-------
1
o
(O
(ft
(ft
tft
.2?
(A
O
I
f— U
il
r— «8
10 *» •
c=§-
O i£
•a w 4J
5 v k
C/> «« *8
ID (A
ja « •»-
U O.T3
•f x e>
•§°^
o> o
-j= tn
O 4J Q>
*ia)je ex
QMS
SL(Ur—
X •»- flS
o-o >
TJ V V
S^g
S (ft O
o^ "
je*o e
*J (B
3 k 4-»
g
4l
•P C «
42| g »
«A u v 5 O *»
V O 4-» > (ft & k k **
T-_$W«>«'>«ftX * 3 O
C C 3r—
•=. O
••= o
1
*•> c»— «-o o> i
«0 V O -^ * >»««
JB JD "O
ctwte. *J «>** y^r-
^lal^i
«»C«*-C4S(O3«>(n«-«
•> u o o > Q. k JC J3 4-» «^-
~ C. Co O.*J 3 JB
O O Q> (ft W **
•StJ*-!0 .5^*tu
3 > 4J S W -M •— (ft«*- 3
-oi-c
pu. 10
k >
3 3 a>
O
k « C 0)
b j= o k v 5 °*£ S. S» £ *3
*- -Mt- C(ftJB 3« U
4-*t>3CUO)U4->(ftS
3 eo-p jc «>«»•»-
kt9 >>"^" OW«OO>Ov><«-
13 10
e
O
(ft
.
>t x
C? 4>
4> V
• x: •
,«OO4-> 4J 4-» C 4-» (ft
«- OUO> Uv- «f- Or— i -a s a> -a 4-> a> >«
o. o> «> m e M JE 5 •»- *• •—
o>oot.o4jC*«*-«
O-»-3 ID
* •> e-o
4J JC O C (ft
£« k«^ o"&+»-~ «ii-
O) C 4-» JB
: a> a> 3X a> e«H-«m4->
I 4-> M >)JC l-> O k CO
_j IB «f- o • +* «> 3 o» e
lloS.«ft*)^--po«>S^^-
r—-r- X VJC X V T-'»J'0'-3
o «<>—
t>kkO4->k jQJtf O-^U
JC«)3i- CX*»-
II?
(/>
42? ,
» «*• k e
** .«„-;= 2£ «
+* »r— «> IQ t> k V IB 3
k (ft k 9 WO •— M
3 O 3 (ft 10 S 3 O +*
C ^ ^»»
s*. «, .0 . - «•
•^ *^ <
5tt,
•p •• 3 • ie « » e «> (ft
v o> en S e u «ft «OT> __ «
*j £ -w Co »- C * o» •
C-^eO«H-f-«)«>>-l4->p k O^ 4J k > JCJEkO
MS-? A. -o 4-» o «o ex X f—»—•»- > **
klO-MWS JQ««M«0«
v o eni
^ «*- k « «f
o >>«*• k a. fc e <-»
«o u u kf*i- e
a u
f—JBU>— V • (ft 3T3 r—TJ •>
SJ*-42§lg~£-g.Saslfe1w
^•Si^gSSls^M^-s
on
64
-------
u •—
4.» >»••
O O
.— M 4-» J= «
^ *! g^S
•a « *» x
•*•», 3 o><-» JE v>
gtt£5|xj.£
To ^
1- 4-* «> « O
4Q 'O *0* C» ^* *^
•— o. -o o
W 3-0 3 f- V f-
— U E 4-» 4J»— B
o *° *" o * t.
fc. O O.B W
i* s 10 t. p o ja
w 6
Q. l/> fc« «-•
x-»-
• V O C f- •«— IO ©•—
1^' >4Jf «T O'
I ^C IO +•> I- O
2«j CS.O WO
4) 3 V) ja B
. «> S W W -t- -^.
• *-»s. w S*- <
0 0
§Z S «*" !?'=1
O U I
fll 41 O* B C
u -P-
O uro
$. §^*^
*- JCO»
^- 3 • -
s- « x-o o>
(O TO E
ft) C MTJ-^"
>l^> S J=
^4J«f- fc 4J
a» *j © s. is
.E « 3 a>
4J «S *• W C
O W.O
v>. §> *« o
4J fc U «
S3 «B «
^i^^l
gl^i*
O.4I 4J Q.<-
s^gss
W S 4Jf
^*
s
to
w ^» •» w ^^ i
4-> U V U I
t» •— e U
»«>©&>
v>
p §^-.©5 °ii'1" s. t.
v- 8^*5 B* ? ««^ 8
fc c « • s e « a> u
J«»i3-a'"°e5o0
MMF-W3t-4J©4->4->>)
«^ V O 'O ^J O fl9 *^ 4^ U ^^
^™* ^O Q fc» ^ T3 C 'O 41 S
— -^""ScIS ^o-^«»«»
>£OBeH-»UfUA>
•»- 4=» © O © © 3 U
U-^-5- « W» 3
««
SO* « X
Of 3 U<—
O)W»—
e fc. * * .^e
3«e 3v>
ovvi«>
•? > 0 J-
© >>TS • >
*j i— .ev>4J-i-
«0*J
CLf
$-
o L> o
4JO«5i
v~ e ^a^> 3 x v o. e
o-o «eoEi-5>3§
«•>- 0) t. > W O> 4-1 © OMB ** 4J
© O.i— *«««•*- *M <^ «
S. 3 3 cBf&4VO>B Ol
-*5s*S28^>,3>
!
iS ">
^«s
»'v
la
65 .
-------
V>
4J
M
1
t
V)
S
8
4J O . O
BJC S-H-
tfl V 4-» O C
O tft 3
t/> C/S f— 0
3 3 10 B
>» « «0
V)
.So££iV
lUl^O.^^
0) -D^
4J • O B I I
«A O «rt ^ »-l
S-T3 M JC*-IJC
O C C/) S • 4-1
o«*-
U CO C^ «Stf JC O Qt
E cr> _ «A v» jc
(O rl Kf- W 4-»
o p t- «« a>
B C 0) 3 >i
o) «<+-,— 3f— ja
CZ Q} f& ^™ lO
3 ST3 3 *D
r— 4J O> T3 > V
=s°*
_u5|||
"p BjajS"0*- s
exS t>^ o B o
^» «•# *•• *•» ^*t^"^
*>-a ft o o 3 o
H- O S. Q.M O «/»
«r- C O> 41 •»- O W
-J Q.X £.-0 5 B
B Jrf
fl (A
81 I
S.8"«o
Of- x S.w
«j ^*» JC O1««- S 3
B4J e e «« o
•^ O f O V)
•a w o»—•»-
> »•) Old O
>»0 X CO
r- 4-» « >»«> S.
*» «« x
!•§£
«/>
o o o w '
•a-f -o $- :
B 4-» 4-» B 3 i
Q. S. V O J
o> 3-0 E 5.:
19 3*0 s- V 4^
4-» 3 Q.f-
T3 O «/» 3 JC
•P E4J O^g*-'
S» W 3 B B 4-»
<0 <0 M O , «C
k (A fc. X
a> «e 4-» o
«
t> jc •> ja
g^
rM •
§
§
°-gr2!
2?tJS
«|«8L
• m x
•1 «A O» V
S19«
«Atn «
gggl*
1°-^^
•? k- R«*-
•««'-
•a *o
B tn
•e • jc oo
. c «2
fc> »• IB «i
O « 'B
C >t t. «
**-
* *r B 3
"5^
o> u
ja s o
las
f** €rt €^
10 «^ M M
4-1 *O •> O
OO*5
4J €> X
§
Mr—
S
1
S
66
-------
population risks were reported in column K of Figures 4 through 19. All assumptions and
caveats associated with these population risks are stated in the footnotes and comments of the
referenced figures (Figures 4 through 19).
Figures 2 and 3 should be interpreted by considering both their uses and their
limitations as follows: ' ; , -,-,.-
• Lifetime risks to individuals as shown" in Figure 2 are highly dependent on "
local exposure conditions and activity patterns. The results of broadly
focused studies by others are intended for comparison purposes and do not
cover the full range of all local exposure conditions. Despite efforts to
equally compare the indoor poilutants, differences are bound to exist based
on varying assumptions by different researchers.
• Data on nationwide risks as presented in Figure 3 are important for broad
policy considerations but are not widely .reported by researchers. Additional
work needs to be done to expand beyond the relatively few reported studies
and to perform sensitivity analyses to look at the range of outputs given
various input assumptions.
Figures 2 and 3 should be used in combination to draw conclusions. For
example, the 5,000 to 20,000 estimated deaths due to radon (Figure 3) are
greater than reported estimates for other pollutants, yet individual lifetime
risks due to environmental tobacco Asmoke and some organic compounds can
be very high as shown in Figure 2. .'" .
For comparisons of the assumptions implicit in the summary data shown,
readers are encouraged to follow the pathway: leading to greater and greater
detail. Summary data in Figures 2 and 3 come from column I and
column K of Figures 4 through 19. These figures clearly show key
measurements or assumptions in Columns A through K. They are further
described in comments below each risk characterization figure. When
additional explanation is required, it is provided in the text near each figure.
Finally, since no figure and brief analysis can cover all aspects of the
complex studies analyzed, readers are- encouraged to refer to the original
references cited and listed at the end of this report.
67
-------
REVIEW AND ANALYSIS OF STUDIES
The following sections present a detailed review of risk characterization studies for
radon, ETS, volatile organics, formaldehyde, and asbestos. Each of the studies reviewed has
a corresponding figure which presents the components of the risk characterizations. The
discussion is organized by pollutant category.
RADON
There are numerous studies in the literature that report risk estimates associated with
exposure to radon and radon decay products. Most of the risk estimates are based on
epidemiological studies of underground miners. Miners are typically healthy males, working
in an occupation that requires a much-higher-than-average amount of physical activity and
breathing air containing a pollutant mix (airborne participates, chemical vapors, and a
relatively high concentration of radon and radon decay products) that is quite different from
most work or home environments. As a result, the uncertainty associated with extrapolation
from the dose-response factors typically derived for miners to a dose-response factor for the
general population is high (Ellett and Nelson, 1985; Nazaroff and Nero, 1988).
The information presented in this section is centered around an attempt to compare risk
estimates from three recent studies: the U.S. Environmental Protection Agency's Radon
Reference Manual (1987c), the most recent report from the National Academy of Sciences'
. Committee on the Biological Effects of Ionizing Radiations (National Research Council,
1988), and the National Council on Radiation Protection and Measurement's Report 78
(NCRP, 1984). Although other risk estimates have been reviewed, these are presented as
representative of the best available risk estimates. Some of the terms used in this discussion
are defined in Appendix A.
