RiskAnalysis, Vol. 11, No. I , 1991
Cancer Fatalities from Waterborne Radon (Rn-222) Douglas J. Crawford-Brown' Received March 7, 1990; revised July 27, 1990
A model of the biokinetics of radon in the human body following ingestion is developed from existing data. Calculations of the probability of cancer fatality from use of radon-laden water in the home then are presented. The pathways of emanation and ingestion are examined and shown to lead to roughly equal risks. The probability of fatal cancer resulting from lifetime use of water at a radon concentration of 1 pCi/L is shown to be 1 x with a reasonable range between The allowed concentration consistent with an excess risk of then is 2 x lo-' and 5 x approximately 100 pCi/L, which is exceeded in a significant fraction of U.S. water supplies. The lifetime number of premature deaths due to waterborne radon in the U.S. is estimated to lie between 5000 and 125,000, with a best estimate of 25,000. KEY WORDS: Radon; cancer; water; biokinetics.
probability of cancer fatality resulting from use of waterborne radon in U.S. homes. The model and resulting calculations will be used here to propose regulatory limits consistent with selected levels of risk. As will be demonstrated, the fatalities imposed by direct ingestion of radon are predicted to be of roughly equal importance to the regulator. Since the fatalities resulting from emanated radon have been reviewed p r e v i ~ u s l y , ( ~these ,~) will simply be summarized here. The models developed for the case of ingestion are, however, previously unpublished and will be developed in detail. Throughout the initial sections of the paper, the risks associated with a unit concentration (1pCi/L) will be reported. The fatalities in the U.S. population at the current EPA estimate of the population-averaged radon concentration then will be developed in the final discussion.
1. INTRODUCTION Throughout the 1980s, there has been a rise in concern over the health effects of environmental radon. While interest has been directed primarily to exposures resulting from airborne radon (or progeny), substantial efforts also have been mounted to quantify the effects from radon contained in water supplies. This latter effort arises from calculations indicating the large contribution of radon to the total estimated risk from all forms of waterborne pollutants, both natural and manmade.(') It has been estimated, for example, that waterborne radon will cause between 1000 and 30,000 premature deaths from cancer in a cohort the size of the current U.S. population.(2) The Office of Drinking Water in the U.S. Environmental Protection Agency has, therefore, undertaken to develop regulatory limits on the allowed concentration of radon in U.S. water supplies. The following sections present a model for the biokinetics of radon in the human body following ingestion, and detail calculations of the expected excess lifetime
2. EMANATED RADON
There are two primary routes of exposure to radon entering a structure through water supplies. The first is emanation of radon into the air, followed by decay to charged radioactive progeny. These progeny then are inhaled, depositing in the passageways of the lung and
Department of Environmental Sciences and Engineering, University of North Carolina at Chapel Hill, Chapel Hill, North Carolina 275997400.
135
0272-4332/91~300-0135106.50/1 OD 1991 Socicty for Risk Analysis
Crawford-Brown
136
irradiating adjacent tissue. The primary concern here is with the induction of lung cancer within the bronchial epithelium.(4)The second route of exposure is the direct ingestion of the radon into the stomach, followed by translocation and/or decay. Stomach cancer is the primary effect of concern here, although other organs will be shown to contribute significantly to the number of fatalities. The present section explores the health effects from emanation. Other sections focus on direct ingestion. As described in previous reports,(2J) calculations of the probability of fatality from lung cancer following emanation require specification of several factors. The first is the concentration, C, (in pCi/L), of radon in the water supply, taken here to be 1 pCi/L. The second factor is the equilibrium concentration ratio for air and water, given as CJC,. The third is the equilibrium factor for radon progeny, f, which describes the ratio of the total potential alpha energy from progeny in air to that obtained under conditions of complete equilibrium. The product of C,, C,/C,, f, and 0.01 (the conversion factor from pCi/L to working levels) then equals the working level concentration of progeny in air, WL. Multiplying this latter concentration by T (the number of hours spent in this environment throughout a year), and dividing by 170 (the number of hours per “working month”), yields the annual exposure in units of working level months, WLM. Values for CJC, (dimensionsless if both concentrations are in pCi/L) have been measured and have been calculated using models of the attachment of progeny to particulate and room surfaces in homes. Two recent compendia report values ranging between 0.17 x and 3.5 x 10-4(5)and 0.23 x A and 1.87 x mean value of approximately 1x for U.S. homes is adopted here. Measured values o f f in dwellings range between 0.25 and 0.8,(7-9) with various authors employing mean values of between 0.5 and 0.7 for purposes of risk estimati~n.(~,’~) A mean value of 0.5 is adopted here. The value of T divided by 170 is set equal to 52 as a matter of regulatory policy (implying continuous exposure). As a result, the conversion factor from the concentration of radon in water (pCi/L) to the annual cumulative exposure (in WZM) is 1x x 0.5 x 0.01 x 52 or 2.6 x WLM per pCi/L. Values for other exposure intervals, T, may be calculated through scaling proportionately. Risk factors for the induction of lung cancer are taken from the recent BEIR IV report,(4)which estimates 350 excess cancers/million personslWLM. This value was obtained from the fitting of a linear dose-response function to the available epidemiological data on mining ex-
posures. The risk factor is corrected for the higher unattached fraction in homes (7%) relative to mines (4%), and for the difference in breathing patterns between the two exposed groups. As estimated by the BEIR comrnittee,c4) the excess lifetime probability of fatal lung cancer assuming continuous exposure at 1 WLMlyear in assuming a latency of 20 years. the home is 1.7 x The conversion factor from radon concentration in water (pCi/L) to excess lifetime probability of fatal lung cancer x (1.7 X or approximately then is (2.6 x 4 x lo-’. For comparison, Cross et ~ 1 . ( estimate ~) between 3 and 8 x lo-’. Previous risk analyses for waterborne radon have relied on one of two approaches. In the first,c2) it was assumed that the risk from ingested radon is insignificant compared to emanation. In the limited data from earlier studies on whole-body retention of radon were used to estimate the risk following ingestion. Since these earlier studies did not measure concentrations in individual organs, however, it was necessary to combine all organs into a single compartment. The exception was the stomach, for which partial data were available. Through use of the second approach, Cross et estimated a lifetime cancer risk (emanation plus ingestion) per pCi/L, with the risk from of between 4 and 8 x ingestion being an order of magnitude below that from emanation.
3. INGESTED RADON More accurate assessments of risk may be obtained through explicit modeling of doses delivered to the separate organs and tissues of the body. Ingested radon enters the body through the stomach, followed by rapid movement to the other body organs. Measurements of the retention of radon have been reported by Hursh et al.,(”) von Dobeln and Lindell,(12)Anderson and Nilsand Suomela and Kahlos.(14)These measurements relied primarily on the detection of radon progeny in the whole body, yielding little information on organspecific doses. These data do, however, indicate that radon is removed from the body with a primary halftime of between 30 and 50 min, with a smaller component characterized by a half-time on the order of several hours. Hursh er ul.(I1) also demonstrated that the radon is removed from the body primarily through exhalation via the lung. Nussbaum and Hursh(I5)report the relative concentrations of radon in rat tissues at equilibrium. They note values of 1 (omental fat), 0.08 (venous blood), 0.06 (kidneys), 0.05 (heart), 0.04 (testis), and 0.3 (muscle).
