Preliminary Beryllium Risk Assessment

August 3, 2010

Preliminary Beryllium Risk Assessment

Introduction

The Occupational Safety and Health (OSH) Act and court cases arising
under it have led OSHA to rely on risk assessment to support the risk
determinations required to set a permissible exposure limit (PEL) for a
toxic substance in standards under the OSH Act. Section 6(b)(5) of the
OSH Act states that “The Secretary [of Labor], in promulgating
standards dealing with toxic materials or harmful agents under this
subsection, shall set the standard which most adequately assures, to the
extent feasible, on the basis of the best available evidence, that no
employee will suffer material impairment of health or functional
capacity even if such employee has regular exposure to the hazard dealt
with by such standard for the period of his working life” (29 U.S.C.
65(b)(5)). 

In Industrial Union Department, AFL-CIO v. American Petroleum Institute
(Benzene), 448 U.S. 607 (1980), the United States Supreme Court ruled
that the OSH Act requires that, prior to the issuance of a new standard,
a determination must be made that there is a significant risk of
material impairment of health at the existing PEL and that issuance of a
new standard will significantly reduce or eliminate that risk. The Court
stated that “before [the Secretary] can promulgate any permanent
health or safety standard, the Secretary is required to make a threshold
finding that a place of employment is unsafe in the sense that
significant risks are present and can be eliminated or lessened by a
change in practices” (448 U.S. at 642). The Court also stated “that
the Act does limit the Secretary’s power to require the elimination of
significant risks” (488 U.S. at 644), and that “OSHA is not required
to support its finding that a significant risk exists with anything
approaching scientific certainty” (488 U.S. at 656).  

OSHA’s approach for the risk assessment incorporates both a review of
the recent literature on populations of workers exposed to beryllium at
and below the current Permissible Exposure Limit (PEL) of 2 µg/m3 and a
statistical exposure-response analysis. A number of recently published
epidemiological studies evaluate the risk of sensitization and CBD for
workers exposed at and below the current PEL and the effectiveness of
exposure control programs in reducing risk. OSHA also conducted a
statistical analysis of the exposure-response relationship for
sensitization and CBD at the current PEL and alternate PELs the Agency
is considering. For this analysis, OSHA used data provided by National
Jewish Medical and Research Center (NJMRC) on a population of workers
employed at a beryllium machining plant in Cullman, AL. 

The review of the epidemiological studies and OSHA’s own analysis show
substantial risk of sensitization and CBD among workers exposed at and
below the current PEL of 2 µg/m3. They also show substantial reduction
in risk where employers have implemented a combination of controls,
including stringent control of airborne beryllium levels and additional
measures such as respirators, dermal personal protective equipment
(PPE), and strict housekeeping to protect workers against dermal and
respiratory beryllium exposure. 

The Health Effects section discussed the role of particle
characteristics and beryllium compound solubility in the development of
sensitization and CBD among beryllium-exposed workers. Respirable
particles small enough to reach the deep lung are responsible for CBD.
However, larger inhalable particles that deposit in the upper
respiratory tract may lead to sensitization. The weight of evidence
indicates that both soluble and insoluble forms of beryllium are able to
induce sensitization and CBD. Insoluble forms of beryllium that persist
in the lung for longer periods may pose greater risk of CBD while
soluble forms may more easily trigger immune sensitization. Although
these factors potentially influence the toxicity of beryllium, the
available data are too limited to reliably account for solubility and
particle size in the Agency estimates or risk. The qualitative impact on
conclusions and uncertainties with regard to risk are discussed in a
later section. 

Lung Cancer 

OSHA considers lung cancer to be an important health endpoint for
beryllium-exposed workers. The International Agency for Research on
Cancer (IARC), National Toxicology Program (NTP), and American
Conference of Governmental Industrial Hygienists (ACGIH) have all
classified beryllium as a known human carcinogen. The National Academy
of Sciences (NAS), Environmental Protection Agency, the Agency for Toxic
Substances and Disease Registry (ATSDR), the National Institute of
Occupational Safety and Health (NIOSH), and other reputable scientific
organizations have reviewed the scientific evidence demonstrating that
beryllium is associated with an increased incidence of cancer. OSHA also
has performed an extensive review of the scientific literature regarding
beryllium and cancer. This includes an evaluation of human
epidemiological, animal cancer, and mechanistic studies described in the
Health Effects Section. Based on the weight of evidence, the Agency has
preliminarily determined beryllium to be an occupational carcinogen. 

Although epidemiological and animal evidence support a conclusion of
beryllium carcinogenicity, there is considerable uncertainty surrounding
the mechanism of carcinogenesis for beryllium. The evidence for direct
genotoxicity of beryllium and its compounds has been limited and
inconsistent (NRC, 2008; IARC, 1993; EPA, 1998; NTP, 2005; ATSDR, 2002).
One plausible pathway for beryllium carcinogenicity described in the
Health Effects Section includes a chronic, sustained neutrophilic
inflammatory response that induces epigenetic alterations leading to the
neoplastic changes necessary for carcinogenesis. The National Cancer
Institute estimates that nearly one-third of all cancers are caused by
chronic inflammation (NCI, 2009). This mechanism of action has also been
hypothesized for crystalline silica and other agents known to be human
carcinogens but have limited evidence of genotoxicity. 

OSHA’s review of epidemiological studies of lung cancer mortality
among beryllium workers found that most did not characterize exposure
levels sufficiently for exposure-response analysis. However, one NIOSH
study evaluated the association between beryllium exposure and lung
cancer mortality based on data from a beryllium processing plant in
Reading, PA (Sanderson et al., 2001a). As discussed in the Health
Effects section, this case-control study evaluated lung cancer incidence
in a cohort of workers employed at the plant from 1940 to 1969 and
followed through 1992. For each lung cancer victim, 5 age- and
race-matched controls were selected by incidence density sampling, for a
total of 142 lung cancer cases and 710 controls. 

Between 1971 and 1992, the plant collected close to 7,000 high volume
filter samples consisting of both general area and short-term,
task-based breathing zone measurements for production jobs and
exclusively area measurements for office, lunch, and laboratory areas
(Sanderson et al., 2001b). In addition, a few (< 200) impinger and
high-volume filter samples were collected by other organizations between
1947 and 1961, and about 200 6-to-8-hour personal samples were collected
in 1972 and 1975. Daily-weighted-average (DWA) exposure calculations
based on the impinger and high-volume samples collected prior to the
1960s showed that exposures in this period were extremely high. For
example, about half of production jobs had estimated DWAs ranging
between 49 and 131 ug/m3 in the period 1935 – 1960, and many of the
“lower-exposed” jobs had DWAs of approximately 20 – 30 ug/m3
(Table II, Sanderson et al., 2001b). Exposures were reported to have
decreased between 1959 and 1962 with the installation of ventilation
controls and improved housekeeping, and following the passage of the OSH
Act in 1970. While no exposure measurements were available from the
period 1961-1970, measurements from the period 1971-1980 showed a
dramatic reduction in exposures plant-wide. Estimated DWAs for all jobs
in this period ranged from 0.1 µg/m3 to 1.9 µg/m3.
Calendar-time-specific beryllium exposure estimates were made for every
job based on the DWA calculations and were used to estimate workers’
cumulative, average, and maximum exposures. Exposure estimates were
lagged by 10 and 20 years in order to account for exposures that did not
contribute to lung cancer because they occurred after the induction of
cancer. 

Results of a conditional logistic regression analysis showed an
increased risk of lung cancer in workers with higher exposures when dose
estimates were lagged by 10 and 20 years (Sanderson et al., 2001a). The
authors noted that there was considerable uncertainty in the estimation
of exposure in the 1940’s and 1950’s and the shape of the
dose-response curve for lung cancer. NIOSH later reanalyzed the data,
adjusting for potential confounders of hire age and birth year
(Shubauer-Berigan et al., 2008). The study reported a significant
increasing trend (p<0.05) in the odds ratio with increasing quartiles of
average [log transformed] exposure lagged 10 years. However, it did not
find a significant trend for quartiles of cumulative (log transformed)
exposure lagged 0, 10, or 20 years.  

 0.1 μg/m3 and 2 μg/m3. The majority of case and control workers in
the Sanderson et al. case-control analysis were first hired during the
1940s when exposures were extremely high (estimated DWAs > 20 µg/m3 for
most jobs). The cumulative, average, and maximum beryllium exposure
concentration estimates for the 142 known lung cancer cases were 46.06
± 9.3µg/m3-days, 22.8 ± 3.4 µg/m3, and 32.4 ± 13.8 µg/m3,
respectively. About two-thirds of cases and half of controls worked at
the plant for less than a year. Thus, a risk assessment based on this
exposure-response analysis would need to extrapolate from very high to
very low exposures, based on a working population with extremely short
tenure. While OSHA risk assessments must often make extrapolations to
estimate risk within the range of exposures of interest, the Agency
acknowledges that these issues of short tenure and extremely high
exposures would create substantial uncertainty in a risk assessment
based on this study population. 

<11.2 μg/m3) and may have had a significant lung cancer risk. This
issue would introduce further uncertainty in lung cancer risks estimated
from this epidemiological study. 

 In developing regulatory risk assessments, OSHA routinely must work
with studies that have significant sources of uncertainty and may
extrapolate from high to low exposures or from short-term to long-term
exposures. The Agency acknowledges that extrapolation can introduce
uncertainty into the results of the risk analysis, particularly when the
mechanism of disease for the chemical of interest is unknown. In most
cases, OSHA proceeds with risk assessment despite these uncertainties
because the Agency is statutorily obligated to determine whether a
significant risk of material impairment exists based on best evidence
available. 

For this rulemaking OSHA has performed a risk analysis based on studies
of beryllium sensitization and CBD in worker populations principally
exposed to levels below the current PEL of 2 μg/m3, which is the range
of interest to OSHA. The risk assessment shows substantial risk and risk
reduction directly observed in working populations exposed to beryllium
below the current PEL. In contrast, the association between beryllium
exposure and lung cancer reported in the NIOSH study involved short-term
workers exposed to much higher beryllium levels than the current or
alternate PELs. While this exposure–response relationship could
potentially be used to develop quantitative estimates of risk for lung
cancer, the results would be highly uncertain and of limited value given
that the Agency’s risk assessment for beryllium sensitization and CBD
shows, with much greater certainty, substantial risks of sensitization
and CBD from exposure at and well below the current PEL. Therefore,
preliminarily the Agency has elected to rely on its risk assessment for
sensitization and CBD to make its findings regarding the significance of
risk at the current PEL and several alternate PELs that OSHA is also
considering. 