U.S. Environmental Protection Agency (1987c)
Figure 4 presents a review of the characterization of risk attributable to exposure to
radon and radon decay products. The derivation of EPA's risk estimates involves
assumptions relating to four steps in the estimation process: (1) determination of radon decay
product concentrations from radon concentrations; (2) estimation of cumulative radon decay
68
-------
product exposure; (3) conversion of individual cumulative exposure to lifetime risk; and
(4) projection of individual lifetime risks to the entire population. With regard to the
relationship between radon and radon decay product concentration, it has been found that the
equilibrium fraction (defined in Appendix A) ranges from 0.3 to 0.7, with an average of
approximately 0.5. Using the average equilibrium fraction of 0.5, a ratio of 200 picocuries
per liter of radon to 1 working level (WL) of decay products is considered to be fairly typical
for residential environments. EPA used this ratio to convert radon concentrations to the
average working level of 0.004 shown in column C of Figure 4.
The estimation of average lifetime dose of radon and radon decay products requires
several assumptions in order to convert cumulative residential exposures to working level
months (WLMs), defined as the integrated exposure a miner receives during 170 hours in a
one-working-level (WL) environment. First, an adjustment factor for exposure duration is
needed. EPA assumed that a resident is exposed to a given radon level 75 percent of the
time (shown in Figure 4, column D), as opposed to 170 hours per month. Second, an
adjustment factor for inhalation rate is required since, on average, the inhalation rate of a
miner is considerable higher than trie rate for the general population as a result of the miner's
increased physical activity. EPA assumed the average breathing rate of an adult is 15.3 liters
per minute while the rate for a miner is about 30 liters per minute. This ratio (15.3 to 30)
was used by EPA to correct the estimate of cumulative residential exposure expressed in
WLM (shown in Figure 4, column F). EPA's estimates of cumulative exposure do not
account for potential variations in lung cell sensitivity or differences in sensitivity of different
individuals.^
Estimates of the risk of lung cancer associated with exposure to radon and radon decay
products are obtained primarily from epidemiological studies of underground miners. Since
miners in the study are still living, the risk observed during part of a lifetime must be
extrapolated to a whole lifetime. EPA has adopted a relative risk model to estimate an
increase in risk per WLM of one percent to four percent. EPA's calculations are based on an
assumed linear dose-response relationship and a minimum latent period of 10 years.
EPA has projected lifetime risk to the entire population resulting in predictions of 5,000
to 20,000 deaths per year from radon exposures. The following equation was used:
69
-------
Total LCDs from Indoor Radon = CR X T X
x TCR X POP
where:
CR = average (mean) lifetime indoor radon decay product concentration
= 0.004 WL-life
T = average interval of lifetime exposure in hours, following a 10 year
minimum induction period during which no lung cancer will be
observed, assuming 75% occupancy and 73.88 year life span (1980
vital statistics)
= .75 X (73.88-10) x 365 X 24 = 419,691.6 hours/life
FWLM = factor converting average cumulative indoor exposure in WL hours
to working level months (WLM) for a miner (since risk estimates are
based on miner data), accounting for 170 hours per month exposure
period per WLM (by definition), and the difference in breathing rate
between the average adult (15.3 liters per minute) and a miner
(30 liters per minute)
= 1/170 x 15.3/30 = 0.003 WLM per hour
RRRM = relative lung cancer risk for lifetime exposure to radon, per WLM,
using relative risk model
= 1% to 4% per WLM \
TCR = underlying annual average of U.S. lifetime lung cancer risk (1980
vital statistics)
= 4.584 x 10'4 per person.
POP = 1980 U.S. population
= 226,545,805
National Research CouncU (1988)
A relative-risk, time-since-exposure model is used by the National Research Council's
Committee on the Effects of Ionizing Radiation (National Research Council, 1988) for
computing an age-specific lung cancer mortality rate attributable to exposure to radon and
radon decay products. Components, with appropriate conversion of units to fit the generic
Risk Characterization Framework, are shown in Figure 5.
The National Research Council (1988) report uses the time-since-exposure (TSE) model
to generate tables of gender-specific lifetime risks of lung cancer mortality by age exposure
70
-------
started and age exposure ended and at annual exposures ranging from 0.10 WLM/year to
20.00 WLM/year.
The mathematical form of the TSE model is:
• . « •
r(a) = r0(a) [I+/8«y(a) (^ + O.SW^]
where:
r(a) = age-specific lung cancer mortality rate from all causes
r0(a), = age-specific background lung cancer mortality rate from all causes other than
radon • -• •
/} = the slope of the dose-response relation
(a) = effect of age on risk (1.2 for ages < 55, 1.0 for ages 55 to 64, and 0.4 for
ages > 65 years) ,
Wt = WLM of exposure incurred between 5 and 15 years before current age
f
The steps used in applying the TSE model are:
1. Each year of the period of interest is placed in the appropriate interval:
a. 5-15 years before current age,
b. > 15 years before current age;
2. The total annual risk for the person's age is calculated using the TSE risk equation
shown above (using an appropriate background age-specific risk r0(a));
3. The calculated value of r(a) is multiplied by the chance of surviving all causes of
death to that age (including the increased risk due to exposure).
The choice of an appropriate age-specific background rate is dependent on proper treatment
of smoking as well as gender and calendar time.
National Research Council (1988) estimates the unit lifetime risk (response factor) of
lung cancer mortality due to a lifetime exposure to radon progeny for males at 5.06 X 10"4
per WLM of exposure and for females at 1.86 x 10"4 per WLM of exposure. This estimate
is based on a lifetime (69.7 years for males and 76.4 years for females) exposure to
0.1 WLM/year of radon decay products. Conversion of unit lifetime risk to average lifetime
individual risk is shown in Figure 5, column I.
71
-------
National Council on Radiation Protection and Measurements (1984)
The NCRP risk model relies on the Harley and Pasternack (1981) Model B of lung
cancer excess due to radon progeny. Components, with appropriate conversion of units to fit
the generic Risk Characterization Framework, are shown in Figure 6.
The mathematical form "of the model is:
85
LR(t0) = 2A(t|t0)
where:
LR(to) = the lifetime risk from a single annual exposure at time t0
A(t| y = attributable annual tumor- rate at age t (t > 40) due to a single annual
exposure at t0. If exposure occurs after age 35, risk commences at
t0 + 5 years, if exposure occurs before age 35, risk commences at
age 40.
The equation used to calculate A(t| t0) is the following:
where:
A(t|g = RC (Pt/Pto)
RC =
risk coefficient; assumed 10 x 10"4 cases/yr/WLM, which is an
average value of all epidemiology studies reviewed
= life-table correction to account for death from other causes
= probability that an individual will be alive at age t0.
= probability that an individual will be alive at age t.
= decrease in rate of risk expression due to repair, cell death, or
unspecified mechanisms.
12
-------
The model includes the following assumptions:
1. Based on a review of epidemiologic data on underground miners, estimates
of attributable risk range from l.f> to 45 X 10"6 with a rounded average
value of 10 x 10"6 cases per year per WLM of exposure. This average
value was used in the calculations in the National Council on Radiation
Protection and Measurements (1984) report.
2. Following a latent period, the tumor rate is an exponentially decreasing
function of the time since exposure.
3. Disease rate excess associated witira single exposure increases with age at
exposure.
4. Lung cancer is rare before the age of 40 years.
5. Median age of lung cancer among miners is about 60 years in nonsmokers
and 50 years or older in smokers,
6. The minimal time for tumor growth, from initial cell transformation to
clinical detection, is 5.years. :,.:-^:v:!>;;
The model specifies a 5-year latency period for persons first exposed at age 35 or older and a
(40 - u)-year latency period for persons under the age of 35 years. To obtain lifetime risk
due to a single exposure at age u, the equation is integrated over t from age 40 to maximal
assumed life (age 85 years). To obtain the excess risk at t due to all previous exposure, the
equation is integrated over years of exposure,-uj,...,un. To obtain lifetime excess risk from
all exposures, the equation is integrated over t and u. , „ •
The NCRP model predicts 130 lung-cancer/deaths per 106 persons per WLM of
exposure. The estimate'is based on an average lifespan of 70 years and a lifetime risk for
lifetime exposure starting at age 1 of 9.1 X 10"3. Conversions to average lifetime;individual
risk estimates are shown in Figure 6, column I.
ENVIRONMENTAL TOBACCO SMOKE (ETS)
A number of studies have been completed which estimate the annual number of deaths
attributable to exposure to environmental tobacco smoke (ETS) and the lifetime individual
risks attributed to such exposures. Six of these studies are reviewed using the Risk
73
-------
Characterization Framework. These reviews are presented in Figures 7 through 12.
Although different methodologies and source data are employed by each of the authors, most
of the estimates for both lifetime individual risk and annual death attributable to exposure to
ETS are similar. The predominant health effect that has been assessed is lung cancer.
However, Fong (1982) includes lung cancer and emphysema together in his population risk
estimates; Russell et al. (1986) include heart disease, bronchitis, emphysema, and lung cancer
together in their risk estimates; and Wells (1988) provides separate risk estimates for other
cancers and for heart disease.
Robins (1986)
The most mathematical analysis of risk attributable to exposure to ETS is Robins'
(1986) appendix to the National Research Council's (1986) report entitled "Environmental
Tobacco Smoke: Measuring Exposures and Assessing Health Effects", which presents a
sensitivity analysis approach to risk assessment that considers both the epidemiological data
and some measures of exposure to the constituents of ETS. Study components are shown in
the Risk Characterization Framework in Figure 7. Robins' estimates of the lifetime risk of
lung cancer attributable to ETS in a nonsmoker with moderate ETS exposure are developed
from an analysis that considers 30 different "exposure histories", three different estimates of
the coefficients of the dose-response model, and two different estimates of the overall
summary rate ratio (i.e., average relative risk) which is assumed to be the ratio of the "true"
relative risk in exposed subjects to that in unexposed subjects. ("True" is used here to mean
that this is, in fact, an assumed true value and not the mean of a sample.)
