Document wreozrpG68jo3OKqBM2O4yrzD

MICHIGAN DIOXIN EXPOSURE STUDY EFFECT OF AGE AND HISTORICAL INTAKE ON BLOOD DIOXIN CONCENTRATIONS: PHARMAKOKINETIC MODELING TO SUPPORT STATISTICAL ANALYSES Jolliet O1, Wenger Y1, Milbrath MO1, Chang C-W1, Chen Q1, Franzblau, A, Garabrant D1, Towey T 2, Adriaens P2 and Gillespie, BW31Department of Environmental Health Sciences, University of Michigan School of Public Health, 109 S Observatory, Ann Arbor, MI 48109; 2Department of Civil and Environmental Engineering, University of Michigan College of Engineering, 1351 Beal, Ann Arbor, MI 48109; 3Department of Biostatistics, University of Michigan School of Public Health, 109 S Observatory, Ann Arbor, MI 48109. Abstract We develop a method to convert historical intake of dioxin-like compounds of a given individual into a 2005 adjusted intake value. Three successive correction factors need to be applied. 1) Present food intakes are corrected to account for the change of intake with age, as a child eats less than an adult. 2) The second factor accounts for historical changes in dioxin concentration in the food chain, with a peak around 1968. 3) The half-life of each congener in the body is used to decay past intake over time. This approach is used as a pre-treatment of food intake variables - expressed in a 2005 actualized number of meals over lifetime - before performing the statistical regression analysis. It also provides insights on an adequate statistical model for predicting blood as a function of food intake and age. Introduction A central goal of the University of Michigan Dioxin Exposure Study (UMDES) is to determine the factors explaining variation in serum congener levels of polychlorinated dibenzodioxins (PCDDs), polychlorinated dibenzofurans (PCDFs), and polychlorinated biphenyls (PCBs). To understand the effect of intake on current serum concentration levels, it is necessary to account for historical intake of persistent congeners and to decay past concentrations over time as a function of their half-lives. High autocorrelation between the answers to questionnaires for intake in last 5 years and for intake in previous periods makes it difficult to differentiate between the effect of subsequent periods. Moreover, it is difficult to account for the effect of age in a cross-sectional study for persistent organic pollutants with long half-lives in humans as it is confounded with the influence of historical change in concentrations. There is, therefore, a need to develop an innovative and robust approach - inspired by Pharmacokinetic modeling - to account for intakes prior to the cross sectional study, decay in the body between the intake and the sampling time, and variations of the concentration of dioxins in the environment, as a pre-treatment of certain inputs of the statistical analysis. Materials and Methods The main idea of this pre-treatment is to convert past food intake into a 2005 adjusted food intake, analogous to the financial conversion from a past value into a present value (actualized value), accounting for inflation. The 2005 actualized intake of a given congener j through a given food item k (e.g. k=beef) taken in during a previous calendar year t can be calculated as a function of three unitless corrections factors (CF) that modify the eaten quantity in 2005 as reported in the UMDES food questionnaire ( Qi,2005 e.g. the number of meals of beef per year reported in 2005 by an individual i). food ,k Qi ,2005 actualized k , j ( t) = = Qi ,2005 food ,k CFfiood ,k Q CFi,2005 food ,k i overall,k , ( age( t j( age,t )) ) CFconc,k , j ( t ) CFdiecay, j ( age,t ) (1) with 1) CFfood :correction factor for the changes in amount of food consumed at different ages. 2) CFconc : correction factor for the relative change in congener concentration in the food with calendar year. 3) CFdecay :correction factor for the metabolic decay in the body between year of intake t and year of blood measurement1,2. The present study explains each of these correction factors in detail and carries out a short sensitivity study on how the actualized intake is affected depending on the considered calendar year and age of the considered person in 2005. Organohalogen Compounds Vol 69 (2007) O-052 226 MICHIGAN DIOXIN EXPOSURE STUDY Determination of correction factors a) Food correction, accounting for age-dependent generic variations in quantities of food consumed The standardized quantity of food item k, eaten by a generic individual i, as a function of his or her age has been determined on the basis of the EPA Exposure Factors Handbook to reflect the food intake per kg bodyweight for