BREATHING UNEASY

Fiberglass dust, the government says, is `reasonably anticipated' to cause cancer. But the debate over what that label means suggests that modern science can detect environmental risks much better than it can assess them

By Peter Perl and Li Fellers
from The Washington Post Sunday Magazine
Sunday, June 22, 1997

• Everything is a poison
• Why file this?
• An equal and opposite Ph.D.
• The hamsters of Switzerland
• Like hanging the accused
• "We never talk about it"

The guys in the machine shop at the Owens-Corning Fiberglas factory had a little game they would play from time to time.

Whenever one of their colleagues celebrated his retirement from the nation's largest producer of fiberglass insulation, usually after a career of 20, 30, even 40 years, they would take bets on how soon he would die.

It just seemed to Danny Humphrey and his buddies at the Newark, Ohio, plant that the retirees would live maybe a year or two, three tops, except for a few old buzzards who lasted forever. Often, they'd base their bets on how badly the retiring guy was breathing.

"We had guys who couldn't walk to the front door without oxygen," says Humphrey, whose gray hair, lined face and slumped posture make him look old for his age, which is 53. Sometimes, those guys would sneak a hit of pure oxygen from the welding equipment.

Nobody was sure why some men seemed to die quickly. The only things for certain were that many of them had breathing problems and almost everybody smoked cigarettes. The word around the shop was usually that some big guy like Jimmy or Frank had fallen away to nothing and died of cancer or something. But the newspaper in Newark, a town of 47,000 just east of Columbus, didn't print causes of death in its obituaries.

Occasionally, the Owens-Corning workers would speculate about the product they made. The glass wool fibers made you itch terribly when you first started work, but then you got used to it. The endless rolls of fibrous wool blanket generated a fine dust during processing. Sometimes you might cough up some fibers. Some guys got nosebleeds, but maybe it was allergies or something. Nothing you'd call in sick over. Once in a while, you'd hear that somebody was coughing blood, like Humphrey himself, a pack-a-day Pall Mall man whose X-rays detected a spot on his lung.

A few of the guys started talking about "pinklung," likening it to coal miners' black lung or cotton-mill workers' brown lung. But some of the other guys said that was a lot of crap, that maybe the dust was bad for you but nobody knew for sure. The company had been providing face masks and respirators for years, though most people didn't use them. Especially the ones who already had trouble breathing.

Besides, there were really very few fibers in the air inside the factory these days. In recent years, the sprawling Newark plant was the cleanest it had ever been. Most of the old-timers, though, like Humphrey, who'd spent half his life at Owens, could remember when the air would get so thick with glass dust that the sunlight streaming through the windows lit up the place like a crystalline snowstorm.

Everything Is a Poison

On a December day in 1993, more than 25 of the federal government's top public-health experts huddled in a windowless conference room at the Department of Health and Human Services. The key item on their agenda was deciding whether the government should declare that microscopic shards of glass wool fiber could cause cancer when inhaled.

They had been considering the same question through nine such meetings over four years. Twice -- first during the Bush Administration, and again under Clinton -- the companies that make fiberglass insulation had persuaded HHS to delay any decision. By now, many of these scientists were tired, frustrated, even privately angry at having to discuss the question yet again.

Fiberglass dust had caused cancer in laboratory rats and hamsters, and there was some evidence that it might be responsible for higher cancer deaths among workers in factories like the Owens-Corning plant in Newark. The World Health Organization had declared fiberglass to be "possibly" carcinogenic in 1988. Based on that finding, American insulation manufacturers were already labeling their product as a "possible" carcinogen. So to some of the scientists, the decision seemed annoyingly obvious.

This time, there was intense discussion, but no more delay. When the question was called to a vote, the six voting members of the panel made it unanimous: Yes, fiberglass was "reasonably anticipated " to cause cancer, and the United States government should say so.

In theory, the vote should have clarified yet another hazard to the public health. In practice, though, it underscored how hazardous such judgments have become.

Reasonably anticipated . . . The label itself is not a model of precision. What is "reasonable" when it comes to cancer? Is that based on a scientific calculation, or cautionary fear? And just how would you "anticipate" the disease? Who would be susceptible? Under what circumstances? By how long would you have to anticipate cancer to escape it? As a warning, the phrase is not terribly helpful.

Yet "reasonably anticipated" is one of the terms of art in this kind of risk analysis. It was the best the government scientists could do.

