Introduction

    Citizen volunteers from the Carlsbad, New Mexico area were monitored for internally deposited radionuclides through a project entitled "Lie Down and Be Counted" (LDBC). This project is provided as an outreach service to the public to support education about naturally occurring and man-made radioactivity present in people and the environment prior to the opening of the Waste Isolation Pilot Plant (WIPP). The data collected prior to the opening of the WIPP facility will provide information for future studies and serve as a baseline for operational monitoring. In addition, information obtained from these measurements will be used to evaluate and reduce the uncertainties associated with bioassay methodologies. It is important to note that these data represent an interim summary of an ongoing study.

Methods

    Following the commissioning of the Center's in vivo monitoring facility, 272 citizen volunteers were assayed during July 1997 to September 1998. These volunteers were recruited through presentations to local community groups and businesses. When a citizen volunteer arrived at the Center for a bioassay, he or she viewed a short video explaining the measurement protocol, and completed a lifestyle questionnaire which included questions regarding age, sex, ethnicity, occupation, foreign travel, wild game consumption, smoking habit and any nuclear medicine procedures (Table 12). In addition, the subject's height and weight were recorded.

    The counting shield consists of a large shielded room measuring 2.7 m wide, 3.0 m long, and 2.7 m high. It is constructed from 25 cm-thick cast iron obtained from pre-World War II iron. A graded-Z liner (Z represents the charge of the liner element) consisting of 64 mm of lead, 32 mm of tin, and 32 mm of stainless steel was added to the inside of the iron walls of the shield to attenuate photons produced within the shield walls.

    Radionuclides analyzed for in lungs included ²³²Th, ¹⁴⁴Ce, natural U, ²³⁵U, ²²⁶Ra, ²³³U, ¹⁵⁵Eu, ²¹⁰Pb, ²³⁷Np, Pu isotopes, ²⁴¹Am, ²⁴⁴Cm, and ²⁵²Cf. Radionuclides analyzed for in the whole body included ⁴⁰K, ⁵¹Cr, ⁵⁴Mn, ⁵⁸Co, ⁵⁹Fe, ⁶⁰Co, ⁶⁵Zn, ⁸⁸Y, ⁹⁵Zr, ¹⁰³Ru, ¹⁰⁶Ru, ¹²⁵Sb, ¹³¹I, ¹³³Ba, ¹³³I, ¹³⁴Cs, ¹³⁷Cs, ¹⁴⁰Ba, ¹⁴¹Ce, ¹⁵²Eu, ¹⁵⁴Eu and ¹⁹²Ir (Table 13). A 30-min count time was used for each subject.

    The identification of radionuclides of interest was accomplished using a library-driven peak search and a library-driven region of interest (ROI) analysis. The concept of 'library driven' means that only the portions of the spectrum that would contain the photons of interest (usually listed in a library) are examined.

    Background and net counts for each ROI were calculated using a step-background computation. This computation calculates the background counts for each channel (i) in the ROI of interest (B_i) using the following equation:

where Y_j is the total counts in channel j, L is the left starting channel of the ROI, R is the right ending channel of the ROI, L_avg is the average value of the Compton continuum on the left side of the ROI determined from the sum of four channels to the left of L, and R_avg is the average value of the Compton continuum on the right side of the ROI determined from the sum of four channels to the right of R.Next,the total background counts for the region of interest (B_T) were calculated by summing B_i from channels R to L. Finally, the net count rate in the ROI was calculated by subtracting B_T from the observed gross counts and dividing by the count time.

where C_gross is the total gross counts in the region of interest and t is the count time in seconds.

