcoming, a ban on Tris-treated garments is justified."

"It may well be that a majority, or perhaps all, of the commissioners have reached the same conclusion," he said in a March 23 letter to Robert H. Harris and Robert J. Raunch of the non-profit Environmental Defense Fund (EDF).

The agency's general counsel is to discuss legal aspects of possible courses of action at a closed meeting with the commissioners this afternoon.

A sales ban was urged on March 21 by Prof. Bruce N. Ames, a biochemist at the University of California at Berkeley, who devised a famed preliminary screening test to determine whether chemicals are carcinogens, or cancer-causing agents. He found that about 9 times out of 10, a carcinogen is also mutagen, meaning that is has genetic effects.

Tris is a mutagen "likely to cause genetic birth defects in the offspring of the children exposed to it" as well as "a strong carcinogen" in male and female rats and moce, Ames and research associate N. Kim Hooper said in a letter to commission Chairman S. John Byington. "It is therefore likely to cause cancer in the children exposed to . . . even small doses," they said.

The EDF has asked the U.S. District Court to compel the commission to take strong regulatory action. The request is in a suit filed on March 23, a year after the EDF petitioned the agency to act on Tris in "a reasonable time."

Commissioner R. David Pittle, in a phone interview, said that the agency has given high priority to efforts to assemble evidence on which informed regulatory actions could be taken. He now is convinced that Tris is a carcinogen in animals and "gets into the human system," he said.

EDF official Harris attributed the agency's delay not to any bad faith, but to uncertainty about what action to take.

In children's sleepwear, Tris (2,3-dibromoproply phosphate) is in heavy concentrations, usually accounting for between 5 per cent and 10 per cent of a garment's weight. The possible risk varies widely, depending upon such factors as whether underclothes curb absorption through the skin and whether an infant sucks on the garment.

The EDF has urged the commission to require Tris-treated apparel to carry a label warning of its presence and saying, "Should be washed at least three times prior to wearing." Sleep garments made of cotton and nylon are not treated with Tris.

Commission scientists, in a report on March 17, estimated that an "average" exposure to Tris sleepwear could cause between 1,848, and 3,300 new cases of cancer in the approximately 3 million children born each year. In animals treated with Tris, cancer occurs in the kidneys.

Ames and Hooper had a much higher estimate. "We calculate that one year of exposure (from age 0-1) may cause cancer in about 1 per cent of the children," they said.