They represented 10 agencies, but gathered as part of an obscure but significant process called the National Toxicology Program, the nation's official early warning system on substances that could cause cancer. The NTP's primary mission is to compile the Annual Report on Carcinogens -- the government's definitive list. And the ARC admits only two types of toxins: those "known" to cause cancer and those "reasonably anticipated" to do so.

Since it was introduced, in 1978, the ARC has listed 23 known carcinogens -- asbestos, for example, and arsenic and mustard gas. In each case, the toxin in question has been proved in human studies to cause cancer. On the other hand, the ARC also lists about 150 "reasonably anticipated" carcinogens -- not only fiberglass, but also saccharin, DDT and formaldehyde. These substances have shown "limited evidence" of carcinogenicity in humans, or "sufficient evidence" in multiple cases of cancer in animal experiments.

These listings carry no regulatory weight of their own, but they have powerful implications for businesses, state and federal government agencies, and the public health. For known carcinogens, the response is obvious: They are restricted in their production and use; they are withdrawn from the market; they are removed from homes and offices; they are buried deep in the earth. For the others, however, the response is complicated by the question of risk, and the question of risk is complicated by the limits of science. Over the last few decades, the science has become both more exact and less conclusive. Society's capacity to detect environmental threats has outstripped its ability to define them.

"One of the big problems we face today is that we learned to measure in parts per billion, parts per trillion," says D.A. Henderson, former dean of the Johns Hopkins University School of Public Health and a highly respected former White House science adviser. As deputy assistant secretary of HHS, he sat in on the meeting at which the NTP's executive committee voted to list fiberglass in the ARC, and he still has reservations about whether the vote was right. "The question is, how do you determine -- if you give a dose of X to an animal and it gets cancer, but a human is going to be exposed to only one one-thousandth of that -- is there a cancer likely? How about one ten-thousandth? . . . Is it reasonable to err towards the side of public health? Well, yes, but I think we can make huge mistakes."

In the case of fiberglass, the government's declaration has settled very little: The stuff is made and sold exactly as before. No new regulations are in the offing. Glass wool is already one of the most studied substances on Earth, the subject of some 50 major studies and more than 400 scientific papers -- and yet a host of questions remain under investigation. Some of them may be answered soon, but others circle back to the conundrum that Paracelsus, the Swiss physician widely regarded as the first toxicologist, articulated in the 16th century. "Everything is a poison," he said. "It all depends on the dose."

In the meantime, the government struggles to manage the risks that scientific studies suggest. Some 70,000 substances have been introduced into American workplaces this century. The Occupational Safety and Health Administration has 25 legal standards covering about 400 of them.

It takes about a decade for the government to assess a risk, gather data and get a regulation passed, says Joseph Dear, who resigned in January as the assistant secretary of labor in charge of OSHA."It is frustrating, particularly when you have a four-year political term," he says. "How do you act in the face of uncertainty or doubt when it can expose people to something that can harm them or kill them?"

Science alone, in his view, cannot provide the ultimate answer. "You do the best you can and try to balance it out," he says. ". . . These issues ultimately revolve around values."

Why File This?

Fiberglass is so useful -- lightweight, flexible and terrifically insulating -- that it has become ubiquitous. It lines the attics and walls of the majority of American houses. It forms ceiling tiles in many office buildings. It is an integral part of cars and planes, and dishwashers and stoves, and shingles and water pipes and air filters and heating and cooling systems. No single substance could easily replace it.

Owens-Corning invented glass wool in the 1930s, and in 1938 perfected a method of producing it by pouring streams of molten glass into rotary metal spinners with thousands of tiny holes that spew out hot filaments. These filaments are blasted by suctioned air that stretches them into microscopic strands, and these strands are sprayed with chemicals (and, in Owens's case, a trademarked pink dye) to form glass wool. This wool is then rolled, cut and processed into its many forms.

Today, roughly 30,000 Americans work in fiberglass manufacturing plants; some 200,000 more install fiberglass insulation in houses and offices, and millions more handle varieties of fibrous glass while turning out countless consumer products. The scientific community has achieved consensus on at least one issue -- that the consumers who use those products are under no threat. But the workers who inhale fiberglass dust in the process of making the wool, cutting it, installing or removing it may be at risk.