    The value of MDA indicates the ability of a facility to detect a radionuclide in a person. The MDA represents the amount of a radionuclide that, if present, would be detected 95% of the time under routine operation of a facility. The MDA is used to measure the efficacy of a facility, but it should not be used to decide if a specific radiobioassay has or has not detected activity within a person. MDA was calculated using the following equation (HPS N13.30, Performance Criteria for Radiobioassay, May 1996):

where K_0.05 is the calibration factor, taking into consideration counting efficiency and self absorption, that represents the 5^th percentile in distribution of individual specific calibration factors; U is a conversion factor taking into account photon yield, radioactive decay during the counting interval and unit conversion; t is the count time in seconds; S_B1 is the standard deviation, including Poisson and other random error components in the count of a subject, determined by the routine measurement procedure, where the subject contains no actual analyte activity above that of the appropriate blank; and S_B0 is the standard deviation, including Poisson and other random error components, in the unadjusted count of an appropriate blank. The term S_B1 was determined from the measurement of 272 and 271 individuals for radionuclides deposited in lungs and whole body, respectively (a single whole body count was invalidated due to instrument malfunction). The term S_B0 was determined from 20 measurements of a bottle manikin absorption (BOMAB) phantom filled with de-ionized water.

    It is important to note that the use of K_0.05 applies to lung counting where individual-specific calibration factors vary with subject chest wall thickness. For example, in this study individual-specific calibration factors for ²³⁸Pu ranged from 4E-5 to 4E-7. For radionclides deposited in the whole body, a single calibration factor per photon energy is applied to all individuals regardless of anthropomorphic characteristics. Historically, the variability associated with K in lung counting has been ignored, resulting in unrealistically low estimates of MDA for radionuclides deposited in lungs.

    Calibration factors for lung counting were determined as a function of photon energy and muscle equivalent chest wall thickness (MCWT) using the humanoid torso phantom developed by Lawrence Livermore National Laboratory, 100% muscle equivalent chest plates, and ²³⁸Pu and ²⁴¹Am/¹⁵²Eu lung sets. Calibration factors for whole body counting were determined using a BOMAB phantom filled with a muscle equivalent epoxy containing ¹⁵²Eu and ⁴⁰K.

    For plutonium isotopes, ²⁴⁴Cm and ²⁵²Cf, the MCWT was determined from the subject's height, weight and age using Equation 4 (Sumerlin, T. J. and S. P. Quant, 1982, Radiation Protection Dosimetry, 3, 203):

where W_a is the subject's weight (kg), H_a is the subject's height (m) and A is the subject's age. The total thickness of tissue composing the chest wall (CWT) was determined for the remaining radionuclides in lungs using Equation 5, which represents a composite of the work performed by Fry (Fry, F. A., 1980, Health Physics, 39, 89), Garg (Garg, S. P., 1977, Health Physics 32, 54) and Dean (Dean, P. N., 1973, Health Physics, 24, 439):

where W_b is the subject's weight (lb) and H_b is the subject's height (inch). It is important to note that this predictive equation is not normalized to muscle equivalent thickness and will result in a conservative estimate of calibration factors (Vickers, L. R., 1996, Health Physics, 70, 346). The calibration factors selected for the MDA calculation for plutonium isotopes, ²⁴⁴Cm and ²⁵²Cf (photon energies less than 20 keV) corresponded to the 95^th percentile of the MCWT (K is inversely proportional to MCWT) determined from Equation 4. The calibration factors selected in the calculation of MDA for the remaining radionuclides in lungs (photon energies greater than 20 keV) corresponded to the 95^th percentile of CWT determined from Equation 5. This value of CWT was then converted into muscle equivalent chestwall thicknesses for each photon energy greater than 20 keV (MCWT_{>20 keV}) using Equation 6 (Krammer, G. H., et. al. 1998, Health Physics, 74, 594):

where m_msc,i is the linear attenuation coefficient at photon energy i for muscle (cm^-1), m_adp,iis the linear attenuation coefficient at photon energy i for adipose (cm^-1) and AMF is the adipose mass fraction. The AMF for male and females were assumed to be 0.34 and 0.64, respectively (Vickers, L. R., 1996, Health Physics, 70, 346). A single value of AMF, weighted by the proportion of males and females in the cohort, was used for the calculation of MCWT_{>20 keV}.

Results and Discussion

    The difference in background rates between the human subjects and the BOMAB phantom were greater for lung counts than whole body counts, where the BOMAB phantom provided a reasonable estimate of human subject background at energies less than 700 keV. The energy response or shape of the background spectrum was well approximated by the BOMAB phantom, especially for a lung count (Fig. 39). For both lung and whole body counts, repeated measures of the BOMAB phantom underestimated the variance in human subject background (Fig 40).

