National Academies Press: OpenBook
« Previous: 1 Introduction
Suggested Citation:"2 Evaluation of Latent and Presumptive Periods." Institute of Medicine. 2004. Veterans and Agent Orange: Length of Presumptive Period for Association Between Exposure and Respiratory Cancer. Washington, DC: The National Academies Press. doi: 10.17226/10933.
×
Page 14
Suggested Citation:"2 Evaluation of Latent and Presumptive Periods." Institute of Medicine. 2004. Veterans and Agent Orange: Length of Presumptive Period for Association Between Exposure and Respiratory Cancer. Washington, DC: The National Academies Press. doi: 10.17226/10933.
×
Page 15
Suggested Citation:"2 Evaluation of Latent and Presumptive Periods." Institute of Medicine. 2004. Veterans and Agent Orange: Length of Presumptive Period for Association Between Exposure and Respiratory Cancer. Washington, DC: The National Academies Press. doi: 10.17226/10933.
×
Page 16
Suggested Citation:"2 Evaluation of Latent and Presumptive Periods." Institute of Medicine. 2004. Veterans and Agent Orange: Length of Presumptive Period for Association Between Exposure and Respiratory Cancer. Washington, DC: The National Academies Press. doi: 10.17226/10933.
×
Page 17
Suggested Citation:"2 Evaluation of Latent and Presumptive Periods." Institute of Medicine. 2004. Veterans and Agent Orange: Length of Presumptive Period for Association Between Exposure and Respiratory Cancer. Washington, DC: The National Academies Press. doi: 10.17226/10933.
×
Page 18
Suggested Citation:"2 Evaluation of Latent and Presumptive Periods." Institute of Medicine. 2004. Veterans and Agent Orange: Length of Presumptive Period for Association Between Exposure and Respiratory Cancer. Washington, DC: The National Academies Press. doi: 10.17226/10933.
×
Page 19
Suggested Citation:"2 Evaluation of Latent and Presumptive Periods." Institute of Medicine. 2004. Veterans and Agent Orange: Length of Presumptive Period for Association Between Exposure and Respiratory Cancer. Washington, DC: The National Academies Press. doi: 10.17226/10933.
×
Page 20
Suggested Citation:"2 Evaluation of Latent and Presumptive Periods." Institute of Medicine. 2004. Veterans and Agent Orange: Length of Presumptive Period for Association Between Exposure and Respiratory Cancer. Washington, DC: The National Academies Press. doi: 10.17226/10933.
×
Page 21
Suggested Citation:"2 Evaluation of Latent and Presumptive Periods." Institute of Medicine. 2004. Veterans and Agent Orange: Length of Presumptive Period for Association Between Exposure and Respiratory Cancer. Washington, DC: The National Academies Press. doi: 10.17226/10933.
×
Page 22
Suggested Citation:"2 Evaluation of Latent and Presumptive Periods." Institute of Medicine. 2004. Veterans and Agent Orange: Length of Presumptive Period for Association Between Exposure and Respiratory Cancer. Washington, DC: The National Academies Press. doi: 10.17226/10933.
×
Page 23
Suggested Citation:"2 Evaluation of Latent and Presumptive Periods." Institute of Medicine. 2004. Veterans and Agent Orange: Length of Presumptive Period for Association Between Exposure and Respiratory Cancer. Washington, DC: The National Academies Press. doi: 10.17226/10933.
×
Page 24
Suggested Citation:"2 Evaluation of Latent and Presumptive Periods." Institute of Medicine. 2004. Veterans and Agent Orange: Length of Presumptive Period for Association Between Exposure and Respiratory Cancer. Washington, DC: The National Academies Press. doi: 10.17226/10933.
×
Page 25
Suggested Citation:"2 Evaluation of Latent and Presumptive Periods." Institute of Medicine. 2004. Veterans and Agent Orange: Length of Presumptive Period for Association Between Exposure and Respiratory Cancer. Washington, DC: The National Academies Press. doi: 10.17226/10933.
×
Page 26
Suggested Citation:"2 Evaluation of Latent and Presumptive Periods." Institute of Medicine. 2004. Veterans and Agent Orange: Length of Presumptive Period for Association Between Exposure and Respiratory Cancer. Washington, DC: The National Academies Press. doi: 10.17226/10933.
×
Page 27
Suggested Citation:"2 Evaluation of Latent and Presumptive Periods." Institute of Medicine. 2004. Veterans and Agent Orange: Length of Presumptive Period for Association Between Exposure and Respiratory Cancer. Washington, DC: The National Academies Press. doi: 10.17226/10933.
×
Page 28
Suggested Citation:"2 Evaluation of Latent and Presumptive Periods." Institute of Medicine. 2004. Veterans and Agent Orange: Length of Presumptive Period for Association Between Exposure and Respiratory Cancer. Washington, DC: The National Academies Press. doi: 10.17226/10933.
×
Page 29
Suggested Citation:"2 Evaluation of Latent and Presumptive Periods." Institute of Medicine. 2004. Veterans and Agent Orange: Length of Presumptive Period for Association Between Exposure and Respiratory Cancer. Washington, DC: The National Academies Press. doi: 10.17226/10933.
×
Page 30
Suggested Citation:"2 Evaluation of Latent and Presumptive Periods." Institute of Medicine. 2004. Veterans and Agent Orange: Length of Presumptive Period for Association Between Exposure and Respiratory Cancer. Washington, DC: The National Academies Press. doi: 10.17226/10933.
×
Page 31
Suggested Citation:"2 Evaluation of Latent and Presumptive Periods." Institute of Medicine. 2004. Veterans and Agent Orange: Length of Presumptive Period for Association Between Exposure and Respiratory Cancer. Washington, DC: The National Academies Press. doi: 10.17226/10933.
×
Page 32
Suggested Citation:"2 Evaluation of Latent and Presumptive Periods." Institute of Medicine. 2004. Veterans and Agent Orange: Length of Presumptive Period for Association Between Exposure and Respiratory Cancer. Washington, DC: The National Academies Press. doi: 10.17226/10933.
×
Page 33
Suggested Citation:"2 Evaluation of Latent and Presumptive Periods." Institute of Medicine. 2004. Veterans and Agent Orange: Length of Presumptive Period for Association Between Exposure and Respiratory Cancer. Washington, DC: The National Academies Press. doi: 10.17226/10933.
×
Page 34
Suggested Citation:"2 Evaluation of Latent and Presumptive Periods." Institute of Medicine. 2004. Veterans and Agent Orange: Length of Presumptive Period for Association Between Exposure and Respiratory Cancer. Washington, DC: The National Academies Press. doi: 10.17226/10933.
×
Page 35
Suggested Citation:"2 Evaluation of Latent and Presumptive Periods." Institute of Medicine. 2004. Veterans and Agent Orange: Length of Presumptive Period for Association Between Exposure and Respiratory Cancer. Washington, DC: The National Academies Press. doi: 10.17226/10933.
×
Page 36
Suggested Citation:"2 Evaluation of Latent and Presumptive Periods." Institute of Medicine. 2004. Veterans and Agent Orange: Length of Presumptive Period for Association Between Exposure and Respiratory Cancer. Washington, DC: The National Academies Press. doi: 10.17226/10933.
×
Page 37
Suggested Citation:"2 Evaluation of Latent and Presumptive Periods." Institute of Medicine. 2004. Veterans and Agent Orange: Length of Presumptive Period for Association Between Exposure and Respiratory Cancer. Washington, DC: The National Academies Press. doi: 10.17226/10933.
×
Page 38
Suggested Citation:"2 Evaluation of Latent and Presumptive Periods." Institute of Medicine. 2004. Veterans and Agent Orange: Length of Presumptive Period for Association Between Exposure and Respiratory Cancer. Washington, DC: The National Academies Press. doi: 10.17226/10933.
×
Page 39

Below is the uncorrected machine-read text of this chapter, intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text of each book. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.

2 Evaluation of Latent and Presumptive Periods A s discussed in Chapter 1, the committee is charged with evaluating the presumptive period between exposure to the herbicides used in Vietnam and their contaminants on the one hand, and risk of respiratory cancer on the other, and with determining whether it is possible to identify a time after cessation of exposure to those compounds beyond which a presumption of ser- vice connection for respiratory cancer could not be warranted. Many issues must be taken into consideration in evaluating the period over which a disease can be presumed to be associated with a given exposure. This chapter discusses those issues. It begins by defining concepts of latent period and presumptive period and then discusses factors that can affect the latent period. A discussion of issues to consider in evaluating latent period and presumptive period in epidemiology studies and of statistical methods to use in their analysis follows. Finally, the chapter discusses the time course of respiratory cancer after exposure to chemi- cals known to be associated with it, and how coexposure to the chemicals might affect the presumptive period for 2,3,7,8-tetrachlorodibenzo-p-dioxin (TCDD) and respiratory cancer. LATENT PERIOD VS PRESUMPTIVE PERIOD When quantifying the relationship between a chemical exposure and a dis- ease, epidemiologists are interested in the rate of disease among exposed people (that is, people exposed to concentrations of the chemical greater than back- ground exposure) compared with the rate expected if people had not been ex- posed (that is, people exposed only to a background concentration). They are interested in either the relative rate or the excess rate of disease as the measure of 14