Exactly how, and to what extent, remains uncertain. In a sea of data, the essential fact -- does fiberglass cause cancer or not? -- remains in dispute. So far, this debate has been haunted by the example of asbestos, which ranks as one of the worst disasters in the annals of American industry.

Since 1940, an estimated 27 million workers have been exposed to asbestos dust in American workplaces. The companies that sold asbestos insulation -- particularly Johns Manville, the industry leader -- had known it was dangerous since the previous decade, but they concealed their knowledge, suppressed scientific studies, censored trade journals and kept on selling their product. It was not until the 1970s -- when those companies and their insurers began to lose hundreds of millions of dollars in damages in a wave of lawsuits -- that the government acted. After years of paralysis, it effectively banned most uses of asbestos.

In all, health experts calculate, more than 400,000 Americans will die of diseases related to asbestos fibers they inhaled. The body count is still rising, because the fibers often take 30 years or more to cause cancer, but it is projected to equal or surpass all U.S. military deaths during World War II. Owens-Corning projects that it will pay $4.8 billion in asbestos-related claims; Johns Manville's liability of more than $1 billion bankrupted the company, which was resuscitated as Schuller International and then, only last month (May, 1997), renamed Johns Manville.

A debacle of this magnitude casts a long shadow -- some scientists repeatedly invoked the "lessons of asbestos" during the NTP's deliberations over fiberglass. But there are subtle yet important ways in which the lessons of asbestos may not apply.

Like asbestos, fibrous glass is potentially lethal only when its most microscopic fibers become airborne and are inhaled. Scientists believe that only the tiniest of them pose any danger -- fibers less than about 3 microns in diameter but more than about 5 microns in length (the diameter of a human hair is about 70 microns). Only those fibers -- considerably less than 5 percent of the total -- are sufficiently aerodynamic to be inhaled deeply enough into the lungs to cause cancer.

Unlike asbestos, however, fiberglass may not remain in the body. Asbestos is a mineral whose fireproof nature inspired ancient Greeks to give it a name that means "inextinguishable" once inhaled, it can remain permanently buried in lung tissue. One key point that the fiberglass industry is seeking to prove is that because glass fibers dissolve relatively quickly, they are harmless.

Industry documents show that concern about fiberglass-related illnesses and liability goes back decades. In a memorandum from December1955, an adviser to Owens president Harold Boeschenstein began: "For the first time, medical literature now carries two articles which indicate that glass fibers can be harmful to human lungs. These articles can be cited effectively in damage suits against us." Three experts consulted by the company, the adviser said, concluded that the articles were "loaded with statements that could be damaging in a suit."

The memo, which recommended further research, was stamped, "WHY FILE THIS?"

But the memo was filed (and unearthed during asbestos litigation two decades later), and the industry did pursue further research. Companies stepped up their studies of dust levels in the workplace and the effects of the fibers on animals. After several studies produced alarming results, manufacturers redoubled their research efforts and launched educational and public relations campaigns to answer workers' and consumers' questions. In the last decade alone, the industry has spent more than $20 million on research.

That sponsorship is why fiberglass is one of the most studied substances on Earth. It is also one reason why the debate over the studies has become so politicized.

An Equal and Opposite Ph.D.

The body of research grew larger, but not much clearer.

Laboratory scientists used a variety of methods to experiment on mice, rats and hamsters. In 11 separate studies, animals that inhaled fibers did not develop cancer at a higher rate than normal. But in 13 other experiments -- in which fibers were injected or implanted around the lung -- the test animals did contract excess cancers.

These studies, coming just as the waves of asbestos litigation were beginning to rock that industry, lent considerable urgency to the question of what, if anything, fiberglass was doing to humans.

To find out, a biostatistician in the late '70s began one of the largest occupational mortality studies ever undertaken. Philip Enterline of the University of Pittsburgh's School of Public Health, on contract with Owens-Corning, began gathering data on workers at 11 fiberglass plants nationwide, including the Owens-Corning plant in Newark, Ohio. Their death certificates and employment records were microfilmed and fed into a computerized database. Their survivors were interviewed in an attempt to document the workers' overall health and smoking histories. Their work environments were reconstructed, department by department dating back to the 1940s, using old records and experiments to simulate how many fibers they might have breathed in years past.

In all, Enterline examined data on 11,380 workers who died from 1946 to 1982. And in 1987, he reported that these workers had "statistically significant" excess death rates for lung cancer and for nonmalignant respiratory diseases. In other words, Danny Humphrey's retired colleagues in Newark and elsewhere had died of breathing maladies more often than they should have.