    The variance to mean ratio (Fig. 41) in human subject background for a lung and whole body count was consistently greater than one. This has important implications to whole body and nuclear counting, since it is often assumed that background is characterized by a Poisson distribution. With a Poisson distribution, the variance to mean ratio is one, and the best estimate of the mean response is that observed (Knoll, G. F., 1989, Radiation Detection and Measurement, John Wiley & Sons Inc., New York). If there are no other sources of variability, then the mean, variance and standard deviation of a count result (e.g. background) can be estimated from a single measurement. For these data, if the Poisson assumption is applied and the human background was estimated from the measurement of single subject or phantom (often the case), the variability in background would have been underestimated by as much as a factor of 10, but in all cases at least a factor of 2. Underestimating the variability of background will result in an unrealistically low estimate of measurement sensitivity.

    MDAs calculated as described herein for radionuclides in lungs ranged from 6.2E+00 Bq for ²³⁵U to 3.1E+04 Bq for ²³⁹Pu (Table 13). MDAs for radionuclides in the whole body ranged from 1.4E+01 Bq for ⁸⁸Y to 1.9E+02 for ⁴⁰K. Values of MDA reported herein are a factor of 2 to 12 greater than those reported in the CEMRC 1997 Report. The reasons for the increases are threefold. First, values of MDA reported in 1997 were based on repeated measures of a BOMAB phantom filled with 140 g of K (reference man level). Although this is standard practice in bioassay programs throughout the U.S., the data reported herein demonstrated that such methodology underestimates the variability in human background (Fig. 40 and 41). Based on a priori assumptions, a CWT value of 2.4 was used for calculations in the CEMRC 1997 Report to determine K_0.05for MDA calculation of radionuclides in lungs. Finally, the previous calibration for radionuclides in lungs was based on phantom overlays consisting of 50% muscle and 50% adipose equivalent materials. A calibration based on 50% muscle/50% fat overlays has less photon attenuation because of the adipose content than that used for the current data (100% muscle overlays).

    Values of MDA reported herein are more realistic than the MDAs as typically calculated, because the variability and magnitude of human subject background and population based calibration factors (K_0.05) have been considered. It is important to note that most facilities, especially those for lung counting, do not calculate MDA in this manner. Typically, an average or more ideal value of CWT (corresponding to K) is selected for the MDA calculation. However, such methodology can be quite misleading when applied to a population of individuals whose body shapes and sizes are unknown a priori. For example, the reported MDA applies only to those individuals with a CWT less than or equal to that assumed in the MDA calculation. If the assumed value of CWT was the average of a population, then 50% of the individuals in the population would have a CWT greater than the assumed value, resulting in a non-detect lung count for a lung burden equal to the reported MDA. In contrast, for this study only 5% of the population would be expected to have a CWT thickness greater than that assumed for the MDA calculation (in this case MCWT). Thus, the MDA reported herein would result in only 5% of the population having a non-detect lung count for a lung burden equal to the reported MDA at the 95% confidence level.

    Using the ROI methodology, the percentage of results greater than L_C were consistent with a 5% random false positive error rate, at the 95% confidence level, for all radionuclides except ²³²Th via ²¹²Pb, ²³⁵U / ²²⁶Ra, ⁶⁰Co, ¹³⁷Cs, ⁴⁰K, ⁵⁴Mn, ¹⁰³Ru, ²³²Th via ²²⁸Ac and ⁶⁵Zn (Table 14). Five of these (²³²Th via ²¹²Pb, ⁶⁰Co, ⁴⁰K, ⁵⁴Mn (²²⁸Ac interference), and ²³²Th via ²²⁸Ac) are part of the shield-room background and positive detection would be expected at low frequency. The percentage of results greater than L_C  for ¹⁰³Ru and ⁶⁵Zn were below the 95% confidence interval for the random false positive error rate and may be statistical anomalies (38 comparisons to the confidence interval were made). ⁴⁰K is a naturally occurring isotope of an essential biological element, so detection in all individuals is expected. ¹³⁷Cs and ²³⁵U / ²²⁶Ra are not components of the shielded room background and were observed at frequency greater than the 95% confidence interval for the false positive error rate (discussed in more detail later). In addition to false positive rates, the ROI methodology appeared to be effective with respect to false negative error rates (Table 15). For example, 24-hour measurements of the BOMAB phantom were used to determine an equivalent human body burden for shield contamination from the Th series and ⁶⁰Co. Theoretically, if a subject had a body burden equal to L_C, the detection of that radionuclide at L_C would be missed, when actually present (false negative), 50% of the time. The false negative error rate would then increase as the body burden becomes smaller relative to L_C until eventually no activity would be detected. In all cases, shield contamination was equivalent to body burdens at levels below L_C and false negative error rates consistent with theory were observed.