Review of Epidemiological Literature on Sensitization and Chronic
Beryllium Disease From Occupational Exposure

As discussed in the Health Effects section, studies of beryllium-exposed
workers conducted using the beryllium lymphocyte proliferation test
(BeLPT) have found high rates of beryllium sensitization and CBD among
workers in many industries, including at some facilities where exposures
were primarily below OSHA's PEL of 2 µg/m3 (Kreiss et al., 1992,
Schuler et al., 2005). In the mid-1990s, some facilities using beryllium
began to aggressively monitor and reduce workplace exposures. Three
plants where several rounds of BeLPT screening were conducted before and
after implementation of new exposure control methods provide the best
currently available evidence on the effectiveness of various exposure
control measures in reducing the risk of sensitization and CBD. The
experiences of these plants, a copper-beryllium processing facility in
Reading, PA; a beryllia ceramics facility in Tucson, AZ; and a machining
facility in Cullman, AL; show that efforts to prevent sensitization and
CBD by using engineering controls to reduce workers’ beryllium
exposures to median levels at or around 0.2 µg/m3 had only limited
impact on risk. However, exposure control programs implemented more
recently, which drastically reduced respiratory exposure to beryllium
via a combination of engineering controls and respiratory protection,
controlled dermal contact with beryllium using PPE, and employed
stringent housekeeping methods to keep work areas clean and prevent
transfer of beryllium between work areas, sharply curtailed new cases of
sensitization among newly-hired workers. A discussion of the experiences
at these three plants follows.

Reading Plant

Schuler et al. conducted a study of workers at a copper-beryllium
processing facility in Reading, PA, screening 152 workers with the BeLPT
(Schuler et al., 2005). Exposures at this plant were believed to be low
throughout its history due to the low percentage of beryllium in the
metal alloys used, and the relatively low exposures found in general
area samples collected starting in 1969 (sample median < 0.1 µg/m3, 97%
< 0.5 µg/m3). The reported prevalences of sensitization (6.5 percent)
and CBD (3.9 percent) showed substantial risk at this facility, even
though airborne exposures were primarily below OSHA’s current PEL of 2
µg/m3. 

Personal lapel samples were collected in production and production
support jobs between 1995 and May 2000. These showed primarily very low
airborne beryllium levels, with a median of 0.073 µg/m3. The wire
annealing and pickling process had the highest personal lapel sample
values, with a median of 0.149 µg/m3. Despite these low exposure
levels, cases of sensitization continued to occur among workers whose
first exposures to beryllium occurred in the 1990s. Five (11.5 percent)
workers of 43 hired after 1992 who had no prior beryllium exposure
became sensitized, including four in production work and one in
production support (Thomas et al., 2009; evaluation for CBD not
reported). Two (13 percent) of these sensitized workers were among 15
workers in this group who had been hired less than a year before the
screening. 

After the BeLPT screening was conducted in 2000, the company began
implementing new measures to further reduce workers’ exposure to
beryllium. Requirements designed to minimize dermal contact with
beryllium, including long-sleeve facility uniforms and polymer gloves,
were instituted in production areas in 2000. In 2001 the company
installed local exhaust ventilation (LEV) in die grinding and polishing.
Personal lapel samples collected between June 2000 and December 2001
show reduced exposures plant-wide. Of 2,211 exposure samples collected,
59 percent were below the limit of detection (LOD) and 98 percent were
below 0.2 ug/m3 (Thomas et al., 2009). Median values below 0.03 ug/m3
were reported for all processes except the wire annealing and pickling
process. Samples for this process remained somewhat elevated, with a
median of 0.1 ug/m3. In January 2002, the plant enclosed the wire
annealing and pickling process in a restricted access zone (RAZ),
required respiratory PPE in the RAZ, and implemented stringent measures
to minimize the potential for skin contact and beryllium transfer out of
the zone. While exposure samples collected by the facility were sparse
following the enclosure, they suggest exposure levels comparable to the
2000-01 samples in areas other than the RAZ. Within the RAZ, required
use of powered-air purifier respirators (PAPRs) indicates that
respiratory exposure was negligible. 

To test the efficacy of the new measures in preventing sensitization and
CBD, in June 2000 the facility began an intensive BeLPT screening
program for all new workers. The company screened workers at the time of
hire; at intervals of 3, 6, 12, 24, and 48 months; and at 3-year
intervals thereafter. Among 82 workers hired after 1999, three cases of
sensitization were found (3.7 percent). Two (5.4 percent) of 37 workers
hired prior to enclosure of the wire annealing and pickling process were
found to be sensitized within 3 and 6 months of beginning work at the
plant. One (2.2 percent) of 45 workers hired after the enclosure was
confirmed as sensitized. Among these early results, it appears that the
greatest reduction in sensitization risk was achieved after median
exposures in all areas of the plant were reduced to below 0.1 µg/m3 and
PPE to prevent dermal contact was instituted. 

Tucson Plant

Kreiss et al. conducted a study of workers at a beryllia ceramics plant,
screening 136 workers with the BeLPT in 1992 (Kreiss et al., 1996).
Full-shift area samples collected between 1983 and 1992 showed primarily
low airborne beryllium levels at this facility. Of 774 area samples, 76
percent were at or below 0.1 µg/m3 and less than 1 percent exceeded 2
µg/m3. A small set (75) of personal lapel samples collected at the
plant beginning in 1991 had a median of 0.2 µg/m3 and ranged from 0.1
to 1.5 µg/m3. However, area samples and short-term breathing zone
samples also showed occasional instances of very high beryllium exposure
levels, with extreme values of several hundred ug/m3 and 3.6 percent of
short-term breathing zone samples in excess of 5 µg/m3. 

Kreiss et al. reported that eight (5.9 percent) of 136 workers tested
were sensitized, six (4.4 percent) of whom were diagnosed with CBD.
Seven of the eight sensitized employees had worked in machining, where
general area samples collected between October 1985 and March 1988 had a
median of 0.3 µg/m3, in contrast to a median value of less than 0.1
µg/m3 in other areas of the plant. Short-term breathing zone
measurements associated with machining had a median of 0.6 µg/m3,
double the median of 0.3 µg/m3 for breathing zone measurements
associated with other processes. One sensitized worker was one of 13
administrative workers screened, and was among those diagnosed with CBD.
Exposures to administrative workers were not well-characterized, but
were believed to be among the lowest in the plant. Of three personal
lapel samples reported for administrative staff during the 1990s, all
were below the then detection limit of 0.2 µg/m3 (Thomas et al., 2009).

Following the 1992 screening, the facility reduced exposures in
machining areas by enclosing machines and installing HEPA filter exhaust
systems. Personal samples collected between 1994 and 1999 had a median
of 0.2 µg/m3 in production jobs and 0.1 in production support. In 1998,
a second screening found that 9 percent of tested workers hired after
the 1992 screening were sensitized, of whom one was diagnosed with CBD.
All of the sensitized workers had been employed at the plant for less
than two years (Henneberger et al., 2001). 

Following the 1998 screening, the company continued efforts to reduce
exposures and risk of sensitization and CBD by implementing additional
engineering and administrative controls and PPE. Respirator use was
required in production areas beginning in 1999, and latex gloves were
required beginning in 2000. The lapping area was enclosed in 2000, and
enclosures were installed for all mechanical presses in 2001. Between
2000 and 2003, water-resistant or water-proof garments, shoe covers, and
taped gloves were incorporated to keep beryllium-containing fluids from
wet machining processes off the skin. The new engineering measures did
not appear to substantially reduce airborne beryllium levels in the
plant. Personal lapel samples collected between 2000 and 2003 had a
median of 0.18 µg/m3, similar to the 1994-1999 samples. However,
respiratory protection requirements were instituted in 2000 to control
workers’ airborne beryllium exposures.

To test the efficacy of the new measures instituted after 1998, in
January 2000 the company began screening new workers for sensitization
at the time of hire and at 3, 6, 12, 24, and 48 months of employment
(Cummings et al., 2007). These more stringent measures appear to have
substantially reduced the risk of sensitization among new employees. Of
97 workers hired between 2000 and 2004, one case of sensitization was
identified (1 percent). This worker had experienced a rash after an
incident of dermal exposure to lapping fluid through a gap between his
glove and uniform sleeve, indicating that he may have become sensitized
via the skin. 

Cullman, AL

Newman et al. conducted a series of BeLPT screenings of workers at a
precision machining facility between 1995 and 1999 (Newman et al.,
2001). A small set of personal lapel samples collected in the early
1980s and in 1995 suggest that exposures in the plant varied widely
during this time period. In some processes, such as engineering,
lapping, and electrical discharge machining (EDM), exposures were
apparently low ( <0.1 µg/m3). Personal lapel samples from all machining
processes combined had a median of 0.33 µg/m3. The majority of these
samples were collected in the high-exposure processes of milling (median
0.3 µg/m3), grinding (median 1.05 µg/m3), and lathing (median 0.35
µg/m3). Newman et al. reported 22 (9.4 percent) sensitized workers
among 235 tested, 13 of whom were diagnosed with CBD within the study
period. 

After a sentinel case of CBD was diagnosed at the plant in 1995, the
company began BeLPT screenings to identify workers at increased risk of
CBD and implemented engineering and administrative controls and PPE
designed to reduce workers’ beryllium exposures in machining
operations. Between 1995 and 1997, the company built enclosures and
installed or updated local exhaust ventilation (LEV) for several
machining departments, removed pressurized air hoses, and required the
use of company uniforms. In contrast to the Reading and Tucson plants,
gloves have never been required at this plant.  

Personal lapel samples collected extensively between 1996 and 1999 have
an overall median was 0.16 µg/m3, showing that the new controls
achieved a marked reduction in machinists’ exposures during this
period. Nearly half of the samples were collected in milling (median =
0.18 µg/m3). Exposures in other machining processes were also reduced,
including grinding (median = 0.18 µg/m3) and lathing (median = 0.13
µg/m3). However, cases of sensitization and CBD continued to occur. 

 At the time that Newman et al. reviewed the results of BeLPT screenings
conducted in 1995-1999, a subset of 60 workers had been employed at the
plant for less than a year. Four (6.7 percent) of these workers were
found to be sensitized, of whom two were diagnosed with CBD and one with
probable CBD (Newman et al., 2001). All four had been hired in 1996. Two
(one CBD case, one sensitized only) had worked only in milling, and had
worked for approximately 3-4 months (0.3 – 0.4 yrs) at the time of
diagnosis. One of those diagnosed with CBD worked only in EDM, where
lapel samples collected between 1996 and 1999 had a median of 0.03
µg/m3. This worker was diagnosed with CBD in the same year that he
began work at the plant. The last CBD case worked as a shipper, where
exposures in 1996-1999 were similarly low, with a median of 0.09 µg/m3.

Beginning in 2000, exposures in all jobs at the machining facility were
reduced to extremely low levels. Personal lapel samples collected in
machining processes between 2000 and 2005 had a median of 0.09 µg/m3,
where more than a third of samples came from the milling process (n =
765, median = 0.09 µg/m3). A later publication on this plant by Madl et
al. reported that only one worker hired after 1999 became sensitized.
This worker had been employed for 2.7 years in chemical finishing, where
exposures were roughly similar to other machining processes (n = 153,
median = 0.12 µg/m3). Madl et al. did not report whether this worker
was evaluated for CBD. 