The question then became: What, exactly, was killing them?

Enterline said he and other researchers identified various "confounding" factors -- factors other than glass wool -- that could account for the higher death rate. In addition to smoking cigarettes, the workers had been exposed to numerous chemicals used in production, and some older workers also had handled asbestos. An unknown factor, Enterline noted, was whether the workers who had the poorest general health to begin with tended to get the dirtiest and most dangerous jobs.

Overall, Enterline found a 16 percent excess in lung cancer deaths, compared with the rate expected nationally. Comparing his data with those for the industrialized areas around each factory, he still found an excess of 9 percent.

Enterline, who is now retired, said he initially considered the mortality "very high," particularly after he learned that the fiber counts in the air at the glass wool plants appeared so low that few excess cancers would be expected. The high mortality was further troubling, he said, because working populations, such as his cohort, usually are healthier than those without jobs.

Enterline says he and his colleagues realized that they would need to lengthen and deepen the study to determine whether there was a genuine "dose-response" relationship between fibrous glass and lung cancer. He says his initial study was well-intentioned but, in hindsight, not necessarily accurate in assessing risk because the average fiber counts were too low in places like Newark.

"The problem is that exposures are not high enough -- even if it was asbestos -- to see something," he says. "If we had seen how low the fiber levels were, I would probably not study it."

And now, after launching one of the largest occupational mortality studies ever undertaken, Enterline defines the risk from fiberglass as this: "not insignificant, but I feel it is nothing like the asbestos problem."

A more telling study, he says, would be of construction workers who handle batts of fiberglass or installers who use air jets to blow loose fiber insulation into walls and attics. Their exposure levels, according to various studies, are hundreds of times higher than those for the plant workers. So far, however, no one has tried any such study. Nor, in the opinion of various scientists, is anyone likely to: The turnover among construction workers and installers is so high that it would be impossible to track the mortality history of a large cohort, as Enterline did.

Still, Enterline's study did not end with his retirement. The North American Insulation Manufacturers Association (NAIMA) has spent half a million dollars a year to continue it. Gary Marsh, a student of Enterline's who took over the study, says the latest data show a 12 percent overall elevation in lung cancer deaths compared with local populations. Marsh characterizes this excess as "very small," because it could easily be caused by confounding factors.

"I am not aware of any practicing epidemiologist who would place much significance on a 12 percent elevation," Marsh says. Fiberglass, in his view, is not the culprit because the deaths do not clearly correlate to the length of employment or intensity of exposure.

Some practicing epidemiologists disagree with Marsh, and strongly. But the mortality data will not yet prove either proposition. Richard Leman, a retired epidemiologist for the National Institute for Occupational Safety and Health -- and the scientist who nominated fiberglass for listing in the Annual Report on Carcinogens -- says this part of the debate is hauntingly similar to the early stages of the debate about asbestos: "For every PhD," he says, "there is an equal and opposite PhD."

Meanwhile, the dueling PhDs aren't likely to get much help from the data coming out of the laboratory.

The Hamsters of Switzerland

Every weekday morning at 8, teams of technicians report for duty at a laboratory run by the Research & Consulting Co. in Basel, Switzerland. They take showers, pull on sterile surgical gowns, and then, amid strains of classical music, start plugging hamsters into plastic nose cones.

More than 700 of the rodents -- Syrian golden hamsters, specially bred in a Canadian laboratory -- spend the next six hours confined in individual tubes on floor-to-ceiling carousels, breathing one of six types of forced air. One group breathes clean filtered air. A second group breathes an aerosol laden with fiberglass insulation fibers. A third breathes a finer, but more durable, specialty glass fiber. Three more groups breathe various levels of asbestos.

The hamsters, which normally live about two years, die intermittently or are "sacked" by lethal injection at intervals. Then their right lungs are cut up and microscopically examined for disease and fiber residue. Their left lungs are shipped to a Manville research facility in Colorado, where they are burned at low temperatures to dissolve the tissue so scientists can measure how much fiber remains trapped within.

This $6 million experiment, run by Manville and NAIMA, is the largest fiber-inhalation study ever conducted. It represents the industry's latest, best effort to demonstrate that inhaling glass fibers does not make people sick, on the theory that the body's mucus and cilia intercept and expel most inhaled fibers and the relatively few that lodge deep in the lungs dissolve harmlessly within weeks.

