Investigating Carcinogen Risk Analysis Through Benzene

Exploring carcinogen risk analysis through benzene
1 / 118
0
0
1401 days ago, 507 views
PowerPoint PPT Presentation

Presentation Transcript

Slide 1

Investigating Carcinogen Risk Analysis Through Benzene Image from Matthew J. Dowd Department of Medicinal Chemistry Virginia Commonwealth University  2002 David M. Hassenzahl

Slide 2

Objective Use benzene as a case for investigating Toxicology Epidemiology Uncertainty Regulatory Science  2002 David M. Hassenzahl

Slide 3

Toolbox Building Likelihood Maximization Curve fitting Bootstrapping Z-Scores Relative Risk Dose-Response extrapolation  2002 David M. Hassenzahl

Slide 4

Overview of benzene Fairly normal hydrocarbon Manufacturing Petroleum items Strongly presumed human cancer-causing agent Animal measures Many epidemiological studies Leukemia as critical endpoint  2002 David M. Hassenzahl

Slide 5

Benzene structure Image from Matthew J. Dowd Department of Medicinal Chemistry Virginia Commonwealth University  2002 David M. Hassenzahl

Slide 6

Benzene Data in Should We Risk It? Toxicological Data, p. 175 et seq. Epidemiological Data p 211 – 216 But numerous other information sets Other toxicological information (uncommon) Chinese laborers Turkish specialists  2002 David M. Hassenzahl

Slide 7

Toxicology Data Set  2002 David M. Hassenzahl Crump and Allen 1984

Slide 8

What are dangers from benzene? Hazard as intensity times presentation How would we decide power? Extrapolate from creature information? Extrapolate from epidemiological information? How wrong will we be? What are "genuine" exposures? What are impacts at these levels?  2002 David M. Hassenzahl

Slide 9

Toxicology Paracelsus "the dosage makes the toxic substance" Regulatory presumptions! This is not Dr. Gerstenberger's Toxicology!  2002 David M. Hassenzahl

Slide 10

Reading SWRI Chapter 5 US EPA Proposed rules (US EPA 1996) Cox 1996  2002 David M. Hassenzahl

Slide 11

General thought Applied measurements Greater specificity about presentation than the study of disease transmission Observed impacts Artificial control of introduction  2002 David M. Hassenzahl

Slide 12

Physiologically Based Pharmacokinetics PBPK Investigate streams of materials through bodies System elements models More on these in introduction address  2002 David M. Hassenzahl

Slide 13

Studies Animals Rarely people Parts Cell tissue  2002 David M. Hassenzahl

Slide 14

Effects Chronic growth casualty expanding enthusiasm for different issues lead and insight in kids. Intense Reversible Irreversible  2002 David M. Hassenzahl

Slide 15

Crump and Allen Benzene information set Animals at different fixations Four information focuses "Fashioner" mice  2002 David M. Hassenzahl

Slide 16

Relevance to Humans How to get from abnormal state, lifetime investigations of creatures to expected low measurement impacts in people?  2002 David M. Hassenzahl

Slide 17

Questions about benzene Is benzene a mouse cancer-causing agent? Is benzene a human cancer-causing agent? Assuming this is the case, how powerful?  2002 David M. Hassenzahl

Slide 18

Benzene information set I Crump and Allen information set (Crump and Allen 1984) Note: the genuine measurements are not expressed effectively here. See "notes for more data  2002 David M. Hassenzahl

Slide 19

Benzene information set II 1.0 0.8 0.6 P(cancer) 0.4 0.2 0 25 50 75 100 Dose (mg/kg/day) Crump and Allen information set .  2002 David M. Hassenzahl

Slide 20

Uncertainty Pervades Often downplayed Creates (or possibly delays) strife Think as we go! (Some portion of Homework PS 2)  2002 David M. Hassenzahl

Slide 21

Animal Test Issues Interspecific correlation Statistical vulnerability Heterogeneity Extrapolation Dose Metric  2002 David M. Hassenzahl

Slide 22

Interspecific examination Mouse-human Metabolism as an element of body weight Dose human = sf  Dose mouse sf = (BW human/BW mouse ) 1 b is exactly determined as 0.75 an a. See SWRI page 177.  2002 David M. Hassenzahl

Slide 23

Interspecific correlation Lifetime of human = lifetime mouse? Mice age 30 days for every human day Total mouse lifetime is much shorter Analogous organs or procedures? Do mice have growth focuses we don't? Do we have tumor focuses mice don't? a. See SWRI page 177.  2002 David M. Hassenzahl

Slide 24

Interspecific examination 1. Hallenbeck, 1993 2. Finley et al., 1994  2002 David M. Hassenzahl

Slide 25

Interspecific examination sf = (BW human/BW mouse ) 1-b sf = (70/0.03) 0.25 = 7.0 Dose human = 7.0  Dose mouse  2002 David M. Hassenzahl

Slide 26

Interspecific correlation Crump and Allen information set, changed over to people  2002 David M. Hassenzahl

Slide 27

Animal Test Issues Interspecies examination Statistical vulnerability Heterogeneity Extrapolation Dose Metric  2002 David M. Hassenzahl

Slide 28

Binomial Distribution 50 hereditarily "indistinguishable" mice… binomial conveyance? Can utilize this to produce "probability capacity" to look at the probability that any given likelihood is  2002 David M. Hassenzahl