EVALUATION OF LATENT AND PRESUMPTIVE PERIODS 15 comparison. Because cancer can take a long time to develop (years or even decades), an analysis of the effects of an exposure must take into account the latent period, a complex concept that can be defined as the time between an exposure and the occurrence of disease related to it. Depending on the circum- stances, determining the length of the latent period can be simple or complex. The discussion below begins with a simple example of a latent period and progresses to complex examples. The latent period is simplest in the scenario illustrated in Figure 2-1a: an exposure occurs at a single time, and a disease has a relatively quick onset, as in a sudden outbreak of acute illness in a fixed population with common exposure to contaminated food, water, or air. For instance, if mayonnaise at the church picnic was contaminated with salmonellae and many picnickers became acutely ill with diarrhea and vomiting 1­3 days later, the latent period would be 1­3 days and would correspond to the incubation period (the time during which the microorganisms multiplied). As is evident from the range of days, the latent period can vary even in such simple cases. Factors that affect the length of latency--such as age, health status, and genetic susceptibility--are discussed later. Because of the variation of the latent period within an exposed popula- tion--which results in different "latencies" that depend on the various factors involved--epidemiologists usually express latency as a range of intervals be- tween exposure and disease onset. The latent period is more difficult to identify when an exposure occurs over an extended period, not at just one time. An example is cigarette-smoking. Expo- sure to smoke usually does not occur at a single time, but over many years. Therefore, it is difficult to know exactly when the exposure caused the cancer. Typically, the latent period is measured from the time of first exposure, that is, when a person began smoking. Evidence that early exposure is most important in determining the risk of respiratory cancer makes such a measurement of latency valid. But there is also evidence that duration of smoking contributes to overall risk, with increasing duration causing greater elevation in risk. To evaluate the period over which a disease can be presumed to be associ- ated with a given exposure, time since start of exposure and time since cessation of exposure need to be determined. For exposures of short duration, times since start and cessation of exposure are often easy to define; this situation holds for environmental exposures in industrial accidents, particularly if an exposure in- volves chemicals that do not remain in the environment or are quickly eliminated from the body. If the exposure is protracted, like the exposure of pesticide appliers who repeatedly apply pesticides or of production workers employed for months or years, time since exposure is more difficult to quantify because there may have been many individual incidents of exposure. It is also important to distinguish between cessation of external exposure and cessation of exposure of the target organs from persistent elevation of TCDD body burden. Some chemicals, including TCDD and many other chlorinated

16 VETERANS AND AGENT ORANGE herbicides, are retained in some body tissues for a long time (even decades), so target organ exposure continues even after external exposure ceases. Even a brief external exposure, such as that which occurred in Seveso, Italy (Bertazzi et al., 1989a,b; 1997), can involve protracted exposure of many organs. This more- complex scenario is illustrated in Figure 2-1b. In that scenario, disease induc- tion, the point at which initiation of the disease occurs (note that the disease might not be apparent or detectable at this point), might even occur after external exposure ceases. Given the difficulty in establishing the true latent period in studies in which exposure is protracted, several surrogate measures of latency are often estimated. In epidemiologic studies, they can include the time between first exposure and the time of peak relative risks compared with those among the nonexposed, the time between cessation of exposure and the point when disease rates increase above those among the nonexposed, and the time after cessation of exposure when disease rates fall back to those among the nonexposed. It is the third interval that is of interest to this committee. The committee's charge is to assess the length of time after cessation of exposure beyond which respiratory cancer could no longer presumably have been caused by the exposure--the "presump- tive period" for respiratory cancer and exposure to TCDD. The relationship between exposure and chronic diseases can be viewed as a multistage process (see Figure 2-1c) (Checkoway et al., 1990). First, exposure of sufficient duration or intensity begins a disease process. The time between first exposure and the occurrence of the initial steps towards the disease (that is, disease induction) can be called the induction period; this is illustrated in Figure 2-1c as the time from A to B. The induction period can depend on the dose and on other cofactors, such as genetic susceptibility, overall health, and diet. Sec- ond, the manifestation of the disease process (when the disease is detected or observed, which could be when symptoms appear, or when subclinical tests indicate positive evidence of early disease) occurs some time after the induction period. The time between induction and disease manifestation (from B to D in Figure 2-1c) is the true, biologic latent period. In practice, however, it is difficult to distinguish the induction period from the latent period. Unless there is a pre- cursor lesion, which serves as a marker, the disease process usually has begun before the disease is manifest. Therefore, in epidemiologic studies, B (disease induction) in Figure 2-1c is not observed and is usually unknowable, and the entire period between first exposure and disease manifestation is often referred to as the latent period. Hence, in the remainder of this report, we will refer to this period, from first exposure to disease manifestation (the time from A to D in Figure 2-1c) as the latent period, and data shown in Chapter 3 have been based on this definition of latency. The period of concern to the Department of Veterans Affairs, referred to as the presumptive period, is the time from cessation of external exposure to dis- ease manifestation (from C1 to D in Figure 2-1c); termination of service in Viet-

EVALUATION OF LATENT AND PRESUMPTIVE PERIODS 17 a) External Exposure Latent Period Disease Detection A D Time b) External External Body Burden Disease Disease Exposure Exposure Reaches Induction Detection Begins Ends Background A B C1 C2 D Body Burden Time True Latent Period c) Induction Period Unrelated Exposure Presumptive Period External Disease External Disease Exposure Induction Exposure Detection Begins Ends A B C1 D True Latent Period FIGURE 2-1 Schematic time courses for exposure and manifestation of disease. a) Sim- ple model of single, acute exposure to chemical with short half-life in body and no accumulation. b) More complex model of accumulation of chemical in body. c) Model of onset of disease process (modified from Checkoway et al., 1990). Unrelated exposure refers to continued exposure that is not related to the induction of the disease. Although the figure depicts disease induction (B) as occurring prior to the end of external exposure, disease induction can occur at any point (during or after exposure) before disease detec- tion. Note that the latent period depicted in a) through c) is the true latent period, that is, the time from disease induction to disease detection. Typically that period is difficult to determine, therefore, the time between first exposure and the time of elevated relative risk is used as a surrogate measure for the true latent period in epidemiologic studies.

18 VETERANS AND AGENT ORANGE nam is a surrogate for cessation of external exposure. Disease induction by a chemical (B in Figure 2-1c), however, may or may not precede cessation of exposure (C1 in Figure 2-1c). It is possible that B, the effective exposure, could occur at any point during the external exposure (time A to C1 in Figure 2-1c) or at any point after the end of the external exposure at which the body burden is still above background concentration (between C1 and C2 in Figure 2-1b). If the manifestation of disease is contingent on exposures incurred near the end of the exposure period, the induction period might not end until some time after expo- sure ends. It should also be noted that the period from B (disease induction) to D (disease manifestation) is influenced by many factors besides the biology of the disease or the aggressiveness of tumor development, such as access to care, quality of screening, health-care use, comorbidity, and tolerance of symptoms. The period from B to D can also depend on the exposure that initiated the disease process; for instance, the characteristics of the exposure can affect progression of a tumor or the speed at which cells proliferate. In the case of TCDD, addi- tional exposure may affect tumor growth and hence latency. Finally, in light of the persistence indicated by the long half-life of TCDD, it is necessary to visual- ize the various periods with the more complex exposure scenario presented in Figure 2-1b, where point C1 represents the end of external exposure, but the period from C1 to C2 represents the time it takes until exposure declines to that among the general population (who did not receive the high exposure from A to C1). During the period from C1 to C2, the body burden remains high, and expo- sures of organs may contribute to any stage in the disease process. Thus, point B may actually occur around the time of C2. The preceding paragraphs have discussed latency as it may occur in a single individual. Earlier in this chapter, it was pointed out that latency in a population is a range representing the latencies in all the individuals. The range occurs because of variability in the induction period and in the interval from disease induction to diagnosis. When exposure is protracted, the exposures that are rel- evant to disease in a particular person are generally unknown, and this adds uncertainty to variability. These phenomena apply to the period between cessa- tion of exposure (whether external or from elevated body burdens) and disease detection. To ensure that any veteran who develops respiratory cancer that could be ascribed to exposures incurred during service in Vietnam is taken into ac- count, the presumptive period must be the maximal interval between cessation of exposure and detection of disease ascribable to that exposure. FACTORS THAT AFFECT TIME COURSE OF DISEASE Chemical Persistence and Duration of Exposure Previous Veterans and Agent Orange (VAO) reports have concluded that there is "limited/suggestive" evidence of an association between respiratory can-