    In contrast, the peak search methodology did not perform as well as the ROI methodology with respect to false positive and false negative error rates and low activity radionuclide detection (Table 14). While valuable to the practice of bioassay, this observation has little impact on this study because the ROI methodology was used for the detection and quantification of radionuclide activity.

    ⁴⁰K results were positive for all individuals, ranging from 2308 to 6513 Bq with an overall mean (±SE) of 3293 (±784) Bq. Such results are expected since K is an essential biological element contained primarily in muscle, and a constant fraction of all naturally occurring K is the radioactive isotope ⁴⁰K. The mean ⁴⁰K value for males (±SE), was 4730
(± 46) Bq, which was significantly greater (P < 0.0001) than that of females, which were 3507 ± 39 Bq (mean ± SE). This result was also expected since, in general, males tend to have larger body sizes and greater muscle content than females.

    In 27.8% of individual measurements, the value for ¹³⁷Cs was greater than L_C  ranging from 6 to 25 Bq. This percentage was significantly higher than the distribution-free confident interval for a 5% random false positive error rate (2.2 to 7.4%), suggesting a low frequency baseline occurrence of ¹³⁷Cs in the local population. From these data, detectable ¹³⁷Cs is present in 22% to 33% (95% confidence level) of citizens living in the Carlsbad area. These results are consistent with preliminary conclusions that were reported in the CEMRC 1997 Report and not unexpected, since ¹³⁷Cs is an abundant, long-lived fission product. Because of its abundance, mobility, and physiological properties, ¹³⁷Cs is widely distributed throughout the biosphere and has been detected previously in many organisms including humans (Whicker, F.W. and V. Schultz, 1982, Radioecology: Nuclear Energy and the Environment 1, CRC Press, Inc., Florida).

    Individual ¹³⁷Cs results were then compared to two sources of demographic data to determine whether the presence of ¹³⁷Cs was dependent on a particular demographic or lifestyle parameter using a Chi-square test of independence (L. Ott, 1988, An Introduction to Statistical Methods and Data Analysis, PWS-Kent Publishing Company, Boston, MA; Table 16). The presence of ¹³⁷Cs was independent of gender, ethnicity, age, radiation work history, consumption of wild game, nuclear medical treatments and European travel. Occurrence of detectable ¹³⁷Cs was slightly associated (p = 0.032) with smoking habit, where smokers had a higher prevalence of ¹³⁷Cs relative to non-smokers (21.3 to 11.2%, respectively). These data are interesting because ¹³⁷Cs is often assumed to be correlated to the consumption of wild game, but this pattern did not appear in these data. The association of ¹³⁷Cs with smoking habit is also interesting, and could be related to the presence of fallout ¹³⁷Cs in tobacco. However, the statistical significance of this dependence was weak, and further study is warranted.

    The percentage of results greater than L_C  for ²³⁵U/²²⁶Ra (9.9%) was significantly (although slightly) higher than the distribution-free confidence interval for a 5% random false positive error rate (2.2 to 7.4%). These data are not nearly as suggestive, when compared to ¹³⁷Cs, of low frequency baseline occurrence of ²³⁵U/²²⁶Ra. It is important to note that ²³⁵U and ²²⁶Ra are naturally occurring and these two radionuclides cannot be distinguished via gamma spectroscopy. Therefore, any positive signal could be the result of either or both radionuclides. This effect was not apparent in the initial data reported in the CEMRC 1997 Report and requires a larger sample size to support or reject the apparent pattern.