Cancer Fatalities from Waterborne Radon
137
These data indicate the ability of fat to sequester radon, suggesting that the half-time for removal from the whole body is dominated by fat at long times after ingestion. Ellis et aZ.(16) measured the kinetics of inhaled krypton in human subjects. Their measured half-time for wholebody retention correlated highly with total body fat, supporting the concept of body fat as a storage site. Kirk et aZ.(17)studied the solubility of krypton in the tissues of guinea pigs and found weighted partition coefficients (equal approximately to the relative concentrations at equilibrium) of 0.421 in omental fat, 0.405 in substaneous fat, 0.259 in thymous, and 0.134 in bone marrow. Tobias et aZ.(18) and Winston and Wilson(19)each have reported measurements of the absorption of gases from the GI tract. Their results suggest little movement from the stomach to the venous blood. The major site of absorption was the small intestine, with the gas moving from the stomach to the small intestine and, finally, to the bloodstream. The clearance of radon from the bloodstream is rapid compared to the half-time of removal from the whole body. Lindellc20)and Underwood and Diaz@l)report blood clearance half-times on the order of minutes. Tobias el a1.(18)also confirmed that radon contained in blood and entering the lung is removed to the lung airspace in a single pass, supporting the contention that the lung is the primay route of removal from the body. As a result, it may be assumed that radon is transferred rapidly from the blood to the systemic organs and tissues. This simplifies calculations since decay in the bloodstream will be negligible.
t
The most complete set of data on radon biokinetics has been reported by Correia et U Z . ( ~ ~ ) In their study, xenon was used as a chemical analogue for radon. The concentration of xenon in each organ or tissue was measured through external imaging following ingestion of approximately 2 mCi of xenon in 100 cc of saline solution. These data then were converted to an estimate of the equivalent radon concentration through use of the relative solubility of radon and xenon in human tissue. The means of measurement data on 22 individuals are provided in Figs. 1-7. Organs imaged include fat, stomach, small intestine, ascending colon, descending colon, liver/portal blood, and whole-field. The points for the general tissue curve were calculated for the present study by subtracting the radon contents of the stomach, small intestine, ascending colon, descending colon, liver, lung, and torso blood from the whole-field content. The results in Figs. 1-7 have been normalized to yield a total activity of 2 mCi in the stomach at the start of measurements. Error bars on the measurements were not reported by the original investigators. In the same study, Correia et aZ.(22)performed measurements of the half-time for removal of xenon from body organs following inhalation. These results are presented in Table I and will be employed in the biokinetics model developed here. It will be noted that the retention half-time in fat is elevated greatly above that for other organs. Figure 8 displays the biokinetics model developed in the present study. The compartment termed general tissue includes the mass of intestinal and stomach epi-
00001
OOooOl(!40 i 0 I i O I$O O 2;
O 2;
2;)O 340 Time (minutes)
3k 4hO 4AO O4;
O 5;
&O 6 h O
Fig. 1. The data (0),fitted exponential functions (--) and biokinetics model predictions (-) for the general body tissue compartment. The total intake of radon is 2 millicuries into the stomach at I = 0. The concentration then is in units of millicuriedcc. Figures 1-7 employ the same key.
Crawford-Brown
138
01 -
0001 -
001 -
Fig. 2. The same key as Fig. 1. The organ is the stomach.
Fig. 5. The same key as Fig. 1. The organ is the descending colon.
i0 i0 O ;l
I$O 2&l
40 2AO 3;O
3LO 4b0 4b0 4AO O 5;
560 6b0
Time (minutes)
oooo10
Fig. 6. The same key as Fig. 1. The organ is the livedportal blood. 40
80 120 I60 2 0 24C 280 320 360 400 440 480 520 560 600 llm(mmulms)
Fig. 3. The same key as Fig. 1. The organ is the small intestine.
Fig. 4. The same key as Fig. 1. The organ is the ascending colon.
thelium, into which radon is recirculated after entering the bloodstream. The assumed masses for the various compartments may be found in Table I1 for adults and are taken from the ICRP Reference Man report.(23) Rate constants for translocation were obtained through simultaneous fitting to the concentrations in Figs. 1-7, with the exception of pathways F,G,I,K, and M. These latter constants have been assigned on the basis of Table I, which was generated from the inhalation/ organ removal studies. All rate constants are displayed in Table 111. The model fits may be found as solid lines in Figs. 1-7 and were obtained by assuming an intake of 2 mCi of radon. The original data in Figs. 1-7 also were approximated by multi-exponential functions describing the concentration in each organ as a function of time. These concentration functions are presented in Table IV and have been normalized to an intake of 1 pCi of radon.
Cancer Fatalities from Waterborne Radon
139
‘1 \
01-‘,
0‘ ,
5
0,
5
0001 -
lim(minutrn)
Fig. 7. The same key as Fig. 1. The organ is the whole body (0)or fat
(a).Only the biokinetics model predictions for fat are displayed (-).
Table I. Summary of Xenon Removal Half-Times for Organs Following Inhalation
Organs
Half-times (Min)
Liver, spleen, kidneys Stomach, small intestine Lower and upper intestine Fat tissue
1.3 1.6 4.2 362
A
The doses calculated using these simplified approximations agree with doses calculated using the biokinetics model, with the values agreeing to within 10% for all organs or tissues except fat (which has no significance in risk estimates). The resulting functions are also displayed in Figs. 1-7 as dashed lines, again assuming an initial intake of 2 mCi of radon.
F
E V
B
+. N
I4
V
A.C.
FAT
4. CALCULATIONS OF INGESTION DOSES
The dose to each organ or tissue following ingestion of 1 pCi was obtained through integration of the model predictions as a function of time. The integral organ burden, B,(pCi-seclg), is equal to I
Bi =
Ci(t)dt
(1)
0
where Ci(t)is the concentration (pCi/g) in organ i at time
Fig. 8. The biokinetics model for ingested radon developed in the present study. Rate constant$ are provided in Table 111.
Crawford-Brown
140 Table 11. Compartmental Masses and Volumes for the Biokinetics Model Displayed in Fig. 8 -
Organ Stomach w/contents Small intestine Upper large intestine Lower large intestine General tissuea Blood (plasma) Blood (total) Liver Lung airspace volume Lung minute volume Fat Muscle
Assumed mass or volume
400 g 1040 g
430 g 295 g 32,035 g 3100 g 5500 g 1800 g 4300 cc 7500 cc/min 13,500 g 28.000 e
"Taken to be the difference between the total body mass (without bone) and the sum of the specific organ masses listed in this table. The epithelium of the stomach and intestines is included in the general tissue compartment mass.
Table 111. A Summary of Transfer Rate Constants for the Biokinetics Model Displayed in Fig. 8 Pathway A
Rate constant (inverse min) 0.04 0.005 0.005
B C D E F G
1.5 0.16 0.4 0.4
H
0.13 0.0073 0.057
I J K L M N P
0.0019
0.083 0.4 1.8 0.49 1.74
Q
t. Assuming a tissue density of 1 g/cc, the dose, Di (in
rads), then is:
Di= Bi x 0,037 x 19.2 x 1.6 x
(2)
In this expression, 19.2 is the total alpha energy (MeV) released by decays of Rn-222, Po-218, and Po-214. It is assumed that the progeny decay at the site of the radon is a conversion factor decay. The constant 1.6 x from MeVIg to rad, and 0.037 is a conversion factor from pCi to disintegrations per sec.