Preliminary Conclusions 

The published literature on beryllium sensitization and CBD shows that
risk of both can be substantial in workplaces in compliance with
OSHA’s current PEL (Kreiss et al., 1992, Schuler et al., 2005). The
experiences of several facilities in developing effective industrial
hygiene programs have shown that minimizing both airborne and dermal
exposure, using a combination of engineering and administrative
controls, respiratory protection, and dermal PPE, has substantially
lowered workers' risk of beryllium sensitization. In contrast,
risk-reduction programs that relied primarily on engineering controls to
reduce workers’ exposures to median levels in the range of 0.1 – 0.2
µg/m3, such as those implemented in Tucson following the 1992 survey
and in Cullman during 1996-1999, had only limited impact on reducing
workers’ risk of sensitization. The prevalence of sensitization among
workers hired after such controls were installed at the Cullman plant
remained high (Newman et al. (6.7 percent) and Henneberger et al. (9
percent)). A similar prevalence of sensitization was found in the
screening conducted in 2000 at the Reading plant, where the available
sampling data show median exposure levels of less than 0.2 µg/m3 (6.5
percent). The risk of sensitization was found to be particularly high
among newly-hired workers (< 1 year of beryllium exposure) in the
Reading 2000 screening (13 percent) and the Tucson 1998 screening (16
percent).

Cases of CBD have also continued to develop among workers in facilities
and jobs where median exposures were below 0.2 µg/m3. One case of CBD
was found in the Tucson 1998 screening among nine sensitized workers
hired less than two years previously (Henneberger et al., 2001). At the
Cullman plant, at least two cases of CBD were found among four
sensitized workers screened in 1995-1999 and hired less than a year
previously (Newman et al., 2001). These results suggest a substantial
risk of progression from sensitization to CBD among workers exposed at
levels well below the current PEL, especially considering the extremely
short time of exposure and follow-up for these workers. Six of 10
sensitized workers identified at Reading in the 2000 screening were
diagnosed with CBD. The four sensitized workers who did not have CBD at
their last clinical evaluation had been hired between one and five years
previously; therefore, the time may have been too short for CBD to
develop.

 In contrast, more recent exposure control programs that have used a
combination of engineering controls, PPE, and stringent housekeeping
measures to reduce workers' airborne and dermal exposures have
substantially lowered risk of sensitization among newly-hired workers.
Of 97 workers hired between 2000 and 2004 in Tucson, where respiratory
and skin protection was instituted for all workers in production areas,
only one (1 percent) worker became sensitized, and in that case the
worker’s dermal protection had failed during wet-machining work
(Thomas et al., 2009). In Reading, where in 2000-2001 airborne exposures
in all jobs were reduced to a median of 0.1 µg/m3 or below (measured as
an 8-hour TWA) and dermal protection was required for production-area
workers, two (5.4 percent) of 37 newly hired workers became sensitized.
After the process with the highest exposures (median 0.1 µg/m3) was
enclosed in 2002 and workers in that process were required to use
respiratory protection the remaining jobs had very low exposures
(medians ~ 0.03 µg/m3). Among 45 workers hired after the enclosure, one
was found to be sensitized (2.2 percent). 

Madl et al. reported one case of sensitization among workers at the
Cullman plant hired after 2000. The median personal exposures were about
0.1 µg/m3 or below for all jobs during this period. Several changes in
the facility’s exposure control methods were instituted in the late
1990s that were likely to have reduced dermal as well as respiratory
exposure to beryllium. For example, the plant installed change/locker
rooms for workers entering the production facility, instituted
requirements for work uniforms and dedicated work shoes for production
workers, and implemented annual beryllium hazard awareness training that
encouraged glove use, and purchased high efficiency particulate air
(HEPA) filter vacuum cleaners for workplace cleanup and decontamination.


The results of the Reading and Tucson studies show that reducing
airborne exposures to below 0.1 µg/m3 and protecting workers from
dermal exposure, in combination, have achieved a substantial reduction
in sensitization risk among newly-hired workers. The reduction is
particularly evident when comparing newly-hired workers in the most
recent Reading screenings (2.2 - 5.4 percent), and the rate of
sensitization found among workers hired within the year before the 2000
screening (13 percent). There is a similarly striking difference between
the rate of prevalence found among newly-hired workers in the most
recent Tucson study (1 percent) and the rate found among workers hired
within the year before the 1998 screening at that plant (16 percent).
These results are echoed in the Cullman facility, which combined
engineering controls to reduce airborne exposures to below 0.1 µg/m3
with measures such as housekeeping improvements and worker training
reduced dermal exposure.  

The surveys examining the efficacy of these more recent programs to
reduce workers' risk of sensitization and CBD were conducted on
populations with very short exposure and follow-up time. Therefore, they
could not address the question of how frequently workers who become
sensitized in environments with extremely low airborne exposures (median
<= 0.1 µg/m3) develop CBD. In the studies discussed above, clinical
evaluation for CBD was not reported for sensitized workers identified in
the most recent Tucson and Reading studies. In Cullman, however, two of
the workers with CBD had been employed for less than a year and worked
in jobs with very low exposures (median 8-hour personal sample values of
0.03 - 0.09 ug/m3). 

The body of scientific literature on occupational beryllium disease also
includes case reports of workers with CBD who are known or believed to
have experienced minimal beryllium exposure, such as a worker employed
only in shipping at a copper-beryllium distribution center (Stanton et
al., 2006), and workers employed only in administration at a beryllium
ceramics facility (Kreiss et al., 1996). 

Review of Community-Acquired CBD Literature

	The literature on community-acquired chronic beryllium disease (CA-CBD)
includes additional sources of information on CBD among individuals
exposed to low levels of airborne beryllium. Cases of CA-CBD were first
reported among residents of Lorain, OH, and Reading, PA, who lived in
the vicinity of beryllium plants. More recently, BeLPT screening has
been used to identify additional cases of CA-CBD in Reading. While case
reports and the CA-CBD literature provide little information on how
frequently workers exposed to very low airborne levels develop CBD, they
demonstrate that individuals exposed to very low levels of airborne
beryllium can develop CBD.

Lorain, OH

In 1948, the State of Ohio Department of Public Health conducted an
X-ray program surveying more than 6,000 people who lived within 1.5
miles of a Lorain beryllium plant (Eisenbud, 1949; Eisenbud, 1982;
Eisenbud, 1998). This survey, together with a later review of all
reported cases of CBD in the area, found 13 cases of CBD. All of the
residents who developed CBD lived within 0.75 miles of the plant, and
none had occupational exposure or lived with beryllium-exposed workers.
Seven residents diagnosed with CBD (1.4 percent) lived within 0.25 miles
of the plant, where about 500 people lived. Five cases were diagnosed
among residents living between 0.25 and 0.5 miles from the plant, one
case was diagnosed among residents living between 0.5 and 0.75 miles
from the plant, and no cases were found among those living farther than
0.75 miles from the plant (total populations not reported) (Eisenbud,
1998). 

Beginning in January 1948, air sampling was conducted using a mobile
sampling station to measure atmospheric beryllium downwind from the
plant. An approximate concentration of 0.2 µg/m3 was measured at 0.25
miles from the plant’s exhaust stack, and concentrations decreased
with greater distance from the plant, to 0.003 µg/m3 at a distance of 5
miles (Eisenbud, 1982). A 10-week sampling program was conducted using
three fixed monitoring stations within 700 feet of the plant and one
station 7,000 feet from the plant. Interpolating the measurements
collected at these locations, Eisenbud and colleagues estimated an
average airborne beryllium concentration of between 0.004 and 0.02
µg/m3 at a distance of 0.75 miles from the plant. Accounting for the
possibility that previous exposures may have been higher due to
production level fluctuations and greater use of rooftop emissions, they
concluded that the lowest airborne beryllium level associated with
CA-CBD in this community was somewhere between 0.01 µg/m3 and 0.1
µg/m3 (Eisenbud, 1982).

Reading, PA

Thirty-two cases of CA-CBD were reported in a series of papers published
in 1959-1969 concerning a beryllium refinery in Reading (Lieben and
Metzner, 1959; Metzner and Lieben, 1961; Dattoli et al., 1964; Lieben
and Williams, 1969). The plant, which opened in 1935, manufactured
beryllium oxide, alloys and metal, and beryllium tools and metal
products (Maier et al., 2008, Sanderson et al., 2001b). In a follow-up
study, Maier et al. presented eight additional cases of CA-CBD who had
lived within 1.5 miles of the plant (Maier et al., 2008). Individuals
with a history of occupational beryllium exposure and those who had
resided with occupationally exposed workers were not classified as
having CA-CBD.

The Pennsylvania Department of Health conducted extensive environmental
sampling in the area of the plant beginning in 1958. Based on samples
collected in 1958, Maier et al. stated that most cases identified in
their study would typically have been exposed to airborne beryllium at
levels between 0.0155 and 0.028 µg/m3 on average, with the potential
for some excursions over 0.35 µg/m3 (Maier et al., p. 1015, 2008). To
characterize exposures to cases identified in the earlier publications,
Lieben and Williams cited a sampling program conducted by the Department
of Health between January and July 1962, using nine sampling stations
located between 0.2 and 4.8 miles from the plant. They reported that 72
percent of 24-hour samples collected were below 0.01 µg/m3. Of samples
that exceeded 0.01 µg/m3, most were collected at close proximity to the
plant (e.g., 0.2 miles from the plant). 

In the early series of publications, cases of CA-CBD were reported among
people living both close to the plant (Maier et al., 2008; Dutra, 1948)
and up to several miles away. Of new cases identified in the 1968
update, all lived between 3 and 7.5 miles from the plant. Lieben and
Williams suggested that some cases of CA-CBD found among more distant
residents might have resulted from working or visiting a graveyard
closer to the plant (Lieben and Williams, 1969). For example, a milkman
who developed CA-CBD had a route in the neighborhood of the plant.
Another resident with CA-CBD had worked as a cleaning woman in the area
of the plant, and a third worked within a half-mile of the plant. 

At the time of the final follow-up study (1968), 11 residents diagnosed
with CA-CBD were alive and 21 were deceased. Among those who had died,
berylliosis was listed as the cause of death for three, including a
10-year-old girl and two women in their sixties. Fibrosis, granuloma or
granulomatosis, and chronic or fibrous pneumonitis were listed as the
cause of death for eight more of those deceased. Histologic evidence of
CBD was reported for nine of 12 deceased individuals who had been
evaluated for it. In addition to showing radiologic abnormalities
associated with CBD, all living cases were dyspneic. 

Following the 1969 publication by Liebman and Williams, no additional
CA-CBD cases were reported in the Reading area until 1999, when a new
case was diagnosed. The individual was 72-year-old woman who had had
abnormal chest x-rays for the previous six years (Maier et al., 2008).
After the diagnosis of this case, Maier et al. reviewed medical records
and/or performed medical evaluations, including BeLPT results for 16
community residents who were referred by family members or an attorney. 