EVALUATION OF LATENT AND PRESUMPTIVE PERIODS 19 cer and exposure to Agent Orange. Epidemiologic studies suggest that risk is related to dose. In this report, the committee evaluates whether there is evidence of a presumptive period after which risk has decreased to the level among those with background exposure. It is also important to determine how that presump- tive period might be related to exposure history and to the dose to which the respiratory system is exposed. To assess the relationship between body burden and the presumptive period, it is important to know the initial exposure and the initial internal concentration, the course of exposure over time, the duration of exposure, the persistence of the chemical, and, on the basis of those, the cumulative body burden at any time. The routes and rates of uptake, tissue distribution, and transformation of a toxi- cant, and its elimination from the body determine the amounts of a particular chemical that potentially reach and persist in organs or cells, thereby influencing the toxicity and the frequency of genetic changes or nongenetic carcinogenic events in those organs or cells. The known toxicokinetic behavior of a chemical of interest is central in reconstructing exposure history. On the basis of the literature reviewed in the Veterans and Agent Orange series, the chemical in herbicides sprayed in Vietnam that is of greatest concern with respect to respiratory cancer is TCDD (IOM, 2003). The distribution of TCDD and other chlorodibenzo-p-dioxin congeners has been examined exten- sively in animal models and to a smaller extent in humans throughout the last 2 decades. The toxicokinetic behavior of TCDD has been discussed in greater detail in VAO and updates up to the most recent, Update 2002 (IOM, 2003). As summarized in Update 2002, TCDD is distributed to all compartments of the body, although the distribution is not uniform and the proportions accumulated differ from organ to organ. Properties of the chemical, properties of the organs and cells, and the route of exposure affect partitioning, absorption, and accumu- lation. The concentration of chemical in a given organ or tissue depends on the dose to which one is exposed and on the absorption, lipid content, and metabo- lism in the organ of concern. In addition, accumulation in one organ can be influenced by processes in other organs. TCDD is a highly hydrophobic chemical that, like other hydrophobic chemi- cals, readily crosses cell membranes and accumulates in lipid-rich organs. Lipid content is a major factor in the accumulation of TCDD in different organs and in the body as a whole. Biologic processes, especially metabolism, are less well characterized for humans than for animals. It is clear, however, that chemical stored or sequestered internally will be mobilized to maintain an equilibrium between the blood and lipid-rich organs, constituting a source of the chemical for other organs. Target organ exposure to a chemical may persist long after increased exposure from external sources has ceased. At any given time, the combination of tissue concentrations and blood concentrations represents the body burden. In practice, blood concentrations are used as a surrogate for body burden, particularly for quantifying TCDD, for which blood concentrations are

20 VETERANS AND AGENT ORANGE standardized to lipid concentrations. Blood concentrations over time determine the cumulative exposure of organs, including the lung. Estimates of the half-life of TCDD in humans have been derived from long- term studies of blood concentrations in Vietnam veterans and other exposed populations. Studies have consistently confirmed that TCDD is highly persistent in the body, with a half-life in humans averaging about 7.5 years. However, as discussed in Update 2002, recent data now indicate that elimination of TCDD is biphasic; a faster phase occurs very early, and the curve of this elimination may be much steeper in people who are highly exposed. Variation in half-life has been demonstrated to be associated with a number of variables, including body- mass index (BMI), weight, initial dose, time after exposure, and age (IOM, 2003). Differing half-lives of the biphasic elimination of TCDD might complicate the back-extrapolation from serum measurements to body burden after initial expo- sure but might not greatly affect the determination of cumulative dose resulting from persistence or continuing exposure. Determination of cumulative dose re- quires multiple body-burden measurements. The timing of those measurements combined with the nonlinearity in elimination could cause errors in extrapolation to initial peaks (times when exposures were highest) and hence affect the esti- mate of cumulative dose. As outlined above, because TCDD and chlorinated herbicides are retained in some body tissues for a long time (such as decades), target organ exposure continues after external exposure ceases. Thus, even an acute external exposure can result in protracted exposure throughout the body. One can think epidemio- logically of the effect of exposure in the past as the change in risk today that is ascribable to prolonged exposure of target organs; tissue concentrations may decline slowly, and blood concentrations may be maintained over time by redis- tribution from one or more storage sites. Determining at what point a past exposure no longer influences disease induction, or what might be considered the period during which an exposure could be presumed to be associated with a disease requires knowledge of both the beginning and the end of the exposure. In the case of chemicals that are rapidly eliminated from the body, it is reasonable to assume that exposure ends with the termination of contact (for example, the end of a work shift for occupa- tional exposures). In the case of TCDD, as with any chemical that can be stored in the body, a person's exposure can be said to continue as the chemical is released from compartments in the body where it is stored. This continuing elevation in body burden increases the overall duration of exposure of the target organs until the body burden, measured as serum concentration, declines to the background value, that is, until it is indistinguishable from the concentrations in populations that have not had an unusual external exposure. Although external exposure to herbicides in Vietnam had a finite duration, the presumptive period (i.e., the interval between external exposure cessation and disease detection) for respiratory cancer due to increased TCDD concentra-

EVALUATION OF LATENT AND PRESUMPTIVE PERIODS 21 tion after service in Vietnam will be extended by virtue of the prolonged eleva- tion in the body burden of TCDD as reflected in the serum concentration. Fol- lowing that logic, "last exposure" or "end of exposure" would not occur until serum TCDD in Vietnam veterans reaches the background value (C2 in Figure 2-1b). It is also known that loss of weight or body fat, whether intentional or through inanition or cachexia, releases stores of TCDD from adipose tissue. For those reasons, serum concentration at any given time, although correlated with persistent exposure from body burden and external sources combined, does not necessarily reflect the pattern of concentrations through time; consequently, a single measurement cannot reflect overall dose over time, which might be the relevant measurement for risk assessment related to the persistent TCDD. Mechanism of Carcinogenicity The multistage carcinogenesis model is often used to evaluate the carcino- genic properties of a chemical by a mechanistic approach (Barrett, 1993; Pitot, 1986). According to that model, a chemical might initiate, promote, or alter the progression of a neoplasm. A chemical that initiates the neoplastic process might do so in a single exposure. Initiation involves a heritable change in the genome of a normal cell whereby it becomes an "initiated" cell. Chemicals that promote the clonal expansion of initiated cells into a histologically visible population are defined as promoters. Promotion might require multiple exposures for a given duration. Often, the first end product of tumor promotion is a benign cellular lesion that is reversible. In progression, initiated cells undergo further changes and expansion into a tissue mass that may progress irreversibly to malignancy. Some chemicals may act specifically in the progression phase of carcinogenesis (Barrett and Wiseman, 1987). For example, an animal can be treated with a chemical that induces initiated cells in the mammary gland; later exposure to another chemical may promote development, clonal expansion, and progression of those cells into a cancer. Multistage models have also been used in experi- mental animals for the development of pulmonary cancers; such models use N- nitrosodimethylamine or urethan as initiators of the tumors (Beebe et al., 1995; Blakley et al., 1992). Multistage models of carcinogenesis recognize that alterations of multiple independent genes (either by direct or epigenetic actions) are involved in the carcinogenic process (Barrett, 1993). The altered expression of proto-oncogenes can result in positive proliferative signals, and the modulation of tumor- suppressor genes can block the neoplastic growth of cells (Boyd and Barrett, 1991). Chemicals can influence the gene expression of important regulators of clonal expansion and cell proliferation that affect not only induction of neopla- sia but also time to full tumor development. Latent periods for different chemicals depend on differences in the mecha- nisms by which the chemicals act, that is, on their contribution to the initiation

22 VETERANS AND AGENT ORANGE and promotion phase. They also vary within a population because of individual differences in the biochemical pathways (some of which are induced by such chemicals as TCDD) that regulate the kinetics and metabolism of the chemicals and in the stages of the carcinogenic process at which exposure occurs. Carcino- genic chemicals that are relatively persistent in the body may initiate or promote the carcinogenic process over a long period. In such cases, latency can vary and may be short or long from the time of initial exposure. For example, substantial individual differences in latency might be observed for persistent tumor promot- ers, such as TCDD, depending on the time of the exposure relative to the time of exposure to the initiating chemicals or events (for example, mutations). For chemicals with relatively short half-lives, such as 2,4-dichlorophenoxyacetic acid (2,4-D) and 2,4,5-trichlorophenoxyacetic acid (2,4,5-T), initiation or promotion might be expected to occur only during a very short period when the body burden is high. As mentioned earlier, however, the latent period may also be affected by individual differences in biochemical and biologic pathways that affect the carcinogenic process. Such differences generally are based on genetic polymorphisms that affect the function or expression of proto-oncogenes and tumor-suppressor genes or affect biochemical pathways that regulate cellular differentiation and proliferation, including signaling processes that regulate the cell cycle. The relative expression of those pathways is also likely to be tissue- specific. Furthermore, the relative activities of processes that repair mutated DNA and of processes, such as immune surveillance mechanisms, that are responsible for removal of cancerous cells may vary individually and affect not only the ability to develop a tumor but the latent period. Such processes may be consider- ably less efficient in older people or in people who have compromised health status and potentially can lead to shorter average latent periods. Metabolism of a chemical may affect the carcinogenic process, as well as the latent period for tumor development. If a chemical requires metabolism to an active intermediate for initiation of a tumorigenic process, tissue-, cell-, and age- specific characteristics of its metabolism may determine the relative amounts of active carcinogen present. That may be more important for chemicals to which there is protracted exposure or that are persistent. In such cases, changes in metabolic processes that occur with changes in health status, coexposure to other chemicals (such as therapeutic drugs or chemicals in cigarette smoke), and aging could affect the amount of the active metabolite produced at any particular time during the exposure period and alter the apparent timing of disease induction or detection. The amounts of carcinogenic metabolites of most chemicals usually depend on dose. At low doses, metabolism by metabolizing enzymes may effi- ciently limit the accumulation of carcinogenic metabolites; at higher doses, meta- bolic processes may be overwhelmed and this can result in a greater abundance of active metabolites. Such properties may be responsible for suggestions that dose-response relationships for many carcinogens are nonlinear and have thresh-