The dose to the walls of the stomach and intestines require a slight modification due to the concentration profile between the lumen and the bloodstream. This concentration equals that in the lumen contents at the surface of the epithelium and decreases to very small values (essentially zero) at the depth of the capillaries. The epithelial layer varies in thickness due to undulations, but generally averages between 30 and 100 mic r o n ~ . ( The ~ ~ ) radiosensitive crypt cells tend to lie most deeply within this layer. In the following calculations, it will be assumed that the radiosensitive cells lie at a depth of 30 microns into a 40 micron epithelial layer. The dosimetry for this arrangement has been reported by Crawford-Brown and Shyr(2s)and will not be reviewed here. The net result of these dosimetric considerations is that the dose calculated using Eq. (2) must be divided by a factor of 3. This reduction does not apply to recirculated radon contained within the general tissue compartment, which distributes uniformly throughout the epithelial layer. Unfortunately, the data of Correia el al.(22)do not permit a separation of the radon in lung tissue from that contained in the air of the lung. The researchers do report, however, that the concentration in lung tissue is approximately equal to that in the liver. For purposes of calculation here, the lung tissue will be assumed in equilibrium with the concentration in the liver. As discussed elsewhere,(28)the dose to the lung includes a component from radon decaying in the airspaces prior to exhalation. This radon (and its progeny) will irradiate the lung during exhalation. It is assumed that essentially all of the ingested radon is removed from the body via the lungs. The details of the calculation required to estimate the dose delivered by exhaled radon are given in Ref. 26 and will not be repeated here. It is assumed that the radon is mixed uniformly in the alveolar air and that the progeny migrate rapidly to the walls of the region in which the radon decays. The doses to cells in the alveolar region from in situ decay may be found in Column 1 of Table V, and the alveolar doses from the exhalation of radon may be found in Column 3 of that table. Similar calculations for various generations of the bronchial tree may be found in Column 4, but will not be used here due to their insignificance in risk estimates. A summary of the organ doses calculated in the present report may be found in Table VI. The dependence of doses on age are obtained through scaling in inverse proportion to body mass. The choice to scale in this manner is driven by the finding of Suomela and Kah10s"~)that the half-time for removal of radon is identical for humans and rats [see also Andree~(*~)]. Since
Cancer Fatalities from Waterborne Radon
141
Table IV. A Summary of Approximate Concentration Functions for the Body Organs Following Ingestion of 1 pCi of Radon-All Concentrations Are in Units of pCi/cc ~~~
Stomach Small intestine Ascending colon Descending colon Liver Muscle Fat Whole body General tissue
C ( f )= C(f) = C(t) = C(f)= C(r) = C ( f )= C(t) =
C(f) = C(t) =
~
~
0.002[0.998 exp( -0.029t) + 0.002 exp( -0.0026t)l 0.0002[0.9 exp(-0.02) + 0.1 exp(-0,0045t)J (1 - exp(-O.O99f)} 0.0001[0.85 exp( -0.015t) + 0.15 exp( -0.006t)J (1 - exp( -0.0991)) O.O0003[exp( -O.O048r)] (1 - exp( -0.099f)) 0.00003[0.87 exp( -0.022!) + 0.13 exp( -0.0039f)J 0.00001[exp( -0.0029t)l (1 - exp(0.231)) 0.00001[exp( -0.005t)l (1 - exp(0.099t)) 0.00001[0.92 exp( -0.022) 0.08 exp( -0.0035t)J 0.0000051[exp( - 0.005t)J (1 - exp(0.099t))
+
Table V. A Comparison of Alveolar and Basal Cell Doses (rads) from In Situ and Exhalation Decays Following Ingestion of 1.0 pCi of Radon in Water
Age (years) 0 2 5
10 Adult
P
2b
3‘
4d
2(E-8) 7(E-9) 4(E-9) 2(E-9) 1(E-9)
2(E-9) 7(E-10) 3(E-10) 2(E-10) 1(E-10)
4(E-9) 7(E-10) 1(E-9) 7(E-10) 3(E-10)
4(E-9) 1(E-9) 7(E-10) 4(E-10) 2(E-10)
5’ 3(E-9) 7(E-10) 7(E-10) 4(E-10) 2(E-10)
6f 1(E-9) 3(E-10) 4(E-10) 3(E-10) 1(E-10)
nAlveolar region doses, and all basal cell doses, from in situ decay. The concentration in lung cells assumed equal to liver. bAlveolar region doses, and all basal cell doses, from in situ decay. The concentration in lung cells assumed equal to general tissue. ‘Alveolar region doses from exhalation decays. dDosesto the basal cells of generation 5 from exhalation decays. ‘Doses to the basal cells of generation 10 from exhalation decays. Boses to the basal cells of generation 16 from exhalation decays.
radon is removed primarily through the lung, this equivalence of half-times may be attributed to the fact that the ratio of lung-surface-area-to-body-massis similar in humans and rats. Since this ratio also is roughly invariant with age in humans,(26)it is suggested that the wholebody removal half-time is invariant with age. Doses resulting from a unit intake of radon will, therefore, be inversely proportional to body mass. The exception is the dose to the stomach and intestinal walls. Recirculated radon in the epithelium should follow the pattern of age-dependent dose noted above. The dose from radon transported from the lumen to the capillarieswill, however, depend only on the transit time across the epithelium. There is no evidence that this transit time varies with age. As a result, the age-dependence will apply only to the recirculated component of the dose to stomach and intestinal epithelium.
5. RISK FROM INGESTED RADON
For the purpose of estimating risk, the methodology of the USEPA(28)will be adopted. This methodology utilizes risk factors developed by the National Academy of Sciences(29)and UNSCEAR(30)and is based on lifetable analyses of competing risks. Their lifetime risk factor for fatalities from all cancers due to low LET, whole body, irradiation is given as 392 per million person-rad. This risk factor would be doubled if the BEIR V 3 1 ) risk estimates are employed. The same report(28)gives site-specific incidence risk coefficients for a variety of organs. From Table 6.6 of that report, the age-adjusted ratio of the incidence of stomach cancer to that of all sites is 0.12, while that for intestinal cancer, liver, and lung is 0.06,0.13, and 0.18, respectively. Multiplying these values by the risk factor of 392 per million person-rad yields a lifetime risk factor of 46 per million person-rad for stomach cancer, 23 per million person-rad for the total intestinal tract, 50 per million person-rad for liver, and 70 per million personrad for lung. It is assumed here that the three intestinal sections possess identical radiosensitivities, indicating that each should be assigned a risk factor of approximately 8 per million person-rad. If a quality factor of 20 is adopted,(32)the risk factors for the stomach and separate intestinal sections are 920 per million person-rad and 150 per million person-rad, respectively. For the liver and lung, the risk coefficients are 1000 and 1400 per million person-rad, respectively. These values will be assumed here. These organs dominate the risk from ingested radon. The USEPA assigns a value of 2 Wday to the daily rate of ingestion of water; of this, between 200 and 1200 ml are ingested untreated.c3) A value of 660 ml is adopted here, consistent with the choice by the EPA. This implies an annual rate of ingestion equal to 220 L. At a
Crawford-Brown
142
Table VI. Summary of Dose per Unit Ingestion (rad/pCi), Assuming 100% of the Ingested Radon Is Taken into the Stomach (the Results of Column 5 Are Used in This Report)
Y
Organ
1" Age 0
2" Age 2
3" Age 5
4" Age 10
5" Adult
Ref. 11
7= Ref. 12
8' Ref, 13
Ref. 14
Stomach Liver Small intestine Ascending colon Descending colon General tissue Lung alveoli Whole body
2(E-8) 2(E-8) 3(E-8) 4(E-8) 6(E-8) 9(E-9) 3(E-8) NA
2(E-8) 7(E-9) 1(E-8) 1(E-8) 2(E-8) 3(E-9) 8(E-9) NA
2(E-8) 4(E-9) 9(E-9) 1(E-8) 1(E-8) 2(E-9) 5(E-9) NA
2(E-8) 2(E-9) 6(E-9) 7(E-9) 6(E-9) 1(E-9) 3(E-9) NA
2(E-8) 1(E-9) 5(E-9) 7(E-9) 4(E-9) 7(E-10) 2(E-9) NA
2(E-8) 7(E-10) NA NA NA NA 4(E-10) NA
NAd NA NA NA NA NA NA 2(E-10)
NA NA NA NA NA NA NA 7(E-10)
4(E-8) NA NA NA NA NA NA NA
6b
"Assumes all progeny decay at the site of radon decay. For the dose to the alveolar cells, doses in columns 1 and 3 are summed. The concentration of radon in lung tissue is assumed equal to that of the liver. For the stomach and intestines (short-term retention components only), radiosensitive cells are located at 30 microns in a 40 micron wall. bSee Ref. 11. Stomach doses include only decays of radon and Po-218. Concentration of radon in the wall is the same as in the lumen. 'Uses measurements of radon progeny. dNA, Not availablc.