Slide 29

Likelihood Maximization More suitable than Least Squares when you know something about probabilities "Bootstrapping" technique required We will work through probability amplification  2002 David M. Hassenzahl

Slide 30

Statistical Uncertainty Can figure standard deviation utilizing the binomial Recall that two standard deviations to either side speaks to a 95% certainty interim, and...  2002 David M. Hassenzahl

Slide 31

Statistical Uncertainty 1.0 0.8 0.6 P(cancer) 0.4 0.2 0 175 350 525 700 Human Dose (mg/kg/day) Crump and Allen information set, connected to people  2002 David M. Hassenzahl

Slide 32

Animal Test Issues Interspecies examination Statistical vulnerability Heterogeneity Extrapolation Dose Metric  2002 David M. Hassenzahl

Slide 33

Heterogeneity Epidemiology and toxicology Genetically indistinguishable mice contrasted with various people Predictable versus flighty helplessness Male and female contrasts (watched growth rates are distinctive)  2002 David M. Hassenzahl

Slide 34

Heterogeneity Genetic assorted qualities among people Early bits of knowledge into tumor instrument: subpopulation conceived with one of two "stages" contended Variability as a component of age  2002 David M. Hassenzahl

Slide 35

Animal Test Issues Interspecies examination Statistical instability Heterogeneity Extrapolation Dose Metric  2002 David M. Hassenzahl

Slide 36

Extrapolation Theoretical or "Robotic" models: one-hit two-hit two-organize Empirical Cox "information driven, demonstrate free bend fitting" EPA Proposed Guidelines  2002 David M. Hassenzahl

Slide 37

Overestimation Tautological effects Thresholds Hormesis, or "Vitamin" impact Underestimation Saturation Synergistic impacts Susceptibility Omission Extrapolation Concerns  2002 David M. Hassenzahl

Slide 38

 2002 David M. Hassenzahl

Slide 39

After EPA (1996)  2002 David M. Hassenzahl

Slide 40

Statistical Uncertainty 1.0 0.8 0.6 P(cancer) 0.4 0.2 0 175 350 525 700 Human Dose (mg/kg/day) Crump and Allen information set, connected to people  2002 David M. Hassenzahl

Slide 41

1.0 LED(10) = 100 mg b/kg/day 0.8 0.6 P(cancer) 0.4 0.2 0 175 350 525 700 Human Dose (mg/kg/day)  2002 David M. Hassenzahl

Slide 42

Extrapolation If LED(10) = 100 mg/kg/day, then LED(10 - 6 ) = 100  10 - 6/0.1 = 1  10 - 4 mg/kg/day  2002 David M. Hassenzahl

Slide 43

Animal Test Issues Interspecies examination Statistical vulnerability Heterogeneity Extrapolation Dose Metric  2002 David M. Hassenzahl

Slide 44

Dose Metric Assumption: presentation is unessential to impact Area under the bend/expected esteem. Lifetime measurements prompts to normal every day dosage. Especially risky if there are limit impacts or outrageous impacts  2002 David M. Hassenzahl

Slide 45

Risk to Humans? Lifetime malignancy hazard 40 hours for every week 50 weeks for every year 30 years Average 10 ppm(v) introduction?  2002 David M. Hassenzahl

Slide 46

Calculate Risk 10ml benzene/liter air 0.313 ml/mg 20m 3 air/day 1000 liters/m 3 70kg man  2002 David M. Hassenzahl

Slide 47

Cancer Risk Lifetime Cancer Probability is a component of Dose and Potency Assume aggregate measurements Use Daily Dose per kg body weight, found the middle value of over lifetime Potency normally given as q* Additional hazard per unit dosage  2002 David M. Hassenzahl

Slide 48

Cancer Risk: Exposure Term  2002 David M. Hassenzahl

Slide 49

Computed Exposure Terms  2002 David M. Hassenzahl

Slide 50

Computed Exposure Terms  2002 David M. Hassenzahl

Slide 51

Cancer Risk  2002 David M. Hassenzahl

Slide 52

"Administrative Science" Issues Neither a straightforward question nor a thoughtless approach (albeit regularly expressed this way)  "Human wellbeing traditionalist" versus "Substantial hand of preservationist suppositions?" May be overestimates May be disparages  2002 David M. Hassenzahl

Slide 53

Regulatory Toxicology "Genuine hazard" is a reified chance ALL evaluations, including focal propensities, are likely wrong More science does not ensure "less hazard" "less instability"  2002 David M. Hassenzahl

Slide 54

Likelihood Maximization A bend fitting strategy  2002 David M. Hassenzahl

Slide 55

Binomial Distribution 50 hereditarily "indistinguishable" mice… binomial circulation? Can utilize this to produce "probability work" for an anticipated result given a watched result  2002 David M. Hassenzahl

Slide 56

Likelihood Maximization More fitting than Least Squares when you know something about probabilities "Bootstrapping" technique required  2002 David M. Hassenzahl

Slide 57

Statistical Uncertainty Can compute standard deviation utilizing the binomial Recall that two standard deviations to either side speaks to a 95% certainty interim, and...  2002 David M. Hassenzahl

Slide 58

Statistical Uncertainty 1.0 0.8 0.6 P(cancer) 0.4 0.2 0 100 200 300 400 Human Dose (mg/kg/day) Crump and Allen information set, connected to people  2002 David M. Hassenzahl

Slide 59

Counting Rules What is the probability of g

SPONSORS