EVALUATION OF LATENT AND PRESUMPTIVE PERIODS 23 olds. If clearance of active compounds is also affected, apparent latency might be altered. CARCINOGENICITY OF TCDD Research has been conducted with the herbicides sprayed in Vietnam and with TCDD to investigate the mechanisms by which they might induce respira- tory cancer. Most of the experimental data on respiratory cancer are for TCDD, not the herbicides themselves. Therefore, TCDD is the focus of this discussion, however, herbicides are also discussed below. Development of cancer in experimental animals usually depends on the dose of and duration of exposure to a chemical. Cancers can develop during the expo- sure period (for example, during the 2 years of dosing in lifetime-exposure stud- ies in rodents) or at some distant time (latency) after a non-lifetime-exposure regimen. To determine whether a chemical is carcinogenic, under some proto- cols laboratory rodents are exposed to incremental doses of a chemical for a lifetime and monitored for the development of cancer. To assess latency, animals are exposed to a carcinogen for a defined duration and then observed for the rest of their life. TCDD is a known carcinogen in rats and mice and is considered to be a carcinogen in humans. There is no evidence to indicate that TCDD is genotoxic. All the available evidence indicates that it acts as a promoter through multiple pathways in the regulation of cell proliferation and differentiation.1 In addition, TCDD is known to alter the relative levels of enzymes that metabolize other chemicals to genotoxic metabolites. Liver tumors have been consistently ob- served in animals after TCDD treatment, and increases in skin cancer, lung can- cer, and thyroid and adrenal cancers have been seen in some studies. Decreases in uterine, pancreatic, pituitary, mammary, and adrenal cancers have also been seen, but most of these decreases occurred only at high doses and were associ- ated with decreases in body-weight gain; the decrease in mammary tumors was seen in only one study (see IOM, 2003 for review). TCDD has been assessed for tumor-promoting activity in a mouse-lung model. Lung tumors were statistically significantly increased in mice initiated with N-nitrosodimethylamine (NDMA) when promoted with low doses of TCDD (Beebe et al., 1995), but not higher doses, which may have caused pulmonary toxicity. The study lasted 52 weeks. It was suggested that the TCDD-elicited induction of cytochrome P4501A1 in the same model system is correlated with and possibly causally involved in the promotion of tumors (Anderson et al., 1The term "initiator" and "promoter" are used in this report to describe the observed behavior of agents in the classical two-stage carcinogenesis model. The committee notes, however, that TCDD can be carcinogenic in rats in the absence of other known exposures (Kociba et al., 1978).

24 VETERANS AND AGENT ORANGE 1991). In another study with the NDMA model, it was determined that TCDD might promote tumors by contributing to the down-regulation of the K-ras proto- oncogene and the stimulation of raf-1 (Ramakrishna et al., 2002). Lung lesions were evaluated in rats treated with TCDD for 14, 30, or 60 weeks (Tritscher et al., 2000); a statistically significant increase in alveolar-bronchiolar metaplasia was observed in the rats treated for 60 weeks but not in those treated for 30 weeks and observed for another 30 weeks, and this suggested that the develop- ment of lesions in this model system required the longer duration of exposure or perhaps a higher cumulative exposure to TCDD. Despite extensive research in this area, the mechanism underlying TCDD's carcinogenicity remains unknown. Although TCDD is the main chemical of interest for respiratory cancer, it is important to consider the herbicides used in Vietnam and their possible role in respiratory cancer. Cacodylic acid was present in an herbicide used in Vietnam. Cacodylic acid is dimethylarsinic acid (DMA), which is also a metabolite of inorganic arsenic in humans. As discussed elsewhere (see US EPA, 2004), inor- ganic arsenic has been associated with lung cancer and other cancers in occupa- tional settings and in studies of individuals exposed to elevated arsenic in drink- ing water (NRC, 2001), and is classified as "known to be a human carcinogen" by the National Toxicology Program (NTP, 2002). Inorganic arsenic has not been shown to be able to induce tumors in laboratory animals, but exposure of rodents to high concentrations of DMA increased bladder tumors in male Fischer 344 rats (Wei et al., 1999) and pulomary tumors in A/J mice (Hayashi et al., 1998). DMA has also been shown to promote urinary, bladder, kidney, liver, and thyroid tumors in rats and lung tumors in mice (see Kenyon and Hughes, 2001 for review). In a recent study, however, pulmonary neoplasms did not develop in rats exposed to DMA at up to 200 ppm in drinking water for 104 weeks (Wei et al., 2002). Therefore, DMA does not appear to be a potent pulmonary carcinogen in those strains of rats and mice. Furthermore, the mechanism of DMA-induced neoplasia is unknown. Finally, although DMA is formed in humans after expo- sure to inorganic arsenic, it has not been established and cannot be inferred that the effects seen after that exposure occur after exposure to cacodylic acid. Epidemiology studies of agricultural workers and chemical-industry work- ers exposed occupationally to those chemicals have been conducted, but it is difficult in many of those studies to determine which chemical (e.g., 2,4-D, 2,4,5-T or its contaminant TCDD, or other pesticides or chemicals) underlies any effects seen. Those studies are discussed in the Veterans and Agent Orange reports (see IOM, 2003). Any of those studies that are relevant to the presump- tive or latent periods and respiratory cancer are discussed in Chapter 3 of this report. There is no strong evidence of tumorigenic potential of 2,4-D, 2,4,5-T, or picloram in laboratory animals or cell-system assays (see IOM, 2001;2003, for reviews). As discussed in VAO (IOM, 1994) and reviewed in Update 2002 (IOM, 2003), three studies have looked at the carcinogenicity of picloram, of which two were negative and one had equivocal results for which a contaminant

EVALUATION OF LATENT AND PRESUMPTIVE PERIODS 25 (hexachlorobenzene) was thought to mediate the carcinogenicity. The experi- mental data for the carcinogenicity of 2,4-D and 2,4,5-T have demonstrated a lack of carcinogenic effects (IOM, 2003). Given the tumor-promoting potential of TCDD and the persistent nature of this chemical, it is possible that, as indicated above, a cancer could be detected either a short or long time after the external exposure to TCDD because promot- ing effects might occur at any time while the body burden of TCDD is high. It is also possible that enzymes which metabolically activate procarcinogens would be chronically induced by persistent elevation of TCDD concentrations. Expo- sure to initiating chemicals that are bioactivated could occur at any time while the body burden of TCDD is elevated, contributing to the apparent latency. It is known, for example, that TCDD modulates the metabolism of chemicals in to- bacco smoke. ANALYSIS OF LATENT PERIOD AND PRESUMPTIVE PERIOD Time-Related Factors As discussed in Update 1998 (IOM, 1999), the analysis of latency in epide- miologic studies can be complicated for a variety of reasons. Many of them are related to the correlations among various time-related factors. Specifically, dura- tion of exposure, time since exposure, age at start of exposure, and exposure magnitude may be correlated, so an observed pattern of one of these factors could actually be due to correlation with one or more of the others. To disen- tangle the confounding of time-related variables, it is commonly necessary to stratify the analysis or model multiple predictors of risk. Realistically, however, most datasets do not have sufficient information to explore those issues; for the most part, they merely serve as a cautionary note in the interpretation of results. In Box 2-1, hypothetical situations help to illustrate how various time-related factors are intertwined. The exposure of interest is referred to as "the agent". The examples in Box 2-1 are fairly typical of what happens in occupational settings. They are presented primarily to show that examining the effects of one time- related factor may be difficult without information about the others. Such rela- tionships, which are presented in Figure 2-1c, can be summarized as: Age at end of study = age at start of exposure + duration of exposure + time since end of exposure As can be seen in Figure 2-1c, the interrelations are also related to the persis- tence of the chemical in the body and to protracted exposures. Although it can be difficult to disentangle those interrelated effects, it is not always impossible. Many occupational studies, for instance, have shown that the effects of industrial chemicals on cancer are stronger 10­20 years after initiation

26 VETERANS AND AGENT ORANGE BOX 2-1 Hypothetical Situations That Illustrate How Various Time-Related Factors Are Connected · Two people are born in 1945. Both are exposed continuously to an agent for 10 years. One person begins exposure at age 20 and continues to age 30; the other begins at age 30 and continues to age 40. In 1995, at age 50, the first person has gone 30 years since first exposure, and the second only 20 years. Thus, age at first exposure and time since first exposure, for a given duration of exposure and followup, are interrelated; time since last exposure differs between the two. · One person begins exposure to an agent in 1970; a second person begins exposure in 1975. They are the same age at first exposure and exposure contin- ues until 1985. When evaluated in 1995, the first person has both a longer duration of exposure (15 years vs 10 years) and a longer time since first exposure (25 years vs 20 years); this illustrates the relationship between the two factors for a given age at first exposure. In this scenario, time since last exposure is the same for the two persons even though duration of exposure is different. · Two people are born in 1945. One begins exposure at age 20; the other at age 30. Exposure stops for both at age 40. The first person has a lower age at first exposure, a longer duration of exposure, and a longer time since first exposure; all three factors are potentially interrelated. · Two people are exposed to the same concentration of an agent. The person with the longer duration of exposure will, by definition, have a higher cumulative exposure. of exposure than in the first 10 years, after age and calendar time have been controlled. And several radiation-related cancers--including leukemia, gastro- intestinal cancers, and breast cancer--show age at start of exposure to be a strong determinant of risk (NAS, 1990). In general, the ability to investigate the issue of timing of exposure in a given dataset depends on the quality of the exposure measure, the accuracy of the timing-of-exposure information, the number of people who have the disease in question, and the variation of exposure over time within the study group. Because of the potential for high correlations among time-related factors, changes in the magnitude of exposure over time are especially important in attempts to identify latent periods or presumptive periods because those periods are likely to be dose-dependent (Enterline and Henderson, 1973; Peto, 1985; Thomas, 1987). Analytic Methods for Addressing Latency Several analytic methods have been developed over the years for addressing latency issues in assessing exposure-time-response relationships. Most of the epidemiologic studies reviewed in VAO and subsequent updates have assumed a