concentration of 1 pCi/L, the annual rate of intake for radon then will be 220 pCi. From Table VI, the dose to the stomach from ingestion of 1 pCi (incorporating the effect of age) is approximately 2 x rad, while that for the small intestine, ascending colon, descending colon, liver, and lung are 7 x 10 x 6x 2x and 2 x respectively. The lifetime probability of fatal stomach cancer from ingestion of 220 pCi/ year of radon then is 2.4 x The lifetime (70-year) probability of fatal cancer of the small intestine, ascending colon and descending colon, liver, and lung are 2.1 x 2x 1x 3x and 7 x lo-&, respectively. The sum of these six probabilities then is 4 x lo-'. The Office of Radiation Programs of the USEPA recommends a quality factor (or Relative Biological Effectiveness) of 8 for the alpha emissions from ingested radon. If that quality is employed here, the total lifetime probability of fatal cancer from ingested radon is 1.6 x 10-7. 6. DISCUSSION As mentioned previously, Cross et ~ 1 . (estimate ~) a risk coefficient of between 4 and 8 x per pCi/L, 90% of which is due to emanation. From the present analysis, water containing radon at a concentration of 1 pCi/L poses an excess lifetime probability of cancer fatality equal to approximately 4 x for emanated radon. The excess lifetime probability of fatal cancer from at the same concentration. ingested radon is 4 x
These results indicate that the risk of fatal cancer from ingestion is approximately equal to that from emanation. Since the risk from ingested radon is dominated by stomach cancers, any future changes in the risk factor for stomach cancer will heavily affect this result. The current population-averaged concentration of radon in U.S. public drinking water supplies is estimated to be 150 pCi/L.(*) Multiplying this concentration by the above risk factors yields an excess lifetime probability of fatal cancer equal to 6 x for emanated radon and 6X for ingested radon. The total excess lifetime conprobability of fatal cancer then is equal to 1x sidering both pathways of exposure. For a population of approximately 250 million persons, the number of excess fatalities due to cancer over a lifetime is approximately 25,000. The uncertainty due to the risk from emanation is on the order of a factor of 5.(2)Assuming a similar uncertainty for the composite risk, the number of excess fatalities over a lifetime probability lies between 5000 and 125,000. These results may now be used to estimate the concentration of radon corresponding to various excess lifetime probabilities of fatal cancer. For a probability of the corresponding concentration would be 1 pCi/ L. For a probability of the corresponding concentration would be 10 p C i L Finally, for a probability of the corresponding concentration would be 100 pCi/ L. The latter concentration falls within the range of values being considered by the EPA in setting a regulation. These concentrations may be compared with the cumulative frequency of water supplies containing radon (see
Cancer Fatalities from Waterborne Radon
curve C of Fig. 1 in Ref. 2). A regulatory standard of 1 pCi/L would require remediation in essentially all water supplies. A standard of 10 pCi/L would require remediation of the water used by approximately 90% of the population. A standard of 100 pCi/L would require remediation of the water used by approximately 30% of the population. Clearly, for any regulation aimed at an excess lifetime probability of or less, the number of supplies requiring remediation will be large. ACKNOWLEDGMENTS
The research reported here was funded through a grant from the USEPA, but does not represent the views of the agency. Gratitude is extended to Greg Helms of the EPA for administering the grant and to Neil Nelson of the EPA for extensive reviews of the research. The author also wishes to thank Ms. Delores E. Plummer for her excellent assistance in preparing the document. REFERENCES 1. C. Cothern and W. Marcus, “Estimating Risk for Carcinogenic Environmental Contaminants and Its Impact on Regulatory Decision Making,” Regulatory Toxicology and Pharmacology 4,265274 (1984). 2. D. Crawford-Brown and C. Cothern, “A Bayesian Analysis or Scientific Judgement of Uncertainties in Estimating Risk Due to Rn-222 in U S . Public Drinking Water Supplies,” Health Physics 53, 11-21 (1987). 3. F. Cross, N. Harley, and W. Hofmann, “Health Effects and Risk from Rn-222 in Drinking Water,” Health Physics 48, 649-670 (1985). 4. National Academy of Sciences, “Health Risks of Radon and Other Internally Deposited Alpha Emitters,” BEIR IV (National Academy Press, Washington, D.C., 1988). 5. U.S. Environmental Protection Agency, “Evaluation of Waterborne Radon Impact on Indoor Air Quality and Assessment of Control Options” (final report) (Industrial Environmental Research Laboratoly, Research Triangle Park, North Carolina, 1984). 6. W. Nazaroff, S. Doyle, A. Nero, and R. Sextro, “Potable Water as a Source of Airborne Rn-222 in U.S. Dwellings: A Review and Assessment,” Health Physics 52, 281-290 (1987). 7. F. Steinhausler, W. Hofmann, and E. Pohl, “Local and Temporal Distribution Pattern of Radon and Daughters in an Urban Environment and Determination of Organ Dose Frequency Distributions,’’ in T. Gesell and W. Lowder (eds.), The Natural Radiation Environment Ill, CONF-1780422(Springfield, Virginia, National Technical Service, 1980), pp. 1145-1155. 8. E. Letourneau, R. McGregor, and H. Taniguchi, “Background Levels of Radon and Radon Daughters in Canadian Homes” (presented at the Nuclear Energy Agency Meeting on Personal Dosimetry and Area Monitoring Suitable for Radon and Radon Daughter Products, Paris, November 1978). 9. United Nations Scientific Committee on the Effects of Atomic Radiation, “Ionizing Radiation: Sources and Biological Effects” (United Nations, New York, 1982).
143 10. National Council on Radiation Protection and Measurement, “Environmental Exposures to Radon and Radon Daughters in the United States,” NCRP Report No. 78 (NCRP, Bethesda, Maryland, 1984). 11. J. Hursh, D. Morken, T. Davis, and A. Loovas, “The Fate of Rn-222 Ingested by Man,” Health Physics 11, 465476 (1965). 12. W. Von Dobeln and B. Lindell, “Some Aspects of Rn-222 Contamination Following Ingestion,” Arkiv fir Fysik 27, 531-572 (1964). 13. I. Anderson and I. Nilsson, “Exposure Following Ingestion of Water Containing Rn-222,” in Assessment of Radioactivity in Man (Vienna, IAEA, 1964), p. 3117. 14. M. Suomela and M. Kahlos, “Studies on the Elimination Rate and the Radiation Exposure Following Ingestion of Rn-222 Rich Water,” Health Physics 23, 641-652 (1972). 15. E. Nussbaum and J. Hursh, “Rn-222 Solubility in Rat Tissues,” Science 125, 552-553 (1957). 16. K. Ellis, S . Cohn, H. Susskind, and H. Atkins, “Kinetics of Inhaled Krypton in Man,” Health Physics 33, 515-522 (1977). 17. W. Kirk, P. Parish, and D. Morken, “In-Vivo Solubility of Kr85 in Guinea Pig Tissues,” Health Physids 28, 249-261 (1975). 18. C. Tobias, H. Jones, J. Lawrence, and J. Hamilton, “The Uptake and Elimination of Krypton and Other Inert Gases by the Human Body,” Journal of Clinical Investigation 28, 1375-1385 (1949). 19. M. Winston and S . Wilson, “Intestinal Perfusion Studies Using Xe-133 During Correction of Mesenteric Vascular Insufficiency,” Journal of Nuclear Medicine 18, 269-271 (1977). 20. B. Lindell, “Ingested Radon as a Source of Human Radiation Exposure,” in Proceedings of the First International Congress of Radiation Protection, (Pergamon Press, New York 1968), pp. 719-725. 21. N. Underwood and J. Diaz, “A Study of the Gaseous Exchange Between the Circulating System and the Lungs,” American Journal of Physiology 13, 88-95 (1941). 22. J. Correia, S . Weise, R. Callahan, and H. Strauss, “The Kinetics of Ingested Rn-222 in Humans Determined from Measurements with Xe-133” (Massachusetts General Hospital, Boston, 1987). 23. International Commission on Radiological Protection, Report of the Task Group on Reference Man, ICRP Publication 23 (Pergamon Press, Oxford, 1979). 24. M. Sleisinger and J. Fordtran, Gastrointestinal Disease (W.B. Saunders, Philadelphia, 1973). 25. D. Crawford-Brown and L. Shy, “The Relationship Between Hit Probability and Dose for Alpha Emissions Under Selected Geometries,” Radiation Protection Dosimetry 20, 155-168 (1987). 26. D. Crawford-Brown, “Age-Dependent Lung Doses from Ingested Rn-222 in Drinking Water,” Health Physics 52, 149-156 (1987). 27. C. Andreev, “The Techniques of Determining the Radiation Doses Absorbed During Whole Body Alpha Therapy and Some Results from Peroral Administration of Rn-222 Water,” Medical Radiology 8, 69-76 (1963). 28. U.S. Environmental Protection Agency, “Risk Assessment Methodology, Environmental Impact Statement for NESHAPS Radionuclides,” US E.P.A. Report EPA 520/1-89-005 (September 1989). 29. National Academy of Sciences, “The Effects on Populations of Exposure to Low Levels of Ionizing Radiation,” BEIR 111 (National Academy Press, Washington, D.C., 1980). 30. United Nations Scientific Committee on the Effects of Atomic Radiation, “Sources, Effects and Risks of Ionizing Radiation” (United Nations, New York, 1988). 31. National Academy of Sciences, “The Effects on Populations of Exposure to Low Levels of Ionizing Radiation,” BEIR V (National Academy Press, Washington, D.C., 1990). 32. International Commission on Radiological Protection, Recommendations of the International Commission on Radiological Protection, ICRP Publication 26 (Pergamon Press, New York, 1977).
