Chronic beryllium disease (CBD) is by far the most interesting and humbling problem I encountered in my 32-year career at the U.S. Department of Energy (DOE). In the 1980s, my colleagues included industrial hygienists who had started working in the atomic weapons complex during the Manhattan Project. From them I heard stories about the industrial hygienists who had studied CBD and published the standards and methods for prevention in the 1940s. These were legends in the profession like Merrill Eisenbud and Al Breslin of the Atomic Energy Commission’s Health and Safety Laboratory and Harry Schulte of Los Alamos National Laboratory. AEC had been in the middle of the 1970s controversy over whether OSHA should regulate beryllium as a carcinogen, and I was an eyewitness to the events of the 1980s and 1990s when CBD made its comeback. This article recounts beryllium’s complex history as a workplace hazard and the difficult problems industrial hygienists encountered when trying to establish protective programs.

ACUTE BERYLLIUM DISEASE By the early 1940s, wartime demand for beryllium-containing materials and labor shortages meant that lost workdays and high turnover from illnesses in beryllium production plants needed to be addressed. As explained in Beryllium: Biomedical and Environmental Aspects, the U.S. Public Health Service issued a bulletin in 1943 stating that beryllium was non-toxic based on its animal toxicology studies and concluded that the illnesses were the result of exposure to the fluoride acid radicals that were also present where beryl ore was being processed. That year Dr. Howard Van Ordstrand of the Cleveland Clinic published a description of the chemical pneumonia occurring among workers exposed to beryllium oxide and where fluorides were not present. By the late 1940s, more systematic animal toxicology studies of beryllium compounds showed that solubility was the key determinant of dose required to produce the effect. As Eisenbud described in his 1990 book An Environmental Odyssey, the more soluble the compound, the more acutely toxic it was. However, the USPHS did not withdraw its bulletin until the 1970s, and it continued to be used by skeptics to downplay the hazards of beryllium.

There were two versions of chemical pneumonia: an acute version caused by single, very high exposures, and a subacute version that would occur after weeks or months of cumulative exposure. Meanwhile, beryllium skin diseases occurring in these plants were of three distinct types: a contact dermatitis due to the irritant effect of soluble beryllium; an eczema allergic reaction to soluble beryllium; and granulomatous ulcers from implantation of beryllium slivers or wounds contaminated with beryllium oxide. This raised the question of whether the subacute pneumonia was an allergic response. However, the simpler explanation prevailed: an exposure rate exceeding the biologic removal rate led to an accumulation of large-enough doses in the lung to cause the disease. Finally, in 2007, a report in Environmental Health Perspectives on two subacute disease cases exposed to beryllium fluoride established that the results of bronchoalveolar lavage were a cell-mediated immune response. This provides a plausible explanation of why CBD was not identified among production plant workers in the 1940s: genetically susceptible workers who developed skin- and sub-acute lung disease were advised by physicians to not go back to that plant because next time they would get sicker sooner. Median durations of employment in these plants were less than 30 days. Susceptible workers probably didn’t remain employed long enough for an association between employment and the longer latency disease to be noticed.

CHRONIC BERYLLIUM DISEASE In the mid-1940s, it was discovered that 17 young women in a cluster of sarcoidosis cases in Salem, Massachusetts, all worked at a Sylvania phosphorescent light tube plant. As explained in a 1946 paper in The Journal of Industrial Hygiene and Toxicology, the affected workers were handling a beryllium-containing silicate phosphor. The phosphor in this plant contained 12 percent unreacted beryllium oxide, and no cases occurred among employees hired after the beryllium oxide content was reduced to less than 2 percent, as Eisenbud and coauthor JH Sterner would show in a 1951 paper. These were the first cases of CBD reported in the literature. Physicians treating patients from the Lorain, Ohio plant began to consider the possibility that five fatal cases among neighbors of the plant originally diagnosed as sarcoidosis were in fact CBD.

AEC’s New York Operations Office managed beryllium procurement contracts for the metal and oxide products being used in nuclear weapons and reactor development. Eisenbud, who worked for the New York Operations Office Health and Safety Laboratory (HASL), became a central player in organizing a response. The Ohio Health Department offered x-ray examinations to residents in Lorain, and HASL initiated studies of ambient air pollution levels in the neighborhood.