EVALUATION OF LATENT AND PRESUMPTIVE PERIODS 27 specified latent period and have analyzed exposure-response relationships by using exposure that is lagged by the specified latent period. That method has the advantage of simplicity, but it also has a great potential for model misspecifica- tion when the assumed latency structure is wrong. When dealing with protracted exposures, the method is almost certain to misrepresent the effect of latency. As discussed earlier, the increased risk associated with exposure will usually change not only with variation in exposure but also with time since exposure. Hence, it is desirable to use methods that allow for relative-risk estimates to vary with latency, that is, with time since exposure. Several analytic techniques that have been developed to assess latency are briefly described below. They are presented in chronologic order of development and increasing complexity. They are based on different statistical models, but they all have the general form of allowing the relative-risk estimates to vary with time since exposure. One could therefore think of the latency effect as a weight function of the dose-response relationship. In describing the various methods, we assume that people are subjected to protracted exposure. Cumulative-Exposure Model This method, which assumes that risks at various times since exposure have equal weight, estimates a single relative risk for the sum of protracted expo- sures by assuming that the total cumulative dose is the effective dose. Given the published evidence on latency and lung cancer, which suggests that risk starts at null, rises to a peak, and can then decline to null (Langholz et al., 1999), this method might not properly depict the real exposure-time-response relationship. In some cases, risk does not decline to null; in the case of smoking, for ex- ample, former smokers retain a higher risk of lung cancer than people who never smoked. Stratified Model This method estimates a separate relative risk for each of several contiguous intervals of time since exposure. For instance, models that lag exposure by 10 years in effect assume no increased risk in the first 10 years. A general pattern in the relative-risk estimates as a function of time since exposure could be gleaned from such analysis, but the disjoint character of the model does not lend itself to a continuous assessment of latency effects. Moreover, the pattern in the relative- risk estimates could be strongly affected by the choice of time-since-exposure intervals. This modeling approach was used in some of the occupational and environmental studies relevant to the charge of the committee, including the National Institute for Occupation Safety and Health (NIOSH) study of chemical production workers (Fingerhut et al., 1991) and the Seveso study (Bertazzi et al., 1989a,b; 1997). When additional covariates (such as confounders and cofactors)

28 VETERANS AND AGENT ORANGE are in the model, adjustments for them are stratum-specific, and this makes the interpretation of the change in relative risk (as a function of latent period) more difficult. Those three studies usually limited their analysis to time since first exposure and did not address the end of exposure. Piecewise-Constant Model This method fits a simple piecewise-constant function over "time windows" (Finkelstein, 1991; Langholz et al., 1999). Unlike the stratified model, in which adjustment for all covariates is stratum-specific, this method allows for com- mon overall adjustment of covariates, which leads to a better (more robust) adjustment for confounders. Of course, stratum-specific adjustment for con- founders could be easily accommodated in the piecewise-constant model, when- ever necessary. This method is useful as a first pass in understanding the gen- eral shape of the latency function. However, the discrete nature of the weight function for latency may not give a realistic depiction of a latency curve that is more likely to be continuous. Because the exposures in adjacent time-window intervals are likely to be correlated, the estimated heights of the time-window specific relative-effect estimates can be statistically unstable, particularly when the time windows are narrow. Here, the time-window-specific relative-effect estimates can be interpreted as excess relative risks per unit dose received dur- ing the corresponding time window. From those estimates, a total excess rela- tive risk associated with an exposure history is obtained as the sum of all time- window-specific estimates. Bilinear Model This model fits a latency-weight function that is constructed by two lines that are attached at an inflection point (usually constrained to be 1) that gives the estimate of the peak in relative effects. It is constructed to give estimates of three important parameters that address the following questions: How long does it take before there is an effect of exposure? At what time is the effect of exposure maximal? How long does the effect last? This method improves on the piece- wise-constant model by giving a continuous latency-weight function, which leads to better handling of the correlation between adjacent exposures. Langholz et al. (1999) gives a complete description of the model with variations that accommo- date different scenarios, such as whether there is a lag period before effects are first manifested and whether the effect decreases to null after reaching a maxi- mum at the inflection point. It assumes a linear relationship in time, both during the rise and during the fall in the relative risk.

EVALUATION OF LATENT AND PRESUMPTIVE PERIODS 29 Exponential-Decay Model This method is a variant of the bilinear model discussed above. The impor- tant difference is that it fits a straight-line latency-weight function up to the inflection point but fits an exponential decay curve afterward. The method al- lows for the fact that the exposure agent of interest (such as TCDD) may be retained in the body and slowly released internally. In such scenarios, the expo- nential decay curve is affected by the number of years required for the effect to be reduced by one-half, the so-called half-life (Langholz et al., 1999). Spline-Based Model This method fits piecewise polynomials within a series of time-window intervals and then smoothly joins the piecewise curves by putting appropriate continuity constraints on the estimation process. The resulting latency curve is called a spline. Hauptmann et al. (2000) discuss the details of the method. It has also been applied to the Colorado Plateau uranium-miner data on radon expo- sures (Hauptmann et al., 2001). Mechanistic Models Unlike the various empirical models discussed above, mechanistically based statistical models assess latency under an assumed theory of carcinogenesis, such as the Armitage-Doll multistage model (Armitage and Doll, 1961) or the Moolgavkar-Knudsen two-stage model (Moolgavkar and Venzon, 1979). If the assumed model of carcinogenesis is correct, mechanistic models provide a pow- erful means for dealing with exposure-time-response relationships. Implications of Model Choice Methods for addressing how long it takes to detect an increase in disease risk after exposure and how long the effects of exposure last are discussed below. Time After Exposure to Detect Increase in Disease Risk To determine how long it takes after an exposure (either initial exposure or cessation of exposure) to detect an increase in disease (that is, latency), one must examine the pattern of relative risk over time, looking for the earliest detection of an increase in risk in an exposed population relative to a nonexposed compari- son group. For protracted exposure, it is customary to examine relative risk by time since first exposure because the earliest detectable increase in relative risk may be a manifestation of the earliest exposure. In fact, relative risk related to

30 VETERANS AND AGENT ORANGE specific times since first exposure is often the only measure of latency reported in studies of protracted exposure to herbicides. The critical data items required are the date of first exposure of each subject and the dates of diagnosis of relevant outcomes. With that information, an inves- tigator can determine the contribution of each subject's time-since-first-exposure to the latent period for the study population. If full exposure histories are avail- able, more sophisticated analyses are possible; to account adequately for dose and duration, a full exposure history is needed. With that information, an investi- gator can determine the contribution from each subject in each exposure cat- egory. Several studies provide excellent examples of how a full exposure his- tory, based on employment histories or some combination of external or internal measurements, can be used in a detailed analysis of latency and other time- related factors. One such study is the Colorado Plateau uranium-miner cohort described and analyzed in detail by Langholz et al. (1999). In that study, bilinear and exponential decay models of latency are evaluated, as well as how to fit those models to various types of data. Duration of Effects of Exposure Relative risk related to specific intervals of time since last exposure are used to address the question of how long the effect of an exposure lasts. The pattern of relative risk is examined for the latest indication that the relative risk is greater than 1. The critical data items required for addressing the question are the dates of each start and stop of exposure and the intensity of exposure. Those are needed to classify subjects' time spent in each time-since-first-exposure category. Again, if full exposure histories are available, more sophisticated analyses (for example, time-windows analyses) are possible. However, if the critical issue is time since last exposure, multiple starts and stops present greater complexity in assignment of dose. Data on last exposure before an event are used to determine how long an effect can persist. Comparison of Methods In addressing how long it takes after an exposure to detect an increase in disease risk, the earliest indication of an increase in relative risk is difficult to measure and will be refined as more data are collected. Latent period varies among individuals, so risk in a population changes continuously rather than suddenly jumping from "normal" to "above normal". The simpler stratified and piecewise-constant models may not be able to depict that reality. In contrast, the bilinear and spline-based models allow greater flexibility in accounting for the continuity in the latency-weight function and may yield a more realistic picture of the underlying exposure-time-response relationship.