This is no simple proposition, however. Critics argue that the experiment design is flawed because rats and hamsters are "nose-only" breathers whose narrow respiratory passages don't take in fibers as easily as those of humans. They also say that the aerosol mixtures are faulty because static electricity makes airborne fibers cling together, and these clumps are too big and too heavy to be breathed deeply enough into the lung to where most cancers would arise.

The majority of World Health Organization and U.S. government scientists who voted to list fibrous glass believed that injecting or implanting fibers in or near the lung gives a more accurate picture of the likelihood of human cancer.

This contention is strongly disputed by Eugene McConnell, who co-chaired the WHO carcinogens committee in 1988 but has since become one of the foremost critics of the implantation studies. He says implantation delivers unrealistically high doses of fiber to the test animals. When a fibrous mass is implanted, he says, "the minute it gets there, the cells are attracted to it. They wrap around a foreign body, and now it can't dissolve. So the [cancerous] lesion you produce in the peritoneal cavity has no rhyme or reason in the body. It could never be produced" except by injection. In real-life settings, he says, "when you are exposed to fibers, it is a given level over a lifetime. You don't get one big blast."

His word on this subject is authoritative, but not universally accepted. McConnell was the research and testing director of the National Toxicology Program from 1978 to 1988, in which capacity, he says, he supervised more animal cancer studies than anyone in the world. He then left the government to work as a consultant to the industry. Since then he has come under criticism from his former colleagues, but he dismisses the suggestion that he has changed his views to suit his new employer as "childish" and "cheap shots."

McConnell says he had unresolved doubts about the WHO vote in 1988 and has come to believe that it was wrong. He acknowledges that some man-made fibers, such as durable ceramics used in extreme high-temperature insulation for aerospace, have been shown to cause excess cancers when inhaled by animals. But glass insulation fibers, he contends, have not.

"Some of these critics do not understand," he says. "A fiber is not a fiber is not a fiber."

Thomas Hesterberg, a Manville scientist who supervised earlier inhalation studies, is also overseeing the experiment in Basel. He says the industry has spent millions at the best lab available to answer criticism of the earlier inhalation studies. One change was to switch from rats to hamsters, which are considered more sensitive to the fibers. Another was to refine the "aerosolizing" technique to deliver a more controlled dosage. This latest study, he says, will give the truest reading yet. The results are expected late this year.

But even then, the issue isn't likely to be settled.

On this point, Peter Infante, director of the office of standards review at OSHA, is adamant: Regardless of the outcome in Basel, the current evidence that fiberglass has caused cancer in animals is "overwhelming," primarily because the implantation studies consistently showed high cancer rates at multiple sites in and around the lung.

In 1994, Infante and three other government scientists published an article in the American Journal of Industrial Medicine. They offered not only a defense of the listing of fiberglass in the ARC -- on which three of the authors had worked -- but also a critique of the industry's inhalation studies. They contended that no industry-sponsored study has demonstrated that it has delivered sufficient fiber to the target area of the lung to yield meaningful results.

In their article, Infante and his coauthors, James Huff, Loretta Schuman and John Dement, also took on Gary Marsh's analysis of the human mortality data. They said a 12 percent increase in lung cancer, the most common form of the disease, was cause for serious concern and could not be explained by confounding factors such as smoking. They said fiberglass workers studied by Enterline and Marsh were exposed to such low fiber counts that asbestos workers receiving a similar dose would not develop cancers. The fact that even low doses of fiberglass appeared to elevate the cancer rate, they said, suggested that "on a fiber-per-fiber basis, glass fibers may be as potent [as] or even more potent than asbestos."

McConnell and Hesterberg replied with strong rebuttals. They said Infante and his coauthors were very selective in their use of data, gave little weight to confounding factors and ignored the findings of several scientific bodies that said implantation was not a realistic way to test fibers.

Another challenge to Infante's conclusions came from Leonard Chiazze, a professor of occupational health studies at Georgetown University Medical Center. Chiazze led an industry-funded team that reconstructed estimated levels of smoking at the Newark plant, the surrounding city and the nation during the decades of the Enterline-Marsh study. Factoring in the smoking habits of 751 Newark workers, the Chiazze study concluded that the excess lung cancer rate at the plant was only 5 to 10 percent over the national figure -- and attributable not to fiber exposure, but to "some unknown set of social, demographic or chance factors."