Two different average relative risk ratios are used in this analysis. A weighted average
of the average relative risk ratios from 13 epidemiological studies presented in the National
Research Council report is calculated to be roughly 1.3. In the analysis, the author assumes
that a weighted average of 1.3 is causally related to differences in ETS exposure between
"exposed" and "unexposed" individuals and not to bias. A second average relative risk ratio
of 1.14 was calculated based on the subset of these epidemiological studies that were
identified as U.S. studies. The assumption underlying the use of these two different average
relative risk ratios is that the true average relative risk ratio is most likely 1.3 and at least
1.14.
74
-------
: In this risk assessment, the most important assumption was that the average relative risk
was caused by ETS exposure. This assumption is basic to the equations used for estimating
the true relative risk, the dose (expressed as the carcinogen-equivalent number of actively
smoked cigarettes inhaled daily), and the.subsequent estimates of lifetime risk of death from
lung cancer. The following is a presentation of the general equations used by the author:
where:
EXCESS^) = tf16 excess of lung cancer deaths at age t due to a specific
exposure history : ...,..;;
70(t) = the incidence of lung cancer death at age t in the absence of
all exposure to ETS '••-'•''
RREXCESS(t) = the excess relative risk for lung cancer due to a particular
exposure history. Mathematically, it can be expressed in the
following way:
- 1
where RR(t) is the true relative risk at age t for the exposed group compared to a completely
unexposed group. The terms in this model are described mathematically by the author in his
article (see National Research Council, 1986, pages 325-326),
One method the author employs in the analysis. is to estimate the coefficients of a
five-stage cancer process under the assumption that cigarette smoke influences the first and
fourth stages of the process. The author further assumes that the ratio of the effect (on a
multiplicative scale) on stage four to that of stage one is the same for ETS and mainstream
smoke. Thus, an estimate of this ratio can be obtained by fitting a five-stage cancer model to
data on the lung cancer experience of active smokers. : The author calculates this ratio based
on two different data sets and calculates a third ratio based on' an adjustment to one of the
data sets. .
In addition to the two average relative risk ratios (i.e., 1.3 and 1.14) and the three
different sets of values for the coefficients of the multistage model, the author developed
75
-------
30 exposure histories for the exposed and unexposed individuals. The 30 exposure histories
describe the percentage of time individuals in various age groups were exposed to various
ETS dose rates (relative to the current ETS exposure of an adult nonsmoker without a
smoking spouse). The author states that smaller differences postulated between the lifetime
ETS exposure of "exposed" and "unexposed" individuals will be associated with larger
estimates of the true relative risk. Thus, some exposure histories were selected that would
modestly underestimate the true difference in exposure between the "exposed" and
"unexposed" study subjects, and others were selected that would modestly overestimate this
difference. Shown in Figure 7 are the maximum and minimum estimates from the range of
lifetime risk estimates for males and females covering the 30 exposure histories and three
response factors for the relative risk ratio of 1.3. Also presented by the author but not shown
in this figure were the same calculations using a relative risk ratio of 1.14. The range of
estimates of lifetime individual risk (shown in Figure 7, column I) for nonsmokers is based '
on three separate estimates of the coefficients in the multistage model and the range in
exposure calculated for the 30 different exposure histories. Population risks expressed as the
total number of lung cancer deaths per year among nonsmoking women (or men) (AN)
attributable to ETS in 1985 were calculated by Robins (1986) using the following equation:
AN = 2 AF (t) X I0(t) X N(t)
t '
where:
N(t) = the number of nonsmoking women (or men) at risk at age t in 1985
I0(t) = the age-specific death rate among nonsmoking women (or men) in
1985
AF(t) = the age-specific fraction of lung cancer deaths due to ETS exposure in
nonsmoking women (i.e., the average relative risk minus 1, divided by
the age-specific relative risk)
76
-------
..;:, In order to estimate age-specific relative risk among nonsmoking women (or men), age-
specific estimates of the probability of being married to a smoker and the true relative risk in
exposed and unexposed subjects are needed. The age-specific estimates of the probability of
being exposed are taken from Garfinkel et al. (1985). Relative risk estimates were .calculated
in the same way as described for lifetime individual risk estimates.
Robins (1986) provides estimates of ETS attributable lung cancer deaths among U.S.
nonsmokers in 1985 based on epidemiologic data, as shown in column K of Figure 7.
Estimates are presented for females and males separately with a maximum and minimum
across all exposure histories and five choices of dose-response coefficients. As shown in,
Figure 7, the maximum and minimum estimates calculated based on epidemiological data are
3,220 and 1,768 for females and 1,942 and 721 for males respectively.
Similar sensitivity analyses are presented by Robins (1986)' for ex-smokers andr : ,>
continuing smokers for an alternative true relative risk ratio of 1.1.4 and for dosimetrie
measurements rather than epidemiological studies,: For example, the author estimates the;
individual lifetime risk of lung cancer attributable to other people's cigarette smoke for an.
ex-smoker who smoked one pack per day from age 18 to 45 and was moderately exposed to
other people's cigarette smoke lies between 5.2 x 10"3 and 2,03 x 10"2. Note these risks
are about 30 percent higher than the estimates shown in Figure 7, column I, for lifetime
nonsmokers. Similarly, use of a relative risk ratio of 1.14 rather than 1.3 provides a range
of estimates of annual lung cancer deaths for female nonsmokers of 935 to 1,730 and for
male nonsmokers of 360 to 980. These estimates are approximately 50 percent lower than
the estimates shown in Figure 7, column K.
Repace and Lowrey
The Repace and Lowrey (1985) estimates of annual risk of limg cancers attributable to
exposure to ETS are based on estimates that U.S. nonsmokers are exposed to between 0 and
14 mg of tobacco tar per day with the typical nonsmoker exposed to 1.4 mg i>er:day. A
calculation based on age standardized differences in lung cancer .mortality rates between
Seventh-Day Adventists (SDAs) who never smoked and demographically comparable
non-SDAs who never smoked (from studies by Phillips et al., 1980) yields a passive smoking
risk rate of 7.4 X 10"5 per year. This was computed from an estimate that about 4,700 lung
77
-------
cancer deaths per year were attributable to exposure to ETS among the 62.4 million U.S.
i
nonsmokers. Results are shown in Figure 8. Note that data from the authors' reference are
in columns I, J, K, and E. Column H was back-calculated in this work.
Seventh-Day Adventists have a lifestyle which avoids smoking and is oriented toward
socialization with co-religionists. Moreover, 40 percent of the study group worked for
church-run organizations; so the SDA lifestyle involves less exposure to ETS than
demographically comparable nonsmokers from the general population.
The phenomenological exposure-response relationship derived by Repace and Lowrey
(1985) was shown by them to predict to within 5 percent the lung cancer mortality rate and
mortality ratio reported in the American Cancer Society cohort studied by Garfinkel (1981) in
his study of passive smoking and lung cancer, as well as to explain the differences between
the study of Garfinkel (1981) and the cohort studied by Hirayama (1981). Repace and
t
Lowrey (1986), in a refinement of their estimates to adjust for sex and standardized to the
total nonsmoking population, calculated a total of about 4,891 lung cancer deaths per year
from passive smoking, of which 1,441 were estimated to be male and 3,450 female. These
estimates are shown in parentheses in Figure 8, column K. These refined estimates are
within 5 percent of their earlier estimates and are consistent with the results of a later
case-control study by Garfinkel et al. (19.85) to within 5 percent and to within 5 percent for
the weighted lung cancer mortality ratios of the 13 epidemiologic studies of passive smoking
and lung cancer as analyzed by the National Research Council report in 1986 (Repace and
Lowrey, 1987).
The methodology for calculating these adjusted estimates of lung cancer deaths
attributable to passive smoking is to start with the age-specific estimates for both sexes
combined. These age-specific estimates are then separated into male and female estimates of
lung cancer deaths, assuming 28.6 percent occur in men as presented in Arundel et al.
(1986). The percent of over- or under-estimation for each sex in each age group is derived
from Arundel et al. (1986). These percentages are used to adjust the estimates of lung cancer
deaths for each sex in each age group. For three of the age groups (e.g., 35-44, 45-54, and
55-64) both male and female estimates of lung cancer deaths were adjusted downward while
for the remaining two age groups (65-74 and 74+) estimates of lung cancer deaths for both
sexes were adjusted upward.
78
-------
Russell et al. (1980)
Urinary nicotine concentrations in smokers and. nonsmokers are used to estimate
exposure and risk to nonsmokers from ETS in Russell et al. (1986). This analysis is
summarized in Figure 9. Urinary nicotine levels were chosen as a chemical marker for
smoke intake because nicotine is specific to tobacco smoke, and urinary nicotine is stable and
thus is suitable as a measure of tobacco smoke exposure over several hours. ,The levels easily
differentiate "exposed" and "non-exposed" nonsmokers.. The average concentration in
nonsmokers who reported some exposure was 15.5 ng/ml, three times the value of 5.2 ng/ml
for those who reported no recent exposure. The average urinary nicotine concentration of
188 nonsmokers from four studies was 10.8 ng/ml compared with 1,471 ng/ml in a combined
sample of 229 smokers. On the basis of these measurements they estimated that nonsmokers
receive on average about 0.7 percent of the nicotine dose of active smokers.
The authors assume that the dose-response effect at low levels is linear. This leads to a
further assumption that a fair .estimate of deaths due to passive smoking can be based on a
proportion of deaths attributed to active smoking. In other words, the risk of death from
passive smoking was assumed by the authors to be approximately 0.7 percent of that due to
active smoking. The authors calculate the number of premature deaths per year among
nonsmokers is estimated to be about 1,000 in Britain and over 4,000 in the U.S. as shown in
column K of Figure 9. . ( r
Wigle et al. (1987) ;
Information on the proportion of people who had never smoked among victims of lung
cancer m Canada is compiled in Wigle et al. (1987). This analysis is summarized in Figure,
10. The total number of lung cancer deaths per year attributable to ETS in Canada was. .
estimated to be 330. This number was derived by applying the age- and sex-specific rates of
death from lung cancer attributable to ETS that were estimated by Repace and Lowrey to the
Canadian population at risk as determined by the 1983 survey on smoking habits of .