Among those referred, eight cases of definite or probable CBD were
identified between 1999 and 2002. All eight were women who lived between
0.1 and 1.05 miles from the plant, beginning between 1943-1953 and
ending between 1956-2001. Five of the women were considered definite
cases of CA-CBD, based on an abnormal blood or lavage cell BeLPT and
granulomatous inflammation on lung biopsy. Three probable cases of
CA-CBD were identified. One had an abnormal BeLPT and radiography
consistent with CBD, but granulomatous disease was not pathologically
proven. Two met Beryllium Case Registry epidemiologic criteria for CBD
based on radiography, pathology and a clinical course consistent with
CBD, but both died before they could be tested for beryllium
sensitization. One of the probable cases, who could not be definitively
diagnosed with CBD because she died before she could be tested, was the
mother of both a definite case and the probable case who had an abnormal
BLPT but did not show granulomatous disease.  

The individuals with CA-CBD identified in this study suffered
significant health impacts from the disease, including obstructive,
restrictive, and gas exchange pulmonary defects in the majority of
cases. All but two had abnormal pulmonary physiology. Those two were
evaluated at early stages of disease following their mother’s
diagnosis. Six of the eight women required treatment with prednisone, a
step typically reserved for severe cases due to the adverse side effects
of steroid treatment. Despite treatment, three had died of respiratory
impairment from CBD as of 2002 (Maier et al., 2008). The authors
concluded that “low levels of exposures with significant disease
latency can result in significant morbidity and mortality” (Maier et
al., p. 1017). The studies conducted in Lorain and Reading demonstrate
that long-term exposure to levels as low as 0.1 µg/m3 and below, with
sufficient disease latency, can lead to serious or fatal CBD. 

Exposure-Response Literature on Beryllium Sensitization and CBD

To further examine the relationship between exposure level and risk of
both sensitization and disease, we next review exposure-response studies
in the CBD literature. Many publications have reported that exposure
levels correlate with risk, including a small number of
exposure-response analyses. Most of these studies examined the
association between job-specific beryllium air measurements and
prevalence of sensitization and CBD. This section focuses on two studies
that included a more rigorous historical reconstruction of individual
worker exposures in their exposure – response analyses. 

Viet et al. – Rocky Flats Facility

In 2000, Viet et al. published a case-control study of participants in
the Rocky Flats Beryllium Health Surveillance Program (BHSP), which was
established in 1991 to screen workers at the Department of Energy's
Rocky Flats nuclear weapons facility for beryllium sensitization and
evaluate sensitized workers for CBD (Viet et al., 2000). The program,
which at the time of publication had tested over 5000 current and former
Rocky Flats employees, had identified a total of 127 sensitized
individuals as of 1994 when Viet et al. initiated their study.

Workers were considered sensitized if two BLPT results were positive,
either from two blood draws or from a single blood draw analyzed by two
different laboratories. All sensitized individuals were offered clinical
evaluation, and 51 were diagnosed with CBD based on positive lung LPT
and evidence of noncaseating granulomas upon lung biopsy. The number of
sensitized individuals who declined clinical evaluation was not
reported. Two cases, one with CBD and one who was sensitized but not
diagnosed with CBD, were excluded from the case-control analysis due to
reported or potential prior beryllium exposure at a ceramics plant.
Another sensitized individual who had not been diagnosed with CBD was
excluded because she could not be matched by the study's criteria to a
non-sensitized control within the BHSP database. Viet et al. matched a
total of 50 CBD cases to 50 controls who were negative on the BeLPT and
had the same age (+ 3 years), gender, race and smoking status, and were
otherwise randomly selected from the database. Using the same matching
criteria, 74 sensitized workers who were not diagnosed with CBD were
age-, gender-, race-, and smoking status-matched to 74 control
individuals who tested negative by the BeLPT from the BHSP database.

Viet et al. developed exposure estimates for the cases and controls
based on daily beryllium air samples collected in one of 36 buildings
where beryllium was used at Rocky Flats, the Building 444 Beryllium
Machine Shop. Over half of the approximately 500,000 industrial hygiene
samples collected at Rocky Flats were taken from this building. Air
monitoring in other buildings was reported to be limited and
inconsistent and, thus, not utilized in the exposure assessment. The
sampling data used to develop worker exposure estimates were exclusively
Building 444 fixed airhead (FAH) area samples collected at permanent
fixtures placed around beryllium work areas and machinery. 

Exposure estimates for jobs in Building 444 were constructed for the
years 1960-1988 from this database. Viet et al. worked with Rocky Flats
industrial hygienists and staff to assign a “building area factor”
(BAF) to each of the other buildings, indicating the likely level of
exposure in a building relative to exposures in Building 444. Industrial
hygienists and staff similarly assigned a job factor (JF) to all jobs,
representing the likely level of beryllium exposure relative to the
levels experienced by beryllium machinists. For example, administrative
work and vehicle operation were assigned a JF of 1, while machining,
mill operation, and metallurgical operation were each assigned a JF of
10. Estimated FAH values for each combination of job, building and year
in the study subjects' work histories were generated by multiplying
together the job and building factors and the mean annual FAH exposure
level. Using data collected by questionnaire from each BHSP participant,
Viet et al. reconstructed work histories for each case and control,
including job title and building location in each year of their
employment at Rocky Flats. These work histories and the estimated FAH
values were used to generate a cumulative exposure estimate (CEE) for
each case and control in the study. A long-term mean exposure estimate
(MEE) was generated by dividing each CEE by the individual's number of
years employed at Rocky Flats.

Viet et al.'s statistical analysis of the resulting dataset included
conditional logistic regression analysis, modeling the relationship
between risk of each health outcome and log-transformed CEE and MEE.
They found highly statistically significant relationships between
log-CEE and risk of CBD (coef = 0.837, p = 0.0006) and between log-MEE
(coef = 0.855, p = 0.0012) and risk of CBD, indicating that risk of CBD
increases with exposure level. These coefficients correspond to odds
ratios of 6.9 and 7.2 per 10-fold increase in exposure, respectively.
Risk of sensitization without CBD did not show a statistically
significant relationship with log-CEE (coef = 0.111, p = 0.32), but
showed a nearly-significant relationship with log-MEE (coef = 0.230, p =
0.097). 

Kelleher et al. – Cullman Facility

μm or less) at the worker breathing zone that made up greater than 50
percent of the beryllium mass. Kelleher et al. used the dataset of 100
personal lapel samples collected by Martyny et al. and other NJMRC
researchers in 1996, 1997, and 1999 to characterize exposures for each
job in the plant. Following a statistical analysis comparing the samples
collected by NJMRC with earlier samples collected at the plant, Kelleher
et al. concluded that the 1996-1999 data could be used to represent
job-specific exposures from earlier periods. 

Detailed work history information gathered from plant data and worker
interviews was used in combination with job exposure estimates to
characterize cumulative and long-term lifetime weighted (LTW) average
beryllium exposures for workers in the surveillance program. In addition
to cumulative and LTW exposure estimates based the total mass of
beryllium reported in their exposure samples, Kelleher et al. calculated
cumulative and LTW estimates based specifically on exposure to particles
< 6 µm and particles < 1 µm in diameter.

To analyze the relationship between exposure level and risk of
sensitization and CBD, Kelleher et al. performed a case-control analysis
using measures of both total beryllium exposure and particle
size-fractionated exposure. The analysis included sensitization cases
identified in the 1995-1999 surveillance and 206 controls from the group
of 215 non-sensitized workers. For nine workers, they could not
reconstruct complete job histories. Logistic regression models using
categorical exposure variables showed positive associations between risk
of sensitization and the six exposure measures tested: total CEE, total
MEE, and variations of CEE and MEE constructed based on particles < 6
µm and < 1 µm in diameter. None of the associations were statistically
significant (p < 0.05); however, the authors noted that the dataset was
relatively small, with limited power to detect a statistically
significant exposure-response relationship. 

Although the Viet et al. and Kelleher et al. exposure-response analyses
provide valuable insight into exposure-response for beryllium
sensitization and CBD, both studies have difficulties that limit their
suitability as a basis of quantitative risk assessment. Their
limitations primarily involve the exposure data used to estimate
workers' exposures. Viet et al.'s exposure reconstruction was based on
area samples from a single building within a large, multi-building
facility. Where possible, OSHA prefers to base risk estimates on
exposure data collected in the breathing zone of workers rather than
area samples, because data collected in the breathing zone more
accurately characterize workers' exposures. Kelleher's analysis, on the
other hand, was based on personal lapel samples. However, the samples
Kelleher et al. used were collected between 1996 and 1999, after the
facility had initiated new exposure control measures in response to the
diagnosis of a case of CBD in 1995. OSHA believes that industrial
hygiene samples collected at the Cullman plant prior to 1996 better
characterize exposures prior to the new controls. In addition, since the
publication of the Kelleher study, the population has continued to be
screened for sensitization and CBD. Data collected on workers hired in
2000 and later, after most exposure controls had been completed, can be
used to characterize risk at lower levels of exposure than have been
examined in many previous studies. 

To better characterize the relationship between exposure level and risk
of sensitization and CBD, OSHA elected to develop an independent
exposure-response analysis based on a dataset maintained by NJMRC on
workers at the Cullman, AL, machining plant. The dataset includes
exposure samples collected between 1980 and 2005, and has updated work
history and screening information for several hundred workers through
2003.

 OSHA’s Exposure-Response Analysis

OSHA evaluated exposure and health outcome data on a population of
workers employed at the Cullman machining facility. NJMRC researchers,
with consent and information provided by the facility, compiled a
dataset containing employee work histories, medical diagnoses, and air
sampling results and provided it to OSHA for analysis. OSHA’s
contractors gathered additional information from (1) two surveys of the
Cullman plant conducted by OSHA's contractor (ERG, 2003 and ERG, 2004a),
(2) published articles of investigations conducted at the plant by
researchers from NJMRC (Kelleher et al., 2001; Madl et al., 2007;
Martyny et al., 2000; and Newman et al., 2001), (3) a case file from a
1980 OSHA complaint inspection at the plant, (4) comments submitted to
the OSHA docket office in 1976 and 1977 by representatives of the metal
machining plant regarding their beryllium control program, and (5)
personal communications with the plant’s current industrial hygienist
(ERG, 2009b) and an industrial hygiene researcher at NJMRC (ERG, 2009a).