BERYLLIUM EXPOSURE LIMITS AEC established three beryllium exposure limits: the 30-day ambient air limit of 0.01 μg/m3, a 30-minute limit of 25 μg/m3, and a daily weighted average (DWA) limit of 2 μg/m3. The DWA limit, according to Eisenbud, was a back-of-the-envelope calculation based on the assumption that beryllium was as toxic as mercury or lead, adjusted for its lower molecular weight and with an uncertainty factor of 10 applied. To provide a better basis for the DWA, Eisenbud established animal toxicology studies and a disease case registry. Finally, the AEC established an advisory committee to review these standards every year. The toxicology program provided good animal models for understanding the acute disease but not the chronic granulomatous disease. The case registry never succeeded in collecting exposure monitoring data as had been hoped. After a few years the advisory committee disbanded because of the lack of any new information to review.

Looking back, it is easy to see that today’s threshold limit value of 0.05 μg/m3 was sitting in front of them the whole time. Rounding up the ambient air limit, multiplying by 168 hours in a 7-day, 24-hour week, and dividing by a 40-hour work week gets you there. But the levels in production plants were orders of magnitude higher than this, and even after some CBD cases among former workers were identified, people thought that the aerosol in the Lorain plant must be much less toxic than the aerosols that caused the neighborhood cases. And there was a difference: emissions from the ore processing operations that contained hydrogen fluoride were emitted from a tall stack and not contributing to the aerosols being measured. The source of emissions affecting the neighborhood was fumes from the master copper alloy-casting operations, which were recovered for recycling with rooftop baghouse filters. In their 1951 paper, Sterner and Eisenbud had focused on beryllium oxide as the cause of CBD because phosphors with only trace levels of unreacted beryllium oxide didn’t cause the disease and cases from AEC labs were being reported from fabrication of beryllium metal and ceramic components.

The next change came in 1959 when ACGIH adopted the 2 μg/m3 limit as its 8-hour time-weighted average TLV. At this time, government contractors were subject to DOL safety and health standards enforced under the Walsh-Healy Act. DOL adopted TLVs as its OELs, which were enforced by AEC. Thus, the TLV became the enforceable standard not only at AEC facilities but at any facility with a government contract.

There is an important difference between an 8-hour TWA and a DWA. At the time, air samples for beryllium were being collected with line-powered, high-volume samplers operating at 200–400 liters per minute on a 4-inch filter. Industrial hygienists would monitor tasks by holding the sampler in the worker’s breathing zone and then asking workers to estimate how often they performed the task. The DWA was calculated every quarter by time-weighting task and general area monitoring results to estimate quarterly mean exposure levels. Thus, the TLV was quite a bit more stringent than the AEC DWA limits they replaced. A barrier to adoption of the TLV was that analytical methods had yet to be developed that could achieve reporting limits less than the TLV in the volume of air collected with wearable, battery-powered pumps.

By this time the 2 μg/m3 DWA was perceived to have eliminated CBD and there was resistance to adoption of the TLV. In ANSI’s 1970 Z37.29 standard, Acceptable Concentrations of Beryllium and Beryllium Compounds, 2 μg/m3 was once again a DWA, 5 μg/m3 was the 8-hour TWA, and 25 μg/m3 was a 30-minute limit. The newly formed OSHA adopted these limits into its Z2 table, which confused two generations of industrial hygienists by labeling 2 μg/m3 an 8-hour TWA, 5 μg/m3 a ceiling limit, and 25 μg/m3 a 30-minute peak limit. Note that the ANSI committee felt that a DWA of 2 μg/m3 and an 8-hour TWA of 5 μg/m3 were roughly equivalent, which is consistent with other published estimates of the relationship between compliance with a limit and long-term arithmetic mean exposure.

Along the way to today’s standards, ACGIH added 15-minute short-term exposure limits (STELs) of 25 μg/m3 from 1976 to 1978 and 10 μg/m3 from 1998 to 2008. In the mid-1970s, NIOSH published a recommended exposure limit of 0.5 μg/m3 as an 8-hour TWA based on beryllium’s carcinogenicity. DOE issued its Chronic Beryllium Disease Prevention Program rule (10 CFR 850) in 2000 with an action level of 0.2 μg/m3 as an 8-hour TWA OEL, which was also adopted by the state of California. In 2009, ACGIH dropped the STEL and lowered the 8-hour TWA TLV to 0.05 μg/m3. Finally, in 2020, OSHA established its 8-hour TWA PEL of 0.2 μg/m3, an action level of 0.1 μg/m3, and a 15-minute STEL of 2 μg/m3. Because the TLV-TWA precludes any 15-minute period in a shift from exceeding 1.6 μg/m3, it is the most stringent standard and compliance with it ensures compliance with all other limits.