Lacking any clear scientific evidence, the people who represent workers' interests are drawing their own conclusions.

"We have not notified our members about dangers, because we don't see people getting sick or dying from this," said Joseph Galvin Jr., research director for the Glass, Molders, Pottery, Plastics & Allied Workers, a union representing about 8,000 fiberglass workers. "We have cancers, and we look closer and they are usually smokers." After attending more than 30 government and industry meetings on the issue, he says, the strongest cancer evidence -- the animal implantation studies -- do not impress him: "None of our workers are going to have it put into his stomach like that."

Other unions, such as the Sheet Metal Workers and the Laborers International, are not so sure. For one thing, their members, who work with fiberglass in thousands of factories and construction sites, are exposed to higher fiber counts than workers in glass wool fabrication plants like Newark's. For another, the estimated latency period for fiber-related lung diseases is 20 to 30 years.

"The jury is out," said Gary Batykefer of the Sheet Metal Occupational Health Institute, which has commissioned a medical study of union members. "This is in its infancy."

Like Hanging the Accused

In the meantime, houses and offices are being built, appliances are being manufactured, airplanes are rolling off the assembly line. Amid the uncertainty, people still need some rational basis for responding to the available evidence. The government has tried to provide this basis through both the ARC listings and various regulations, but neither approach has been totally successful.

Even as the scientists from the National Toxicology Program were trying to decide whether to list fiberglass in the ARC, they knew they were laboring under outmoded criteria dating from the program's origin in 1978. The ARC criteria called for a substance to be listed as a "reasonably anticipated" carcinogen based on even a single "positive" finding of cancer, regardless of the method of exposure, such as inhalation versus injection.

Kenneth Olden, the head of the NTP, made it a priority to change these criteria when he took over in 1991. Olden and others believed the rules hamstrung scientists because, among other things, a few cancer findings could trigger a listing even if a larger emerging body of evidence showed no danger.

"I was changing the criteria because of the evolution of science" says Olden, the former chief of the Howard University Hospital cancer center.

D.A. Henderson, the ranking HHS official directly involved, believed parts of the ARC process were an "idiotic" way to conduct science because the criteria were too rigid. But changing the criteria would take considerable time, and Henderson and Olden both believed that under the existing law and regulations, the case of fiberglass was handled properly.

Executives in the industry, however, were livid at the prospect that fiberglass would be labeled a possible carcinogen while the government was reviewing how to make that judgment. "In our view," wrote Kenneth Mentzer, the executive director of NAIMA, in a letter to HHS, "listing [glass wool] first and reviewing the decision criteria second is a bit like hanging the accused first, then having a trial."

In February 1993, industry lawyer Harrison Wellford wrote HHS Secretary Donna Shalala, his former Carter administration colleague, two letters on the same day. "Dear Donna," one began. "Again, congratulations! I remember when we speculated about roles for you in the second Carter term. I never dreamed we would have to wait 13 years!" The letter discussed the "obscure" ARC listing process and, along with the more formal second letter, informed her that the industry might sue HHS to stop the listing of fiberglass.

The industry intensified its lobbying against the ARC listing -- at one point mobilizing 19 senators and 28 members of the House to write to Shalala. In the end, though, these and many, many other letters did not change Shalala's mind. She approved the listing in June 1994.

Olden, in retrospect, is annoyed at the level of pressure exerted. "They knew I was moving as quickly as I could. I got tremendous pressure . . . We had bent over backwards" to listen to the industry's viewpoint beforehand, he said. "I feel good about it. I can sleep at night."

New criteria were adopted last year, and Olden believes they will make the process more scientifically rigorous. The new process includes a board of counselors from government, industry and academe to provide an extra review of scientific findings. The criteria also will allow greater flexibility in weighing evidence, because a listing will not necessarily be triggered by only one or two positive studies. Instead, the agency will be able to better evaluate the total weight of all evidence and the specific "mechanism" by which cancer is caused in animal studies or other new forms of testing. It will also include a way to "de-list" a substance if new data become available.

Ultimately, there will no longer be ARCs. The process was so cumbersome that HHS had issued only seven annual reports in 18 years. Acknowledging how time-consuming the process is, HHS has approved a switch to Biennial Reports on Carcinogens, so they'll be BRCs.