Plant Operations

The Cullman plant is a leading fabricator of precision machined and
processed materials including beryllium and its alloys, titanium,
aluminum, quartz, and glass (ERG, 2009b). The plant has approximately
210 machines, primarily mills and lathes, and processes large quantities
of beryllium on an annual basis. The plant provides complete fabrication
services including ultra precision machining; ancillary processing
(brazing, ion milling, photo etching, precision cleaning, heat treating,
stress relief, thermal cycling, mechanical assembly, and chemical
milling/etching); and coatings (plasma spray, anodizing, chromate
conversion coating, nickel sulfamate plate, nickel plate, gold plate,
black nickel plate, copper plate/strike, passivation, and painting).
Most of the plant's beryllium operations involve machining beryllium
metal and high beryllium content composite materials (beryllium
metal/beryllium oxide metal composites called E-Metal or E-Material),
with occasional machining of beryllium oxide/metal matrix (such as
AlBeMet, aluminum beryllium matrix) and beryllium-containing alloys.
E-Materials such as E-20 and E-60 are currently processed in the E-Cell
department. 

The 120,000 square-foot plant has two main work areas: a front office
area and a large, open production shop. Operations in the production
shop include inspection of materials, machining, polishing, and quality
assurance. The front office is physically separated from the production
shop. Office workers enter through the front of the facility and have
access to the production shop through a change room where they must don
laboratory coats and shoe covers to enter the production area.
Production workers enter the shop area at the rear of the facility where
a change/locker room is available to change into company uniforms and
work shoes. Support operations are located in separate areas adjacent to
the production shop and include management and administration, sales,
engineering, shipping and receiving, and maintenance. Management and
administrative personnel include two groups: those primarily working in
the front offices (front office management) and those primarily working
on the shop floor (shop management). 

evel of 0.2 μg/m3, installed change/locker rooms for workers entering
the production facility, eliminated pressurized air hoses, discouraged
the use of dry sweeping, initiated biennial medical surveillance using
the BLPT, and implemented annual beryllium hazard awareness training. 

In 1996, the company instituted requirements for work uniforms and
dedicated work shoes for production workers, eliminated dry sweeping in
all departments, and purchased high efficiency particulate air (HEPA)
filter vacuum cleaners for workplace cleanup and decontamination. Major
engineering changes were also initiated in 1996, including the purchase
of a new local exhaust ventilation (LEV) system to exhaust machining
operations producing finer aerosols (e.g., dust and fume versus metal
chips). The facility also began installing mist eliminators for each
machine. Departments affected by these changes included cutter grind
(tool and die), E-cell, EDM, flow lines, grind, lapping, and optics. Dry
machining operations producing chips were exhausted using the existing
LVHV exhaust system (ERG, 2004a). In the course of making the
ventilation system changes, old ductwork and baghouses were dismantled
and new ductwork and air cleaning devices were installed. The company
also installed Plexiglas enclosures on machining operations in
1996-1997, including the lapping, deburring, grinding, EDM, and tool and
die operations. In 1998, LEV was installed in EDM and modified in the
lap, deburr, and grind departments. 

Most exposure controls were reportedly in place by 2000 (ERG, 2009a). In
2004, the plant industrial hygienist reported that all machines had LEV
and about 65 percent were also enclosed with either partial or full
enclosures to control the escape of machining coolant (ERG, 2004b). Over
time, the facility has built enclosures for operations that consistently
produce exposures greater than 0.2 μg/m3. The company has never
required workers to use gloves or PPE.

Air Sampling Data

The NJMRC dataset includes industrial hygiene sampling results collected
by the plant (1980-1984 and 1995-2005) and NJMRC researchers (June 1996
to February 1997 and September 1999), including 4,370 breathing zone
(personal lapel) samples and 712 area samples (ERG, 2004b). Limited air
sampling data is available before 1980 and no exposure data appears to
be available for the 10-year time period 1985 through 1994. A review of
the NJMRC air sampling database from 1995 through 2005 shows a
significant increase in the number of air samples collected beginning in
2000, which the plant industrial hygienist attributes to an increase in
the number of air sampling pumps (from 5 to 23) and the purchase of an
automated atomic absorption spectrophotometer. 

ERG used the personal breathing zone sampling results contained in the
sample database to calculate exposure statistics for each year and for
several-year periods, disaggregated by process name. In each case, ERG
calculated the arithmetic mean, geometric mean, median, minimum,
maximum, and 95 percentile value. Prior to generating these statistics
ERG made several adjustments. After consultation with researchers at
NJMRC, four particularly high exposures were identified as probably
erroneous and excluded from calculated statistics. In addition, a 1996
sample for the HS (Health and Safety) process was removed from the
sample calculations after ERG determined it was for a non-employee
researcher visiting the facility. 

Most samples in the sample database for which sample times were recorded
were long-term samples. For example, 2,503 of the 2,557 (97.9 percent)
breathing zone samples with sample time recorded had times greater than
or equal to 400 minutes. No adjustments were made for sampling time,
except in the case of four samples for the “maintenance” process for
1995. These results show relatively high values and exceptionally short
sample times consistent with the nature of much maintenance work, marked
by short-term exposures and periods of no exposure. The four 1995
maintenance samples were adjusted for an eight-hour sampling time
assuming that the maintenance workers received no further beryllium
exposure over the rest of their work shift. 

After reviewing the sample statistics for individual years, exposure
monitoring data for primary machining jobs were grouped into four time
periods that correspond to stages in the plant’s approach to beryllium
exposure control. These time periods were used to construct the exposure
history of workers contained in the work history database and are
outlined below. 

1980-1995, a period of relatively minimal control prior to the 1995
discovery of a case of CBD among the plant's workers; 

1996-1997, a period during which some major engineering controls were in
the process of being installed on machining equipment;

1998-1999, a period during which most engineering controls on the
machining equipment had been installed;

2000-2003, a period when installation of all exposure controls on
machining equipment was complete and exposures very low throughout the
plant.

Exposure results from 1996-1997 were not found to be consistently
reduced in comparison to the 1985-1995 period for primary machining
jobs. Therefore, these two periods were grouped together in the job
exposure matrix. Exposure monitoring for jobs other than the primary
machining operations were not further subdivided but represented by a
single mean exposure value for 1980-2003. The risk assessment background
document describes construction of the job exposure matrix in more
detail. 

 , where e(i) is the exposure level for job (i), and t(i) is the time
spent in job (i). Cumulative exposure was divided by total exposure time
to estimate each worker’s long-term average exposure. For workers with
beryllium sensitization or CBD, exposure estimates excluded exposures
following diagnosis.

Workers who were employed for long time periods in jobs with low-level
exposures tend to have low average and cumulative exposures due to the
way these measures are constructed, incorporating the worker’s entire
work history. However, higher-level exposures such as those encountered
in machining jobs may be more relevant to risk of sensitization than
exposures experienced in administrative or other low-exposure work. To
explore this possibility, ERG constructed a third type of exposure
estimate. This measure reflects the exposure level associated with the
highest-exposure job (HEJ) and time period experienced by each worker.
This exposure estimate (HEJ), the cumulative exposure estimate, and the
average exposure were used in the quartile analysis and statistical
analyses.

Prevalence of Sensitization and CBD

In the database provided to OSHA, seven workers were reported as
sensitized only. Sixteen workers were listed as sensitized and diagnosed
with CBD upon initial clinical evaluation. Three workers, first shown to
be sensitized only, were later diagnosed with CBD. Tables 1 and 2 below
present the prevalence of sensitization and CBD cases across several
categories of lifetime-weighted (LTW) average and cumulative exposure.
Exposure values were grouped by quartile. Note that all workers with CBD
are also sensitized. Thus, the columns “Total” and “Total %”
refer to all sensitized workers in the dataset, including workers with
and without a diagnosis of CBD.

Table 1: Prevalence of Sensitization and CBD by LTW Average Exposure
Quartile

   Average Exposure  (μg/m3)	Group Size	Sensitized only	CBD	Total	Total
%	CBD %

0.0 – 0.080	91	1	1	2	2.2%	1.0%

0.081 - 0.18	73	2	4	6	8.2%	5.5%

0.19 - 0.51	77	0	6	6	7.8%	7.8%

0.51 - 2.15	78	4	8	12	15.4%	10.3%

Total	319	7	19	26	8.2%	6.0%



Table 2: Prevalence of Sensitization and CBD by Cumulative Exposure
Quartile

Cumulative Expo sure (μg/m3-yrs)	Group Size	Sensitized only	CBD	Total
Total %	CBD %

0.0 – 0.147	81	2	2	4	4.9%	2.5%

0.148 – 1.467	79	0	2	2	2.5%	2.5%

1.468 – 7.008	79	3	8	11	13.9%	8.0%

7.009 – 61.86	80	2	7	9	11.3%	8.8%

Total	319	7	19	26	8.2%	6.0%



Table 3: Prevalence of Sensitization and CBD by Highest-Exposed Job
Exposure Quartile

μg/m3)	Group Size	 Sensitized only	CBD	Total	Total %	CBD %

0.0 - 0.086	86	1	0	1	1.2%	0.0%

0.091 - 0.214	81	1	6	7	8.6%	7.4%

0.387 - 0.691	76	2	9	11	14.5%	11.8%

0.954 - 2.213	76	3	4	7	9.2%	5.3%

Total	319	7	19	26	8.2%	6.0%

Table 1 shows increasing prevalence of total sensitization and CBD with
increasing LTW average exposure, measured both as average and cumulative
exposure. The lowest prevalence of sensitization and CBD was observed
among workers with average exposure levels less than or equal to 0.08
μg/m3, where two sensitized workers (2.2 percent) including one case of
CBD (1.0 percent) were found. The sensitized worker in this category
without CBD had worked at the facility as an inspector since 1972, one
of the lowest-exposed jobs at the plant. Because the job was believed to
have very low exposures, it was not sampled prior to 1998. Thus,
estimates of exposures in this job are based on data from 1998-2003
only. It is possible that exposures earlier in this worker’s
employment history were somewhat higher than reflected in his estimated
average exposure. The worker diagnosed with CBD in this group had been
hired in 1996 in production control, and had an estimated average
exposure of 0.08 μg/m3. He was diagnosed with CBD in 1997.

The second quartile of LTW average exposure (0.81 - 0.18 μg/m3) shows a
marked rise in overall prevalence of beryllium-related health effects,
with six workers sensitized (8.2 percent), of whom four (5.5 percent)
were diagnosed with CBD. Among six sensitized workers in the third
quartile (0.19 - 0.50 μg/m3), all were diagnosed with CBD (7.8
percent). Another increase in prevalence is seen from the third to the
fourth quartile, with 12 cases of sensitization (15.4 percent),
including eight (10.3 percent) diagnosed with CBD.

μg/m3-yrs, 0.148 - 1.467 μg/m3-yrs). The upper bound on this
cumulative exposure range, 1.467 μg/m3-yrs, is the cumulative exposure
that a worker would have if exposed to beryllium at a level of 0.03
μg/m3 for a working lifetime of 45 years; 0.15 μg/m3 for ten years; or
0.3 μg/m3 for five years.