EVALUATION OF LATENT AND PRESUMPTIVE PERIODS 31 Mechanistic models constitute a more structured way of handling the under- lying process and could be informative if the presumed model of carcinogenesis is reasonable. Actual changes in relative risk probably would occur earlier than indicated by the analysis; but because of limitations in study designs, such changes might not be detectable. In other words, the degree to which an in- creased risk is statistically detectable depends on the size of the particular dataset, on the magnitude of the background risk (which in turn depends on the age distribution of study subjects), and on the magnitude of the increase in risk (which in turn depends on exposure, variation in susceptibility, length of follow- up, and distribution of latent periods among the exposed population). It should be noted that if the latent periods are highly variable among individuals, analysis by time since first exposure may be insensitive because an increase in risk will appear slight and occur gradually. In addition, if the effect of time since first exposure is modified by intensity of exposure or age at exposure, these other factors would have to be accounted for in the analysis. For example, latency might be longer after a small exposure than after a larger one, in which case a study that examined only time since first exposure might encounter greater vari- ability in latency and hence have less ability to assess how long it takes to observe an effect of exposure. Such effect modification would also limit the degree to which results of one study can be generalized to a different population or to another exposure scenario. If exposure is protracted, time since last exposure must be analyzed in the proper time-dependent fashion to address how long the effect of an exposure lasts (Clayton and Hills, 1993). Adjustment for age is also necessary. To achieve adequate power and precision, a study must use a sufficient number of subjects with a long period since cessation of exposure. If exposure has been protracted, much longer follow-up is needed to determine the presumptive period than the latent period. Latency and Interpretation of Epidemiologic Literature Measurement Error Measurement error in assignment of exposure or timing of exposure could increase observed variation in latent period or presumptive period. If extended elevation of body burden after external exposure ceases was not explicitly taken into account in the analysis, the resulting error would be considered an error in measurement. Mortality and Incidence Studies for Examining Latency A chemical with carcinogenic activity may increase the chance of cancer, or it may accelerate development of cancer so that it occurs at an earlier age than it

32 VETERANS AND AGENT ORANGE otherwise would have. The agent may also influence the likelihood that a cancer will result in death, or it may shorten the time between occurrence of the cancer and death caused by it. Which of those processes occurs may depend not only on the agent but also on the site of cancer. For example, lung cancer tends to be fatal in a very high percentage of cases, and death usually comes swiftly. For lung cancer, therefore, a study of mortality is unlikely to provide different results from a study of incidence. In a contrasting example, prostate cancer is fatal in a fairly small proportion of cases; incidence is 5 times higher, or more, than mor- tality (Merrill and Brawley, 1997). Therefore, a study of prostate cancer mortal- ity would be less likely to detect the effect of a carcinogenic agent than would a study of prostate cancer incidence, unless the agent increased the severity of disease. But because prostate cancer is so common and occurs with increasing frequency as men age, any study of prostate cancer incidence should examine whether those exposed to the agent of interest develop the cancer at an earlier age than those not exposed. That type of analysis could be accomplished by using age-specific rates. Caution would have to be exercised in interpreting inci- dence studies because of the recent introduction of prostate-specific-antigen (PSA), a marker for prostate tumors that are not clinically detectable, as a screen- ing tool. Differences among populations in the extent to which PSA is used could confound results (Gann, 1997). In the investigation of cancer latency, changes in relative risk with time since exposure will occur later in mortality studies than in incidence studies by an amount approximately equal to the average time from cancer occurrence to death. Given the short survival of lung cancer patients, it is likely that the pattern with time since exposure will be similar in a mortality study, but the latent period will be longer than an incidence study. The same would be true of the presumptive period. As a result, at any given point in the follow-up period, a mortality study will record fewer events than a study of incidence and so will have lower statistical power even if the exposed and nonexposed cases have the same prognosis. Similarly, if there is a substantial group in the population that is genetically susceptible to the effects of a carcinogen or has an acquired state of susceptibil- ity, there may be a phenomenon similar to "exhaustion of susceptibles", which is more commonly observed in infectious disease. In an infectious-disease out- break, people lacking immunity or natural resistance develop the disease; when this pool of susceptible people is exhausted, the incidence in the population declines, perpetuated only by outbreaks among new entrants. In a chronic dis- ease, such as cancer, there may be a pool of susceptible people who, when challenged by exposure to a carcinogen, get cancer at an earlier age than they would otherwise, and that can cause an apparent decline in incidence in the older age groups. However, one would expect conspicuous excess mortality in the younger groups for this effect to explain a later dip in incidence. A further consideration is competing mortality. When people who are at risk

EVALUATION OF LATENT AND PRESUMPTIVE PERIODS 33 for a cancer die earlier than they otherwise might, they are not available for causes of death later. If the diseases are linked by an exposure, one disease may "hide" the other. For example, a person who smokes heavily may develop lung cancer or chronic obstructive lung disease. The two are related, and there is evidence that a common trait of individual susceptibly plays a role in both. A person who develops lung cancer, which is usually fatal, is not available later in life to develop chronic obstructive pulmonary disease, which is less often fatal in younger people. The sparseness of empirical data available on Vietnam veterans makes it difficult to ascertain whether those complicated effects are occurring. Most of the relevant epidemiologic data for TCDD and herbicides reviewed to date have used the simplest models based on time since exposure. LATENCY AND RESPIRATORY CANCER As discussed in Update 1998 (IOM, 1999), a substantial body of literature explores issues of timing of exposure and respiratory cancer, especially for some agents whose carcinogenic properties in the respiratory system are well studied. This section discusses briefly what is known about the latent period and pre- sumptive period of well-studied respiratory carcinogens (gamma rays, radon daughters, smoking, arsenic, and asbestos) and about factors known to be con- founders and effect modifiers for respiratory cancer. Time Course for Respiratory Carcinogens Other Than TCDD In an investigation of latency issues in radiation exposure of atomic-bomb survivors, it was found that the relative risk of lung cancer began to rise 5­10 years after exposure and reached a plateau about 15 years after exposure to gamma rays. Thirty years after exposure, there was no evidence of a decrease in relative risk (Land, 1987). In addition, the effects of age at exposure are quite pronounced for some sites, such as leukemia, digestive cancers, and breast can- cer (NAS, 1990). In miners exposed to radon daughters (radon decay products), the relative risk of lung cancer was seen to peak 5­10 years after first exposure and then to decline slowly, although the risk appears to be increased even 30 years after exposure (Lubin et al., 1994; Thomas et al., 1994). In addition, the effect of exposure varies with age at exposure: a given exposure results in a lower relative risk in older workers than in younger workers. Langholz et al. (1999) analyzed data on Colorado Plateau miners exposed to radon and concluded that the risk returns to the background value after 34 years. Hazelton et al. (2001) analyzed data from a historical cohort of Chinese tin miners, investigating the contribu- tions of arsenic, radon, cigarette smoke, and pipe smoke to lung cancer risk. With respect to radon, their analyses indicate that the hazard posed by radon

34 VETERANS AND AGENT ORANGE increases sharply 45 years after first exposure and does not appear to decrease with increasing time since last exposure. Analyses of lung cancer indicate that the relative risk posed by smoking begins to rise substantially about 20 years after the initiation of cigarette- smoking. Among ex-smokers, the relative risk declines to about half that of smokers by 12 years after cessation but then remains fairly constant--and higher than in those who never smoked (IOM, 2001). Analyses by Ockene et al. (1990) indicate that it takes as long as 20 years for the benefit of cessation of smoking (a decrease in risk of lung cancer) to be seen. Burns (2000) found that the risk of lung cancer remains increased up to 20 years after cessation of smoking, and Ebbert et al. (2003) found no decrease in risk 30 years after cessation. The analyses for smoking by Hazelton et al. (2001) of the Chinese tin miners show risk remaining increased for at least 30 years. In a cohort of workers exposed to arsenic from a copper smelter in Montana, relative risk of lung cancer was observed to increase with time after exposure, reaching a maximum 15­20 years after first exposure, after which it slowly declined (Breslow and Day, 1987). There was little change in relative risk with age at first exposure. Brown and Chu (1983, 1987) observed a stronger effect of time since last exposure than of time since first exposure, although a model with both gave an excellent fit. The statistical models of arsenic exposure in Chinese tin miners were very similar to those of radon exposure, with a sharp rise in hazard 45 years after first exposure and no apparent drop in risk more than 50 years after last exposure (Hazelton et al., 2001). In a cohort of workers exposed briefly to high concentrations of asbestos during World War II, the relative risk of lung cancer rose sharply 5­10 years after exposure, after which it remained constant up to 40 years after exposure (US EPA, 1986). Relative risk was independent of age at exposure. More recent data on asbestos confirm that the risk of lung cancer after asbestos exposure remains increased for many years, although the risk decreased to less than one- half of the peak 20 years after cessation of exposure (Hauptmann et al., 2002). Thus, after some exposures, relative risks reached a plateau or peaked within 5­10 years; but after most exposures, it took at least 20 years after exposure began for relative risks to peak. In addition, most of the data on those exposures indicate that risks remain increased for many decades after cessation. Data on some of the better-studied lung carcinogens indicate that lung cancer might be attributable to exposure to them for 20, 30, 40, or even more than 50 years after exposure has ended. Potential for Confounding and Effect Modification In any person, multiple factors can contribute to or cause respiratory can- cer. Those factors can include genes that confer a predisposition to cancer, age, and various exogenous factors, such as diet, smoking, and other environmental