Cancer Fatalities from Waterborne Radon (Rn-222) Douglas J. Crawford-Brown' Received March 7, 1990; revised July 27, 1990
A model of the biokinetics of radon in the human body following ingestion is developed from existing data. Calculations of the probability of cancer fatality from use of radon-laden water in the home then are presented. The pathways of emanation and ingestion are examined and shown to lead to roughly equal risks. The probability of fatal cancer resulting from lifetime use of water at a radon concentration of 1 pCi/L is shown to be 1 x with a reasonable range between The allowed concentration consistent with an excess risk of then is 2 x lo-' and 5 x approximately 100 pCi/L, which is exceeded in a significant fraction of U.S. water supplies. The lifetime number of premature deaths due to waterborne radon in the U.S. is estimated to lie between 5000 and 125,000, with a best estimate of 25,000. KEY WORDS: Radon; cancer; water; biokinetics.
probability of cancer fatality resulting from use of waterborne radon in U.S. homes. The model and resulting calculations will be used here to propose regulatory limits consistent with selected levels of risk. As will be demonstrated, the fatalities imposed by direct ingestion of radon are predicted to be of roughly equal importance to the regulator. Since the fatalities resulting from emanated radon have been reviewed p r e v i ~ u s l y , ( ~these ,~) will simply be summarized here. The models developed for the case of ingestion are, however, previously unpublished and will be developed in detail. Throughout the initial sections of the paper, the risks associated with a unit concentration (1pCi/L) will be reported. The fatalities in the U.S. population at the current EPA estimate of the population-averaged radon concentration then will be developed in the final discussion.
1. INTRODUCTION Throughout the 1980s, there has been a rise in concern over the health effects of environmental radon. While interest has been directed primarily to exposures resulting from airborne radon (or progeny), substantial efforts also have been mounted to quantify the effects from radon contained in water supplies. This latter effort arises from calculations indicating the large contribution of radon to the total estimated risk from all forms of waterborne pollutants, both natural and manmade.(') It has been estimated, for example, that waterborne radon will cause between 1000 and 30,000 premature deaths from cancer in a cohort the size of the current U.S. population.(2) The Office of Drinking Water in the U.S. Environmental Protection Agency has, therefore, undertaken to develop regulatory limits on the allowed concentration of radon in U.S. water supplies. The following sections present a model for the biokinetics of radon in the human body following ingestion, and detail calculations of the expected excess lifetime
2. EMANATED RADON
There are two primary routes of exposure to radon entering a structure through water supplies. The first is emanation of radon into the air, followed by decay to charged radioactive progeny. These progeny then are inhaled, depositing in the passageways of the lung and
Department of Environmental Sciences and Engineering, University of North Carolina at Chapel Hill, Chapel Hill, North Carolina 275997400.
135
0272-4332/91~300-0135106.50/1 OD 1991 Socicty for Risk Analysis
Crawford-Brown
136
irradiating adjacent tissue. The primary concern here is with the induction of lung cancer within the bronchial epithelium.(4)The second route of exposure is the direct ingestion of the radon into the stomach, followed by translocation and/or decay. Stomach cancer is the primary effect of concern here, although other organs will be shown to contribute significantly to the number of fatalities. The present section explores the health effects from emanation. Other sections focus on direct ingestion. As described in previous reports,(2J) calculations of the probability of fatality from lung cancer following emanation require specification of several factors. The first is the concentration, C, (in pCi/L), of radon in the water supply, taken here to be 1 pCi/L. The second factor is the equilibrium concentration ratio for air and water, given as CJC,. The third is the equilibrium factor for radon progeny, f, which describes the ratio of the total potential alpha energy from progeny in air to that obtained under conditions of complete equilibrium. The product of C,, C,/C,, f, and 0.01 (the conversion factor from pCi/L to working levels) then equals the working level concentration of progeny in air, WL. Multiplying this latter concentration by T (the number of hours spent in this environment throughout a year), and dividing by 170 (the number of hours per “working month”), yields the annual exposure in units of working level months, WLM. Values for CJC, (dimensionsless if both concentrations are in pCi/L) have been measured and have been calculated using models of the attachment of progeny to particulate and room surfaces in homes. Two recent compendia report values ranging between 0.17 x and 3.5 x 10-4(5)and 0.23 x A and 1.87 x mean value of approximately 1x for U.S. homes is adopted here. Measured values o f f in dwellings range between 0.25 and 0.8,(7-9) with various authors employing mean values of between 0.5 and 0.7 for purposes of risk estimati~n.(~,’~) A mean value of 0.5 is adopted here. The value of T divided by 170 is set equal to 52 as a matter of regulatory policy (implying continuous exposure). As a result, the conversion factor from the concentration of radon in water (pCi/L) to the annual cumulative exposure (in WZM) is 1x x 0.5 x 0.01 x 52 or 2.6 x WLM per pCi/L. Values for other exposure intervals, T, may be calculated through scaling proportionately. Risk factors for the induction of lung cancer are taken from the recent BEIR IV report,(4)which estimates 350 excess cancers/million personslWLM. This value was obtained from the fitting of a linear dose-response function to the available epidemiological data on mining ex-
posures. The risk factor is corrected for the higher unattached fraction in homes (7%) relative to mines (4%), and for the difference in breathing patterns between the two exposed groups. As estimated by the BEIR comrnittee,c4) the excess lifetime probability of fatal lung cancer assuming continuous exposure at 1 WLMlyear in assuming a latency of 20 years. the home is 1.7 x The conversion factor from radon concentration in water (pCi/L) to excess lifetime probability of fatal lung cancer x (1.7 X or approximately then is (2.6 x 4 x lo-’. For comparison, Cross et ~ 1 . ( estimate ~) between 3 and 8 x lo-’. Previous risk analyses for waterborne radon have relied on one of two approaches. In the first,c2) it was assumed that the risk from ingested radon is insignificant compared to emanation. In the limited data from earlier studies on whole-body retention of radon were used to estimate the risk following ingestion. Since these earlier studies did not measure concentrations in individual organs, however, it was necessary to combine all organs into a single compartment. The exception was the stomach, for which partial data were available. Through use of the second approach, Cross et estimated a lifetime cancer risk (emanation plus ingestion) per pCi/L, with the risk from of between 4 and 8 x ingestion being an order of magnitude below that from emanation.
3. INGESTED RADON More accurate assessments of risk may be obtained through explicit modeling of doses delivered to the separate organs and tissues of the body. Ingested radon enters the body through the stomach, followed by rapid movement to the other body organs. Measurements of the retention of radon have been reported by Hursh et al.,(”) von Dobeln and Lindell,(12)Anderson and Nilsand Suomela and Kahlos.(14)These measurements relied primarily on the detection of radon progeny in the whole body, yielding little information on organspecific doses. These data do, however, indicate that radon is removed from the body with a primary halftime of between 30 and 50 min, with a smaller component characterized by a half-time on the order of several hours. Hursh er ul.(I1) also demonstrated that the radon is removed from the body primarily through exhalation via the lung. Nussbaum and Hursh(I5)report the relative concentrations of radon in rat tissues at equilibrium. They note values of 1 (omental fat), 0.08 (venous blood), 0.06 (kidneys), 0.05 (heart), 0.04 (testis), and 0.3 (muscle).