A sharp increase in prevalence of sensitization and CBD and total
sensitization occurs in the third quartile (1.468 - 7.008 μg/m3-yrs),
with roughly similar levels of both in the highest group (7.009 - 61.86
μg/m3-yrs). Cumulative exposures in the third quartile would be
experienced by a worker exposed for 45 years to levels between 0.03 and
0.16 μg/m3, for 10 years to levels between 0.15 and 0.7 μg/m3, or for
five years to levels between 0.3 and 1.4 μg/m3. 

 When workers’ exposures from their highest-exposed job are
considered, the exposure-response pattern is similar to that for LTW
average exposure in the lower quartiles (Table 3). The lowest prevalence
is observed in the first quartile (0.0 - 0.86 ug/m3), with sharply
rising prevalence from first to second and second to third exposure
quartiles. The prevalence of sensitization and CBD in the top quartile
(0.954 - 2.213 ug/m3) decreases relative to the third, with levels
similar to the overall prevalence in the dataset. 

Statistical Modeling

μg/m3 on average), the level of detail presented in the published
studies limits the Agency’s ability to characterize risk at all the
alternate PELs OSHA is considering. To better understand these risks,
OSHA used the dataset described in the previous section to characterize
risk of sensitization and CBD among workers exposed to each of the
alternate PELs under consideration for the revised beryllium rule. 

OSHA's contractor used statistical modeling to characterize
exposure-response in this dataset. Asymptotic logistic regression models
provided an initial look at the dataset. Logistic models are often used
by researchers to analyze prevalence data in epidemiological studies.
For example, Viet et al. and Kelleher et al. both employed logistic
regression models in their exposure – response assessments (see
earlier section). Logistic modeling was followed up with a more rigorous
statistical analysis using a proportional hazards model to develop
quantitative risk estimates for sensitization and CBD at each of the
alternate PELs under consideration. 

Logistic Regression Model

The preliminary logistic regression models were fit to two health
endpoints: (1) CBD and (2) sensitization, including sensitized
individuals with and without CBD. Since only seven workers were reported
as sensitized without CBD, this health outcome was not included in the
statistical modeling. Predictors tested in the models included
cumulative and time-weighted average exposure measures, as well as total
duration of exposure. Models were fit with both continuous and
categorical exposure measures, where seven categories were used for
cumulative and average exposures, and four for years of exposure. The
results of continuous exposure models are reported here. Results of the
categorical analyses were similar and are available in the companion
background document to this section. 

The results of the simple logistic regression indicate a significant
positive association between the odds of sensitization and cumulative
exposure to beryllium, total years of exposure, and the average exposure
weighted by years spent at each job, as shown in Table 4 below. 

 

Table 4. Summary of Logistic Regression Results - Sensitization Endpoint

Exposure Measure	Coefficient	P-value	95% confidence interval

Cumulative Exposure (μg/m3-years)	0.033	0.031	0.003 - 0.625

Average Exposure (μg/m3)	0.917	0.014	0.187 - 1.646

HEJ Exposure (μg/m3)	0.481	0.085	- 0.658 - 1.028

Exposure Duration (years)	0.041	0.034	0.003 - 0.785



As shown in Table 5 below, the odds of CBD also increased with total
years of exposure (p = 0.020), and cumulative beryllium exposure (p =
0.043). Unlike beryllium sensitization, the odds of CBD was positively,
but non-significantly, associated with average exposure (p = 0.117).
This is not surprising since CBD is believed to respond to the
accumulated burden of beryllium retained in the lung over an extended
period of time (Newman et al., 1996). Lung burden would be expected to
primarily track cumulative exposure and years exposed and, to a lesser
extent, long-term average exposure.

Table 5. Summary Logistic Regression Results - CBD Endpoint

Exposure Measure	Coefficient	P-value	95% confidence interval

Cumulative Exposure (μg/m3-years)	0.034	0.043	0.001 - 0.067

Average Exposure (μg/m3)	0.694	0.117	-0.174 - 1.562

HEJ Exposure (μg/m3)	0.293	0.389	-0.374 - 0.960

Exposure Duration (years)	0.052	0.020	0.008 - 0.096



OSHA’s contractor performed the logistic regressions as a simple first
analysis of the data. Although relatively simple to interpret, standard
logistic regression has substantial limitations in the present setting
and is not a fully appropriate approach for analysis of this data set.
Logistic regression models the probability of an event in a fixed period
of time. As such, logistic regression cannot account for essential
features of these data, specifically, differential time at risk and
changing exposures over time. OSHA therefore did not use the results of
the logistic regression to create quantitative risk estimates for
beryllium sensitization and CBD. OSHA’s contractor developed a
proportional hazards model, a more appropriate model form for this data
set, to serve as the basis for the Agency’s quantitative risk
estimates.

Proportional Hazards Model

OSHA's contractor performed a complementary log-log proportional hazards
model. The proportional hazards model is an extension of logistic
regression that allows for time-dependent exposures and differential
time at risk. The proportional hazards model accounts for the fact that
individuals in the dataset are followed for different amounts of time,
and that their exposures change over time. In addition, where logistic
regression yields odds ratios, the proportional hazards model provides
hazards ratios, which estimate the relative risk of disease at a
specified time for someone with exposure level 1 compared to exposure
level 2. To perform this analysis, OSHA's contractor constructed
exposure files with cumulative and average exposures for each worker in
the data set in each year that a case of sensitization or CBD was
identified. Workers were included in only those years after they started
working at the plant and continued to be followed. Sensitized cases were
not included in analysis of sensitization after the year in which they
were identified as being sensitized, and CBD cases were not included in
analyses of CBD after the year in which they were diagnosed with CBD.

The results of the discrete proportional hazards analyses are summarized
in Tables 6-9 below. All coefficients used in the models are displayed,
including the exposure coefficient, the model constant for diagnosis in
1995 and additional exposure-independent coefficients for each
succeeding year (1996-1999) of diagnosis fit in the discrete time
proportional hazards modeling procedure. Model equations and variables
are explained more fully in the companion risk assessment background
document. 

Relative risk of sensitization increased with cumulative exposure (p =
0.05). A positive, but not statistically significant association was
observed with LTW average exposure (p = 0.09). The association was much
weaker for exposure duration (p = 0.31), consistent with the expected
biological action of a immune hypersensitivity response where onset is
believed to be more dependent on the concentration of the sensitizing
agent at the target site rather than the number of years of occupational
exposure. The association was also much weaker for highest-exposed job
(HEJ) exposure (p = 0.3).

Table 6.  Proportional Hazards Model: Cumulative Exposure and
Sensitization

Variable	Coefficient	95% confidence interval	P-value

Cumulative Exposure (μg/m3 -yrs)	0.031	0.00 to 0.063	0.05

constant	-3.48	-4.27 to -2.69	0.00

1996	-1.49	-3.04 to 0.06	0.06

1997	-0.29	-1.31 to 0.72	0.57

1998	-1.59	-3.15 to -0.02	0.05

1999	-1.60	-3.16 to -0.04	0.05



Table 7.  Proportional Hazards Model: LTW Average Exposure and
Sensitization

Variable	Coefficient	95% confidence interval	P-value

Average Exposure (μg/m3)	0.54	-0.09 to 1.17	0.09

constant	-3.55	-4.42 to -2.69	<0.001

1996	-1.48	-3.03 to 0.07	0.06

1997	-0.29	-1.31 to 0.72	0.57

1998	-1.54	-3.09 to 0.01	0.052

1999	-1.53	-3.08 to 0.03	0.054



Table 8.  Proportional Hazards Model: Exposure Duration and
Sensitization

Variable	Coefficient	95% confidence interval	P-value

Exposure Duration (years)	0.03	-0.03 to 0.08	0.31

constant	-3.55	-4.57 to -2.53	0.00

1996	-1.48	-3.03 to 0.70	0.06

1997	-0.30	-1.31 to 0.72	0.57

1998	-1.59	-3.14 to -0.04	0.05

1999	-1.62	-3.17 to -0.72	0.04



Table 9.  Proportional Hazards Model: HEJ Exposure and Sensitization

Variable	Coefficient	95% confidence interval	P-value

HEJ Exposure (μg/m3)	0.31	-0.27 to 0.88	0.30

constant	-3.42	-4.27 to -2.56	<0.001

1996	-1.49	-3.04 to 0.06	0.06

1997	-0.31	-1.32 to 0.70	0.55

1998	-1.59	-3.14 to -0.04	0.04

1999	-1.60	-3.15 to -0.05	0.04

  

The proportional hazards models for the CBD endpoint (Tables 10-13
below) showed positive relationships with cumulative exposure (p = 0.09)
and duration of exposure (p = 0.04). However, the association with the
cumulative exposure metric was not as strong as that for sensitization,
probably due to the smaller number of CBD cases. As found with the
simple logistic regression model, LTW average exposure and HEJ exposure
were not closely related to relative risk of CBD (p-values > 0.5). 

Table 10.  Proportional Hazards Model:  Cumulative Exposure and CBD

Variable	Coefficient	95% confidence interval	P-value

Cumulative Exposure (μg/m3 -yrs)	0.03	-.01 to 0.07	0.09

constant	-3.77	-4.67 to -2.86	<0.001

1997	-0.59	-1.86 to 0.68	0.36

1998	-2.01	-4.13 to 0.11	0.06

1999	-0.63	-1.90 to 0.64	0.33

2002	-2.13	-4.25 to -0.01	0.049



Table 11.  Proportional Hazards Model:  LTW Average Exposure and CBD

Variable	Coefficient	95% confidence interval	P-value

Average Exposure (μg/m3)	0.24	-0.59 to 1.06	0.58

constant	-3.62	-4.60 to -2.64	<0.001

1997	-0.61	-1.87 to 0.66	0.35

1998	-2.02	-4.14 to 0.10	0.06

1999	-0.64	-1.92 to 0.63	0.32

2002	-2.15	-4.60 to -2.64	0.047



Table 12.  Proportional Hazards Model:  Exposure Duration and CBD

Variable	Coefficient	95% confidence interval	P-value

Exposure Duration (yrs)	0.05	-0.01 to 0.11	0.04

constant	-4.18	-1.84 to 0.69	< 0.001

1997	-0.53	-4.13 to 0.11	0.38

1998	-2.01	-1.94 to 0.60	0.06

1999	-0.67	-4.34 to -0.10	0.30

2002	-2.22	-5.40 to -2.96 	0.04



Table 13.  Proportional Hazards Model:  HEJ Exposure and CBD

Variable	Coefficient	95% confidence interval	P-value

HEJ Exposure (μg/m3)	0.03	-0.70 to 0.77	0.93

constant	-3.49	-4.45 to -2.53	<0.001

1997	-0.62	-1.88 to 0.65	0.34

1998	-2.05	-4.16 to 0.07	0.06

1999	-0.68	-1.94 to 0.59	0.30

2002	-2.21	-4.33 to -0.09	0.04



In addition to the models reported above, comparable models were fit to
the upper 95 percent confidence interval of the HEJ exposure;
log-transformed cumulative exposure; log-transformed average exposure;
and log-transformed HEJ exposure.  Each of these measures was positively
but not significantly associated with sensitization.

 to 2 μg/m3 for 10 years, the models predicts 56 workers would be
sensitized over the time interval from 1985 to 1995 and 12 workers over
the time interval from 1989 to 1999. The other model projections for
this exposure scenario fall between these estimates. OSHA considers the
range of estimates to be a reasonable model descriptor of sensitization
risk per 1000 workers exposed to 2 μg/m3 over a 10-year period.
Model-predicted cases of sensitization per 1000 workers for the current
PEL and four alternate PELs over exposure durations from five to 45
years are shown below. The corresponding risks for CBD are presented in
Table 15.   