representative consumers, using the following formula (coefficients in table 1, figure 2a). Q ST food ,k ( age ) = k - k age +k (1- e( -( age k ) k ) ) [g/day] (2) Table 1. Coefficients of equation 2, describing the standard intake of food with age, for the 50% percentile food intake (higher percentile for fish to be representative for fish consumers) k=Beef k=Poultry k=Pork k=Fish k [g/day] k[g/day/year] k [g/day] k [g/day]] k [-] 17.61 0.3279 41.75 14.5 1.812 7.814 0.0574 18.89 10.29 1.809 2.42 0.0265 4.412 12.81 2.485 6 0.005 15 15 4 The standard corrective factor for food is then calculated for each year as a ratio of the considered food food consumption in calendar year t, divided by his consumption of the same food item in 2005 (figure 2b). The correction is relatively small as the factor ranges between 0.4 and 1.6. Quantity of food [g/day-pers] Food correction factor Standard ingested quantities of different food items as a function of age 60 50 40 30 20 10 0 0 20 40 60 80 100 Age Beef Poultry Pork Fish Correction factor for beef intake as a function of calendar year, for different age in 2005 1.8 1.6 1.4 1.2 1 0.8 0.6 0.4 0.2 0 1900 1920 1940 1960 1980 Calendar year for Intake 2000 k=Be ef 10 30 50 70 202090 Figure 2a. Variation in quantities of food eaten as a function of age Figure 2b. Corrective factor for beef (CFbieef , unitless) as a function of the calendar year of intake, for individuals of 10 to 90 years old in 2005. b) Correction for changes in food concentration with calendar year Wenger et al.3 fitted measured variation in dioxin concentration in fish from the Great-Lakes adapting the functional form suggested by Pinsky and Lorber4. A similar relationship was fitted for meat and milk as described by Towey et al. 5, based on Winters et al.6 TEQ data, combined with more recent data from Ferrario et al.7,8 and latest USDA survey9: Cconc,k , j ( t ) = ak + ( bk - ak )e( -cbk ( t peak -t )dk ) while t < tpeak (3a) Cconc,k , j ( t ) = ak + ( bk - ak )e( -cak ( t-t peak )dk ) while t tpeak (3b) The resulting correction factor is equal to the ratio of the concentration of congener j in calendar year t of ingestion divided by the concentration of the same food item in 2005 (figure 3). Organohalogen Compounds Vol 69 (2007) O-052 227 MICHIGAN DIOXIN EXPOSURE STUDY Concentration correction tpeak ak bk cbk cak Terrestrial 1968 0.3 2.2 0.006 0.015 Fish 1968 0 1.29 0.00401 0.00401 d k 1.9 1.81 Table 2. Parameters of the TCDD historical function (equ. 3a and 3b). Relative change in TCDD concentration 2005 equivalent C(t)/C(2005) 35 30 25 20 15 10 5 0 1880 1900 1920 1940 1960 1980 2000 2020 Calendar year Terrestrial Fish Figure 3. Correction factors describing the relative change in TCDD concentration in the food with calendar year. c) Correction for decay in the body Based on a comprehensive review of the literature1,2, the main factors influencing the half-life of dioxins and dioxin-like compounds are a high TEQ body burden, smoking, age, gender, and body fat. The third correction factor describes the elimination of the congener j in the body between a calendar year t and the year of blood measurements: CFmi et , j ( t ) = e - k i B , j ( 2005 - t ) (4) where k i B , j [1/year] is the average apparent body elimination rate of the standardized individual i for congener j, from age(t) until age in 2005. The determinants of the elimination rates are discussed in a detailed literature review and analysis by Milbrath et al.2 (equation (1) and (2) and table 1), assuming a linearly increasing half life with age (0.4 years for an infant and 7.2 years at 50 years old for TCDD) and a correction with body fat, smoking status and number of month breastfeeding: Figure 4 shows that for young adults, TCDD decay is rapid and only the last years contribute. However, for a 70 to 90 year old person, decay is much slower, leading to high contribution of past exposures since 1960. Decay correction factor relative to 2005 [-] CF overall Average decay correction factor as a function of intake calendar date for different age in 2005 1.E+00 1.E-01 1.E-02 1.E-03 1.E-04 1905 1925 1945 1965 Year of intake 1985 2005 10 30 50 70 90 Overall correction factor for beef as a function of year of intake, for individuals of different age in 2005 4 3.5 3 2.5 2 1.5 1 0.5 0 1900 1920 1940 1960 1980 2000 2020 year of intake 10 30 50 70 90 Figure 4. Correction factor to adjust for decay of Figure 5. Correction factors to adjust a past beef meal to a past intakes in the body, for individuals of 10 to 90 2005 actualized number of meals, for individuals of 10 to 90 years old in 2005. years old in 2005. Results: Overall and cumulative correction factor for an intake in calendar year t Multiplying the three correction factors (provided in figures 2b, 3 and 4), we obtain an overall correction factor used to calculate a 2005 actualized number of meals (figure 5). The points along each curve are then summed up from birth up to 2005 to calculate a 2005 actualized cumulative number of meal over lifetime as a function of the age of each individual in 2005: 2005 2005 Qi,2005 lifetime adjusted k , j ( age ) = Q i ,2005 adjusted k,j(t ) = Q i ,2005 food ,k CFoiverall,k , j ( t ) t =birthdate t =birthdate (5) Organohalogen Compounds Vol 69 (2007) O-052 228 MICHIGAN DIOXIN EXPOSURE STUDY Figures 6a and 6b compare this lifetime cumulate adjustment factor for TCDD and TCDF., showing that: - The TCDD lifetime actualized number of meals is 10 times the number of meals eaten in 2005 for an individual who is 40 years old in 2005. It increases to more than 50 times for an 80 year old person. - As TCDF has a much shorter half-life in humans than TCDD (2 years for a 50 year old versus 7.2 years for TCDD), its lifetime actualized number of meals is 10 times smaller. As a consequence, intake prior to the 1980s has little influence. Cumulative correction factor for lifetime actualized number of meals Cumulative correction factor for lifetime actualized number of meals Cumulative correction factor [-] Cumulative correction factor [-] 100 50 20 10 5 2 1 0 20 40 60 80 Age of the individual in 2005 100 k=Beef Poultry Pork Dairy Fish Figure 6a. Cumulative lifetime adjustment factor as a function of the age of the individual in 2005. TCDD (log scale) 100 50 k=Beef 20 Poultry 10 Pork 5 Dairy Fish 2 1 0 20 40 60 80 100 Age of the individual in 2005 Figure 6b. Cumulative lifetime adjustment factor as a function of the age of the individual in 2005. TCDF (log scale) Discussion of consequences for the statistical analysis Figures 6a and 6b show that the overall relationship between the logarithm of the integrated factor and age is approximately linear between 20 and 80 years of age. This suggests that an adequate statistical model for predicting blood as a function of food intake should be composed of an exponential term for age (and possibly for factors affecting decay rate) that multiplies a multi-linear sum of diet intakes. Gillespie et al.10 applied the present method to illustrate how the cumulative correction factors can be used as a pre-treatment of food intake variables - expressed in a 2005 actualized cumulative number of meals over lifetime - before performing the statistical regression analysis. The lifetime cumulative number of meals clearly increases with age. It is likely to be a better predictor of blood concentration than the 2005 number of meals as blood concentration is observed to increase with age. Acknowledgements Financial support for this study comes from the Dow Chemical Company through an unrestricted grant to the University of Michigan. The authors acknowledge Ms. Sharyn Vantine for her continued assistance and Drs. Linda Birnbaum, Ron Hites, Paolo Boffetta and Marie Haring Sweeney for their guidance as members of our Scientific Advisory Board. References 1. Milbrath MO, Wenger Y, Gillespie BW, Jolliet O. Submitted to Organohalogen Compounds 2007. 2. Milbrath MO, Chang C-W, Emond C, Franzblau A, Garabrant D, Gillespie BW, Wenger Y, Adriaens P and Jolliet O. Organohalogen Compounds 2007, (forthcoming). 3. Wenger Y, Adriaens P, Franzblau A, Milbrath MO, Garabrant D, Gillespie BW, Towey T, Jolliet O. Organohalogen Compounds 2007, (forthcoming). 4 Pinsky P, Lorber M. Journal of Exposure Analysis and Environmental Epidemiology, Vol. 8(2): 187-206, 1998. 5. Towey T, Wenger Y, Adriaens P, Chang S-C, Hedgeman E, Demond A, Jolliet O. Organohalogen Compounds 2006. 6 . Winters DL, Anderson S, Lorber, M, Ferrario J, Byrne C.Organohalogen Compounds.38:75-77, 1998. 7 Ferrario, J, Byrne, Ch, McDaniel, D and Dupuy, A. Anal. Chem. 1996, 68(4):647-652. 8. Ferrario J.; Byrne C.; Lorber, M.; Saunders P.;Leese W.; Dupuy A.;Winters D.;Cleverly D.; Schaum J;Pinsky P.;Deyrup C.;Ellis, R.; Walcott J. Organohalogen Compounds, Volume 32: 245-251. 9. USFDA, 2004. PCDD/PCDF. Exposure Estimates. Center for Food Safety and Applied Nutrition. Available from: <http://www.cfsan.fda.gov/~lrd/dioxee.html> 10. Gillespie BW, Jolliet O, Chang C-W, Milbrath MO, Wenger Y, Chen Q, Towey T, Lee S-Y, Hong B, Garabrant D, Franzblau A, Hedgeman E, Knutson K, Adriaens P, Demond A, Trin H, Lepkowski J, Olsen K, Luksemburg W. Organohalogen Compounds 2007, (forthcoming).. Organohalogen Compounds Vol 69 (2007) O-052 229