Canadians. The U.S. population is approximately 10 times the population, of Canada; ,
therefore 3,300 deaths from lung cancer were estimated in the U.S. based on .this study. ,
79
-------
Wells (1986)
The epidemiological literature on ETS and adult lung cancer, other cancers, and heart
disease is reviewed in Wells (1986). The primary model that was used by the author to
estimate risk combined relative risks from the various studies that pertained to a given sex
and disease and assumed that the combined relative risk was constant with age. A combined
relative risk was calculated using the following equation:
Rcb = exp
Wco In Rco +
In Rcc
W + W
"co ~ vvcc
where:
w =
co
w =
cc
= the combined relative risk
= the relative risk for cohort studies
= the relative risk for case control studies
the weight for cohort study
the weight for case control study
The weights for the cohort and case control studies are the inverses of the respective
variances. The excess death rates attributable to lung cancer for never smokers for passive
smoking (Dpx) for each sex are shown in Figure 11, column K. These estimates were
calculated for each sex and five-year age range from never smoker death rates (Dns) by the ,
following equation:
where:
Fp = the fraction of the population that is exposed
R = the combined relative risk (Rcb from above)
80
-------
Deaths were then calculated by multiplying the passive smoking excess death rate by the
exposed population for each sex and five-year age interval, then summed. The sum for males
and females is shown in Figure 11, column K.
For female lung cancer three cohort studies and 14 case control studies were combined.
The overall combined relative risk is 1.44 with a 95 percent confidence limit of 1.3 to 1.7.
The male lung cancer relative risk was based on a combination of two cohort studies and
seven case control studies. The overall combined relative risk is 2.1 with a 95 percent
confidence limit of 1.3 to 3.2.
In addition to lung cancer risk, the author estimates passive smoking related deaths for
other cancers and for heart disease. By disease, the total consists of 700 lung cancer deaths,
11,000 other cancer deaths, and 32,000 deaths due to heart disease. For each million of the
total population, the deaths by disease are 13 for lung cancer, 46 for other cancers, and 134
for heart disease. , ,
Fong(1982)
The hazard of ambient cigarette smoke to nonsmokers is compared to the hazard of
primary smoke to smokers in Fong (1982). Figure 12 provides a summary of his analysis.
Fong assumed the nonsmoker worked in an office for 11 hours per day where 37 percent of
the workers smoke and commuted on an unventilated bus 1 hour each day where 37 percent
of the occupants smoke. The remaining 12 hours of each day was assumed to be unexposed.
The ratio of the exposure of nonsmokers to the exposure of smokers is represented by the
following equation:
R = r1 X r2 x r3 X r4
where:
R = the ratio of exposure of a nonsmoker to a smoker
rj = the ratio of the density of secondary smoke to primary smoke
r2 = the ratio of the times of exposure of the two cases
81
-------
r3 = the ratio of toxicities of the two types of smoke
r4 = the ratio of the amounts of the ambient smoke (secondary plus side stream)
a nonsmoker is exposed to and the primary smoke a smoker is exposed to
The maximum value of the ratio R calculated for this scenario is 1/13.
Since the effect of smoking in the U.S. has been stated as an excess of deaths of
350,000 per year, the maximum risk of ambient smoke to nonsmokers in the U.S. was
estimated at 50,000 excess deaths per year (based on a smoker to nonsmoker ratio of 37:63).
Fong (1982) assumes that 16 percent of the total deaths due to smoking are caused by lung
cancer and 4 percent are due to emphysema. The remaining 80 percent are caused by heart
disease and other cancers. The author notes, however, that other factors such as diet
confound the effects of ETS on heart disease and cancers at other sites. He, therefore,
assumes that only 20 percent of the risk of death from all causes attributed to ETS exposure
as calculated by the above methodology is the realistic state-of-the-art estimate of the excess
risk of death attributable to ETS. Thus the excess risk of death from ETS exposure is 1/60
that of primary smoke. This corresponds to an excess of 10,000 deaths per year.
VOLATILE ORGANICS
This section includes risk estimates of indoor exposure to volatile organic compounds
other than formaldehyde, which is addressed separately in the next section. Three studies are
included in this review: Tancrede et al. (1987), Wallace (1986), and McCann et al. (1986).
EPA's Total Exposure Assessment Methodology (TEAM) Study provides indoor exposure
concentration data used in all three studies. However, the first two studies use TEAM Study
data exclusively to estimate exposure, whereas the McCann et al. (1986) study uses multiple
literature sources as the basis for determining exposure concentrations to be used in the risk
analyses.
A major difference among the three studies is the methodology employed to estimate
carcinogenic potency. Wallace (1986) uses unit risk values developed by EPA's Carcinogen
Assessment Group (CAG). As part of a sensitivity analysis, he compares these unit risk
estimates with those developed by the "Harvard Group" which explicitly incorporates
uncertainty into the estimate of potency. The potencies estimated by the Harvard Group can
82
-------
be several orders of magnitude higher than the CAG estimates. Tancrede et al. (1987) use
the Harvard Group methodology to estimate carcinogenic potency. Because this methodology
is different from the methodology employed by GAG, Tancrede et al. estimate potencies for
52 chemicals, of which only nine have corresponding CAG values. McCann et al. (1986)
provide a sensitivity analysis on carcinogenic potency for 24 chemicals, based on four
methodologies. Nine of the chemicals have corresponding GAG values. -Figures 13, 14, and
15 provide a detailed review of these three studies.. .,
Tancrede et al. (1987) , —
This study by Tancrede et al. uses indoor air pollutant exposure data from the Total
Exposure Assessment Methodology (TEAM) Study conducted in Bayonne and Elizabeth, NJ,
and Los Angeles, CA, as well as data from a study by Lebret et al. (1984) in the Netherlands
to predict cancer risk from exposure to 52 organic chemicals. Median estimates of cancer
potency were calculated from human data (benzene) or animal bioassay data (20 chemicals)
when such information was available. Cancer potency was estimated from toxicity and .
mutagenicity data and other information on biological activity for the remaining 31 chemicals
using the analogical theoretical methodology developed by Fiering and Wilson (1983). For
most of the 31 chemicals, studies of comparative toxicities and activities in promotion and
co-carcinogenesis experiments of the chemicals and related compounds were used to estimate
potencies. For a few of these chemicals, potencies were estimated from an inhalatipn or oral
LD50 taken from the literature. ,r \ .""".,-..
The review presented in Figure 13 focuses on Tancrede's risk estimates using exposure
data from the Los Angeles, CA, TEAM study. These data are based on personal exposure
monitoring of 200 people in the Los Angeles area during the winters of 1983 and 1984.
Although not presented in Figure 13, the authors present a similar analysis based on the New
Jersey TEAM study and the study by Lebret. et al. (1984) of organic chemicals in four Dutch
homes over a six-month period. Figure 13 also presents only those risk estimates for which
the potency factor was based on human or animal bioassay data, and excludes all; risk
estimates for which the potency factor was based on the analogical theoretical methodology
since this methodology is not in accord with the EPA Risk Assessment Guidelines (U.S. -
Environmental Protection Agency, 1987a). Tancrede et al. 's (1987) analysis of lifetime..
83
-------
median, mean, and 98th-percentile risk to residents of Los Angeles, CA, for individual
organic chemicals is presented on separate framework pages in Figure 13.
The median risk estimate (pages 1 to 3 of Figure 13) is based on 'median exposure and
potency estimates, without incorporating uncertainty, using the following equation:
R = 3 x KT4 ~fi ~d
where:
R = the median estimate of lifetime risk
ft = the median estimate of carcinogenic potency
d = the median estimate of exposure dose
3 X 10"4 is a dosimetry factor which accounts for a breathing rate of the exposed individual
of 20m3 per day; a 100 percent absorption rate; a 70 kg average body weight; and the
conversion of micrograms to milligrams.
The mean risk estimates (pages 4 to 6 of Figure 13) incorporate uncertainty information
into the risk calculation by using median estimates of the potency factor and dose along with
an estimate of a2 in the following way:
R = 3 X KT4 ~0 'd exp(a2/2)
where:
)8, d and 3 x 10"4 are defined above.
The authors assume that potency and dose are lognormally distributed, and thus risk is
assumed to be lognormally distributed with a variance a2 given by adding the variance of
each of the following four factors:
84
-------
where:
a =
= the standard deviation of the risk estimate
the uncertainty in measuring potency in animal experiments
ff =
the uncertainty in extrapolating from animals to humans, assumed to be
loge(4.5) = 1.5
the uncertainty due to different routes of administration, is assumed to
beloge(5)=1.6
ffz = uncertainty (variability) in dose
This results in a mean risk estimate which is exp(a2/2) greater than the median estimate.
The 98th-percentile risk estimate (pages 7 to 8 of Figure 13) incorporates uncertainty
into the risk calculation in a similar manner using the following equation:
= 3 X 10"4 ~p H exp(2.0537o) j
Thus, the 98th-perceritile risk estimate is exp(2.0537a) greater than the median' estimate. A
result of this methodology is that estimates with the most uncertainty (largest standard
deviation) in the input parameters have the highest estimated risk.
Two key assumptions are made in this analysis. First, exposure is assumed to be
continuous for a 70-year lifetime. This assumption is based on evidence that individuals
spend 80 to 90 percent of their time indoors. However, the exposure data used in the
analysis are from overnight personal air samples taken during the winter months in each of
the three locations. Thus, a second assumption is that the overnight personal air sample data
for the home are reasonably representative of air concentrations from all indoor micro-
environments, over long time periods.
Tancrede et al. (1987) also provide estimates of the mean annual total risk for the
mixture of contaminants measured in the TEAM Study at each of three locations: Bayonne,
NJ; Elizabeth, NJ; and Los Angeles, CA. The basic assumption used to derive these
estimates is that the risk from the mixture of organics presented in the analysis is the sum of
the lifetime individual risk posed by each of the chemicals separately. Thus, risks are
85
-------
assumed to be additive. These risk estimates are not presented in Figure 13 since they
include many compounds other than those shown here.
Wallace (1986)
This study uses EPA's TEAM Study data to construct a rough nationwide risk
assessment for the following six chemicals: benzene, chloroform (both air and water),
carbon tetrachlbride, trichloroethylene, tetrachloroethylene, and ^ara-dichlorobenzene. The
TEAM study measured personal air and drinking water exposures to these six chemicals and
approximately 15 other chemicals for 600 residents of New Jersey, North Carolina, North
Dakota, and California between 1981 and 1984. This sample represented a total population
of ~ 700,000 in seven cities. The author cautions that although this is the largest personal
t
exposure data set available, its use in providing rough estimates of population risks should be
considered in light of the many and great uncertainties in these estimates.