μg/m3)	Cumulative (μg/m3-yrs)	cases/1000	μg/m3-yrs	cases/1000
μg/m3-yrs	cases/1000	μg/m3-yrs	cases/1000

2.0	10.0	8.6 - 41.3	20.0	11.7 - 56.0	40.0	21.8 - 102.3	90.0	100.0 -
403.5

1.0	5.0	7.3 - 35.4	10.0	8.6 - 41.3	20.0	11.7 - 56.0	45.0	25.4 - 118.5

0.5	2.5	6.8 - 32.8	5.0	7.3 - 35.4	10.0	8.6 - 41.3	22.5	12.6 - 60.4

0.2	1.0	6.5 - 31.3	2.0	6.7 - 32.3	4.0	7.1 - 34.3	9.0	8.3 - 40.0

0.1	0.5	6.4 - 30.8	1.0	6.5 - 31.3	2.0	6.7 - 32.3	4.5	7.2 - 34.9



μg/m3)	Cumulative (μg/m3-yrs)	cases/1000	μg/m3-yrs	cases/1000
μg/m3-yrs	cases/1000	μg/m3-yrs	cases/1000

2.0	10.0	3.7 - 30.6	20.0	5.0 - 45.8	40.0	9.1 - 73.7	90.0	39.9 - 290.4

1.0	5.0	3.2 - 26.4	10.0	3.7 - 34.1	20.0	5.0 - 41.1	45.0	10.5 - 85.1

0.5	2.5	3.0 - 24.5	5.0	3.2 - 29.5	10.0	3.7 - 30.6	22.5	5.4 - 44.3

0.2	1.0	2.8 - 23.5	2.0	2.9 - 27.0	4.0	3.1 - 25.7	9.0	3.6 - 29.7

0.1	0.5	2.8 - 23.1	1.0	2.8 - 26.2	2.0	2.9 - 24.2	4.5	3.1 - 26.0



μg/m3 for an exposure duration of 45 years (90 μg/m3-yr). The
predicted risks are more than ten-fold less for a 45 year exposure at
the lowest alternate PEL, 0.1 μg/m3 (4.5 μg/m3-yr). On the other hand,
the model predicts relatively little change in risk (<35 percent) over
the cumulative exposure range from 0.5 (e.g., 0.1 μg/m3 for five years)
to 10 μg/m3-yr (e.g., 2.0 μg/m3 for five years). The model-predicted
risks at these exposures are 0.6 to 4.1 percent for sensitization and
0.3 to 3.1 percent for CBD. While these risks are still of an
appreciable magnitude, they tend to be lower than the 2.5 to 14 percent
prevalence observed in the Cullman data set over similar cumulative
exposures (Table 2). However, the higher model estimates at 1.0 – 2.0
μg/m3-years (e.g. 2.3 to 3.2 percent) appear to be more congruent with
the prevalence of sensitization and CBD (2.5 percent) found among
workers in the second quartile of cumulative exposure (0.148 – 1.467
μg/m3-years). While the reason for the discrepancies between the
modeling results and the prevalence analyses is not entirely clear, the
model estimates assume a group without turnover, so they are not
directly comparable to prevalence values discussed in previous sections

08 μg/m3-years at the end of their follow-up, were sensitized. However,
the model results predict sensitization in only 1.2 - 5.6 percent of
workers with cumulative exposure of 20 μg/m3-years, more than twice the
upper limit of that quartile. As discussed in the background document
for this analysis, most workers in the data set had low cumulative
exposures (roughly half below 1.5 μg/m3-years). Model estimates for
cumulative exposures greater than 40 μg/m3-years are well beyond the
observed range for the vast majority of Cullman exposures, and,
therefore, should be regarded with a greater degree of uncertainty. 

Due to limitations including size of the dataset, relatively limited
exposure data from the plant’s early years, study size-related
constraints on the statistical analysis of the dataset, and limited
follow-up time on many workers, OSHA must interpret the risk estimates
presented in Tables 14 and 15 with caution. The Cullman study population
is a relatively small group and can support only limited statistical
analysis. For example, its size precludes inclusion of multiple
covariates in the exposure-response models or a two-stage
exposure-response analysis to model both sensitization and the
subsequent development of CBD within the subpopulation of sensitized
workers. The limited size of the Cullman dataset is characteristic of
studies on beryllium-exposed workers in modern, low-exposure
environments, which are typically small-scale processing plants (up to
several hundred workers, up to 20-30 cases). However, these recent
studies also have important strengths: they include workers hired after
the institution of stringent exposure controls, and have extensive
exposure sampling using full-shift personal lapel samples. In contrast,
older studies of larger populations tend to have higher exposures, less
exposure data, and exposure data collected in short-term samples or
outside of workers’ breathing zones.   

Another limitation of the Cullman dataset, which is common to recent
low-exposure studies, is the short follow-up time available for many of
the workers. While in some cases CBD has been known to develop in short
periods (< 2 years), it more typically develops over a longer time
period. Sensitization occurs in a typically shorter time frame, but new
cases of sensitization have been observed in workers exposed to
beryllium for many years. OSHA expects that the dataset does not fully
represent the risk of sensitization and particularly CBD among workers
at this facility. The Agency believes the short follow-up time to be a
significant source of uncertainty in the statistical analysis, a factor
likely to lead to underestimation of risk in this population.

Particle size, particle surface area, and beryllium compound solubility
are believed to be important factors influencing the risk of
sensitization and CBD among beryllium-exposed workers. The workers at
the Cullman machining plant were primarily handling insoluble beryllium
compounds, such as beryllium metal and beryllium metal/beryllium oxide
composites. Particle size distributions from a limited number of
airborne beryllium samples collected just after the 1996 installation of
engineering controls indicate worker exposure to a substantial
proportion of respirable particulates. There was no available particle
size data for the 1980 to 1995 period prior to installation of
engineering controls when total beryllium mass exposure levels were
greatest. Particle size data was also lacking from 1998 to 2003 when
additional control measures were in place and total beryllium mass
exposures were lowest. For these reasons, OSHA was not able to
quantitatively account for the influence of particle size and solubility
in developing the risk estimates based on the Cullman data set. However,
it is not unreasonable to expect the CBD experienced by this cohort to
generally reflect the risk from exposure to beryllium that is relatively
insoluble and enriched with respirable particles. As explained
previously, the role of particle size and surface area on risk of
sensitization is more difficult to predict.  

to alter its qualitative conclusions with regard to the risk at the
current PEL and at alternate PELs as low as 0.1 μg/m3. The existing
studies provide clear evidence of sensitization and CBD risk among
workers exposed to a number of beryllium forms as a result of different
processes such as beryllium machining, beryllium-copper alloy
production, and beryllium ceramics production. The Agency believes all
of these forms of beryllium exposure contribute to the overall risk of
sensitization and CBD among beryllium-exposed workers.

Preliminary Conclusions

As described above, OSHA’s risk assessment for beryllium sensitization
and CBD relied on two approaches: (1) review of the literature and (2)
analysis of a dataset provided by NJRMC. First, the Agency reviewed the
scientific literature, to ascertain whether there is substantial risk to
workers exposed at and below the current PEL and to characterize the
expected impact of more stringent controls on workers’ risk of
sensitization and CBD. This review focused on facilities where exposures
were primarily below the current PEL, and where several rounds of BeLPT
and CBD screening had been conducted to evaluate the effectiveness of
various exposure control measures. Second, OSHA investigated the
exposure-response relationship for beryllium sensitization and CBD by
analyzing a dataset that NJMRC provided on workers at a prominent,
long-established beryllium machining facility. Although
exposure-response studies have been published on sensitization and CBD,
OSHA believes the nature and quality of their exposure data
significantly limits their value for the Agency’s risk assessment.
Therefore, OSHA developed an independent exposure-response analysis
using the NJMRC dataset, which was recently updated, includes workers
exposed at low levels, and includes extensive exposure data collected in
workers’ breathing zones, as is preferred by OSHA. 

x

{

É

x

D

	”ÿì

	”ÿì

	”ÿì

	¿

	¿

kd+

	¿

愀Ȥ摧噈Ý欀

D

F

p

r

t

z

ˆ

Š

¨

¬

®

´

Â

Ä

Ò

Ü

Þ

V

È

ò

þ

愀Ȥ摧噈Ý欀।

-

"

&

,

8

D

b

h

l

p

t

~

ˆ

–

ž

¢

¨

®

¸

Â

HÄ

  h

”ÿ^

	Ê

húk

	@

	V

	@

	V

	@

	V

	@

	V

	@

	V

	@

	V

h+

ÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿ

愀Ĥ摧rᘀ

hay

h

hr

u

Mrams that primarily used engineering controls to reduce airborne
exposures to median levels at or around 0.2 μg/m3 had only limited
impact on workers’ risk. Cases of sensitization continued to occur
frequently among newly hired workers, and some of these workers
developed CBD within the short follow-up time. 

In contrast, industrial hygiene programs that minimized both airborne
and dermal exposure substantially lowered workers' risk of sensitization
in the first years of employment. Programs that drastically reduced
respiratory exposure via a combination of engineering controls and
respiratory protection, minimized the potential for skin exposure via
dermal PPE, and employed stringent housekeeping methods to keep work
areas clean and prevent transfer of beryllium between areas, sharply
curtailed new cases of sensitization among newly-hired workers. For
example, studies conducted at copper-beryllium and beryllia ceramics
facilities show that reduction of exposures to below 0.1 μg/m3 and
protection from dermal exposure, in combination, achieved a substantial
reduction in sensitization risk among newly-hired workers. However, even
these stringent measures did not protect all workers from sensitization.
   

The most recent epidemiological literature on programs that have been
successful in reducing workers' risk of sensitization have had very
short follow-up time; therefore, they cannot address the question of how
frequently workers sensitized in very low-exposure environments develop
CBD. Clinical evaluation for CBD was not reported for workers at the
copper-beryllium and ceramics facilities. However, cases of CBD among
workers exposed at low levels at a machining plant and cases of CA-CBD
demonstrate that individuals exposed to low levels of airborne beryllium
can develop CBD, and over time, can progress to severe disease. This
conclusion is also supported by case reports within the literature of
workers with CBD who may have been minimally exposed to beryllium, such
as worker(s) employed only in administration at a beryllium ceramics
facility (Kreiss et al., 1996).