EVALUATION OF LATENT AND PRESUMPTIVE PERIODS 35 chemical or physical agents. Substances known to be associated with lung can- cer are nickel, chromium, arsenic, cadmium, polycyclic aromatic hydrocarbons, radon, gamma rays, and asbestos. There is great concern with smoking in con- nection with lung cancer, but it is important to remember that although it is estimated that about 90% of lung cancers in males are the result of smoking tobacco (ACS, 2003), only 10% of smokers will ever develop this cancer. Thus, smoking is not deterministic, and other factors clearly play a role in increasing the risk of cancer. In estimating the risk of respiratory cancer associated with exposure to a given agent, ideally the extent of exposure to those other agents associated with respiratory cancer would be known. If any other such agents are associated with exposure to the herbicides used in Vietnam, they could confound the relationship with respiratory cancer unless adjusted for in statistical analysis. If such con- founding exists, the risk of respiratory cancer associated with exposure to the herbicides or TCDD might be over- or under-estimated. The most pertinent question for this report, however, is whether other respi- ratory carcinogens might affect the duration of latent periods and presumptive periods and, if so, how. Specifically, what is the effect of coexposures on the longest latency to be expected from TCDD? A cofactor, such as smoking, could have a different distribution of latencies; as a result, the apparent distribution of latencies associated with the exposure of interest would be distorted (it could be shorter or longer on the average, and it could show more or less variability). In addition, if the exposure to the cofactor was shorter or longer than the exposure to the main agent of interest, further distortion in the observed latent periods could take place. To evaluate the length of time that a respiratory cancer might be presumed to be associated with exposure, it is necessary to address how coexposures might alter the presumptive period. If other factors alter the rate of cellular processes that affect carcinogenesis (apoptosis, cellular transformation, etc.) or alter the internal doses of herbicides or their contaminants (lengthening or shortening half-lives, blocking absorption), then latent periods or presumptive periods could be shortened or lengthened according to the extent of exposure to the co-factor and the strength of its association with either carcinogenesis or dose modifica- tion. That is, the other exposures could be modifiers of the latent period or presumptive period. Whether smoking or other factors that contribute to respiratory carcinogen- esis could modify the distribution of latent periods over which TCDD or herbi- cides used in Vietnam could cause respiratory cancer is unclear from the empiri- cal evidence; much of the literature reviewed in Chapter 3 of this report regarding the carcinogenicity of TCDD, however, comes from studies in occupational set- tings or of the population exposed environmentally to TCDD after an industrial accident in Seveso, Italy (Bertazzi et al., 1989a,b). It should also be recognized that the observed latent or presumptive periods

36 VETERANS AND AGENT ORANGE could be confounded. For instance, if people exposed to TCDD at the highest levels or for the longest time were also more highly exposed to cigarette smoke or other environmental agents, some of the earliest or latest effects could be mostly or wholly attributable to the other exposures. Since the specific causes of individual cases of disease are not easily identified, confounding of this type would be exceedingly difficult to assess. To address the question of how coexposures alter the latent period and presumptive period, smoking data would be needed, in addition to detailed exposure histories. However, many studies of TCDD that are the most useful for examining the time between exposure and respiratory cancer appear not to include individual-level smoking data in their analyses. Therefore, the possibil- ity that smoking could alter the presumptive period must be considered. Studies that conduct analyses of presumptive periods stratified by smoking status are needed to address that issue. SUMMARY AND CONCLUSIONS Of the chemicals sprayed in Vietnam, TCDD is of greatest concern for the development of respiratory cancer. In evaluating the time course between expo- sure to TCDD and respiratory cancer, it is important to differentiate between the true latent period (time from the induction of disease to disease detection), the latent period typically measured in epidemiologic studies (time from beginning of exposure to disease detection), and what can be referred to as the presumptive period (time from cessation of exposure to disease detection). The presumptive period can be affected by the duration of exposure and the mechanism of carcinogenicity. The effects of chemicals that act early in the carcinogenic process (initiators) generally end earlier than the effects of chemi- cals that act later in the carcinogenic process (promoters or tumor-progression factors). Evidence indicates that TCDD is not an initiator but has tumor- promoting activity. Target organ exposure due to release of chemicals from stores in the body must be taken into account in considering the duration of exposure to chemicals, such as TCDD, that are not rapidly eliminated from the body. A number of statistical methods are available to assist in the analysis of the temporal relation between exposure and disease. If mechanistic data are available, models that incorporate them can be informative. In general, data on chemicals other than TCDD that are known to be associ- ated with respiratory cancer indicate presumptive periods of at least 20 years. In some studies, the risk of respiratory cancer had not dropped to background val- ues even 50 years after cessation of exposure. It is possible that exposure to chemicals other than TCDD, such as by smoking, could modify the length of the presumptive period for TCDD and respiratory cancer, but there are no data on what any such modification might be.

EVALUATION OF LATENT AND PRESUMPTIVE PERIODS 37 REFERENCES ACS (American Cancer Society). 2003. Smoking. http://www.cancer.org/downloads/COM/ November_09_03.pdf (accessed December 18, 2003). Anderson LM, Beebe LE, Fox SD, Issaq HJ, Kovatch RM. 1991. Promotion of mouse lung tumors by bioaccumulated polychlorinated aromatic hydrocarbons. Experimental Lung Research 17(2):455­471. Erratum in: Exp Lung Res 1992 May­Jun;18(3):433. Armitage P, Doll R. 1961. Stochastic models for carcinogenesis. In Neyman J ed. Proceedings of the 4th Berkeley Symposium on Mathematical Statistics and Probability. Berkeley: University of California Press. Barrett JC. 1993. Mechanisms of multistep carcinogenesis and carcinogen risk assessment. Environ- mental Health Perspectives 100:9­20. Barrett JC, Wiseman RW. 1987. Cellular and molecular mechanisms of multistep carcinogenesis: relevance to carcinogen risk assessment. Environmental Health Perspectives 76:65­70. Beebe LE, Anver MR, Riggs CW, Fornwald LW, Anderson LM. 1995. Promotion of N-nitrosodim- ethylamine-initiated mouse lung tumors following single or multiple low dose exposure to 2,3,7,8-tetrachlorodibenzo-p-dioxin. Carcinogenisis 16(6):1345­1349. Bertazzi PA, Zochetti C, Pesatori AC, Guercilena S, Sanarico M, Radice L. 1989a. Mortality in an area contaminated by TCDD following an industrial incident. Medicina Del Lavoro 80:316­ 329. Bertazzi PA, Zocchetti C, Pesatori AC, Guercilena S, Sanarico M, Radice L. 1989b. Ten-year mor- tality study of the population involved in the Seveso incident in 1976. American Journal of Epidemiology 129:1187­1200. Bertazzi PA, Zochetti C, Guercilena S, Consonni D, Tironi A, Landi MT, Pesatori AC. 1997. Dioxin exposure and cancer risk: a 15-year mortality study after the "Seveso accident." Epidemiology 8(6):646­652. Blakley BR, Gagnon JM, Rousseaux CG. 1992. The effect of a commercial 2,4-D formulation on chemical- and viral-induced tumor promotion in mice. Journal of Applied Toxicology 12(4):245­249. Boyd JA, Barrett JC. 1991. Role of tumor suppressor genes in a multistep model of carcinogenesis. Basic Life Sciences 57:183­194; discussion 194­196. Breslow NE, Day NE. 1987. Statistical Methods in Cancer Research, Vol II: The Design and Analy- sis of Cohort Studies. Lyon: International Agency for Research on Cancer. Oxford University Press. Brown CC, Chu KC. 1983. A new method for the analysis of cohort studies: implications of the multistage theory of carcinogenesis applied to occupational arsenic exposure. Environmental Health Perspectives 50:293­308. Brown CC, Chu KC. 1987. Use of multistage models to infer stage affected by carcinogenic expo- sures: example of lung cancer and cigarette smoking. Journal of Chronic Diseases 40 (Suppl. 2):171S­179S. Burns DM. 2000. Primary prevention, smoking, and smoking cessation: implications for future trends in lung cancer prevention. Cancer 89 (Suppl 11):2506­2509. Checkoway H, Pearce N, Hickey JL, Dement JM. 1990. Latency analysis in occupational epidemiol- ogy. Archives of Environmental Health 45(2):95­100. Clayton D, Hills M. 1993. Statistical Models in Epidemiology. New York: Oxford University Press. Ebbert JO, Yang P, Vachon CM, Vierkant RA, Cerhan JR, Folsom AR, Sellers TA. 2003. Lung cancer risk reduction after smoking cessation: observation from a prospective cohort of wom- en. Journal of Clinical Oncology 21(5):921­926. Enterline P, Henderson V. 1973. Type of asbestos and respiratory cancer in the asbestos industry. Archives of Environmental Health 27:312­317.