The review presented in Figure 14 focuses on the author's estimates of individual
lifetime risk and population risk for the six chemicals for both metropolitan and
nonmetropolitan residents. The equation used by Wallace (1986) for calculating the risk of
excess cancer incidence (y^ for an individual from exposure to a carcinogen is the following:
= aX
where:
a =
potency of the carcinogen, in units of excess cases per unit concentration
normalized to a 70 kg male with a 70-year lifetime (i.e., unit risk factor)
X; = mean lifetime concentration
The equation used to estimate aggregate or population risk (Y) where the sample size is N
and the sum of the weights equals a target population (P) is the following:
Y = Pa wx
86
-------
where:
a = potency as defined above
wx .== the arithmetic mean of the population-weighted exposure distribution
The author notes that these equations ignore differences in individual susceptibility; the
possible multistage nature of cancer; the different effects of early exposure versus late
exposure; and the different actions of initiators, promoters, and carcinogens, among others.
The potency estimates used by,the author for five of the chemicals were unit risk
estimates [expressed as (/fg/m3)"1] developed by EPA's Carcinogen Assessment Group
(GAG)., For benzene, the unit risk estimate is based on human epidemiological data and are
a best (maximum likelihood) estimate. For tetra chloroethylene, trichloroethylene,
chloroform, and carbon tetrachloride, the unit risk estimates are based on animal studies and
are calculated from the 95-percent upper confidence limit on the potency factor. The unit
risk estimate for ^ara-dichlorobenzene is nota'CAG estimate but was calculated by the
author based on a National Toxicology Program mouse bioassay. For comparative purposes,
the author also presents mean and 95-percent upper confidence limit potency estimates
developed by a group at Harvard. Their methodology incorporates uncertainties of various
types into the calculation, A large variance exists between the CAG 95-percent upper
confidence limit estimates and the Harvard estimates.
The author calculates both lifetime individual risk and population.risk for those living in
U.S. metropolitan areas and nonmetropolitan areas. The author assumed that the TEAM
Study data for New Jersey and California are representative exposures for 178 x 106
metropolitan residents, and the North Carolina and North Dakota data are representative
exposures of 56 x 106 nonmetropolitan residents.
In addition to what is shown in Figure 14, the author presents similar estimates for
outdoor exposures to these chemicals and comparative risk estimates using the potency factors
developed by the Harvard group. '
87
-------
McCann et al. (1986)
The literature for indoor air concentrations for 140 organic compounds is reviewed in
McCann et al. (1986). In Figure 15 the risk estimates calculated using both maximum and
mean concentrations combined with the maximum likelihood (MLE) and 95-percent upper
confidence level (UCL) unit risk factors are shown for eight carcinogens. These eight
chemicals (plus formaldehyde shown in the next section) were included in Figure 15 because
they are considered carcinogens by EPA and have unit risk factors estimated by EPA's
Carcinogen Assessment Group. The MLE estimate is shown because in most cases it is the
lowest or similar to the lowest unit risk estimate of the four. Similarly, the 95-percent UCL
is similar to the highest unit risk estimate, in most cases. However, EPA unit risk estimates
are the highest for several compounds including benzene, dichloromethane,
dimethylnitrosoamine, and vinylidenechloride. Figure 15 gives the general range of risk
estimates made by the authors.
The primary sources for the full list of compounds examined in this study were the
published literature and presentations made at the 1984 Indoor Air Quality meetings in
Sweden. Based on this literature review, the authors performed preliminary assessments of
cancer risks for 24 chemicals. All concentration data presented by the authors are direct field
measurements in homes and public buildings that the authors judged to reflect everyday
exposure in normal, noncomplaint homes and offices. They did not include, for example,
concentrations of formaldehyde in UFFI homes or concentrations measured in traditional
occupational settings. When available, maximum and median or mean concentration
measurements were recorded by the authors.
Four different approaches were employed to estimate unit risk values for the 24
chemicals. Maximum likelihood estimates and 95-percent upper confidence level values were
calculated using the multistage model based on the dose-response data given by Gold et al.
(1984, 1986). The authors do not state which animal study these estimates are based upon
when multiple studies are reported for the same chemical. Risks were also calculated using
the most potent TD50 estimated by Gold et al. (1984, 1986). The authors assumed this value
is a point on a linear dose-response curve and have divided 0.5 by the TD50 (estimated as the
(•>
equivalent dose in ^g/nr) to obtain the risk per unit dose. When Gold et al. (1984, 1986)
reported that curves were nonlinear, the authors modified the risk estimates by using a less
88
-------
than (<) or greater than (>) sign. Also, unit risk factors reported in EPA's Health
Assessment Documents (U.S. Environmental Protection Agency, 1985a-f, 1986) were used.
Nine of the 24 chemicals have EPA unit risk numbers and eight of these chemicals are shown
in Figure 15; formaldehyde is shown separately in Figure 18.
FORMALDEHYDE
Interagency Regulatory Liaison Group (1981)
The Interagency Regulatory Liaison Group's Task Group on Formaldehyde (1981)
published an extensive summary of studies on the exposure and resulting health effects,
environmental effects, and health risks attributable to formaldehyde. A summary is presented
in Figure 16. The report provides a detailed breakdown for exposure of subpopulations
including four occupational categories, three consumer categories (e.g., residents of mobile
homes, conventional homes), and ambient air and water. The risk estimates shown in
Figure 16, column I have been calculated by the Interagency Regulatory Liaison Group
(1981) with potency factors that were estimated from an animal bioassay. No scaling factors
were used in estimating human potency; therefore, it is difficult to present an accurate •
relationship between columns G, H, and I in Figure 16.
Tancrede et al. (1987)
Tancrede et al. (1987) present median, mean, and 98th-percentile individual risk
estimates attributable to exposure to formaldehyde in the average U.S. home. The authors
assume formaldehyde exposure follows a lognormal distribution with a median indoor
concentration of 0.05 ppm and a logarithmic standard deviation of 1.8. This is an assumed
value in agreement with several studies for all U.S. houses. The authors use a carcinogenic
potency factor derived by Zeise et al. (1984) of 6a = 0.11 (mg/kg-day)"1 with a standard
deviation of 0.18. The authors also assume the standard deviation of human response
associated with this potency factor to be 1.5. This analysis is shown in Figure 17.
89
-------
McCann et al. (1986)
Formaldehyde was included in the McCann et al. (1986) study on indoor air
concentrations and preliminary carcinogenic risk assessment for 24 organic compounds. This
study has been previously mentioned in the section on volatile organic compounds.
The authors used four different approaches to obtain unit risk values. Maximum
likelihood estimates (MLE) and 95-percent upper confidence level values (UCL) were
estimated from experimental results and tumor data used by Gold et al. (1984, 1986) to
calculate the most potent TD50. A response factor was also calculated from the TD50 value,
which is assumed to be a point on a linear dose-response curve. The fourth method (EPA)
was to use the unit risk factor reported in EPA Health Assessment Documents (Environmental
Protection Agency, 1986). The unit risk factors were applied to both average and maximum
concentrations reported in literature to obtain the eight lifetime risk estimates which are
shown in Figure 18. Lifetime individual risk using maximum measured concentrations
ranged from a high of 9.0 X 10"2 to a low of 3.1 x 10"3. (Figure 18, page 1, column I).
Corresponding values for the mean measured concentrations (page 3, column I) ranged from
a high of 9.5 X 10~3 to a low of 3.7 X 10~6. Due to the nonlinearity of the dose-response
curve for formaldehyde, the potency factor estimates are also dependent on the concentration.
Thus, corresponding values in column H on pages 1 and 3 of Figure 18 are not identical. At
lower concentrations there is a nonlinear change in risk (column I) for MLE and UCL
estimates. Linearity is assumed in EPA and TD50 potency estimates.
Additionally, the authors estimated the percentage of population at "high risk" from
formaldehyde exposure. "High risk" was defined as lifetime individual risks of greater than
one in 1000 (10"3). Populations at risk were calculated based on geometric mean and
geometric standard deviation of concentration data by DeBortoli et al. (1985). Those at high
risk ranged from less than 0.01 percent of the total U.S. population for the MLE to greater
than 99 percent for TD50. These values have been included in Figure 18, column K, page 1,
for estimates of carcinogenic risk for lifetime exposures to maximum measured
concentrations. However, no attempt was made to quantify the number of excess cancers
expected.
90
-------
ASBESTOS
Several risk assessments for occupational exposures to asbestos have been performed
(Hogari and Hod, 1981; Nicholson et al., 1982; and Nicholson, 1983). These assessments
focus on occupational exposures that have occurred in the past. The recent analysis by
Mauskopf (1987), which is reviewed in this monograph," focuses on occupational and
nonoccupational exposures to asbestos from products manufactured under current regulations
for asbestos.and asbestos products. This analysis draws on the work of Nicholson and others
and was viewed as one of the few analyses addressing nonoccupational exposure. Risk
characterization components are shown in Figure 19.
Mauskopf (1987)
This study presents an analysis of cancer risk attributable to exposure to asbestos from
products manufactured during the years 1985 to 2000-for nine product categories (shown in
column B of Figure 19). This analysis takes into account current government regulation of
asbestos and projects cancer risk for the next 100 years attributable to exposure to products
manufactured during the 16-year period, 1985 to 2000. Five occupational and four
nonoccupational exposures identified in the lifecycle of each of the nine product categories
are included in the analysis: occupational and nonoccupational exposure during primary
manufacture, secondary manufacture (occupational only), installation, use, and .
repair/disposal.
Exposure data used by the author vary in duration from 1 to 30 years among the
categories and therefore cannot be presented directly in Figure 19. A weighted yearly
lifetime average exposure was calculated and is presented in Figure 19, column E in units of
fibers per year.' These exposure values must be converted, using the dosimetry factor in
Figure 19, column F, to units applicable to the risk models used (fibers per ml), normalized
for an occupational work exposure (Figure 19, column G). The risk models used were
developed from occupational studies. The unit of measure for dose-in the model is in fibers
per ml over a typical work year. Since data from occupational exposure studies indicate that
excess mortality is proportional to both the level and duration of exposure, it is reasonable to
normalize indoor air exposure to an occupational equivalent. The total exposure in fibers per
year were normalized to an occupationally equivalent concentration, as if that exposure
91
-------
occurred during a typical work year of 40 hours per week, 50 weeks per year with a
o
breathing rate of 1 nr/hr.