The Agency’s analysis of the Cullman dataset provided by NJMRC showed
strong exposure-response trends using multiple analytical approaches,
including examination of sensitization and disease prevalence by
exposure categories and a proportional hazards modeling approach. In the
prevalence analysis, cases of sensitization and disease were evident at
all levels of exposure. The lowest prevalence of sensitization (2.0
percent) and CBD (1.0 percent) was observed among workers with LTW
average exposure levels below 0.1 μg/m3, while those with LTW average
exposure between 0.1 - 0.2 μg/m3 showed a marked increase in overall
prevalence of sensitization (9.8 percent) and CBD (7.3 percent).
Prevalence of sensitization and CBD also increased with cumulative
exposure. 

μg/m3 and each of the alternate PELs under consideration: 1 μg/m3, 0.5
μg/m3, 0.2 μg/m3, and 0.1 μg/m3. To estimate risk of CBD from a
working lifetime of exposure, the Agency calculated the cumulative
exposure associated with 45 years of exposure at each level, for total
cumulative exposures of 90, 45, 22.5, 9, and 4.5 μg/m3-years. The risk
estimates for sensitization and CBD ranged from 100 - 403 and 40 - 290
cases, respectively, per 1000 workers exposed at the current PEL of 2
μg/m3. The risks are projected to be substantially lower for both
sensitization and CBD at 0.1 μg/m3 and range from 7.2 - 35 cases per
1000 and 3.1 - 26 cases per 1000, respectively.  In these ways, the
modeling results are similar to results observed from published studies
of the Reading, Tucson, and Cullman plants and the OSHA analysis of
sensitization and CBD prevalence within the Cullman plant.

μg/m3, tightly controlled both respiratory and dermal exposure, and
incorporated stringent housekeeping measures have substantially reduced
risk of sensitization within the first years of exposure. These
conclusions are supported by the results of several studies conducted in
state-of-the-art facilities dealing with a variety of production
activities and physical forms of beryllium.  In addition, these
conclusions are supported by OSHA’s statistical analysis of a dataset
with highly detailed exposure and work history information on several
hundred beryllium workers.  While there is uncertainty regarding the
precision of model-derived risk estimates, they provide further evidence
that there is substantial risk of sensitization and CBD associated with
exposure at the current PEL, and that this risk can be substantially
lessened by stringent measures to reduce workers’ beryllium exposure
levels.  

References

Bailey DO, Thomas CA, Deubner DC, Kent MS, Kreiss K, Schuler CR.  (2010)
Evaluation of a Preventative Program to Reduce Sensitization at a
Beryllium Metal, Oxide, and Alloy Production Plant.  JOEM 52: 505-12  

Dattoli JA, Lieben J, Bisbing J. (1964) Chronic Beryllium Disease. A
Follow-Up Study. J Occup Med. 6:189-94.

Dutra FR. (1948)  The pneumonitis and granulomatosis peculiar to
beryllium workers. Am J Pathol. 24(6):1137-65.

Eisenbud, M (1998)  The Standard for Control of Chronic Beryllium
Disease.  Appl Occup Environ Hyg 13(1): 25-31.

Eisenbud M, Lisson J. (1983)  Epidemiological aspects of
beryllium-induced

nonmalignant lung disease: a 30-year update.  J Occup Med.
25(3):196-202.

Eisenbud M. (1982)  Origins of the standards for control of beryllium
disease

(1947-1949). Environ Res. 27(1):79-88. 

Eisenbud M, Wanta RC, Dustan, C, Steadman, LT, Harris, WB, Wolf, BS
(1949)  Non-occupational Berylliosis.  J Ind Hyg Toxicol. 31(5):282-94.

 

Henneberger PK, Cumro D, Deubner DD, Kent MS, McCawley M, Kreiss K.
(2001)    HYPERLINK
"http://www.ncbi.nlm.nih.gov/pubmed/11355290?itool=EntrezSystem2.PEntrez
.Pubmed.Pubmed_ResultsPanel.Pubmed_RVDocSum&ordinalpos=3"  Beryllium
Sensitization and Disease Among Long-Term and Short-Term Workers in a
Beryllium Ceramics Plant.   Int Arch Occup Environ Health 74(3):167-76.

Lieben J, Williams RR. (1969)  Respiratory Disease Associated With
Beryllium Refining And Alloy Fabrication.  1968 Follow-up.  J Occup Med.
11(9):480-5.

Kelleher PC, Martyny JW, Mroz MM, Maier LA, Ruttenber AJ, Young DA,
Newman LS. (2001)    HYPERLINK
"http://www.ncbi.nlm.nih.gov/pubmed/11285872?itool=EntrezSystem2.PEntrez
.Pubmed.Pubmed_ResultsPanel.Pubmed_RVDocSum&ordinalpos=1"  Beryllium
Particulate Exposure and Disease Relations in a Beryllium Machining
Plant.   J Occup Environ Med. 43(3):238-49.

Kreiss K, Mroz MM, Newman LS, Martyny J, Zhen B. (1996)    HYPERLINK
"http://www.ncbi.nlm.nih.gov/pubmed/8837677?itool=EntrezSystem2.PEntrez.
Pubmed.Pubmed_ResultsPanel.Pubmed_RVDocSum&ordinalpos=3"  Machining Risk
of Beryllium Disease and Sensitization With Median Exposures Below 2
micrograms/m3.   Am J Ind Med. 30(1):16-25.

Lieben J, Metzner F. (1959)  Epidemiological Findings Associated With
Beryllium

Extraction. Am Ind Hyg Assoc J. 20:494-9.

Madl AK, Unice K, Brown JL, Kolanz ME, Kent MS. (2007) 
Exposure-Response Analysis For Beryllium Sensitization and Chronic
Beryllium Disease Among Workers in a Beryllium Metal Machining Plant. J
Occup Environ Hyg. 4(6):448-66.

Maier LA, Martyny JW, Liang J, Rossman MD. (2008)  Recent Chronic
Beryllium Disease in Residents Surrounding a Beryllium Facility. Am J
Respir Crit Care Med. 1;177(9):1012-7. 

Martyny JW, Hoover MD, Mroz MM, Ellis K, Maier LA, Sheff KL, Newman LS.
(2000)

Aerosols Generated During Beryllium Machining.  J Occup Environ Med.
42(1):8-18.

Metzner FN, Lieben J. (1961)  Respiratory Disease Associated With
Beryllium Refining and Alloy Fabrication; A Case Study.  J Occup Med.
3:341-5.

Newman LS, Mroz MM, Balkissoon R, Maier LA. (2001)    HYPERLINK
"http://www.ncbi.nlm.nih.gov/pubmed/11285871?itool=EntrezSystem2.PEntrez
.Pubmed.Pubmed_ResultsPanel.Pubmed_RVDocSum&ordinalpos=3"  Efficacy of
Serial Medical Surveillance For Chronic Beryllium Disease in a Beryllium
Machining Plant.   J Occup Environ Med. 43(3):231-7.

Newman LS, Llyody J, Daniloff E.  (1996)  The natural history of
beryllium sensitization and chronic beryllium disease.  Environ Health
Perspect. Oct;104 Suppl 5:937-43

Sanderson WT, Ward, EM, Steenland, K, Petersen MR. (2001a)  Lung Cancer
Case Control Study of Beryllium Workers.  Am J Ind Med. 39(2):133-44.

Sanderson WT, Petersen MR, Ward EM. (2001b)  Estimating Historical
Exposures Of Workers In A Beryllium Manufacturing Plant.  Am J Ind Med.
39(2):145-57.

Schubauer-Berigan MK, Deddens, JA, Steenland, K, Sanderson, WT, Peterson
MR. (2008)  Adjustments for Temporal Confounders in a Reanalysis of a
Case Control Study of Beryllium and Lung Cancer.  Occup Environ Med
65(6): 379-83 

Schuler CR, Kent MS, Deubner DC, Berakis MT, McCawley M, Henneberger PK,
Rossman MD, Kreiss K. (2005)    HYPERLINK
"http://www.ncbi.nlm.nih.gov/pubmed/15712254?itool=EntrezSystem2.PEntrez
.Pubmed.Pubmed_ResultsPanel.Pubmed_RVDocSum&ordinalpos=2" 
Process-Related Risk of Beryllium Sensitization and Disease in a
Copper-Beryllium Alloy Facility.   Am J Ind Med. 47(3):195-205.

  h

# h

h5

 h

Stanton ML, Henneberger PK, Kent MS, Deubner DC, Kreiss K, Schuler CR.
(2006)

Sensitization and Chronic Beryllium Disease Among Workers in
Copper-Beryllium

Distribution Centers.  J Occup Environ Med. 48(2):204-11.

Thomas CA, Bailey RL, Kent MS, Deubner DC, Kreiss K, Schuler CR. (2009) 
 HYPERLINK
"http://www.ncbi.nlm.nih.gov/pubmed/19618813?itool=EntrezSystem2.PEntrez
.Pubmed.Pubmed_ResultsPanel.Pubmed_RVDocSum&ordinalpos=1"  Efficacy of a
Program to Prevent Beryllium Sensitization Among New Employees at a
Copper-Beryllium Alloy Processing Facility.   Public Health Rep.124
Suppl 1:112-24.

Viet SM, Torma-Krajewski J, Rogers J. (2000)  Chronic Beryllium Disease
and Beryllium Sensitization at Rocky Flats: a Case-Control Study. 
AIHAJ. 61(2):244-54.

THIS INFORMATION IS DISTRIBUTED SOLELY FOR THE PURPOSE OF
PRE-DISSEMINATION PEER REVIEW UNDER APPLICABLE INFORMATION QUALITY
GUIDELINES. IT HAS NOT BEEN FORMALLY DISSEMINATED BY OSHA. IT DOES NOT
REPRESENT AND SHOULD NOT BE CONSTRUED TO REPRESENT ANY AGENCY
DETERMINATION OR POLICY.

THIS INFORMATION IS DISTRIBUTED SOLELY FOR THE PURPOSE OF
PRE-DISSEMINATION PEER REVIEW UNDER APPLICABLE INFORMATION QUALITY
GUIDELINES. IT HAS NOT BEEN FORMALLY DISSEMINATED BY OSHA. IT DOES NOT
REPRESENT AND SHOULD NOT BE CONSTRUED TO REPRESENT ANY AGENCY
DETERMINATION OR POLICY.

 PAGE   50 

THIS INFORMATION IS DISTRIBUTED SOLELY FOR THE PURPOSE OF
PRE-DISSEMINATION PEER REVIEW UNDER APPLICABLE INFORMATION QUALITY
GUIDELINES. IT HAS NOT BEEN FORMALLY DISSEMINATED BY OSHA. IT DOES NOT
REPRESENT AND SHOULD NOT BE CONSTRUED TO REPRESENT ANY AGENCY
DETERMINATION OR POLICY.

  PAGE  51 