38 VETERANS AND AGENT ORANGE Fingerhut MA, Halperin WE, Marlow DA, Piacitelli LA, Honchar PA, Sweeney MH, Greife AL, Dill PA, Steenland K, Suruda AJ. 1991. Cancer mortality in workers exposed to 2,3,7,8-tetra- chlorodibenzo-p-dioxin. New England Journal of Medicine 324:212­218. Finkelstein MM. 1991. Use of "time windows" to investigate lung cancer latency intervals at an Ontario steel plant. American Journal of Industrial Medicine 19(2):229­235. Gann PH. 1997. Interpreting recent trends in prostate cancer incidence and mortality. Epidemiology 8:117­120. Hauptmann M, Wellmann J, Lubin JH, Rosenberg PS, Kreienbrock L. 2000. Analysis of exposure- time-response relationships using a spline weight function. Biometrics 56(4):1105­1108. Hauptmann M, Berhane K, Langholz B, Lubin J. 2001. Using splines to analyse latency in the Colorado Plateau uranium miners cohort. Journal of Epidemiology and Biostatistics 6(6):417­ 424. Hauptmann M, Pohlabeln H, Lubin JH, Jockel KH, Ahrens W, Bruske-Hohlfeld I, Wichmann HE. 2002. The exposure-time-response relationship between occupational asbestos exposure and lung cancer in two German case-control studies. American Journal of Industrial Medicine 41(2):89­97. Hayashi H, Kanisawa M, Yamanaka K, Ito T, Udaka N, Ohji H, Okudela K, Okada S, Kitamura H. 1998. Dimethylarsinic acid, a main metabolite of inorganic arsenics, has tumorigenicity and progression effects in the pulmonary tumors of A/J mice. Cancer Letters 125(1-2):83­88. Hazelton WD, Luebeck EG, Heidenreich WF, Moolgavkar SH. 2001. Analysis of a historical cohort of Chinese tin miners with arsenic, radon, cigarette smoke, and pipe smoke exposures using the biologically based two-stage clonal expansion model. Radiation Research 156:78­94. IOM (Institute of Medicine). 1994. Veterans and Agent Orange: Health Effects of Herbicides Used in Vietnam. Washington, DC: National Academy Press. IOM. 1999. Veterans and Agent Orange: Update 1998. Washington, DC: National Academy Press. IOM. 2001. Veterans and Agent Orange: Update 2000. Washington, DC: National Academy Press. IOM. 2003. Veterans and Agent Orange: Update 2002. Washington, DC: The National Academies Press. Kenyon EM, Hughes MF. 2001. A concise review of the toxicity and carcinogenicity of dimethylars- inic acid. Toxicology 160(1-3):227­236. Kociba RJ, Keyes DG, Beyer JE, Carreon RM, Wade CE, Dittenber DA, Kalnins RP, Frauson LE, Park CN, Barnard SD, Hummel RA, Humiston CG. 1978. Results of a two-year chronic toxic- ity and oncogenicity study of 2,3,7,8-tetrachlorodibenzo-p-dioxin in rats. Toxicology Applied Pharmacology 46(2):279­303. Land C. 1987. Temporal distributions of risk for radiation-induced cancers. Journal of Chronic Diseases 40:45­58. Langholz B, Thomas D, Xiang A, Stram D. 1999. Latency analysis in epidemiologic studies of occupational exposures: application to the Colorado Plateau uranium miners cohort. American Journal of Industrial Medicine 35(3):246­256. Lubin J, Boice J, Edling C, Hornung R, Howe G, Kunz E, Kusiak R, Morrison H, Radford E, Samet J, Tirmarche M, Woodward A, Xiang Y, Pierce D. 1994. Radon and Lung Cancer Risk: A Joint Analysis of 11 Underground Miners Studies. US Department of Health and Human Services, Public Health Service, National Institutes of Health, Bethesda. Merrill RM, Brawley OW. 1997. Prostate cancer incidence and mortality rates among white and black men. Epidemiology 8:126­131. Moolgavkar S, Venzon D. 1979. Two-event models for carcinogenesis: incidence curves for child- hood and adult tumors. Mathematical Biosciences 47:55­77. NAS (National Academy of Sciences), National Research Council, Committee on the Biological Effects of Ionizing Radiation (BEIR V). 1990. Health Effect of Exposures to Low Levels of Ionizing Radiation. Washington, DC: National Academy Press.

EVALUATION OF LATENT AND PRESUMPTIVE PERIODS 39 NRC (National Research Council). 2001. Arsenic in Drinking Water: 2001 Update. Washington, DC: National Academy Press. NTP (National Toxicology Program). 2002. Report on Carcinogens, Tenth Edition; US Department of Health and Human Services, Public Health Service, National Toxicology Program, Decem- ber. Ockene JK, Kuller LH, Svendsen KH, Meilahn E. 1990. The relationship of smoking cessation to coronary heart disease and lung cancer in the Multiple Risk Factor Intervention Trial (MRFIT). American Journal of Public Health 80(8):954­958. Peto J. 1985. Some problems in dose-response estimation in cancer epidemiology. In Voug V, Butler G, Hoel D, Peakall D, eds. Methods for Estimating Risk of Chemical Injury: Human and Non- Human Biota and Ecosystems. New York: Wiley. Pitot HC. 1986. The molecular determinants of carcinogenesis. Symposium on Fundamental Cancer Research 39:187­196. Ramakrishna G, Perella C, Birely L, Diwan BA, Fornwald LW, Anderson LM. 2002. Decrease in K- ras p21 and increase in Raf1 and activated Erk 1 and 2 in murine lung tumors initiated by N- nitrosodimethylamine and promoted by 2,3,7,8-tetrachlorodibenzo-p-dioxin. Toxicology and Applied Pharmacology 179:21­34. Thomas DC. 1987. Pitfalls in the analysis of exposure­time­reponse relationships. Journal of Chron- ic Disease 40:70­78. Thomas DC, Pogoda J, Langholz B, Mack W. 1994. Temporal modifiers of the radon-smoking interaction. Health Physics 66:257­262. Tritscher AM, Mahler J, Portier CJ, Lucier GW, Walker NJ. 2000. Induction of lung lesions in female rats following chronic exposure to 2,3,7,8-tetrachlorodibenzo-p-dioxin. Toxicologic Pa- thology 28:761­769. US EPA (United States Environmental Protection Agency). 1986. Airborne Asbestos Health Assess- ment Update. Research Triangle Park, NC: Environmental Criteria and Assessment Office. US EPA. 2004. Arsenic, inorganic (CASRN 7440-38-2). Available at http://www.epa.gov/iris/subst/ 0278.htm (accessed February 2). Wei M, Wanibuchi H, Yamamoto S, Li W, Fukushima S. 1999. Urinary bladder carcinogenicity of dimethylarsinic acid in male F344 rats. Carcinogenesis 20(9):1873­1876. Wei M, Wanibuchi H, Morimura K, Iwai S, Yoshida K, Endo G, Nakae D, Fukushima S. 2002. Carcinogenicity of dimethylarsinic acid in male F344 rats and genetic alterations in induced urinary bladder tumors. Carcinogenesis 23(8):1387­1397.

Next: 3 Epidemiologic Studies »
Veterans and Agent Orange: Length of Presumptive Period for Association Between Exposure and Respiratory Cancer Get This Book
×
Buy Paperback | $29.00 Buy Ebook | $23.99
MyNAP members save 10% online.
Login or Register to save!
Download Free PDF

From 1962 to 1971, US military forces sprayed herbicides over Vietnam to strip the thick jungle canopy that helped conceal opposition forces, to destroy crops that enemy forces might depend on, and to clear tall grasses and bushes from the perimeters of US base camps and outlying fire-support bases. Mixtures of 2,4-dichlorophenoxyacetic acid (2,4-D), 2,4,5-trichlorophenoxyacetic acid (2,4,5-T), picloram, and cacodylic acid made up the majority of the herbicides sprayed. Agent Orange was a 50:50 mixture of 2,4-D and 2,4,5-T. At the time of the spraying, 2,3,7,8-tetrachlorodibenzo-p-dioxin (TCDD, one form of dioxin) was an unintended contaminant from the production of 2,4,5-T and was present in Agent Orange and some other formulations sprayed in Vietnam.

In 1991, because of continuing uncertainty about the long-term health effects on Vietnam veterans of the herbicides sprayed, Congress passed Public Law 102-4, the Agent Orange Act of 1991. In response to the request from the VA, IOM extended the service of the Committee to Review the Health Effects in Vietnam Veterans of Exposure to Herbicides that was responsible for Update 2002 to address the question of presumptive period and respiratory cancer. The charge to the committee was to undertake a review and evaluation of the evidence regarding the period between cessation of exposure to herbicides used in Vietnam and their contaminants (2,4-D, 2,4,5-T and its contaminant TCDD, cacodylic acid, and picloram) and the occurrence of respiratory cancer.

  1. ×

    Welcome to OpenBook!

    You're looking at OpenBook, NAP.edu's online reading room since 1999. Based on feedback from you, our users, we've made some improvements that make it easier than ever to read thousands of publications on our website.

    Do you want to take a quick tour of the OpenBook's features?

    No Thanks Take a Tour »
  2. ×

    Show this book's table of contents, where you can jump to any chapter by name.

    « Back Next »
  3. ×

    ...or use these buttons to go back to the previous chapter or skip to the next one.

    « Back Next »
  4. ×

    Jump up to the previous page or down to the next one. Also, you can type in a page number and press Enter to go directly to that page in the book.

    « Back Next »
  5. ×

    To search the entire text of this book, type in your search term here and press Enter.

    « Back Next »
  6. ×

    Share a link to this book page on your preferred social network or via email.

    « Back Next »
  7. ×

    View our suggested citation for this chapter.

    « Back Next »
  8. ×

    Ready to take your reading offline? Click here to buy this book in print or download it as a free PDF, if available.

    « Back Next »
Stay Connected!