Two health effects were evaluated, lung cancer and mesothelioma. The author used a
life table risk assessment model which utilized actual data on the time, level, and duration of
exposure and which provided estimates of individual lifetime risk (Figure 19, column I) and
expected cancer cases (column K) among total exposed population (column J). Estimates
were also generated for when in the future those cancer cases would occur and the mean ages
of those persons contracting these diseases as a result of asbestos exposures. These types of
estimates are not shown in Figure 19.
Cancer cure rates and annual death rates after diagnosis for the terminal patients were
assumed to be constant across demographic groups, exposure categories, and products.
Values estimated arid used in the analysis were the following: cure rates of 8 percent and
2 percent for lung cancer and mesothelioma respectively, and, for those dying from the
diseases, annual death rates after diagnosis of 81 percent and 71 percent for lung cancer and"
mesothelioma respectively. The following dose response factors from Selikoff et al. (1979)
were used: KL = 1.0 x 10"2 (fibers/milliliter)"1 for the lung cancer relative risk model and
Km = 1.5 x 10"8 for the mesothelioma absolute risk model. These are not the same as the
response factors shown in Figure 19, column H. The calculated response factors shown in
Figure 19, column H represent the average lifetime risk from asbestos exposure of an
occupationally normalized exposure of 1 fiber/ml. These response factors were
back-calculated by dividing the lifetime individual risk (Figure 19, column I) by the
calculated yearly average lifetime dose (Figure 19, column G).
The author conducted a sensitivity analysis using alternative dose-response factor.
Upper bound estimates were from Seidman et al. (1979) and obtained from a study of
f\
workers exposed less than a year. Seidman values were KL = 6.8 x 10 and Km =
5.7 X 10'8. Values of KL = 3:1 X 10'3 (Hughes and Weif, 1980) and K^ = 0.7 X 10'9
(Peto, 1980) were used to obtain lower bound estimates. The author predicted a total of
1,541 cancer cases, and the sensitivity analysis revealed a range of 268 to 8,090 total cases.
92
-------
GONCLUSIONS/REGOMMENDATIONS
CONCLUSIONS
• , Few complete quantitative risk characterizations appear in the literature.
While.many scientific studies were found which deal with individual components of risk
assessment, only a few studies combine exposure assessment data with dose-response
evaluations to'produce quantitative risk characterizations (as defined in the U.S.
Environmental Protection Agency Risk Assessments Guidelines of 1986).
• Most of the risk characterization studies which were found and reviewed focus on
cancer risk. - \ ...-'-
For radon, formaldehyde, and asbestos, the cancer risk estimates are for lung cancer. The
analysis of cancer risk from exposure to asbestos includes mesothelioma as a separate
health effect in addition to lung cancer. The analysis of cancer risk from exposure to
specific volatile organics does not specify any particular type of cancer. Three of the ETS
.analyses address diseases other than cancer including heart and lung, diseases.
• Few published risk characterizations are strong hi all components of risk assessment.
Often a rigorous dose-response evaluation has been combined with a very cursory
exposure assessment (or vice versa).
• Pollutants may be ranked for comparison by individual lifetime risk or by risk to
exposed populations.
The two approaches (Figures 2 and 3) give different results and are best analyzed in
combination and with regard to weight-of-evidence evaluations. These summary figures
provide useful insights into the relative significance of indoor air pollutants such as radon,:
asbestos, organic compounds, and environmental tobacco smoke.
93
-------
• A large quantity of data essential to future risk characterizations is emerging from
%
the current work.
While few completed risk characterization studies were found, a large number of studies
containing one or more essential components of risk characterization were identified.
• The review of risk characterization studies and analyses presented in this work must
be considered preliminary.
While considerable effort was expended to identify risk characterization studies (from EPA
and RTI data bases, technical libraries, and computerized library resources), the authors
cannot be sure that the universe of all relevant studies has been captured.
RECOMMENDATIONS
• Continued work should focus on:
— refining the systematic review and presentation of the complex components of risk
assessment.
This should be done so that critical assumptions and data strengths and weaknesses can
be further understood by more people. The Risk Characterization Framework
(presented in Figure 1 and utilized in Figures 4 through 19) provides a start in that
direction.
— combining existing data on exposure and dose-response relationships to make new
risk estimates.
Such estimates should be based on a methodology arid assumptions designed to provide
risk estimates that are as comparable as possible between pollutants. A best currently
available risk characterization for a particular pollutant could be developed by
combining the most scientifically defensible information on hazard identification,
exposure assessment, and dose-response relationships.
94
-------
performing sensitivity analyses.
This would provide an understanding of the importance of input risk parameters to the
resulting risk characterizations. Careful study of the uncertainties in the risk
characterization process will lead to improved risk estimates and greater confidence in
those estimates.
95
-------
REFERENCES
Albert, R. E.; Sellakumar, A. R.; Laskin, S.; Kuschner, M.; Nelson, N.; Snyder. C. A. (1982) Gaseous
formaldehyde and hydrogen chloride induction of nasal cancer in the rat. JNCI J. Natl. Cancer Inst. 68:
597-603.
Anderson, E. L.; Albert, R. E.; McGaughy, R.; Anderson, L.; Bayard, S.; Bayliss, D.; Chen, C.; Chu. M.;
Gibb, H.; Haberman, B.; Hiremath, C.; Singh, D.; Thorslund, T. (1983) Quantitative approaches in use
to assess cancer risk. Risk Anal. 3: 277-295.
Arundel, A.; Irwin, T.; Sterling, T. (1986) Nonsmoker lung cancer risks from tobacco smoke exposure: an
evaluation of Repace and Lowrey's phenomenological model. Environ. Carcinog. Rev. C 4: 93-118.
Dally, K. A.; Hanrahan, L. P.; Woodbury, M. A.; Kanarek, M. S. (1981) Formaldehyde exposure in
nonoccupational environments. Arch. Environ. Health 36: 277-284.
Ellett, W. H.; Nelson, N. S. (1985) Epidemiology and risk assessment: testing models for radon-induced lung
cancer. In: Gammage, R. B.; Kaye, S. V.; Jacobs, V. A., eds. Indoor air and human health: proceedings
of the seventh life sciences symposium; October 1984; Knoxville, TN. Chelsea, MI: Lewis Publishers,
Inc.; pp. 79-107.
Federal Register. (1986) Guidelines for carcinogen risk assessment. F. R. (September 24) 51: 33992-37003.
Fiering, M.; Wilson, R. (1983) Attempts to establish a risk by analogy. Risk Anal. 3: 207-216.
Fong, P. (1982) The hazard of cigarette smoke to nonsmokers. J. Biol. Phys. 10: 65-73.
Garfinkel, L. (1981) Time trends in lung cancer mortality among nonsmokers and a note on passive smoking.
JNCI J. Natl. Cancer lost. 66: 1061-1066.
Garfinkel, L.; Auerbach, O.; Joubert, L. (1985) Involuntary smoking and lung cancer: a case-control study.
JNCI J. Natl. Cancer Ihst. 75: 463-469.
Gold, L. S.; Sawyer, C. B.; Magaw, R.; Backman, G. M.; de Veciana, M.; Levinson, R.; Hooper, N. K.;
Havender, W. R.; Bernstein, L.; Peto, R.; Pike, M. C.; Ames, B. N. (1984) A carcinogenic potency
database of the standardized results of animal bioassays. Environ. Health Perspect. 58: 9-319.
Gold, L. S.; de Veciana, M.; Backman, G. M.; Magaw, R.; Lopipero, P.; Smith, M.; Blumenthal, M.;
Levinson, R.; Bernstein, L.; Ames, B. N. (1986) Chronological supplement to the carcinogenic potency
database: standardized results of animal bioassays published through December 1982. Environ. Health
Perspect. 67: 161-200.
Harley, N. H.; Pasternack, B. S. (1981) A model for predicting lung cancer risks induced by environmental
levels of radon daughters. Health Phys. 40: 307-316.
Hawthorne, A. R.; Gammage, R. B.; Dudney, C. S.; Matthews, T. G.; Erdman, D. J. (1984) Formaldehyde
levels in forty East-Tennessee homes. In: Berglund, B.; Lindvall, T.; Sundell, J., eds. Indoor air:
proceedings of the 3rd international conference on indoor air quality and climate, v. 3, sensory and
hyperreactivity reactions to sick buildings; August; Stockholm, Sweden. Stockholm, Sweden: Swedish
Council for Building Research; pp. 17-22. Available from: NTIS, Springfield, VA; PB85-104206.
96
-------
Hirayama, T. (1981) Non-smoking wives of heavy smokers have a higher risk of lung cancer: a study from
Japan. Br. Med. J. 282: 183-185.
Hogan, M. D.; Hoel, D. G. (1981) Estimated cancer risk associated with occupational asbestos exposure. Risk
Anal. 1: 67-76.
Interagency Regulatory Liason Group. (1981) Integrated risk assessment: formaldehyde. Washington, DC: IRLG
members: Consumer Product Safety Commission, Food and Drug Administration, Food Safety and
Quality Service, Occupational Safety and Health Administration, and U. S. Environmental Protection
Agency; non-IRLG members: U. S. Department of Energy and U. S. Department of Housing and Urban
Development. „
International Agency for Research on Cancer. (1982) IARC monographs on the evaluation of the carcinogenic
~ risk of chemicals to humans: some industrial chemicals and dyestuffs; v. 29. Lyon, France: World Health
Organization.
Lebret, E.; van de Wiel, H. J.; Bos, H. P.; Noij, D.; Boleij, J. S. M. (1984) Volatile hydrocarbons in Dutch
homes. In: Berglund, B.; Lindvall, T.; Sundell, J., eds. Indoor air: proceedings of the 3rd international
conference on indoor air quality and climate, v. 4, chemical characterization and personal exposure;
August; Stockholm, Sweden. Stockholm, Sweden: Swedish Council for Building Research; pp. 169-174.
Available from: NTIS, Springfield, VA; PB85-104214.
Mauskopf, J. A. (1987) Projections of cancer risks attributable to future exposure to asbestos. Risk Anal. 7:
477-486.
McCann, J.; Horn, L.; Girman, J.; Nero, A. V. (1986) Potential risks from exposure to organic compounds in
indoor air. Berkeley, CA: Lawrence Berkeley Laboratory, Indoor Environment Program; report no.
LBL-22473. Available from: NTIS, Springfield, VA; DE89-013403.