Cancer Fatalities from Waterborne Radon
137
These data indicate the ability of fat to sequester radon, suggesting that the half-time for removal from the whole body is dominated by fat at long times after ingestion. Ellis et aZ.(16) measured the kinetics of inhaled krypton in human subjects. Their measured half-time for wholebody retention correlated highly with total body fat, supporting the concept of body fat as a storage site. Kirk et aZ.(17)studied the solubility of krypton in the tissues of guinea pigs and found weighted partition coefficients (equal approximately to the relative concentrations at equilibrium) of 0.421 in omental fat, 0.405 in substaneous fat, 0.259 in thymous, and 0.134 in bone marrow. Tobias et aZ.(18) and Winston and Wilson(19)each have reported measurements of the absorption of gases from the GI tract. Their results suggest little movement from the stomach to the venous blood. The major site of absorption was the small intestine, with the gas moving from the stomach to the small intestine and, finally, to the bloodstream. The clearance of radon from the bloodstream is rapid compared to the half-time of removal from the whole body. Lindellc20)and Underwood and Diaz@l)report blood clearance half-times on the order of minutes. Tobias el a1.(18)also confirmed that radon contained in blood and entering the lung is removed to the lung airspace in a single pass, supporting the contention that the lung is the primay route of removal from the body. As a result, it may be assumed that radon is transferred rapidly from the blood to the systemic organs and tissues. This simplifies calculations since decay in the bloodstream will be negligible.
t
The most complete set of data on radon biokinetics has been reported by Correia et U Z . ( ~ ~ ) In their study, xenon was used as a chemical analogue for radon. The concentration of xenon in each organ or tissue was measured through external imaging following ingestion of approximately 2 mCi of xenon in 100 cc of saline solution. These data then were converted to an estimate of the equivalent radon concentration through use of the relative solubility of radon and xenon in human tissue. The means of measurement data on 22 individuals are provided in Figs. 1-7. Organs imaged include fat, stomach, small intestine, ascending colon, descending colon, liver/portal blood, and whole-field. The points for the general tissue curve were calculated for the present study by subtracting the radon contents of the stomach, small intestine, ascending colon, descending colon, liver, lung, and torso blood from the whole-field content. The results in Figs. 1-7 have been normalized to yield a total activity of 2 mCi in the stomach at the start of measurements. Error bars on the measurements were not reported by the original investigators. In the same study, Correia et aZ.(22)performed measurements of the half-time for removal of xenon from body organs following inhalation. These results are presented in Table I and will be employed in the biokinetics model developed here. It will be noted that the retention half-time in fat is elevated greatly above that for other organs. Figure 8 displays the biokinetics model developed in the present study. The compartment termed general tissue includes the mass of intestinal and stomach epi-
00001
OOooOl(!40 i 0 I i O I$O O 2;
O 2;
2;)O 340 Time (minutes)
3k 4hO 4AO O4;
O 5;
&O 6 h O
Fig. 1. The data (0),fitted exponential functions (--) and biokinetics model predictions (-) for the general body tissue compartment. The total intake of radon is 2 millicuries into the stomach at I = 0. The concentration then is in units of millicuriedcc. Figures 1-7 employ the same key.
Crawford-Brown
138
01 -
0001 -
001 -
Fig. 2. The same key as Fig. 1. The organ is the stomach.
Fig. 5. The same key as Fig. 1. The organ is the descending colon.
i0 i0 O ;l
I$O 2&l
40 2AO 3;O
3LO 4b0 4b0 4AO O 5;
560 6b0
Time (minutes)
oooo10
Fig. 6. The same key as Fig. 1. The organ is the livedportal blood. 40
80 120 I60 2 0 24C 280 320 360 400 440 480 520 560 600 llm(mmulms)
Fig. 3. The same key as Fig. 1. The organ is the small intestine.
Fig. 4. The same key as Fig. 1. The organ is the ascending colon.
thelium, into which radon is recirculated after entering the bloodstream. The assumed masses for the various compartments may be found in Table I1 for adults and are taken from the ICRP Reference Man report.(23) Rate constants for translocation were obtained through simultaneous fitting to the concentrations in Figs. 1-7, with the exception of pathways F,G,I,K, and M. These latter constants have been assigned on the basis of Table I, which was generated from the inhalation/ organ removal studies. All rate constants are displayed in Table 111. The model fits may be found as solid lines in Figs. 1-7 and were obtained by assuming an intake of 2 mCi of radon. The original data in Figs. 1-7 also were approximated by multi-exponential functions describing the concentration in each organ as a function of time. These concentration functions are presented in Table IV and have been normalized to an intake of 1 pCi of radon.
Cancer Fatalities from Waterborne Radon
139
‘1 \
01-‘,
0‘ ,
5
0,
5
0001 -
lim(minutrn)
Fig. 7. The same key as Fig. 1. The organ is the whole body (0)or fat
(a).Only the biokinetics model predictions for fat are displayed (-).
Table I. Summary of Xenon Removal Half-Times for Organs Following Inhalation
Organs
Half-times (Min)
Liver, spleen, kidneys Stomach, small intestine Lower and upper intestine Fat tissue
1.3 1.6 4.2 362
A
The doses calculated using these simplified approximations agree with doses calculated using the biokinetics model, with the values agreeing to within 10% for all organs or tissues except fat (which has no significance in risk estimates). The resulting functions are also displayed in Figs. 1-7 as dashed lines, again assuming an initial intake of 2 mCi of radon.
F
E V
B
+. N
I4
V
A.C.
FAT
4. CALCULATIONS OF INGESTION DOSES
The dose to each organ or tissue following ingestion of 1 pCi was obtained through integration of the model predictions as a function of time. The integral organ burden, B,(pCi-seclg), is equal to I
Bi =
Ci(t)dt
(1)
0
where Ci(t)is the concentration (pCi/g) in organ i at time
Fig. 8. The biokinetics model for ingested radon developed in the present study. Rate constant$ are provided in Table 111.
Crawford-Brown
140 Table 11. Compartmental Masses and Volumes for the Biokinetics Model Displayed in Fig. 8 -
Organ Stomach w/contents Small intestine Upper large intestine Lower large intestine General tissuea Blood (plasma) Blood (total) Liver Lung airspace volume Lung minute volume Fat Muscle
Assumed mass or volume
400 g 1040 g
430 g 295 g 32,035 g 3100 g 5500 g 1800 g 4300 cc 7500 cc/min 13,500 g 28.000 e
"Taken to be the difference between the total body mass (without bone) and the sum of the specific organ masses listed in this table. The epithelium of the stomach and intestines is included in the general tissue compartment mass.
Table 111. A Summary of Transfer Rate Constants for the Biokinetics Model Displayed in Fig. 8 Pathway A
Rate constant (inverse min) 0.04 0.005 0.005
B C D E F G
1.5 0.16 0.4 0.4
H
0.13 0.0073 0.057
I J K L M N P
0.0019
0.083 0.4 1.8 0.49 1.74
Q
t. Assuming a tissue density of 1 g/cc, the dose, Di (in
rads), then is:
Di= Bi x 0,037 x 19.2 x 1.6 x
(2)
In this expression, 19.2 is the total alpha energy (MeV) released by decays of Rn-222, Po-218, and Po-214. It is assumed that the progeny decay at the site of the radon is a conversion factor decay. The constant 1.6 x from MeVIg to rad, and 0.037 is a conversion factor from pCi to disintegrations per sec.