Even so, the changes may not do much to resolve the regulatory uncertainty over fiberglass. Although fiberglass remains in the ARC, no new federal regulations -- if, indeed, someone decides new federal regulations are necessary -- are likely to appear soon.

Since the listing, four federal agencies have begun to consider new standards concerning glass wool: the Occupational Safety and Health Administration, the Environmental Protection Agency, the Consumer Product Safety Commission, and the National Institute of Occupational Safety and Health. So have about 10 state governments, some foreign governments and the European Union, according to the industry, which points out that Canada has reviewed similar data but decided not to list glass wool as possibly carcinogenic.

OSHA and EPA have studied fiberglass intermittently since the 1970s, but efforts within both agencies to tighten standards have stalled. EPA internally classifies glass fiber as a possible carcinogen. At OSHA, fibrous glass is still regulated as if it were any other form of "nuisance dust," though the agency has put it on a "priority" list for
future regulation. Ultimately, however, the agency's basic ability to draft any new limits has been stymied by a court challenge unrelated to fiberglass. It will likely be years before OSHA takes any action, officials said.

For now, both agencies have joined CPSC and NIOSH to form a "fiber" task force to assess the actual risks, if any, for users of various man-made fibers.

Meanwhile, the industry's greatest fear -- that consumers would overreact to reports of the ARC listing -- has not been realized. Since 1993, sales have risen steadily, the industry says, influenced by fluctuations in housing construction rather than any consumer anxiety. On the other hand, the industry can't stop worrying: Just last September, for example, the Natural Resources Defense Council reviewed scientific literature on competing insulation materials and reported that fiberglass had the "most well-documented" risks. The documentation, of course, is to a large degree the result of the industry's $20 million worth of research.

We Never Talk About It

Danny Humphrey and his sister, Mary Maharg, sat at her kitchen table in Newark, reminiscing about family and friends who have worked at Owens-Corning for generations.

As kids, they remembered, they would get itchy from the fibers they'd pick up from their father's work clothes and they'd break out in rashes that looked like measles. Their father, Sewell, worked at Owens for 45 years, and they recalled him sometimes sitting at the kitchen table pressing his hands to his chest, as if he were in pain. He developed multiple tumors and died of prostate cancer.

Their mother, Tollie, also worked in the fiberglass plant in the years before she had kids. They remember her coughing blood and being told she had scars on her lungs, but the cause was never pinpointed. Mary's husband, George, worked at Owens 10 years and also coughed blood occasionally and had trouble breathing, she said, until he left.

Their four brothers have worked at Owens, as have their uncles, aunts, brothers-in-law, and many friends -- most of whom were also smokers.

"Uncle Anton's 70. Smoked a few years. Lungs congested. He's got breathing problems," Maharg said. "Our neighbor Chuck is in a comatose state. Lung cancer, age 44. Another guy was 47 and he died, that's nothing new. Another neighbor, he's got cancer and he's working there part time . . . This other guy, he was 47, too. No cause of death listed."

Joel Bender, Owens's vice president for health and safety, who supervises a staff of 35, said the ongoing studies of Newark convince him that smoking is the primary culprit in the excess cancers. Owens is now a smoke-free workplace where employees are given regular health screenings and X-rays, he said, and their troubles are inevitably traced either to cigarettes or to causes other than glass fiber, such as allergies, nasal dryness or temporary irritations.

Bender said he is firmly convinced that fibrous glass cannot be blamed for lung disease. "This industry has tested its products. As technology changes, we have changed our products and we have changed our testing," Bender said. The ongoing animal inhalation experiments are yielding new knowledge about many of the 3,000 varieties of man-made fibers. "We have come a long way in the last five years, to the point where we know which formulations are dangerous" -- such as high-temperature ceramic fibers,
which, he said, the company does not produce.

With the added testing, he said, "we have moved from `All fibers are bad' to `All fibers are good' to somewhere in between, and I think it's going to serve the public well."

In the meantime, the Humphreys, the Mahargs and others in Newark are also caught somewhere in between. After the cancer warning labels in the 1980s, people took notice, at least a little, Danny Humphrey said. "We know it's not good for you, but we never talk about it . . . Well, us workers, we talked about it, but what else can you do about it?"

Humphrey said he can't believe that all the illness and death he has seen is caused by everything else other than the fibers. "It makes you wonder," he said, "but I don't know how you would prove it."

Peter Perl is a staff writer for the Magazine. Li Fellers is a freelance writer based in Washington. D.C.