National Council on Radiation Protection and Measurements. (1984) Evaluation of occupational and
environmental exposures to radon and radon daughters in the United States: recommendations of the
National Council on Radiation Protection and Measurements. Bethesda, MD: National Council on
Radiation Protection and Measurements; NCRP report no. 78.
National Research Council. (1981) Indoor pollutants. Washington, DC: National Academy Press.
National Research Council. (1986) Environmental tobacco smoke: measuring exposures and assessing health
effects. Washington, DC: National Academy Press.
National Research Council. (1988) Health risks of radon and other internally deposited alpha^emitters: BEIR-IV.
Washington, DC: National Academy Press. ,
Naugle, D. F.; Pierson, T. K.; Mauskopf, J. M. (1990) Indoor air - assessment: development of a risk
characterization framework. Research Triangle Park, NC: U. S. Environmental Protection Agency, Office
of Health and Environmental Assessment, Environmental Criteria and Assessment Office; EPA report no.
EPA/600/8-90-040.
Nazaroff, W. W.; Nero, A. V., Jr. (1988) Radon and its decay products in indoor air. New York, NY: John
Wiley & Sons.
Nicholson, J. (1983) Quantitative risk assessment for asbestos-related cancers. Washington, DC: U. S.
Department of Labor, Occupational Safety and Health Association; contract no. J-9-f-2-0074.
97
-------
Nicholson, W. J.; Perkel, G.; Selikoff, I. J. (1982) Occupational exposure to asbestos: population at risk and
projected mortality—1980-2030. Am. J. Ind. Med. 3: 259-311.
Peffizzari, E. D.; Perritt, K.; HartweU, T. D.; Michael, L. C.; Sparacino, C. M.; Sheldon, L. S.; Whitmore,
R.; Leninger, C.; Zelon, H.; Handy, R. W.; Smith, D. (1987a) The total exposure assessment
methodology (TEAM) study: Elizabeth and Bayonne, New Jersey, Devils Lake, North Dakota and
Greensboro, North Carolina: volume H. Parts 1 and 2. Washington, DC: U. S. Environmental Protection
Agency, Office of Research and Development; EPA report no. EPA-600/6-87-002b. Available from:
NTIS, Springfield, VA; PB88-100078/REB.
Pellizzari, E. D.; Perritt, K.; HartweU, T. D.; Michael, L. C.; Whitmore, R.; Handy, R. W.; Smith, D.; Zelon,
H. (1987b) The total exposure assessment methodology (TEAM) study: selected communities in Northern
and Southern California: volume IE. Washington, DC: U. S. Environmental Protection Agency, Office
of Research and Development; EPA report no. EPA-600/6-87-002c. Available from: NTIS, Springfield
VA; PB88-100086/REB.
Phillips, R. L.; Garfinkel, L.; Kuzma, J. W.; Beeson, W. L.; Lotz, T.; Brin, B. (1980) Mortality among
California Seventh-Day Adventists for selected cancer sites. JNCI J. Natl. Cancer Inst. 65: 1097-1107.
Repace, J. L.; Lowrey, A. H. (1985) A quantitative estimate of nonsmokers' lung cancer risk from passive
smoking. Environ. Int. 11: 3-22.
Repace, J. L.; Lowrey, A. H. (1986) A rebuttal to criticism of the phenomenological model of nonsmokers' lung
cancer risk from passive smoking. J. Environ. Sci. Health: Environ. Carcinog. Rev. C4: 225-235.
Repace, J. L.; Lowrey, A. H. (1987) Predicting the lung cancer risk of domestic passive smoking. Am. Rev
Respir. Dis. 136: 1308.
Research Triangle Institute. (1985) Regulatory impact of controls on asbestos and asbestos products, volume 1,
technical report. Washington, DC: U. S. Environmental Protection Agency, Office of Pesticides and
Toxic Substances.
Robins, J. (1986) Appendix d: risk assessment—exposure to environmental tobacco smoke and lung cancer. In:
National Research Council. Environmental tobacco smoke: measuring exposures and assessing health
effects. Washington, DC: National Academy Press; pp. 294-337.
Russell, M. A. H.; Jarvis, M. J.; West, R. J. (1986) Use of urinary nicotine concentrations to estimate exposure
and mortality from passive smoking in non-smokers. Br. J. Addict. 81: 275-281.
Samet, J. M.; Marbury, M. C.; Spengler, J. D. (1987) Health effects and sources of indoor air pollution. Part I.
Am. Rev. Respir. Dis. 136: 1486-1508.
Seidman, H.; Selikoff, I. J.; Hammond, E. C. (1979) Short-term asbestos work exposure and long-term
observation. Ann. N. Y. Acad. Sci. 330: 61-89.
Seifert, B.; Esdorn, H.; Fischer, M.; Rueden, H.; Wegner, J., eds. (1987) Indoor air '87: proceedings of the 4th
international conference on indoor air quality and climate; August; West Berlin, Federal Republic of
Germany. Berlin, Federal Republic of Germany: Institute for Water, Soil and Air Hygiene.
Selikoff, I. J.; Hammond, E. C.; Seidman, H. (1979) Mortality experience of insulation workers in the United
States and Canada, 1943-1976. Ann. N. Y. Acad. Sci. 330: 91-116.
Surgeon General of the United States. (1979) Smoking and health. Washington, DC: U. S. Department of
Health, Education, and Welfare, Public Health Service; publication no. (PHS) 79-50066.
98
-------
Surgeon General of the United States. (1986) The health consequences of involuntary smoking: a report of the
Surgeon General. Rockville, MD: U. S. Department of Health and Human Services", Office on Smoking
and Health; publication no. DHHS (CDQ87-8398.
Tancrede, M.; Wilson, R.; Zeise, L.; Crouch, E. A. C. (1987) The carcinogenic risk of organic vapors indoors:
a theoretical survey. Atmos. Environ; 21: 2187-2205.
U. S. Environmental Protection Agency. (1985a) Health assessment document for tetrachloroethylene
(perchloroethylene). Final report. Research Triangle Park, NC: Office of Health and Environmental
Assessment, Environmental Criteria and Assessment Office; EPA report no. EPA-600/8-82-005F.
Available from: NTIS, Springfield, VA; PB85-249704.
U. S. Environmental Protection Agency. (1985b) Health assessment document for vinylidene chloride. Final
report. Washington, DC: Office of Health and Environmental Assessment, Environmental Criteria and
Assessment Office; EPA report no. EPA/600/8-83-03 IF. Available from: NTIS, Springfield, VA;
PB86-100641/AS. ->
U. S. Environmental Protection Agency. (1985c) Health assessment document for trichloroethylene. Research
Triangle Park, NC: Office of Health and Environmental Assessment, Environmental Criteria and
Assessment Office; EPA report no. EPA-600/8-82-006F. Available from: NTIS, Springfield, VA;
PB85-249696.
U. S. Environmental Protection Agency. (1985d) Addendum to the health assessment document for
dichloromethane (methylene chloride): updated carcinogenicity assessment of dichloromethane (methylene
chloride). Washington, DC: Office of Health and Environmental Assessment; EPA report no.
EPA-600/8-82-004FF. Available from: NTIS, Springfield, VA; PB86-123742.
U. S. Environmental Protection Agency. (1985e) Health assessment document for 1,2-dichloroethane (ethylene
dichloride). Research Triangle Park, NC: Office of Health and Environmental Assessment, Environmental
Criteria and Assessment Office; EPA report no. EPA/600/8-84/006F. Available from: NTIS, Springfield,
VA; PB86-122702/AS. :
U. S. Environmental Protection Agency. (1985f) Health assessment document for dichloromethane (methylene
chloride). Final report. Washington, DC: Office"of Health and Environmental Assessment, Environmental
Criteria and Assessment Office; EPA report no. EPA/600/8-82-004F. Available from: NTIS, Springfield,
VA; PB85-191559/AS.
U. S. Environmental Protection Agency. (1987a) EPA indoor air quality implementation plan - a report to
Congress under Title IV of the Superfund Amendments and Reauthorization Act of 1986: radon gas and
indoor air quality research. Washington, DC: Office of Air and Radiation; EPA report no. ., :.
EPA-600/8-87-031. Available from: NTIS, Springfield, VA; PB87-210720/REB. ,
U. S. Environmental Protection Agency. (1987b) Radon reference manual.. Washington, DC: Office of Radiation
Programs; EPA report no. EPA-520/1-87-20. Available from: NTIS, Springfield, VA; PB88-196654.
U. S. Environmental Protection Agency. (1987c) The risk assessment guidelines of 1986. Washington, DC:
Office of Health and Environmental Assessment; EPA report no. EPA/600/8-87-045. Available from:
NTIS, Springfield, VA; PB88-123997/AS.
Wallace, L. A. (1986) Cancer risks from organic chemicals in the home. In: Environmental risk management: is
analysis useful? Proceedings of an APCA international specialty conference; April; Chicago, IL.
Pittsburgh, PA: Air Pollution Control Association; publication no. SP-55; pp. 14-24.
*>'•" • . - - ' •
Wells, A. J. (1986) Misclassification as a factor in passive smoking risk. Lancet 8507: 638.
99
-------
Wells, A. J. (1988) An estimate of adult mortality in the United States from passive smoking. Environ. Int. 14:
249-265.
Wigle, D. T.; Collishaw, N. E.; Kirkbride, J.; Mao, Y. (1987) Deaths in Canada from lung cancer due to
involuntary smoking. Can. Med. Assoc. J. 136: 945-951.
Zeise, L.; Wilson, R.; Crouch, E. (1984) Use of acute toxicity to estimate carcinogenic risk. Risk Anal. 4:
187-199.
100
-------
.APPENDIX A
GLOSSARY
Absolute risk projection model: a model which estimates the risk of exposure beyond
the years of observation of a studied population by projecting the average observed
number of excess cancers per unit dose into the future years of risk (U.S. ,
Environmental Protection Agency, 1987).
Activity: the mean number of decays per unit time of a radioactive nuclide. Units:
becquerel (Bq), curie (Gi) (National Research Council, 1988).
Alpha energy: the energy released when an alpha particle emitted during radioactive
decay is halted by collision with a substance (e.g., lung tissue). The amount of energy
depends on the velocity of the alpha particle, which in turn depends on the source of
radioactive decay (e.g., decay of U-238 versus Ra-226) (U.S. Environmental Protection
Agency, 1987).
Alpha particle: two neutrons and two protons bound as a single particle that is emitted
from the nucleus of certain radioactive isotopes in the process of decay or disintegration
(National Research Council, 1988).