CBD “DISAPPEARS” By the 1960s, CBD case-finding efforts ceased functioning. Exposure controls largely eliminated acute cases, and the chronic cases did not attract the attention of the medical community. The beryllium case registry evolved into a program of research on diagnosis and treatment of CBD, including the first use of corticosteroids to treat an illness.

Like other industrial hygienists, I became convinced that workplaces were safe. One factor in this belief was that controlling exposures to levels less than 2 μg/m3 was always going to be a reach. A great deal of effort went into designing the exhaust ventilation systems needed to achieve these results, which took years to implement. I think the level of effort created the expectation that conditions would be protective and the exceedances that continued to happen were discounted as outliers. Another factor is that the accuracy of the DWA method was highly dependent on industrial hygienists identifying tasks that created significant near-field exposures. This method misses unplanned events; over time, predictable sources of exposure will be controlled, leaving the less predictable, and creating a more skewed distribution of exposure monitoring results. Outliers in a normal distribution are not outliers in a lognormal distribution, and these few high results are the main source of risk. Because the DWA was a quarterly mean, I think individual exceedances were not investigated in a timely manner. In contrast, an 8-hour TWA encourages a more immediate investigation that is more likely to identify interventions that can prevent reoccurrence.

At DOE’s Rocky Flats plant near Denver, beryllium exposures were being monitored with fixed devices plumbed to a vacuum system that operated at about 1 cfm (28 lpm) and collected the sample on a 37 mm filter. The monitors were located next to machine tools and other aerosol-producing process equipment. Measuring levels at the source was thought to provide a conservative overestimate of workers’ actual exposure. The first set of personal samples in the mid-1980s was therefore a big surprise, exceeding levels at the source by about a factor of 6. In retrospect, the fixed monitoring method reflected overconfidence in IHs’ ability to identify sources of exposure for control through their senses alone.

CBD REAPPEARS Of course, CBD was there the whole time. As explained in a 1985 paper in the British Journal of Industrial Medicine, the cases were attributed to exposures well above limits. The ability to study the prevalence of CBD didn’t become possible until the beryllium lymphocyte proliferation test (BeLPT) was developed. Studies published in 1993 by the Journal of Occupational Medicine and the American Review of Respiratory Disease established the usefulness of the BeLPT for workers in a ceramics plant and a nuclear weapons facility. It became obvious that cases among bystander workers at DOE’s Rocky Flats plant couldn’t be explained by non-compliance with the existing limits. Cases included a secretary and a security guard from the self-reported no-or-minimal-exposure group.

A prevalence study at DOE’s Y-12 plant published in Toxicology in 2003 produced similar results to the 1993 studies of workers fabricating ceramic and metal components. In addition, the first large set of personal breathing zone monitoring results supported the conclusion that working conditions had met the usual definition of compliance with fewer than 5 percent of results exceeding the 2 μg/m3 limit. The experiences at Rocky Flats and Y-12 became the basis for the publication of 10 CFR 850, which extended use of the BeLPT to all DOE sites. In addition, the rule lowered the OEL by a factor of ten because that is what you do when the only thing you know is that the old limit didn’t work.

By shining a light on CBD, the BeLPT provided industrial hygienists an opportunity to further reduce exposures. Of the 64,842 exposure monitoring results reported to the Beryllium-Associated Worker Registry (BAWR) in the ten-year period of 2008–2017, 4.8 percent of results were above laboratory reporting limits and 0.6 percent exceeded the 0.2 μg/m3 limit. In addition, the BeLPT allows workers to be removed from exposure, reducing the chances that a patient will progress to disabling disease.

BERYLLIUM CARCINOGENESIS In October 1975, OSHA published a proposed rule for beryllium as part of its policy objective of reducing exposures to carcinogens to levels as low as technically feasible. At the time, this level for beryllium was judged to be 1 μg/m3. The beryllium industry, DOE, and the Department of Defense submitted testimony challenging the technical basis for the proposed rule. It was withdrawn after the 1978 Supreme Court benzene decision that overturned OSHA’s carcinogen policy.