The dose to the walls of the stomach and intestines require a slight modification due to the concentration profile between the lumen and the bloodstream. This concentration equals that in the lumen contents at the surface of the epithelium and decreases to very small values (essentially zero) at the depth of the capillaries. The epithelial layer varies in thickness due to undulations, but generally averages between 30 and 100 mic r o n ~ . ( The ~ ~ ) radiosensitive crypt cells tend to lie most deeply within this layer. In the following calculations, it will be assumed that the radiosensitive cells lie at a depth of 30 microns into a 40 micron epithelial layer. The dosimetry for this arrangement has been reported by Crawford-Brown and Shyr(2s)and will not be reviewed here. The net result of these dosimetric considerations is that the dose calculated using Eq. (2) must be divided by a factor of 3. This reduction does not apply to recirculated radon contained within the general tissue compartment, which distributes uniformly throughout the epithelial layer. Unfortunately, the data of Correia el al.(22)do not permit a separation of the radon in lung tissue from that contained in the air of the lung. The researchers do report, however, that the concentration in lung tissue is approximately equal to that in the liver. For purposes of calculation here, the lung tissue will be assumed in equilibrium with the concentration in the liver. As discussed elsewhere,(28)the dose to the lung includes a component from radon decaying in the airspaces prior to exhalation. This radon (and its progeny) will irradiate the lung during exhalation. It is assumed that essentially all of the ingested radon is removed from the body via the lungs. The details of the calculation required to estimate the dose delivered by exhaled radon are given in Ref. 26 and will not be repeated here. It is assumed that the radon is mixed uniformly in the alveolar air and that the progeny migrate rapidly to the walls of the region in which the radon decays. The doses to cells in the alveolar region from in situ decay may be found in Column 1 of Table V, and the alveolar doses from the exhalation of radon may be found in Column 3 of that table. Similar calculations for various generations of the bronchial tree may be found in Column 4, but will not be used here due to their insignificance in risk estimates. A summary of the organ doses calculated in the present report may be found in Table VI. The dependence of doses on age are obtained through scaling in inverse proportion to body mass. The choice to scale in this manner is driven by the finding of Suomela and Kah10s"~)that the half-time for removal of radon is identical for humans and rats [see also Andree~(*~)]. Since
Cancer Fatalities from Waterborne Radon
141
Table IV. A Summary of Approximate Concentration Functions for the Body Organs Following Ingestion of 1 pCi of Radon-All Concentrations Are in Units of pCi/cc ~~~
Stomach Small intestine Ascending colon Descending colon Liver Muscle Fat Whole body General tissue
C ( f )= C(f) = C(t) = C(f)= C(r) = C ( f )= C(t) =
C(f) = C(t) =
~
~
0.002[0.998 exp( -0.029t) + 0.002 exp( -0.0026t)l 0.0002[0.9 exp(-0.02) + 0.1 exp(-0,0045t)J (1 - exp(-O.O99f)} 0.0001[0.85 exp( -0.015t) + 0.15 exp( -0.006t)J (1 - exp( -0.0991)) O.O0003[exp( -O.O048r)] (1 - exp( -0.099f)) 0.00003[0.87 exp( -0.022!) + 0.13 exp( -0.0039f)J 0.00001[exp( -0.0029t)l (1 - exp(0.231)) 0.00001[exp( -0.005t)l (1 - exp(0.099t)) 0.00001[0.92 exp( -0.022) 0.08 exp( -0.0035t)J 0.0000051[exp( - 0.005t)J (1 - exp(0.099t))
+
Table V. A Comparison of Alveolar and Basal Cell Doses (rads) from In Situ and Exhalation Decays Following Ingestion of 1.0 pCi of Radon in Water
Age (years) 0 2 5
10 Adult
P
2b
3‘
4d
2(E-8) 7(E-9) 4(E-9) 2(E-9) 1(E-9)
2(E-9) 7(E-10) 3(E-10) 2(E-10) 1(E-10)
4(E-9) 7(E-10) 1(E-9) 7(E-10) 3(E-10)
4(E-9) 1(E-9) 7(E-10) 4(E-10) 2(E-10)
5’ 3(E-9) 7(E-10) 7(E-10) 4(E-10) 2(E-10)
6f 1(E-9) 3(E-10) 4(E-10) 3(E-10) 1(E-10)
nAlveolar region doses, and all basal cell doses, from in situ decay. The concentration in lung cells assumed equal to liver. bAlveolar region doses, and all basal cell doses, from in situ decay. The concentration in lung cells assumed equal to general tissue. ‘Alveolar region doses from exhalation decays. dDosesto the basal cells of generation 5 from exhalation decays. ‘Doses to the basal cells of generation 10 from exhalation decays. Boses to the basal cells of generation 16 from exhalation decays.
radon is removed primarily through the lung, this equivalence of half-times may be attributed to the fact that the ratio of lung-surface-area-to-body-massis similar in humans and rats. Since this ratio also is roughly invariant with age in humans,(26)it is suggested that the wholebody removal half-time is invariant with age. Doses resulting from a unit intake of radon will, therefore, be inversely proportional to body mass. The exception is the dose to the stomach and intestinal walls. Recirculated radon in the epithelium should follow the pattern of age-dependent dose noted above. The dose from radon transported from the lumen to the capillarieswill, however, depend only on the transit time across the epithelium. There is no evidence that this transit time varies with age. As a result, the age-dependence will apply only to the recirculated component of the dose to stomach and intestinal epithelium.
5. RISK FROM INGESTED RADON
For the purpose of estimating risk, the methodology of the USEPA(28)will be adopted. This methodology utilizes risk factors developed by the National Academy of Sciences(29)and UNSCEAR(30)and is based on lifetable analyses of competing risks. Their lifetime risk factor for fatalities from all cancers due to low LET, whole body, irradiation is given as 392 per million person-rad. This risk factor would be doubled if the BEIR V 3 1 ) risk estimates are employed. The same report(28)gives site-specific incidence risk coefficients for a variety of organs. From Table 6.6 of that report, the age-adjusted ratio of the incidence of stomach cancer to that of all sites is 0.12, while that for intestinal cancer, liver, and lung is 0.06,0.13, and 0.18, respectively. Multiplying these values by the risk factor of 392 per million person-rad yields a lifetime risk factor of 46 per million person-rad for stomach cancer, 23 per million person-rad for the total intestinal tract, 50 per million person-rad for liver, and 70 per million personrad for lung. It is assumed here that the three intestinal sections possess identical radiosensitivities, indicating that each should be assigned a risk factor of approximately 8 per million person-rad. If a quality factor of 20 is adopted,(32)the risk factors for the stomach and separate intestinal sections are 920 per million person-rad and 150 per million person-rad, respectively. For the liver and lung, the risk coefficients are 1000 and 1400 per million person-rad, respectively. These values will be assumed here. These organs dominate the risk from ingested radon. The USEPA assigns a value of 2 Wday to the daily rate of ingestion of water; of this, between 200 and 1200 ml are ingested untreated.c3) A value of 660 ml is adopted here, consistent with the choice by the EPA. This implies an annual rate of ingestion equal to 220 L. At a
Crawford-Brown
142
Table VI. Summary of Dose per Unit Ingestion (rad/pCi), Assuming 100% of the Ingested Radon Is Taken into the Stomach (the Results of Column 5 Are Used in This Report)
Y
Organ
1" Age 0
2" Age 2
3" Age 5
4" Age 10
5" Adult
Ref. 11
7= Ref. 12
8' Ref, 13
Ref. 14
Stomach Liver Small intestine Ascending colon Descending colon General tissue Lung alveoli Whole body
2(E-8) 2(E-8) 3(E-8) 4(E-8) 6(E-8) 9(E-9) 3(E-8) NA
2(E-8) 7(E-9) 1(E-8) 1(E-8) 2(E-8) 3(E-9) 8(E-9) NA
2(E-8) 4(E-9) 9(E-9) 1(E-8) 1(E-8) 2(E-9) 5(E-9) NA
2(E-8) 2(E-9) 6(E-9) 7(E-9) 6(E-9) 1(E-9) 3(E-9) NA
2(E-8) 1(E-9) 5(E-9) 7(E-9) 4(E-9) 7(E-10) 2(E-9) NA
2(E-8) 7(E-10) NA NA NA NA 4(E-10) NA
NAd NA NA NA NA NA NA 2(E-10)
NA NA NA NA NA NA NA 7(E-10)
4(E-8) NA NA NA NA NA NA NA
6b
"Assumes all progeny decay at the site of radon decay. For the dose to the alveolar cells, doses in columns 1 and 3 are summed. The concentration of radon in lung tissue is assumed equal to that of the liver. For the stomach and intestines (short-term retention components only), radiosensitive cells are located at 30 microns in a 40 micron wall. bSee Ref. 11. Stomach doses include only decays of radon and Po-218. Concentration of radon in the wall is the same as in the lumen. 'Uses measurements of radon progeny. dNA, Not availablc.