Attached radon decay product: a radon decay product that is attached to a particle of
dust or other material in air (U.S. Environmental Protection Agency, 1987).
Bequerel (Bq): the SI unit of radioactivity equal to one disintegration per second (U.S.
Environmental Protection Agency, 1987).
Beta particle: a negatively charged subatomic particle (electron) emitted from a
nucleus during some types of radioactive decay (U.S. Environmental Protection
Agency, 1987).
Cohort: a large homogeneous group of people tested in epidemiological or
socioeconomic studies. EPA's lung cancer estimates are based on calculations for a
cohort of 100,000 people (U.S. Environmental Protection Agency, 1987).
Constant-relative-risk model: a risk model which assumes that, after a certain time,
the ratio of the risk at a specific dose to the risk in the absence of the dose does not
change with time (National Research Council, 1988).
Cumulative working level month (CWLM): a unit of cumulative radon exposure.
the sum of lifetime exposure to radon working levels in total working level months
(U.S. Environmental Protection Agency, 1987).
101
-------
Curie (Ci): a unit of radioactivity, defined as that quantity of any radioactive nuclide
which spontaneously undergoes 3.7 x 1010 disintegrations per second. One gram of
radium-226 has an activity of one curie (U.S. Environmental Protection Agency, 1987).
Decay series: the consecutive members of a family of radioactive isotopes formed by
sequential radioactive decay. A complete series commences with a long-lived parent
such as U-238 and ends with a stable element such as Pb-206 (U.S. Environmental
Protection Agency, 1987).
Electron volt (eV): a unit of energy = 1.6 x 10"12 ergs = 1.6 x 10~19 J; 1 eV is
equivalent to the energy gained by an electron in passing through a potential difference
of 1 V; 1 KeV = 1,000 eV; 1 MeV = 1,000,000 eV (National Research Council,
1988).
Equilibrium factor: an adjustment used in converting from picocuries per liter (pCi/L)
to working-level concentration (WL), which takes into account the possible absence of
radioactive equilibrium between radon and its decay products (U.S. Environmental
Protection Agency, 1987).
Equilibrium fraction: ratio of the concentration of radon decay products to
concentration of radon in the same sample of air at radioactive equilibrium. Typical
measured values range from 0.3 to 0.7, with an average of about 0.5 (U.S.
Environmental Protection Agency, 1987).
Gamma ray: short-wavelength electromagnetic radiation of nuclear origin (range of
energy, 10 keV to 9 MeV) (National Research Council, 1988).
Ionizing radiation: subatomic particles or photons that have sufficient energy to
produce ionization directly in their passage through a substance (U.S. Environmental
Protection Agency, 1987).
Latent period: the minimum period of time between exposure and expression of the
disease. After exposure to a dose of radiation, there is delay of several years (the latent
period) before any cancers are seen (National Research Council, 1988).
Lifetime risk: the lifetime probability of contracting a specific disease (National
Research Council, 1988).
Lifetime risk ratio: the ratio of the lifetime risk (Rg) of an exposed person to the
lifetime risk of an unexposed person (R0). This number minus 1 is the proportional
increased risk associated with exposure (Re/R0 - 1) (National Research Council, 1988).
Linear dose model: this model postulates that the excess risk is linearly proportional to
the dose (National Research Council, 1988).
102
-------
Linear energy transfer (LET): average amount of energy lost per unit track length
(National Research Council, 1988).
Kcocuries per'liter (PCi/L): a unit of measurement of activity concentration. One
picocurie per liter is equal to 10"12 curies per liter (U.S. Environmental Protection
Agency, 1987).
Progeny: the decay products resulting after a series of radioactive decays. Progeny can
also be radioactive, and the chain continues until a stable nuclide is formed. Radon
progeny of primary concern to the public health community are 218-Pp, 214-Pb,
214-Bi, and 214-Po (National Research Council, 1988).
Radiation dose: the total amount of ionizing radiation absorbed by material pr tissues,
in the sense of absorbed dose (expressed in rads), exposure (expressed in roentgens), or
dose equivalent (expressed in rems) (U.S. Environmental Protection Agency, 1987).
Radioactive equilibrium: a state in which the rate of formation of atoms by decay of a
parent radioactive isotope is equal to its rate of disintegration by radioactive decay, so
that the activity of the parent and the decay product assume a constant proportion. This
proportion is equal to one, if the parent has only one mode of radioactive decay.
Because radon decay products tend to attach readily to surfaces, equilibrium between
radon and its decay products is seldom reached/Using the average value of the
equilibrium fraction (0.5), a ratio of about 200 pCi/L of radon to 1 WL of radon decay
products is fairly typical for residences (U.S. Environmental Protection Agency, 1987).
Radioactive secular equilibrium: ideal steady state situation in which radon decay
products would be formed and would decay at the same rate. Under such ideal
conditions, the ratio of concentrations (or activities) of radon to decay products would
be such that 100 pCi/L of radon and 1 WL of decay products would be present in the
same sample of air. Secular equilibrium is never achieved because some of each of the
decay products is removed from the, air (due to attachment to walls, floors, etc.) before
undergoing decay (U.S. Environmental Protection Agency, 1987; National Research
Council, 1988). ? -..-.,;.•
Relative risk: the ratio of the rate of disease in exposed to unexposed populations
(U.S. Environmental Protection Agency, 1987).
Relative risk projection model: a model which estimates the risk of exposure beyond
the years of observation of a study's population by projecting the currently observed
percentage increase in cancer risk per unit dose into the future years (U.S.
Environmental Protection Agency, 1987).
103
-------
Risk coefficient: the increase in the annual incidence or mortality rate per unit dose:
(1) absolute risk coefficient is the observed, minus the expected number of cases per
person-year at risk for a unit dose; (2) the relative-risk coefficient is the fractional
increase in the baseline incidence or mortality rate for a unit dose (National Research
Council, 1988).
Risk estimate: the number of cases (or deaths) that are projected to occur in a
specified exposed population per unit dose for a defined exposure regime and expression
period—for radon, the number of cases per person cumulative working-lever month
(National Research Council, 1988). ;
Synergistic model: a form of cancer causality whereby two Or more carcinogens (for
example, radon and tobacco smoke) act synergistically to cause cancer with a greater
probability than if each were acting alone (U.S. Environmental Protection Agency,
1987).
Threshold hypothesis: the assumption that no radiation injury occurs below a specified
dose (National Research Council, 1988).
Time-since-exposure (TSE) model: a model in which the relative or absolute risk is
not constant but varies with the time after exposure (National Research Council, 1988).
Unattached radon decay product: a radon decay product that is not electrostatically
attached to dust or particles in the air. Capable of attaching to lung tissue if inhaled
(U.S. Environmental Protection Agency, 1987).
Working level (WL): any combination of short-lived radon progeny in 1 liter of air
that will result in the ultimate emission 1.3 X 105 MeV of potential alpha energy. This
number was chosen because it is approximately the alpha energy released from the
decay of progeny in equilibrium with 100 picocuries of 222Ra (National Research
Council, 1988).
Working-level month (WLM): exposure resulting from inhalation of air with a
concentration of 1 working level of radon progeny for 170 working hours (National'
Research Council, 1988).
104
-------
APPENDIXB
Concentration Units and Conversion Factors for Radon and Radon Progeny
The quantity of radioactive species present in a given volume of air is typically specified
by activity rather than by mass. The standard international (SI) unit for activity is the
Becquerel (Bq), which is defined as 1 radioactive decay per second,; The Curie (Ci) and
picocurie (pCi) represent 3.7 X 1010 Bq and 0.037 Bq, respectively, and are often used in
the U.S. to express quantity of radon. . ; ,
Concentration of radon in air is typically expressed in Becquerels per cubic meter
(Bq/m3) or in picocuries per liter (pCi/L). Concentration of radon progeny (218Po, 214Pb,
214Bi, and 214Po) in air is typically expressed in working levels (WL). A working level is
any combination of radon progeny which has the potential to release 1.3 x iO mega-electron
volts of alpha energy per liter of air (MeV/L). Concentrations of radon and radon progeny in
indoor air are, of course, related, but the relationship is dependent upon a number of factors.
Such factors include the rate at which progeny are removed from the air (by attachment to
aerosols, dust particles, walls, ceilings, and floors). In indoor environments,'one working
level of progeny corresponds to about 150-300 pCi/L of radon, with 200 pCi/L representing
about the average conversion factor for radon in homes. ,
Radon itself, because of its behavior as essentially a chemically inert gas .with no
affinity for attachment to other materials (including aerosols, dust particles, and lung tissue),
does not appear to pose a serious threat to human health. Radon progeny, however, attach
: •' • * - ' • '"•-"'- - - "
readily to other materials, including lung tissues. Because of this, epidemiological studies are
usually concerned with exposure to radon progeny rather than with exposure to radon. '
Exposure to radon progeny is typically measured in working-level months (WLMs), where
one WLM is equivalent to 170 hours of exposure to one WL of progeny.
105
-------
TABLE OF CONVERSIONS
Description
Conversions
Activity: mean number of radioactive decays per unit
time. Typically expressed in Becquerels (Bq), Curies
(Ci), or picocuries (pCi).
1 Bq = 1 decay/sec
1 Ci = 3.7 X 1010 Bq
1 pCi = 0.037 Bq
Concentration: activity per unit volume. Typically
expressed in Becquerels per cubic meter (Bq/m3) or
picocuries per liter (pCi/L) for radon and in working
levels (WL) for radon progeny. A WL is defined in
terms of potential alpha energy (from radon progeny).
per liter of air. Some assumptions regarding relative
concentrations of radon and progeny are required for
conversions between pCi/L (or Bq/m3) and WL.
Progeny concentration may also be expressed in
energy units of mega-electron volts per liter (MeV/L)
or joules per cubic meter (J/m3) with proper
conversion of units.
1 pCi/L = 37 Bq/m3
1 WL = 200 pCi/L
(typically, in a home)
1 WL = 3.7 X 103Bq/m3
(at radioactive equil.)
1 WL = 1.3 X 105 MeV/L
1 WL = 2.1 x 10-5J/m3
Exposure: exposure concentration multiplied by the
length of time exposed. Exposure to radon progeny is
typically expressed in working-level months (WLM).
A WLM is equivalent to exposure to a concentration
of 1 WL for 170 hours.
1 WLM = 170 WL-h
1 WLM = 3.5 X 10"3
J-h/m3
106
-------