The proposed rule was disputed on two main issues. The first objection concerned the feasibility of achieving the proposed permissible exposure limit of 1 μg/m3. More problematic were criticisms by a group of industry-supported university professors of the NIOSH epidemiology study, which reviewed mortality of employees at seven beryllium production plants. Lung cancer deaths were elevated but the standardized mortality ratios were small. It was an odd occupational cohort in that the majority were employed fewer than 30 days, and these short-term employees had higher lung cancer risk than employees of longer duration—a seemingly backward dose-response relationship. The study was also criticized for weaknesses in the methods used to control for cigarette smoking and for work in other industries.

NIOSH continued to follow this beryllium production plant cohort with periodic updating by collecting additional death certificates. Because NIOSH persisted through the years in updating and refining its studies and expanding it to a case control study of selected plants, a very good no-observed-adverse-effects level in humans has been established. In the 2001 case control study, the NOAEL of 4 μg/m3 is a grand working lifetime average. In the 2011 cohort mortality study, the NOAEL of 10 μg/m3 is an average for the job with the highest exposures among those cases. Because these estimates are based on DWA measurements, they have some additional conservatism built in.

HIDDEN COSTS OF OCCUPATIONAL DISEASE Popular suspicions of a nuclear weapons complex created scrutiny of the response when CBD made its reappearance in the late 1980s and early 1990s. In DOE this was playing out in the context of the collapse of the Soviet Union, the work of the Department shifting from weapons production to clean-up, and the workforce shifting from predominantly industrial to predominantly building trades. In the late 1980s, a top legislative initiative of the AFL-CIO was the creation of a medical surveillance program to be operated by NIOSH and offered to workers separated from employment but still at risk for occupational disease. When this failed to garner support, a demonstration project for DOE workers was included in the Defense Authorization Act of 1993.

A just-completed Rocky Flats study had been supported by the local union but had been under the radar of the AFL-CIO’s national leaders. DOE planned to follow up with screening for Rocky Flats workers who hadn’t had a chance to participate and to replicate the study at Y-12. Initially the unions were worried that DOE beryllium workers who tested positive on the BeLPT would be fired. As a result, DOE received a great deal of help from unions in designing the informed consent process and job protections for current employees.

As former worker projects expanded nationwide, some states’ statutes of limitation on workers’ compensation meant that in a few instances the projects were screening people for CBD without the ability to refer them on for definitive diagnosis and care. DOE sought to address CBD compensation through legislation that Congress expanded to include the radiation-relatable cancers. This gained bipartisan support across the country. The in-person testimony of workers suffering significant disability from CBD, an indisputable occupational disease, was a major factor in the legislation passing. To date, over $20 billion in compensation, medical benefits, and survivor benefits have been paid for by this program.

LEARNING FROM HISTORY Few, if any, workplace hazards have presented as many challenges as beryllium. What can we learn from this history? Here are four major lessons for industrial hygienists:

People get sick. Progress in preventing death and disability from one cause results in the rise of another. With beryllium, the cause of occupational morbidity and mortality evolved from acute disease to cancer and now immune-system mediated effects. Continued progress in extending life expectancy means that occupational exposures become increasingly important attributable risk factors for leading causes of death such as heart disease, cancer, lung disease, and dementia. The burden of occupational exposure on the public’s health is not being measured but probably is growing.

The more you monitor, the less you know. As the operation of an exposure monitoring and control program succeeds in eliminating the predictable sources of exposure, you are left with the unpredictable. With beryllium we exhaust the ability to use our senses, knowledge, and skills to make useful predictions well before meeting safe levels. The inability to make accurate predictions requires more monitoring with fewer assumptions to avoid errors.

Industrial hygienists can be their own worst enemy. We work on the loss-control side of the ledger, and if there are no losses, loss control becomes useless overhead. It is a mistake, then, to position ourselves as solely responsible for workplace protections. Ideally, the organization takes ownership of its health protection program, and the industrial hygienist supports its efforts to set goals and measure progress in reducing the uncertainty.

Outrage is a cost multiplier. In occupational health, there is an imbalance because risks are accepted by managers but taken by workers. Disease cases are concrete confirmation of suspicions and distrust created by this imbalance. If an enterprise can’t be trusted with protecting the health of their own employees, how can they be trusted with anything else?

Send feedback to The Synergist.