concentration of 1 pCi/L, the annual rate of intake for radon then will be 220 pCi. From Table VI, the dose to the stomach from ingestion of 1 pCi (incorporating the effect of age) is approximately 2 x rad, while that for the small intestine, ascending colon, descending colon, liver, and lung are 7 x 10 x 6x 2x and 2 x respectively. The lifetime probability of fatal stomach cancer from ingestion of 220 pCi/ year of radon then is 2.4 x The lifetime (70-year) probability of fatal cancer of the small intestine, ascending colon and descending colon, liver, and lung are 2.1 x 2x 1x 3x and 7 x lo-&, respectively. The sum of these six probabilities then is 4 x lo-'. The Office of Radiation Programs of the USEPA recommends a quality factor (or Relative Biological Effectiveness) of 8 for the alpha emissions from ingested radon. If that quality is employed here, the total lifetime probability of fatal cancer from ingested radon is 1.6 x 10-7. 6. DISCUSSION As mentioned previously, Cross et ~ 1 . (estimate ~) a risk coefficient of between 4 and 8 x per pCi/L, 90% of which is due to emanation. From the present analysis, water containing radon at a concentration of 1 pCi/L poses an excess lifetime probability of cancer fatality equal to approximately 4 x for emanated radon. The excess lifetime probability of fatal cancer from at the same concentration. ingested radon is 4 x
These results indicate that the risk of fatal cancer from ingestion is approximately equal to that from emanation. Since the risk from ingested radon is dominated by stomach cancers, any future changes in the risk factor for stomach cancer will heavily affect this result. The current population-averaged concentration of radon in U.S. public drinking water supplies is estimated to be 150 pCi/L.(*) Multiplying this concentration by the above risk factors yields an excess lifetime probability of fatal cancer equal to 6 x for emanated radon and 6X for ingested radon. The total excess lifetime conprobability of fatal cancer then is equal to 1x sidering both pathways of exposure. For a population of approximately 250 million persons, the number of excess fatalities due to cancer over a lifetime is approximately 25,000. The uncertainty due to the risk from emanation is on the order of a factor of 5.(2)Assuming a similar uncertainty for the composite risk, the number of excess fatalities over a lifetime probability lies between 5000 and 125,000. These results may now be used to estimate the concentration of radon corresponding to various excess lifetime probabilities of fatal cancer. For a probability of the corresponding concentration would be 1 pCi/ L. For a probability of the corresponding concentration would be 10 p C i L Finally, for a probability of the corresponding concentration would be 100 pCi/ L. The latter concentration falls within the range of values being considered by the EPA in setting a regulation. These concentrations may be compared with the cumulative frequency of water supplies containing radon (see
Cancer Fatalities from Waterborne Radon
curve C of Fig. 1 in Ref. 2). A regulatory standard of 1 pCi/L would require remediation in essentially all water supplies. A standard of 10 pCi/L would require remediation of the water used by approximately 90% of the population. A standard of 100 pCi/L would require remediation of the water used by approximately 30% of the population. Clearly, for any regulation aimed at an excess lifetime probability of or less, the number of supplies requiring remediation will be large. ACKNOWLEDGMENTS
The research reported here was funded through a grant from the USEPA, but does not represent the views of the agency. Gratitude is extended to Greg Helms of the EPA for administering the grant and to Neil Nelson of the EPA for extensive reviews of the research. The author also wishes to thank Ms. Delores E. Plummer for her excellent assistance in preparing the document. REFERENCES 1. C. Cothern and W. Marcus, “Estimating Risk for Carcinogenic Environmental Contaminants and Its Impact on Regulatory Decision Making,” Regulatory Toxicology and Pharmacology 4,265274 (1984). 2. D. Crawford-Brown and C. Cothern, “A Bayesian Analysis or Scientific Judgement of Uncertainties in Estimating Risk Due to Rn-222 in U S . Public Drinking Water Supplies,” Health Physics 53, 11-21 (1987). 3. F. Cross, N. Harley, and W. Hofmann, “Health Effects and Risk from Rn-222 in Drinking Water,” Health Physics 48, 649-670 (1985). 4. National Academy of Sciences, “Health Risks of Radon and Other Internally Deposited Alpha Emitters,” BEIR IV (National Academy Press, Washington, D.C., 1988). 5. U.S. Environmental Protection Agency, “Evaluation of Waterborne Radon Impact on Indoor Air Quality and Assessment of Control Options” (final report) (Industrial Environmental Research Laboratoly, Research Triangle Park, North Carolina, 1984). 6. W. Nazaroff, S. Doyle, A. Nero, and R. Sextro, “Potable Water as a Source of Airborne Rn-222 in U.S. Dwellings: A Review and Assessment,” Health Physics 52, 281-290 (1987). 7. F. Steinhausler, W. Hofmann, and E. Pohl, “Local and Temporal Distribution Pattern of Radon and Daughters in an Urban Environment and Determination of Organ Dose Frequency Distributions,’’ in T. Gesell and W. Lowder (eds.), The Natural Radiation Environment Ill, CONF-1780422(Springfield, Virginia, National Technical Service, 1980), pp. 1145-1155. 8. E. Letourneau, R. McGregor, and H. Taniguchi, “Background Levels of Radon and Radon Daughters in Canadian Homes” (presented at the Nuclear Energy Agency Meeting on Personal Dosimetry and Area Monitoring Suitable for Radon and Radon Daughter Products, Paris, November 1978). 9. United Nations Scientific Committee on the Effects of Atomic Radiation, “Ionizing Radiation: Sources and Biological Effects” (United Nations, New York, 1982).
143 10. National Council on Radiation Protection and Measurement, “Environmental Exposures to Radon and Radon Daughters in the United States,” NCRP Report No. 78 (NCRP, Bethesda, Maryland, 1984). 11. J. Hursh, D. Morken, T. Davis, and A. Loovas, “The Fate of Rn-222 Ingested by Man,” Health Physics 11, 465476 (1965). 12. W. Von Dobeln and B. Lindell, “Some Aspects of Rn-222 Contamination Following Ingestion,” Arkiv fir Fysik 27, 531-572 (1964). 13. I. Anderson and I. Nilsson, “Exposure Following Ingestion of Water Containing Rn-222,” in Assessment of Radioactivity in Man (Vienna, IAEA, 1964), p. 3117. 14. M. Suomela and M. Kahlos, “Studies on the Elimination Rate and the Radiation Exposure Following Ingestion of Rn-222 Rich Water,” Health Physics 23, 641-652 (1972). 15. E. Nussbaum and J. Hursh, “Rn-222 Solubility in Rat Tissues,” Science 125, 552-553 (1957). 16. K. Ellis, S . Cohn, H. Susskind, and H. Atkins, “Kinetics of Inhaled Krypton in Man,” Health Physics 33, 515-522 (1977). 17. W. Kirk, P. Parish, and D. Morken, “In-Vivo Solubility of Kr85 in Guinea Pig Tissues,” Health Physids 28, 249-261 (1975). 18. C. Tobias, H. Jones, J. Lawrence, and J. Hamilton, “The Uptake and Elimination of Krypton and Other Inert Gases by the Human Body,” Journal of Clinical Investigation 28, 1375-1385 (1949). 19. M. Winston and S . Wilson, “Intestinal Perfusion Studies Using Xe-133 During Correction of Mesenteric Vascular Insufficiency,” Journal of Nuclear Medicine 18, 269-271 (1977). 20. B. Lindell, “Ingested Radon as a Source of Human Radiation Exposure,” in Proceedings of the First International Congress of Radiation Protection, (Pergamon Press, New York 1968), pp. 719-725. 21. N. Underwood and J. Diaz, “A Study of the Gaseous Exchange Between the Circulating System and the Lungs,” American Journal of Physiology 13, 88-95 (1941). 22. J. Correia, S . Weise, R. Callahan, and H. Strauss, “The Kinetics of Ingested Rn-222 in Humans Determined from Measurements with Xe-133” (Massachusetts General Hospital, Boston, 1987). 23. International Commission on Radiological Protection, Report of the Task Group on Reference Man, ICRP Publication 23 (Pergamon Press, Oxford, 1979). 24. M. Sleisinger and J. Fordtran, Gastrointestinal Disease (W.B. Saunders, Philadelphia, 1973). 25. D. Crawford-Brown and L. Shy, “The Relationship Between Hit Probability and Dose for Alpha Emissions Under Selected Geometries,” Radiation Protection Dosimetry 20, 155-168 (1987). 26. D. Crawford-Brown, “Age-Dependent Lung Doses from Ingested Rn-222 in Drinking Water,” Health Physics 52, 149-156 (1987). 27. C. Andreev, “The Techniques of Determining the Radiation Doses Absorbed During Whole Body Alpha Therapy and Some Results from Peroral Administration of Rn-222 Water,” Medical Radiology 8, 69-76 (1963). 28. U.S. Environmental Protection Agency, “Risk Assessment Methodology, Environmental Impact Statement for NESHAPS Radionuclides,” US E.P.A. Report EPA 520/1-89-005 (September 1989). 29. National Academy of Sciences, “The Effects on Populations of Exposure to Low Levels of Ionizing Radiation,” BEIR 111 (National Academy Press, Washington, D.C., 1980). 30. United Nations Scientific Committee on the Effects of Atomic Radiation, “Sources, Effects and Risks of Ionizing Radiation” (United Nations, New York, 1988). 31. National Academy of Sciences, “The Effects on Populations of Exposure to Low Levels of Ionizing Radiation,” BEIR V (National Academy Press, Washington, D.C., 1990). 32. International Commission on Radiological Protection, Recommendations of the International Commission on Radiological Protection, ICRP Publication 26 (Pergamon Press, New York, 1977).
